Study On Optimal Adhesion Control Of High-speed Train Based On Adhesion Slip Characteristics

In recent years,with the sustaining acceleration and optimization of high-speed trains,there is no doubt that the requirement for adhesion performance of trains is gradually increasing.In the process of train running,the degree of adhesion between wheel and rail is very sensitive to the surface contact of wheel-rail,which will be affected by many factors.If the train has a slip/slide phenomenon without adhesion control in time,it can lead to poor wheel-rail contact,wheel-rail heating and abrasion,and even serious accidents such as derailment,which will endanger the safety of the entire people.Therefore,it is of great practical significance to study the problem of wheel-rail adhesion of high-speed trains.This paper mainly aims at studing the adhesion control method of high-speed train,and the ultimate goal is to improve the adhesion utilization ratio.The optimal adhesion control strategy based on the adhesion slip characteristics of high-speed train wheel-rail is explored.The main contents of research are as follows:Firstly,starting from the traction mechanism of high-speed train and the adhesion mechanism between wheel and rail,this paper analyzes the essence of adhesion force between wheel and rail,the creep phenomenon and adhesion characteristics of wheel-rail.Besides,the main factors affecting the adhesion coefficient are also introduced;Secondly,the dynamic equation of train traction system is obtained through analysis,and the CRH2 A train is taken as an example to build its uniaxial dynamic model in Matlab/simulink environment,which is driven by 10 N level traction force,and compared with the actual operating characteristics.The validity of the model is verified,which lays a foundation for subsequent research;Thirdly,aiming at the problem that the adhesion coefficient is difficult to obtain in real time,a full-dimensional state observer is constructed,and the poles are properly configured to estimate the load torque more accurately.the wheel-rail adhesion coefficient is finally estimated indirectly.The simulation results show that the observer can effectively solve the problem of sensitivity of common observer to the external disturbance and improve the observation accuracy and has better stability;Fourthly,aiming at the problem that the slope of adhesion slip curve is difficult to be calculated,this paper introduces the Recursive Least Squares(RLS)with forgetting factor to estimate the slope of adhesion slip characteristic curve without direct derivation.Besides,the gradient descent method is used to estimate the optimal reference creep velocity,aiming at the problem that the traditional fixed step gradient method has a slower estimation speed near the extremum due to the smaller slope,so the sigmoid function will be introduced into the step length extremum search.At the same time,the search step is set as a function related to theslope.Furthermore PID contronl method is combined to the real-time fast automatically seek the adhesion peak point under the current rail surface condition,the simulation results verify the effectiveness of this method when the wheel-rail surface changes suddenly and the accuracy in the process of searching the extremum,and it is better than the traditional gradient descent search algorithm.Fifthly,in order to solve the problem that traditional PID controller’s parameters are difficult to set,the RBF neural network control algorithm is added to adjust the PID controller’s parameters and optimize its control performance.The comparison between the RBF neural network PID control and traditional PID adhesion control methods shows that the optimal adhesion control is feasible and effective and satisfy the purpose of preventing idling.At the same time,it is verified that this method is superior to the traditional PID adhesion control,which can make the high-speed train achieve the optimal adhesion quickly and obtain the higher adhesion utilization ratio,and its average adhesion utilization ratio is about 97%.Meanwhile,it can respond to different road conditions in time with strong stability and basically no overshoot.