Research On The Health Management Of The Whole Life Cycle Of High-speed Railway Track Circuits In Complex Environment

Track circuit is an important basic equipment of train control system in China.Its working state not only affects the efficiency of railway transportation,but also relates to the safety of train operation.Therefore,to ensure the safe operation of the train has become the focus of the development of China’s railway transportation.China has a vast territory and has a variety of complex terrain.Some railway lines have been in a bad environment for a long time,and their working state is easily affected by the external environment.Therefore,it has important practical significance and research value to accurately and timely predict the track circuit fault and scientific health management in the complex environment.The main research object of this thesis is ZPW-2000 A track circuit which is widely used in our country.Firstly,the basic theory of fault diagnosis is elaborated,and the principles of track circuits,main equipment composition and functions are analyzed.Combined with case studies,the fault modes of track circuits in complex environments and the impact of complex environments on the state of track circuits are analyzed.A track circuit shunt state model is constructed and its effectiveness and reliability are verified.In response to issues such as resource wastage caused by manual inspections and the current poor effectiveness of fault diagnosis,a track circuit fault diagnosis method based on PSO-SOM-LVQ hybrid neural network is proposed.The process of SOM(Self-Organizing Map)neural network is optimized by PSO(Particle Swarm Optimization Algorithm)to weaken the influence of the inherent shortcomings of SOM neural network on the fault diagnosis results,The LVQ(Learning Vector Quantization)network is used as the final classification network to output the diagnosis results.The simulation results show that the hybrid neural network has higher diagnosis accuracy and improves the efficiency of track circuit fault diagnosis.On the basis of the above fault diagnosis analysis,given the high difficulty of predicting rail circuit faults and the inadequacies of traditional methods in data feature extraction and handling imbalanced fault type samples,a hidden Semi Markov model based on DBN(Deep Belief Networks)feature extraction is proposed to predict the fault of track circuit equipment.In order to solve the problem of unbalanced fault type samples in traditional data collection,SMOTE(Synthetic Minority Oversampling Technology)algorithm and linear interpolation algorithm are used to expand the fault type data with small number of samples to prepare for feature extraction and improve the efficiency of machine learning.The data features of track circuit are extracted by using deep belief network,and the data feature set is integrated.HSMM(Hidden Semi Markov model)is trained by using data features.Combined with microcomputer monitoring data,the dwell time and state transition probability are calculated by using Hidden Semi Markov model,so as to predict the residual life of track circuit in its whole life cycle,The simulation results show that compared with the original HSMM model,the accuracy of state recognition based on DBN-HSMM is improved by nearly 20%,which has higher practical application value and theoretical innovation significance.Finally,on the basis of life prediction and condition based maintenance decision,the changes of track circuit health level in various complex environments are analyzed,and different maintenance methods are adopted for different situations,so as to ensure the high efficiency of train operation.This thesis focuses on the health management of the whole life cycle of high-speed railway track circuit and studies the theoretical models and algorithms of reliability prediction,equipment fault detection and prediction,so as to control the track circuit fault status in time.Taking into account the complex environmental factors in the whole life cycle of health management,this thesis constructs the PHM diagnostics and health management(PHM)track circuit control model and optimization algorithm,which plays a guiding role in improving the fault detection,prediction means and maintenance efficiency of high-speed railway track equipment and escorts the safety of high-speed railway.