Engineering design of localised synergistic production systems

By

Melissa Yuling LEUNG PAH HANG

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Centre for Environment and SustainabilityFaculty of Engineering and Physical Sciences

University of SurreyApril 2017

Summary

Addressing a number of critical challenges caused by centralised production and large scale

distribution infrastructures, local production systems designed in a synergistic manner could

offer a possible pathway towards sustainability. The thesis focuses on the technical design of

local production systems integrating local heterogeneous processes to satisfy local demands

through efficient use of locally available renewable resources within technical and ecological

constraints.

A conceptual and quantitative multi-level framework, based on the Cumulative Exergy

Resource Accounting methodology, was first developed for a better understanding of a local

production system by considering the production and consumption of products or services as

well as ecological processes. A general design framework comprising an optional preliminary

design stage followed by a simultaneous design stage based on mathematical optimisation

was then developed for solving the design problem towards minimum overall resource

consumption. The preliminary design stage considers each supply subsystem individually and

allows insights into the potential interactions between them. The simultaneous design stage

has the capacity to include all design integration possibilities. A second, insight-based

approach was further developed, which offers a new hierarchical and iterative decision and

analysis procedure and incorporates design principles and ability to examine design

decisions.

The multilevel resource accounting framework was demonstrated on ethanol production from

cane and successfully revealed how decisions at one level would affect other levels of the

system. Both design approaches were illustrated on a case study for the design of local food-

energy-water nexus. It showed the advantages of an integrated design of a system which

makes use of local resources to meet its demands over a system relying on centralised

supplies and over a design without considering integration opportunities between subsystems.

The insight-based approach was also found to produce a comparable design to the

simultaneous design approach while offering more valuable insights for decision makers.

Declaration of Originality

This thesis and the work to which it refers are the results of my own efforts. Any ideas, data,

images or text resulting from the work of others (whether published or unpublished) are fully

identified as such within the work and attributed to their originator in the text, bibliography or

in footnotes. This thesis has not been submitted in whole or in part for any other academic

degree or professional qualification. I agree that the University has the right to submit my

work to the plagiarism detection service TurnitinUK for originality checks. Whether or not

drafts have been so-assessed, the University reserves the right to require an electronic version

of the final document (as submitted) for assessment as above. 

Melissa Yuling LEUNG PAH HANG

24th April 2017

Table of Contents

SummaryDeclaration of OriginalityTable of ContentsList of TablesList of FiguresAcknowledgementsChapter 1: Introduction..........................................................................................................................12

1.1 Localised production as an enabler of sustainable development.................................................121.2 Rationales for the design of local production systems................................................................151.3 Scope, aim and objectives............................................................................................................161.4 Overview of thesis.......................................................................................................................17

PART I: Towards a coherent multi-level framework for resource accounting.....................................21Chapter 2: A conceptual framework for resource accounting...............................................................21

2.1 The need for resource accounting................................................................................................212.2 Existing work on resource accounting........................................................................................22

2.2.1 Mass based resource accounting..........................................................................................232.2.2 Energy and Exergy based resource accounting...................................................................242.2.3 Emergy based resource accounting......................................................................................262.2.4 Multi-level framework for resource accounting..................................................................26

2.3 Aim and objectives for the conceptual framework......................................................................292.4 A conceptual framework for resource accounting.......................................................................30

2.4.1 Basic concepts of system, environment, process and flow..................................................302.5 Resource flows and their accounting principle...........................................................................32

2.5.1 Accounting for resource flows from Type-I processes........................................................322.5.2 Accounting for resource flows from Type-II processes......................................................33

2.6 Resource consuming processes...................................................................................................342.6.1 Flow making, capacity making, transportation and storage................................................352.6.2 Recycle, exchange and repair processes..............................................................................362.6.3 Environmental remediation processes.................................................................................36

2.7 Multilevel structure of a system..................................................................................................372.7.1 Onion model........................................................................................................................372.7.2 Significance of multi-level view..........................................................................................39

2.8 Summary of conceptual framework for resource accounting......................................................40Chapter 3: Algebraic quantification of resource accounting.................................................................41

3.1 Resource accounting algebra.......................................................................................................413.2 Resource accounting at unit level................................................................................................423.3 Resource accounting at process level..........................................................................................433.4 Resource accounting at inter-process level.................................................................................443.5 Resource accounting at production-consumption level...............................................................463.6 Summary of the resource accounting algebra..............................................................................50

Chapter 4: Case study on multi-level framework for resource accounting using algebras...................514.1 Overview of case study on ethanol production from sugarcane..................................................514.2 Ethanol production at the unit level.............................................................................................524.3 Ethanol production system at the process level...........................................................................534.4 Production of ethanol at the inter-process level...........................................................................544.5 Interaction between production and consumption of ethanol......................................................544.6 Comparative analysis...................................................................................................................554.7 Summary of Part I: a coherent multi-level framework for resource accounting.........................58

PART II: Design approach for integrated local production systems.....................................................54Chapter 5: Systematic approach for designing locally integrated production systems based on mathematical programming...................................................................................................................61

5.1 Rationales for shifting to localisation..........................................................................................615.1.1 Design problem statement and quantification of resource consumption..............................655.1.2 Overview of the proposed approach.....................................................................................665.1.3 Conceptual construction of superstructures..........................................................................685.1.4 Constructing the mathematical optimisation model for each subsystem..............................705.1.5 Preliminary design analysis..................................................................................................715.1.6 Constructing and solving a simultaneous design model.......................................................72

5.2 Building design models for food-energy-water nexus.................................................................735.2.1 Building superstructures.......................................................................................................735.2.2 Superstructure for food production subsystem.....................................................................745.2.3 Superstructure for water production subsystem....................................................................745.2.4 Superstructure for energy production subsystem..................................................................755.2.5 Superstructure for simultaneous food, energy and water design..........................................76

5.3 Mathematical formulation for the preliminary design analysis...................................................775.3.1 Mathematical formulation of food production system.........................................................815.3.2 Mathematical formulation of water production system........................................................855.3.3 Mathematical formulation of energy production network....................................................88

5.4 Mathematical formulation for the simultaneous design..............................................................925.4.1 Objective function.................................................................................................................935.4.2 Cross-subsystem flows.........................................................................................................93

5.5 Case study....................................................................................................................................965.5.1 Preliminary design analysis: Food production subsystem....................................................975.5.2 Preliminary design analysis: Water production subsystem................................................1005.5.3 Preliminary design analysis: Energy production subsystem...............................................1025.5.4 Simultaneous approach results............................................................................................104

5.6 Summary of preliminary and simultaneous design approaches to LIPS...................................111Chapter 6: An insight-based approach for the design of integrated local food-energy-water systems.............................................................................................................................................................112

6.1 Rationales for an insight-based design approach.......................................................................1126.2 Aim and Objectives...................................................................................................................1136.3 Methodology for insight-based approach..................................................................................114

6.3.1 Overview of methodological framework for insight-based design approach.....................1146.3.2 Design goal and resource gain............................................................................................1156.3.3 LIPSOM: Locally Integrated Production System Onion Model........................................1166.3.4 Principles for designing individual subsystems..................................................................1196.3.5 Cumulative exergy consumption of local products............................................................120

6.4 Sequential synthesis of multiple subsystems.............................................................................1226.4.1 Synthesis sequence.............................................................................................................1226.4.2 Inter-subsystem resource allocation....................................................................................1236.4.3 A sequential synthesis procedure........................................................................................123

6.5 The integration stage: resource cascading, recycling and regeneration.....................................1256.5.1 Quality of a resource...........................................................................................................126

6.6 Summary of the methodology for insight-based approach........................................................1306.7 Case study: designing the food-energy-water system for an eco-town.....................................131

6.7.1 Initial design of food subsystem.........................................................................................1326.7.2 Initial design of water subsystem........................................................................................1326.7.3 Initial design of energy subsystem......................................................................................1336.7.4 Iterative design....................................................................................................................1346.7.5 Integration: water reuse and regeneration...........................................................................1366.7.6 Integration: energy reuse....................................................................................................1366.7.7 Comparative analysis and final assessment........................................................................1376.7.8 Summary of insight-based approach...................................................................................139

Chapter 7: Robustness analysis and robust design of LIPS under uncertainties................................1407.1 Rationales for designing LIPS under uncertainties and type of uncertainties...........................1407.2 Approaches to handling uncertainties in design........................................................................140

7.3 Methodology for addressing uncertainties in design of LIPS...................................................1417.3.1 Post-design uncertainty assessment....................................................................................1437.3.2 Uncertainty-embedded design through two-stage stochastic programming.......................145

7.4 Case study on design of local food production system..............................................................1477.4.1 Mathematical model for deterministic design of local food production system.................1487.4.2 Post-design uncertainty assessment of food production system design.............................1507.4.3 Embedded design uncertainty.............................................................................................1537.4.4 Stochastic programming of food production system..........................................................1567.4.5 Concluding remarks for robustness analysis and design under uncertainties.....................1587.4.6 Summary of systematic approaches to the design of localised integrated production systems (LIPS)...........................................................................................................................................159

Chapter 8: Conclusions........................................................................................................................1608.1 Main research contributions and conclusions............................................................................1608.2 Wider implications of research..................................................................................................1648.3 Future research avenues.............................................................................................................165

Appendix A..........................................................................................................................................167A.1 Cumulative Exergy Consumption for cane agronomy..............................................................167

A.1.1 Cumulative Exergy Consumption for fertilisers................................................................167A.1.2 Cumulative Exergy Consumption for pesticides, insecticides and fungicides..................168A.1.3 Exergy consumption for ecosystem inputs........................................................................168A.1.4 Cumulative Exergy Consumption for land use..................................................................169A.1.5 Cumulative Exergy Consumption for human labour.........................................................169A.1.6 Cumulative Exergy Consumption for lubricants...............................................................170A.1.7 Cumulative exergy consumption for capital resources......................................................171A.1.8 Cumulative Exergy Consumption for environmental remediation (CO2 emissions).........172A.1.9 Total exergy consumption for cane agronomy..................................................................174

A.2 Cumulative exergy consumption for diesel for cane transportation.........................................174A.3 Cumulative exergy consumption for industrial cane processing..............................................174

A.3.1 Cumulative exergy consumption of electricity and imbibition water for cane milling.....174A.3.2 Cumulative exergy consumption of lime and steam for juice treatment...........................175A.3.3 Amount and exergy of bagasse..........................................................................................175A.3.4 Allocation factor between raw cane juice and bagasse......................................................175A.3.5 Cumulative exergy consumption of operating resources for fermentation........................176A.3.6 Cumulative exergy consumption of operating resources for distillation...........................176A.3.7 Cumulative exergy consumption for vinasse treatment.....................................................177A.3.8 Cumulative exergy consumption for operating resources for molecular sieve..................178A.3.9 Cumulative exergy consumption for operating resources for azeotropic dehydration......178

A.4 Cumulative Exergy Consumption for power station................................................................179A.4.1 Electricity and steam production from power house.........................................................180A.4.2 Cumulative exergy consumption for water for power house.............................................182A.4.3 Cumulative exergy consumption for electricity for power house......................................182A.4.4 Allocation factor for bagasse.............................................................................................182

A.5 Total exergy consumption at process level with intra-recycling flows....................................182A.6 Total exergy consumption at inter-process level with recycling and exchange flows.............183A.7 Summary of total exergy consumption of each unit in ethanol production and consumption. 185A.8 Comparative analysis for resource consumption for the 3 scenarios........................................186

Appendix B..........................................................................................................................................187B.1 Cumulative exergy resources for food production subsystem..................................................187

B.1.1 Cumulative exergy resources for bread production...........................................................188B.1.2 Cumulative exergy resources for beef production.............................................................190B.1.3 Cumulative exergy resources for pork production.............................................................190B.1.4 Cumulative exergy resources for potatoes production.......................................................191

B.2 Cumulative exergy resources for water production subsystem.................................................192B.2.1 Wastewater production from food production subsystem.................................................193B.2.2 Water demand and wastewater production from residential..............................................193

B.2.3 Rainwater collected in Eco-Town......................................................................................193B.2.4 Water demand and wastewater production from energy production subsystem................194B.2.5 COD of water sources........................................................................................................195B.2.6 Cumulative exergy resources for wastewater treatment....................................................195B.2.7 Cumulative exergy resources for groundwater..................................................................196

B.3 Energy production system.........................................................................................................196B.3.1 Electrical efficiency...........................................................................................................196B.3.2 Cumulative exergy of energy input....................................................................................197B.3.3 Cumulative exergy of energy production...........................................................................198B.3.4 Variability of energy sources.............................................................................................201B.3.5 Land requirement for energy production...........................................................................202

Appendix C..........................................................................................................................................206C.1 Evaluation of resource regeneration options.............................................................................206C.2 Water regeneration....................................................................................................................209

Appendix D..........................................................................................................................................212D.1 Initial design of food production subsystem.............................................................................212D.2 Initial design of water production subsystem...........................................................................217D.3 Initial design of energy production subsystem.........................................................................219D.4 Iterative design of local production system..............................................................................221

D.4.1 1st iterative design of food subsystem................................................................................221D.4.2 1st iterative design of water subsystem...............................................................................222D.4.3 1st iterative design of energy subsystem.............................................................................223D.4.4 2nd iterative design of local production system..................................................................224

D.5 Process integration of food-energy-water local production system..........................................227D.5.1 Integration options for water reuse and regeneration........................................................227D.5.2 Integration options for energy reuse..................................................................................231

References............................................................................................................................................239

List of Tables

Table 2-1: Decisions making at different levels of analysis..................................................................22Table 2-2: Decisions making at different levels of analysis..................................................................39Table 3-1: Description of indices used in Figures 3-1 to 3-4................................................................41Table 3-2: Classification of flows..........................................................................................................42Table 3-3: Description of the notations used in Equations (3.2) - (3.9)................................................48Table 4-1: Resource consumption for scenarios 1, 2 and 3...................................................................56Table 5-1: Key features of IS and EIPS, regional supply chain and LIPS............................................64Table 5-2: Specificities of Whitehill-Bordon eco-town........................................................................96Table 5-3: Preliminary design analysis for food production system.....................................................99Table 5-4: Contribution analysis of resource consumption for each locally produced food...............100Table 5-5: Preliminary design analysis for water production system..................................................102Table 5-6: Preliminary design analysis for energy production system................................................104Table 5-7: Detailed results from simultaneous design...........................................................................94Table 6-1: Examples of intended purposes and indicators of quality of some resources....................127Table 6-2: Outcome of 1st Iteration......................................................................................................134Table 7-1: Results of deterministic design of food production system...............................................150Table 7-2: Results of robustness analysis............................................................................................152Table 7-3: Fixed and flexible decision variables in the food production system................................154Table 7-4: Results of scenario based simulations with partial optimisation........................................155Table 7-5: First stage and second stage decision variables.................................................................156Table 7-6: First stage and second stage equations...............................................................................156Table 7-7: Results of stochastic design over deterministic design......................................................157Table A-1: Fertiliser input to cane agronomy......................................................................................167Table A-2: Pesticides, insecticides and fungicides input to cane agronomy.......................................168Table A-3: Exergy of flows from Type-II processes...........................................................................169Table A-4: Total exergy flows for surface water and lubricants for cane agronomy..........................170Table A-5: Cumulative exergy flows for capital resources.................................................................172Table A-6: Carbon dioxide released into the atmosphere due to cane agronomy...............................173Table A-7: Cumulative exergy consumption for carbon dioxide absorption......................................174Table A-8: Cumulative exergy consumption of electricity and imbibition water for cane milling.....174Table A-9: Cumulative exergy consumption of lime and steam for juice treatment...........................175Table A-10: Cumulative exergy consumption of operating resources for fermentation.....................176Table A-11: Cumulative exergy consumption of operating resources for distillation........................176Table A-12: Cumulative exergy consumption of operating resources for vinasse treatment..............177Table A-13: Cumulative exergy consumption of operating resources for dehydration......................178Table A-14: Cumulative exergy consumption for production of cyclohexane...................................178Table A-15: Total exergy consumption without recycling flows........................................................185Table A-16: Total exergy consumption with recycling flows.............................................................185Table A-17: Total exergy consumption with recycling and exchange flows......................................186Table A-18: Comparative resource consumption for the 3 scenarios..................................................187Table B-1: Food demand by local population in the eco-town............................................................187Table B-2: Cumulative exergy of imported food used in Chapter 5...................................................188Table B-3: Specificities for local bread manufacture including wheat cultivation.............................189Table B-4: Specificities for local beef manufacture............................................................................190Table B-5: Specific cumulative exergy of resources used in pork production....................................191Table B-6: Specificities for local pork manufacture........................................................................... 191Table B-7: Specificities for local potatoes production........................................................................192Table B-8: Wastewater produced from food processes.......................................................................193Table B-9: Water demand and wastewater generated from residential...............................................193Table B-10: Seasonal rainwater collected...........................................................................................194Table B-11: Water demand and wastewater generated from energy production subsystem...............195

Table B-12: Quality of water source....................................................................................................195Table B-13: Operating flows for wastewater treatment.......................................................................195Table B-14: Resources for groundwater..............................................................................................196Table B-15: Heat and electrical efficiency of CHP.............................................................................197Table B-16: Cumulative exergy of energy input.................................................................................197Table B-17: Total cumulative exergy consumption of energy technology.........................................198Table B-18: Carbon dioxide emissions from CHP..............................................................................199Table B-19: Variability of energy sources...........................................................................................202Table B-20: Land use of energy sources.............................................................................................203Table B-21: Inlet temperature of waste heat........................................................................................204Table B-22: Temperature required by heat sinks.................................................................................204Table B-23: Seasonal residential heat demand....................................................................................204Table D-1: Cumulative exergy of imported food used in Chapter 6...................................................212Table D-2: Specific cumulative exergy of conventional sources of water and energy.......................213Table D-3: Specific cumulative exergy of resources for bread production.........................................213Table D-4: Specificities for local bread manufacture..........................................................................214Table D-5: Specificities for local potatoes production........................................................................214Table D-6: Specificities for local beef manufacture............................................................................215Table D-7: Specificities for local pork manufacture...........................................................................215Table D-8: Specific resource gain of each food type...........................................................................216Table D-9: Initial design of food production subsystem.....................................................................216Table D-10: Amount of locally produced and imported bread per season..........................................217Table D-11: Parameters for treating groundwater...............................................................................218Table D-12: Quality of water sinks......................................................................................................218Table D-13: Initial design of the water production subsystem............................................................219Table D-14: Wastewater generated from initial design of water subsystem.......................................219Table D-15: Cumulative exergy consumption of associated energy technology................................220Table D-16: Initial base design of energy subsystem..........................................................................221Table D-17: Specific resource gain of food products for 1st iterative design of food subsystem........221Table D-18: 1st iterative design of food subsystem..............................................................................222Table D-19: 1st iterative design of water subsystem............................................................................223Table D-20: Wastewater generated from 1st iterative design of water subsystem..............................223Table D-21: 1st iterative design of energy subsystem..........................................................................224Table D-22: 2nd iterative design of water subsystem...........................................................................225Table D-23: Wastewater generated from 2nd iterative design of water subsystem..............................225Table D-24: 2nd iterative design of energy subsystem.........................................................................226Table D-25: Base design of local production system..........................................................................227Table D-26: Availability of water sources and their quality...............................................................228Table D-27: Water sinks and their quality...........................................................................................229Table D-28: Resource gain of water sources after regeneration..........................................................230Table D-29: Stream data for winter.....................................................................................................232Table D-30: Stream data for summer..................................................................................................233Table D-31: Stream data for autumn...................................................................................................234Table D-32: Stream data for spring.....................................................................................................235Table D-33: Heat recovery for each season.........................................................................................237

List of Figures

Figure 1-1: Local production system.....................................................................................................14Figure 2-1: System and environment framework..................................................................................32Figure 2-2: Overall depiction of the Environment and Societal System...............................................35Figure 2-3: Detailed depiction of the product provision subsystem......................................................36Figure 2-4: Onion model for structural representation of Environment-Society System......................38Figure 3-1: A representation of the unit level........................................................................................43Figure 3-2: A representation of the process level..................................................................................43Figure 3-3: A representation of the inter-process level.........................................................................45Figure 3-4: A representation of the product-consumption level............................................................47Figure 4-1: The case study on sugarcane ethanol production................................................................52Figure 4-2: Ethanol production at the unit level....................................................................................52Figure 4-3: Ethanol production system at the process level..................................................................53Figure 4-4: Production of ethanol at the inter-process level..................................................................54Figure 4-5: The interaction between production and consumption of ethanol......................................54Figure 4-6: Overall resource consumption for the three scenarios excluding Type-II flows................57Figure 5-1: Methodological framework for designing a localised synergistic production system........67Figure 5-2(a): Illustrative superstructure of a single (sub-) system.......................................................69Figure 5-2(b): Generic superstructure representation of combined systems.........................................70Figure 5-3: Superstructure for food production subsystem...................................................................74Figure 5-4: Superstructure for water production subsystem..................................................................75Figure 5-5: Superstructure for electricity production............................................................................75Figure 5-6: Superstructure for heat production……………………………………………….............76Figure 5-7: Superstructure for integrated food, energy and water system.............................................77Figure 5-8: Proportion of resource consumption for each locally produced food...............................100Figure 5-9: Resource consumption by each water source...................................................................101Figure 5-10: Resource consumption by energy source........................................................................103Figure 5-11: Results of simultaneous design.......................................................................................105Figure 5-12: Net resource consumption for each scenario..................................................................108Figure 5-13: Cumulative exergy of food subsystem for all 3 subsystems...........................................108Figure 5-14: Cumulative consumption of resources (a) chemicals, (b) heat, (c) electricity, (d) capital resources by all 3 scenarios for each water source..............................................................................109Figure 5-15: Cumulative consumption by each technology for energy production in all scenarios.. .110Figure 6-1: Locally Integrated Production System Onion Model (LIPSOM).....................................117Figure 6-2: A sequential synthesis procedure......................................................................................124Figure 6-3: Methodological framework for insight-based design approach........................................131Figure 6-4: Base design of local production system............................................................................135Figure 6-5: External CExC for each scenario using insight-based approach......................................138Figure 6-6: External CExC for all subsystems of insight-based and simultaneous approaches..........139Figure 7-1: Methodological framework for addressing uncertainties in design..................................142Figure 7-2: Variation in objective function with uncertainties............................................................151Figure 7-3: Robustness analysis..........................................................................................................152Figure 7-4: Monte-Carlo Simulation results........................................................................................153Figure A-1: Condensing Extraction Steam Turbine............................................................................179Figure C-1: Generic water pinch diagram...........................................................................................209Figure D-1: Grand composite curve for winter...................................................................................232Figure D-2: Grand composite curve for summer.................................................................................233Figure D-3: Grand composite curve for autumn..................................................................................234Figure D-4: Grand composite curve for spring....................................................................................235

Acknowledgements

Foremost, I would like to thank my supervisors Matthew Leach and Aidong Yang for their

constant support throughout my PhD study. I am forever grateful to Aidong for his constant

availability for precious guidance especially on the technical part of the research, critical

feedback and valuable comments at each stage of the project. I am also highly indebted

towards Matthew for his mentorship, constructive criticisms and prompt help with any

administrative issues. I could not have wished for better supervisors and I aspire to become

like them one day. I want to also thank Elias Martinez Hernandez for speeding up my

learning curve, encouraging me to think critically, generous time and timely advice

throughout my PhD.

I would like to express my gratitude to the Leverhulme Trust and Overseas Research

Scholarship from University of Surrey for financial support and making this research

possible.

Words cannot express how grateful I am to Moira, our dear CES administrator, for all her

help and support. Thank you for making me and everyone else feel so welcome and making

CES such a friendly and enjoyable place. My earnest thanks also go to all CES students as

you have all directly or indirectly helped me in the realisation of this research project. Thanks

to all my officemates and friends: Ida, Nini, Punch, Richard, Tyler, Anna, Claire, Mercio,

Marcio, Kok Siew, Nittida and Xin. Your friendships mean a lot to me and I will miss all the

good times we spent together.

I wish to also thank my best friend Deepti whose moral support was paramount in helping me

to achieve my dream of getting a PhD. My heartfelt and biggest thanks go to my parents for

their unconditional love, affection and support. Thanks to all my brothers who are a constant

source of motivation for me to strive higher in all my endeavours. This thesis is dedicated

fully to my family. Last but none the least, I am grateful to God for bestowing upon me

health, wisdom, strength and perseverance throughout my studies.

Chapter 1: Introduction

1.1 Localised production as an enabler of sustainable development

The rapid increase in industrialisation and a growing world population that is expected to

reach 9.6 billion by 2050 (UN, 2013) have led to mounting pressure on global demands for

material and energy resources. Production activities (e.g. industrial, construction and

agricultural) have increased considerably in the recent decades in order to meet the demands

of rising economies such as China and India and that of an ever-growing standards of living.

Such activities require huge and constant supply of energy and materials resources. For

thousands of years preceding the industrial revolution, resources were extracted traditionally

from locally available renewable resources such as biomass, hydro and wind power and were

then processed locally through distributed small scale production. With the advent of

industrialisation and the subsequent widening exploitation of energy-dense fossil fuels,

production has been diverted rapidly to centralised systems based primarily on fossil

resources, accompanied by large-scale distribution infrastructures. While the large scale

economies of these centralised systems have been beneficial to the society in certain respects,

continued reliance on geographically concentrated fossil resources for energy and most

materials, coupled with population growth and rising economies, has caused a range of severe

issues and challenges facing the world today, such as energy supply security, detrimental

environmental consequences with global dimensions, and social-economic injustice. Though

these challenges are global, they result from an aggregation of local problems that may affect

each locality differently. The scale of these problems also means that there is an

unprecedented urgent need to shift to alternative sustainable systems of production,

distribution and consumption of energy and material products and services.

Another set of resource shifts is thus expected alongside changes in the locations and scale of

production, distribution and consumption activities and their infrastructures leading to the

introduction of a large number of small-scale localised production activities operating on

local renewable resources and which could be owned and operated by local people to meet

local demands (e.g. food, energy, water and materials demand) (Martinez-Hernandez et al,

2016). The benefits of such a resource shift and re-localisation have been widely acclaimed

by various groups such as economics schools of “eco-localism” (Curtis, 2003) or “distributed

economy” (Johansson et al., 2005), grass-root social movements such as the Transition

Towns Network (Middlemiss and Parrish, 2010) and Royal Academy of Engineering which

particularly identifies the positive benefits for localised energy and water supply with respect

to resilience (The Royal Academy of Engineering, 2011). Moreover, the UK Government has

recently initiated several projects to promote its vision of localism and decentralisation of

governing power such as the implementation of an anaerobic digestion strategy for the

production of energy from locally available wastes (DECC, 2015a).

In a similar context, Martinez-Hernandez et al. (2016) define the scope of a local production

system as one which focuses on the co-location of resource extraction from the local

environment, processing and consumption by local population. A local production system can

be viewed as a network of heterogeneous processes, including both technological and

ecological, integrated in a synergistic manner to achieve a high degree of resource efficiency,

potentially leading to improved economic viability while preserving the ecosystem

(Martinez-Hernandez et al., 2016). Figure 1-1 illustrates the key concepts of a local

production system. A local system comprises the local environment including ecological

processes and man-made components for production and consumption activities all co-

existing within a specified local geographical boundary. The latter can be considered as that

of an area under the direct governance of a local or regional planning body. Therefore, its

geographical boundary could encompass existing lower levels of public governance such as a

town, a city, or a county where decision making may be pursued based on sufficient local

details. Figure 1-1 depicts that the basic needs of a local population such as nutrition,

sanitation and thermal comfort are the drivers of a local population and these needs are met

through the consumption of resources (e.g. food, energy and water) produced from the

production processes. Ecological processes from the local environment set the constraints for

the production processes such as agricultural, industrial and municipal processes which in

turn provide the intermediate flows that are consumed by the local population to satisfy their

needs. The production and consumption processes are closely linked to the local

environment. Locally available renewable resources are extracted from the ecological

processes in the local environment, processed by production processes to generate a final

product or service that is consumed by the local population; releasing harmful effluents and

solid wastes into the local environment and potentially affecting its capacity to sustainably

supply resources and regulate the ecological processes.

Ecological processesWater cycleWaste Consumption/

Local needsNutrition (food)SanitationThermal comfortMobilityHousingWater

Local environmentLandWater bodiesClimatic conditions (wind/irradiance/temperature/precipitatio

Figure 1-1: Local production system

The establishment of local production systems has been identified as a possible pathway

towards sustainable development (Martinez-Hernandez et al., 2016). These systems offer the

possibility to facilitate more effective use of renewable resources which can be captured or

produced locally to meet demands of a local population. The use of renewable resources is

known to be highly beneficially to the environment and the ecosystem. Life-cycle global

warming emissions from extraction to processing to decommissioning for most renewable

resources are insignificant (IPCC, 2011). In contrast, 44% of the world’s carbon dioxide

emissions are generated from coal, 34% from oil and the rest from oil with a negligible

contribution from other sources (IEA, 2015). The Millennium Ecosystem Assessment also

advocates the benefits of local management strategies in order to preserve the sustainability

of ecosystem services which have been jeopardised by harmful effluents from large

centralised production systems (MEA, 2005). Replacing a fossil-fuel based economy with a

renewable one reduces water and air pollution and thus contributes to improving on

ecosystem and human health (Rizk, 2013). Furthermore, distributed and modular renewable

energy systems are more reliable, resilient and less prone to large scale failure as compared to

the traditional centralised distribution systems (NREL, 2014). According to IRENA (2016)

renewable resources can also potentially improve revenues and create more employment

given the labour-intensive nature of the distributed renewable energy sector.

In addition to promoting the use of renewables, local production systems have the potential to

enable symbiotic integration of multiple distinct production processes (e.g. for provision of

food, energy and water) within the same locality in order to increase resource efficiency and

sustainability. Any wastes or by-products that are generated from the production processes as

well as used products from consumption will seek to be looped back within the local system

Ecological processesWater cycleWaste Consumption/

through symbiotic arrangements based on the principles of industrial ecology (Chertow and

Ehrenfeld, 2012). Local production system offers thus a pathway towards shifting towards a

more cyclical and closed loop one where wastes disposal is minimal as they are viewed as

valuable resources to be used again in the production system; therefore greatly reducing the

need for input of virgin resources.

Besides, centralised production faces the challenge of delivering the resources to any demand

locations which could lead to logistic as well as political issues (Klemes and Varbanov,

2013). Local production systems have the advantage of avoiding large transportation

distances and the associated distribution losses and risks.

1.2 Rationales for the design of local production systems

Given the multi-disciplinary problems (e.g. environmental, social and economic) that local

production systems aim to target, they will be internationally interesting and relevant to

study, although they will require different implementations depending on the social,

economic and local environment settings pertaining to the local region. As compared to

conventional production systems that generally manufacture only one type of product in bulk

(e.g. plants producing bulk chemicals and oil refineries) and which often belong to a rather

linear supply chain and to which one or very few technical designs are universally adopted

regardless of their locations, a local production system will comprise a non-linear value chain

that will require its design and that of its components to be highly tuned and adapted to the

local settings. As pointed out by Wilbanks and Kates (1999), while many frameworks have

been developed to generate insights from the interactions occurring at the global scale, these

do not necessarily represent the conditions at the local scale. It is thus pivotal to develop a

systematic design framework for local production systems and to offer a holistic engineering

approach to the sustainable provision of multiple goods/services under a range of conditions

such as resource availability, population needs and ecological and technical constraints. Such

an approach should suggest how to formulate a conceptual design problem for such systems

under different conditions, what tools to use to solve these problems using key performance

indicators and quantitative models and what design rules and principles can be used to

provide guidance on a variety of potential design alternatives and their associated trade-offs

so as to support decision making on the suitable design for implementation.

1.3 Scope, aim and objectives

The broad aim of this PhD project is to develop systematic tools for the sustainable design of

local production systems. Its specific objectives are as follows:

Proposing a conceptual and quantitative multi-level framework for a better

understanding of a local production system by considering not only the production

and consumption of products or services but also the ecological and technological

processes.

Formulating the problem of synthesising local production systems under different

circumstances and local settings using appropriate case studies.

Developing systematic approaches for solving the design problem towards optimal

technical performance.

Developing a set of preliminary guidance, design rules and principles to practices

related to the design of local production systems.

The PhD research work will focus primarily on the technical design of a local production

system, while acknowledging the need to integrate technical, economic and social

perspectives in guiding the development in practice. The engineering oriented research

carried out in this work is intended to develop solid “physics” to support future research that

emphasizes on the social, political and cultural aspects of this area. The problem statement

for the (technical) design of such system can be defined as one of selecting and arranging

industrial and agricultural processes based on the type and volume of input flows (i.e. feed)

and output flows (e.g. products and services), technological options, geographical location

and the associated infrastructure components as part of the supporting capacity. In addition,

the design of a local production system will aim at optimising performance indicators within

the constraints imposed by the physics of processes involved (e.g. technological efficiency

limit) and capacity limitations (e.g. groundwater abstraction limit). Two primary types of

design approaches will be examined in this PhD project to handle the heterogeneity and

complexity of local production systems:

Mathematical programming approaches using superstructure modelling to represent

possible solutions and then numerical optimisation algorithms to solve the

superstructure model and identify the optimal solution(s). The optimisation tools and

methods are already fairly mature and have successfully been applied in process

systems engineering (Klatt and Marquardt, 2009) to solve a wide range of problems

such as the design of a bioenergy network (Beck et al., 2008) and that of energy

supply chains (Almansoori and Shah, 2012). This work will make use of existing

optimisation approaches but focus the new research on formulating optimisation

problems that appropriately represent the nature of the task of designing integrated

local production systems.

Physics-based approaches which are based on the second law of thermodynamics and

comprising techniques such as pinch analysis and resource cascading (Varbanov and

Klemes, 2011; Geldermann et al., 2006). These approaches have already been applied

to the design of conventional chemical processing systems. In the context of using

physics-based approaches for the design of local production system, special attention

will be given to (i) formulating a unified performance indicator to measure the true

‘cost’ of a production process that encompasses the operating, capital and resources

consumed for environmental remediation of harmful effluents (ii) accommodating

processes with very diverse natures (e.g. manufacturing, agricultural and

municipal/utility production processes) (iii) handling the intermittency and seasonality

of the supply of renewable resources.

1.4 Overview of thesis

This research encompasses two inter-connected parts. Part I (Chapters 2-4) develops a

thorough conceptual and multi-level framework for resource accounting that can be applied

to local production systems for analysing their performance. Using the resource accounting

approach recommended in Part I, Part II develops approaches for the integrated design of

local production systems (Chapters 5 and 6) and includes a section on handling uncertainties

in the design of these systems (Chapter 7). Given this structure, relevant literature is

critically reviewed within each of the main chapters.

The main contributions resulting from this PhD research work are described as follows:

Developing a coherent framework for resource accounting. Conceptually, the

framework represents the key aspects of a system such as system boundary, types of

flows and processes. The principles for a concise, holistic resource accounting that at

the same time avoid ambiguity and double-counting of resources, and multiple levels

of analysis of a local system are presented in Chapter 2 and form part of a paper

entitled “Towards a coherent multi-level framework for resource accounting”

published in Journal of Cleaner Production. As compared to existing multi-level

framework studies (Hanes and Bakshi, 2015a, 2015b), this conceptual framework

offers a more thorough multilevel analysis of the processes and flows pertaining to

resource consumption at the various levels of a particular system. Such analysis is

required to reveal how a resource, before and after being processed at different stages,

flows within the system, which is essential for the identification of potential

synergistic integration with flows linked to other products or services in the system.

From the conceptual framework, an algebraic quantitative approach to resource

accounting based on the concept of cumulative exergy consumption as key

performance indicator has been developed at each level of the framework in Chapter

3 and also forms part of the publication mentioned above. Previous studies have not

focused on a holistic quantitative study encompassing at the same time ecosystems

(i.e. natural processes), production, and consumption of desired product or service

(i.e. human systems), as pointed out also in Martinez-Hernandez et al. (2016). The

developed framework fills this gap and provides support for decision making at

specific technical levels of interest with respect to resource consumption.

The conceptual and quantitative framework for resource accounting was applied and

demonstrated on a case study on ethanol production from sugarcane in Chapter 4 and

can also be found in the published article in the Journal of Cleaner Production. The

application of the developed framework on this case study illustrates how to use the

framework to assess the full impacts on resource consumption for design decisions at

all levels, allowing design options to be explored to find the most efficient option. The

proposed framework has provided powerful insights into how reduction/increase of

resource consumption can occur at different levels. It also offers the potential to

identify key components and flows that can be either removed or improved through

integration and linkage with other flows or components in the system.

A systematic approach to the design of local production system based primarily on

mathematical programming was proposed in Chapter 5 and forms part of a

publication in the Journal of Cleaner Production, entitled “Designing integrated local

production systems: a study on the food-energy-water nexus”. The chapter presents a

preliminary design analysis tool that is useful when dealing with existing

infrastructure and the design is more for retrofitting purposes or when systems are

implemented separately in stages with a view to develop systems integration in the

future. Chapter 5 also describes the optimal integrated design of local production

system through a simultaneous mathematical modelling approach based on

superstructure modelling and optimisation. In such an approach, the superstructures

for all the production processes are combined into a single superstructure and solved

in one mathematical optimisation considering all integration opportunities. In

comparison to the preliminary design approach, this approach considers all design

integration options simultaneously across all subsystems. This approach is essential

for revealing the benefits of an integrated local production system on resource

efficiency and circularity as compared to the practice of designing distinct subsystems

in silos.

Chapter 6 is about developing a systematic insight-based approach for the design of

local production systems. Such an approach is required as simultaneously designing

multiple distinct processes (e.g. agricultural, industrial and municipal) can prove to be

too complex to address in one big step. It also offers a piecewise and incremental

approach with the appropriate balance between capturing complexities while keeping

the algorithm simple yet robust. It is a practical tool that realistically allows feedback

from users at any design stage to generate insights exploring the design options that

are more aligned to their core interests; thus enabling them to make better informed

decisions. Chapter 6 was submitted for publication in Environmental Science &

Technology.

Chapters 5 and 6 of the thesis focus on the tools and methodologies developed for the design

of a local production system. The application of these tools is illustrated on a case study on

the integrated design of the local food-energy-water nexus based on an eco-town in the UK.

Food-energy-water nexus is an emerging area of research since its importance has been

highly recognised for sustainable development and national security by various global

organisations such as UN and FAO. Chapter 7 illustrates how the existing approaches for

handling uncertainties in design can be applied to the design of local production systems.

This thesis ends with Chapter 8 that synthesises the main contributions of this research work

and their implications in the wider context of sustainable development. Chapter 8 also

discusses future research avenues.

PART I: Towards a coherent multi-level framework for resource accounting

Chapter 2: A conceptual framework for resource accounting2.1 The need for resource accounting

Resource scarcity and environmental impacts of production and processing of resources are

two main rationales behind resource accounting. Natural resources are the ultimate source of

all the goods and services to meet human needs (e.g. food, energy, water). With world

population at 7.4 billion people (PRB, 2016) and overall standards of living rising, there is

inevitably a subsequent increase in the consumption of natural resources globally. There are

mounting concerns that the supply of key resources such as energy, water and materials

would not be sufficient anymore to meet the needs of a rising world population. Since 1970,

the world population has almost doubled while global economy and global material

extraction have almost tripled over nearly four decades according to UNEP (2016a). From

2000 to 2010, with the exception of biomass extraction which remained constant at 2%, the

rate of extraction of all other materials increased. Fossil fuels consumption increased on

average by 2.9%, metal ores by 3.5% and non-metallic minerals by 5.3% (UNEP, 2016).

Allwood et al. (2011) also predicted that demands for engineering materials used for the

construction of buildings, infrastructure, equipment and products are expected to double in

the next 40 years. Moreover, resource scarcity has also led some commodity prices to rise

significantly while depleting fossil fuels have contributed to soaring oil prices (Krautkraemer,

2005). Furthermore, inefficient use and over-exploitation of resources have adverse impacts

on the health of human beings as well as on the environment and contributes to climate

change and global warming. Allwood et al. (2011) reported that the negative environmental

impacts of producing and processing materials have driven the promotion of material and

resource efficiency in policies while Huijbregts et al. (2010) have demonstrated that a number

of emission-related impacts are strongly linked to resource use.

Improving resource efficiency by producing, processing and consuming Earth's limited

resources in a sustainable manner while minimising impacts on the environment from the

overall life cycle of the resource (EC, 2013; UNEP, 2012), can bring significant economic

benefits and boost competitiveness (EC, 2013). There is an urgent need to develop new

design methods for reducing resource use, minimise waste, improve management of resource

stocks, change consumption patterns, optimise production processes, management and

business methods, and improve logistics. Efficient use of resources can help in identifying

superior technological options, increasing employment in the rapidly evolving green

21

technology field, creating new export markets as well as benefiting consumers through more

environmental friendly and sustainable products (EC, 2013). Appropriate tools and

techniques are required for the realisation of these benefits. Resource accounting becomes

thus an important approach that can be used to assist decision making and system design,

gain insights on the performance of a production system and devise options for improving

resource efficiency in order to optimise utilisation of available resources while minimising

impacts on the environment. By monitoring and assessing resource consumption, the effect of

retrofitting or introducing a new component into a system can be analysed and can serve as a

guide for the selection of those components that improve the performance of a system.

2.2 Existing work on resource accounting

Different approaches using mass, energy, exergy and emergy, summarised in Table 2.1, for

resource accounting exist (Ukidwe and Bakshi, 2005).

Table 2-1: Resource accounting approaches

Resource accounting approaches Advantages Disadvantages

Mass based

-They can be the basis of a good database for developing other more

comprehensive methods (Ukidwe and Bakshi, 2005)

-Mass based methods are based on material weight only and do not

give information on the quality of materials or impact on ecosystems

that are interacting (EC, 2013)

Energy based-Established quantity for physical

quantification of resources (Sfez et al., 2017)

- Different resources cannot be compared and aggregated based

only on their energy content as their quality might be different and so

they cannot be substituted for each other (Bakshi, 2013)

Exergy based: Cumulative Exergy

Consumption (CEC) by Szargut et al. (1988)

-Well established exergy based method for resource accounting adopting a

Life Cycle Assessment (LCA) approach to account for material and energetic inputs from extraction to

industrial manufacture of the product/service

- Does not account for non-energetic resources such as money,

labour and environmental remediation resources

-Does not account for ecological resource consumption

Exergy based: Industrial Cumulative Exergy

Consumption (ICEC) (Ukidwe and Bakshi, 2007; Zhang et al.,

2010)

-Well established exergy based method similar to CEC but focuses on

industrial systems

-Does not explicitly take into consideration resource consumption in other parts of the value chain of a

product/service- Does not account for ecological

resource consumptionExergy based: Extended

Exergy Accounting (EEA)

-Is an extension of Szargut’s CEC and additionally accounts for non-energetic

resources such as money, labour and environmental remediation costs for

zero environmental impact by technological processes (Sciubba,

-EEA is still a relatively young methodology

- Potential double counting and cost allocation for different products-Does not account for ecological

resource consumption

22

2001)

Ecological Cumulative Exergy Consumption

(ECEC)

-Based on ICEC but extends its boundary to account for the total exergy consumed in ecological

processes for the production of natural resources as well as for assimilating pollutants (Hau and Bakshi, 2004)

-Accounts for the use of ecological resources based on emergy which is

a controversial quantity.

Exergy based: Cumulative Exergy Extraction from the

Natural Environment (CEENE) developed by

Dewulf et al. (2007)

-Account for ecological resource consumption and offers a more

comprehensive accounting of all natural resources including land use

-Avoid any double counting by setting correct system boundaries

-Offers comprehensive resource accounting for ecological resources

especially land use but does not offer a holistic quantitative study encompassing at the same time

ecosystems (i.e. natural processes), production, and consumption of desired product or service (i.e.

human systems).

Emergy based

-An attempt to analyse ecological and economic systems and to account for

ecological goods and services in a common unit of solar energy required

to produce them

-Not easily understood and controversial quantity with quantitative and algebraic

challenges

The key features and characteristics of each of these resource accounting approaches are

further detailed in sections 2.2.1 to 2.2.3.

2.2.1 Mass based resource accounting

Mass based methods have been widely used for reporting resource consumption especially at

the level of the entire economy (Adriaanse et al., 1997; Matthews et al., 2000). More

precisely, material resource has been popularly measured using the Material Flow Analysis

(MFA) metric which is used for setting targets for material use at the macroscopic level.

MFA describes the flow of materials in the economy in physical terms with total inputs and

total outputs measured by weight using the mass balance principle based on a period of one

year (EC, 2012) and provides an account of the aggregated physical amount of extracted raw

materials as well as that of imports and exports (EC, 2013). The main categories of materials

that have been considered in MFA studies are biomass, non-metallic minerals, metals as well

as fossil-fuels (EC, 2012). Targets set by MFA for material use typically include Domestic

Extraction (DE), Domestic Material Input (DMI) and Domestic Material Consumption

(DMC) which are direct flows to a system. MFA also considers indirect flows to a system

such as Raw Material Equivalent (RME), Raw Material Consumption (RMC) and Total

Material Requirement (TMR) and Total Material Consumption (TMC). However, data for

calculating these indicators are not usually readily available. The Material Input per Service

(MIPS) is another material resource consumption metric that was developed to account for all

23

the material resources on a life cycle basis to produce a product or service (Ritthoff et al.,

2002).

2.2.2 Energy and Exergy based resource accounting

Mass accounting methods are not able to account for all energy carriers, typically wind

energy and electricity (Sfez et al., 2017). However, similar to using mass, there are some

caveats to using energy for resource accounting. Energy can neither be created nor destroyed

but its ability to do work decreases in real processes. Exergy, defined as the maximum

available energy to do useful work, is a thermodynamic measure of energy quality and a more

insightful indicator of resource consumption as compared to energy (Amini et al., 2006).

Extensive work has been done on resource accounting based on exergy by researchers such as

Wall (1977, 1999, 2002, 2011), Zaleta-Aguilar et al. (1998), Gong and Wall (2000), Valero et al. (2002), Chen (2005, 2006), Chen and Ji (2007), Huang et al. (2007), Valero (2008) and Jiang et al. (2009) as reviewed by Gaudreau (2009). Exergy based methods are preferred for resource accounting because

they embody both the first and second laws of thermodynamics and can capture a wide range

of material and energy streams. Also, exergy is an established thermodynamically rigorous

quantity with a solid quantitative formulation (Dewulf et al., 2007; Sciubba and Wall, 2007).

Moreover, exergy is a more universal quantity that can also capture the contribution of non-

energetic resources (e.g. labour), environmental impacts of pollutants including those on the

behaviour of ecosystems (Jorgensen, 1997) as opposed to mass and energy based methods

(Ukidwe and Bakshi, 2004). In comparison, resource consumption cannot be fully quantified

using matter or energy because both are always conserved (Wall, 1977; Cornelissen, 1997; Gong and Wall, 2000; Rosen et al. 2008; Valero, 2008). Connelly and Koshland (2001) and Cornelissen and Hirs (2002) argued that resource consumption cannot be defined from a first law of thermodynamics principle.

A literature search showed that popular exergy based methods that adopt a life cycle approach for resource accounting include Cumulative Exergy Consumption (CExC), Industrial Cumulative Exergy Consumption (ICEC), Extended Exergy Accounting (EEA), Ecological Cumulative Exergy Consumption (ECEC) and Cumulative Exergy Extraction from the Natural Environment

24

(CEENE). More recent applications of exergy based methods include LCA to assess

alternative soil remediation technologies (Rocco et al., 2015), and attempt to include

economic and environmental factors in ECEC of industrial processes (Yang et al., 2015) and

the extension of the classical Economic Order Quantity (EOQ) model to take into account

sustainability factors such as labour, capital and environment based on EEA approach (Jawad

et al., 2015). Szargut et al. (1988) first introduced the concept of CExC where all exergy

resource inputs along the production chain from extraction to industrial manufacture of the

product or service are added together. ICEC is based on the CExC developed by Szargut et al.

(1988) but focuses only on the total cumulative exergy consumption in industrial systems

(Ukidwe and Bakshi, 2007; Zhang et al., 2010). Moreover, Sciubba (2005) introduced the

concept of EEA. This is an extension of Szargut’s cumulative exergy consumption and

accounts not only for the material and energetic flows but also for non-energetic resources

such as money, labour and the environmental remediation costs for zero environmental

impact by technological processes. However, EEA is still a relatively young methodology

that will require the use of other supporting tools as well as further validation to be widely

accepted in engineering analysis (Rocco et al., 2013). Moreover, there are some relevant

issues about EEA such as potential double counting and cost allocation for different products

that still need to be addressed. Indeed, Rocco et al. (2013) argued that due to the nature of

EEA and its holistic approach, it might be subjected to double counting issues especially if it

is not well supported by disaggregated database. Furthermore, they concluded that the

definitions of the exergy equivalent of labour and capitals still require more investigation and

that EEA’s novel technique for estimating the exergy equivalence of monetary unit is

controversial as there is currently no general agreement on the rationale for estimating the

exergy equivalence of capitals. Rocco et al. (2013) also reported that the general principle

used in EEA for the allocation of inputs in a multiproduct system is still unclear and needs

systematization.

CExC, ICEC and EEA do not encompass the resource consumption by ecological

processes. To address this shortcoming, the concept of ECEC has been developed by Hau and

Bakshi (2004a). ECEC is similar to ICEC but extends its boundary to account for the total

exergy consumed in ecological and natural processes for the production of natural resources

such as fossil fuels, ore and renewable energy as well as for assimilating pollutants (Hau and

Bakshi, 2004a). Taking the work done by nature for granted can cause severe degradation of

ecological goods and services that are so paramount for human survival and the sustainability

25

of its activities. Similarly, CEENE developed by Dewulf et al. (2007) is a resource

accounting methodology that specifically accounts for all the natural resources derived from

the ecosystem. The novelty of CEENE is that it offers a more comprehensive accounting of

all natural resources including land use; the latter has been overlooked in the other resource

accounting methodologies. It offers a comprehensive database for the cumulative exergy

taken from the ecosystems for the provision of atmospheric resources, land resources, water

resources, minerals, metal ores, nuclear energy, fossil fuels and renewable resources (Dewulf

et al., 2007). Moreover, the work done on CEENE by Dewulf et al. (2007) intended to avoid

any double counting by setting correct system boundaries. One of the main differences

between the different exergy based accounting techniques seems to be the system boundary.

The determination of the appropriate system boundary in resource accounting appears to be

an important matter of concern. However, it can be noticed that not much consideration has

been given to the consumption of final product or service in the existing work as opposed to

the inclusion of the ecosystem and the production stage of the product or service. This

appears to be a significant shortcoming as responsible consumption is also essential in

contributing to long term sustainability and should not be overlooked.

2.2.3 Emergy based resource accounting

Emergy based methods (Odum, 1996) have also been used to analyse ecological and

economic systems and to account for ecological goods and services. Emergy is a quantitative

analysis technique that has been developed to estimate the values of resources, services and

commodities in a common unit of solar energy required to produce them. However, emergy

has encountered much resistance and criticism from the wide community of economists,

physicists and engineers (Hau, 2005). Emergy analysis faces a lot of quantitative and

algebraic challenges while its broad claims about ecological and economic systems are highly

controversial. Moreover, emergy analysis of the economy disaggregated to the level of

industrial sectors is presently lacking and uses only a single emergy to money ratio to

estimate emergy for the whole economy (Ukidwe and Bakshi, 2005). Furthermore, estimating

the emergy of the economy through the emergy to money ratio might not be very accurate and might also lead to double counting (Ayres, 1998; Cleveland et al., 2000). Also, this approach seems to contradict Odum’s claim that that money is not a complete measure of wealth (Hau and Bakshi, 2004b).

26

2.2.4 Multi-level framework for resource accounting

Previous studies especially those by Hau and Bakshi (2004), Yi et al. (2004) and Liao et al.

(2012) have recognised the need for a multilevel analysis for resource accounting, based on

exergy and with emphasis on the system boundary, as opposed to a narrow analysis focused

on individual processes which might shift the resource consumption impacts to other parts of

the value chain of the product or service. Yi et al. (2004) analysed resource accounting at four

different levels namely process, life cycle scale, economy scale and ecosystem scale. The

process scale is the lowest scale and analyses resource accounting around a process or

equipment. The system boundary is further extended in the life cycle scale to include key

processes in the life cycle of a product or service. However, the life-cycle scale still ignores

many of processes in the life cycle and could consequently lead to significant inaccuracy in

the results. The economy scale includes the activities that are relevant in the whole economy

to satisfy the requirements of the selected processes in the life cycle scale and combines

economic input-output LCA with process LCA. The cumulative exergy consumption at this

stage is facilitated by using an ICEC to money ratio as given by Equation (2.1)

ICEC i=mi Ci R ICEC ,i (2.1)

where,

ICECi is the industrial cumulative exergy consumption for product i,

mi is the mass flow of the product i,

C i is the price of product per unit of mass,

R ICEC,i is the ratio of ICEC to money in the economic sector corresponding to product i.

The ecosystem scale proposed by Yi et al. (2004) further extends the boundary of the analysis

to capture the contribution of the ecological goods and service. The ecosystem scale is also

facilitated and conveniently determined by using an ECEC to money ratio as illustrated in

Equation (2.2)

ECEC i=miC i RECEC , i (2.2)

where,

ECEC i is the cumulative exergy consumption for product i,

mi is the mass flow of product i,

C i is the price of product per unit of mass,

27

RECEC ,i is the ratio of ECEC to money in the economic sector corresponding to product i

However, the multi-level view proposed by Yi et al. (2004) and other researchers such as

Liao et al. (2012) do not offer detailed physical quantification of the processes and flows

pertaining to the resource consumption at the different levels. A detailed multi-level analysis

is required to unfold how a resource, before and after being processed at different stages,

flows within the totality of the system, which is essential for the identification of important

flows that can be either removed or improved through integration with flows associated with

other products or services in the system.

Most recently, a methodological framework has also been developed by Hanes and Bakshi

(2015a, 2015b) to address analyses at different scales. However, it can be remarked that there

are still some confusions and unsettled aspects in the existing work. While the system

boundary is an important consideration in all previous studies, there is no single detailed and

holistic quantitative study encompassing at the same time ecosystem (i.e. natural processes),

production as well as consumption of a final product or service (i.e. human systems). In

particular, the consumption side of a product or service has largely been overlooked. In terms

of scope, it can be observed that the resource burdens of constructing plant, equipment and

machineries have been largely overlooked in existing resource accounting methods and that

there are no detailed studies explicitly acknowledging the importance of quantitatively

accounting for these resources. Additionally, though recent studies on resource accounting

include a wide range of resources there are still controversial aspects related to the

admissibility of the inclusion of capitals and potential double counting of labour and money

resources (Rocco et al., 2013). There have also been many studies done around the design of

biorefineries using an LCA approach. Fahd et al (2012) also developed an LCA-based

sustainability multi-scale multi-method approach for integrated assessment of material,

embodied energy, environmental impact and economic flows and performance. Alvarado-

Morales et al. (2009) presented a cost effective, operation and sustainability approach for the

design and analysis of biorefineries to generate new alternatives with respect to wastewater

reduction and efficient downstream separation. Other papers by Ojeda et al. (2011), Bao et al.

(2011), Heyne and Harvey (2013), Akgul et al. (2012) and (Brehmer et al., 2009; Fahd et al.,

2012) as reviewed by Martinez et al (2013) in the area of biorefinery process design,

integration and sustainability indicators do not offer a differential environmental impact

analysis of the smallest element (such as a stream associated with a unit operation) to the

28

largest element (such as a whole system) by means of a unified framework. To cover this gap,

Martinez et al (2013) developed a new methodology that provides insights into the

differential economic and environmental performances of individual elements in a process

network, directing to network hot spots analysis. However, previous studies do not offer a

detailed multilevel quantitative analysis of the processes and flows pertaining to resource

consumption at the various levels of a particular system from the unit production to the point

of local consumption of the product or service inclusive while taking also into consideration

environmental remediation processes in the system boundary.

2.3 Aim and objectives for the conceptual framework

The aim of the work presented in this chapter is to develop a coherent framework for resource

accounting to support decision making during the evaluation of alternatives for human

production and consumption activities. This is with the specific objectives to fill gaps of

existing exergy-based frameworks including:

Proposing a structural and quantitative multi-level understanding of a system by

considering not only the production and consumption of products or services but also

the ecological and industrial/technological processes. This provides a more holistic

approach to resource accounting by accounting for all types of resources including

renewable resources and non-energetic resources while avoiding double counting and

a simpler approach to accounting for ecosystem and natural processes.

Developing a methodology that can support the analysis work where resource

efficiency or resource consumption can be used as an indicator. By monitoring and

assessing resource efficiency or consumption resource, the effect of retrofitting or

introducing a new component into a system can be analysed and can serve as guide

for the selection of those components that improve the performance of a system. A

resource accounting methodology that can guide the design or retrofit of components

in a consumption-production system by using resource efficiency or resource

consumption as an objective function to be optimised will be especially useful.

Therefore, this work presents a unique adaptation of the Cumulative Exergy Resource

Accounting, CERA, methodology based on a structural and quantitative multilevel

understanding of production and consumption of product or service within a defined system

boundary. The proposed multilevel analysis and the resource accounting methodology can be

29

used as a framework and basis for comparing different design options at any particular level.

Also, in addition to resource consumption and efficiency, any other metric/indicator could be

used with the proposed multilevel framework. The following section outlines the conceptual

framework developed for resource accounting and is aimed at understanding resource

consumption at different levels. A conceptual framework that identifies the key aspects of a

system such as system boundary, types of flows and processes, principles for determining

resource consumption while avoiding ambiguity and double-counting at multiple levels of

analysis is presented in this chapter. Building on the conceptual framework, a quantitative

approach to resource accounting based on the concept of cumulative exergy consumption will

be described in the next chapter. The scope of the proposed framework is to assess resource

consumption from a technical perspective and aims to provide support for decision making at

the technical level of interest to process engineers and inform decision-makers in industry,

government or non-government organisations particularly for the purpose of strategic

planning, product design or redesign. While the framework does not directly contain business

logics or management principles for commercial operations, the systematic approach on

physical resource accounting has the potential to provide a solid basis for informing the

relevant stakeholders with respect to the resource impact of their decisions.

One of the major limitations of the approach is the uncertainty associated with the cumulative

exergy of the data used. Data uncertainty appears to be a common limitation to holistic

approaches to resource assessment (Brown and Ulgiati, 2010; Yang et al., 2010). In a

practical application, it may be addressed by a careful combination of quality sources of

cumulative exergy consumption data, possibly supplemented by other types of data sources

such as LCA databases. Besides, this approach currently does not take into account the full

range of environmental impacts such as climate change effects, toxicity and impacts of

monoculture on biodiversity and offers no indication on resource depletion. For instance, the

approach does not tell if groundwater is depleted, a river is polluted or the system is a

monoculture with little biodiversity. The approach needs to be completed with other

approaches for a comprehensive sustainability accounting or assessment and a part of a wider

multi-criteria evaluation. By combining the proposed system characterisation and modelling

of resource flows with other approaches such as LCA, the wider environmental implications

of a system can be assessed. These limitations must be considered when interpreting the

results obtained from this approach. The proposed framework will be demonstrated through a

case study on the production and consumption of sugarcane bioethanol, in Chapter 4.

