Research On Transmission Lines Fault Analysis Based On Big Data Algorithm

With the development of the scale,capacity and coverage in the smart grid,the transmission lines are playing an important role in the national economy life,thus the faults of power system will cause significant economic losses for the society.And because the transmission line has the complex characteristics such as the long distances,wide areas,and complex geographical environments,which are easy to be affected by the natural environments and human factors,there are many difficulties in the operation and maintenance of transmission lines.Thus it has become a hot research topic how to effectively improve the quality of the operation and maintenance of transmission lines for the scholars and power system.Recently,the modern information society has entered the era of big data.The big data technologies have rapidly become the hot spots in the area of the academia and industry,and also have been widely applied into many fields.Big data analysis methods can be utilized to find out the potential models and rules from the massive data of the power system,and provide the decision support for the decision makers.As an auxiliary tool to handle the accidents,the transmission line fault analysis based on the big data algorithms can play an important role in shortening the time to address accidents and preventing the expansion of accidents.First,the thesis analyzes the common faults of transmission lines,the application fields of big data analysis methods,and discusses the application of big data algorithms and their shortcomings in the current intelligent diagnosis of transmission lines.Thus,the thesis provides a rich theoretical basis and new ideas to solve the issues of transmission lines' fault analysis via big data algorithms.Second,this thesis proposes a novel transmission lines' fault analysis model on the basis of existing research results.According to the characteristics of k-nearest neighbor classification(KNN)and the fuzzy theory,the thesis proposes a fuzzy KNN algorithm.In particular,when the category boundaries of data classification are not obvious,the application of fuzzy theory can effectively solve the deviation problem produced by the KNN algorithm.The experimental results show that the fuzzy KNN algorithm greatly improves the analysis accuracy of the ordinary KNN algorithm.In addition,the thesis implements a real-time fault analysis platform of transmission lines based on the Spark to satisfy the requirements of real-time fault analysis.Third,in order to solve the problem of the low accuracy rate of fuzzy KNN algorithm for processing the hybrid data,this thesis develops a multi-classification model based on density-based logistic regression(MCDLR),which can greatly increase the accuracy of transmission lines' fault data analysis.Finally,the thesis has done a large number of experiments to verify the feasibility and effectiveness of the proposed model.The proposed models can not only satisfy the requirements of fault analysis of transmission lines,but also realize the application ofbig data analysis algorithms in the power system,which have the good creation and practice in the society.