43 simulink model for economic analysis and

692 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 20, NO. 2, MAY 2005

Simulink Model for Economic Analysis andEnvironmental Impacts of a PV With Diesel-Battery

System for Remote VillagesRichard W. Wies, Member, IEEE, Ron A. Johnson, Ashish N. Agrawal, Student Member, IEEE, and

Tyler J. Chubb, Student Member, IEEE

Abstract—This paper discusses the economic analysis andenvironmental impacts of integrating a photovoltaic (PV) arrayinto diesel-electric power systems for remote villages. MATLABSimulink is used to match the load with the demand and appor-tion the electrical production between the PV and diesel-electricgenerator. The economic part of the model calculates the fuelconsumed, the kilowatthours obtained per gallon of fuel supplied,and the total cost of fuel. The environmental part of the modelcalculates the 2, particulate matter (PM), and the x

emitted to the atmosphere. Simulations based on an actual systemin the remote Alaskan community of Lime Village were performedfor three cases: 1) diesel only; 2) diesel-battery; and 3) PV withdiesel-battery using a one-year time period. The simulation resultswere utilized to calculate the energy payback, the simple paybacktime for the PV module, and the avoided costs of 2, x,and PM. Post-simulation analysis includes the comparison ofresults with those predicted by Hybrid Optimization Model forElectric Renewables (HOMER). The life-cycle cost (LCC) and airemissions results of our Simulink model were comparable to thosepredicted by HOMER.

Index Terms—Energy payback period, greenhouse emissions,hybrid power system, photovoltaic (PV) array, power systemmonitoring, remote terminal unit.

I. INTRODUCTION

THE NEED for energy-efficient electric power sources inremote locations is a driving force for research in hybrid

energy systems. Power utilities in many countries around theworld are diverting their attention toward more energy- efficientand renewable electric power sources. Reasons for this interestinclude the possibilities of “taxes” or other penalties for emis-sions of greenhouse gases as well as other pollutants plus the fi-nite supply of fossil fuels. The use of renewable energy sourcesin remote locations could help reduce the operating cost throughthe reduction in fuel consumption, increase system efficiency,and reduce noise and emissions [1], [2]. In some remote villages,including Lime Village, Alaska, stand-alone hybrid power sys-tems are often more cost effective than utility grid extensions,mainly due to the high cost of transmission lines.

Manuscript received March 26, 2004; revised August 20, 2004. This workwas supported by the Arctic Energy Technology and Development Laboratory(AETDL) under Grant G00000406 with the United States Department of En-ergy. Paper no. TPWRS-00167-2004.

The authors are with the Electrical and Computer Engineering Depart-ment, University of Alaska, Fairbanks, AK 99775-5915 USA (e-mail:[email protected]; [email protected]).

Digital Object Identifier 10.1109/TPWRS.2005.846084

Based on energy consumption studies compiled by the U.S.Department of Energy, Alaska spends $28.71 per million BTUfor electrical energy versus $19.37 per million BTU for the restof the United States [3]. It is very expensive to transport fuelfor diesel electric generators (DEGs) in some villages of Alaska[4] due to the remoteness of the site. Furthermore, there are is-sues associated with oil spills and storage of fuels [5]. There-fore, photovoltaic (PV), wind, and other renewable sources ofenergy are being integrated with DEGs to help reduce the fuelconsumed by the DEGs.

This paper presents a model based on an existing hybrid elec-tric power system for a remote location in the Alaskan commu-nity of Lime Village. The input data to the model are acquiredusing a remote terminal unit (RTU) that must first be installed atthe site. The RTU allows for remote data collection and systemcontrol while also providing information necessary for mod-eling the hybrid power system. The information from the RTUcan be processed using the model described in this paper. In thisway, the RTU and the model can be used to optimize the perfor-mance of the hybrid power system.

