Adaptive Online-Learning Volt-Var Control for Smart Inverters Using Deep Reinforcement Learning

The increasing penetration of the power grid with renewable distributed generation causes significant voltage fluctuations. Providing reactive helps balancing in grid. This paper proposes a novel adaptive volt-var control algorithm on basis deep reinforcement learning. learning agent is an online-learning deterministic policy gradient that applicable under real-time conditions smart inverters for management. only uses input data from connection point inverter itself; thus, no additional communication devices are needed and it can be applied individually to any proposed successfully simulated at various points 21-bus low-voltage distribution test feeder. resulting behavior analyzed systematic reduction observed both static environment dynamic environment. enables flexible adaption changing environments through continuous exploration during process and, contributes decentralized, automated future grids.