university, denmark jannick/icta.uab.cat/ecotech/sudoe/advanced_course/_input-output lca, js 8th...

Input-output LCA

Jannick H SchmidtBarcelona 8th November 2012

http://people.plan.aau.dk/~jannick/

CEO, 2.-0 LCA consultants Associate Professor, PhD, Aalborg University, Denmark

www.lca-net.com

2

Terminology of life cycle inventoryThe very basic of LCAProcess LCA and IO-LCAUses of IO-LCADerivation of supply-use tables and IO-tableTrade-linkingSupply use data and Mass flow analysisFuture challenges

Now

Content of the lecture

From product flow diagram to matrix- Product system

3

B C

A

D

2 kg 5 kg

6 kg

3 kg

3 kg CO2

1 kg CO2

2 kg CO2

4 kg CO2

System boundary between technosphere and nature

Functional unit/reference flow

Output: product (reference flow)________________ Output: co-products Input: products from other processes Output: waste sent to treatment/recycling________________ Emissions Resources

From product flow diagram to matrix- What is an LCA process?

4

Other processes canlink to this flow

Inputs/outputs from other processes

Exchanges with the environment

LCA activity = LCA-process = activity with inputs and outputs

From product flow diagram to matrix- no co-products

5

B C

A

2 kg 5 kg

6 kg

3 kg CO2

1 kg CO2

2 kg CO2

D

3 kg

4 kg CO2

Functional unit/reference flow

System boundary between technosphere and nature

Process AOutputProcess AOutput 6

InputsABCD

InputsAB 2C 5D

EmissionsCO2

EmissionsCO2 1

B C D2 5 3

3

2 3 4

NormalisedProcess A B C DOutput 1 1 1 1

InputsAB 0.33C 0.83D 1.50EmissionsCO2 0.17 1.00 0.60 1.33

Z~Direct requirement coefficient matrix

2 3 4EmissionsCO2 1

InputsAB 2C 5D

From product flow diagram to matrix- with co-products

6

3

Direct requirement coefficient matrix depends on allocation/system expansion

Process AOutputsABCD

Process AOutputsA 6BCD

B C DB C D

251 3

7


Now


8

The very basics of IO-LCA

The inventory problem: Sector A) 1 USD A, uses 0.6 USD B Sector B) 1 USD B, uses 0.39 USD A I want to analyse 1 USD B – how much of

sector A and B is needed? Solution: Use an LCA software

9

LCA software: SimaPro

1 USD B: Result in SimaProSector A: A Sector B: B

10

What is the underlying calculation?

Convention: Output = positive values, and input = negative values Our inventory problem: two equations with two unknowns:

Sector A: Sector B: Demand

B USD1A USD0

B USD1A USD39.0

B USD0.6A USD1

2

2

1

1

ss

ss

Scaling factors Solve equation:

1.3s and 51.0 21 s

11

Let's formalise the notion...

B USD1A USD0

B USD1A USD39.0

B USD0.6A USD1

2

2

1

1

ss

ss

B USD1B USD1 0.6-A USD39.0A USD1

A

Technology matrix (A) Final demand vector (f)

B USD1A USD0

f

2

1

2

2

22

12

1

1

21

11

ff

ss

aa

ss

aa

fAs

fAs -1

3.151.0

10

3.178.051.03.1

B USD1A USD0

B USD1B USD0.6-A USD39.0A USD1

s

s

s

fAs-1

-1

12

EmissionsEmissions matrix (B)

ion arsenic, kg 0ion arsenic, kg 11E0.4SO kg 0.0012SO kg 0

CO kg 0.85CO kg 0

22

22

B

SectorA Sector B

Resulting emissions (g)

fABBsg -1

Formalised formula

ion arsenic, kg 11-2.0ESO kg 0.0016

CO kg 1.1

ion arsenic, kg 0SO kg 0.0012

CO kg 0.85

ion arsenic, kg 11-4.0ESO kg 0CO kg 0

2

2

3

2

1

3

2

1

2

2

2

2

2

1

1

1

2

2

ggg

ggg

sss

sss

13


Now


14

Process LCA and Input-Output LCA In process LCA we have the technology matrix (A)

