Adaptive Dynamic Programming for Gas-Power Network Constrained Unit Commitment to Accommodate Renewable Energy With Combined-Cycle Units

Abstract

Due to the fast response capabilities, combined-cycle units could provide valuable flexibilities to cope with uncertainties from the renewable energy resources. However, combined-cycle units intensify the coupling between power and gas networks, so operational constraints of the two networks should be explicitly considered simultaneously. Indeed, neglecting the impact of any network may lead to infeasible solutions that will violate physical operational constraints. Compared with the previous stochastic unit commitment model with only combined-cycle units, this paper establishes a stochastic gas-power network constrained unit commitment model considering both combined-cycle units and gas network. Furthermore, an adaptive dynamic programming (ADP) to prevent the curse of dimensionality is proposed. Numerical results verify the effectiveness of the proposed model.