Load prediction method and apparatus for power system

Abstract

The invention discloses a load prediction method and apparatus for a power system. Historical load data of a power system are obtained; decomposition and single-branch reconstruction are carried out on the historical load data by wavelet transform, thereby obtaining wavelet decomposition data of loads with different frequencies; a BP neural network model is established; wavelet decomposition data are trained by using the BP neural network model and a network parameter is optimized by using a crisscross optimization algorithm with an elitism selection strategy, and an optimal network parameter is determined; with the optimized BP neural network, load components obtained by single-branch reconstruction in the wavelet decomposition data are predicted; and prediction values of all load components are superposed and a practical prediction result is determined. According to the method and apparatus provided by the invention, on the basis of the wavelet transform and the crisscross optimization algorithm, the load prediction model of the neural network is optimized; and the neural network parameter is optimized by using the crisscross optimization algorithm. Therefore, defects that the BP neural network is vulnerable to local optimum and poor generalization can be overcome, so that the prediction precision of a region having lots of impact loads can be improved effectively.

Description

technical field [0001] The invention relates to the technical field of power systems, in particular to a method and device for load forecasting of a power system. Background technique [0002] Power system load forecasting is an important part of power system planning. More accurate forecasting data is helpful for dispatchers to make reasonable start-up and stop arrangements for hydrothermal power generation units, coordination of peak and valley power, load arrangements for transmission line tap points, and maintenance of power transformation and distribution equipment. The prediction accuracy directly affects the safety, economy and power supply quality of the power system operation. [0003] At present, the commonly used short-term load forecasting methods include classical forecasting methods represented by time series method and regression analysis method, and artificial intelligence methods represented by expert system method, neural network and support vector machine....

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method and device for power system load forecasting, to solve the shortcomings of the existing BP neural network, whi

Images