Study On Condition Monitoring And Fault Diagnosis Of SPMT Based On Data-driven

Self-propelled modular transporters are widely used in various large-scale heavyload transportation occasions due to their advantages of large load capacity,flexible use and high stability.In this paper,the state monitoring and fault diagnosis system of the transporter are studied for the safety and fault maintenance problems in the use of the self-propelled modular transporter.First,a certain type of self-propelled modular transporters is used as an object to introduce the basic composition,functions,and application characteristics of the transporter.And a detailed analysis has been done,which was about the structures and the hydraulic system of its main components,drive system,suspension system,and steering system.Based on the existing monitoring system,a state monitoring system plan including basic parameter monitoring,safe state monitoring and fault state monitoring is proposed.And the realization process of the basic parameter monitoring and safe state monitoring in the system,and preliminary analysis of the failure status of transport vehicles are completed.Then the method of the system’s fault diagnosis to achieve its fault status monitoring has being studied,and due to the limitation of time,the fault diagnosis method research only focuses on the suspension hydraulic system.To this end,a complete model of the suspension hydraulic system is established in the simulation software AMESim.The normal and fault conditions of the suspension system are simulated,and the impact of different types and degrees of hydraulic component failures on the system performance is initially analyzed,which provides basis and data source for subsequent fault diagnosis.Finally,based on the introduction and comparison of existing hydraulic fault diagnosis methods,and the characteristics of the system itself,a fault diagnosis scheme has been designed in this paper,which is based on wavelet packet analysis and multiclassification support vector machine.In this scheme,the collected flow signal samples are analyzed by wavelet packet and the energy spectrum of the sub-signal is extracted.Combined with the results of simulation analysis and comparative experiments,suitable features are selected for various types of faults.Through analysis and comparison of various support vector machine methods,a multi-class support vector machine for fault recognition based on binary tree are designed.The samples obtained from the simulation model are processed by the above wavelet packet analysis,and then used to train and test the support vector machine model.With an improved grid search method,the best parameters for model are found.This scheme finally gets a good classification effect and realizes the fault diagnosis of the suspended hydraulic system.So far,this paper has completed the main research work in the condition monitoring system of the self-propelled modular transporter.