Analysis And Research On Short-term Forecast Of Power System Load

In the new situation of power system reform, the price competition mechanism is introduced into the electricity market, the system operators and market participants pay more and more attention to the short-term load forecasting. Rapid and accurate forecasting facilitates appropriate plan of electricity transactions, developing appropriate operational plans and bidding strategies. With the development of new theories and new technologies, the research on the new method of load forecasting is still deepening.Support vector machine is a common method of data prediction, compared with other methods, it shows a better performance, it can achieve the methods of structural risk minimization better and it is used in pattern recognition and dealing with regression problems and many other areas.In this paper, based on the advantages of support vector machine (SVM) in nonlinear learning and prediction performance, the non-linear characteristics of short-term load forecasting are analyzed, and the support vector machine are difficult to be used in the selection of parameters and some defective problems exist in the traditional method. At the same time,because of the different performance of different kernel functions, some problems still exist in the kernel function selection of support vector machines. The optimization method of short -term load forecasting in power system based on support vector machine is proposed. In this paper, an improved bilinear search method and an improved mesh search method are proposed for the existence of bilinear search method and grid search method to compare with the bilinear search method, the mesh search method, and the experimental method is used to verify the accuracy and timeliness of the improved method. In addition, in order to solve the problem that the weight coefficient of mixed kernel function is difficult to be selected accurately, particle swarm optimization is used to optimize the weight. Aiming at the problem that the particle swarm algorithm is easy to fall into the local optimal problem, an improved particle swarm method is proposed to dynamically adjust the inertia weight, local acceleration constant and global acceleration constant according to the fitness value in the iterative process. The experimental results show that the improved kernel function of the improved particle swarm has higher prediction accuracy than the single kernel function.