电力市场非凸成本补偿机制对发电商竞价策略的影响分析OA

A rational pricing mechanism is the basis for the effective operation of electricity markets,with the participation rationality being a basic requirement for mechanism design.For the three-part bidding mechanism designed to facilitate the participation of resources with start-up and no-load costs,as well as intertemporal constraints such as coal-fired and gas-fired power plants in the market,the locational marginal pricing approach based directly on Lagrange multipliers is no longer applicable.Internationally,this issue is typically addressed through full-cost compensation mechanisms.However,in China,considering the coupling relationship between spot markets and medium-and long-term markets,partial compensation mechanisms are mostly adopted.To investigate the impact of partial compensation mechanisms on the strategic behavior of market participants,a multi-agent simulation framework based on deep reinforcement learning is constructed to precisely characterize the three-part bidding strategies of generating units,thereby enabling the quantitative revelation of the strategic behavior characteristics of market participants and their risk transmission pathways,as well as validating the effectiveness of market mechanism design.Based on this framework,simulation analyses are conducted for typical compensation mechanisms,including no compensation,start-up cost compensation only,time-of-use compensation,and full-cost compensation.The results indicate that under locational marginal pricing mechanisms,inadequate compensation tends to induce the strategic price increase among generating units to avoid losses,thereby increasing the market electricity prices and reducing social welfare,further affecting the power balance of the power system.In contrast,a full-cost compensation mechanism helps safeguard the revenues of power producers,lower electricity prices,and incentivize the provision of regulation capabilities.