A Review of Solar Photovoltaic Levelized Cost of Electricity

Renewable and Sustainable Energy Reviews 15 (2011) 4470 4482

Contents lists available at SciVerse ScienceDirect

Renewable and Sustainable Energy Reviews

jo ur n al hom ep a ge: www.elsev ier .com/ locate / rser

A review of solar photovoltaic levelized cost of electricity

K. Brankera, M.J.M. Pathaka, J.M. Pearcea,b,

a Department of Mechanical and Materials Engineering, Queens University, Kingston, Canadab Department of Materials Science & Engineering and Department of Electrical & Computer Engineering, Michigan Technological University, Houghton, MI, USA

a r t i c l

Article history:Received 29 MAccepted 5 JulAvailable onlin

Keywords:PhotovoltaicLevelized costLCOEGrid paritySolar economi

advantageous source of electricity over expanding geographical regions. 2011 Elsevier Ltd. All rights reserved.

Contents

1. Introd2. Revie

2.1. 3. LCOE 4. Addre

4.1. 4.2. 4.3. 4.4. 4.5.

5. Nume6. Discu7. Concl

AcknoRefer

Correspon601 M&M Buil

E-mail add

1364-0321/$ doi:10.1016/j.uction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4471w of the cost of electricity and LCOE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4471Estimates for solar PV LCOE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4472

methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4472ssing major misconceptions and assumptions in LCOE for solar PV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4475Discount rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4475System costs, nancing and incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4475System life for solar PV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4475Degradation rate and energy output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4476Grid parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4477

rical example in Ontario, Canada. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4477ssion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4478usions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4479wledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4480

ences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4480

ding author at: Department of Materials Science & Engineering, Department of Electrical & Computer Engineering, Michigan Technological University,ding, 1400 Townsend Drive, Houghton, MI 49931-1295, United States. Tel.: +1 906 487 1466.ress: [email protected] (J.M. Pearce).

see front matter 2011 Elsevier Ltd. All rights reserved.rser.2011.07.104 e i n f o

arch 2011y 2011e 15 September 2011

cs

a b s t r a c t

As the solar photovoltaic (PV) matures, the economic feasibility of PV projects is increasingly being eval-uated using the levelized cost of electricity (LCOE) generation in order to be compared to other electricitygeneration technologies. Unfortunately, there is lack of clarity of reporting assumptions, justications anddegree of completeness in LCOE calculations, which produces widely varying and contradictory results.This paper reviews the methodology of properly calculating the LCOE for solar PV, correcting the mis-conceptions made in the assumptions found throughout the literature. Then a template is provided forbetter reporting of LCOE results for PV needed to inuence policy mandates or make invest decisions.A numerical example is provided with variable ranges to test sensitivity, allowing for conclusions to bedrawn on the most important variables. Grid parity is considered when the LCOE of solar PV is comparablewith grid electrical prices of conventional technologies and is the industry target for cost-effectiveness.Given the state of the art in the technology and favourable nancing terms it is clear that PV has alreadyobtained grid parity in specic locations and as installed costs continue to decline, grid electricity pricescontinue to escalate, and industry experience increases, PV will become an increasingly economically

K. Branker et al. / Renewable and Sustainable Energy Reviews 15 (2011) 4470 4482 4471

1. Introduction

It is technically feasible for renewable energy technologies(RETs) to replace the present fossil fuel electricity infrastructure[1,2]; however, economic barriers remain the primary impedimentto a renewable-powered society. Solar photovoltaic (PV) technol-ogy, which converts sunlight directly into electricity, is one of thefastest growing RETs in the world [3,4]. PV is considered a clean,sustainable, renewable energy conversion technology that can helpmeet the energy demands of the worlds growing population, whilereducing the adverse anthropogenic impacts of fossil fuel use [57].From 2000 to 2010, global solar PV deployment has increasedfrom 0.26 G 1

than 40% [3reduced maincentives f

Despite tainable forsupply contion is cons[1721] giving while Plifetime genrable with grid [13,15average fortricity pricevalidity depused to calcity. In additof retail prfeasibility ovarious me(LCOE) genity generatitechnologienately, the is a lack ofing understwhich prodconcept of ship betwewhich depe[11,13,17,1assumptionestimated feliminatingcustomer ch[39]. Reporin not only misguide pocase for exatechnologythe long terof solar PV hachieving g

1 Units used1 MW = 1000 kWp, Watts pethe manufactukilowatt-hourthe system). AThe capacity fover a period o100% of the tim

exist for LCOE, this paper reviews the methodology of calculatingthe LCOE for solar PV, correcting the misconceptions made in theassumptions and provides a template for better reporting neededto inuence the correct policy mandates. A simple numerical exam-ple is proviconclusions

2. Review

A clear feasibility omining eneelectricity p

by tres [city aariece th

or tits fo

grid sal elet of

LCOnchmrent

mad con

priclude

diffe) takeaid fake a. Rathis, wed awn ifognivity tl yeaer poas poand oks atrue (electhly tood

shoume f

theave

of e LCOeraleratexity,51]. cessaeld

use the nd dissionies (nW to 16.1 GW [8] with an annual growth rate of more,911], due to both technological innovations that havenufacturing costs by 100 times and various governmentor consumers and producers [3,4,1115].increased incentives and the demand for more sus-ms of energy, PV has still not become a major energytributor [3,16]. The tipping point for solar PV adop-idered to be when the technology achieves grid parityen that conventional-powered electricity prices are ris-V installed prices are falling. Grid parity refers to theeration cost of the electricity from PV being compa-the electricity prices for conventional sources on the,1720,2224] often graphically given as the industry

solar PV electricity generation against the average elec- for a given country. While this is a useful benchmark, itsends on the completeness and accuracy of the methodulate the lifetime generation cost of solar PV electric-ion, claims of grid parity at manufacturing cost insteadice have contributed to confusion [15]. The economicf an energy generation project can be evaluated usingtrics [15,2528], but the levelized cost of electricityeration is most often used when comparing electric-on technologies or considering grid parity for emergings such as PV [9,11,13,15,17,19,22,24,2832]. Unfortu-LCOE method is deceptively straightforward and there

clarity of reporting assumptions, justications show-anding of the assumptions and degree of completeness,uces widely varying results [3,10,15,25,30,3238]. Thegrid parity for solar PV represents a complex relation-en local prices of electricity and solar PV system pricends on size and supplier, and geographical attributes9,21]. Different levels of cost inclusion and sweepings across different technologies result in different costsor even the same location. In addition, the trend of

avoidable costs for consumers and folding them intoarges can mask real costs of conventional technologies

ting the wrong LCOE values for technologies can resultsub-optimal decisions for a specic project, but can alsolicy initiatives at the local and global scale. In the solarmple, it is still a common misconception that solar PV

has a short life and is therefore extremely expensive inm [20,21,40,41]. Yet, depending on the location, the costas already dropped below that of conventional sourcesrid parity [3,18,2022,42,43]. Since varying estimates

in solar PV industry: W, Watt (measure of power); 1 kW = 1000 W,W; 1 GW = 1000 MW, used in capacity rating of energy technologies.ak (measure of nominal or rated power of solar PV system as perrer); kWh, kilowatt-hour (measure of electrical energy); kWh/kW/yr,s per kilowatt per year (annual energy produced per rated power of

solar insolation value with these units accounts for capacity factor.actor (CF) is the ratio of actual power output to nameplate capacityf time since power systems do not generate at maximum efciency,e.

