Short Term Load Forecasting Of Power System Based On GA-BP Neural Network

Short-term load forecasting of power system is a major foundation for power attemper department to set down generating electricity plan and arrange attemper plan, power supply plan and bargaining plan under market environment. With the rapid development of the power industry market process, the precision of forecasting of Short-term load forecasting of power system has a direct effect on economic benefit of industry department. Short-term load forecasting, a major foundation for the research of power system planning and power system economic operation and automatic dispatch, is an important task of the modern power system operation research.This paper is concerned with a series of problems in the application of forward-back neural network(BP network) to the short-term load forecasting of power system, including optimizing the initial network parameters using genetic algorithm(or GA). The main achievements are as follows:1. The load of power system is an unsteady stochastic process, whose observed value may exist some "unhealthy data" due to the effect of various factors. These unhealthy data, participating the training of neural network intermingled with normal data, badly affect the accuracy of load forecasting. This paper finds out the mean value and variance of load sequence in a period of time based on statistics and then works out the bias ratio of every point in load sequence using corresponding calculation formula and at the same time compares it with threshold value so that "unhealthy data" can be removed and accurate and effective load forecasting can be ensured.2. Through the analysis of the regularity of historical load data, the conclusion is drawn that the variance of performance capacity is due to the rule that "large cycle period" overlaps "small cycle period". As to the selection of neural network input node, not only is related historical load was introduced as ?the drilling sample, but also influence of temperature and weather sensitive factors to the load variance is considered.4. To improve the slowly convergence speed of forward-back neural network, an improved BP algorithm with changeable step and variable factor a is adopted; To avoidentering into local minimum point for improper selection of initial parameters value of forward-back neural network, genetic algorithm is introduced into determine initial parameters value of network and a Short-term load forecasting method of power system based on GA-BP is presented. Compared with traditional neural network ,the method presented in this thesis can quicken the learning speed of the network and imp rove the predicting precision. In this method, GA is used to optimize connection weights of forward-back neural network until the learning error has tended to stability, Then we use SP algorithm with optimized weights to finish short-term load forecasting process. Experimental results show that this method can W e also consider the influence of climate fo r the sho r(2term load and make it as one of the input fo r the BP. Experimental results show that can quicken the learning speed of the network and imp rove the predicting precision.