Renewable power generation costs in 2017

Transcript

1 Renewable Power Generation Costs in 2017

2 Copyright © IRENA 2018 Unless otherwise stated, this publication and material herein are the property of the International Renewable Energy Agency (IRENA) and are subject to copyright by IRENA. Material in this publication may be freely used, shared, copied, reproduced, printed and/or stored, provided that all such material is clearly attributed to IRENA and bears a notation of copyright (© IRENA) with the year of copyright. Material contained in this publication attributed to third parties may be subject to third-party copyright and separate terms of use and restrictions, including restrictions in relation to any commercial use. ISBN 978-92-9260-040-2 IRENA (2018), Renewable Power Generation Costs in 2017 Citation: , International Renewable Energy Agency, Abu Dhabi. ABOUT IRENA The International Renewable Energy Agency (IRENA) is an intergovernmental organisation that supports countries in their transition to a sustainable energy future, and serves as the principal platform for international co-operation, a centre of excellence, and a repository of policy, technology, resource and financial knowledge on renewable energy. IRENA promotes the widespread adoption and sustainable use of all forms of renewable energy, including bioenergy, geothermal, hydropower, ocean, solar and wind energy, in the pursuit of sustainable development, energy access, energy www.irena.org security and low-carbon economic growth and prosperity. ACKNOWLEDGEMENTS This report benefited from the reviews and comments of numerous experts, including Ana Andrade (DGEG Portugal), Volker Berkhout (Fraunhofer IWES), Rina Bohle (Vestas), Heather Brent (EPRI), Henrik Breum (Danish Energy Agency), Luis Crespo (Estelasolar), Jürgen Dersch (DLR), Anthony Drummond (EPRI), Morten Dyrholm (Vestas), Lisa Ekstrand (Vestas), Pilar Gonzalez (Iberdrola), Claudia Grotz (Siemens Windpower), Daniel Gudopp (deea Solutions), Tomas Kåberger (Renewable Energy Institute), Keiji Kimura (Renewable Energy Institute), Simon Le Clech (Aldwych International Ltd), Roberto Lacal Arántegui (JRC), Eckhard Lüpfert (DLR), Christoph Pfister (Fraunhofer IWES), Simon Price (Exawatt), Mathis Rogner (IHA), with comments from EDF, Mott Macdonald and Sarawak Energy, Andreas Wade (First Solar) and Michael Waldron (IEA). Dolf Gielen (Director, IRENA Innovation and Technology Centre) also provided valuable input to the study. Andrei Ilas, Pablo Ralon, Asis Rodriguez and Michael Taylor (IRENA). Contributors: For further information or to provide feedback: [email protected] This report is available for download: www.irena.org/publications DISCLAIMER This publication and the material herein are provided “as-is”, for informational purposes. All reasonable precautions have been taken by IRENA to verify the reliability of the material featured in this publication. Neither IRENA nor any of its officials, agents, data or other, third-party content providers or licensors provides any warranty, including as to the accuracy, completeness, or fitness for a particular purpose or use of such material, or regarding the non- infringement of third-party rights, and they accept no responsibility or liability with regard to the use of this publication and the material therein. The material contained herein does not necessarily represent the views of the Members of IRENA, nor is it an endorsement of any project, product or service provider. The designations employed and the presentation of material herein do not imply the expression of any opinion on the part of IRENA concerning the legal status of any region, country, territory, city or area, or their authorities, or concerning the delimitation of frontiers or boundaries.

3 Renewable Power Generation Costs in 2017 3

4 RENEWABLE POWER GENERATION COSTS oday, countries around the world are more firmly committed than ever to accelerating renewable energy deployment. Technological T innovation, enabling policies and the drive to address climate change have placed renewables at the centre of the global energy transformation. Yet alongside these developments, the chief driver of renewable energy is its strong business case, which offers increasingly exciting economic opportunities. With rapidly falling renewable power generation costs, policy makers and investors need to confront the economic opportunities, as well as challenges, arising from a scale-up of renewable energy. Informed decision-making about the role of renewables in future electricity systems depends on reliable cost and performance data. In this context, the International Renewable Energy Agency (IRENA) has developed one of the most comprehensive datasets available on renewable power generation technology costs and performance. This detailed cost data confirms latest auction prices, showing renewables to be cost-competitive in a growing array of markets and conditions. The rate of cost reduction has been wholly impressive. Solar photovoltaic (PV) modules are more than 80% cheaper than in 2009. The cost of electricity from solar PV fell by almost three-quarters in 2010-2017 and continues to decline. Wind turbine prices have fallen by around half over a similar period, depending on the market, leading to cheaper wind power globally. Onshore wind electricity costs have dropped by almost a quarter since 2010, with average costs of USD 0.06 per kilowatt-hour in 2017. Such cost reductions are driven by continuous technological improvements, including higher solar PV module efficiencies and larger wind turbines. Industrialisation of these highly modular technologies has yielded impressive benefits, from economies of scale and greater competition to improved manufacturing processes and competitive supply chains. Simultaneously, various new cost reduction drivers are emerging. Competitive procurement, notably auctions, has resulted in more transparent costs, while global competition has brought the experience of a myriad of project developers to new markets. Their combination of expertise, purchasing power and access to international financial markets is further driving down project costs and risks, and a string of record-low auction prices for solar PV, concentrating solar power (CSP), onshore wind and offshore wind power were set in 2016-2017. The trend is clear: by 2020, all mainstream renewable power generation technologies can be expected to provide average costs at the lower end of the fossil-fuel cost range. In addition, several solar PV and wind power projects will provide some of the lowest-cost electricity from any source. 4

5 2017 As renewables go head-to-head with fossil-based power solutions to provide new capacity without financial support, key opportunities exist to open cost-effective technology pathways. This is especially true in developing countries, where much of the world’s future energy demand growth will occur. Renewable energy increasingly makes business sense for policy makers and investors. For this reason, renewables will continue driving the global energy transformation, while benefiting the environment and our collective future. Adnan Z. Amin Director-General International Renewable Energy Agency 5

6 RENEWABLE POWER GENERATION COSTS CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 08 Figures Tables and boxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Key findings EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 . INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 . . . . . 1 R enewable energy cost analysis at IRENA ... 26 1.1 Cost metrics for renewable power generation technologies ... 27 1.2 The IRENA Renewable Cost Database 1.3 30 ... COST TRENDS IN GLOBAL RENEWABLE POWER GENERATION . . . . . . . . . . . 33 2 2.1 T he new cost reduction drivers: Competitive procurement, international competition and improved technology ... 36 2.2 Renewable electricity cost trends by region and technology ... 40 2.3 The cost of renewable electricity to 2020: Insights from project data and auctions ... 46 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SOLAR PHOTOVOLTAICS 59 3 3.1 I nstalled cost trends ... 61 Capacity factors ... 66 3.2 3.3 Operation and maintenance costs ... 68 3.4 Levelised cost of electricity ... 69 CONCENTRATING SOLAR POWER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4 4.1 I nstalled cost trends ... 80 4.2 Capacity factors ... 83 4.3 Operation and maintenance costs ... 83 85 4.4 Levelised cost of electricity ... 6

7 WIND POWER 89 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 W ... 90 5.1 ind power technology trends ... Wind turbine costs 92 5.2 ... 5.3 Total installed costs onshore 94 99 ... 5.4 Total installed costs offshore Capacity factors ... 102 5.5 Operation and maintenance costs 104 5.6 ... Levelised cost of electricity 109 5.7 ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 HYDROPOWER 6 I nstalled cost trends ... 116 6.1 Capacity factors ... 120 6.2 2 6.3 peration and maintenance costs O ... 12 6.4 evelised cost of electricity 2 12 ... L 7 BIOENERGY FOR POWER 127 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 127 B 7.1 iomass feedstocks ... 128 7.2 Installed cost trends ... 7.3 Operation and maintenance costs 130 ... 130 Capacity factors and efficiency 7.4 Levelised cost of electricity ... 133 7.5 8 137 . . . . . . . . . . . . . . . . . . . . . . . . . . . . GEOTHERMAL POWER GENERATION nstalled cost trends 138 ... 8.1 I 138 Capacity factors 8.2 ... ... 140 8.3 Levelised cost of electricity 144 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES ANNEXES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 7

8 RENEWABLE POWER GENERATION COSTS FIGURES Global levelised cost of electricity Figure ES.1 Global weighted average total installed Figure 2.6 from utility-scale renewable power costs, capacity factors and LCOE for ... 17 generation technologies, 2010-2017 44 ... onshore wind, 2010-2017 Figure ES.2 The levelised cost of electricity for Figure 2.7 Global weighted average total installed projects and global weighted costs, capacity factors and LCOE for average values for CSP, solar PV, ... 45 bioenergy for power, 2010-2017 onshore and offshore wind, Global weighted average total installed F igure 2.8 ... 20 2010-2022 costs, capacity factors and LCOE for Learning curves for the global Figure ES.3 45 ... geothermal power, 2010-2017 weighted average levelized cost Figure 2.9 Global weighted average total installed of electricity from CSP, solar PV costs, capacity factors and LCOE for and onshore and offshore wind, ... 46 CSP, 2010-2017 ... 22 2010-2020 Global weighted average total installed Figure 2.10 igure ES.4 F lobal weighted average total G costs, capacity factors and LCOE for installed costs and project percentile offshore wind, 2010-2017 ... 47 ranges for CSP, solar PV, onshore and offshore wind, 2010-2017 23 ... igure 2.11 F roject LCOE ranges and weighted P averages for China and India, OECD Cost metrics analysed to calculate igure 1.1 F and rest of the world, 2016 and 2017 49 ... 27 the levelised cost of electricity. ... lobal levelised cost of electricity and G igure 2.12 F Figure 1.2 istribution of projects by technology D auction price trends for onshore wind and country in the IRENA renewable and solar PV, 2010-2020 ... 50 cost database and auctions database. 31 .. F igure 2.13 Global levelised cost of electricity and lobal levelised cost of electricity G igure 2.1 F auction price trends for offshore wind from utility-scale renewable power and CSP from project and auction generation technologies, 2010-2017 ... 34 data, 2010-2020 ... 52 F igure 2.2 Renewables are experiencing a virtuous G lobal weighted average CSP, solar F igure 2.14 cycle of technology improvement and PV, onshore and offshore wind project 36 cost reduction ... LCOE data to 2017 and auction price D evelopment of patents data for Figure B2.1 53 ... data to 2020, 2010-2020 renewable energy technologies, igure 2.15 F gional total installed cost ranges for Re 39 ... 2010-2016 onshore wind and solar PV, Figure 2.3 gional weighted average levelised Re 2016 and 2017 55 ... cost of electricity by renewable power lobal levelised cost of electricity and G igure 2.16 F generation technology, 2016 and 2017 . 4 0 auction price trends for solar PV, CSP, lobal weighted average total G Figure 2.4 onshore and offshore wind from installed costs, capacity factors and LCOE project and auction data, 2010-2022 .. 56 forsolar PV, 2010-2017 ... 42 igure 3.1 F Yearly added and cumulative global PV F igure 2.5 G lobal weighted average total installed capacity by region, 2006-2016 60 ... costs, capacity factors and LCOE for 43 ... hydropower, 2010-2017 8

9 2017 Solar PV module production: Capacity Figure 3.2 C F igure B3.1 ommercial solar PV total installed 61 .. and volume by technology, 2010-2016 cost and levelised cost of electricity 74 ... by country or state, 2009-2017 A verage monthly European solar PV F igure 3.3 module prices by module technology F igure 4.1 D evelopment of the cumulative and manufacturer, March 2010—May installed CSP capacity by region, 2017 and average yearly module prices ... 78 2006-2016. ... 62 by market in 2015 and 2016 S tatus of planned PTC and ST projects F igure 4.2 Figure 3.4 T otal installed costs for utility-scale 79 ... registered since 2015 solar PV projects and the global F igure 4.3 Installed costs and capacity factors 64 ... weighted average, 2010-2017 of CSP projects by their quantity of U tility-scale solar PV total installed F igure 3.5 80 ... storage, 1984-2016. cost trends in selected countries, C SP installed costs by project size, F igure 4.4 ... 6 2010-2017 5 collector type and amount of storage, Figure 3.6 E stimated utility-scale solar PV ... 81 2009-2016 system costs: China compared to Figure 4.5 S torage hours of planned CSP other countries, 2015-2016. . . . . . . . . . . . 66 projects with operational status D etailed breakdown of utility-scale Figure 3.7 ... 82 updates in 2015-2017 67 solar PV costs by country, 2016 ... H eat-transfer fluid use in operational F igure 4.6 F igure 3.8 A verage total installed costs of and planned projects with operational residential solar PV systems by 82 ... status updates in 2015-2017 ... 68 country, Q2 2007-Q1 2017 F igure 4.7 C apacity factor trends for CSP plants, G lobal weighted average capacity F igure 3.9 ... 84 2009-2016 factors for utility-scale PV systems, D irect normal irradiance levels for CSP F igure 4.8 69 ... 2010-2016 projects by year of commissioning and F igure 3.10 L evelised cost of electricity from ... 84 technology, 2009-2016 utility-scale solar PV projects, global F igure 4.9 T he levelised cost of electricity for CSP weighted average and range, ... 85 projects, 2009-2016 70 ... 2010-2016 L evelised cost of electricity and auction F igure 4.10 U tility-scale solar PV: Electricity F igure 3.11 ... 86 price trends for CSP, 2010-2022 cost trends in selected countries, 71 ... 2010-2017 Figure 5.1 W eighted average rotor diameter and nameplate capacity evolution, F igure 3.12 L evelised cost of electricity from ... 91 2010-2016 residential solar PV systems by 72 ... country, Q2 2007-Q1 2017 W ind turbine price indices and price F igure 5.2 trends in the United States and China, L evelised cost of electricity from F igure 3.13 9 3 ... 1997-2017 residential PV: Average differentials between Germany and other Figure 5.3 Total installed costs of onshore wind ... 73 countries, 2010-2017. projects and global weighted average, 1983-2017 94 ... 9

10 RENEWABLE POWER GENERATION COSTS Onshore wind weighted average Figure 5.4 Full-service (initial and renewal) O&M Figure 5.15 total installed costs in 12 countries, pricing indexes, the weighted average 1983-2016 95 ... O&M revenues of two manufacturers and O&M costs in Denmark, Germany, otal installed costs ranges and T igure 5.5 F ... 106 Ireland and Sweden, 2008-2017 weighted averages for onshore wind farms by country/region, 2010-2016 96 ... Figure 5.16 Project level O&M cost data by component from a subset of the IRENA Figure 5.6 ost breakdown of onshore wind C database compared to the BNEF O&M farms by country and region, ... 107 index range, 2008-2016 ... 1998 -2016 97 T he global weighted average levelised Figure 5.17 Average total installed cost igure 5.7 F cost of electricity of onshore wind, reduction by source for onshore ... 11 0 1983-2017 98 ... wind, 2010-2014/15 and 1998-2012 Figure 5.18 he weighted average LCOE of T Figure 5.8 istribution and weighted average D commissioned onshore wind projects share of onshore wind total installed in 12 countries, 1983-2016 111 ... costs by source for China and India, and rest of the world, 2006-2017 ... 99 gional weighted average LCOE Re igure 5.19 F and ranges of onshore wind in ffshore wind farm projects and O igure 5.9 F 2010 and 2016 12 1 ... distance from port, 2001-2017 00 1 ... Figure 5.20 he LCOE of commissioned and T Figure 5.10 Total investment costs for proposed offshore wind projects and commissioned and proposed offshore auction results, 2000–2022 13 ... 1 101 projects, 2000-2018 ... otal installed costs by project and T Figure 6.1 Global weighted average capacity Figure 5.11 global weighted averages for factors for new onshore and offshore 17 hydropower, 2010-2017 ... 1 wind power capacity additions by year ... 102 of commissioning, 1983–2017 Figure 6.2 otal installed cost ranges and weighted T averages for hydropower projects by Figure 5.12 H istorical onshore wind capacity country/region, 2010-2016 8 ... 11 03 1 ... factors in a sample of 12 countries otal installed cost ranges and capacity T Figure 6.3 ountry-specific weighted average C Figure 5.13 weighted averages for small and large capacity factors for new onshore hydropower projects by country/region, wind projects, 2010 and 2016 04 1 ... 2010-2016 119 ... Figure 5.14 lobal weighted average hub height, G Figure 6.4 otal installed cost breakdown T rotor diameter and capacity factors, by component and capacity weighted and cumulative capacity for onshore averages for 25 hydropower projects wind, 1983-2016 05 1 ... in China, India and Sri Lanka, 2010-2016 ... 12 0 ydropower project capacity factors H Figure 6.5 and capacity weighted averages for large and small hydropower projects 1 by country/region, 2010-2016 ... 12 10

11 2017 TABLES Figure 6.6 Hydropower O&M cost breakdown by Ta ble 5 .1 O&M costs of onshore wind in project for a sample of 25 projects in OECD countries... selected 108 .. 123 China, India and Sri Lanka,2010-2016 The weighted average LCOE Ta ble 5 . 2 Levelised cost of electricity and Figure 6.7 reduction of commissioned onshore weighted averages of small and wind projects in 12 countries... 111 projects by large hydropower Table 7.1 Fixed and variable O&M costs for 125 ... country/region, 2010-2016 bioenergy power... 131 Figure 7.1 T otal installed costs of biomass-fired generation technologies by project 129 ... capacity Total installed costs of biomass-fired Figure 7.2 generation technologies by 1 30 country/region ... Figure 7.3 P roject capacity factors and weighted averages of biomass-fired electricity generation systems by country ... 132 and region Levelised cost of electricity by Figure 7.4 BOXES project and weighted averages of bioenergy-fired electricity generation A C Box 1 autionary Tale: When is an LCOE by feedstock and country/region, ... 38 not a FiT or a PPA Price? 2000-2016 ... 1 34 T racking Innovation trends: A look B ox 2 Figure 7.5 Levelised cost of electricity factors of 38 ... at patent data for renewables bioenergy-fired projects, 2000-2016 ... 1 35 Box 3 S olar PV cost trends in the ... 74 commercial sector G eothermal power total installed Figure 8.1 costs by project, technology and P ... 1 24 umped hydro storage B ox 4 139 ... capacity, 2007-2020 Figure 8.2 C apacity factors of new geothermal power plant by technology and ... 1 40 project size, 2007-2020 E lectricity generation and capacity Figure 8.3 factor of an 88.2 MW geothermal plant 1 ... in California, 1989-2017 41 Figure 8.4 Levelised cost of electricity of geothermal power projects by technology and size, 42 2007-2020 1 ... 11

12 RENEWABLE POWER GENERATION COSTS ABBREVIATIONS independent power producer IPP ACP Alternative Compliance Payment International Renewable Energy IRENA CAD Canadian dollar Agency Caribbean Community CARICOM IRP integrated resource plan CCS carbon capture and storage kW kilowatt Co uncil of European Energy Regulators CEER kilowatt-hour kWh Contract for Difference CfD LSE load-serving entities CSP concentrating solar power MDG Millennium Development Goal DNI direct normal irradiance Ministry of Energy, Mines, Water and MEMEE European Council EC Environment (Morocco) Economic Community of West African ECOWAS M iddle East and North Africa MENA States m illion tonnes of oil equivalent Mtoe EJ exajoule meg awatt MW ropean Union EU Eu m MWh egawatt-hour EUR euro N NDRC ational Development and Reform feed-in tariff FIT Commission British pound GBP NREL ational Renewable Energy Laboratory N (US) gross domestic product GDP O rganisation for Economic OECD lobal Status Report GSR G Co-operation and Development igawatt GW g PPA Power Purchase Agreement GWh igawatt-hour g Su SDG stainable Development Goal igawatt-thermal GWth g terawatt-hour TWh ILUC indirect land-use change VRE Variable Renewable Electricity INR Indian rupee 12

13 2017 13

14 RENEWABLE POWER GENERATION COSTS KEY FINDINGS • • After years of steady cost decline for solar Three key cost reduction drivers are becoming and wind technologies, renewable power is increasingly important: becoming an increasingly competitive way to 1. technology improvements; meet new generation needs. 2. competitive procurement; • For projects commissioned in 2017, electricity 3. costs from renewable power generation have a large base of experienced, internationally continued to fall. active project developers. • • - Bioenergy-for-power, hydropower, geother Continuous technological innovation remains mal and onshore wind projects commissioned a constant in the renewable power generation in 2017 largely fell within the range of gener - market. With today’s low equipment costs, 1 ation costs for fossil-based electricity. Some however, innovations that unlock efficiencies projects undercut fossil fuels, data collected in manufacturing, reduce installed costs or by the International Renewable Energy Agency improve performance for power-generation (IRENA) shows. equipment will take on increasing significance. • • The global weighted average cost of electricity These trends are part of a broader shift across was USD 0.05 per kilowatt-hour (kWh) from the power generation sector to low-cost new hydropower projects in 2017. It was As competitive procurement drives renewables. 0.06/kWh for onshore wind and 0.07/kWh USD costs lower, a wide range of project developers for bioenergy and geothermal projects. are positioning themselves for growth. • • The fall in electricity costs from utility-scale The results of recent renewable power auctions solar photovoltaic (PV) projects since 2010 has – for projects to be commissioned in the been remarkable. The global weighted average coming years – confirm that cost reductions levelised cost of electricity (LCOE) of utility- are set to continue through 2020 and beyond . scale solar PV has fallen 73% since 2010, to Auctions provide valuable price signals about USD 0.10/kWh for new projects commissioned future electricity cost trends. in 2017. • Record low auction prices for solar PV in Dubai, Mexico, Peru, Chile, Abu Dhabi and Saudi Arabia 1. The fossil fuel-fired power generation cost range for G20 countries in 2017 was estimated to be between USD 0.05 and USD 0.17/kWh. 14

15 2017 - auction and project-level cost data, global aver in 2016 and 2017 confirm that the LCOE can be 0. 05/kWh age costs could decline to about USD reduced to USD 0.03/kWh from 2018 onward, 0. 06/kWh for solar for onshore wind and USD given the right conditions. PV. • Onshore wind is one of the most competitive • Recent sources of new generation capacity. Auction results suggest that concentrating solar auctions in Brazil, Canada, Germany, India, - power (CSP) and offshore wind will provide elec Mexico and Morocco have resulted in onshore and 0.06 0.10/kWh USD tricity for between USD wind power LCOEs as low as USD 0.03/kWh. by 2020. • • The lowest auction prices for renewable Falling renewable power costs signal a real power reflect a nearly constant set of key paradigm shift in the competitiveness of These include: different power generation options. This competitiveness factors. a favourable regulatory and institutional includes cheaper electricity from renewables as framework; low offtake and country risks; a a whole, as well as the very low costs now being attained from the best solar PV and onshore strong, local civil engineering base; favourable wind projects. taxation regimes; low project development costs; and excellent resources. • Sharp cost reductions – both recent and • anticipated – represent remarkable deflation Electricity from renewables will soon be rates for various solar and wind options. consistently cheaper than from most fossil 2 Learning rates By 2020, all the renewable power fuels. for the 2010-2020 period, based generation technologies that are now in on project and auction data, are estimated at commercial use are expected to fall within the 14% for offshore wind, 21% for onshore wind, fossil fuel-fired cost range, with most at the 30% for CSP and 35% for solar PV. lower end or undercutting fossil fuels. • Reductions in total installed costs are driving • the fall in LCOE for solar and wind power The outlook for solar and wind electricity costs technologies to varying extents. This has been to 2020 presages the lowest costs yet seen for most notable for solar PV, CSP and onshore - these modular technologies, which can be de wind. ployed around the world. Based on the latest 2. The learning rate is the percentage cost reduction experienced for every doubling of cumulative installed capacity. 15

16 RENEWABLE POWER GENERATION COSTS EXECUTIVE SUMMARY the global weighted average LCOE was around For new projects commissioned in 2017, elec - 0.07/kWh. USD tricity costs from renewable power generation have continued to fall. After years of steady cost The fall in electricity costs from utility-scale solar decline, renewable power technologies are be - photovoltaic (PV) projects since 2010 has been coming an increasingly competitive way to meet Driven by an 81% decrease in solar PV remarkable. new generation needs. - module prices since the end of 2009, along with re - ductions in balance of system (BoS) costs, the glob In 2017, as deployment of renewable power al weighted average LCOE of utility-scale solar PV generation technologies accelerated, there has 0.10/kWh. fell 73% between 2010 and 2017, to USD been a relentless improvement in their com - Increasingly, this technology is competing head- petitiveness. Bioenergy for power, hydropower, to-head with conventional power sources – and - geothermal and onshore wind projects commis doing so without financial support. sioned in 2017 largely fell within the range of fossil fuel-fired electricity generation costs (Figure ES.1), Offshore wind power and concentrated solar data collected by the International Renewable power (CSP), though still in their infancy in terms Energy Agency (IRENA) shows. Indeed levelised of deployment, both saw their costs fall between 1 cost of electricity (LCOE) for these technologies 2010 and 2017. The global weighted average LCOE was at the lower end of the LCOE range for fossil of offshore wind projects commissioned in 2017 was 2 fuel options. 0.14/kWh, while for CSP, it was USD 0.22/kWh. USD However, auction results in 2016 and 2017, for The global weighted average LCOE of new CSP and offshore wind projects that will be hydropower plants commissioned in 2017 was commissioned in 2020 and beyond, signal a around USD 0.05 per kilowatt-hour (kWh), while for step-change, with costs falling to between onshore wind plants it was around USD 0.06/kWh. USD 0.06 and USD 0.10/kWh for CSP and offshore For new bioenergy and geothermal projects, wind. 1. T he LCOE of a given technology is the ratio of lifetime costs to lifetime electricity generation, both of which are discounted back to a common year using a discount rate that reflects the average cost of capital. In this report, all LCOE results are calculated using a fixed assumption of a real cost of capital of 7.5% in OECD countries and China, and 10% in the rest of the world, unless explicitly mentioned. All LCOE calculations exclude the impact of any financial support. The fossil fuel-fired electricity cost range in 2017 was estimated to range from a low of USD 0.05 per kilowatt-hour (kWh) to a high 2. USD 0.17/kWh, depending on the fuel and country. 16

17 2017 Figure ES.1 Global levelised cost of electricity from utility-scale renewable power generation technologies, 2010-2017 Concentrating Onshore BiomassGeothermalHydroSolar Oshore photovoltaic wind solar power wind 0.4 0.36 0.33 0.3 0.22 0.2 Fossil fuel cost range 0.17 2016 USD/kWh 0.14 0.1 0.10 0.08 0.07 0.07 0.07 0.06 0.05 0.05 0.04 20102017201020172010201720102017201020172010201720102017 300 ≥ 1 100200 ≥ Capacity (MW) Source: IRENA Renewable Cost Database. Note: T he diameter of the circle represents the size of the project, with its centre the value for the cost of each project on the Y axis. The thick lines are the global weighted average LCOE value for plants commissioned in each year. Real weighted average cost of capital is 7.5% for OECD countries and China and 10% for the rest of the world. The band represents the fossil fuel-fired power generation cost range. 17

18 RENEWABLE POWER GENERATION COSTS These trends are part of a larger dynamic across Three main cost reduction drivers have the power generation sector, prompting a rapid emerged for renewable power: 1) technology transition in the way the industry functions. improvements; 2) competitive procurement; and 3) a large base of experienced, internationally In many parts of the world, renewable power technologies now offer the lowest cost source of active project developers. new power generation. In the past, typically, there Historically, technology improvements have was a framework offering direct financial support, been vital to the performance increases and often tailored to individual technologies (e.g., solar installed cost reductions which have (in addition PV) and even segments (e.g., varying support for to industrialisation of the sector and economies of residential, commercial and utility-scale sectors, scale) made solar and wind power technologies sometimes differentiated by other factors such as competitive. Competitive procurement — amid whether they are building-integrated or not). Now, globalisation of the renewable power market — this is being replaced by a favourable regulatory has emerged more recently as another key driver. and institutional framework that sets the stage Along with this comes the emergence of a large for competitive procurement of renewable base of experienced medium-to-large project power generation to meet a country’s energy, developers, actively seeking new markets around environmental and development policy goals. the world. The confluence of these factors is Around the world, medium-to-large renewable increasingly driving cost reductions for renewables, project developers are adapting to this new reality with effects that will only grow in magnitude in and increasingly looking for global opportunities 2018 and beyond. to expand their business. They are bringing, not only their hard won experience, but access to Continuous technology innovation remains a international capital markets. In competition with constant in the renewable power generation market. their peers, they are finding ways to continuously Indeed, in today’s low equipment cost era, technology reduce costs. innovations that unlock efficiencies in manufacturing, as well as power generation equipment — in terms The results of recent renewable power auctions of performance improvements or installed cost – for projects to be commissioned in the coming reductions — will take on increasing importance. years – confirm that cost reductions are set to Bigger wind turbines with larger swept areas harvest continue to 2020 and beyond. more electricity from the same resource. New In addition to the IRENA Renewable Cost Database, solar PV cell architectures offer greater efficiency. which contains project level cost data for around Real-time data and ‘big data’ have enhanced 15 predictive maintenance and reduced operation 00 utility-scale projects, IRENA has compiled a 0 and maintenance (O&M) costs. These are just a database of auction results and other competitive few examples of the continuous innovation driving 000 projects. procurement processes for around 7 reductions in installed costs, unlocking performance Although care must be taken in comparing the improvements and reducing O&M costs. Technology results of these two databases, as an auction price improvements, therefore, remain a key part of the is not necessarily directly comparable to an LCOE 4 cost reduction potential for renewable power. At the calculation, analysis of the results of the two same time, the maturity and proven track record of databases provides some important insights into renewable power technologies now reduces project the likely distribution of renewable electricity costs 3 risk, significantly lowering the cost of capital. over the next few years. 3. T he generally low cost of debt since 2010 has combined to enhance this effect as not only have risk margins fallen, but the base cost of debt as well. 4. At a minimum, the weighted average cost of capital (WACC) is not going to be the same. For an LCOE calculation, the WACC is a fixed and known value, whereas the WACC of a project in an auction is unknown and subsumed in the range of other factors that determined the price bid by an individual project developer. 18

19 2017 low costs depends on supporting factors, such Record low auction prices for solar PV in 2016 and as access to low-cost finance, a conducive policy 2017 in Dubai, Mexico, Peru, Chile, Abu Dhabi and environment and good auction design. The key Saudi Arabia have shown that an LCOE of USD 0.03/ policy drivers (IRENA, 2017e, Renewable Energy kWh is possible from 2018 and beyond, given the Auctions: Analysing 2016) are confirmed by the right conditions. These include: a regulatory and latest auction results. institutional framework favourable to renewables; low offtake and country risks; a strong, local civil Electricity from renewables will soon be engineering base; favourable taxation regimes; consistently cheaper than from fossil fuels. By low project development costs; and excellent solar 2020, all the power generation technologies resources. that are now in commercial use will fall within the fossil fuel-fired cost range, with most at the Similarly, very low auction results for onshore wind lower end or even undercutting fossil fuels. in countries such as Brazil, Canada, Germany, India Mexico and Morocco have shown that onshore Even by 2020, projects contracted via competitive wind is one of the most competitive sources of new procurement will represent a relatively small generation capacity. For CSP and offshore wind, subset of annual new renewable power generation 2016 and 2017 have been breakthrough years, as capacity additions – and trends in auction results auction results around the world have confirmed may not remain representative of LCOE trends at a that a step change in costs has been achieved and project level. Nevertheless, recent auction results will be delivered in projects commissioned in 2020 show that cost reductions will continue for CSP, and beyond. Indeed, auction results in 2016 and solar PV, onshore and offshore wind through 2020 2017 suggest that projects commissioned from and beyond. While the validity of comparing LCOE 2020 onwards for both technologies could fall in and auction prices for individual projects must be the range USD 0.10/kWh. 0.06 and USD done with caution, the volume of data available and the consistent trends between the two datasets Competitive procurement, particularly auctions, provide some confidence in the overall trend. is spurring further cost reductions for power from solar and wind power technologies. Still, achieving 19

20 RENEWABLE POWER GENERATION COSTS available do accurately represent global deploy - Analysing the the trends in the LCOE of projects ment trends, then by 2019 or 2020, the average and auction results to 2020 suggests that LCOE for solar PV may fall to below USD 0.06/kWh, average costs for onshore wind could fall from converging to slightly above that of onshore wind, USD 0.06/kWh in 2017 to USD 0.05/kWh by 2020. at USD 0.05/kWh. The recent auction results for offshore wind from 2016 and 2017 in Belgium, Denmark, the Kingdom The outlook for solar and wind electricity costs of the Netherlands, Germany and the United to 2020, based on the latest auction and project- Kingdom suggest that for projects that will be level cost data, presages the lowest costs yet commissioned in 2020 and beyond, costs could fall seen for these modular technologies that can be in the USD 06 to USD 10/kWh range. Indeed, in 0. 0. deployed around the world. Germany, two projects that will be commissioned By 2019, the best onshore wind and solar PV in 2024 and one in 2025 won with bids that did not projects will be delivering electricity for an LCOE ask for a subsidy over market rates. A similar story equivalent of USD 03/kWh, or less, with CSP 0. has emerged for CSP, where a project in South and offshore wind capable of providing electricity Australia to be commissioned from 2020 will have very competitively, in the range of USD 0.06 to a cost of USD 0.06/kWh, while in Dubai, a project 10/kWh from 2020 (Figure ES.2). Already 0. USD that will be commissioned from 2022 onwards will today, and increasingly in the future, many have a cost of USD 0.07/kWh. renewable power generation projects can Solar PV auction data needs to be treated with undercut fossil fuel-fired electricity generation, somewhat more caution. This is because the without financial support. With the right regulatory distribution of projects is concentrated in higher- and institutional frameworks in place, their - irradiation locations than recent capacity-weight competitiveness should only further improve. ed deployment. Even so, if the auction results The levelised cost of electricity for projects and global weighted average values for CSP, solar PV, Figure ES.2 onshore and offshore wind, 2010-2022 Concentrating solar Oshore wind Solar PV Onshore wind power 0.4 Auction database LCOE database 0.3 0.2 2016 USD/kWh 0.1 Fossil fuel cost range 0.0 2012 2012 2012 2012 2018 2018 2018 2108 2016 2016 2016 2016 2014 2014 2014 2014 2010 2010 2010 2010 2022 2022 2020 2020 2020 2020 Source: IRENA Renewable Cost Database and Auctions Database. Note: Each circle represents an individual project or an auction result where there was a single clearing price at auction. The centre of the circle is the value for the cost of each project on the Y axis. The thick lines are the global weighted average LCOE, or auction values, by year. For the LCOE data, the real WACC is 7.5% for OECD countries and China, and 10% for the rest of the world. The band represents the fossil fuel-fired power generation cost range. 20