eratedmeasuelectripliers vTo redua xedaccounto the the nthe cos

Theas a beof diffetion ismethomate anot incfor theiff (FITto be psarily thurdleanalysproducbe dra

Recsensitiseveraconsidrately Carlo) that lowhile index a monunders[31]. Itthe sanancetems hqualityby theare genity gencompl[18,19not nein the agentstion ofplant acommsubsidded with variable ranges to test sensitivity, allowing for to be drawn on the most important variables.

of the cost of electricity and LCOE

understanding of the relative cost-effectiveness andf different energy technologies is paramount in deter-rgy management policies for any nation. The actualrices depend on the marginal cost of electricity gen-

he given power plant and market-based or regulatory26,44,45]. Various power plants can compete to supplyt different bids, such that the electricity price from sup-s depending on the accepted bid and technology [26,46].is volatility, calculations are used by retailers to assumeered system that is predictable for consumers and thatr any volatility in the supplied electricity price, upgradesystem and other administrative duties [26,39,44]. Thusctricity price paid by consumers will be different from

generation [19,47].E methodology is an abstraction from reality and is usedarking or ranking tool to assess the cost-effectiveness

energy generation technologies [19,27,32]. The abstrac-de to remove biases between the technologies. Thesiders the lifetime generated energy and costs to esti-e per unit energy generated. The method usually does

risks and different actual nancing methods availablerent technologies [26,32,48]. For example, a feed in tar-s away the price risk for RETs by guaranteeing the priceor energy generated by the source, but does not neces-way the nancing risk for the technology, which is still aer all technologies should be given the same economic

ith the only difference being the actual costs, energynd lifetime [27]. Conceptual parallels with reality can

the scenarios closest to reality are chosen.zing that LCOE is a benchmarking tool, there is higho the assumptions made, especially when extrapolatedrs into the future [27,30,32,41,49,50]. Thus, if used tolicy initiatives, assumptions should be made as accu-ssible, with respective sensitivity analysis (e.g. Montejustications [30]. Ordinarily, LCOE is a static measuret a snapshot in deriving the price per generated energy,markets prices are dynamic. The SolarBuzz solar pricetricity, system and module prices) attempts to reportdynamic LCOE, although the assumptions should be

and it represents an average for specic circumstancesld be stressed that the type of nancing is usually keptor all technologies, even though real markets wouldm differently. In addition, economic and nancial sys-a large impact on the price of electricity, although thelectricity rarely changes, which is often not reectedE. Finally, the technological assumptions often usedized for the given equipment setup. Costs and electric-ed can vary based on location, capacity for generation,, efciency, operation, plant lifetime and other factorsThe efciencies and lifetime are taken as given, but dorily reect the actual specications and performance

. The usual criticisms of the misuse of the LCOE is thatoutdated data, do not consider the real plant utiliza-technology, do not capture the correct lifetime of theo not account for the full costs of the plant, such as de-ing, carbon and other environmental costs, insuranceuclear) and fuel subsidies (fossil) [32,52,53].

4472 K. Branker et al. / Renewable and Sustainable Energy Reviews 15 (2011) 4470 4482

Improvements to the LCOE for solar PV can be made oncerealistic assumptions and justications are given, real nancingvariability is considered, and consideration is made for techno-logical and geographical variability. Understanding the true costs,energy production and system specications would improve thecapabilities of LCOE software like the Solar Advisor Model (SAM).2

2.1. Estima

In generhigh comp[3,4,10,11,1these studidropped drOne of the California Elumps solarinclude a raforward looinvestor-owconsistent limitations.strated thatsources in plants. Anoreport, whiall technolothin-lm anof incentive

The Onta(through a as the pricgenerator ttion, operaover the plsion costs [6report by Gimethod incthat the meysis. It shoupart of the ignores biowith differe

Table 1 America sinspecicatioordered froshowing thwith 30 yeathe LCOE renot fully coreporting ofor the relatfrom elsewbeing reprethe assump

3. LCOE m

In this pa correct mwhere few

2 https://ww

considering energy management strategies [37,63]. Calculating theLCOE requires considering the cost of the energy generating systemand the energy generated over its lifetime to provide a cost in $/kWh(or $/MWh or cents/kWh) [27,30,32,34,49]. Many have noted thatLCOE methothat it is cuaccount for

s expomport e

sumted s27,32tion f the

outpt or .

LCOE

(1 +

nginer y

, theentd, op

or sown iancinniti

it mnnects ackives w

=

hat wan ar

in a lied

with by msolaW/y

the nr thef kW

majt of

depng coves mlar Pimathe m,67], st cotes for solar PV LCOE

al, estimates for LCOE for solar PV tend to be fairlyared to alternatives based on common assumptions4,15,18,19,25,3234,36,37,41,49,5462]. Note thates are all highly time dependent as the cost of PV hasamatically in the last several years [9,11,18,19,49].most clear recent LCOE reports was completed by thenergy Commission in 2010 [14]. Although the report

PV technologies with a life of only 20 years, its meritsnge of cost estimates, projections for variables allowingking values, a range of project types (Merchant; IOU,ned utilities; POU, publicly owned utilities) and aset of assumptions with detailed justications and

It should be noted that this report has already demon- solar PV can be less expensive than traditional energyCalifornia when considering peak power natural gasther recent reliable report is Lazards LCOE consultingch lists all the key assumptions made in the analysis ofgies (PV is split into the two dominant technologies,d crystalline silicon), considering price ranges, effects and effect of carbon emission costs [58].rio Power Authority (OPA) in Canada considered LCOEmethod called levelized unit electricity cost or LUEC)e (escalating with ination) that would be paid to ahat equals the present value direct costs (construc-tion and decommissioning) for the energy generatedants lifetime and included connection and transmis-3]. Apart from having no estimates made for solar PV, abson et al. [53] outlined several deciencies in the LUECluding not fully capturing current and future costs sothod cannot be considered an all-inclusive cost anal-ld also be pointed out that the OPA LUEC analysis asIntegrated Power System Plan like many other LCOEsphysical, social and economic externalities associatednt supply mix options [53].summarizes several solar PV LCOE results in Northce 2004 for variables including technology, year, plantns, lifetime, loan and incentives, and location roughlym best to worst in terms of reporting and methodology,at solar PV gets a 2025 year lifespan in most studiesrs considered for projections. As can be seen in Table 1,sults vary by more than a factor of four and many dover assumptions. From the survey, it is clear that betterf LCOE assumptions and justication is required evenively few variables chosen. Some studies quote a valuehere without restating the major assumptions or casesented [3,4,33,36,37]. This paper attempts to improvetions used and the clarity of the LCOE methodology.

ethodology

aper, the LCOE of solar PV is reviewed and claried andethodology is demonstrated for a case study in Canada,LCOE calculations have been done for solar PV when

w.nrel.gov/analysis/sam/.