30

2.4 A conceptual framework for resource accounting

2.4.1 Basic concepts of system, environment, process and flow

A system for which resource accounting is considered is defined by (i) the resource-

embedded incoming flows that enter the system from its environment, (ii) the process or

processes that convert the flows from the environment and (iii) the outgoing flows produced

by the process(es) that leave the system and enter its environment. Resource accounting can

be carried out for systems of different levels or scales. In particular, the global level and the

local level can be distinguished here; a further elaboration of system levels is presented in

Section 2.7.

At the global level, the system comprises all human-driven processes as well as the processes

in the natural eco-system which can potentially be affected by human-driven processes. In

other words, it covers all processes where human decisions and actions could make an

impact, thus forming a scope within which resource accounting, with the purpose of

evaluating and shaping human decisions and actions, remains relevant. Following this

principle, this scope does not include processes which will occur in the future (e.g. formation

of solar energy) or have occurred historically (e.g. formation of fossil fuels) independently

from human intervention. Such processes essentially form the environment of the global

system. To facilitate the subsequent discussions in this thesis, processes within the (global)

system are referred to as Type-I processes and those within its environment as Type-II

processes.

In contrast with the global system, a local system comprises human-driven processes as well

as natural processes that can be affected by human decisions and actions at the corresponding

local level, be it of a village, a company, an industrial park, a country or a region. As such,

this local system exchanges flows with its environment, which typically includes both other

local systems and natural processes that do not form part of any local systems. A general

representation of the system and environment framework considered in this resource

accounting analysis is illustrated in Figure 2-1. The system boundary separates the system

from the environment. The study proposed in this report will focus mainly on local system as

opposed to global system.

31

Figure 2-1: System and environment framework

2.5 Resource flows and their accounting principleThis section details the potential input flows to the system from Type-I processes and Type-II

processes and their accounting principle. The input flows considered in this study for a local

system include the followings:

Material and energy flows from the natural processes

Material and energy flows from human/industrial/technological processes

Human labour.

2.5.1 Accounting for resource flows from Type-I processes

Material and energy flows from either human-driven processes or natural processes that are

affected by human activities are accounted for by their cumulative exergy. Labour is also

accounted by its cumulative exergy and this can be achieved most conveniently by an exergy

to labour conversion factor. Several approaches have also been developed to evaluate the

exergy equivalence of labour input (Gong and Wall, 1997; Kotas, 1985; Sciubba, 1995; Wall,

1999). The human labour input to the system has been popularly determined in exergy

equivalent by using Equation (2.3):

BH=NH KH (2.3)

32

Environment

Output flows

System boundary

Input flowsSystem

where,

BH is the total exergy equivalence of human labour

N H is the number of work hours

K H is the conversion factor for converting human labour into exergy equivalence and can be

evaluated by using Equation (2.4):

K H=Cumulative Exergy Consumptionof the societyNumber of workers∗work hours / year (J/work hour) (2.4)

In principle, labour input to a system should be re-created by consuming resources generated

outside the system, to avoid double counting.

Also, money will not be considered as additional flows to the system but will rather be used

to facilitate the cumulative exergy of other flows. Money will be used as a basis to estimate

the cumulative exergy cost associated with physical resource flows which are difficult to

estimate directly. For example, the cumulative exergy resource consumption for

manufacturing and provision of equipment (e.g. agricultural machinery, chemical reactors,

wind turbines, solar panels) can be estimated indirectly by using their capital cost if

determining their cumulative exergy resource consumption directly becomes too cumbersome

and impractical. This assumes a direct correlation between economic costs and resource

consumption. Monetary flows can be converted into their exergy equivalence by using

Equation (2.5) given by Hau and Bakshi (2004).

Bc=Fc K c (2.5)

where,

Bcis the total exergy equivalence of money flow

F c is the flow of money into the system

K c is the conversion factor for converting money into exergy equivalence and can be

evaluated by using Equation (2.6) adapted from Hau and Bakshi (2004):

K c=Cumulative Exergy Consumption of the society

Economic Gross Product of Society (J/£) (2.6)

2.5.2 Accounting for resource flows from Type-II processes

Input flows from Type-II processes will only be accounted for by their exergy content as

opposed to a cumulative exergy value. Only the exergy content of the material fossil fuels

and ores is considered and not the cumulative exergy consumption of the ecological

33

processes required for their natural formation. This is because their formation is considered

historical burden and is not relevant to any future human activities and will be considered as

part of the environment. It is reasonable not to account for the formation of fossil fuels as

different technologies for the production of same goods and services will be compared within

a relatively short time period, e.g. a few years or decades, while the timescale for

replenishing fossil fuels by ecological processes is much longer, e.g. millennia. Since human

activities will not alter the generation of fossil fuels and ores within the time scale of interest,

they will be accounted for only by their input flows’ exergy content into the system, not the

cumulative exergy consumption.

Similarly, the cumulative exergy consumption of the ecological processes for the formation

of wind and solar radiation (sunlight) will not be taken into consideration. These renewable

resources are present whether they are being used or not and there might be no real benefit to

account for their cumulative exergy, under the realistic assumption that human activities will

not noticeably affect the future formation of sunlight or wind. Moreover, in some studies,

renewables are often differentiated from non-renewables by not accounting for their exergy

inputs (Wall, 2011). In our framework, the resource value (in the form of exergy content) of

renewables is accounted for because these resources in principle have alternative uses. It

should be noted that biomass is a special case of a renewable resource. Human activities can

affect the production of biomass, which is thus considered as a Type-I process. Therefore,

when there exists a biomass flow as an input to a system, its resource value is quantified not

by its exergy context, but rather by the cumulative exergy consumption to produce the

biomass using water, land, sunlight and natural and synthetic nutrients.

2.6 Resource consuming processes

Following the discussions in the previous subsections of section 2, the system boundary

defines the starting point of cumulative exergy which will apply only to resource consuming

Type-I processes. A detailed overall depiction of the environment and the resource

consuming processes occurring in the society is illustrated in Figure 2-2. The society is the

system over which resource accounting is conducted. It comprises two main subsystems

namely the product provision subsystem and the product consumption subsystem. A process

in Figure 2-2 represents a sequence of processing units which can produce at least one final

product or intermediate products that can be further processed by other temporally and

34

spatially separate processes. Apart from the main flow being processed, the additional

material and energy inputs required for or generated from the processing of recycle,

exchange, reuse and discharge flows have been omitted in Figure 2-2 for simplicity. The

product provision subsystem comprises all the processes that are required to manufacture a

product or service.

Figure 2-2: Overall depiction of the Environment and Societal System

2.6.1 Flow making, capacity making, transportation and storage

Figure 2-3 further breaks down the product provision subsystem into flow making, capacity

making and transport and storage processes. Flow making processes include resource

extraction, agriculture and industrial/manufacturing systems which eventually deliver the

products for consumption and also provide for the energy and materials required by all types

of processes in the system. When resource extraction processes are part of the system, the

cumulative exergy consumption will be the aggregation of exergy content in the resource and

the exergy consumption from extraction to the point of the process unit where the resource is

used. Capacity making processes are the industrial systems that provide machineries,

equipment and consumables other than feedstock (e.g. catalysts, buildings and

infrastructures) to enable flow making and transport and storage components. In economic

terms, capital cost is an important factor together with operating cost. Similarly, in resource

terms, it is sensible to also account resource consumption for capital making processes if their

capital costs have not been neglected. For a holistic resource accounting analysis, it is

35

reasonable to examine the resources involved in capacity making processes especially for

renewable energy systems. For example, solar and wind energy are considered as unlimited

input resources while resources for capacity making components (e.g. metals, rare earth

elements for wind turbines and solar panels) are limited. In this study, capital resources will

refer to all resource consumption for capacity making processes while operating resources

will be the resources that are used directly to operate the flow making processes for the

production of the desired flows. Transport and storage processes serve both flow and capacity

making processes and acts as the interface between the product provision subsystem and the

product consumption subsystem.

Figure 2-3: Detailed depiction of the product provision subsystem

2.6.2 Recycle, exchange and repair processes

Recycle of flows is a resource consuming process that can exist for flow making processes,

capacity making processes as well as for transportation and storage. Flows can also be

exchanged between the different processes. The exchanged flows might need to be processed

before being used as input flows to other processes, hence consuming resources. The final

product from the product consumption subsystem can either be repaired or recycled back into

the product provision subsystem. Both repair and recycling of final product will incur some

resource consumption that needs to be accounted for in the analysis. For simplicity, the

resource flows for recycling, exchange and repair are not shown in Figure 2-3.

2.6.3 Environmental remediation processes

All the environment pollutant flows produced within the system need to be processed and

treated before they can be discharged into the environment. The environment remediation

36

processes for treating pollutants before they are released into the environment form part of a

system and these processes can be either industrial/technological processes or natural

processes or a combination of both. The environmental remediation cost is defined in this

study as the cumulative exergy consumption of the environmental remediation processes

required to treat the environmental pollutants to the extent that, in principle, no harm is made

to the environment, or, in practical terms, a certain set of environmental regulations are met.

2.7 Multilevel structure of a system

2.7.1 Onion model

In an attempt to provide a better understanding of a system and support the development of a

resource accounting methodology, a system is conceptualised as a hierarchical structure

having the following levels:

Unit level which involves a single conversion step where input to the unit is processed

to output with no recycle and reuse involved. For example, the unit level could

involve a molecular sieve unit for ethanol dehydration.

Process level where intra process recycling and reuse (i.e. recycling and reusing flows

between the processes) can occur. For example, the process level could involve

different units such as from cane milling to ethanol dehydration units connected

together with intra process water flows from ethanol distillation recycled back into the

cane milling unit to reduce freshwater consumption for the purpose of producing

ethanol from cane.

Inter-process level with exchange of flows between two or more processes. For

example, the inter-process level could involve different processes for the production

of different products/services such as cane ethanol from an ethanol plant and

electricity production from a power plant with exchange of flows between them (i.e.

bagasse from cane ethanol plant and heat/electricity from the power plant).

Production-consumption level with reuse and recycle of products; this level includes

consumption by the society. For example, the production-consumption level could

involve taking into consideration all the resources consumed during the consumption

stage of the product or service such as resources for environmental remediation from

ethanol consumption in vehicles.

37

The structural conceptualisation of a system can be represented by the onion model shown in

Figure 2-4. The onion model represents the different levels at which a system can be

analysed. From the onion model, the sequence for resource consumption accounting starts

from the unit level and continues towards the production-consumption level.

Figure 2-4: Onion model for structural representation of Environment-Society System

The physical quantification of resource consumption might be different at each system level,

depending on the processes, recycling, exchange and repair flows that become more

prominent. As such, at each level different key decisions can be made. At the unit level,

decision making would involve choosing the most appropriate and resource efficient unit

operation. At the process level, the focus of decision is on selecting among the best process

design to adopt. Key intra-process recycling flows that can significantly decrease resource

consumption are identified at the process level. Moreover, analysing resource consumption at

the inter-process level will give an indication of the industrial synergies to promote or adopt.

Flows that can potentially be exchanged between the processes are identified at this level. At

the production-consumption level, it could help to identify links between production

processes, the ecosystem and consumption society with the aim of achieving a cradle to

cradle resource model, similar to the ‘circular economy concept’ where the resources are

continuously recycled and used within the system, thus increasing sustainability. In concrete

terms, specific decisions at the production-consumption level could involve choosing

38

Process

Unit

Inter-Process

Production-Consumption

Environment

between repairs or recycling of final products after they have been consumed. Table 2-2

summarises briefly the main decisions at the different levels of analysis.

Table 2-2: Decisions making at different levels of analysis

Level DecisionUnit Technology development/Innovation

Process Plant/factory designInter-process Industrial symbiosis

Production-consumption Circular economy, repair versus recycling

2.7.2 Significance of multi-level view

One of the rationales behind adopting a multi-level view for resource accounting is that

adopting a narrow view by analysing only individual processes might lead to shifting the

resource consumption impacts to other parts of the value chain of the product and/or service.

Reduction/increase of resource consumption at one level might lead to increased/reduced

resource consumption at a higher level. For instance, a lower conversion rate adopted by a

chemical reactor will result in lower resource consumption at this unit level due to the lower

energy and capital making cost for the smaller size reactor. However, this might adversely

cause higher process recycling cost as the separation to enable recycling has to work harder;

which in turn might lead to a higher overall resource cost at the whole process level. Also, at

the production-consumption level, it might be worthwhile to spend more resources in

manufacturing a new product if that will lead to the significant reduction of resource

consumption by-product recycling/repairing at a later stage in the product life cycle.

A multi-level view gives a clear insight and understanding of the whole system. It serves the

purpose of reminding decision makers of the implication to other levels when making choices

at a particular level, and offers the potential to identify key components and flows that can be

either removed or improved through integration and linkage with other flows or components

in the system. The structural depiction through a multi-level view illustrates how a resource,

before and after being processed at different levels, flows within the totality of the

environment and society system. This depiction serves to provide a basis for assessing

resource accounting quantitatively and also for improving resource utilisation and

39

consumption at the different levels. It is visually similar to the conceptual design approach

first developed by Douglas (1988) who originally evaluated process hierarchically starting

from the process unit itself and gradually expanding the system boundaries as successive

scales/levels are added. Douglas (1988) pointed out that a hierarchical approach facilitates the

evaluation procedure by starting with simple systems and increasing complexity gradually as

successive layers of information are added.

2.8 Summary of conceptual framework for resource accounting

This chapter presented a holistic and comprehensive framework for assessing resource

consumption in industrial production processes and their interactions with the environment

and the consumption system. Fundamental concepts of system, flow, process and

environment were introduced and resource-generating processes categorised as either Type-I

or Type-II processes. A multi-level structure was then also developed for resource accounting

at unit, process, inter-process and production-consumption taking into consideration intra-

level and inter-level connections. A resource accounting algebra will next be formulated in

the ensuing chapter based on the multi-level structure.

40

Chapter 3: Algebraic quantification of resource accounting

3.1 Resource accounting algebra

Following the conceptual framework for resource accounting, a quantitative assessment of

resource consumption at each of the levels is proposed in this chapter; also published in

Leung Pah Hang et al. (2016a). It proposes quantities to support the evaluation of important

resource decisions such as intra-level recycling, inter-level exchange, and repair and

recycling of used products. The general resource accounting equation (3.1) that can be

applied to each system level can be expressed as:

Total resource consumption = total operating resource consumption + total capital resource

consumption + total resource consumption for environmental remediation (3.1)

In Equation (3.1), the resource consumption is expressed in terms of exergy. The embedded

resource consumption of flows from Type-I processes will be accounted by the cumulative

exergy consumption incurred during their production. Flows resulting from Type-II processes

will be accounted for by their respective exergy content to acknowledge that they have

alternative competing uses. Sections 3.2 to 3.5 detail the resource accounting algebra/model

from unit to production-consumption levels and Table 3.1 describes the indices used in

Figures 3-1 to 3-4.

Table 3-2: Description of indices used in Figures 3-1 to 3-4

Indices Descriptionu Unit, u Ui Input flows, i I (other than capital and environmental remediation resource flows)mc Capital resource flows, mc MCw Environmental remediation resource flows, w Wp Process, p Pr Resource flows require to process intra recycling flows, r Rim Intermediate flows, im IMir Fresh input flows replaced by the intra recycled flows, ir IRjrp Output flows from the units that could be used as intra recycling flows, jrp JRPαi Proportion of output flows from the units actually used as intra recycling flows

ac Flows that would have been required for treating flows to be discharged to the environment if they were not recycled, ac AC

ip Inter-process, ip IP

41

r,p Resource flows require to process recycling flows in process pi,p Input flows from process pj,p Output flows from process pex Flows exchanged between different processes p, ex EXei Fresh input flows that the exchange flows replaced, ei EI

enx Avoided flows that would have been required for treating the discharged flows if they were not exchanged between different processes, enx ENX

pc Production-consumption, pc PC

rcs Flows required for processing recycled flows from product consumption to product-provision subsystem, rcs RCS

re Flows required for repair, re RE

j,p,t Resource flows required for transporting the desired output j from process p to the point of consumption

end Resources consumed in the use phase of the product, end END

rc Fresh flows avoided with recycling flows from product consumption to product-provision subsystem, rc RC

np Fresh flows avoided for making a new product if it is not repaired, np NP

enr Resources for disposing the used-products if they are not recycled back into the product- provision subsystem, enr ENR

A formal classification of flows and their corresponding definition as used in this Chapter is

also given in Table 3-2.

Table 3-2: Classification of flows

Type of flow Description

Fresh Imported or locally produced flows from other production systems to be used as raw materials

Desired

Flows that were required to be produced at unit, process, inter-process, or production-

consumption level (e.g. excluding any waste and by-products)

WasteFlows that could potentially be harmful to the environmental and would require treatment before being released into the environment

Internal/Intra recycled Flows produced and used again within the units of a same process

Exchange Flows produced and used again between different processes

Input flows for transportation

Flows/resources required for transportation to take place (e.g. fuel, capital resources for

manufacturing the vehicle, resources for treating effluents from the vehicle)

3.2 Resource accounting at unit level

Figure 3-1 is a simple representation of the unit level. The unit level comprises the unit itself

and also includes an environmental remediation process for treating any environmentally

harmful pollutants and waste flows produced by the unit. Remediation can be achieved by

technological or natural processes to a level which is harmless to the environment or within

42

limits set by environmental regulations before they are released to the environment. Note that

in Figure 2-1, the environmental remediation has been represented as a dotted box as it will

be present in addition to the unit itself if there are any harmful effluents from the unit that

require treatment before release into the environment. This also applies for the process, inter-

process, production-consumption levels and for simplification purposes, the environmental

remediation will not be represented in Figures 3-2 to Figure 3-4.

Figure 3-5: A representation of the unit level

Consider unit u from Figure 3-1, the total exergy consumption by this unit (ExCu) for the

production of the desired output flow is the sum of three elements: (i) total exergy

consumption of its inputs i=1 to I, each with ExCi , u as its specific cumulative exergy content,

representing resource consumptions by flows from upstream units and operating resources (

∑i=1

I

ExC i ,u Fi , u¿, (ii) total exergy consumption for providing the capital resources mc = 1 to

MC for unit u (∑mc=1

MC

ExCmc ,u Fmc,u) and (iii) total exergy consumption for all the

environmental remediation processes associated with unit u(∑w=1

W

ExCw ,u Fw , u) , as given

algebraically in Equation (3.2).

ExCu=∑i=1

I

ExCi , ,u Fi , u + ∑mc=1

MC

ExCmc ,u Fmc ,u + ∑w=1

W

ExC w ,u Fw ,u ∀u U (3.2)

3.3 Resource accounting at process level

The process level is represented in Figure 3-2 which illustrates a process with two units and

intra-process recycling.

43Fim

Fmc

Fmc

Fw

Treated flows Environmental remediation Unit u

F i

Desired flows Boundary at unit level

Fmc

Fmc

Unit 2Unit 1

Boundary at process level Fr

Fj

Internal recycled flows αi Fjrp

Fi

Figure 3-6: A representation of the process level

The total exergy consumption accounting for producing the desired flows at the level of

process p (ExC p) can be determined by the difference between (i) the sum of the total exergy

consumption of all units u=1 to U (∑u=1

U

ExC u , p ¿ and total exergy consumption of all the flows

required to process the internal/intra recycling flows r = 1 to R (∑r=1

R

ExC r , p F r , p ¿ and (ii) the

sum of the total exergy consumption of all the intermediate flows (i.e. flows produced within

the process and flowing from one unit to another and represented by a dotted arrow in Figure

3.2) im = 1 to IM (∑ℑ=1

ℑ

ExCℑ , p Fℑ , p), total exergy consumption associated with all the fresh

input flows replaced by the recycled input flows ir = 1 to IR (∑ir=1

IR

ExCir , p F ir , p) and total

exergy consumption of all the flows ac = 1 to AC that would have been required for treating

flows to be discharged to the environment if they were not recycled (∑ac=1

AC

ExCac , p Fac , p ¿. The

resource accounting algebra at process level is expressed in Equation (3.3).

ExC p=∑u=1

U

ExCu , p+∑r=1

R

ExCr , p Fr , p - ∑ℑ=1

ℑ

ExCℑ , p Fℑ , p - ∑ir=1

IR

ExCir , p F ir , p - ∑ac=1

AC

ExCac , p Fac , p

∀ p P (3.3)

The resource benefit of process recycling (i.e. does process recycling reduce fresh resource

consumption or more resources are consumed in order to use the recycled flows),ɳ pr, can be

expressed as the ratio of (i) the sum of the total exergy consumption of all the fresh input

flows replaced by the recycled input flows and the total exergy consumption that might be

required for treating the flow before discharging if there was no process recycling arranged,

to (ii) the total exergy consumption of recycling as shown in Equation (3.4):

44

ɳ pr = ∑ir=1

IR

ExCir , p Fir , p+∑ac=1

AC

ExC ac , p Fac , p

∑r=1

R

ExCr , p F r , p

∀ p P (3.4)

The higher the ratio, the better will be the recycling efficiency and the reduction in fresh

resource consumption.

3.4 Resource accounting at inter-process level

A schematic representation of the inter-process level with two processes is shown in Figure

3-3 with recycling as well as exchange flows.

Figure 3-7: A representation of the inter-process level

The resource accounting at the inter-process level (ExCip) is given by the net difference

between the resources incurred and gained from exchange of flows. Thus, it is given by the

45

Fmc,p=2

Fi,p=1

Fmc,p=1

Process p=1

Fr,p=2 Flows associated with recycling in process p=2 Flows for processing

exchange flows, Fex,1,2

Process p= 2Desired flows, Fj,p=2

Fr,p=1 Flows associated with recycling in process p=1

#Internal recycling flow, αiFjrp

Exchange flows, Fe, 1,2

Fi,p=2

Fj,p=1

Flows for processing recycled or exchanged flows

Flow Boundary Processing of recycled or exchanged flows

difference between (i) the sum of the total exergy consumption at the inter-process level (

∑p=1

P

ExC p ,ip) and the total exergy consumption of processing exchange flows ex = 1 to EX at

the inter-process level (∑p '=1

P

∑p=1

P

∑ex=1

EX

ExCex , p , p' ,ip Fex , p , p' , ip) and (ii) the sum of the total exergy

consumption of fresh input flows ei = 1 to EI that the exchange flows replaced (

∑p=1

P

∑ei=1

EI

ExC ei , p ,ip F ei , p , ip) and the total exergy consumption of the avoided flows enx = 1 to

ENX that would have been required for treating the discharged flows if they were not

exchanged between different processes ( ∑enx=1

ENX

CExCenx ,ip F enx ,ip). This accounting is expressed

in Equation (3.5).

ExCip=∑p=1

P

ExC p ,ip+∑p '=1

P

∑p=1

P

∑ex=1

EX

ExCex , p , p' ,ip Fex , p , p' , ip - ∑p=1

P

∑ei=1

EI

ExC ei , p ,ip F ei , p , ip -

∑enx=1

ENX

CExCenx ,ip F enx ,ip ∀ ip IP (3.5)

The resource benefit of exchange flows (i.e. does exchange of flows reduce fresh resource

consumption or more resources are consumed in order to use the exchange flows),ɳ IPE, can be

expressed as the ratio of (i) the sum of the total exergy consumption of all the fresh input

flows that the exchanged input flows replaced and the total exergy consumption that might be

required for treating the flow if it was not exchanged, to (ii) the total exergy cost of

processing any exchange flows at the inter-process level, as given in Equation (3.6):

ɳ IPE = ∑p=1

P

∑ei=1

EI

ExCei , p ,ip Fei , p ,ip+ ∑enx=1

ENX

ExCenx , ip Fenx , ip

∑p '=1

P

∑p=1

P

∑ex=1

EX

ExC ex , p , p' , ip Fex , p , p' ,ip

∀ ip IP (3.6)

The higher the ratio, the better is the resource benefit of exchange flows and thus the

reduction in fresh resource consumption.

3.5 Resource accounting at production-consumption level

The schematic representation of the production-consumption level is shown in Figure 3-4 and

captures the life of the desired product and/or service after it has been produced. The product

46

can be repaired, recycled or simply disposed of after it has been consumed. As given in

Equation (3.7), the resource consumption at the product-consumption level, ExC pc, including

repair and recycling of desired products based on the boundary shown in Figure 3-4 can be

expressed as the difference between two groups of terms. The first group, representing

resource expenditures, includes (i) the total exergy consumption at the inter-process level

(product-provision subsystem) (ExC IP), (ii) the total exergy resource consumption for

transporting the desired products to the point of consumption (product-consumption

subsystem) (∑t =1

T

∑p=1

P

∑j=1

J

F j , p ,t , pc ExC j , p ,t , pc), and (iii) the total exergy resource consumption

associated with (iii-a) processing of the recycled flows from the product consumption

subsystem (∑rcs=1

RCS

F rcs , pc ExC rcs, pc¿ ,(iii-b) processing of the repaired flows from the product

consumption subsystem (∑ℜ=1

ℜ

Fℜ , pc ExCℜ , pc), and (iii-c) resources consumed in the use phase

of the product ( ∑end=1

END

F end , pc ExC end , pc). Note that the last term might include resources

consumed for using the product or service, treating any effluents during product use, and end-

of-life disposal. Each of these might apply to some products but not others. For instance, a

washing machine consumes resources during its use, which does not apply to ethanol as a

fuel. The second group, representing avoided consumptions due to resource saving measures

at the production-consumption level, includes resource (i) for providing fresh flows avoided

with recycling (∑rc=1

RC

F rc , pc ExC rc , pc), (ii) for making a new product if it is not repaired (

∑np=1

NP

Fnp, pc ExCnp , pc), and (iii) for disposing the used-products if they are not recycled back

into the product provision subsystem ( ∑enr=1

ENR

F enr , pc ExCenr , pc).

ExC pc ¿ ExC IP+

∑t=1

T

∑p=1

P

∑j=1

J

F j , p ,t , pc ExC j , p ,t , pc+ ∑rcs=1

RCS

Frcs , pc ExC rcs, pc+∑ℜ=1

ℜ

Fℜ , pc ExCℜ , pc+ ∑end=1

END

F end , pc ExC end , pc−∑rc=1

RC

Frc , pc ExCrc , pc−∑np=1

NP

Fnp , pc ExC np , pc

- ∑enr=1

ENR

F enr , pc ExCenr , pc ∀ pc PC (3.7)

47

Figure 3-8: A representation of the product-consumption level

The resource benefit of product recycling (i.e. does product recycling reduce fresh resource

consumption or more resources are consumed in order to do product recycling), ɳrecycling , can

be expressed as the ratio of (i) the sum of the total exergy consumption of all the fresh input

flows replaced by the recycled input flows and the total exergy consumption for disposing the

products if they are not recycled back into the product provision subsystem, over (ii) the total

exergy consumption for product recycling; given in Equation (3.8).

ɳrecycling=∑rc=1

RC

Frc , pc ExCrc , pc+ ∑enr=1

ENR

Fenr , pc ExC enr , pc

∑rcs=1

RCS

Frcs , pc ExC rcs, pc

∀ pc PC (3.8)

The resource benefit of product repairing, (i.e. does product repairing reduce fresh resource

consumption or more resources are consumed in order to do product repairing), ɳrepair , can be

48

Transportation

Product-provision subsystem

Input flows for transportation

F rcs, flows for processing recycled flows

Product Consumption Subsystem (PCS)

F r, flows recycled back from PCS

Fℜ, flows required for repair

Treated flows discharged to environment

Flows for processing recycled or exchanged flows

Reconditioning/Repairing for reuse

expressed as the total exergy consumption of the extended service life of the product brought

by repairing, as given by Equation (3.9). The total exergy consumption of the extended

service life can be calculated as a proportion of the total exergy consumption for the

manufacture of a new product.

ɳrepair= ExC pc x Extended service life

standard service life

∑ℜ=1

ℜ

F ℜ, pc ExCℜ, pc

∀ pc PC (3.9)

The higher the ratio of ɳrecycling and ɳrepairthe better is the resource benefit of product recycling

and repairing respectively and thus the reduction in fresh resource consumption.

Table 3-3 gives the physical interpretation of each of the terms used in Equations (3.2)-(3.9).

All of these notations are parameters/pre-defined quantities.

Table 3-3: Description of the notations used in Equations (3.2) - (3.9)

Notation DescriptionEquation (2)ExCu Total exergy consumption of producing a product or service at the unit u level (e.g. MJ h-1)

ExCi Specific (i.e. per unit input) exergy consumption of input i (e.g. MJ kg-1), i=1 to I

F i Flow rate of input i (e.g. kg h-1 in the case of a material input)

ExC w Specific exergy consumption of input w for environmental remediation (e.g. MJ kg-1), w = 1 to W

Fw Flow rate of input w required for environmental remediation from unit u (e.g. kg h-1)

ExC mc Specific total exergy consumption of capital resources corresponding to input mc (e.g. MJ kg-

1), mc = 1 to MCFmc Flow rate of capital resources input mc (e.g. kg h-1)Equation (3)ExC p Total exergy consumption at the level of process p (e.g. MJ h-1)

ExC r Specific total exergy consumption of all the flows r required to process the internal recycling flows (e.g. MJ kg-1), r=1 to R

F r Flow rate of internal recycling flows r (e.g. kg h-1)

ExCℑ Specific exergy consumption of all the intermediate flows im (e.g. MJ kg-1), im=1 to IM

Fℑ Flow rate of intermediate flows im (e.g. kg h-1)

ExCir Specific exergy consumption associated with all the fresh input flows ir replaced by the recycled input flows (e.g. MJ kg-1), ir=1 to IR

F ir Flow rate of all the fresh input flows ir replaced by the recycled input flows (e.g. kg h-1)

ExC ac Specific exergy consumption of all the flows ac that would have been required for treating flows to be discharged to the environment if they were not recycled (e.g. MJ kg-1), ac=1 to AC

Fac Flow rate of flows ac that would have been required for treating flows to be discharged to the environment if they were not recycled (e.g. kg h-1)

49

Equation (5)ExCip Total exergy consumption at the inter-process level ip (e.g. MJ h-1)

ExC ex , p , p ' Specific exergy consumption of processing any exchange flows, ex, from process p’ to p at the inter-process level (e.g. MJ kg-1), ex = 1 to EX, p (p’) =1 to P

F ex , p , p' Flow rate of the exchange flows, ex, from process p’ to p at the inter-process level (e.g. kg h-1)

ExC ei , p Specific exergy consumption of fresh input flows, ei, that the exchange flows replaced for process p (e.g. MJ kg-1), ei=1 to EI, p =1 to P

F ei , p Flow rate of the fresh input flows ei that the exchange flows replaced for process p (e.g. kg h-1)

ExC enx Specific exergy consumption of flows enx that would have been required for treating the discharged flows if they were not exchanged between different processes (e.g. MJ kg-1), enx= 1 to ENX

F enx Flow rate of avoided input flows enx that would have been required for treating the discharged flows if they were not exchanged between different processes (e.g. kg h-1)

Equation (7)ExC j , p , t Specific exergy consumption associated with transport t of final products j from process p (e.g.

MJ kg-1), j =1 to J, p =1 to P, t = 1 to TF j , p , t Flow rate of input flows required for transportation t of output j from process p (e.g. kg h-1)

ExC rcs Specific exergy consumption associated with the input flows rcs required for processing of the recycled flows from the consumption subsystem (e.g. MJ kg-1), rcs = 1 to RCS

F rcs Flow rate of input flows rcs required for processing of the recycled flows from the consumption subsystem (e.g. kg h-1)

ExCℜ Specific exergy consumption of input flows re required for the processing of the repair flows from the consumption subsystem (e.g. MJ kg-1), re=1 to RE

Fℜ Flow rate of the input flows re required for the processing of the repair flows from the consumption subsystem (e.g. kg h-1)

ExC rc Specific exergy consumption associated with fresh flows rc avoided with recycling (e.g. MJ kg-1), rc = 1 to RC

F rc Flow rate of fresh flows rc avoided with recycling (e.g. kg h-1)

ExC np Specific exergy consumption of flows np required for making a new product if it is not repaired (e.g. MJ kg-1), np = 1 to NP

Fnp Flow rate of flows input np required for making a new product if it is not repaired (e.g. kg h-1)

ExC end Specific exergy consumption of all the flows end consumed in the use and end-of-life phases of the product (e.g. MJ kg-1), end = 1 to END

F end Flow rate of input flows end consumed in the use and end-of-life phases of the product (e.g. kg h-1)

ExC enr Specific exergy consumption of flows enr for disposing the products if they are not recycled back into the product provision subsystem (e.g. MJ kg-1), enr=1 to ENR

Fenr Flow rate of input flows enr used for disposing the products if they are not recycled back into the product provision subsystem (e.g. kg h-1)

3.6 Summary of the resource accounting algebra

In this chapter, a resource accounting algebra was formulated according to the multi-level

structure framework, offering key equations for quantitatively assessing different design

options at different level and supporting the evaluation of important resource decisions such

as intra-level recycling, inter-level exchange, and repair and recycling of used products. The

next chapter demonstrates its application on a case study for the production of ethanol from

sugarcane.

50

Chapter 4: Case study on multi-level framework for resource accounting

using algebras

4.1 Overview of case study on ethanol production from sugarcane

The developed conceptual framework for resource accounting from Chapters 2 and 3 is

demonstrated through a case study on the production of ethanol from sugarcane for a typical

plant with a capacity of 50,000 tonnes of ethanol per year, as shown in Figure 4-1. Biofuels

have been advocated as an important alternative for energy supply especially as a substitute

for fossil fuels (Pereira and Ortega, 2010) and Brazil is one of biggest ethanol producer in the

world and produces most of its ethanol from sugarcane. The increase in the demand for

biofuel as a renewable substitute for gasoline has intensified the need for more efficient

means of production (Dias et al., 2010). As such, the analysis of ethanol production and the

identification of key components that could potentially lead to huge reductions in resource

consumption are required. The case study illustrates how natural (e.g. photosynthesis),

51

agricultural (e.g. sugarcane cultivation) and industrial/manufacturing processes (e.g.

bioethanol plant) as well as the consumption of the final desired product are included in the

resource accounting from a life cycle thinking or total systems perspective. More specifically,

the application of the adapted cumulative exergy consumption -Cumulative Exergy Resource

Accounting-, CERA, methodology is shown at the unit, process, inter-process and

production-consumption levels. At the unit level, resource accounting is applied to cane

agronomy, cane transportation, cane milling, juice treatment, fermentation, distillation and

dehydration, distribution of ethanol and its final consumption. At the process level, the

industrial manufacture of ethanol and the production of steam and electricity from bagasse

are analysed as two separate processes and the resource savings due to internal recycling is

also assessed. At the inter-process level, resource accounting for the combined processes of

ethanol manufacture and steam and power generation and the benefit of exchange flows are

mainly investigated. At the production-consumption level, the resource costs for the

distribution and consumption of ethanol are additionally accounted for while the resource

cost benefit of product recycling and repair area analysed.

The basis of the resource accounting is one tonne of ethanol. The resource accounting does

not include land use for the ethanol plant as well as the resources used for site development

and plant installation. Moreover the human input for industrial manufacture of ethanol has

not been accounted for in the study following common practice in life-cycle studies (Iribarren

and Vázquez-Rowe, 2013). On the environmental remediation resource consumption, only

those for treating Vinasse, carbon dioxide emissions and methane emissions resulting from

bagasse decomposition have been considered. In this case study, the Carbon dioxide and

methane are removed through the natural ecological process of photosynthesis where plants

absorb carbon dioxide and use energy from sunlight to produce food (i.e. glucose) while

Vinasse is treated through technological processes. Sections 4.1 to 4.4 summarise the results

of the adapted CERA for the different units of ethanol production and consumption and at the

different levels of analysis. A comprehensive dataset for CExC for various resources used in

sugarcane ethanol production has been derived and presented in Appendix A. Data have been

taken mainly from studies on exergy consumption from sugarcane ethanol carried out by

Bastianoni and Marchettini (1996) and Palacios-Bereche et al. (2012, 2013) together with the

cumulative exergy database developed by Szargut et al. (1988), which are adapted to the

framework and the system boundary considered in this study.

52

Figure 4-9: The case study on sugarcane ethanol production

4.2 Ethanol production at the unit level

Figure 4-10: Ethanol production at the unit level

Figure 4-2 represents resource accounting at the unit level. A unit could be cane agronomy,

cane transportation, cane milling, cane juice clarification, fermentation, distillation,

dehydration, distribution, and consumption of ethanol. The resource accounting algebra is

illustrated here at the unit level to assess two alternative technologies for ethanol dehydration,

namely molecular sieve and azeotropic distillation. Using Equation (3.2) for resource

accounting at the unit level, the total exergy resource consumption for the dehydration of

hydrous ethanol to anhydrous ethanol using azeotropic distillation was determined to be

5.86×103 MJ/tonne ethanol more than that for molecular sieve distillation. Though, the CExC

of the upstream flows to both the molecular sieve and azeotropic distillation unit was similar,

the CExC of the operating resources (i.e. steam) and capital resources in terms of equipment

required for azeotropic distillation were found to be respectively 2.8 and 1.3 times higher

than that for molecular sieve. Moreover, azeotropic distillation uses cyclohexane as a

consumable with a CExC of 478 MJ/tonne ethanol.

53

Cane

CaneAgronomy

BagasseCO2

By-products

CO2

CO2

Transportation

Energy

Ethanol plant

Wastewater treatment

Transportation

Power plant

Land

Rain

Wind

Sun

CO2

Soil

Chemicals

Energy

Nutrients Pesticides

Water

Water

Absorption by plants

Absorption by plants

Electricity

Vinasse

Ethanol

Biogas

Energy

Energy Capital resources

Capital resources

Chemicals Capital resourcesWater

Operating flows

Output flows

Capital flows

Unit Upstream flows

4.3 Ethanol production system at the process level

Figure 4-11: Ethanol production system at the process level

Figure 4-3 illustrates the production of ethanol at the process level without any intra-process

recycling flows. At this level, the resource accounting algebra is used to assess a) the

recycling of water flows from the distillation unit to be used as imbibition water for cane

milling to reduce freshwater consumption and b) the recycling of reject flow (containing

mostly water and residual ethanol) from the molecular sieve dehydration units to the

distillation unit to recover more ethanol and thus reduce the amount of cane processed to

produce the same amount of final ethanol product. However, the recycled flows might require

some processing before being pumped back as input flows to a unit. Consequently, a proper

resource accounting at the process level can determine if intra recycling flows will have a

positive impact overall on resource consumption.

The ethanol/water flow from the regeneration bed can replace about 15% of the ethanol from

the fermentation beer; leading to a saving of 15% on all the resources used before the

distillation unit. Part of the water produced from the distillation unit can be recycled back

internally to the cane milling unit to fully satisfy its imbibition water requirements. Using

Equation (3.3) for resource accounting at the process level, it was determined that it would be

possible to reduce the total exergy resource consumption for the production of ethanol by

1.23×104 MJ/tonne ethanol through the implementation of these internal recycling flows.

4.4 Production of ethanol at the inter-process level

54

Steam/Electricity

Bagasse

Cane Milling Juice clarification

Fermentation

Cane from cane

agronomy

Distillation Dehydration

Filter cake Bagasse

Ethanol

Cane transportation

Wastewater CO2

water

Ethanol/water

Ethanol production process

Power and steam production process

Figure 4-12: Production of ethanol at the inter-process level

Figure 4-4 illustrates the production of ethanol at the inter-process level with exchange and

recycled flows. At the inter-process level, synergies between different types of processes;

including heterogeneous processes like ecological and technological processes, can be

investigated and their overall impact on resource consumption determined. For ethanol

production, the bagasse produced as a by-product of cane milling can be considered as a

useful resource and can be exchanged with the power plant so as to produce the steam and

electricity required by the ethanol plant. Using Equation (3.5) for resource accounting at the

inter-process level, it was inferred that an additional resource saving of 8.88×104 MJ/tonne

ethanol could be achieved by implementing exchange flows together with the internal

recycling flows.

4.5 Interaction between production and consumption of ethanol

Figure 4-13: The interaction between production and consumption of ethanol

Figure 4-5 shows the overall system of production and consumption of ethanol with exchange

and recycled flows. In principle, the resource benefit of practices that can potentially promote

sustainable consumption such as product recycling and repair after the product has been

consumed can be investigated at this level. However, ethanol is an immediate product of

consumption; and as such repair or recycling of ethanol is not applicable in this case. The

consumption unit for the ethanol case study would include upstream resource flows for

ethanol production, resources required for the environmental remediation of harmful

emissions (i.e. carbon dioxide) released during ethanol transportation from the ethanol plant

to fuelling stations and from ethanol combustion in vehicles; capital resources for

manufacturing the distribution infrastructures; and operating resources (i.e. diesel) used for

ethanol transportation in tank cars. Using Equation (3.7) for resource accounting at the

production-consumption level, the total exergy consumption for ethanol production and

55

Power and steam Production

Ethanol production

Transportation and distribution

Consumption of ethanol

consumption was determined to be 1.26×104 MJ/tonne ethanol excluding exergy flows from

Type-II processes.

4.6 Comparative analysis

The impacts of intra-process recycling and inter-process exchange of flows on resource

consumption have been further illustrated and analysed in three different scenarios. Scenario

analyses the production and consumption of ethanol without any recycling flows. Scenario 2

is on the production and consumption of ethanol from cane using intra-recycling flows, i.e.

recycled water flows from the distillation unit to be used as imbibition water for cane milling

and recycled ethanol/water flows from the molecular sieve dehydration units to the

distillation unit. It is assumed that these intra-recycling flows do not require any processing

before being used. In both the scenarios 1 and 2, energy is supplied externally and bagasse,

the by-product of cane milling, is not used. Scenario 3 analyses the significance of using

exchange and recycling flows, including the use of bagasse for energy supply. It is assumed

in scenario 3 that the bagasse exchange flow does not require any processing prior to being

used in the power station. The bagasse does not need any drying pre-treatment before being

sent to the boiler as most boilers nowadays can burn bagasse with moisture content of up to

50%. An efficient milling process usually produces bagasse with a moisture content of about

48% (BioEnergy Consult, 2014).

The total exergy flows from Type-II processes and Type-I processes were of different orders

of magnitude. For cane agronomy, the total operating exergy flows in terms of exergy content

from Type-II processes (e.g. sunlight, wind and land use for cane agronomy) was estimated

to be approximately 1.33×107 MJ/tonne ethanol as compared to the total operating exergy

flows from Type-I processes determined to be about 2.76×103 MJ/tonne ethanol based on

cumulative exergy consumption for scenario 1. Due to this huge disparity, the combination of

these two types of flows would hinder a practical analysis of the modifications done mainly

to improve performance within the boundary of the sugarcane processing, which involves

flows from Type-I processes only. Therefore, the detailed resource consumption analysis

focus primarily on flows from Type-I processes. The resource consumption for each of the

unit of scenarios 1, 2 and 3 is given in Table 4-1 and the detailed calculations can be found in

Appendix A of the thesis. Figure 4.6 shows the overall comparison of resource consumption

for the three scenarios excluding flows from Type-II processes. It can be observed that

56

environmental remediation accounts for most of the resource consumption followed by

operating resources and capital resources.

Table 4-1: Resource consumption for scenarios 1, 2 and 3

Resources Operating resources Capital resources Environmental remediation resources

Scenario Scenario 1

Scenario 2

Scenario 3

Scenario 1

Scenario 2

Scenario 3

Scenario 1

Scenario 2

Scenario 3

Cane agronomy 2757 2343 2079 0.03 0.03 0.02 -97,686 -83,033 -73,650

Cane burning before

harvesting0 0 0 0 0 0 40,738 34,627 30,714

Cane transportatio

n2780 2363 2096 0.23 0.20 0.17 493 419 372

Cane milling 3257 2023 0 38 32 29 10,0015 85,012 0Juice

clarification 10,546 8964 89 13 11 10 0 0 0

Fermentation unit 5259 4470 159 14 12 10 8676 7375 6541

Distillation unit 14,861 14,861 0 8 8 7 352 352 309

Dehydration unit

molecular sieve

2934 2934 0 8 8 7 0 0 0

Power house 0 0 95 0 0 38 0 0 25,038Consumption 730 730 730 0.04 0 0.04 17,956 17,956 17,956

Total (MJ/tonne ethanol)

43,123 38,688 5248 81 71 101 70,544 62,709 7280

Scenario 1 (No recycling flows)

Scenario 2 (Intra-recycling flows)

Scenario 3 (Exchange & recycle flows)

0

20000

40000

60000

80000

100000

120000Environmental remediation resources

Capital resources

Operating resources

Scenario

Resource consump-tion/MJ/

tonne ethanol

Figure 4-14: Overall resource consumption for the three scenarios excluding Type-II flows

57

The resource consumption from environmental remediation is relatively high in both

scenarios 1 and 2 as bagasse is considered as a waste and is simply disposed of. It was

determined that it would incur about 1.00×105 MJ/tonne ethanol of cumulative exergy

consumption to treat the methane emissions generated from decomposition of bagasse. Using

global warming impact factors, the CO2 equivalent emission was determined and it was

assumed that such amount of CO2 is absorbed using photosynthesis. In scenario 3, the use of

bagasse for energy production led to a 90% decrease in environmental remediation resource

consumption as compared to scenario 1. This is because carbon dioxide released during

bagasse burning requires 25 times less resources for environmental remediation than methane

emissions. The operating resource consumption in scenario 3 has decreased by 90% as

compared to scenario 1. This is mainly because the total resource consumption for ethanol

production is now being shared with bagasse with an allocation factor of 0.507 based on

exergy content. Besides, the energy requirements of the ethanol plant, which contribute to

most of its operating resource consumption, are now being fully satisfied by steam and

electricity from the combustion of the bagasse exchange flow in the power station. In

addition, the surplus electricity generated from bagasse burning in the power station has an

overall positive impact on resource consumption for ethanol production. The surplus

electricity is a valuable resource that can be exported to the grid and hence the resource

consumption for ethanol produced from the ethanol plant is now shared with the surplus

electricity with an exergy based allocation factor of 0.887.

Capital resources have negligible contribution to the cumulative exergy consumption for the

three scenarios. The capital burdens of plant equipment, machineries and transportation

vehicles have been estimated by spreading the total cumulative exergy consumption for their

capital resources to unit output of their output. The high usage factor of plant equipment and

an assumed operational life span of 15 years for the agricultural machinery, plant equipment

and transportation vehicles lead to low capital costs compared to the operating and

environmental remediation costs. Scenario 3 has slightly higher capital resource cost than the

other two scenarios especially due to capital resource consumption associated with the power

station. The capital burdens of plant equipment, machineries and transportation vehicles have

been estimated by spreading the total cumulative exergy consumption for their capital

resources to unit output. The high usage factor of plant equipment and an assumed

operational life span of 15 years for the agricultural machinery, plant equipment and

58

transportation vehicles have led to negligibly low capital resource consumption compared to

the operating and environmental remediation resource consumption.

4.7 Summary of Part I: a coherent multi-level framework for resource

accounting

Part I of the thesis presented a comprehensive framework for assessing resource consumption

with the following key elements:

(1) At the starting point, a conceptual framework introduces fundamental concepts such

as system, flow, process and environment. Various types of resource flows, such as

material, energy and human labour are considered. Resource-generating processes are

distinguished into Type-I and Type-II processes, to allow appropriate resource

accounting principles to be applied to flows originated from these processes. Resource

consumption by processes for both production and environmental remediation is taken

into account.

(2) With the basic concepts introduced above, a multi-level structure is presented for

resource accounting at different technical levels, which include unit, process, inter-

process, and production-consumption, with intra-level and inter-level connections

specified.

(3) Finally, a resource accounting algebra is formulated according to the multi-level

structure, offering key equations for quantitatively assessing different design options

at different levels. It also proposes quantities to support the evaluation of important

resource decisions such as intra-level recycling, inter-level exchange, and repair and

recycling of used products.

By using the proposed multilevel resource accounting framework to a case study on ethanol

production and consumption from sugarcane, it showed what scenarios can be the best

solution and to help assess the full effects on resource efficiency of design decisions at all

levels, allowing exploration to find the most resource efficient option. It also offers the

potential to identify key components and flows that can be either removed or improved

through integration and linkage with other flows or components in the system. In particular,

the framework demonstrated the effects of design decisions at the various levels, such as

choosing between molecular sieve and azeotropic distillation at the unit level, adoption of

water recycling at the process level, and bagasse exchange flows at inter-process level. In

59

addition, by explicitly accounting for all types of resources including operating and capital

resources, environmental remedial resources, labour and ecological goods and services, this

method could be used as an insightful tool for evaluating resource utilisation and

consumption.

Overall, while sharing some common principles with early industrial ecology approaches and

accounting (Chertow and Ehrenfeld, 2012) (the framework considers not only the industrial

processes but also the ecological processes that occur in the environment of the industrial

operations) and exergy-based approaches such as CEC (Szargut et al., 1988), ICEC (Ukidwe

and Bakshi, 2007; Zhang et al., 2010), EEA (Sciubba, 2001), ECEC (Hau and Bakshi, 2004),

CEENE (Dewulf et al. (2007), the holistic and coherent systematic approach has given

insights into how changes in resource consumption occur at different levels. By revealing the

resource consumption through each system layer, the framework provides a robust and

transparent way to capture effects of decision making during design or retrofitting of

processes in order to find the most resource efficient design options. The framework could

also be applied to support further environmental research and (for instance, it could be

combined with other approaches such as LCA so that he wider environmental implications of

a system can be assessed) in other areas such as those addressing the social, cultural and

business perspectives of resource management. The next steps in the research work, tackled

in part II of the thesis, involve developing a method for optimal design of local production

systems where resource consumption is used as an objective function to be optimised and as

an indicator to guide the design or retrofit.

60

PART II: Design approach for integrated local production systems

61

Chapter 5: Systematic approach for designing locally integrated

production systems based on mathematical programming

5.1 Rationales for shifting to localisation

With the advent of industrialisation, the supply of energy and materials to meet human needs

has been driven primarily by centralised production, harnessing economies of scale, based on

fossil fuels and large scale distribution infrastructures. However, continuation of this mode of

production coupled with growing population has led to a range of issues such as climate

change, energy supply insecurity, deterioration of ecosystems and depletion of resources.

Local production systems have been regarded as one possible pathway towards sustainability

(Royal Academy of Engineering, 2011). Though the challenges are global, they have local

impacts and may affect each local system differently. This calls for the engineering of

human-made systems with a focus on the rational use of locally available resources. Such

systems require new design tools to allow decision makers to explore the roles of local details

such as the significance of local resource use and the opportunities for interactions between

co-located subsystems.

A local production system or Locally Integrated Production System (LIPS) as it will be

mainly referred to onward, considers all types of production processes that can occur at a

local scale for the production of products (e.g. food) or services (e.g. heat) to satisfy local

demands such as food, energy, water and material demands that are required to meet local

needs (e.g. nutrition, sanitation, thermal comfort, mobility and housing). While these

processes differ in technical natures, they share the following characteristics desirable from

sustainability perspectives; it is precisely this set of common characteristics that is to be

explored by this work. First of all, these systems offer the possibility to use renewable

resources which can be captured or produced locally to meet demands of the local population

(see Figure 1-1). They also have the advantage of avoiding large transportation distances and

the resulting impact on energy consumption and the environment. Furthermore, a localised

paradigm allows the processes and technologies to be developed or adapted according to

local conditions. More importantly, the main opportunity that arises from LIPS is the

potential for symbiotic integration of multiple and distinct subsystems (e.g. water, energy,

food, and ‘wastes’ arising from their supply and use) within the same locality in order to

increase efficiency and sustainability.

62

The aim of the work presented in this chapter is to propose a systematic approach to the

design of local production system based primarily on mathematical programming.

Mathematical programming is an established approach used in process integration for better

utilisation and savings regarding energy, water and other resources; which thus suit the

design problem formulated in this chapter. It has been used broadly for the design of mass

exchanger networks (El-Hawagi and Manousiouthakis, 1990), combined mass and heat

exchanger networks (Srinivas and El-Halwagi, 1994), the integration of batch chemical

processes (Smith, 2005) and more recently, the design of heat exchanger networks using a

two-step optimisation procedure incorporating detailed exchanger design (Short et al., 2016).

Mathematical programming approach allows different synergies between different processes

to be explored so as to generate a truly integrated design. This advantage has been exploited

for the integrated design of local renewable resources for energy supply (Kostevsek et al.,

2015) and the simultaneous design of energy and water networks within the same production

system (Martin and Grossmann, 2015) but not for the integration between multiple

production systems (e.g. food production, water treatment, energy production). The

conventional way of designing production systems rarely explores the potential for

integration with other production systems to satisfy local demands in the most sustainable

manner, but this could be addressed by local production systems. On the other hand, insight-

based approaches (Foo, 2007) which include techniques such as pinch analysis for heat

integration (Linnhoff and Hindmarsh, 1983) and mass integration (El-Halwagi and

Manousiouthakis, 1989) have also been proved useful for integration of production processes

and will be explored further in the next Chapter for solving LIPS. These methods have also

been combined with mathematical programming into hybrid methods (Luo et al., 2009).

More recent developments on insight-based and mathematical programming have been

reviewed by Foo et al. (2012) who compiled process design and optimisation techniques

recently developed for improving sustainability of industrial processes. Klemes et al. (2013)

have also given an overview of achievements and future challenges in process integration

while the review by Foo and Tan (2015) emphasized on approaches for the reduction of

carbon emissions and environmental footprint.

Localised production is closely related to Eco-Industrial Parks (EIPs) under the broad concept

of Industrial Symbiosis (IS) which advocates the leveraging of the synergies between

geographically co-located industrial processes (Chertow and Ehrenfeld, 2012). Mathematical

optimisation methods have proved useful for designing EIPs (Boix et al., 2015). More

63

specifically, an optimisation approach has been formulated for the cost effective design of

water and wastewater treatment among industries in an EIP (Lovelady and El-Halwagi,

2009), on the maximisation of economic performance for the design of bioenergy-based

industrial symbiosis system (Ng et al., 2014), the optimisation of material flow by-products

as feedstock for other industrial processes (Cimren et al., 2011) and more recently on the

fuzzy optimisation of waste-to-energy among several plants contained in an EIP (Takshiri et

al., 2015) and the design of a cost efficient renewable electricity integration at the regional

level through a mixed integer linear programming optimisation model (Dominguez-Ramos et

al, 2016). On the other hand, some other studies have focused on regional supply chain

optimisation such as the optimisation of a regional renewable energy supply chain from

biorefinery operations (Lam et al., 2010b), cellulosic ethanol supply chain optimisation at the

county level (You et al., 2011) and the optimisation of water supply chain across different

regions (Aviso et al., 2011). Local integrated production systems, as explored in this work,

clearly should share the beneficial features of EIPs or IS systems in general, such as the

exchange of wastes, energy and water between industrial processes. However, the concept of

local production system has a distinctive emphasis on local resources and demands (see

Figure 1-1) and on the holistic consideration of all types of agricultural, industrial and

municipal processes to take place at the locality of concern. In contrast, the work on IS

systems including EIPs generally does not have a focus on the "local" dimension in terms of

using locally available resources that occurs naturally (e.g. sunlight, wind, biomass) and local

societal demands such as basic local demands in food, energy and water and rarely considers

agricultural processes with the focus being more on industrial processes and no synergies

considered between designing agricultural and industrial processes. The design of regional

supply chains, on the other hand, often addresses one or very a few specific products such as

energy and fuels, not aiming to explore the synergies in meeting other regional demands.

Therefore, while sharing commonalities with existing work on eco-industrial parks and

regional supply chains in terms of pursuing higher resource efficiencies through optimisation

and integration, the design of local production systems as studied in this work represents a

rather different decision problem in terms of aim and scope that are meant to support the

expected economic paradigm shift towards localised production, with the potential benefits of

a sustainable development path as highlighted earlier. Table 5-1 summarises the key features

IS and EIPS, regional supply chain and LIPS.

64

Table 5-1: Key features of IS and EIPS, regional supply chain and LIPS

Features IS& EIP Regional Supply Chain LIPS

Aim at higher resource efficiencies through

optimisation and integration

Synergies between geographically co-located industrial

processes (i.e. exchange of wastes,

energy and water between industrial

processes)

×

Local focus and dimension (i.e. distinct

emphasis on local resources and

demands)

× ×

Local heterogeneous processes (i.e. holistic

consideration of all types of agricultural,

industrial and municipal processes to

take place at the locality of concern)

× ×

LIPS will comprise a non-linear structure, i.e. with waste and by-products looped back into

the system and synergies exploited, which is in comparison to the traditional linear

production system where most often there is no recycle and exchange of flows between the

different units of the system, and will require the design of the system and its components to

be highly tuned to the local settings. In this chapter, a systematic mathematical programming

approach published in Leung Pah Hang et al. (2016b), is proposed for designing local

production systems that, given a set of locally available resources, selects and integrates a

combination of production or treatment processes to meet given local population demands. It

adopts a life cycle approach accounting for resource consumption using cumulative exergy

consumption as an indicator of resource intensity for the imported flows as well as for capital

resources and environmental remediation efforts. Furthermore, this is the first time that such

a systematic approach is applied for designing the food-energy-water nexus at the local

scale. To-date, a range of methods and tools has been developed for investigating the

interconnected food, energy and water systems, including those for modelling and assessment

65

(Foo, 2007, 2013; Klemes et al., 2013) and those for optimal design and planning (Linnhoff,

1993; Wang and Smith, 1994; Nelson and Liu, 2008). Most of the existing work however

addresses larger (e.g. national and regional) scales.

5.1.1Design problem statement and quantification of resource consumption

The design of local production systems considers the production of multiple products and

services to satisfy local demands within the capabilities of the local environment and

ecosystems (e.g. groundwater abstraction limit). Due to the different nature of the resources

used in a system that integrates heterogeneous components, it is desirable to adopt a unifying

quantity such as exergy (Sciubba and Wall, 2007); defined as the available energy of a

resource to do useful work. In this work, Cumulative Exergy Consumption (CExC) will be

used, which is an approach also applied in other contexts (Allwood et al., 2011). CExC in

delivering a service is the sum of the exergy of all types of resources required from extraction

to the point where they are used. The problem of designing local production systems can be

generally stated as:

Given a set of demands (e.g. food, energy and water) by the population in a locality and the

availability of local and external resources, determine the combination of a set of processes

and activities which can meet such demands so that the total cumulative exergy consumption

is minimised while satisfying all necessary constraints.

LIPS will be designed with a strong focus on using locally available resources, yet with the

recognition that not all resources can or should necessarily be provided locally, also

considering the possibility of having production surpluses for export and discharges to the

environment. Therefore, the designs are expected to generally result in a mixture of local,

imported and exported resource flows that allow satisfying local demands in a resource

efficient manner.

Following the above principle, the design objective (for minimisation) can be stated as the

sum of the CExC of every flow that goes into (i) the local production system and (ii) the

technological or environmental processes required for treating the effluents of (i) to the extent

that, in principle, no harm is made to the environment, or, in practical terms, a certain set of

environmental regulations are met. When the production system exports a valuable product,

66

its resource content, valued by the average CExC of the product of the same nature as

available in the external market, is treated as resource “credit” of the system. This credit is

deducted from the total resource consumption by the local production system, leaving the

design objective as to minimise the net resource consumption for meeting the local demand.

