Research On Flexible Load Feature Extraction And Control Strategy For Wind Power Consumption

In the context of the smart grid and the ubiquitous power Internet of things,the accurate extraction of power load is not only beneficial to the safety and stability of the power system but also helpful to the demand management and the load forecast.This paper studies the extraction method of pure trough heat storage load inflexible load,the two algorithms of KNN and SVM are combined with K-means clustering to improve the extraction accuracy of flexible loads.In recent years,the rapid development of wind power generation technology has positive significance for the rational use of resources and the protection of the environment.However,there is a serious problem of wind power consumption,especially in the "three north" areas of China.In winter,the use of electric heat storage technology is an important means of wind power consumption,which can greatly improve the utilization rate of wind power and reduce wind curtailment.In addition,the battery energy storage has the characteristics of rapid regulation,bidirectional flow,and time-shifting energy,which has plays an important role in achieving peak load shifting of the grid and reducing wind curtailment.The main work of this paper is as follows:(1)Summarized the significance of load clustering extraction in power system and the current status of wind power development in China,as well as the domestic and foreign research status of power load clustering and the use of heat storage and energy storage to reduce wind curtailment.(2)The K-means clustering algorithm is studied and analyzed,and the K-means is used to extract the pure trough heat storage load in the flexible load.The results show that this algorithm requires multiple clustering and low accuracy.The method of combining KNN,SVM,and K-means is proposed,and the features extraction of pure trough heat storage load is again performed.By comparing with the results of directly k-means clustering,more accurate extraction is achieved,and the three validity of combination.(3)Design the display interface of the heat storage system.In order to observe the operation of trough heat storage load more intuitively,the feature extraction method combined with the three algorithms mentioned above is used to extract the trough heat storage load of the Liao Ning province based on the grid full-service data center platform.Finally,it shows in the form of a heat storage interface.(4)To reduce the wind abandonment,the centralized storage heat and distributed storage are first presented in the paper and then investigated using thermal storage system,storage systems given to abandon the control strategy has carried on the research and analysis of the wind.At last,explain the effectiveness of the two to improve wind power consumption.