Application And Study On Railway Anti-Rush System Based On Fuzzy Neural Network

The train anti-intrusion system is the core equipment to ensure the safe operation of the train.At the beginning of its design,it only considered the mechanism of emergency braking under the train protection curve to achieve the purpose of safe stopping,so it was difficult to meet the increasingly diversified train operation requirements under the operating conditions of complex lines.Aiming at the above problems,in order to meet the requirements of train comfort and energy saving on the basis of ensuring safety and accurate stopping,this paper studies the precise modeling of train braking process,multi-factor optimization of train running curve and optimization of train speed control algorithm.1.Studied on the nonlinear modeling problem of train braking process.Based on the fuzzy neural network algorithm has the ability to model the nonlinear complex system,and combined with the description of the events of the fuzzy system and the advantages of the neural network self-learning,the discrete motion equation of the train braking is modeled and the parameters are identified by using the fuzzy neural network.;2.Studied on the multi-factor optimization of train running curve.Firstly,based on the force analysis of the single mass point of the train,the velocity-distance curve formula is obtained,and then the gaussian linear fitting is used to soften the formula.Secondly,based on the study of the characteristics of traditional particle swarm optimization(pso),an improved pso based on the multi-factor evaluation index of train operation is proposed,and the speed-distance curve is converted into speed-time optimization curve through this algorithm.3.Combined with the speed control of the train,the controller based on the traditional PID algorithm is analyzed.Although PID algorithm has the advantages of good overall control effect and simple algorithm implementation,it also has the disadvantages of poor responsiveness and curve fluctuation when multiple models are switched.At the same time,the traditional PID algorithm and BP-PID(neural network PID)are compared and analyzed.Although it has obvious advantages in multi-model switching responsiveness,it is difficult for the controlled system to meet the control requirements due to the limitation and randomness of the initial weight selection of the neural network algorithm.4.A speed controller based on PSO-BP-PID algorithm is designed.In order to solve the problem of multi-model switching response and initial weight selection,combining the advantages and disadvantages of traditional PID algorithm and BP-PID algorithm,particle swarm optimization algorithm is used to optimize the initial weight parameters of BP-PID algorithm,so as to design a speed controller based on PSO-BP-PID algorithm.In this essay,based on the test line data of China railway rolling stock corporation,the speed controller based on PSO-BP-PID algorithm is used to simulate the running process of the train anti-intrusion system in the active intervention mode,and a number of performance evaluation indexes are selected to analyze the simulation results.The simulation results show that the proposed anfis-based multi-factor PSO-BP-PID train anti-intrusion system has obvious advantages over the traditional algorithm in the simulation of multi-factor evaluation indexes and dynamic performance of the controller system.