Research On Short-term Load Forecasting Model Of Power System In Leshan

Short-term load forecasting of power system generation scheduling and transmission is to develop programs primarily based on its level of prediction accuracy, and it directly affects the operation of security and economy in power system. With the power system market continued in-depth, short-term load forecasting in power system is becoming more important. In this paper, principal component analysis based on least squares Support Vector Machines combined short-term load forecasting model to Leshan Electric Power Bureau historical load data and weather data-based, mainly on short-term load characteristics, historical data pre-processing and support for Vector Machine parameter selection, the construction of the nuclear function study. Principal component analysis is used for the input variables affecting the load characteristics of the data extraction component, select fewer variables as much as possible the integrated variables to reflect the original information, and select radial basis function as the least squares support vector machine kernel function, this can better achieve the non-linear fitting to improve prediction accuracy. Using this short-term load forecasting model in Leshan, and making a comparison case analysis with other methods, experimental results show that the model has good predictive results and it can be applied to other regions of the grid short-term load forecast.