Coordination of Heat and Electricity Systems Under Uncertainty

The transition from fossil fuel-based to weather-dependent renewable energy-based generation introduces challenges for the reliable operation of power system. The uncertainty and variability of renewable energy generation requires system operators to procure additional operational flexibility. At the same time, increased interactions between heat and electricity systems due to the increasing number of flexible units at the interface provide new opportunities for system operators to unlock flexibility from the heat system. In this context, the objective of this thesis is to improve the coordination of heat and electricity systems in the presence of uncertain renewable energy sources.

In the current operational framework, both systems operate sequentially and independently. This framework prevents harvesting the untapped flexibility from the operational synergies between heat and electricity. Coordination of heat and electricity systems aims to enhance the overall system’s operational efficiency by introducing new mechanisms and tools yet preserving the current operational framework. The improved coordination leads to an increase in the overall social welfare or utilization of the available renewable energy generation. This thesis introduces new concepts and mathematical models for heat and electricity system coordination to unlock the potential cross-sectoral flexibility by applying uncertainty-aware optimization.

Co-optimization model of heat and electricity systems is introduced and results in an extreme coordination approach that provides the most efficient operation of the systems. This approach is not practical for implementation but helpful in studying the effect of uncertainty propagation among energy systems. Combined heat and power units are located at the interface of the energy systems. In the model, they are used as a valve to control the uncertainty propagation from the electricity to the heat side by whether re dispatching heat production or not. Affine control policies and chance-constrained programming are proposed to model the uncertainty propagation and response of flexible units and the heating network to renewable energy uncertainty. The proposed approach illustrates how to access the district heating system flexibility by modeling and controlling the uncertainty propagation in the coupled system.

Moving from an ideal towards pragmatic coordination, soft-coordination approaches respect the current sequential operation of energy systems. In particular, market-based coordination is proposed based on self-scheduling combined heat and power units. This mechanism aims to improve the efficiency of the overall system utilizing current physical and economic interactions between heat and electricity systems. Combined heat and power units, possessing accurate information about heat and electricity market prices, increases the flow of information between heat and electricity markets. The improved coordination is quantified in terms of reduced cost of heat and electricity production and increased renewable production utilization.

From agents’ perspective, combined heat and power producers face the uncertainty of electricity prices defined by the market-clearing outcome. The uncertainty could prevent the producers from participating in the electricity market since potential losses could be incurred. The increasing amount of historical data of market prices makes it possible to employ efficient data-driven optimization approaches to model the market participation strategy. To address the market price uncertainty, data-driven uncertainty- and heat network-aware market participation strategy for combined heat and power producers is proposed. An accurate representation of the heat transmission dynamics is modelled to unlock the energy storage potential of the heating network. Convex relaxations and linearization of the heat system network model is introduced to achieve tractability in the optimization problem. The model reveals the benefit of accurate uncertainty representation and the value of the heating network flexibility in improved out-of-sample performance.