Short-term Power Load Forecasting Based On Data Mining And Improved Whale Algorithm To Optimize Support Vector Machine

Power load forecasting is a hot topic that has been studied a lot at present.Accurate power load forecasting can ensure the stable operation of the power grid and is also of great significance for ensuring the safety of the power system.This paper mainly studies short-term power load forecasting,including preprocessing historical load data,analyzing load-related influencing factors,selecting appropriate training sets and establishing power load forecasting models.And use a variety of data mining techniques to conduct research.First of all,for how to solve the power load data preprocessing,here we use an improved fuzzy C-means clustering algorithm.The clustering center and the number of clusters for fuzzy C-means are determined by the subtractive clustering algorithm to extract the daily load characteristic curve,and then determine the bidirectional detection threshold according to the normal distribution theory.Smoothness,identify the abnormal data of the daily load curve,and finally correct the abnormal data.Secondly,the analysis of the relevant factors and the selection of similar days on the load mainly analyze the periodic characteristics of the load itself,and at the same time use the gray correlation analysis method to analyze the relationship between the weather factors and the load.The main relevant factors affecting the load are selected to form the load feature vector,and the rough set of similar days is selected based on this.Finally,the similar day is selected by the fuzzy clustering gray correlation analysis method.Finally,on the condition of the methods in the previous chapters,the support vector machine model optimized by the improved whale algorithm is adopted as the short-term electric load forecasting model in this paper.The power load data of Putian City,Fujian Province is selected for short-term load forecasting,and then the prediction model in this paper is compared with the support vector machine-based power load forecasting model.The following two conclusions are drawn through the results of the calculation example:comparing the model in this paper with the support vector machine model,it is concluded that the prediction model in this paper has higher prediction accuracy;the pre-processed data and the unprocessed data are respectively adopted in this model Prediction,it is obtained that the data predicted by the method pre-processed in this paper has better accuracy.