When quantifying the CExC of flows, two different types of processes from which the flows

are originated have been distinguished (Leung Pah Hang et al., 2016a) to avoid unnecessary

complexity while maintaining consistency. In order to facilitate subsequent discussions in this

Chapter, flow-generating processes have been classified into two types. Type-I processes are

defined as those that can be affected by human decisions, while Type-II processes are those

that typically are not under human control. These flows classifications have also been

presented earlier in Chapter 2 of the thesis. Flows from Type-I processes (e.g. grid electricity)

would be accounted by their full CExC while flows from Type-II processes (e.g. wind,

sunlight and ores) by their exergy content fully defined by their physical nature and any

further exergy consumption for their extraction and processing. The resource value of flows

from Type-II processes should be considered when there is a need for recognising that these

resources have alternative competing uses. However, the full CExC for the formation in the

natural environment of these flows will not be taken into consideration as they occur

independently from human intervention.

The scope of the proposed method is to optimise resource consumption from a technical

perspective; cumulative exergy consumption is used because exergy is a unifying quantity

that can represent material, energy and non-energetic streams. Nevertheless, the modelling

approach to capturing the interconnections between different processes and subsystems while

taking into account local resources and demands may be applied with other objective

functions and constraints pertaining to economic costs, social benefits, regulatory

considerations and broader environmental impacts.

5.1.2 Overview of the proposed approach

Figure 5-1 depicts the steps in the proposed methodological framework for the design of

locally integrated production systems (LIPS) based on mathematical programming. Given a

set of demands (e.g. food, energy and water) by the population in a locality and the

availability of local and external resources, the first step is to construct conceptual

superstructures (section 5.1.3), to identify possible processes and flows to introduce within

individual subsystems and the possible exchanges between these subsystems. Mathematical

models are then constructed (section 5.1.4) according to the conceptual superstructures with

67

the aim of minimising cumulative exergy resource consumption. With the models of

individual subsystems, an optional preliminary design analysis (section 5.1.5) can be carried

out, if it is desirable by the decision-makers to gain an initial understanding of the design

alternatives. The last step of the approach (section 5.1.6) solves the mathematical model of

the entire system to achieve integrated optimal design.

Figure 5-15: Methodological framework for designing LIPS

As stated above, the design procedure starts with a pre-defined geographical scope, i.e. that of

the locale of concern. This will in turn determine (i) the population for which the demands to

be met and (ii) the natural resources to be tapped in, thus defining the system boundary. In

practice, the scope of the targeted locale will depend on the intention of the decision makers

or their perception of the feasibility for implementing an integrated design. This may settle to

an area under the direct governance of a local or regional planning body or one under direct

influence of a community group, e.g. a village, a town, or a county in the UK context. As the

local scale offers geographical proximity between different processes, optimisation can not

only consider general inter-dependencies between the food, energy and water systems, as is

68

Construct conceptual

superstructures

Optimal design of localised production system

Simultaneously optimise all the

subsystems in one mathematical model

No

 Insights on individual sub-systems and their interactions

Yes

Solve the model of each subsystem separately and perform scenario based

analysis on single subsystems

Preliminary design to be performed?

Construct mathematical design

models

typically done by work on larger scales, but also explore symbiotic resource (e.g. heat,

organic waste) reuse opportunities between specific facilities that can realistically be

physically connected only at the local scale. While the proposed approach does not determine

the system boundary, the decision makers could apply the approach to alternative scopes to

assess the impact, leading to an optimal scope for designing and eventually implementing a

local integrated production system. Moreover, the optimisation problem with the conceptual

superstructure and mathematical models will need to be reformulated depending on the key

characteristics and strategic priorities of the locale under consideration. For instance, if one

was to design a nexus consisting of agriculture-energy-water to align with the strategic design

objectives and availability of agricultural land of the locale under consideration, a wider

range of agricultural activities might to be considered to meet not only the local food

demands of the local population but also the external food demands; resulting in more

interactions with other localities.

5.1.3 Conceptual construction of superstructures

Superstructures are used to represent design options by means of sources, sinks and their

connections. Based on the generally accepted definition of source and sink in process

integration (El-Halwagi, 2011), in this work, a source refers to a material or energy flow,

while sinks are defined as those components of the system that can receive flows, which

either process them to generate new flows or act as terminating points for flows (e.g.

consumption by local population or discharge into the environment). As presented before, a

local production system will be made up of various interconnected subsystems. These

subsystems can be for example the food production subsystem, the energy subsystem, the

water subsystem and so on. Each of these subsystems contains production or treatment

processes. The design of superstructures involves the following steps:

(1) Identify all subsystems in the local area, to ensure that there is scope for exchange and

integration between different subsystems.

(2) Determine the possible sources and sinks based on the availability of resources and

demands to be satisfied in each subsystem. Figure 5-2(a), illustrating the

superstructure of sources and sinks in a single subsystem, shows that a source (i) can

be can be an external incoming flow (e.g. i=1), an internal resource flow from the

local environment (e.g. i=2), a discharge flow to the environment (e.g. i=6), a flow

exchanged between two internal processes (e.g. i=3, 4), or an export flow of surplus

product to external systems (i.e. i = 5). On the other hand, a sink (j) can represent a

process (e.g. j=1, 2), local consumption (e.g. j=5), the local environment as the

69

destination of discharge (j=4) or an external system as the destination of export (e.g.

j=3). Note that the processes are those taking place within the system boundary of the

local system leading to products that can satisfy local demands.

(3) Establish integration opportunities within and between subsystems by means of

exchanges of various sources between different sinks, as illustrated in Figures 5-2(a)

and 5-2(b). Figure 5-2(b) shows the potential sources and sinks in a system

comprising two subsystems connected together. It particularly illustrates how

exchanged streams between the subsystems become potential sources for subsystem A

(e.g. i=5, 9) and subsystem B (i.e. i’= 2, 3). Consider specifically options for recycling

and exchange of locally available flows based on their content (e.g. agricultural

residues can be used as energy feedstock due to high calorific value or as feedstock

for livestock due to its nutritional content).

Figure 5-2(a): Illustrative superstructure of a sub system

70

External system (j =3)

i=7, Local resources

Local environment (j=4)

i =5, Export

i=4, Exchanged resources

i=3,Exchanged resources

i=2, Local resources

i: source j: sink

j=5Local

consumption

i=6, Discharge

j=2

Internal process

 

j=1Internal process

i=8 Final products

i=1, Imported resource

Figure 5-2(b): Generic superstructure representation of combined systems

The conceptual construction of superstructures is exemplified further for the food-energy-

water nexus in Section 5.3.

5.1.4 Constructing the mathematical optimisation model for each subsystem

An optimisation model is formulated for each subsystem based on their superstructure and

should consist of:

1) An objective function that minimises the net resource consumption for each

subsystem.

2) A set of equations that describe the technical constraints of the processing units and

the interconnections between various sinks as expressed in the superstructure.

3) Ecological constraints that limit the use of locally available resources and the

discharge of waste streams to the environment within the ecosystem’s capacity for

resource re-generation and waste assimilation.

4) Time slice/period and storage to provide a suitable treatment of the temporal

variations in supply and demand. Local storage may be required to reconcile the

temporal mismatch between varying local supply and demand, affected by the

connectedness of the locale with other local systems or a central system for

distribution. The temporal variations within the system need to be properly handled

71

i’ =8, Export External system (j’=3)

i’=6, Exchanged resources

i’=3, Exchanged resources

i=7, Local resources

i=4, Exchanged resources

i=3, Exchanged resources

i’=7,Final

products

i’=1, Imported resource

Subsystem A

j’=2, Internal

process j’=1, Internal process

j=1

Internal process

 

Subsystem B

i’=5, Local resources

j’=5Local

consumption

i=6, Discharge i=2, Local resources

i=8, Final products

i=1, Imported resource

i=5, Exchanged resources

i’=4, Discharge

i’=3, Exchanged resources

i’: source for subsystem B j’: sink for subsystem B 

i: source for subsystem A j: sink for subsystem A  

i’=2, Exchanged resources

i=9, Exchanged resources

j=4Local

consumption

Local environment (j=3)

j=2

Internal process

 

Local environment

(j’=4)

by the mathematical model, possibly through time slicing/period of time (Becker and

Marechal, 2011) to allow variations to occur between different time slices. The size of

time slice, which represents the model’s temporal resolution, should match with the

intended use of storage and may vary between as short as hourly (e.g. storing wind or

solar electricity for a local system not connected to a central grid) and as long as

seasonally (e.g. rain water collection where supply and demand variations are largely

seasonal). When different sets of supply and demand have diverse time

characteristics, the finest size of time slice will be adopted for the whole system.

The construction of the optimisation model is illustrated in section 5.4.

5.1.5 Preliminary design analysis

Once the mathematical models for individual subsystems have been built, a preliminary

design analysis could be carried out if it is desirable to gain an initial understanding about

possible designs. This could also be useful when dealing with existing infrastructure and the

design is more for retrofitting purposes, or when systems are implemented separately in

stages with a view to develop systems integration in the future. Any of these cases may

benefit from an incremental understanding of the improvement potential. The preliminary

design analysis broadly includes using the separate optimisation models for individual

subsystems and performing scenario analysis. The specific stages in the preliminary design

analysis could include:

(1) Use conventional input flows such as grid electricity and natural gas derived heat in

the initial design of a subsystem and report its objective function and decision

variables.

(2) Vary the source of input flows (e.g. energy supply from biomass CHP instead of grid

electricity) to the subsystem and analyse the impacts on its objective function and

decision variables.

(3) Design the other subsystems by using the same source of input flows as in the

previously designed subsystem where relevant. If the other subsystems have a logical

connection with the previous subsystem, consider satisfying the new demands

resulting from the previous subsystem in the design of the other subsystems. For

instance, if the food subsystem is designed first, the water and energy demands of the

food processes need to be considered in the subsequent design of the water and energy

subsystems respectively, in addition to other local water and energy demands.

72

(4) Perform a scenario analysis by repeating (2) and (3) with different sets of input, and

analyse the outcome of these scenarios.

Such preliminary analyses could help to understand the interactions between the various

components of a local production system and how these might affect the overall resource

consumption. In particular, the results obtained from the preliminary design analysis may

provide insights into the balance of exchange flows between individual subsystems and the

trade-offs between using imported conventional resources and locally available resources.

5.1.6 Constructing and solving a simultaneous design model (solving the mathematical

models for each subsystem all at once)

To eventually identify the optimal, integrated design, all the superstructures of the

subsystems are combined into a holistic superstructure by considering the integration

opportunities. Based on the combined superstructure, illustrated earlier in Figure 5-2(b), one

mathematical optimisation model is then formulated and solved. The elements of the

integrated system model are similar to those of individual subsystems as presented in section

5.1.4. However, the objective function now is to minimise total net resource consumption and

does not include any intermediate flows (internally exchanged) between subsystems. Besides,

the quantity of demand of each subsystem for the other subsystems and the characteristics of

the supply from one subsystem to the others become unknown and will be determined via

optimisation.

The simultaneous approach will generate the optimal design of a local integrated production

system as a whole. In comparison to the preliminary analysis, it considers all design

integration options simultaneously across all subsystems. This approach is essential for

revealing the benefits of a local integrated production system on resource efficiency and

circularity as compared to the practice of designing distinct subsystems in silos. Since the

simultaneous approach solves the mathematical models for all the subsystems pertaining to

the local integrated production system at once in one mathematical model, it involves more

complex and large mathematical models. However, it can solve complex design problems

more accurately and produce design results faster as it does not involve any iteration as

compared to the preliminary design analysis.

73

5.2 Building design models for food-energy-water nexus

The methodology for constructing the design models is illustrated by the food-energy-water

nexus. The nexus concept has been broadly used to identify the issues arising from the

interconnectedness between food, energy and water subsystems. In addition, the nexus has

been used as a framing for systems analysis albeit mostly at the regional, national and global

scales (FAO, 2014). The importance of this nexus has been clearly recognised for sustainable

development and national security in the UN (UN WATER, 2104), FAO, governments and

organisations around the world (NEXUS, 2015). Though these three sectors are inextricably

linked, as actions in one sector have impacts in one or both of the others, these areas have too

often been considered in isolation (FAO, 2014). Thus, there is a clear need to look at them

holistically and develop tools that consider their interdependencies (Machell et al., 2015).

This opens up opportunities for Process System Engineering (PSE) research (Garcia and You,

2016). PSE deals with the design, operation, optimisation and control of processes through

computer-aided approaches and one of the challenges in PSE research is to develop concepts,

methodologies and models for decision-making for a design system. While existing tools

such as modelling and Life Cycle Assessment have been applied to study various parts of the

nexus, new frameworks are required to analyse complex relationships embedded in such

systems (Keairns et al., 2016), using a systematic approach to address the resulting challenges

associated with risks from a supply chain perspective (Irabien and Darton, 2015) and

highlighting the role of systems integration (Wolfe et al., 2016). In our work, this is the first

time that such a systematic view is applied for designing a food-energy-water nexus at the

local scale with many nexus related studies to-date focusing on the larger scales (Foo, 2007,

2013; Klemes et al., 2013; Linnhoff, 1993; Wang and Smith, 1994; Nelson and Liu, 2008).

Therefore, nexus in this work refers to a system that takes advantage of opportunities for

synergy and integration arising from closely connected and geographically co-located

subsystems for food, energy and water production.

5.2.1 Building superstructures

The models presented in this section have assumed specific food types and energy generation

technologies, originated from a case study to be presented in Section 5.5, to aid the

understanding of the conceptual superstructure. With some adaptions, the mathematical

models can be applied to cope with arbitrary system components (see section 5.1.2).

74

5.2.2 Superstructure for food production subsystem

Figure 5-3: Superstructure for food production subsystem

The superstructure for the food production subsystem is shown in Figure 5-3. The food types

chosen as examples are bread, potatoes, beef and pork and have been selected based on local

food preferences in Whitehill and Bordon eco-town. The annual consumption by the local

population is given in Table 5-2 in section 5.5 and was determined based on the average daily

consumption of these food types according to DEFRA (2014). The sources include imported

fertilisers, animal feed and locally produced nutrient flows from manure or crop residues

while internal processes (as sinks) are those for bread, potatoes and beef and pork production.

5.2.3 Superstructure for water production subsystem Figure 5-4 illustrates the superstructure for water production subsystem. The possible water

sources include water of varying quality (e.g. COD from food and energy production

subsystems, treated residential wastewater, groundwater and rainwater. The sinks considered

are the food and energy subsystems and the residential sector. Water treatment operations (as

“intermediate” sinks) were included and acted as regeneration processes before water sources

could be made available for use in the “final” sinks.

75

Imported fertiliser

External systemImported animal

feed

Imported fertiliser

Residues

Manure

Manure

Residues

Residues Manure

Imported animal feed

PotatoesPork

BeefBread

ResiduesManure

Potato consumption

Potato cultivation and processing

Cattle breeding and processing

Beef consumption

Pork consumption

Pig breeding and processing

Wheat cultivation and processing

Bread consumption

Energy sinks

Energy sources

Figure 5-4: Superstructure for water production subsystem

5.2.4 Superstructure for energy production subsystem

Figure 5-5: Superstructure for electricity production

76

Grid

Local environment

Wind powergeneration Solar power

generation

Surplus power

Power

Export sink

Combined heat and power

Natural gas

Biomass wood

External system

Wind Solar

Water subsystem Residential Food subsystem

Organic waste

Energy subsystem

Energy wastewater

 Energy wastewater

treatment

Treated water

Treated water

Water sources

Water regeneration processes

Food processing wastewater

Water subsystem

Discharge

 

Residential wastewater

Local environment

Water sinks

Residential wastewatertreatment

Rainwater  

Groundwater

Rainwater treatment

Groundwater treatment

Discharge Residential Cattle and pig breeding and processing

Wheat and potato

cultivation and processing

Energy

production

Food wastewatertreatment

External systemImported water

Figure 5-6: Superstructure for heat production

Figures 5-5 and 5-6 represent respectively the electricity and heat energy sources and sinks

considered for the design of the energy subsystem. The energy sources are grid electricity,

electricity from wind and solar sources, heat from natural gas boilers, and heat and power

from Combined Heat and Power (CHP) based on biomass, organic waste or natural gas.

Waste heat sources considered were Low Temperature (LT) waste heat available from all

CHPs (apart from the main heat flows produced) and food production processes which will

be lost if not recovered. The sinks were food and water production processes and the

residential sector and the possibility for export of electricity to the grid.

5.2.5 Superstructure for simultaneous food, energy and water design

Figure 5-7 shows the superstructure comprising representative sources and sinks for all the

three subsystems. The important exchange of flows between the three subsystems include

energy flows from the energy subsystem to the food and water subsystems and water flows

from the water subsystem to the food and energy subsystems. Wastewater generated from the

food and energy subsystems could be treated in the water subsystem and organic waste from

food subsystem could be used as a potential energy source for the energy subsystem.

77

Main heat 

Local environment

External system Natural gas

Natural gas boilers

LT Waste heat 

Organic waste

Organic waste

Combined heat and power

Pork production

Beef production

Water subsystem

Residential

 

Bread production

Biomass wood

Energy sinks

Figure 5-7: Superstructure for integrated food, energy and water system

5.3 Mathematical formulation for the preliminary design analysis

In this section the model formulation for a preliminary design analysis is presented. The time

period of one year is chosen as basis for the model. Note that though 1 year has been chosen

for illustration only but in practical, one design must be arrived by considering time

variations of parameters over a much longer period of time. The local system studied is well

connected to the grid, so energy storage was not considered, and thus a seasonal time slice/

[period that can match seasonal storage for agricultural crops and rainwater collection has

been adopted for the design. The starting season for design was taken to be summer. The sets,

parameters and variables used in the mathematical models are given below:

Sets

a A Water sources

ag AG Agricultural commodities

b B Water sinks

78

External system

Surplus electricity 

Organicwaste  

Water

Energy 

Water Energy 

Export (Grid)

Natural gas

Biomass wood Boilers

Water

Wastewater

Wastewater from energy processes

Wastewater from food processes

Rainwater and groundwater

Water regeneration processes

 

Local environment

Local foodconsumption

Imported animal feed

Pork Beef

Potatoes Bread

Imported fertiliser

Manure

Residues Pig and cattle breeding and processing

 

Residential  

Combined heat and power

 

Discharge

Wheat and potatoes

cultivation and processing

 

Energy subsystem

Water subsystem

Food subsystem

b’ B’ Regenerator water sinks

c C Crops

d D Food types

i I Nutrient sources

i’ I’ Imported nutrient sources

i’’ I’’ Locally produced nutrient sources

j J Food sinks

l L Livestock

o O Operating flows

r R Energy raw material

s S Seasons

x X Energy sources

y Y Energy sinks

Parameters

cbc ,d Conversion factor from crop c to food d

cf l ,d Conversion factor from livestock l to food d

codb Maximum allowable COD of water sink b, g COD/kg

eo , j Specific cumulative exergy of operating flows o to nutrient sink j, MJ/kg or MJ/MJ

er Specific cumulative exergy of raw material r for energy production, MJ/kg

ecw Specific cumulative exergy of chemicals per unit wastewater, MJ/kg

eelw Specific cumulative exergy of electricity per unit wastewater, MJ/kg

ehew Specific cumulative exergy of heat per unit wastewater, MJ/kg

e ie Specific cumulative exergy of imported energy, MJ/MJ

e iel Specific cumulative exergy of total imported flows for producing electricity, MJ/kg

e ihe Specific cumulative exergy of total imported flows for producing heat, MJ/kg

exel Specific cumulative exergy for producing electricity from source x, MJ/kg

exhe Specific cumulative exergy for producing heat from source x, MJ/kg

edimp Specific cumulative exergy of imported food d, MJ/kg

e i ’ , jimp Specific cumulative exergy of imported nutrient flows i’ to nutrient sink j, MJ/kg

E y, sdem Electricity demand at sink y per season s, GJ

ELDd Electricity demand per unit food d, MJ/kg

Fd , sdem Demand of food d in season s, t

FC Nominal size of storage facility, t

79

H i ' ' Harvest recovery rate of locally produced nutrient sources i’’

H y , sdem Heat demand at sink y per season s, GJ

H Max Maximum heat load in waste heat, GJ

HEDd Heat demand per unit food d, MJ/kg

Lr , x Land use per unit raw material r from source x, ha/MJ

Lagri Total amount of agricultural land available, ha

Len Land available for energy production, ha

M r , sAv Availability of raw material r in season s, MJ

N j ,sdem Demand of nutrient sink j in season s, kg

nc i ' ' Nutrient content of locally produced nutrient sources i’’, kg N

yc , s Yield of crop c per season s

y l Yield of livestock l

RAag Amount of residues or manure per unit of agricultural commodity, kg/kg

Ref COD removal efficiency of treatment plant, %

RW s Amount of rainwater collected in season s, t

SEDWA Electricity demand for treating unit wastewater, MJ/kg

SHDWA Heat demand for treating unit wastewater, MJ/kg

SL Number of years of service life of storage facility, y

t Time period over which heat is transferred, y

T x '¿ Inlet temperature of heat source x’ before heat exchange, °C

T x 'out Outlet temperature of heat source x’ after heat exchange, °C

T y¿ Temperature of heat sink y before heat exchange, °C

T yout Temperature of heat sink y after heat exchange, °C

TD Minimum temperature difference, °C

TEc Specific cumulative exergy of operating resources per unit accumulated crop, MJ/kg

UTD Upper bound for temperature difference, °C

W b ,sdem Water demand of sink b in season s, t

WCd Amount of water required for agriculture per unit food d, kg/kg

WE Amount of water required per energy produced, kg/MJ

WEG Amount of wastewater generated per energy produced, kg/MJ

WGPd Amount of wastewater generated per unit food d, kg/kg

℘d Amount of water required for industrial processing per unit food d, kg/kg

80

ηx ,rel Electrical efficiency of source x for raw material r

ηx ,rhe Heat efficiency of source x for raw material r

Variables

Aag , s Amount of agricultural commodity ag produced during season s, t

Ac , s Amount of crop c locally produced in season s, t

ACc , s Amount of crop c accumulated at season s, t

ACc , s−1 Amount of crop c accumulated from season s-1, t

ARs−1 Amount of rainwater accumulated from season s-1, t

AW s Amount of rainwater available for consumption in season s, t

CAc Capital exergy resources for storage of crop c, GJ

CArw Total capital exergy resources for rainwater storage, GJ

codb ' , s COD of treated wastewater from treatment plant sink b’ in season s, g COD/kg

CPx ' ,s Heat capacity flow rate of source x’ for season s, GJ/season

CS y , s Heat capacity flow rate of sink y for season s, GJ/season

E x ,grid , s Amount of electricity from source x exported to grid in season s, GJ

E x , y , s Amount of electricity from source x to sink y in season s, GJ

ELDsFD Total electricity demand of food processes in season s, GJ

ELDsWA Total electricity demand of water processes in season s, GJ

Fd , scrop Amount of locally produced food d from crop in season s, t

Fd , simp Amount of imported food d in season s, t

Fd , slive Amount of locally produced food d from livestock in season s, t

Fd, slocal Amount of locally produced food d in season s, t

H x, y , s Amount of heat from source x to sink y in season s, GJ

H x' , y, s Amount of heat exchanged between waste heat source x’ and sink y in season s, GJ

HED sFD Total heat demand of food processes in season s, GJ

HED sWA Total heat demand of water processes in season s, GJ

Lc Land use for production of crop c, ha/t

Ll Land use for production of livestock l, ha/t

M r , s Amount of raw material r in season s, t

N i ’, j , simp Amount of imported nutrient flows i’ to nutrient sink j in season s, t

N i ' ' , j , slocal Amount of locally produced nutrient j from source i’’ in season s, t

81

OPc ,s Operating exergy resources for storage of crop c in season s, GJ

OSR Optimal size of the rainwater storage tank, t

Q x' , y, s Amount of heat energy from waste heat sourcex ' to sinks y in season s, GJ

TC Total capital resource exergy for storage facility, GJ

U o , j , si mp Amount of imported operating flows o to nutrient sink j in season s, t or GJ

W a ,s Total amount of water from source a in season s, t

W a ,b , s Amount of water from source a to sink b in season s, t

W a ,b ' ,s Amount of wastewater from water sources a to the treatment plant sink b’ in season s, t

W b ' , b ,s Amount of wastewater from the treatment plant b’ to water sinks b in season s, t

W rw ,b ,s Rainwater consumed by water sink b in season s, t

WCc ,s Amount of locally produced crop c that is used in the same season s, t

WST c Size of crop c storage facility, t

W sEN Total water requirement of energy processes in season s, t

W sFD Total water requirements of food processes in season s, t

WG sEN Total wastewater generation from energy processes in season s, t

WG sFD Total wastewater generation from food processes in season s, t

zx ' , y , s Binary variable for heat integration between waste heat source x’ to sink y in season s

5.3.1 Mathematical formulation of food production system

The optimisation problem for the design of a food production subsystem is to minimise the

total cumulative exergy consumption for meeting the local food demand. This objective

function comprises resource consumption associated with imported food, nutrients, operating

flows and crop storage, as formulated in Equation (5.1).

Minimise objective function:

Total cumulative exergy consumption for food production subsystem (TEF) = Total

cumulative exergy for imported food + Total cumulative exergy for imported nutrients +

Total cumulative exergy for operating flows + Total cumulative exergy for capital resources

for crop storage + Total cumulative exergy for operating resources for crop storage

TEF=∑s S∑d D

edimp Fd , s

imp+∑sS

∑j J∑i ’I ’

ei ’ , jimp N i ’, j , s

imp +∑s S

∑j J∑o O

eo, j U o , j ,simp +∑

c CCAc+∑

s S∑c C

OPc, s(5.1)

82

TEF is the total cumulative exergy consumption for the food production subsystem, edimp is the

specific cumulative exergy of imported food d, Fd , simp the amount of imported food d in season

s, e i ’, jimp the specific cumulative exergy of imported nutrient flow i’ to sink j, N i ’ , j , s

imp the amount

of imported nutrient flow i’ to sink j in season s, eu , j the specific cumulative exergy of

operating flow o to sink j,Uo , j , simp the amount of operating flow o to sink j in season s, CActhe

capital exergy resource for storage of crop c storage and OPc ,sthe operating exergy resource

for crop storage in season s.

The optimisation is subject to the following constraints:

1) Final food demand balance

The food demand balance for each season by the local population is given in Equation (5.2).

Locally produced food in a particular season (Fd , slocal) + Imported food in a particular season

(F ¿¿d , slocal)¿ = Demand of food in that particular season (F ¿¿d , sdem)¿

Fd , simp+Fd , s

local=Fd ,sdem∀ d D , s S (5.2)

Fd, slocal is the amount of locally produced food d in season s, Fd , s

imp the amount of imported food

d in season s and Fd , sdem the demand of food d in season s. Fd , s

local can be produced from either

livestock (e.g. it will come from livestock if beef is produced locally) or crop (e.g. it will

come from crop (i.e. wheat) if bread is produced locally).

Fd , slive is the amount of locally produced food d from livestock l and can be determined from

the yield of livestock,y l, the conversion factor cf l ,dfrom livestock l to food d, and the land use

for livestock production Llas given by Equation (5.3):

Locally produced food from livestock in a particular season (Fd , slive¿= Land use for livestock

production (Ll ¿ × yield of livestock (y l ¿ × conversion factor from livestock to food (cf l ,d)

Fd , slive=Ll y lcf l ,d ∀ l L , d D , s S (5.3)

The amount of crop c to be locally produced in any season, Ac , s, can be calculated as follows:

Locally produced crop in a particular season ( Ac , s,) = Land use for livestock production (Ll ¿

× crop yield in that particular season ( yc , s¿ A c, s =Lc y c, s ∀ cC , s S

with Lc being the land use for crop production and yc , s the crop yield per season s.

83

The amount of locally produced food product d from a particular crop during season s, Fd , scrop,

can be calculated through the amount of the locally produced crop c that is used in the same

season s,WCc ,s , and the conversion factor cbc ,d from crop to food product as shown:

Locally produced food from crop during any season (Fd , scrop¿ = Conversion from crop to food (

cbc ,d ¿ × locally produced crop in that particular season (WCc ,s ¿

Fd , scrop=cbc , dWC c, s∀ c C, sS ,d D

2) Land availability constraint

The land occupied by cropsLc and livestockLl must not exceed the total amount of

agricultural land available Lagrias given in Equation (5.4).

Total land occupied by crops (∑c C

Lc) + total land occupied by livestock (∑l L

Ll) ≤ Total

agricultural land available (Lagri ¿

∑c C

Lc+∑l L

Ll ≤ Lagri (5.4)

3) Nutrient requirement for crop and livestock

The sum of the imported and locally produced nutrient flows (denoted by i) should be equal

to the total nutrient demand of each sink j in each season s as shown in the nutrient balance

for crops and livestock in Equation (5.5).

Total imported nutrients to food sink j in a particular season (∑i ' I '

N i' , j , simp

) + Total locally

produced nutrients to food sink j in a particular season (∑i ' ' I ' '

N i ' ' , j , slocal

) = Total nutrient demand

for food sink j in that particular season¿¿)

∑i ' I '

N i' , j , simp + ∑

i ' ' I ' 'N i' ' , j ,s

local =N j , sdem∀ j J , s S (5.5)

with N i ' ' , j , slocal being the amount of locally produced nutrient from source i’’ (e.g. from crop

residues or manure) in season s and N j ,sdem the demand of sink j in season s. N i ' ' , j , s

local can be

determined through Equation (5.6).

84

Totally locally produced nutrient from source i’ ’in a particular season (∑j J

N i' ' , j ,slocal

) = Nutrient

content of locally produced nutrient (nc i ' ' ¿× locally produced agricultural commodity in a

particular season ( Aag , s) × locally produced nutrient per agricultural commodity (RAag) ×

harvest rate of locally produced nutrient in that particular season (H i ¿)

∑j J

N i' ' , j ,slocal ¿nc i} {A} rsub {ag,s} {RA} rsub {ag} {H} rsub {i ∀ i I , ag AG ,ℜℜ , s S(5.6)

with nc i ' ' being the nutrient content of locally produced nutrient i’’, Aag , s the amount of

agricultural commodity (i.e. crop or livestock) ag produced locally, RAag the ratio of amount

of residues or manure generated per unit output of ag and H i ' 'the harvest recovery rate of

locally produced nutrient i’’ taking into account that some residues need to be left in the field

to maintain the nutrient soil balance.

4) Crop storage considerations

The amount of the locally produced crop c that is used in season s, WCc ,s, should not exceed

the sum of the amount of crop c produced in that season, Ac , s ,and the amount of crop c

accumulated from previous seasons, ACs−1 as given in Equation (5.7).

Locally produced crop in a particular season (WCc , s) ≤ Amount of crop produced in that

particular season¿¿) + amount of crop accumulated from previous season ( AC¿¿ s−1)¿

WCc ,s ≤ A c ,s+ AC s−1∀ cC , s S (5.7)

The accumulation of crop c by the end of any season s, ACc , scan be determined by the

difference between the sum of the amount of crop c (i) produced in season s,Ac , s , (ii)

accumulated in the previous season,ACc , s−1, and (iii) consumed in season s, WCc ,s as given in

Equation (5.8).

Accumulated crop by end of any season ( ACc , s) = Amount of crop produced in that season (

Ac , s) + Amount of crop accumulated from previous season ( ACc , s−1) – amount of crop

consumed in that season

ACc , s = Ac , s+AC c ,s−1−WCc , s∀ c C , s S (5.8)

85

Since summer is the starting season for design, it is assumed that there is no crop

accumulated from the previous season at the beginning of summer. Thus, ACc , s−1= 0 for s

denoting summer.

The size of crop storage facilityWST c , should accommodate the maximum accumulated crop

level during the year as given in the inequality constraint (5.9).

WST c ≥ ACc , s∀ s S , c C (5.9)

The capital exergy resources for crop storage for a year, CAcr , can be calculated as follows:

CAcr=TCSL

WSTFC

with TC being the total capital resource exergy for the storage facility with a nominal size of

FC, and SL is the number of years of the service life of the storage facility.

The operating exergy resources for crop storage,OPc ,s , is given in Equation (5.10).

Operating exergy resources for crop storage in a particular season (OPc ,s ¿ = Specific

cumulative exergy of operating resources per unit accumulated crop × accumulated crop in

that season (ACc , s¿

OPc ,s=TEc ACc , s ∀ cC , s S (5.10)

TEc is the specific cumulative exergy of operating resources per unit accumulated crop and is

assumed to be the same for all seasons.

5.3.2 Mathematical formulation of water production system

The optimisation problem for the design of a water production subsystem is to meet local

water demands while minimising total exergy resource consumption of electricity, heat and

chemicals used for treating water sources to sinks as given in Equation (5.11).

Minimise objective function:

Total cumulative exergy consumption for water production subsystem (TEW) = Total

cumulative exergy of electricity + total cumulative exergy of heat + total cumulative exergy

of chemicals + total capital exergy resources for rainwater storage (CArw ¿

86

TEW=∑s S

∑b B

∑a A

(e¿¿elw+ehew+ecw)W a ,b , s+¿CArw (5.11)¿¿

TEW is the total cumulative exergy consumption for the water production subsystem, eelw is

the specific exergy resource of electricity, ehew is the specific exergy resource of heat andecw

is the specific exergy resource of chemicals per kg of wastewater treated. W a ,b , sis the amount

of water from water source a to water sink b in season s and CArw the total capital exergy

resources for the rainwater storage facility.

The optimisation is subject to the following constraints:

1) Mass balance around water sources

The total amount of water that can be supplied from source water a to the water sink b is

given by Equation (5.12).

∑b B

W a ,b , s=W a , s∀a ϵ A , s S(5.12)

where W a ,b , sis the amount of water from water source a to water sink b in season s and W a ,s

the total amount of water from water source a in season s.

2) Concentration balance with respect to chemical oxygen demand (COD) levels

The plant treats wastewater generated from various sources before they can be used in the

sinks as each of the water sinks can only accept water of a certain level of COD. The

concentration balance around the wastewater treatment plants (regenerators) adapted from

Sadhukhan et al. (2014) is given in Equation (5.13). (1- Removal efficiency of COD by

wastewater treatment plant) × Concentration of wastewater from effluents going into

wastewater treatment plant = × Concentration of treated water going out of wastewater

treatment plant(1−Ref )∑a A

coda W a , b ' , s=codb ' , s∑b B

W b ' ,b , s∀b ' B' , sS (5.13)

where Ref is the COD removal efficiency of the treatment plant, coda the COD of the

wastewater from source a, W a ,b' ,s the amount of wastewater from source a to the treatment

plant sink b’ in season s, W b ' , b ,s the amount of wastewater from the treatment plant b’ to sink

b in season s and codb ' , s the COD of the treated wastewater from treatment plant sink b’ in

season s. Equation (5.13) introduces non-linearities in the water model as a result of COD

87

contaminant mixing which gives rise to a bilinear term from the multiplication of unknowns

of the outlet flow (i.e. W b ' , b ,s ¿and concentration from the treatment plant (codb ' , s).

3) Quality constraint

The maximum allowable COD of each sink b,codb should not be exceeded by the COD level

resulting from the mixing of various supplying sources, as indicated by Equation (5.14).

∑a A

codaW a ,b , s ≤codb W b , sdem ∀bB , s S(5.14)

where W b ,sdemis the water demand of sink b in season s.

4) Mass balance around water sinks

The total amount of water supplied from source a to sink b in season s should balance its

water demand in that season, W b , sdem , as given in Equation (5.15).

∑a A

W a ,b , s=W b , sdem∀b B , s S (5.15)

5) Rainwater storage considerations

The amount of rainwater available for consumption in season s, AW scan be determined

through Equation (5.16).

AW s=RW s +ARs−1 ∀ s S(5.16)

where RW sis amount of rainwater collected in season s and ARs−1 is the amount of rainwater

accumulated from the previous season s-1.

Rainwater accumulated by the end of season s, ARs can be determined through Equation

(5.17) by the difference between (i) the sum of rainwater collected in season s (RW s ¿and the

amount of rainwater accumulated from the previous season s-1 (ARs−1) and (ii) the rainwater

consumed in season s.

ARs=AW s−∑b B

W rw, b ,s ∀ s S (5.17)

88

where W rw ,b ,s is the amount of rainwater supplied to sink b in season s.

It is assumed that stored rain water is always fully consumed within the year of storage. Thus

for summer, which is the starting season for design, ARs−1 = 0.

The optimal size of the rainwater storage tank, OSR, should accommodate the maximum

rainwater level available for consumption at any season during the year ( AW s ¿as given in the

inequality constraint (5.18).

OSR≥ AW s ∀ s S (5.18)

The capital exergy resources for rainwater storage CArw , can be calculated as follows:

CArw=TCSL

OSRFC

where TC is the total capital resource exergy for the storage tank with a nominal size ofFC,

and SL is the number of years of the service life of the tank.

5.3.3 Mathematical formulation of energy production network

The optimisation problem for the design of an energy production subsystem is to minimise

the net total cumulative exergy consumption meeting the local energy demand, comprising

resource consumption associated with raw material, capital and operating resources minus the

cumulative exergy consumption avoided by exporting any surplus local power generation to

the grid as formulated in Equation (5.19). Capital resources (i.e. those consumed for building

equipment and production facilities) for CHPs, wind turbines and solar panels were included

as these technologies consume relatively negligible operating resources; making their capital

resources relatively significant. With technologies such as the co-generation of heat and

power (CHP), it is reasonable to assume the design of the energy system will seek to meet the

local heat demand, hence possibly leading to surplus electricity for export to grid. As the

benefit of avoiding the cumulative exergy consumption associated with grid electricity

through local power export is included in the objective function, there is an incentive for the

model to choose options which will export.

Minimise objective function:

89

Total net cumulative exergy consumption for energy production subsystem = Total

cumulative exergy of raw material + Total cumulative exergy for producing electricity locally

(includes both capital and operating resources) + Total cumulative exergy for producing heat

locally (includes both capital and operating resources) - Total cumulative exergy of surplus

electricity

TEE=∑sS

∑r R

er M r ,s+∑sS

∑yY∑x X

exel E x , y , s+∑

s S∑y Y

∑x X

exhe H x , y , s−∑

s S∑x X

egrid Ex , grid ,s(5.19)

where TEE is the total net cumulative exergy consumption for the energy production

subsystem, er is the specific cumulative exergy of the raw material (i.e. energy input

including grid electricity) r,M r , s the amount of raw material r in season s, exel the specific total

(i.e. operating and capital) cumulative exergy consumption for producing electricity from

energy source x, E x , y , s the amount of electricity from energy source x to energy sink y in

season s, exhe the specific total (i.e. operating and capital) cumulative exergy consumption for

producing heat from energy source x, H x, y , s the amount of heat from source x to sink y in

season s, egrid the cumulative exergy of grid electricity and E x ,grid , s the amount of electricity

from source x exported to grid.

The optimisation is subject to the following constraints:

1) Electricity demand constraint for each sink

The supply of electricity from all electricity energy sources for local consumption should

balance the electricity demand at energy sink y per season s, E y, sdem , as shown in Equation

(5.20).

∑x X

E x , y, s=E y , sdem∀ y Y , s S(5.20)

2) Heat demand constraint for each sink

The supply of heat from all heat energy sources should balance the heat demand at energy

sink y per season s,H y , sdem , as given in Equation (5.21).

∑x X

H x , y , s=H y, sdem ∀ yY , sS (5.21)

90

3) Raw material availability constraint

The raw material availability constraint is given in Equation (5.22).

∑x X

M r , x, s ≤ M r , sAv ∀ r R , s S (5.22)

where M r , x , s is the amount of raw material r consumed for energy source x in season s, M r , sAv

the availability of raw material r in season s. Equation (5.22) applies to biomass, organic

waste, solar and wind energy available for energy production, assuming seasonal averages are

suitable for quantifying availability.

4) Land availability constraint

The total land use by the energy processing technologies should not exceed the land available

for energy production,Len , as shown in Equation (5.23).

∑sS

∑r R

∑x X

Lx M r , x, s ≤ Len(5.23)

Lx is the land use per unit raw material for energy source x and M r , x , s is the amount of raw

material r for energy source x in season s.

5) Heat integration constraints

Low temperature waste heat from food and energy production processes as a potential energy

source was considered. The constraints governing the use of waste heat for energy integration

are as follows:

a) Heat exchange between waste heat and heat sink

The amount of heat exchanged between the waste heat source x’ and energy sink y,H x' , y, s , is

constrained by Equation (5.24).

H x' , y, s ≤ H Max zx ' , y , s ∀ x ' X ' , y ϵ Y , sS (5.24)zx ' , y , s is a binary variable introduced to denote the possible existence of heat exchange

between the waste heat source (hot stream) and energy sink (cold stream) (Floudas and

Grossmann, 1986). The presence of this variable renders the energy model as a mixed-integer

91

linear program (MILP). H Max is the maximum heat load in the waste heat source for the

gradient between inlet and outlet temperature, which ensures that Equation (5.24) always

holds true.

b) Temperature difference constraint on the hot side

It is assumed that waste heat is exchanged through a counter current flow heat exchanger. A

minimum temperature difference TD between the inlet temperature of the hot flow from the

energy source x’,T x '¿ and the temperature required by the heat sink y, T y

out is required to

prevent temperature crossing in the heat exchanger. This constraint is given in Equation

(5.25).

TD ≤ (T x'¿ −T y

out )+ M (1−zx ' , y , s) ∀ x ' X ' , y Y , s S (5.25)

M is the upper bound for temperature difference; it should be high enough to make the

constraint holds with any value of the binary variable.

c) Temperature difference on the cold side

The same constraint applies for the minimum temperature difference TD between the outlet

temperature of the heat source after exchange T x 'out and the temperature of the heat sink before

exchangeT yi n as given in Equation (5.26).

TD ≤ (T x'out−T y

¿ )+M (1−zx ' , y , s) ∀ x ' X ' , y Y , s S (5.26)

d) Heat load availability of waste flows

The heat load Q x' , y, s ,representing the amount of heat energy that can be transferred from the

waste heat sourcex ' to the heat sink y is given in Equation (5.27).

∑yY

Q x ' , y , s=(T ¿¿ x ' ¿−T x'out )CP x ' , s t ∀ x ' X ' , s S (5.27)¿

CPx ' ,s is the heat capacity flow rate of waste heat source x’ for season s and t is the time

period over which heat is transferred.

e) Heat load required by heat sinks

92

The heat load constraint required for the heat sink y is given in Equation (5.28).

∑x X

Q x ' , y , s=(T ¿¿ yout−T y¿ )CS y ,s t ∀ y ϵ Y ,∀ sϵ S (5.28)¿

CS y , sis the heat capacity flow rate of the heat sink y for season s.

6) Electricity production

The total electricity produced from each source x can be determined through Equation (5.29).

∑y ϵ Y

Ex , y ,s=¿∑r ϵ R

ηx, rel M r , x ,s∀ x ϵ X , sϵ S (5.29)¿

ηx ,rel is the electrical conversion of source x for raw material r.

7) Heat production

The total heat generated from each source x can be determined through Equation (5.30).

∑y ϵ Y

H x , y, s=∑r ϵ R

ηx , rhe M r , x, s∀ x ϵ X , sϵ S ∀ x ϵ X , sϵ S (5.30)

where ηx ,rhe is the heat conversion of source x for raw material r.

5.4 Mathematical formulation for the simultaneous design

The simultaneous model will include all the constraint equations from section 5.4. In

addition, since the three subsystems are now to be designed simultaneously, certain known

parameters in the preliminary design analysis now become variables. Specifically, the

electricity demand of food and water sinks in Equation (5.20), heat demand of food and water

sinks in Equation (5.21), heat flow of waste heat sources in Equation (5.26), heat demand of

food and water sinks in Equation (5.28) and water demand of food and energy sinks in

Equation (5.15) are variables instead of known parameters.

93

5.4.1 Objective function

The optimisation problem for the simultaneous design is to minimise the total net cumulative

exergy consumption for meeting the local food, water and energy demands:

Minimise objective function:

Total net cumulative exergy resource consumption (NEC) = Total cumulative exergy

consumption for food subsystem (TEF) + total cumulative exergy consumption for water

subsystem (TEW) + total cumulative exergy consumption for energy subsystem (TEE)

TEF can be determined from Equation (5.1) where all the imported flows will be from

outside the boundary of the whole system. TEW can be determined from Equation (5.11)

where eelw+ehew will be the specific cumulative exergy of imported electricity and heat

respectively. TEE can be obtained from Equation (5.19) with eeland eelbeing the specific

cumulative exergy of imported resources for producing electricity and heat, respectively and

er is the specific cumulative exergy of only the imported raw material to the whole system.

5.4.2 Cross-subsystem flows

All the exchanges of flow occurring between the three subsystems and shown previously on

Figure 5-7 need to be represented in the integrated model.

1) Water requirements

The water requirements for producing food d, W sFD , can be determined by Equation (5.31).

W sFD=∑

d D(WCd+℘d)Fd , s

local∀ s S(5.31)

where WCd and ℘dare the amount of water required for agriculture and industrial processing,

respectively, per unit of produced food d.

The water requirement for producing energy e, W sEN ,is determined by Equation (5.32).

W sEN=WE∑

y ϵ Y∑x ϵ X

Ex, y ,s ∀ s S (5.32)

94

with WE being the amount of water required per amount of energy produced and E x , y , s the

amount of energy from source x to sink y in season s.

2) Wastewater generation

The wastewater generation from the food subsystem, WG sFD , in season s was determined by

Equation (5.33).

WG sFD=∑

d DWGP d Fd , s

local∀ s S (5.33)

where WGPdis the amount of wastewater generated per unit of produced food d.

Equation (5.34) determines wastewater generated from the energy production processes,

WG sEN.

WG sEN=WEG∑

y ϵ Y∑x ϵ X

Ex , y , s∀ sS (5.34)

with WEG being the amount of wastewater generated per unit energy produced.

3) Energy demand

The electricity demand for the food subsystem ELDsFD is determined through Equation (5.35).

ELDsFD=∑

d DELDd Fd ,s

local ∀ s S (5.35)

whereELDd is the electricity demand per unit of food d.

The heat demand for the food subsystem,HED sFD is determined through Equation (5.36)

HED sFD=∑

d DHEDd Fd , s

local ∀ sS (5.36)

whereHEDd is the heat demand per unit of food d.

95

The electricity demand for the water subsystem, E LDsWA is determined through Equation

(5.37).

ELDsWA=SEDWA∑

dDWGPd Fd ,s

local ∀ s S (5.37)

where SEDWA is the electricity demand for treating unit wastewater.

The heat demand for the water subsystem, HED sWA , is determined through Equation (5.38).

HED sWA=SHDWA∑

d DWGPd Fd , s

local∀ s S(5.38)

where SHDWA is the heat demand for treating unit wastewater.

The heat capacity flow rate of waste heat sources satisfies Equation (5.39), which is based on

Equation (5.27).

CPx ' ,s=Q x' , y ,s

(T x '¿ −T x '

out) (5.39)

where heat integration occurs, it is assumed that heat removed from the hot flows (i.e. waste

heat sources) equals the heat gained by the cold flows (i.e. heat sinks). Thus, the heat capacity

flow rate of the cold flows can be determined by Equation (5.40) based on Equation (5.28).

CW y , s=Qx ' , y , s

(T ¿¿ yout−T y¿ )¿

(5.40)

96

5.5 Case study

The methodology for the design of local production systems is illustrated by a case study for

the integrated design of the food-energy-water nexus in Whitehill and Bordon, an area

identified for the development of an eco-town in the UK. The specificities of this eco-town

are given in Table 5-1 and are based primarily from data given in the master plans for the

eco-town (Whitehill and Bordon, 2012) and DEFRA (2014).

Table 5-3: Specificities of Whitehill-Bordon eco-town

Specificities ValuePopulation 17,000

Agricultural land 17 haGroundwater abstraction limit 14,875,942 t/y

Residential water demand 2,347,470 t/yResidential electricity demand 90,254 GJ/y

Land availability for energy production 70 haRainwater availability (t)

Winter 62,899Spring 52,880

Summer 52,880Autumn 68,141

Residential heat demand (GJ/y)Winter 112,128Spring 85,848

Summer 82,344Autumn 96,360

Food demand (t/y)Bread 224Potato 403Pork 46Beef 88

Availability of energy sources (PJ/y)Wood chips 1.66

Organic waste(Animal manure and food waste) 0.10

Wind 0.40Solar 15.8 GJ/y/m2

Source: Whitehill and Bordon (2012) and DEFRA (2014).For simplicity, a small selection of foods typically consumed and with potential for local

production was chosen. The land availability for energy production includes land that can be

used for solar and wind power generation and the installation of CHP plants. It excludes land

for biomass sources as these areas are already part of the fixed geographical setting and are

not to be optimised. All data used in the design can be found in Appendix B.

97

The objective is to select the food, energy and water production processes and to determine

the flow rates of source flows to sinks that will minimise total resource consumption while

observing a set of local ecological and technical constraints for satisfying local demands for

food, energy and water.

The exergy content of flows from Type-II processes (i.e. processes that are typically not

under human control) was not considered as it is assumed that they do not have alternative

competing uses in this case study. Flows from Type-I processes (i.e. processes that can be

affected by human decisions) were accounted by their cumulative exergy consumption as

normal. Note that wheat can be planted either in autumn or spring but harvested in late

summer (UK Agriculture, 2014a) while potatoes are harvested in summer and autumn in UK

(UK Agriculture, 2014b). The models were solved using GAMS (Rosenthal, 2015), with

CPLEX as the mixed-integer linear model solver and BARON as the mixed-integer nonlinear

model solver. Note that Equation (5.13) is the only non-linear equality constraint that renders

the mathematical model non-linear.

5.5.1 Preliminary design analysis: Food production subsystem

Five scenarios were analysed for the food subsystem as summarised in Table 5-3. Scenario

F1 investigates the impact of using conventional energy and water sources for the local food

production subsystem. Scenarios F2 and F3 respectively analyse the impact of supplying all

the energy demands for food production from wood chip CHP and from organic waste CHP

respectively. Scenario F4 is similar to F2 but additionally investigates the impact of

supplying its water demand by a different water source namely collected rainwater for crop

cultivation and the rest of water demand by groundwater. It is assumed that there is enough

collected rainwater available for crop cultivation with no CExC associated with it. Scenario

F5 uses organic waste CHP and rainwater for crop cultivation.

As compared to F1, the specific CExC of electricity decreases by 10.6% and that of heat by

2% in F2. The decrease in specific CExC of heat and electricity from F1 to F2 did not affect

the result of the food subsystem design. However, the more substantial decrease in specific

CExC of electricity by 86% and 63% heat from F1 to F3 led to significant changes to the

result of the food design as 100% of potatoes and 28% of bread demand are satisfied locally

in F3 compared to 75% potatoes and 36% bread demand in F1. More specifically, Table 5-3

shows that from F2 to F3 more electricity is used but less heat and water is consumed. Figure

98

5-8, derived from Table 5-3 on the contribution analysis of resource consumption for each

locally produced food, illustrates that, as the CExC of electricity is reduced in F3, water

becomes more important for potato production. From Table 5-3, electricity consumption

decreases when only bread is produced. This means potato production was more dependent

on this input due to the relatively high electricity demand required for potato storage. In F4,

lower CExC water resource (rainwater) combined with high CExC energy source (and potato

production being much more dependent on electricity), changes the design to the production

of bread only, despite that this design requires higher heat, water and fertiliser consumption.

From these insights, it is not surprising that in F5, with low CExC water and energy sources,

bread production is also the only locally produced food.

In summary, the adoption of energy and water sources with relatively high CExC produces a

design of mixed food production with potatoes being favoured. In contrast, either high or low

CExC energy source combined with relatively low CExC water source favours bread

production. Note that this is also because assuming a value of CExC equal to zero for

rainwater (it is assumed that the crops are naturally rain fed and no rainwater storage is

needed) favours water intensive bread production.

This example illustrates how the interactions between subsystems can be systematically

analysed in order to obtain insights which are not intuitively obvious, hence demonstrating

the value of the preliminary design analysis.

99

Table 5-3: Preliminary design analysis for food production system

Proposed design ScenarioF1 F2 F3 F4 F5

Energy source

CExC of grid electricity = 5.97 MJ exergy/MJ

electricityCExC of natural gas

heat from boiler = 2.05 MJ exergy/MJ heat

Electricity from biomass CHP, CExC =

5.34 MJ exergy/MJ electricity

Heat from biomass CHP, CExC =

2.01 MJ exergy/MJ heat

Electricity from organic waste CHP, CExC = 0.83 MJ exergy/MJ electricityHeat from organic waste CHP, CExC = 0.76 MJ

exergy/MJ heat

Electricity from biomass CHP, CExC =

5.34 MJ exergy/MJ electricity

Heat from biomass CHP, CExC =

2.01 MJ exergy/MJ heat

Electricity from organic waste CHP, CExC = 0.83 MJ exergy/MJ electricityHeat from organic waste CHP, CExC = 0.76 MJ

exergy/MJ heat

Water source CExC of groundwater= 0.06 MJ/kg

CExC of groundwater= 0.06 MJ/kg

CExC of groundwater= 0.06 MJ/kg

CExC of groundwater= 0.06 MJ/kg

CExC of untreated rainwater = 0 MJ/kg

CExC of groundwater= 0.06 MJ/kg

CExC of untreated rainwater = 0 MJ/kg

Total electricity consumption(GJ energy/y)

73.8 73.8 81.1 52 52

Total heat consumption(GJ energy/y)

81.2 81.2 63.5 134 134

Total water consumption

(t/y)194,726 194,726 156,374 309,784 309,784

Total imported fertilisers

(t/y)3.35 3.35 3.29 3.53 3.53

% food demand satisfied locallyPotatoes 75 75 100 0 0

Bread 36 36 28 60 60Pork 0 0 0 0 0Beef 0 0 0 0 0

Total CExC (GJ/y) 135,260 135,168 134,247 130,283 129,747

100

Table 5-4: Contribution analysis of resource consumption for each locally produced food

Scenarios Scenario F1 Scenario F2 Scenario F3Scenario

F4

Scenario

F5

Resource

consumed

(MJ/h)

Bread Potatoes Bread Potatoes Bread Potatoes Bread Bread

Water 11281 441 11281 441 8828 588 13143 13143

Electricity 188 501 168 448 15 98 278 30.8

Heat 205 0 186 0 27 0 362 73.6

Fertilisers 70 40 70 40 55 53 115 115

Storage capital

resources0.11 0.51 0.11 0.51 0.04 1 0.34 0.34

Total resource

consumption11745 982 11706 929 8925 740 13899 13364

Bread Potatoes Bread Potatoes Bread Potatoes Bread Potatoes Bread Potatoes0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Resource consumption by locally produced food Storage capital resources Fertilisers Heat Electricity Water%

Scenario F1 Scenario F2 Scenario F3 Scenario F4 Scenario F5

Figure 5-8: Proportion of resource consumption for each locally produced food

5.5.2 Preliminary design analysis: Water production subsystem

Four scenarios (W1, W2, W3 and W4) were analysed for the water subsystem, with results

presented in Table 5-5. The scenarios are chosen to compare a traditional water production

subsystem (i.e. no usage of rainwater and use of grid electricity and natural gas boiler for

101

heat) with alternative subsystems that use rainwater for crop cultivation and renewable

energy such as biomass wood chips CHP and organic waste CHP. Scenario W1 illustrates a

traditional water production subsystem that uses conventional energy and handles water

demands from food scenario F1. Scenario W2 is similar to W1 but handles water demand

from food scenario F4. Scenario W3 investigates the impact of using energy from biomass

(wood chips) CHP and water demand of food scenario F4. Scenario W4 of the water

subsystem is similar to W3 except that energy is supplied by organic waste CHP.

The resource consumption of every water source in each scenario is given in Figure 5-9. It

reveals that electricity contributes the most to the total CExC of each water source and that

the use of groundwater is highly dependent on electricity consumption. Increasing the water

demands of food processes by 4.5% from scenario W1 to W2 did not have any significant

impact on the water subsystem design as the percentage of each water source in the water

supply is fairly consistent with scenario W1, still with groundwater dominating the water

supply. The overall decrease in specific CExC of the energy supply by 8.6% coupled with the

increase in water demand by 4.5% from scenario W2 to W3 did not impact on the design of

the water subsystem either; showing the weak link between the water and the energy

subsystems in this particular case (the energy sources such as biomass CHP requires

relatively little water supply). This weak connection is further demonstrated by the fact that

no alternation to the water subsystem design was triggered by the overall noteworthy 91.6%

decrease in the specific CExC of energy supply from scenario W1 to W4.

Chemica

lsHea

t

Electr

icity

Storag

e

Chemica

lsHea

t

Electr

icity

Storag

e

Chemica

lsHea

t

Electr

icity

Storag

e

Chemica

lsHea

t

Electr

icity

Storag

e

0

20000

40000

60000

80000

100000

120000

140000

160000 Resource consumption for each water source

Groundwater

Rainwater

Treated domestic wastewater

Cumulative exergy (GJ/y)

Scenario 1Scenario 2

Scenario 3

Scenario 4

Figure 5-9: Resource consumption by each water source

102

Table 5-5: Preliminary design analysis for water production system

Proposed design

ScenarioW1 W2 W3 W4

Energy source

CExC of grid electricity = 5.97 MJ exergy/MJ

electricityCExC of natural

gas heat from boiler= 2.05 MJ exergy/MJ heat

CExC of grid electricity = 5.97 MJ exergy/MJ

electricityCExC of natural

gas heat from boiler= 2.05 MJ exergy/MJ heat

CExC of electricity from biomass CHP

= 5.34 MJ exergy/MJ electricity

CExC of heat from biomass CHP =

2.01 MJ exergy/MJ heat

CExC of electricity from organic waste

CHP = 0.83 MJ exergy/MJ electricity

CExC of heat from organic waste CHP

= 0.76 MJ exergy/MJ heat

Water demandResidential

Scenario 1 of food subsystem

ResidentialScenario 4 of

food subsystem

ResidentialScenario 4 of food

subsystem

ResidentialScenario 5 of food

subsystem% of water supply by different water sources

Groundwater 49.7 50.8 50.8 50.8Treated

residential wastewater

13.6 14.0 14.0 14.0

Rainwater 36.8 35.2 35.2 35.2Total CExC

(GJ/y) 214,124 219,205 200,251 44,325

However, the total CExC for scenarios W3 and W4 decrease respectively by 6.5% and 79.3%

as compared to scenario W1. Overall, the lower specific cumulative exergy consumption

associated with biomass and organic waste CHP energy supply has a net positive benefit on

resource consumption despite the increase in water demands.

