Research On Identification Of Power Equipment Family Defect Based On Text Information

In this intelligent era,with the development of information technology,data and information are increasing exponentially.In current technological revolution,a breakthrough has to be achieved in terms of how to extract valuable information from these numerous and vogue data and convert it to effective knowledge.Grid companies in all the provinces have accumulated a great number of electric power data texts about the defect characteristics of electric equipment,status evaluation information and equipment operation log.Valuable information about the operation,overhaul,fault and customer needs in these data is of great guiding significance for grid companies to improve the management of stable operation of electric system.Based on data mining technology,texts about the family quality defects of voltage transformer,current transformer,circuit breaker and disconnecting switch in a certain province in recent years have been studied to solve the problem that the potential value of many grid-related texts wasn’t utilized.Defect texts were preprocessed using jieba particle,and interrelation of the key words of family defects was analyzed with Apriori algorithm.It was found that family defects of different electric equipment were interrelated.Then,an auto recognition model for family defects of electric equipment through Logistic regression and XGBoost algorithm was built and the recognition effect of each algorithm was evaluated.It was found that XGBoost had a definite advantage in recognizing family defects of electric equipment.Defect data of electric equipment are basically non-linear data,so the non-linear model performs better than the linear model.In the end,with the study aforesaid,the recommendation algorithm theory was applied and the application of family defect maintenance scenario of electric equipment was realized.Feasible overhaul suggestions about upcoming defects of electric equipment were proposed on the basis of electric power database.Thus the efficiency of equipment maintenance is improved and cost is reduced.