The power system load modeling based on Recurrent RBF Neural Network

The accuracy of the load model has great effects on power system stability analysis and control. In order to solve the problem of the difficulty of establishing accurate load model and the complexity of the modeling the non-linear properties of dynamic load, this paper proposes a methodology based on the RRBFNN (Recurrent RBF Neural Network) on modeling load from field measurements. New method is proposed to model power system load, which consists of recurrent network (RNN) and radial basic function (RBF) network and uses the ability of RNN for learning time series and the property of RBF with self-structuring and fast convergence. This new method which is tested by computer simulations on benchmark New England test system and applied in model identification of composite load for power system, has been proved its validity and accuracy.