Planning tools for long-term power system expansion with a higher share of variable renewable energy. Asami Miketa November 29, 2017, Amman, Jordan

Planning tools for long-term power system expansion with a higher share of variable renewable energy Asami Miketa November 29, 2017, Amman, Jordan

About IRENA International organization supported by governments of member countries» HQ: Abu Dhabi, UAE» Members: 180 countries (153 ratified)» Mandate: Sustainable deployment of six RE resources (Biomass, Geothermal, Hydro, Ocean, Solar, Wind)

Power sector transformation at IRENA Market design, regulation, business models Long term, least cost capacity expansion plan Find the optimal pathway for power sector transformation Unit commitment and economic dispatch Grid studies Work with countries in conducting planning studies with renewables using power sector modelling tools Develop guidelines and tools to support better planning practices and disseminate them Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies

Key questions addressed in this presentation What are the characteristics of VRE that require change in the long-term planning process? How can these changes be implemented in planning tools?

Agenda Planning processes and tools VRE s planning impacts Generation expansion planning Check point 1 Definition of time Check point 2 Adequate firm capacity Check point 3 Flexibility of a system Check point 4 Transmission investment needs Check point 5 Stability related operation constants Planning solutions summary IRENA 2017

Power system planning: fundamentals How much electricity demand will there be? How much and what type of generation is needed to serve this demand? What enhancements to the network are needed to ensure the reliable supply of electricity? Energy/power system models are used to answer these questions while taking into account economic and technical consequences of alternative choices.

Power sector planning: Four focus areas for techno-economic analysis Capacity (MW) 5,000 4,000 3,000 2,000 1,000 - Generation expansion planning Ministry of Energy Planning agency Utility Dispatch simulation Utility Regulators TSO Geo-spatial planning Government planning office Planning agency Utility TSO Technical network studies TSO Regulator Project developer

2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 Capacity (MW) 1. Generation expansion planning 5,000 4,000 3,000 2,000 1,000 - Future energy mix and investment path Compliance with long-term energy policy goals Political consensus making Linked often with non-power sector planning Department of Energy Regulatory commission Utility Specialized agency 8

2. Geo-spatial planning Zone identification for investment promotion Generation siting and long-term transmission development needs High-level screening scenarios for transmission network development Source: ERCOT 9

3. Dispatch simulation Fuel and operation cost calculation Maintenance scheduling Economic power flow Market and regulation design VRE integration study 10

C2 XX MW 4. Technical network studies TRANSMISION SYSTEM OVERVIEW 2016 Zone 8 Zone 6 N3 XX MW Area 1 Peak demand: 500 MW Low demand: 350 MW Installed Capacity: 600 MW Zone 13 Zone 4 N6 XX MW N2 XX MW S1 45 MW Zone 9 N1 XX MW N7 XX MW Total Transfer Capacity 500 MW Zone 12 N4 XX MW N8 XX MW Zone 2 N9 S2 155 MW N5 XX MW Area 2 Peak demand: 700 MW Low demand: 400 MW Installed Capacity:1500 MW Zone 5 Zone 11 Total Transfer Capacity 564 MW Area 3 Peak demand: 1500 MW Low demand: 565 MW Installed Capacity: 1200 MW N10 S3 314 MW Zone 7 Zone 1 C1 XX MW C3 XX MW Zone 10 C4 XX MW Zone 3 Load flow analysis Simulate power flow of a given network under a challenging situation Identify network enhancement needs VRE integration study Stability assessment Simulation of frequency and voltage response in a network to a contingency event VRE integration study 11

Time dimensions of power sector planning Typical time resolution Seasonally to sub-daily Generation expansion planning Typical time frame 20-40 years Seasonally to sub-daily (Static) Hourly to sub-hourly Geo-spatial planning Dispatch simulation 5-20 years Weeks-years Sub-hourly to subseconds Technical network studies Snapshot Near-term Planning time horizon Long-term Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 12

Planning tools Generation expansion models (time resolution: hours seasons) Low detail/ Wide scope Geo-spatial planning models (time resolution: hours seasons) Dispatch simulation models (time resolution: minutes hours) Generation and network capacities Network topology Static grid models (time resolution: single point) Dynamic grid models (time resolution: milliseconds minutes) Highly-resolved dispatch and operational details Steady state grid currents and voltages High detail/ Narrow scope Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 13

