Power system's security improvement through ANN-based decentralized secondary voltage control

In the paper, issues of electrical power system (EPS) security are investigated. The EPS equipped with a decentralized modular secondary voltage and reactive power control based on artificial neural networks (ANNs) is presented. Each generator is equipped with an automatic secondary voltage controller with ANNs that requires only local measurements for its operation. The ANNs were trained on optimal power flows (OPF). The EPS should remain secure during topology changes, such as generator's line outage, as well as during an outage of the voltage controller. The EPS states before and after the voltage control actions were compared to the OPF solutions and statistically evaluated.