A wavelet neural network short-term load prediction method based on improved harmony search optimization

Abstract

The embodiment of the invention provides a wavelet neural network short-term power load prediction method with improved harmony search optimization, which comprises the following steps: preprocessinghistorical load sample data; using a recursive feature cancellation method to select the features of the hysteretic historical load data and determine the input of wavelet neural network; introducingtwo updating strategies of differential evolution algorithm into the updating process of a harmony search algorithm by linear decreasing rules, and selecting the harmony of mutation operator by a tournament selection strategy; constructing a wavelet neural network and selecting the optimal initial weights of the wavelet neural network by using the improved harmony search algorithm; carrying out short-term load prediction using the wavelet neural network. According to the technical scheme provided by the embodiment of the invention, the accuracy of short-term electric power load prediction canbe improved.

Description

【Technical field】 [0001] The invention relates to a wavelet neural network short-term power load forecasting method based on improved harmony search optimization, which belongs to the technical field of power system load forecasting. 【Background technique】 [0002] In the field of short-term power load forecasting, forecasting methods are mainly divided into traditional methods based on statistics and methods based on machine learning. The traditional methods are mainly time series method, regression analysis, Kalman filter method, etc. These methods mostly judge the future trend of load by fitting historical data. The method is simple, but the disadvantage is that it cannot reflect the nonlinearity of load. characteristic. Machine learning methods mainly include support vector machine (SVM), BP neural network, etc. Although the prediction accuracy of these methods is higher than that of traditional methods, there are also problems such as complex process, poor stability, a...

AI Technical Summary

Problems solved by technology

However, the harmony search algorithm still has the defects of unstable convergence and insufficient local search ability.

Images