Modelling software indicative coverage Zone 8 Zone 6 Area 1 Peak demand: 500 MW Low demand: 350 MW Installed Capacity: 600 MW Zone 13 Zone 4 Zone 9 N1 XX MW Total Transfer Capacity 500 MW Zone 12 N8 XX MW Zone 2 N9 Zone 11 Area 2 Peak demand: 700 MW Low demand: 400 MW Installed Capacity:1500 MW Zone 5 Total Transfer Capacity 564 MW Area 3 Peak demand: 1500 MW Low demand: 565 MW Installed Capacity: 1200 MW N10 Zone 7 C2 XX MW Zone 1 C3 XX MW Zone 10 Zone 3 MESSAGE Quantum GIS MARKAL/TIMES ArcGIS NEPLAN PLEXOS-LT PLEXOS-ST BALMOREL Power Factory Grid-View PSSE OPT-GEN WASP SDPP WASP GT-MAX GT-Max TRANSMISION SYSTEM OVERVIEW 2016 N3 XX MW N6 XX MW N2 XX MW N7 XX MW N4 XX MW N5 XX MW C1 XX MW S3 314 MW C4 XX MW S2 155 MW S1 45 MW Cap expansion Geo-spatial Dispatch Static Dynamic

Example of the tools used in the region Generation planning WASP (IAEA), OPTGEN (PSR), EGEAS (EPRI), Aurora (EPIS) Renewable / geospatial planning ArcGIS (ESRI), Patro Solar, SAM (NREL) Operational planning EMS, SPPD (PSR) Transmission planning Power factory (Digsilent), PSS/E (Siemense)

Impacts of VRE on planning VRE PROPERTIES SYSTEM PROPERTIES PLANNING STAGES Non- Synchronous Locationconstrained & distributed Uncertainty Frequency & Voltage response provision Voltage control needs Transmission capacity needs Flexibility needs Geo-spatial planning & Technical network studies Geo-spatial planning & Technical network studies Dispatch simulation Variability Flexibility needs Firm capacity provision Generation expansion planning Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 17

Long-term investment implications of VRE s system properties Generation Networks Adequacy Security of operation Firm capacity Flexibility Transmission capacity Voltage control capability Stability (frequency response and voltage response) Most relevant High relevance System-specific Near-term relevance Source: IRENA (2017), Planning for the Renewable Future 18

With variable renewable energy Feedback from all the levels Long-term energy planning models (time resolution: Firm capacity hours seasons) Low detail/ Wide scope Geo-spatial planning models (time resolution: hours seasons) New transmission Dispatch simulation models (time resolution: Flexibility minutes hours) Generation and network capacities Network topology Static grid models (time resolution: single point) Transmission enhancement Dynamic grid models Stability Highly-resolved dispatch and operational details Steady state grid currents and voltages High detail/ Narrow scope Two solutions: institutional and modelling

Institutional solutions Better coordination at different planning stages across planning bodies Clearer link with longer-term goals E.g., Generation planning and transmission planning E.g., network studies and generation planning

Modelling solutions Creation of a super-model Model coupling Representation of the key VRE features in a simplified manner - Input data preparation (better temporal and spatial resolution RE generation data) - Constraints Part 2 of the report serves as a catalogue of modelling methodologies

Generation capacity expansion planning Commonly used modeling software BALMOREL OptGen MESSAGE PLEXOS-LT WASP OSeMOSYS MARKAL/TIMES etc Key differences: model scopes, interfaces, update frequency, user support, and cost 23

Notes on choice of software Key differences: model scopes, interfaces, update frequency, user support, and cost The choice of software is a secondary issue; more important is how to better use them! Difficult to make an objective assessment on desirability of one software than others Discuss with the software developer and the key software issues for VRE are summarized as five check points 24

Agenda Planning work at IRENA Planning processes and tools Generation expansion planning Check point 1 Definition of time Check point 2 Adequate firm capacity Check point 3 Flexibility of a system Check point 4 Transmission investment needs Check point 5 Stability related operation constants Planning solutions summary IRENA 2017

Checkpoint 1: Definition of time Source: EIA (2015) 26

The choice of the right temporal resolutions The results of the analysis is highly dependent to the temporal resolution of the models 27 PV cost 1$/KW PV cost 0.5 $/KW (Source: Merrick, 2016)

