Appliciation Of Support Vector Machines In Power System Load Forecasting

Power system load forecast, which is the base of power market technique supporting system, is an essential part of power system operation, control and planning .The level of load forecasting is one of the measures of modernization of power system management. Because power system load forecasting is a uncertain, non-linear, dynamic and complicated system, it was difficult to describe such a nonlinear character of this system by traditional methods, So the load forecasting could not be accurately forecasted.Support Vector Machines (SVM) is a new machine learning technology, which is based on Statistical Learning Theory. It can solve small-sample learning problems better by using Experiential Risk Minimization in place of Structural Risk Minimization. Moreover, this theory can change the problem in non-linearity space to that in the linearity space in order to reduce the algorithm complexity by using the kernel function idea. Because it has quite perfect theoretical properties and good learning performance, SVM theory becomes the new research hotspot after the research of Artificial Neural Net and pushes the development in machine learning theory and technology.Support vector machine is a relatively recent machine learning technique. There are still many unexplored areas and unanswered questions of both theoretical and practical nature in this field, such as training speed, parameters selection, etc. In order to study and solve these problems, this thesis mainly focuses on the support vector machine algorithm.On the basis of analyzing the parameter performance of support vector machine, an immune support vector machines method for short-term load forecasting is presented in which the parameters in SVM method are optimized by immune algorithm. The calculation results from load forecasting example of actual power network show that the presented immune SVM method can offer more accurate forecasting result than SVM method.A load forecasting method uses the hybrid model of wavelet transform and support machine is given. At first, based on wavelet multi-resolution analysis the load series is decomposed into the series with different frequency characteristics, then according to the features of the decomposed components different SVMs are constructed to forecast the components, finally the forecasted signals of the components are reconstructed to obtain the ultimate forecasting result, Experimental results show that the given forecasting method is effective.