In Input-Output LCA we have an IO-table also called coefficient transaction matrix ( )

specifies the input per unit of output

If A has only ones on the diagonal (normalised), then:

For IO-LCA we now have:

elec.kWh 1elec.kWh 0.6-coal kg 39.0coal kg 1

A

Z~

Z~

elec.elec./kWh kWh 0coal elec./kgkWh 0.6elec.coal/kWh kg 39.0coal coal/kg kg 0~Z

AIZ ~

fZ-Is

fAs-1

-1

~

fZIBg

fABBsg-1

-1

~

Format in process LCA

Format in IO LCA

15

Input-Output LCA: Units of the IO-table Units of the coefficient transaction matrix ( )

Typically monetary units, e.g.

If different units: Hybrid unit IO-model Example where coal is in physical units:

are related via prices, e.g. USD/kg coal

Z~

$/$ 0$/$ 0.5$/$ 1.0$/$ 0~Z

elec. elec./$ $ 0coal elec./kg $ 0.5elec. coal/$ kg 1.0coal coal/kg kg 0~Z

PZZ physicalmonetary *.~~

elec.$/kWh .0700elec.$/kWh 0.070coal $/kg 0.081coal $/kg 0.081

P

physicalmonetary ZZ ~ and ~

16

Process LCA versus IO-LCA

Z~

Z~

Z~

So far: IO-LCA = process-LCA What is the difference?

IO-LCA: Process-LCA: A

How do we derive A in process LCA? Follow physical flows in process-diagram Buttom-up approach

How do we derive in IO-LCA? Using information provided by statistical agencies: National accounts (supply

and use tables) Top-down approach

Difference = the way A and are derived

17

Process LCA versus IO-LCA – Data sources Process data (bottom-up)

Enterprise green accounts / direct data collection Emission inventories Technical handbooks and risk assessment estimates Industry LCA databases Commercial LCA databases

Input-Output data (top-down) Statistical agencies (based on industry data) National projects to complete the emissions data