MATLAB Simulink is used to model the system and apportionthe electrical production between the PV array and diesel-electricgenerator. In general, the Simulink model can be used to studythe performance of any hybrid power system. Using Simulink,other renewable energy sources, dynamic operation, and controlsystem strategies can be easily incorporated into the existing hy-brid power system model to study the overall performance of thesystem. Simulations are performed for three cases: 1) diesel only;2) diesel-battery; and 3) PV with diesel-battery using a one-yeartime period. The results of the simulations are used to performan economic analysis and predict the environmental impacts ofintegrating a PV array into diesel-electric power systems for re-mote villages. The economic part of the model calculates the fuelconsumed, thekilowatthoursobtainedpergallonof fuel supplied,and the total cost of fuel. The environmental part of the model cal-culates the CO , particulate matter (PM), and the NO emitted tothe atmosphere. These results are then utilized to calculate the en-ergypayback, thesimplepaybacktimefor thePVmodule,and theavoided costs of CO , NO , and PM.

II. HYBRID POWER SYSTEM MODEL

A. General Model

In general, when two or more different sources of electricityare connected to a common grid and operate hand in hand to

0885-8950/$20.00 © 2005 IEEE

WIES et al.: SIMULINK MODEL FOR ECONOMIC ANALYSIS AND ENVIRONMENTAL IMPACTS 693

Fig. 1. General hybrid power system model.

supply the desired load, the system becomes a hybrid electricpower system. A simple block diagram of a hybrid power systemis shown in Fig. 1. The sources of electric power in this hy-brid system consist of a diesel generator, a battery bank, a PVarray, and a wind generator. The diesel generator is the mainsource of power for many of the remote villages in Alaska [5]and around the world. The output of the diesel generator is reg-ulated ac voltage, which supplies the load directly through themain distribution transformer.

The battery bank, the PV array, and the wind turbine are in-terlinked through a dc bus. The RTU regulates the flow of powerto and from the different units, depending on the load. The in-tegration of a RTU into a hybrid power system is important toenhance the performance of the system [6]. The overall purposeof the RTU is to give knowledgeable personnel the ability tomonitor and control the hybrid system from an external controlcenter. Since the hybrid systems of interest in this research arelocated in remote areas, the ability for external monitoring andcontrol is of utmost importance. The RTU is interfaced with avariety of sensors and control devices located at key locationswithin the hybrid system. The RTU processes the data fromthese sensors and transmits it to a control center. In addition, theRTU is also capable of receiving control signals and adjustingparameters within the system without the physical presence ofthe operating personnel.

B. Lime Village Model

This paper investigates the integration of a PV array with adiesel-battery hybrid electric power system located in Lime Vil-lage, Alaska. The hybrid power system of Lime Village consistsof 21- and 35-kW diesel generators, 100 kWh (95 two-volt cells)of valve-regulated lead acid batteries, and a 12-kW PV array.The PV array consists of 8 kW of BP275 solar panels and 4kW of Siemens M55 solar panels. Wind generation is not a vi-able renewable energy source for Lime Village due to the lowwind speeds in this area. A 30-kVA bidirectional power con-verter/controller is used to supply power to and from the bat-

teries and from the PV array. Figures from the Alaska EnergyAuthority (AEA) show that the operating cost of fuel suppliedfor the generators of Lime Village ranges from $2.80 per gallonin summer to $4.80 per gallon in winter [4]. Due to the high costof fuel, it is desired that the diesel generators operate efficientlyand economically. The use of renewable energy in the form of aPV array combined with regulated battery storage helps in con-straining the use of the diesel generator while optimizing theefficiency and economics of the system. Efforts are already un-derway to install an RTU at Lime Village, further enhancing theperformance of the system.

III. SIMULATION MODEL

A model of a hybrid power system of Lime Village wasdesigned using MATLAB Simulink. The Simulink model wasdeveloped so that it can be used to study the performance of anyhybrid power system. Using the -function in Simulink, blocksrepresenting other renewable energy sources can be easilyincorporated into the existing hybrid power system model.Simulink also allows the dynamic operation and the controlsystem strategy to be incorporated into the hybrid power systemmodel to study the dynamic performance of the system. Theoverall block diagram of the current system is shown in Fig. 2.The model consists of nine different subsystems contained inblocks. The Input Parameters block includes data files obtainedfrom the site. After the installation of the RTU, the modelwill acquire the data directly from the RTU. This data can beused by engineers and operators to evaluate and optimize theperformance of the system.

