Short-term power load prediction method based on similar days and RBF neural network

Abstract

The invention discloses a short-term power load prediction method based on similar days and an RBF neural network, and relates to the field of power load prediction, a proper similar day is selected as a training sample in short-term power load prediction, so that the training process can be simplified and the prediction precision can be improved; and in order to reduce the influence of summer accumulated temperature effect on similar day selection, the similarity of the temperature and other load influence factors is calculated respectively, so that a similar day is selected according to thecalculated comprehensive similarity; besides, the prediction effect of the RBF neural network is improved; the training samples are clustered by using subtractive clustering, the initial value of fuzzy c-means clustering is set according to the clustering result, the hidden layer parameters of the RBF neural network are optimized by applying fuzzy c-means clustering, and finally short-term power load prediction is carried out by combining the similar day and the improved RBF neural network, so that the accuracy of short-term power load prediction can be obviously improved.

Description

technical field [0001] The invention relates to the technical field of power load forecasting, in particular to a short-term power load forecasting method based on similar day and RBF neural network. Background technique [0002] The main purpose of short-term power load forecasting is to predict future load conditions based on historical load and influencing factors. The forecast results can be used as a reference for dispatching and distribution of the power sector; in short-term power load forecasting, forecasting based on similar day data can use less training The data achieves a high prediction accuracy, so it is particularly important to select the appropriate similar day; Li Xiaocong, Li Chuntao, and Cong Lanmei published short-term load forecasting based on the dynamic weight similar day selection algorithm based on the dynamic weighted sum of various influencing factors Select similar days; Wang Jianfeng, Xiang Tieyuan, Xu Fuxiang, etc. published a short-term load f...

AI Technical Summary

Problems solved by technology

[0002] The main purpose of short-term power load forecasting is to predict future load conditions based on historical load and influencing factors. The forecast results can be used as a reference for dispatching and distribution of the power sector; in short-term power load forecasting, forecasting based on similar day data can use less training The data achieves a high prediction accuracy, so it is particularly important to select the appropriate similar day; Li Xiaocong, Li Chuntao, and Cong Lanmei published short-term load forecasting based on the dynamic weight similar day selection algorithm based on the dynamic weighted sum of various influencing factors Select similar days; Wang Jianfeng, Xiang Tieyuan, Xu Fuxiang, etc. published a short-term load forecast based on the multi-objective similar day selection method using the concept of virtual similar days to select similar days for multiple load periods in a day; Liu Yifeng, Zhou Guopeng, In the short-term load forecasting method based on intelligent similar day identification and deviation correction published by Liu Xin et al., the similar day is intelligently identified by constructing a characteristic matrix of relevant factors, and a real-time meteorological deviation correction strategy is established to perform secondary correction on the load curve; The selection of days has certain significance, but they only select similar days by predicting the correlation analysis between the load influencing factors of the day and the historical daily load influencing factors. Because the temperature factor is different from other influencing factors, in the case of continuous high temperature The temperature change of several days will also affect the selection of similar days

[0003] The RBF neural network has powerful nonlinear mapping capabilities, and is a commonly used method in short-term power load forecasting. Optimizing its hidden layer parameter settings can improve the overall performance of the model; In the short-term load forecasting of the power system, the alternating gradient method is used to alternately train the RBF hidden layer parameters and output layer weights; Hui Lichuan, Yu Miao, Liang Zhirui, etc. published the application of the nearest neighbor propagation algorithm to improve the RBF short-term load forecasting. Select the hidden layer center of RBF; Guan Shuo, Gao Junwei, Zhang Bin, etc. published K-means clustering algorithm based on RBF neural network traffic flow prediction, using k-means clustering algorithm to select RBF hidden layer parameters; the above various improvements Both methods improve the prediction accuracy of the RBF model, but they are all improved for the RBF model with a large number of sample data, and do not improve the small sample RBF model based on similar daily data.

Images