5.5.3 Preliminary design analysis: Energy production subsystem

Three scenarios were analysed for the energy subsystem as shown in Table 5-6. Scenario E1

considers energy demands from food scenario F1 and water scenario W1. Scenario E2

analyses the impact of switching the energy demands to that of food scenario F4 and water

scenario W3 where wood chip was used. Scenario E3 investigates how the energy demands

from food scenario F5 and water scenario W4, which considered organic waste CHP for

satisfying energy demand, affect the design of the energy subsystem. Figure 5-10 illustrates

the resource consumption by each energy source for heat and electricity.

103

Biomass CHP Power

Biomass CHP heat

Organic w

aste C

HP power

Organic w

aste C

HP heatWind

Solar

Biomass CHP Power

Biomass CHP heat

Organic w

aste C

HP power

Organic w

aste C

HP heatWind

Solar

Biomass CHP Power

Biomass CHP heat

Organic w

aste C

HP power

Organic w

aste C

HP heatWind

Solar

0

100000

200000

300000

400000

500000

600000

Resource consumption by each energy sourceEnvironmental remediation capital resources Local water Raw material

Cumulative exergy (GJ/y)

Scenario 2 Scenario 3Scenario 1

Figure 5-10: Resource consumption by energy source

The CExC of each energy source is attributed mainly to its raw material followed by

resources consumed for environmental remediation. It can also be inferred that changes in the

sources of water supply would not affect the energy subsystem as the contribution of water in

the total resource consumption for each energy source is negligible. Though total CExC of

energy generation from organic waste CHP was estimated to be the lowest, its contribution to

energy production is severely constrained by its relatively poor feedstock availability in the

Eco-Town and as such did not contribute much to satisfying local energy demand.

Due to the higher energy demand for food and water processes in comparison with scenario

E1, the total CExC of scenario E2 increases by 4%. Since total energy demand from scenario

E3 is similar to that in scenario E2, both yielded similar results for the design of the energy

subsystem. The percentage of heat supply is the same in all scenarios while the electricity

mix for satisfying local demands is rather consistent in all 3 scenarios; indicating the

relatively weak dependence of the energy subsystem on the food and water subsystems in this

particular case. There is a slight increase in the contribution of solar power in favour of wind

and biomass CHP power in scenarios E2 and E3 as compared to scenario E1. Though wind

power has a lower total CExC in comparison with solar and biomass CHP, its share in the

electricity mix is more constrained by its availability and land use. Overall, the energy

104

subsystem behaves fairly linearly with increase in total CExC as total local energy demand

increases.

Table 5-6: Preliminary design analysis for energy production system

Proposed design ScenarioE1 E2 E3

Energy sources considered:Electricity from biomass, organic waste and natural gas CHP, solar and wind

Heat from biomass, organic waste and natural gas CHP and natural gas boilersWaste heat from food, water and energy processes

Energy demand

ResidentialScenario 1 of food

subsystemScenario 1 of water

subsystem

ResidentialScenario 4 of food

subsystemScenario 3 of water

subsystem

ResidentialScenario 5 of food

subsystemScenario 4 of water

subsystemSurplus electricity

(GJ/y) 105,777 104,989 104,989

% electricity supply by each sourceWind 43.4 40.5 40.5Solar 32.0 35.9 35.9

Biomass 24.6 23.6 23.6% heat supply by each source

Wood chip CHP 70.9 70.9 70.9Organic waste CHP 12.7 12.7 12.7

Waste heat 16.4 16.4 16.4Total CExC (GJ/y) 123,337 128,217 128,217

5.5.4 Simultaneous approach results

The results of the simultaneous optimisation for a design period of one year are illustrated in

meet the water demand of the eco-town. The eco-town is also self-sufficient in its electricity

and heat supplies through the use of locally available resources of organic waste, wood chips,

solar and wind, and with waste heat recovery providing for 16.4% of its heat demand.Figure

5-11 and reported in Table 5-7. The total CExC was determined to be 273,901GJ/y. Though

bread and potatoes were produced in the food scenarios of the preliminary analysis of the

food subsystem, the simultaneous design indicates that 17.6% of potatoes and 86% of pork

demand can be satisfied locally; suggesting that pork will offer better compromise on

resource consumption for the overall food-energy-water local design. 14% treated residential

wastewater, 50% groundwater and 36% rainwater would be used to meet the water demand

of the eco-town. The eco-town is also self-sufficient in its electricity and heat supplies

through the use of locally available resources of organic waste, wood chips, solar and wind,

and with waste heat recovery providing for 16.4% of its heat demand.

105

Figure 5-11: Results of simultaneous design

106

Table 5-7: Detailed results from simultaneous approach

Source Sink Locally produced food (t)Winter Spring Summer Autumn

Local pork Local consumption 10 10 10 10

Local potatoes Localconsumption 0 0 36 36

Imported animal feed Pig rearing 9 9 0 0

Potato residues Pig rearing 0 0 9 9Pig manure Potato cultivation 0 0 0.14 0.14

Source Sink Water supply (t)Winter Spring Summer Autumn

Water flows @ ≤0.010 g COD/kg

Groundwater: 50.1%

Rainwater: 35.9%Treated residential wastewater: 14.0%

Residential 586,867 586,867 586,867 586,867Food processes (cultivation and

processing)59,552 59,552 60,403 60,403

Energy processes 5257 4031 3868 4522

Water flows @ ≤0.10 g COD/kg Discharge 276,475 278,727 278,162 276,794

Source Sink Energy supply (GJ)Winter Spring Summer Autumn

ElectricityBiomass CHP:

23.5%Wind: 40.6%Solar: 35.9%

Waterprocesses 7685 7992 7966 8011

Residential 22,566 22,566 22,566 22,566Food

processes 15.9 15.9 30.5 30.5

Grid(export of surplus

electricity)18,757 22,484 29,767 34,136

Heat from biomass CHP Residential 82,161 60,199 57,270 68,392

Heat from biomass CHP Water processes 0.03 0.03 0 605

Heat from organic waste CHP Residential 11,463 11,436 12,045 12,045

Heat from organic waste CHP Food processes 0 14.9 0 0

Heat from organic waste CHP Water processes 582 594 0.17 0

LT waste heat from biomass CHP Residential 16,095 11,804 11,229 13,514

LT waste heat from biomass CHP Food processes 14.9 0 0 14.9

LT waste heat from organic waste CHP Residential 2409 2409 1800 2409

LT waste heat from organic waste CHP Water processes 0 0 594 0

LT waste heat from organic waste CHP Food processes 0 0 14.9 0

107

In order to fully investigate the benefits of the simultaneous approach for producing an

integrated design of the local system, two further reference scenarios were developed and

compared with the integrated design. The first scenario, termed “centralised supply”,

assumed that all the local demands of the eco-town were met by imported food and

conventional utility sources of grid electricity, heat from natural gas boilers and groundwater.

The second scenario, termed “design in silos” involves designing each subsystem separately

and independently without considering the synergies between them. The food subsystem is

designed considering only grid electricity and heat from natural gas boilers; groundwater and

any wastewater generated is treated within the food subsystem. The water subsystem is

designed to supply the water demand of the residential sector. It considered only the options

for using water sources of different quality available within this subsystem, such as rainwater

and treated residential wastewater (i.e. COD concentration). In addition, the energy

subsystem is also designed to only meet the residential energy demand, without considering

heat recovery options between the subsystems but allowing for choice from the full range of

energy sources.

A comparison of the net CExC of all three scenarios for food, water and energy subsystems is

given in Figure 5-12. There is a general decrease in the CExC of the food, water and energy

subsystems of the integrated design as compared to the other two scenarios. Overall, the total

CExC of the centralised supply and that of design in silos were determined to be about 6 and

2 times respectively higher than the integrated design. Figure 5-13 indicates that there is not

much difference in the resource consumption of the food subsystem for the three scenarios.

Though imported food dominates resource consumption in the food subsystem, producing

food locally consumes high volumes of water which reinforces the need to exploit water re-

use between the subsystems for local food production. Interestingly, imported fertilisers

account for only a negligible percentage of total CExC in all three scenarios; suggesting that

for this particular case study coupling between the subsystems (e.g. re-use of organic residues

from water and energy processes) for satisfying nutrient demands, not considered in this

study, will have a negligible impact on the total resource consumption of the food subsystem.

108

Food subsystem Water subsystem Energy subsystem0

200000

400000

600000

800000

1000000

1200000

1400000

Net cumulative resource consumption for the 3 scenarios

Integrated design

Centralised supply

Design in silos

Cumulative exergy (GJ/y)

Figure 5-12: Net resource consumption for each scenario

Integrated design Centralised supply Design in silos 110000

115000

120000

125000

130000

135000

140000

Resource consumption of food subsystem for the 3 scenarios by resource type

Wheat storage

Potatoes storage

Imported fertilisers

Imported animal feed

Heat

Electricity

Water

Imported food

Cumulative exergy (GJ/y)

Figure 5-13: Cumulative exergy of food subsystem for all 3 subsystems

In addition, the CExC of all the resources associated with each water source in all scenarios is

analysed in Figures 5-14(a)-(d). The CExC of chemicals for wastewater treatment was higher

for the integrated design due to higher volumes of wastewater generated from the food and

energy subsystems. Both the integrated and design in silos scenarios require capital resources

for rainwater storage as compared to the centralised supply scenario.

109

Integrated design Centralised supply

Design in silos0

1000

2000

3000

4000

5000

6000

7000

(a) Consumption of chemicals in all 3 scenarios for each water source

Total Rainwater supplyGroundwater supplyEnergy wastewater treatment Residential wastewater treatment & supply

Cumulative exergy (GJ/y)

Integrated design Centralised supply Design in silos0

5001000150020002500300035004000

(b) Consumption of heat in all 3 scenarions by each water source

Total Rainwater supplyGroundwater supplyEnergy wastewater treatment Residential wastewater treatment & supply

Cumulative exergy (GJ/y)

Integrated design Centralised supply Design in silos

0

50000

100000

150000

200000

250000

(c) Consumption of electricity in all 3 scenarios by each water source

Total Rainwater supplyGroundwater supplyEnergy wastewater treatment Residential wastewater treatment & supply

Cumulative exergy (GJ/y)

Integrated design Centralised supply Design in silos0

5000

10000

15000

20000

25000

(d) Capital resource consumption in all 3 scenarios by each water source for storage

Total Rainwater supplyGroundwater supplyEnergy wastewater treatment Residential wastewater treatment & supply

Cumulative exergy (GJ/y)

Figure 5-14: Cumulative consumption of resources (a) chemicals, (b) heat, (c) electricity, (d) capital resources by all 3 scenarios for each water source

However, the higher CExC for chemicals (cf. Figure5-14(a)) and rainwater storage (cf.

Figure5-14(d)) in the integrated design is offset by much less resource intensive energy

consumption (cf. Figures5-14(b) and 5-14(c)). As compared to the centralised supply and

design in silos scenarios which consume conventional energy for their water subsystems,

solar, wind, wood chip and organic waste CHP as well as low temperature waste heat

generated from the energy subsystem are used to satisfy the energy demands of the water

subsystem of the integrated design. Figure 5-14(c) also indicates that using rainwater to

satisfy part of the water demand in the eco-town, as is the case in the integrated design and

design in silo scenarios, will also contribute to lower total resource consumption as it has

very low CExC as compared to groundwater supply. Besides, the CExC associated with

groundwater (cf. Figures 5-14(a-c) is not insignificant; re-using part of the treated residential

wastewater instead of discharging them to the local environment will reduce the consumption

of fresh water resource and contribute to lower total resource consumption.

110

Figure 5-15 illustrates the CExC for production of electricity, heat and surplus electricity in

all three scenarios by each energy source (i.e. natural gas boiler, wood chips CHP, organic

waste CHP, grid, solar and wind power). The major improvement in resource consumption of

the energy subsystem in the integrated design as compared to the centralised supply is due to

the significantly lower CExC associated with solar, wind and organic waste CHP than that of

grid electricity and natural gas boilers. For the design in silos, the resource available for

organic waste CHP is constrained to be from the residential sector alone, and the renewable

energy options are not required, with the majority of demand met by biomass CHP. The

integrated design allows use of organic wastes from the food and water sectors as well, and

also allows more than 16% of the heat demand in water and food sub systems to be met by

low temperature waste heat recovered from the energy subsystem. The total surplus

electricity generated from the integrated design is higher by about 60% as compared to that of

the design in silos, further contributing to a lower overall net CExC.

Electricity Heat Surplus electricity

Electricity Heat Surplus electricity

Electricity Heat Surplus electricity

-1000000

-500000

0

500000

1000000

1500000

2000000 Resource consumption for all 3 scenarios by each energy technology

Wind power

Solar power

Grid power

Organic waste CHP

Wood chips CHP

Natural gas boilers

Cumulative exergy (GJ/y)

Integrated design

Centralised sup-ply

Design in silos

Figure 5-15: Cumulative consumption by each technology for energy production in all scenarios

Overall, about 50% resource savings were achieved by the integrated design of the local

food-energy-water nexus as compared to the design in silos approach, which is in line with

the reported benefits by integrated design as applied to other systems, e.g. 29% reduction in

total cost with material by-products exchange (Cimren et al., 2011) and more than 80%

111

savings in total energy cost with the implementation of waste heat recovery in industrial

parks (Chae et al., 2010). The results were the outcome of applying the specific set of

parameters, which nevertheless have shown the key (and likely representative) mechanisms

for resource savings by a locally integrated system (use of renewables, local resource

cascading use etc.).). Throughout the case study, a number of quantitative results has been

presented, which have been obtained based on a specific set of parameter values mostly

adopted from literature. In reality, there are inevitably issues arising from data quality and

uncertainties, which may impact on the reliability of individual results and consequent

decision recommendations. There are established approaches to deal with these issues, such

as sensitivity analysis and optimization with uncertainties, and these have in fact been applied

along with the work on the optimal design of local FEW system in chapter 7 of the thesis, to

allow this chapter to focus on the main design approach.

5.6 Summary of preliminary and simultaneous design approaches to LIPS

This chapter has proposed a methodology for the design of local production systems, and

developed a superstructure-based optimisation model specifically for design of the food-

energy-water nexus in a local context. Based on the optimisation models of individual

subsystems, the preliminary design analysis allows insights to be gained about the design

alternatives. It reveals interdependencies between subsystems and the how the inter-

subsystem coupling options would affect substantially the choice made for the internal design

of individual subsystems. The simultaneous approach, utilising models of all subsystems and

accounting for all the possible interactions between the subsystems, offers an optimal

solution for the integrated, whole-system design. It was shown that designing simultaneously

the subsystems of the food-energy-water nexus allows capture of all the integration options

and exploitation of emerging synergies and opportunities for circularity. Mathematical

programming based design approaches do not offer much insight to practitioners. A main

limitation of such an approach is that the solution process is entirely one of mathematically

solving an optimization model: although a decision maker can be involved in constructing the

model, what is presented to him or her is merely the final design, which makes it difficult to

understand the impact of different factors, options, inter-subsystem interactions and trade-

offs. In the next chapter, an insight-based design approach that can give a wider range of

design options to guide decision makers based on their preference and core business will be

presented.

112

Chapter 6: An insight-based approach for the design of integrated

local food-energy-water systems

6.1 Rationales for an insight-based design approach

A variety of different process systems engineering approaches have been used such as

mathematical programming (MP) techniques (Grossmann et al., 1999), insight-based

techniques (Foo, 2007; Foo, 2013) and hybrid techniques (Luo et al., 2009; Martin and

Grossmann, 2015) in the design of processes for improving resource efficiency (Klemes et

al., 2013). Insight-based approaches are more readily adopted by industries and practicing

engineers because they allow physical insights used routinely by engineers and designers to

be included in the design problem formulation (Klemes et al., 2013). One of the unique

benefits of the insight-based approach as compared to mathematical programming approaches

is that it allows decision variables to remain tractable throughout the design of the local

production system. The generation of intermediate results will allow more insights to be

incorporated into the design of such production systems. These results can be analysed,

interpreted and used to inform the implications of a given design for the operation of the

whole system, thus making the insight-based approach a more practical tool for decision-

makers and local planners to use. More importantly, with MP-based approaches, any insights

are generated only at the end of the decision-modelling process; instead, an incremental

procedure would allow the decision makers to gain such insights at a series of intermediate

steps, where they can also scrutinize, approve or reject the “optimal” options identified from

the suggested design principles and provide further information such as new design options

based on their preference and core business and adapted parameter values that can be

incorporated in the subsequent stages of the design process.

Insight-based methods such as pinch analysis techniques have been successfully implemented

for the analysis of different resources such as heat (Linnhoff, 1993), water (Wang and Smith,

1994) and hydrogen (Nelson and Liu, 2008), to name a few. The well-established pinch

analysis methods can be used to set utility targets and design energy systems at unit, process

and inter-process level. Energy integration has been illustrated through the concept of locally

integrated energy sector (LIES) (Perry et al., 2008; Kostevsek et al., 2015) where heat and

renewable energy sources are integrated and exchanged between diverse industrial processes.

113

Mass pinch analysis, especially water pinch analysis in new-build and existing water

networks (Tan et al., 2007; Manan et al., 2006; Foo et al., 2006) is also well established and

has been applied within and between different processes and more recently for targeting

carbon emissions (Foo and Tan, 2015). However, insight-based approaches have so far been

applied mainly within the same production system and among conventional production

systems of energy and water. These conventional systems overlook the potential for

integration with other types of production systems such as agricultural production to satisfy

local demands in the most sustainable manner. Locally integrated production systems (LIPS)

for the satisfaction of local demands have the potential to offer a more sustainable path

towards development (Leung Pah Hang et al., 2016b; Martinez-Hernandez et al., 2016). The

definition for LIPS and its subsystem components and the rationales for designing such a

system have been detailed in Chapter 1 and in Martinez-Hernandez et al. (2016).

This chapter, based on a research article on “An insight-based approach for the design of

integrated local food-energy-water systems’ submitted for publication in Environmental

Science & Technology, is about developing a systematic insight-based approach for the

design of LIPS. Such an approach is required as it can give better insights to practitioners. A

systematic insight-based approach also offers an incremental approach with an appropriate

balance between:

Capturing complexities while keeping the algorithms/methods simple but robust.

Capability to apply mathematical modelling for solving sub-problems.

Ability to verify the outcome of the design at any stage of the approach and to embed

feedback from users.

The following sections detail the developed insight-based approach for the design of a LIPS

and its application on a case study on food-energy-water production based on an eco-town in

the UK. The design problem in this work considers (i) the local dimension of resources and

demands, (ii) a diverse range of industrial, manufacturing and agricultural processes and (iii)

the seasonal variability of the amount of resources available.

6.2 Aim and Objectives

The design problem is to determine the combination of a set of processes and activities which

can meet a set of demands (e.g. food, energy and water to satisfy local basic needs) by a

population in a locality within the availability of local resources so that total resource

114

consumption is minimised while observing a set of technical (e.g. conversion efficiency of

processes), environmental (e.g. discharge limits to rivers) and ecological constraints (e.g.

biomass growth rate). The main novelty of the present work comprises a systematic approach

for generating insights into the design of a localised production system which

unconventionally involves distinct resources (e.g. residues, intermittently available wind and

solar, heat) and processes with very diverse natures (e.g. industrial, agricultural and

municipal).

More specifically, the various aspects in the novelty of this work include:

(1) Developing design rules that will be used to generate a basic design of a local

production system focused on meeting local demands based on locally available

resources.

(2) Adaptation of existing insight-based methods for the design of an integrated local

production system. Existing methods such as pinch analysis will be used to integrate

heat and water resources not only within subsystem but also across subsystem (e.g.

across food, energy and water production processes).

(3) Developing the resource gain indicator to guide local resource allocation, by offering

a useful metric for decision making between purposing, regeneration and re-purposing

a particular resource.

The work presents for the first time an insight-based design approach aimed at designing

locally integrated production systems (LIPS) for food, energy and water, which is able to

generate insights on comparison between design options, impact of constraints, and

interconnections between subsystems. It covers both (i) the generation of the basic design of

individual subsystems taking into consideration their interactions and (ii) cross-subsystem

integration to maximise the whole-system resource efficiency. Its application is demonstrated

through a case study on an eco-town in the UK.

6.3 Methodology for insight-based approach

6.3.1 Overview of methodological framework for insight-based design approach

The design of LIPS generally involves the production of several products and services to

meet local demands taking into consideration locally available resources and ecological

constraints (e.g. groundwater abstraction limit). Given a set of demands by a local population

115

and the availability of local and external resources, the design task is to determine the

combination of processes and activities to meet such demands so that the total net cumulative

exergy resource consumption is minimised while observing all necessary constraints (Leung

Pah Hang et al., 2016b). The insight-based approach for LIPS design consists of two main

stages, namely synthesis and integration, guided by a Locally Integrated Production System

Onion Model (LIPSOM). The synthesis stage produces a base design of LIPS based on

principles for designing individual subsystems and a sequential synthesis procedure for

converging the design of subsystems into that of the entire LIPS. In the integration stage,

options for cascading, regeneration and re-purposing of resource streams are considered.

Design decisions in both stages make use of the concept of “resource gain”, to eventually

achieve the goal of minimising resource consumption.

6.3.2 Design goal and resource gain

As mentioned earlier in section 6.2, the design goal is to minimise the total cumulative

exergy consumption (CExC) of LIPS in meeting the local demands and while satisfying

resource and ecological constraints. CExC is defined as the sum of exergy of all resources

consumed along the supply chain of a product/service (Szargut et al., 1988). The approach

developed by Szargut et al. (1988) for reporting CExC is based on the principles of

attributional LCA which accounts for the environmentally relevant physical flows to and

from a life cycle and its subsystems (Martin et al., 2015;Hertwich, 2015;Sadhukhan et al.,

2014). Consequential LCA expands the boundary of the attributional LCA and is a system

modelling approach which describes how relevant environmental flows will change in

response to possible decisions without taking into account whether these changes take place

within or outside of the cradle-to-grave system being investigated (ISO, 14040).

Attributional/change oriented and consequential/accounting LCA have different purposes

aimed to answer different sets of sustainability questions; while the former traces a specific

aspect of the product back to its contributing unit processes based on allocation rules and is

useful in making comparisons between products prospectively, the latter is a decision support

tool that provides a comparison of a decision made today with existing ones retrospectively

and involves changing an existing product/service with a new one that provides the same

functionality (Sadhukhan et al., 2014). As a key aspect of considering LIPS is to assess its

comparative performance against conventional options, e.g. satisfying local demands with

external imports from centralised production, a resource gain (RG) indicator based on the

principles of consequential LCA is introduced to guide design decisions, which is defined as

116

the net avoided CExC due to substitution of a reference (often conventional) option by a

different, alternative option to be evaluated:

RG=CExCref−CExC alt (6.1)

CExCref and CExCalt are the CExC by adopting the reference option and the alternative

option, respectively. The nature of a design option being assessed depends on that of the

corresponding decision to be made, which could be, for example:

producing locally or importing (e.g. local or imported bread);

selection of design technology/process options (e.g. wind or solar for energy supply);

selection of resource for a production process (e.g. local or imported; fresh or re-

generated), or

resource allocation between competing uses (e.g. land for food and/or energy crops).

The last decision task requires the use of a slightly adapted concept, namely specific resource

gain (SRG). A resource as provided by the local environment may serve multiple competing

purposes and satisfy various demands. For example, crop residues, as a resource, can either

be used for animal feed or energy generation. Choosing one or another purpose may result in

lower or greater resource consumption. In order to decide which purpose should be satisfied

with priority, SRG is defined in Equation (6.2) as resource gain per resource quantity

allocated to a particular purpose:

SRG=(CExC¿¿ ref −CExCalt)/ F r ¿ (6.2)

where F r is the resource quantity allocated. As such, SRG for fulfilling a certain purpose in

connection with a particular resource allows this purpose to be gauged with other competing

ones, to support resource prioritisation which may take place within (see Section 6.3.4) or

between (see Section 6.4.2) subsystems.

6.3.3 LIPSOM: Locally Integrated Production System Onion Model

The design activities in the proposed insight-based approach are organised with the guidance

of the LIPSOM, as shown in Figure 6-1.

117

Figure 6-1: Locally Integrated Production System Onion Model (LIPSOM)

LIPSOM is based on the onion model developed by Douglas (1988) as a conceptual design

approach for process synthesis, in which a hierarchical procedure is followed starting from a

reactor and gradually expanding the system boundaries as successive levels are added. The

onion model highlights the sequence of design steps where each layer involves making

design decisions that will affect those to be made in the successive layers of the model. As an

adaptation to the original onion model, LIPSOM consists of five sequential layers and aims at

capturing important interdependencies between them. The model starts with the ecological

layer, to develop considerations on the local ecosystem consisting of components such as

water bodies, atmosphere and land. The main purpose of design considerations at this layer is

gathering ecological information such as the resources available and their constraints (e.g.

biomass yield, land availability). This will offer important inputs to the subsequent design

particularly with respect to resource selection and allocation, and identification of ecosystem

components that could be later interconnected to the production processes (e.g. wetland for

wastewater treatment).

Next, the agricultural layer addresses processes for producing food and non-food crops and

livestock. In this layer, one can establish what agricultural processes can take place in the

locale of concern, design options for each process, and the resource cost for each option.

Locally available resources identified from the ecological and agricultural layers can be

classified into: a) basic resources from existing ecosystems and the environment such as

solar radiation, wind, forest biomass, that are unprocessed and unmanaged; and b) managed

118

Agricultural

Ecological

Industrial

Resource cascading

Environmental remediation

resources such as cultivated biomass, food crops and livestock. As suggested earlier (Leung

Pah Hang et al., 2016b), the cost of the basic resources is simply measured by their exergy

content, while that of the managed resources is quantified by the cumulated exergy

consumption across all the steps and activities for producing such resources.

The next layer is the industrial layer where industrial (including municipal) processing units

are considered, such as those for food processing (e.g. bread production), energy conversion

(e.g. solar panels), and water processing (for clean water supply or wastewater treatment).

Process options, judged feasible based on the resource availability identified in the ecological

and agricultural layers, will be identified and their resource consumption evaluated.

Combined with the learning from the agricultural layer, this will enable the base design of

subsystems producing a final product to meet local demand (e.g. electricity or bread

production); the principles for such design are outlined in Section 6.3.4.

In the resource cascading layer, any used or residual resources generated from the agricultural

and industrial production processes (e.g. wastewater), which have not been fully utilised in

the basic design, are considered for further utilisation. As the process integration part of the

insight-based approach, this layer takes into consideration all the process integration

opportunities within and across the subsystems, including options for regeneration and re-

purposing of resources. The last layer comprises the design of environmental remediation

units where either a technological or an ecological (e.g. wetland) option can be considered for

the treatment of each of the resources identified from the cascading layer that cannot be

further used for any purpose. The application of the process integration stages (i.e. last two

layers of LIPSOM) may lead to two different outcomes:

(i) There is no change in the base case local production system and the result of

process integration will simply make the basic system more efficient, or

(ii) Process integration leads to significant reduction of resource cost by changing

design options and evolving into new alternative designs. In the latter case, some

iteration between the layers of synthesis and integration of the LIPSOM will be

required. Given a base design, a sensitivity analysis can reveal whether the change

in a specific resource flow would lead to a very different outcome; then in the

process integration stage special attention should be given to such resource

streams and the synthesis re-done whenever process integration does lead to a

considerable change to this resource stream.

119

A series of steps (guidelines) and rules have been devised in this work for each layer of the

LIPSOM in order to aid in the design of LIPS. Design rules have been used extensively by

process engineers who rely on their personal experience in designing similar systems and on

insights into the physical and chemical phenomena relating to unit operations and offer a way

to quickly locate one or several acceptable solutions (Nishida et al., 1981). The main

advantage of using design rules (also known as heuristics or rules of thumb), particularly as

compared to a mathematical programming design approach which solves the design problem

by using a monolithic optimization model and any insights are generated only at the end of

the decision-modeling process, is that the decision makers and local planners can keep

control of the basic decisions and interact as the design develops. By staying in control of the

basic decisions, the intangibles of the design can be included in the decision making (Smith,

2005).

6.3.4 Principles for designing individual subsystems

A single subsystem of LIPS may involve components from ecological, agricultural and

industrial layers. For example, a food subsystem could make use of local land and rain water

to grow crops which are further processed into food products. Typically, the design of such a

subsystem needs to incorporate considerations pertaining to all these layers, following several

steps:

1. Identify the availability of local and external resources.

2. Identify technical options for agricultural or industrial processes required.

3. Determining resource implications for each resource or technical option.

4. Establish a reference system comprising conventional options, against which

alternative, potentially more advantageous options could be compared.

5. Determine the design by choosing advantageous options based on resource gains.

Several details are given below for the final decision making step:

Firstly, at this step, RG is typically calculated for each alternative option, according to

Equation (6.1), and an option with a positive RG, or one with the greatest positive RG when

multiple alternatives exist, will be adopted. In some special cases where a retrofitting

decision is to be made on whether a new and operationally more efficient technical

component (e.g. an anaerobic digestion reactor for energy production) should be introduced,

120

the operational RG, calculated as operational gain per unit time without considering resource

consumptions needed for building and installing the new component - referred to as capital

resources - can be first calculated. Next, the period of time for the payback of capital

resources is determined as CExCcap

RG , with CExCcap denoting the CExC for capital resources

associated with the new facility in question. This payback can then be compared with the

decision maker’s expectation or the service life span of the component, to determine its

favourability.

Secondly, if multiple purposes (e.g. in agriculture, local production of bread and pork) within

the subsystem compete for a limited resource (e.g. agricultural land), resource allocation will

be based on the SRG (introduced in Section 6.3.2) of each of these purposes. In principle, the

resource is to be used first to satisfy the purpose with the highest SRG, until the demand for

that purpose is met either fully or to the extent possible. If spare resource is still available,

remaining purposes are satisfied in the descending order of SRG.

Thirdly, the design needs to ensure that all the ecological limits identified in the ecological

layer are met. In particular, the selection of a resource flow i needs to satisfy

c i+x i

y i<1 ,wherec i is the current (existing) consumption, x i is the (additional) quantity of the

resource consumed for the design of the subsystem and y i is the maximum quantity of the

resource available locally according to either the natural replenishment rate or any authorised

constraint (e.g. water abstraction limit set by water authorities).

As a final remark, the above rules and principles could be sufficient for “manually” designing

a subsystem where not too many options are to be assessed. However, when the number of

options increases to a level which makes a manual design prohibitively complex, a

mathematical programming problem could be formulated at the subsystem level to minimise

total CExC of the subsystem while observing resource and ecological constraints as stated

above.

6.3.5 Cumulative exergy consumption of local products

The main caveat of using these decision rules is that in practice there will be many other

factors influencing the choice between different technologies locally. However, this work

121

emphasized on the resource implications of choices. Applying the design rules across the first

3 layers of the LIPSOM will help to select the resources and processing units and to

determine the total cumulative exergy consumption, CExC local for satisfying a particular

demand locally. In order to compare and select the best use of available resources for a

particular product calculations have to be done independently for each product option. Thus,

perform the calculation of CExC local for each product demand as if all the selected resources

were available for producing that product. Using the general resource accounting framework

presented previously in Chapter 3, the resource accounting algebra for producing a product

demand locally using the LIPSOM principles is given in Equation (6.3).

CExC local=∑C ExCEL+∑ CExC AL+¿ ∑CExC IL (6.3)

where,

∑C ExC ELis the total cumulative exergy of ecological resources,

∑CExC AL is the total cumulative exergy consumption for producing agricultural products,

∑CExC IL is the total cumulative exergy consumption for producing industrial products.

∑C ExC EL does not include capital and environmental remediation resources but takes into

consideration the quantification of Type-I and Type-II process flows. ∑CExC ALand

∑CExC IL include cumulative exergy consumption for operating, capital and environmental

remediation resources following the conceptual framework for resource accounting and

quantitative assessment of resource consumption at each level of the multi-level framework

as presented in Chapter 2 and 3 respectively. The agricultural products or resources from the

agricultural layer can be processed further in the industrial layer to produce the industrial

products. Care should be taken to avoid double counting ∑C ExC EL and ∑CExC AL in the

determination of ∑CExC ALand ∑CExC IL respectively. For example, the total cumulative

exergy of ecological resources should be excluded when determining the total cumulative

exergy of agricultural products.

The cumulative exergy consumption for importing the product demand from other localities

or other countries is also determined and compared with CExC local . The following design

rules can then be applied:

(1) If CExC local >CExC imp, import the product.

122

(2) If CExC local<CExC imp, produce the product locally if it does not have any competitors

(i.e. other products that also satisfy CExClocal<CExC imp with resource requirements

overlapping with those for the product in question). Otherwise, proceed with the

determination of its specific resource gain (per unit amount of competing resource) to

inform resource allocation. For example, if there are two food products suitable for

local product but competing for land, determine their specific resource gain for land

allocation. At this point, the methodology allows looking at this kind of interactions

between resource uses and allows early identification of trade-offs and ruling out

those options that are resource inefficient as compared to the conventional options.

6.4 Sequential synthesis of multiple subsystems

Beyond the design of individual subsystems, much of the complexity of synthesising a whole

LIPS lies in handling the connections and interactions between subsystems. This involves the

determination of the sequence for designing the individual subsystems and the handling of

inter-subsystem stream connections and resource allocation. In this section, principles for

these tasks and an iterative design procedure for the complete process synthesis are presented.

6.4.1 Synthesis sequence When designing a LIPS that involves multiple subsystems, one needs to decide the sequence

in which subsystems should be considered to make the design process more efficient, by

reducing the need for design iterations caused by subsystems’ interconnections. As a general

principle, the place of a subsystem in a design “queue” should be in accordance with the

degree to which its design is affected by the design decisions of the other subsystems. The

design of a subsystem typically depends on the following factors which are affected by the

inter-subsystem connections:

(1) Availability of required resources

(2) Cost of required resources

(3) Product demand

If a subsystem requires a resource from another subsystem, then (1) and (2) could be affected

by the latter subsystem via its capacity and efficiency respectively. If a resource is to be

shared by other subsystems, (1) will be affected due to inter-subsystem competition. If the

output of the current subsystem is input to another subsystem, (3) will be affected. For a

food-energy-water system, each of the three subsystems involves interconnections affecting

123

(1) and (2), but the food subsystem is simpler because it is not affected by (3), compared to

the other two. As such, it is more independent, and therefore can be chosen to be designed

first. However, the order between the two other subsystems is rather difficult to determine

generally, and should be decided on a case-by-case basis. In the following, the design

approach is explained assuming the water subsystem is more independent than the energy

subsystem and hence should be designed first. It should be noted that the degree of

independency of each subsystem may change, depending on the context. For example, if in a

particular system the energy generation happens to heavily depend on the waste streams from

food production, theoretically the required output from the food subsystem may be affected

by the energy subsystem, deviating from the more common case described above. In this and

other situations where there is no subsystem which is significantly more independent than the

others, one may randomly choose a subsystem as the starting point of the iterative design

process. Note that the sequence of designing the subsystems affects the effort needed for

reaching a convergence; it is not expected to change the converged design outcome.

6.4.2 Inter-subsystem resource allocation

In Section 6.3.4, a principle was presented for allocating a shared resource between multiple

purposes in a single subsystem. If a resource is potentially shared by multiple subsystems, its

availability to each subsystem needs to be updated when moving from the design of one

subsystem to another. An example of this consideration is land allocation between growing

food crops and energy crops. When the food subsystem is designed prior to the energy

subsystem, all the agricultural land is considered as available for food production. When the

energy subsystem is designed in a later step, if a technical component consuming energy

crops (e.g. biomass CHP) is identified as a preferred option, and there is no sufficient land

available for supplying energy crops due to the occupation by the food subsystem, its SRG

needs to be compared with those of the options chosen by the food subsystem design. If the

former turns out to be higher, land will be re-allocated from the food subsystem to the energy

subsystem; and the former will be re-designed (in the next iteration, see Section 6.4.3) based

on the updated land availability.

6.4.3 A sequential synthesis procedure

Implementing a determined design sequence and handling inter-subsystem connections, a

sequential synthesis procedure, illustrated in Figure 6-2, is proposed in the form of an

iterative procedure for synthesising a complete LIPS.

124

Figure 6-16: A sequential synthesis procedure

At the start of the procedure (i=0), synthesis will assume conventional utilities, without

considering their alternative sources that the local system can potentially offer. For example,

the food subsystem uses grid electricity (as opposed to power from local renewables) and

standard ground water (as opposed to rain water), as the cost information of these alternatives

is yet to be produced later when the energy and water subsystems are designed. Following the

design sequence determined in Section 6.4.1, the food subsystem is first synthesised, using

the principles stated in Section 6.3.4, in isolation from the other subsystems. Then, the water

subsystem is synthesised but considering the water demands of the food subsystem just

designed. Specific water streams that can be used preferably by the food subsystem and other

local water sinks (i.e. residential users) and their associated CExC will be determined. Lastly,

the energy subsystem is synthesised taking into account the energy sources and sinks arising

from the initial design of the food and water subsystems, to determine specific energy

streams that can be generated locally from preferred sources or obtained from centralised

supply to meet the demands.

125

Yes

No

Yes

i > 0

Terminate

Set i =i+1,use new solution for

water and energy supply parameters;

update shared resource availability

|Ci˗Ci-1|/ (Ci-1) ≤ ε

Input energy requirements

Set i =0 Assume conventional energy

and water sources

Synthesise water subsystem

Synthesise energy subsystem

Synthesise food subsystem

Input energy requirements

Input water requirements

No

Determine the local demands to be met (e.g. food, energy

and water)C i=(∑CExC food+∑CExC water+∑CExCenergy )i

Following the initial pass of design, the first iteration (i.e. i=1) is carried out. The water and

energy supply parameters from the initial pass (i.e. i=0) are used and the same design

sequence repeated. The iterative process stops when the quantity of resource streams (e.g.

rainwater, solar power) exchanged between the three subsystems and their corresponding

specific CExC become stable, and no further adjustment is needed for inter-subsystem land

allocation. At this point of convergence, the overall cumulative (external) exergy

consumption (Ci) for production of food, water and energy systems will stabilise and satisfy

the convergence criterion |Ci˗Ci-1|/ (Ci-1) ≤ ε, where ε represents the criterion for convergence.

6.5 The integration stage: resource cascading, recycling and regeneration

Following the synthesis stage, resource cascading can be carried out upon the synthesised

base system. Sirkin and Houten (1994) were among the very first to explore in depth the

concept of resource cascading and defined it as the process of optimising resource utilisation

through a sequential re-use of the remaining resource quality from previously used

commodities or substances. It can be viewed as a systematic technique for implementing

strategies to maximise resource reuse and recycling through integration of resource-using

processes (Sirkin and Houten, 1994). Resource cascading has often been applied in the

context of resource scarcity as a resource conservation technique. In our work, cascading use

of a resource refers to the multiple re-use of a resource stream, resulting in quality

degradation in each time of use without quality up-lifting between uses. Two cases of

cascading use of a resource can be identified:

(1) Same-purpose cascading use where the successive uses are based on the same trait of

the resource, i.e. the same indicator of resource quality for that purpose. For example,

satisfying water demands of water sinks by reusing water based on its COD

concentration level.

The application of cascading use of resources for a single purpose is relatively well

established and has been widely applied to chemical process integration where process units

generate or demand a flow of resource (e.g. heat and water) at various qualities (e.g.

temperature, concentration, etc.).

(2) Re-purposing where the next use of the resource is based on a trait of the resource

different from that of its previous use. For example, repurposing might involve

recovering wastewater for energy production rather than for water supply.

126

Wang and Smith (1994) define resource recycling, in the context of water, as allowing the

resource to be used in the same process that previously generated it. If the used resource is to

be an input flow in the same process that generated it and to satisfy the same purpose, the

quality of this used resource will have to be uplifted before it can be used again in the same

process. This is because the resource quality will generally degrade through its use in the

process. Thus, recycling refers to the repeated use of a resource stream in the same

application but with quality upgrading between the uses. While regeneration refers to the

quality-upgrading practice to a resource that has already experienced one or multiple uses,

recycling is a special case of regeneration where it refers to the repeated use of a resource

stream in the same application (single step use of a resource as compared to multiple uses of

a resource) but with quality upgrading between the uses. As such, regeneration and recycling

are key enablers of resource cascading by process integration in the present insight-based

design approach. This is because regeneration and recycling allow resources that would

otherwise be wasted to be used again through the same process or through different type of

processes. Both regeneration and recycling will be generally undertaken after resource

cascading.

6.5.1 Quality of a resource

The quality of a resource is pivotal to the process of cascading, recycling and regeneration of

the resource. Sirkin and Houten (1994) defined resource quality as being an expression of the

ability of the resource to serve several purposes or the same purpose repeatedly but at

different degrees of difficulty (e.g. COD of water increases after each use and water has to be

upgraded to fulfil the same purpose repeatedly). Resource quality can be expressed in several

ways including as a function of the quantity of embodied energy, the degree of structural

organization and the chemical composition of a given resource, substance or material. It can

also be described as a function of the effort required to produce or reproduce the quality.

In this work, it refers to the multiple re-use of a resource stream, resulting in quality

degradation in each time of use without quality up-lifting between uses. The quality of a

resource will be defined based on its intended purpose. Examples of the intended purpose and

quality of some resources are given in Table 6-1.

127

Table 6-4: Examples of intended purposes and indicators of quality of some resources

Resource Intended purpose Quality of resource

Heat

High pressure steam (e.g. for power generation, co-generation)

Steam at pressure 82 bars and temperature 525 °C

Medium pressure steam (e.g. for industrial purposes) Steam at pressure of 10 bars and temperature 120°C

Low pressure steam/hot water (e.g. district heating) Hot water at pressure of 1 bar and temperature 60°C

Mass (general)

Mass exchange with solvents (e.g. ethanol) % composition of ethanol in a solution

Mass exchange with reactants (e.g. hydrogen) % composition of hydrogen

Water

CleaningAgriculture (irrigation)

DrinkingDomestic (e.g. cooking, cleaning, shower)

Industrial processes

Each of the intended purpose has a maximum allowable COD or TS

content for accepting the sources. This maximum limit is set by law and

regulations.Source: Adapted from Foo (2010)

A high quality resource is usually either expensive to produce or scarce. Therefore, the design

of a series of cascading use of a certain resource generally aims to minimise resource costs by

matching the quality “grades” of the resource with the levels of demand for quality. In fact,

this is the common principle shared by the existing pinch analysis and design approaches

(e.g. Linnhoff (1993), Foo et al. (2006)). These approaches are adopted for designing the

cascading use of resources in LIPS.

Quality of land

Land is a resource that is consumed (or occupied) when it is used for a certain purpose and is

similar in nature to resources such as water and heat in that land is still available once it has

fulfilled its purpose but at a lower quality. It is distinct from resources such as nutrients

which once consumed for their intended purposes, are no more physically available for any

other purposes. There are two types of decisions to make around land: allocation and

regeneration.

The optimal allocation of land for energy production has been extensively studied by Lam et

al. (2009a), Lam et al. (2009b) and Lam et al. (2010a) who developed a regional resources

128

management composite curve (RRMCC) to graphically represent the relationship between

land use and energy production and consumption. It provides a tool for analysing the trade-

off between land use and biomass generation in a region through a single plot, aiding regional

planners in analysing the optimal land use and the management of the energy surpluses and

deficits. It uses the principle idea of a Grand Composite Curve (Townsend and Linnhoff,

1983) and adapts it to the problem of regional resource management where the horizontal

axis represents the energy supply/demand profile and the vertical axis represents the

cumulative land area for the regional area. The limitation of the developed RRMCC is that it

does not offer any indication of the quality of the land and how the land should be prioritised

between different purposes. MAFF (1988) argue that land should be prioritised for its

intended purposes based on its quality, and they propose 5 different such qualities However a

simpler categorisation is also in common usage, with land grades from 1 to 5 in decreasing

order of agricultural land quality: grade 1 is the highest quality land for agricultural purposes

and can be used to grow all types of crops while grade 5 is the poorest quality land for

agricultural purposes and is usually used for grazing. For different purposes that require the

same quality of land, the resource gain indicator can be used to allocate the land to fulfil the

purposes in decreasing order of resource gain.

However, land can be cascaded and regenerated to be used again either for the same purpose

or for a different purpose once it has fulfilled its intended purpose. The quality of the land

will usually degrade after it has been used for its first purpose. This degradation of the land

means that after time (e.g. years) its productivity for its intended use will decrease and the

land will need to be used for other purposes. It is also possible to regenerate/restore the land

either by technological or ecological/natural means. For instance, if the initial purpose of the

land was to produce wheat crops; due to land degradation associated with fulfilling this

purpose, after some years the land will need to be used for other less demanding purposes

such as the growing of energy crops that are low maintenance harvest. Land degradation can

be measured by loss in soil and decrease in crop yield as formulated by Lal (1981) through

Equation (6.4).

Y=C e−βx (6.4)

with Y is the yield in tons per hectare per year, C is the yield on un-eroded (newly cleared)

land in tons per hectare per year, β is the decline coefficient,x is the cumulative soil loss per

hectare per year.

129

The values of x and β will depend on the amount of rainfall, natural slope and soil

management. The decline coefficient varies depending on the crop and the slope. For

instance, for slopes varying from 1,5,10, 15%, β ranges from 0.002 to 0.036 for cowpea and

0.003 to 0.017 for maize (Lal, 1981). There are some existing studies on values of β and x for

different types of crops under different conditions of rainfall, natural slope and soil

management (Lal, 1981). As the yield of the crop declines, the quality of the land will also

decline (e.g. from grade 3 to grade 4) and it will become more suitable for other purposes

such as energy crops plantation.

Since more in depth research is still needed to find an appropriate universal quantity for the

quality of land and its degradation after different uses, this work will treat land as a resource

that is occupied once it is used and will focus on how best to allocate this resource to serve

different purposes such as crop plantation for human consumption or energy crop plantation

for energy production. Thus, opportunities for cascading and regeneration of land will not be

considered in this work.

At the end of a series of cascade steps, reprocessing could be undertaken to allow resources

that would otherwise be wasted to be used again through the same process or through

different type of processes. Re-processing may materialise in (i) upgrading or regeneration to

improve the quality of a used resource to enable recycling, i.e. reuse of the resource in the

same application, or (ii) any treatment needed to repurpose the resource. A review on pinch

analysis including water pinch is included in Appendix C.

Options for recycling or repurposing are ranked according to their RG, evaluated by adapting

equation (6.5):

R G=CExCavoided+CExC treatment−CExCrep (6.5)

where,

CExCavoided is the CExC of fresh or other resources avoided by recycling or repurposing;

CExC treatment is the CExC for treating waste that would arise from the used resource if it was

not reused through recycling or repurposing;

CExCrep is the CExC needed by the reprocessing to prepare for recycling or repurposing.

130

For each used resource considered, the recycling or repurposing option with the highest RG is

adopted.

6.6 Summary of the methodology for insight-based approach

Figure 6-3 summarises the different steps in the methodological framework for the insight-

based approach for designing LIPS, which combines the synthesis stage and the various

decisions in the integration stage. The application of design steps and rules can be quite

tedious for complex subsystems. In such cases, a mathematical programming model can also

be formulated to aid resource allocation and technology selection.

In order to satisfy local demands, the base design will usually result in a combination of local

production routes as well as possibility of importing. For each purpose/task, both the quality

and quantity of the resource required could be estimated based on the nature of the task and a

mass or energy balance as appropriate around the task. If the purpose allows reuse of the

resource, cascade the resource using pinch analysis methods so as to reduce the amount of

imported or fresh local resource to the system which will need to be re-assessed.

Next, identify any remaining flows of the resource that cannot be re-used. Options for

regeneration are then considered and ranked using resource gain indicator for regeneration for

all these flows/streams. The regeneration option with highest resource gain is then compared

with the re-purposing option with the highest resource gain. Thereafter, various insightful

principles can then aid decision making between implementing regeneration and re-purposing

for design. The decision for re-purposing a stream may be tied to the configuration of the

affected subsystem which will thus require the design of the whole system to be adjusted. For

example, if energy is generated from the re-purposing of a stream of wastewater, this will

affect the current design configuration of the energy subsystem and the process synthesis

stage will need to be carried out again. Note that for the case of re-purposing wastewater into

energy production, this can also be considered as regeneration as wastewater is also being

treated in the re-purposing process and can thus be used for meeting water demands. The

implementation of the regeneration option will require the re-assessment of the amount of

imported or fresh resource. If the resource in question is of a limited availability and involves

competing uses of multiple tasks, the change in the fresh resource demand of one purpose

131

Determine the local demands to be satisfied by the LIPS.

Follow the sequential synthesis procedure to generate a base design.

Identify resources that could be re-processed

Generate recycling and repurposing options

Determine the RG for each option and identify the option with the highest RG,

i.e. for each resourceUse to select option to treat stream before discharge into environment

Cascade the resources using pinch analysis

method and principles

Re-assess amount of fresh resource required

Apply recycling/re-pursing option

For each resource,is > ?

 

Any resources available for

reuse?

No

Yes

Yes

No

(due to regeneration) will affect resource allocation between different purposes, hence

requiring re-adjustment of the design. Any excess unused streams that were not regenerated,

due to poor resource gain value or the fact that source-demand has already been satisfied, will

have to be treated before being discharged into the environment. If different environmental

remediation options exist for treating the unused streams, determine their CExCenv(Chapter 3)

and choose the one with the lowest CExCenv, CExCenvmin.

Figure 6-3: Methodological framework for insight-based design approach

6.7 Case study: designing the food-energy-water system for an eco-town

This case study demonstrates the application of the insight-based approach to design a LIPS

for food, energy and water supply, specifically through the following aspects:

Using RG based design principles to generate the base design of the water and food

subsystems, and a linear programming assisted design of the energy subsystem.

Using the sequential design procedure to produce a complete whole-system base

design.

Cascading use of wastewater and waste heat within and across subsystems.

Assessing regeneration options for wastewater.

132

To simplify the case study, regeneration for waste heat source and re-purposing options for

both wastewater and waste heat were not included. As no retrofitting decision was

considered, the design principles based on the payback of capital resources were not

demonstrated. Section 6.7.1 presents the main results from the various stages of the case

study; all the detailed results, calculations and assumptions can be found in Supporting

Information. In Section 6.7.2, comparisons are made between the resource consumption of

the LIPS designed by the insight-based approach and that of a system relying on external,

centralised supply and between the results of the insight-based approach presented in this

work and those of the mathematical programming approach from our earlier work, to offer an

overall assessment.

The selected case study locale is similar to the one presented in chapter 5 but with slightly

different data and assumptions. Note that Chapter 6 was carried out after Chapter 5 and thus

some of the data and assumptions used in the former chapter have been updated. Refer to

Appendix D for all the data and assumptions taken as well as for the detailed results of the

iterative design of LIPS using LIPSOM and the sequential synthesis procedure.

6.7.1 Initial design of food subsystem

The food products considered are bread, potatoes, pork and beef and have been chosen based

on local food preferences in the eco-town. These food choices also give a good representation

of a human being’s dietary requirements in carbohydrate, protein and fats. With a total

availability of 17 ha, agricultural land is a limited resource and as such the specific resource

gain SRG of each food type, with the imported food as the reference option, was determined

using Equation (6.2). The food products with SRG values in decreasing order were bread,

potatoes, beef and pork, at 31.7×103 MJ/ha, 9.29×103 MJ/ha, 6.02×103 MJ/ha, 1.02×103

MJ/ha respectively. Using the design principle for resource allocation, growing wheat for

bread was given priority to receive land, and it turned out that all the land was to be allocated

for this purpose and 60% of the bread demand could be satisfied by local wheat; while all

other food demands needed to be imported due to limited agricultural land availability,

despite their positive RG.

133

6.7.2 Initial design of water subsystem

The initial design of the water subsystem needed to satisfy the water demand by bread

manufacture from the food subsystem and the residential sector, with groundwater and

rainwater as the sources available. The potential uses of rainwater at this stage were limited to

wheat cultivation and non-potable domestic water uses. Using Equation (6.1) and taking

groundwater as the reference resource, the RG for rainwater was determined to be -0.048

MJ/kg. The negative RG was due to the relatively high capital resources to implement a local

rainwater harvesting system leading to groundwater being overall a more resource efficient

alternative for water supply. Furthermore, groundwater has also an abstraction limit in the

eco-town which is sufficiently high (Whitehill and Bordon, 2012); therefore its use was not

limited by its availability.

6.7.3 Initial design of energy subsystem

Alternative electricity generation sources from wood chip biomass CHP, organic waste CHP,

natural gas CHP, solar and wind and heat generation from wood chip biomass boiler, wood

chip biomass CHP, organic waste CHP and natural gas CHP were considered alongside grid

electricity and heat from natural gas boilers as conventional energy sources. Agricultural

residues produced from the initial design of the food subsystem are available in summer for

energy production, which is assumed to merge with organic waste from municipal operations

to feed into a CHP facility. As using the RG based principles to design this subsystem would

be too complex given the number of options to be considered, linear programming (LP) was

adopted to generate a fast optimum design. The objective function of the LP problem was to

minimise the net total CExC of this subsystem while meeting local heat and electricity

demands, allowing the generation of surplus electricity if a CHP facility was selected.

When the difference between the CExC of the locally generated surplus electricity and the

CExC of grid electricity was taken as resource credit in the objective function, the optimal

design of the energy subsystem suggested to produce most of its electricity output from wood

chip biomass CHP (35,000 MWh), followed by wind power (14,000 MWh), solar (12,000

MWh) and organic waste CHP (9500 MWh). Encouraged by the resource credit assigned to

the exported electricity, virtually all the power generated by CHP was surplus. The total heat

demand for the eco-town was met fully by the wood chip biomass CHP (91,000 MWh) and

organic waste CHP (14,000 MWh). This electricity mix corresponds to an average specific

CExC of 1.90 MJ/MJ electricity; a significant 68% decrease from grid electricity. The

134

average specific CExC of supplying heat was determined to be about 1.80MJ/MJ heat; which

corresponds to a 10% decrease from that of supplying heat from conventional natural gas

boilers. Furthermore, taking agricultural residues as feed for energy production resulted in a

reduction of the total CExC required by 4% compared to not considering the residues.

If the resource credit of producing surplus power was not included in the objective function

of the LP energy model, the energy supply for satisfying heat demand includes 13% organic

waste CHP, 9.5% biomass CHP and 77.6% from wood chip biomass boiler while the

electricity mix would comprise 50% wind, 14% wood chip biomass CHP and 36% organic

waste CHP with no surplus electricity generated. The average CExC was determined to be

540,150 GJ/y compared to the average CExC of -41,361 GJ/y in the initial energy production

subsystem; indicating that the inclusion of a credit for avoiding the CExC associated with

grid electricity through local electricity export decreases resource consumption significantly;

far offsetting the amount of resources spent to produce energy locally. The use of agricultural

residues did not also have any noticeable impact on the total CExC of such an energy model.

6.7.4 Iterative design

Following the initial pass of the sequential synthesis procedure, further design iterations were

carried out. Table 6-2 shows the outcome of the 1st iteration, along with the insights gained

from the change in the design outcome resulting from this iteration. The 2nd iteration was

subsequently conducted, which did not alter the selection of design options, but the 15%

decrease in energy consumption for groundwater use from the 1st iteration led to the change

in the RG ranking of the food products, with beef now being more efficient to produce than

potatoes, which suggests that beef production cost is rather sensitive to water supply. Also,

although there were changes in the energy demand from the food and water subsystems, these

changes were not significant enough to modify the optimal energy mix supplied by the

energy subsystem, suggesting the overall design was becoming stable. In fact, the

convergence criterion was met following the 3rd iteration, marking the completion of the

synthesis stage. The final base design is illustrated in Figure 6-4.

Table 6-2: Outcome of 1st Iteration

Subsystem Design Outcome InsightsFood -Bread still has the highest RG at

4.63×104 MJ/ha followed by pork at 1.16×104 MJ/ha, potatoes at 9.29×103

-With similar water but cheaper energy supply to its initial design, pork becomes the second

cheapest food type to produce locally;

135

MJ/ha and beef at 3.63×103 MJ/ha.- 10% and 68% decrease in specific

CExC for heat and electricity respectively.

indicating that pork is very sensitive to resource cost for energy supply.

-Between beef and pork, the local production of the latter would still consume more energy than the former despite the change in energy

cost per unit (kg), as the quantity of pork produced locally based on land available is

higher than beef, with a pork to land ratio of 1.2 ha/tonne compared 4.7 ha/tonne for beef.

Water

-Specific CExC groundwater was 0.051MJ/kg (a 15% decrease) while RG

for rainwater was -0.06 MJ/kg.- New water demand and wastewater

generation from energy production from i=0

-More resource efficient using groundwater to satisfy all water demands in the eco-town

rather than using rainwater due to negative RG and relatively high abstraction limit for

groundwater.

Energy

-Electricity demand supplied by 50.1% wind, 43.3% solar and 6.5% wood chip

biomass CHP.-Average specific CExC of 2.02 MJ/MJ

electricity; 6.5% higher than in i=0-Same heat source mix as in i=0

- Contribution of wood chip biomass CHP, which has a relatively high specific CExC, in

the electricity mix increased by 3.5% compared to in i=0 due to the need to satisfy

higher energy demands.

136

Food wastewatertreatment

Groundwatertreatment

  Residential

Wheat cultivation

Residential wastewater treatment

 

Wheat

Wastewater 

Water production subsystem  

Food production subsystem  

Wheat processing

 

Biomass CHP 

Bread 

   

Energy wastewatertreatment

Organic waste CHP

 Solar

 

Imported fertiliser

 

Wastewater 

Water 

Water 

Surplus electricity to grid 

Energy production subsystem  

Wind 

Electricity  

Wastewater 

Imported food 

Discharge 

Discharge 

Discharge 

Agricultural residues 

Figure 6-4: Base design of local production system

6.7.5 Integration: water reuse and regeneration

After the base design was generated, the system was optimised by considering integration

options for resource reuse. As all the water sources considered had a COD level much higher

than the COD requirements of any water sinks, from a quality perspective the cascade use of

the water sources through the application of pinch analysis would result in no possible

recovery. As direct re-use was infeasible, options of regeneration (to enable reuse) were

evaluated. The RG for regenerating different water sources of residential wastewater,

wastewater from food production, wastewater from energy production up to the desired

quality of the water sinks were assessed using Equation (6.3). Taking into account resource

cost for environmental discharge of un-regenerated wastewater and the cost for fresh water

avoidable by using regenerated water, the net specific CExC for regeneration was determined

to be positive, hence supporting the regeneration option. The use of regenerated wastewater

was limited to wheat cultivation, non-potable residential purposes and energy production, but

not for food processing, in line with health and safety regulations (UN Water, 2013). It was

determined that about 61% of the eco-town’s water demands could be satisfied by

regenerated water sources, hence significantly reducing the consumption of groundwater. The

average specific CExC of water supply was reduced by 60% from its original value in base

design. The reduction in total CExC of the water subsystem did not impact the design

decisions of the energy subsystem as water was not a significant component (less than 1%) of

the specific CExC of the energy options. Any remaining unused streams were to be treated

before environmental discharge.

