Approximating multi-purpose AC Optimal Power Flow with reinforcement trained Artificial Neural Network

Solving AC-Optimal Power Flow (OPF) problems is an essential task for grid operators to keep the power system safe use cases such as minimization of total generation cost or infeed curtailment from renewable DERs (Distributed Energy Resource). Mathematical solvers are often able solve AC-OPF problem but need significant computation time. Artificial neural networks (ANN) have a good application in function approximation with outstanding computational performance. In this paper, we employ ANN approximate solution multiple purposes. The novelty our work new training method based on reinforcement learning concept. A high-performance batched flow solver used physical environment training, which evaluates augmented loss and numerical action gradient. consists objective term each case penalty constraints violation. This enables without reference OPF integration discrete decision variable transformer tap changer position constrained optimization. To improve optimality approximation, further combine approach supervised labeled by OPF. Various benchmark results show high quality proposed while achieving efficiency cases.