Research On Model-free Adaptive Control About Locomotive Adhesion

The rapid development of economy and society puts forward higher requirement on the volume and speed of railway transportation,which aggravates the deterioration between locomotive wheels and rails,leading to a reduction in the life of related equipment,affecting the running quality of the locomotive,and even endangering running safety.The purpose of locomotive adhesion control is to improve wheel-rail adhesion condition and effectively exert the locomotive traction/braking force.However,the accurate identification of parameters during locomotive operation and a reasonable and effective adhesion controller are the basis for achieving good adhesion control effect of locomotive.Therefore,this thesis starts with the realtime filtering estimation of adhesion coefficient and the design of the adhesion controller,and conducts the research on the model-free adaptive control about locomotive adhesion.The main details of this thesis are as follows.(1)Through the illustration on the wheel-rail adhesion mechanism,the wheel-rail adhesion characteristics and related influencing factors are analyzed.Based on the single-particle model,the force condition during locomotive operation is analyzed,and a simplified single wheelset dynamics model of locomotive is established.(2)Aiming at the time-varying parameters identification problem during locomotive operation and the parameters identification problem under noise interference,the SVD-UKF algorithm is proposed.The Singular Value Decomposition(SVD)is used to improve Unscented Kalman Filter(UKF),which is insensitive to system disturbances and has strong stability.It overcomes the limitation of easily falling into the abnormal state for matrix solution and realizes the real-time filtering estimation of the adhesion coefficient.(3)For the problem to optimize parameters of the filter and the adhesion controller,an improved pigeon-inspired optimization(IPIO)is proposed.The beetle antennae algorithm is introduced into the compass operator,which has better global search ability,and the chaotic disturbance strategy is introduced into the landmark operator to improve the local search accuracy.The standard test function is used to compare and analyze the search speed and search accuracy of the IPIO,PIO and PSO.The analysis shows that the improved pigeon-inspired optimization(IPIO)has better optimization performance,which verifies the correctness and effectiveness of the improved scheme.(4)In view of the unmodeled dynamics and robustness problem in the controller design process,three kinds of model-free adaptive controllers that only need I/O data of the controlled system are designed.Based on Model-Free Adaptive Control(MFAC)method,the simplified locomotive wheelset dynamics model is taken as the controlled object,and three-form modelfree adaptive controllers of locomotive adhesion are designed,which is compact format,partial format and full forma respectively.Besides,the optimal adhesion control strategy for locomotives containing 5 modules is also proposed.(5)Combined with the fitness functions constructed,the IPIO is used to optimize the parameters of the SVD-UKF and the adhesion controllers,and the optimal adhesion control strategy of the locomotive is simulated and verified.The analysis results show that the threeform adhesion controllers designed have good and stable control performance.Meanwhile,the IPIO has good optimizing ability,under the guidance of the fitness functions constructed,which optimizes the parameters of the filter and the adhesion controller.This can effectively avoid the influence of subjective experience on the overall control effect.