Ride Comfort Standard Evaluation Studies Based On Genetic Neural Network Trains

High speed is a very important direction for China railway's modernization. High speed railway should be developed; meanwhile, the ride comfort should not be decreased as the speed increases. However, the factors that affect the ride comfort were not paid sufficient attention to. Lack of comfort and stability testing methods and measurement standards, it is difficult to ensure passengers travel in comfort.International Organization for Standardization and International Union of Railways had made appropriate standards for the value of comfort. Through special researches, especially through plenty of tests, Britain and Germany have revised and drawn up specific plans suitable for their own applications founded.This paper established the model of train vibration through the SIMPACK simulation software, and then got lots of vibration acceleration data by inputing each track irregularity power spectrum. After then, comfort index value was calculated accurately in MATLAB environment.The method based on genetic algorithm (GA) was employed for optimizing the genetic algorithm back propagation (BP) neural network. But the performance of BP neural network is not excellent. This paper proposed another method for optimizing the RBF neural network based on hierarchical genetic algorithm (HGA). The simulation results clearly showed validity of hierarchical genetic algorithm.Two models based on HGA-RBF neural network model and GA-BP neural network model were built up. The connection between various ride comfort indices of railway vehicles was analyzed by these models. The results show that HGA-RBF model has higher prediction accuracy and stronger generalization ability than traditional GA-BP prediction model.