21 2017 The sharp cost reductions for CSP, solar PV, Decreasing electricity costs from renewables as onshore and offshore wind – both recent and a whole, and the low costs from the best solar anticipated – represent remarkable deflation PV and onshore wind projects, represent a real rates. paradigm shift in the competitiveness of different Solar and wind power power generation options. Conventional wisdom has been a poor guide in will provide very affordable electricity, with all estimating the rate of cost reductions from solar and the associated economic benefits. Furthermore, wind power technologies. It has underestimated their low costs mean that previously uneconomic the capacity of technology improvements, the strategies in the power sector can become industrialisation of manufacturing, economies profitable. Curtailment – previously an unthinkable of scale, manufacturing efficiencies, process economic burden for renewables – could become innovations by developers and, competition in a rational economic decision, maximising variable supply chains to all continuously drive down costs renewable penetration and minimising overall faster than expected in the right regulatory and system costs. policy setting. Similarly, very low prices in areas with excellent The decline in the cost of electricity experienced solar and wind resources could open-up the from 2010 to 2017, and signalled for 2020 from economic potential of “power-to-X” technologies auction data, is plotted against cumulative installed (e.g., power to hydrogen or ammonia, or other capacity in Figure ES.3 for the four main solar and energy dense, storable mediums). At the same wind technologies. A log-log scale is used to allow time, low prices make the economics of electricity easy interpretation as learning curves. The learning storage more favourable. This could turn a rate for offshore wind (i.e. the LCOE reduction potential drawback of electric vehicles (EVs) – their for every doubling in global cumulative installed potentially high instantaneous power demand for capacity) could reach 14% over the period 2010- recharging – into an asset. In effect, EVs can take 2020, with new capacity additions over this period advantage of cheap renewable power when it is estimated to be around 90% of the cumulative available, while potentially feeding electricity back installed offshore wind capacity that would be into the grid when needed. 5 deployed by the end of 2020. This, however, needs to be balanced against the For onshore wind, the learning rate for 2010 to increased costs of integrating variable renewables 2020 may reach 21%, with new capacity added and the increased flexibility required to manage over this period covering an estimated 75% of systems with very high levels of variable renewable cumulative installed capacity at the end of 2020. energy (VRE). To date, these integration costs have CSP has a higher estimated learning rate of remained modest, but they could rise as very high 30%, with deployment between 2010 and 2020 VRE shares are reached (IRENA, 2017f, Chapter 3 in representing an estimated 89% of cumulative Perspectives for the Energy Transition), especially 6 installed capacity by the end of that period. Solar without complementary policies across the power PV has the highest estimated learning rate – 35% sector. For instance, if transmission expansions fail between 2010 and 2020 – with new capacity to keep pace with deployment, renewable power additions over this timescale that are estimated to sources could face curtailment. be 94% of cumulative capacity by its conclusion. 5. G lobal cumulative installed capacity of CSP is projected to be 12 GW by 2020, for offshore wind 31 GW, solar PV 650 GW and onshore wind 712 GW. This is based on IRENA (2017a), GWEC (2017), WindEurope (2017), SolarPower Europe (2017), and MAKE Consulting, 2017a. 6. Extending the horizon to 2022 to take into account the likely commissioning of the Dubai Electricity and Water Authority project increases uncertainty over total deployment values, but in most scenarios would not materially change the learning rate. 21

22 RENEWABLE POWER GENERATION COSTS Learning curves for the global weighted average levelized cost of electricity from CSP, solar PV Figure ES.3 and onshore and offshore wind, 2010-2020 0.500 0.400 2011 2012 2010 2011 2016 0.300 2010 2013 2015 2012 2013 0.200 2015 2016 2013 2014 2010 0.150 2017 2014 2015 2012 2016 0.100 2017 2013 2010 2016 USD/kWh 2020 2020 2016 0.070 2015 2020 Fossil fuel cost range 2020 0.050 0.040 0.030 0.020 0.015 0.010 50 000 20 000 10 000 5 000 500 000 200 000 100 000 1 000 000 1 000 2 000 Onshore wind O‚shore wind CSP PV Cumulative deployment (MW) Based on IRENA Renewable Cost Database and Auctions Database; GWEC, 2017; WindEurope, 2017; MAKE Consulting, 2017a; and SolarPower Europe, 2017a. Note: Each circle represents an individual project, or, in some cases, auction result where there was a single clearing price at auction. The centre of the circle is the value for the cost of each project on the Y axis. The thick lines are the global weighted average LCOE or auction values by year. For the LCOE data, the real WACC is 7.5% for OECD countries and China, and 10% for the rest of the world. The band represents the fossil fuel-fired power generation cost range. 22

23 2017 storage. Wherever renewable power technologies Onshore wind is one of the technologies with the can be modular, scalable and replicable, decision longest histories of available cost data. Data in the makers can be confident that industrialisation IRENA Renewable Cost Database shows that the and the opening of new markets will yield steady learning rate for the cost of electricity from this cost reductions in the right regulatory and policy source is higher for the period 2010-2020 than the environment. learning rate estimated for the period 1983-2016. This will, in all probability, be in part due to a lower Reductions in total installed costs are driving WACC from the auction results than is used in the the fall in the LCOE for solar and wind power LCOE calculations. This is unlikely to explain all of technologies, but to varying extents. They have the difference, however. The data therefore tends been most important for solar PV, CSP and to suggest that the learning rate for onshore wind, onshore wind. at least, is currently higher than the long-term On the back of price declines for solar PV modules, average. the installed costs of utility-scale solar PV projects The modular, scalable nature solar and wind power fell by 68% between 2010 and 2017, with the LCOE generation technologies, and the replicability of for the technology falling 73% over that period. their project development process, rewards stable The total installed costs of newly commissioned support policies with continuous cost reductions. CSP projects fell by 27% in 2010-2017, with a 33% This has already made onshore wind and solar LCOE reduction overall. Installed costs for newly PV highly competitive options for new generation commissioned onshore wind projects fell by 20%, capacity. Auction results suggest that CSP and with a 22% reduction in LCOE. For offshore wind, offshore wind should follow a similar path. A the total installed costs fell by 2%, with a 13% comparable process is playing out for electricity reduction in LCOE over the same period. Figure ES.4 Global weighted average total installed costs and project percentile ranges for CSP, solar PV, onshore and offshore wind, 2010-2017 Solar photovoltaicConcentrating solar power O shore wind Onshore wind 10 000 9 000 8 000 7 583 7 000 th percentile 95 6 000 5 564 5 000 4 394 4 331 2016 USD/kW 4 000 4 239 3 000 th 1 843 percentile 5 2 000 1 477 1 388 1 000 0 2017 2017 2017 2017 2010 2010 2010 2010 Source: IRENA Renewable Cost Database. 23

24

25 1. INTRODUCTION biomass for power, hydropower, geothermal and he electricity sector is undergoing a period onshore wind projects consistently provided new of rapid, unprecedented change in the scale T electricity at competitive rates – compared to and breadth of deployment of renewable power fossil fuel-fired power generation – excluding the generation technologies. Since 2012, these have impact of any financial support. accounted for more than half of new electric power generation capacity additions, worldwide. At the It is growth in the “new” renewable power generation end of 2016, total renewable power generation technologies of solar and wind, however, that has capacity surpassed 2 000 GW, meaning that pushed renewable power generation capacity it had more than doubled in the space of nine additions to record levels. The levelised cost of years (IRENA, 2017a). New capacity additions of electricity (LCOE) of solar PV fell 73% between G G W W, with 36 renewables in 2016 reached 162 2010 and 2017, making it increasingly competitive of new hydropower capacity added, 51 GW of wind at the utility scale. Technology improvements and GW of solar photovoltaic (PV) capacity, power, 71 installed cost reductions have made onshore wind 9 GW of bioenergy power generation capacity and one of the most competitive sources of new power a combined 1 GW from concentrating solar power generation. Despite the fact that CSP and offshore (CSP), geothermal and marine energy. wind are in their deployment infancy, these technologies have seen their costs come down. This growth is set to continue, with accelerating Tender and auction results in 2016 and 2017 show deployment of renewables, notably for solar PV increasingly that even without financial support, in China, set to continue. Global solar PV capacity these technologies will be able to compete directly additions in 2017, in all probability, will flirt with, or with fossil fuels from 2020 onwards if the right exceed, 90 GW, while new capacity additions for policy and regulatory frameworks are in place. GW, setting the wind power are likely to exceed 40 scene for another record year for renewable power Crucially, the drivers behind lower equipment and generation deployment. installed costs – and performance improvements – have not yet run their course, either. Continued cost - Renewable power generation is currently benefit reductions for solar and wind power technologies ting from a virtuous cycle, in which policy support can therefore still be expected (IRENA, 2016a). for renewable power generation technologies - leads to accelerated deployment, technology im The renewable energy industry thus has a track provements and cost reductions, with these then record of delivering on cost reductions. These reducing the cost of electricity from renewable have been achieved by unlocking economies-of- power generation technologies and encouraging scale, investing in more efficient manufacturing greater uptake of these technologies. In 2016, processes, improving the efficiency of technologies, in many regions of the world, the commissioned 25

26 RENEWABLE POWER GENERATION COSTS and by demonstrating a technological maturity that reduces financing costs and drives down costs Renewables increasingly in supply chains. Auction results around the world provide electricity at costs in 2016 and 2017 for future delivery graphically highlight this. Record low prices for solar PV in Abu competitive with, or lower Dhabi, Chile, Dubai, Mexico, Peru and Saudi Arabia than, fossil-based power highlight just how far renewables have come, with results around USD 0.03/kWh on an LCOE basis now setting the benchmark. The full cost of some onshore and offshore wind out to 2025 (IRENA, onshore wind and solar PV projects that will come 2016a), along with a regional report on solar PV online in 2018 and beyond will be less than only costs in Africa (IRENA, 2016b). IRENA has also lev - the variable costs of many existing fossil fuel-fired eraged its cost data to provide analytical products generators. that support policymakers in understanding the Yet, the public debate around renewable energy implications of cost trends, including the IRENA - often continues to suffer from an outdated per Cost and Competitiveness Indicators for Rooftop - ception that renewable energy is not competi Solar PV (IRENA, 2017b). In 2017, IRENA also tive. This report demonstrates that the blanket released its analysis of electricity storage costs assumption that renewable power generation is and markets out to 2030 (IRENA, 2017c). This rep - expensive is outdated given that renewable power resents the beginning of IRENA’s efforts to analyse generation is increasingly providing electricity at the cost and performance of the technologies that costs that are competitive, or even lower than, will help facilitate the energy transition. IRENA has fossil fuel-fired power generation costs. also started to analyse the flow of cost and per - formance data that is becoming available from the increased use of auctions to competitively procure 1.1 R ENEWABLE ENERGY COST ANALYSIS renewable power generation capacity. AT IRENA This analysis has contributed to more transparent Since 2012, IRENAs cost analysis programme cost data in the public domain, allowing policy has been collecting and reporting the cost makers, key decision makers, industry players, and performance data of renewable energy researchers and the media to have a better technologies. Having reliable, transparent, up-to- understanding of the true costs for renewable date cost and performance data from a reliable energy today and their continued cost reduction source is vital, given the rapid growth in installed potential. Given the rapid cost reductions being capacity of these technologies. The associated experienced, especially by solar and wind power cost reductions mean that data from even one or technologies, the importance of this data being in two years ago can be significantly erroneous, and, the public domain should not be underestimated, indeed, in the case of solar PV, in some markets, as there is a significant amount of perceived even data six months old can significantly overstate knowledge about the cost and performance of the costs. renewable power generation that is not accurate IRENA has previously reported on costs in the and can even be misleading. This problem is often power generation sector (IRENA, 2012a-e; IRENA, compounded by a lack of transparency in the 2013a; IRENA, 2015) and the transport sector methodology and the assumptions used by many (IRENA, 2013b). IRENA analysis is not restricted commentators in their cost calculations, which to historical costs or global analysis, either. It is can lead to confusion about the comparability of also increasingly focused on answering questions data. This report, based on the IRENA Renewable about the future cost and competitiveness of Cost Database – with its a large global dataset – renewables and their cost structures in new and provides one of the most comprehensive overviews emerging markets. IRENA has released reports on of renewable power generation costs using a the cost reduction potential for solar PV, CSP and consistent methodology and set of assumptions. 26

27 2017 yield a lower return on capital, or even a loss (see C 1.2 OST METRICS FOR RENEWABLE Annex One for a detailed discussion of the LCOE POWER GENERATION TECHNOLOGIES and other cost metrics). The LCOE of renewable The cost of power generation technologies can energy technologies varies by technology, country be measured in a number of ways, and each way and project, based on the renewable energy of accounting for the cost brings its own insights. resource, capital and operating costs, and the IRENAs work in this report focuses on analysing efficiency/performance of the technology. The the impact of technology and market development approach used to assess the LCOE in this report is on the LCOE. To understand the drivers of these based on a simple discounted cash flow analysis. changes requires an analysis of the equipment This method of calculating the cost of electricity costs, total installed costs, performance (capacity is based on discounting financial flows (annual, factors), operation and maintenance (O&M) costs quarterly or monthly) to a common basis, taking and weighted average cost of capital (WACC) into consideration the time value of money. Given (Figure 1.1). It also requires an analysis of trends the capital-intensive nature of most renewable in technology developments and their market power generation technologies and the fact that share, manufacturing innovations and supply chain fuel costs are low-to-zero, the WACC (or discount capacities, and an understanding of developments rate) used to evaluate the project has a critical in the drivers of the different markets for each impact on the LCOE. technology. The total installed cost for projects in the IRENA The LCOE is an indicator of the price of electricity Renewable Cost Database represent all of the required for a project where revenues would equal costs of developing a project. They thus differ costs, including making a return on the capital from “overnight” capital costs in that they include invested equal to the discount rate or WACC. An interest during construction (including on working electricity price above this would yield a greater capital needs), project development costs and any return on capital, while a price below it would upfront financing costs. Cost metrics analysed to calculate the levelised cost of electricity. Figure 1.1 Project development Site preparation Operation & maintenance Gird connection WACC Working capital Resource quality Auxiliary equipment Transport cost Capacity factor Non-commercial cost Import levies Life span Working capital, etc. Total installed On site Factory gate LCOE equipment equipment cost LCOE Levelised cost of electricity (Discounted lefetime cost divided by discounted lifetime generation) Source: IRENA. 27

28 RENEWABLE POWER GENERATION COSTS There are a number of important points to The analysis is designed to inform policy makers remember when interpreting the data presented and decision makers about the recent trends in the in this report: relative costs and competitiveness of renewables. It therefore excludes the impact of government • MW The analysis is for utility-scale projects (>1 incentives, or financial support for renewables. The for solar PV, >5 W for onshore wind, >50 W M M analysis also excludes any system balancing costs, MW for offshore wind), unless for CSP and >200 or benefits associated with variable renewables, explicitly mentioned. Projects below these size and any system-wide cost savings from the merit levels may have higher costs than those quoted 1 order effect. in this report. Furthermore, the analysis does not take into • All cost data in this report from the IRENA account any CO pricing, or the benefits of 2 Renewable Cost Database refers to the year in renewables in reducing other externalities, such which the project was commissioned, unless as reduced local air pollution or contamination of explicitly mentioned otherwise. For data from the natural environment. Similarly, the benefits of the Auction Database, a standard assumption renewables being insulated from volatile fossil fuel of technology for the time from auction prices have not been quantified. These issues are announcement to commissioning is used, important, but are covered by other programmes unless a specific date is available. of work at IRENA. • All data are in real 2016 USD – that is to say, it is The starting point for the analysis presented in this corrected for inflation. report is the IRENA Renewable Cost Database. • This contains information on the installed costs, When average data is presented, it consists capacity factors and LCOEs of over 15 0 00 utility- of weighted averages based on new capacity scale renewable power generation projects around deployed in that year unless explicitly stated the world. This project-level data covers around otherwise. half of all installed renewable power generation • Data for costs and performance for 2017 is capacity, but where data gaps for an individual preliminary and subject to change. Revisions technology in an individual year and country exist, are almost certain for most countries and national secondary sources of data are used to technologies as additional data is reported. ensure a comprehensive result. • Cost data in the IRENA Renewable Cost In addition to calculated LCOEs based on project Database used for calculating LCOEs excludes level data, IRENA has also collected data from any financial support by governments (national auction results to complement the LCOE data. or subnational) to support the deployment They are not necessarily directly comparable to of renewables, or to correct for non-priced LCOE values, given that key assumptions relative to externalities. their calculation will differ (e.g., the remuneration • period, cost of capital, project specific operations The raw data in the IRENA Auctions Database and maintenance costs, etc.). The database includes the impact of financial support policies contains auction results for almost 6 0 00 auctions/ that reduce the price required by a project projects and complements the project database, developer to make its expected rate of return while also providing forward-looking indicators (e.g. it includes the impacts of tax credits in of future commissioned project costs, with the the United States or other favourable taxation caveat already mentioned regarding the potential treatment). difference between LCOE and auction prices. 1. T he merit order effect, is the impact zero marginal cost renewables have on lowering wholesale electricity market prices by displacing higher marginal cost plant (typically fossil fuel-fired). 28

29 2017 • providing ancillary grid services. This is not typically The impact of grid constraints and curtailment the case for stand-alone variable renewable is not accounted for in this analysis. This is a technologies, but improved technology for solar market issue beyond the scope of this report. and wind technologies is making these more grid • The WACC is fixed over the period examined in friendly. Hybrid power plants, with storage, or this report. other renewable power generation technologies, • plus the creation of “virtual” power plants that The LCOE of solar and wind power technologies mix generating technologies, can all transform is strongly influenced by resource quality; the nature of variable renewable technologies into higher LCOEs don’t necessarily mean inefficient more stable and predictable generators. capital cost structures. Thus, although LCOE is a useful metric as a starting It should be clear from this presentation that point for deeper comparison, it is not necessarily given the complexities involved in collecting and the most useful indicator of cost between different reporting the cost data presented in this report, power generation technologies. Nor is the LCOE care should be taken in interpreting the results. necessarily the most useful tool in identifying the As already mentioned, different cost measures optimal role of each renewable power generation provide different information. These measures technology in a country’s energy mix, over the therefore need to be considered in the context medium- to long-term. Over the year, electricity of what question is being asked. For instance, systems need a balance of resources to meet comparing the installed cost of an individual overall demand and minute-by-minute variation, in technology across different markets can highlight the most economic way. To meet peaks, a system cost differentials, but not identify the causes may therefore simultaneously need to add a base of these variations. Higher costs in one market low-cost source of electricity at the same time as do not necessarily imply cost “inefficiency”, but needing a plant that will only run for a few hundred may be due to structural factors, such as greater hours each year, at costs perhaps four or even ten distances to transmission networks, or higher times higher, in LCOE terms. This is would be the material and labour costs. Only a detailed country- lowest cost solution to minimising the average cost specific analysis, supported by very detailed of electricity over the year. This highlights not only cost breakdowns, can hope to provide fuller the importance of system modelling in capacity explanations for cost variance. expansion, but also the critical importance of using the correct input assumptions for different Similarly, although the LCOE is a useful metric for cost metrics that are provided in this report. The a first-order comparison of the competitiveness cost data in this report represent the building of projects, it is a static indicator that does not blocks for a robust, dynamic modelling of the take into account interactions between generators electricity system that can take into account all the in the market. The LCOE does not take into specificities of demand and the network, as well as account either that the profile of generation of the existing generators’ costs. This report provides a technology may mean that its value is higher a robust dataset that includes current, as well as or lower than the average market price it might near-future costs of renewable power generation receive. As an example, CSP with thermal energy technologies. These can be used in dynamic storage has the flexibility to target output in high energy sector models to ensure that the many cost periods of the electricity market, irrespective complexities of operating an electricity grid are of whether the sun is shining. The LCOE also fails adequately assessed in determining the potential to take into account other potential sources of future role of renewables. revenue or costs. For example, hydropower can earn significant revenue in some markets from 29

30 RENEWABLE POWER GENERATION COSTS for commissioning between 2018 and 2025 (not T 1.3 HE IRENA RENEWABLE COST shown in Figure 1.2). The database contains data DATA BA S E on hydropower projects going back to 1961, with The data presented in this report is predominantly significant numbers of onshore wind projects from drawn from the IRENA Renewable Cost Database 2004 and solar PV from 2008. 2 and IRENA Auctions Database. The IRENA The IRENA Auction Database includes a number Renewable Cost Database contains the project level of projects that overlap with the main IRENA details for almost 15 000 utility-scale renewable Renewable Cost Database, so the totals are not power generation projects around the world, from additive. The Auctions Database contains a total large GW-scale hydropower projects to small solar G W of projects around the world. Of this, of 293 PV projects (those down to 1 W). The database M 92 GW (32% of the total) of the projects are in Brazil, also covers small-scale rooftop solar PV in the G W of the projects are in the United States (26%), 78 residential sector and larger rooftop systems in 48 G G W in W in India (16%), 10 GW in Chile (3%), 6.5 W category) M the commercial sector (in the sub-1 e G W (2%) ach), 5.5 Argentina and South Africa (2% with aggregate results derived from over one in the United Kingdom and around 5 GW (2% each) million installed systems amongst Organisation for in China and Germany. In terms of technologies, Economic Co-operation and Development (OECD) onshore wind is the largest contributor, with data member states. G W (39%) to date. The for projects totalling 114 The data available by project varies, but always next largest contributors are solar PV with 85 GW contains the total installed costs and lifetime G W (15%), biomass and (29%), hydropower, with 44 3 capacity factor. The IRENA Auctions Database offshore wind with 9 GW (3%) each, CSP with 4 GW tracks the results of competitive procurement of GW. Where fossil fuels and geothermal with 0.1 renewable power generation capacity, as well as have also been auctioned, this data has also been other power purchase agreements (PPAs) that collected and the database contains 28 GW of fossil are in the public domain. The Auctions Database fuel-fired projects. contains information on successful individual In this report, where auction data is compared to projects, or bundled projects when results are LCOE data, auction prices are corrected for the not individually disclosed, including information impact of financial support that directly reduces on the project, technology, price of winning bids, the price required by project developers (e.g. the currency for payment, remuneration period and wind production tax credit in the United States) or indexation. Not all this information is available from the data is excluded from the discussion where an all auctions, but the maximum detail available has accurate correction is not easily calculated. been collected. The Auctions Database currently contains auctions results for around 7 0 00 Given that the data for 2017 is still coming in and 4 projects. that a full and robust assessment of cost trends for 2017 is not yet possible for all technologies Figure 2 presents an overview of the two data - and all individual countries, data for 2017 is only bases. The IRENA Renewable Costing Database’s presented at a global level for each technology. nearly 15 0 G 0 17 00 projects account for 1 W of Where IRENA has assessed that the data available capacity, or around half of all renewable power for 2017 is already representative at a country level generation capacity installed up to the end of and unlikely to be significantly revised as new data 2016. In addition to these already commissioned becomes available, however, country or regional projects, the database also contains an additional 5 level data is also provided. 37 G W of as-yet unrealised project proposals 2. T his database includes results from a range of competitive procurement processes, including auctions, tenders, power purchase agreements (PPA), contracts for differences, etc. For simplicity, and given auctions are the dominant competitive procurement process, the database has been called the “Auctions Database”. 3. Projects without even this basic level of data are not included in the main database. 5. G iven that final deployment numbers for each technology by country and region for 2017 were not available at the time of this analysis, this is by necessity a qualitative judgement by IRENA based on current expectations for deployment in 2017. I 4. n some cases where there are multiple individual winners that are not disclosed, the database entry is not a single project, but the average result. 30

31 2017 renewable cost database Distribution of projects by technology and country in the IRENA Figure 1.2 and auctions database. Number Number GW of projects GW of projects IRENA Auctions IRENA Renewable 7 057298 14 8871 017 Database Cost Database In azil Br dia India Br azil United States China 7% 9% 31%26%16% 38% 400 100 0 1 0001 200 0200400600 200 800 300 GW GW 681 Number of projects 1 833 Number of projects 100 2 000 4 000 5 500 Onshore wind Onshore wind Hydro Solar PV Hydro 39% 15% 30% 56% 26% 100 0 200300400 02004006008001 0001 200 GW GW Based on IRENA data 31

32

33 2. OST TRENDS C IN GLOBAL RENEWABLE POWER GENERATION products that best suit local market and s deployment of renewable power generation resource conditions. technologies accelerates, a continuous and A relentless improvement in their competitiveness • Technology improvements that are raising has also been maintained throughout 2016 and capacity factors and/or reducing installed 2017. This has led to the fact that in virtually costs. every region of the world, bioenergy for power, • hydropower, geothermal and onshore wind Experienced project developers that projects commissioned in 2016 and 2017 largely have standardised approaches to project fell within the range of fossil fuel-fired electricity development and who have minimised project generation costs. development risks. • With very rapid reductions in solar PV module Optimised O&M practices and the use of and balance of system costs, utility-scale solar real-time data to allow improved predictive PV is now increasingly competing head-to-head maintenance, reducing O&M costs and with alternatives – and without financial support. generation loss from planned and unplanned Offshore wind power and CSP, despite having outages. significantly lower installed capacity compared to • Low barriers to entry and a plethora of other renewable technologies, have also seen their experienced medium- to large-scale developers costs fall, with auction results in 2016 and 2017 competing to develop projects, worldwide. indicating that they too are on track to achieve • cost competitiveness for projects commissioned Falling or low cost of capital, driven by between 2020 and 2022. supportive policy frameworks, project de- risking tools and the technological maturity of These cost reductions are being driven by: renewable power generation technologies. • Increasing economies of scale in manufacturing, All of this has been taking place against a vertical integration and consolidation among backdrop of increasing competitive pressure that manufacturers. is driving innovation in technology, but also in • business models. With the newer solar and wind Manufacturing process improvements that technologies benefiting from support policies, reduce material and labour needs, while there has been a steady – and sometimes dramatic optimising the utilisation of capital. – increase in their deployment in the last 10 years. • More competitive, global supply chains that This has been accompanied by growth in the are increasingly optimised to provide tailored number of markets for solar and wind. 33

34 RENEWABLE POWER GENERATION COSTS The period 2010-2017 saw the global-weighted The global weighted-average LCOE of utility-scale average cost of electricity from onshore wind fall solar PV projects commissioned in 2017 was 73% 1 by 23%. Indeed, wind power has experienced a lower than those commissioned in 2010 (Figure 2.1). somewhat unnoticed revolution since 2008-09 This was driven by an 81% reduction in solar PV as wind turbine prices have declined. Between module prices since the end of 2009, with learning 2 2008 and 2015, a virtuous cycle of improved rates for solar PV modules in the range of 18-22%, turbine technologies, as well as higher hub or even higher, if only more recent deployment is heights and longer blades with larger swept taken into account (Theologitis & Masson, 2015). areas, has increased capacity factors for a given Balance of system costs have also fallen, but not wind resource. As a result, the global weighted to the same extent, meaning the global weighted- average capacity factor for newly commissioned average total installed costs of newly commissioned projects increased from an average of 27% in 2010 projects fell by 68% between 2010 and 2017. Figure 2.1 Global levelised cost of electricity from utility-scale renewable power generation technologies, 2010-2017 Onshore Concentrating BiomassGeothermalHydroSolar Oshore wind wind solar power photovoltaic 0.4 0.36 0.33 0.3 0.22 0.2 Fossil fuel cost range 0.17 2016 USD/kWh 0.14 0.1 0.10 0.08 0.07 0.07 0.07 0.06 0.05 0.05 0.04 20102017201020172010201720102017201020172010201720102017 300 1 ≥ 100200 ≥ Capacity (MW) Source: IRENA Renewable Cost Database. Note: Each circle represents an individual project in the IRENA Renewable Cost Database, with the centre of the circle representing the LCOE value on the Y-axis and the diameter of the circle the size of the project. The lines represent the global weighted average LCOE value for a given years newly commissioned projects, where the weighting is based on capacity deployed by country/year. 1. All cost data in this chapter, unless explicitly mentioned otherwise, is from the IRENA Renewable Cost Database or Auctions Database. All references to a specific year for equipment costs, total installed costs, capacity factors or LCOE refer to the data associated with newly commissioned projects (e.g. new additions only) in that year unless explicitly stated otherwise. L earning rates for technologies are the average percentage cost or price reduction that occurs for every doubling in cumulative 2. installed capacity of that technology. 34

35 2017 processing plants. In such cases, biomass power to 30% in 2017, with many countries experiencing generation projects can produce electricity for as much more dramatic increases than the average. 0.03/kWh, when waste heat is also little as USD Installed cost reductions have been driven by used for productive purposes in combined heat and declines in wind turbine prices which, which fell by power plants (CHP). The global weighted-average between 39-58%, depending on the market, from LCOE for biomass-fired power generation projects their peaks in 2007-2010. The balance of project fell slightly between 2010 and 2017 to just below costs for onshore wind have also declined, with th th USD 0.07/kWh. The 5 these factors all driving down the LCOE of wind and 95 percentiles for and spurring increased deployment. projects have typically ranged between USD 05 0. 0.13/kWh. However, deployment is quite and USD Hydropower has historically produced some thin and this varies significantly by year. of the lowest-cost electricity of any generation technology – and continues to do so, where By the end of 2016, geothermal global cumulative untapped economic resources remain. The LCOE installed capacity was still relatively modest of large-scale hydro projects at excellent sites can at 12.6 GW and was surpassed in installed be as low as USD 0.02/kWh, with the majority of capacity terms by offshore wind in that year. 0.10/kWh. projects falling between this and USD Geothermal electricity generation is a mature, Schemes for electrification in remote areas can see baseload generation technology that can higher costs, however. A shift to more challenging provide very competitive electricity where high- projects with higher civil engineering and project quality resources are well-defined. The LCOE development costs has pushed up the global of conventional geothermal power varies from weighted average total installed cost in recent USD 0.04/kWh to around USD 0.13/kWh for years. This has in turn driven up the global weighted recent projects. average cost of electricity for hydropower, with this Offshore wind and CSP had cumulative installed 0. 0. 046/kWh 036/kWh to USD rising from USD G capacity at the end of 2016 of around 14 W and between 2010 and 2017. 5 W respectively, and have higher costs than the G Small-scale hydropower can also be very economic, other more mature technologies. Costs are falling, although typically it has higher costs than large however, and between 2010 and 2017 the cost of scale and is sometimes more suitable as an option electricity of newly commissioned CSP projects fell for electrification, providing lower-cost electricity by 33% to USD 0.22/kWh and those for offshore to remote communities, or for the local grid. 0.14/kWh. The years 2016 and wind by 13% to USD 2017 saw a breakthrough for both technologies, with Biomass-generated electricity can be very auction results for projects to be commissioned competitive where low-cost feedstocks are from around 2020 onwards anticipated to have available onsite at industrial, forestry or agricultural significantly lower LCOEs than in 2017. 35

36 RENEWABLE POWER GENERATION COSTS HE NEW COST REDUCTION DRIVERS: 2.1 T Renewables are experiencing a virtuous Figure 2.2 COMPETITIVE PROCUREMENT, cycle of technology improvement INTERNATIONAL COMPETITION and cost reduction AND IMPROVED TECHNOLOGY The power sector is currently undergoing a transformation that represents the beginning of the transition to a renewables-dominated, truly sustainable power sector. This is required in Policy Cost order to avoid the dangerous effects of climate Reductions Support change. The power generation sector is leading this transformation, with renewables estimated to have added around half or more of global new capacity required every year, from 2012 onwards (IRENA, 2017d). At the end of 2001, the total global capacity of solar PV was less than 1 GW; by end Technology of 2016, it had surpassed 291 W and by the end G Improvements W. G of 2017 should have grown to around 381-386 Similarly, wind power capacity at the end of 2001 was 24 G W, but by the end of 2016 had reached 467 W. Meanwhile, annual new capacity additions G of renewable power generation technologies GW in 2001 to 167 GW in 2016, a increased from 16 Yet it has not just grown, but also experienced ten-fold increase, with total new capacity additions a welcome broadening in geographical scope. in 2017 likely to surpass this record. In some cases, this has been accompanied by slowing or stagnant markets for new projects in The virtuous cycle of long-term support policies mature markets (notably Europe), resulting in a accelerating the deployment of renewables – large number of very experienced medium- and which in turn leads to technology improvements large-scale developers now increasingly looking and cost reductions (Figure 2.2) – has led to the for international opportunities. increased scale and competitiveness of markets for renewable technologies. The transformation of This confluence of factors has been driving recent the power generation sector is therefore an active cost reduction trends for renewables, with effects one, where the policy support for renewables that will only grow in magnitude in 2018 and to meet countries’ long-term goals for secure, beyond. The three main emerging drivers that are reliable, environmentally friendly and affordable starting to increasingly drive cost reductions are: energy is bearing fruit. • Competitive procurement of renewable power As equipment costs for solar and wind power generation : As renewable power generation technologies have fallen, notably for solar PV technologies have matured and cost reductions modules and onshore wind turbines, a shift have exceeded expectations, there is a growing in emphasis in cost reduction drivers is also shift towards auctions and other competitive emerging. As equipment costs fall, the importance procurement processes (IRENA, 2017e). In of addressing balance of system costs, improving mature markets, with limited volumes on the performance of the technologies, reducing offer, this has led to intense competition for O&M costs and driving down the cost of capital projects and has resulted in falling costs. all start to take on greater importance (IRENA, Similarly, reduced support levels have also 2016a). At the same time, markets and business forced developers to implement best practices models are not standing still. In recent years, as the in terms of project development, utilise newer compelling case for renewable power generation’s innovative technology solutions, and generally competitiveness has grown, so too has deployment. reduce margins. 36

37 2017 • architectures with greater efficiency are Increasing international competition for projects : helping to reduce module installed costs and With the sustained growth in renewable power balance of system components. These are but generation deployment, a large number of very a few examples of the constant innovation that experienced medium- to large-scale project is helping to drive down costs. developers have emerged around the world. Many have seen their original markets slow These trends are not new, but their importance and have looked to new markets to maintain a has grown significantly in recent years. They pipeline of projects and grow their businesses. are part of a larger dynamic across the power This has allowed new markets to benefit from generation sector, driven by the fact that in many previous, hard-won business acumen in the regions of the world, renewable power generation field of renewable project development. In technologies often offer the lowest cost source conjunction with local partners, in many cases, of new power generation. The industry is thus to help navigate the local regulatory and rapidly transitioning. In the past, typically, there business landscape; these project developers was a framework offering direct financial support, are enabling even new markets to achieve very often tailored to individual technologies (e.g., solar competitive pricing, where the regulatory and PV) and even segments (e.g., varying support for policy framework is conducive to renewables. residential, commercial and utility-scale sectors, • sometimes differentiated by other factors such as : As Continuous technology innovation whether they are building-integrated or not). Now, economies of scale in manufacturing and this is being replaced by a favourable regulatory and institutional framework that sets the stage for competitive procurement of renewable Technology providers and power generation to meet countries energy, project developers have environmental and development policy goals. reduced costs to remain In many parts of the world, utilities, industry players, project developers and asset owners have competitive rapidly embraced this new dynamic and are finding ways to profitably navigate this new landscape. In the absence of direct financial support, project materials efficiency have been unlocked developers are also using new business models in recent years, continued cost reductions to grow. Companies are identifying strategies that are beginning to be more heavily driven by will allow subsidy-free projects to be profitable improvements in technology. This is particularly in different markets. Examples of this range true for wind, where larger turbines with larger from utilising corporate or utility PPAs to provide swept areas are harvesting more electricity revenue certainty, or merchant solar PV plants for the same resource. Larger turbines also being built in certain locations where wholesale enable the amortising of project development market forecasts support their economics. Other costs over greater capacities and allow greater examples include looking at new opportunities, economies of scale in O&M. At the same such as also including storage to better access time, wind turbine manufacturers are offering peak prices and potentially achieve new revenue an increasing range of products to allow streams by providing ancillary services to the grid. optimisation for individual wind sites, while the This section will now examine their impact on utilisation of real-time data and “big data” to recent cost trends, according to each technology, enhance predictive maintenance and reduce through 2017 and beyond, using data both from O&M costs and lost energy from downtime the IRENA Renewable Cost Database and the are also playing a role. For solar PV modules, Auctions Database. the continued efforts to commercialise cell 37

38 RENEWABLE POWER GENERATION COSTS A Cautionary Tale: When is an LCOE not a FiT or a PPA Price? Box 1 The LCOE metric used in this report represents an indicator of the price of electricity required for a project in which revenues would equal costs over the life of an asset. This includes making a return on the capital invested equal to the discount rate, while excluding the impact of existing government incentives or financial support mecha - nisms. For solar and wind technologies in particular, various power purchase agreement (PPA) prices have been announced recently in different locations. With such developments, it can become harder to distinguish between these “record” prices and the LCOE concept as discussed in this report. Though these very low PPA prices point to the increasing competitiveness of renewable energy sources compared to fossil fuel alternatives, they often cannot be directly compared to the LCOE, nor necessarily to feed-in tariffs - (FiTs). The end auction or PPA prices depend on a set of obligations and contract-defined terms that are very de pendent on the specific market situation of the project setting. Assumptions made to calculate these prices usually differ from the more standardised ones used for the LCOE indicator calculations in this report. There is also the chance that if these conditions are not fulfilled, the PPA price may not materialise – if, for example, the independent - power producer (IPP) does not fulfil the output requirements or electricity quality. In extreme cases, the deficien cies in the initial winning bid may see a developer walk away from the project, as the financial penalties incurred are lower than the expected loss if the project is completed. As an example of the potential differences between auction and PPA prices compare to LCOEs, in 2015 a United States solar PV developer agreed to sell power at a record low headline price of USD 0.0387/kWh from a 100 MW solar plant to utility NV Energy. It was not widely quoted, however, that this price included a 3% escalation clause and that according to a filing with the Public Utilities Commission of Nevada, the LCOE of the project was estimated at about USD 0.047/kWh after the Investment Tax Credit (Public Utilities Commission of Nevada, 2015). Allowing 0.066/kWh (70% higher for the impact of the 30% Investment Tax Credit raises the electricity price to around USD than the headline value). In the case of FiTs, they are also not directly comparable to the PPA contract set prices. For instance, in Germany the current FiT for solar PV is nominal and payable for a period of 20 years, below the economic life of 25 years. The starting point for any comparison of an LCOE metric against a FiT or PPA price should therefore be one that assumes they are not directly comparable. The exception would be one where the weighted average cost of capital (WACC) of a project equals that assumed for the LCOE calculation, the remuneration period equals the economic life of the asset, the remuneration is “complete” in terms of the fact no other revenue streams are available (e.g. potential revenue from green certificates or capacity payments that are not included in the headline remuneration figure), and that remuneration is indexed to inflation. It should therefore be clear that a lower PPA price than the LCOE may not necessarily represent a lower cost of project. Care should thus be taken in comparing LCOE, FiT levels and auction/PPA prices, as they can be very different cost metrics. Tracking Innovation trends: A look at patent data for renewables Box 2 The past decade has seen robust growth of innovation and inventions for renewable energy technologies. Patents are an important mechanism to foster such innovation. They support revenue generation (through licences), en - courage partnerships, and can create market advantages while balancing the interests of inventors and the general public (IRENA, 2013c). Reliable patent data provides a means to track renewable energy innovation worldwide, heightening the key role of patents in the technology life cycle and new technology uptake. In order to facilitate such global tracking, IRENA has developed a web-based tool, INSPIRE (www.irena.org/inspire), that facilitates such global tracking and helps to assess trends in research, development and demonstration. 38

