Regional Control Strategy For Power Quality Of Distribution Network Based On Data Driven

With the access of new energy and power electronic load,the power quality problem of distribution network becomes more and more serious.Due to the uncertainty of distributed generation output and the variability of power electronic equipment operation,the application of power system is complex and changeable and the network structure is constantly adjusted.It is difficult to build pollution model and grid structure model accurately.In addition,the research on power quality management strategies mainly focuses on single power quality control,resulting in a large number of governance areas and increasing operation and maintenance costs.Therefore,in this paper,how to make full use of data information is studied to obtain a comprehensive partition governance strategy for harmonic and voltage problems.(1)Aiming at the problem that the number of single power quality regionalized control method divides many partitions and the control equipment is difficult to make full use of,a comprehensive partition control strategy of power quality based on grey correlation degree weight is proposed in this paper.Based on the preprocessed data set,the grey correlation weight is used to reflect the influence of each index on the overall power quality level at each sampling time,and the comprehensive power quality time series is constructed.Based on the comprehensive power quality data set,the grey correlation analysis method is used to obtain the comprehensive correlation degree.Then,according to the zoning judgment,the region is divided,dominant nodes and the compensation rate of the control devices are determined.Finally,the number of partitions is reduced,and the power quality control device installed which can control the problems of harmonics and voltage deviations simultaneously at the dominant node.So,the power quality of the distribution network is improved.The rationality and effectiveness of the proposed control strategy are verified by simulation analysis.(2)Aiming at the problem that the limitation of time series geometric similarity coefficient leads to unreasonable regional division results,a comprehensive regionalized control strategy of power quality based on improved grey correlation analysis is proposed.Based on the comprehensive power quality data set,the numerical distance similarity correlation degree and the change trend similarity coefficient are combined to obtain the improved grey correlation similarity coefficient.Then,the improved grey correlation degree is obtained.According to the zoning criteria,the region is divided,the dominant nodes and the compensation rate of the control devices are determined.Finally,the rationality of regional division and the results of partition governance are further improved.The rationality and effectiveness of the proposed control strategy are verified by simulation analysis.(3)Aiming at the problem that the piecewise linear representation method is difficult to screen out the random disturbance data in the dimensionality reduction sequence,which affects the accuracy of correlation degree and the quality of region division,a power quality partition governance strategy based on improved piecewise linear representation is proposed in this paper.Based on the traditional piecewise linear representation method,the corresponding threshold parameters according to the variation of adjacent and interval sequence points are determined in this strategy to remove the disturbance data and reduce the dimension to construct the effective feature point time series.Based on the effective feature point data set,the improved grey correlation analysis theory is used to obtain the correlation degree.Then,the region is divided according to the partition criterion to determine the dominant nodes and the compensation rate of the control devices.Finally,the disturbance data are removed and the partition governance results are improved.The rationality and effectiveness of the proposed control strategy are verified by simulation analysis.