Energy Analytics & Markets at the Technical University of Denmark

Transcription

1 Energy Analytics & Markets at the Technical University of Denmark Pierre Pinson et al. Technical University of Denmark DTU Electrical Engineering - Centre for Electric Power and Energy mail: ppin@dtu.dk - webpage: (with ackn. to all funding sources, data providers, collaborators and students, for material and ideas) 1 / 23

2 Outline Who are we?... and what we do Some interesting problems towards an healthy integration of renewables, e.g. Renewable energy analytics Stochastic market clearing and probabilistic offering Towards fully controllable power systems (and globally!) Demand response & energy system integration Give us more data! Outlook... you may find a number of short videos (5-7mins) describing our works on the DTU CEE Youtube Channel 2 / 23

3 1 Who are we?... and what we do 3 / 23

4 Center for Electric Power and Energy (CEE, at DTU) Established 15 August 2012 as a merger of existing units (Lyngby + Risø) One of the strongest university centers in Europe with approx. 100 employees A single, clear interface for our external collaboration partners Provides cutting-edge research, education and innovation in the field of electric power and energy to meet the future needs of society regarding a reliable, cost efficient and environmentally friendly energy system Bachelor and Master programs: Electrical Engineering, Wind Energy, Sustainable Energy Main competences: Components Power systems Distributed energy resources Market modeling, forecasting and optimization Center agreement (i.e., direct support): Energinet.dk, Siemens, DONG Energy, Danfoss DTU ranked world 2nd-7th in Energy Science and Engineering(!) 4 / 23

5 Who are we? - The Energy Analytics & Markets group One of the 5 groups of the Center for Electric Power and Energy, Dpt. of Electrical Engineering Resources: ( 10 nationalities) Faculty: 1 Prof, 2 Assist. Profs. Junior: 3 post-doc fellows, 9 Ph.D. students (+3-4 not at DTU), xx research assistants + student helpers, and Ph.D. guests from, e.g., China, Brazil, US, Spain, France, Italy, The Netherlands, Germany, etc. Projects (examples, active in 2017): EU: BestPaths Danish: MULTI-DC, 5s, EcoGrid 2.0, The Energy Collective, CITIES, EnergyLab Nordhavn, CORE Danish-Chinese: PROAIN Education: Various courses on analytics, optimization, forecasting, game theory, renewables and electricity markets (hopefully) recognized leading expertise in energy analytics and markets 5 / 23

6 What we really do... Forecasting Clustering & Profiling Open-source (repository) Big data Data-driven analytics Advanced Computational Techniques Energy Analytics & Markets HVDC Design Power System Stability Energy markets Market Integration System models & Optimization Uncertainty, variability & flexibility Large-scale optimization Modelling & Simulation Offering strategy & trading Equilibrium models Stochastic optimization 6 / 23

7 Book on renewables in electricity markets Available from Dec (sorry, not open access!) > e-copies to date...! Topics: Renewable energy modeling and forecasting Day-ahead market-clearing with significant share of renewables Balancing markets Managing uncertainty (rewarding flexibility) Impact of renewables on market quantities Trading stochastic power generation Virtual power plants (meaningful association and operation) Demand-side aspects Jalal also co-authored a book released more recently 7 / 23

8 2 Some interesting problems towards an healthy integration of renewables 8 / 23

9 Renewable energy analytics as a breakthrough The MIT Technology Review: founded at MIT in 1899 daily review/analysis of technological innovation worldwide impact: members and website visitors per month! The 10 breakthrough technologies 2014: genome editing microscale 3D printing neuromorphic chips brain mapping etc. renewable energy analytics, especially short-term forecasting (0-72h) and its dynamic uncertainty [See link: MIT Technology Review - Smart Wind and Solar Power] 9 / 23

