Design And Implementation Of Data-driven Motor Fault Diagnosis And Prediction System

The motor is a kind of industrial equipment with the largest consumption and the widest coverage,which occupies a very important position in industrial production.Once the motor fails,the economic loss will be unpredictable,however,the failure of the motor during its life cycle is inevitable.In order to avoid economic losses and accidents caused by motor failures,it is of great significance to perform motor failure prediction and health management.Data-driven PHM collects characteristic parameters related to motor attributes based on advanced sensor technology,and correlates these characteristic parameters with useful information,uses machine learning and data mining algorithms to perform system state monitoring,fault diagnosis and fault prediction,and gives motor faults The degree of evolution and remaining life to provide decision support for the maintenance,repair and management of industrial motors.Therefore,it is planned to take the motor fault prediction and health management as the direction,and apply industrial big data and data mining technology to design and implement the motor predictive maintenance system.In this paper,the data-driven motor fault diagnosis and fault prediction method is used to reduce the dependence on the physical model of the motor.The motor's key performance indicators are used to identify the motor state.The KPCA method based on the contribution rate graph is used to diagnose the motor fault and the particle filter is used.Motor fault prediction,and the introduction of self-feedback and self-learning mechanisms to improve the accuracy and adaptability of fault prediction.Compared with most of the fault diagnosis and fault prediction methods,the method presented in this paper does not need a large number of fault data to establish a model,can accurately carry out motor fault diagnosis,can accurately predict the remaining service life of the motor after the fault,has strong feasibility and high practicability.In addition,this paper also designs and implements a data-driven motor fault diagnosis and prediction system.First,the functional requirements of the system are analyzed,and other non-functional requirements of the system are also explained.Then,the big data computing framework of the system is developed,the overall architecture of the system is designed,the related implementation technologies are introduced in detail,and the core is finally displayed.The fault diagnosis and fault prediction function.