Research On Failure Data Mining And Visualization Method Of Refining And Chemical Instruments

The raw materials and products of the refining system are both flammable and explosive materials,and there are great risks in the process.As an important part of safety instrument system,refining instrument is an important equipment to ensure the safe operation of refining equipment and avoid accidents.This paper collects the failure data of refining instrument and establishes a case database of refining instrument failure.Research on the failure control method of refining instrument based on knowledge ontology.Multi-sample algorithm is introduced into the failure analysis of refining instrument to calculate the failure rate of similar instruments with multiple samples.Analyze the failure data of refining instrument,and establish a visualization platform of refining instrument failure data.(1)Based on the failure data of refining instrument,analyze the failure characteristics of refining instrument and construct the knowledge body of refining instrument failure.Based on knowledge ontology to establish a case library of refining instrument failure.Achieve similarity matching of target cases and source cases through improved expert scoring method.Obtain similar source cases to provide reference for risk control of target cases.(2)Introduce multi-sample algorithm into failure data analysis of refining instruments.Calculate the failure rate of similar instruments based on a multi-sample algorithm to solve the problem of querying instrument failure rates from the database.Based on the multi-sample failure rate algorithm,the safety integrity level assessment of the functional units of the refining instrument system is carried out.(3)Study the spatiotemporal distribution of failure data of refining instrument.Summarize and classify the failure data of refining instruments.Using the Vue framework and visualization library,establish a Web-based refining instrument failure data visualization platform,and realize data visual browsing.