Sensors on the system are used to gather information, suchas the amount of sunlight incident upon the PV arrays, chargelevel of the batteries, and important operating parameters of thediesel generator. The voltage or current signals from these sen-sors are transmitted to signal conditioning devices that convertthe signals to an instrumentation level. These signals are thenpassed to analog input modules of the RTU and digitized forprocessing. The processing consists of scaling the inputs and


Fig. 2. Simulink model of the Lime Village power system.

converting them to a meaningful unit. The data is then savedwithin the memory of the RTU and unloaded to a database on acentral server at a location outside of the village at a user-speci-fied time interval. The data are transferred through TCP/IP con-nections and are usually accomplished through dial-up/ethernetconnections with the RTU. At this point, the data are placed in adatabase and accessed via a web page or other methods and areavailable as input to the model [7].

The input data files to the model are the system electrical load,solar insolation values, ambient temperature, and the kilowattratings of the different energy components. The Simulink modeldeveloped here uses data from the manufacturer to calculate theefficiency and the amount of fuel used for the DEG. Knowingthe above parameters, the Simulink model can be used to studythe performance of any hybrid power system.

After being processed by the Input Parameters block of themodel, this information is used by all of the other subsystems tocalculate efficiency, fuel consumption, and total cost of fuel.

The PV Model block is the model of the 12-kW PV arrayinstalled at Lime Village. This block calculates the power avail-able from the PV array, depending on the intensity of sunlight.The -function written in MATLAB performs the followingtasks.

1) The total power available from the PV array (aligneddue south and tilted at a 15 angle) is calculated usingthe solar insolation values, the total area of the col-lector, and the efficiency of the solar collector. Thesolar insolation values were obtained as the input ofthe PV Model from the output of the Input Parame-ters block. These input values were obtained using asolar map developed by the National Renewable En-ergy Laboratory (NREL). This map utilizes extrapo-lations of 30-yr data from measurements at other lo-cations combined with satellite data on cloud cover[8]. The total collector area for the PV array was ob-tained from the manufacturer data sheet. The efficiencyof commercially available solar collector is about 15%[8]. In this project, a collector efficiency of 12% is as-sumed.

2) The model compares the calculated PV power to therequired load. If the PV power is more than the load onthe system, the model checks the battery kilowatthours.If the battery kilowatthours is less than 95% of its ratedkilowatthours, the model will send the excess availablepower to charge the battery bank. On the other hand,if the kilowatthours rating of the battery is more than95% of its rated kilowatthours, the model will sendthe excess power to the dump load. The dump loadconsists of resistive banks that can adsorb excess poweravailable from the PV array, which can subsequently beused to provide space heating. Lime Village does notcurrently have dump load. If the PV power is less thanthe load on the system, all of the power available fromthe PV array will go to the load. The battery bank willsupply the remaining load. If the battery bank is unableto supply the rest of the load, the load is passed to thediesel generator. The diesel generator then supplies theload and charges the battery bank simultaneously.

The hybrid power system is designed in such a way that thePV array has the highest priority to supply the load. If the loadis not met by the PV power, the battery bank is used to supplythe required load. If the battery bank is less than 20% charged,the controller sends the signal to start up the diesel generator.The diesel generator is then used to supply the desired load andcharge the battery bank simultaneously. On the other hand, ifthere is excess power available from the PV array, the excesspower is used to charge the battery bank. If the battery bank is95% charged, the excess power is sent to a resistive dump load,which can be used for space-heating purposes. In the Simulinkmodel, the roundtrip efficiency of the rectifier/inverter and theinternal loss in the battery bank per cycle was considered as90%.

The Battery Model block consists of the battery bank andcontroller. The Battery Model has the second highest priorityto supply the load. Once the RTU is installed at Lime Village,it will regulate the power output of the diesel generator, the PVarray, and the battery bank through digital/analog output capa-bilities that enable equipment to be switched “on” and “off.” Thecontrol settings and set point configurations are programmed


into the memory of the RTU. These set points of the RTU can bechanged while the simulation is in progress in order to furtheroptimize the system.