LCOE imore cand Sh

Thegenerafrom [calculaning oenergytial cosfrom 1

Tt=0

(

Rearravalue p

LCOE =

Finallyinvestmnancecosts fas shofor ntial deprices,and coprojecincent

LCOE =

Note tis just eratedmultipenergyminedsolar inkWh/kplyingper yeaunits o

Thethe cosis veryfacturiinitiatitial sothe esthave t[11,38tion codology is very sensitive to the input assumptions, suchstomary to perform a sensitivity analysis [30,32,65] to

any uncertainty. The general calculation method forressed by Eqs. (1)(3) [18,27,30,32,34,35,49,66] whilelicated expressions can be pursued in Darling et al. [30]t al. [27]. Table 2 summarizes the nomenclature.

of the present value of LCOE multiplied by the energyhould be equal to the present valued net costs (adapted,49]) in Eq. (1). It should be noted that the summationstarts from t = 0 to include the project cost at the begin-rst year that is not discounted and there is no systemut to be degraded. Other methods can include the ini-

down payment outside the summation, with t starting

t

r)t Et)

=T

t=0

Ct

(1 + r)t(1)

g, the LCOE can be found explicitly assuming a constantear in Eq. (2).Tt=0Ct/(1 + r)

t

Tt=0Et/(1 + r)

t(2)

net costs will include cash outows like the initial (via equity or debt nancing), interest payments if debteration and maintenance costs (note: there are no fuel

lar PV) and cash inow such as government incentivesn Eq. (3). As such, the net cost term can be modiedg, taxation and incentives as an extension of the ini-on [30,65]. If LCOE is to be used to compare to gridust include all costs required (including transmissiontion fees if applicable) and must be dynamic with futurenowledged in the sensitivity analysis. In this paper, noill be considered.

Tt=0(It + Ot + Mt + Ft)/(1 + r)

t

Tt=0Et/(1 + r)

t

Tt=0(It + Ot + Mt + Ft)/(1 + r)

t

Tt=0St(1 d)

t/(1 + r)t(3)

hile it appears as if the energy is being discounted, itithmetic result of rearranging Eq. (1). The energy gen-given year (Et) is the rated energy output per year (St)by the degradation factor (1 d) which decreases the

time. The rated energy output per year can be deter-ultiplying the system size/capacity in kW by the local

tion that takes capacity factor into account in the units:ear1. Traditionally, this value is determined by multi-umber of days in the year by average number of hours

solar PV system operates by system size to get the nalh/year.or generation cost for solar PV is the upfront cost andnancing the initial investment, which means the LCOE

endent on the nancing methods available and manu-st reductions. Thus it has been argued that policy andust focus on this hurdle to make distributed residen-

V affordable [8,9,15,19,28,49,55,56]. When surveyinges as seen in Table 1, residential PV systems tend toore expensive LCOE due to lacking economies of scaledespite amortization facilities and lack of interconnec-mpared to utility scale PV [19]. The majority of this paper

K.

Branker et

al. /

Renew

able and

Sustainable Energy

Review

s 15 (2011) 4470 4482

4473

Table 1Summary of LCOE estimated from various sources in North America.

EstimatedLCOE ($/kWh)

Technology Year Plant specications Life Financing and incentives Location and solar resource Ref.

0.280.46 Solar PV (includingtracking 0.5%/year degr.)

2008 Residential ($7.5/W, CF 1433%) 30 No subsidies (30 year mortgage,100% nanced, 6% IR, 6% DR, 35%TR)

Various cities in USA(10002500 kWh/m2/year)

[11]

0.200.32 Solar PV (includingtracking 0.5%/year degr.)

2008 Residential ($7.5/W, CF 1433%) 30 With subsidies covering 30% initialcost (30 year mortgage, 100%nanced, 6% IR, 6% DR, 35% TR)

Various cities in USA (10002500kWh/m2/year)

[11]

0.150.80 Solar PV single axis 2009 25 MW (CF 27%, $4.55/Wp) 20 With and without tax benets, andother incentives (merchant, IOU,POU)

CA, USA [California EnergyCommission]

[14]

0.150.20 Solar PV-crystalline 2009 10 MW (CF 2027%, $5/Wp) 20 Lower price includes incentives USA [58]0.120.18 Solar PV-thin lm 2009 10 MW (CF 2023%, $4/Wp) 20 Lower price includes incentives USA [58]

0.16 (year 1) Solar PV 2010 Large scale ($3.00/W, CF 21%) 20/100 20 year, 6% IR, no incentives or tax USA Southwest [49]

0.3160.696 Solar PV January 2011 2 kW ($7.51/W) 20 5% cost of capital (tax andincentives excluded)

Global [used 5.5 sun-hours and 2.5sun-hours as high and low sites]

[64]

0.1690.372 Solar PV January 2011 500 kW ($3.98/W) 20 5% cost of capital (tax andincentives excluded)


[64]

0.3190.702 Solar PV December2010

2 kW ($7.61/W) 20 5% cost of capital (tax andincentives excluded)


[31]

0.1710.376 Solar PV December2010

500 kW ($4.07/W) 20 5% cost of capital (tax andincentives excluded)


[31]

0.15 Solar PV (1%/yeardegr.)

2011 4.5 kW residential ($5/W, 10 yearinverter life)

35 Not considered (SAM used) Phoenix, USA [19]


2011 150 kW commercial ($4/W, 15 yearinverter life)



2011 12 MW single axis at tilt ($3.9/W,15 year inverter life)



2011 12 MW two-axis conc. ($4.3/W, 15year inverter life)



2005 4 kW (residential) ($8.47/W) 30 SAM (low values if unnanced)effects of incentives, nancing andtax considered

Phoenix, USA [55]


2005 150 kW (commercial) ($6.29/W) 30 SAM (low values if unnanced) Phoenix, USA [55]

0.150.22 Solar PV (1%/yeardegr.)

2005 10 MW (utility scale) ($5.55/W) 30 SAM (low values if unnanced) Phoenix, USA [55]

0.30 Solar PV (no degr.) 2007 Residential ($8.5/Wp) 30 Home equity loan/mortgage, 90%debt, 6% IR, 28% TR, 30 year loanwith government incentives

USA (average maps with statevalues given) (SAM used)

[56]

0.062 Solar PV 2006 3.51 MW, Utility Scale Pv xed atplate ($5.40/Wp, CF 19.5%)

30 No nancing cost due to pay-as-goequity (IOU), includes tax credits

Springerville, Tucson, AZ, USA(1707 kWh/kW/year)

[59]

0.166 Solar PV 2003 5 MW ($4.16/W, CF 24%) 40 5% DR, no nancing USA [26]0.269 Solar PV 2003 5 MW ($4.16/W, CF 24%) 40 10% DR, no nancing USA [26]

0.248 Solar PV 2010 Roof top PV (projected) 25 Weighted average cost of capital(6.4%)

AZ, USA (1700 kWh/kWp) [18]

0.294 Solar PV 2008 Roof top PV ($5.2/W) 25 Weighted average cost of capital(6.4%)

AZ, USA (1700 kWh/kWp) [18]

4474K.

Branker et

al. /

Renew

able and

Sustainable Energy

Review

s 15 (2011) 4470 4482

Table 1 (Continued)

EstimatedLCOE ($/kWh)

Technology Year Plant specications Life Financing and incentives Location and solar resource Ref.


2009 Commerical ($6.7/W, CF 18%) 30 7% DR, no incentives (nancingunclear)

USA [38,10]


2009 Rooftop ($7.20/Wp, CF 17%) 25 5%10% DR, no incentives(nancing unclear)

AZ, USA [10]


2009 80 MW ($6.7/Wp, CF 19%) 30 5%10% DR, no incentives(nancing unclear)

AZ, USA [10]


2009 Rooftop ($7.20/Wp, CF 12%) 25 5%10% DR, no incentives(nancing unclear)

NJ, USA [10]

0.198 Concentrated solarPV (CSP)

2007 65 MW ($3.7/W, CF 22%) 30 7% DR, no subsidies (higher O&Mthan roof top) (nancing unclear)

NV, USA [10]

0.170.249 Concentrated solarPV (CSP)

2009 80 MW ($4.4/W, CF 29%) 30 5%10% DR, no incentives(nancing unclear)

USA [10]

0.1220.192 Concentrated solarPV (CSP)

2009 500 MW ($3.9/W, CF 23%) 30 5%10% DR, no incentives(nancing unclear)

USA [10]

0.250.40 Solar PV(12%/year degr.)

2003 Utility Scale PV or residential($6.209.50/W)

20 With and without subsidies, taxes,etc. (nancing uncertain)

CA, USA (2000:kWh/m2/year) [61] otherprojectionsmade

0.49 Solar PV 2010 1 kW (CF 20%, $8.73/Wp) 25 Residential amortization USA [15]

0.1380.206 Solar PV thin-lm 2009 Large scale 20 MW (CF 1827%,$3.74.0/W)

20? With and without incentives,nancing?

