Trading of Pumped Hydro Storages in ID Markets

Transcription

1 Trading of Pumped Hydro Storages in ID Markets Continuous Optimization ID Price Forecasts Market Insights Strommarkttreffen Berlin,. Juni 28 Sebastian Braun, Senior Analyst Power Markets

2 Introduction Education Studied Industrial Engineering at KIT (Karlsruhe/Germany) Project Management at NTNU (Trondheim/Norway) Ph.D. Hydro Power Storage Optimization and Trading Considering Short-Term Electricity Markets at TUM (Munich/Germany) Work experience ICIS Senior Analyst Power Markets (Karlsruhe/Germany) EnBW Trading and Asset Optimization (Karlsruhe/Germany) EnBW Risk Management of RES (Istanbul/Turkey) DLR on Concentrated Solar Power (Almeria/Spain) Free Time Ski touring Running copyright 27 ICIS analytics.icis.com 2

3 Structure Presentation Strommarkttreffen Lünersee Lünersee machine Motivation to revise the pumped hydro power scheduling problem Challenges of continuously traded intraday markets Approach for pump storage optimization in intraday markets Continuous ID price forecast Trading strategies to exploit the spreads between ID high and low Latschau reservoir copyright 27 ICIS analytics.icis.com 3

4 Motivation for Flexibilization of Pumped Storages Comparison of the average hourly day-head auction price from Monday to Friday in 25 with 25 and the average quarterhourly intraday price from Monday to Friday in 25, data derived from EPEX Spot (27) Example: 5MW pumped hydro storage with efficiency of 8% and grid charges of 4 /MWh, 25 on average: spread: 32.2 EUR pumping: 9 hours generating: 7 hours 25 on average: spread: 2.34 EUR pumping: 6 hours generating: 5 hours 25 (qh) on average: spread: 2.34 EUR pumping: 9 hours generating: 7 hours copyright 27 ICIS analytics.icis.com 4 4

5 Intraday Trading Source: Dissertation Sebastian Braun copyright 27 ICIS analytics.icis.com 5

6 Intraday Screen Screenshot of an exemplary intraday orderbook from September 6 th, 23 at am including some explanations Screenshot of the price development for one product on the continuous intraday market from September 6 th, 23 at 7 am including some explanations copyright 27 ICIS analytics.icis.com 6

7 General Optimization Problem Wehr Oberbecken Minimize the costs for a specific period of time min c u,t (x u,t ) x u,t The load need to be covered at all time s.t. x u,t = l t u t Reservoir balancing equation Ramp-up and down gradient SQ s,t = SQ s,t + SQinflow s,t SV s,t x u + x u pump, t, s u turbines/ Min and max production power Max storage filling level Start and End filling levels x u,t x u,t gmax u,t u, t pmin u,t x u,t pmax u,t u, t SQ s,t SQmax s t, s SQ s.t= = SQstart s t, s SQ s,t=t = SQend s s Wehr Unterbecken copyright 27 ICIS analytics.icis.com 7

8 Challenges Optimization not quick enough for ID Shadow price steering difficult for daily pumped hydro storages Connection between dual variable water value shadow price not always clear How much water should be released at the level of the shadow price? How to consider order book data? copyright 27 ICIS analytics.icis.com 8

9 Reservoir filling level in MWh PFC in /MWh PFC in /MWh Intraday Algorithm for Pumped Hydro Power Storages λ g = λ p = t p t g quarter-hours unconstrained Time in quarter-hours Pumping Marginal Cost in /MWh Generating Marginal Cost in /MWh [-52] constrained [53-55] [56-26] copyright 27 ICIS analytics.icis.com 9

10 How to get a price forecast for the intraday market Which price fits for the optimization? VWAP (Volume Weighted Intraday Average Price) Last trade for each product Average of the first orders in the order book (two prices for bid and ask side?) Price forecast At that time we calculated the VWAP of the last trades for each our in the future. In case there were no deals we used the results of the Intraday Auction copyright 27 ICIS analytics.icis.com

11 ICIS Continuous Intraday Forecast Input Model Output wind forecasts PV forecasts demand forecasts available capacities market data (e.g. tick data, order books, bidding curves, balancing prices) statistical models fundamental models Forecast for the next 7 hours combination Updates once per hour, planned to be every few minutes forecast for hourly ID price range for ID forecast for quarter-hourly ID quarter-hourly price range for ID balancing price copyright 27 ICIS analytics.icis.com

12 Intraday Forecast copyright 27 ICIS analytics.icis.com 2

13 Intraday Forecast copyright 27 ICIS analytics.icis.com 3

14 price duration curves of hourly German ID continuous markt in /MWh Hourly ID Price Distribution high 25 high 26 high 27 high 28 high 24 low 25 low 26 low 27 low 28 low 24 price 25 price 26 price 27 price 28 price Average spread between ID high and low in /MWh copyright 27 ICIS analytics.icis.com 4

15 price duration curves of quarter-hourly German ID continuous markt in /MWh Quarter-Hourly ID Price Distribution high 25 high 26 high 27 high 28 high 24 low 25 low 26 low 27 low 28 low 24 price 25 price 26 price 27 price 28 price Average spread between ID high and low in /MWh copyright 27 ICIS analytics.icis.com 5

16 percentage per class percentage per class Development of ID Price Spreads spreads between hourly intraday high and low price [-5] [5-] [-5] [5-2] [2-25] [25-3] [3-35] [35-4] [4-45] [45-5] [5-55] [55-6] [6-4] spreads between quarter-hourly intraday high and low price [-5] [5-] [-5] [5-2] [2-25] [25-3] [3-35] [35-4] [4-45] [45-5] [5-55] [55-6] [6-4] copyright 27 ICIS analytics.icis.com 6

17 percentage per class Cumulated probability percentage per class Cumulated probability Development of ID Price Spreads spreads between hourly intraday high and low price [-5] [5-] [-5] [5-2] [2-25] [25-3] [3-35] [35-4] [4-45] [45-5] [5-55] [55-6] [6-4] cum. 24 cum. 25 cum. 26 cum. 27 cum. 28 spreads between quarter-hourly intraday high and low price [-5] [5-] [-5] [5-2] [2-25] [25-3] [3-35] [35-4] [4-45] [45-5] [5-55] [55-6] [6-4] cum. 24 cum. 25 cum. 26 cum. 27 cum. 28 copyright 27 ICIS analytics.icis.com 7

18 installed capacity in GW share of direct marketing in % Growth of RES Production and Trading installed capacity of renewable energy sources hydro power landfill gas, sewage and mine gas biomas geo thermal onshore wind energy offshore wind energy solar radiation energy direct marketing copyright 27 ICIS analytics.icis.com 8

19 Intraday Forecast copyright 27 ICIS analytics.icis.com 9

20 Renewable Production copyright 27 ICIS analytics.icis.com 2

21 Thank you for you attention! ICIS Tschach Solutions GmbH Stephanienstr ICIS Karlsruhe/Germany Tschach Solutions GmbH Sebastian Steinhäuserstrasse Braun 9 Senior 7635 Analyst Karlsruhe Power / Germany Markets Sebastian.braun@icis.com direct: +49 () ()