39 2017 The tool aims to support technology and innovation strategies among IRENA’s Member States by furnishing com - prehensive, reliable, regularly updated information on renewable energy patents and technical standards. Such information facilitates standardisation, quality management and technology risk reduction as countries pursue the transition to renewables. In developing the INSPIRE platform, IRENA worked closely with the European Patent Office to shed light on trends in climate-change mitigation technologies, as reflected in recent renewable energy patents. As the resulting data showed, the total number of renewable energy patents filed worldwide at least tripled between 2006 and 2016. This represents compound annual growth of 17%, with more than half a million patents filed for these technologies by the end of 2016 (www.irena.org/inspire). Along with the intensification of inventive activity, renewable energy has achieved sharp cost reductions and sustained deployment growth. For example, solar PV-related patent filings reached 183 00 while cumulative deployment for the technology barely exceeded 290 GW. A more mature tech - 0 nology, hydropower, had a more modest 36 000 patent filings, despite its much higher cumulative deployment of 250 GW (Figure B2.1). about 1 Figure B2.1 Development of patent data for renewable energy technologies, 2010-2016 1 400 550 500 2016 1 200 450 2010 1 000 400 2006 350 800 300 GW 600 250 2016 200 400 Number of patents (thousands) 150 2016 200 100 2010 2016 2006 2010 2010 50 2016 2006 2006 2006 0 2016 2016 2010 0 50 200 0 100 150 20062016 Number of patents (thousands) Geothermal Ocean Solar PV Hydro (with pumped storage) Bioenergy Solar thermal Wind energy Solar PV Based on INSPIRE web platform (www.irena.org) and IRENA (2017a). Solar PV held the largest share of patents among all renewable energy technologies at the end of 2016, following a five-fold increase – also the fastest patent growth in renewables – since 2006. Solar PV and solar thermal technol- ogies together account for more than half of patents filed, while wind patents contribute another fifth of the total, and bioenergy just over one sixth, followed by hydropower and other technologies with smaller shares. 39

40 RENEWABLE POWER GENERATION COSTS Asia stands out as a region with particularly R 2.2 ENEWABLE ELECTRICITY COST competitive average costs across all of the TRENDS BY REGION AND TECHNOLOGY technologies. This is due to a mixture of excellent Figure 2.3 highlights the regional weighted average resource endowment and lower than average LCOE by technology for an average of 2016 and installed costs, notably for solar PV and onshore 2017 to ensure maximum representativeness for wind in China and India, which dominate all technologies and regions. While the range deployment in the region. of these projects’ individual electricity costs In solar PV, what has been truly remarkable is spans around these points, the chart serves to that rapid declines in module prices and installed highlight just how competitive renewable power costs have resulted in an increasing number of generation technologies have become. For regions having weighted average LCOEs that bioenergy, geothermal, hydropower and onshore are increasingly competitive at the utility-scale, wind all regions with meaningful deployment without financial support. These projects now have weighted averages within the range of fossil fall within the fossil fuel-fired cost range. This is fuel-fired power generation costs. Only CSP, solar a truly impressive transition, given that in 2010 PV and offshore wind still see weighted averages the regional weighted average LCOE of solar PV by region outside the fossil fuel-fired cost range. projects ranged from 65% higher than the upper As will become apparent when examining the range of fossil fuel-fired costs in North America, LCOE data and the impact of auction results on to 236% higher in Africa – albeit where expensive upcoming project costs, however, this will very projects in more remote areas had raised costs. The soon be a thing of the past. weighted average LCOE by region for utility-scale gional weighted average levelised cost of electricity by renewable power generation technology, Re Figure 2.3 2016 and 2017 0.2 0.1 2016 USD/kWh Fossil fuel cost range 0.0 Concentrating Solar BiomassGeothermalHydro O‚shore windOnshore wind solar power photovoltaic Middle East Africa North America Oceania Asia Eurasia Europe Central America and the Caribbean South America Source: IRENA Renewable Cost Database. 40

41 2017 are typically economically supportable, as the solar PV projects that were installed in 2016 and savings in diesel costs and, sometimes, improving 2017 ranged from a low of around USD 09/kWh in 0. electricity network reliability, make the projects 0.17/kWh in Eurasia. In Central Asia to a high of USD economic. America, the Caribbean and South America the 13/kWh. Projects are now being 0. average was USD Capacity factors for utility-scale solar PV projects built with an LCOE of as low as USD 05/kWh, and 0. have been edging higher through time, with a as presented in Figure 2.1, with the costs continuing global weighted average increase of 28%, from to fall, the global weighted average for 2017 alone 14% on average in 2010 to 18% in 2017. This is has fallen to USD 0.10/kWh. While even lower predominantly due to a shift in deployment to values are going to be seen in the coming years, as areas with better solar resources, rather than as a the record breaking auction results in Dubai, Chile, result of an increase in the use of tracking or other Abu Dhabi, Mexico and Saudi Arabia come online. technology improvements. There have been some 0.03/kWh or lower. These are all at around USD system performance improvements in this time as well, notably in terms of improving the overall Focussing on the global weighted average trends efficiency of the array and inverters to reduce for new utility-scale solar PV projects by year losses, but these are minor contributors to the (Figure 7), the LCOE reduction of 73% between overall improvement. 2010 and 2017 is put in context. By far the main driver has been the reduction in total installed The overall result of the contribution of these costs for utility-scale solar PV, with a 68% two factors playing out at a project level was reduction in total installed costs between 2010 the dramatic fall in LCOE of utility-scale solar PV and 2017. But there has not just been a reduction between 2010 and 2017. Within this, two distinct in average costs that has been significant, there periods are visible: between 2010 and 2013, the has also been a shift in the distribution of projects global weighted average LCOE fell by around 20% around the weighted average that has occurred as each year. After 2014, when the decline was 10%, the weighted average has shifted to the lower end the fall was more variable, as 2015 saw a 20% th h of the 5t and 95 percentile ranges. decline, 2016 a 10% reduction, and 2017 a 17% decline. The compound annual rate of decline was The global weighted average total installed cost 21% per year for 2010-2013 and 14% per year for for utility-scale solar PV fell from USD 94/kW 3 4 th 2013-2017. 1 388/kW in 2017, with a 5 in 2010 to USD and th 95 3 percentile of USD 754/ 898/kW and USD Hydropower produces some of the lowest-cost kW. The distribution of project costs for solar PV electricity of any generation technology and is the remains wide and is skewed towards a long tail of largest source of renewable electricity generation more expensive projects. In part, this reflects the today (3 TWh in 2015). The LCOE of large-scale 996 natural variation in project costs for renewable hydro projects at excellent sites can be as low as projects; however, there are two other significant 02/kWh, while average costs have risen 0. USD drivers. The first is that there remain a number in recent years and in 2016 the global weighted of markets with persistently higher costs than in average reached USD 0. 053/kWh. In 2017, it fell back other markets, with the United States and Japan 0.047/kWh. Developments in Asia, where to USD being two notable examples. Historically, though, good untapped economic resources still remain, this has also been the case for new markets that saw weighted average LCOEs of USD 0.04/kWh in have yet to establish mature and competitive local 2016-2017, with South America having a weighted supply chains and developers. Secondly, solar PV is 05/kWh and North America average of USD 0. extremely modular and is often increasingly being USD 06/kWh. Africa, Eurasia and the Middle 0. deployed in remote locations (e.g., in the interior 0. 07/kWh. Developments were East averaged USD of African countries, islands, or other isolated somewhat more expensive in Central America and locations), where logistical costs are significantly 10/kWh, and in Europe, at 0. the Caribbean, at USD higher than in areas close to ports and with USD 0.12/kWh. supporting infrastructure. Here, the higher costs 41

42 RENEWABLE POWER GENERATION COSTS Figure 2.4 Global weighted average total installed costs, capacity factors and LCOE for solar PV , 2010-2017 Total installed cost Levelised cost of electricity Capacity factor 5 500 0.40 0.4 0.36 000 5 th 95 0.35 percentile 4 394 500 4 0.30 0.3 4 000 0.28 3 663 500 3 0.25 0.22 3 066 3 000 0.20 0.2 0.18 2 500 2 424 0.18 0.17 0.16 2 224 0.16 2016 USD/kW Capacity factor 0.17 0.16 2016 USD/kWh 0.14 0.15 0.16 000 2 1 749 0.13 0.14 0.12 500 1 510 1 0.10 0.1 1 388 0.10 000 1 th 5 percentile 0.05 500 0 0.00 0.0 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. as the weighted average LCOE fell 14% in 2017 (to As deployment has accelerated in regions USD 0.046/kWh), compared to 2016. that have previously had significant untapped potential, notably in Asia and South America, Small-scale hydropower can also be very economic, recent development has had to start depending although typically it has higher costs than large- on projects at more challenging sites, with higher scale projects. This is partly due to economies of project development costs and civil engineering scale, but is often because it is being deployed in costs, either due to conditions at the dam location, remote areas, as it can provide low-cost electricity or in terms of more expensive infrastructure and to isolated communities or locations. logistics for the project. This means that projects’ Onshore wind now rivals hydropower, geothermal total installed costs have started to rise (Figure 2.5). and biomass as a source of low-cost electricity, To some extent, this was offset by an increase without financial support. Capacity factors have in the weighted average project capacity factor, increased as performance has improved, installed which went up from around 44% for projects in costs have fallen and O&M costs have reduced 2010 to 50-51% in 2014-2016, although in 2017 this all serving to drive down the LCOE. The global fell back to 48%. weighted average LCOE for onshore wind fell In terms of LCOE, projects in Asia and South by 22% between 2010-2017 and is now around America are clearly moving up the cost curve as 0. 06/kWh (Figure 2.6). The weighted USD deployment continues. Hydropower remains one average regional LCOE of onshore wind has of the most competitive sources of new electricity, also narrowed in recent years. In 2016/17 Asia, however, and significant untapped potential still Eurasia, North America and South America all remains for sustainable hydropower development, 06/kWh or less, while the 0. averaged around USD notably in Africa, but also in Asia and the Americas. weighted average was USD 0. 08/kWh in Europe The global weighted average LCOE of hydropower 09/kWh in the Middle East 0. and Oceania, USD projects increased from an average of around and Africa, and USD 10/kWh in Central America 0. USD 04/kWh in 2010 to USD 05/kWh in 2016 0. 0. and the Caribbean. Where excellent resources and and 2017, with a decline between 2016 and 2017 42

43 2017 Global weighted average total installed costs, capacity factors and LCOE Figure 2.5 for hydropower, 2010-2017 Levelised cost of electricity Capacity factor Total installed cost 0.15 500 5 1.0 5 000 0.9 500 4 th 95 0.8 percentile 000 4 0.7 0.10 500 3 0.6 0.51 000 3 0.50 0.51 0.5 0.46 0.50 2 500 0.48 0.46 0.44 0.4 2016 USD/kW Capacity factor 2016 USD/kWh 0.05 000 2 1 780 0.05 0.04 521 1 0.3 1 427 0.040.04 0.05 500 1 0.03 208 1 0.04 1 535 1 535 0.04 0.2 1 233 000 1 171 1 th 5 percentile 0.1 500 0 0.00 0.0 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. Biomass-generated electricity can be very low-cost structures exist, wind power projects are competitive where low-cost feedstocks are now routinely achieving costs of just USD 0.04/ available onsite at industrial, forestry or agricultural kWh, without any financial support, and in some processing plants. In such cases, biomass power 0.03/kWh. currently exceptional cases, USD th th generation projects can produce electricity for as The 5 percentile range for the LCOE of and 95 little as USD 0.06/kWh in the OECD countries, and newly commissioned onshore wind projects was 03/kWh in developing countries. 0. as low as USD 12/kWh in 2017, 04 and USD 0. 0. between USD The typical LCOE range for biomass-fired power which is wider than in 2010 as new markets have generation projects is between USD 0.04 and developed broadening the deployment of onshore USD 0.19/kWh, but can fall outside that range wind from traditional markets. for some projects. The weighted average LCOE Globally, onshore wind total installed costs fell by by region in 2016/17 varied from a low of around an average of 20% between 2010 and 2017, notably 0.06/kWh 0.05/kWh South America to USD USD as deployment in China and India grew, given their in Asia, and to between USD 0. 0. 11/kWh 07 to USD relatively low-cost structures. The global weighted in other regions. average capacity factor increased by around 11% Deployment of new bioenergy projects for power over the same period, from 27% to 30%, conversely (and often heat generation at the same time) is being slowed by the increased share of China and smaller than for hydropower, solar PV and onshore India, which have only average resources and are wind and results in more year-to-year volatility lagging somewhat in the deployment of the latest in the characteristics of newly commissioned turbine technologies. Changes in the shares of projects. With a shift to more sophisticated, deployment by country between 2010 and 2013, bioenergy plants capable of performing with a despite total installed costs in individual countries range of heterogenous feedstocks, the global continuing to decline, combined to yield relatively weighted average total installed cost increased modest global weighted average reduction in the between 2010 and 2014 before falling in 2015 and LCOE of just 2%, before reductions of 19% between 2013 and 2017. 43

44 RENEWABLE POWER GENERATION COSTS Figure 2.6 for onshore wind, 2010-2017 Global weighted average total installed costs, capacity factors and LCOE Total installed cost Capacity factor Levelised cost of electricity 0.15 0.6 500 3 th 95 3 000 percentile 0.5 500 2 0.10 0.4 0.08 2 000 868 1 0.080.08 843 1 0.07 0.290.29 684 1 0.08 0.28 0.3 1 828 538 1 0.300.30 752 1 0.07 500 1 0.07 0.270.270.27 1 540 2016 USD/kW 0.06 1 477 Capacity factor 2016 USD/kWh 0.2 0.05 000 1 th 5 percentile 0.1 500 0 0.0 0.00 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. benefit from past experience with a geothermal 2016 (Figure 2.7). Data for 2017 is preliminary, reservoir and can not only reduce risks, but but suggests more capital intensive plants took existing infrastructure in place can reduce a larger share of deployment that year. With a engineering and grid-connection costs, as well corresponding increase in capacity factors, due as spread O&M maintenance costs over greater to the anticipated wider range of feedstocks capacity. Significantly, geothermal projects carry available at low cost, the impact on LCOE was a very different risk profile than other renewable muted, however. technologies, given that the dynamics of managing Geothermal electricity generation is a mature, geothermal reservoirs over the life of a project baseload generation technology that can 3 present some unique challenges. provide very competitive electricity where high- quality resources are well-defined. The LCOE The two main CSP systems that have been of conventional geothermal power varies from deployed commercially are parabolic trough and 0. 0. 13/kWh for recent projects. Yet 05 to USD USD solar towers. Deployment of these is still modest, the LCOE can be as low as USD 0. 04/kWh for the however, and until recently was concentrated in most competitive projects, such as those which Spain and the United States. Between 2009 and utilise excellent, well-documented resources and 2011, the LCOE of projects varied from around are brownfield developments. 0. 0. 30 to USD USD 47/kWh as generous support policies provided little incentive to drive down costs, Many recent projects have been based on well with installed costs remaining high. Since 2012, surveyed fields, helping to reduce development these have been falling, as deployment has shifted risks and keep installed costs towards the lower away from the traditional markets of Spain and end of the cost range. Brownfield projects can 3. G iven field dynamics and uncertainty about how the reservoir will react to different operating regimes, operational experience is always adding to the base of knowledge that allows for optimal reservoir management. 44

45 2017 for bioenergy for power, Global weighted average total installed costs, capacity factors and LCOE Figure 2.7 2010-2017 Levelised cost of electricity Capacity factor Total installed cost 000 10 0.20 1.0 9 000 0.9 0.86 0.8 000 8 0.75 0.75 0.73 0.15 7 000 0.7 0.66 th 95 0.69 0.67 percentile 6 000 0.6 0.61 0.5 000 5 0.10 0.080.08 0.4 000 4 2016 USD/kW Capacity factor 2016 USD/kWh 0.07 0.06 0.070.07 756 753 2 2 0.07 0.3 2 668 3 000 0.06 th 5 0.05 percentile 774 1 0.2 000 2 2 348 057 2 966 1 608 1 0.1 1 000 0 0.00 0.0 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. for geothermal power, Global weighted average total installed costs, capacity factors and LCOE Figure 2.8 2010-2017 Total installed cost Levelised cost of electricity Capacity factor 9 000 0.15 1.0 th 95 percentile 0.87 0.9 0.87 0.87 8 000 0.84 0.83 0.82 0.8 7 000 0.79 6 351 0.7 0.10 6 000 0.09 0.6 5 000 0.5 0.07 0.06 0.06 0.06 4 000 3501 0.4 2016 USD/kW Capacity factor 3 193 2016 USD/kWh 0.06 3 3663 424 0.05 3 000 0.3 2 959 0.05 2 452 2 000 0.2 th 5 percentile 1 000 0.1 0.0 0 0.00 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2015 2015 2014 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. 45

46 RENEWABLE POWER GENERATION COSTS Belgium, Denmark, Germany, the Kingdom of the the United States. Greater competitive pressures Netherlands and the United Kingdom. In addition, have reduced installed costs, with projects also China has also built some inter-tidal projects and benefitting from higher solar resources in new the United States is joining the ranks of offshore markets like Chile, Morocco and the United Arab wind power producers. 16 and Emirates. LCOEs ranged between USD 0. USD 29/kWh in 2016-2017. Recent auction results, 0. Costs for offshore wind in the early 2000s however, have heralded an acceleration in cost climbed, as deployment accelerated and projects reductions, as supply chains have become more moved into deeper waters, further offshore – competitive, a wider range of project developers raising foundation and installation expenditure. have had experience developing multiple projects Costs have since peaked, however, and have come and projects have been more often sited in regions down significantly in recent years. Nonetheless, with excellent solar resources, but still with access the weighted average LCOE by region remains to low-cost finance. These results remain to be 0.15/kWh. As with CSP, 0.14 to USD around USD confirmed by a broader set of auction results or though, the recent auction results from 2016 and project announcements beyond Australia and 2017 in Belgium, Denmark, the Kingdom of the Dubai, but the initial indications are that the Netherlands, Germany and the United Kingdom all competitiveness of CSP will fundamentally change show that offshore wind will be a very competitive for plants commissioned beyond 2020. source of new generation capacity in Europe for projects that will be commissioned in 2020 and Offshore wind, like CSP, has relatively modest beyond. Indeed, in Germany, 2 projects that will be levels of cumulative installed capacity with just commissioned in 2024 and 1 in 2025 won with bids 13 GW installed at the end of 2016. Deployment 4 that did not ask for a subsidy over market rates. has been concentrated in Europe, notably in for CSP , 2010-2017 Global weighted average total installed costs, capacity factors and LCOE Figure 2.9 Total installed cost Capacity factor Levelised cost of electricity 000 12 0.45 0.6 000 11 0.40 10 000 0.35 th 95 0.35 percentile 8 764 9 000 0.42 0.330.32 670 654 7 7 000 8 0.4 0.30 0.27 0.34 7 583 0.25 7 000 0.32 0.25 0.30 0.25 0.25 588 6 0.33 6 000 106 6 0.30 0.22 5 803 0.29 0.20 564 5 5 000 0.27 th 2016 USD/kW Capacity factor 5 2016 USD/kWh percentile 4 000 0.2 0.15 000 3 0.10 2 000 0.05 1 000 0 0.0 0.00 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. 4. For more details see: https://ore.catapult.org.uk/download/subsidy-free-offshore-wind/ 46

47 2017 Figure 2.10 for offshore wind, 2010-2017 Global weighted average total installed costs, capacity factors and LCOE Total installed cost Capacity factor Levelised cost of electricity 0.6 0.30 7 000 6 000 0.25 th 95 452 5 percentile 0.42 4 883 5 000 0.41 0.39 0.4 0.20 0.19 331 4 0.18 0.39 458 4 411 4 0.17 0.37 0.16 000 4 239 4 200 4 0.36 0.34 0.34 782 3 0.15 0.15 0.15 0.14 0.14 000 3 2016 USD/kW th Capacity factor 5 2016 USD/kWh percentile 0.2 0.10 000 2 0.05 1 000 0 0.00 0.0 2017 2012 2013 2015 2017 2017 2016 2012 2012 2013 2013 2014 2015 2015 2011 2016 2016 2010 2014 2014 2011 2011 2010 2010 Source: IRENA Renewable Cost Database. T 2.3 HE COST OF RENEWABLE wind power must also be taken into consideration ELECTRICITY TO 2020: INSIGHTS in system modelling to arrive at the least-cost FROM PROJECT DATA AND AUCTIONS combination of technologies. However, as previous IRENA analysis has highlighted, the additional The range of costs for renewable power generation environmental costs of fossil fuels and estimates technologies between regions is wide for a given of the additional costs of variability of solar and technology – and even for a given technology wind may broadly offset each other (IRENA, 2015). within a particular region, due to site-specific However, estimates of both these cost groups is cost drivers. It is striking, though, that virtually all country specific and evolving over time as a better renewable power generation technologies now understanding of the various impacts of each not only include significant numbers of projects is achieved through operational experience and which offer very competitive electricity costs, but additional research. that renewable power generation technologies are This section examines in more detail some of the also increasingly overlapping towards the low-end high-level trends that are behind the convergence of the fossil fuel-fired electricity cost range. This is in LCOE, for commissioned projects up to 2017 despite the fact that fossil fuels still do not pay for and for proposed projects up to 2020. It will look the local and global environmental damage they at all the major contributors to new capacity – cause, or their negative health impacts. Including hydropower, onshore and offshore wind, solar these costs would significantly improve the photovoltaics and CSP – and outline five key economics of renewable power generation costs, 5 messages from the data: in comparison with the figures presented here. As already discussed, the variability of solar PV and 5. F or a more detailed discussion of the costs of local and global pollutants see IRENAs analysis in “Perspectives for the energy transition: Investment needs for a low-carbon energy system” (IRENA, 2017f). 47

48 RENEWABLE POWER GENERATION COSTS • In 2017, weighted average electricity costs for - In 2017, a significant number of newly commis bioenergy for power, geothermal, hydro, onshore sioned bioenergy for power, hydropower, geo - wind and solar PV all fell within the range of thermal, onshore wind and, increasingly, solar fossil fuel-fired electricity and are often the PV projects competed head-to-head with fos - cheapest source of new generation needs. The sil-fuels without financial support. Offshore fossil fuel-fired electricity generation cost range wind and CSP projects to be commissioned in 0.05 to for G20 countries spans the range USD - the period from 2020 onwards will also com 6 USD 0.17/kWh (IRENA, forthcoming). pete in this fashion. • A remarkable convergence in the global Figure 2.11 shows the weighted average LCOE by weighted average cost of electricity from each technology and region/country grouping, as well th th as the 5 technology has been signaled to 2020 by recent percentile ranges for projects and 95 auction results. Installed cost differentials commissioned in 2016 and 2017. In China and between countries persist for onshore wind India, hydropower remains the most competitive and solar PV in particular, however, highlighting source of electricity, on average coming in below cost reduction opportunities. the lowest fossil fuel-fired option. The weighted average LCOEs for bioenergy for power and • Cost reductions for solar and wind are continuing onshore wind are only slightly higher than the at a steady pace and between 2010 and 2020 lowest fossil fuel-fired cost option, while solar PV represent remarkable rates of cost reduction, 0.08/kWh and is also has fallen to around USD significantly beating long-term forecasts. increasingly competitive. • Renewable power generation technologies In 2016/2017, in the OECD countries, onshore wind are increasingly not just competitive without was the cheapest renewable power generation financial support, but out-compete fossil fuel- option, with an average USD 0. 065/kWh. fired power. Hydropower and bioenergy for power were on • average only slightly more expensive, while solar The cost of electricity from onshore wind PV was more expensive, but still well within the and solar PV is reaching extremely low levels, range of the LCOE of fossil fuel-fired electricity. only achieved in the past by the very best In the rest of the world, a similar pattern exists, hydropower projects. 6. I n 2017 IRENA collected project level cost data for fossil fuel-fired power stations in the G20 countries, as well as data on actual capacity factors, O&M costs, operational efficiency and fuel costs, this analysis is forthcoming and will be published in 2018. 48

49 2017 Figure 2.11 ranges and weighted averages for China and India, OECD Project LCOE and rest of the world, 2016 and 2017 Rest of the World China & India OECD 0.3 0.2 Fossil fuel cost range 2016 USD/kWh 0.1 0.0 Hydro Hydro Hydro Biomass Biomass Biomass solar power solar power solar power Geothermal Geothermal Onshore wind Onshore wind Onshore wind Concentrating Concentrating Concentrating Solar photovoltaic Solar photovoltaic Solar photovoltaic Source: IRENA Renewable Cost Database. 2017), care should be taken in interpreting this close with very competitive weighted average LCOEs relationship. What is clear from the trend in auction for bioenergy for power, geothermal, hydropower results for projects that will be commissioned and, to a lesser extent, onshore wind. The weighted between 2018 and 2020 however, is that recent average solar PV LCOE remained close to the cost reductions identified from project-level data upper end of the fossil fuel-fired LCOE range. look set to continue at a steady pace. This presumes Figure 2.12 highlights the continued cost reductions that the recent relationship between the two for onshore wind and solar PV that have been datasets is maintained over this period, although as experienced. Since 2013, the weighted average can be seen, there are slight deviations in trends in LCOE trends from the IRENA Renewable Cost individual years. Yet the direction of travel is clear. If Database and Auctions Databases have followed current trends continue, in 2019 or 2020, the global a similar path and level. Given that competitive weighted average LCOE for solar PV may fall to procurement represents a relatively small below USD 0. 06/kWh, converging to slightly above percentage (10-15%) of recently commissioned 0.05/kWh. that of onshore wind at USD utility-scale onshore wind and solar PV (IEA PVPS, 49

50 RENEWABLE POWER GENERATION COSTS Global levelised cost of electricity and auction price trends for onshore wind and solar PV, 2010-2020 Figure 2.12 Onshore wind Solar PV 0.4 0.4 Auction database LCOE database 0.3 0.3 0.2 0.2 2016 USD/kWh 0.1 0.1 0.0 0.0 2017 2017 2012 2012 2013 2013 2015 2015 2018 2018 2016 2016 2019 2019 2014 2014 2011 2011 2020 2010 2010 Source: IRENA Renewable Cost Database and Auctions Database. Each circle represents an individual project or auction result, while the solid line is the capacity-weighted average from Note: each database. as all IRENA LCOE calculations are. For 39% of There are a number of caveats to a comparison the projects in the Auction Database, it was not of LCOE results and auction prices, however. The clear from the announcements if the project was two metrics are rarely equivalent and cannot indexed or not. For onshore wind, where data necessarily be compared at an individual project was available, 80% of projects were fully indexed, level. The reasons for this are manifold. Firstly, but for solar PV, this dropped to 30%, with 70% it is rare that the auction or tender terms reflect appearing not to be indexed to inflation. the same assumptions for the calculation of an LCOE. The length of remuneration may not match Other issues are that the remuneration may the economic life of the asset. For instance, in the cover only a fraction of the project’s output and IRENA Auction Database, where contract length the balance may be contracted bilaterally at an was disclosed, around 15% of the onshore wind and undisclosed value. The project may also benefit two-thirds of the solar PV projects had terms that from free land under the auction and/or share O&M matched the 25-year assumption IRENA uses for costs over a number of projects in a development their economic life. Yet this is only a partial view, zone. Another significant differentiator of prices as 60% of the onshore wind projects in the Auction can be if an existing (or to be built) grid connection Database did not have their contract length is provided to the developer, or the developer is disclosed with the announcement of the price required to construct its own. This has a significant (although this falls to 16% for solar PV projects). difference on the auction prices seen for offshore Another important issue is that the auction price wind in Denmark and the Netherlands, where recent may not be indexed to inflation, or may be partially auctions included grid connections, while in the indexed, meaning the price is not in real terms, 50

51 2017 Taking these limitations into account, though, it is United Kingdom, the project developer has to pay clear that cost reductions will continue for onshore for this work. Although these issues are also present wind and solar PV out to 2020 and beyond. Even in project-level data from the IRENA Renewable if the validity of comparing LCOE and auction Cost Database, they highlight the need to have prices for individual projects is often difficult or large volumes of data to draw robust conclusions inadvisable, the volume of data available and on trends and the dangers of comparing individual the consistent trends between the two datasets project without full knowledge of the terms and suggest that its possible to feel some confidence conditions under which it will be developed. in the overall trend. In addition, there are a number of auction design CSP and offshore wind had cumulative installed choices that can greatly affect the risk profile of capacity of just 5 GW and 13 GW respectively a project. These can include whether the winners at the end of 2016, while the cost of electricity will be remunerated in local currency or USD, or from recently commissioned projects for these if the offtake party has a government guarantee/ technologies is higher than for other renewable partial guarantee or not, amongst other factors. power generation technologies. Yet costs are The final complication is that the LCOE calculation coming down. For both technologies, 2016 and assumes a single value for WACC, effectively 2017 have been breakthrough years, as auction controlling for this variables impact on costs, results around the world have confirmed that while the auction price is explicitly dependent on a step change in costs has been achieved. The the, unknown, WACC of the individual project and 7 estimated global weighted average offshore project developer. This is an important point, as wind project LCOEs between 2010 and 2017 recent auction experience suggests that very low varied between USD 0.14 and USD 0.19/kWh costs of capital are playing an important role in (Figure 2.13). Auction results in 2016 and 2017 the most competitive auction results. Policies to suggest, however, that projects commissioned reduce the perceived risks of project development 0. 06 from 2020 onwards will fall in the range USD are therefore an important part of the overall to USD 09/kWh, excluding grid connection 0. framework required to achieve very low costs. 0. 0. 10/kWh, including 07 to USD costs, and USD Finally, there are other complications. In many 8 grid connection costs. The progression for CSP instances, the full details of the auction or tender appears to be equally, if not more spectacular. conditions are not publicly disclosed, making any Although the estimated weighted average LCOE judgement about the relative level of remuneration of projects fell significantly between 2010 and highly speculative. Sometimes “headline” prices 2017 for commissioned projects, they were still announced do not represent the full remuneration estimated to average USD 22/kWh in 2017 – 0. to the project under the agreement. For instance, albeit in a relatively thin year for deployment. The only the off-peak remuneration may be quoted, successful bidder for the recent Dubai auction or additional capacity payments that are not heralded a new price paradigm, however, while remunerated by kWh may be left out. Australia has also announced a highly competitive 9 There may also be additional sources of revenue project in South Australia. With slightly longer available to the project that are not clear. In the lead times for commissioning, notably for the recent Mexican auctions, for example, much has MW Dubai Electricity and Water Authority 700 0. 02/kWh results. been made of the sub-USD (DEWA) project, by 2022, CSP will be providing Yet this excludes the value of the clean energy electricity in the USD 0.07/kWh range, while the certificates that will be associated with the South Australian Port Augusta project is expected projects, with the value of these still unclear today. to be online in 2020 and delivering electricity at a price of USD 0.06/kWh. 7. T his makes a project-by-project comparison of costs difficult, but also represents an opportunity. Future work by IRENA will look at trying to use auction data to identify WACC spreads in different markets based on auction results. In some markets, offshore wind farm developments have been co-ordinated in zones, so as to share grid infrastructure which is 8. provided by the grid operator. Such projects do not therefore include these costs in their bids. In other markets, however, notably the UK, this is not the case. 9. https://www.premier.sa.gov.au/index.php/jay-weatherill-news-releases/7896-port-augusta-solar-thermal-to-boost-competition-and- create-jobs 51

52 RENEWABLE POWER GENERATION COSTS Global levelised cost of electricity and auction price trends for offshore wind and CSP from project Figure 2.13 and auction data, 2010-2020 Concentrating solar power Oshore wind 0.4 0.4 0.3 0.3 0.2 0.2 2016 USD/kWh 0.1 0.1 Auction database LCOE database 0.0 0.0 2017 2017 2012 2013 2012 2015 2013 2015 2020 2018 2021 2019 2016 2016 2019 2014 2020 2014 2011 2011 2022 2010 2010 Source: IRENA Renewable Cost Database and Auctions Database. publications, but to highlight just how much solar Thus, cost reductions for onshore wind, solar PV, PV – and to a lesser extent onshore wind – have offshore wind and CSP are continuing unabated. continuously exceeded expectations. Erring on the Despite the increasing maturity of the markets side of caution in terms of cost reduction potential for onshore wind and solar PV, too, further cost can therefore be a major error. reductions are being carved out. As a result, these technologies have significantly exceeded previous Indeed, solar and wind technologies highlight just predictions for cost reduction. It is also worth how poor a guide conventional wisdom can be in highlighting just how wrong previous projections estimating the continued capacity for technology or assumptions have sometimes been. In 2017, the improvement, industry efforts to improve global weighted average installed cost of utility- manufacturing, the impact of competition on supply 3 1 88/kW. This was around scale solar PV was USD chains and the benefits of experienced project 30% lower than the 2050 estimated value from the developers in driving down contingencies to wafer 2004 United States Solar PV Industry Roadmap thin margins. This process is also beginning to play and only slightly higher than the roadmap module out in other areas of the energy transition – notably only cost for 2030 (Moner-Girona, Kammen and in electricity storage (IRENA, 2017c). Margolis, 2018). More recent estimates have also The cost declines experienced from 2010 to 2017 been exceeded, too, with the 2017 installed cost and signalled for 2020 thus represent a remarkable numbers already lower than the projected values rate of change, and have enormous implications for 2031-2035 made in the International Energy for the competitiveness of renewable power World Energy Outlook (IEA, 2012). Agency’s 2012 generation technologies. This is not meant to denigrate the efforts of these 52