CA, USA [25]

0.1350.219 Solar PV crystallinesingle axis tracking

2009 Large scale 20 MW (CF 2328%,$7.047.15/W)

20? With and without incentives,nancing?

CA, USA*done for different project zones

[25]

0.456 Solar PV (xed atplate)

2008 20 MW ($7.98/W, CF 26%) 30? Weighted cost of captial after tax5.9%, 15 year accelerated Depr?

USA [41]

0.200.80 Solar PV 2007 Rooftop PV (25 kW) 20? No subsidies Worldwide range for25001000 kWh/m2 solarinsolation -quoted from range ofreports

[33]

0.200.50 Solar PV 2009 Rooftop (25 kW) ? No subsidies/incentives World average quoted fromrange of reports

[3]

0.150.40 Solar PV 2008 Different applications (?) ? Variable including taxes for USA (?) Different locations, USA (?) see [58] [4]

0.19 Solar PV 2007 Large scale 20 Independent power producernancing (no incentives)

Pacic north west, USA [60]

0.220.24 Solar PV 2007 Small scale 20 Independent power producernancing (no incentives)

Pacic north west, USA [60]

0.255 Solar PV (solar cell) 2008 5 MW ($5.782/W, CF 21%) ? No incentives, nancing for IPP USA [57]

0.200.50 Solar PV 2006 Varies at consumer level 20? No incentives Canada [36]

0.20, 0.31 Solar PV 2004 2003 prices ? DR 10% and 15% (Sandia Model,GenSim)

Chicago, USA [62]

0.3370.526 Solar PV-crystalline

2008 (2005price)

5 MW ($6.31%7.81/W, CF 1525%) 20 ? ? [34]

0.392 Solar PV 2008 5 MW ($7/W, CF 20%) ? ? Minera Escondida Limitada coppermine (off-grid) South America

[34]

0.25 Solar PV 2010 2006 prices, includes storage ? ? USA [54]

0.150.78 Solar PV 2003 ? ? ? Canada, taken from US studies andconverted to Canadian $

[37]

degr., degradation rate; CF, capacity factor; DR, discount rate; IR, interest rate; TR, tax rate; Depr, depreciation; IPP, independant power producer; IOU, investor-owned utilities; POU, publicly owned utilities; W, Wp assumed asmeaning the rated system power (units displayed as referred in the sources); SAM, Solar Advisor Model (NREL).


Table 2LCOE calculation nomenclature.

Nomenclature

T Life of the project [years]t Year tCt Net cost of project for t [$]Et Energy produced for t [$]It Initial investment/cost of the system including construction,

installation, etc. [$]Mt Maintenance costs for t [$]Ot Operation costs for t [$]FtrStd

will considepapers like

4. AddressLCOE for so

The maichoice of daverage systhe lifetime

4.1. Discou

Firstly, ttainty and tof discountvaries by ciFurthermortechnologieThe choicenologies whsector favoubut these msocial endetructure ansocial discosocial benecount rate (3.54.5% [6inal discoun[30].

4.2. System

In genercosts assocadministranancing ccosts associin all powemental andapart from cdent on the

3 Negative include carbonFor example, lutants from c[124].

dwelling. For example, in general, a thin-lm system is less costlyper unit power than a crystalline silicon system [68]. Inverters havevariable prices, types and lives and the type of racking and installa-tion needed depends on the house. Nonetheless, most LCOE studiesreport an avtechnologyneeds to beand how cogeneral, the[67], but st

pracng wes an

are rovidffer

costsdrastarizeor reUniteendixatiing cnt innanct paying tolar e cann be iatioge, aup tolar Plar Pnergircumevel

loanand Sthe a

conctee pymehe LCesenethoreasing tthe ethoratesInterest expenditures for t [$]Discount rate for t [%]Yearly rated energy output for t [kWh/year]Degradation rate [%]

r costs in the context of residential systems while other[30] cover utility scale.

ing major misconceptions and assumptions inlar PV

n assumptions made in the LCOE calculation are theiscount rate, average system price, nancing method,tem lifetime and degradation of energy generation over.

nt rate

he choice of discount rate comes with ample uncer-his is dealt with using sensitivity analysis. The concept

rate puts a value on time preference on money, whichrcumstance, location, and the time period considered.e, some investors vary their discount rate betweens to reect their perception of its nancial risks [26].

of discount rate can largely affect the energy tech-ich are relatively more competitive [49]. The privaters higher discount rates to maximize short-term prot,ay be too high to capture the benets of long-term

avours undertaken in the public sector, such as infras-d energy projects [49]. Governments often estimate aunt rate for rating public projects that have long-termt. For example, in Ontario, Canada, the real social dis-SDR) range used is 28%, with an individuals SDR being3]. Finally, there is a distinction between real and nom-t rate where ination is included in the nominal rate

costs, nancing and incentives

al, for the solar PV system costs, there are the projectiated with actual system, its design and installation;tive costs such as insurance and interconnection;osts associated with the nancing method and public

to bestdeclinifacilitirantiesmay pmay sulation not as summcosts fin the

Deptrys taFinancernmeDebt interesspreadif the sincomtem cadeprecmortgaloans (of a soogy, soother ein all csome dFIT.

TheSingh while wrongguaranated paallow tthe prloan mual incextendWhile loan mlation ated with taxes. However, what is not often consideredr generation technologies are the economic, environ-

health cost of negative externalities.3 The system price,apacity and manufacturing variability, is highly depen-

type of solar PV system and location and type of the

externalities for conventional electrical generation technologies dioxide emissions, thermal and air pollution and habitat disruption.

there are costs due to health problems associated with the air pol-oal-red generation [122,123] and for global climate destabilization

certain nashould be nessarily whconstructs mandate.

