| During the operation of power equipment,various defects will occur.If these defects are not detected and eliminated in time,it will endanger the operation safety of the power system.In order to improve the accuracy of defect diagnosis and elimination,it is necessary to analyze the factors affecting the occurrence of defects in power equipment and the law of defects.At present,the details of power equipment defects are often described in text form.The workload of manual analysis and statistics is large,and the analysis is inaccurate due to differences in knowledge and experience.Therefore,it is urgent to use a computer to automatically perform text mining to extract the key from the defect description text information to achieve fine analysis and statistics of power equipment defects.The research work in this paper is mainly aimed at the text of power equipment defects,researching text mining technology,and realizing the refined analysis of power equipment defects.The main research contents of this article include:1.Analyzing the preprocessing technology of power equipment defect text.Firstly,the characteristics of defect texts of power equipment are analyzed,and it is pointed out that the defect texts include structured and unstructured information.The mining object of this article is the unstructured part,that is,the detailed description of the defects.Then the preprocessing technology of the defect texts,including the defect text word segmentation technology and ontology dictionary construction technology,as well as methods for removing stop words;then briefly introduces common text representation methods,including vectorized representation models,tree / graph structure models;and finally proposes quality analysis of defective texts based on knowledge graph technology.The method and retrieval error checking process can provide qualityguaranteed text for subsequent text mining.2.A text mining method for power equipment defects based on semantic framework is proposed.After analyzing a large number of defect texts,it can be found that a text often describes multiple defect conditions of different parts of a device.If it is not separated,it will cause confusion in understanding and reduce the statistical accuracy of defects.Therefore,we choose to establish semantics framework and semantic slot methods are analyzed.First,the power semantic framework and semantic slot are defined,and then the slot filling and semantic framework construction process is proposed.The semantic close-up matching method is used to construct the semantic framework.The word string merge is used to automatically improve the ontology dictionary.Finally,the defect text mining results are obtained.The application in reliability statistics was studied.Calculation examples show that the proposed mining technology is feasible and effective for automatic classification and statistics of power grid defects.3.A method of power equipment defect analysis based on text mining technology is proposed.After the unstructured information of the defective text is structured through the semantic frame method,the structured information of the defective text is merged.First,a power equipment defect analysis model based on text mining technology is established to decompose the factors that cause the defects into internal factors and external factors,and gives the definition of defect rate.Then through an example,the defect rate of a single factor and the association rules and degrees of correlation between multiple factors are analyzed.The analysis results can not only summarize the occurrence rules of power equipment defects,but also provide guidance for the selection,elimination and maintenance of future equipment significance. |