Research On Bayes Analysis System Of Maintenance Data

Maintenance data refers to all relevant data that generated or used in the activity of equipment recovery,maintenance and improvement.The maintenance data is rich in connotation and various data types,the main components include: econfiguration data(the basic configuration of equipment includes the configuration of its sub-system and parts,technical performance and other parameters),Technical status and failure data,maintenance process data(repair record or maintenance sheet),supply support data(equipment supply information,etc.),planning management data(equipment management related information,such as maintenance unit,equipment affiliation).The maintenance data provides the basis for mastering the actual technical state of the equipment,ensuring the law of resource consumption and maintaining the capability and level of maintenance.It is great significance to improve the pertinence,accuracy and scientific level of equipment maintenance by stufficient analysis,mining and utilization of these data.So in order to improve the maintenance data analysis ability,improve the utilization rate of maintenance data information and data analysis to analysis personnel requirements,reduce maintenance in this paper,based on the bayesian method to analyze maintenance data,combing theory algorithm,design a modular system,and validation of the system.The main contents are as follows:(1)Maintenance data bayesian model finishing.Through the in-depth reading of the related bayesian data analysis literature,the paper summarizes the typical algorithms as the theoretical support of the system.The model is divided into discrete model,continuous model and hierarchical model by data type.(2)Maintenance data bayesian analysis system design.With the support of theoretical model,system software development is carried out in Python language,where PyMC3 is the core library of bayesian analysis and PyQt4 as the core library of Gui construction.(3)Maintenance data bayesian analysis system application.The final link of software system development is applied to the actual application.In this paper,the actual cases are selected to apply the test to the system,and discrete data and continuous data are selected respectively.(4)Maintenance data bayesian analysis system validation.The system verification is the key link to guarantee the reliability of the results.This paper validates the accuracy of the system by comparing and verifying the results.