4.3. System

The nabe the man[75,76]. Hoels is well erage for solar PV, not distinguishing between different types and balance of system (BOS) costs. If averaging

made for simplication, then the assumptions mademmon they are should be reported (such as in [67]). In

BOS and labour costs represent 50% of the system costrategies are being developed to halve these comparedtice [69]. Solar manufacturing prices have been rapidlyith economies of scale through turn-key manufacturingd industrial symbiosis [68,70,71]. Inverter life and war-being extended to 10 years [11,72] and micro-inverterse an economical choice for residential systems, whichfrom partial shading challenges [73,74]. Finally, instal-

will decrease with technological experience, althoughically [15]. Recent estimated installed system costs ared in Table 3. It should be noted that average installedsidential systems are lower in Germany and Japan thand States [67].ng on an individuals credit history and the coun-on system, different nancing methods can be used.an come in the form of loans, a second mortgage, gov-centives, third party nancing and equity nancing.ing (loans or mortgages) is usually preferable sincements are non-taxable in some systems and it allows

he cost of the system over a longer period. Furthermore,PV system is recognized by a feed-in-tariff program, the

be recognized as business activities for which the sys-used against taxes via the capital cost allowance in assetn [67]. Finally, although many are adverse to a secondmortization allows for a longer loan term than usual

40 years). This is important given the long working lifeV system (greater than 20 years). As a proven technol-V should be able to obtain similar nancing methods asy technologies, although this is not necessarily the casestances as was recently shown in the difculties for

opers to nd nancing for projects under the Ontario

method effect on LCOE was recently considered byingh [28]. They indicated that the LCOE value is static,ctual cost of electricity increases, which results in thelusions for grid parity. Further, the loan period is for theeriod and not the working life of the PV system. A gradu-nt instead of an equated payment loan was suggested toOE of the solar PV to escalate like grid electricity. Thus,

t day LCOE would be lower than with the traditionald, increasing as the standard of living of the individ-ed. The new loan method was suggested since simplyhe loan term did not reduce the LCOE signicantly [28].analysis was not done for a specic system, the newd was done for different terms, interest rates and esca-, illustrating that grid parity could occur today underncial circumstances with the new method. Finally, itoted that what is mathematically feasible in not nec-at is socially feasible based on the current economicof society and such an approach would require a policy

life for solar PV

nceable life for a solar PV system is usually considered toufacturers guarantee period which is often 2025 yearswever, research has shown that the life of solar PV pan-beyond 25 years; even for the older technologies, and


Table 3Summary of recent solar PV installed system costs.

Solar PV technology Installed cost [$/Wp] Project scale

Crystalline ( UtilityCrystalline (Crystalline (Thin-Film CdThin-Film a-Crystalline aCrystalline aCrystalline aCrystalline aCrystalline aCrystalline a

a Estimate bb Average ofc Average of

Table 4Effect of degra

Degradation

0.2% 0.5% 0.6% 0.7% 0.8% 1.0%

current oneyear lifetim[28] expliciworking lifeguarantee pthe loan terwould still was plottedloan term, awould be susidered [49which it cothat the opethe asset. Sicosts rise, this considerelife, the asseexpensive tare due to rcleaning ancosts that wthe life of mrated since nitely theconsidered able energywhich thereof degradatthe system(Pmax).

Finally, tfor differen(c-Si) PV mestablishedprovided th20 years oriimmature tels are belo

.3% f the lbut mtemhe ual prd th

es th gain

grad

ermis onn pron wire ofation

[49] PV ster

[77]es ha

year.Europe)a 5.00 China)a 4.42 Japan)a 5.02 S/CdTea 4.28 Si/-Sia 3.52 nd thin lm (USA)b 7.50 nd thin lm (Germany)c 7.70 nd thin lm (Japan)c 4.70 nd thin lm (USA)c 5.90 nd thin lm (CA,USA)b 7.30 nd thin lm (CA,USA)b 6.10

ased on module prices [68]. installed systems [67]. installed systems excluding sales taxes [67].

dation rate and performance requirement on system life.

rate Lifetime to 80%Pmax [years]

Lifetime to 50%Pmax [years]

100 25040 10033 8329 7125 6320 50

s are likely to improve lifetime further [7781]. A 30e or more is becoming expected [82]. Singh and Singhtly called for scientists to give an authentic gure on the

of solar PV systems to improve condence for the loaneriod [28]. An important consideration is that even ifm was shorter, the energy output from the PV panelscontinue at a negligible cost. If the LCOE for each year

over time, with different equations before and after thedjusting for the annualized loan cost, the yearly LCOEbstantially less after the loan term than currently con-]. In general, the working life of an asset is the life forntinues to perform its tasks effectively. It is often trueration and maintenance (O&M) costs rise with the age ofnce annual capital costs tend to decline and annual O&M

was 17deneevent, the sysises tmaterision, anlifetimedge is

4.4. De

Detdependsulatioand catems adegradrantiessiliconthat fainitelymodulthe 1%Table 5ere is a minimum average cost per year at which point itd the economic life of the asset [83,84]. At the economict is then replaced or refurbished, since it becomes moreo run the asset thereafter. For solar PV, the O&M costseplacing inverters (usually every 10 years), occasionald electrical system repairs [49,85], which are relativeill decrease with time. It should also be noted thatany conventional power plants is much longer than

they tend to be refurbished or re-commissioned inde- same could be true of solar PV plants [49]. Thus, what isthe economic life of the system depends on the accept-

output, which depends on the degradation rate (rate at is a reduction in output). Table 4 illustrates the effection rate and acceptable performance on the lifetime of

in terms of a percentage of maximum power output

he lifetime and reliability of solar PV can be consideredt solar PV technologies. Crystalline silicon wafer basedodules offer the best in-eld data being the technology

on the market for the longest time. Skoczek et al. [77]e results for c-Si PV modules in the eld for more thanginally characterized between 1982 and1986 (relativelyechnologies). In their ndings, more than 65.7% of pan-w the 1% per year degradation rate (mean power loss

In anothnology) andof 0.2%, alththe earlier dFurthermorat least 15 years [78]. actual in-ecluded thatconsidered

It shouldsilicon (a-S[87]. In a-Siof exposurereached [88Si:H PV are value, ignorcompound gies, the outby conventogy, becausperformancbeen shownnometers cUtilityUtilityUtilityUtilityCapacity weighted average (2009)Residential (25 kW) (2009)Residential (25 kW) (2009)Residential (25 kW) (2009)Residential 10 kW (2010)>100 kW (2010)

or 21 years average). For PV technology, it is difcult toifetime since ordinarily there is no single catastrophicore gradual aging and degradation. The end of life of

has not been reached once the power output still sat-ser. Gradual degradation occurs due to chemical andocesses associated with weathering, oxidation, corro-ermal stresses [78,77]. Current research would improverough greater quality in production processes as knowl-ed about failure mechanisms [78,86].

ation rate and energy output

ning the energy output of solar PV over its lifetime assumed degradation rate of the panels. Module encap-tects against weather factors, moisture and oxidation

thstand mechanical loads (e.g. wind and hail). PV sys-ten nanced based on an assumed 0.51.0% per year

rate [65] although 1% per year is used based on war-. This rate is faster than some historical data given for[77,78,86]. In a study on c-Si modules, it was founddegradation occurs earlier and then it stabilizes indef-. In the study, more than 70% of 1923-year-old c-Sid an annual degradation rate of 0.75%, still less than

assumed [77]. The failure sources are summarized in

er study, c-Si PVs installed in 1982 (much older tech- tested in 2003 had an annual power degradation rateough this rate was faster in the latter 4 years [78]. Thus,egradation must have been slower than 0.2% per year.e, accelerated aging tests indicated that the panels hadyears more of acceptable performance beyond the 21Finally, another study indicated that the degradation ofld c-Si cells is 0.20.5% per year [86]. It can thus be con-

in general, a degradation rate of 0.20.5% per year is reasonable given technological advances.

be noted here that there is a special case for amorphous

i:H) PV, which suffers from light-induced degradation

PV cells, performance degrades rapidly in the rst 100 h to 1 sun illumination until a degraded steady state is90]. The effect has of yet not been eliminated, but a-

sold with warranties valued at the degraded steady stateing the above specied initial performance. To furtherthe appropriate calculations of such thin lm technolo-put of a-Si based solar cells is generally under-predictedional techniques developed on c-Si-based PV technol-e of the superior a-Si:H temperature coefcients ande in diffuse light conditions [9193]. In addition, it has

that the use of integrating photometers such as pyra-an directly introduce errors in the prediction of a-Si PV


Table 5Summary of power loss results for 204 modules installed in 19821986 with 1923 years [77].

