Design And Implementation Of Equipment Fault Datebase Based On Text Semantic Analysis

Modern factories mainly rely on production equipment to support large-scale production,the operation status of equipment directly affects the production plan of enterprises.In general,the maintenance of equipments requires considerable expertise and experience of maintenance workers.In case of an equipment failure,the knowledge and experience of the maintainer will result in a difference in the quality of the work and time to solve the problem.Although problem solving is based on fault manuals,experienced and skilled professionals can quickly diagnose and take actions by applying personal knack.But these knacks are difficult to pass it on completely to a successor,only a large number of maintenance records are left.For this reason,people have been trying to transform text knowledge into data-based knowledge,but the development of systems for researching related equipment or extracting text features to search for similar instances is currently not perfect.This topic proposes an equipment fault database management system that provides an action guide for emerging failures by using equipment maintenance information and maintenance experts’ experience,takes maintenance records as an example of semantic matching.This paper collects the text data of the equipment failure of the welding automation production line in the aluminum car body workshop of a rail car manufacturing enterprise from2015 to 2020,and establishes the case base.This paper proposes a method for text topic mining of equipment failure cases that combines Chinese word segmentation technology,TFIDF and LDA topic model algorithm.Semantic matching is carried out between problem and subject case set by a random forest algorithm.After matching,cluster the subject case collection,and extract the maintenance case with the best clustering effect as a guide for measures.In this study,the intelligent semantic matching of new cases is carried out,and at the same time,the measureing guide is put forward.Combining with fault tree method,the Bayesian network model of system reliability evaluation is established,and the reliability of the system is analyzed through the unique bidirectional reasoning of Bayesian networks,which is convenient to find the weak links of the system.The equipment fault database management system based on text semantic analysis realizes the functions of full-text retrieval of equipment faults,feature extraction of equipment faults,analysis of frequent parts of equipment faults,analysis of key equipment faults,correlation analysis of accidents,etc.The practice has proved that the research results of this paper can provide practical basis and guidance for technicians and experts working in the actual field,and at the same time provide feasible ideas for the construction of the faulty database management system in the future.