Research On Interval Observer-based Fault Diagnosis Method Of Traction Motors For High-Speed Trains

As a symbol of China,high-speed trains provide convenience to people,and their safety and security issues are also of concern to everyone.Traction motor is the core of high-speed rail,and its stability is related to the safety of high-speed rail.Fast and accurate fault diagnosis of traction motors is the guarantee for the safe operation of high-speed rail.Therefore,it is very valuable to study the fault detection and estimation of traction motors for high-speed trains both in theory and engineering applications.Because of the nonlinearity of the traction motor of the high-speed train,the large number of variables and the strong coupling,this puts higher requirements on the diagnostic algorithm.In this paper,the interval observer based fault detection and fault estimation method is applied to the CRH5 high-speed traction motor model(6FJA3257A).Firstly,the fault diagnosis of high-speed trains is introduced from the meaning and background of the subject.Then,several commonly used algorithms for fault diagnosis of traction motors and the application extension of these algorithms by domestic and foreign scholars are introduced.The readers are introduced to the reader through the frequently used fault diagnosis observers.Based on the research status of the fault diagnosis using the observer,the domestic and foreign situations of the interval observer method used in this paper are analyzed.Then,the traction motor system of the high-speed train,the dynamic model of the traction motor,the common coordinate system,the coordinate transformation and motor model in the coordinate system are introduced.Finally,the motor model under the dq coordinate system is given.Secondly,the model of stator winding fault is established,and then the correlation algorithm based on interval observer is introduced into the fault detection of traction motor.The proposed interval observer design method has the following characteristics: the stable operation system is always located in the upper and lower bounds of the interval observer without the occurrence of faults,so that the interval observer can track the original system in real time after the motor runs stably.If the motor system fails,the original system is no longer in the upper and lower bounds of the interval observer,and deviates from the interval observer.A linear transformation method is proposed to solve when the interval observer design conditions are not met.The fault diagnosis algorithm designed in this paper is verified by MATLAB/SIMULINK module.In the simulation,the stator system and stator fault are injected into the motor system.Thirdly,considering the failure of the shaft stator current sensor in the traction motor,the fault estimation algorithm is continuously studied based on the previous fault detection.In order to apply the state estimation method to achieve fault estimation,the state quantity and sensor fault of the system are augmented into a new vector.After simplifying the model,an equivalent state equation is obtained,and the unknown input observer is designed on the augmented model.The advantage is that the disturbance is completely isolated from the structure at the time of design.Then the interval observer is used to track the state,and the fault estimation algorithm is used to estimate the fault of the stator current sensor.When the augmented error system does not meet the positive system design conditions,the linear transformation method is used to construct the positive system.At the same time,how to improve the fault estimation performance is studied.Fourthly,an interval fault estimation observer based on adjustable parameters is designed for high-speed train traction motor,and the fault estimation performance is enhanced by adding adjustable items.The proposed method has the same application range as the augmented fault estimation design,but is more flexible in the fault estimator design than the augmented fault estimation design.Then,based on the augmented model,the upper bound and the system and the lower bound and system error dynamic equations are constructed.Then the existence of the interval observer is proved by theory,and then the traction motor system and the interval observer are equivalently transformed to make it conform to the positive system.The design conditions introduce constant faults and time-varying faults in the simulation part.When the circular domain configuration is the same,the adjustable parameters are changed to study the influence on the fault estimation effect.Different circular domains are studied under the condition that the adjustable parameters are consistent.The impact of the configuration on the results is verified by MATLAB/SIMULINK.Finally,the paper summarizes all the contents of this thesis,and puts forward the prospect of the topic,pointing out the areas where research needs to be strengthened.