A secondary voltage control based on artificial neural networks

The use of intelligent voltage controllers makes the operation of the electrical system simpler for system operators. This work aims to propose a methodology to find voltage control actions that simultaneously keep the voltage levels within the operating limits and manage the participating generators’ reactive power reserves. For this purpose, an intelligent secondary voltage control, called SVC-neural, is proposed. The proposed methodology combines optimal power flow solutions with artificial neural networks. The latter is responsible for gathering the knowledge and characteristics of the system and proposing control solutions quickly and adaptively. The application of the proposed SVC-neural resulted in reduced total system losses and adequate control of voltage and reactive power. The most significant loss reduction using the controller was 2.24%, and the smallest reduction was 0.5%. SVC-neural results show the ability to generalise and not need to solve the optimisation problem, which can often become a complex task and require a great deal of computational effort. The results also validate the proposed SVC-neural concerning its ability to reduce losses and motivate future research on intelligent controllers.