Dynamic unbalance identification and quantitative diagnosis of cardan shaft in high-speed train based on improved TQWT-RBFNN-NSGA-II method

Abstract

In order to identify and quantitatively diagnose the dynamic unbalance of the cardan shaft equipped in high-speed trains, firstly, a numerical simulation model of the vehicle considering the cardan shaft dynamic unbalance is established. Secondly, the vibration and acceleration response of the motor and gearbox under different cardan shaft offset conditions are calculated. In order to better extract the feature signal related to the shaft unbalance, a novel algorithm based on the improved tunable Q-factor wavelet transform (TQWT) method is proposed. On this basis, in order to build the relationship between the unbalance and the characteristic frequency amplitude intensity (AI) index, radial basis function neural network (RBFNN) is used. In addition, non-dominated sorting genetic algorithm II (NSGA-II) is used to find the optimum Pareto solution of a combination of offsets to quantitatively analyze the cardan shaft dynamic unbalance. The results show that the proposed method can identify dynamic unbalance effectively and its magnitude quantitatively with an error of less than 15 %.