10 Revisiting market-clearing procedures Day-ahead market-clearing insures an optimal match of production and demand, a fair amount of time prior to operations, regardless of the nature of offers/bids With increased variability and uncertainty to be dealt with, the system should be placed in a state permitting to optimally cope with whatever could realistically happen The alternative schools of thoughts: Conventional sequential market-clearing(s), where day-ahead aspects and balancing are decoupled, but possibly using smart(er) reserve products Stochastic optimization, accounting for expected costs of balancing The more practical, and still efficient, solutions: Conventional market-clearing with improved dispatch of stochastic production Robust optimization based dispatch... virtual bidding? This is what we are here to talk about! 10 / 23

11 Conceptually, 3 alternative proposals Placing ourselves in an energy-only context (i.e., reserves are disregarded) The alternatives to market clearing we use in the following include In short Link day-ahead/balancing Forecast input Optimization problem ConvD sequential deterministic 2 (MI)LPs StochD integrated probabilistic stochastic program ImpD integrated probabilistic bilevel program ConvD: StochD/ImpD: 11 / 23

12 Towards fully controllable power systems (globally!)... and Spyros is a strong supporter of the Global Grid (see his paper in Renewable Energy) 12 / 23

13 Demand response in EcoGrid EU Key aspects: Nearly 1900 households/commercial/industry installations participating Design and evaluation of a complete market concept Commitment from academia and relevant industry partners, e.g., Energinet, Østkraft, IBM, Siemens, etc. A lot of sweat and stress / 23

14 Demand-response: the elastic demand background The market itself: An additional market aiming to mobilize demand-side flexibility (near) real-time Demand is then enabled as a service provider, e.g., for balancing or congestion management Demand is part of a pool among other potential service providers Important considerations on the demand side: The approach should scale nicely, and directly reward demand for its flexibility However, such an approach relies on the conditional dynamic elasticity of electricity consumers Forecasting it is key, and even then the actual response may still be uncertain 14 / 23

15 EcoGrid EU: The resulting market concept A few personal thoughts: Some might say that this is a twisted market, since demand never offered its flexibility (we do it now!) The concept is packed with analytics concepts - it could be very sensitive to the quality of models employed Forecasting conditional dynamic elasticity is a difficult but fun problem 15 / 23

16 Towards the future in København... EnergyLab Norhavn... And also EcoGrid 2.0! 16 / 23

17 Integrated energy markets: complete coupling For a mathematical point of view, we can write and solve fully integrated markets for el-gas, el-heat, el-gas-heat... but / 23

18 Integrated energy markets: loose coupling What do we mean by loose coupling? 18 / 23

19 Integrated energy markets: loose coupling What do we mean by loose coupling? respecting organizational aspects of the energy system, e.g., heat and el management are separated, the system operator is not taking care of day-ahead electricity market clearing, etc. profit of existing levies for impacting dispatch, costs, etc. A practical example: heat and el interaction through Varmelast 19 / 23

20 Give us more data! Let us build a (very) large-scale dataset for the whole European system Coal Fuel Oil Hydro Lignite Natural Gas Nuclear Unknown Coal Fuel Oil Hydro Lignite Natural Gas Nuclear Unknown Fully open access (CC-BY, on zenodo.org), companion paper soon in Nature, Scientific Data 20 / 23

21 The big picture... The grand forecasting challenge : predict renewable power generation, dynamic uncertainties and space-time dependencies at once for the whole Europe...! Linkage with future electricity markets: Monitoring and forecasting of the complete Energy Weather over Europe Provides all necessary information for coupling of various existing markets (e.g., day-ahead, balancing), and deciding upon optimal cross-border exchanges 21 / 23

22 Outlook There is still so much to be done in this field! E.g., understanding the impact of renewables on power system operations, electricity markets, players, investment, and overall social welfare propose and deploy new concepts and computational techniques to adapt to this new reality, with variability, uncertainty, high-dimensional, distribut, etc. simply make it happen, by bridging the gap between research ideas and real-world deployment... etc. It is difficult to tell you about all we do in a few slides only. Our group at DTU (and affiliates) aims at producing scientifically excellent and high-impact contributions having a strong link to real-world problems keeping it fun for us and our students Do not hesitate to pay us a visit! 22 / 23

23 Thanks for your attention! 23 / 23