Scheduling of AA-CAES Based Energy Hub with Wind Power Generation: A Multi-stage Robust Optimization Approach

The development of renewable is becoming a new trend to cope with the shortage of fossil fuels and environmental pollution, but the inherent uncertainty and volatility pose great challenges to power system operation with high share of renewables. Energy hub which incorporates multiple energy production, conversion, and storage facilities can play a noteworthy role in renewable accommodation and energy saving. This paper proposes a multi-stage robust optimization model for day-ahead scheduling of energy hub composed of wind generation, advanced adiabatic compressed air energy storage (AA-CAES) system, and heat pump in the presence of wind uncertainty. To overcome the computational difficulty, the multistage model is decomposed into two stages. The first stage minimizes the operation cost under the predicted scenario, and the second stage determines real-time dispatch in response to the observed wind power generation. Dynamic lower and upper bound variables for two correlated state-of-charges (SoC) are added as the first-stage decision variables with ramping limit constraints, so that the second-stage decision can be made period-by-period without jeopardizing robust feasibility. This is an important difference compared to the two-stage robust optimization whose second-stage decision is anticipative and thus non-causal. Finally, the day-ahead scheduling problem is reformulated as a mixed-integer linear program (MILP) guaranteeing the robustness and non-anticipativity of real-time dispatch. Case studies verify the effectiveness of the proposed method.