The -function in the Battery Model block performs the fol-lowing tasks.

1) The total battery voltage is calculated using the numberof battery cells ( ) and the voltage per cell as follows:

battery volt volt per cell (1)

where volt per cell is obtained from the output of theInput Parameters block.

2) The model then compares the required load with themaximum capacity of the two generators. If the re-quired load exceeds the capacity of the two genera-tors, then the model displays the message that the loadcannot be supplied with the available generators. If theload is less then the maximum capacity of the two gen-erators, then the model checks for the available kilo-watthours and the mode (charging or discharging) ofthe battery bank. If the available kilowatthours of thebattery bank is greater than 20% and the battery is inthe discharging state, then the battery energy will beused to supply the load. If the available kilowatthoursof the battery bank is less than 20% of its rated kilo-watthours or if the battery bank is in the charging stage,then the energy from the diesel generator will be usedto supply the load and charge the battery bank simul-taneously.

The Generator Model block contains the manufacturer’sspecifications for the efficiency of the electric generator.Knowing the efficiency and the load on the generator, the powerinput to the generator can be calculated as

(2)

where is the load on the diesel generator.The Generator Model block is designed in such a way that the

diesel generators are always operating at 95% of their kilowattrating while operating in conjunction with the battery bank andthe PV array. This way, the generators operate at their maximumefficiencies and also give better displacement power factor. Ifone generator is insufficient to supply the load, the second gen-erator is turned “on.” In Lime Village, one generator is alwayssufficient to supply the load, while the other generator acts as aback-up generator. If the model is used for other villages wheretwo generators are used to supply the load, the percentage loadon both the generators is the same. Therefore, both generatorsoperate at 95% of their kilowatt rating.

The Fuel Consumption Model block calculates the amount offuel required by the diesel engine to supply the load. The fuelconsumed by the engine depends on the load and the electricalefficiency of the generator. The electrical efficiency is depen-dent on the displacement power factor of the load. If there aretwo generators operating, the block will calculate the fuel re-quired by each engine and also the total fuel required to supplythe load. The plot for the fuel consumption obtained from the

manufacturer’s data sheet can be mathematically interpreted asfollows:

lbs (3)

gallonsFuel consumed lbs

(4)

where is the input power to the generator given in kilo-watts; 7.1 is the factor that converts pounds (lbs) to gallons, de-pending on the type of fuel that is used. For different types ofgenerators, the fuel consumption can be obtained from the man-ufacturer’s data sheet.

The Error block calculates the difference between the sup-plied power and the required power. The error within the modelis calculated by

Error power (5)

where is the power supplied by the battery bank, is thepower supplied by the diesel generators, is the power sup-plied by the PV array, is the power delivered to the load, and

is the power delivered to the dump load.The Error block also calculates the rms value of the error

power. The rms value of the error will depend on the time in-terval over which the simulation is performed, the time incre-ment between the two simulation steps, and the fluctuation inload. The rms value of error is given by

Error power rmsInstantaneous value

(6)

where is the ratio of the simulation time to the time incrementbetween the two simulation steps.

The Calculate Other Parameters block calculates the param-eters such as the total kilowatthours per gallon supplied by thegenerator, fuel consumed in pounds and gallons, the total costof fuel (in U.S. dollars), the , particulate matter (PM), andthe emissions. The kilowatthours per gallon and total fuelcost are calculated as

kWhgallon

kWh(7)

Total cost USDcost

gallon(8)

where kWh is the total kilowatthours supplied by the dieselgenerator, and is the total fuel consumed (in gallons).

The Display Parameters block is used to display all the calcu-lated parameters, including the fuel consumption, the total costof fuel, the kilowatthours per gallon, and the amount of green-house gases emitted to the environment.

IV. MODEL VALIDATON

In Alaska, there is less sunlight available during the wintermonths and, therefore, very few diesel-hybrid power systemsthat incorporate PV arrays. As a result, field data are not easilyavailable for the PV with diesel-battery system to validate the


Fig. 3. Annual load profile for Lime Village.