6.7.6 Integration: energy reuse

Pinch analysis (Linnhoff and Hindmarsh, 1983) was applied to optimise the use of heat

available in each season. Low temperature (LT) waste heat available from organic waste CHP

and wood chip biomass CHP was candidate for reuse to meet heat demands from industrial

bread production, wheat storage, wastewater treatment plant and residential. It was

determined that adopting 3 heat exchangers placed above the pinch would allow for

maximum heat recovery of 2.79×107 MJ/y. The reuse of LT waste heat contributes to

137

Heat 

satisfying about 10% of the total heat demand. The proportion of high and medium

temperature heat produced from wood chip and organic waste CHPs in the heat energy

supply mix decrease from 87% and 13% to 78.4% and 11.7% respectively. With this new

heat energy supply mix, the average specific CExC of heat was reduced by 10% from its

value from the base design. However, the new specific CExC for heat coupled with cheaper

water supply did not alter the order of the SPG for the different food options, though SPG for

pork followed very closely that of bread. Similarly, the design of the water subsystem was

not changed.

Together, water and energy reuse design showed that the decisions in the integration stage

did not lead to a qualitatively different design from the synthesis stage in this particular case.

However, the base design was made more efficient, through the reduction in energy

production required from the CHPs and the consumption of groundwater.

6.7.7 Comparative analysis and final assessment

To show the extent to which a LIPS, designed following the insight-based approach, can

achieve resource savings, the results presented in Section 6.7.6 is compared with the resource

consumption for meeting the same local demands by another two scenarios: “centralised

supply”, which imports food and utility (grid electricity and natural gas) and ‘design-silo’

which involves designing each scenario independently with no exchange of resources across

the subsystems. Figure 6-5 illustrates the external resource consumption of each subsystem

for each scenario and clearly demonstrates the resource advantages of the LIPS. When

considering the credit of surplus electricity, the LIPS consumes less than 10% and 90% of

resources (measured in CExC) needed by the centralised supply and design in-silo scenarios

respectively; a 39% and 12% saving is still achieved from both scenarios even without the

aforementioned credit.

The total external CExC of the food and water subsystems of the integrated design were

lower across all the scenarios considered. However, since the energy subsystem of the

integrated design (ID) has to meet a greater demand (energy demand for wastewater

treatment, residential and food production), its net CExC was found to be 36% higher than

that for the design in-silo. Compared to the latter scenario, the main resource reduction in the

integrated design was driven by cheaper locally produced bread which uses energy produced

locally while the design in-silo uses imported energy and treats any food wastewater within

138

the food subsystem. In addition, water is also supplied and treated at cheaper resource costs

using locally available energy while in design in-silo, although the demand to be satisfied

was lower (residential water and treatment of residential wastewater) and treated residential

wastewater supplied about 60% of water demand based on LIPSOM, the use of imported

energy sources made its water subsystem overall 32% less resource efficient.

Food subsystem Water subsystem Energy subsystem Electricity export Total net resource consumption

-1500000

-1000000

-500000

0

500000

1000000

1500000

2000000

External CExC for each scenario using insight-based approach

Integrated design

Centralised supply

Design in silos

Cumulative exergy (GJ/y)

Figure 6-5: External CExC for each scenario using insight-based approach

The result of the insight-based design approach was also compared to that of the

mathematical programming (MP) approach developed earlier (Leung Pah Hang et al., 2016b),

which was adapted to include all assumptions and options for food, water and energy

subsystems considered in insight-based approach. The comparative analysis between the two

design approaches based on external CExC is shown in Figure 6-6. The overall resource

consumptions of the two designs are similar, with the result of the MP approach being

slightly better. A closer inspection of the design decisions by the MP approach (see SI)

revealed that both approaches suggested qualitatively identical designs in terms of the food

product and technical options selected for local production. Quantitatively, both designs

suggested the same decision for the food subsystem, i.e. using all the land to meet 60% of

bread demand. There were minor differences in the percentage of groundwater replacement

by regenerated wastewater and in the energy mix proportions, and there was a noticeable

(6%) increase in waste heat recovery identified by the MP approach. These quantitative gains

139

by the MP approach are not surprising, given its adoption of rigorous and simultaneous

mathematical optimisation.

Food subsystem Water subsystem Energy subsystem Electricity export Total net resource consumption

-1000000

-800000

-600000

-400000

-200000

0

200000

400000

600000

800000

1000000

External CExC for insight-based and simultaneous approaches

Insight-basedSimultaneous

Cumulative exergy (GJ/y)

Figure 6-6: External CExC for all subsystems of insight-based and simultaneous approaches

6.7.8 Summary of insight-based approach

Overall, it can thus be inferred that the insight-based approach offers sensible and comparable

design solutions in this case study. Compared to the MP approach, the largely design rules

based and incremental nature of the insight-based approach makes it easier for the decision

makers to use, by allowing them to keep control of the decision process and enrich their

understanding of the implications of various design options and their interactions, particularly

those across different subsystems. Therefore, it offers an effective approach for practitioners

to discover the superior designs of an integrated local production system for supplying food,

energy and water to achieve significant resource savings than relying on centralised supply,

through rational utilisation of local renewables and resource sharing and exchange between

different local production processes. The adoption of the approach, however, will require the

overcoming of practical limitations in data availability, an issue to be resolved by the

combination of literature data sources (particularly for cumulative exergy consumption) and

locale-specific data gathering.

140

Chapter 7: Robustness analysis and robust design of LIPS under uncertainties

7.1 Rationales for designing LIPS under uncertainties and type of uncertainties

It is most important to manage the uncertainties in design as high level of uncertainties

complicates the assessment of different design alternatives for decision making by

practitioners. This generates the need for tools that can assess the robustness of a given

design decision by evaluating which uncertainties might have major impacts on the system

design and also to generate solutions that are robust to them. Handling uncertainties is a

common challenge for design and the purpose of this chapter is simply to show how the

existing approaches can be applied to the design of local production systems through an

illustration on the food production subsystem.

The design of a system can involve broadly two types of uncertainties namely

information/factual and operational uncertainties. Information uncertainties can arise due to

imprecise measurements, average or out-dated data using proxies and incomplete data and

several assumptions such as linear correlations and averaged data over time and across

regions for estimating the value of the data (Sadhukhan et al., 2014). An example of

information uncertainty could be the specific cumulative exergy value of imported electricity

from the grid. The uncertainties embedded in this value will remain the same at any point in

time. On the other hand, operational uncertainties are related to the uncertainties that might

happen in the future; such as after the design decision has already been made and

implemented. For instance, operational uncertainties could involve changes in supply due to

severe weather conditions that can affect crop yields, changes in demand of a particular

product and changes in technical efficiency and cost of a technology due to technology

learning and advancement.

7.2 Approaches to handling uncertainties in design

The different approaches that can be used to systematically address uncertainties in design

are presented in this chapter. These approaches have been adapted from previous literature

review and will be tested through a case study on the design of a localised food production

system in the UK using mathematical optimisation. Overall, two fundamentally distinct ways

for addressing uncertainties in design have been proposed in this chapter as follows:

141

(1) Post-design uncertainty assessment

(2) Uncertainty-embedded design

Post design uncertainty assessment involves assessing the robustness of a given design to

uncertainties and evaluates the impact of uncertainties to a design decision which has already

been made. Most recent studies include the robustness to demand variations of a biomass

processing network for biofuel production (Kim et al, 2011) and a robust optimisation

approach to closed-loop supply chain network (Pishvaee et al., 2011) and developing a new

robust optimisation approach for integrated multi-echelon, multi-product, multi-period supply

chain network design under process uncertainty (Akbari and Karimi, 2015).

Uncertainty-embedded design attempts at optimising the system design given a set of

uncertainties so as to produce a robust design, i.e. a design decision which best can cope with

the uncertainties. This approach is based on two-stage stochastic programming technique and

has been used in many recent studies such as the two-stage stochastic programming model

developed by Zhou et al. (2013) for the optimal design of distributed energy systems and the

two-stage stochastic programming of a supply chain model for the production of biodiesel

through wastewater treatment.

Due to the nature of these two different tasks, different approaches are required to handle

these design uncertainties. The post-design uncertainty assessment could be done following

any of the two design approaches developed in chapters 5 and 6 of this thesis while the

uncertainty-embedded design is a variation of the mathematical programming based

approach.

7.3 Methodology for addressing uncertainties in design of LIPS

The overall methodological framework adopted in this research work to address uncertainties

in the design of LIPS is illustrated in Figure 7-1. It illustrates how based on the uncertainty

task, different systematic approaches on robustness analysis and stochastic programming, can

be used to either assess the performance of a certain system design with uncertainties which

in turn could provide feedback possibly leading to the alternation of the original design (not

discussed in this work) or produce a robust design given uncertainties in the system.

Decision variables are defined as a set of quantities that needs to be determined in order to

solve the problem. The variables can typically represent the amount of resources to use or the

level of some activity (FRM, 1998). For example, a variable might represent the hectares of

142

land devoted to crop plantation or the size of a generator to install in a power station.

Decision variables can be considered within specific time frame of the design. To exemplify

this, a time frame of one year can be considered for agriculture and a time frame in the range

of 15-20 years can be generally considered for industrial systems.

Within the considered time frame, the decision variables can be further divided into two types

namely fixed and flexible decision variables. The fixed decision variables are those variables

that once determined, have to remain fixed during the operational period assumed by the

design as it will not be sensible or realistic to change their values before the next round of

design is considered. Examples include the land area dedicated to crop production or

industrial activities and the type and capacity of equipment to install in an industrial plant. In

comparison, flexible variables are those variables that can be adjusted throughout the

operational phase. For instance, decisions on how much of a certain type of food to import

can be adjusted to cope with a sudden decrease in local crop yield. The operational load of a

plant, such as steam and electricity production, can also be adjusted to cope with changes in

demand.

143

Design Problem Statement

Yes

Presence of factual uncertainties?

Select how to address the uncertainties in design

Terminate

Yes

Robustness analysis results

Post design uncertainty assessment -Assessing the

robustness of a given design to uncertainties

No

Uncertainty-embedded design -Given uncertainties,

produce a robust design

Perform nominal deterministic system design

Build key scenarios based on factual uncertainties

Run scenario based simulations

Presence of operational uncertainties?

No

Presence of fixed decision variables?

No Terminate

Run scenario based simulation -optimisation

Build key scenarios based on operational uncertainties

Perform two-stage programming considering all uncertainties

Robust design

Yes

Figure 7-1: Methodological framework for addressing uncertainties in design

Post-design uncertainty assessment, further described in section 7.3.1 evaluates the

robustness of a given design where the basis of the design is subject to known uncertainties

which are represented through the definition of a set of scenarios. Two different types of

scenario based robustness analysis could be undertaken depending on the type of

uncertainties and decision variables present as detailed in the next section. As illustrated in

Figure 7-1, scenario based simulations are used to test how the variation in the uncertain

factual information could affect the performance of a design which already produced with

fully determined fixed and flexible decision variables. In contrast, scenario based simulations

with operational re-optimisation are performed for operational uncertainties. This type of

analysis is needed only when the design contains both fixed and flexible decision variables,

where the difference in re-determined flexible decision variables in different scenarios can

indicate the varying consequences of adopting the (previously determined) fixed decision

variables between these scenarios, an insight not available from the original deterministic

design.

Embedded uncertainty design, as an alternative to post-design uncertainty assessment, aims

to produce a robust design to incorporate the knowledge about uncertainties during the design

process, can be implemented using two-stage stochastic programming, as further described in

section 7.3.2.

7.3.1 Post-design uncertainty assessment

Post-design uncertainty assessment can be used to assess the robustness of a given design to

uncertainties. Based on the types of uncertainties and decision variables present in the system

design, two main post-design uncertainty assessments can be undertaken. The first step for

both types of post-design uncertainty assessment is to undertake a nominal deterministic

design whereby all the parameters are assumed to take their nominal values. The result of the

nominal deterministic design will suggest values for both fixed and flexible decision

variables. When this design is implemented, fixed decision variables remain fixed throughout

the assumed operational period under all circumstances. However, practitioners might

realistically adjust the flexible decision variables to cope with operational disturbances, i.e.

operational uncertainties. Such adjustments might thus lead to a change in the value of the

objective function of the nominal deterministic design.

144

The impact of factual uncertainties on a system implemented according to a certain design

can be evaluated by running scenario-based simulations based on the dominant parameters

with factual uncertainties. No adjustment to the operational variables need to be triggered and

the simulations will simply involve re-determining the value of the objective function based

on the values of all the decision variables fixed by the original design; i.e. the simulations

will be done based on original values of fixed and flexible decision variables from the

nominal deterministic design. An alternative to running a fixed set of scenarios is to run a

Monte-Carlo simulation analysis.

To evaluate the impact of operational uncertainties on the performance of a nominal

deterministic design, key scenarios are defined according to the operational uncertainties and

these scenarios are re-optimised in order to determine the optimal adjustment to the flexible

variables (e.g. amount of potatoes to import) given a certain operational disturbance (e.g.

drop in local potato yield) while keeping constant the fixed variables as they are assumed to

be fixed during the full operational period (e.g. land for growing local potatoes for this

particular year). The outcome of this partial re-optimisation of these scenarios gives insights

on how the operational uncertainties will affect the actual performance of the system

designed according to the fixed decision variables determined by the nominal deterministic

design. Similarly, one alternative to running a fixed set of scenarios is also to run a Monte-

Carlo simulation analysis.

As mentioned earlier in this section, the first step for the post-design uncertainty assessment

is to undertake a nominal deterministic design whereby all the parameters are assumed to take

their nominal values. A sensitivity analysis can then be performed to determine which

parameters with an embedded uncertainty influence more the value of objective function

while keeping the values of the decision variables of the nominal design fixed. Each

parameter is varied individually at selected points within a range of values for the parameters.

The dominant parameters, defined as those with the highest contributions to both the positive

and negative deviations of the objective function from the nominal scenario, are then

selected. The number of dominant parameters, m, are then combined to form 2m scenarios

which are constructed based on possible combination of the variations of the parameters

within their high and low ends of their expected range of variation (Kim et al., 2011). This

results in 2m simulations where any changes in the original nominal deterministic design are

noted.

145

7.3.2 Uncertainty-embedded design through two-stage stochastic programming

The outcome of the 2nd approach is a robust design and yields a design decision which can

best cope with uncertainties. This approach involves the use of two-stage stochastic

programming. Two-stage stochastic programming has been widely used to model

uncertainties (Birge and Louveaux, 2011). Deterministic optimisation is formulated with

known parameters. A robust optimisation such as stochastic programming takes into account

that some parameters are known only within certain bounds which can be defined by their

probability distribution. The aim of the robust optimisation is to find a solution that will be

both feasible and optimal for all such data.

A standard two-stage stochastic programming model consists of decision variables that are

divided into two groups; first stage and second stage variables. First stage variables are

decided upon before the actual realisation of the random parameters. Once the uncertain

events unfold, further operational adjustments can be made to the design through values of

the second-stage or alternatively referred to as recourse variables (Al-Qahtani and Elkamel,

2010).

Stochastic programming has become a significant problem area. With current standard off-

the-shelf software including modelling systems such as AMPL and GAMS, powerful large-

scale general-purpose solvers such as CPLEX and specialised stochastic programming

solvers namely OSL-SE, EMP: DE, LINDO and DECIS, users can develop realistic

stochastic programming models and solve them using standard desktop (Kalvelagen, 2003).

This report will focus on solving two-stage stochastic programming in GAMS using the

DECIS solver. DECIS solver was chosen because it has been used successfully for the

solution of a variety of very large problems, is easy to use and suits the purpose of the

stochastic model presented in this chapter as it is an established solver for solving programs,

which include parameters (coefficients and right-hand sides) that are not known with

certainty, but are assumed to be known by their probability distribution (Infanger, 1997).

The steps in the two-stage stochastic programming using DECIS solver in GAMS are adapted

from Infanger (1997) and can be briefly summarised as follows:

1. The deterministic core model

2. Specify the decision stages

3. Specify the distribution of the uncertain parameters

4. Set DECIS as the solver to be used for the optimisation.

146

The first step involves constructing a deterministic model. The deterministic model is

extended to stochastic model by firstly specifying the decision stages. The variables and

constraints belonging to the first and second stage need to be specified. Next, the independent

random variables in the stochastic model are specified. DECIS solver works only with

discrete variables and any continuous distributions has to be approximated by discrete

distributions. The framework for using DECIS in GAMS specifies the set stoch for labelling

outcome named "out" and probability named "pro" of each independent random parameter.

The stochastic parameters of the model are defined by writing a file, the GAMS stochastic

file, using the put facility of GAMS.

DECIS can be used when the coefficient and RHS parameters are not known with certainty

and that they assume a probability distribution. DECIS can be used in two ways:

The optimization mode for solving stochastic problems

The evaluation mode for evaluating a given solution for the stochastic problem.

There are also 4 approaches to solving by DECIS (Infanger, 1997):

Universe problem

Expected value problem

Monte Carlo Sampling using Benders decomposition algorithm

Monte Carlo Pre-Sampling using Crude Monte Carlo only

The chosen approach can be adopted by changing the strategy in the parameter file. The

universe problem approach solves all the possible outcomes and solves the corresponding

problem exactly using the Benders decomposition algorithm (Infanger, 2007). This approach

is not always feasible as there might be too many possible realisations. The expected value

problem approach involves replacing the stochastic parameters by their expected values. It

can be used as a benchmark to compare the solution obtained from solving the stochastic

problem and it also gives a good starting point for solving the stochastic problem.

The Monte Carlo Sampling using Benders decomposition algorithm is used mainly when the

possible realisations are too large to be solved by the Universe problem. In this approach,

DECIS does not determine the expected cost and the coefficients and the RHS of the Benders

cuts exactly. It estimates an independent sample from the distribution of random parameters

147

from each number of iterations. In comparison to the Monte Carlo Sampling approach, the

Monte Carlo Pre-Sampling takes a random sample from the distribution of the random

parameters and then generates the approximate stochastic problem defined by the sample

instead of using Monte Carlo sampling in each number of iteration of the decomposition

(Infanger, 2007). The default sampling size when using Monte Carlo pre-sampling is 100.

7.4 Case study on design of local food production system

A food production system can be described as a system of inter-connected food production

processes including agricultural activities and industrial food processing. The primary

sources of food are the agricultural activities which can be mainly interlinked synergistically

by exchange of nutrient flows. Therefore, the design of a food production system is based

essentially on the integration of food production processes that could use the nutrient

resources in an efficient manner. The aim of designing such kind of food systems comprises

selecting the food production processes as well as determining the exchange flow rate

between the processes.

The design problem for the food production system is thus formulated as follows:

A given set of final local food demands (d= 1, 2…Ndemands) with total flow rate Fd is to be

supplied by food production processes (e.g. crop cultivation and processing, animal breeding

and processing) which are sinks (j= 1, 2...Nsinks) consuming nutrients from available sources

(i=1, 2…Nsources). A source can be a nutrient flow from the food production processes (e.g.

organic manure and agricultural residues) with total flow rate F i which can be exchanged

between those processes at flow rate Fi,j. A nutrient flow can also be imported to supplement

the locally available nutrient sources (e.g. imported fertilisers). Similarly, food can be

imported to supplement locally produced food. Due to the constraint of land availability, the

land dedicated by the producer to crop and livestock production has to be determined.

The objective is to design the food production system by firstly deciding on how much land

to devote for each of crop and livestock production, then selecting the food production

processes and determining the flow rates from source to sink that will minimise total resource

consumption while observing a set of constraints for satisfying local human food demands.

Therefore, the design questions to be answered are:

148

1) What amount of land to devote for crop and livestock production?

2) Which food production processes to include in the food production system?

3) Are there any possible exchange of flows from locally available source to sink, Fi,j? If

so, what should be the flow rate of these exchange flows?

7.4.1 Mathematical model for deterministic design of local food production system

The optimisation problem for the deterministic design of a food production system based on

minimisation of resource consumption has been formulated in Equation (7.1). Compared to

section 5.3.1 in Chapter 5, the food models presented in this section have been further

simplified and do not include seasonality and crop storage capabilities for the purpose of

illustrating the approaches to handling uncertainties in design.

Minimise the objective function for deterministic design of food subsystem:

TEF=∑d D

edimp Fd

imp+∑j J∑i ’ I ’

e i ’ , jimp N i ’ , j

imp +∑j J∑oO

eo , jU o , jimp+(7.1)

TEF is the total cumulative exergy consumption for the food production subsystem, edimp is the

specific cumulative exergy of imported food d, Fdimp the amount of imported food d, e i ’ , j

imp the

specific cumulative exergy of imported nutrient flow i’ to sink j, N i ’ , jimp the amount of imported

nutrient flow i’ to sink j, eu , j the specific cumulative exergy of operating flow o to sink j,U o , jimp

the amount of operating flow o to sink j.

s.t.

1) Final food demand balance Fd

imp+Fdlocal=Fd

dem(7.2)

Fdlocal is the amount of locally produced food d, Fd

imp the amount of imported food d and Fddem

the demand of food d.

Fdlive is the amount of locally produced food d from livestock l and can be determined from

the yield of livestock,y l, the conversion factor cf l ,dfrom livestock l to food d, and the land use

for livestock production Llas given by Equation (7.3):

Fdlive=Ll y lcf l ,d ∀ l L (7.3)

The amount of crop c to be locally produced in any season, Ac, can be determined through

Equation (7.4).

149

Ac =Lc y c ∀ cC (7.4)

with Lc being the land use for crop production and yc the crop yield.

The amount of locally produced food product d from a particular crop, Fdcrop, can be

determined through the amount of the locally produced crop c,WCc , and the conversion

factor cbc ,d from crop to food product as shown in Equation (7.5).

Fdcrop=cbc , dWC c∀ cC (7.5)

2) Land availability constraint

The land occupied by livestockLl and cropsLc must not exceed the total amount of

agricultural land available Lagrias given in Equation (7.6).

∑c C

Lc+∑l L

Ll ≤ Lagri (7.6)

3) Nutrient requirement for crop and livestock

The sum of the imported and locally produced nutrient flows (denoted by i) should be equal

to the total nutrient demand of each sink j as shown in the nutrient balance for crops and

livestock in Equation (7.7).

∑i ' I '

N i' , jimp+∑

i ' ' I ' 'N i ' ' , j

local=N jdem∀ j J (7.7)

with N i ' ' , jlocal being the amount of locally produced nutrient from source i’’ (e.g. from crop

residues or manure) and N jdem the demand of sink j. N i ' ' , j

local can be determined through Equation

(7.8).

∑j J

N i' ' , jlocal¿nci } {A} rsub {ag} {RA} rsub {ag} {H} rsub {i ∀ i I , ag AG, ℜℜ(7.8)

with nc i ' ' being the nutrient content of locally produced nutrient i’’, Aag the amount of

agricultural commodity (i.e. crop or livestock) ag produced locally, RAag the ratio of amount

of residues or manure generated per unit output of ag and H i ' 'the harvest recovery rate of

150

locally produced nutrient i’’ taking into account that some residues need to be left in the field

to maintain the nutrient soil balance.

The results of the deterministic design of the local food production system done in GAMS are

reported in Table 7-1. The input and output flow rates are on a per year basis.

Table 7-1: Results of deterministic design of food production system

Category Value UnitObjective function (resource consumption in

terms of exergy) 1.04×108 MJ/y

Amount of imported bread 1.61×105 kg/yAmount of imported beef 8.76×104 kg/y

Amount of imported pork meat 4.63×104 kg/yAmount of imported fertiliser for bread 1.67×103 kg/y

Amount of imported fertiliser for potatoes 1.61×103 kg/yNutrients supplied for livestock production 0 kg/y

Amount of bread produced locally 6.33×104 kg/yAmount of potatoes produced locally 4.03×105 kg/y

Amount of beef produced locally 0 kg/yAmount of pork meat produced locally 0 kg/yAmount of wheat crop produced locally 5.56×104 kg/y

Amount of potatoes crop produced locally 4.03×105 kg/yLand dedicated to wheat production 8.05 ha

Land dedicated to potatoes production 8.95 ha

7.4.2 Post-design uncertainty assessment of food production system design

A post-design uncertainty assessment for both the factual and operational uncertainties

present in the design of the local food production system. The 8 key factual parameters used

in the design are listed as follows:

(1) Specific cumulative exergy of imported bread (SCB)

(2) Specific cumulative exergy of imported potatoes (SCP)

(3) Specific cumulative exergy of imported beef (SCBE)

(4) Specific cumulative exergy of imported pork meat (SCPM)

(5) Specific cumulative exergy consumption factor of utilities per amount of processed bread

(SCFB)

(6) Specific cumulative exergy consumption factor of utilities per amount of processed potatoes

(SCFP)

(7) Specific cumulative exergy consumption factor of utilities per amount of processed beef

(SCFBE)

(8) Specific cumulative exergy consumption factor of utilities per amount of processed pork meat

(SCFPM)

151

A sensitivity analysis was performed on the 8 identified factual parameters and the change in

the objective function value using discrete values of the parameters at the nominal plus or

minus given percentage changes (−50%, −30%, −10%, 10%, 30%, and 50%) was noted.

Figure 7-2 shows the results of the sensitivity analysis.

-30% -10% 0% 10% 30% 50%60000000

70000000

80000000

90000000

100000000

110000000

120000000

130000000

140000000

150000000

Variation in objective function with uncertainties

SCBSCPSCBESCPMSCFBSCFPSCFBESCFPM

% variation

Objective function MJ/y

Figure 7-2: Variation in objective function with uncertainties

The 3 most dominant parameters which trigger a significant change in the value of the

objective function in decreasing order of impact were found to be specific cumulative exergy

of imported beef (SCBE), specific cumulative exergy of imported bread (SCB) and specific

cumulative exergy of imported pork meat (SCPM). These 3 parameters form the basis for the

scenario set: S = [SCBE, SCB, SCPM]. The number of scenarios generated is 23; which gives 8

scenarios. Each scenario is created by varying the parameters by ±20%; an assumed range

which could be adapted for any specific design where the upper and lower bounds of the

uncertain parameters are known either through experience or historical data. The scenarios

are then numbered using a binary encoding scheme adapted from Kim et al (2011) with the

number given by ∑i

Bi× 2i where i refers to the position of the parameter in the list S in order

andBi =1 if the parameter is at +20% and 0 if at -20%. The possible scenarios based on the

dominant parameters with factual uncertainties are illustrated through the following matrix:

152

Possible scenarios = (0 0 00 0 10 1 01 0 01 0 11 1 00 1 11 1 1

)The next step involves running the 8 simulations based on the possible scenarios while

keeping all the design decision variables (i.e. fixed and flexible decision variables)

determined through the original nominal deterministic design. The results of the robustness

analysis are given in Table 7-2 and illustrated ion Figure 7-3.

Table 7-2: Results of robustness analysis

Scenario Scenario type Objective function (MJ/year)S0 (Nominal design) - 1.04×108

S1 000 8.50×107

S2 001 9.04×107

S3 010 9.23×107

S4 100 1.11×108

S5 101 1.16×108

S6 110 1.18×108

S7 011 9.77×107

S8 111 1.24×108

S1 S2 S3 S4 S5 S6 S7 S880,000,000.0085,000,000.0090,000,000.0095,000,000.00

100,000,000.00105,000,000.00110,000,000.00115,000,000.00120,000,000.00125,000,000.00130,000,000.00

Robustness analysis

Scenario

Objective function (MJ/y)

Figure 7-3: Robustness analysis

Table 7-2 and Figure 7-3 show how the value of the objective function changes with the

different possible scenarios arising from a combination of possible values for the parameters

with factual uncertainties. The robustness analysis of the nominal design based on the

dominant factual uncertainties reveals that the value of the objective function varies within -

153

18% to 18%. In the worst case scenario (S8) the value of the objective function was

determined to be 1.24×108 MJ/y while in the best case scenario (S1) the objective function

was found to be 8.50×107 MJ/y.

A Monte-Carlo simulation was undertaken in Excel as an alternative to running a fixed set of

scenarios. The results of the Monte-Carlo simulation with a sample size of 5000 and a 95%

confidence interval for the mean are summarised in Figure 7-4 with a histogram and

cumulative probability curve.

Figure 7-4: Monte-Carlo Simulation results

Figure 7-4 illustrates the possible values of the objective function (i.e. bins) and their

frequency of occurrence (i.e. count). The mean value of the objective function was

determined to be 104,251,671 MJ/y with a standard deviation of 7,891,596 MJ/y and a mean

standard error of 111,604 MJ/y. The median value was found to be 104,200,508 MJ/y and the

minimum and maximum values of the objective function were determined to be 86,024,751

MJ/y and 122,266,037 MJ/y.

7.4.3 Embedded design uncertainty

An embedded design uncertainty robustness analysis was also performed to assess the impact

of operational uncertainties on the nominal food production system design. Due to the

154

presence of fixed decision variables in the design, scenario based simulations with partial

optimisation of the flexible variables were undertaken. Similar to assessing the robustness of

the nominal design against factual uncertainties, the first step was to construct the scenarios

based on the dominant operational uncertainties. Three key operational uncertainties namely

final food demand by the local population (FDF), local crop yield (YC) and yield of livestock

(YL) were identified. The 3 dominant parameters form the basis for the scenario set: S =

[FDF, YC, YL]. The number of scenarios generated is 23; which gives 8 scenarios. Each

scenario is created by varying the parameters by an assumed range of ±20%.

The possible scenarios based on the dominant parameters with operational uncertainties are

illustrated through the following matrix:

Possible scenarios = (0 0 00 0 10 1 01 0 01 0 11 1 00 1 11 1 1

)The 8 scenarios (S1 to S8) simulations with partial optimisation are then run in GAMS and

the fixed decision variables are kept constant while the flexible variables are allowed to re-

optimise to cope with the operational uncertainties. The fixed and flexible decision variables

for the design of the food production system are summarised in Table 7-3.

Table 7-3: Fixed and flexible decision variables in the food production system

Fixed decision variables Flexible decision variablesLand dedicated to wheat crop production Flow rate of imported food

Land dedicated to potatoes production Flow rate of locally supplied nutrient for crop production

Land dedicated to cattle rearing Flow rate of locally supplied nutrient for livestock production

Land dedicated to pig rearing Flow rate of locally produced food from cropsFlow rate of locally produced food from livestock

Table 7-4 summarises the results from the scenario based simulations with partial

optimisation of the food production system with operational uncertainties. The objective

function of the scenarios varies within ± 20% of the nominal deterministic objective function;

comparable to the robustness analysis for the factual uncertainties in the food production

system. Change in final demand by the local population (FDF) triggered more uncertainties in

155

the objection function, i.e. resource consumption of the food production system.

Uncertainties in the local crop yield (YC) did not have a significant impact on the objective

function but it did allow the flexible decision variables to adjust so as to cope with the

uncertainties. The robustness analysis based on operational uncertainties indicated that

uncertainties in yield of livestock (YL), within the tested range, did not impact on the results

of the food production system.

Table 7-4: Results of scenario based simulations with partial optimisation

Decision variables S1 S2 S3 S4 S5 S6 S7 S8Land dedicated to

wheat crop production (ha)

8.05 8.05 8.05 8.05 8.05 8.05 8.05 8.05

Land dedicated to potatoes

production (ha)8.95 8.95 8.95 8.95 8.95 8.95 8.95 8.95

Land dedicated to cattle rearing (ha) 0 0 0 0 0 0 0 0

Land dedicated to pig rearing (ha) 0 0 0 0 0 0 0 0

Flow rate of imported bread

(kg/y)

128,428 128,428 103,094 217,976 217,976 192,642 103,094 192,641

Flow rate of imported potatoes

(kg/y)6 6 0 161,072 161,072 8 0 8

Flow rate of imported beef

(kg/y)64,743 64,743 63,434 105,132 105,132 97,115 63,434 97,115

Flow rate of imported pork meat

(kg/y)0 0 0 2618 2618 0 0 0

Flow rate of locally supplied nutrient for wheat crop

(kg/y)

1333 1333 2000 1333 1333 2000 2000 2000

Flow rate of locally supplied nutrient for potatoes crop

(kg/y)

1289 1289 1289 1289 1289 1933 1289 1933

Flow rate of locally supplied nutrient for cattle (kg/y)

14,521 14,521 18,078 0 0 21,783 18,078 21,782

Flow rate of locally supplied nutrient

for pig (kg/y)33,823 33,823 33,823 48,344 48,344 50,733 33,823 50,733

Flow rate of locally produced bread

(kg/y)50,670 50,670 76,004 50,670 50,670 76,004 76,004 76,004

Flow rate of locally produced potato

322,128

322,128 322,134 322,128 322,128 483,192 322,134 483,192

156

(kg/y)Flow rate of locally

produced beef (kg/y)

5345 5345 6654 0 0 8017 6654 8017

Flow rate of locally produced pork

meat (kg/y)37,066 37,066 37,066 52,980 52,980 55,598 37,066 55,598

Objective function (MJ/y) 8×107 8×107 8×107 1.3×108 1.3×108 1.3×108 8.3×107 1.3×108

7.4.4 Stochastic programming of food production system

Stochastic programming using DECIS solver in GAMS was performed to tackle the

embedded design uncertainty problem and to generate a robust design of the local food

production system considering all uncertainties. For simplicity, only the operational

uncertainties namely uncertainties associated with yields of crops (wheat and potatoes), yield

of livestock (cattle and pig) and the final local demand of food (bread, potatoes, beef and

pork) were considered. The first and second stage variables for the stochastic design of the

local food production system are summarised in Table 7-5.

Table 7-5: First stage and second stage decision variables

First stage decision variables Second stage decision variablesLand dedicated to wheat crop production Flow rate of imported food

Land dedicated to potatoes production Flow rate of locally supplied nutrient for crop production

Land dedicated to cattle rearing Flow rate of locally supplied nutrient for livestock production

Land dedicated to pig rearing Flow rate of locally produced food from cropsFlow rate of locally produced food from livestock

The first stage decision variables essentially comprise the fixed decision variables while the

second stage decision variables constitute the flexible decision variables. Next, the first and

second stage equations and constraints are also specified. First stage equations comprise only

first stage variables while the second stage equations contain both first and second stage

variables. Table 7-6 gives the first and second stage equations used for the stochastic design

of the local food production system.

Table 7-6: First stage and second stage equations

First stage equations Second stage equationsLand use for wheat production Final food demand for bread

Land use for potatoes production Final food demand for potatoesLand use for cattle rearing Final food demand for beefLand use for pig rearing Final food demand for pork

157

Crop production (wheat and potatoes)Livestock production (cattle and pork)

Nutrient balance for cropsNutrient balance for livestock

Food produced locally from cropsFood produced locally from livestockNutrient flows for crops and livestock

The uncertainties were specified using discrete values approximated to a uniform continuous

distribution with their mean being the nominal value used in the deterministic model. The

method adopted by DECIS to solve the stochastic model was specified in the DECIS options

file with istrat = 3 and nsamples = 100 which solves the expected value problem combined

with using Monte Carlo importance sampling.

Table 7-7: Results of stochastic design over deterministic design

Category Deterministic Stochastic UnitObjective function (resource

consumption in terms of exergy) 1.043×108 1.043×108 MJ/y

Amount of imported bread 160,536 243,015 kg/yAmount of imported potatoes 0 15,889 kg/y

Amount of imported beef 87,610 94,331 kg/yAmount of imported pork meat 46,332 51,447 kg/y

Amount of imported fertiliser for bread 1667 0 kg/y

Amount of imported fertiliser for potatoes 1611 1761 kg/y

Nutrients supplied for livestock production 0 0 kg/y

Amount of bread produced locally 63,336 0 kg/yAmount of potatoes produced locally 402,667 440,246 kg/y

Amount of beef produced locally 0 0 kg/yAmount of pork meat produced locally 0 0 kg/y

Amount of wheat crop produced locally 55,558 0 kg/y

Amount of potatoes crop produced locally 402,667 440,246 kg/y

Land dedicated to wheat production 8 0 haLand dedicated to potatoes production 9 17 ha

The results of the stochastic programming of the local food production system are

summarised in Table 7-7 and compared with the results of the nominal deterministic model.

Due to the relative linearity of the food production model and the uniform distribution of the

uncertainties around their nominal values, it is not surprising that the stochastic model

resulted in same objective function as the deterministic model. However, different results for

158

the decision variables were obtained which potentially indicate that multiple design solutions

exist that lead to the same objective function value.

7.4.5 Concluding remarks for robustness analysis and design under uncertainties

Post-design uncertainty assessment and uncertainty-embedded design are the two

fundamental ways that have been identified to address uncertainties in design and offer

respectively powerful insights into how robust a certain design is and the most robust design

that can cope with all uncertainties so as to help practitioners in their decision making

process. Post design uncertainty evaluates the performance of a given design with

uncertainties through a robustness analysis while embedded design uncertainty generates a

robust design in a robust optimisation given all the uncertainties. A clear distinction has also

been made in this report about the different types of uncertainties and decision variables that

can be present in a design and how they affect the approaches used to tackle the uncertainties.

Both post-design uncertainty assessment and uncertainty-embedded design were then

illustrated on a case study on the design of a localised food production system in Eco-Town

UK. A robustness analysis including a Monte Carlo simulation was performed to assess the

robustness of the given food design and a robust stochastic optimisation in GAMS using the

DECIS solver gave a design of the localised food production system that can best cope with

all the uncertainties. This illustration can be used as a basis for assessing uncertainties in

more complex design such as the simultaneous design of a localised food, water and energy

network.

The objective function for the stochastic optimisation was based on minimising the expected

value of the performance indicator (i.e. total cumulative exergy resource consumption) for the

design of the localised food production system. As a future step, the standard deviation of the

objective function could instead be adopted as a performance indicator to be minimised. The

minimisation of standard deviation is often adopted in financial portfolio optimisation.

Minimising the standard deviation of the objective function reduces its range of possible

values and is a very useful indicator that has the added benefit of enabling practitioners to

better manage risks in their design. Both the objective function on total cumulative exergy

resource consumption and its standard deviation can also be minimised in a multiple-

objective optimisation. In a multi-objective optimisation framework, Pareto-optimal

augmented ɛ-constraint, where one of the objective functions becomes a constraint in the

optimisation, is a well-known method that can be used.

159

7.4.6 Summary of systematic approaches to the design of localised integrated

production systems (LIPS)

Part (II) of the thesis presented two systematic approaches, presented in Chapters 5 and 6, for

the design of LIPS. The main novelty of Chapter 5 consists of a mathematical programming

based approach for designing local production systems which involve processes with very

diverse natures (e.g. manufacturing, agriculture and municipal). This approach is expected to

be capable of capturing integration opportunities and handling the characteristics of local

resources, such as seasonality of renewable resource supply. More specifically, the various

aspects in the novelty of this work include:

Capturing the integration opportunities not only within subsystems but also across

subsystems.

A life cycle approach accounting for resource consumption using cumulative exergy

consumption as an indicator of resource intensity for the imported flows as well as for

capital resources and environmental remediation efforts.

Consideration of local ecosystem limits that restrict the use of local resources.

Focusing on meeting local demands based on locally available resources.

Furthermore, this is the first time that such a systematic approach was applied for designing

the local food-energy-water nexus. The approach was illustrated using a case study on the

Whitehill and Bordon eco-town in the UK.

A systematic insight-based approach for the design of LIPS was then presented in Chapter 6.

It was envisaged that such an approach can give better insights to practitioners. Due to the

incremental nature of such an approach, it has the ability to capture complexities while

offering a relatively simpler but robust algorithm, apply mathematical modelling for solving

sub-problems and flexibility in shaping the design of the LIPS through feedback from users at

any stage of the design process. The insight-based approach was also tested on a case study

on the local design of food-energy-water nexus and compared with that from the

simultaneous design approach. The results indicated that though the simultaneous design

approach captured more resource integration opportunities, the insight-based approach was

not significantly worse off with both approaches suggesting qualitatively identical designs in

local food product and technical options for local production.

160

Chapter 8: Conclusions

8.1 Main research contributions and conclusions.

Locally integrated production systems (LIPS) have the potential to address many of the

critical challenges caused by centralised production and large scale distribution

infrastructures. If designed in a synergistic manner, LIPS can offer a possible pathway

towards sustainability. Its design is aimed at optimising the local heterogeneous processes

(e.g. agricultural, industrial and municipal) to satisfy the local demands by making the best

most efficient use of locally available renewable resources within technical and ecological

constraints. The design of LIPS is distinct from the design of conventional monolithic

production systems (e.g. plants producing bulk chemicals such as ammonia and ethylene, oil

refineries, car factories) which are often part of a relatively linear supply chain and to which

one or very few technical designs are universally adopted regardless of their locations. In

contrast, the LIPS will consist of a non-linear structure with any wastes and by-products

recycled back into the system and synergies within and across its different processes and

components to be exploited; thus requiring LIPS to be designed according to the local

settings and environment.

Before addressing the design problem itself, a proper indicator is needed to account for

system performance at every level of decision making. Therefore, the first step of this

research work was to define a clear design criterion namely resource consumption that

defines the objective of design and allows transparent comparison between design

alternatives. Resource accounting is an important approach that can assist decision making

and system design and contribute to a more sustainable path to development through

appropriate utilisation of resources along the whole value chain of a product or service. The

first main novelty of this research work presented in Chapter 2 of the thesis was then the

development of a comprehensive conceptual framework of a system for resource accounting

that encompasses production and consumption of products or services as well as the

ecological processes. The framework identifies the main characteristics of a system, which

can be either local or global, such as system boundary and types of flows and processes and

its multi-level nature allows for a better understanding of the analysis of the performance of a

system.

161

One of the main novelties of the research work was also to present a unique adaptation of the

Cumulative Exergy Accounting (CERA) methodology based on the conceptual multi-level

framework. Through CERA, the developed framework provides a more holistic and simpler

approach to resource accounting by using the unifying quantity of exergy which can account

for all types of resources including ecological, renewables, non-material and non-energetic

resources but also resources for both ecological and technological environmental remediation

of harmful effluents. CERA determines the cumulative exergy consumption during all the

processes leading to final products due to the consumption of material and energy. The use of

CERA has been motivated in this research by the goal of establishing a proper way to

quantify the “true” costliness of products and services by means of a physical quantity which

can then be used for assessing design options and act as an objective function to be

minimised in design along other considerations. Chapter 3 details the algebraic quantitative

approach developed as a key performance indicator at each level of the conceptual

framework. The quantitative multi-level framework was then applied for the first time on a

case study for the production of ethanol from sugarcane in Chapter 4. The framework

successfully demonstrated how the choice of design components and processes at one level

can have an impact on the other levels of the framework and affect the overall resource

consumption for producing the final product or service. It also proved useful in identifying

the resources that can be recycled and exchanged between and across the different levels of a

system. Due to its engineering oriented nature the developed framework aims to provide

support for decision making from a physical or technical perspective. The multi-level nature

of the resource accounting framework means that it would be of interest to stakeholders at

different levels such as engineers designing chemical production systems and planners of

industrial complexes and regional production-consumption systems. While the framework

does not directly contain business logics or management principles for commercial

operations, the systematic approach on physical resource accounting has the potential to

provide a solid basis for informing the relevant stakeholders with respect to the resource

impact of their decisions and to support research in other areas such as business for a

comprehensive assessment or sustainable accounting and to appeal to a wider audience

including corporate managers. In addition, the generic nature of the technical framework

entails that any other design criterion or performance indicator, other than from resource

consumption, could be used accordingly based on the interest of the decision makers.

162

Once the framework that characterises and quantifies resources in a local production system

has been established, another significant contribution of this research work, presented in

Chapter 5, was the development of a systematic approach to its design based primarily on

mathematical programming. In this context, a preliminary design analysis algorithm for LIPS

that could be carried out if it is desirable to gain an understanding of the interactions between

the various components of a local production system and how these might affect the overall

resource consumption was first presented. It is an optional design step that adds great value in

enriching the understanding of the linkages between subsystems. It is a useful tool when

dealing with existing system infrastructure, retrofitting design or when the subsystem

components of the local production system need to be designed and implemented separately

in stages with a view to develop system integration in the future. Next, a simultaneous

approach based purely on mathematical programming was proposed for designing a local

production system. Compared to the optional preliminary design approach, the simultaneous

one designs all the subsystem components of the system at once while accounting for all

possible interactions between the subsystems and local constraints for designing a local

system such as seasonality of resources to generate an optimal solution for the integrated

whole system design. Coupled with the novelty of the design approach, it was also the first

time that such approaches were being applied for the technical design of a local food-energy-

water nexus. For this purpose, the Whitehill and Bordon eco-town in the UK, a peri-urban

locale where residential/municipal facilities and local industry are key issues and that seeks to

transform existing facilities and industries to low carbon options by means of low renewable

resources, was specifically chosen as its vision matches that of this research work. It was

found that the simultaneous approach offered a superior design that was 6 and 2 times

respectively lower in resource consumption to that of a centralised supply design where

demands are satisfied by imported resources and designing the subsystems in silos, i.e.

individually without considering any synergies between the subsystems by considering all

integration options, circularity opportunities and emerging synergies including the exchange

of waste heat, treated domestic wastewater and rainwater between subsystems and the re-use

of organic residues between different the heterogeneous processes.

Chapter 6 presented the development of an insight-based design approach to generate

powerful insights into the design alternatives for LIPS. The robust approach offers a unique

iterative algorithm that can capture the complexities of designing a local production system.

It is also resource gain oriented and based on the concept of CExC developed in Chapter 2

163

and 3. The incremental nature of the algorithm makes it an interactive tool for decision

makers to generate insights into their preferred design alternatives. Another main contribution

of the developed insight-based approach was the establishment of a Locally Integrated

Production System Onion Model (LIPSOM) used to guide not only design but also the

process of information gathering. LIPSOM provides a conceptual approach to designing and

integrating unconventional and heterogeneous processes that usually constitute a local

production system. The insight-based approach was also demonstrated through a case study

on the design of local food-energy-water nexus similar to that presented in Chapter 5 though

slightly different data, assumptions and options for food, water and energy subsystems were

considered. This case study was then applied again to the simultaneous design approach to

reflect any changes made to the original case study presented in Chapter 5 and the results

compared with that from the insight-based approach. Both approaches indicated qualitatively

identical designs in terms of the food product and technical options selected for local

production. However, quantitative gains in the design of water and energy production

systems were captured in the simultaneous design approach; which can be attributed to its

rigorous and simultaneous mathematical optimisation nature.

The research thesis has aimed to make relevant contributions to the engineering of locally

integrated production systems by adopting and developing systematic tools for resource

accounting and process integration that could guide the efficient and confident generation of

sustainable LIPS designs. The main research question addressed in this thesis was how to

engineer sustainable locally integrated production systems using renewable resources to meet

local human needs under a range of conditions such as ecological and technical constraints.

More specifically, this thesis has aimed at addressing (i) how to characterise a system and

measure its technical performance and (ii) how to formulate these localised production

synthesis problems under different circumstances and how to solve these problems. In order

to address the former, Part (I) of this thesis, published in Leung Pah Hang et al. (2016a) has

focused on:

1) The development a conceptual framework for characterising any production-

consumption system (i.e. either local or global/external system).

2) A holistic and comprehensive multi-level framework that was used alongside a

resource consumption technical performance indicator developed for the evaluation of

resource accounting (including all types of resources from renewable, natural

resources from ecosystem processes, material and non-energetic resources such as

164

labour and capital resources) at multiple levels with the potential to reveal how

decisions at one level would affect other levels of the system.

3) Based on (1) and (2), the formulation a resource accounting algebra using exergy as a

unifying quantity, called ‘CERA’, for the quantitative assessment of resource

consumption at the different levels of a system.

Part (II) of the thesis, which comprises Chapter 5 published in Leung Pah Hang et al. (2016b)

and Chapter 6 submitted for publication in Leung Pah Hang et al. (2017), addressed the latter

research questions and has focused on:

Formulating the design problem for synthesising a local production system under

different circumstances and local settings taking the example of a local food-energy-

water nexus based on the Whitehill and Bordon locale in the UK.

Developing systematic approaches through the preliminary and simultaneous

mathematical programming approaches for solving the design problem towards

optimal technical performance based on the objective function of minimising total

resource consumption based on CERA.

Developing a set of preliminary guidance, design rules and principles to practices

related to the design of local production systems through the insight-based design

approach.

8.2 Wider implications of research

With the new economic paradigm shift from centralisation towards localisation, chemical

engineers must be ready with tools that allow a clear understanding of the new challenges

associated with the engineering of local production systems. This thesis has aimed to make

relevant contributions to the field of local production system by adopting and developing

insight-based and mathematical programming tools for process integration that could

facilitate the generation of resource efficient local systems. Notable contribution has been

made in this research to the field of process systems engineering (PSE) in the emerging area

of localised production systems engineering by providing a design methodological

framework to address the challenges and complexities posed by the planning and design of

sustainable locally integrated production systems from a systematically integrated perspective

that accounts for both intra- and inter-heterogeneous process integration opportunities

together with their interactions with the external world (e.g. other locales, regions, countries

etc.) while taking a life cycle approach.

165

Significant technical contributions have also been made to the emerging area of research of

food-energy-water nexus. Food, energy and water are essential needs to sustain human life.

The complex interactions and interdependencies among the infrastructures and decision

making processes involved in the provisioning of such needs have recently generated great

debate under the umbrella of the Food-Energy-Water (FEW) nexus (also refer to as Water-

Energy-Food nexus). As society becomes increasingly aware that resources available to

satisfy the needs of a growing population are finite, there is a call to look at the interactions

and start creating more efficient solutions to manage the wide range of nexus scenarios,

especially under conditions of climate change, increased urbanisation and waste generation.

The nexus may manifest in unique ways in different localities. Similarly, the global

conditions will affect each location differently. Therefore, it is urgent to develop holistic

analytical tools that could inform decision making for managing the interdependencies of the

FEW nexus at the local scale. This is one of the ways chemical engineers are able to

contribute with solutions by applying their skills of modelling processes and managing

constraints associated with an interconnected system. The design approaches developed in

this research work are expected to have wide application by engineers working across the

nexus system components, planners for urban or rural development, policy makers and

decision makers in general to understand, manage and create sustainable solutions to the

nexus. The thesis demonstrated that designing a set of interconnected heterogeneous

processes from resource extraction, agricultural to industrial/manufacturing processes at the

local scale by exploiting the well-established research areas of process integration and

industrial ecology could become complex could become complex compared to designing

traditional monolithic production system.

8.3 Future research avenues

One of the main novelties in this research work that was presented in Chapters 2 and 3 was

the development of CERA- a conceptual and quantitative framework for resource accounting

based on cumulative exergy consumption (CExC). Though, the inclusion of environmental

remedial resource costs in CExC allows for environmental protection to be taken into

consideration, the proposed resource accounting methodology does not address the issue of

depleting non-renewable resources or over-consumption of renewable resources. A new

concept could be adopted, similar to that of exergy replacement cost of mineral resources by

Valero et al. (2013), which accounts for the total exergy required for restoring used mineral

166

resources into the same state in which they were supplied by the ecosystems with the

available technology, in order to comprehensively address the issue of resource sufficiency as

well as efficiency.

The research done around CERA has also opened some future opportunities to account for

the true resource cost of a product or service. While this research has provided a rigorous

accounting procedure for operating and environmental remediation, including both ecological

and technological resources, it has built upon existing resource accounting work to provide

for an estimate of the cumulative exergy cost of capital resources. A more rigorous

systematic procedure for estimating capital resource consumption and understanding when

capital resource consumption becomes a significant component of total resource consumption

of a product/service could prove useful in generating a set of principles that could guide the

inclusion of capital resources in CERA so as to provide an even more rigorous resource

accounting methodology that would reflect the true costliness of the product/service.

The design approaches presented in Part (II) of the thesis have been useful in generating

design options for LIPS that demonstrate technical excellence. These approaches have further

established a practical framework that can be used in the future for the evaluation of the

economics and the social impacts of LIPS. The economic analysis could focus on estimating

the cost for running industrial and agricultural processes and building/maintaining system

capacity. Economic consideration can also be given to assess the impacts of increased intra-

region exchange and reduced inter-region exchange. Analysing the social impact, as once

pointed out by Johansson et al. (2005), of LIPS can define indicators with respect to

diversification of needs and wants, retaining of social capital, renewed producer-consumer

relationships and collaborative spirit within communities, drawing on Value Chain Analysis

methods. The outcomes of such economic and social evaluation could provide sound input

for decisions regarding the integration of technical, fiscal, political and social instruments

should LIPS be promoted to form part of the rebalanced economy.

167

Appendix AThis appendix contains a comprehensive list of all the data used in the case study on the

multi-level framework for resource accounting using algebras for the production of ethanol

from sugarcane for a typical plant with a capacity of 50,000 tonnes of ethanol per year.

A.1 Cumulative Exergy Consumption for cane agronomy

Basis of calculation: One tonne of ethanol.

Ethanol yield from sugar cane = 84.8 L/tonne (Junqueira et al., 2010).

Sugarcane yield = 65 tonne cane/ha/y (Perez, 1997).

Ethanol = 0.789 kg/L

Volume of ethanol produced = 1000/ 0.789 = 1267 L

Amount of sugarcane required to produce one tonne of ethanol = 15 tonne cane/tonne ethanol

A.1.1 Cumulative Exergy Consumption for fertilisers

Equation (A.1) can be used to calculate cumulative exergy consumption from fertilisers,

pesticides, insecticides and fungicides (flows from Type-I processes)

CExC i=CExC i ×F i

Y c×Y e (A.1)

where CExC i is the cumulative exergy consumption associated with resource input i to cane

agronomy (MJ/tonne ethanol)

CExC i is the specific cumulative exergy consumption of resource input i to cane agronomy

(MJ/kg)

F i is the amount of resource input i per area per year (kg/ha/y)

Y c is the cane yield (tonne cane/ha/y)

Y e is the ethanol yield (tonne cane/tonne ethanol)

Table A-1 summarises total exergy fertiliser input to cane agronomy.

Table A-1: Fertiliser input to cane agronomy

Fertiliser inputResource

requirement, F i (kg/ha/y)

CExC(MJ/kg)

CExC i

(MJ/tonne ethanol)

References

Nitrogen fertilizer 1.09×102 32.7 8.23×102 CSO 2010, Wittmus et al. 1975

Phosphorus fertiliser 4.77×101 7.52 8.28×101Odum 1995,

Wittmus et al., 1975

168

Potassium fertiliser 1.91×102 4.56 2.01×102 Odum 1995, Pimentel 1991

A.1.2 Cumulative Exergy Consumption for pesticides, insecticides and fungicides

Using Equation (A.1), the cumulative exergy consumption for pesticides, insecticides and

fungicides flows are summarised in Table A-2.

Table A-2: Pesticides, insecticides and fungicides input to cane agronomy

Resource inputResource

requirement, F i (kg/ha/y)

CExC(MJ/kg)

CExC i

(MJ/tonne ethanol)

References

Pesticides 1.42×102 3.68×102 1.20×102Odum 1995, Özilgena and

Sorgüven, 2011

Insecticides 7.29×10-1 3.44×102 5.77×101Odum 1995, Özilgena and

Sorgüven, 2011

Fungicides 8.40×10-3 2.56×102 4.9×10-1

AGRECO Consortium 2006,

Özilgena and Sorgüven, 2011

It was inferred from the study conducted by Özilgena and Sorgüven (2011) that the boundary

for the estimation of the specific cumulative exergy consumption for pesticides, insecticides

and fungicides encompasses their production, transportation and application. Their average

specific cumulative exergy consumption is used in this study.

A.1.3 Exergy consumption for ecosystem inputs

The ecosystem inputs are from Type-II processes and only their exergy content is accounted

for. The approach for this type of inputs started from the emergy values because these are the

type of information available.

Thus, Equation (A.2) adapted from the transformity equation by Odum (1995) is used to

obtain the exergy values:

EmY c

Y e=T × EX (A.2)

where,

Em is the emergy or available solar energy used up directly and indirectly to make the flow

(sej/ha/y)

T is the transformity in emergy per unit available energy/exergy (sej/J)

169

EX is the exergy is the exergy content/value of the flow (MJ)

Table A-3summarises the exergy of flows from Type-II processes.

Table A-3: Exergy of flows from Type-II processes

Ecosystem Input Emergy (sej/ha/y) Transformity (sej/J) Exergy (MJ/tonne ethanol)

Sunlight 5.54×1013 1.00 1.28×107

Rain chemical potential 8.18×1014 1.82×104 1.04×104

Rain geo-potential 0.12×1014 1.05×104 2.53×102

Wind 2.48×1015 1.50×103 3.82×105

Earth cycle 2.55×1014 2.55×104 2.30×103

Loss of topsoil 1.27×1016 7.38×104 3.96×104

Source: Bastianoni and Marchettini, 1996; Odum, 1995; Brown and Arding, 1991

A.1.4 Cumulative Exergy Consumption for land use

Exergy equivalence of land use, considered as flow from Type-II processes, for the plantation

of sugarcane can be estimated by using the exergy to land conversion proposed by Dewulf et

al. (2007), exergy to land = 68.14 MJ/m2/y

Amount of land required to produce 15 tonne of cane for the production of one tonne of

ethanol = (15 tonne cane/tonne ethanol)/ (65 tonne cane/ha/y) = 0.23 ha/y/tonne ethanol

Exergy of land required for production of one tonne of ethanol

= 0.23 ha/y/tonne ethanol × 10 000 m2/ha × 68.14 MJ/m2/y

= 1.57 × 105 MJ/tonne ethanol

Exergy of surface water = 2.26×1010 J/ha/y (Bastianoni and Marchettini, 1996)

Total exergy of surface water

= (2.26×1010 J/ha/y)/ (65 tonne cane/ha/y) × 15 tonne cane/tonne ethanol

= 5.20×109 J/ tonne ethanol

Total exergy of flows from Type-II processes including land use for cane agronomy and

surface water = 1.33×107 MJ/tonne ethanol

A.1.5 Cumulative Exergy Consumption for human labour

Bastianoni and Marchettini (1996) estimated the cumulative exergy labour input by the

product of labour working hours and an emergy to labour working hour conversion.

Cumulative exergy consumption equivalence of human labour for production of cane

170

= 5.20×107 J/ha/y (Bastianoni and Marchettini, 1996)

Cumulative exergy consumption equivalence of human labour for case study

= [(5.20×107 J/ha/y)/ (65 tonne cane/ha/y)] × 15 tonne cane/tonne ethanol

= 1.20 ×107 J/ tonne ethanol

A.1.6 Cumulative Exergy Consumption for lubricants

Cumulative exergy consumption for lubricants which are flows from Type-I processes were

determined based on Equation (A.3),

C ExC i=C Ei

Y c×Y e (A.3)

where,

C ExC i is the Cumulative Exergy Consumption of resource input i (MJ/tonne ethanol)

CEi is the Cumulative Exergy Consumption of resource input i to cane agronomy per area per

year (MJ/ha/y)

Y c is the cane yield (tonne cane/ha/y)

Y e is the ethanol yield (tonne cane/tonne ethanol)

Table A-4 summarises the result obtained from using Equation (A.3) on lubricants.

