Simplified Deep Reinforcement Learning Based Volt-var Control of Topologically Variable Power System

The high penetration and uncertainty of distributed energies force the upgrade volt-var control (VVC) to smooth voltage var fluctuations faster. Traditional mathematical or heuristic algorithms are increasingly incompetent for this task because slow online calculation speed. Deep reinforcement learning (DRL) has recently been recognized as an effective alternative it transfers computational pressure off-line training timescale reaches milliseconds. However, its offline speed still limits application VVC. To overcome issue, paper proposes a simplified DRL method that simplifies improves operations in DRL, avoiding invalid explorations reward Given problem network parameters original topology not applicable other new topologies, side-tuning transfer (TL) is introduced reduce number needed be updated TL process. Test results based on IEEE 30-bus 118-bus systems prove correctness rapidity proposed method, well their strong applicability large-scale variables.