53 2017 installed capacity out to 2020. CSP has a higher Figure 2.14 plots the LCOE evolution of the four, main solar and wind technologies against learning rate of 30%, with deployment between cumulative installed capacity. A log-log scale 2010 and 2020 representing an estimated 89% of 11 cumulative installed capacity in 2020. is used to allow easy interpretation as learning Solar PV curves. The learning rate for offshore wind (i.e. has the highest learning rate – 35% between 2010 the LCOE reduction for every doubling in global and 2020 – with new capacity additions over this cumulative installed capacity) is expected to period that are estimated to be 94% of cumulative reach 14% over the period 2010 to 2020, with new capacity in 2020. capacity additions over this period estimated to Solar and wind power generation technologies be 90% of the cumulative installed offshore wind have entered a phase of rapid scale up and 10 capacity that would be deployed out to 2020. increasing technological and industry maturity For onshore wind, the learning rate for 2010- that in many ways mirrors the theory of industry 2020 is 21%, with new capacity added over this lifecycles (Utterback and Abernathy, 1975). As period covering an estimated 75% of cumulative Global weighted average CSP, solar PV, onshore and offshore wind project LCOE data to 2017 and Figure 2.14 auction price data to 2020, 2010-2020 0.500 0.400 2011 2012 2010 2011 2016 0.300 2010 2013 2015 2012 2013 0.200 2015 2016 2013 2014 2010 0.150 2017 2014 2015 2012 2016 0.100 2017 2013 2010 2016 USD/kWh 2020 2020 2016 0.070 2015 2020 Fossil fuel cost range 2020 0.050 0.040 0.030 0.020 0.015 0.010 20 000 10 000 1 000 000 500 000 200 000 100 000 50 000 1 000 2 000 5 000 CSP O‚shore wind Onshore wind PV Cumulative deployment (MW) Based on IRENA Renewable Cost Database and Auctions Database; GWEC (2017), MAKE Consulting (2017a), SolarPower Europe (2017), and WindEurope (2017). 10. G lobal cumulative installed capacity of CSP is projected to be 12 GW by 2020, for offshore wind 31 GW, solar PV 650 GW and onshore GW. This is based on IRENA (2017a), GWEC (2017), WindEurope (2017), SolarPower Europe (2017) and MAKE Consulting wind 712 (2017a) 11. Extending the horizon to 2022 to take into account the likely commissioning of the DEWA project increases uncertainty over total deployment values, but would be unlikely to greatly alter the learning rate. 53

54 RENEWABLE POWER GENERATION COSTS and technologies, while ongoing R&D efforts will such, rather than a focus on product differentiation, push those boundaries out even further. At the industry is increasingly having to focus on cost same time, for solar and wind, there still remain competitiveness. It is doing this by unlocking significant installed cost differences between economies of scale and optimising manufacturing countries. The convergence of installed costs and delivery processes to ensure an optimised low- towards best practice in many countries therefore cost product that meets the full range of customer still represents a significant cost reduction needs. It is also resulting a in a focus on improving potential, in addition to the underlying competitive the efficiency of the overall technology system and technology drivers acting to drive down the (e.g., reducing PV module and inverter losses, costs of best-in-class projects. wind availability focussing on MWh lost, not just downtime for O&M, etc.). This focus is facilitated th th Figure 2.15 highlights the 5 percentile and 95 by the highly modular and replicable nature of ranges for the total installed costs of onshore renewable power generation technologies. wind and solar PV projects by region. There exist This is not to imply that renewable energy significant differences within regions, due to site technologies are simple or not continuing to evolve. specific factors, but also market maturity, while The ongoing R&D efforts and sophistication of there are also significant differences between current solar PV panels, wind turbines, gearboxes, regions. For wind, China and India have different blade designs, control software etc. is undoubtable. cost structures to the rest of the world. These are The advantage comes from the completeness of not easily replicable, given their lower labour, raw the product as it leaves the factory, and the basic material and commodity costs and their access to construction skills then required for installation. cheap, local manufacturing hubs. However, that is When combined with the volume of individual not to say that individual projects in other regions projects, renewable technologies represent can't achieve these installed costs, just that the technologies and processes that can benefit from average is likely to remain higher. For most of standardisation, replicability and adaptability. countries and regions, however, shifting towards The latter is important, once local technical best practice in terms of today’s installed cost specificities (e.g., cold or hot climate operation, structures still represents one of the largest cost typhoon strengthening, etc.), regulatory, legal reduction opportunities available (IRENA, 2016a). and environmental processes are adapted to, then The trend of convergence towards best practice new markets can rapidly benefit from experienced installed costs is already underway and is likely to project developers replicating projects. continue in the period out to 2020 and beyond, This has been evident in recent years, as solar given the current evidence from auctions and and wind auctions in Mexico, Argentina, Saudi ongoing competitive pressures. What has been a Arabia and elsewhere have seen very competitive remarkable trend in the successful bids from recent results in countries without a significant history auctions has been the emergence of results in the in deployment of solar or wind technologies. 0. 0. 03 to USD 04/kWh range in Australia, USD The open question is how long this period of Canada, Chile and Turkey and elsewhere, for both rapid cost reduction will continue before the solar PV and onshore wind. industry experiences a slowing in the rate of cost For onshore wind, recently commissioned projects reductions. Given the relatively narrow deployment havepreviously achieved these levels of LCOE and of the majority of solar and wind power capacity are part of the reason why the global weighed to date – relative to the global potential – there is average has been declining. Yet these projects no reason to think that there will be a slowing in have typically been concentrated in locations with the average rate of cost reduction at a global level the best wind resources. What has been just as in the short- to medium-term. There still remain impressive, therefore, are the bids seen in more important technology improvements that are mature markets with significantly poorer wind already signalled by today's best-in-class projects 54

55 2017 gional total installed cost ranges for onshore wind and solar PV, 2016/2017 Figure 2.15 Re Onshore wind Solar PV 4 000 4 000 3 000 3 000 2 000 2 000 2016 USD/kW 1 000 1 000 0 0 Asia Asia Asia Asia India India Africa Africa Europe Europe Eurasia Eurasia Oceania Oceania Middle East Middle East North America North America South America South America Central America Central America and the Caribbean and the Caribbean Source: IRENA Renewable Cost Database. regulatory and policy framework, low offtake risk, resources. These include Germany and India, exchange rate risk and country risk are all essential where projects have been bid in the range around to unlocking low cost finance. Governments can 0.04 to USD 0.05/kWh. USD go a long way in ensuring these factors come Solar PV is not being left behind in trend towards together, but in some cases they will need the aid very low electricity costs. For 2018 and 2019, of development partners. One example of this is the auction results announced in 2016 and 2017 Zambia, where the clear benefits unlocked from suggest that cost reductions are set to continue the World Bank’s “Scaling Solar” programme apace, as deployment starts to accelerate in reduced country risk, offtake risk and exchange regions with excellent solar resources. The series of rate risk allowing a successful bid around half that world record low successful bids in 2016 and 2017 12 of results in a neighbouring country. for solar PV capacity in Abu Dhabi, Chile, Dubai, Peru, Mexico and Saudi Arabia has shown that Elsewhere, the recent auction in Mexico has very low solar PV costs are possible, particularly potentially seen values of around USD 0. 02/kWh where there are excellent resources, strong local being locked in for solar PV and onshore wind, civil engineering sectors, a regulatory and policy although these results are undoubtedly counting structure that inspires confidence in the stability on additional revenue from the clean energy of a project’s cashflows, and there is access to low certificates that will accompany the project. Given cost finance. current and likely near-term equipment costs, bids 0. in the USD 02/kWh or lower range are extremely This latter point is extremely important in achieving unlikely to represent an LCOE equivalent value, very competitive solar PV, even with low capital with additional revenue streams likely already costs and excellent resources. Having the right factored in. This allows the headline price to be at 12. For more details see: www.scalingsolar.org 55

56 RENEWABLE POWER GENERATION COSTS will be consistently undercutting fossil fuel-fired a discount to what an LCOE, even with very low electricity generation, without financial support WACC, would look like. and despite the fact that fossil fuel projects do not Figure 2.16 presents the range of LCOE and pay for their full local and global environmental auction price data from the IRENA Renewable costs. The global average cost of electricity from Cost Database and Auction Database for onshore onshore wind and solar PV will be flirting with the wind, solar PV, offshore wind and CSP for the lowest cost value for fossil fuel-fired electricity, period 2010-2021, as well as the weighted average while CSP projects and offshore wind will be at the trend for these sources and the fossil fuel-fired lower end of that cost range and offer competitive cost range. By 2019-2022, depending on the new generation capacity. technology, solar and wind power generation The outlook for solar and wind power electricity technologies will not only be providing competitive costs to 2020 presages historically low costs for electricity where new generation is needed, but new, renewable electricity. The overall average, individual projects will be increasingly providing but especially the very low electricity costs for the some of the lowest cost electricity available, best solar PV and onshore wind projects represent substantially undercutting fossil fuel-fired power a real paradigm shift in the competitiveness of generation LCOEs. renewables. Given these low costs, previously By 2019, the best onshore wind and solar PV uneconomic strategies for the electricity projects that will be commissioned will be and energy sector could become profitable. delivering electricity for an LCOE equivalent of Curtailment – previously an unthinkable economic USD 0.03/kWh or less, with CSP and offshore burden for renewables – may become a rational wind providing very competitive electricity from economic decision, maximising variable renewable 2020 onwards. Today and increasingly in the penetration and minimising overall system costs. future, many renewable power generation projects Global levelised cost of electricity and auction price trends for solar PV, CSP, onshore and offshore Figure 2.16 wind from project and auction data, 2010-2022 Concentrating solar Solar PV Onshore wind Oshore wind power 0.4 Auction database LCOE database 0.3 0.2 2016 USD/kWh 0.1 Fossil fuel cost range 0.0 2012 2012 2012 2012 2018 2018 2018 2108 2016 2016 2016 2016 2014 2014 2014 2014 2010 2010 2010 2010 2022 2022 2020 2020 2020 2020 Source: IRENA Renewable Cost Database and Auctions Database. 56

57 2017 processes. This is in stark contrast with the norm Similarly, such low prices in areas with excellent for large civil engineering projects today, where solar and wind resources open up the economic cost overruns and time delays are common (Adam potential of power-to-X technologies (e.g., power et.al; 2017). to hydrogen or ammonia, or other energy dense, storable mediums). At the same time, these low As a result, the LCOE of electricity from onshore prices make the economics of electricity storage wind, offshore wind, solar PV and CSP are now more favourable, potentially turning a drawback converging on very competitive levels. By 2020- of electric vehicles (EVs) – their potentially high 2022, the LCOE of electricity from solar and wind instantaneous power demand for recharging technologies will fall solidly within the range of – into an asset, as EVs can take advantage of USD 0.03 to USD 0.10/kWh. There will, however, cheap renewable power when it is available and be a range of projects that fall outside this range. potentially feed electricity back into the grid if Recent auction results have already signaled that needed later on. This, however, has to be balanced there could be projects that, in future, fall below by the increased costs of integrating variable this range. At the same time, a range of projects renewables and the increased flexibility required in new markets or challenging development to manage systems with high Variable Renewable environments, such as in remote locations or on Energy (VRE) – although noting that low costs islands, will continue to fall above this range. help make that challenge less costly. To date, these There is a lesson here for the rest of the energy integration costs have remained modest, but they sector’s transformation, too. With the right policy will rise as very high shares of VRE are reached and regulatory settings, renewable technologies (IRENA, 2017f). can scale to provide cost-effective solutions to There is a clear pattern to the evolution of the countries energy, environmental, economic and cost of electricity from solar and wind power social goals. Crucially, once sufficient momentum technologies. It is a template that has been driven in the sector is achieved, they will often exceed by support policies that have unlocked technology expectations as industrialisation and scale effects improvements and cost reductions and has begin to take hold. Yesterday’s insurmountable resulted in a virtuous cycle. As market deployment challenges, in terms of cost competitiveness, are has grown, economies of scale have followed. More falling by the wayside and there is a template in competitive supply chains and improvements in this for addressing tomorrow’s challenges. The manufacturing processes have come as the markets lesson of the last 10 years from solar and wind for these technologies have been industrialised. technologies is clear: a long-term vision, with the Onshore wind and solar PV have both benefitted right support policies and regulatory frameworks, from this process of industrialisation, and now can allow industry to scale, competition to play offshore wind and CSP are benefitting from the its part and the right technology solutions to same development. Industrialisation has been be brought to market faster and cheaper than facilitated, too, by the relative simplicity of the conventional wisdom suggests is possible. components, their modularity, scalability and the replicability of the construction and installation 57

58

59 3. SOLAR PHOTOVOLTAICS The bulk of PV production capacity continues to be he global PV market has grown rapidly in the situated in Asia, where China is the world leader in last decade. Cumulative global installed PV T PV production. China and Japan together account - capacity grew from 6.1 GW at the end of 2006 ed for around 70% of global module production in to 291 GW at the end of 2016 (IRENA, 2017a). both 2015 and 2016. Manufacturing capacity and From 2010 to 2016, net additions grew about production is also growing in other countries in the 28% annually on average and additions in the Asia-Pacific & Central Asia regions and countries time period account for about 94% of the total in these regions accounted for about a tenth of the capacity that was installed between 2006 and modules produced globally in 2016 (Fraunhofer 3.1). 2016 (Figure ISE, 2016, 2017). First and second-generation tech - Recent growth in the Asian PV market has more 1 nologies account for virtually all production, while than compensated for the decrease in new capacity crystalline silicon-based photovoltaics currently additions in Europe in recent years, as growth in 3.2). continues to dominate the market (Figure China and Japan has increased. These countries Crystalline silicon module production accounted GW between 2014 and together installed about 88 for about 94% of production during 2016, up 2016 alone. At the end of 2016, China was home from 93% in 2015. (Fraunhofer ISE, 2016, 2017; to 27% of cumulative installed capacity globally. GlobalData, 2017). Growth in other regions has also continued. For example, through steady growth in recent years, In the last decade, crystalline silicon wafer based the United States has become a large PV market commercial module average efficiencies have with 11% of the global cumulative installed capacity increased from about 12% to a range of 17% to at the end of 2016. 17.5%. Best performing modules in the laboratory can currently reach up to 24.4% efficiency. Current Yearly installations in Europe have declined since crystalline module efficiencies are typically at GW of new their highest historical value of 22 least 2% lower than efficiencies at the cell level capacity additions in 2011. In both 2014 and 2015, due to losses caused by various factors such as: new additions did not exceed 8 GW, and in 2016 the module border, cell spacing, cover reflection 5 W were installed in the region. Europe’s share G and cell interconnection. However, cell and of total global cumulative capacity declined from module efficiencies are intrinsically linked and around three quarters over the period 2009 to 2011, current developments in best cell efficiency levels to 44% in 2015. This was the last year when Europe suggest that continued improvements in the held the leading position in respect to cumulative average efficiency of modules will continue for the capacity. In 2016, this share decreased to 35%. ore detailed discussion of this solar PV technology categorisation can be found in IRENA, 2016a. 1. A m 59

60 RENEWABLE POWER GENERATION COSTS Yearly added and cumulative global PV capacity by region, 2006-2016 Figure 3.1 New grid connected capacity 70 60 50 40 GW 30 20 10 0 20062007200820092010201120122013201420152016 Cumulative grid connected capacity 300 250 200 GW 150 100 50 0 20062007200820092010201120122013201420152016 Region EuropeAsiaNorth AmericaOceaniaAfricaSouth AmericaCentral America and the Caribbean Middle EastEurasia Source: IRENA, 2017a. 60

61 2017 Figure 3.2 Solar PV module production: Capacity and volume by technology, 2010-2016 2010 20122013201420152016 2011 100 80 60 GW 40 20 0 100 % 50 % % of Total GW 0% Production capacity, crystallineProduction volume, crystalline Production capacity, thin-film Production volume, thin-film Based on GlobalData, 2017 and Fraunhofer ISE, 2016, 2017. foreseeable future. For instance, by 2024, industry 3.1 INSTALLED COST TRENDS expectations place the range of stabilised cell Recent module costs trends efficiency for mass production of crystalline silicon based cells at 19.8-25% depending on cell type and Solar PV module prices in Europe decreased by architecture up from a current range of 18.8-23.5% 83% from the end of Q1 2010 to the end of Q1 (ITRPV, 2017; Fraunhofer ISE, 2017). 2017 (Figure 3.3). Module costs declined 80% between the end of 2010 and the end of 2016, a The two most deployed thin-film technologies are period over which 87% of the cumulative global Cadmium-Telluride (CdTe) and Copper-Indium- PV capacity installed at the end of 2016 occurred. Gallium-Selenide (CIGS). First Solar (the largest Solar PV module costs fell rapidly until 2013, but CdTe manufacturer) reported fleet average have experienced more modest cost reductions efficiencies increasing from 12.9% in 2012 to in recent years as PV module manufacturers 16.6% in 2016 for their CdTe modules (First Solar, made efforts to return profit margins to more 2017). For CdTe cells, module efficiency record sustainable levels and various trade disputes for the moment is 18.6%. The best CIGS reported affected minimum prices in different markets. efficiencies so far were 17.5% for modules (Green Average monthly solar PV module prices in Europe et al., 2017). Solar Frontier reports current CIGS in 2016, were 13% lower than in 2015, while the module efficiencies between 12.2%-13.8% for their decline in average prices across a range of markets CIGS modules (Solar Frontier, 2017). (right side of Figure 3.3) was 18% between 2015 and 2016. As a result of import treatment and individual market preferences for particular module types, there are a wide range of module prices depending on the market. Figure 3.3 highlights that although PV modules are relatively 61

62 RENEWABLE POWER GENERATION COSTS to improvements in the production processes homogeneous technologies, they are not entirely and to efficiency gains associated with increased interchangeable commodities. In 2016, average adoption of newer cell designs (although the growth selling prices in China were around USD 0.43/W, in cumulative deployment and manufacturing while California became one of the highest priced scale still plays a role in achieving low costs). 61/W, though major markets with prices of USD 0. all analysed markets experienced a decreasing On the processing side, previous IRENA work has cost trend between 2015-2016. These are average reported on the growing market presence of the values and a range of prices around these values diamond wafer cutting method (IRENA, 2016a). occur. In 2017, module prices have dipped as low as Diamond wire sawing provides opportunities to USD 3/W, but are somewhat higher for modules 0. reduce costs through reduction of material losses from Chinese majors and good quality modules during slicing. During 2016, these costs were about 4/W can now be produced sustainably for USD 0. 2 a fifth lower than for the traditional method. Since or less (Exawatt, 2017). 2016, this wafer slicing technique is already prevalent Rather than being driven primarily by substantial in the monocrystalline segment, and by the end of capacity and deployment upsurge and their 2017 90%+ of monocrystalline wafers worldwide associated economies of scale, recent and near will be being cut with this method and around half future module cost reductions relate more closely of multicrystalline (Exawatt, 2017). This occurred Figure 3.3 Average monthly European solar PV module prices by module technology and manufacturer, March 2010—May 2017 (left) and average yearly module prices by market in 2015 and 2016 (right) 0.75 4.0 3.5 0.70 3.0 0.65 2.5 0.60 2.0 0.55 2016 USD/W 2016 USD/W 1.5 0.50 1.0 0.45 0.5 0.0 0.40 2016 2015 Mar 17 Mar 12 Mar 13 Mar 15 Sep 12 Sep 13 Sep 15 Mar 16 Mar 14 Sep 16 Sep 14 Mar 10 Mar 11 Sep 10 Sep 11 South Africa Mexico Japan United Kingdom Brazil China Thin film a-Si/u-Si Crystalline Europe (Germany) or Global price index Ontario Malaysia India Crystalline Japan (Q4 2013 onwards) Germany California Thin film CdS/CdTe Thin film a-Si Crystalline China Source: GlobalData, 2017; pvXchange, 2017; Photon Consulting, 2017. n the longer term, the advantage may be greater and by 2027, industry expects diamond slicing technology’s kerf losses to decline I 2. m for slurry (ITRPV, 2017). μ μ m, compared to 120 to 60 62

63 2017 wire cut multi-crystalline wafers based on black as manufacturers transitioned from the traditional silicon, as well as an increasing market share taken method involving abrasive powdered silicon carbide by PERC cells and their associated cost reductions slurries. Industry announcements confirm a trend are to be expected (pv magazine, 2017). towards higher shares of diamond wafer slicing technology use in the multi-crystalline segment With the cell architecture shifting towards PERC as well, as this has the potential to lower silicon cells, makers are also developing technologies consumption (ITRPV, 2017; Bernreuter Research, to address the light-induced degradation (LID) 2017; GCL-Poly, 2017; CanadianSolar, 2016). problem which affects them. On the mono- crystalline side, LONGi Solar is developing a At the same time, effective texturing processes Light-Induced Regeneration (LIR) technology, are necessary for diamond wire sliced cells. This is jointly developed with the University of New South in order to avoid issues with high surface texture Wales. LONGi Solar claims that by controlling reflectivity – which can affect cell performance – the degradation through the LIR technology, the resulting from the slicing process. In this respect, energy yield at the PV plant level can be enhanced there is a trend towards increased use of ‘black by 1%. The company has also announced that it silicon’ in wafer texturing as an anti-reflection is willing to open-up this technology to industry. measure for solar cells and in combination with (LONGi Solar, 2017). If realised, this step could diamond wire slicing processes. The term ‘black have important implications for the race in cell silicon’ refers to a silicon surface which has been architecture technology, given the leading market covered with a nanostructured surface layer in presence of the company. Though LID is a well- order to boost its light absorption properties (Liu known issue for mono-crystalline wafer cells, et al., 2014). it also affects multi-crystalline cells and PERC Various approaches to black silicon fabrication are cells. Research and industry efforts are underway available. For example, metal assisted chemical to better understand this phenomenon and to etching methods are able to provide an efficient minimise the performance losses associated way of producing high efficiency ‘black’ multi- with this (Luka et al., 2016; Kraus et al., 2016; crystalline cells (Ying et al., 2016). Other etching Padmanabhan et al., 2016). methods, such as the reactive ion etching method With recent and expected cost and performance (Shim et al., 2012), are also being researched and developments, a definitive PV module technology used for this purpose. While industry opinion strategy remains difficult to predict. What is certain seems divided regarding the most adequate is that competitive pressures in the PV module etching method, several multi-crystalline industry market will remain intense, with technology players have placed their attention on black silicon innovations crucial to module manufacturers technologies in an effort to improve the cost ability to remain profitable in a rapidly evolving performance ratio of multi-crystalline wafers cut market. with diamond wires (EnergyTrend, 2017). Newer etching-texturing technologies are therefore Total installed costs expected to continue to gain market share over Though solar PV technology has matured and more traditional, standard acidic etching methods. and more countries are starting to deploy solar PV In terms of the uptake of novel cell designs, there at scale, regional cost differences persist. Different is a trend towards increased adoption of both domestic market maturity levels (as, for example, multi- and mono-passivated emitter rear (PERC) evidenced in project developer’s experience), cell architectures (IRENA, 2016a). Such a trend, as well as differences in local labour and alongside a shift towards a more widespread use manufacturing costs and different support policy of black silicon, is allowing multi-crystalline cells to structures can all influence competitiveness. Some move into the higher efficiency segment, has been detailed research comparing individual markets has confirmed. While the definitive technology path been published (Seel et al., 2014; Friedman et al., remains uncertain, wider adoption of diamond 2014; Kimura and Zissler, 2016; Strupeit, 2016), yet 63

64 RENEWABLE POWER GENERATION COSTS include reducing the administrative hurdles much more research of this nature would benefit associated with gaining permits or incentives, or the understanding of why cost differentials persist those that slow connection application processes. and how they might be most effectively reduced to best practice levels. In addition, on-going Between 2010 and 2017, the global capacity research on the topic is necessary, since gaining weighted average total installed cost of newly a deep understanding of the reasons behind the commissioned utility-scale PV projects decreased cost differences in the different markets can be by 68%, with a 10% decrease in 2017 from 2016 extremely valuable in informing policy making for levels (Figure 3.4). Projects in newer markets are cost reduction targeting. being developed at costs that are increasingly 3 at par, and sometimes even cheaper than the As balance of system (BoS) costs, discussed in averages in more cost mature markets. more detail in Annex I, contribute more and more to total system cost reductions, adopting policies Rapid installed cost declines in China, Japan and that can bring down soft costs provides the the United States – and the rapid emergence of an opportunity to improve cost structures towards increasing number of cost competitive projects in best practice levels. Examples of such policies Total installed costs for utility-scale solar PV projects and the global weighted average, 2010-2017 Figure 3.4 8 000 th 95 percentile 7 000 6 000 5 000 2016 USD/kW 4 000 th percentile 5 3 000 2 000 1 000 Weighted average investment cost 20102011201220132014201520162017 ≥ 1100200 ≥ 300 Capacity MW Source: IRENA Renewable Cost Database. 3. Balance of system costs in this chapter do not include inverter costs, which are treated separately. 64

65 2017 competitive pressures, markets in Australia, Chile, India (often at best-in-class cost levels), as well as France, Jordan and the United Kingdom have all in the newer markets – have been the main driver seen rapid installed cost reductions that have in the increasing competitiveness of utility-scale reduced the differential from China in the period PV. During the period 2010 to 2017, utility-scale 2015 to 2016. total installed cost reductions in many markets have exceeded 70% (Figure 3.5). Between 2010 During 2016, the percentage difference of total and 2017, the United States saw utility-scale total installed costs for utility-scale systems compared installed costs reduce the least, at 52%, with Italy to Chinese levels ranged between -6% and 77%. experiencing the largest reduction of 79%. This is a significantly narrower span than in 2015, when they ranged between 10% and 136% above Despite the generalised reduction in installed costs the Chinese level. across all markets, significant cost differentials between markets remain. Using China as a base Figure 3.7 highlights the major reasons for these for an index, Figure 3.6 shows that for a range of cost differentials by providing a detailed breakdown countries, the cost differentials compared to China of utility-scale total installed costs by country in have been declining. Cost differences among 2016. The markets that significantly reduced the markets are expected to continue to decline, as differential over Chinese installed costs did so by the least mature markets gain more experience driving down BoS costs towards more competitive during their growth (IRENA, 2015). With greater Figure 3.5 Utility-scale solar PV total installed cost trends in selected countries, 2010-2017 Japan Germany United States China 8 000 6 000 4 000 2016 USD/kW 2 000 -52% -77% -71% -70% 0 20172010 20172010 20172010 2017 2010 India Italy France United Kingdom 8 000 6 000 4 000 2016 USD/kW 2 000 -76% -77% -75% -79% 0 20172010 2017 20172010 20172010 2010 Source: IRENA Renewable Cost Database. 65

66 RENEWABLE POWER GENERATION COSTS Figure 3.6 Estimated utility-scale solar PV system costs: China compared to other countries, 2015-2016 Japan Australia USA ChinaGermanyIndiaUKFranceJordanChile 136% 140% Total installed cost as a percentage dierence of Chinese total installed costs 120% 110% 111% 100% 80% 77% 80% 73% 72% 60% 44% 40% 22% 20% 19% 16% 20% 12% 7% % Dierence in total installed cost (2016 USD/kW) 1 1101 168 0% Total installed cost as in China -6% (2016 USD/kW) -9% -20 % 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 Source: IRENA Renewable Cost Database. with a wide span of installed cost levels between levels. Countries with competitive installed cost markets. California has become the most expensive levels have, on average, balance of system costs residential solar PV market for which IRENA has (excluding the inverter) that make up about half data, with total installed costs of USD 0/kW 4 55 of the total installed cost. Soft cost categories for in Q1 2017, more than three times higher than India the displayed countries make up a third of these and double the costs in Germany. BoS costs, and about 17%, on average, of the total installed costs. 3. 2 CAPACIT Y FACTORS Residential PV system total installed costs have also declined sharply in a wide range of countries 4 of The global weighted average capacity factor since 2010. The range of residential solar PV total utility-scale PV systems increased by 28% be- system costs in the markets with the longest tween 2010 and 2017, from an average of 13.7% to historical data shown in Figure 3.8 decreased from 17.6%. This has been driven by three major fac- between USD 6 700 USD 100/kW in Q2 and USD 11 tors, the trend towards greater deployment in re- 1 050 and USD 0/kW 4 55 2007 to between USD gions with higher irradiation levels, the increased in Q1 2017 (a decline of 47-78%). Since 2013, with use of tracking and improvements in the perfor- the broadening of the residential solar PV market, mance of systems as losses have been reduced more data has become available for a much (e.g., though improvements in inverter efficiency). wider selection of markets. For this wide range of Data from the United States, for instance, high- emerging and OECD economy markets, the total lights the increased use of trackers, with these installed costs of residential PV systems fell by making up 39% of new capacity additions in 2014, between 18-66% between Q2 2013 and Q1 2017, but 70% in 2015 and 79% in 2016 (Bolinger and Seel, 2016 and Bolinger et al., 2017). 4. T he capacity factor for PV in this chapter is reported as an AC/DC value. For other technologies in this report, the capacity factors are expressed in AC-to-AC terms. More detailed explanations on this can be found in Bolinger and Weaver, 2014; Bolinger et al., 2015. 66

67 2017 Detailed breakdown of utility-scale solar PV costs by country, 2016 Figure 3.7 2 000 1 500 1 000 2016 USD/kW 500 0 IndiaGermanyItaly China FranceJordanChileUKAustraliaJapanUnited States Soft costs Installation Hardware Customer acquisition Electrical installation Module Inverter Inspection Financing costs Cabling/ wiring Mechanical installation Incentive application Grid connection Margin Permitting Monitoring and control System design Racking and mounting Safety and security Source: IRENA Renewable Cost Database. 67

68 RENEWABLE POWER GENERATION COSTS Figure 3.8 Average total installed costs of residential solar PV systems by country, Q2 2007-Q1 2017 Q2 2007-Q1 2017 Japan US non-California (10-kW) California (0-10 kW) Germany 10 000 5 000 2016 USD/kW -47% -78% -73% -62% 0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q2 2007 Q2 2007 Q2 2007 Q2 2007 Q2 2013-Q1 2017 India France China AustraliaBrazil Korea 5 000 2 500 -26% 2016 USD/kW -66% -39% -50% -41% -54% 0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 MalaysiaSouth AfricaSpainSwitzerlandThailandUK (0-10kW) 5 000 2 500 -26% -36% 2016 USD/kW -18% -40% -44% -48% 0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2013 Q1 2013 Q1 2013 Q1 2013 Q1 2013 Q1 2013 Source: IRENA Renewable Cost Database. being developed in higher irradiation regions and This increase in global weighted average utility- the increased use of tracking seem to be driving scale solar PV capacity factors is despite the trend increases in the global weighted average capacity in some markets towards higher inverter load 5 factor, offsetting any reductions caused by ratios (ILR) . This ratio of DC module capacity to increasing ILRs in recent years (Figure 3.9). AC inverter capacity (also known as DC/AC ratio) is a project design consideration and raising it can reduce the LCOE in some contexts (Good and 3.3 OPERATION AND MAINTENANCE COSTS Johnson, 2016). All things being equal, increasing the ILR reduces the AC/DC capacity factor. In the Historically, solar PVs O&M costs have not been United States, for example, the capacity weighted considered a major challenge to their economics. average ILR of utility-scale projects increased 9% Yet, with the rapid fall in solar PV module and between 2010 and 2016 to a value of 1.31 (Bolinger installed costs over the last five years, the share of et al., 2017; Fiorelli and Zuercher-Martinson, 2013). O&M costs in the LCOE of solar PV in some markets Globally, the trend towards more PV projects has climbed significantly. T 5. he Inverter Load Ratio (or DC/AC ratio) describes the ratio of a module array’s DC rated output and the inverters size expressed in AC power terms. 68

69 2017 Global weighted average capacity factors for utility-scale PV systems, 2010-2016 Figure 3.9 30% th 95 Percentile 20% Capacity factor 10% th Percentile 5 0% 2011201220132014201520162017 2010 Source: IRENA Renewable Cost Database. 3.4 LEVELISED COST OF ELECTRICITY O&M costs in some OECD markets, such as Germany and the United Kingdom, now account Rapid declines in installed costs and increased 25% of the LCOE (STA, 2014; deea, 2016). In - for 20 capacity factors have improved the economic terms of the breakdown of O&M costs, data for the competitiveness of solar PV around the world. United Kingdom in 2014 suggested maintenance The global weighted average LCOE of utility-scale costs accounted for 45% of total O&M costs, land PV plants is estimated to have fallen by 73% lease for 18%, local rates/taxes for 15%, insurance 36 to 0. between 2010 and 2017, from around USD for 7%, site security and administration costs for 4% USD 0.10/kWh. Between 2010 and 2013, the global each, and utilities (including purchased electricity) weighted average LCOE declined by about 20% for 2% (STA, 2014). O&M costs for utility-scale per year, although it experienced a more modest plants in the United States have been reported 8% decline between 2013 and 2014, as the market to be between USD 10 and USD 18/kW per year experienced a shift away from traditionally low cost (Bolinger and Seel, 2015; Fu, et al., 2015). markets towards higher cost markets, such as Japan Land lease costs are very site- and market-specific. and the United States (IRENA, 2015). Between They can be extremely low where land values are 2014 and 2015 the LCOE declined again, by around minimal (e.g., in deserts or other uninhabited areas a fifth, while the descent between 2015-2016 without other productive uses) or can even be zero was 11%. The estimated decline between 2016 and when no land fees are charged as an incentive for 2017 was 15% 6 the project developer to minimise costs. This is in th th The 5 and 95 percentile range of the utility- stark contrast to markets where land constraints scale LCOE declined from between USD 0.18 and are an important challenge, such as in densely USD 07 and 60/kWh in 2010 to between USD 0. 0. populated locations, where land-use costs can be th th 31/kWh in 2017. The 5 0. USD percentile and 95 very significant. values declined by 58% and 48% respectively during the same period (Figure 3.10). his can be the case where regional or central governments have land in their possession that can be used for solar PV projects and can T 6. reduce the procurement costs for the electricity offered by project developers. 69

70 RENEWABLE POWER GENERATION COSTS Levelised cost of electricity from utility-scale solar PV projects, global weighted average and range, Figure 3.10 2010-2016 0.6 th 95 percentile 0.4 2016 USD/kWh 0.2 th percentile 5 Weighted average LCOE 0.0 20102011201220132014201520162017 ≥ 1100200 ≥ 300 Capacity MW Source: IRENA Renewable Cost Database. that the LCOE of utility-scale projects in the United The downward trend in the LCOE of utility-scale States is not significantly higher than in other solar PV by country is presented in Figure 3.11. markets. Between 2010 and 2017, the weighted average LCOE of utility-scale solar PV declined by between The LCOE of residential systems has also declined 40-75% depending on the country. The Italian at a very fast pace. For example, based on the market experienced the largest percentage LCOE assumption of a 7.5% cost of capital, the LCOE of reduction between 2010 and 2017, driven by residential PV systems in Germany declined 73% module price reductions, but also by significant 2 0. 55 007 and Q1 2017 from USD between Q2 reductions in BoS costs across the board. Italy has to USD 2010 0.15/kWh (the decline from Q1 now reduced soft costs and other hardware costs 7 2017 was 58% to Q1 ). Data since 2013 from to very low levels (Figure 3.7). In the United States, India, China, Australia and Spain shows that in stubbornly high BoS costs across the board have these countries, which have better irradiation resulted in slower cost reductions than in other conditions, and where installed costs have markets. However, excellent solar resources mean 7. A ssuming a weighted average cost of capital of 5% the LCOE decline in Germany between Q2 2007 and Q1 2017 would have been 72% (from USD 0.46 to USD 0.13/kWh). From Q1 2010-Q1 2017 it would have been 56% (from USD 0.30 to USD 0.13/kWh). 70

71 2017 Figure 3.11 Utility-scale solar PV: Electricity cost trends in selected countries, 2010-2017 Germany Japan United States China 0.6 0.4 2016 USD/kWh 0.2 -73% -40% -64% -72% 0 2017 20172010 20172010 20172010 2010 Italy United Kingdom India France 0.6 0.4 0.2 2016 USD/kWh -70% -67% -71% -75% 0 20172010 20172010 20172010 2017 2010 Source: IRENA Renewable Cost Database. poor solar resource. Figure 3.13 shows the average become increasingly competitive, lower LCOEs yearly LCOE estimates for residential PV in Germany, than the above German example can be achieved as well as the percentage difference of the LCOE in even if installed costs are sometimes higher. In other markets to the German LCOE for a given year. these low-cost markets, the LCOE range was From this point of view, it is noticeable that due to 0.20/kWh in Q2 2013, 0.15 and USD between USD total installed cost reductions, traditionally high- falling to between USD 0.08 and USD 0.12/kWh cost markets have started to converge around the in Q1 2017 (Figure 3.12), a decline of between 34% German level. At the same time, for markets with and 45% during the period. very competitive installed costs and good irradiation In higher cost markets, reductions have continued conditions, LCOE estimates have continued to as well. In France, for example, residential PV LCOEs fall and indeed have opened up a larger gap with declined 61% between Q2 2 2 013 and Q1 017, while Germany. Australia is a notable example, despite in the United Kingdom, they declined 38% during higher installed costs, the excellent solar resource the same period. The LCOE estimates in these three meant that the estimated residential LCOE in 2010 0. 22/kWh during Q1 countries did not exceed USD in Australia was only 7% higher than in Germany 2017, however, this is still 46% higher than the costs and around the same in 2011. Since then, continued in the more mature market of Germany. installed cost reductions in Australia saw the LCOE gap compared to Germany widen. In 2016 the LCOE Historically, Germany was a major driver of the estimate was 30% lower than in Germany and in growth in residential solar PV over the last ten years 2017 it was 31% lower. and has highly competitive installed costs, but a 71