Table A-4: Total exergy flows for surface water and lubricants for cane agronomy

Resource inputCEi

(MJ/ha/y)CExC i (MJ/tonne

ethanol)References

Lubricants 3.00×102 6.90×101 Bastianoni and Marchettini 1996

Volume of diesel required to operate cane agricultural tractors

= 1.1 L/tonne cane (Leung Pah Hang, 2012)

Total diesel requirement to operate the agricultural machineries = 1.1 L/tonne cane × 15

tonnes of cane/tonne ethanol =16.5 L/tonne ethanol

Density of diesel = 0.832 kg/L (JRC, 2007)

Specific cumulative exergy consumption for diesel = 53.2 MJ/kg (Szargut et al., 1988)

Cumulative exergy consumption for diesel for the production of one tonne of ethanol

= 16.5 L/tonne ethanol × 0.832 kg/L×53.2 MJ/kg = 7.30×102 MJ/ tonne ethanol

171

A.1.7 Cumulative exergy consumption for capital resources

The accounting of capital resource consumption involves essentially spreading the total

resource consumption for providing a manufacturing capital, e.g. a piece of equipment, over

the amount of “jobs” this equipment carries out; the latter may be expressed in terms how

much material or energy this equipment processes during its operation, linking to its capacity.

The cumulative exergy consumption for capital resources,CExCmc, can be estimated by

Equation (A.4)

CExCmc=A

B × C (A.4)

where,

A is the total capital resource cost for a piece of equipment, determined by its total economic

cost and a money to exergy conversion factor (e.g. MJ),

B is the service life of the equipment (e.g. years),

C is the processing or other functional capacity of the equipment per year (e.g. kg/year), for

instance:

For material transportation, C is the tonne of material transported per year

For agricultural machinery, C is the tonne of sugarcane harvested per year

For industrial equipment, C is the capacity that the equipment can deliver per year

The total exergy of capital resources for agricultural tractor used in cane agronomy was

determined as follows:

Cost of agricultural tractor for sugarcane= 4408.02 Euros (Alibaba.com)

Average harvest rate for the cane agricultural tractor = 37.5 tonne per hour (Meyer, 2006)

Average estimated operating time the harvester per year = 569 hours/y (Beer et al., 1989)

Exergy to Money conversion for United States = 2.85 MJ/Euro (Sciubba, 2011)

Using Equation (A.4) for capital resources for the agricultural tractor:

where,

A= 4408.02 Euros x 2.85 MJ/Euro = 12563 MJ

B= 15 years (Assuming that the tractor has a service life of 15 years)

C= (37.5 tonne cane/hour) × 569 hours/y = 21337.5 tonne cane/y

Fmc = 15 tonne cane/tonne ethanol

172

CExCmc Fmc = 0.589 MJ/tonne ethanol

Similarly, other capital resources used in ethanol production and consumption were estimated

by using Equation (A.4) and the results are presented in Table A-5. The Chemical

Engineering’s equipment cost index was used to estimate the cost of equipment used in the

ethanol plant. The plant was assumed to be running 180 days per annum.

Table A-5: Cumulative exergy flows for capital resources

Capital resource A (MJ) B (years) CCExCmc Fmc (MJ/tonne ethanol)

Reference

Lorry for cane transportation 4.11×104 15 1.76×105

tonne cane/y 2.03×10-1 Beer et al 1989

Cane miller 2.74×107 15 7.20×105

tonne cane/y 3.80×101Dias et al

2010, Fulmer 1991

Clarifier 9.46×106 15 7.20×105

tonne cane/y 1.31×101Dias et al

2010, Fulmer 1991

Fermenter 9.93×106 15 7.20×105

tonne cane/y 1.37×101 Fulmer 1991

Distillation unit 5.84×106 15 7.20×105

tonne cane/y 8.12 Fulmer 1991

Molecular sieve 4.98×106 15 5.40×107 L ethanol/y 7.79 Bastidas et al

2010Azeotropic distillation 6.34×106 15 5.40×107 L

ethanol/y 9.91 Bastidas et al 2010

Anaerobic digester 9.09×107 15 2.53×109

m3/y 3.6×10-2Seckin and Bayulken

2013

Power house 1.47×103 15 4.32×103

kWh/y 4.14×101 Dean 1997

Tank car for ethanol

transportation3.04×105 15 5.84×105 L

ethanol/y 4.40×10-2

Stevens 2014, COM 2013, Beer et al,

1989

A.1.8 Cumulative Exergy Consumption for environmental remediation (CO2 emissions)

For simplicity, only CO2 emissions are considered for environmental remediation. Carbon

dioxide emissions are treated naturally through photosynthesis. The amount of carbon dioxide

released per annum from cane agronomy is summarised in Table A-6.

173

Table A-6: Carbon dioxide released into the atmosphere due to cane agronomy

Resource input CO2 released (kg/ha/y) CO2 released (kg/tonne ethanol) References

Pre-harvest burning 1.94×104 4.46×103 Dias de Oliveira et al, 2012

Loss of topsoil 7.62×102 1.75×102 Biondi, Panaro and Pellizi, 1989

Phosphate 4.77×101 1.1×101 Biondi, Panaro and Pellizi, 1989

Potash 1.91×102 4.39×101 Biondi, Panaro and Pellizi, 1989

Insecticides 7.29×10-1 1.68×10-1 Biondi, Panaro and Pellizi, 1989

Pesticides 1.42 3.27×10-1 Biondi, Panaro and Pellizi, 1989

Diesel 1.35×102 3.11×101 Biondi, Panaro and Pellizi, 1989

Lubricants 5.83 1.34 Biondi, Panaro and Pellizi, 1989

Human labour 3.3×10-2 8.0×10-3 Tiezzi, 1982Subtotal - 4.69×103 -

Average exergy consumption of net photosynthesis for all types of plant = 8 kCal/g carbon

(Odum, 1995)

Exergy of net photosynthesis in MJ/kg of carbon dioxide

= [(8 kCal/g carbon) × (4184 J/kCal) × (12 g carbon /44 g carbon dioxide)] × (1000g/kg) × (1

MJ/1000000 J)

= 9.13 MJ/kg carbon dioxide

Hence the total ecological cumulative exergy consumption for absorbing carbon dioxide

emissions related to cane agronomy = 4692 kg/tonne ethanol× 9.13 MJ/kg = 42838 MJ/tonne

ethanol. Similarly, carbon dioxide absorbed (negative as credit) and released along the supply

chain of ethanol production and consumption are summarised in Table A-7. Methane released

during bagasse decomposition is assumed to be treated through the natural process of

photosynthesis through its conversion to carbon dioxide equivalents on the basis of their

relative global warming potential (1 kg CH4 = 25 kg CO2 equivalents).

174

Table A-5: Cumulative exergy consumption for carbon dioxide absorption

Resource input CO2 (kg/tonne ethanol)

Cumulative exergy consumption

(MJ/tonne ethanol)Reference

Cane growth -1.38×104 -1.26×105 NSWS, 2014Cane transportation 5.40×101 4.93×102 Biondi et al., 1989

Methane 1.10×104 1.00×105 Vivekanand et al, 2014

Fermentation 9.50×102 8.68×103 StoichiometryBagasse burning 3.64×103 3.32×104 Stoichiometry

Ethanol combustion 1.91×103 1.75×104 StoichiometryEthanol transportation 5.40×101 4.93×102 Biondi et al., 1989

A.1.9 Total exergy consumption for cane agronomy

Following Equation (3.2) in Chapter 3 for total exergy consumption at the unit level, ExC for

cane agronomy excluding flows from Type-II processes was determined as follows:

ExCu=∑i=1

I

ExCi Fi + ∑mc=1

MC

ExCmc Fmc + ∑w=1

W

ExCw Fw

= 2757 MJ/tonne ethanol + 0.03 MJ/tonne ethanol + -97686 MJ/tonne ethanol

= -9.49×104 MJ/tonne ethanol

A.2 Cumulative exergy consumption for diesel for cane transportation

Using Equation (3.3) from the main text, the cumulative exergy consumption for diesel which

is considered a flow from Type-I processes was determined to be 2.78×103 MJ/tonne ethanol

(Odum, 1995)

A.3 Cumulative exergy consumption for industrial cane processing

A.3.1 Cumulative exergy consumption of electricity and imbibition water for cane

milling

The results of the cumulative exergy consumption of electricity and imbibition water for cane

milling are summarised in Table A-8.

Table A-6: Cumulative exergy consumption of electricity and imbibition water for cane milling

Resource input Flow rateSpecific

cumulative exergy

CExC i

(MJ/tonne ethanol)

Reference

Electricity 2.31×102

kWh/tonne ethanol

2.86 MJ exergy/MJ

electrical energy2.38×103 Macedo et al 2007;

Dewulf et al 2000

Imbibition water 4.32×103 kg/tonne 3.76 MJ 8.77×102 Palacios- Bereche

175

ethanol exergy/MJ wateret al, 2012;

Johnson and Seebaluck, 2012

A.3.2 Cumulative exergy consumption of lime and steam for juice treatment

The results of the cumulative exergy consumption of lime and steam for juice treatment are

summarised in Table A-9.

Table A-9: Cumulative exergy consumption of lime and steam for juice treatment

Resource input Flow rate (kg/tonne ethanol)

Specific cumulative

exergy (MJ/kg)

CExC i

(MJ/tonne ethanol)

Reference

Lime 6.69×10-1 1.01×101 1.01×102 MSIRI 2010; Szargut et al 1988

Steam 2.71×103 3.86 1.04×104Palacios- Bereche

et al, 2012; Dewulf et al, 2000

A.3.3 Amount and exergy of bagasse

The outputs of cane milling are raw mixed juice and bagasse produced from imbibition water

and cane inputs to the milling equipment.

Total raw mixed juice produced = 1.02 tonne/tonne cane (Seebaluck et al., 2008)

Total amount of raw mixed juice = 1.02 tonne/tonne cane ×15 tonne cane/tonne ethanol

= 15.3 tonne/tonne ethanol

Specific exergy of raw cane juice = 2697 kJ/kg (Palacios-Bereche et al., 2012)

Total exergy of raw cane juice

= 2697 kJ/kg ×15.3 tonne/tonne ethanol × 1000 kg/tonne × 1 MJ/1000 kJ

= 4.13×104 MJ/tonne ethanol

Total amount of bagasse/tonne ethanol = Total amount of cane/tonne ethanol + Total amount

of imbibition water/tonne ethanol – Total amount of raw mixed juice/tonne ethanol

= 15 tonne cane/tonne ethanol + 4.32 tonne imbibition water/tonne ethanol - 15.3 tonne raw

mixed juice/tonne ethanol

=4.02 tonne bagasse/tonne ethanol

Specific exergy of bagasse = 9979 kJ/kg (Palacios-Bereche et al., 2012)

Hence total exergy of bagasse produced = 4.02 tonne bagasse/tonne ethanol × 9979 kJ/kg ×

1000 kg/tonne bagasse × 1 MJ/1000 kJ = 4.01×104 MJ/tonne ethanol

176

A.3.4 Allocation factor between raw cane juice and bagasse

The cumulative exergy consumption for cane milling is allocated between raw cane juice and

bagasse based on their exergy content.

Allocationfactor = Exergy content of ra w mixed juice

Totalexergy content of raw canemixed∧bagasse

Allocationfactor = 41264 MJ

41264 MJ +40116 MJ

=0.507

A.3.5 Cumulative exergy consumption of operating resources for fermentation

The results of the cumulative exergy consumption of operating resources for fermentation are

summarised in Table A-10.

Table A-10: Cumulative exergy consumption of operating resources for fermentation

Resource input Flow rateSpecific

cumulative exergy

CExC i

(MJ/tonne ethanol)

Reference

Yeast 5.07×10-3 kg/tonne ethanol

3.46×10-2

MJ/kg 1.75×10-2 EIA 2011, Marques et al 1997

Sodium hydroxide 1.05×101 kg/tonne ethanol 1.45×101 MJ/kg 1.53×102 Langer, 2006

Szargut et al, 1988

Steam 1.27×103 kg/tonne ethanol 3.86 MJ/kg 4.89×103 Finguerut 2003,

Dewulf et al, 2000

Electricity 1.84×101

kWh/tonne ethanol

2.86 MJ exergy /MJ electrical

energy1.89×102 Jacques 2003,

Dewulf et al, 2000

Sulphuric acid 2.40 kg/tonne ethanol 1.11×101 MJ/kg 2.67×101

Macedo, 2004 cited in Langer,

2006, Szargut et al, 1988

A.3.6 Cumulative exergy consumption of operating resources for distillation

The results of the cumulative exergy consumption of operating resources for distillation are

summarised in Table A-11.

Table A-7: Cumulative exergy consumption of operating resources for distillation

Resource input Flow rate Specific cumulative

exergy

CExC i

(MJ/tonne ethanol)

Reference

Electricity 1.84×101 kWh/tonne 2.86 MJ exergy /MJ 1.89×102 Jacques 2003

177

ethanol electrical energy

Steam 3.80×103 kg/tonne ethanol 3.86 MJ/kg 1.47×104 Finguerut 2003,

Dewulf et al, 2000

A.3.7 Cumulative exergy consumption for vinasse treatment

The results of the cumulative exergy consumption of operating resources for vinasse

treatment are summarised in Table A-12.

Table A-12: Cumulative exergy consumption of operating resources for vinasse treatment

Resource Flow rate Specific cumulative exergy

CExC i

(MJ/tonne ethanol)

Reference

Electricity 8.14×10-1

kWh/tonne ethanol 2.86 MJ/MJ 8.38Dewulf et al.,

2000, Khan et al, 2011, Jordao, 2010

Calcium carbonate

1.45×102 kg/tonne ethanol 1.0×101(MJ/kg) 1.46×103 Souza, 1986;

Meneses, 2008

Iron (III) chloride

4.8×101 kg/tonne ethanol 1.86×101(MJ/kg) 8.92×102

Seckin and Bayulken 2013;

Szargut et al, 1988

Volume of vinasse produced = 12 L per L of ethanol (Smith, 2006)

Total volume of vinasse formed = 12 L/L ethanol×1267 L ethanol/tonne ethanol × 1 m3/1000

L = 15.204 m3 /tonne ethanol

Chemical Oxygen Demand, COD of Vinasse characteristics (EIA, 2011) = 29000 mg/L

COD characteristic of Vinasse in kg COD per L ethanol

=

= 0.348 kg COD per L ethanol

Total COD characteristic of vinasse in kg of COD per tonne ethanol

= 0.348 kg COD/L ethanol × 1267 L ethanol /tonne ethanol

= 440.9 kg COD/tonne ethanol

178

L ethanol L

1000 L

1 m3

0.012 m3 10-6 kg1 mg

29 000 mg

Taking an average of 0.20 m3 of methane to be liberated per kg COD removed (Ramjeawon,

1995),

Total methane biogas released for vinasse treatment = 0.20 m3/kg COD × 295 kg COD/tonne

ethanol = 59 m3/tonne ethanol

Specific exergy content of methane biogas = 34 MJ/ m3 (Szargut et al., 1988)

Total exergy output of methane biogas = 34 MJ/ m3 ×59 m3/tonne ethanol = 2006 MJ/tonne

ethanol

Net cumulative exergy consumption for vinasse wastewater treatment

= Cumulative exergy consumption electricity + Cumulative exergy consumption CaCO3 +

Cumulative exergy consumption FeCl3 + Cumulative exergy consumption capital resources

for Vinasse wastewater treatment - exergy output of methane biogas

= 8.38 MJ/tonne ethanol + 1457.25 MJ/tonne ethanol + 892.8 MJ/tonne ethanol + 0.036

MJ/tonne ethanol - 2006 MJ/tonne ethanol

= 3.52×102 MJ/tonne ethanol

A.3.8 Cumulative exergy consumption for operating resources for molecular sieve

The results of the cumulative exergy consumption of operating resources for dehydration are

summarised in Table A-13.

Table A-13: Cumulative exergy consumption of operating resources for dehydration

Resource input

Flow rate(kg/tonne ethanol)

Specific cumulative exergy (MJ/kg)

CExC i

(MJ/tonne ethanol)

Reference

Steam 7.60×102 3.86 2.93×103 CTC 2005, Dewulf et al, 2000

A.3.9 Cumulative exergy consumption for operating resources for azeotropic

dehydration

Cyclohexane is used in the dehydration of ethanol. The resource inputs for the production of

cyclohexane are given in Table A-14.

Table A-14: Cumulative exergy consumption for production of cyclohexane

Resource input Inputs (kg/kg cyclohexane)

Cumulative exergy consumption (MJ/ kg

cyclohexane)Reference

Benzene 9.3×10-1 5.49×101 Zhang 2008; Szargut,et al 1988

179

Hydrogen 7.8×10-2 1.95×101 Zhang 2008; Dewulf et al., 2000

Steam 1.0×10-1 3.86×10-1 Zhang 2008; Dewulf et al., 2000

Cumulative exergy consumption for producing 1 kg cyclohexane = 54.87 MJ/kg of

cyclohexane + 19.5 MJ/kg of cyclohexane + 0.386 MJ/kg of cyclohexane = 74.8 MJ/kg of

cyclohexane

The amount of cyclohexane required per L ethanol has been estimated from Bastidas et al.,

(2010) to be around 5.04 × 10-3 kg cyclohexane per L ethanol

Hence cumulative exergy consumption for cyclohexane

= 5.04 × 10-3 kg cyclohexane/L ethanol × 1267 L/tonne ethanol × 74.8 MJ/kg cyclohexane

= 4.78×102 MJ/tonne ethanol

Amount of steam required for azeotropic distillation = 1.7 kg /L (Finguerut, 2003)

Hence total amount of steam required = 1.7 kg /L × 1267 L/tonne ethanol = 2154 kg

steam/tonne ethanol

Specific cumulative exergy consumption of steam = 3.86 MJ/kg (Dewulf et al., 2000)

Cumulative exergy consumption for steam for production of one tonne of ethanol

= 3.86 MJ/kg × 2154 kg/tonne ethanol

= 8.31×103 MJ/tonne ethanol

A.4 Cumulative Exergy Consumption for power station

Power house produces steam and electricity. It is assumed that a state of the art boiler and 82

bars and 525 °C condensing extraction steam turbine are used for the production of steam and

electricity. The power house meets the energy requirements of the ethanol plant and any

surplus of electricity is exported to the grid.

Figure A-1: Condensing Extraction Steam Turbine

180

Source: Lau, 2008

A.4.1 Electricity and steam production from power house

Total amount of bagasse was calculated to be 4.02 tonne bagasse/tonne ethanol

The steam to bagasse ratio is the amount of steam generated per unit of bagasse burned and is

the superheated live steam produced from the 82 bars and 525°C.

Taking the steam to bagasse ratio = 2.4 (Hau, 2008)

The superheated live steam has a specific enthalpy of 3445 kJ/kg as per the steam table.

h1 = 3445 kJ/kg

Moreover, the process steam produced from the condensing extraction steam turbine is at

150°C and 2.0 bars. At this temperature and pressure, the bled process steam is slightly

superheated while the vapour leaving the condenser part of the condensing extraction steam

turbine is at 42°C and 0.08 bars and has an assumed dryness fraction of 0.92 (Lau, 2008).

From steam table,

h2= 2770 kJ/kg

h3= 2384 kJ/kg

Using the steam to bagasse ratio of 2.4;

Total amount of superheated steam generated = 2.4 × 4020 kg/tonne ethanol = 9648 kg/tonne

ethanol

Total amount of process steam required to be bled from turbine (kg)/tonne ethanol

= Total steam consumption for juice treatment + Total steam consumption for fermentation +

Total steam consumption for distillation + Total steam consumption for dehydration

= 2760 kg/tonne ethanol + 1267 kg/tonne ethanol + 3801 kg/tonne ethanol + 760 kg/tonne

ethanol

= 8534 kg/tonne ethanol

Total mass of vapour going to condenser

= Total amount of superheated steam - Total amount of process steam

= 9648 kg/tonne ethanol - 8534 kg/tonne ethanol

= 1114 kg/tonne ethanol

181

Using the first law of thermodynamics, the electrical output, W, from the Condensing

Extraction Steam Turbine is calculated by the following formula:

W =η× {mtotal live steam × (h1 – h2) + (mtotal live steam – msteam to process) (h2 – h3)}/3600

where,

W is the total electrical output of the turbine in kWh/tonne ethanol

h1 is the enthalpy of live steam in kJ/kg

h2 is the enthalpy of process steam in kJ/kg

h3 is the enthalpy of steam leaving the condenser in kJ/kg

mtotal live steam is the mass of live steam in kg/tonne ethanol

mSteam to process is the mass of process steam in kg/tonne ethanol

mVapour is the mass of vapour leaving condenser of CETA in tonne

η is the combined overall mechanical and electrical efficiency of the condensing extraction

steam turbine and it is assumed to be 0.95.

W = η × {mtotal live steam x (h1 – h2) + (mtotal live steam – msteam to process) (h2 – h3)}/3600

= 0.95 × {9648 kg/tonne ethanol × (3445 – 2770) + 1114 kg/tonne ethanol (2770 –

2384)}/3600

= 1832 kWh/tonne ethanol

Total electrical output from turbine = 1832 kWh/tonne ethanol

Total surplus electricity that can be exported to the grid

= Total electrical output from turbine – Total electrical input for ethanol plant – Total

electrical input for power house

= 1832 kWh/tonne ethanol – (231 kWh/tonne ethanol + 18.37 kWh/tonne ethanol + 18.37

kWh/tonne ethanol) – 322.5 kWh/tonne ethanol

= 1242 kWh/tonne ethanol

Total exergy content of electricity is the same as the total energy content of electricity since

in theory all the electricity can be converted into work.

Total exergy content of the surplus electricity

= 1242 kWh/tonne ethanol × 3.6 MJ/kWh

= 4.47× 103 MJ/tonne ethanol

182

A.4.2 Cumulative exergy consumption for water for power house

Amount of water required per kWh of electrical output =1.140 kg/ kWh (Macknick et al.,

2011).

Total amount of water = 1.140 kg/kWh ×2205 kWh/tonne ethanol = 2514 kg/tonne ethanol

Specific exergy of water = 50 kJ/kg (Szargut et al., 1988)

Total exergy of water = 50 kJ/kg × 2514 kg/tonne ethanol = 1.26× 102 MJ/tonne ethanol

A.4.3 Cumulative exergy consumption for electricity for power house

The electrical consumption in a typical bagasse cogeneration plant equipped with

electrostatic precipitator for particulate matter control is estimated to be around 21.5 kWh per

tonne cane (Lau, 2008). Electricity is required in the power house for running the electrical

equipment and for compressing the air before it is supplied to the combustion chamber.

Total electricity requirement of the power house = 21.5 kWh/tonne cane × 15 tonne

cane/tonne ethanol = 322.5 kWh/tonne ethanol = 1.16×103 MJ/tonne ethanol

This electricity can be supplied internally by recycling part of the electricity produced by the

power house.

A.4.4 Allocation factor for bagasse

The allocation factor for production of the bagasse can be calculated by the following

equation:

Allocationfactor = Exergy content of bagasse

Totalexergy content of cane juice∧bagasse

Allocationfactor = 40116 MJ

41264 MJ +40116 MJ

= 0.49

A.5 Total exergy consumption at process level with intra-recycling flows

It is assumed that the intra-recycling flows do not need any processing before being used as

input flows to a unit. From a backward mass balance for ethanol production and using

recovery rates and purity data from Bastidas et al. (2010), the ethanol/water flow from the

regeneration bed can replace about 15% of the ethanol from the fermentation beer.

Additionally, from a backward mass balance for ethanol production and an assumed purity

ethanol of 10% from the fermentation beer and 98% recovery rate from the distillation unit,

the amount of water produced from the distillation unit was estimated to be about 10.74 tonne

of water. Part of this water can be recycled back internally to the cane milling unit to fully

satisfy its imbibition water requirements.

183

The total exergy consumption for ethanol production with recycled flows from the distillation

and dehydration units can be determined by using Equation (3.3) from Chapter 3 where,

∑ir=1

IR

CExC ir F ir

= Sum of the cumulative exergy consumption of the fresh feed flows that recycled

ethanol/water mixture from dehydration replaced and cumulative exergy consumption of the

fresh feed flows that recycled water from distillation replaced

∑ir=1

IR

CExC ir F ir

= 0.15 ×CExC fermentation F fermentation + 0.85×CExC imbibition water Fimbibition water

¿ (0.15× 76900 MJ/tonne ethanol) + (0.85 × 877 MJ/tonne ethanol)

= 1.23×104 MJ/tonne ethanol

∑r=1

R

CExC r F r =0 J/tonne ethanol, assuming the recycling flows do not require any processing,

∑ac=1

AC

CExCac Fac = 0 MJ/tonne ethanol as the disposal resource costs for the ethanol/water

mixture form the dehydration unit and condensate water flows from the distillation unit were

both ignored in this study.

∑u=1

U

CExCu - ∑ℑ=1

ℑ

CExC ℑ Fℑ = 9.51×104 MJ/tonne ethanol,

Hence, using Equation (3.3), CExC p=¿8.28×104 MJ/tonne ethanol

A.6 Total exergy consumption at inter-process level with recycling and exchange

flows

The total exergy consumption for the production of ethanol at the inter-process level with

recycled and exchanged flows can be determined by using Equation (3.5) from Chapter 3

where,

∑p=1

P

CExC p = (CExC p )ethanol production+ (CExC p )Steama nd electricity generation

∑p=1

P

∑ei=1

EI

CExC ei , p Fei , p =0.85 × (CExCelectricity milling F electricitymilling +

CExCsteam juice treatment F steam juice treatment+¿

184

CExCsteam fermentation F steamfermentation+CExC electricity fermentation F electricity fermentation ¿ +

CExCsteam distillation F steam distillation+CExCelectricity distillation F electricity distillation +

CExCsteam dehydration F steamdehydration+ 0.85×CExCbagasse Fbagasse

Cumulative exergy consumption of avoided methane emissions, ∑enx=1

ENX

CExCenx F enx = 1.00×105

MJ/tonne ethanol

Since 15% less bagasse is sent to the power house with recycled flows from the regeneration

bed of the dehydration unit,

Total surplus electricity = 0.85× 1242 kWh/tonne ethanol = 3.80×103 MJ/tonne ethanol

Specific exergy of ethanol = 30 MJ/kg (Szargut et al., 1988)

Total exergy of ethanol = 30 MJ/kg x 1000 kg/tonne ethanol = 3.00×105 MJ/tonne ethanol

Since two useful products electricity and ethanol are being produced;

The allocation factor for the production of ethanol can be calculated as follows based on

exergy content:

Allocationfactor = Exergy content of ethanol

Totalexergy content of ethanol∧electricity

= 30 000 MJ / tonne ethanol

30 000 MJtonne

ethanol+3800.52 MJtonne

ethanol

=0.887

Assuming that the same amount of diesel is consumed for the transportation of ethanol to the

fuelling station as that for the transportation of cane from field to the ethanol plant, the

cumulative exergy consumption for diesel is 7.30×102 MJ/ tonne ethanol.

Using Equation (3.2) from the main text, the total exergy consumption for ethanol

consumption was determined to be 1.87×104 MJ/tonne ethanol. Hence, total exergy

consumption for production and consumption of ethanol taking into account the allocation

factor of 0.887 was determined to be 1.26×104 MJ/tonne ethanol excluding flows from Type-

II processes.

185

A.7 Summary of total exergy consumption of each unit in ethanol production and

consumption

Using Equation (3.2) from the main text, the total exergy consumption for each unit,

excluding flows from Type-II processes, in ethanol production and consumption if there are

no recycling and exchange flows is summarised in Table A-15.

Table A-15: Total exergy consumption without recycling flows

Unit Operating resources(MJ/tonne ethanol)

Capital resources(MJ/tonne ethanol)

Environmental remediation resources

(MJ/tonne ethanol)Cane agronomy 2.76×103 3.00×10-2 -9.77×104

Pre-harvest burning 0.00 0.00 4.07×104

Cane transportation 2.78×103 2.3×10-1 4.93×102

Cane milling 3.26×103 3.8×101 1.00×105

Juice clarification 1.05×104 1.3×101 0.00Fermentation 5.26×103 1.37×101 8.68×103

Distillation 1.49×104 8.16 3.52×102

Dehydration (molecular sieve) 2.93×103 7.79 0.00

Consumption 7.30×102 4.4×10-2 1.80×104

Similarly, the total exergy consumption for each unit, excluding flows from Type-II

processes, in ethanol production and consumption with recycling flows but no exchange

flows is summarised in Table A-16.

Table A-16: Total exergy consumption with recycling flows

Unit Operating resources(MJ/tonne ethanol)

Capital resources(MJ/tonne ethanol)

Environmental remediation resources

(MJ/tonne ethanol)Cane agronomy 2.34×103 3.00×10-2 -8.30×10-4

Pre-harvest burning 0.00 0.00 3.46×104

Cane transportation 2.36×103 2.0×10-1 4.19×102

Cane milling 2.02×103 3.23×101 8.50×104

Juice clarification 8.96×103 1.11×101 0.00Fermentation 4.47×103 1.16×101 7.38×103

Distillation 1.49×104 8.16 3.52×102

Dehydration (molecular sieve) 2.93×103 7.79 0.00

Consumption 7.30×102 4.40×10-2 1.80×104

186

The total exergy consumption for each unit, excluding flows from Type-II processes, in

ethanol production and consumption with recycling flows and exchange flows is summarised

in Table A-17.

Table A-17: Total exergy consumption with recycling and exchange flows

Unit Operating resources(MJ/tonne ethanol)

Capital resources(MJ/tonne ethanol)

Environmental remediation resources

(MJ/tonne ethanol)Cane agronomy 2.08×103 2.30×10-2 -7.37×104

Pre-harvest burning 0.00 0.00 3.07×104

Cane transportation 2.10×103 1.70×10-1 3.72×102

Cane milling 0.00 2.87×10-1 0.00Juice clarification 8.95×10-1 9.88 0.00

Fermentation 1.59×102 1.03×101 6.54×103

Distillation 0.00 7.23 3.09×102

Dehydration (molecular sieve) 0.00 6.91 0.00

Power station 9.50×101 3.76×101 2.50×104

Consumption 7.30×102 4.40×10-2 1.80×104

A.8 Comparative analysis for resource consumption for the 3 scenarios

The total exergy resource consumption, excluding flows from Type-II processes, for the 3

scenarios- ethanol production and consumption without recycling, ethanol production and

consumption with recycling and, ethanol production and consumption with recycling and

exchange flows are summarised in Table A-18.

Table A-18: Comparative resource consumption for the 3 scenarios

Resources (MJ/tonne ethanol) Scenario 1 Scenario 2 Scenario 3

Operating 4.31×104 3.87×104 5.25×103

Capital 8.11×101 7.13×101 1.01×102

Environmental remediation 7.05×104 6.27×104 7.28×103

187

Appendix BThis section contains all the data and assumptions used for Chapter 5 of the thesis and is

based on the supporting information document for Leung Pah Hang et al. (2016b).

B.1 Cumulative exergy resources for food production subsystem

The food products considered are bread, potatoes, pork and beef and have been chosen based

on local food preferences in Whitehill and Bordon eco-town. These food choices also give a

good representation of a human being’s dietary requirements in carbohydrate, protein and

fats. The annual consumption by the local population is given in Table B-1 and was

determined based on the average daily consumption of these food types from DEFRA (2014).

Table B-8: Food demand by local population in the eco-town

Food product Demand (t/y)Bread 224

Potatoes 403Beef 88Pork 46

Table B-2 shows the cumulative exergy consumption associated with imported food. The

cumulative exergy consumption values were determined based on data from various literature

sources and include cumulative exergy consumption for transporting the food into the UK.

Beef is assumed to be imported from Ireland (AHDB, 2013) and pork from Denmark

(AHDB, 2014). In addition, the distances for importing the beef and pork to UK were

estimated using a Food Miles Calculator online tool. It was also determined that 0.6 MJ

exergy is spent on transporting 1 tonne of food over a distance of 1 km; a value derived from

the cumulative exergy of diesel which is 53.2 MJ/kg (Szargut et al., 1988) and the diesel fuel

efficiency of 5 MJ per tonne km (IPCC, 2007). Only the distance used to transport the food

by freight trailer from other European countries to UK was taken into account. The distances

for transporting the food within the UK into Eco-Town were not considered in the estimate.

The cumulative exergy of groundwater used in the production of the imported food products

was estimated to be about 0.06 MJ/kg from section B.2 of the Appendix. In addition, it was

assumed that conventional energy sources such as heat from natural gas boilers and grid

188

electricity were used in the production of the imported food. The specific cumulative exergy

intensity of heat produced from natural gas boilers was estimated to be about 2.01 MJ exergy

per MJ heat energy while that of grid electricity was determined to be about 5.97 MJ exergy

per MJ electricity in section B.3 of the Appendix.

Table B-2: Cumulative exergy of imported food used in Chapter 5

Food type Specific cumulative exergy (MJ/kg) Source

Bread 150IME (2014);

Nielsen and Nielsen (2003);Andersson and Ohlsson (1999)

Potatoes 6.0 Zhang et al. (2012)Beef 950 EA (2009), IME (2014)

Pork 370 IME (2014), Gerbens-Leenes et al. (2013); Williams et al. (2006)

B.1.1 Cumulative exergy resources for bread production

The specific cumulative exergy values of resources (i.e. per kg of the resource) other than

utilities used for local food production are reported in Table B-3. Table B-3 details the

amount of each resource considered in bread production. It is assumed that there is no

cumulative exergy associated with the fodder (i.e. agriculture residues) produced from wheat

crops as they were considered to be available as a “free” input since they are side product of

crop cultivation. The harvest recovery ratio, defined as the ratio of the amount of residues

harvested to the total amount of residues produced, was assumed to be 0.80. Agricultural

residues can be used as a local source of protein for animal feed. The protein fraction of

fodder was taken to be 0.2 (Panday and Mishra, 2011). The cumulative exergy associated

with land use was not taken into account in this case study. Furthermore, only the capital

resource consumption for crop storage was considered. The capital resource consumption for

agricultural machineries and food processing plants were not included.

The loss in wheat grain from moving grain into and out of storage and loss in quality of the

wheat grain due to shrinkage during storage was neglected as it is usually a negligible factor

in the range of 0.5 to 1% (Edwards, 2015). The capital cost of wheat crop storage facility is

about 33 Euro per tonne wheat crop capacity (FAO, 2015). A service life of 20 years for the

wheat crop storage facility and a cumulative exergy to capital cost of 2.85 MJ/Euro which

depends on the socio-economic conditions at a particular time period in the UK (Sciubba,

189

2011) were used. Using the general Equation (5.12) presented in Chapter 5 for cumulative

exergy of capital resources for the crop storage facility and substituting the values gives:

CAwheat=33 ×2.85

20WST

where, WST is the size of the wheat storage facility

In addition to the capital exergy of the wheat storage facility, wheat storage also requires

operating exergy resources. The operating resources vary linearly with the amount of wheat

grain stored in each season. The moisture content of wheat grain is about 25% (McNeil et al.

(2010) and needs to be artificially dried to 15% for storage (Maier and Bakker-Arkema,

2002). The temperature of thermal treatment required as a pest control technique to disinfect

the wheat grains was taken to be about 60 °C (FAO, 2011). The cumulative exergy

consumption of operating resources for wheat storage was determined to be about 0.52MJ

heat energy per kg wheat grain (Maier and Bakker-Arkema, 2002) and includes the heat

energy for aerating and decreasing the moisture content of wheat from 25% to 15% in a dryer

operating with a thermal efficiency of 4.64 MJ per kg of water removed from a wheat flow

rate of 82,000 kg/h. From Marques et al. (1997) the exergy of yeast is estimated to be

34.57kJ/kg. The cumulative exergy consumption for producing the yeast as per the system

boundary considered in this study was approximated to be its total exergy content due to lack

of available data on yeast production.

Table B-3: Specificities for local bread manufacture including wheat cultivation

Resources Quantity SourceWheat cultivation

Water 701.8 kg water/kg bread IME (2014)Heat for wheat storage 0.52 MJ/kg wheat stored Maier and Bakker-Arkema

(2002)Fodder from wheat 7 t/ha Cowell and Parkinson (2003)

Fertiliser 0.0264 kg N/kg bread DK (2014)Pesticides 1.566 kg/ha Audsley et al. (2009)

Diesel 0.0144 kg/kg bread DK (2014)Wheat processing into bread

Wheat yield 6.9 Mg dry matter/ha/year Williams et al. (2006)Water 1608 kg water/kg bread IME (2014)Heat 1 MJ/kg bread Nielsen and Nielsen (2003)

Electricity 0.388 MJ /kg bread Nielsen and Nielsen (2003)Yeast demand 0.0035 kg/kg bread Ruskins (2015)

Wheat crop to bread 1.14 kg bread/kg wheat crop DK (2014)

190

B.1.2 Cumulative exergy resources for beef production

The resources considered for local beef manufacture are detailed in Table B-4. It is assumed

that manure produced from the cattle will not have any specific cumulative exergy associated

with it. It is assumed that there is no cumulative exergy associated with the organic manure

produced from cattle. The harvest recovery ratio of manure, defined as the ratio of amount of

manure collected to the total amount of manure produced from cattle, was taken to be 0.80.

For simplicity, the animal feed was characterised based on its protein content to satisfy the

protein requirements of the cattle. About 1.44 t protein is required per t live cattle weight per

year (Panday and Mishra, 2011). The mass per cattle was taken to be 0.68 t (DM, 2013).

Table B-4: Specificities for local beef manufacture

Resources Quantity SourceCattle breeding

Area required per tonne meat beef 4.7 ha/t Cowells and Parkinson (2003)

Manure per kg live cattle per year 25 kg/kg cattle/y USDA (2009)

Nitrogen content in manure 0.005 kg N/kg manure ECOCHEM, 2015Fodder required per live cattle

per day 13.38 kg fodder/cattle/day Panday and Mishra (2011)

Protein required per amount of live cattle per year 1.44 t/t cattle/year Panday and Mishra (2011)

Cattle processing into beefBeef to cattle ratio 0.36 t/cattle Cowells and Parkinson (2003)

Water demand (including water demand for cattle breeding) 15,415 L/kg beef IME (2014)

Electricity 2.56 MJ electricity/kg cattle AHDB (2013)Heat 1195 MJ heat/kg cattle EA (2009)

B.1.3 Cumulative exergy resources for pork production

The specific cumulative exergy of the resources used in the production of pork is detailed in

Table B-5 while the amount of resources required for pork manufacture is given in Table B-6.

Similar to manure produced from cattle; it is assumed that there is no cumulative exergy

associated with the organic manure produced from pig. The harvest recovery ratio of manure

from pig, defined as the ratio of amount of manure collected to total amount of manure

produced from pig, was assumed also to be 0.80. The animal feed was characterised based on

its protein content to satisfy the protein requirements of pigs. About 0.73 t protein is required

per t live pig weight per year (FF, 2015). The mass of one pig was taken to be 0.090 t

(Lawlor, 2010).

191

Table B-5: Specific cumulative exergy of resources used in pork production

Resources Specific cumulative exergy (MJ/kg) Sources

Calcium 10.05 Szargut et al. (1988)Iron 28.63 Szargut et al. (1988)

Sulphur 30.21 Szargut et al. (1988)

Szargut et al. (1988) generally considered in their analysis the sum of all the exergy

consumed along the supply chain of a product from extraction of the natural resources to the

industrial manufacture of the final product. However, their cumulative exergy database does

not include the cumulative exergy for environmental remediation processes and does not

offer a comprehensive accounting of all the resources, such as capital resources and labour,

used in the production of the final product.

Table B-6: Specificities for local pork manufacture

Resources Quantity SourcePig breeding

Pork to pig ratio 0.072 t/pig Cowells and Parkinson (2003)Area required per tonne meat

pork 1.2 ha/t Cowells and Parkinson (2003)

Manure per kg live pig per year 25 kg/kg pig/y USDA (2009)Nitrogen content in manure 0.006 kg N/kg manure ECOCHEM, 2015

Protein required per amount of live pig per year 0.73 t/t pig/year FF (2015)

Fodder required per amount of live pig per day 0.01 t/t pig/day FF (2015)

Pig processing into porkWater demand (including water

demand for pig breeding) 5988 L/kg pork IME (2014)Gerbens-Leenes et al. (2013)

Heat 1.51 MJ/kg pork EA (2009)Electricity 1.58 MJ/kg pork ERM (2009)Calcium 0.0057 kg/kg pork Williams et al. (2006)

Iron 0.0041 kg/kg pork Williams et al. (2006)Sulphur 0.0057 kg/kg pork Williams et al. (2006)

B.1.4 Cumulative exergy resources for potatoes production

The specificities required for local potatoes production are given in Table B-7. Similar to

bread production from wheat, it is assumed that there is no cumulative exergy associated with

the agricultural residues produced from potatoes. The heat demand for potatoes was not

considered in this case study as the potatoes production process involves potatoes cultivation

and thoroughly washing the potatoes before being sold to the local population. The energy

192

content of pesticides was taken to be 809 MJ/kg (Audsley et al., 2009) and was used to derive

its cumulative exergy in terms of MJ exergy per kg pesticides.

Table B-7: Specificities for local potatoes production

Resources Quantity SourcePotatoes cultivation

Annual potatoes yield 45 t/ha UK Agriculture (2014)Fertiliser 175.5 kg N/ha Williams et al. (2006)

Residue to potatoes ratio 1.62 t residue/t crop Jata et al. (2011)Electricity demand for irrigation 142 MJ/tonne Gulati and Singh (2011)Electricity for potatoes storage 270 MJ/t potatoes stored Kneeshaw (2006)Water demand for cultivation 18,000 L/t Gulati and Singh (2011)

Diesel 158 MJ/t Gulati and Singh (2011)Pesticides 0.11 kg/t Audsley et al. (2009)

Potatoes processingWater demand for processing 6000 L/t Gulati and Singh (2011)

As was the case for wheat production, storage is also considered for potatoes. Potatoes are

usually planted in April but harvested through summer and autumn in a year (UK

Agriculture, 2014). The capital cost of wheat crop storage tank is about 55.7 Euro per tonne

potatoes crop capacity and a service life of 30 years for the potatoes storage facility were

used (Patterson, 2007). The cumulative exergy to capital cost was taken to be 2.85 MJ/Euro

(Sciubba, 2011). Using the general Equation (5.12) presented Chapter 5 for cumulative

exergy of capital resources for the crop storage facility and substituting the values gives:

CA potatoes=55.730

WST

where, WST is the size of the potatoes storage facility

Potatoes need to be stored in a cool dark place at about 4°C. Electricity requirement for

refrigeration, recirculation fans and humidification has been determined to be about 75 kWh

per tonne potatoes (Kneeshaw, 2006).

B.2 Cumulative exergy resources for water production subsystem The water requirements of the food, energy production processes and residential sector in

eco-town have been considered in the design of the water production system. The cumulative

exergy consumption of capital resources associated with wastewater treatment such

wastewater treatment plant equipment and water distribution infrastructure (e.g. piping) to the

193

locality were not considered in this study. However, the cumulative exergy of operating

resources of energy and chemicals and the energy required for pumping groundwater were all

considered.

B.2.1 Wastewater production from food production subsystem

The wastewater generated from the food processes are detailed in Table B-8. The following

assumptions were made to estimate the amount of wastewater produced from the food

processes:

The density of all the wastewater produced from bread, potatoes, pork and beef

production is similar to the density of water (i.e. 1000 kg/m3).

Table B-8: Wastewater produced from food processes

Food processes Quantity SourceBeef production 5.28 kg wastewater/kg beef EPA (2008a)

Potatoes production 6 m3/tonne Xu et al. (2014)Bread production 1.42kg wastewater/kg bread AssumedPork production 3.67 kg wastewater/kg pork EPA (2008a)

B.2.2 Water demand and wastewater production from residential

Approximately 350-500 litres of water is required for domestic purposes per person on a

daily basis while the amount of wastewater generated is about 200-300 litres per person per

day (DEHP, 2014). Based on these values and a population of 17,000 (Whitehill and Bordon,

2012), the total water requirements and the total wastewater generated from residential have

been determined as given in Table B-9.

Table B-9: Water demand and wastewater generated from residential

Resources Quantity per season (t)Water demand 586,867

Wastewater 365,126

B.2.3 Rainwater collected in Eco-Town

Table B-10 shows the estimated rainwater collected per season in eco-town. An example

illustrating how rainwater collected for summer is as follows:

Number of houses in eco-town = 4000 (Whitehill and Bordon, 2012)

Average roof area of a house in UK = 91 m2 (BCD, 2008)

Total surface area for rainwater collection = 4000×91m2 = 364,000 m2

Average amount of rainfall in summer = 0.1937 m (Met Office, 2012)

194

Total volume of rainfall in summer = 0.1937 m×364,000 m2 = 70,506.8 m3

An efficiency of 75% is assumed due to evaporation and leaks (BCD, 2008).

Thus, total volume of rainfall that can be collected in summer

= 70,506,800 L x 0.75

= 52,880,100 L

The same method was used to determine rainwater collected in spring, autumn and winter.

Table B-10: Seasonal rainwater collected

Season Volume of rainwater (L)Spring 52,880,100

Summer 52,880,100Autumn 68,140,800Winter 62,899,151

Rainwater is assumed to be collected from the residential roof area in eco-town. In line with

the Master plan for Eco-Town, the rainwater harvesting is done at a centralised eco-town

level. Thus, all collected rainwater is brought to a common location where it is treated before

being distributed to the town to be used for industrial, residential and agricultural purposes

(Whitehill and Bordon, 2012). The following data were used to estimate the total capital

exergy resources for rainwater storage:

Capital cost of rainwater storage tank is 0.65 Euro per litre (ECOSURE, 2015).

A service life of 20 years for the rainwater storage tank was assumed.

Cumulative exergy to capital cost is 2.85 MJ/Euro (Sciubba, 2011).

Substituting the values in Equation (5.22) presented in Chapter 5 for total cumulative exergy

consumption for capital resources for rainwater storage tank gives:

CArw=0.65 ×2.8520

OSR

B.2.4 Water demand and wastewater production from energy production subsystem

From a study by Rasmussen (2011), it was determined that the water requirement by biomass

CHP, organic CHP and natural gas CHP was about 0.074 L per MJ energy. It was assumed

that 85% of the water requirements for the energy production processes are converted into

wastewater (IEA, 2012). Hence, wastewater generated from the CHPs was estimated to be

0.063 L wastewater per MJ energy. Table B-11 summarises the water demand and

wastewater production from the energy production subsystem.

195

Table B-11: Water demand and wastewater generated from energy production subsystem

Resources Quantity SourceWater demand 0.074 kg/MJ energy Rasmussen (2011)

Wastewater 0.063 kg wastewater/MJ energy IEA (2012)

B.2.5 COD of water sources

The COD concentration of the wastewater generated from food processes, energy processes,

residential as well as the COD concentration of groundwater and rainwater is given in Table

B-12.

Table B-12: Quality of water source

Water source COD concentration (g COD/kg wastewater) Source

Wastewater from bread manufacture 6.5 Chen et al. (2006)Wastewater from pork manufacture 14 EPA (2008a)

Wastewater from potatoes production 11 Sayed et al. (2005)Wastewater from beef production 18 EPA (2008a)

Domestic wastewater 0.750 Henze and Comeau (2008)Groundwater 0.035 CEMEX, 2014

Rainwater 0.006 Ward (2010)Wastewater from biomass CHP 15 Huber (2015)

Wastewater from organic waste CHP 14 de Mes et al. (2003)Wastewater from natural gas CHP 0.750 de Mes et al. (2003)

The maximum quality limit allowed for the water requirements for food, energy and

residential purposes was assumed to be 0.010 g COD per kg water (Enderlein et al., 2014)

while that for discharge into water bodies such as river was taken to be 0.10 g COD per kg

water (NEA, 2014).

B.2.6 Cumulative exergy resources for wastewater treatment

In this case study, treatment of wastewater generated from food, energy and residential as

well as treatment of rainwater and groundwater before they are used for water consumption

are considered. The capital resources used for wastewater treatment were not considered. The

operating resources required for wastewater treatment are given in Table B-13. The

efficiency of wastewater treatment was taken to be 92% (Saad, 2009).

Table B-13: Operating flows for wastewater treatment

Resources Quantity SourceHeat 0.641×10-3 MJ/L EPA (2011)

Electricity 0.81×10-3 MJ/L Menendez (2009)

Chemicals 0.49 kg calcium carbonate/kg COD removed Souza (1986)

196

Considering the heat is sourced from natural gas boilers, the specific cumulative exergy of

heat for treating wastewater was found out to be 6.4×10-4 MJ exergy per kg wastewater.

Assuming electricity is sourced from the grid; the specific cumulative exergy of electricity

for treating wastewater was determined to be 7.8×10-4 MJ exergy per kg wastewater.

B.2.7 Cumulative exergy resources for groundwater

The resources considered in determining the cumulative exergy of groundwater are detailed

in Table B-14.

Table B-14: Resources for groundwater

Resources Quantity SourceHeat for treatment of

groundwater 0.641×10-3 MJ/L EPA (2011)

Electricity for treatment of groundwater 0.81×10-3 MJ/L Menendez (2010)

Chemicals for treatment of groundwater

0.49 kg calcium carbonate/kg COD removed Souza (1986)

Electricity for pumping groundwater 1.98×10-3 MJ/L King et al. (2008)

Considering only conventional energy sources (i.e. heat from natural gas boilers and grid

electricity) and including both cumulative exergy for pumping and treating the groundwater,

the specific cumulative exergy of groundwater was determined to be 0.06 MJ/kg.

B.3 Energy production system

This section contains all the data and assumptions used to derive the cumulative exergy

consumption of producing energy from different energy sources.

B.3.1 Electrical efficiency

The heat and electrical efficiency of biomass, organic waste and natural gas CHP, solar panel

and wind turbines are given in Table B-15. The biomass combined heat and power (CHP)

considered works by direct combustion of the wood chips biomass where the heat generated

from combustion is transferred to a fluid that is used in an organic rankine cycle.

In the natural gas CHP, natural gas is fed to a gas turbine to produce the energy while in the

organic waste CHP biogas is first generated from anaerobic digestion (AD) of the organic

waste and then fed into a gas turbine. The efficiency of the organic waste CHP stated in Table

197

B-15 refers to the efficiency of converting the biogas produced from (AD) into heat and

electricity.

Table B-15: Heat and electrical efficiency of CHP

CHP Efficiency (%) SourceHeat efficiency of biomass CHP 51 Whitehill and Bordon (2012)Electrical efficiency of biomass

CHP 19 Whitehill and Bordon (2012)

Heat efficiency of natural gas CHP 50 EPA (2008b); Whitehill and Bordon (2012)

Electrical efficiency of natural gas CHP 35 EPA (2008b); Whitehill and

Bordon (2012)Heat efficiency of organic waste

CHP(conversion of biogas to heat)

50 Assume same as natural gas CHP

Electrical efficiency of organic waste CHP

(conversion of biogas to electricity)35 Assume same as natural gas

CHP

Electrical efficiency solar panel 20 MacKay (2009)Electrical efficiency of wind

turbine 44 NREL (2010)

Thermal efficiency of natural gas boilers 85 Stark (2015)

B.3.2 Cumulative exergy of energy input

The cumulative exergy of biomass wood chips, organic waste, natural gas, wind and solar is

given in Table B-16 in MJ exergy per MJ of energy content. Organic waste was taken to have

no cumulative exergy associated with it as it is a recycled waste product within the local

system. Additionally, the exergy contents of wind and solar which are flows from Type-II

processes are not considered in this case study as it is assumed that these flows do not have

alternative competing uses in eco-town.

Table B-16: Cumulative exergy of energy input

Energy sourceCumulative exergy

consumption per energy input (MJ/MJ)

Source

Biomass 1 Chen and Chen (2009)Organic waste 0 Assumed

Natural gas 1.32 Szargut et al. (1988)Solar 0 AssumedWind 0 Assumed

198

B.3.3 Cumulative exergy of energy production

The total cumulative exergy associated with the production of energy from solar panel, wind

turbines and biomass, organic waste and natural gas CHP are summarised in Table B-17. The

values for CHPs were derived from Carbon Trust (2013), UoS (2001), Rasmussen (2011) and

Szargut et al. (1988) and include cumulative exergy of capital resources, water, raw material

and resources for environmental remediation of carbon dioxide emissions. Only the capital

resources were considered in the estimation of the total cumulative exergy consumption for

solar panels and wind turbines. The resources associated with the piping and infrastructure

for district heating and for distributing the energy to end-users in eco-town were not taken

into consideration. It was also assumed that distribution losses for energy were negligible

within the town. The service life of the solar panels and wind turbines was taken to be 20

years (CAT, 2015; VESTAS, 2015). The cost of a 2 kW capacity solar panels with an annual

electricity output of 1700 kWh is about £4500 (Vasili, 2015). The cost of a 1 kW capacity

wind turbine is $1940 (AWEA, 2012) with an annual electricity output of about 2182 kWh

(EWEA, 2015). The service life of the natural gas boiler was assumed to be 15 years. The

capital cost of the boiler was taken to be £129/kW with an average of 4000 operating hours

per year (DECC, 2014; EST, 2014).

Table B-17: Total cumulative exergy consumption of energy technology

Energy technology Total cumulative exergy consumptionBiomass heat 2.01 MJ exergy/MJ heat

Biomass electricity 5.34 MJ exergy/MJ electricityNatural gas CHP heat 3.37 MJ exergy/MJ heat

Natural gas CHP electricity 4.75 MJ exergy/MJ electricityNatural gas boilers 2.01 MJ/MJ heatOrganic waste heat 0.409 MJ exergy/MJ heat

Organic waste electricity 0.58 MJ exergy/MJ electricitySolar 2.23 MJ exergy/MJ electricityWind 1.42 MJ exergy/MJ electricity

Table B-18 gives the amount of carbon dioxide released for the production of heat and power

from biomass, organic waste and natural gas CHP per MJ of energy from a life cycle

perspective.

199

Table B-18: Carbon dioxide emissions from CHP

CHP Carbon dioxide emissions SourceBiomass heat 0.0045 kg CO2/MJ heat Carbon Trust (2013)

Biomass electricity 0.0089 kg CO2/MJ electricity Carbon Trust (2013)Natural gas CHP heat 0.078 kg CO2/MJ heat Carbon Trust (2013)

Natural gas CHP electricity 0.107 kg CO2/MJ electricity Carbon Trust (2013)Natural gas boilers heat 0.05 kg/MJ heat EPA (2015)

Organic waste heat 0.044 kg CO2/MJ heat IPCC (2006)Organic waste electricity 0.063 kg CO2/MJ electricity IPCC (2006)

An example illustrating how the total cumulative exergy associated with producing heat and

electricity from wood biomass CHP was determined is as follows:

Cumulative exergy of wood biomass = 1 MJ/MJ energy

Efficiency of converting biomass into electricity = 0.19

Efficiency of converting biomass into heat = 0.51

Cumulative exergy of biomass for producing electricity = 1

0.19 = 5.26 MJ/MJ electricity

From section D.2, total amount of water required for both electricity and heat generation from

CHP = 268 litres/MWh (Rasmussen, 2011)

Amount of water required for electricity production from CHP

= 268 × 0.19(0.19+0.51)

=72 litres /MWh

Specific cumulative exergy of groundwater was estimated in section B.2 to be 0.06 MJ/kg.

Assuming conventional groundwater is used, cumulative exergy of water for electricity

production from biomass CHP

= 72 kg/MWh × 0.06 MJ/kg × 1 MWh/3600 MJ = 0.0012 MJ/MJ

The following Equation (B.1) can be used to estimate the total capital resources associated

with energy production.

CExCc=A

B × C (B.1)

where,

200

CExCc (e.g. MJ/MJ) is the cumulative exergy consumption of capital resources per unit of

energy produced,

A (e.g. MJ) is the total capital resource cost for the CHP, determined by its total economic

cost and a money to exergy conversion factor,

B (e.g. years) is the service life of the CHP,

C (e.g. kg/year) is the processing or other functional capacity of the equipment per year

The capital cost of CHP was taken to be about €1750/kW (Simet, 2012). The money to

cumulative exergy consumption was assumed to be 2.85 MJ/€ (Sciubba, 2011). It is also

assumed that the CHP operates 7000 hours per year (Danon et al., 2012) and that the CHP

plant has a life of 20 years (EPA, 2009). The electrical efficiency of biomass CHP was taken

to be 19% (Whitehill and Bordon, 2012). Substituting the values in Equation (D.1) gives:

CExCc=1750 ×2.85

7000 ×20×3.6 = 0.010 MJ/MJ

Thus, cumulative exergy of capital resources for producing electricity from biomass CHP

= 0.01 MJ/MJ × (0.19

(0.19+0.51))

= 0.003 MJ/MJ

The cumulative exergy for treating carbon dioxide emissions through the ecological process

of photosynthesis was estimated to be 9.13 MJ/kg of CO2 (Odum, 1995).

Hence, cumulative exergy for treating carbon dioxide emissions from production of

electricity from biomass CHP

= 9.13 MJ/kg CO2 × 0.0089 kg CO2/MJ electricity

= 0.08 MJ/MJ electricity

Total cumulative exergy of producing electricity from biomass CHP

= Cumulative exergy of raw material + cumulative exergy of water + cumulative exergy of

capital resources + cumulative exergy of treating carbon dioxide

= (5.26 + 0.0012 + 0.003 + 0.08) MJ/MJ electricity

= 5.34 MJ/MJ electricity

The cumulative exergy of biomass for producing heat

201

= 1

0.51

= 1.96 MJ/MJ heat

Amount of water required for heat production from CHP

= 268 litres/MWh-72 litres/MWh = 196 litres/MWh

Assuming conventional groundwater is used, exergy of water required for heat production

from biomass CHP = 196 kg/MWh × 0.06 MJ/kg × 1 MWh/3600 MJ= 0.003 MJ/MJ

Cumulative exergy for treating carbon dioxide emissions from production of heat from

biomass CHP

= 9.13 MJ/kg CO2 × 0.0045 kg CO2/MJ heat

= 0.04 MJ/MJ heat

Cumulative exergy of capital resources for producing heat from biomass CHP

= 0.01 MJ/MJ - 0.003 MJ/MJ

= 0.007 MJ/MJ

Thus, total cumulative exergy of producing heat from biomass CHP

= Cumulative exergy of raw material + cumulative exergy of water + cumulative exergy of

treating carbon dioxide + cumulative exergy of capital resources

= (1.96 + 0.003 + 0.04 + 0.007) MJ/MJ heat

= 2.01 MJ/MJ heat

The same approach is applied for determining the total cumulative exergy of producing

electricity and heat from natural gas and organic waste CHP.

The cumulative exergy of grid electricity was determined to be 3.40MJ/MJ electricity by

Szargut et al. (1988) which take into account a conventional power plant fed with bituminous

coal at the consumption place. However, this value does not include the resources consumed

for treating the carbon dioxide emissions. For consistency, the carbon dioxide emissions from

coal-fired power station are about 0.28 kg CO2 per MJ electricity (Carbon Trust, 2013) or

2.57 MJ exergy per MJ electricity. Hence, the cumulative exergy of imported electricity from

grid was determined to be about 5.97MJ/MJ electricity.

202

B.3.4 Variability of energy sources

The variability of the energy sources is given in Table B-19. It is assumed that the availability

of biomass and organic waste is constant for all seasons. The organic waste in eco-town

comprises food and animal waste generated in the local system. The availability of wood

chips biomass and biogas from anaerobic digestion of organic waste in eco-town has been

estimated by Whitehill and Bordon (2012) in their detailed energy feasibility study. The

formula for determining the wind energyWE (e.g. MJ/h) available in eco-town for the

production of electricity is given in Equation B.2.