Averagelosses (%)

Std dev (%) Reasons

Power loss inatio

Loss in VOC (terminal)

f subs

Loss in ISC (sdelivered)

le agintercon

Loss in Fill Fto maximu

le agintercondation

system outa-Si:H PV ddepending Si:H coversspectrum wa-Si:H PV wwider specthave been wson generalPV modulesmodules [9

4.5. Grid pa

As mentfor the costof solar PVsupplied elelectricity iLCOE is notfor the totathe realisticto back calcto be to atta

Yang [15would suggcost-effecticlaimed. Thwere not amwrongfully installed coCamstars Aper kWh of applied to trecycling at

5. Numeri

In Cana$0.17/kWh the LCOE fothe simplitions, the LCvariables toa realistic sare declininthe relativefor the base[11,78,86], uof 1.5% of tcompanies)the total ins

7,72gh loent Ktor oe ranin th

1Awith

is 1ere d 10

OE dend int fave LCOs $0.7. Noto haours

2, waryinnt raould red [e it irrant

3 sstemfromscouWh/rease

$2.2ptionpeteime, easet eco

4 ill on are prmin17.3 23.5 Comb

open circuit voltage potential across 10.6 18.5 Loss o

hort circuit current maximum current 5.8 20 Moducell in

actor, FF (ratio of maximum actual powerm theoretical power)

9.1 22 Moducell indegra

put, and over the course of a year, the output from anevice can vary by 1020% due to this spectral effect,on seasonal and locational effects [94102]. Because a-

only a small fraction of this range, differences in theill have an amplied effect on the performance of anhen compared to c-Si PV devices, which cover a muchral range because of its smaller bandgap. These effectsidely documented [94,95,97103] and is the nal rea-

ly attributed to the well-documented claims that a-Si:H will produce more energy per rated power than c-Si PV5].

rity

ioned before, grid parity is considered a tipping point effectiveness of solar PV, and entails reducing the cost

electricity to be competitive with conventional grid-ectricity. For parity, the total cost to consumers of PVs compared to retail grid electricity prices. Although the

the same as retail electrical prices, it is used as a proxyl price to be paid by consumers, adding in as many of

costs as possible. The LCOE methodology is then usedulate what the required system and nance costs needin grid parity.] determined that a realistic examination of grid parityest that solar PV is much further away from becomingve in distributed (residential) systems than is normallye main problem Yang identied was (1) many analystsortizing all of the cost to the end consumers and (2)

considering $1/Wp manufactured cost instead of retailst when calculating grid parity [15]. However, applyingdvanced Product Quality model suggests that the costthe solar industry can be shifted down by 1317% whenhe manufacturing supply chain from design to system

end of life [23].

cal example in Ontario, Canada

da, electricity prices range from $0.06/kWh toin major cities [51] so that as a proxy for grid parity,r residential solar would need to be in this range. Usinged method outlined in Section 3 and improved assump-

[4,11,6althourepresity faca viablstates

Fig.varies outputrates w4.5% anthe LCterm ais mosest, thLCOE ilationsyears tof cont

Fig.with vdiscourate wcompabecausPV wa

Fig.the syvaries real di1270 klife inccost ofassumto comof lifetto decrcurren

Fig.outputoutputto deteOE was calculated for Ontario, Canada using ranges of test sensitivity as an example. As shown in Table 5,tarting fully installed system price is $5/Wp1 as pricesg and thin-lm PV would show better performance inly cloudy region of Ontario [9698]. Other assumptions

example case include: a degradation rate of 0.5%/yearsing 100% debt nancing, an operating (insurance) cost

he total system cost (average quotes from 3 insurance and a maintenance (inverter replacement) cost of 9% oftalled system cost (ranged from 6 to 9% in US for 2009)

installed syand 1270 kWcost of


Fig. 1. LCOE idiscount ratesdegradation ra

6. Discussi

Table 1 gand inconsito be addregiven, the fon $/kWh for solar PV system varying interest rates, loan terms and (A, 0%; B, 4.5%; C, 10%) assuming initial installed system cost of $5/Wp,te of 0.5%/year and energy output of 1270 kWh/kW/year.

on

ives an example of the existing varying LCOE estimatesstency of reporting assumptions. Thus, the rst pointssed is the reporting of LCOE. With the value or rangellowing assumptions must be provided and justied:

Fig. 2. LCOE inassuming zeroenergy output

Fig. 3. LCOE fzero interest loutput of 1270

1. Solar PV 0.5%/yea

2. Scale, siz(resident

3. Indicatiographic l

4. Lifetime sarily eq

5. Financiacost of ca

6. Additiontion, carbshould b

A simple yeto calculate $/kWh for solar PV system varying system costs and discount rates interest loan, 30 year lifetime, degradation rate of 0.5%/year and

of 1270 kWh/kW/year.

or lifetime of solar PV system versus initial cost of the system for aoan, discount rate of 4.5%, degradation rate of 0.5%/year and energy

kWh/kW.

technology and degradation rate (e.g. c-Si or a-Si:H, andr degradation rate).e and cost of PV project [including cost breakdown]ial, commercial, utility scale/# kW, # MW, $/Wp).n of solar resource: capacity factor, solar insolation, geo-ocation, and shading losses.of project and term of nancing (these are not neces-ual).l terms: nancing (interest rate, term, equity/debt ratiopital), discount rate.al terms: ination, incentives, credits, taxes, deprecia-on credits, etc. (these need not be in the analysis, but ite stated whether or not these are included).

t correct methodology with clear assumptions was used the LCOE for solar PV in Ontario, Canada. The results


Fig. 4. LCOE floan, discount

as presentesidering theconstrainininclude Mospecic varcosts and ecircumstantechnologic

The highdle to adopFigs. 14, locount rate nancing istime of thePositive disin the near far term. Ifcoal-red pwhich has na positive dConsumptivfuel costs ththat requirnegligible cment, capitconcept is l

Fig. 3 decould resulfor residentprovide eitit was as ifgovernmenallowing goa greater scmeant to renance), whicontracts gucial instituthistories an

The miscdiscussed hteed and ne

facilities do not acknowledge this extended time, the LCOE shouldstill consider the working life for the operation and maintenancecosts and energy production [49]. In the case of degradation rate,

nctions may be needed in the LCOE calculation to recognizer som

have modes usally, a

feasstani andpply r pro imolar

be mid past in811orescn, b

beins oldn athey c

neecultntinuable ater ives 1]. Cer c

g BOd adoverationemerm

le. Focity ttabng re

be r, for or energy output versus initial cost of the system for a zero interest rate of 4.5%, degradation rate of 0.5%/year and 30 year lifetime.

d in Figs. 14 with contours give a useful way of con- LCOE for various systems and specications withoutg the assumptions. Other sensitivity techniques wouldnte Carlo Simulations [30]. Furthermore, the effect ofiables can easily be seen once the calculations take allnergy into consideration. For Canada, under specicces, solar PV LCOE grid parity is a reality once certainal, pricing and policy hurdles are addressed.

initial upfront cost of solar PV still seems to be a hur-tion, despite declining cost of systems. As shown inwer interest rates, longer term loans and higher dis-

are preferred in combination. The preference for debt due to the ability to spread out the cost over the life-

system, and is highly inuenced by the discount rate.count rates mean cash inows (benets) are preferredterm, whilst cash outows (costs) are preferred in the

comparing a consumptive technology like nuclear orlants to a capital intensive technology like solar PV,o fuel cost that is susceptible to price uctuation risk,iscount rate biases towards consumptive technologies.e technologies involve long installation times and highat would seem preferred over a capital intensive plant

step futhat fowouldfailurethe rat

Finnot becircumHawaithe suinvertetinue tsome sshoulding grto inve[15,10pact long rution ofmore aPV, eveual if tenergyis a difstill corenew