72 RENEWABLE POWER GENERATION COSTS Levelised cost of electricity from residential solar PV systems by country, Q2 2007-Q1 2017 Figure 3.12 Q2 2007-Q1 2017 US non-California (10-kW) California (0-10 kW) Japan Germany 0.6 0.4 -73% 0.2 -44% -58% 2016 USD/kWh -70% 0.0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q2 2007 Q2 2007 Q2 2007 Q2 2007 Q2 2013-Q1 2017 AustraliaBrazil China France India Korea 0.6 0.4 0.2 -34% -45% -24% -45% -61% -33% 2016 USD/kWh 0.0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 MalaysiaSouth AfricaSpainSwitzerlandThailandUK (0-10kW) 0.6 0.4 0.2 -23% -38% -15% -36% -32% 2016 USD/kWh -41% 0.0 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q1 2017 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Q2 2013 Source: IRENA Renewable Cost Database. 72

73 2017 Levelised cost of electricity from residential PV: Average differentials between Germany Figure 3.13 and other countries, 2010-2017. Average residential solar PV LCOE in Germany % Dierence from Germany 0.35 100% Australia 0% 0.30 -100% 100% 0.25 France 0% -100% 0.20 100% India 0% 0.15 -100% 2016 USD/kWh 100% 0.10 Japan 0% -100% 0.05 100% UK 0% (0-10 kW) 0.00 -100% Note: 2017 up to Q1 20102011201220132014201520162017 20102011201220132014201520162017 Source: IRENA Renewable Cost Database. 73

74 RENEWABLE POWER GENERATION COSTS Solar PV cost trends in the commercial sector Box 3 As more companies and businesses turn toward solar PV for electricity generation due to attractive economic returns under net metering or feed-in-tariff schemes, the commercial PV market has seen significant growth in recent years. The commercial segment is more heterogenous in class sizes among countries and economic sectors than the residential market. This and the diverse point in time at which the data has become available can make a comparison of cost trends between markets challenging. However, to shed more light into the global trends of this PV market segment, IRENA has compiled a dataset of commercial PV costs for systems up to 500 kW of capacity from markets for which data is readily available. Commercial solar PV total installed cost and levelised cost of electricity by country or state, 2009-2017 Figure B3.1 Total installed costs of commercial PV (up to 500kW) and percentage change between first and last available quarter value Massachusetts China New York Italy Arizona UK Japan Germany Australia California 2009 Q2 2009 Q2 10 000 2009 Q2 2009 Q2 2009 Q2 8 000 2009 Q2 6 000 2012 Q2 -53% 2011 Q1 4 000 2016 USD/kW 2011 Q2 2014 Q2 -67% 2017 Q2 -66% 2017 Q2 2 000 2017 Q2 -69% -69% -62% -82% 2017 Q2 -44% -81% 2017 Q2 -62% 2016 Q2 2017 Q2 2017 Q2 0 2016 Q2 2017 Q2 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 Levelised cost of electricity of commercial PV (up to 500 kW) and percentage change between first and last available quarter value Australia Italy Japan UK China California New York Massachusetts Germany Arizona 2009 Q2 1.0 0.8 2009 Q2 2009 Q2 0.4 2009 Q2 2009 Q2 2012 Q2 2009 Q2 0.4 2011 Q1 2016 USD/kWh 2011 Q2 -57% 0.2 2014 Q2 -61% -63% -77% -79% -54% -38% -65% 2017 Q2 2016 Q2 -67%-51% 2017 Q2 2017 Q2 2016 Q2 2017 Q22017 Q2 0 2017 Q22017 Q22017 Q2 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 20092017 Source: IRENA Renewable Cost Database. The total installed costs of commercial sector solar PV for system sizes up to 500 kW have often followed a similar downward trend as has been in evidence in the utility-scale solar PV sector. The lowest average total installed costs for commercial PV can be found in Germany and China, at USD 1 º 1 00/kW and 1 150/kW, respectively. The highest 650/kW. In terms of the LCOE of commercial solar cost market remains California with total installed costs of USD 3 PV, the lowest average LCOE was around USD 0.10/kWh in Australia Q2 2017, after having decreased 38% between Q2 2014 and Q2 2017. 74

75 2017 75

76

77 CONCENTRATING 4. SOLAR POWER Though much less deployed, Fresnel collectors oncentrating solar power (CSP) relies on are another type of technology in linear focusing concentrating the sun’s rays through the use of C CSP plants. These are similar to PTCs, but they use mirrors to create high temperature heat to drive a an array of almost flat mirrors (reflectors) instead steam turbine. In the majority of today’s systems, of parabolic trough-shaped mirrors – although the sun’s energy is transferred to a fluid, which in they are designed to approximate the PTC’s turn is passed through heat exchangers to run a form. In Fresnel systems, mirrors concentrate the traditional electricity steam cycle, similar to the sun’s rays onto elevated linear receivers that are one used in conventional thermal power plants. not directly connected to them, but are located CSP plants can also have thermal storage systems. several metres above the primary mirror field. Often, a two-tank molten salt storage system is Solar towers are currently the most used point used, but designs vary. According to the way solar focal system currently deployed. Often also known collectors concentrate the solar irradiation, CSP as ‘power towers’, solar tower CSP systems use systems can be divided into line-concentrating a ground based array of large mirrors that track and point focussing systems. the sun individually in two axes and which are Parabolic trough collectors (PTC) are the more commonly known as heliostats. In solar towers, widely deployed linear concentrating technology. the heliostats concentrate solar irradiation onto a PTCs consist of parabolic trough shaped mirrors receiver mounted at the top of a tower. The central (collectors) that concentrate the solar radiation receiver absorbs the heat through a heat transfer along a heat receiver tube (absorber). This tube 1 medium, which is then used to generate electricity, is thermally efficient and placed in the collectors’ typically through a water-steam thermodynamic focal line. Single axis sun tracking systems are cycle. Solar towers can achieve very high solar typically used in PTC systems to orient the solar concentration factors (above 1000 suns) and reach collectors, together with the receiver tubes, higher operating temperatures than PTC plants, towards the sun and increase energy absorption. which can allow for low-cost thermal energy Through the use of a heat transfer fluid (often storage and higher capacity factors and efficiency thermal oil) and a heat transfer fluid system these levels compared to PTC plants. individual solar collectors are connected in a loop and deliver the heat to heat exchangers, where CSP has the advantage that it can be equipped superheated steam is produced. The steam typically with low-cost thermal energy storage. This allows drives a steam turbine electricity generator. CSP to provide dispatchable renewable power. CSP therefore can offer advantages, such as allowing ome solar tower designs aim at avoiding the use of the heat transfer medium, however, and instead directly produce steam. 1. S 77

78 RENEWABLE POWER GENERATION COSTS Chile and China. Compared to other technologies, for generation to be shifted to times when the sun China’s share of CSP installations is quite modest is not shining or to maximising generation at peak th – and ranked 10 demand times. CSP with integrated storage can in the world at the end of 2016 thus be a cost effective, flexible option in different (IRENA, 2017a). The country has announced plans locations, especially in the context of increasing to increase CSP deployment, however, with the shares of VRE. (Lunz et al., 2016; Mehos et al., G W of CSP by 2020. This is half goal of installing 5 2015). a previously released goal of 10 W, though. In G September 2016, China released information on a Cumulative CSP capacity grew tenfold worldwide first group of CSP demonstration projects, some between 2006 and 2016 (Figure 4.1). Growth rates of which have already been implemented, albeit have been linked in the past to incentive schemes at slower pace than expected (SolarPV.TV, 2016; in key markets. During the 2000s, support policies SolarPACES, 2016; Wang et al., 2017). drove early CSP expansion, primarily in the United States and Spain, and these two countries account GW Globally, at the end of 2016, an estimated 4 for more than 80% of the total cumulative installed of CSP projects were under construction or under G W of CSP capacity between them. At about 5 development (SolarPACES, 2017a). This data cumulative installed capacity, compared with other should be treated with caution, as projects can be renewable energy technologies, CSP deployment abandoned or delayed in the planning or project remains modest. development stages for a variety of reasons. As an example, the subset of projects in the SolarPACES Since 2013 in particular, new projects and plans database for which the planning status has been have started to proliferate in new and emerging recently revised (that is to say their status was markets. Many of these have high irradiation 2 revised in the period 2015-2017) is close to 3 GW. levels, or major renewable energy adoption plans Figure 4.2 shows the capacity of these more that include CSP, or both. These markets include recent projects, broken down by technology and India, South Africa, Morocco, the UAE, Australia, operational status. Development of the cumulative installed CSP capacity by region, 2006-2016. Figure 4.1 5 100% 90% 4 80% 70% 3 60% 50% GW 2 40% 30% 20% 1 10% 0% 0 2012 200620082010201220142016 200620082010 20142016 Middle East Eurasia North America Oceania Europe Asia Africa Source: IRENA, 2017a. 2. A t the time of writing, information as to whether some of the earlier projects categorised under these headings will be realised was . unavailable 78

79 2017 Status of planned PTC and ST projects registered since 2015 Figure 4.2 3.0 2.5 2.0 1.5 1.0 Gross capacity (GW) 0.5 0.0 Parabolic trough, under construction Total, under construction Power tower, under construction Total, under development Power tower, under development Parabolic trough, under development Source: IRENA analysis based on SolarPACES, 2017b. 79

80 RENEWABLE POWER GENERATION COSTS very early plants built in California in the 1980s, 4.1 INSTALLED COST TRENDS capital costs for PTC without storage started to Total installed costs for CSP plants that include increase as projects shifted to Spain. Projects from thermal energy storage tend to be higher than those this era in the IRENA Renewable Cost Database, without, but storage also allows for higher capacity range in costs between USD 300/kW 3 650 and 11 factors. For example, for parabolic trough systems (Figure 4.4) for the period of 2009-2013. (the technology with the highest share of installed There was also a strong capital cost increase for projects so far), total installed plant costs can PTC without storage during the period 2008-2011. range between USD 11 550 and USD 2 5/kW for 26 This increase could in part be explained by the systems with no storage. Adding four to eight hours comparatively lower solar resources in the project of storage, however, can see this range increase locations in Spain, but analysis allowing for Direct to between USD 50/kW for 50 and USD 6 1 13 0 Normal Irradiance (DNI) suggests that at least 65% projects for which cost data is available in IRENAs of the cost increase ought not to be attributed Renewable Cost Database for the period 1984-2016. to the lower solar resources, but to fundamental (Figure 4.3). cost increases in the configuration (Lilliestam et A time series of such projects from 2009-2016 al., 2017). Figure 4.4 shows the narrower range of shows that PTC and ST CSP capital costs for 000/kW that can 2 55 0 and USD 7 between USD systems with no storage displayed a wide range be observed for the ‘no storage’ configuration in during the period, varying between USD 2 0 55 the IRENA Renewable Cost Database for more 300/kW. The majority of these projects and USD 11 recent PTC and ST plants, installed since 2014. started operating between 2009 and 2013 in Spain Parabolic trough and solar tower projects with and still benefitted from, or where conceived up to four hours of storage show a range of under, the generous Spanish FiT incentive of that total installed costs between USD 3 5 00 and time that kickstarted this second phase of CSP 000/kW (though projects of this kind with USD 9 development. After a downward trend from the Installed costs and capacity factors of CSP projects by their quantity of storage, 1984-2016. Figure 4.3 18 000 Capacity MWe 1 100 200 300 377 12 000 Type Linear Fresnel Parabolic trough 2016 USD/kW Solar tower 6 000 Storage (hours) No storage 0-4 4-8 8+ 0 0.15 0.50 0.200.250.300.350.400.45 0.550.600.65 Capacity factor Source: IRENA Renewable Cost Database. 80

81 2017 Figure 4.4 CSP installed costs by project size, collector type and amount of storage, 2009-2016 18 000 Type Linear Fresnel Parabolic trough Solar tower 12 000 Storage (hours) No storage 0-4 4-8 2016 USD/kW 6 000 8+ 0 20092010201120122013201420152016 Capacity MWe 100 300 1 200 377 Source: IRENA Renewable Cost Database. increased use of molten salt as the HTF, compared MW of capacity for which costs data larger than 50 to the subset of projects in operation. Though data is available were only installed from 2015 onwards). is not available for all projects, it seems that some In the case of PTC and ST plants with four to ST plants are planned to operate with a water- or eight hours of storage, capital costs ranged from steam-based HTF configuration, these can provide 600/kW. Between 2013 and 6 050 and 12 USD efficiency gains, but are not suitable for use with 2015, PTC and ST projects with storage capacities large-scale storage. Most solar tower plants ‘under larger than eight hours were installed at a range of construction’ or ‘under development’, however, costs between USD 7 300 and 11 300/kW. Despite are poised to continue to use of molten salt as the a somewhat irregular market growth, a trend HTF. The dataset also suggests that about 10% of towards plant designs with higher hours of storage PTC planned capacity is also going to use molten can be inferred from the IRENA dataset. It can also salt as its HTF with its associated benefits of be confirmed by analysing the storage design higher operating temperatures (thermal oil is not configuration for projects ‘under construction’ or 3 suitable for operating temperatures in excess of ‘under development’ in the SolarPACES database. 400°C) and hence higher steam cycle efficiencies For PTC projects, an average 7.6 hours of storage compared to when mineral oils are used as the HTF is planned, while for solar towers, project designs 4.6). (Figure are for nine hours of storage or more (Figure 4.5). Even though CSP deployment has been somewhat The SolarPaces database also provides some limited compared to other renewable power insight into trends in heat transfer fluid usage for generation technologies, there exist significant the two main CSP technologies. Data for planned opportunities for cost reductions as deployment projects with recently updated operational grows (IRENA, 2016a). These cost reduction status in the database suggests a trend towards 3. That is to say, where the project status information was updated during the 2015-2017 period. 81

82 RENEWABLE POWER GENERATION COSTS Storage hours of planned CSP projects with operational status updates in 2015-2017 Figure 4.5 Parabolic trough Solar tower 10 9.4 8 7.6 6 4 Storage capacity (hours) 2 0 Under development Under development Under construction Under construction IRENA analysis based on SolarPACES, 2017b. Figure 4.6 Heat-transfer fluid use in operational and planned projects with operational status updates in 2015-2017 Under construction & Under construction & Operational Operational under development under development 100% 3 80% 2 60% 40% Gross capacity (GW) 1 20% 0 0% Parabolic Parabolic Parabolic Solar Solar Parabolic Solar Solar trough tower trough tower tower trough tower trough Not avaible Dipheny/Biphenyl oxide (synthetic oil) Water & water/steam Molten salt IRENA analysis based on SolarPACES, 2017b. 82

83 2017 A clear trend towards higher Direct Normal potentials will enable CSPs market presence to grow Irradiance (DNI) values of commissioned CSP and for this technology to contribute substantially projects can be observed after 2012, albeit from to the global energy transition towards a low relatively thin deployment data. For instance, the carbon future. Technological improvements in capacity weighted average DNI value for projects solar field elements, such as collectors and mirrors, for which data is available increased 11% between reduced costs in installation and engineering, 2 2012 and 2013 and exceeded 2 h/m kW 800 and cost reductions in specific components are /year expected for CSP. The technology is also expected in both 2014 and 2015. During 2016 it remained to experience declines in its indirect costs and the about one fifth higher than in 2012. owner’s cost elements, with slightly higher cost reduction potential in these items for solar towers, 4.3 OPERATION AND MAINTENANCE COSTS compared to PTC. This can be explained with reference to the lower deployment of solar towers CSP O&M costs are a significant component of the so far. With larger deployment, the risk margins of overall LCOE of CSP projects (IRENA, 2016a). They suppliers and EPC contractors would also fall, as have been falling through time and are significantly developers and other players gain more experience lower today than the original, pioneering Solar (IRENA, 2016a). Electricity Generating System (SEGS) plants that were built between 1982 and 1990. The SEGS plants Learning rates (the cost decrease with every were estimated to have had O&M costs of around doubling of cumulative capacity) for CSP have USD 0.04/kWh (Cohen, 1999), with expenditure been previously estimated to be between 10% for replacement receivers and mirrors being one and 12% (Neij, 2008); (Haysom et al., 2015); of the largest cost components, as a result of glass (Fraunhofer ISE, 2013). However, recent analytical breakage. work suggests higher learning rates for CSP since 2013, with an estimated learning rate above 20% Advances in materials and new designs have (Lilliestam et al., 2017; Pitz-Paal, 2017). If the helped to reduce the failure rate for receivers, to the auction results for Dubai and South Australia are point where mirror receiver breakage is no longer factored in, then for the period 2010-2022 the a large cost component. However, the cost of learning rate could reach 30%. mirror washing, including water costs, is However, the significant. Plant insurance can also be an important expense, with its annual cost potentially 4. 2 CAPACIT Y FACTORS 1% of the initial capital outlay. Even ‑ between 0.5 higher costs are possible in particularly unsecure The evolution of the capacity factors in the locations. IRENA Renewable Cost Database is presented in 4.7. Capacity factors have increased over Figure More recent projects built in Spain, the United time as a shift towards newer technologies, with States and elsewhere are estimated to have lower larger thermal storage capacities has coincided O&M costs than those of the SEGS plants, however. with a trend towards the growth of markets in On the basis of available, bottom-up, engineering higher irradiation locations. The dominance of estimates (e.g., Turchi, 2010a and Turchi, 2010b) Spanish CSP projects, often with no storage and recent proposed projects (Fichtner, 2010), capacity, has given way to projects with significant O&M costs can be estimated to be in the range of levels of storage, often in locations with higher 04/kWh (including insurance). 02 to USD 0. 0. USD DNI than in Spain, notably as projects in Morocco, The IRENA CSP cost analysis used in this report Chile, South Africa and the United Arab Emirates assumes an insurance-included average O&M cost have come online. The evolution of DNI of projects range of USD 03/kWh for PTC and 02 to USD 0. 0. is presented in Figure 4.8. For CSP plants, the 0. 0. 03 to USD 04/kWh for ST (IRENA, 2016a). USD irradiation level at the plant location (typically referenced by the DNI metric) is inversely correlated to the LCOE (IRENA, 2015). 83

84 RENEWABLE POWER GENERATION COSTS Capacity factor trends for CSP plants, 2009-2016 Figure 4.7 0.7 Capacity MWe 1 Weighted average capacity factor 100 200 0.6 300 377 0.5 Type Linear Fresnel Parabolic trough 0.4 Solar tower Capacity factor Storage (hours) 0.3 No storage 0-4 4-8 0.2 8+ 0.1 20092010201120122013201420152016 Source: IRENA Renewable Cost Database. Direct normal irradiance levels for CSP projects by year of commissioning and technology, 2009-2016 Figure 4.8 3 000 2 800 2 600 /year 2 2 400 Weighted average kWh/m 2 200 2 000 1 800 20092010201120122013201420152016 Capacity MWe 300 100 1 200 377 Solar tower Parabolic trough Based on Lilliestam et al., 2017. 84

85 2017 4.4 LEVELISED COST OF ELECTRICITY Higher levels of irradiance The LCOE of CSP plants stayed relatively stable were likely the main factor between 2009 and 2012. Significant deployment during 2012, primarily in Spain (at least 800 MW), behind lower levelised costs along with a couple of projects in the United States and in a few other countries, coincided, however, with a widening of the LCOE range During 2016 the capacity weighted average in that year as more competitive plants were LCOE of CSP plants was estimated to be also commissioned (Figure 4.9). A downward 0.27/kWh (a fifth lower than in 2009) USD trend in LCOE started in 2012. Indeed, during although IRENA data suggests that the LCOE, 2013 and 2014, the LCOE estimates were, on 0.22/kWh. although about 18% during 2017 to USD average, about one fifth lower than those of The LCOE estimates discussed in this section the 2009-2012 period. This decrease, coincided assume a 25-year economic life and a WACC with a geographical shift away from Spain to of 7.5% in OECD countries and China, and 10% newer markets with higher solar resources and elsewhere. Apart from increased DNIs at project sometimes, lower installed costs. Higher levels of locations between 2012 and 2014, the downward Direct Normal Irradiance (DNI), were , however, LCOE trend observed during 2012 and 2014 can be likely the main factor behind lower levelised explained by a similar upward trend in the capacity costs during that period (Lilliestam et al., 2017). factor of plants. The growth in the capacity factors Learning effects and technology improvements of CSP during this period is not only related to have not yet, therefore been the main driver of higher solar resource availability, but also due to cost reductions, leaving significant cost reduction plant configurations with higher storage capacities potentials to be unlocked as already highlighted and dispatching abilities. (IRENA, 2016a). The levelised cost of electricity for CSP projects, 2009-2016 Figure 4.9 0.6 Type Linear Fresnel Parabolic trough 0.5 Solar tower Storage (hours) No storage 0.4 0-4 4-8 0.3 8+ 2016 USD/kWh 0.2 20122013201420152016 200920102011 Capacity MWe 1 100 300 377 200 Weighted average LCOE Source: IRENA Renewable Cost Database. 85

86 RENEWABLE POWER GENERATION COSTS These results should be treated with caution, as From 2014-2016, thin deployment makes it difficult to come to definitive conclusions regarding the a direct comparison with project level LCOEs is complicated given PPA prices and LCOE often LCOE trend, but a range between USD 0.14 and do not represent a like-for-like comparison due USD 0.35/kWh can be observed for projects for to auction prices being dependent on a set of which cost data is available in the IRENA Renewable obligations and terms in the contract that can Cost Database. The LCOE of most projects in be very market and project specific. The other 0.30/kWh (Figure 4.9). this period is below USD important point to take into consideration is Recent announcements and analysis of planned that these prices apply to projects that will be projects seems to predict a clear downward trend, commissioned in the period 2020-2022 and too, starting in 2017 (Lilliestam et al., 2017). Indeed, beyond. However, these announcements do recently, very low bids for CSP projects have been point towards the increased competitiveness of announced. Examples include the USD 073/kWh 0. renewable energy projects compared to fossil fuel bid announced by the Dubai Electricity and Water alternatives and that by 2020 commissioned CSP Authority (DEWA) for a 700 MW plant at the plants will increasingly be delivering electricity at Mohammed bin Rashid Al Maktoum Solar Park a cost that is within the lower end of the fossil fuel- (DEWA, 2017) and the Port Augusta CSP project fired cost range (Figure 4.10). 0.06/kWh. in Australia, at around USD Figure 4.10 Levelised cost of electricity and auction price trends for CSP, 2010-2022 0.4 0.3 2016 USD/kWh 0.2 0.1 0.0 201020112012201320142015201620172018201920202022 United Arab Emirates India Morocco South Africa France LCOE Database Kuwait China Spain Auctions Database United States Australia Note: Each bubble represents a renewable energy project. The center of the bubble is the winning bid price in that year. Source: IRENA Renewable Cost Database and Auctions Database. 86

87 2017 87

88

89 5. WIND POWER costs have fallen as wind turbine prices have n 1979, Danish and German manufacturers Vestas, declined from their peak in 2008-09. Balance of Nordtank, Kuriant and Bonus ushered in wind I project costs have also declined, with these factors power’s modern era with the mass production of all driving down the LCOE of wind and spurring large wind turbines to produce electricity. These increased deployment. Yet there are significant cost early wind turbines had small capacities by today’s differentials between countries. Comprehensive ‑ standards – 10 30 ut they have scaled up kW – b data on installed costs and market performance rapidly, as the modern wind power industry has are crucial to understanding the current cost grown and matured. of electricity and opportunities for future cost Wind power technologies have two main reductions from performance improvements and characteristics: the axis of the turbine and the installed cost reductions. location. The axis of the turbine can be vertical From 2000 to 2016, cumulative installed wind or horizontal and the location can be onshore or capacity increased at a compound annual rate of offshore. Virtually all onshore wind turbines are 15%, and by the end of 2016, total installed wind horizontal axis turbines, predominantly using three G G W, with 454 W capacity had reached 467 blades and with the blades “upwind”. The utility- onshore (IRENA 2017b). China has the largest scale market for wind technologies uses almost share of this – 32% at the end of 2016 – followed by exclusively horizontal axis turbines, both onshore United States (17%), Germany (11%), India (6%) and and offshore. Spain (5%). China accounted for 38% of new annual The amount of electricity generated by a wind capacity additions in 2016, followed by United turbine is determined by nameplate capacity (in States (17%), Germany (10%), India (7%), Brazil kW or MW), the quality of the wind resource, the (4%) and France (3%). Net additions of wind power height of the turbine tower, the diameter of the were 21% lower in 2016 than in 2015, a record year rotor and the quality of the O&M strategy. Wind G W was added to global capacity. This in which 65 turbines typically start generating electricity at a was mainly due to policy changes in China, which wind speed of 3-5 metres per second (m/s), reach drove a rush of installation before the expiration maximum power at 11-12 m/s and generally cut out of a policy support scheme at the end of 2015. at a wind speed of around 25 m/s. China added 42% less capacity in 2016 compared to 2015, accounting for almost all the global Wind power has experienced a somewhat difference between 2016 and 2015. The range of unheralded revolution since 2008-09. Between expected yearly additions in the next 3-5 years 2008 and 2017, improved technologies – such W. China, the United States, Germany, G is 40-50 as higher hub heights and larger areas swept by India, and France are expected to account for the blades – have increased capacity factors for a majority of new additions (MAKE, 2017). given wind resource. At the same time, installed 89

90 RENEWABLE POWER GENERATION COSTS number of offerings in their portfolio since 2010, 5.1 WIND POWER TECHNOLOGY TRENDS with each now offering over 20 models. This also The largest share of the total installed cost of a wind helps to reduce costs below what they would project is related to the wind turbines. Contracts for otherwise be, as utilising the same structural these typically include the towers, installation, and components across a given platform can mean up delivery, except in China. The range of the share of to 50% of the turbine components are identical, wind turbines in total installed costs has historically significantly reducing development costs and varied from 64-84% for onshore wind and 30-50% unlocking supply chain efficiencies (MAKE for offshore wind (IRENA Renewable Cost Consulting, 2015b). Database; Blanco, 2009; EWEA, 2009; Douglas- One of the key drivers of the increasing Westwood, 2010; and MAKE Consulting, 2015a). In competitiveness of wind power has been major markets, as costs have fallen, the share of continued innovation in wind turbine design wind turbines has tended towards the higher end and operation (IRENA, 2016a). There has been of this range. a continuous increase in the average capacity of Five major cost categories drive the total installed turbines, hub-heights and swept areas as blade costs of a wind project: lengths have grown. These trends work together • in synergy to reduce the cost of electricity from Turbine cost: Rotor blades, gearbox, generator, wind power. Higher hub-heights allow turbines nacelle, power converter, transformer and 2 to access higher wind speeds, tower. while larger swept areas from longer blades also increase the • Construction works for the preparation of the yield of a wind turbine. Higher turbine capacities site and foundations for the towers allow larger projects, which can amortise project • development costs over a larger output. The Grid connection: Includes transformers and trade-off for these developments is that taller substations and connection to the local towers supporting greater weight typically cost distribution or transmission network. more, so the impact in some markets may be • Planning and project costs: Depending on cost-neutral for installed costs, but result in a project complexity, these can represent a lower LCOE due to the higher yields. The other significant share of the balance of project costs challenge is that longer blade lengths come with 1 (i.e. the non-turbine costs). additional engineering challenges, as loads on • turbines increase significantly with longer blades, Land: Cost of land represents one of the smallest thus necessitating a different structural design. shares of total costs. Land is usually leased They also present a logistical challenge onshore, through long-term contracts in order to diminish given their sheer length. Research into very long the high administrative costs associated with segmented blades is therefore ongoing, but for land ownership, but it is sometimes purchased large projects road upgrades may prove a cheaper outright. option than investing in segmented blades. One of the important trends in the wind market Demand for the latest turbine technologies is is the larger range of wind turbines offered by being driven by Europe, where space constraints manufacturers to allow developers to choose and siting challenges mean that profitability designs that yield the lowest LCOE for the site rests heavily on using the highest performing constraints they are facing. General Electric, technologies. Crucially, taller towers in European Siemens and Vestas have all roughly doubled the 1. T hese include costs such as: development costs and fees, licenses, financial costs, development and feasibility studies, legal fees, right of way, insurance, debt service reserves, and construction management not associated with the engineering, procurement and construction contract. 2. W ind farm economics are significantly enhanced by accessing higher wind speeds given that the yield increases by a power of three as a function of wind speed. 90

91 2017 having increased average nameplate capacity by markets allow for the exploitation of marginal wind 79% between 2010 and 2016 and rotor diameter sites and existing forested land that is available for by 53%. Canada, and to a lesser extent, the United development (MAKE Consulting, 2013 & 2017b). The States, are interesting examples of markets that rapid development of wind turbine technologies have increased the rotor diameter faster than has seen the most advanced turbine designs the nameplate capacity. Between 2010 and available change rapidly. In 1985, typical turbines 2016, the rotor diameter of newly commissioned had a capacity of 50 nd a rotor diameter of kW a projects increased by 47% in Canada and 22% in 15 metres (UpWind, 2015). In 2016, offshore wind the United States, while the growth in nameplate W capacity with a rotor diameter M turbines of 8 capacity was 7% and 13% respectively. Overall, of 164 metres were in operation, while a 9.5 W M the largest increases in rotor diameter occurred version of the same turbine is now available. in Ireland (53%), Canada (47%) and Germany Figure 5.1 presents the evolution of wind turbine (36%). In percentage terms, the largest increase rotor diameter and nameplate capacity between in nameplate capacity was observed in Ireland, 2010 and 2016 for countries where data is available. followed by Germany (42%) and Denmark (42%). The ongoing trend towards larger turbines with greater swept areas is clear. Ireland stands out, Weighted average rotor diameter and nameplate capacity evolution, 2010-2016 Figure 5.1 120 100 2016 80 2010 Rotor diameter (m) 60 40 2 3 1 Nameplate capacity (MW) Brazil China Denmark Germany France Canada United States India Turkey Sweden Ireland Sources: Wiser and Bollinger, 2017; Global Data, 2016; Danish Energy Agency, 2017; IEA Wind, 2015; CanWEA, 2016. 91

92 RENEWABLE POWER GENERATION COSTS 95 metres declined by 53% between 2009 and 5.2 WIND TURBINE COSTS 2017, while the index for diameters greater than Wind turbine prices fluctuate with demand and etres declined by 41%. This value is in line m 95 supply, as well as with economic cycles. The latter with the decline observed in the average selling can affect the cost of the materials used in wind price for Vestas wind turbines over the period, at turbine manufacturing, as these have a significant 48%, and close to values observed in the United exposure to commodity prices – notably those States, for the vast majority of contracts. Chinese of copper, iron, steel and cement – given these wind turbine prices peaked in 2007 and have fallen account for a sizeable part of the final cost of a 37% between 2007 and 2016 – but started from wind turbine. lower levels, thus having slightly less room for cost 3 declines. Wind turbine prices reached a low in the period The decline in turbine prices globally 2000-2002, but prices then increased, as has occurred at the same time as improved wind commodity prices rose, turbine supply tightened turbine technology: rotor diameters, hub heights, and the growth in larger, higher performing and nameplate capacity have all increased turbines accelerated. During 2000-2002, the markedly. average turbine price in the United States Provisional data for 2017 indicates that average wind was at its lowest, at around USD 800/kW turbine prices across most, if not all, markets were 2 100/kW 000 to 2 and peaked at around USD below USD 1 000/kW by the year’s end. The last in 2008, (Wiser and Bollinger, 2017). In Europe, time this happened, in 2002, was when the most 900/kW for 1 average prices peaked at around USD kW 000 common installed turbine was in the 750-1 contracts signed in 2008/2009 (BNEF, 2017). range. Contracts for onshore wind turbines signed Depending on the market and technology in 2017 were for a weighted average turbine rating segment, wind turbine prices peaked between of 2 400-2 NEF, 2017;a;b;c). This is in 800 kW (B 2007 and 2010 before starting to decline addition to the fact that more favourable terms 5.2). The cost increase was driven by three (Figure are now often being extracted from turbine factors. Firstly, the increase in construction costs, manufacturers. These can include shorter delivery with materials (e.g. steel, copper, cement), labour lead times, more generous initial O&M contracts, and civil engineering costs all rising prior to the better performance guarantees and a reduced 2009 financial crisis. Secondly, for a few years, need for the order to be part of a larger framework demand outstripped supply as many countries agreement (Wiser and Bollinger, 2017). adopted policies favourable to wind deployment. The drivers of wind turbine price declines since This allowed manufacturers to operate with higher 2007-2010 have been falling commodity prices, margins, as they struggled for a time to meet rising greater supply chain competition, manufacturing demand. Lastly, technology improved markedly; a economies of scale and process improvements; trend that has continued ever since: wind turbine transforming the global market into one more manufacturers introduced larger, more expensive favourable for buyers. Competition has also turbines, with higher hub heights. As a result, more increased in the wind turbine market. In 2016, the capital-intensive foundations and towers were manufacturer with the largest share of global new needed, but helped deliver higher energy outputs, capacity installed accounted for just 16.5% of total largely offsetting the higher installed costs and installations (BNEF, 2017). Indeed, competition has hence delivering a lower LCOE. heightened to such an extent in the last few years Bloomberg New Energy Finance's (BNEF) index that consolidation in the sector is gathering pace for turbines with rotor diameters of less than (Reuters, 2015, 2016 and Bloomberg, 2015). 3. C hinese wind turbine prices are not directly comparable to the other indices, as they don’t include towers or transportation, which are included in their engineering procurement and construction contracts. 92

93 2017 Wind turbine price indices and price trends, 1997-2017 Figure 5.2 2 500 2 000 1 500 2016 USD/kW 1 000 500 0 Feb-17 Feb-12 Feb-13 Feb-15 Feb-18 Feb-16 Feb-14 Feb-10 Feb-01 Feb-11 Feb-97 Feb-07 Feb-98 Feb-99 Feb-02 Feb-03 Feb-05 Feb-08 Feb-09 Feb-06 Feb-04 Feb-00 BNEF WTPI United States <5 MW Chinese turbine prices United States 5-100 MW BNEF WTPI <95m Ø United States >100 MW BNEF WTPI >95m Ø Vestas average selling price Period% Decrease BNEF WTPI 2009-2017 49% BNEF WTPI <95m Ø 2009-2017 53% 2009-2017 41% BNEF WTPI >95m Ø 2007-2016 Chinese turbine prices 37% United States 5-100 MW2010-201544% United States <5 MW2008-201121% United States >100 MW2008-201656% Vestas average selling price2008-201748% Sources: Wiser and Bollinger, 2017; CWEA, 2013; BNEF, 2016; Global Data, 2014 and Vestas Wind Systems, 2005-2017. 93

94 RENEWABLE POWER GENERATION COSTS installed cost reductions of different countries 5.3 TOTAL INSTALLED COSTS ONSHORE show a range of declines, from 30-68%. As is clear In the past 30 years, onshore wind installed costs from this comparison, on an installed cost basis, have declined significantly, according to IRENAs the shift in deployment to the most competitive database of onshore wind power project costs countries has resulted in a larger global weighted 4 from 1983-2016. The estimated global weighted average cost reduction than has been seen in any average fall in total installed cost of wind farms one country (Figure 5.4). For those countries with between 1983 and 2017 was 70%, as costs fell data available from 1983-2016, installed costs fell from 477/kW. This represents 880 to USD 1 4 USD by the most (68%) in the United States and the a learning curve of 9% for total installed costs for least (53%) in Denmark. For the group of countries every time installed capacity doubled, worldwide that started deployment at the end of the 1980s, (Figure 5.3). the fall in installed costs for the period 1989-2016 ranged from a high of a 52% reduction in Spain to a Depending on the country, the start date for first low of a 30% reduction in the United Kingdom. For commercial deployment varies, complicating a the group of countries where data is available for simple comparative analysis. Nonetheless, the Total installed costs of onshore wind projects and global weighted average, 1983-2017 Figure 5.3 6 000 5 000 4 000 3 000 2016 USD/kW 2 000 1 000 0 1987 1997 1983 1992 1985 1993 2017 1995 2012 2013 1988 2015 1989 1998 1986 1990 1991 2007 1984 2001 1996 1999 2000 2016 1994 2002 2014 2003 2011 2005 2008 2010 2009 2006 2004 ≤ 1 100200 ≤ 300 Capacity MWe Source: IRENA Renewable Cost Database. This dataset covers more than 85% of all onshore wind capacity installed at the end of 2016. 4. 94

