Research On Fault Diagnosis Method Of Traction Gear Based On Volterra Series

With the progress of society and the continuous improvement of living standards,higher requirements have been put forward for the safe development of railway transportation.The traction gear transmission system is an important part of the locomotive.Timely and effective detection of early failures of the traction gear(such as crack failures)is crucial for the realization of locomotive “conditional” maintenance to ensure locomotive safety.Therefore,it is of great practical significance and engineering application value to carry out research on fault diagnosis methods for traction gears of locomotives.There is a complex nonlinear coupling relationship between the traction gear crack fault and the symptom,making the effective diagnosis of traction gear crack faults difficult.Based on this,taking advantage of Volterra series which can fully reflect the dynamic characteristics of nonlinear input / output of the system,the fault diagnosis method of traction gear based on Volterra series is studied in this paper.The main research content is as follows:(1)The dynamic model of the traction gear system of the HXD1 B locomotive was established by considering the time-varying meshing stiffness,meshing error and meshing damping.According to the energy method,the formulas for time-varying meshing stiffness under normal and crack fault conditions are deduced,and the formation mechanism of cracks in traction gears is studied;the system dynamics model is solved to provide data support for Volterra time-domain nuclear identification in subsequent chapters.(2)The traction gear crack fault diagnosis method based on recursive least squares algorithm to identify the Volterra time domain kernel is studied.The third-order Volterra series model of the traction gear system under normal and crack failure conditions is established.The Volterra time-domain kernel is identified by recursive least-squares algorithm.Taking the second-order and third-order Volterra time-domain kernel as the main criterion,it is analyzed an diagnosed whether the teeth of the traction gear are cracked.Simulation experiments show that the Volterra series model can accurately reflect the changes of the nonlinear dynamic characteristics of the system.The second-order and third-order time domain nuclei under normal and crack failure conditions have obvious differences,and the fault characteristics of traction gear cracks can be extracted accordingly to realize crack fault diagnosis.(3)A fault diagnosis method of traction gear crack based on particle swarm optimization algorithm for identifying Volterra time-domain kernel to extract fault features is studied.The particle swarm optimization algorithm is used to identify the Volterra time-domain kernel.And Second-order and third-order Volterra time domain kernels are used as the main judging criteria to analyze and diagnose whether the teeth of the traction gear are cracked.The simulation experiments show that when performing Volterra time-domain kernel identification,the particle swarm optimization algorithm has higher recognition accuracy than the recursive least-squares algorithm,and can more accurately reflect the nonlinear dynamic characteristics of the system.Therefore,the diagnosis accuracy of cracks in the traction gear can be improved.