부산대학교 도서관

The increasing penetration of distributed generation on power grids will increase the number of voltage violations because of voltage rise phenomenon. Thus, there is a need for efficient utilization of reactive power resources available on the grid including using local voltage control of DGs for voltage regulation. One of such is Q(U) control. However, the question is how to tune its control parameters to guarantee good performances as the grid condition keeps changing. This paper proposes a two stage offline methodology based on genetic algorithm (GA) optimization and artificial neural network (ANN) to adapt the control parameters. The objective of the ANN is to develop a fitting function correlating the Thevenin impedance seen at the point of common coupling and optimal control parameters obtained from the GA optimization. This scheme is implemented on 16 United Kingdom generic test distribution network. Results demonstrated the effectiveness of this method in regulating the voltage as compared to the base where the control parameters of the Q(U) control are fixed.