95 2017 of how efficient total installed cost structures the period 1991-2016, total installed cost reductions are (e.g., due to logistics and installation, where ranged from a high of 67% in India to a low of a a shortage of specialised equipment can raise reduction of 38% in Brazil. costs). Cost ranges also represent the natural In terms of trends in total installed costs, there variation of renewable power projects, given the is still a wide range of individual project costs site-specific characteristics that can influence within a region. In some cases, this represents the total installed costs. These characteristics include differences between countries, where the maturity items such as the level of existing infrastructure of local markets can be an important determinant to enable access to sites, the distance from ports Figure 5.4 Onshore wind weighted average total installed costs in 12 countries, 1983-2016 1983-2016 United States Denmark Germany Sweden 6 000 4 000 -50% -53% 2 000 2016 USD/kW -53% -68% 0 1983 2016 1983 20161987 20161984 2016 1989-2016 Canada Spain Italy United Kingdom 6 000 4 000 -32% -30% -52% 2 000 2016 USD/kW -44% 0 20161990 20161990 1989 2016 2016 1989 1991-2016 China India France Brazil 6 000 4 000 -38% 2 000 -48% 2016 USD/kW -67% -56% 0 1991 20161991 20161996 20162001 2016 Source: IRENA Renewable Cost Database. 95

96 RENEWABLE POWER GENERATION COSTS competitive onshore wind installed costs, with or manufacturing hubs, the distance from a major a weighted average of USD 1 775/kW in 2016. grid-interconnection point, labour costs, and many Between 2010 and 2016, costs fell by 36% in others. Overall, however, from 2010-2016, total Oceania, 22% in North America, 19% in Europe and installed costs decreased significantly across the between 13% and 19% in other regions (Figure 5.5). nine geographical regions covered, while ranges also diminished across all regions. In association with its Renewable Costing Alliance partners, IRENA conducted additional research in The lowest installed costs for onshore wind 2017 to collect cost breakdown data and O&M data projects are to be found in China and India, with for a range of projects in the IRENA Renewable weighted average total installed costs estimated to Cost Database. This has yielded a subset of projects be USD 45/kW and USD 21/kW respectively in 2 1 1 1 for onshore wind based on a consistent collection 2016 which translates into a decline of 11% and 16% methodology. The subset includes data for 448 respectively from 2010. Weighted average installed onshore wind projects commissioned between costs have declined in Brazil from USD 90/kW 3 2 GW of the 2006 and 2017. These represented 17.2 94/kW in 2016. In terms of regions, 9 in 2010 to 1 capacity deployed in 15 countries, stretching from Asia (excluding China and India), Oceania, Central Asia to South America. The data collected should be America and the Caribbean and South America treated with caution, as it may not be representative (excluding Brazil) are the most expensive regions, for all countries and regions in each year. with weighted averages of between USD 884 1 and USD 2 256/kW in 2016. North America has Total installed costs ranges and weighted averages for onshore wind farms by country/region, Figure 5.5 2010-2016 4 000 3 000 2 000 2016 USD/kW 1 000 0 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 ChinaIndiaBrazil Africa Asia (excl. China and India) Central America and the Caribbean South America (excl. Brazil) North America Europe Eurasia Oceania Source: IRENA Renewable Cost Database. 96

97 2017 ranged from 5-23% in China, 8-12% in India and Figure 5.6 presents the evolution in cost 8-10% in the rest of the world. Grid connection breakdown for onshore wind projects in India, costs also accounted for a significant share of China, Germany and the rest of the world, but only costs, with a maximum of 11% in the rest of the for this subset of data. Wind turbines represent world and a minimum of 3% in China. The figures the highest share of costs in all regions, ranging for Germany are presented using slightly different from 66-84%, with this changing little over time, categories, as data was not available on the same except in China. There, wind turbines accounted 5 basis as that of other countries and regions. In for 84% of costs in 2010, falling to 72% in 2014. Germany, wind turbines ranged from 75-78% of Civil works associated with the development costs, depending on the year, while the share of of the wind farm, access and grid connection, grid connection costs diminished over time, from represent the second highest share of costs in 10% in 1998 to 5% in 2012. onshore wind projects. From 2010-2014, this share Cost breakdown of onshore wind farms by country and region, 1998-2016 Figure 5.6 India China 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% 20092010201120122013201420152016 2011201220132014 2010 Rest of the world Germany 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% 2008 201020112012201320142015 199820012010 2012 Civil works Grid connection Development Grid connection Foundation Wind turbine Other Wind turbine Land Planning Planning Source: IRENA Renewable Cost Database and DWG, 2015. his sharp change in the share in 2011 is unlikely to be statistically significant and maybe the result of a dataset for 2011 that is not T 5. representative of the overall market. IRENA will endeavour to identify additional data for 2011 to confirm or reject this hypothesis. 97

98 RENEWABLE POWER GENERATION COSTS Figure 5.8 presents the distribution and weighted Figure 5.7 uses the data available for cost average of five main cost items in onshore wind breakdowns from Figure 5.6, with the overall projects in the subset of data covering 448 projects weighted average change in the total installed for which IRENA has been able to collect consistent cost of each of the markets from the complete data. This is stated as a share of the total installed set of country data in the IRENA Renewable Cost costs for the project, where cost breakdown data Database to examine what has been driving the was available. The weighted average share of wind reduction in total installed costs. For India and turbine costs in China and India was around 73%, the rest of the world, this is for the period 2010- falling to 69% in the rest of the world data sub-set. 2015, while for China, it is for 2010-2014, and for The share of civil works in total installed costs was Germany, from 1998-2012. Over these time periods, higher in the rest of the world (15%) than in China total onshore wind costs declined by 29% in India, and India (12%). Most significantly, on average, the 13% in China, and 34% in the rest of the world. In share of grid connection costs was almost twice Germany, prices increased slightly, by 3%, due as high in rest of the world at 9% than in China to higher prices for the more advanced turbines and India at 5%. The cost of land in China and India being used in Germany, when compared to the makes up a modest share of total installed costs at other markets, but also because this subset of data 3% in China and India and 1% in rest of the world. for Germany ends before the cost reductions seen in 2013 to 2016. The main drivers in China, India and the “Rest of the World” were declines in wind turbine costs, followed by declines in civil works, grid connection and planning and other project development costs. Figure 5.7 Average total installed cost reduction by source for onshore wind, 2010-2014/15 and 1998-2012 Rest of the world China Germany India +3% 100% -13% 80% -29% -34% 60% 40% 20% 0% 2015 1998 2010 2010 2010 2015 2012 2014 Other Planning Civil works Civil works Civil works Foundation Cost of land Cost of land Cost of land Development Wind turbines Wind turbines Wind turbines Wind turbines Grid connection Grid connection Grid connection Grid connection Planning and other Planning and other Planning and other Source: IRENA Renewable Cost Database, 2017. 98

99 2017 Distribution and weighted average share of onshore wind total installed costs by source for China Figure 5.8 and India, and rest of the world, 2006-2016 Rest of the world China and India 100% th 95 percentile 80% th 5 percentile 60% 40% 20% 0% Grid Grid Planning Planning and other and other Civil works Civil works connection connection Cost of land Cost of land Wind turbine Wind turbine Source: IRENA Renewable Cost Database, 2017. projects can also significantly increase the costs 5.4 TOTAL INSTALLED COSTS OFFSHORE of construction, as well as grid connection due In comparison to onshore wind projects, offshore to the expense of deploying undersea cables and wind farms have significantly higher lead times. working further from a port on the installation. Planning for offshore wind farms is more complex Increased costs to protect equipment and and construction even more so, increasing total installations from the harsh marine environment installed costs. Given their offshore location, they also add to the final costs. These can be profitable also have higher grid connection and construction incremental investments if they mitigate costly costs. Offshore wind project installed costs rose unplanned maintenance interventions. Operation in the period to around 2012-13, as projects were and maintenance costs are higher for offshore wind sited farther from shore and have been using more than for onshore wind, because of the complexity advanced technology. of servicing offshore wind turbines and the more challenging environment at sea. Yet, on average, Wind turbines for offshore wind projects account offshore wind projects harvest more energy than for somewhere between 30% and 50% of total onshore wind projects, notably in Europe, due installed costs, while foundations are also a to the availability of better wind resources, less significant part of total installed costs (IRENA, turbulence and steadier winds, overall. 2016a). The specific location of offshore wind 99

100 RENEWABLE POWER GENERATION COSTS with larger rotors and higher hub heights. These The global cumulative installed capacity of were increasingly specifically designed by offshore wind was 14 W at the end of 2016, or G manufacturers for the offshore market and the 3% of total installed wind capacity. Over 2013 to harsh marine environment in which they operate. 2016, the annual new capacity installed was above W, as the offshore market picked up a solid G 2 The reason for this trend towards larger turbines pace, predominantly in Europe. and longer blades designed for offshore operations was to increase capacity factors, as As the industry matured after 2000, projects developers accessed better quality wind resources moved to deeper water and farther from shore further offshore. Larger turbines can also help (Figure 5.9). Since 2009, most projects have been reduce installation costs and amortise project sited in water depths greater than 15 metres and development costs over larger wind farm capacities at a distance of 20-80 km from the nearest port. for the same physical area. Cost reductions begun The average size of grid-connected offshore wind to be unlocked as the industry increasingly farms in Europe in 2016 was 380 W, while the M standardised new wind turbines and industrialised average water depth of completed, or partially the manufacturing process. Installation methods completed, wind farms was 29 metres, with an and offshore construction vessels also became average distance to the nearest port of 44 km more sophisticated and more efficient, reducing (WindEurope, 2017). Developers also began using the time, and hence costs, of installation. larger turbines during the period in question, Figure 5.9 Offshore wind farm projects and distance from port, 2001-2017 Year of commissioning 2015 2001 80 60 40 Average distance from port (km) 20 0 10 30 40 0 20 Average water depth (m) 65 234 Turbine rating (MW) Source: IRENA Renewable Cost Database. 100

101 2017 Total installed costs of offshore wind From 2001-2010, most of the wind turbines for offshore projects were in the 2-3.6 W range. M Figure 5.10 presents the evolution of total installed After 2011, due to improvements in technology, costs for offshore wind projects from 2000-2016. the range increased significantly, to 3.6-6.15 MW. These rose in the period 2000-2010, as the shift Projects also became larger after 2011, thus allowing towards deeper waters and locations farther developers to benefit from economies of scale and from ports took place. Installed costs appear to to offset some the cost increases related to siting have peaked around 2012-2013, although better projects further ashore and in deeper waters. The wind resources accessed by better technologies average size of a European offshore wind farm moderated the impact of increasing installed costs was slightly below 200 W in 2011 while in 2016 M between 2000 and 2012-2013 on LCOE. Between MW, a 90% increase over the it had risen to 380 2010 and 2016, global weighted average installed entire period (EWEA, 2012; WindEurope, 2017). costs increased by 4%, up from USD 30 to 4 4 This trend has led to greater economies of scale, USD 4 487/kW. more competitive supply chains and O&M benefits In 2016, average installed costs for a European that have been part of the drivers of recent cost offshore wind farm were slightly higher than reductions. the global weighted average, at USD 4 697/kW Figure 5.10 Total investment costs for commissioned and proposed offshore projects, 2000-2018 7 000 6 000 5 000 4 000 2016 USD/kW 3 000 2 000 1 000 0 2020 2012 2004 2016 2000 2008 ≥ 800 600 1200400 Capacity MWe Source: IRENA Renewable Cost Database. 101

102 RENEWABLE POWER GENERATION COSTS 5.5 CAPACIT Y FACTORS (IRENA, 2016a). The turbine rotor and nacelle was estimated to account for 38% of total installed The capacity factors of wind projects are costs, construction and installation for 19%, the determined by the quality of the wind resource support structure and foundations for 18%, grid and the technology employed. There has been a connection/transmission for 13%, the turbine tower trend towards the use of more advanced turbine for 6% and project development and wind farms technologies as previously discussed. As a result, electrical array for 3% each. As is to be expected, there has been a consistent trend towards higher foundations account for a significant percentage capacity factors globally, but with significant of total costs, due to the expense of operating variations by market. This has been driven by the offshore and designing for the harsh marine growth in the average hub height, turbine rating environment. On average, these therefore account and rotor diameters of installed turbines, but also for around 18% of installed costs (IRENA, 2016a). by the trends in resource quality at new projects in This share can vary, however, and is influenced by individual markets. The global weighted average water depth, conditions on the seabed, turbine capacity factor for onshore wind increased loading, rotor and nacelle weight and the speed from around 20% in 1983 to around 29% in 2017 of the rotor. (Figure 5.11) – a rise of about 45%. The global Global weighted average capacity factors for new onshore and offshore wind power capacity additions Figure 5.11 by year of commissioning, 1983-2017 50% 40% 30% Capacity factor 20% 10% 0% 1983198619891992199519982001200420072010201320162019 Oshore WindOnshore Wind Source: IRENA Renewable Cost Database. 102

103 2017 additions commissioned in each year for the weighted average capacity factor for offshore wind 12 countries in IRENAs learning curve analysis. also increased by 56%, but from a higher starting Capacity factors doubled in the United States point. In 2017, the weighted average offshore and increased by more than 60% in Denmark and capacity factor for newly commissioned plants Sweden. The simple average increase in capacity reached around 42%, but given the relatively low factors for these 12 countries was around 43%. The volumes of projects being developed the newly United States stands out, however, as a market commissioned average for a given year has been where the trend towards higher capacity factors has quite variable. been driven not only by technology improvements, Figure 5.12 presents the historical evolution of but also the trend towards the location of projects onshore wind capacity factors for new capacity in areas with the best resources. Figure 5.12 Historical onshore wind capacity factors in a sample of 12 countries 1983-2016 Denmark Sweden Germany United States 40% +116% +64% +60% 30% +30% 20% Capacity factor 10% 0% 20161984 20161983 1983 2016 20161987 1989-2016 Spain Canada Italy United Kingdom 40% +42% +51% +12% +41% 30% 20% Capacity factor 10% 0% 20161989 2016 20161990 1989 20161990 1991-2016 India Brazil France China 40% 50% +32% 40% +21% 30% +31% +9% 30% 20% 20% Capacity factor 10% 10% 0% 0% 2016 1991 20162001 20161991 20161996 Source: IRENA Renewable Cost Database. 103

104 RENEWABLE POWER GENERATION COSTS 5.6 O PERATION AND MAINTENANCE COSTS Figure 5.13 focuses on the change in the weighted average capacity factor of onshore wind projects The global wind power O&M market is expected that were commissioned in 2010 and 2016 in a to grow from USD 12 billion in 2016 to more than range of countries. All countries for which data 2 bi llion by 2026 (MAKE Consulting, 2017c). 7 USD are available experienced a significant increase The biggest markets for O&M services are those in the weighted average capacity factor of newly countries with the greatest installed capacity commissioned projects between 2010 and 2016, – among which are China, the United States, ranging from a low of a 11% increase in the United Germany, India, Brazil and Spain. Operations and Kingdom to a high of 76% in Turkey. maintenance costs, both fixed and variable, are a Figure 5.14 presents the evolution of the global significant part of the LCOE of wind power. Yet, weighted average hub height, rotor diameter data for the actual O&M costs of commissioned and capacity factor. Hub heights increased from projects is not readily available. Where it is, care around 20 metres in 1983 to more than 100 must be taken in extrapolating from historical metres in 2016, while capacity factors increased O&M costs, as significant changes in wind turbine from 23% in 1983 to 28% in 2016 – more than 25% technology over the last decade must be taken over the entire period. This has been achieved as into consideration. Though data for maintenance installed capacity of onshore wind has increased is often available, the cost data for operations is exponentially, growing from 0.2 not systematically and uniformly collected (e.g. GW in 1983 to management costs, insurance, fees, land lease, more than 454 GW at the end of 2016. taxes etc.). Figure 5.13 Country-specific weighted average capacity factors for new onshore wind projects, 2010 and 2016 50% Percentage Country increase China 6% 9% France 40% 11% United Kingdom 12% Germany India 12% 13% Italy Canada 14% 30% 16% Spain Capacity factor 18% Sweden Denmark 20% 24% United States 28% Brazil 20% 38% Netherlands 76% Turkey 2016 2010 Source: IRENA Renewable Cost Database. 104

105 2017 Global weighted average hub height, rotor diameter and capacity factors, and cumulative capacity for Figure 5.14 onshore wind, 1983-2016 140 120 101 101 m m 100 diameter diameter 80 60 Hub height (m) m 51 51 m diameter diameter 40 17 m m 17 20 diameter diameter 0 30% 28% 25% 20% 20% 21% Capacity factor 15% 10% 454.4 400 200 17.3 0.2 0 Cumulative deployment (GW) 2016 2000 1983 Source: IRENA Renewable Cost Database. between 2008 and 2017, while full-service renewal O&M costs measured as initial full-service contracts varied from USD 44/kW/year. 22 to USD contracts are less expensive than full-service In the United States, O&M costs ranged from 5.15). A clear trend renewal contracts (Figure 1 USD U 3 t 6 o SD 7/kW/year in 2016, while the weighted cannot be extracted from the available data on 2 7/kW/year (MAKE Consulting, average was USD these two categories, however, as these costs 2017c). Data from IEA Wind for four countries shows vary depending on the year. Initial full service contracts varied from USD 1 3 4 to USD 0/kW/year 105

106 RENEWABLE POWER GENERATION COSTS Figure 5.15 Full-service (initial and renewal) O&M pricing indexes, weighted average O&M revenues of two manufacturers, and O&M costs in Denmark, Germany, Ireland and Sweden, 2008-2017 100 80 60 2016 USD/kW/year 40 20 0 2012 2008 2018 2016 2014 2010 Initial full-service contracts Weighted average revenues Full-service renewal contracts Ireland Sweden Germany Denmark Sources: BNEF, 2017; Global Data, 2017; IEA Wind, 2017. global average for fleet is slightly over six years old. that O&M costs declined in three out of four markets, Original equipment manufacturers (OEM) had the with high volatility in the Irish market in particular. largest share of the routine turbine O&M market The premium identified in the BNEF full service in 2016, with around 70%. By 2026, however, renewal contract index over the initial contract OEM’s market share is expected to decrease, as offers represents the additional expected costs the trend towards self-operation increases (MAKE as turbines age. This will become an increasing Consulting, 2017c). consideration for wind farm asset owners as wind Figure 5.16 presents the annual range of O&M costs farms begin to age. Given the rapid growth in in China, India and the rest of the world for the deployment, wind turbine fleets are still relatively 448 project subset in the IRENA Renewable Cost young. In 2016, Germany had one of the oldest Database. Reported O&M cost ranges are lower fleets in service, but was still just over 10 years old. in India than in China, but in both countries there In the United States the fleet was 8.5 years old, was a downward trend from 2010-2016. The bulk in China it was only around five years, while the 106

107 2017 Project level O&M cost data by component from a subset of the IRENA database compared Figure 5.16 to the BNEF O&M index range, 2008-2016 60 40 BNEF O&M Index - Full service contracts 2016 USD/kW/year 20 0 Central and South China India Rest of the America world 100% th percentile 95 80% 60% th 5 percentile 40% 20% 0% Salary Maintenance Other Material Source: IRENA Renewable Cost Database, BNEF 2017. 107

108 RENEWABLE POWER GENERATION COSTS In China, converting the USD/kW/year O&M cost of projects in China had O&M costs in the range of calculation to the USD/kWh range yields costs USD 2 to USD 7/kW/year during this period. The 4 2 0.028/kWh, while 0.008 to USD ranging from USD lower values for India are significant, but question the average is USD 0.017/kWh. In India, weighted marks over the comparability of reporting for all 0.005 to average O&M costs range from USD cost categories suggest that Indian data often USD 0.027/kWh. The weighted average O&M costs excludes some operations costs and is composed in the database for Central and South America is mainly of maintenance costs. Further analysis is USD 0.014/kWh, below that observed in China. required to confirm or reject this hypothesis. A Looking at the average share of O&M costs across smaller number of projects reported O&M costs all of the projects in the subset of data for which in other regions, and while this data spans a IRENA has detailed O&M data, the largest share wide range, the quantity of data is not sufficient, of O&M costs is represented by maintenance when compared to the data for China and India, operations, which have a weighted average of to suggest a definitive answer on where the bulk 67%, followed by salaries at 14% and materials at of project O&M costs range. The BNEF O&M index 7% (Figure 5.16). range has been included for comparison, but as already mentioned uncertainty about reporting of Table 5.1 presents data for O&M costs reported cost categories means that care should be taken in for a range of OECD countries. The data is any comparison across the different cost metrics. not consistently reported, however, making comparisons difficult. Averages of USD 0.02 to 0. 03/kWh appear to be the norm, with certain USD exceptions. O&M costs of onshore wind in selected OECD countries Table 5.1 Variable Fixed Country (2016 USD/kWh) (2016 USD/kW/year) Germany 0.03 66 Denmark 0.02 Ireland 74 Norway 0.03 0.00 United States 53 0.04 Austria 41 Finland 50 Italy 76 Japan The Netherlands 0.01 0.03 Spain 0.03 Sweden Switzerland 0.05 Source: IEA Wind, 2011b; IEA Wind, 2015. 108

109 2017 • O&M costs for offshore wind farms are higher than Capacity factor: This is the result of an interplay - of several variables, among which the most those for onshore wind, mainly due to the high er costs of access to the site and of performing important is the nature and quality of the wind resource, followed by wind turbine design and maintenance on towers and cabling. The marine operational availability – including potential environment is harder to operate within than dry land, adding to the overall O&M costs. O&M curtailment. costs for Europe are estimated to be between • Total installed costs: The turbine cost is usually USD 109/kW/year and USD 140/kW/year today, the single largest cost item in a wind project, 9/kW/year by 2025 (IRENA, 7 but could fall to USD though depending on the complexity of the 2016a; IEA Wind, 2016). project, its share can be less important. This is even more so for offshore wind projects. LE 5.7 VELISED COST OF ELECTRICITY • WACC: The cost of debt, the equity premium of the investors, and the share of debt and equity The LCOE of a wind power project is driven by in a project all go towards the final value of the total installed costs, wind resource quality, the WACC . technical characteristic of the wind turbines used, O&M costs, the cost of capital and the economic • Operations and maintenance costs: Operational life of the project. Thus, the LCOE depends largely expenses consist of both fixed and variable on four factors: costs and can represent up to 20%-25% of LCOE. 109

110 RENEWABLE POWER GENERATION COSTS known with any certainty. However, the auction Figure 5.17 presents the evolution of the LCOE data for projects that will be commissioned out to of onshore wind between 1983 and 2017. The 2020 yields a learning rate of 21% for the period global weighted average LCOE declined from 2010-2020, a figure more likely to represent the USD 06/kWh in 2017, 40/kWh in 1983 to USD 0. 0. true learning curve value given the auction results an 85% decline. The data suggests that every include the impact of a lower WACC. time cumulative installed capacity doubles, the LCOE of onshore wind drops by 15%. This trend Figure 5.18 presents the historical evolution of the includes the impact of lower O&M costs over time, LCOE of onshore wind in 12 countries where IRENA but not the impact of a reduced cost of capital, as has the longest time series data. The data needs to technology matures and financial markets become be interpreted with care, however, given that the more comfortable with wind power development. first year for which IRENA has data for a country The rate of 15% is therefore an underestimate of varies. From 2010 to 2016, the greatest decline in the total learning rate for onshore wind, but a LCOE was in Spain, at 48%, followed by the United lack of data means the exact value cannot be States, at 45%, and Italy with 43% (Figure 5.18, Table 5.2). Figure 5.17 The global weighted average levelised cost of electricity of onshore wind, 1983-2017 0.4 0.3 0.2 2016 USD/kWh 0.1 0.0 1983198619891992199519982001 20072010201320162019 2004 ≤ 0 ≤ 300 100200 Capacity MWe Sources: IRENA Renewable Cost Database. 110

111 2017 Figure 5.18 The weighted average LCOE of commissioned onshore wind projects in 12 countries, 1983-2016 1983-2016 0.5 Sweden Denmark United States Germany 0.4 0.3 0.2 2016 USD/kWh 0.1 0.0 20161984 2016 1983 20161987 20161983 1989-2016 0.5 Italy United Kingdom Spain Canada 0.4 0.3 0.2 2016 USD/kWh 0.1 0.0 1989 20161990 20161989 20161990 2016 1991-2016 0.5 France India Brazil China 0.4 0.3 0.2 2016 USD/kWh 0.1 0.0 20161991 20162001 20161996 1991 2016 Source: IRENA Renewable Cost Database The weighted average LCOE reduction of commissioned onshore wind projects in 12 countries Table 5.2 Beginning - 2016 Country Beginning to 2010 2010 - 2016 89% 45% 80% United States 81% Denmark 26% 74% 72% 60% 31% Germany 28% 71% Sweden 79% 63% United Kingdom 66% 10% 70% 48% 42% Spain Italy 49% 43% 71% 27% Canada 56% 68% 69% 42% France 47% 72% 19% 77% India China 65% 19% 71% 39% 57% 29% Brazil Source: IRENA Renewable Cost Database 111

112 RENEWABLE POWER GENERATION COSTS Figure 5.19 presents regional and selected country weighted average LCOEs and ranges Improved technology for onshore wind from 2010-2016. In 2016, the has allowed higher most competitive weighted average LCOEs were observed in China, India, Brazil, Eurasia and North capacity factors America, at USD 06 to USD 07/kWh. These 0. 0. at the same site countries and regions are home to over half of global cumulative installed capacity. The highest weighted average LCOE in 2016 was observed in Europe, at USD 08/kWh, while in 2010 the 0. highest LCOE was observed in Oceania and Asia 11/kWh. (excluding China and India), at USD 0. The rate of LCOE decrease in North America From 2010-2016, the global weighted average between 2010 and 2016 was 30%. In 2010, Eurasia LCOE of offshore wind decreased from USD 17 to 0. had one of the highest weighted averages, at 0.14/kWh, despite total installed costs having USD USD 0.10/kWh, while in 2016 it came second, increased by 8% during this period (Figure 5.20). 0.06/kWh, the highest with an LCOE of USD This has been made possible by improved regional LCOE decrease, at 40%. The second technology that has allowed higher capacity highest regional LCOE decline occurred in Oceania factors that have more than offset the increase in where the LCOE fell by 33% from 2010 to 2016, installed costs observed in this period. The prices 06/kWh. Europe saw its 11 to USD 0. 0. from USD awarded in auctions in 2016 and 2017 for projects weighted average LCOE decrease by 24% over the that will come online by 2020-2022 range from 0.10/kWh to USD 0.08/kWh. period, from USD 0.06 to USD 0.10/kWh. USD Figure 5.19 Regional weighted average LCOE and ranges of onshore wind in 2010 and 2016 0.20 0.15 0.10 2016 USD/kWh 0.05 0.00 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 Asia (excl. China and India) Africa ChinaIndiaBrazil Central America and the Caribbean South America (excl. Brazil) Oceania North America Europe Eurasia Source: IRENA Renewable Cost Database 112

113 2017 The LCOE of commissioned and proposed offshore wind projects and auction results, 2000–2022 Figure 5.20 0.3 0.2 Fossil fuel power cost range 2016 USD/kWh 0.1 0.0 2001 2012 2008 2024 2020 2016 2004 Auction DatabaseProject Database 800 1200400 ≥ 600 Capacity MWe Source: IRENA Renewable Cost Database and IRENA PPA Database 113

114

115 6. HYDROPOWER ydropower is a mature and reliable technology, It is important, however, that hydropower devel- that still dominates total renewable electricity opments respect the three pillars of sustainability: H 1 economic, environmental and social. Sustainable generation. Worldwide, total installed hydropower development of hydropower and early consultation capacity (excluding pumped hydro) was 1 121 GW with stakeholders are crucial in reducing project at the end of 2016, although its share of global lead times and project development risks, and in renewable capacity has been slowly declining. In accelerating the development of hydropower. 2010, it accounted for around 75% of this total, but by 2016, its share was approximately 50%. In terms When hydropower schemes have storage that is of electricity production, hydropower accounted manageable – for example, in the reservoir behind for 81% of electricity from renewable sources, but the dam – hydropower can contribute to the stability by 2016, its share had dropped to 70%. of the electricity system by providing flexibility and grid services. Hydropower can provide Hydropower is an extremely attractive renewable important grid stability services, as spinning technology due to the low-cost of the electricity turbines can be ramped up more rapidly than it produces. Where reservoir storage is available any other generation source to provide additional hydropower is also uniquely placed to provide generation or voltage regulation to ensure that flexibility services to the grid that will, in addition the electricity system operates within its quality to providing low cost electricity in its own right, limits. Hydropower projects are also unique in contribute to integrating higher shares of VRE. Its that they often combine both energy and water usefulness is not restricted to the ability to absorb supply services. Hydropower projects can open VRE when the sun is shining and wind is blowing, up opportunities for irrigation schemes, drought however, as it can also provide other grid services management, municipal water supply, navigation such as frequency or voltage regulation, fast and recreation; thus bringing local social and reserve, etc. Its ability to meet load fluctuations 2 economic benefits. Similarly, hydropower projects minute by minute and operate efficiently at partial can provide important flood control services. The loads, which is not the case for many thermal 3 LCOE analysis in this report does not include an plants, makes it a valuable part of any electricity estimate of the value of these services, however, system. as they are very site-specific. 1. T his section doesn’t discuss the costs and performance of pumped hydro storage, as they are an electricity storage technology, not a generation source. 2. S ome electricity storage devices, such as flywheels, can match this capability, but are more expensive and, in general, the more responsive they are, the less time they can be used before needing to be recharged. 3. A lthough many modern gas-fired plants can operate within one or two percentage points of their design efficiency over a relatively wide load range, this is usually not the case for older plants and coal-fired plants. 115

116 RENEWABLE POWER GENERATION COSTS • Hydropower schemes often have significant Pumped storage hydropower schemes use off-peak electricity to pump water from a flexibility in their design. This enables them to meet lower reservoir to a higher reservoir, so that baseload demand with relatively high capacity the pumped storage water can be used for factors, or to have higher installed capacities and generation at peak times, provide grid stability, a lower capacity factor, but meet a much larger flexibility and other ancillary grid services. share of peak electricity demand. Hydropower can also store energy over weeks, months, seasons or Pumped storage hydropower can also absorb renewable power generation during times of even years, depending on the size of the reservoir. surplus, thus reducing potential curtailment. Hydropower can therefore provide the full range Hydropower is a capital-intensive technology, of ancillary services required to allow high however, with long lead times for development penetration of variable renewable energy sources, and construction. This is due to the requirement such as wind and solar PV. The importance of for significant feasibility assessments, planning, hydropower is therefore likely to grow over time design and civil engineering work. as the shift to a sustainable electricity sector accelerates. Hydropower can therefore provide There are two major costs components for low-cost electricity and, in many cases, some of hydropower projects: the flexibility required to integrate high levels of • variable renewables at minimal costs. The civil works for the hydropower plant construction, including any infrastructure development required to access the site and 6.1 INSTALLED COST TRENDS the project development costs. Hydropower plants can be constructed in a variety • The costs related to electro-mechanical of sizes and with different properties. There are a equipment. range of technical characteristics that affect the choices of turbine type and size, as well as the The largest share of installed costs for large generation profile. These include the height of the hydropower plants is typically for civil construction water drop to the turbine – known as the “head” – works (such as the dam, tunnels, canal and seasonal inflows, potential reservoir size, minimum construction of powerhouse). Following this, costs downstream flow rates, and many other factors. An for fitting out the power house (including shafts and important opportunity offered by hydropower is electro-mechanical equipment, in specific cases) the possibility to add capacity at existing schemes, are the next largest capital outlay, accounting for or install capacity at dams that do not yet have a around 30% of the total costs. The long lead times hydropower plant. for these types of hydropower projects (7-9 years or more) mean that owner costs (including the Hydropower schemes can be broadly classified project development costs) can be a significant into the following categories: portion of the overall costs, due to the need for • working capital and interest during construction. Run-of-river hydropower projects have no, or Additional items that can add significantly to overall very little, storage capacity behind their dams, costs include the pre-feasibility and feasibility with generation almost completely dependent studies, consultations with local stakeholders and on the timing and size of river flows. policy-makers, environmental and socio-economic • Reservoir (storage) hydropower schemes mitigation measures and land acquisition. can store water behind the dams in order to Although electro-mechanical equipment costs de-couple generation from hydro inflows. usually contribute less to the total cost in large- Reservoir capacities can be small or very large, scale projects, the opposite is true of small-scale depending on the characteristics of the site and M W). projects (with installed capacity of less than 10 the economics of dam construction. 116

117 2017 by numerous factors pertaining to the site, the For small-scale projects, the electro-mechanical scale of development and the technological equipment costs can represent 50% or more of the solution that is most economic. Hydropower is a total costs, due to the higher specific costs per kW highly site-specific technology, with each project of small-scale equipment. designed for a particular location within a given The cost breakdown for small hydro projects river basin. This is so that it may meet specific in developing countries reflects the diversity needs for energy and water management, based of hydropower projects and their site-specific on local conditions and inflows into the catchment constraints and opportunities (IRENA, 2013a). basin. Proper site selection and hydro scheme It would require a large dataset to identify the design are therefore key challenges, and detailed specific reasons for the wide variation in project work at the design stage can avoid expensive cost breakdowns and to identify “efficient” levels. mistakes later (Ecofys et al., 2011). Infrastructure costs can account for up to half of The total installed costs for hydropower projects total costs for projects in remote or difficult to typically range from a low of USD 500/kW to access locations. It is also possible to have projects around USD 00/kW (Figure .1). It is not 5 6 4 in remote locations where good infrastructure unusual, however, to find projects outside this exists but there are no transmission lines nearby, range. For instance, adding hydropower capacity resulting in significant grid connection costs. to an existing dam that was built for other purposes The capital costs of large hydropower projects may have costs as low as USD 450/kW. On the other are dominated by the civil works and equipment hand, projects at remote sites, without adequate costs, which can represent between 75% and local infrastructure and located far from existing as much 90% of the total investment costs transmission networks, can cost significantly more (IRENA, 2015). Civil works costs are influenced Figure 6.1 Total installed costs by project and global weighted averages for hydropower, 2010-2017 3 000 2 000 1 780 1 521 1 535 1 427 1 535 1 208 2016 USD/kW 1 171 1 233 1 000 0 20102011201220132014201520162017 300 ≥ 1100200 Capacity MWe Sources: IRENA Renewable Cost Database. 117

118 RENEWABLE POWER GENERATION COSTS Total installed costs are lowest in China and India 4 than USD 500/kW due to higher logistical, civil and highest in Oceania and Central America and engineering and grid connection costs. the Caribbean (Figure 6.2). The range in installed The global weighted average for the total installed costs for hydropower is wide, reflecting the very cost of a hydropower projects has increased site-specific development costs of hydropower in recent years from USD 1 171/kW in 2010 to projects. Hydropower costs are typically lower 80/kW in 2016, before falling back to 1 7 USD in regions with significant remaining economic USD 8/kW in 2017. This trend has mainly been 1 55 potential, like in Asia, as there are likely to be more driven by increases in average total installed ideal sites left to exploit. However, even in higher costs in Asia, Eurasia and North America, while cost regions, the value of other services they can other regions have experienced more volatile provide — such as potable water, flood control, annual weighted averages. Although an analysis irrigation and navigation — which are included of the reasons behind these cost trends is not in the hydropower project costs but are typically yet available, possible explanations include a not remunerated, may mean benefits exceed shift towards hydropower projects in less ideal costs. In addition, this does not take into account sites, with higher project development costs, the additional value of grid services provided projects further from existing infrastructure or by hydropower in terms of short-term flexibility the transmission network, thus requiring higher and long-term energy storage, which may have transport and logistical outlay, as well as boosting significant value over and above a simple LCOE grid connection costs. analysis. Total installed cost ranges and weighted averages for hydropower projects by country/region, Figure 6.2 2010-2016 8 000 6 000 th 95 percentile 4 000 2016 USD/kW 2 000 th 5 percentile 0 ChinaEurasiaEuropeIndia Oceania AfricaBrazil Other Central North Middle Other South America America Asia East America and the Caribbean 1100200 ≥ 300 Capacity MWe Source: IRENA Renewable Cost Database. 118