WE=12

× ρ × A × v3 ×3600 × 1106 × N (B.2)

With ρ the density of air at 1.225 kg/m3, A is the area of the wind turbine blade assumed to

be 1963 m2 with diameter 50 m (VESTAS, 2015), v is the average velocity of the wind

assumed to be 6.5 m/s (Whitehill and Bordon, 2012) and N is the number of wind turbines

that can be installed in Eco-Town. N was estimated to be 3 large wind turbines each of

capacity 2.5 MW for a total land area of 70 ha based on data provided in the master plans for

Eco-Town.

Table B-19: Variability of energy sources

Energy source Supply of energy source per season (GJ) SourceWinter Spring Summer Autumn

Wood chips biomass 413,910 413,910 413,910 413,910

Whitehill and Bordon

(2012)Organic waste(availability of

biogas from organic waste)

24,090 24,090 24,090 24,090Whitehill

and Bordon (2012)

Wind 22,688 18,615 38,115 33,165 Prasad et al (2009)

Solar* 1.26 5.24 6.63 2.71 PVGIS (2014)

* Unit is in GJ of energy per m2 area per year

B.3.5 Land requirement for energy production

The land available for energy generation in Eco-Town is assumed to be 70 ha. The land use

for all the energy sources is given in Table B-20.The area required for solar panels was

estimated from Eco-Town’s planned production of 425,000 kWh of solar power per ha per y

(Whitehill and Bordon, 2012). This is equivalent to 57.3 m2 per MJ per hour as illustrated:

203

Total planned production of solar power

= 425,000kWh/ha/y

= (1 ha × 10,000 m2)/ ((425,000 kWh/y × ha × y × 3.6 MJ)/ (24 × 365 h × 1kWh))

= 57.3m2/MJ/h

From AWEA (2015), an average of 60 acres of land is required for the aerial production of 1

MW of electricity from wind. This is equivalent to 67.4 m2 per MJ per hour. The total area

for biomass energy production in eco-town is estimated at 4000 m2 including a biomass

collection centre. About 4000 MWh of electricity is expected to be produced from this area

per year.

The total area including an anaerobic digestion plant planned in eco-town for production of

energy from organic waste is 5000 m2; out of which 10,400 MWh of electricity is expected to

be produced per year. The land use for producing energy from natural gas CHP is assumed to

be similar to the land use for organic waste CHP.

It was assumed that natural gas boilers can be installed within the residential houses and

industrial processing facilities for food production and wastewater treatment and as such will

not take any significant additional land use in comparison with other energy technologies.

Table B-20: Land use of energy sources

Energy source Land use (m2/MJ/h) SourceSolar 57.3 Whitehill and Bordon (2012)

Wind 67.4 AWEA (2015); Whitehill and Bordon (2012)

Biomass CHP 2.43 Whitehill and Bordon (2012)Organic waste CHP 1.17 Whitehill and Bordon (2012)

Natural gas CHP 1.17 Assumed

The temperatures at which waste heat is available from the food and energy processes are

given in Table B-21. The source of the waste heat for bread production is mainly from the

low temperature heat originating from the oven exhaust gases (Paton, 2003). Refrigeration in

the meat processing industry also produces low temperature waste heat at about 60 °C. Low

temperature waste heat available at an assumed temperature of about 120°C from the exhaust

the CHPs’ turbine is also recovered in this case study. The cumulative exergy of capital

resources for the heat exchangers that are required for heat integration were not considered in

this study.

204

Table B-21: Inlet temperature of waste heat

Waste heat Temperature (°C) SourceBread production 70 Paton (2013)Beef production 60 FAO (2010a)Pork production 60 FAO (2010a)Biomass CHP 120 Assumed

Organic waste CHP 120 AssumedNatural gas CHP 120 Assumed

The temperature at which heat is required for food processes, residential and wastewater

treatment purposes is given in Table B-22.

Table B-22: Temperature required by heat sinks

Waste heat Temperature (°C) SourceBread production 220 Paton (2013)Beef production 85 FAO (2010a), FAO (2010b)Pork production 85 FAO (2010a), FAO(2010b)Wheat storage 65 FAO (2011)

Residential 62.5 Varbanov and Klemes (2011)Wastewater treatment 35 Souza (1986)

Some key assumptions made in the design of the energy production system are as follows:

10% of heat required by bread, beef and pork production is lost as recoverable waste

heat and 10% of heat produced from CHP becomes waste heat (Law et al., 2011).

The temperature of all the waste heat after heat exchange with heat sinks is 30 °C.

The temperature of heat sinks before heat exchange is about 20 °C.

Residential electricity demand is constant throughout the year and is assumed to be

22,566 GJ electricity for each season based on the data from Whitehill and Bordon

(2012).

Heat and electrical demands for food and water processes are also assumed to be

constant throughout the year.

The heat demand for eco-town per season is estimated based on DECC (2015b) and Whitehill

and Bordon (2012) and is given in Table B-23.

Table B-23: Seasonal residential heat demand

205

Season Residential heat demand per season (GJ)Winter 112,128Spring 85,848

Summer 82,344Autumn 96,360

The maximum quantity of heat that can be exchanged between the waste heat sources and the

heat sinks,H Max, was taken to be 10,000. The minimum temperature TD between the inlet

temperature of the waste heat and the temperature required by the heat sink and that between

the outlet temperature of the waste heat after exchange and the temperature of the heat sink

before exchange was assumed to be 10 °C. M is the upper bound for temperature difference

and was taken to be a high value of 300.

206

Appendix CThis appendix contains additional review on resource regeneration options including a review

on pinch analysis as illustrated for water pinch.

C.1 Evaluation of resource regeneration options

Regeneration of resources is considered an essential process that can potentially reduce the

consumption of fresh resource and the generation of waste streams. A waste stream has a

resource cost burden associated with it as it has to be treated before it can be discharged

safely into the environment within the stringent environmental laws and regulations.

Regeneration can take the form of any process that treats a resource that has generally used

already to increase its quality level (i.e. quality upgrading) (Hallale, 2002). In order to

achieve the optimum regeneration process, the following three key principles adapted from

Foo et al. (2006) have to be followed:

(i) Selecting the optimal sequence in terms of starting quality (e.g. starting COD

level) to carry out regeneration of available resources. This requires one to

determine at what quality it is best to regenerate the resource. For instance, is it

better to regenerate water at 30 g COD/kg water or continue its use in the same or

another process unit where it gets dirtier at 40 g COD/kg water?

(ii) Selecting the optimum target quality upgrading. For example, what is the

optimum target COD concentration for regenerating wastewater?

(iii) Selecting the most resource effective regeneration technology scheme.

Regeneration has been extensively applied with pinch analysis techniques to optimise the

design of water networks (Manan et al., 2006) and energy networks (Becker et al., 2011).

As regeneration of a resource involves its quality upgrading, the change in the load of the

resourceΔ mi, defined as the product of the amount/quantity of the resource and its quality

change (i.e. flow rate of the resource multiplied by the change between inlet and outlet

quality) is expressed in Equation (C.1).

Δ mi=F i ΔCR (C.1)

F i is the amount of resource i available for regeneration and ΔCR is the change between the

inlet and outlet quality of source i. The operating resource cost is typically proportional to the

change in load. However, the amount of resource i (F i¿ is inversely proportional toΔCR; this

207

means that a smaller F i, which in turn would imply a smaller size of the regeneration unit,

would need to be processed to regenerate a lower quality resource (ΔCR ¿ than for a relatively

high quality resource (due to lower ΔCR ¿for a fixed amount of contaminant removed (Δ mi).

Following this principle, Foo et al. (2006) reported that to achieve zero discharge with the

minimum capital and operating resource costs for the case of water regeneration, it is

important to adopt the heuristic of purifying water sources in increasing order of its quality. If

COD is a measure of water quality where high COD refers to low quality/purity of water, this

heuristic means that in order to minimise both the capital and operating resource costs of

water regeneration one should first purify the water source at the highest COD concentration

level and continue with sources at the second highest COD concentration level and so on,

until all the water sources have been purified and wastewater eliminated to achieve zero

discharge. An algebraic equation for determining the cumulative exergy resource cost for

regeneration has been formulated through Equation (C.2).

CExCreg=CExC Fi ΔCR (C.2)

CExCreg is the cumulative exergy resource cost for regeneration and includes operating,

capital and environmental remediation resources, CExCis the specific exergy consumption

for regenerating unit amount of degraded resource up to a certain quality level. CExC is

normally not a constant value and is a function of the quality upgrading. The practicality of

using Equation (C.2) is limited by the availability of data for determining the specific exergy

consumption for regenerating unit amount of degraded resource up to a certain quality level

(i.e. CExC ¿.

For the case of water regeneration, Equation (C.2) translates to Equation (C.3).

CExCwreg=CExCw(¿C cod ,initial−C cod , final) Fw ¿ (C.3)

CExCwreg is the cumulative exergy consumption for water regeneration, CExCwis the specific

exergy consumption for regenerating unit amount of degraded water up to a certain quality

level (e.g. MJ per kg COD removed). C cod , initial and C cod , final are the initial and final

concentrations of COD in the water respectively (e.g. kg COD/kg water). Fw is the total flow

rate of water regenerated (e.g. kg/s). According to Foo (2013), CExCw is likely to decrease

when the difference between C cod , initial and C cod , final increases.

208

Equation (C.2) as applied to a heat pump gives Equation (C.4) for heat regeneration.

CExChreg=CExCh Fe(C.4)

CExChreg is the cumulative exergy cost for heat regeneration, CExCh is the specific exergy

consumption for regenerating unit amount of degraded heat up to a certain quality level (e.g.

MJ/ MJ). CExCh is not a constant and is a function of the initial and target temperature of the

heat energy resource. F eis the amount of heat being regenerated at the high temperature (e.g.

MJ).

To better understand the underlying principles behind the regeneration process and why in

principle the specific exergy consumption for regenerating unit amount of degraded resource

is not a constant, a thermodynamic quantity based on the second law of thermodynamics has

been formulated in Equation (C.5). It is based on an adaptation of the study undertaken by

Cerci et al. (2003) on water desalination and employs the chemical potential and Gibbs

functions. Equation (C.5) can be used as a generic equation for minimum work, W min ,∈¿, ¿for

the regeneration of a solution consisting of many componentsi.

Wmin ,∈¿=−T0∑

iRi mf i ln(mf i

M m

M i)¿ (C.5)

where Ri is the gas constant of a component i, M m is the molar mass of the solution, M i is the

molar mass of component i and mf i is the mass fraction of componenti. Equation (C.5) can

then be used with the thermodynamic efficiency to determine the cumulative exergy cost for

resource regeneration (i.e. the actual work) as given in Equation (C.6). The thermodynamic

efficiency is between 0 and 1.

CExCreg=W min ,∈¿

ɳ¿ (C.6)

CExCregis the actual work input or total cumulative exergy cost for resource regeneration as

represented in Equation (C.5). W min ,∈¿¿ is the minimum work input as given in Equation

(B.6). The minimum work input and the second-law efficiency provide a basis for

comparison of actual regeneration processes to the idealised ones and they can be very

valuable tools for assessing the thermodynamics performance of regeneration processes.

Equation (B.6) is less practical to use than Equation (C.5) due to the difficulty in determining

the thermodynamic efficiency; which depends on the type of technology used for the

209

regeneration process as well as the range of working condition such as the initial and target

temperature of heat regenerated as in the case of a heat pump.

C.2 Water regeneration

Water regeneration involves the upgrading of water purity using any purification techniques.

The regenerated water can either be reused in other water-using processes or recycled to the

same process to further reduce fresh water and wastewater flow rates. Different types of

water purification techniques such as filtration, activated carbon, biological treatment and

membranes can be applied independently or in combination (Tan et al., 2007). El-Halwagi et

al. (2003) and Prakash and Shenoy (2005) developed a material recovery pinch diagram

(MRPD) which is a rigorous graphical targeting approach based on well-established

composite curve. A detailed step by step procedure for constructing MRPD can also be found

in Foo (2013). It is useful in locating a material recycle/reuse pinch point; providing

insightful information on the use of fresh resources as well as the discharge of unused

materials. MRPD has traditionally been used internally within conventional industrial

facilities/plants for the design of water networks. As applied to a water network, the MRPD

can give insightful information about targets for fresh water resource (e.g. groundwater),

discharge of wastewater, location of water recycling/reuse pinch point and the relationships

between the water sources and the water sinks.

Figure C-1: Generic water pinch diagram

The water pinch point defines the overall bottleneck for maximum recovery among all the

water sources and sinks in the water network. Regeneration of water sources can be applied

based on the water composite curve in three different options:

(i) Regenerating water sources that are already above the pinch to a higher

quality/purity.

210

Source

Sink

Shifted source composite curve

Load

Flow rate

Pinch point

Maximum resource recovered

Min fresh resource Resource lost

(ii) Regenerating water source(s) that have purities below the pinch to a purity which

is above the pinch.

(iii) Regenerating water source(s) that have purities below the pinch to a purity which

is also below the pinch.

It is well established that regeneration option (ii) is the most resource beneficial one as it

removes water from a region of water surplus (below the pinch) and gives it back to a region

with deficit high purity water (above the pinch). It thus generates the maximum savings on

fresh water consumption and wastewater generation (Hallale, 2002). Regeneration below the

pinch does not affect the amount of fresh water consumption as it is taking water from a

region of water surplus to return it to the same region. This is comparable to the qualitative

rule proposed by Townsend and Linnhoff (1983) for the arrangement of heat pumps which

states that heat pumps for heat regeneration must be placed across the pinch in a system to

reduce utility consumption. This cross-pinch rule holds because there is always a net heat

source below the pinch point and a net heat sink above the pinch point and using a heat pump

for upgrading (i.e. regenerating) heat from intervals that are below the pinch to those above

the pinch can enhance the energy efficiency of a process.

The resource gain indicator can be used to help the decision making process of determining

the most resource efficient option for water regeneration among different alternatives for

initial COD of water sources, target COD and regeneration technology in the design of LIPS.

By using the resource gain indicator with the well-established pinch analysis technique, the

process integration stage of the insight-based design approach for LIPS can be systematically

optimised to ensure that the most efficient resource regeneration option is selected.

Furthermore, the amount of reference resource saved through regeneration can be determined

by reassessing the amount of fresh or imported resource consumption in correspondence to

the regeneration option by applying pinch analysis technique through the MRPD. For the case

of water, pinch analysis is applied again considering the identified regenerated stream to

determine the new minimum flow rate of freshwater to be imported into the system.

Using Equation (C.3) the different regeneration options can be generated by varying one of

the following parameters:

1) Initial quality of the regenerated resource, e.g. C cod , initial

2) Final concentration of the regenerated resource, e.g. C cod , final

211

3) Using different regeneration technologies that have different thermodynamic

efficiency or specific cumulative exergy cost.

Note that in a case where water cascading is adopted, if the unused streams after resource

cascading are used individually to produce the regeneration options, the initial quality of

these streams would be fixed by the terminating quality of cascaded use. If the streams are

combined, then the initial quality can be manipulated by varying proportions. Various

regeneration options can then be produced while keeping the source-demand balance and

selection of final qualities should follow the pinch rules.

212

Appendix DAppendix D contains all the data and assumptions used for demonstrating the insight-based

approach developed in Chapter 6 of the thesis on a case study for the design of an integrated

local food-energy-water production system for Whitehill and Bordon, an eco-town in the UK.

Note that Chapter 6 was carried out after Chapter 5 and thus some of the data and

assumptions used in the former chapter have been updated.

D.1 Initial design of food production subsystem

The food choices and demand considered in Chapter 6 and 7 for the design of the food

production subsystem are similar to those considered in Chapter 5. The specific cumulative

exergy values of imported bread, potatoes, beef and pork are given in Table D-1 along with

the data sources used for estimating these values. The specific cumulative exergy values for

imported bread, potatoes, beef and pork have been revised from the values determined in

Leung Pah Hang et al. (2016b) and updated according to a recent report by Behzadian et al.

(2016). The water consumption for wheat cultivation in the UK is lower than the global

average water consumption for wheat cultivation as the wheat yield is relatively high in the

UK. Furthermore, the specific cumulative exergy for imported potatoes was determined from

Chen and Chen (2009), and the cost of transportation for imported potatoes was calculated

assuming that they are imported from France which is the country from which the highest

quantity of potatoes are imported to the UK (AHDB, 2014).Note that imported food can also

be from another local production system within the same country but for practical reasons

and data availability, an external source was considered.

Table D-1: Cumulative exergy of imported food used in Chapter 6

Food product (Imported) Specific cumulative exergy (MJ/kg) Source

Bread 45IME (2014);

Nielsen and Nielsen (2003);Andersson and Ohlsson (1999)

Potatoes 4 Chen and Chen (2009)Beef 971 EA (2009), IME (2014)

Pork 380 IME (2014), Gerbens-Leenes et al. (2013); Williams et al. (2006)

Conventional sources of utility (electricity, heat and water) have been assumed in the initial

estimate of the cumulative exergy consumption for producing food locally. The cumulative

exergy values associated with these conventional utility sources are given in Table D-2 and

213

have been determined in Appendix B. The specific cumulative exergy values of resources

(i.e. per kg of the resource) other than utilities used for local food production are reported in

Table D-3. Furthermore, due to lack of reliable data, the total cumulative exergy consumption

for environmental remediation of pollutants and harmful effluents generated during the

manufacture of the imported food was not taken into account in the estimation of their

specific cumulative exergy consumption.

Table D-2: Specific cumulative exergy of conventional sources of water and energy

Conventional sources Specific cumulative exergyGroundwater 0.06 MJ/kg

Heat from natural gas boilers 2.01 MJ exergy/MJ heat energyGrid electricity 5.97 MJ exergy/MJ electrical energy

Table D-3: Specific cumulative exergy of resources for bread production

Resources Specific cumulative exergy SourcesFertiliser 32.7 MJ/kg N Wittmus et al. (1975)Pesticides 368 MJ/kg Özilgena and Sorgüven (2011)

Diesel 53.2 MJ/kg Szargut et al. (1988)Yeast 34.57 kJ/kg Marques et al. (1997)

The ecological layer of LIPSOM was first used to select the most resource efficient locally

available resources within ecological limits and the allocation of limited resources such as

land to different activities such as crop production and livestock breeding. Next, the

agricultural layer was used to reinforce decision making on what crops (i.e. wheat or

potatoes) and animals (i.e. cattle or pig) to produce locally and what are the agricultural

processing units required for crop plantation and animal breeding. The industrial layer then

decides on the most resource efficient industrial processing units for converting the

agricultural products. For the initial estimate of the cumulative exergy consumption for

locally producing each food type (CExC local using Equation (6.3) in the main text) imported

chemicals such as fertilisers and pesticides and imported animal feed are considered. It is

assumed that no animal feed is to be provided by purposely grown cereals from the local

area. This is because it is not yet known at this stage if local nutrients such as manure from

livestock rearing and agricultural residues from food crop cultivation will be available as

feedstock for crop cultivation and livestock rearing respectively. The specificities for

producing bread locally are given in Table D-4.

214

Table D-4: Specificities for local bread manufacture

Resources Quantity SourceWheat cultivation

Water 701.8 kg water/kg bread IME (2014)Heat for wheat storage 0.52 MJ/kg wheat stored Maier and Bakker-Arkema

(2002)Fodder from wheat 7 t/ha Cowell and Parkinson (2003)

Fertiliser 0.0264 kg N/kg bread DK (2014)Pesticides 1.566 kg/ha Audsley et al. (2009)

Diesel 0.0144 kg/kg bread DK (2014)Wheat processing into bread

Wheat yield 6.9 Mg dry matter/ha/year Williams et al. (2006)Water 1608 kg water/kg bread IME (2014)Heat 1 MJ/kg bread Nielsen and Nielsen (2003)

Electricity 0.388 MJ /kg bread Nielsen and Nielsen (2003)Yeast demand 0.0035 kg/kg bread Ruskins (2015)

Wheat crop to bread 1.14 kg bread/kg wheat crop DK (2014)

The specificities for locally producing potatoes are given in Table D-5. These data were used

to have an initial estimate of the cumulative exergy of locally producing potatoes to meet the

local demand in potatoes. The specificities for locally producing beef and pork are given

respectively in Table D-6 and Table D-7.

Table D-5: Specificities for local potatoes production

Resources Quantity SourcePotatoes cultivation

Annual potatoes yield 45 t/ha UK Agriculture (2014)Fertiliser 175.5 kg N/ha Williams et al. (2006)

Residue to potatoes ratio 1.62 t residue/t crop Jata et al. (2011)Electricity demand for irrigation 142 MJ/tonne Gulati and Singh (2011)Electricity for potatoes storage 270 MJ/t potatoes stored Kneeshaw (2006)Water demand for cultivation 18,000 L/t Gulati and Singh (2011)

Diesel 158 MJ/t Gulati and Singh (2011)Pesticides 0.11 kg/t Audsley et al. (2009)

Potatoes processingWater demand for processing 6000 L/t Gulati and Singh (2011)

The cumulative exergy consumption for locally producing each food type up to the point that

their local demand is satisfied and assuming conventional utility sources, CExC local and the

corresponding cumulative exergy consumption for importing them, CExC imp are compared. If

CExC local>CExC imp, import the food product. Otherwise, determine the specific resource gain

215

SRG of each food type using Equation (6.3) in Chapter 6 for the allocation of a (limited)

resource to different tasks (such as crop plantation or animal breeding). Using Equation (6.3),

CExCref is the cumulative exergy of the imported food product, CExCalt is the cumulative

exergy of the locally produced food product and F r is the amount of land allocated to each

task. A summary of the results is given in Table D-8.

Table D-6: Specificities for local beef manufacture

Resources Quantity SourceCattle breeding

Area required per tonne meat beef

4.7 ha/t Cowells and Parkinson (2003)

Manure per kg live cattle per year

25 kg/kg cattle/y USDA (2009)

Nitrogen content in manure 0.005 kg N/kg manure ECOCHEM, 2015Fodder required per live cattle

per day13.38 kg fodder/cattle/day Panday and Mishra (2011)

Protein required per amount of live cattle per year

1.44 t/t cattle/year Panday and Mishra (2011)

Cattle processing into beefBeef to cattle ratio 0.36 t/cattle Cowells and Parkinson (2003)

Water demand (including water demand for cattle breeding)

15,415 L/kg beef IME (2014)

Electricity 2.56 MJ electricity/kg cattle AHDB (2013)Heat 1195 MJ heat/kg cattle EA (2009)

Table D-7: Specificities for local pork manufacture

Resources Quantity SourcePig breeding

Pork to pig ratio 0.072 t/pig Cowells and Parkinson (2003)Area required per tonne meat

pork1.2 ha/t Cowells and Parkinson (2003)

Manure per kg live pig per year 25 kg/kg pig/y USDA (2009)Nitrogen content in manure 0.006 kg N/kg manure ECOCHEM, 2015

Protein required per amount of live pig per year

0.73 t/t pig/year FF (2015)

Fodder required per amount of live pig per day

0.01 t/t pig/day FF (2015)

Pig processing into pork

Water demand (including water demand for pig breeding)

5988 L/kg porkIME (2014)

Gerbens-Leenes et al. (2013)Heat 1.51 MJ/kg pork EA (2009)

Electricity 1.58 MJ/kg pork ERM (2009)

216

Table D-8: Specific resource gain of each food type

Food type CExC local (MJ/y) CExC imp (MJ/y) SRG(MJ/ha)Bread 5.08×106 6.02×106 5.53×104

Potatoes 1.70×106 1.79×106 9.29×103

Beef 3.49×106 3.51×106 1.34×103

Pork 5.37×106 5.38×106 1.02×103

The food product with the highest specific resource gain values in decreasing order were

bread, potatoes, beef and pork. The design heuristic rule is to adopt the option with the

highest specific resource gain (SRG) first to the extent where it becomes limited by some

constraints such as when the local demand is fully met and/or when all available land is

allocated. Following this design heuristic rule, bread is produced first. However, due to a

wheat crop yield of 6.9t/ha/y (Williams et al., 2006) and an agricultural land availability of 17

ha for the eco-town, only bread can be produced locally. About 117t/y of wheat crops can be

produced locally in summer; which is the harvesting season for wheat crops in UK (UK

Agriculture, 2014a) using the available agricultural land of 17 ha; from which 134 t/y of

bread can be produced locally per annum and a total of 87 t/y of wheat crop stored in summer

and autumn. Therefore, the initial design indicates that 60% of the bread demand can be

satisfied locally while all other food demands need to be imported to the eco-town due to

limited agricultural land availability. The result of the initial design for the local food

production system is given in Table D-9.

Table D-9: Initial design of food production subsystem

Food type Source Quantity Process/activity in the local production system

Bread

Locally produced bread 134 t -Imported bread 90 t -

Imported fertilisers 3296 kg N FertilisationGroundwater 66,451 t Irrigation for wheat cultivation

Diesel 1930 kg Land preparation-trackersImported pesticides 27 kg Pesticide application

Heat from natural gas boilers) 45,340 MJ Standard storage system for

wheat cropsGroundwater 12,060 t

Assuming standard industrial processing units for production

of bread from wheat

Heat from natural gas boilers 134, 000 MJ

Grid electricity 51,992 MJImported yeast 469 MJ

Potatoes Imported potatoes 1793 GJ (CExC) -Beef Imported beef 86,448 GJ (CExC) -Pork Imported pork 17,480 GJ (CExC) -

217

Different food processing technologies would have different processing efficiencies and

therefore different levels of operating resource consumption. For the purpose of illustrating

the developed approach in this study, standard food processing technologies to process wheat

to bread have been adopted. Wheat is harvested in summer and planted in spring in the UK

(UK Agriculture, 2014). Most of the water for wheat cultivation is required in spring. About

117t of wheat crops are harvested in summer and stored for production of bread in other

seasons. The amount of bread produced locally and imported per season is summarised in

Table D-10. Note that external supply of agricultural crops has not been considered.

Table D-10: Amount of locally produced and imported bread per season

Season Locally produced (t) Imported (t)Summer 56 0Autumn 56 0Winter 21.7 34.3Spring 0 56

D.2 Initial design of water production subsystem

For the initial design of the water production subsystem, the two sources of freshwater

available were groundwater and rainwater. The water sinks considered were bread

manufacture from the outcome of the initial design of the food subsystem and the residential

sector in the eco-town. As only the first 3 layers of LIPSOM are considered in the generation

of a base design of LIPS, the reuse of regenerated wastewater from residential sector and

bread production to satisfy water sinks are not considered at this stage of design.

The capital cumulative exergy consumption per year for rainwater harvesting system was

determined to be 0.11 MJ/kg of rainwater supplied; a value determined from Envirowise

(2016) which reported a total capital cost of £250,000 for a storage tank of 303,000 litres and

assuming that the rainwater harvesting system has a service life of 25 years. The operational

and maintenance resources are reported to be negligible for rainwater harvesting system

(UNEP, 2016b). The capital cumulative exergy consumption for supplying groundwater per

year was estimated to be 0.045 MJ/kg based on estimates of resource consumption for

groundwater extraction and purification (Lyons et al., 2014) while that for wastewater

treatment plant was determined to be 0.66 MJ/kg (GBRA, 2016); higher than that for

groundwater since more rigorous purification technologies such as activated sludge process

are required to treat the more polluted wastewater. The parameters that were taken into

218

consideration for determining the cumulative exergy consumption for treating groundwater

are summarised in Table D-11.

Table D-11: Parameters for treating groundwater

Parameters Unit SourcesElectricity for treating

groundwater 0.0014 kWh/g COD removed Saghafi et al. (2015)

Electricity for pumping groundwater 0.00198 MJ/l King et al. (2008)

Heat for treating groundwater 0.0003 kWh/g COD removed EPA (2011), Singh et al. (2012)

Chemicals for treating groundwater 0.49 kg CaCO3/kg COD removed Souza (1986)

Capital resources 0.045 MJ (exergy)/kg groundwater Lyons et al. (2014)

The specific cumulative exergy for supplying treated groundwater was estimated to be 0.06

MJ/kg water. The quality limits acceptable in terms of COD by the water sinks in the initial

design of the water subsystem are summarised in Table D-12. It is assumed that the

maximum environmental discharge limit is 0.1 g COD/kg water.

Table D-12: Quality of water sinks

Water sinks Quality (mg/L) SourcesWheat cultivation 75 Henry (2015)

Wheat processing into bread 10 FSHT (2014)Residential 10 FSHT (2014)

Rainwater has a COD concentration level of 0.006 g COD/kg water, which is lower than the

highest quality required by the water sinks at 0.01 g COD/kg water and much lower than the

environmental discharge limit of 0.1 g COD/kg water. As such, it does not require any

treatment before being discharged into the environment or treatment for COD removal before

being used to supply the water sinks. Since the other characteristics (such as pH, total

dissolved solids) of rainwater were not considered in this work, the potential uses of

rainwater in the initial design of the water subsystem were limited only to wheat cultivation

and non-potable domestic water uses (i.e. excludes rainwater use for food processing and

water for drinking purposes). Using Equation (6.1) in Chapter 6 and taking the groundwater

as the reference resource, the resource gain for rainwater was determined to be -0.048 MJ/kg.

The negative RG was due to the relatively high capital resources to implement a localised

rainwater harvesting system leading to groundwater being overall a more resource efficient

219

alternative for water supply. Furthermore, groundwater has also a relatively high abstraction

limit of 14,875,942t/y in the eco-town (Whitehill and Bordon, 2012) and can thus be sensibly

used to satisfy all or most water sinks within its abstraction limit. The initial design of the

water production subsystem per season is summarised in Table D-13.

Table D-13: Initial design of the water production subsystem

Processes Spring (t) Autumn (t) Summer (t) Winter (t) SourceBread manufacture 66,312 1462 1462 567 Groundwater

Residential 586,867 586,867 586,867 586,867 Groundwater

The total amount of wastewater generated from bread manufacture and residential sector

given in Table D-14 is simply treated before discharge into the environment.

Table D-14: Wastewater generated from initial design of water subsystem

Processes Wastewater generated (t)Spring Autumn Summer Winter

Bread manufacture 0 80 80 31Residential 365,126 365,126 365,126 365,126

D.3 Initial design of energy production subsystem

The resource gain of each technology, which is determined by the difference between the

specific cumulative exergy of the reference technology and that of the alternative technology,

as formulated in Equation (6.4) in the main text, is summarised in Table D-15. The resource

gain for electricity generation from alternative energy sources such as wood chip biomass

CHP, organic waste CHP, natural gas CHP, solar and wind was determined with the

reference resource being grid electricity. On the other hand, the resource gain for heat

generation from alternative sources such as heat from wood chip biomass boiler, wood chip

biomass CHP, organic waste CHP and natural gas CHP was determined with the reference

resource being heat produced from natural gas boilers. The efficiency of natural gas boiler

was assumed to be 0.90 while that of wood chip biomass boiler was taken to be 0.85. The

efficiencies used for the other technologies are given in Leung Pah Hang et al. (2016b). The

relative high cumulative exergy consumption for producing electricity from wood chip

biomass CHP was due to the low electrical efficiency of 19% of the CHP (Whitehill and

Bordon, 2012). The organic waste includes food waste and garden waste and has an

availability of 99,000 GJ per annum (Whitehill and Bordon, 2012). Any animal waste and

agricultural residues produced can be considered at this stage of the design as by-products of

the food production subsystem that contribute to the availability of organic waste; which are

220

potential energy sources to the base design of the energy subsystem. Note that the allocation

of CExC to heat and electricity for CHP was done based on its heat to power ratio.

For the design of the energy subsystem, using the design rule of resource gain will result in

many possible options. This is a classic issue for any CHP design problem as the energy

option for combined heat and power (CHP) has two resource gain values; one for heat and

one for power generation. Decision on whether the CHP option with the highest electricity

gain or highest heat gain should be chosen depends on the preference of decision makers and

their core business as well as the relative levels of local heat and power demands and the

match of these to outputs from the CHP. A linear programming (LP) optimisation similar to

the energy optimisation model developed in Leung Pah Hang et al. (2016b) is formulated to

generate a fast optimum design option for the energy subsystem. The objective of the LP

optimisation problem for the design of the energy subsystem is to minimise the net total

cumulative exergy consumption meeting the local energy demand, comprising resource

consumption associated with raw material, capital and operating resources minus the

cumulative exergy consumption avoided by exporting any surplus local power generation to

the grid. Capital resources (i.e. those consumed for building equipment and production

facilities) for CHPs, wind turbines and solar panels were included as these technologies

consume relatively negligible operating resources, making their capital resources relatively

more significant. The results of the LP optimisation for the initial base design of the energy

subsystem are given in Table D-16.

Table D-15: Cumulative exergy consumption of associated energy technology

Electricity sourcesCumulative exergy

consumption (MJ/MJ) Resource gain (MJ/MJ)

Grid electricity 5.97 0Electricity from biomass CHP 5.34 0.63Electricity from organic waste

CHP0.58 5.39

Electricity from natural gas CHP 4.75 1.22Electricity from solar 1.42 4.55Electricity from wind 2.23 3.74

Heat from natural gas CHP 3.37 -1.36Heat from natural gas boiler 2.01 0

Heat from biomass boiler 1.23 0.78Heat from biomass CHP 2.01 0

Heat from organic waste CHP 0.41 1.60

221

Table D-16: Initial base design of energy subsystem

Source Sink Energy supply (GJ)Winter Spring Summer Autumn

Total electricity production mixBiomass CHP:

49.2%Wind: 19.9%Solar: 17.2%Organic CHP:

13.7%

WaterProcesses 1262 1277 1263 1263

Residential 22,566 22,566 22,566 22,566Food

Processes 8 0 22 22

Grid(export of surplus

electricity)45,830 33,039 34,955 39,972

Total heat production mix

Biomass CHP: 87%Organic CHP: 13%

Water processes 281 281 285 281Residential 112,128 85,848 82,344 96,360

Food processes 22 0 91 66

D.4 Iterative design of local production system

D.4.1 1st iterative design of food subsystem

Based on the algorithm for the sequential synthesis of multiple subsystems, the results of the

initial design of the energy subsystem are then used to design the food and water subsystems

again in the 1st iterative round of design (i.e. i=1). The resource gain for the different food

product options were re-assessed based on the lower energy cumulative exergy costs of 1.80

MJ/MJ heat and 1.90MJ/MJ electricity but similar cumulative exergy for water. The results

are summarised in Table D-17 in decreasing order.

Table D-17: Specific resource gain of food products for 1st iterative design of food subsystem

Food product SRG(MJ/ha)Bread 6.99×104

Potatoes 9.29103

Pork 6.64×103

Beef 3.63×103

Since bread still has the highest resource gain as compared to other food products, the 1st

iterative design involves only its production locally similar to the initial design of the food

production system. Table D-18 summarises the outcome of the 1st iterative design of the food

subsystem and its new energy sources. The decrease in specific CExC for heat and electricity

of 10% and 68% respectively led to pork being more resource efficient to produce locally

than beef as compared to the initial design of the food production subsystem (i.e. i=0).

Though beef production consumes more heat and electricity per unit beef, the total amount of

energy required to produce pork locally exceeds that for beef as the quantity of pork that can

be produced locally based on the land available is higher than that for beef with a pork to land

222

ratio of 1.2 ha/tonne as compared to beef with 4.7 ha/tonne (Cowells and Parkinson, 2003).

Thus, local pork production in the eco-town is very sensitive to a change in specific CExC of

heat and electricity as compared to beef.

Table D-18: 1st iterative design of food subsystem

Food type Source Quantity Process/activity in the local production system

Bread

Locally produced bread 134 t/y -Imported bread 90 t/y -

Imported fertilisers 3296 kg N Fertilisation

Groundwater 66,451 t Irrigation for wheat cultivation

Diesel 1930 kg Land preparation-trackersImported pesticides 27 kg Pesticide application

Heat from wood chip biomass CHP (87%) and

organic waste CHP (13%)45,340 MJ Storage of wheat crops

Groundwater 12,060 t

Assuming standard industrial processing units for conversion of wheat to

bread

Heat from wood chip biomass CHP (87%) and

organic waste CHP (13%)134, 000 MJ

Electricity from wind (52%), solar (47%) and

wood chip biomass CHP (1%)

51,992 MJ

Imported yeast 469 MJPotatoes Imported potatoes 4. 1793 GJ (CExC) -

Beef Imported beef 86,448 GJ (CExC) -Pork Imported pork 17,480 GJ (CExC) -

D.4.2 1st iterative design of water subsystem

With more resource efficient energy sources, the specific CExC for groundwater reduces to

15% of the original value of 0.06 MJ/kg while the resource gain for rainwater was determined

to be -0.057 MJ/kg. Thus, it is still more resource efficient to use groundwater rather than

rainwater to satisfy water demands in the eco-town; similar to the initial design of the water

subsystem. However, the design of the water subsystem will change as compared to the initial

design with new water demand and wastewater generated from the energy processes which

respectively become new water sinks and water sources to the water subsystem. Since the

total amount of water demand is significantly less than the abstraction limit for groundwater

in the eco-town, groundwater is used to satisfy the new water demands including water

demand for residential, food and energy production for all seasons as given in Table D-19.

223

Table D-19: 1st iterative design of water subsystem

Water sinks Quantity (t)Spring Autumn Summer Winter

Bread manufacture 66,313 1,462 1,462 567

Residential 586,867 586,867 586,867 586,867Energy

production 12,532 14,067 12,039 16,348

With new energy processes in the LIPS (note that imported electricity and heat from natural

gas boilers were used to meet the energy requirements of food and water subsystem in the

initial design of LIPS), the total amount of wastewater to be treated per season increases as

given in Table D-20.

Table D-20: Wastewater generated from 1st iterative design of water subsystem

Processes Wastewater generated (t)Spring Autumn Summer Winter

Bread manufacture 0 80 80 31Residential 365,126 365,126 365,126 365,126

Energy production from wood chip biomass CHP

9,151 10,458 8,657 12,401

Energy production from organic waste

CHP1,518 1,518 1,592 1,518

D.4.3 1st iterative design of energy subsystem

Furthermore, additional energy is required to pump and treat groundwater to satisfy the water

demands of the energy production processes from the initial design of the energy subsystem.

Energy required to treat the wastewater generated from these energy production processes

also needs to be taken into consideration in the 1st iterative design of the energy subsystem.

With these new energy demands, the results of the LP energy optimisation model for the 1st

iterative design of the energy subsystem are summarised in Table D-21.

224

Table D-21: 1st iterative design of energy subsystem

Source Sink Energy supply (GJ)Winter Spring Summer Autumn

Total electricity production mixBiomass CHP:

49.3%Wind: 19.9%Solar: 17.2%Organic CHP:

13.7%

WaterProcesses 2294 2066 2021 2150

Residential 22,566 22,566 22,566 22,566Food

Processes 8 0 22 22

Grid(export of surplus

electricity)46,058 30,052 35,139 40,178

Total heat production mixBiomass CHP:

87%Organic CHP:

13%

Water processes 510 456 453 478Residential 112,128 85,848 82,344 96,360

Food processes 22 0 91 66

D.4.4 2nd iterative design of local production system

Following the same design sequence, a second round of design iteration was carried out

starting with the food subsystem where the specific cumulative exergy consumption for water

(i.e. 0.051 MJ/kg) and energy sources determined from the previous iteration (i.e. i=1) is used

(i.e. 2.02 MJ/MJ electricity and 1.80 MJ/MJ heat). Similar to the initial and 1st iterative

design of the food subsystem, bread has the highest specific resource gain at 1.06×105 MJ/kg,

followed by pork at 5.14×104 MJ/kg, beef at 3.31×104 MJ/kg and potatoes at 9.29×103 MJ/kg.

Further analysis indicates that if the specific CExC of water remains unchanged at 0.06

MJ/kg, the SRG values of the food types follow the same pattern (

SRGbread>SRGpotatoes>SRGpork>SRGbeef ¿ as in its 1st iterative design. With a 15% decrease in

specific CExC of water, pork and beef lead to more resource savings than potatoes which

becomes the least resource efficient food option to produce locally. This suggests that pork

and beef productions are very sensitive and significantly dependent on water inputs to the

food production subsystem. Both pork and beef production consume relatively high volumes

of water at 5988 L/kg pork (IME, 2014, Gerbens-Leenes et al., 2013) and 15,415 L/kg beef

(IME, 2014). On the other hand, only decreasing the value of specific CExC of electricity by

63% while keeping the values of specific CExC of water and heat constant from their

respective values in the previous design (i =1) will still lead to potatoes being more resource

efficient to produce than pork and beef. Local potato production is relatively dependent on

electricity especially as they need to be stored in a cool dry place.

225

Furthermore, with new electricity source mix but unchanged heat sources from the previous

design of the energy subsystem (i.e. i=1), the specific CExC of groundwater remains

unchanged; indicating that electricity is not a significant contributor to resource consumption

for the supply of groundwater. Though groundwater is still better option relative to using

localised rainwater harvesting system, the design of the water subsystem will change again in

its 2nd iterative design with new water demands and amount of wastewater generated from the

energy production processes of the previous iterative design (i.e. i=1). Due to the relatively

high availability of groundwater in the eco-town, it remains to supply all water demands for

each season as shown in Table D-22.

Table D-22: 2nd iterative design of water subsystem

Water sinks Quantity (kg)Spring Autumn Summer Winter

Bread manufacture 66,312,740 1,461,600 1,461,600 566,944

Residential 586,867,440 586,867,440 586,867,440 586,867,440Energy

production 12,557,426 14,096,068 12,063,937 16,381,381

Total amount of wastewater generated per season increases moderately as per Table D-23

with a slight increase in energy demand from the energy subsystem as indicated in Table C-

24.

Table D-23: Wastewater generated from 2nd iterative design of water subsystem

Processes Wastewater generated (kg)Spring Autumn Summer Winter

Bread manufacture 0 79,520 79,520 30,845Residential 365,125,560 365,125,560 365,125,560 365,125,560

Energy production from wood chip biomass CHP

9,173,111 10,483,036 8,678,475 12,428,641

Energy production from organic waste

CHP1,517,670 1,517,670 1,592,174 1,517,670

226

Table D-24: 2nd iterative design of energy subsystem

Source Sink Energy supply (GJ)Winter Spring Summer Autumn

Total electricity production mixBiomass CHP:

49.3%Wind: 19.9%Solar: 17.2%Organic CHP:

13.7%

WaterProcesses 2296 2068 2023 2152

Residential 22,566 22,566 22,566 22,566Food

Processes 8 0 22 22

Grid(export of surplus

electricity)46,058 30,052 35,139 40,178

Total heat production mixBiomass CHP:

87%Organic CHP:

13%

Water processes 511 457 454 479Residential 112,128 85,848 82,344 96,360

Food processes 22 0 91 66

A third round of design iteration was carried out, which led to a stabilised solution (i.e. the 3 rd

iterative design of each subsystem (i.e. i=3) gave similar design outcome to the 2nd iteration

(i.e. i=2). Note that the iteration stops when the demands in each subsystem and the

cumulative exergy consumption to produce the streams exchanged (i.e. heat, electricity and

water) between the three subsystems remain unchanged; meaning that no change in the

design of the subsystem is incurred by further iterations according to the algorithm in

sequential synthesis of multiple subsystems. Since the 3rd iterative design did not incur any

further change in the design of each subsystem with the convergence criteria of |Ci˗Ci-1|/ (Ci-1)

≤ 0, the iteration stops and the final design of a basic food-energy-water local production

system is obtained. The final sinks and sources in the local production system are similar to

those determined in the initial design. However, the quantity of these flows, summarised in

Table D-25, have altered throughout the iterative design before stabilising at the 3rd iteration.

227

Table D-25: Base design of local production system

Source Sink Locally produced food (t)Winter Spring Summer Autumn

Local bread Local consumption 22 0 56 56

Source Sink Water supply (t)Winter Spring Summer Autumn

GroundwaterWater flows @

≤0.010 g COD/kg

Residential 586,867 586,867 586,867 586,867Food processes (cultivation and

processing)567 66,312 1462 1462

Energy processes 16,381 12,557 12,064 14,096

Water flows @ ≤0.10 g COD/kg Discharge 379,103 375,816 375,476 377,206

Source Sink Energy supply (GJ)Winter Spring Summer Autumn

Total electricity production mixBiomass CHP:

49.3%Wind: 19.9%Solar: 17.2%Organic CHP:

13.7%

Waterprocesses 2296 2068 2023 2152

Residential 22,566 22,566 22,566 22,566Food

processes 8 0 22 22

Grid(export of surplus

electricity)46,058 30,051 34,946 40,178

Total heat production mix

Biomass CHP: 87%Organic CHP: 13%

Water processes 511 457 454 479Residential 112,128 85,848 82,344 96,360

Food processes 22 0 91 66

D.5 Process integration of food-energy-water local production system

After the base design of the local production system is generated, the system is optimised by

considering integration options for resource reuse, regeneration and/or options for re-

purposing. The following sections D.5.1 and D.5.2 illustrate integration options for water and

energy resources, respectively.

D.5.1 Integration options for water reuse and regeneration

The two sources of fresh water available in the eco-town are groundwater and rainwater.

However, from the application of the incremental approach to generate a basic design of the

water subsystem, groundwater was found to be a better alternative than rainwater given its

availability and lower cumulative resource exergy consumption to supply water demands.

Water pinch can be applied to optimise the supply of water sources to the water sinks in the

eco-town by reuse and regeneration, with the potential to reduce the amount of fresh

228

groundwater required, if resource gains can be achieved. Water pinch should be applied

across all the four seasons to optimise water supply for the whole year due to varying water

demands and availability of water sources in different seasons. The sources of water available

for integration between the different subsystems and their corresponding quality are

illustrated in Table D-26. In the base design of the local production system, the water sources

considered are groundwater and rainwater, both of which are fresh resources. In the water

pinch analysis the focus is on the re-use of the wastewater sources in order to minimise the

consumption of the two fresh resources.

Table D-26: Availability of water sources and their quality

Sources

Winter Summer Autumn Spring Quality (g

COD/kg water)

ReferenceFlow rate (kg/season)

Residential wastewater 365,125,560 365,125,560 365,125,560 365,125,560 0.75

Henze and Comeau (2008)

Wastewater from bread production

1581 79,520 79,520 0 0.81 Chen et al. (2006)

Wastewater from energy production

from organic waste CHP

1,517,670 1,592,174 1,517,670 1,517,670 0.441 GCCSI (2016)

Wastewater from energy production from wood chip CHP

12,428,641 8,678,475 10,483,036 9,173,111 0.441 GCCSI (2016)

The sinks of water available per season and their quality are illustrated in Table D-27.

229

Table D-27: Water sinks and their quality

SinksWinter Summer Autumn Spring Quality

(g COD/kg)Flow rate (kg/season)

Wheat cultivation 0 0 0 66,312,740 0.075

Wheat processing into

bread566,944 1,461,600 1,461,600 0 0.01

Residential 586,867,440 586,867,440 586,867,440 586,867,440 0.01Energy

production from wood chip CHP

14,598,721 10,193,765 12,313,408 10,774,766 0.06

Energy production

from organic waste CHP

1,782,660 1,870,172 1,782,660 1,782,660 0.06

The quality of the water demand for wheat cultivation is higher than the quality of the water

demand for the other water sinks. In this particular case study, as all the water sources have a

COD level much higher than the COD requirements of any water sinks, from a quality

perspective the cascade use of the water sources through the application of pinch analysis

would result in no possible recovery. As direct re-use is infeasible, options of regeneration (to

enable reuse) need to be assessed.

The resource gain for regenerating the different water sources of residential wastewater,

wastewater from food production, wastewater from energy production up to the desired

quality of the water sinks were assessed using Equation (6.5) in the main text, assuming that

the reference resource is treated groundwater. From the base design, the specific CExC of

groundwater was assumed to be 0.051 MJ/kg. The resource gain for regenerating the

wastewater considers (i) the environmental remediation resource cost for discharging the

wastewater into the environment (environment discharge COD limit is 0.1 g COD/kg water)

if it was not reused and (ii) the cumulative exergy consumption (i.e. capital and operational

resources) for regenerating the wastewater for reuse. Energy required by various treatments

was assumed to be supplied by the energy sources from the basic design of the local

production system with specific cumulative exergy of 2.02 MJ/MJ for electricity and 1.80

MJ/MJ for heat. It is also assumed that the amount of groundwater replaced is the same as the

amount of wastewater treated for reuse.

230

The results of the resource gain for each of the water sources considered are presented in

Table D-28. The resource gain per kg water was determined to be similar and positive for all

the water sources. This is not surprising since the environmental discharge limit (i.e. 0.1 g

COD/kg) and the COD limits to which the water sources were regenerated were similar for

all the water sources. Thus, the order that the regenerated water sources are matched to the

water sinks is not important. The net specific cumulative exergy of the water sources after

regeneration at 0.075 g COD/kg, 0.06 g COD/kg and 0.01 g COD/kg was determined to be

0.426 kJ/kg, 0.681 kJ/kg and 1.533 kJ/kg respectively based on the net difference between the

specific cumulative exergy of regeneration and environmental remediation.

Table D-28: Resource gain of water sources after regeneration

Alternative sourcesResource gain (MJ/kg)

Water sink @ 0.075 g COD/kg

Water sink @ 0.06 g COD/kg

Water sink @ 0.01 g COD/kg

Treated residential wastewater (@0.75 g

COD/kg0.0512 0.0510 0.0501

Treated wastewater from bread manufacture @

0.81g COD/kg0.0512 0.0510 0.0501

Treated wastewater from energy production from wood chip biomass CHP

@ 0.441 g COD/kg

0.0512 0.0510 0.0501

Treated wastewater from energy production from organic waste CHP @

0.441 g COD/kg

0.0512 0.0510 0.0501

The use of the regenerated wastewater is limited to wheat cultivation, non-potable residential

purposes and for energy production. Drinking water usually constitutes 4% of total domestic

water use in the UK (WATERWISE, 2012) and it is assumed that the rest of the domestic

water use can be supplied by treated wastewater. The assumptions made in this case study on

possible water reuse/regeneration represent some extremes of what might be considered

acceptable socially in order to explore the technical possibilities of such a design system

locally. Thus, about 563,393 t of residential water per season can be potentially supplied by

regenerated wastewater. Furthermore, treated or regenerated wastewater is not used for food

processing in this case study; in line with health and safety regulations (UN Water, 2013).

The total water demand for the eco-town was determined to be 2,472,371t/y and about 96%

of it can be potentially satisfied by sources other than fresh groundwater such as the

231

regenerated water sources. Due to its availability in the LIPS, it was determined that only

about 61% of the eco-town’s water demands can be satisfied by regenerated water sources

with the rest being satisfied by groundwater.

As treated wastewater is not allowed to satisfy water demand for industrial food production,

groundwater supplies it with an associated CExC of 0.051 MJ/kg determined from the base

design of the local production system. Thus, the food processing component of the food

subsystem will be unaffected by the results of process integration for water resource.

However, it should be noted that with new water sources for supplying water sinks of wheat

cultivation at 0.075 g COD/kg water, energy production processes at 0.06 g COD/kg water

and non-potable residential uses at 0.01 g COD/kg water; the average specific CExC of water

supply for these water sinks has been reassessed to be about 0.02 MJ/kg which is a significant

61% decrease from the specific CExC determined from the base design of the local

production system.

D.5.2 Integration options for energy reuse

The stream data collected for each season for pinch analysis are given in Tables D-29-D-32.

A stream data gives information about the properties such as temperature of the flow

(commonly referred to as stream in pinch analysis). Low temperature waste heat available

from energy production from organic waste CHP and wood chip biomass CHP can be used to

supply part of the heat energy demand in the eco-town and reduce consumption from other

energy sources. The low temperature heat available from bread production was determined to

have an average heat load of 0.0689 kW, which was discarded in the pinch analysis as it is

too insignificant to be considered for heat recovery. The quantity of heat available from the

different sources will vary across the seasons. The heat demands considered were heat for

industrial bread production, for wheat storage, and the wastewater treatment plant and

residential heat requirements. Note that the minimum temperature difference between the

heat sources and sinks at any point during the heat exchange is taken to be 10 ºC. The grand

composite curves for each season are given in Figures D-1-D-4. As can be seen from the

grand composite curves, the pinch point occurred at 30 ºC.

232

Table D-29: Stream data for winter

Streams Supply temperature (ºC)

Target temperature (ºC) Heat Flow (kW) Stream type

Low temperature waste heat from organic waste

CHP

120 30 38.2 Hot

Low temperature waste heat from

wood chip biomass CHP

120 30 320 Hot

Heat demand for residential purposes

20 62.5 3556 Cold

Heat demand for industrial bread

production20 220 0.689 Cold

Heat demand for wastewater treatment

20 35 28.3 Cold

Figure D-1: Grand composite curve for winter

233

Table D-30: Stream data for summer

Streams Supply temperature (ºC)

Target temperature (ºC) Heat Flow (kW) Stream type

Low temperature waste heat from organic waste

CHP

120 30 38.2 Hot

Low temperature waste heat from

wood chip biomass CHP

120 30 225.7 Hot

Heat demand for industrial bread

production20 220 1.78 Cold

Heat demand for wheat storage 20 62.5 1.12 Cold

Heat demand for residential purposes

20 62.5 2611 Cold

Heat demand for wastewater treatment

20 35 24.9 Cold

Figure D-2: Grand composite curve for summer

234

Table D-31: Stream data for autumn

Streams Supply temperature (ºC)

Target temperature (ºC) Heat Flow (kW) Stream type

Low temperature waste heat from organic waste

CHP

120 30 38.2 Hot

Low temperature waste heat from

wood chip biomass CHP

120 30 270 Hot

Heat demand for residential purposes

20 62.5 3056 Cold

Heat demand for industrial bread

production20 220 1.78 Cold

Heat demand for wheat storage 20 62.5 0.314 Cold

Heat demand for wastewater treatment

20 35 26.5 Cold

Figure D-3: Grand composite curve for autumn

235

Table D-32: Stream data for spring

Streams Supply temperature (ºC)

Target temperature (ºC) Heat Flow (kW) Stream type

Low temperature waste heat from organic waste

CHP

120 30 38.2 Hot

Low temperature waste heat from

wood chip biomass CHP

120 30 237 Hot

Heat demand for residential purposes

20 62.5 2722 Cold

Heat demand for wastewater treatment

20 35 25.2 Cold

Figure D-4: Grand composite curve for spring

After performing pinch analysis and determining the targets of minimum heating and cooling

requirements for each season, the heat exchanger network for waste heat recovery was

synthesized for each season. It was found that 3 heat exchangers placed above the pinch are

required for winter, summer and autumn and 2 heat exchangers placed above the pinch are

required for spring. In the event that different numbers of heat exchangers are obtained for

different seasons, the following design steps and rules, based on resource gains and adapted

to heat exchangers, can be used to aid decision making in choosing the most resource

efficient heat exchanger network option:

236

Step 1: Determine the net saving for hot and cold utilities for adopting each number of heat

exchangers across all seasons throughout the year.

Step 2: Based on step 1, determine the resource gain for installing a particular number of heat

exchangers using Equation (6.4) in Chapter 6 where CExCref is the total cumulative exergy of

providing the recovered heat from conventional sources such as natural gas boilers and

CExCalt is the total cumulative exergy associated with the recovered heat.

CExCalt as applied to the heat exchanger system will comprise mainly of their total capital

resource cost. Note that the recovered heat stream is assumed to have zero cumulative exergy

consumption associated with it as it a waste resource, similar to agricultural residues.

Rule: The number of heat exchangers that gives the highest resource gain for heat recovery

for the whole year is the optimum number of heat exchangers that should be adopted.

Through the heat integration analysis, it was found that there are no external cooling

requirements for any seasons as maximum recovery of the available cold utilities from the

cold streams is possible. Using the developed design rules for heat exchangers in the process

integration stage of the insight-based design approach, it was found that adopting 3 heat

exchangers will allow for maximum heat recovery in all four seasons as that would maximise

the resource gain across the year. The total heat recovered with 3 heat exchangers was found

to be 2.79×107 MJ. From Perera et al. (1998), the specific capital resource cost was found to

be 3.3×10-4 MJ/MJ heat for a shell and tube heat exchanger having a service life of 5 years.

Taking a specific CExC of 2.01 MJ/MJ for producing heat from natural gas boilers (Leung

Pah Hang et al., 2016b), the total cumulative exergy consumption for producing 2.79×107 MJ

heat was calculated to be 5.61×107 MJ while the total cumulative exergy of providing the

recovered heat was found to be 9206 MJ mainly from the capital resource cost of the heat

exchangers. Therefore, the resource gain was determined to be 5.61×107 MJ. The results for

heat recovery synthesis for each season are tabulated in Table D-33.

The total net saving in heat load for a year was determined to be 2.79×107MJ. With a total

heat load of 3.80×108 MJ, the reuse of low temperature waste heat contributes to satisfying

about 10% of the total local heat demand in the eco-town. The proportion of high and

medium temperature heat produced from wood chip and organic waste CHPs in the heat

237

energy supply mix decrease from 87% and 13% to 78.4% and 11.7% respectively. With this

new heat energy supply mix, the average specific CExC of heat was re-evaluated at 1.62

MJ/MJ; which is a 10% decrease from its value from the outcome of the base design of the

local production system. However, this new specific CExC for heat did not have any impact

on the order of the specific resource gain for the different food options considered in the

design of the food subsystem and also did not have any repercussion on the water subsystem.

In conclusion, process integration did not cause any change in the choice of resources and

technologies of the base design for this particular case study but has made it more resource

efficient; significantly decreasing the amount of fresh water and energy fresh resources and

the capacity of energy generating technologies.

Table D-33: Heat recovery for each season

Season Stream Minimum Temperature

Maximum temperature Heat load (MJ)

Winter

Heat for bread production 40 220 19,550

Heat for residential 24.3 62.5 1.01×108

Heat for wastewater treatment

20 35 893,282

Total minimum heat load 1.02×108

Total heat load 1.13×108

Net heat load 1.12×107

Autumn

Heat for bread production 27.8 220 53,829

Heat for wheat storage 20 60 9908

Heat for residential 24.3 62.5 8.66×107

Heat for wastewater treatment

20 35 835,598

Total minimum heat load 8.75×107

Total heat load 9.73×107

Net heat load 9.72×106

Summer

Heat for bread production 27.8 220 53,827

Heat for wheat storage 20 60 35,452

Heat for residential 24.3 62.5 7.40×107

Heat for wastewater treatment

20 35 783,715

Total minimum heat load 7.49×107

238

Total heat load 8.32×107

Net heat load 8.32×106

Spring

Heat for residential 24.3 62.5 7.72×107

Heat for wastewater treatment

20 35 795,346

Total minimum heat load 7.80×107

Total heat load 8.66×107

Net heat load 8.67×106

239

ReferencesAdriaanse, A.; Bringezu, S.; Hammond, A.; Moriguchi, Y.; Rodenburg, E.; Rogich, D.;

Schutz, H. (1997) Resource Flows: The Material Basis of Industrial Economies. World

Resource Institute. Washington D.