Fig. 4. Solar insolation profile for Lime Village.

model. In order to validate the model, simulations were per-formed for the PV with diesel-battery system for the load profileshown in Fig. 3. This load profile was obtained by interpolatingand averaging a 24-h summer load profile and a 24-h winter loadprofile obtained from Lime Village over a one-year time period.Each data point represents a daily average, and a second-orderpolynomial fit to the data is used, as shown in the figure.

The solar insolation profile for Lime Village is shown inFig. 4. A third-order polynomial fit to the data is used, as shownin the figure. It can be observed from this plot that duringsummer days, there is abundant sunlight; hence, the energyavailable from the sun is distributed throughout the day. If thereis any extra power available from the PV array after supplyingthe load, it is utilized to charge the battery bank.

The post-simulation analysis includes the comparison of re-sults from the Simulink model with those predicted by Hy-brid Optimization Model for Electric Renewables (HOMER).HOMER online was released in fall 2001 at NREL. HOMER

is a computer model that designers can use to simulate and op-timize standalone electric power systems. HOMER can modelany combination of wind turbines, solar PV panels, run-on-riverhydro, small modular biomass, conventional generators (diesel,propane, and gasoline), and battery storage. The trial versionand description of HOMER are available in [8].

HOMER can be used to design a stand-alone power systemin a remote village, investigate the cost of powering an off-gridhouse, and assess the potential of renewable energy. The com-parison results of Simulink model with HOMER are describedin more detail in Section V-C.

V. SIMULATIONS AND RESULTS

Simulations were performed for three cases using the LimeVillage model and a one-year time period. The three casesstudied in this work include diesel-only system, diesel-batterysystem, and PV with diesel-battery system.

Table I shows the costs of the different components installedat Lime Village for the three cases. The costs of the differentcomponents were obtained from the various manufacturers. Theengineering cost, commissioning, installation, freight, and othermiscellaneous costs were obtained from a report prepared bythe Alaska Energy Authority (AEA) [4]. Due to the remotenessof the site, the cost for transporting the various components isrelatively high.

Table II shows the results for the three cases. In this model,the roundtrip efficiency of the rectifier/inverter and the internalloss in the battery bank per cycle was considered as 90%. Thecollector efficiency for the PV array is assumed as 12%. As men-tioned in HOMER, the heating value of fuel is assumed to be48.5 MJ/kg, and the density of fuel is assumed to be 840 .The post-simulation analysis includes an economic and environ-mental component illustrating the simple payback and avoidedcost of emission reductions using the PV array.

A. Economic Analysis

The economic analysis part of the simulation model involvescalculation of the simple payback time (SPBT) for the PVmodule and calculation of energy payback time (EPBT) forthe PV array. In most of the remote villages, battery banks areused as back-up sources for power. Therefore, the PV withdiesel-battery system is compared to the diesel-battery systemin the analysis of SPBT. The SPBT is given as

SPBTExcess Cost of PV system

Rate of saving(9)

Using data from Table II

SPBT years

In order to calculate the EPBT for the PV array, it is essen-tial to know the energy required in the construction of the PVarray (also called embodied energy). In [9], Knapp et al. de-scribe a method to calculate the embodied energy of a PV array.In this method, the total energy required is the sum of energies


TABLE ICOMPONENT AND INSTALLATION COSTS FOR LIME VILLAGE

TABLE IICOMPARISON OF RESULTS FOR LIME VILLAGE

required for raw materials and the energy required in the var-ious processes involved to convert the raw materials into the PVarray. The embodied energy of a PV system is given by

kW (10)and

EPBTkW

(11)

where kWh is the embodied energy, kW is the rated power ofthe PV array, and is the energy generation rate of the PV array.