119 2017 The data indicate that for this sample, civil works Figure 6.3 presents the installed costs for small and mechanical equipment comprise the largest (less than 10 W) and large hydro plants by region. M share of costs (Figure 6.4). The share of civil In almost every surveyed region, small hydro plants works in these projects varied from 17% to 65% have higher installed costs compared to large hydro in this particular sample. Mechanical equipment plants, with the exception being of Central America represented the second largest cost, on average, and the Caribbean and of Oceania. The small plants varying in the sample from a minimum of 18% to are 20-80% more expensive on average, outside of a maximum of 66%. Planning costs varied from Central America and the Caribbean and Oceania. 6-29% of total costs for these projects. Grid In the case of Central America and the Caribbean connection can represent a significant cost for and Oceania, where installed costs are higher for the more remote hydropower projects, but are large hydropower plants, the IRENA Renewable sometimes minimal if they represent an expansion Cost Database contains a smaller subset of data of an existing scheme, with grid connection costs than for many other regions and the results should accounting for 1% for projects close to existing grid be treated with caution. nodes to a high of 17% for projects in more remote To understand better the share of different areas. Lastly, land costs represent the smallest cost components in the total installed costs of share of a hydropower project, varying from 1-8%. hydropower projects, IRENA collected cost data Care should be taken in interpreting these values MW, from a sample of 25 projects, totaling 337 given the relatively small sample size, but they do in China, India and Sri Lanka. These projects serve to illustrate the wide ranges of individual were commissioned between 2010-2016 and cost components that are driven by individual site had installed costs of between USD 922 and characteristics. 1 976/kW for all projects, while the installed USD costs for large projects having costs from 389/kW. 1 035 to USD 1 USD Total installed cost ranges and capacity weighted averages for small and large hydropower projects Figure 6.3 by country/region, 2010-2016 ChinaEurasiaEuropeIndia Other Other Central AfricaBrazil North Oceania Middle South America Asia America East America and the Caribbean 000 8 000 6 percentile th 95 000 4 2016 USD/kW 000 2 percentile th 5 0 Small Small Small Small Small Small Small Small Small Small Small Small Large Large Large Large Large Large Large Large Large Large Large Large 300 ≥ 1100200 Capacity MWe Source: IRENA Renewable Cost Database. 119

120 RENEWABLE POWER GENERATION COSTS Total installed cost breakdown by component and capacity weighted averages for 25 hydropower Figure 6.4 projects in China, India and Sri Lanka, 2010-2016 100% 80% th 95 percentile 60% 40% Share of total installed costs 20% th 5 percentile 0% Grid connectionCost of land Planning Civil worksMechanical equipment and other Source: IRENA Renewable Cost Database. expected, given that each hydropower project 6.2 CAPACITY FACTORS has very different site characteristics and that low The global weighted average capacity factor of capacity factors are sometimes a design choice newly commissioned hydropower projects between to size the turbines to help meet peak demand 2010 and 2016 was 49% for small hydropower and provide other ancillary grid services. Average projects and 48% for large hydropower projects, capacity factors for newly commissioned large with most projects in the range of 25-84% hydropower projects are highest in South America (Figure 6.5), Europe being a notable exception to and Brazil, with 62% and 60%, respectively, while this for having a range of projects with capacity their average capacity factors for small hydropower factors lower than 20%. This wide range is to be projects are 66% and 58%, respectively. 120

121 2017 Hydropower project capacity factors and capacity weighted averages for large and small hydropower Figure 6.5 projects by country/region, 2010-2016 North Central Other Other Middle Oceania ChinaEurasiaEuropeIndia AfricaBrazil America Asia South America East America and the Caribbean 100% 80% th 95 percentile 60% 40% Capacity factor th 5 percentile 20% 0% Small Small Small Small Small Small Small Small Small Small Small Small Large Large Large Large Large Large Large Large Large Large Large Large 300 ≥ 1100200 Capacity MWe Source: IRENA Renewable Cost Database. Note: small refers to sub-10 MW plant and large to those above. 121

122 RENEWABLE POWER GENERATION COSTS costs. Figure 6.6 presents the cost distribution PERATION AND MAINTENANCE COSTS 6.3 O of individual O&M items in the sample. As can Annual O&M costs are often quoted as a percentage be seen, operations and salaries take the largest of the investment cost per kW per year. Typical slices of the O&M budget. Maintenance varies from values range from 1-4%. The International Energy 20-61%, salaries from 13-74% of O&M costs, and Agency (IEA) assumes 2.2% for large hydropower materials are estimated to account for around 4%. projects and 2.2-3% for smaller projects, with a The O&M costs reported do not typically cover global average around 2.5% (IEA, 2010). This would the replacement of major electro-mechanical put large-scale hydropower plants in a similar equipment, or the refurbishment of penstocks, range of costs as a percentage of total installed 4 tailraces, etc costs as those for wind, although not as low as the . Replacement of these is infrequent, O&M costs for solar PV. When a series of plants are with design lives of 30 years or more for electro- installed along a river, centralised control, remote mechanical equipment, and 50 years or more management and a dedicated operations team for penstocks and tailraces. This means that the to manage the chain of stations can reduce O&M original investment has been completely amortised costs to low levels. by the time these investments need to be made, and therefore they are not included in the LCOE Other sources, however quote lower or higher analysis presented here. They may, however, values. The Energy Information Agency assumes represent an economic opportunity before the full 0.06 % of total installed costs as fixed annual amortisation of the hydropower project, in order to O&M and 0.003 USD/MWh as variable O&M costs boost generation output. for a conventional hydropower plant of 500 MW that would be commissioned in 2020 (EIA, 2017a). Other studies (EREC/Greenpeace,2010) indicate 6.4 LE VELISED COST OF ELECTRICITY that fixed O&M costs represent 4% of the total Hydropower is a proven, mature, predictable capital cost. This figure may represent small-scale technology and has historically been a low-cost hydropower, but large hydropower plants will source of electricity. Investment costs are highly have significantly lower O&M costs. An average dependent on location and site conditions, which value for O&M costs of 2-2.5% is considered explains the wide range of plant installed costs. the norm for large-scale projects (IPCC, 2011), However, the relatively high initial investment is which is equivalent to average costs of between balanced, by the long economic lifetime of the 20 and USD 60/kW/year for the average USD hydropower plant (with parts replacement) as well project by region in the IRENA Renewable Cost as by the low O&M costs. Thus, the average LCOE Database. This will usually include an allowance from hydropower is typically low, with excellent for the periodic refurbishment of mechanical and hydropower sites offering some of the lowest cost electrical equipment, such as turbine overhaul, electricity of any generating option. generator rewinding and reinvestments in communication and control systems, but exclude Hydropower projects can be designed to perform major refurbishments. very differently from each other, however, which complicates a simple LCOE assessment. A plant The 25 projects that IRENA collected cost with a low installed capacity could run continuously breakdown data for tend to confirm these results, to ensure high average capacity factors, but at the as the average O&M cost was slightly less than 2% expense of being able to ramp up production to of total installed costs per year, with a variation meet peak demand loads. Alternatively, a plant between 1-3% of total installed costs per year. with a high installed electrical capacity and low Larger projects have O&M costs below the 2% capacity factor would be designed to help meet average, while smaller projects approach the peak demands and provide spinning reserve and maximum, or are higher than the average O&M 4. P enstocks are tunnels or pipelines that conduct the water to the turbine, while the tailraces are the tunnels or pipelines that evacuate the water after the turbine. 122

123 2017 Hydropower O&M cost breakdown by project for a sample of 25 projects in China, India and Figure 6.6 Sri Lanka, 2010-2016 100% 80% th 95 percentile 60% 40% Share of total O&M costs th 5 percentile 20% 0% Other Salary Operation costs Material Source: IRENA Renewable Cost Database. 123

124 RENEWABLE POWER GENERATION COSTS the number of kWhs generated relative to the other ancillary grid services. The latter strategy investment. The value of peak generation and would involve higher installed costs and lower the provision of ancillary grid services can thus capacity factors, but where the electricity system have a significant impact on the economics of a needs these services, hydropower can often be 5 hydropower project. the cheapest and most effective solution to these needs. The weighted average country/regional LCOE Deciding which strategy to pursue for any given of all projects, large and small, in the IRENA hydropower scheme is highly dependent on the Renewable Cost Database ranged from a low of local market, the structure of the power generation USD 0.04/kWh in Brazil for all of the projects in pool, grid capacity and constraints, the value of the IRENA Renewable Cost Database to a high of providing water management and grid services, 11/kWh in Europe. Focusing in on the global USD 0. etc. Perhaps more than with any other renewable weighted average LCOE trend by year, in 2017, energy, the true economics of a given hydropower the global weighted average cost of electricity scheme are driven by these factors, not just by from hydropower projects commissioned in that Pumped hydro storage Box 4 Energy storage is becoming an increasingly flexible and cost-effective tool for grid operators to help manage instability on their networks. This is especially so, given with the growing amount of variable renewable energy generation being deployed in major markets worldwide, such as that from solar PV and wind. Energy storage has gained prominence in recent years, and plays a key role in the design of modern electricity grids. According to the Global Energy Storage Database [DOE, 2017], the rated power of operational stationary energy storage reached a total of more than 170 W, globally, by October 2016. More than 96% was provided by G pumped hydro storage, followed by thermal storage (1.9%), electro-chemical batteries (1.0%) and electro-mechan - ical storage (0.9%). Three quarters of all energy storage was installed in the top 10 countries, led by China (18.8 %), Japan (16.7 %) and the United States (14.1 %). Pumped hydro storage is a well-understood and proven technology, with decades of operating experience. Due to this maturity, only slight improvements in cost structure or transformation efficiency can be expected in the next few years. There are, however, many new ideas on how to expand worldwide pumped hydro storage capacities. These include the use of wind turbine structures as upper reservoirs [GE Reports, 2016], existing underground for - mations such as abandoned mines [ESA, n.d. – a], or added weight through rock formations [Heindl Energy, 2016]. These approaches may offer lower-cost pumped hydro storage and/or greatly expand their potential, but given the early stage of development of these approaches significant uncertainty remains as to their likely deployment. Another type of unconventional Pumped Hydroelectric Storage (PHS) that has already been commercialised in a mid-sized storage asset is seawater storage [Fujihara, 1998]. This type of PHS utilises the sea as the lower water res - ervoir, instead of an artificial lake. These storage systems promise to offer comparably lower installation costs due to their single reservoir construction. Yet, the maintenance costs of these storage systems are significantly higher, due to the highly corrosive salt water environment and marine growth on hydraulic structures [ESA, n.d. – b], and suitable geological structures next to seas or lakes are not always available. Promising developments in other energy storage technologies may one day challenge pumped hdyro storage's - near monopoly on low-cost electricity storage (IRENA, 2017c), but for now, pumped hydro is still the only technol ogy offering economically viable large-scale storage. The importance of pumped hydro storage, and indeed reser - voir hydropower, is likely to grow over time as the shift to a truly sustainable electricity sector accelerates, not just for the low-cost storage it provides, but for the flexibility it brings to integrate high levels of variable renewables at minimal cost. 5. T his is without considering the other services being provided by the dam (e.g. flood control) that are not typically remunerated, but are an integral part of a project’s purpose. 124

125 2017 with significant remaining untapped economic year was USD 0.047/kWh. This was slightly lower resources. Europe is something of an exception, as than the weighted average of USD 053/kWh for 0. most of the economic hydropower potential in this projects commissioned in 2016, but substantially region has already been exploited. In Europe, new higher than in 2010, when the weighted average projects are relatively few in number, face long LCOE of newly commissioned projects was lead times to develop and have a higher weighted USD 0.036/kWh. average LCOE, at USD 0.11/kWh. 24 6 Figure 6.7 presents the LCOE of the 3 In terms of the differences between small and hydropower projects contained in the IRENA large hydropower plants, the LCOE of small hydro Renewable Cost Database. It shows that many plants is usually higher than the LCOE of large new hydropower projects are expected to be hydro plants, by 10%-40%, which is somewhat less highly competitive. LCOE data ranges from than the difference in total installed costs for these 0.02/kWh to a high of a low of around USD different projects. Small hydropower projects can USD 30/kWh. Although the range is wide, for 0. be attractive, despite higher LCOEs, either because reasons already discussed, the weighted average they are the least costly supply solution in more LCOE is below USD 10/kWh for almost all regions 0. remote areas, or because they provide valuable and the weighted average remains low in most grid services. regions, typically ranging between USD 04 and 0. 0.06/kWh. This is typically correlated with regions Levelised cost of electricity and weighted averages of small and large hydropower projects Figure 6.7 by country/region, 2010-2016 North AfricaBrazil ChinaEurasiaEuropeIndia Oceania Middle Other Other Central America East South America Asia America and the Caribbean 0.30 0.25 0.20 0.15 0.10 2016 USD/kWh 0.05 0 Small Small Small Small Small Small Small Small Small Small Small Small Large Large Large Large Large Large Large Large Large Large Large Large 300 ≥ 1100200 Capacity MWe Source: IRENA Renewable Cost Database. 125

126

127 7. BIOENERGY FOR POWER technologies are available that can use ower generation from bioenergy can come biomass as a fuel input, but technology risks from a wide range of feedstocks and use a P remain for some of the newer, more innovative variety of different combustion technologies. technologies. Bioenergy power generation technologies range from commercially proven solutions, with a wide The analysis in this report focuses on the costs of range of suppliers, through to those that are only - power generation technologies and their econom just being deployed on a commercial scale. ics, while briefly discussing delivered feedstock - costs. Indeed, one of the most important determi The power generation technologies that are nants of the economic success of biomass projects mature, commercially available and have a long is the availability of a secure and sustainable fuel track record include: direct combustion in stoker supply (i.e. feedstocks) for conversion. This area is boilers; low-percentage co-firing; anaerobic the focus of increasing research by IRENA, given digestion; municipal solid waste incineration; the uncertainty surrounding the global potential landfill gas and combined heat and power. Other and supply of sustainably sourced bioenergy feed - less mature technologies, such as atmospheric stocks (IRENA, 2017f and 2017g). biomass gasification and pyrolysis, are only at the beginning of their deployment. The potential for cost reductions from the technologies in use 7.1 BIOMASS FEEDSTOCKS is therefore very heterogeneous. While only marginal cost reductions can be anticipated in the Biomass is the organic material of recently living short term, there is good, long-term potential for plants, such as trees, grasses and agricultural cost reductions from those technologies that are crops. Biomass feedstocks are very heterogeneous not yet widely deployed. and the chemical composition is highly dependent on the plant species. Ash content, density, and To analyse the use of biomass power generation, particle size and moisture content are all critical the following three components must be examined: issues for the biomass feedstock. These factors • have an impact on the cost of this feedstock per Biomass feedstocks: These come in a variety of unit of energy, its transportation, pre-treatment forms and have different properties that impact and storage costs, as well as the appropriateness their use in power generation. of different conversion technologies. Moreover, • Biomass conversion: This is the process by heterogeneity in quality can also be a problem for which biomass feedstocks are transformed into the conversion process, since some combustion the energy form that will be used to generate technologies require much more homogeneous heat and/or electricity. feedstocks to operate. This can add complexity to • the planning and economic viability of biomass- Power generation technologies: A wide range based power plants. of commercially proven power generation 127

128 RENEWABLE POWER GENERATION COSTS I Thus, unlike wind, solar and hydro, the economics 7.2 NSTALLED COST TRENDS of biomass power generation are dependent Technology options largely determine the cost and upon the availability of a predictable, sustainably efficiency of biomass power generation equip - sourced, low-cost and long-term adequate ment, although equipment costs for individual feedstock supply. The range of costs for feedstocks technologies can vary significantly. Factors affect - is highly variable, too. Waste produced due to ing this depend on the region, feedstock type and industrial processes can have a zero or even availability, and how much feedstock preparation negative cost if it is waste that would otherwise or conversion happens on site. have incurred disposal charges, such as black liquor at pulp and paper mills. Yet there can also be Planning, engineering and construction costs, potentially high prices for dedicated energy crops, fuel handling and preparation machinery, and if productivity is low and transport costs are high. other equipment (e.g. the prime mover and fuel More modest costs are incurred for agricultural conversion system) represent the major categories and forestry residues that can be collected and of total investment costs of a biomass power plant. transported over short distances, or are available Additional costs are derived from grid connection at processing plants as a by-product. Transport and infrastructure (e.g. roads). Combined heat and costs add a significant amount to the costs of power (CHP) biomass installations have higher feedstocks, if the density of the feedstock is lower capital costs, but the higher overall efficiency and the distances become large. Transforming (around 80%-85%) and the ability to produce wet biomass into higher-density forms will help heat and/or steam for industrial processes, or for reduce transportation costs per unit of energy, space and water heating through district heating but the transformation costs must also be taken networks, can significantly improve the economics. into account. There is often a trade-off between Biomass power plants in emerging economies the volume of low-cost feedstock available to a bioenergy power plant as collection radius grows, this can be offset if more cost-effective bulk freight The wide range of deliveries can be made by rail or water. bioenergy-fired power Feedstocks typically account for between 20-50% of the final cost of electricity from biomass generation technologies technologies. Agricultural residues, such as straw and sugarcane bagasse, tend to be the translates into a broad least expensive feedstocks, as they are a harvest or processing by-product. They are, however, range of observed correlated with the price of the primary commodity installed costs from which they are derived and have registered increased costs from 2000 to 2011 as indicated in the World Bank agricultural commodities index (World Bank, 2017). However, the cost of can have significantly lower investment costs agricultural commodities has edged down, after than the cost ranges for OECD-based projects, the peak observed in 2011, with prices down by due to lower local content costs and the cheaper 28% in 2016 compared to 2011. Biomass power equipment allowed, in some cases, by less stringent generation plants that are exposed to feedstocks environmental regulations. that are derived from traded commodities are Figure 7.1 and Figure 7.2 highlight the relatively low therefore exposed to volatile commodity prices, installed cost of biomass combustion technologies unless they have secure supplies or have acquired for projects in Asia and South America, while a long-term contract for their feedstock needs more expensive projects occur mostly in Europe (see IRENA, 2015, for a more detailed discussion of and North America. Although small-scale feedstock costs). 128

129 2017 projects can have higher capital costs, most large the technology, and tend to cluster between projects have total installed costs in the range of USD 000/kW. The less expensive 500 and USD 8 USD 450 to 2 500/kW. The lower range can be projects are in Asia and South America, while the achieved when additional capacity is added to an more expensive ones are in Europe. existing project, as the economics of electricity Figure 7.2 presents the total installed cost range of generation improve. The data to which IRENA biomass fired power in several regional groupings. has access is dominated by steam cycle boiler Biomass installed costs in India are the lowest, systems, although in many cases the technology is 2 600/kW, while in 450 to USD ranging from USD not disclosed. Biomass projects using steam cycle China they range from USD 3 600/kW. 450 USD boilers appear to have the lowest costs, clustering Installed cost ranges are wider in Europe, North between USD 500 and USD 000/kW, while fixed 2 America and the rest of the world category, as the bed gasifiers deployed in Europe and North America technological options used to develop projects are between USD 000/kW. 7 000 and USD 2 are more heterogeneous and on average more Most of the projects in the IRENA Renewable expensive. Cost Database have not, however, disclosed Total installed costs of biomass-fired generation technologies by country/region and project capacity Figure 7.1 8 000 6 000 4 000 2016 USD/kW 2 000 0 55 051015202530 35 40 45 50 Capacity MW OECD China and India Rest of the world Source: IRENA Renewable Cost Database. 129

130 RENEWABLE POWER GENERATION COSTS Figure 7.2 Total installed costs of biomass-fired generation technologies by country/region 000 8 6 000 4 000 th 95 percentile 2016 USD/kW 2 000 th 5 percentile 0 North AmericaRest of the world India Europe China 1100200 ≥ 300 Capacity MWe Source: IRENA Renewable Cost Database. replacement and incremental serving costs are O PERATION AND MAINTENANCE COSTS 7.3 the main components of variable O&M costs. Fixed operations and maintenance (O&M) costs for Unfortunately, the available data often merges bioenergy power plants typically range from 2-6% fixed and variable O&M costs into one number, thus of total installed costs per year, while variable rendering impossible a breakdown between fixed O&M costs are typically relatively low, at around and variable O&M costs. Table 7.1 provides data 0.005/KWh. Fixed O&M costs include labour, for the fixed and variable O&M costs for selected scheduled maintenance, routine component/ bioenergy for power technologies. equipment replacement (for boilers, gasifiers, feedstock handling equipment, etc.), insurance, etc. The fixed O&M costs of larger plants are lower 7.4 CAPACITY FACTORS AND EFFICIENCY per kW due to economies of scale, especially for Technically, bioenergy-fired electricity plants can labour. Variable O&M costs are determined by the achieve capacity factors of 85-95%. In practice, output of the system and are usually expressed most plants do not regularly operate at these as USD/kWh. Non-biomass fuel costs, such as levels. Feedstocks may be a constraint on capacity ash disposal, unplanned maintenance, equipment factors, particularly in cases where systems relying 130

131 2017 Table 7.1 F ixed and variable O&M costs for bioenergy power Variable O&M Fixed O&M (2016 USD/MWh) (% of CAPEX/YEAR) 3.2 4.08 - 5.03 Stoker/BFB/CFB boilers 3 - 6 4.08 Gasifier 2.1 - 3.2 Anaerobic digester 4.49 2.3 - 7 Landfill gas n.a 11 - 20 Source: IRENA, 2015. as well as waste-to-energy plants and those using on agricultural residues may not have year-round forestry or pulp and paper residues available from access to low-cost feedstock, and where buying their year-round operation. alternative feedstocks might make plant operation uneconomical. This is illustrated in Figure 7.3. In Weighted average capacity factors are above 60% Figure 7.3, the lower capacity factors for projects in China, India and the rest of the world, while in in India represent the impact of the many bagasse- Europe and North America they are above 80% fired projects, which operate only during and after (Figure 7.3). Biomass plants relying on landfill harvesting season until they exhaust the available gas and other biogases, wood and wood straws, feedstock supply. In contrast, the higher capacity fuel wood and industrial and renewable municipal factors observed in Europe and North America are waste tend to have higher capacity factors than a consequence of these plants having invested in the regional weighted average. Projects relying on higher-cost technologies that can process a range agricultural inputs, such as bagasse, tend to have of heterogeneous feedstocks, sourcing a steady lower capacity factors, as they depend on seasonal supply of wood pellets and wood waste provided harvesting. by a functional, buyer-driven international markets for such resources (Argus Biomass Markets, 2014), 131

132 RENEWABLE POWER GENERATION COSTS Figure 7.3 Project capacity factors and weighted averages of biomass-fired electricity generation systems by country and region 100% th 95 percentile 80% 60% th percentile 5 40% Capacity Factor 20% 0% Rest of the world Europe North India China America Bagasse Biomass energy Black liquor Energy crops Industrial waste Fuelwood Other biogases from anaerobic fermentation Other primary solid biomass Landfill gas Other vegetal and agricultural waste Pulp and paper residues Renewable municipal waste Straw Wood waste Wood and straw pellets/ briquettes Rice husks Sewage sludge ≥ 1100200 300 Capacity MWe Source: IRENA Renewable Cost Database. maintenance – result in lower overall efficiencies. The assumed net electrical efficiency (after These can be around 25%, but many technologies accounting for feedstock handling) of the prime are available with higher efficiencies, ranging from mover (generator) averages around 30%, but 31% for wood gasifiers to a high of 36% for modern varies from a low of 25% to a high of around well-maintained stoker, circulating fluidised bed 36%. In developing countries, less advanced (CFB), bubbling fluidised bed (BFB) and anaerobic technologies – and sometimes suboptimal digestion systems (Mott MacDonald, 2011). 132

133 2017 4/GJ where additional and rise to as much as USD 7.5 EVELISED COST OF ELECTRICITY L feedstocks are purchased to achieve higher The wide range of bioenergy-fired power capacity factors. These projects, using simple and generation technologies and feedstock costs cheap combustion technologies, can have very translates into a broad range of observed LCOEs for competitive LCOEs (Figure 7.4). Even higher-cost bioenergy-fired electricity. Figure 7.4 summarises projects in certain developing countries, however, the estimated range of costs for biomass power can be attractive, because they provide security generation technologies in countries and regions of supply where brownouts and blackouts can where the IRENA Renewable Cost Database be particularly problematic for the efficiency of has good coverage. Assuming a cost of capital industrial processes. of 7.5%-10% and feedstock costs between Many of the higher cost projects in Europe and USD /GJ (the LCOE calculations in /GJ and USD 9 1 North America use municipal solid waste as a 1.5/GJ), this report are based on an average of USD feedstock. It is therefore worth noting that the the weighted average LCOE of biomass-fired primary objective of these projects is not power 05/kWh in 0. electricity generation is around USD generation, but waste disposal. Capital costs are India and USD 0.06/kWh in China. often higher, as expensive technologies are used The weighted average LCOE in Europe to ensure local pollutant emissions are reduced to and North America is higher, at around acceptable levels. Excluding these projects – which 0.08/kWh-USD 0.09/kWh, reflecting more USD are typically not the largest – reduces the weighted advanced technology choices, but also the more average LCOE in Europe and North America by stringent emissions controls and higher feedstocks around USD 0.01/kWh and narrows the gap with costs. Where capital costs are relatively low – and the LCOE of non-OECD regions. low-cost feedstocks are available – bioenergy Finally, the availability of a continuous and can provide competitively priced, dispatchable affordable stream of feedstock allows for higher electricity generation with an LCOE as low as capacity factors, but does not have a significant 0. 04/kWh. The most competitive around USD impact on LCOE. Projects based on bagasse projects make use of agricultural or forestry and other agricultural residues come with lower residues already available at industrial processing capacity factors, due to the seasonality of the sites where marginal feedstock costs are minimal, available feedstock. The LCOE of these projects, or even zero. Where industrial process steam or however, is comparable to projects relying on heat loads are also required, the ability to integrate more generic woody biomass feedstocks, such as CHP systems can reduce the LCOE for electricity to wood pellets and wood waste that can be more as little as USD 0.03/kWh. readily purchased year round. Thus, access to Low-cost opportunities to develop bioenergy-fired low cost feedstock offsets the impact of lower power plants present themselves at sites where capacity factors on LCOE. Lastly, projects relying low-cost feedstocks and handling facilities are on municipal waste come with very high capacity available to keep feedstocks and capital costs low. factors, but also some of the highest LCOEs, Where this is not the case, or where these feedstocks 0.15/kWh. Given that these projects above USD need to be supplemented by additional feedstocks have been developed mostly to solve waste (e.g. outside seasonal harvesting periods), then management issues, though, and not primarily for competitive supply chains for sustainably-sourced the competitiveness of their electricity production, feedstocks are essential in making biomass-fired this is not necessarily an impediment to their power generation economic. viability. In Europe, they are also sometimes supplying heat either to local industrial users, This is the pattern seen outside Europe and North or district heating networks, the revenues from America, where biomass costs for most projects these sales will reduce the LCOE below what is can range from negligible for agricultural or presented here. 2.25/GJ. forestry processing residues, up to USD They may sometimes exceed these values, too, 133

134 RENEWABLE POWER GENERATION COSTS Levelised cost of electricity by project and weighted averages of bioenergy-fired electricity generation Figure 7.4 by feedstock and country/region, 2000-2016 0.25 0.20 0.15 th percentile 95 2016 USD/kWh 0.10 0.05 th 5 percentile 0.00 Rest of the world India Europe North China America Bagasse Black liquor Energy crops Fuelwood Industrial waste Biomass energy Landfill gas Other biogases from anaerobic fermentation Other primary solid biomass Other vegetal and agricultural waste Pulp and paper residues Renewable municipal waste Straw Wood waste Wood and straw pellets/ briquettes Rice husks Sewage sludge ≥ 300 1100200 Capacity MWe Source: IRENA Renewable Cost Database. 134

135 2017 Levelised cost of electricity by capacity factors of bioenergy-fired projects, 2000-2016 Figure 7.5 0.25 0.20 0.15 2016 USD/kWh 0.10 0.05 0.00 60% 100% 40% 20% 0% 80% Capacity factor Fuelwood Bagasse Biomass energy Industrial waste Energy crops Black liquor Landfill gas Other biogases from anaerobic fermentation Other primary solid biomass Other vegetal and agricultural waste Renewable municipal waste Pulp and paper residues Sewage sludge Straw Wood waste Wood and straw pellets/ briquettes Rice husks ≥ 1100200 300 Capacity MWe Source: IRENA Renewable Cost Database. 135

136

137 GEOTHERMAL 8. POWER GENERATION exploration should be conducted. This is usually an eothermal resources are found in the Earth’s expensive and time consuming process, however, crust, in active geothermal areas on or near G and is one of most important barriers to the uptake its surface and at deeper depths. These resources of geothermal power generation. Poorer than consist of thermal energy, stored as heat in rocks expected results during the exploration phase of the Earth’s crust and interior, at shallow depths might require additional drilling, or wells may hot water or steam maybe be produced from need to be deployed over a much larger area to subterranean water that has come in contact with generate the expected electricity. Globally, around the heated area. In other cases, water will need to 78% of production wells drilled are successful, be injected through wells to harness the the heat with the average success rate improving in recent found in otherwise dry rocks. decades. This is most likely due to better surveying Geothermal deployment reached a total installed technology, which is able to more accurately target G W, globally, at the end of 2016. capacity of 12.7 the best prospects for siting productive wells. A This was 26% up on the 2010 level. Most of this key point is that adherence to global best practices capacity is deployed in active geothermal areas. significantly reduces exploration risks (IFC, 2013). The new capacity added in 2016, 780 MW, was Geothermal plants are very individual in terms of more than twice the capacity added in 2010. the quality of their resources and management Geothermal is a mature, commercially available needs, and therefore specific lessons cannot be technology that can provide low-cost baseload easily inferred. Nonetheless, adherence to best capacity in geographies with very good to excellent international practices for survey and management high-temperature resources that are close to the and thorough data analysis from the project site Earth’s surface. The deployment of geothermal are the best risk mitigation tools available to power outside such areas, however, using the so- developers (IFC, 2013). called “enhanced geothermal” or “hot dry rocks” Once commissioned the management of a approach, is much less mature. In this instance, it geothermal plant and its reservoir evolves over comes with costs that are typically significantly time, as more information becomes available higher, rendering the economics of such projects from operational experience. Once productivity at much less attractive today. existing wells declines, there might also be a need Readily available, extensive geothermal resource for replacement wells to make up for the loss in mapping can reduce the costs of development, productivity. by minimising the uncertainty about where initial 137

138 RENEWABLE POWER GENERATION COSTS 1 9 8 3 00/kW in 2009. Binary 00/kW and USD USD 8.1 NSTALLED COST TRENDS I power plants were more expensive and installed Geothermal power plants are, as with all other costs for typical projects were between USD 2 2 50 renewable technologies, relatively capital- and USD 500/kW that same year (IPCC, 2011). 5 intensive – yet they also come with low and Geothermal power plant costs can be as low predictable operating costs. The costs of 560/kW, however, where capacity is as USD engineering, procurement and construction (EPC) being added to a geothermal reservoir which is of a geothermal power plant follow trends in already well mapped and understood, and where commodity prices and drilling costs. Thus, when existing infrastructure can be used, but these commodity and oil markets are surging, the costs cases are exceptional. Data for recent projects of developing geothermal power plants often also (Figure 8.1) fits within the range of USD 000 to 2 rise. The opposite happens when these markets 000/kW, but there have also been some 5 USD are slowing. small projects in new markets where costs are The total installed costs of a geothermal power higher. Based on the data available in the IRENA plant consist of: Renewable Cost Database, the trend of increased • installed costs up to 2014 seems to have ended exploration and resource assessment costs in 2015, when, on average, costs began declining. • drilling costs for production and re-injection Given the relatively thin market for geothermal costs, as well as additional working capital power generation deployment, this trend, however, given that the success rate for well could vary should be treated with caution. between 60-90% (Hance, 2005; GTP, 2008) • field infrastructure, the geothermal fluid 8.2 C APACITY FACTORS collection and disposal system, and other surface installations; The capacity factors of geothermal power plants vary from around 60% to more than 85%. Using • costs of the power plant data from the IRENA Renewable Cost Database, • Figure 8.2 shows that geothermal plants using project development and grid connection costs. direct steam deliver capacity factors higher than The characteristics of the geothermal field are key 80%, while projects utilising lower temperature to what type of power plant (flash or binary) can resources that require binary plants deliver capacity be used for a given site. These field characteristics factors of 60-80%. Geothermal plants using "flash" 1 will determine well productivity, energy delivery, technologies consistently deliver capacity factors and the economic capacity to provide steam, higher than 80%, with few outliers below that value. given the quality of the geothermal resource and In terms of efficiency of conversion, geothermal its geographical distribution. power plants report a worldwide average of 12% efficiency, while the upper range is situated at 21% In line with rising commodity prices and drilling for a vapour dominated plant (Zarrouk, Moon, 2014). costs, the total installed costs for geothermal plants increased by between 60-70% (IPCC, 2011) It’s important to note that geothermal power between 2000 and 2009. Project development plants need active management of the reservoir costs followed general increases in civil and production profile to maintain production at engineering and EPC costs during that period, and the designed capacity factor. This will frequently cost increases in drilling associated with surging require additional production wells, as over time, oil and gas markets. The total installed costs individual production wells become less productive of conventional condensing “flash” geothermal as reservoir pressure around the production well power generation projects were between drops. This tends to mean capacity factors would 1. T he well productivity and energy delivery will determine the number of wells necessary for a given electrical capacity desired. These factors, and the geographical distribution of wells, will have a significant impact on overall development costs. 138

139 2017 Geothermal power total installed costs by project, technology and capacity, 2007-2020 Figure 8.1 000 10 000 8 6 000 2016 USD/kW 000 4 2 000 0 2011201220132014201520162017201820192020 2007200820092010 Direct steam Enhanced geothermal Flash types n.a Binary 1100200 ≥ 300 Capacity MWe Source: IRENA Renewable Cost Database and Global Data, 2016. 139

140 RENEWABLE POWER GENERATION COSTS Capacity factors of new geothermal power plants by technology and project size, 2007-2020 Figure 8.2 100% 80% 60% Capacity Factor 40% 20% 0% 2011201220132014201520162017201820192020 2007200820092010 Binary Flash types n.a Enhanced geothermal Direct steam 300 1100200 ≥ Capacity MWe Source: IRENA Renewable Cost Database. 8.3 LEVELISED COST OF ELECTRICITY otherwise decrease over time and is why O&M costs are high, as provision for new production The LCOE of a geothermal plant is determined wells needs to be incorporated. Figure 8.3 presents by its installed costs, O&M costs, economic a somewhat extreme example of the historical lifetime and the weighted average cost of MW geothermal electricity generation of an 88.2 capital. Geothermal power projects need careful plant in California. For this plant, the capacity management, as geothermal resources require factor in the first 17 years of its life was 82%, while careful optimisation through time. Following best in the last 18 years, in was 70% – a 15% decrease. practice for field appraisal, project development, drilling and operation is therefore important to ensuring that projects match their anticipated economic performance. 140

141 2017 Figure 8.3 Electricity generation and capacity factor of an 88.2 MW geothermal plant in California, 1989-2017 100 100% 93% 80% 80 65% 60 60% GWh 40% 40 Capacity Factor (%) 20 20% 0 0% January 1997 January 1993 January 2017 January 2013 January 1989 January 2001 January 2005 January 2009 Source: Energy Information Administration. For projects commissioned in 2014 and up to Figure 8.4 presents the LCOE for geothermal 2020, the LCOE of geothermal power plants projects under the following assumptions: appears to be trending downwards, in line with the a 25-year economic life; O&M costs of general decrease in total installed costs observed. 1 10/kW/year; capacity factors based on USD However, given the very thin deployment of project data (or national averages where project geothermal and the very site-specific nature of data is not available); two sets of wells for make- geothermal developments, care needs to be taken up and re-injection over the 25-year life of the in interpreting this trend. Additionally, this cost project; and the capital costs outlined in Figure represents expectations about the lifetime costs of 8.1. Between 2007 and 2014, the trend in LCOE the project and may prove either overly pessimistic was increasingly in line with rises in capital costs. or optimistic for individual projects. During this period, the LCOE varied from as low 0. as USD 04/kWh for second-stage development of an existing field to as high as USD 0. 14/kWh for greenfield developments. 141

