Storage in the energy market

Storage in the energy market Richard Green Energy Transitions 216, Trondheim 1

including The long-run impact of energy storage on prices and capacity Richard Green and Iain Staffell Imperial College Business School Energy Transitions 216, Trondheim2

Demand and Supply /MWh Marginal Cost Nuclear CCGT OCGT GW 3

Demand and Supply /MWh Marginal Cost Nuclear CCGT OCGT GW 4

Costs & Revenues /kw-year Open Cycle Gas Turbine Combined Cycle Gas Turbine Nuclear (French or Chinese) Hours per year 5

Capacity and Load GW OCGT CCGT Nuclear After renewables OCGT CCGT Nuclear Hours per year 6

Storage, Capacity and Load GW OCGT CCGT Nuclear After storage After renewables OCGT CCGT Nuclear Hours per year 7

Optimise dispatch for a year, with starts, no-load and ramping constraints included Assume wholesale prices are equal to marginal cost Generators added or removed until all types break even Making normal profits including a return on capital Storage included at various power capacity levels 4 hours of energy capacity 24-hour storage to be added Simulations with Magic Linear program written in GAMS Imperial College Business School 8

Capacity Mix GB 23, endogenous as storage varies GW 6 5 4 3 2 1 Storage Hydro OCGT Biomass and Waste CCGT Coal Nuclear 2GW 5 GW 1 GW Storage Capacity Imperial College Business School 9

Output Mix GB 23, endogenous as storage varies TWh 2 15 1 5 Storage Hydro OCGT Biomass and Waste CCGT Coal Nuclear 2GW 5 GW 1 GW Storage Capacity Imperial College Business School 1

Price variation over the day Storage capacity: 2 GW, 8 GWh /MWh 55 5 45 4 4 3 35 3 2 25 2 1 15 1 5 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 2 21 22 23 24 9% to 95% 75% to 9% 25% to 75% 1% to 25% 5% to 1% Min to 5% Median Imperial College Business School 11

Price variation over the day Storage capacity: 5 GW, 2 GWh /MWh 55 5 45 4 4 3 35 3 2 25 2 1 15 1 5 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 2 21 22 23 24 9% to 95% 75% to 9% 25% to 75% 1% to 25% 5% to 1% Min to 5% Median Imperial College Business School 12

Price variation over the day Storage capacity: 1 GW, 4 GWh /MWh 55 5 45 4 4 3 35 3 2 25 2 1 15 1 5 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 2 21 22 23 24 9% to 95% 75% to 9% 25% to 75% 1% to 25% 5% to 1% Min to 5% Median Imperial College Business School 13

Price-duration curves GB, 23 with endogenous capacity /MWh 3 2 Storage Capacity: 2 GW, 8 GWh 5 GW, 2 GWh 1 GW, 4 GWh 1-1.1.2.3.4.5.6.7.8.9 1 Imperial College Business School 14

Price-duration curves GB, 23 with endogenous capacity /MWh 3 2 Storage Capacity: 2 GW, 8 GWh 5 GW, 2 GWh 1 GW, 4 GWh 1-1.1.2.3.4.5.6.7.8.9 1 Imperial College Business School 15

Price-duration curves GB, 23 with endogenous capacity /MWh 3 2 Storage Capacity: 2 GW, 8 GWh 5 GW, 2 GWh 1 GW, 4 GWh 1-1.1.2.3.4.5.6.7.8.9 1 Imperial College Business School 16

The impact of storage on prices Endogenous generating capacities /MWh 11 1 9 8 demand-weighted time-weighted offshore wind onshore wind 7 6 2 GW, 8 GWh 5 GW, 2 GWh 1 GW, 4 GWh Imperial College Business School 17

But storage is versatile Power Capacity Energy Arbitrage Energy Balancing Transmission Constraints Distribution Constraints The next few slides draw heavily on work by Goran Strbac and his coauthors Imperial College Business School 18

It s profitable to be charged more Arbitrage only State of charge Imperial College Business School & Department of Electronic & Electrical Engineering 19

It s profitable to be charged more Arbitrage and balancing State of charge Imperial College Business School & Department of Electronic & Electrical Engineering 2

It s profitable to be charged more Arbitrage only Arbitrage and balancing State of charge Imperial College Business School & Department of Electronic & Electrical Engineering 21

Average value of storage Distributed to near customers Opex: save fuel costs from buy cheap, sell dear Capex: save cost of investment in generation, transmission, interconnectors or distribution system as power can be stored when the system is not constrained and released later when it is. Source: Strbac et al (212) Report to Carbon Trust, Fig. 12 (part) 22

Marginal Value of Storage (Strbac et al., 212) /kw-year 5 4 3 2 Opex Savings Generation Transmission Distribution Interconnectors Capex Savings 1 5 1 15 GW -1 (Marginal values derived from and averaged over the capacity intervals in the previous slide)

The value of storage over time /kw-year 5, 4, 3, 2, 1, 23 25 215 5 1 15 2 25 Optimal storage capacity GW 24

What about the competitors? /kw-year 2,5 2, Distributed Storage Storage Interconnection Increasing asset utilisation and efficiency Flexible of operation Generation Flexible Demand 1,5 1, 5 Base Case 5 1 15 Optimal storage capacity GW 25

What about the competitors? /kw-year 2,5 2, Flexible Generation Distributed Storage Storage Interconnection Increasing asset utilisation and efficiency Flexible of operation Generation Flexible Demand 1,5 1, 5 Flexible Generation Base Case 5 1 15 Optimal storage capacity GW 26