Greincent[23,11customportinreduceFITs. Gdardizmanaglong-tetainabelectrithe proroundishouldwouldes high upfront costs in a short installation time, butosts thereafter. In terms of sustainable energy manage-al intensive technologies should be preferable, but thisost in the current economic system.monstrated that a zero interest loan over a long periodt in the lowest LCOE values. Financing still is an issueial systems and incentives should be considered thather zero interest loans or offset interest costs so that

there was no interest. A zero interest loan from thet would work for distributed PV community programs,vernments to meet their renewable energy targets onale. The Ontario FIT Program is the opposite incentiveduce the effect of long-term costs (interest and mainte-le providing some economic return. Again, although FITarantee a price for the energy, as seen in Ontario, nan-ions still consider loans in terms of individuals creditd not the value of the contract.onception about system lifetime and degradation wasere. Solar PV lifetimes will often be greater than guaran-w industry norms will at least be 30 years. If nancing

revenue. Ficleaner andnologies toan income greater inctainability should asseincome claenergy targsupported athey could dtions with creation [12

7. Conclus

As the soof PV projecost of elecother electre systems, more energy is produced in earlier years and a higher weighting with a positive discount rate. Ases and degradation mechanisms are better understood,ed in LCOE can be systematically improved.lthough Yang [15] stated that some system costs wouldible for grid parity, the fact is that it is under certainces grid parity has already been reached in places like

California [107] and much can still be done to improvechain to reduce costs [23,67,70]. Solar module prices,ices, system and component lives and BOS costs con-prove as research and development evolves, puttingmanufacturers at grid parity today [23]. In addition, itentioned that cost effectiveness (or in this case obtain-rity) is not necessarily a sufcient driver for people

any new technology including residential PV systems4]. An example is the adoption of energy efcient com-ent light bulbs (CFLs) that are more economic in theut have a higher upfront cost giving them the percep-g expensive [110]. In Canada, CFLs are being adopteder technology (incandescent) are banned. The LCOE of

grid parity may be of little consequence to an individ-annot reap near term prots (savings) or the required

ds as the next best alternative. Concerning grid parity, it endeavour considering, fossil fuels and nuclear powere to receive larger indirect and direct subsidies than

energy technologies [115119].adoption of solar PV will be driven by governmentand policies and solar PV supply chain innovationonsumers would prefer innovative products, greaterare, increased reliability and quality of panels and sup-S, greater standardization in installation quality andministrative time for government incentives such asnments can monitor and create the policy for stan-

to improve quality and provide training and interfacent education [111]. Government policies need to haveobjectives and certainty so that incentives are sus-r example, encouraging third party sale of solar PVo the grid beyond the FIT at a retail price would increaseility of the system. Furthermore, if public policies sur-tail, insurance and nancing are aligned, then solar PVecognized for its added value, like a swimming poola residential dwelling except that PV would producenally, tax breaks (sale or income) can be considered for

renewable technologies over fossil fuel based tech- encourage their adoption. One study indicated thattax benet for purchase of the technology could haveentive than low interest loan [120]. To ensure sus-of solar PV adoption through incentives, governmentsss the impact of incentives on adoption for differentsses and determine which will be best to meet theirets. Finally, in the same way that governments havend invested in conventional power generation projects,o so for PV manufacturing to be able to reap cost reduc-economies of scale and other social benets like job1].

ions

lar photovoltaic (PV) matures, the economic feasibilitycts is increasingly being evaluated using the levelizedtricity (LCOE) generation in order to be compared toicity generation technologies. A review of methodology


and key assumptions of LCOE for solar PV was performed. The LCOEcalculations and assumptions were claried and a correct method-ology and reporting was demonstrated for a case study in Canada.It was found that lack of clarity in assumptions and justications insome LCOEicy initiativis a need fodistributionset of assumassumptionA higher inccations is resource. Givenancing teity in specigrid electricincreases, Pgeous sourc

Acknowled

The authural Scienchelpful disc

References

[1] Fthenakfeasibili2009;37

[2] Sovacoodesirabl

[3] Renewa2010 gl

[4] KirkegaGlobal iWorkin

[5] Pearce 2002;34

[6] Sims REparisonelectric

[7] Internatnarios aIEA/OEC

[8] Mints PNavigan19.01.20

[9] Internatenergy;

[10] Doty GNsustaina2010, vo

[11] Price S,Renewa

[12] Renewaglobal s

[13] Jogleka policy aof 2008

[14] Klein J. Cnal sta1186.

[15] Yang C. [16] Energy

USA: US[17] Denholm

photovoRenewa

[18] Breyer Cand US dence oProceed44925

[19] Hegeduphotovoscience

[20] Pernick R, Wilder C. Utility solar assessment (USA) study reaching ten percentsolar by 2025. Clean Edge, Inc./Co-op. America Foundation; 2008. p. 176(June).

[21] Song J, Boas R, Bolman C, Farber M, Flynn H, Meyers M, et al. True cost of solarpower: race to $1/W. Boston, MA: Photon Consulting LLC.; 2009.rown ons; 11360

amstanovat10. pke Reructurt ana

lack ahase 2e; 201rganisergy nerat230.ort Wcienboratngh Penewaheynelar Lg/chipek lowarlingergy larBu10 su

elosa April 20enewa07 gl

andyor BHPrana ttp://

emystationamarkealdenergy eptemiser Rsts ofvermoarce Jnservstomunt T.ectoIQutchin. Progtions

. 111uller Botoniaper Tanso

ationaarketunasigton, 395 (

atural09. g.php

ationansum

si/rnrgrchaseibevilgen

aramee 200ydro-Qties; Oavitt ttp://

ibson the O estimates could lead to the wrong outcomes and pol-es. Since the inputs for LCOE are highly variable, therer using sensitivity analysis to represent actual variables so that there is no unreasonable condence in a singleptions. This paper illustrated that the most importants were system costs, nancing, lifetime and loan term.lusivity of costs and reporting assumptions and justi-commended, even if merely using the work of anothern the state of the art in the technology and favourablerms it is clear that PV has already obtained grid par-c locations and as installed costs continue to decline,ity prices continue to escalate, and industry experienceV will become an increasingly economically advanta-e of electricity over expanding geographical regions.

gements

ors would like to acknowledge support from the Nat-es and Engineering Research Council of Canada andussions with B. Purchase and R. Andrews.

is V, Mason JE, Zweibel K. The technical, geographical, and economicty for solar energy to supply the energy needs of the US. Energy Policy:38799.l BK, Watts C. Going completely renewable: is it possible (let alonee?). Electric J 2009;22:95111.ble Energy Policy Network for the 21st century (REN21). Renewablesobal status report. Paris; 2010. p. 180.ard JF, Hanemann T, Weischer L, Miller M. Toward a sunny future?ntegration in the solar PV Industry, World Resources Institute (WRI)g Paper Series; May 2010. p. 166.JM. Photovoltaics a path to sustainable futures. Futures:6374.H, Rogner H, Gregory K. Carbon emission and mitigation cost com-s between fossil fuel, nuclear and renewable energy resources fority generation. Energy Policy 2003;31:131526.ional Energy Agency (IEA). Energy technology perspectives 2008: sce-nd strategies to 2050. Paris, France: International Energy Agency,D; 2008. p. 1650.. The 12-step solar program: towards an incentive-less future.t Consulting on ElectoIQ. [accessed11].ional Energy Agency (IEA). Technology roadmapsolar photovoltaic

October 2010. p. 148., McCree DL, Doty JM, Doty FD. Deployment prospects for proposedble energy alternatives in 2020. In: ASME Conference Proceedingsl. 171. 2010. p. 17182.