142 RENEWABLE POWER GENERATION COSTS Levelised cost of electricity of geothermal power projects by technology and size, 2007-2020 Figure 8.4 0.20 Fossil fuel power cost range 0.15 0.10 2016 USD/kWh 0.05 0% 2011201220132014201520162017201820192020 2007200820092010 Binary n.a Direct steam Enhanced geothermal Flash types ≥ 300 1100200 Capacity MWe Source: IRENA Renewable Cost Database. 142

143 2017 143

144 RENEWABLE POWER GENERATION COSTS REFERENCES BNEF . Adam, A., Josephson, P.-E. B. and Lindahl, G (2017d), H2 2016 Wind O&M Index Report, (2017), ‘Aggregation of factors causing cost Bloomberg New Energy Finance, London. overruns and time delays in large public construction projects: Trends and implications’, M., J. Seel, and K.H. LaCommare (2017), Bolinger, Engineering, Construction and Architectural Utility-Scale Solar 2016: An Empirical Analysis of , vol. 24, no. 3, pp. 393–406 Management Project Cost, Performance, and Pricing Trends in [Online]. DOI: 10.1108/ECAM-09-2015 (Accessed , Lawrence Berkeley National the United States 8 January 2018). Laboratory (LBNL), https://utilityscalesolar. lbl. . gov/ Argus Biomass Markets (2014), “Weekly Biomass Argus Markets News and Analysis”, , No. 14-30. Utility-Scale Solar (2016), Bolinger, M. and J. Seel 2015: An Empirical Analysis of Project Cost, Bernreuter Research (2017), “Silicon consumption Performance, and Pricing Trends in the United to drop to 3.6 grams per watt by 2020”, www. States , Lawrence Berkeley National Laboratory bernreuter.com/en/news /press-releases/ https://emp.lbl.gov/publications/utility- (LBNL), polysilicon-consumption-for-solar-cells.html -empirical . scale-solar-2015 (accessed 4 July 2017). Utility-Scale Solar Bolinger, M. and J. Seel (2015), Blanco, M.I. (2009), “The economics of wind 2014: An Empirical Analysis of Project Cost, energy”, Renewable and Sustainable Energy Performance, and Pricing Trends in the United , Vol. 13, No. 6-7, Elsevier, pp. 1372-1382. Reviews States , Lawrence Berkeley National Laboratory (LBNL), https://emp.lbl.gov/publications/utility- . scale-solar-2015-empirical BNEF (2017a), 2H 2017 Wind Turbine Price Index, Bloomberg New Energy Finance, London. Bolinger, M. and S. Weaver Utility-Scale (2014), Solar 2013: An Empirical Analysis of Project Cost, BNEF (2017b), 3Q 2017 Wind Turbine Contract Performance, and Pricing Trends in the United Order Dataset, Bloomberg New Energy Finance, States , Lawrence Berkeley National Laboratory London. https (LBNL), ://emp.lbl.gov/publications/utility- scale-solar-2013-empirical . (2017c), Vestas reclaims top spot in annual BNEF ranking of wind turbine makers, Bloomberg New .bnef. https://about Energy Finance, London. com/blog/vestas-reclaims-top-spot-annual- -turbine-makers/ ranking-wind 144

145 2017 DOE (2015), “Is DOE Global Energy Storage (2017), Bolinger, M., S. Weaver, and J. Zuboy Database , U.S. Department of Energy, $50/MWh solar for real? Falling project prices Office of Electricity & Energy Reliability, and rising capacity factors drive utility-scale PV www.energystorageexchange.org/projects Progress toward economic competitiveness”, (accessed 4 August 2017). , in Photovoltaics: Research and Applications Vol. 23, No. 12, pp. 1847-1856, http ://dx.doi. org/10.1002/pip.2630 . Kostensituation der Windenergie an (2015), DWG (”Onshore wind energy cost Land in Deutschland (2016), Solar Industry Technolo situation in Germany“) , Deutsche WindGuard, - CanadianSolar www.canadiansolar.com/ Varel. gy Report 2015-2016, - media/canadian_solar-solar _industry_technol ogy_report_2015-2016.pdf . (2017), “DEWA receives lowest international DEWA bid for 4th phase of the Mohammed bin Rashid CanWEA. Al Maktoum Solar Park”, Dubai Water and (2016). Canadian Wind Farm Database. www.dewa.gov. Electricity Authority, Dubai, Canadian Wind Energy Association /news-and-media/press- ae/en/about-dewa and-news/latest-news/2017/06/dewa -receives- Cohen, G.E. (1999), “Final report on the operation (accessed 21 July 2017). lowest-international-bid and maintenance improvement program for concentrating solar power plants, SAND99- Douglas-Westwood (2010), 1290, NM”, Sandia National Laboratories. Offshore Wind Assessment in Norway , The Research Council of Norway, Oslo. Danish Energy Agency (2017), Data on operating and decommissioned wind turbines, Danish Energy Agency, Copenhagen. Ecofys et al. (2011), “Financing renewable energy in the European energy market: Final report”, Ecofys, Utrecht. (2016), Report on Solar PV Balance of System deea deea Solutions, Cost Reduction Potential to 2025, Frankfurt am Main. EIA (2017a), Annual Energy Outlook 2017 , U.S. Energy Information Administration, Washington, DC. 145

146 RENEWABLE POWER GENERATION COSTS (2013), (2017b), Form EIA-923 detailed data, EIA Fiorelli, J. and M. Zuercher-Martinson “How oversizing your array-to-inverter ratio U.S. Energy Information Administration, can improve solar-power system performance”, www.eia.gov/electricity/data . /eia923/ , Vol. 7, pp. 42-48. Solar Power World Energy Storage Association (n.d. – a), Sub- (2017) “First Solar investor overview, First Solar Surface Pumped Hydroelectric Storage, http://files.shareholder.com/ June 2017”, http://energystorage.org/energy-storage/ -surface-pumped- downloads/ technologies/sub FSLR/4693351233x0x947312/38 (accessed 8 January hydroelectric-storage 0F6F6B-A645-456C-8FBF-FC2582FFA057/ 2018). 2017. p df First_Solar_Investor_Overview_June_ (accessed 23 June 2017). (n.d. – b), Surface Energy Storage Association (2017), “Photovoltaics report”, Fraunhofer ISE Reservoir Pumped Hydroelectric Storage, Fraunhofer Institute for Solar Energy Systems http://energystorage.org/energy-storage/ ise.fraunhofer.de/content/ technologies/surface-reservoir-pumped- www. ISE, Freiburg, hydroelectric-storage, (accessed 8 January dam/ise/de/documents/publications/studies/ (accessed 8 January 2018). Photovoltaics-Report.pdf 2018). EnergyTrend (2017), “Multi-Si to challenge Mono- Fraunhofer ISE (2016), “Photovoltaics report”, Si’s market share as black silicon technology Fraunhofer Institute for Solar Energy Systems http://pv.energytrend.com/news/ matures”, www.ise.fraunhofer.de/content/ ISE, Freiburg, _Mono_Sis_Market_ Multi_Si_to_Challenge dam/ise/de/documents/publications/ studies/ Share_as_Black_Silicon_Technology_Matures. . Photovoltaics-Report.pdf (accessed 3 July 2017). html Fraunhofer ISE (2013), “Levelized cost of electricity (European Renewable Energy Council) and EREC www.ise. renewable energy technologies”, Energy [R]evolution Greenpeace (2010), , EREC/ fraunhofer.de/content/dam/ ise/en/documents/ Greenpeace, Brussels. publications/studies/Fraunhofer-ISE_LCOE_ Renewable_Energy_ (accessed technologies.pdf Wind in Power: 2011 European EWEA (2012), 19 July 2017). Statistics , European Wind Energy Association, Brussels. (2014), Friedman, B., R. Margolis and J. Seel “Comparing Photovoltaic (PV) Costs and The Economics of Wind Energy (2009), EWEA , Deployment Drivers in the Japanese and US European Wind Energy Association, Brussels. Residential and Commercial Markets”, NREL, Golden, Colorado. (2017), Personal communication with S. Exawatt Price, 12 December 2017. Economic Competitiveness of Fu, R. et al. (2015), US Utility-Scale Photovoltaics Systems in 2015: (2010), “Technology assessment of Fichtner Regional Cost Modeling of Installed Cost ($/W) CSP technologies for a site-specific project in and LCOE. ($/kWh), Photovoltaic Specialist South Africa: Final report”, The World Bank and Conference (PVSC), 2015 IEEE 42nd, New ESMAP, Washington, DC. Orleans. 146

147 2017 Fujihara, T., H. Imano and K. Oshima Haysom, J.E. et al. (2015), “Learning curve (1998), analysis of concentrated photovoltaic systems: “Development of pump turbine for seawater Progress Concentrated photovoltaic systems”, pumped storage power plant”, Hitachi Review, in Photovoltaics: Research and Applications Vol. 47, pp. 199-202. , Vol. 23, No. 11, pp. 1678-1686, https://dx.doi. . org/ 10.1002/pip.2567 (2017), 2016 Annual Results, GCL-Poly - http://gcl-poly. todayir.com/attach Heindl Energy (2016), “Gravity storage: A new ment/201703311521191714644172_en.pdf www. (accessed 4 July 2017). solution for large scale energy storage”, fileadmin/user_upload/ heindl-energy.com/ Gravity_Storage_overview_2016-us.pdf GE Reports (2016), “This unique combo of wind (accessed 31 August 2017). and hydro power could revolutionize renewable energy”, www.gereports.com/unique-combo- (2012), World Energy Outlook, IEA/OECD, wind-hydro-power-revolutionize-renewable- IEA energy/ (accessed 8 January 2018). Paris. IEA (2010), Energy Technology Perspectives: (2017), “Power generation technologies GlobalData , Scenarios and Strategies to 2050. capacities”, Generation and Markets Database GlobalData, London. IEA/OECD , Paris. Good, J. and J.X. Johnson (2016), “Impact of inverter loading ratio on solar photovoltaic (2017), “IEA PVPS annual report 2016”, IEA PVPS system performance”, Applied Energy , Vol. International Energy Agency Photovoltaic Power http://dx.doi.org/ 10.1016/j. 177, pp. 475-486, Systems Programme, Paris. apenergy.2016.05.134 . IEA Wind , (2017), Onshore Wind Data Viewer Green, M.A. et al. (2017), “Solar cell efficiency International Energy Agency, Paris, https:// Progress in Photovoltaics: tables (version 49)”, community.ieawind.org/task26/dataviewer. , Vol. 25, No. 1, pp. Research and Applications 3-13, https://dx.doi.org/10.1002/pip.2855 . (2016), IEA Wind Task 26, Offshore Wind IEA Wind Farm Baseline Documentation, International (2008), “Geothermal tomorrow 2008”, DOE- GTP Energy Agency, Paris - https://www.nrel.gov/ GO-102008-2633, Geothermal Technologies fy16osti/66262.pdf docs/ Program, US Department of Energy, Washington, DC. IEA Wind . (2015). Task 26: Wind Technology, Cost, and Performance Trends in Denmark, Germany, Global Wind Report: Annual Market GWEC (2017), Ireland, Norway, the European Union, and the , Global Wind Energy Council, Update 2016 United States: 2007 - 2012. IEA Wind Energy Brussels. Systems. Hance, C.N. (2005), “Factors affecting costs of IEA Wind (2011), Task 26: Multinational case study geothermal power development”, Geothermal of financial cost of wind energy, work package 1, Energy Association, US Department of Energy, final report, IEA Wind Energy Systems. Washington, DC. 147

148 RENEWABLE POWER GENERATION COSTS Success of Geothermal Wells: A Global (2013), IFC IRENA (2016a), The Power to Change: Solar , Study , International Finance Corporation, and Wind Cost Reduction Potential to 2025 Washington, DC. International Renewable Energy Agency, Abu Dhabi, www.irena.org/DocumentDownloads/ Publications /IRENA_Power_to_Change_2016. IPCC IPCC Special Report on Renewable (2011), pdf . Energy Sources and Climate Change Mitigation , Intergovernmental Panel on Climate Change, (2016b), IRENA Geneva. Solar PV in Africa: Costs and Markets , International Renewable Energy www.irena.org/-/media/ Agency, Abu Dhabi, Solar and Wind Power (forthcoming), IRENA Files/IRENA/Agency/Publication/2016/IRENA_ in the G20: Cost Reduction Potential and Solar_PV_Costs_Africa_2016.pdf . Competitiveness to 2030, International Renewable Energy Agency, Abu Dhabi. Renewable Power Generation (2015), IRENA , International Renewable Costs in 2014 IRENA (2017a), Renewable Capacity Statistics 2017 , www.irena.org/ Energy Agency, Abu Dhabi, International Renewable Energy Agency, Abu DocumentDownloads/Publications/IRENA_ Dhabi. RE_Power_Costs . _2014_report.pdf IRENA IRENA Cost & Competitiveness (2017b), (2013a), Renewable Power Generation Costs IRENA Indicators: Rooftop Solar PV , International , International Renewable in 2012: An Overview Renewable Energy Agency, Abu Dhabi. Energy Agency, Abu Dhabi. IRENA Electricity Storage and Renewables: (2017c), Road Transport: The Cost of (2013b), IRENA Costs and Markets to 2030 , International Renewable Solutions , International Renewable Renewable Energy Agency, Abu Dhabi. Energy Agency, Abu Dhabi. IRENA (2017d), REthinking Energy 2017 , IRENA (2013c), “Intellectual property rights: The International Renewable Energy Agency, Abu role of patents in renewable energy technology Dhabi. innovation” (working paper), International Renewable Energy Agency, Abu Dhabi, www. (2017e), IRENA Renewable Energy Auctions: irena.org/publications/2013/Jun/Intellectual- , International Renewable Energy Analysing 2016 - Property-Rights-The-Role-of-Patents-in Agency, Abu Dhabi. . Renewable-Energy-Technology-Innovation IRENA (2017f), Chapter 3 in Perspectives for the IRENA (2012a), Renewable Energy Technologies: Energy Transition: Investment Needs for a Low- , Vol. 1, Costs Analysis Series, Biomass for Power , IEA and IRENA. Carbon Energy System International Renewable Energy Agency, Abu Dhabi. Bioenergy from Degraded Land in (2017g), IRENA , International Renewable Energy Agency, Africa Abu Dhabi. 148

149 2017 IRENA (2014), “Black silicon: fabrication Liu, X. et al. (2012b), Renewable Energy Technologies: methods, properties and solar energy Costs Analysis Series, Concentrating Solar power , Energy & Environmental Science , applications”, Vol. 2, International Renewable Energy Agency, https://dx.doi. Vol. 7, No. 10, pp. 3223-3263, Abu Dhabi. . org/10.1039/C4EE01152J Renewable Energy Technologies: IRENA (2012c), (2017), “LONGi Solar willing to LONGi Solar Costs Analysis Series, Hydropower , Vol. 3, share LIR technology with industry”, International Renewable Energy Agency, Abu : // http Dhabi. en.longi-solar.com/Home/Events/press_detail/ id/9 _LONGi_Solar_Willing_to_Share_LIR_ Technology_with_Industry Renewable Energy Technologies: (2012d), IRENA , Vol. 4, Costs Analysis Series, Solar Photovoltaic (2016), Luka, T., C. Hagendorf and M. Turek International Renewable Energy Agency, Abu “Multicrystalline PERC solar cells: Is light-induced Dhabi. degradation challenging the efficiency gain of , Photovoltaics International rear passivation?”, IRENA (2012e), Renewable Energy Technologies: Vol. 32. Costs Analysis Series, Wind Power , Vol. 5, International Renewable Energy Agency, Abu (2016), “Evaluating the value Lunz, B. et al. Dhabi. of concentrated solar power in electricity AIP systems with fluctuating energy sources”, ITRPV International Technology Roadmap (2017), , Vol. 1734, https:// Conference Proceedings , Eighth for Photovoltaic (ITRPV) 2016 Results dx.doi.org/10.1063/1.4949251 (accessed 19 July Edition, International Technology Roadmap for 2017). www.itrpv.net/.cm4all/iproc.php/ Photovoltaic, . ITRPV Eighth Edition 2017.pdf?cdp=a MacDonald, M. (2011), Costs of Low-Carbon Generation Technologies , Committee on Climate Kimura, K. and R. Zissler (2016), “Comparing Change, London. prices and costs of solar PV in Japan and Germany: The reasons why solar PV is more MAKE Consulting (2017a), Q3 2017 – Global Wind expensive in Japan”, Tokyo, Renewable Power Market Outlook Update, Aarhus. www.renewable-ei.org /en/ Energy Institute, images/pdf/20160331/JREF_Japan_Germany_ . solarpower_costcomparison_en .pdf Global Wind Turbine (2017b), MAKE Consulting Trends , MAKE Consulting, Aarhus. Kraus, K. et al. (2016), “Fast regeneration processes to avoid light-induced degradation in MAKE Consulting (2017c), Global Wind Turbine IEEE Journal of multicrystalline silicon solar cells”, , MAKE Consulting, Aarhus. O&M Photovoltaics , Vol. 6, No. 6, pp. 1427-1431, https : // . dx.doi.org/10.1109/JPHOTOV.2016.2598273 (2015a), MAKE Consulting Wind Energy Levelised , MAKE Cost of Electricity: Globally Competitive (2017), “Empirically observed Lilliestam, J. et al. Consulting, Aarhus. learning rates for concentrating solar power and their responses to regime change”, Nature Global Wind Turbine (2015b), MAKE Consulting https://dx.doi.org/10.1038/ , Vol. 2, No. 7, Energy , MAKE Consulting, Aarhus. Trends . nenergy.2017.94 149

150 RENEWABLE POWER GENERATION COSTS (2013), “Global wind turbine MAKE Consulting Reuters (2016), “Siemens, Gamesa to form world’s largest wind farm business”, Reuters, trends 2012: Lowering the levelized cost of article/us-gamesa- electricity”, (research note), MAKE Consulting, London, www.reuters.com/ Aarhus. worlds- m-a-siemens/siemens-gamesa-to-form- largest-wind-farm-business-idUSKCN0Z22JC . (2015), “An assessment of the net Mehos, M. et al. (2015), “GE clears final hurdle to $14 billion value of CSP systems integrated with thermal Reuters energy storage”, Energy Procedia Alstom deal”, Reuters, London, www. reuters. , Vol. 69, https://dx.doi.org /10.1016/j. pp. 2060-2071, com/article/us-alstom-m-a-general-electric-eu/ . egypro.2015.03.219 ge-clears-final-hurdle-to-14-billion-alstom-deal- . idUSKCN0R81QR20150908 Moner-Girona, Magda, et al. (2018), Review of Seel, J., Barbose, G.L. and Wiser, R.H. (2014), Photovoltaic Technology Cost and Performance Projections , self-published. “An analysis of residential PV system price differences between the United States and Germany”, , Vol. 69, pp. 216-226, Energy Policy Neij, L. (2008), “Cost development of future . https://dx.doi.org/10.1016/j.enpol.2014.02.022 technologies for power generation: A study based on experience curves and complementary Shim, J.-M. et al. (2012), “17.6% conversion , bottom-up assessments”, Energy Policy efficiency multicrystalline silicon solar cells https://dx.doi. Vol. 36, No. 6, pp. 2200-2211, using the reactive ion etching with the damage . org/10.1016/j.enpol.2008.02.029 International Journal of removal etching”, , Vol. 2012, pp. 1-6, Photoenergy https://dx.doi. (2016), “Light-induced Padmanabhan, M. et al. . org/10.1155/2012/248182 degradation and regeneration of multicrystalline Physica silicon Al-BSF and PERC solar cells”, (2017), CIS Modules STC Solar Frontier Status Solidi RRL , Vol. 10, No. 12, pp. 874-881, www.solar-frontier.com/eng/ characteristics, https://dx.doi .org/10.1002/pssr.201600173 . solutions/modules/S002210. (accessed 23 html June 2017). Photon Consulting (2017), Data from Photon Consulting’s Demand subscription v 2017, US, (2017a), CSP Projects around the SolarPACES Photon Consulting. World, www.solarpaces.org/csp-technology/ csp-projects-around (accessed 17 -the-world Pitz-Paal, R. (2017), “Concentrating solar power: July 2017). , Vol. Still small but learning fast”, Nature Energy ://dx.doi.org/10.1038/ https 2, No. 7, p. 17095, (2017b), Concentrating Solar Power SolarPACES . nenergy.2017.95 www.nrel.gov/csp/solarpaces/ Projects, NREL, (accessed 18 July 2017). (2017), “2017: A year for PERC & pv magazine , No. 03-2017. pv magazine black silicon”, SolarPACES (2016), “China announces the first group of CSP demonstration projects”, (2017), PV module price index pvXchange www.solarpaces.org/china-announces-the- .com/ www.pvxchange (PV marketplace), first-group-of-csp-demonstration-projects . priceindex/ (accessed 8 January 2018). 150

151 2017 (2016), “Solar in China: NEA officially SolarPV.TV Vestas Wind Systems A/S (2005-2017), Annual releases China’s 13th Five-Year-Plan for solar Financial Reports, Copenhagen. h t t p : // development, confirms 110 GW target”, solarpv.tv/index.php/2016/12/17/solar-in- china- Wang, J. et al. (2017), “Status and future strategies nea-officially-releases-chinas-13th-five-year- Energy for Concentrating Solar Power in China”, solar-development-confirms-110-gw- plan-for- , Vol. 5, No. 2, pp. 100-109, Science & Engineering (accessed 17 July 2017). target/ . https://dx.doi.org/10.1002/ese3.154 Global Market Outlook (2017), SolarPower Europe (2017), Annual Offshore Statistics, WindEurope , SolarPower Europe, for Solar Power: 2017/2021 Brussels. Brussels. (2012), Annual Offshore Statistics, WindEurope Cost Reduction Potential of Large Scale (2014), S TA Brussels. Solar PV , Solar Trade Association, London. Wiser, R. and M. Bolinger (2017), 2016 Wind (2016), “Streamlining photovoltaic Strupeit, L. Technologies Market Report , LBNL, CA. deployment: The role of local governments Energy Procedia , Vol. in reducing soft costs”, 2015 Wind (2016), Wiser, R.H. and M. Bollinger https://dx.doi 88, pp. 450-454, .org/10.1016/j. , LBNL, CA. Technologies Market Report . egypro.2016.06.023 World Bank (2017), World Bank Agricultural Theologitis & Mason (2015), “Potential for Cost Commodities Index, World Bank, Washington, Reduction of PV Technology – Impact of . www.data.worldbank.org DC, CHEETAH Research Innovations”, 31st European Photovoltaic Solar Energy Conference and World Bank Group (2017), World Bank Exhibition , WIP, Munich. Proceedings Commodities Price Data (The Pink Sheet), World Bank Group, Washington, DC, www.worldbank. Turchi et al. Current and Future Costs for (2010a), org/en/research/commodity- markets . Parabolic Trough and Power Tower Systems in , NREL, Boulder. the US Market Ying, Z. et al. (2016), “High-performance black multicrystalline silicon solar cells by a highly Parabolic Trough Reference (2010b), Turchi et al. simplified metal-catalyzed chemical etching Plant for Cost Modeling with the Solar Advisor method”, IEEE Journal of Photovoltaics , Vol. 6, , NREL, Boulder. Model No. 4, pp. 888-893, https://dx.doi.org/10.1109/ . JPHOTOV.2016.2559779 Design Limits and Solutions for (2015), UpWind , EWEA, Brussels. Very Large Wind Turbines Zarrouk, S.J. and H. Moon (2014), “Efficiency of geothermal power plants: A worldwide review”, Utterback & Abernathy (1975), “A dynamic model Geothermics , Vol. 51. of process and product innovation”, Omega 3:639-656. 151

152 RENEWABLE POWER GENERATION COSTS ANNEX I COST METRIC METHODOLOGY power producer or an individual or community ost can be measured in a number of different looking to invest in small-scale renewables. The ways, and each way of accounting for the cost C analysis excludes the impact of government of power generation brings its own insights. The incentives or subsidies, system balancing costs costs that can be examined include equipment associated with variable renewables and any costs (e.g. PV modules or wind turbines), financing system-wide cost-savings from the merit order costs, total installed cost, fixed and variable effect. Furthermore, the analysis does not take operating and maintenance costs (O&M), fuel costs into account any CO (if any) and the levelised cost of energy (LCOE). pricing, nor the benefits of 2 renewables in reducing other externalities (e.g. The analysis of costs can be very detailed, but reduced local air pollution or contamination of for comparison purposes and transparency, the natural environment). Similarly, the benefits of the approach used here is a simplified one that renewables being insulated from volatile fossil fuel focusses on the core cost metrics for which good prices have not been quantified. These issues are data is readily available. This allows greater scrutiny important, but are covered by other programmes of the underlying data and assumptions, improves of work at IRENA. transparency and confidence in the analysis, and also facilitates the comparison of costs by country Clear definitions of the technology categories are or region for the same technologies in order to provided, where this is relevant, to ensure that identify the key drivers in any differences. cost comparisons are robust and provide useful insights (e.g. off-grid PV vs. utility-scale PV). • The five key indicators that have been selected Similarly, functionality has to be distinguished are: from other qualities of the renewable power • generation technologies being investigated (e.g. Equipment cost (factory gate, FOB, and concentrating solar power with and without delivered at site); thermal energy storage). This is important to • Total installed project cost, including fixed ensure that system boundaries for costs are financing costs; clearly set and that the available data are directly • comparable. Other issues can also be important, Capacity factor by project; and such as cost allocation rules for combined heat The levelised cost of electricity. and power plants, and grid connection costs. The analysis in this paper focuses on estimating The data used for the comparisons in this paper the costs of renewables from the perspective of come from a variety of sources, such as IRENA private investors, whether they are a state-owned Renewable Costing Alliance members, business electricity generation utility, an independent 152

153 2017 (DCF) analysis. This method of calculating the journals, industry associations, consultancies, cost of renewable energy technologies is based governments, auctions and tenders. Every effort on discounting financial flows (annual, quarterly has been made to ensure that these data are or monthly) to a common basis, taking into directly comparable and are for the same system consideration the time value of money. Given the boundaries. Where this is not the case, the data capital-intensive nature of most renewable power have been corrected to a common basis using generation technologies and the fact that fuel the best available data or assumptions. This data costs are low, or often zero, the weighted average has been compiled into a single repository – The cost of capital (WACC), often also referred to as IRENA Renewable Cost Database – that includes a the discount rate, used to evaluate the project has mix of confidential and public domain data. a critical impact on the LCOE. An important point is that, although this report There are many potential trade-offs to be tries to examine costs, strictly speaking, the data considered when developing an LCOE modelling available are actually prices, and are sometimes approach. The approach taken here is relatively not even true market average prices, but price simplistic, given the fact that the model needs indicators (e.g. surveyed estimates of average to be applied to a wide range of technologies in module selling prices in different markets). The different countries and regions. However, this difference between costs and prices is determined has the additional advantage that the analysis by the amount above, or below, the normal profit is transparent and easy to understand. In that would be seen in a competitive market. The addition, more detailed LCOE analyses result in rapid growth of renewables markets from a small a significantly higher overhead in terms of the base means that the market for renewable power granularity of assumptions required. This often generation technologies is sometimes no well- gives the impression of greater accuracy, but when balanced. As a result, prices can rise significantly it is not possible to robustly populate the model above costs in the short term if supply is not with assumptions, or to differentiate assumptions expanding as fast as demand, while in times of based on real world data, then the “accuracy” of excess supply, losses can occur and prices the approach can be misleading. may be below production costs. This can make analysing the cost of renewable power generation The formula used for calculating the LCOE of technologies challenging for some technologies renewable energy technologies is: in given markets at certain times. Where costs are significantly above or below what might be expected to be their long-term trend, every effort has been made to identify the causes. Although every effort is made to identify the reasons why costs differ between markets for Where: individual technologies, the absence of the LCOE = t he average lifetime levelised cost of electricity detailed data required for this type of analysis generation; often precludes a definitive answer. IRENA has It = investment expenditures in the year t; conducted a number of analyses focussing on Mt = Operations and maintenance expenditures in the year t; Ft = fuel expenditures in the year t; individual technologies and markets in an effort to Et = electricity generation in the year t; fill this gap (IRENA, 2016a,b). r = discount rate; and The LCOE of renewable energy technologies varies n = life of the system. by technology, country and project, based on the All costs presented in this report are real 2016 renewable energy resource, capital and operating USD; that is to say, after inflation has been taken costs, and the efficiency/performance of the into account unless otherwise stated. The LCOE is technology. The approach used in the analysis the price of electricity required for a project where presented here is based on a discounted cash flow 153

154 RENEWABLE POWER GENERATION COSTS The analysis in this report assumes a WACC for a revenues would equal costs, including making a project of 7.5% (real) in Organisation for Economic return on the capital invested equal to the discount Co-operation and Development (OECD) countries rate. An electricity price above this would yield a and China, where borrowing costs are relatively low greater return on capital, while a price below it and stable regulatory and economic policies tend would yielder a lower return on capital, or even a to reduce the perceived risk of renewable energy loss. projects, and 10% in the rest of the world. These As already mentioned, although different cost assumptions are average values, but the reality measures are useful in different situations, the is that the cost of debt and the required return LCOE of renewable energy technologies is a on equity, as well as the ratio of debt-to-equity, widely used first order measure by which power varies between individual projects and countries generation technologies can be compared. More depending on a wide range of factors. This can have detailed DCF approaches taking into account a significant impact on the average cost of capital taxation, subsidies and other incentives are used and the LCOE of renewable power projects. It also by renewable energy project developers to assess highlights an important policy issue: in an era of the profitability of real world projects, but are low equipment costs for renewables, ensuring that beyond the scope of this report. policy and regulatory settings minimise perceived risks for renewable power generation projects The calculation of LCOE values in this report is can be a very efficient way to reduce the LCOE by based on project specific total installed costs and lowering the WACC. capacity factors, as well as the O&M costs detailed in the individual chapters. The standardised assumptions used for calculating the LCOE include the WACC, economic life and cost of bioenergy feedstocks. Economic life Weighted average cost of capital, real OECD and China Rest of the world Wind Power 25 25 Solar PV CSP 25 10% 7. 5% Hydropower 30 Biomass for power 20 25 Geothermal 154

155 2017 155

156 RENEWABLE POWER GENERATION COSTS ANNEX II IRENA RENEWABLE COST DATABASE wind and CSP, where the relatively small number of he composition of the IRENA Renewable Cost projects can more easily be tracked. The database Database largely reflects the deployment of T for onshore wind and hydropower is representative renewable energy technologies over the last 10- from around 2007, with comprehensive data from 15 years. Most projects in the database are in around 2009 onwards. Gaps for some countries China (388 GW), India (89 GW), the United States (in the top 10 for deployment in a given year) (88 GW), and Brazil (69 GW). It is significantly in some years require recourse, however, to more difficult to collect cost data from OECD secondary sources in order to develop statistically countries, however, due to greater difficulties representative averages. Data for solar PV at with confidentiality issues. The exception is the the utility-scale has only become available more United States, where the nature of support policies recently and the database is representative from leads to greater quantities of project data being around 2011 onwards, and comprehensive from available. After these four major countries, Canada around 2013 onwards. is represented by 26 GW of projects, the Russian Federation by 25 GW, Vietnam by 23 GW, Pakistan The data available so far for 2017 represents by 21 GW, Chile and the United Kingdom by 16 GW roughly 50-60% of what has become available to each, and Germany by 15 GW of projects. IRENA in previous years. At the time of release, for 2017, the total capacity of projects in the database With data for a small number of very large totalled around 56 GW. The main technologies hydropower projects and the more extensive time where data is yet to become available are onshore series available, hydropower is the largest single wind and, to a lesser extent, solar PV. As such, technology represented in the IRENA Renewable data for 2017 needs to be considered preliminary Cost Database. This technology has provided and subject to change. Typically, given previous cost data for 570 GW of projects since 1961, with experience with data collection, over the next one around 90% of those projects commissioned in the to two years, we would expect the volume of data year 2000 or later. The next largest technology available for solar PV (in GW) to double, given represented in the database is onshore wind, with that data has become available for a significant cost data for 268 GW of projects,worldwide. Cost number of projects already. For onshore wind data is available for 118 GW of solar PV projects, power projects, the data in the database could 31 GW of commissioned and proposed offshore grow up to five-fold, given that relatively little data wind projects, 20 GW of biomass for power has currently been finalised. projects and 5 GW each of geothermal and CSP projects. The coverage of the IRENA Renewable Cost Database is more or less complete for offshore 156

157 2017 157

158 RENEWABLE POWER GENERATION COSTS ANNEX III REGIONAL GROUPINGS • • Asia: Afghanistan; Bangladesh; Bhutan; Brunei Eurasia: Armenia; Azerbaijan; Georgia; Russian Federation; Turkey. Darussalam; Cambodia; China; Democratic People’s Republic of Korea; India; Indonesia; • Europe: Albania; Andorra; Austria; Belarus; Japan; Kazakhstan; Kyrgyzstan; Lao People’s Belgium; Bosnia and Herzegovina; Bulgaria; Democratic Republic; Malaysia; Maldives; Croatia; Cyprus; Czech Republic; Denmark; Mongolia; Myanmar; Nepal; Pakistan; Estonia; Finland; France; Germany; Greece; Philippines; Republic of Korea; Singapore; Hungary; Iceland; Ireland; Italy; Latvia; Sri Lanka; Tajikistan; Thailand; Timor-Leste; Liechtenstein; Lithuania; Luxembourg; Malta; Turkmenistan; Uzbekistan; Viet Nam. Monaco; Montenegro; Netherlands; Norway; • Poland; Portugal; Republic of Moldova; Romania; Africa: Algeria; Angola; Benin; Botswana; San Marino; Serbia; Slovakia; Slovenia; Spain; Burkina Faso; Burundi; Cabo Verde; Cameroon; Sweden; Switzerland; the former Yugoslav Central African Republic; Chad; Comoros; Republic of Macedonia; Ukraine; United Congo; Côte d’Ivoire; Democratic Republic of Kingdom of Great Britain and Northern Ireland. the Congo; Djibouti; Egypt; Equatorial Guinea; Eritrea; Ethiopia; Gabon; Gambia; Ghana; • Middle East: Bahrain; Iran (Islamic Republic of); Guinea; Guinea-Bissau; Kenya; Lesotho; Liberia; Iraq; Israel; Jordan; Kuwait; Lebanon; Oman; Libya; Madagascar; Malawi; Mali; Mauritania; Qatar; Saudi Arabia; Syrian Arab Republic; Mauritius; Morocco; Mozambique; Namibia; United Arab Emirates; Yemen. Niger; Nigeria; Rwanda; Sao Tome and Principe; • Senegal; Seychelles; Sierra Leone; Somalia; North America: Canada; Mexico; United States South Africa; South Sudan; Sudan; Swaziland; of America. Togo; Tunisia; Uganda; United Republic of • Oceania: Australia; Fiji; Kiribati; Marshall Islands; Tanzania; Zambia; Zimbabwe. Micronesia (Federated States of); Nauru; New • Zealand; Palau; Papua New Guinea; Samoa; Central America and the Caribbean: Antigua Solomon Islands; Tonga; Tuvalu; Vanuatu. and Barbuda; Bahamas; Barbados; Belize; Costa Rica; Cuba; Dominica; Dominican Republic; El • South America: Argentina; Bolivia (Plurinational Salvador; Grenada; Guatemala; Haiti; Honduras; State of); Brazil; Chile; Colombia; Ecuador; Jamaica; Nicaragua; Panama; Saint Kitts Guyana; Paraguay; Peru; Suriname; Uruguay; and Nevis; Saint Lucia; Saint Vincent and the Venezuela (Bolivarian Republic of). Grenadines; Trinidad and Tobago. 158

159 2017 159

160 Renewable Power Generation Costs in 2017 www.irena.org © IRENA 2018

Related documents

17 8652 GSR2018 FullReport web final

17 8652 GSR2018 FullReport web final

RENE WA BL E S 2018 GLOBAL STATUS REPORT A comprehensive annual overview of the state of renewable energy. 2018

More info »