Smart Grid Online Learning and Autonomous Optimal Control Demonstration System Based on Deep Reinforcement Learning
In order to promote the experience and achievements of smart grid construction under the background of artificial in-telligence,a smart grid online learning and autonomous optimal control demonstration system based on deep reinforcement learning is proposed.The experience pool of power grid operation situation in highly dynamic and complex environment is formed by integrating experience and practice.Through continuous coupling and interaction with highly dynamic and complex environment,online learning of power grid operation situation under positive time sequence is realized.The optimal coordinated control of power grid operation in highly dynamic and complex environment is realized.The application practice verification analysis is carried out for the demonstration system.The results show that the accuracy of power grid operation situation pre-diction in a long period is 93.57％,and the effectiveness of optimal coordinated control of power grid operation in a dynamic en-vironment is 92.81％.It has the online learning function of power grid operation situation,and can realize the optimal control of power grid operation scheme in highly dynamic and complex application scenarios.
deep reinforcement learningsmart gridsituational online learningautonomous optimization controlvisual dem-onstration system
万方数据
维普
