Design Of Petrochemical Equipment Management And Fault Classification System Based On FSM And SVM

In recent years,the petrochemical industry is in an important period of equipment renewal and enterprise reform.The large number of devices,rapid updates,and complex and ever-changing business processes have also brought heavy equipment management system maintenance costs.At the same time,companies lack intelligent solutions for the classification of equipment failures.Most of them still depend on artificial forecasts,which leads to a reduction in the accuracy of classification.The workflow engine has been proven to improve the management capabilities of the information platform to some extent.This paper builds a workflow model based on a finite state machine,and enhances the business expansion capabilities of the model through the Aspect Oriented Programming idea,which improves the scalability of the model.Then state transition algorithm is proposed based on the workflow model.The algorithm implements AOP enhancement processing and workflow scheduling of the model.The engine is used as the middleware of the petrochemical equipment management information platform to provide process scheduling and data scheduling for the petrochemical equipment management system based on the web.To a certain extent,the system’s use threshold and maintenance costs are reduced.In the fault classification module of the equipment management system,this paper collects a petrochemical original operating data,and carries out data augmentation with different intensity based on the SMOTE algorithm for the small data volume category in the data sample to obtain the appropriate data enhancement strategy for the subject.The experimental results show that the classification accuracy of the few data categories after the enhancement has been significantly improved.Then this paper verifies the classification status of the three multi-category algorithms of OVO,OVR and DAG-SVM before and after data augmentation under different kernel functions.Analyze and compare from the experimental results and get the appropriate SVM multi-classification solution in the context of this paper.The results show that the DAG-SVM algorithm performs best on the overall performance of the classification accuracy and model training time-consuming.Finally,this paper applies the experimental conclusions to the fault classification module in the equipment management system.This module generates fault classification reports through the data scheduling of the workflow engine to assist the equipment failure processing.The petrochemical enterprise equipment management and fault classification system designed in this paper has achieved good results in the enterprise and has a certain application prospect.