For Lime Village, the PV array is rated to produce 12 kW, andfrom Table II, the value for is 9445 kWh/yr.

kW kWhand

EPBTkWh

kWhyear

years

B. Environmental Analysis

The environmental analysis part of the model involves thecalculation of the avoided costs for , PM, and . Thefigures for PM and obtained in Table II are based on thevalues obtained from the manufacturer. The emission of3.1 kg /kg fuel is based on the mass balance for the com-bustion of the fuel. In [10], Narula et al. describe a way of cal-culating the avoided costs for . One way of reducing thegreenhouse gas emissions from electric power plants is by re-moving the gases through the use of chemical or other processes.Some DEGs have pollution control equipment to reduce emis-sions. DEGs in most Alaskan villages are not currently requiredto have emissions monitored. The cost associated with the re-moval processes is called removal cost (RC) and is described in[10].

The use of a PV array with the DEGs in Lime Village resultsin decreased emissions. The cost associated with the difference


in the amount of emitted pollution is called the avoided cost.The avoided cost is calculated as

ACCOE COE

(12)

whereAC avoided cost ($/ton);COE cost of electricity from low emission plant;COE cost of electricity from high emission plant;

emission from high emission plant (ton);emission from low emission plant (ton).

To calculate avoided cost, it is essential to calculate the cost ofelectricity for each system. Therefore, it is necessary to know the

ratio for a system, where is the annual payment on a loanwhose principal is at an interest rate for a given period ( ).Details of these calculations are described in [11]. The followingassumptions are used for the analysis presented here.

1) Interest rate .2) Life-cycle period for PV years.3) Life-cycle period for diesel battery system years.4) Life-cycle period for diesel battery system when

operating in conjunction with PV years.The higher life cycle period for the diesel-battery system

when operating in conjunction with the PV array is assumedbecause in the PV with diesel-battery system, about 10% of theload is supplied by the PV array. So, the life of the diesel-batterysystem will increase. The formula for is given as

(13)

for PV array

Similarly, for other cases is calculated and tabulated inTable III.

The annual cost of electricity for different systems is calcu-lated as

COE (14)

and

COE (15)

where is the cost of the PV with diesel-battery system,is the cost of diesel-battery system, and is the annual

cost of fuel.Substituting the values from Table II

COE

COE

Similarly, COE is calculated as $61 735.Using (12), the avoided costs for different emissions of

Table II are calculated and are listed in Table IV.The first cost is in the range of estimates provided by the

Intergovernmental Panel on Climate Change (IPCC) [12], whichhas estimated the cost for capture at power stations to be inthe range of $30–$50 per ton of avoided . The CaliforniaAir Resources Board (CARB) [13] estimated a cost of about

TABLE IIIA=P AND COE FOR VARIOUS CASES

TABLE IVAVOIDED COST OF EMISSIONS

$25 per pound of PM avoided by retrofitting buses with dieselparticle filters (DPFs). CARB [14] also reported $23 and $13per pound for and , respectively, as averages paidfor emissions offsets transactions in 35 California districts.

C. Comparison of Simulink Model With HOMER

Table V shows the comparison of the results of Lime Villageusing the Simulink model with those predicted by HOMER.It was observed that the efficiency of the diesel generatoris higher using the Simulink model. This is because in theSimulink model, the battery bank has a longer charge/dischargecycle. So, whenever the diesel generator is “on,” it operatesat a higher load and, hence, more efficiently. This is achievedby sacrificing the life of the battery bank. So, the life of thebattery bank is less in the Simulink model as compared to thatof HOMER.

In HOMER, the energy generated by the diesel engine ishigher because the battery bank is designed to cycle between40% and 82% of its kilowatthour rating rather than between 20%and 95% in the Simulink model. The inverter and rectifier areoperating with much less efficiency in HOMER as comparedto the Simulink model (about 20% difference). In HOMER, theDEG is loaded anywhere between 6.3–21 kW, with the averageload of 13.4 kW, and, hence, operates with a lower electrical ef-ficiency than in the Simulink model. In the Simulink model, thebattery bank acts as a source of power. So. whenever the DEGis “on,” it operates at 95% of its rated power—therefore, with ahigher electrical efficiency. If the load on the DEG is less than95% of its rated power, the excess power is utilized to charge thebattery bank. It can also be observed that the efficiencies for theDEG-battery and PV-DEG-battery are the same in the Simulinkanalysis. In the above analysis, HOMER has the advantage witha higher net present value (NPV) due to the longer life of the bat-tery bank over the Simulink model. The battery bank is the mostexpensive component of the system.

