Research On Small Fault Diagnosis Of High-speed Railway Traction Drive System Based On ICA And Bayes Algorithm

The safe and stable operation of high-speed railways plays a very important role in the steady development of our society and economy.One of the keys to ensuring the safe and stable operation of high-speed railways is the traction drive system.Due to the high switching frequency of insulated gate bipolar transistors(IGBTs)in the traction drive system of high-speed rail,heat generation and wear are serious,and it is easy to cause the accumulation of incipient faults and cause larger faults.Aiming at the above problems,this thesis studies the incipient fault diagnosis method of the high-speed rail traction drive system based on Independent Component Analysis(ICA)and Bayes algorithm.Firstly,this thesis proposes that the Independent Component Analysis algorithm is used for the preliminary diagnosis of incipient faults of high-speed railway traction transmission system,and then the Bayes algorithm is combined to improve the diagnosis method of incipient faults,so as to improve the efficiency and accuracy of the diagnosis of incipient faults.Specific include:(1)The comparative analysis of the incipient fault diagnosis methods,the data-driven incipient fault diagnosis method is selected for the high-speed rail traction drive system,and the independent component analysis algorithm based on the data-driven is selected based on the actual situation of the high-speed rail.(2)According to the composition of the high-speed railway traction drive system,MATLAB Simulink is used to build a high-speed railway traction drive system simulation model and analyze the circuit of the traction drive system,determine where it is easy to form an incipient fault,and focus on the incipient fault formed by the traction rectifier and traction inverter IGBT in the high-speed rail traction drive system.(3)A preliminary method for diagnosing incipient faults in high-speed rail traction drive systems based on independent component analysis algorithms is proposed.The separation of background noise in traction drive systems is studied.Offline component modeling and online detection are performed using independent component analysis algorithms to simulate incipient faults in traction drive systems.Complete the separation of background noise and analysis of statistical results,the fault points are located based on the fault contribution map.(4)An improved method for incipient fault diagnosis of high-speed railway traction drive system based on independent component analysis and Bayes algorithm is proposed.Two comprehensive statistics,Ensemble Squared prediction error(ESPE)and Ensemble I~2(EI~2),are applied to incipient faults.The detection and simulation analysis were performed to verify the accuracy and feasibility of the proposed improved algorithm for incipient fault diagnosis of high-speed railway traction drive system.