The Research And Application Of Ta Bei Power Network Load Forecasting

The short-term power load forecasting is one of the most important contents of power system. And accurate forecast of the short-term power load plays an important role in the power system's security and economy. Thus, it sets a higher request to the result of short-term load forecast. In this thesis, A new model which combins the wavelet transformation and the artificial neural network for load forecasting is proposed , and the samples are obtained from Ta Bei Area.The article analyzes the present situation, many kinds of forecast methods and the forecast models. The conventional algorithm can not reflect well the influence of conditions, but one of the intelligent algorithms—artificial neural network(ANN) method which has the high nonlinear mapping ability, may consider the factors well for the power network load's influence. This article uses BP (Back Propagation) network even if it still having many shortcomings. Such as: the study restraining speed, the restraining stability, overall minimum and so on. Therefore this text takes the Mallat algorithm of the wavelet theory to the BP network. Decomposing the sequence based on the Mallat algorithm. And then according to the characteristic of various components it structures a neural network model to carry on the forecasting. Finally, the neural network restructures various components of the forecast data through the wavelet to obtain the forecasting result.Thus, this method resolves the questions above. Finally the article uses GUI (Graphical User Interface) to design the user interface which cause the operation to be simpler.The experimental load forecast result of the Ta Bei area indicates that comparing the artificial neural networks forecast method with the wavelet neural network model which has higher forecast precision and stronger compatibility, and in the future it will have the good prospect of application.