Periodic Reporting for period 3 - ICEBERG (Scalable Optimization of Power Systems with Flexible Demand and Renewable Supply)
Berichtszeitraum: 2023-09-01 bis 2025-02-28
The large-scale integration of renewable resources in electric power systems requires the mobilization of flexible consumers who can adapt their consumption to the variable and uncertain fluctuation of renewable supply. The mobilization of demand-side flexibility remains an elusive goal in electric power systems, while the majority of flexible consumers are connected to distribution systems that are currently operated passively. The major obstacles towards the optimal management of demand-side flexibility include the enormous number of flexible consumers (with ensuing challenges for scalable optimization), the presence of uncertainty at all layers of the power grid, and the physical complexity of distribution system power flow.
The first ingredient is a novel approach for planning and simulating the dispatch of the system which exploits the structure of distribution networks and can scale to systems of arbitrary size.
The second ingredient is an original optimization framework for tackling uncertainty and non-convexity at every layer of the system.
The third ingredient is a novel implementation of this optimization framework in parallel and distributed computing infrastructure, which will enable the optimal short-term planning and real-time coordination of resources at all layers of the system.
The vision of ICEBERG is to break down the current barriers to renewable energy integration by mobilizing the as yet untapped flexibility that is present at all layers of the network. This will enable the achieving of ambitious sustainability targets with acceptable infrastructure upgrades and without any deterioration in the quality of electric power service, which consumers currently enjoy.
The first ingredient is a novel approach for planning and simulating the dispatch of the system which exploits the structure of distribution networks and can scale to systems of arbitrary size.
The second ingredient is an original optimization framework for tackling uncertainty and non-convexity at every layer of the system.
The third ingredient is a novel implementation of this optimization framework in parallel and distributed computing infrastructure, which will enable the optimal short-term planning and real-time coordination of resources at all layers of the system.
The vision of ICEBERG is to break down the current barriers to renewable energy integration by mobilizing the as yet untapped flexibility that is present at all layers of the network. This will enable the achieving of ambitious sustainability targets with acceptable infrastructure upgrades and without any deterioration in the quality of electric power service, which consumers currently enjoy.
In terms of academic output, the project has produced 15 journal publications and 3 conference papers.
In terms of team composition, the project has funded the PI, 3 post-doctoral researchers (Daniel Avila, Dimitris Chatzigiannis, Efthymios Karangelos) and 6 PhD students (Nicolas Stevens, Jehum Cho, Jacques Cartuyvels, Daniel Avila, Ruan Zejun, Marilena Zambara).
Main results that the team boasts include:
• The development of a hierarchical TSO-DSO market design and market clearing platform
• The development of methods for pricing service contracts that can be used for mobilizing residential consumers
• The development of methods for hierarchically coupling separate control areas of high-voltage transmission grids
• The development of novel parallelization methods for implementing stochastic dual dynamic programming, and their deployment on high performance computing infrastructure, which has led to outperforming commercial grade SDDP software
• The development of parallel computing methods for supporting the optimal expansion of generation capacity in electric power systems while endogenously accounting for uncertainty
• The development of coherent market design proposals for balancing markets and imbalance settlement in EU real-time markets
• The development of mixed integer programming methods for solving stochastic programs that enable the optimal sizing of reserves on networks
• The development of scalable pricing algorithms in markets with non-convex operating constraints
• The analysis of the implications of zonal pricing for long-term investment incentives
• The analysis of the need / benefit of capacity remuneration mechanisms in markets with binary investment decisions
• The development of coherent market design proposals for European real-time markets with multiple reserve products
In terms of team composition, the project has funded the PI, 3 post-doctoral researchers (Daniel Avila, Dimitris Chatzigiannis, Efthymios Karangelos) and 6 PhD students (Nicolas Stevens, Jehum Cho, Jacques Cartuyvels, Daniel Avila, Ruan Zejun, Marilena Zambara).
Main results that the team boasts include:
• The development of a hierarchical TSO-DSO market design and market clearing platform
• The development of methods for pricing service contracts that can be used for mobilizing residential consumers
• The development of methods for hierarchically coupling separate control areas of high-voltage transmission grids
• The development of novel parallelization methods for implementing stochastic dual dynamic programming, and their deployment on high performance computing infrastructure, which has led to outperforming commercial grade SDDP software
• The development of parallel computing methods for supporting the optimal expansion of generation capacity in electric power systems while endogenously accounting for uncertainty
• The development of coherent market design proposals for balancing markets and imbalance settlement in EU real-time markets
• The development of mixed integer programming methods for solving stochastic programs that enable the optimal sizing of reserves on networks
• The development of scalable pricing algorithms in markets with non-convex operating constraints
• The analysis of the implications of zonal pricing for long-term investment incentives
• The analysis of the need / benefit of capacity remuneration mechanisms in markets with binary investment decisions
• The development of coherent market design proposals for European real-time markets with multiple reserve products
Our goals for the project moving forward include:
• The development of large-scale TSO-DSO coordination test cases and the demonstration of the effectiveness of our developed TSO-DSO coordination platform on these test cases
• A review of the state of play of European TSO-DSO coordination platforms, and an alignment of our proposed hierarchical solution with the institutional constraints of the European market
• The application of SDDP and other cutting plane methods, deployed on high performance computing infrastructure, for supporting the long-term optimal expansion of the European power system under multi-stage uncertainty
• The application of scarcity pricing and reserve deliverability on TSO-DSO coordination
• The analysis of linearizations of optimal power flow on large-scale TSO-DSO coordination models
• The development of long-term capacity expansion models under uncertainty with endogenous expected energy not served constraints
• The analysis of the welfare benefits of co-optimizing energy and reserves on the European electricity market
• The development of coherent market design proposals for imbalance settlement in European balancing markets with 15-minute market time units
• The development of large-scale TSO-DSO coordination test cases and the demonstration of the effectiveness of our developed TSO-DSO coordination platform on these test cases
• A review of the state of play of European TSO-DSO coordination platforms, and an alignment of our proposed hierarchical solution with the institutional constraints of the European market
• The application of SDDP and other cutting plane methods, deployed on high performance computing infrastructure, for supporting the long-term optimal expansion of the European power system under multi-stage uncertainty
• The application of scarcity pricing and reserve deliverability on TSO-DSO coordination
• The analysis of linearizations of optimal power flow on large-scale TSO-DSO coordination models
• The development of long-term capacity expansion models under uncertainty with endogenous expected energy not served constraints
• The analysis of the welfare benefits of co-optimizing energy and reserves on the European electricity market
• The development of coherent market design proposals for imbalance settlement in European balancing markets with 15-minute market time units