18


RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRawCapital

EnergyCapitalRaw

EnergyCapitalRaw

EnergyCapitalRaw

CapitalRawEnergy

Capital

RawRaw

RawEnergyCapital

RawEnergy

CapitalRaw

CapitalEnergy

Raw

RawEnergyCapital

RawEnergy

CapitalRaw

CapitalRawEnergy

Capital

RawRaw

RawEnergyCapital

RawEnergy

CapitalRawRawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRawEnergy

EnergyCapital

EnergyCapital

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw RawEnergyCapital

RawEnergy

CapitalRaw

CapitalRawEnergy

Capital

RawRaw

RawEnergyCapital

RawEnergy

CapitalRaw

CapitalRawEnergy

Capital

RawRaw

RawEnergyCapital

RawEnergy


RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy


RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

Energy

CapitalRaw

EnergyCapitalRaw

EnergyCapitalRaw

Energy

CapitalRawEnergy

Capital

Raw

EnergyCapital

Raw

EnergyCapital

RawEnergyCapital

RawEnergy

CapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw RawEnergyCapital

RawEnergy

CapitalRaw

CapitalRawEnergy

Capital

RawRaw

EnergyRaw

EnergyCapitalRaw

EnergyCapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

Raw

Capital

Raw

Energy

Capital

aw

EnergyCapitalRaw

EnergyCapital

Raw

Energy

awEnergyCapital

RawEnergy

CapitalRaw

EnergyRaw

EnergyCapitalRaw

EnergyCapitalRaw

RawEnergyCapital

RawEnergy

CapitalRaw

Raw

Capital

Raw

nergy

Capitalaw

EnergyCapitalRaw

EnergyCapitalRaw

Energy

Raw

Capital

Energy

CapitalRaw

Energ

Capital

EnergCapital

Capital

Raw

Energy

Capit

Raw

CapitalEnergy

Capital

Raw

Raw

Capital

EnergCapital

CaRaw

Energy

Capit

Raw

CapitalEnergy

Capital

Raw

Capital

Raw

EnergCapital

Capital

Raw

Energy

Raw

CapitalEnergy

CapitalRaw Raw

CapitalEnergy

RawRaw

Capital

Energy

RawRaw

CapitalEnergy

Capital

RawCapital

Capital

Capital

EnergypitalRaw

EnergyCapitalRaw

EnergyCapitalRaw

Raw

Energy

Capital

CapitalEnergyRaw

Raw

Raw

Raw

Energy

Capital

EnergyCapital

Raw

Capital

Raw

CapitalEnergy

pital

aw

EnergyCapital

CapitalRaw

Capital

Raw

CapitalEnergy

pital

Capital

aw

EnergyCapital

Raw

Capital

Capital

Capital

Capital

Capital

Capital

Capital

Raw

RawRawRaw

Raw

Raw

Energy

Raw

EnergyRawCapital

Energy

Raw

Capital

Raw

Raw

Raw

Capital

Energy

Raw

Use

Raw A

Manuf.

Raw B

Disp.

RawEnergy

Capital

Capital

CapitalEnergy

Raw

Do you think we remember it all inprocess-LCA?


IO-LCA: No cut-off Automatical inclusion of 'hard to quantify' processes

Accounting, marketing, use of post services Business travelling Office machinery, printing paper, pencils, furniture etc. Capital goods (buildings, machinery)

19

Process LCA versus IO-LCA – Results

20

Process-data:Ecoinvent and LCAfood

IO-data:Swedish IO-table(FORWAST project)

21


Now


22

Uses of IO-LCA What can IO-LCA be used for?

Complete green accounts Short-cut to screening LCAs Provide inputs to process LCA => increase completeness (hybrid LCA) LCA of nations/regions/municipalities: production and consumption LCA of international trade

23


Now


24

Derivation of the IO-table

How is derived? Supply (V') and use (U) tables

Z~

Products (c)

Monetary supply table Total

Total

V'o

g'

qNc

ImportActivities (a) Monetary use table

Products (c)

Activities (a) Total

Uo

Final uses

Cap

ital f

orm

atio

n

Ec q

Primary inputs

Labour

Taxes

Profit

Use of fixed capital

Total g'

y0

Emissions B

GDP

Final demand

Export

Investments

25

Derivation of the IO-table

Supply (V') and use (U) tables are balanced

Balanced MSUT

Products (c)

Total

U

Needsfulfilment

y Ec q

Primary inputs

Labour

Taxes

Profit

Total g'

Products (c) V'

Activities (a) Import

qNc

Total g'

Export

26

From purchaser’s prices to basic prices What is purchaser’s price? What is basic price? Why have table in basic prices?

q is in purchaser’s price U => purchaser’s price V’ + valuation => purchaser’s

price

Methodology and procedure:Schmidt J H, Weidema B P, and Suh S (2010), Documentation of the final model used for the scenario analyses. Deliverable 6-4 of the EU FP6-project FORWAST. http://forwast.brgm.fr/

27

Derivation of the IO-table (simple case) Simplified case: No co-products

Z~

1-gUZ ˆ~ U normalised by total supply from activities (g)

28

Derivation of the IO-table (normal case) How to deal with co-products?

Two options: Substitution Co-product allocation

29

Derivation of the IO-table (normal case) System expansion (by-product technology assumption)

1-gUZ ˆ~

U normalised by total supply from activities (g)

Derivation of the IO-table (normal case) Substitution (by-product technology assumption)

30

1ˆ~~ VV-UZ

31

Derivation of the IO-table (normal case) Economic allocation (industry-technology assumption)

q

g'

11 ˆˆ~ qV'gUZ

V’ normalised by total supply of products (q)=> average market supply, similar to attributional modelling

U Normalised by total supply from activities (g)=> Economic allocation similar to attributional modelling

