Study Of Short-term Power Load Prediction Based On Optimal PSO-Elman Neuralnetwork Model

With the social and economic development today, China’s electric power sector isexpanding its scale rapidly. The specialty that the electric energy cannot be stored requires thepower-generating sector to conduct investigation in the power market and predict the preciseload of the power system so as to assure the balance of the generation and assumption of electricpower. The studies on advanced prediction method is a hot issue nowadays about the supply anddemand of the electric power and the better working of the society.The paper covers the following four parts:1.The importance of the load prediction in power system is introduced;the analysis andcomparison of long-term,medium-term and short-term load prediction methods with thehighlights on the importance and prospect of short-term load prediction method.2.Two artificial neural network models(ANNS) is introduced: feedforward neural network(BP, back propagation) and feedback neural network(Elman) and algorithm implementationrespectively.3.The basic PSO model and the process of algorithm implementation is introduced,point outthe defect and put forward the optimal methods.4.The optimal methods to the implementation of Elman neural network to predict theshort-term power load is applied and compare the results with the predictions prior bysimulation.