Checkpoint 1: Definition of time How? Representation of VRE generation in the model should be based on meteorological data Data sources: Observation data Global re-analysis data 28

Check point 1 Does the model reflect the solar and wind variability based on meteorological data? 29

Check point 2: Adequate firm capacity VRE PROPERTIES SYSTEM PROPERTIES PLANNING STAGES Non- Synchronous Locationconstrained & distributed Uncertainty Frequency & Voltage response provision Voltage control needs Transmission capacity needs Flexibility needs Geo-spatial planning & Technical network studies Geo-spatial planning & Technical network studies Dispatch simulation Variability Flexibility needs Firm capacity provision Generation expansion planning Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 31

Firm capacity / capacity credits Driven by temporal correlation of VRE and load pattern Sun is not guaranteed to shine when needed Conventional generators are guaranteed to generate when needed 1 MW of solar generators < 1 MW of conventional generators Lower capacity credit Higher capacity credit Lower capacity credit 32

Planning for adequate firm capacity A system needs to have sufficient generation capacity even during the time of high demand / low solar availability How much renewable energy contributes to firm capacity and to planning reserve margin needs to be evaluated Lower capacity credit means lower utilization of the rest of the system Source: Mils and Wiser (2012) 33

Representing adequate firm capacity How are the capacity credits estimated and used in the modelling tools? Detailed methodology based on reliability Using the probabilistic reliability indicators Simplified methodologies Capacity factor during the peak hours Rule-of-thumb US system operators (in 2012) Reliability based 9 Statistical analysis 6 Peak hours 10 Rule-of-thumb 1 Source: Rogers and Porter (2012) EU system operators (in 2014) Reliability based 2 Rule-of-thumb 8 Source: CEER (2014) 34

Check point 2 Is the capacity credit of VRE reflected in the reserve margin requirement in the model, so that long-term generation plans ensure the sufficient generation at all times? 35

Check point 3: Flexibility needs VRE PROPERTIES SYSTEM PROPERTIES PLANNING STAGES Non- Synchronous Locationconstrained & distributed Uncertainty Frequency & Voltage response provision Voltage control needs Transmission capacity needs Flexibility needs Geo-spatial planning & Technical network studies Geo-spatial planning & Technical network studies Dispatch simulation Variability Flexibility needs Firm capacity provision Generation expansion planning Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 37

Flexibility Flexibility requirements» Variability - fast changing VRE output increase in ramping capability may be required» Uncertainty forecast and estimation errors increase in operational reserve may be required Flexibility sources» Ramp rate» Minimum load levels» Start-up times» Storage» Interconnectors» Demand response Lack of flexibility would result in inefficient operation of power systems 38

Planning and modelling flexibility How are they covered in the modelling tools? Flexibility parameters of various technologies Flexibility supply needs to be matched with the demand for flexibility A separate flexibility analysis may be required using a dispatch simulation tool Source: Denholm, P., Hand, M. (2011) 39

168 175 182 189 196 203 210 217 224 231 238 245 252 259 266 273 280 287 294 301 308 315 322 329 MW IRENA flexibility tool Capacity expansion + dispatch optimization tool Assessment flexibility needs of a given capacity mix 1. Input 2. Run the model 3. Results 4. Alternative run Estimated generation mix at 2030 Run dispatch New investments disabled Identify flexibility shortages Check if any other issues Same input data, allow investments Get least cost flexibility solutions Flexibility metrics loss of load, curtailment Flexibility options - Transmission investment, batteries, DSM, investment in new capacity 2030ref, hydro -15% 16000 14000 12000 10000 8000 6000 4000 2000 0 Time loss of lo PV wind Hydro_R Hydro_R Geother ST_bio CC_oil GT_gas ST_coal curtailed

Check point 3 Is the flexibility of a power system properly represented in the model? Do we know how much flexibility would be needed and how much would be met by what? 41

Check point 4: Transmission capacity VRE PROPERTIES SYSTEM PROPERTIES PLANNING STAGES Non- Synchronous Locationconstrained & distributed Uncertainty Frequency & Voltage response provision Voltage control needs Transmission capacity needs Flexibility needs Geo-spatial planning & Technical network studies Geo-spatial planning & Technical network studies Dispatch simulation Variability Flexibility needs Firm capacity provision Generation expansion planning Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 43

