Research on Dynamic Load Modeling Using Back Propagation Neural Network for Electric Power System

It is a well-known fact that load representation can have a significant impact on voltage stability. Accurate load models capturing load behaviors during dynamics are therefore necessary to allow more precise calculations of power system controls and stability limits. Recently artificial neural network (ANN) techniques have been widely used in power system simulation analysis. This paper deals with data recorded during the field experiments in power systems using a kind of multilayer feed forward (MLFP) networks with error back- propagation (BP) algorithm and a kind of aggregate load model with least square identification. The results show that the ANN model with the improved back-propagation learning rule have a satisfactory interpolation and extrapolation ability, and also have the ability to describe the voltage-power non-linear relationship of load dynamic characteristics.