HYDROPOWER OPERATION IN A CHANGING ENVIRONMENT

Transcription

1 HYDROPOWER OPERATION IN A CHANGING ENVIRONMENT Moritz Schillinger 1, Hannes Weigt 1, Michael Barry 2, René Schumann 2 1 University of Basel, 2 HES-SO Valais IEWT Session 1C: Erneuerbare Energien I Wednesday 15th February

2 /MWh Introduction European electricity markets in transition: - Nuclear phase-out - Increasing share of RES Scenario Swiss energy transition 2050: Challenge: - Falling prices challenge Swiss HP owners Subsidies discussed in Swiss politics Average Spot Price Source: Prognos

3 Introduction Market Opportunities Source: Barry et al

4 Introduction Market Opportunities and Challenges: Revenue Prospects in a Low Price Market - How to optimize revenue across markets (spot and balancing markets)?» Historic: What could have been?» Future: Outlook into the next decade - 4 -

5 Hydropower Operation Model Model charateristics - Single plant perspective or Cascaded Structures - Revenue maximization (spot and balancing markets) - Short-term (timeframe = 1 year), 15 minute steps - Deterministic, combining LP and NLP - GAMS - Modular design: - 5 -

6 Data Three generic HP plants Big Medium Small Capacity (MW) Maximum Head (m) Ratio inflow to storage capacity Ratio storage to discharge capacity (h) Full load hours (h)

7 Mio EUR per year Historic Results Spot and balancing markets 25 Energy Prim Sec Ter+ Ter spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all spot all Big Medium Small - Decreasing spot prices = decreasing spot revenues - Significant additional revenues from balancing - Larger storage important - Theoretical maximum upper bound - 7 -

8 0h/0h 6h/ 0.25h 12h/0.25h 24h/0.5h 36h/0.75h 48h/1.25h 60h/1.5h 168h/4h 0h/0h 6h/ 0.25h 12h/0.25h 24h/0.5h 36h/0.75h 48h/1.25h 60h/1.5h 168h/4h 0h/0h 6h/ 0.25h 12h/0.25h 24h/0.5h 36h/0.75h 48h/1.25h 60h/1.5h 168h/4h Revenue relative to 0h/0h Historic Results Spot and balancing markets - Security constraint: - Uncertainty in call-up of balancing energy reserve enough water in the storage for specific timeframe Energy Prim Sec Ter+ Ter- 100% 90% 80% 70% 60% 50% 40% 30% 20% Big Medium Small - Small effect on big plants due to high storage capacity - Plants with small storage capacity suffer from uncertainty in call-up - 8 -

9 Revenue relative to theoretical maximum Historic Results Spot and balancing markets - Maximum accepted bids: - Markets are small and could be satisfied by a small number of HP plants - Reduce size of maximum accepted bid Energy Prim Sec Ter+ Ter- 100% 90% 80% 70% 60% 50% 40% 30% 20% max avg. 10% 5% 2.5% max avg. 10% 5% 2.5% max avg. 10% 5% 2.5% Big Medium Small - Significant decrease in balancing revenues for big plants - Smaller effect on smaller plants due to low bids - 9 -

10 Historic Results Spot and balancing markets - which allows to benefit from weeks with high balancing prices without changing the weekly generation quantity: TRL+: Bid quantitiy neutral in balancing market at opportunity costs (+ margin) based on spot generation and call-up If opportunity costs (+ margin) lower price, bid is accepted SRL/ TRL-: Bid weekly spot generation in balancing market at opportunity costs (+ margin) If opportunity costs (+ margin) lower price, bid is accepted

11 Revenue relative to spot only Historic Results Spot and balancing markets - : Energy Prim Sec Ter+ Ter- 115% 110% 105% 100% 95% 90% Spot only SRL TRL- TRL+ Spot only SRL TRL- TRL+ Spot only SRL TRL- TRL Big Medium Small - 1-2% higher revenue due to TRL+ for big plants - Otherwise no additional revenues in 2015 due to heuristic - does not improve revenues in normal year

12 Revenue relative to spot only Historic Results Spot and balancing markets - : Energy Prim Sec Ter+ Ter- 115% 110% 105% 100% 95% 90% Spot only SRL TRL- TRL+ Spot only SRL TRL- TRL+ Spot only SRL TRL- TRL Big Medium Small % higher revenues due to SRL market (2-3 high price weeks) - 3% higher revenue due to TRL- for small plants - No additional revenues from TRL+ in = lower bound

13 Overview Future Simulation EU Investment model future generation capacity Swissmod (Swiss electricity market model) future electricity and balancing prices - Scenarios: - EU Reference Scenario: as is EU Energy Trends - Base Price 2015 Scenario: what if prices prevail - C+: slow increase of carbon price (35 /t in 2030) - C++: fast increase of carbon price (50 /t in 2030) - F+: slow increase of fuel prices (+50% until 2030) - F++: fast increase of fuel prices (+100% until 2030) - R+: stronger increase in wind and solar (+10% relative to EU Trend) - R-: weaker increase in wind and solar (-10% relative to EU Trend) - Combinations: all combinations of scenarios