Location specificity Egypt, PV zones 128 $/MWh 134 $/MWh Key planning implication: Trade-off between resource quality and transmission investment Source: Lawrence Berkley National Lab, MapRE 44

Planning and modelling transmission capacity Are transmission investment needs taken into account? To which degree site specificity of generation and transmission sites are taken into account? How are they covered in the modelling tools? Cost mark up to generation investment Site specific representation of generation and transmission Source: 45

Check point 4 Is the trade-off between resource quality and transmission investment needs analyzed in the model? Is the resource quality assessed using the georeferenced data? 46

Check point 5: stability constraints VRE PROPERTIES SYSTEM PROPERTIES PLANNING STAGES Non- Synchronous Locationconstrained & distributed Uncertainty Frequency & Voltage response provision Voltage control needs Transmission capacity needs Flexibility needs Geo-spatial planning & Technical network studies Geo-spatial planning & Technical network studies Dispatch simulation Variability Flexibility needs Firm capacity provision Generation expansion planning Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 48

Non-synchronous interface with gird Operating a system with a higher share of non-synchronous generators (e.g., solar PV, and wind) is a challenge as a system requires synchronous generators to provide frequency and voltage response after a contingency event (within a second) to gain stability in a system How are they covered in the modelling tools? Putting a hard constraint on instantaneous penetration limits - it can be rule of thumb or based on a full dynamic study

Check point 5 Do we expect a technical limit to instantaneous penetration of solar and wind? If so, is it a hard limit, or depending on institutional arrangements? Are these limits modelled as scenarios? 50

Summary: modelling solutions A cross-cutting solution Increasing temporal and spatial resolution Representing firm capacity Better calibration of time slice using VRE data Adding capacity credit constraints Representing flexibility Incorporating constraints on flexibility provision Validating flexibility balance in a system Coupling with production cost models Representing transmission capacity Linking grid investment needs with VRE expansion Site specific representation of generation and transmission needs Representing stability constraints 52

Capacity expansion models used in the AUE Tools [developer] Aurora [EPIS] EGEAS [EPRI] OPTGEN/SDDP/NCP [PSR] Ordina [Mercados] Promod [ABB] WASP [IAEA] Countries Oman Egypt, Qatar Morocco Saudi Arabia Saudi Arabia Algeria, Jordan, Libya, Saudi Arabia, Sudan, Tunis, UAE 53

Note on WASP Current version (version IV)» can model up to 12 seasons per year» Uses load duration curve and sequence of loads cannot be modelled» Limited modelling applicability to solar and wind New version (version V):» Hourly simulation of VRE is added as a new feature» Beta-version is avaiable for field testing» Contact: Ms. Ilse Berdellans-Escobar at the IAEA (I.Berdellans- Escobar@iaea.org) 54

Point for discussions Who is responsible for generation planning in your country? Which generation capacity expansion model are you using? How are the VRE s features incorporated in the model? Capacity credit? Flexibility? Site specificity? Stability related operational constraints? What are the recent improvements to represent characteristics of VRE in your country s generation capacity expansion planning (e.g., data collection, modelling etc)? What is your experience with the current modelling software in representing VRE? 55

Asami Miketa, AMiketa@irena.org

Net-load profiles Net load is the demand that must be supplied by conventional generation unless RE is deployed to provide flexibility Implication: Balancing requires more flexibility (i.e., the ability of a power system to respond to change in demand and supply) 57 Source: NREL

Need for flexibility 58 Source: NREL

Two methodological guides from IRENA Long-term energy planning models (time resolution: hours seasons) Geo-spatial planning models (time resolution: hours seasons) Defining long-term capacity mix and transmission infrastructure Production cost models (time resolution: minutes hours) Static grid models (time resolution: single point) Dynamic grid models (time resolution: milliseconds minutes) Grid integration studies for a given capacity and transmission infrastructure 59

Scope of the planning» Planning time horizon» Long-term planning (planning ahead of 20-40 years)» Short-term planning (planning ahead of one year)» Real-time planning» Planning time frame» Years» Weeks» Snap shot» Planning time resolutions» Seasonal» Sub-daily» Hourly» Sub-hourly» Sub-seconds Low detail High detail wider scope narrower scope

