Research On Real-time Reliability Of On-board Equipment Of CTCS-3 Train Control System Based On State Space

The on-board equipment of train control system is essential to ensure the safe and efficient operation of high speed train,which shall be provided with high reliability and high safety.The researches on traditional reliability are based on the statistical law of the same kinds of products,with a certain kind of products as the object of study,while the object of study for the real-time reliability evaluation is the specific and unique equipment.Compared with the traditional reliability evaluation,the real-time reliability of on-board equipment can reflect the ability of on-board equipment during train operation.The real-time reliability evaluation of on-board equipment is conducive to establishing an intelligent condition based maintenance system for the purpose of guaranteeing continuous train operation.According to the inherent characteristics of on-board equipment,a set of methods for evaluation of real-time reliability of on-board equipment based on state space are designed in this thesis.Firstly according to the thought of hierarchical analysis,the real-time evaluation of on-board equipment is divided into two parts:namely system level evaluation and component level evaluation.Secondly FTA,ARIMA and HMM are used to build the real-time reliability model of on-board equipment,where ARIMA is the special expression of state space model,being suitable for the situation where the residual series do not follow normal distribution and being able to predict the performance degradation trend of on-board equipment,and HMM is a Markov chain based on state space,used here to calculate the real-time reliability of on-board equipment.Finally,the software for real-time reliability evaluation of on-board equipment is designed and implemented.The main works of this thesis are as follows:(1)Based on analyzing the structure of the on-board equipment,combined with traditional reliability definition and practical situation,the real-time reliability of on-board equipment is defined,and the indicator system of real-time reliability evaluation is established.(2)The evaluation process of real-time reliability of on-board equipment is realized at the system level and the component level.At the system level,the FTA model is built to calculate the reliability of the on-board equipment;and at the component level the performance degradation indices of on-board equipment component are established by the method of FMEA.By using ARIMA-ARCH model,the degradation path of the component is predicated and the degradation degree of the component is calculated.Then the health state of the component is graded and the improved HMM with multiple observations is used to reveal the law of state transition of the component.Finally real-time reliability of on-board equipment is obtained by state transition matrix and C-K equations.(3)The calculation of the real-time reliability of Balise Transmission Module(BTM)is given as an example.The Lloyd algorithm is used to realize vector quantization of characteristic parameters.The state transition model for performance degradation is built to predict the degradation law of BTM,and the result of real-time reliability of BTM can be obtained by using HMM training.In this method,the performance degradation data is used directly to realize the accurate perception of the hidden state transition process of on-board component.In this way,the distribution assumption of the trajectory parameters can be avoided and a better description of the real-time reliability of on-board component can be achieved.(4)The software for real-time evaluation of on-board equipment of train control system is designed and implemented in this thesis,by adopting mixed programming of Matlab and C#.