Research On Fault Diagnosis Of Single Electromagnet Suspension System Sensor For Medium-speed Maglev Train

Magnetic Suspension(EMS)medium-speed maglev trains as a new type of maglev train combine the advantages of simple low-speed maglev trains which have simple guide structures with high-speed maglev trains.As a key component of medium-speed maglev trains,suspended sensor has a bad working environment and is easy to wear out.Its failure will seriously affect the stable suspension of the train,and even cause accidents of falling or sucking.This paper takes the single electromagnetic suspension system sensor as the object and studies the method of sensor fault diagnosis(FD),including analyzing all typical fault features,adopting the fastest tracking differentiator and Variational Modal Decomposition(VMD)to diagnose the fault.The main research contents and achievements are as follows:(1)Sensor fault characteristics of single electromagnet suspension system is analyzed.The single electromagnet suspension system is considered as a black box.The suspension behavior model of the single electromagnet suspension control system is developed considering the influence of the feedback control.The typical failures of the sensors are analyzed,and the failure signals of the three sensors of gap,acceleration and current are obtained.Based on the analysis of fault signal time domain features.The results show that partial failure of three sensors have similar features.A simple and fast method of diagnosing the amplitude domain is proposed for faults with distinct characteristics.The faults with similar characteristics cannot be diagnosed by simple time domain analysis methods.(2)Based on the sampling control of the optimal tracking differentiator,the sensor faults with obvious time-domain characteristics are diagnosed.In this paper,the gap signal is differentiated by the optimal tracking differentiator.The fault is diagnosed using the contrast residual of the speed signal of the differential speed signal and the acceleration integral.The residual threshold is set by the Bayesian decision theory.The four failure categories(e.g.,additive gap fault,multiplicative fault,acceleration sensor additive fault,and zero output fault)are diagnosed accurately applying to the value domain analysis.(3)The sensor failure with insignificant time-domain characteristics is diagnosed based on combination between Variational Mode Decomposition and Fourier transform.In this paper,the time-frequency domain of accelerometer and current sensor multiplicative fault signals is analyzed by using VMD and Fourier transform.The results show that the mid-high frequency decomposition signal can be better characterized the sensor in operation and the category of fault.(4)In this paper,the VMD method in fault analysis is improved based on the modal function and chaotic firefly algorithm considering the shortcomings of the stability and timeliness,which is more suitable for real-time diagnosis of maglev train sensor faults.Based on the analysis of sensor frequency domain signal characteristics,in this paper,a method for selecting the fixed frequency of sensor fault diagnosis is proposed.Modal function of amplitude modulation and constant frequency is introduced in the adaptive decomposition process of VMD,which improves the stability of the diagnosis.Then the chaotic firefly algorithm is used to optimize the parameters of the improved VMD method,and a parameter evaluation function is proposed according to the diagnostic requirements.The computational efficiency of the VMD method in sensor diagnosis is improved.The two diagnostic methods and the fault determination criteria proposed in this paper can be effectively applied to the sensor fault diagnosis of a single electromagnetic suspension system.The tracking differential method can quickly diagnose the faults that have serious impact on the train.The improved VMD method can be used to diagnose unobvious failure in time.