Key regional development on renewable policies Solar PV 2.3 GW (Egypt, 2022); 13.6 GW (Algeria, 2030); 5 GW (UAE, 2030); 4.6 GW (Kuwait, 2030); 2 GW (Syria, 2030); 16 GW (KSA, 2040) Solar CSP 5.7 GW (Kuwait, 2030); 2 GW (Algeria, 2030); 1.3 GW (Syria, 2030); 25 GW (KSA, 2040) Pan Arab Strategy for the Development of Renewable Energy Applications (2010-2030), prepared by The Arab ministerial Council for Electricity and Adopted by League of Arab States in 2013 Wind 7.2 GW (Egypt, 2022); 5 GW (Algeria, 2030); 4.2 GW (Morocco, 2030); 9 GW (KSA, 2040)

Scope of the challenges Targets Wind PV CSP Biomass Geothermal % MW MW MW MW MW in generation in capacity Target year Algeria 5010 13575 2000 1000 15 27 37 2030 Bahrain 5 2030 Egypt 7200 2300+ 20 2022 Iraq 300 1 2020 Jordan 800 800 100 50 2020 Kuwait 700 4600 5700 15 2030 Lebanon 400 150-100 12* 2020 Libya 1000 844 375 10 2025 Morocco 4200 4560 52* 2030 Palestine 44 45 20 21 10 2020 Qatar 20 2030 Saudi Arabia 9000 16000 25000 3000 1000 30 2040 Syria 1000 2000 1300 250 30 2030 Tunisia 1755 1510 460 30 2030 UAE - Abu Dhabi 7 2020 UAE - Dubai 5000 25 2030 Yemen 400 8.25 100 6 200 15 2025 * Numbers for Lebanon and Morocco include hydro.

With PPA results for future plants converging for solar & wind Source: IRENA renewable cost analysis 63

Drivers for the PV penetrations Source: IRENA renewable cost analysis 64

Key recommendations Three key recommendation for modelers: Understanding temporal aspects of the VRE resource availability Starting with simple methodologies and scale up to more complex solutions depending on the data availability Coordination at different planning stage

How Reality Surpassed Expert Projections WEO: World Energy Outlook from IEA Source: Metayer et. al (2016), The projections for the future and quality in the past of the World Energy Outlook for solar PV and other renewable energy technologies; and Gilbert et. al (2016), Looking the wrong way: Bias, renewable electricity, and energy modelling in the United States

How Reality Surpassed Expert Projections: e.g. Solar PV AEO: Annual Energy Outlook from US EIA Source: Metayer et. al (2016), The projections for the future and quality in the past of the World Energy Outlook for solar PV and other renewable energy technologies; and Gilbert et. al (2016), Looking the wrong way: Bias, renewable electricity, and energy modelling in the United States

Features of generation expansion planning model Cost minimization over a longtime horizon Capacity build up with time steps of 1-5 years Limited time resolution Limited spatial resolution Winter Spring Summer Autumn Work WE 1 week 1 year Night Day time Peak Evening Total of 32 slices 1 day Example of models with advanced approaches Model name GEMS +CEEM DIMENSION +INTRES Region No. of time slices Germany 432 Europe 192 DIMENSION Europe 7200 US-REGEN US 50 LIMES-EU+ Europe & Middle East and 49 North Africa URBS-EU Europe 8064 - Texas (US) 696 68

Key features of solar and wind» Firm capacity / capacity credit» Flexibility» Transmission investment needs» Stability consideration Typically not well covered in traditional generation expansion planning models and methodologies 69

With variable renewable energy... Feedback from all the levels Long-term energy planning models (time resolution: Firm capacity hours seasons) Low detail/ Wide scope Geo-spatial planning models (time resolution: hours seasons) New transmission Dispatch simulation models (time resolution: Flexibility minutes hours) Generation and network capacities Network topology Static grid models (time resolution: single point) Transmission enhancement Dynamic grid models Stability Highly-resolved dispatch and operational details Steady state grid currents and voltages High detail/ Narrow scope Key elements of the subsequent steps can be pre-analyzed in a simplified manner Source: IRENA (2017), Planning for the Renewable Future: Long-term modelling and tools to expand variable renewable power in emerging economies 70

It is important to do it right from the beginning! How? Coordinated planning across planning bodies Improve long-term energy planning modeling methodologies by incorporating key VRE features Key elements of the subsequent steps can be pre-analyzed in a simplified manner 71