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.
关键词
深度强化学习/智能电网/态势在线学习/自主优化控制/可视演示系统
Key words
deep reinforcement learning/smart grid/situational online learning/autonomous optimization control/visual dem-onstration system
维普
万方数据