Margolis R. Solar technologies market report. Energy Efciency &ble Energy, US Department of Energy, 2010; January 2008. p. 1131.ble Energy Policy Network for the 21st Century (REN21). Renewablestatus report 2009 update, Paris; 2009. p. 132.NR, Graber-Lopez ES. A countdown towards solar power at grid parity:nalysis based on the evolution of price-performance. In: Proceedings

ISDSI international conference. 2008. p. 118.omparative costs of California Central Station electricity generation:ff report. Sacramento, CA: California Energy Commission; 2010. p.

Reconsidering solar grid parity. Energy Policy 2010;38:32703.Information Administration (EIA). International energy outlook 2010.

Department of Energy; 2010. p. 1338. P, Margolis RM, Ong S, Roberts B. Break-even cost for residential

ltaics in the United States: key drivers and sensitivities. Nationalble Energy Laboratory (NREL) technical report; 2009. p. 133., Gerlach A, Mller J, Behacker H, Milner A. Grid-parity analysis for EUregions and market segmentsdynamics of grid-parity and depen-n solar irradiance, local electricity prices and PV progress ratio. In:ings of 24th European photovoltaic solar energy conference. 2009. p.00.s S, Luque A. Achievements and challenges of solar electricity fromltaics. In: Luque A, Hegedus S, editors. Handbook of photovoltaicand engineering. 2nd ed. John Wiley and Sons Ltd.; 2011. p. 138.

[22] Bti90

[23] Cin20

[24] Pistke

[25] Bpte

[26] OEnge1

[27] ShefLa

[28] SiR

[29] Csoorse

[30] Den

[31] So20

[32] VA

[33] R20

[34] Bfo

[35] G.RB, Wineld M, Markvart T, Gaudreau K, Taylor J. An analysisntario Power Authoritys consideration of environmental sustain-


ability in electricity system planning. Studies in Ontario electricity policyseries paper no. 2. University of Waterloo and York University; 2008.p. 1173.

[54] Electric Power Research Institute (EPRI). The power to reduce CO2emissionsthe full portfolio. Palo Alto, CA: Electric Power Research Institute;2007. p

[55] U.S. Depyear pro

[56] U.S. Depyear pro

[57] Kaplan Washin

[58] Lazard Lazard L

[59] Moore Lphotovo

[60] Oregonelectric

[61] Simonsbenets

[62] Tolley Gsity of C

[63] OntariorenewaD & E; 2

[64] SolarBu.

S, Roman P, Selkirk C. Project managementan engineering eco-perspective. 2nd custom ed. Pearson Custom Publishing; 2007 (for

University).M, Jewkes EM, Bernhardt I, Tajima M. Engineering economics in

. 3rd ed. Toronto, Ontario: Pearson Hall; 2006.t Consulting. A review of PV inverter technology cost and perfor-rojections. Golden, CO: National Renewable Energy Laboratory; 2006.e D, Realini A, Cereghetti N, Rezzonico S, Bura E, Friesen G. Analy-eather c-Si PV modules. In: Proceedings of 3rd world conference onltaic solar energy conversion. 2003. p. 5.

DL, Wronski CR. Reversible conductivity changes in discharge-d amorphous Si. Appl Phys Lett 1977;31:292.

co[90] R

bclti2

[91] Janco

[92] Japfe

[93] Jasip

[94] Rm1

[95] HfaSo

[96] Gso2

[97] Gapp

[98] Gsiam

[99] Ru

[100] Jaefce

[101] Cce

[102] MamM

[103] AmRTI

[104] Frin01

[105] H.

S, McKenney DW, Poissant Y, Morris R, Lawrence KM, Campbell KL,e development of photovoltaic resource maps for Canada. In: Annualnce of the solar energy society of Canada. 2006.. Gaining on the grid, BP Globalreports and publications; 19 (2007)http://www.bp.com/sectiongenericarticle.do?categoryId=9019305&Id=7035199>.M, Denkenberger D, Zielonka H. Accelerating applied sustainability byg return on investment for energy conservation measures. Int J Energy

Econ 2009;17:6180.e AV, Duo E. The economic lives of the poor. USA: Bureau of Researchnomic Analysis; 2006, 143 (Working paper no. 135).a C, Ferriera S, Lazarova E. Shedding light on the light bulb puzzle:

of attitudes and perceptions in the adoption of energy efcient lightcott J Politic Econ 2010;57:4867.. Beyond technology-push and demand-pull: lessons from Califor-

lar policy. Energy Econ 2008;30:282954.Demirguc-Kunt A, Martinez PMS. Reaching out: access to and use of

services across countries. Washington: World Bank; 2005., Miller JB, Wang YD, Byrne JB. Energy-micronance intervention foroverty line households in India. Energy Policy 2009;37:1694712.n M. The micronance revolution: sustainable nance for the poor.gton: World Bank; 2001. p. 199215.. Investors renewables growth is slower but steady. RenewableEner-d.com. [accessed

10].g M. Federal energy subsidies: not all technologies are created equal.gton, DC: World Bank; 2000. p. 120 (Research report no. 11, Renew-ergy policy project).

. The federal energy subsidy scorecard: how renewables stack up.ble energy world.com.


rea/news/article/2009/11/the-federal-energy-subsidy-scorecard-how-renewables-stack-up> [accessed 3.10.2009].

[118] Pearce JM. Increasing PV velocity by reinvesting the nuclear energy insurancesubsidy into large scale solar photovoltaic production. In: 34th proceedingsof IEEE photovoltaic specialists conference. 2009. p. 133843.

[119] Zelenika-Zovko I, Pearce JM. Diverting indirect subsidies from the nuclearindustry to the photovoltaic industry: energy and economic returns. EnergyPolicy 2011;39:262632.

[120] Chandrasekar B, Kandpal TC. Effect of nancial and scal incentives on theeffective capital cost of solar energy technologies to the user. Solar Energy2005;78:14756.

[121] Branker K, Pearce JM. Financial return for government support of large-scale thin-lm solar photovoltaic manufacturing in Canada. Energy Policy2010;38:4291303.

[122] Boyd DR, Genuis SJ. The environmental burden of disease in Canada: respira-tory disease, cardiovascular disease, cancer, and congenital afiction. EnvironRes 2008;106:2409.

[123] Canadian Medical Association (CMA). No breathing roomnational illnesscosts of air pollution; August 2008. p. 145.

[124] Stern N. The economics of climate change: the Stern review. Cambridge, UK:Cambridge University Press; 2007.

A review of solar photovoltaic levelized cost of electricity1 Introduction2 Review of the cost of electricity and LCOE2.1 Estimates for solar PV LCOE

3 LCOE methodology4 Addressing major misconceptions and assumptions in LCOE for solar PV4.1 Discount rate4.2 System costs, financing and incentives4.3 System life for solar PV4.4 Degradation rate and energy output4.5 Grid parity

5 Numerical example in Ontario, Canada6 Discussion7 ConclusionsAcknowledgementsReferences

Documents

A Review of Solar Photovoltaic Levelized Cost of Electricity