D. Overall Results

After performing the simulations for the three cases, it wasobserved that case 3 provided superior results in terms of fuelconsumption for the diesel generator and the greenhouse emis-sions. It was observed that the diesel generator operates mostefficiently for case 3, while the diesel-battery system in case 2has the highest kilowatthours per gallon. In case 1, the entire


TABLE VCOMPARISON OF RESULTS FOR LIME VILLAGE WITH HOMER

load was supplied without the PV array and the battery bank,leaving the load to be supplied by the diesel generator. Since thediesel generator operates with the lowest load for the diesel-onlysystem, it is the least efficient system and has the lowest kilo-watthours per gallon. In case 2, when the battery bank is dis-charged, the diesel generator is used to charge the battery bank,so eventually, the entire load is supplied with the help of thediesel generator. In case 3, part of the load was supplied usingthe PV array. As a result, there is substantial saving in the fuelconsumption by the diesel generator due to use of the batterybank and the PV array with the diesel-only system. The LCCand air emissions results of our Simulink model were compa-rable to those predicted by HOMER.

VI. CONCLUSION

The preliminary results reported here demonstrate that theintegration of a PV array into a diesel-battery stand-alone hy-brid power system reduces the operating costs and the green-house gases and particulate matter emitted to the atmosphere.The hybrid system has been in reliable operation since July2001. A Simulink model of the PV with diesel-battery hybridpower system installed at Lime Village, AK, was developed inthis project. The Simulink model can be used to study the per-formance of any PV with diesel-battery hybrid power systemif the operating characteristics of the power system are known.With few modifications, the model can be extended to incor-porate other renewable energy sources. The incorporation ofadditional renewable sources of energy, such as wind turbinesin this system, could further reduce fuel consumption. The dy-namic performance and the control system strategy of the powersystem can also be incorporated into the model.

The model was validated by comparing the results for sup-plying an annual load profile with those predicted by HOMER.The LCC and air emissions results of our Simulink model werecomparable to those predicted by HOMER. Although there is asignificant capital investment to purchase a PV system for thisapplication, the PV system may have acceptable 20-yr LCCsfor many remote locations. Furthermore, over its life cycle, thePV hybrid power system will consume less fuel and emit less

, , and PM than the diesel-only system. If the externalcosts associated with these emissions are taken into account, thePV system discounted payback period will further decrease. Hy-brid energy systems, which result in more economical and ef-ficient generation of electrical energy, would not only enhancethe capability of automated and precision generation systems,but would also help to extend the life of nonrenewable energysources.

ACKNOWLEDGMENT

The authors would also like to thank E. Baumgartner of Mc-Grath Power and Light for providing the information and datafrom Lime Village, G. Hanson of Marathon Electric for pro-viding the design specifications for the diesel-electric generator,John Deere for providing the diesel engine specifications, andGNB Industrial Power for providing the specifications for thebattery bank.

REFERENCES

[1] W. S. Fyfe, M. A. Powell, B. R. Hart, and B. Ratanasthien, “A globalcrisis: Energy in the future,” Nonrenewable Resources, pp. 187–195,1993.

[2] F. P. Dawson and S. B. Dewan, “Remote diesel generator with photo-voltaic cogeneration,” in Proc. Solar Conf., Denver, CO, Sep. 1989, pp.269–274.

[3] (2003). U.S. Department of Energy, Energy Information Authority. [On-line]. Available: http://www.eia.doe.gov/emeu/sep/ak/frame.html

[4] Lime Village Power System Alternatives, Alaska Energy Authority, An-chorage, AK, 2001.

[5] Performance and Economic Analysis of the Addition of Wind Power tothe Diesel Electric Generating Plant at Wales, Alaska, National Renew-able Energy Laboratory, Golden, CO, 1997.