32

Allocation problems in the SUT framework Years of discussions on difficulties in substitution is now solved by

using a simple matrix operation Allocation versus substitution: Two different matrix formulas

Suh S, Weidema B, Schmidt J H and Reinout H (2010), Generalized Make and Use Framework for Allocation in Life Cycle Assessment. Journal of Industrial Ecology 14(2): 335-353 < DOI: 10.1111/j.1530-9290.2010.00235.x>

33


Now


34

Tradelinking Example of tradelinked DK, EU27 (ex DK) and RoW IO-table

IO-table:DK

Import to DK from EU27

Import to DK from RoW

IO-table:EU27 (ex DK)

IO-table:RoW

DK NAMEA:GHG-emissions

EU27 NAMEA:GHG-emissions

RoW NAMEA:GHG-emissions

Import to EU27 (ex DK)

from RoW

Import to EU27 (ex DK)

from DK

Import to RoW from

EU27 (ex DK)

Import to RoW from

EU27 (ex DK)

35


Now


36

FORWAST: Project under the 6th European Union Framework Programme Purpose: Overall mapping of physical flows and stocks of resources to

forecast waste quantities in Europe and identify life-cycle environmental stakes of waste prevention and recycling

Technical lead: 2.-0 LCA consultants Partners: 6 partners throughout the EU Duration: March 2007 to December 2009 Project homepage: http://forwast.brgm.fr/

37

CREEA: Project under the 7th European Union Framework Programme Purpose: Compiling and Refining Economic and Environmental Accounts MFA / waste lead: 2.-0 LCA consultants Partners: 12 partners throughout the EU Duration: April 2011 to April 2014 Project homepage: http://creea.eu

38

CREEA WP4: Waste and MFA accounts

UP

D0*UP

Gc

diag(f0*rT)

Feedstock present in supply of products

rT

gR

S+WV

Emissions

Use of products

Ressource inputs

Emissions

wUUse of waste

diag(re0*wU)

gW

Emissions

How to calculate waste generation

39

Physical supply and use table (dry matter mass)- preliminary draft of use table (UP)- waste rows = 0 Product material

composition (water content)

a) D

omes

tic p

rodu

ctio

n

b) Im

port

Dom

estic

use

= a

+b-c

c) E

xpor

t

Physical supply PhysicaluseIm

port

Fina

l use

Expo

rt

Physical data (dry mass)

Inve

stm

ents

Statistical data

Fina

l use

Inve

stm

ents

PSUTMSUT


40

Auxiliary data: Feedstock specification (D1)

Resource efficiency (f0) Unused extraction rate

Recycling efficiency (e0)

D1

1

0.8

1

1

1

1

1

Activities

Prod

ucts


41

Calculation of waste generation: feedstock efficiency (D0): From D1 to D0

waste+stock addition (WV+S)

split waste from stock addition

e.g. use investment info in use table or general data from FORWAST

=> supply of waste table WV

WRUU0UT0TP0TV ggGw*rewr*frU*DUWΔS ˆˆˆˆ diagdiag

Pij,1

P0,1ij,1U0T0Pij0,

ij1ij0,

ij0,

ij,1

ij,1

ij,1

ij

U*Di'

./UDi'w*rei'r*fi'V'i'iDDD

D

DDD

D

1

*

0

11,0

0

Was

te(p

rodu

ct c

lass

ifica

tion)


42

Waste (material for treatment) in the supply-use tabels


43

Input-output LCA – future challenges Disaggregation

The most detailed IO-table: 480 products, but 100 is normal Ecoinvent (process LCA): >4000 processes

Waste treatment is normally not correctly modelled Virgin production and recycling are mixed Trade with e.g. steel scrap is classified as trade with steel

It takes time to create IO-tables EU27 IO-table for 2003 is available in SimaPro USA table for 2002 Exiopol table for 2000

Tradelinked IO-tables are rare

If you want to know more…Choose among our courses atwww.ilca.es

44

university, denmark jannick/icta.uab.cat/ecotech/sudoe/advanced_course/_input-output lca, js 8th...

Documents