Research On Speed Estimation Of Power Centralized EMU

As the power concentrated EMU to replace the 25 T passenger trains of the existing railway lines,the adhesion control is an important device to ensure the traction efficiency of the EMU,just like other passenger and freight locomotives.The adhesion between wheel and rail determines the running condition of EMU,and the adhesion control ensures that the serious friction of wheel is avoided due to idling of wheel.For the adhesion control system,the speed is the key parameter of control,but in general,the speed can not be directly obtained,and each axle can be obtained through the sensor of the corresponding axis of the wheel angular speed.In view of the nonlinearity of the adhesive system,this paper adopts the nonlinear system state estimation method to study the speed estimation of power centralized EMU,which mainly includes the following contents:Firstly,the speed estimation method of power concentrated EMU based on fuzzy system and extended Kalman filter algorithm is designed.Based on the multi-axis traction dynamics model,a multi-model extended Kalman filter algorithm is designed for each axis with dry and wet track surfaces.The estimated speed of each axis is obtained by speed selection method.The weight coefficients of each axis are determined by fuzzy logic.The reference speed is obtained by data fusion.The speed can be accurately estimated when the wheelset of EMU is idling or sliding due to the change of track surface.Then,the speed estimation method of EMU based on interactive multi-model nonlinear filtering algorithm is designed.Based on the uniaxial traction dynamics model,an interactive multi-model nonlinear filtering algorithm is designed considering the three track surfaces and the nonlinearity of the adhesive system.With the improved Markov matrix,the exact speed in the adhesive zone without the need of additional track surface judgment can be obtained.An interactive multi-model nonlinear smoothing algorithm is designed to improve the overall speed estimation accuracy.The target is to realize the precise tracking of the speed of the power concentrated EMU under variation of different rail surfaces.Finally,an interactive multi-model filtering method for speed estimation based on type-2fuzzy system is designed.Based on the multi-axis traction dynamics model,the estimated speed of each axis is obtained by interactive multi-model nonlinear filtering algorithm.The weight coefficients of each axis are determined by the type-2 fuzzy system.The accurate estimation of speed of power concentrated EMU is realized by data fusion.On the premise of the adaptive tracking between rails,to overcome the problem in which the estimation error of the filter algorithm is increased due to high uncertainty between wheel and rail.