Power system load modeling by learning based on system measurements

This paper is concerned with an investigation of a methodology using intelligent learning techniques based on system measurements to construct power system load models alongside with distribution network reduction. A comprehensive load model is proposed to represent the loads in an area of a power system. A population diversity-based genetic algorithm (GA) is developed to obtain the structure and parameters of the load model. Simulation results on a five-bus power system and an IEEE 30-bus power system are given to show the potential of this new methodology of power system modeling.