What about the competitors? /kw-year 2,5 2, Flexible Generation Distributed Storage Storage Interconnection Increasing asset utilisation and efficiency Flexible of operation Generation Demand Flexible Response 1,5 1, 5 Flexible Demand Flexible Generation Base Case 5 1 15 Optimal storage capacity GW 27

What about the competitors? /kw-year 2,5 2, Flexible Generation Interconnection Distributed Storage Storage Interconnection Increasing asset utilisation and efficiency Flexible of operation Generation Demand Flexible Response 1,5 1, 5 Flexible Demand Flexible Generation Interconnection increases Base Case 5 1 15 Optimal storage capacity GW 28

Prosumage in the electricity market Richard Green Imperial College Business School Energy Transitions 216, Trondheim29

Making up words Consumer + Producer = Prosumer Consumer + Producer + Storage = Prosumage Imperial College Business School 3

Why be a prosumager? Consume more of your electricity Arbitrage between selling price to grid for surplus power and buying price for consumption Leave the grid entirely Imperial College Business School 31

Modelling the effects Take demand profiles from the DESSTINEE model for 25 Start from energy service demands, plus technology split, plus efficiency, to get annual TWh demands Assign to daily profiles for each electricity use Scenario replicating IEA 2 degrees scenario Many Electric Vehicles, little electric heating (in UK) Solar output profiles provided by Stefan Pfenniger of ETH Zurich, available on beta.renewables.ninja Imperial College Business School 32

Storage over the year kwh 2 PV capacity 1.1 kw 15 kw 2 kw 9 15 85 1 8 1-Jan 3-Jan 5-Jan 13 5 125 12 12-Jul 14-Jul 16-Jul 1-Jan 11-Jan 22-Jan 1-Feb 12-Feb 23-Feb 5-Mar 16-Mar 27-Mar 6-Apr 17-Apr 27-Apr 8-May 19-May 29-May 9-Jun 2-Jun 3-Jun 11-Jul 21-Jul 1-Aug 12-Aug 22-Aug 2-Sep 13-Sep 23-Sep 4-Oct 14-Oct 25-Oct 5-Nov 15-Nov 26-Nov 7-Dec 17-Dec 28-Dec Imperial College Business School 33

Storage over the year kwh 2 PV capacity 1.1 kw 15 kw 2 kw 15 1 5 1-Jan 11-Jan 22-Jan 1-Feb 12-Feb 23-Feb 5-Mar 16-Mar 27-Mar 6-Apr 17-Apr 27-Apr 8-May 19-May 29-May 9-Jun 2-Jun 3-Jun 11-Jul 21-Jul 1-Aug 12-Aug 22-Aug 2-Sep 13-Sep 23-Sep 4-Oct 14-Oct 25-Oct 5-Nov 15-Nov 26-Nov 7-Dec 17-Dec 28-Dec Imperial College Business School 34

Storage over the year kwh 2 PV capacity 1.1 kw 15 kw 2 kw 15 1 5 1-Jan 11-Jan 22-Jan 1-Feb 12-Feb 23-Feb 5-Mar 16-Mar 27-Mar 6-Apr 17-Apr 27-Apr 8-May 19-May 29-May 9-Jun 2-Jun 3-Jun 11-Jul 21-Jul 1-Aug 12-Aug 22-Aug 2-Sep 13-Sep 23-Sep 4-Oct 14-Oct 25-Oct 5-Nov 15-Nov 26-Nov 7-Dec 17-Dec 28-Dec Imperial College Business School 35

How much storage do you need? Days of Storage 9 8 7 6 5 4 3 2 1 UK 1. 1.25 1.5 1.75 2. 2.25 2.5 2.75 3. PV Output relative to Annual Load Imperial College Business School 36

How much storage do you need? Days of Storage 9 8 7 6 5 4 3 2 1 Germany UK 1. 1.25 1.5 1.75 2. 2.25 2.5 2.75 3. PV Output relative to Annual Load Imperial College Business School 37

How much storage do you need? Days of Storage 9 8 7 6 5 4 Germany 3 2 1 UK Spain 1. 1.25 1.5 1.75 2. 2.25 2.5 2.75 3. PV Output relative to Annual Load Imperial College Business School 38

System-wide effects Households with an EV, a 3 kw panel (2441 kwh p.a.) and 1 kwh of storage Use a simple algorithm to allocate solar output to the EV If I can half-charge my EV from my solar output, I spread that output in proportion to my hourly EV demand and buy the rest from the grid If I only need half my solar output to charge my EV, I export half of each hour s output What will this do to the market? Imperial College Business School 39

Prosumage and the Load-duration Curve GW 9 8 7 6 Smart EV charging: no prosumage 6 million EVs 5 4 3 2 1 1, 2, 3, 4, 5, 6, 7, 8, Hours per year Imperial College Business School 4

Prosumage and the Load-duration Curve GW 9 8 7 6 Smart EV charging: no prosumage 6 million EVs 5 4 3 2 1 1, 2, 3, 4, 5, 6, 7, 8, Hours per year Imperial College Business School 41

Making it happen Policy makers Recognise increased risk and complexity associated with innovation and the deployment of energy storage technologies Market, regulatory and commercial regime design Facilitate multi-service provision and market for flexibility Facilitate investment under uncertainty Storage scientists and technologists Halve the price and double the lifetime (please) Imperial College Business School 42