Frequency Control And Economic Dispatch For Integrated Energy Systems Based On Deep Reinforcement Learning

Maintaining the active power balance between electricity generation and consumption is a basic task in the operation of power systems.Frequency control during active power imbalance and economic dispatch during active power balance are two important measures to achieve stable and economical operation of power systems.With the large-scale integration of distributed renewable energy and the close coupling of different forms of energy networks,integrated energy systems have been formed.Traditional frequency control and economic dispatch strategies have certain limitations in integrated energy systems.Therefore,this paper conducts research on the frequency control and economic dispatch issues of integrated energy systems,and uses multi-agent deep reinforcement learning algorithms to design optimal frequency coordinated control and optimal decentralized dispatch strategies for multi-area integrated energy systems.This will solve the problems of frequency instability caused by fluctuations in new energy output and load,the coupling relationship between multiple forms of energy,as well as the difficulty in solving economic dispatch strategies due to diverse energy transformations.The main innovative work of this paper is as follows:(1)A multi-region interconnected comprehensive energy system integrated frequency control strategy based on a multi-agent proximal policy optimization(MAPPO)algorithm is proposed to address the problem of AGC coordination caused by inconsistent control objectives between load frequency control(LFC)and generation command dispatch(GCD).It reduces system frequency deviation and frequency regulation generation cost,thus achieving frequency stability control of the multi-region integrated energy system.The dynamic optimization performance and coordinated control performance of the proposed method are verified through case simulation under large-scale random disturbances.(2)The economic dispatch problem of integrated energy system with active power balance and frequency stability is addressed by proposing a multi-regional integrated energy system economic dispatch strategy based on MAPPO,which uses the self-learning characteristics of MAPPO for distributed optimization of multi-region comprehensive energy system dispatch strategies.A multiregion comprehensive energy system optimization dispatch model considering energy trading and carbon trading is established for multi-region integrated energy systems containing electric-thermal systems under source-load uncertainty and multi-energy coupling.The MAPPO algorithm is used as the solution method,and the optimal dispersed dispatch strategy for the multi-region integrated energy system is obtained through training with a large amount of historical data.Finally,the proposed method was verified by simulation,which shows that it can adaptively control the energy output of each region under the influence of random fluctuations of distributed power sources and loads,achieving optimal decentralized scheduling for multi-area integrated energy systems.