Agriculture and Horticulture Development Board (AHDB), (2013) UK Yearbook 2013 –

Cattle [Online], Available from:

http://www.eblex.org.uk/wp/wp-content/uploads/2014/02/m_uk_yearbook13_Cattle110713.p

df, [Accessed 17 September 2015]

Agriculture and Horticulture Development Board (AHDB), (2014) UK Pig Meat Imports

[Online], Available from: http://pork.ahdb.org.uk/prices-stats/imports-exports/uk-pig-meat-

imports/ [Accessed 30 October 2014]

Akbari, A.A. and Karimi, B. Int J Adv Manuf Technol (2015) 79: 229. doi:10.1007/s00170-

015-6796-9

Allwood, J.M, Ashby, M.F, Gutowski, T.G., Worrell, E., (2011) Material efficiency: a white

paper. Resources, Conservation and Recycling 2011, 55(3):362–81.

Almansoori, A. and Shah, N. (2012) Design and operation of a stochastic hydrogen supply

chain network under demand uncertainty, International Journal of Hydrogen Energy, 37,

3965-3977

Amini, S.H, Remmerswaal, J.A.M, Castro, M.B, Reuter, M.A (2006) Quantifying the quality

loss and resource efficiency of recycling by means of exergy analysis, Journal of Cleaner

Production 15 (2007) 907-913

Andersson, K and Ohlsson, T. (1999) Life Cycle Assessment of Bread Produced on Different

Scales, International Journal of LCA, 4 (1) 25-40.

240

Audsley, E., Stacey, K., Parsons, D.J., Williams, A.G. (2009) Estimation of the greenhouse

gas emissions from agricultural pesticide manufacture and use [Online], Available from:

dspace.lib.cranfield.ac.uk/bitstream/1826/3913/1/Estimation_of_the_greenhouse_gas_emissi

ons_from_agricultural_pesticide_manufacture_and_use-2009.pdf, [Accessed 10 October

2014]

Aviso, K.B., Tan, R.R., Culaba, A.B., Cruz Jr., J.B., 2011. Fuzzy input-output model for

optimizing eco-industrial supply chains under water footprint constraints. J. Cleaner. Prod.

19, 187-196.

AWEA, (2012) U.S. Department of Energy report: Wind power costs near record low

[Online], Available from: www.aweablog.org/u-s-department-of-energy-report-wind-power-

costs-near-record-low/, [Accessed 20 November 2015]

Ayres, RU, (1998) Eco-thermodynamics: economics and the second law, Ecological

Economics 26:189-209

Bakshi, B. (2013) Energy, Sustainability and Life Cycle Assessment [Online], Available

from: web.mit.edu/ebm/www/250s_2013/Session%207.pdf [Accessed 23 November 2013]

Bastianoni, S and Marchettini, N (1996) Ethanol production from biomass: Analysis of

process efficiency and sustainability, Biomass and Bioenergy, Vol. 11, No. 5, pp. 41 I-418,

1996

Bastidas, P., Gil, I., Rodríguez, G. (2010) Comparison of the main ethanol dehydration

technologies through process simulation, 20th European Symposium on Computer Aided

Process Engineering–ESCAPE20

Beck, J., Kempener, R., Cohen, B., Petrie, J., (2008) A complex systems approach to

planning, optimisation, and decision making for energy networks, Energy policy 36, 2795-

2805

241

Becker, H.C., Maréchal, F. (2011). Targeting industrial heat pump integration in multi-period

problems. 11th International Symposium on Process Systems Engineering. ISSN: 1570-7946,

vol. 31, p. 415-419

Beer, A.G., Boast, M., Worlock, B. (1989) The agricultural consequences of harvesting

sugarcane containing various amounts of tops and trash, Proceedings of The South African

Sugar Technologists' Association - June 1989

Behzadian, K., Farmani, R., Butler, D., (2016) Water Feasibility Project for Local Nexus

Network of Food, Energy and Water [Online], Accessed from:

localnexus.org/wp-content/uploads/2015/04/Water-in-Bread-draft-report.pdf, [Accessed 3rd

July 2016]

BioEnergy Consult (2014) Salient Features of Sugar Industry in Mauritius [Online],

Available from: http://www.bioenergyconsult.com/tag/mauritius/, [Accessed 12 May 2014]

Biondi, P., Panaro, V. and Pellizi, G. 1989 (Eds), Le richieste di energia de1 sistema agricolo

italiano. ENEA-PFE (Progetto Finahzzato Energetica), Rome

Boix, M., Montastruc, L., Azzaro-Pantel, C., Domenech, S. (2015). Optimization methods

applied to the design of eco-industrial parks: a literature review. J. Cleaner. Prod. 87 (2015)

303-317

Brehmer, B. (2008) Chemical biorefinery perspectives: The valorisation of functionalised

chemicals from biomass resources compared to the conventional fossil fuel production route

[Online] Available from: http://www.lcacenter.org/InLCA2006/Brehmer-abstract.pdf;

[Accessed 4 May 2014]

Brown, M. T. and Arding, J. (1991) Transformity Working Paper. Centre for Wetlands,

University of Florida, Gainesville, FL

Brown, M.T., Ulgiati, S., (2010). Updated evaluation of exergy and emergy driving the

geobiosphere: a review and refinement of the emergy baseline. Ecological Modelling. 221,

2501-2508.

242

Building Code Division (BCD), (2008) Rainwater harvesting [Online], Available from:

http://www.bcd.oregon.gov/pdf/3660.pdf, [Accessed 8 March 2015]

Cane Technology Centre (CTC) (2005) Biomass power generation: sugar cane bagasse and

trash. Report to UNDP/MCT/GEF, Project BRA/96/G31

Carbon Trust, (2013) Biomass heating-A practical guide for potential users [Online],

Available from: https://www.carbontrust.com/media/31667/ctg012_biomass_heating.pdf,

[Accessed 2 June 2015]

CEMEX UK Cement Ltd, (2014) Surrender Site Condition Report EPR/BK09731K/S007

[Online], Available from: www.scambs.gov.uk/sites/default/files/documents/Appendix

%2011.3%20Surrender%20Site%20Condition%20Report%20Issue.pdf, [Accessed 10

November 2015]

Centre for Alternative Technology (CAT), (2015) How long do solar electric PV panels last?

[Online], Available from: http://info.cat.org.uk/questions/pv/life-expectancy-solar-PV-panels

[Accessed 3 June 201]

Chae, S.H., Kim, S.H., Yoon, S.-G., Park, S., 2010. Optimization of a waste heat utilization

network in an eco-industrial park. Appl. Energy. 87, 1978-1988

Chen, G. Q. (2005). Exergy consumption of the earth. Ecological Modelling 184(2-4): 363-

380.

Chen, G. Q. (2006). Scarcity of exergy and ecological evaluation based on embodied exergy.

Communications in Nonlinear Science and Numerical Simulation 11(4): 531-552.

Chen, P.J., Yang, L., Bai, R. (2006), Bakery Waste Treatment, Handbook of Industrial and

Hazardous Wastes Treatment.

Chen, C.Q. and Chen, B. (2009) Extended-exergy analysis of the Chinese Society, Energy 34

(2009) 1127–1144

243

Chertow, M., and Ehrenfeld, J. (2012) Organizing self-organizing systems, toward a theory of

industrial symbiosis, Journal of Industrial Ecology, 16(1):13-27

Cimren, E., Fiksel, J., Posner, M.E., Sikdar, K. (2011) Material flow optimization in by-

product synergy networks. J. Ind. Ecol. 15, 315-332.

Cleveland, C.J., Kaufmann, R.K. and Stern, D.I., (2000). Aggregation and the role of energy

in the economy. Ecol.Econ., 32: 301-317.

Commonwealth of Massachusetts (COM) (2013) Module 3: Transportation and Transfer of

Ethanol-Blended Fuels [Online], Available:

ethanolresponse.com/wp-content/uploads/2016/01/Participant-Guide-Mod3-1.pdf, [Accessed

15 April 2014]

Connelly, L. and C. P. Koshland (2001). Exergy and industrial ecology-art 1: An exergy-

based definition of consumption and a thermodynamic interpretation of ecosystem evolution.

Exergy, An International Journal 1(3): 146-165.

Cornelissen, R. L. (1997). Thermodynamics and sustainable development - The use of exergy

analysis and the reduction of irreversibility. Mechanical Engineering. Groningen, The

Netherlands, University of Groningen. Ph.D. Mechanical Engineering: 170.

Cornelissen, R. L. and G. G. Hirs (2002). The value of the exergetic life cycle assessment

besides the LCA. Energy Conversion and Management 43(9-12): 1417-1424.

Cote, RP. and Hall J. (1995) Industrial parks as ecosystems. Journal of Cleaner Production

1995; 3 (1, 2):41–6.

Cowell, S., Parkinson, S. (2003) Localisation of UK food production: an analysis using land

area and energy as indicators, Agriculture, Ecosystems and Environment, 94 (2003) 221–236

Curtis, F. (2003), Eco-localism and sustainability, Ecological Economics, 46, 83-102

244

Danon, G., Furtula, M., Mandic, M. (2012) Possibilities of implementation of CHP

(combined heat and power) in the wood industry in Serbia, Energy, 48 (2012) 169-176

de Mes, T.Z.D., Stams, A.J.M., Reith, J.H. and Zeeman, G. (2003) Methane production by

anaerobic digestion of wastewater and solid wastes [Online], Available from:

http://es.ircwash.org/sites/default/files/Reith-2003-Bio.pdf#page=59, [Accessed 22

September 2014]

Dean, P.E. (1997) Economical Condensing Turbines? [Online], Available from:

https://repository.tamu.edu/bitstream/handle/1969.1/91264/ESL-IE-97-04-51.pdf?

sequence=1, [Accessed 12 May 2014]

DECC (Department of Energy and Climate Change) (2014) Gas Boiler Cost Data [Online],

Available from: http://2050-calculator-tool-wiki.decc.gov.uk/cost_categories/82, [Accessed 6

November 2015]

Department of Energy and Climate Change (DECC) (2015a), Association of Decentralised

Energy Speech [Online], Available from:

https://www.gov.uk/government/speeches/association-of-decentralised-energy-speech,

[Accessed 16 July 2017]

DECC (Department of Energy and Climate Change (DECC) (2015b), Energy consumption in

the UK [Online], Available from:

www.gov.uk/government/uploads/system/uploads/attachment_data/file/449134/

ECUK_Chapter_3_-_Domestic_factsheet.pdf, [Accessed 3 March 2015]

Department for Environment Food and Rural Affairs (DEFRA), (2010) Developing an

Anaerobic Digestion (AD) Framework Document [Online], Available from:

//archive.defra.gov.uk/environment/waste/ad/documents/anaerobic-digestion-framework-

101130.pdf, [Accessed 30 September 2014]

Department for Environment Food and Rural Affairs (DEFRA) (2014) Family food datasets

[Online], Available from: www.gov.uk/government/statistical-data-sets/family-food-datasets,

[Accessed 2 November 2014]

245

Department of Environment and Heritage Protection (DEHP), (2014) Wastewater [Online],

Available from: https://www.ehp.qld.gov.au/water/monitoring/wastewater.html, [Accessed

January 2015]

Dewulf, J., Bosch, M.E., Demeester, B., Vandervorst, G., Vanlangenhove, H,. Hellweg, S.,

and Huijbregts, M.A., (2007). Cumulative Exergy Extraction from the Natural Environment

(CEENE): a comprehensive Life Cycle Impact Assessment method for resource accounting.

Environmental Science Technology. 41, 8477–8483

Dewulf, J., Van Langenhove, H., Dirckx, J. (2000) Exergy analysis in the assessment of the

sustainability of waste gas treatment systems, The Science of the Total Environment 273

(2001) 41-52

DK (2014) LCA food [Online], Available from:

http://www.lcafood.dk/products/crops/bread.htm, [Accessed 7 September 2014]

Dias, M.O.S., Cunha, M.P., Jesus, C.D.F., Scandiffio, M.I.G., Rossell, C.E.V., Filho, R.M.,

Bonomi, A. (2010) Simulation of ethanol production from sugarcane in Brazil: economic

study of an autonomous distillery, Computer Aided Chemical Engineering, 28 (2010) pp.

733-738, 20th European Symposium on Computer Aided Process Engineering

Dias de Oliveira, M.E., Vaughan, B., Rykiel, E. (2012) Ethanol as Fuel: Energy, Carbon

Dioxide Balances, and Ecological Footprint [Online], Available from:

https://biogas.ifas.ufl.edu/BESTS/files/deOliveira.pdf, [Accessed 5 February 2015]

DM (2013) How much do cows weight? [Online], Available from:

http://www.dairymoos.com/how-much-do-cows-weight/ [Accessed 11 October 2014]

Dominguez-Ramos, A., Triantafyllidis, C., Samsatli, S., Shah, N. and Irabien, A.,

(2016). Renewable electricity integration at a regional level: Cantabria case study. Computer

Aided Chemical Engineering, 38, pp. 211-216.

Douglas, J.M., (1988) Conceptual Design of Chemical Processes. McGraw Hill, New York.

246

ECOCHEM (2015) Manure is an excellent fertiliser [Online], Available from:

http://www.ecochem.com/t_manure_fert.html, [Accessed 5 September 2014]

ECOSURE (2015) Rainwater storage tank [Online], Available from:

http://www.ecosure.co.uk/, [Accessed 28 September 2015]

Edwards, W. (2015) Cost of storing grain [Online], Available from:

http://www.extension.iastate.edu/agdm/crops/html/a2-33.html [Accessed 30 August 2015]

El-Halwagi, M.M., Manousiouthakis, V. (1989) Synthesis of mass-exchange networks. Am.

Inst. Chem. Eng. J. 35, 1233–1244

El-Hawagi, M.M., Manousiouthakis, V. (1990) Automatic synthesis of mass-exchange

networks with single-component targets. Chem Eng Sci. 45(9):2813-2831

El-Halwagi, M. M. (2011) Sustainable Design through Process Integration: Fundamentals

and Applications to Industrial Pollution Prevention, Resource Conservation, and

Profitability Enhancement, Butterworth-Heinemann Ltd. Elsevier, Oxford, UK

Enderlein, S., Enderlein, R. and Williams, P. (2014) Water Quality Requirements [Online],

Available from: http://www.who.int/water_sanitation_health/resourcesquality/wpcchap2.pdf,

[Accessed 11 October 2014]

Environment Agency (EA), (2009) How to comply with your environmental permit

Additional guidance for: The Red Meat Processing (Cattle, Sheep and Pigs) Sector (EPR

6.12) [Online], Available from:

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/298054/

geho0209bpja-e-e.pdf, [Accessed 2 October 2014]

Environment Impact Assessment (EIA) (2011) Installation and Operation of a Distillery and

Concentrated Molasses Solids (CMS) Fertilizer Blending Plant [Online], Available from:

http://www.gov.mu/portal/goc/menv/files/dist_CMS/toc.pdf, [Accessed 21 December 2011]

247

Environmental Protection Agency (EPA), (2008a) Technical Development Document for the

Final Effluent Limitations Guidelines and Standards for the Meat and Poultry Products Point

Source Category (40 CFR 432) [Online], Available from:

http://water.epa.gov/scitech/wastetech/guide/mpp/upload/2008_07_15_guide_mpp_final_tdd

06.pdf, [Accessed 11 October 2014]

Environmental Protection Agency (EPA) (2008b) Combined heat and power partnerships

[Online], Available from: http://www.epa.gov/chp/documents/catalog_chptech_intro.pdf

[Accessed from 10 October 2014]

Environmental Protection Agency, (EPA) (2011) Opportunities for Combined Heat and

Power at Wastewater Treatment Facilities: Market Analysis and Lessons from the Field

[Online], Available from: http://www.epa.gov/chp/documents/wwtf_opportunities.pdf,

[Accessed 2 October 2014]

Environmental Protection Agency, (EPA) (2015) Natural gas combustion [Online], Available

from: www3.epa.gov/ttn/chief/ap42/ch01/final/c01s04.pdf, [Accessed 6 November 2015]

ERM (2009) Life Cycle Assessment of Pork [Online], Available:

http://www.bpex.org.uk/prices-facts-figures/documents/LifeCycelAssmntofPorklaunchversio

n.pdf, [Accessed 3 October 2014]

Energy Saving Trust (EST), (2014) Our calculations [Online], Available from:

www.energysavingtrust.org.uk/content/our-calculations, [Accessed 6 November 2015]

European Wind Energy Association (EWEA), (2015) Wind energy's frequently asked

questions (FAQ)[Online], Available from: www.ewea.org/wind-energy-basics/faq/,

[Accessed 20 November 2015]

European Commission (EC), (2012) Assessment of resource efficiency indicators and targets

[Online], Available from:

ec.europa.eu/environment/enveco/resource_efficiency/pdf/report.pdf [Accessed 20 November

2013]

248

European Commission (EC), (2013). Online Resource Efficiency Platform (OREP) [Online],

Available from: ec.europa.eu/environment/resource_efficiency/ [Accessed 20 November

2013]

Finguerut, J. (2003) Ethanol production–Research and Development (Brazil), PowerPoint

presentation presented by Copersucar Technology Centre at the ISSCT Co-Product

Workshop, Piracicaba, Sao Paolo, Brazil, 14-18 July 2003.

Floudas, C.A. and I.E. Grossmann, Synthesis of Flexible Heat Exchanger Networks for

Multiperiod Operation, Computers and Chemical Engineering, 10, 153 (1986).

Fodder Feed (FF), (2015) Fodder Feed [Online], Available from:

http://www.fodderfeed.org/Index.html, [Accessed 1 October 2014]

Food and Agriculture Organisation of the United Nations (FAO) (2010a) Cleaning and

sanitation in meat plants [Online], Available from:

http://www.fao.org/docrep/010/ai407e/ai407e26.htm, [Accessed 2 March 2015]

Food and Agriculture Organisation of the United Nations (FAO) (2010b) Heat treatment of

meat products [Online], Available from:

http://www.fao.org/docrep/010/ai407e/AI407E08.htm, [Accessed 3 April 2015]

Food and Agriculture Organisation of the United Nations FAO (2011) Harvest and storage

management of wheat [Online], Available from:

http://www.fao.org/docrep/006/y4011e/y4011e0u.htm, [Accessed 30 September 2015]

Food and Agriculture Organisation of the United Nations (FAO) (2014) The Water-Energy-

Food Nexus: A new approach in support of food security and sustainable agriculture

[Online], Available from: www.fao.org/nr/water/docs/FAO_nexus_concept.pdf, [Accessed

27 July 2015]

Food and Agriculture Organisation of the United Nations (FAO), (2015) Costs of bulk

storage [Online], Available from: http://www.fao.org/docrep/t1838e/T1838E1c.htm,

[Accessed 2 January 2015]

249

Foo, D.C.Y., Manan, Z.A., Tan, Y.L. (2006) Use cascade analysis to optimise water

networks [Online], Available from: www.geocities.ws/foodominic/CEP_WCA_Proof.pdf,

[Accessed 26 July 2016]

Foo, D. (2007). Water cascade analysis for single and multiple impure fresh water feed.

IChemE, Vol. 85 (A8) 1169-1177

Foo, D.C.Y., El-Halwagi, M.M., Tan, R.R. (2012) Recent Advances in Sustainable Process

Design and Optimization. World Scientific Publishing, Singapore

Foo, D. (2013). Process integration for resource conservation. Taylor & Francis Group,

ISBN 9781439860489

Foo, D.C.Y., Tan, R.R. (2015). A review on process integration techniques for carbon

emissions and environmental footprint problems. Process Saf. Environ. Prot,

http://dx.doi.org/10.1016/j.psep.2015.11.00

Fulmer, M. (1991) Electricity-Ethanol Co-production from sugarcane: A technical and

economic assessment, Master thesis, University of Princeton

Garcia, D.J., You, F. (2016) The water-energy-food nexus and process systems engineering:

A new focus. Comput. Chem. Eng. (In Press), doi:10.1016/j.compchemeng.2016.03.003

Gaudreau, K., (2009). Exergy Analysis and Resource Accounting, Master Thesis, University

of Waterloo, Ontario, Canada

Geldermann, J., Treitz, M., Rentz, O. (2006) Integrated technique assessment based on the

pinch analysis approach for the design of production networks, European Journal

Operational Research, 171, 1020–1032

Gerbens-Leenes, P.W., Mekonnen, M.M., Hoekstra, A.Y. (2013) The water footprint of

poultry, pork and beef: A comparative study in different countries and production systems,

Water Resources and Industry, 1–2 (2013) 25–36

250

Guadalupe-Blanco River Authority (GBRA) (2016), Section 5: Cost estimating process

[Online], Available from: http://www.gbra.org/documents/studies/calhoun/05-

costestimating.pdf, [Accessed 21 August 2016]

Gulati, S. and Singh, M. (2011) Energy requirement and management in a potato production

system, Potato Journal, 38 (1): 61-66, 2011

Gong, M. and G. Wall (2000) On exergy and sustainable development - Part 2 -Indicators and

methods, Exergy, an International Journal 1(4): 17.

Grossmann, I.E., Caballero, J.A., Yeomans, H., (1999) Mathematical programming

approaches to the synthesis of chemical process systems, Korean Journal of Chemical

Engineering, Vol. 16, Issue 4, pp. 407-426

Hanes, R. J., Bakshi, B. R. (2015a) Process to planet: A multiscale modeling framework

toward sustainable engineering. AIChE J. DOI: 10.1002/aic.14919

Hanes, R. J., Bakshi, B. R. (2015b) Sustainable process design by the process to planet

framework. AIChE J. DOI: 10.1002/aic.14918

Hau, J and Bakshi, B (2004a) Expanding Exergy Analysis to Account for Ecosystem Services

and Products, Environmental Science Technology 2004, 38, 3768-3777

Hau, J.L., Bakshi, B.R. (2004b) Promise and Problems of Emergy Analysis. Ecological

Modelling, 178(1-2), 215-225.

Hau, J.L. (2005) Toward environmentally conscious process systems engineering via joint

thermodynamic accounting of industrial and ecological systems, PhD dissertation, Ohio State

University.

Henze, M. and Comeau, Y. (2008) Wastewater Characterization [Online], Available from:

ocw.unesco-ihe.org/pluginfile.php/462/mod_resource/content/1/Urban_Drainage_and_Sewer

251

age/5_Wet_Weather_and_Dry_Weather_Flow_Characterisation/DWF_characterization/

Notes/Wastewater%20characterization.pdf, [Accessed 23 October 2014]

Hertwich, E. Understanding the climate mitigation benefits of product systems: comment on

Using Attributional Life Cycle Assessment to Estimate Climate-Change Mitigation. J Ind

Ecol. 2014; 18(3):464–5.

Huang, L. Q., G. Q. Chen, et al. (2007). Exergy as a unified measure of water quality.

Communications in Nonlinear Science and Numerical Simulation, 12(5): 663-672.

Huber (2014) Wood industry: COD and solids reduction as first pre-treatment for wastewater

to be reused [Online], Available from:

http://www.huber.de/huber-report/ablage-berichte/sludge-treatment/wood-industry-cod-and-

solids-reduction-as-first-pre-treatment-for-wastewater-to-be-reused.html [Accessed 23

October 2014]

Huijbregts, M.A.J., Hellweg, S., Frischknecht, R., Hendriks, H.W.M., Hungerbuhler, K.,

Hendriks, A.J., (2010). Cumulative energy demand as predictor for the environmental burden

of commodity production. Environmental Science Technology 44:2189-2196

International Energy Agency (IEA) (2012), Water for Energy- Is energy becoming a thirstier

resource? [Online], Available from:

www.worldenergyoutlook.org/media/weowebsite/2012/WEO_2012_Water_Excerpt.pdf,

[Accessed 10 November 2015]

International Energy Agency (IEA), (2015) CO2 emissions from fuel combustion, Highlights

[Online], Available from:

www.iea.org/.../CO2EmissionsFromFuelCombustionHighlights2015.pdf, [Accessed 8

September 2016]

IME (2014) How much water is needed to produce food and how much do we waste?

[Online], Available from: www.theguardian.com/news/datablog/2013/jan/10/how-much-

water-food-production-waste#data, [Accessed 10 October 2014]

252

IPCC (2006) Emissions from waste incineration [Online], Available from: www.ipcc-

nggip.iges.or.jp/public/gp/bgp/5_3_Waste_Incineration.pdf, [Accessed 3 April 2015]

IPCC (2007) Emissions Scenarios [Online], Available from:

www.ipcc.ch/ipccreports/sres/emission/index.php?idp=70, [Accessed 3 September 2014]

Intergovernmental Panel on Climate Change (IPCC), (2011) IPCC Special Report on

Renewable Energy Sources and Climate Change Mitigation, Cambridge University Press,

Cambridge, United Kingdom and New York, NY, USA, pp. 1075 (Chapter 9).

International Renewable Energy Agency (IRENA), (2016) Renewable energy benefits:

Measuring the economics [Online], Available from:

www.irena.org/DocumentDownloads/Publications/IRENA_Measuring-the-

Economics_2016.pdf, [Accessed 9 September 2016]

Iribarren, D., Vázquez-Rowe, I. (2013) Is Labor a Suitable Input in LCA + DEA Studies?

Insights on the Combined Use of Economic, Environmental and Social Parameters, Soc. Sci.

2013, 2, 114–130; doi:10.3390/socsci2030114

ISO 14040 (2017) The ISO 14040 standards for consequential LCA [Online], Available from:

https://consequential-lca.org/clca/why-and-when/the-iso-14040-standards-for-consequential-

lca/ [Accessed 25 October 2017]

Jacques, K.A. (2003) The Alcohol Textbook: A reference for the beverage, fuel and industrial

alcohol industry. 4th Ed, Nottingham University Press

Jata, S.K., Nedunchezhian, M., Misra, R.S. (2011) The Triple ‘f’ (food, fodder and fuel) Crop

Sweet Potato [Ipomoea batatas (L.) Lam. [Online], Available from: http://odisha.gov.in/e-

magazine/Orissareview/2011/Dec/engpdf/83-93.pdf, [Accessed 4 September 2014]

Jawad, H., Jaber, M.Y., Bonney, M., (2015). The Economic Order Quantity model revisited:

an Extended Exergy Accounting approach. Journal of Cleaner Production, 105 (2015) 64-73

Jiang, M. M., J. B. Zhou, et al. (2009). Ecological evaluation of Beijing economy based

253

on emergy indices. Communications in Nonlinear Science and Numerical Simulation, 14(5):

2482-2494.

Jorgensen, S.E., (1997). Integration of Ecosystem Theories: a Pattern; Kluwer Academic

Publishers: Boston, MA

Johansson, A., Kisch, P., Mirata, M. (2005) Distributed economies-A new engine for

innovation, Journal of Cleaner Production, 13, 971-979

Johnson, F and Seebaluck, V. (2012) Bioenergy: For sustainable development and

international competitiveness, Routlegde, New York

Joint Research Council (JRC) (2007) Well-to-Wheels Analysis of Future Automotive Fuels

and Powertrains in the European Context [Online], Available from:

http://ies.jrc.ec.europa.eu/uploads/media/TTW_Report_010307.pdf [Accessed 4 May 2014]

Jordao, E.P. (2010) Cogeneration of electrical and thermal energy from biogas in wastewater

treatment plants – The case of Brazil [Online], Available from:

www.globalmethane.org/documents/events_combined_20101111_perspectives_from_brazil.

pdf, [Accessed 8 February 2012]

Junqueira, T.L., Dias, M.O.S., Jesus, C.D.F., Mantelatto, P.E., Cunha, M.P., Cavalett, O.,

Filho, R.M., Rossell, C.E.V., Bonomi, A. (2010) Simulation and Evaluation of Autonomous

and Annexed Sugarcane Distilleries [Online], Available from:

www.aidic.it/pres11/webpapers/71Junqueira.pdf, [Accessed 21 February 2014]

Keairns, D.L., Darton, R.C., Irabien, A. (2016) The Energy-Water-Food Nexus. Annu. Rev.

Chem. Biomol. Eng. 2016. 7:9.1–9.24

Khan, A.A., Gaur, R.Z., Mehrotra, I., Kazmi, A.A. (2011) UASB-CFID System: An Energy

Efficient Technology for Sewage Treatment and Reuse [Online], Available from:

http://www.iwawaterwiki.org/xwiki/bin/view/Articles/EnergyEfficientTreatmentSystems,

[Accessed 8 February 2012]

254

King, C., Holman, A., Webber, M.E. (2005) Thirst for energy, Nature Geoscience 1, 283–

286, DOI: 10.1038/ngeo195

Klatt, K.U., and Marquardt, W. (2009) Perspectives for process systems engineering-Personal

views from academia and industry, Computers and Chemical Engineering, 33, 536–550

Klemes, J.J., Varbanov P. (2013), Integration of energy and resource flows, Chemical

Engineering Transactions, 34, 7-12 DOI: 10.3303/CET1334002

Klemes, J.J., Varbanov, P.S., Kravanja, Z. (2013) Recent developments in Process

Integration. Chem. Eng. Res. Des.91, 2037–2053

Kneeshaw, A. (2006) Energy status report: GB potato storage for British Potato Council

[Online], Available from:

potatoes.ahdb.org.uk/sites/default/files/publication_upload/FECenergyReport.pdf, [Accessed

20 November 2015]

Kostevsek, A., Petek, J., Cucek, L., Klemes, J.J., Varbanov, P.S. (2015) Locally Integrated

Energy Sectors supported by renewable network management within municipalities. Appl.

Therm. Eng., 89, 1014-1022

Krautkraemer, J., (2005). Economics of Natural Resource Scarcity: The State of the Debate

[Online], Available from www.rff.org/files/sharepoint/WorkImages/Download/RFF-DP-05-

14.pdf, [Accessed 24 December 2015]

Lal, R. (1981) Soil erosion problems on alfisols in Western Nigeria. VI. Effects of erosion on

experimental plots. Geoderma, 25: 215-230.

Lam, H.L., Varbanov, P., Klemes, J. (2009a). Regional resource management composite

curve. Chem Eng Trans, 18:303–8.

Lam, H.L., Varbanov, P., Klemes, J. (2009b) Regional renewable energy and resource

planning. In: Special session: integrating waste and renewable energy to reduce the carbon

footprint locally integrated energy sectors, SEDEWES 09, Dubrovnik; 2009. p. 565

255

Lam, H.L., Varbanov, P., Klemes, J. (2010) Minimising carbon footprint of regional biomass

supply chains. Resour Conserv Recycl, 54:303–9.

Lam, H.L., Varbanov, P.S., Klemes, J.J. (2010b) Optimisation of regional energy supply

chains utilising renewables: P-graph approach. Comput. Chem. Eng. 34 (2010) 782–792

Langer, T. (2006) Simplified Life Cycle Assessment study of the substitution of 5 % of Swiss

gasoline by Brazilian bio–ethanol [Online], Available from:

http://www.ekosbrasil.org/media/file/Ethanol%20LCA_Instituto_Ekos_Brasil.pdf, [Accessed

17 March 2014]

Lau, A.F. (2008) An assessment of the renewable energy export potential of the Mauritian

sugar cane industry with new practices and prospective technologies. MSc Thesis. University

of Ulster, United Kingdom

Lawlor, P. (2010) What is the optimum slaughter weight for pigs? [Online], Available from:

www.teagasc.ie/pigs/articles/farming_independent/2010/Optimum_slaughter_weights_May2

010.pdf, [Accessed 4 September 2014]

Law, R., Harvey, A., Reay, D. (2011) Opportunities for Low-Grade Heat Recovery in the UK

Food Processing Industry [Online], Available from:

research.ncl.ac.uk/pro-tem/components/pdfs/SusTEM2011/T1S4_01_Newcastle_RLAW_Op

portunities_for_Low-Grade_Heat_Recovery_in_the_UK.pdf, [Accessed 10 November 2015]

Leung Pah Hang, M., Seebaluck, V., Ragen, A. (2012) Resource requirements of the cane

agro-industry, Undergraduate thesis, University of Mauritius, Mauritius

Leung Pah Hang, M., Martinez-Hernandez, E., Leach, M., and Yang, A. (2015) Engineering

Design of Localised Synergistic Production Systems, Computer Aided Chemical

Engineering, 37. pp. 2363-2368

Leung Pah Hang, M.Y., Martinez-Hernandez, E., Leach, M., Yang, A. (2016a) Towards a

coherent multi-level framework for resource accounting, Journal of Cleaner Production, 125

(2016), pp. 204–215

256

Leung Pah Hang, M.Y., Martinez-Hernandez, E., Leach, M., Yang, A. (2016b) Designing

integrated local production systems: A study on the food-energy-water nexus, Journal of

Cleaner Production, 135 (2016) 1065-1084

Leung Pah Hang, M.Y., Martinez-Hernandez, E., Leach, M., Yang, A. (2017) An insight-

based approach for the design of integrated local food-energy-water systems, Submitted for

publication in Environmental Science & Technology

Liao, W., Heijungs, R., Huppes, G., (2012). Thermodynamic resource indicators in LCA: a

case study on the titania produced in Panzhihua city, southwest China. International Journal

Life Cycle Assessment (2012) 17:951–961

Linnhoff, B., Hindmarsh, E. (1983). The pinch design method for heat exchanger networks.

Chem. Eng. Sci. 38, 745–763.

Linnhoff, B. (1993) Pinch Analysis-A state of the art overview. Trans IchemE, 1993;

71(A):503.

Lovelady, E.M., El-Halwagi, M.M., 2009. Design and integration of eco-industrial parks for

managing water resources. Environ. Prog. Sustain. Energy. 28, 265-272.

Luo, X., Wen, Q.Y., Fieg, G., 2009. A hybrid genetic algorithm for synthesis of heat

exchanger networks. Comput. Chem. Eng.33 (6), 1169–1181

Lyons, E., Zhang, P., Benn, T., Costanza, M., Li, K. (2014) Life Cycle Assessment of Three

Water Scenarios: Importation, Reclamation, and Desalination [Online], Available from:

http://www2.bren.ucsb.edu/~keller/energy-water/3-3%20John%20Crittenden.pdf, [Accessed

6 September 2014]

Macedo, I.C., Seabra, J.E.A. and Silva, J.E.A.R., 2007, Greenhouse gases emissions in the

production and use of ethanol from sugarcane in Brazil: The 2005/2006 averages and a

prediction for 2020. Biomass and Bioenergy 32 (2008) 582–595

257

Macknick, J., Newmark, R., Heath, G., Hallett, K.C. (2011) A Review of Operational Water

Consumption and Withdrawal Factors for Electricity Generating Technologies, Golden, CO:

National Renewable Energy Laboratory.

Machell, J., Prior, K., Allan, R., Andresend, J.M. (2015). The water energy food nexus-

challenges and emerging solutions. Environ. Sci.: Water Res. Technol. 2015, 1, 15-16. DOI:

10.1039/C4EW90001D

MacKay, D. (2009) Sustainable energy- Without the Hot Air, UIT Cambridge, ISBN-13:

9780954452933

Maier, D.E., Bakker-Arkema, F.W. (2002) Grain Drying Systems [Online], Available from:

http://www.uwex.edu/energy/pubs/GrainDryingSystems_GEAPS2002.pdf [Accessed 30

September 2015]

Manan, Z.A., Wan Alwi, S.R., Ujang, Z. (2005) Water pinch analysis for an urban system: a

case study on the Sultan Ismail Mosque at the Universiti Teknologi Malaysia (UTM),

Desalination 194 (2006) 52–68

Marques, J.C, Pardal, M.A, Nielsen, S.N, Jorgensen, S.E. (1997) Analysis of the properties of

exergy and biodiversity along an estuarine gradient of eutrophication, Ecological Modelling

102 (1997) 155-167

Martin, M., Grossmann, I.E. (2015) Water–energy nexus in biofuels production and

renewable based power. Sustainable Prod Consumption. 2(2015) 96–108

Martin, E.W., Chester, M.V. & Vergara, S.E. Curr Sustainable Renewable Energy Rep

(2015) 2: 82. https://doi.org/10.1007/s40518-015-0034-9

Martinez-Hernandez, E., Leung Pah Hang, M.Y., Leach, M., Yang, A. (2016) A Framework

for Modeling Local Production Systems with Techno‐Ecological Interactions, Journal of

Industrial Ecology, In press, DOI: 10.1111/jiec.12481

258

Marufuzzaman, M., Eksioglu, Y.H., (2014) Two-stage stochastic programming supply chain

model for biodiesel production via wastewater treatment, Computers & Operations Research,

Vol 49, Pages 1-17

Matthews, E.; Amann, C.; Bringezu, S.; Fischer-Kowalski, M.; Huttler, W.; Kleijn, R.;

Moriguchi, Y.; Ottke, C.; Rodenburg, E.; Rogich, D.; Schandl, H.; Schutz, H.; Van Der Voet,

E.; Weisz, H. (2000) The Weight of Nations: Material Outflows from Industrial Economies.

World Resource Institute. Washington D.C.

Mauritius Sugar Industry Research Institute (MSIRI) (2010) Annual Report 2009 -2010,

Republic of Mauritius

McNeill, S., Overhults, D., Montross, M. (2010) Harvesting, Drying and Storing Wheat

[Online], Available from: www2.ca.uky.edu/agc/pubs/id/id125/10.pdf, [Accessed 20

November 2014]

Menendez, M. (2009), M. Menendez, How We Use Energy at Wastewater Plant and How We

Can Use Less [Online],

www.ncsafewater.org/Pics/Training/AnnualConference/AC10TechnicalPapers/

AC10_Wastewater/WW_T.AM_10.30_Menendez.pdf [Accessed 2 September 2014]

Meneses, B. (2008) Biogas Production with Vinasse, a Feasible Alternative to Contribute to

the Development of Bioenergy [Online], Available from:

http://www.sugarjournal.com/articles/active_subs/2008/oct08/Bioga_%20Production.pdf,

[Accessed 20 December 2011]

Met Office (2012) UK Climate [Online], Available from:

http://www.metoffice.gov.uk/public/weather/climate/, [Accessed 3 March 2015]

Meyer, E. (2006) A review of the harvesting, loading, transport and mill receiving operations

of the South African Sugar Industry [Online], Available from:

www.cenicana.org/pdf/otros/foro_cosecha_transporte_2006/6_cosecha_transporte_recepcion

_sudafrica_may9-2006.pdf, [Accessed 3 May 2014]

259

Middlemiss, L., Parrish, B.D. (2010) Building capacity for low-carbon communities: The role

of grassroots initiatives, Energy Policy, 38, 7559-7566.

Millennium Ecosystem Assessment (MEA) (2005) Ecosystems and human wellbeing:

synthesis. Island Press, Washington DC

Ministry of Agriculture, Fishery and Food (MAFF) (1988) Agricultural land classification of

England and Wales [Online], Available from:

webarchive.nationalarchives.gov.uk/20130402151656/http:/archive.defra.gov.uk/foodfarm/

landmanage/land-use/documents/alc-guidelines-1988.pdf, [Accessed 18 April 2016]

National Renewable Energy Laboratory (NREL) (2010), Cost and Performance Assumptions

for Modeling Electricity Generation Technologies [Online], Available from:

http://www.nrel.gov/docs/fy11osti/48595.pdf, [Accessed 3 September 2014]

National Renewable Energy Laboratory (NREL), (2014) Distributed solar PV for electricity

system resiliency, policy and regulatory considerations [Online], Available from:

www.nrel.gov/docs/fy15osti/62631.pdf, [Accessed 8 September 2016]

NEA (2014), Water pollution control [Online], Available from: http://app2.nea.gov.sg/anti-

pollution-radiation-protection/water-pollution-control/allowable-limits, [Accessed 4

September 2014]

Nelson, A.M., Liu, Y.A. Hydrogen pinch analysis made easy. Chem Eng J, 2008; 115(6):56–

61.

New South Wales Sugar (NSWS) (2014) The carbon dioxide cycle [Online], Available from:

http://www.nswsugar.com.au/index.php?

option=com_contentandview=articleandid=48:caring-for-the-environmentandcatid=33:nsw-

sugar-industryandItemid=202 [Accessed 14 April 2014]

NEXUS (2015) The Water, Energy and Food Security Resource Platform [Online], Available

from: www.water-energy-food.org/en/home.html, [Accessed 7 December 2015]

260

Nielsen, A., Nielsen, P.H. (2003) Industrial baking of bread [Online], Available from:

http://www.lcafood.dk/processes/industry/baking.htm, [Accessed 3 September 2014]

Nishida, N., Stephanopoulos, G., Westerberg, A.W., (1981) A review of process synthesis,

AIChE J, 27: 321–351. doi:10.1002/aic.690270302

Ng, R., Ng, D., Tan, R., El-Halwagi, M. (2014) Disjunctive fuzzy optimization for planning

and synthesis of bioenergy based industrial symbiosis systems. J. Environ. Chem. Eng. 2,

652-664

Odum, H. T. (1995) Emergy and Public Policy, Part I-II. Environmental Engineering

Sciences, University of Florida, Gainesville, FL; Wiley, New York

Odum, H.T. (1996) Environmental Accounting: Emergy and Environmental Decision

Making. John Wiley & Sons, New York, USA, 52, 75-80

Özilgena, M., Sorgüven, E. (2011) Energy and exergy utilization, and carbon dioxide

emission in vegetable oil production, Energy 36 (2011) 5954-5967

Palacios-Bereche, R., Mosqueira-Salazar, K.J., Modesto, M., Ensinas, A., Nebra, A., Serra,

L., Lozano, M (2012) Exergetic analysis of the ethanol production by enzymatic hydrolysis

process from sugarcane biomass, 3rd International Conference on Contemporary Problems of

Thermal Engineering CPOTE 2012, 18-20 September 2012, Gliwice, Poland Institute of

Thermal Technology

Palacios-Bereche, R, Mosqueira-Salazar, K.J., Modesto, M., Ensinas, A., Nebra, A., Serra,

L., Lozano, M. (2012) Exergetic analysis of the integrated first- and second-generation

ethanol production from sugarcane, Energy (2012) 1-16

Palacios-Bereche, R, Mosqueira-Salazar, K.J., Modesto, M., Ensinas, A., Nebra, A., Serra,

L., Lozano, M (2013), Exergetic analysis of the integrated first- and second-generation

ethanol production from sugarcane, Energy 62 (2013) 46-61

261

Panday, R. and Mishra, A. (2011) Livestock fodder requirements and household

characteristics in rural economy of hilly region, Uttarakhand, Himalayan Ecology, 19, 1

Patterson, P.E. (2007) Potato Storage Costs [Online], Available from:

www.cals.uidaho.edu/potatoes/Research&Extension/Topic/Storage/PotatoStorageCosts-

07.pdf, [Accessed 20 November 2015]

Paton, J. (2013) Energy utilisation in commercial bread baking, PhD thesis, School of

Mechanical Engineering, University of Leeds

Prasad, R.D., Bansal, R.C., Sauturaga, M. (2009) Some of the design and methodology

considerations in wind resource assessment, IET-Renewable Power Generation 3 (2009) 53-

64

Perera, C., Peng, C., Lee, S., Peters, T. (1998) Cost estimation of a heat exchanger [Online],

Available from:

wps.prenhall.com/wps/media/objects/148/151801/internet.../heat_exchanger.ppt, [Accessed 4

December 2015]

Pereira, C, Ortega, E (2007) Sustainability Assessment of Ethanol Production from

Sugarcane, 1st International Workshop on Advances in Cleaner Production

Perry S., Klemeš J., Bulatov I., 2008. Integrating waste and renewable energy to reduce the

carbon footprint of locally integrated energy sectors, Energy, 33(10), 1489-1497.

Pimentel, D. (1991) Ethanol fuels: energy, security, economics, and the environment. J of

Agricultural and Environmental Ethics 1991; 4:1-13

Pishvaee, M.S., Rabbani, M., Torabi, S.A. (2011) A robust optimisation approach to closed-

loop supply chain network design under uncertainty. Applied Mathematical Modelling, 35

637–649

Photovoltaic Geographical Information System (PVGIS), Available from:

http://re.jrc.ec.europa.eu/pvgis/apps4/pvest.php#

262

Ramjeawon, T. (1995) Integrated Management of cane–sugar factory waste waters in

Mauritius using the up-flow anaerobic sludge blanket (UASB) process. Ph.D. Thesis, Faculty

of Engineering, University of Mauritius

Rasmussen, U. (2011) Water Consumption in the Energy Sector and Energy Consumption in

the Water-Sector in a Danish Municipality, The Journal of Trans disciplinary Environmental

Studies, vol. 11

Rizk, M., (2013) Economic value of US fossil fuel electricity health impacts, Environment

International, 52, 75-80

Ritthoff, M., Rohn, H., Liedtke, C. (2002) Calculating MIPS Resource productivity of

products and services [Online] Available from:

www.econstor.eu/bitstream/10419/59294/1/485276682.pdf [Accessed 18 December 2013]

Rocco, M.V., Cassetti, G., Gardumi, F., Colombo, E., (2015). Exergy Life Cycle Assessment

of soil erosion remediation technologies: an Italian case study. Journal of Cleaner Production

112 (2016) 3007-3017

Rosen, M. A., I. Dincer, et al. (2008). Role of exergy in increasing efficiency and

sustainability and reducing environmental impact. Energy Policy 36(1): 128-137.

Rosenthal, R. (2015) GAMS- A User’s Guide [Online], Available from:

www.gams.com/help/topic/gams.doc/userguides/GAMSUsersGuide.pdf, [Accessed 25

November 2015]

Royal Academy of Engineering (2011) Infrastructure, Engineering and Climate Change

Adaptation –ensuring services in an uncertain future, ISBN 1-903496-61-6

Ruskins and Associates, (2015) How much yeast to use when baking bread? [Online],

Available from: lighterside.complianceofficer.com/ratio-yeast-flour-bread.html [Accessed 17

September 2014]

263

Saad, A. (2009) COD and BOD Reduction of Domestic Wastewater using Activated Sludge,

Sand Filters and Activated Carbon in Saudi Arabia. Biotechnology, 8: 473-477.

Sadhukhan, J., Ng, K.S., Martinez-Hernandez, E., 2014. Biorefineries and Chemical

Processes, UK: John Wiley &Sons Ltd. p147-148.

Sayed, K.I., El-Ezaby, K.H., Groendijk, L. (2005) Treatment of potato processing wastewater

using a membrane bioreactor, Ninth International Water Technology Conference, IWTC9

2005, Sharm El-Sheikh, Egypt

Sciubba, E. (2011) A revised calculation of the econometric factors and for the Extended

Exergy Accounting Method, Ecological Modelling 222 (2011) 1060–1066

Seckin, C., Bayulken, A. (2013) Extended Exergy Accounting (EEA) analysis of municipal

wastewater treatment-Determination of environmental remediation cost for municipal

wastewater, Appl. Energy 110 (2013) 55–64

Seebaluck, V., Mohee, R., Sobhanbabu, P.R.K., Rosillo-Calle, F., Leal, M.R.L.V. & Johnson,

F.X. (2008) Bioenergy for sustainable development and global competitiveness: The case of

sugarcane in Southern Africa [Online], Available from:

http://www.carensa.net/tr/CARENSA-TR2-industry_final.pdf, [Accessed 23 August 2011]

Sciubba, E. (2005) From Engineering Economics to Extended Exergy Accounting: A

possible path from Monetary to Resource-Based Costing, Journal of Industrial Ecology,

Volume 8

Sciubba, E. and Wall, G. (2007), A Brief Commented History of Exergy from the Beginnings

to 2004, International Journal of Thermodynamics, vol. 10 pp. 1-26

Sciubba, E (2011) A revised calculation of the econometric factors and for the Extended

Exergy Accounting Method, Ecological Modelling, 222 (2011) 1060–1066

Sfez, S., Dewulf, J., DE Soete, W., Schaubroeck, T., Mathieux, F., Kralisch, D., DE Meester,

S. (2017) Toward a Framework for Resource Efficiency Evaluation in Industry:

264

Recommendations for Research and Innovation Projects, Resources 2017, 6, 5;

doi:10.3390/resources6010005

Short, M., Isafiade, A.J., Fraser, D.M., Kravanja, Z (2016) Synthesis of heat exchanger

networks using mathematical programming and heuristics in a two-step optimisation

procedure with detailed exchanger design. Chem. Eng. Sci., 44, 372–385

Simet, A. (2012) Global Costs for Biomass Power [Online], Available from:

biomassmagazine.com/articles/8344/global-costs-of-biomass-power, [Accessed 10

November]

Sirkin, T., and Houten, M. T. (1994) The Cascade Chain–A theory and tool for achieving

resource sustainability with Applications for Product Design. Resources, Conservation and

Recycling, 10 (1994) 213-277

Smith, R (2005) Chemical Process Design and Integration, West Sussex, England: John

Wiley & Sons

Smith, B. (2006) Anaerobic Digestion of Vinasse for the Production of Methane in the Sugar

Cane Distillery [Online], Available from: http://www.smithbaez.com/Download%20page

%20files/MethaneProductionfromVinasse.pdf, [Accessed 14 April 2014]

Srinivas, B.K., El-Halwagi, M.M. (1994) Synthesis of combined heat and reactive mass-

exchange networks. Chem Eng Sci. 49:2059-2074, 1994

Souza, M.E. (1986) Criteria for the utilisation, design and operation of UASB reactors.

Wat.Sci.Tech, 18 (12), 55–69

Stark, C. (2015) Reducing Energy Cost Through Reducing Energy Cost Through Boiler

Efficiency [Online], Available from: www.ncsu.edu/project/feedmill/pdf/E_Reducing

%20Energy%20Cost%20Through%20Boiler%20Efficiency.pdf, [Accessed 6 November

2015]

265

Stevens, L., 2014, BNSF Railway Boosts Safety Efforts [Online], Available from:

http://online.wsj.com/news/articles/SB10001424052702304275304579394983087734524,

[Accessed 15 April 2014]

Szargut, J, Morris, D, Steward, F (1988) Exergy analysis of thermal, chemical, and

metallurgical processes, Hemisphere Publishing Corporation, New York

Tan, Y.L., Manan, Z.A., Foo, D.C.Y. (2007) Retrofit of water network with regeneration

using water pinch analysis, Process Safety and Environmental Protection, Vol. 85 (B4) 305–

317

Taskhiri, M.S., Behera, Tan, R.R., Park, H-S. (2015) Fuzzy optimization of a waste-to-

energy network system in an eco-industrial park, J Mater Cycles Waste Manag (2015)

17:476–489 DOI 10.1007/s10163-014-0259-5

The Royal Academy of Engineering, (2011), Infrastructure, Engineering and Climate

Change Adaptation-Ensuring services in an uncertain future [Online], Available from:

www.raeng.org.uk/publications/reports/engineering-the-future, [Accessed 6 September

2016], ISBN 1-903496-61-6

Townsend, D.W., Linnhoff, B. (1983). Heat and power networks in process design. Part II:

Design procedure for equipment selection and process matching. AIChE J, 29(5):748–71.

UK Agriculture (2014a) Wheat – Farming and production [Online], Available from:

www.ukagriculture.com/crops/wheat.cfm, [Accessed 17 September 2014]

UK Agriculture (2014b) Potatoes in the UK [Online], Available from:

www.ukagriculture.com/crops/potatoes_uk.cfm, [Accessed 17 September 2014]

Ukidwe, N.U., Bakshi, B.R. (2004). Thermodynamic accounting of ecosystem contribution to

economic sectors with application to 1992 U.S. economy. Environmental Science Technology

15; 38(18):4810-27

266

Ukidwe, N.U., Bakshi, B.R. (2005). Flow of Natural versus Economic Capital in Industrial

Supply Networks and its Implications to Sustainability, Environmental Science Technology,

2005, 39 (24), pp. 9759–9769

Ukidwe, N.U and Bakshi, B.R (2007) Industrial and ecological cumulative exergy

consumption of the United States via the 1997 input-output benchmark model, Energy 32

(2007) 1560–1592

United Nations Environment Programme (UNEP) (2012). Global Outlook on Sustainable

Consumption and Production Policies: Taking action together [Online], Available from:

www.unep.fr/shared/publications/pdf/DTIx1498xPAGlobalOutlookonSCPPolicies.pdf,

[Accessed 28 November 2013]

United Nations Environment Programme (UNEP) (2016a) Global Material Flows and

Resource Productivity-Assessment Report for the UNEP International Resource Panel

[Online], Available from: unep.org/documents/irp/16-

00169_LW_GlobalMaterialFlowsUNEReport_FINAL_160701.pdf, [Accessed 12 December

2014]

United Nations Environment Programme (UNEP), (2016b) An Introduction to Rainwater

Harvesting [Online], Available from: gdrc.org/uem/water/rainwater/introduction.html,

[Accessed 4 July 2016]

UN (2013), World population projected to reach 9.6 billion by 2050–UN report [Online],

Available from: www.un.org/apps/news/story.asp?NewsID=45165#.V88FGPl97IU,

[Accessed 6 September 2016]

UN Water (2013) Safe use of wastewater in agriculture [Online], Available from:

http://collections.unu.edu/eserv/UNU:2661/proceedings-no-11_WEB.pdf, [Accessed 31

August 2016]

UN WATER, (2014) WATER, FOOD AND ENERGY NEXUS [Online], Available from:

http://www.unwater.org/topics/water-food-and-energy-nexus/en/, [Accessed 7 December

2015]

267

University of Strathclyde (UoS), (2001) The Green Islands Project [Online], Available from:

www.esru.strath.ac.uk/EandE/Web_sites/01-02/green_islands/m-economic.html, [Accessed

20 October 2014]

United States Department of Agriculture (USDA), (2009) Agricultural Waste Management

Field Handbook [Online], Available from:

www.nrcs.usda.gov/wps/portal/nrcs/detailfull/national/landuse/crops/npm/?

cid=stelprdb1045935, [Accessed 20 November 2015]

Valero, A., L. Ranz, et al. (2002). Exergetic evaluation of natural mineral capital (1)

Reference environment methodology. ECOS 2002, Berlin

Valero, A. (2008). Exergy Evolution of the Mineral Capital on Earth. Mechanical

Engineering, University of Zaragoza. Ph.D. Mechanical Engineering: 481.

Valero, A., Dominguez, A., Valero, A. (2013) Exergy replacement costs of mineral

resources, Journal of Environmental Accounting and Management, DOI:

10.5890/JEAM.2013.05.004

Varbanov, P. S., and Klemes, J. J. (2011) Integration and management of renewables into

Total Sites with variable supply and demand, Computers and Chemical Engineering, 35,

1815–1826.

Vasili, A. (2015) How Much Electricity Does the Average Solar Panel System Generate?

[Online], Available from: www.theecoexperts.co.uk/how-much-electricity-does-average-

solar-panel-system-generate, [Accessed 21 November 2015]

VESTAS (2015) Wind energy [Online], Available from:

ftp://ftp.campbellsci.co.uk/pub/outgoing/info/Vista%20Datavision/their%20website

%20content%20files/Help/db.data.browser/wind_energy.htm, [Accessed 8 March 2015]

Vivekanand, V., Olsen, E., Eijsink, V., Horn, S. (2014) Steam-exploded bagasse,

BioResources 9(1), 1311-1324

268

Wall, G. (1977) Exergy - a useful concept within resource accounting, Goteborg, Institute of

theoretical physics

Wall G. (1999), Conditions and tools in the design of energy conversion and management

systems of a sustainable society. In: Proc ECOS’99, Tokyo. 1999. p. 1231–8.

Wall, G. (2002). Conditions and tools in the design of energy conversion and management

systems of a sustainable society. Energy Convers. Manage. 43 (2002) 1235–1248.

Wall, G. (2011). Tools for Sustainable Energy Engineering, World Renewable Energy

Congress 2011-Sweden

Wang, Y.P. and Smith, R., (1994) Wastewater minimisation. Chem. Eng. Sci. 49, pp. 981-

1006

Ward, S. (2010) Rainwater harvesting in the UK: a strategic framework to enable transition

from Whitehill and Bordon. (2012) Eco-town Master plan. Hampshire, UK. novel to

mainstream, PhD thesis, University of Exeter.

WATERWISE (2016) Water-The facts [Online], Available from:

www.waterwise.org.uk/data/resources/25/Water_factsheet_2012.pdf, [Accessed 21 August

2016]

Wilbanks, T. J. and Kates, R.W. (1999) Global change in local places: how scale matters.

Climatic change 43(3): 601-628.

Williams, A.G., Audsley, E. and Sandars, D.L. (2006) Determining the environmental

burdens and resource use in the production of agricultural and horticultural commodities,

Main Report, Defra Research Project IS0205. Bedford: Cranfield University and Defra

Available on www.silsoe.cranfield.ac.uk and www.defra.gov.uk,

Wittmus, H., Olson, L., Lane, D (1975) Energy requirements for conventional versus

minimum tillage. J of Soil and Water Conservation 1975; 3:72-5

269

Whitehill and Bordon (2012). Eco-town Masterplan. Hampshire, UK.

Wolfe, M.L., Ting, K.C., Scott, N., Sharpley, A., Jones, J.W., Verma, L. (2016) Engineering

solutions for food-energy-water systems: it is more than engineering. J Environ Stud Sci,

6:172–182

Xu, S., Bai, Z., Jin, B., Xiao, R., Zhuang, G., (2014) Bioconversion of wastewater from sweet

potato starch production to Paenibacillus polymyxa bio fertilizer for tea plants, Scientific

Reports 4, DOI:10.1038/srep04131

Yang, Z.F., Jiang, M.M., Chen, B., Zhou, J.B., Chen, G.Q., Li, S.C. (2010). Solar emergy

evaluation for Chinese economy. Energy Policy. 39, 875-886.

Yang, S., Yang, S., Qian, Y. (2015). The inclusion of economic and environmental factors in

the ecological cumulative exergy consumption analysis of industrial processes. Journal of

Cleaner Production, 108 (2015) 1019-1027

Yi, H-S., Hau, J.L., Ukidwe, N.U., Bakshi, B.R. (2004). Hierarchical Thermodynamic

Metrics for Evaluating the Environmental Sustainability of Industrial Processes. Environ

Prog. Vol. 23, No.4, DOI: 10.1002/ep.10049

You, F., Tao, L., Graziano, D.J., Synder, S.W. (2011) Optimal Design of Sustainable

Cellulosic Biofuel Supply Chains: Multiobjective Optimization Coupled with Life Cycle

Assessment and Input-Output Analysis. AIChE J. 2012 Vol. 58, No. 4

Zaleta-Aguilar, A., L. Ranz, et al. (1998). Towards a unified measure of renewable resources

availability: the exergy method applied to the water of a river. Energy Conversion and

Management, 39(16-18): 1911-1917.

Zhang, Y. (2008) Ecologically-based LCA an approach for quantifying the role of natural

capital in product life cycles, PhD thesis, The Ohio State University

270

Zhang, Y., Singh, S., Bakshi, B. R. (2010). Accounting for Ecosystem Services in Life Cycle

Assessment, Part I: A Critical Review. Environ Sci Technol. 44, 7, 2232-2242

Zhang, B., Peng, B., Liu, M. (2012) Exergetic Assessment for Resources Input and

Environmental Emissions by Chinese Industry during 1997–2006, The Scientific World

Journal, Vol. 2012

Zhou, Z., Zhang, J., Liu, P., Li, Z., Georgiadis, M.C., Pistikopoulos, E.N. (2013) A two-stage

stochastic programming model for the optimal design of distributed energy systems, Applied

Energy 103 (2013) 135–144

271