[6] “ Rural Alaska Energy Plan Initiatives Aimed at Improving Rural EnergyEfficiency and Reliability,” Alaska Energy Authority, Anchorage, AK,2002.

[7] Wales, Alaska High Penetration Wind-Diesel Hybrid Power System:Theory of Operation, National Renewable Energy Laboratory, Golden,CO, 2002.

[8] (2003). National Renewable Energy Laboratory, Golden, CO. [Online].Available: http://rredc.nrel.gov/solar

[9] PV Payback, 2000.[10] R. G. Narula, H. Wen, K. Himes, and B. Power, “Incremental cost of

CO reduction in power plants,” in Proc. ASME Turbo Expo., 2002, pp.1–7.

[11] Stand-Alone Photovoltaic Systems: A Handbook of RecommendedDesign Practices (Revised), Sandia National Labs, Albuquerque, NM,1995.

[12] Climate Change 2001: Working Group III: Mitigation: 3.8.4.4 Tech-nical CO Removal and Sequestration (2003, Nov.). [Online]. Avail-able: http://www.grida.no/climate/ipcc

[13] Staff Analysis of PM Emission Reductions and Cost-Effectiveness, Ap-pendix F, California Air Resources Board 2002 (2003, Nov.). [Online].Available: http://www.arb.ca.gov/regact/bus02/appf.pdf

[14] G. Kats, “The Costs and Financial Benefits of Green Buildings,” Cali-fornia’s Sustainable Building Task Force, 2003.


Richard W. Wies (S’92–M’99) received the B.S., M.S., and Ph.D. degrees inelectrical engineering from the University of Wyoming, Laramie, in 1992, 1995,and 1999, respectively.

He was a Department of Energy AWU Graduate Fellow at the University ofWyoming Electric Motor Training and Testing Center during the 1993–1994academic year, a Department of Energy EPSCOR Research Fellow during the1995–1996 academic year, and an AWU Graduate Fellow at Pacific NorthwestNational Laboratories during the summer of 1996. He has been an AssistantProfessor of Electrical and Computer Engineering, University of Alaska Fair-banks, Fairbanks, since 1999. He teaches undergraduate and graduate coursesin electric power systems, electric machines, power electronics, digital signalprocessing, and controls. His current research involves the application of hybridelectric power systems for remote locations and the application of small-signalstability analysis techniques for power systems.

Dr. Wies is a member of the IEEE Power Engineering Society, Tau Beta Pi,and the Society of Automotive Engineers.

Ron A. Johnson received the B.Sc. degree from Brown University, Providence,RI, in 1965 and the M.S. degree in 1966 and the Ph.D. degree in 1969 fromCornell University, Ithaca, NY, all in aerospace engineering.

He was with the Avco Systems Division in Wilmington, MA, from 1969 untilhis arrival at the University of Alaska, Fairbanks (UAF) in 1976. He is currentlya Professor of Mechanical and Environmental Engineering at UAF. His currentresearch interests include indoor air quality and sustainable energy systems.

Ashish N. Agrawal (S’02) received the B.E. degree in from Pune University,Pune, India, in 1999 and the M.S. degree in from the University of Alaska Fair-banks, Fairbanks,in 2003, both in electrical engineering. He is currently workingtoward the Ph.D. degree in electrical engineering at the University of AlaskaFairbanks.

His current thesis research involves the development of a hybrid power systemmodels for cold climate applications. His work also includes contributions toThe Power Electronics Handbook (Boca Raton, FL: CRC Press, 2002).

Tyler J. Chubb (S’03) received the B.S. degree in electrical engineering fromthe University of Wyoming, Laramie, in 2001 and the M.S. degree in electricalengineering from the University of Alaska Fairbanks, Fairbanks, in 2004.

During his studies for the B.S. degree, he completed internships at PacificNorthwest National Laboratories in Richland, WA. He also was an Engineerat the Golden Valley Electric Association in Fairbanks, Alaska. He is currentlywith the Alaska Village Electric Cooperative, Anchorage.

43 simulink model for economic analysis and

Documents