14 Revenue relative to 2015 Future results Spot EUTrend BasePrice2015 C+ C++ F+ F++ R+ R- C+F+ C+F++ C+R+ C+R- C++F+ C++F++ C++R+ C++R- F+R+ F+R- F++R+ F++R- C+F+R+ C+F+R- C+F++R+ C+F++R- C++F+R+ C++F+R- C++F++R+ C++F++R- 160% 140% 120% 100% Spot and balancing 80% 60% Spot 40% Decrease in total revenue until 2020, afterwards scenario dependent Highest revenue path = fast increase in fuel/ carbon prices + weak RES growth Lowest revenue path = strong RES growth while fuel/ carbon prices on current level B i g

15 Revenue relative to 2015 Future results Spot EUTrend BasePrice2015 C+ C++ F+ F++ R+ R- C+F+ C+F++ C+R+ C+R- C++F+ C++F++ C++R+ C++R- F+R+ F+R- F++R+ F++R- C+F+R+ C+F+R- C+F++R+ C+F++R- C++F+R+ C++F+R- C++F++R+ C++F++R- 160% 140% 120% 100% Spot and balancing 80% 60% Spot 40% Variance in revenues further increases for small HP plants Higher carbon prices more important for smaller units (large units are peak price driven) Reservoir size influence uncertainty in future revenues S m a l l

16 EUR/MWh Mio EUR per year Future results Revenue by market (average across scenarios): Energy Prim Sec Ter+ Ter Big Medium Small Revenues close to 2015 level in 2025 and % higher revenues due to balancing Rush to balancing markets would reduce current potential (more important for large units) av price summer winter

17 Conclusion Historic: - Decreasing HP profitability in the last years due to decreasing spot prices - In theory, balancing markets can significantly increase revenues - In practice, potential of balancing markets reduced due to uncertainties/ market characteristics but cross-market optimization still profitable Future: - Market price prospects for coming decade low to modest, existing EU capacity structure likely to remain stable Global fuel markets, ETS and RES as decision makers for HP - Balancing market benefits significantly reduced if full Swiss HP aims for balancing, more important for larger units Optimized operation helps, but is limited in scale

18 Thank you for your attention! Questions? Contact: Moritz Schillinger Forschungsstelle Nachhaltige Energie- und Wasserversorgung Universität Basel Wirtschaftswissenschaftliche Fakultät Peter Merian-Weg 6 moritz.schillinger@unibas.ch

19 Backup

20 Market Characteristics Source: Abrell

21 Hydropower Operation Model Objective max Revenue = Revenue DayAhead + Revenue PRL + Revenue SRL + Revenue TRL + +Revenue TRL Revenue SRL = cap p t,srl Cap t,i,srl + energy+ p t,srl + G t,i,srl energy p t,srl G t,i,srl t,i t,i t,i Storage S r,t+1 = S r,t + i r,t+1 net R i,t+1 map r,i Spill r,t+1 + Spill rr,t+1 map r,rr r, t i Capacity constraints Cap t,i,dayahead + Cap t,i,prl + Cap t,i,srl + Cap t,i,trl + cap i max Cap t,i,dayahead Cap t,i,prl Cap t,i,srl Cap t,i,trl cap i min i, t i, t

22 Hydropower Operation Model Head and head loss H net t,i = H gross loss t,i H t,i i, t H gross t,i = h constant i + H loss net t,i = a i R 2 i,t i, t r D t,r map r,i i, t Security constraint S r,t s r min + t security i water (Cap water t,i,prl +Cap water t,i,srl + Cap t,i,trl + )map r,i r, t

23 Hydropower Operation Model Solution algorithm Yearly LP (with constant head) Seasonality storage Weekly storage values from LP Weekly NLP (with head effects)

24 Investment Model and Swissmod EU Investment Model - 20 countries: - Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, UK - 5 year steps up to Output = generation capacity by technology, year and country Swissmod - Swiss electricity market model by Schlecht and Weigt (2014, 2015)/ Savelsberg and Mendes (2017) - Classical dispatch model based on DC-Load-Flow Approach - Output = Electricity and balancing prices

25 Revenue relative to 2015 Future results Spot EUTrend BasePrice2015 C+ C++ F+ F++ R+ R- C+F+ C+F++ C+R+ C+R- C++F+ C++F++ C++R+ C++R- F+R+ F+R- F++R+ F++R- C+F+R+ C+F+R- C+F++R+ C+F++R- C++F+R+ C++F+R- C++F++R+ C++F++R- 160% 140% 120% 100% Spot and balancing 80% 60% Spot 40% M e d i u m Similar picture for medium plants but variance in revenues increases compared to large plants