Soft Sensing Research Of Basic Parameters For ZPW-2000A Jointless Track Circuit

At present,ZPW-2000A jointless track circuit is widely used in China railway system,and its abnormal basic parameters have great influence on track circuit.A low ballast resistance will lead to "red light" malfunction and a high shunt resistance will lead to "shunt malfunction",both of which are inertia incidents of track circuit.They can easily cause some serious security incidents such as forcing open of the point,derailment,side conflict,train collision and so on to influence driving safety.These inertia incidents can be avoided by the precise measurement of basic parameters which can ensure traffic safety.However,the direct measurement of basic parameters is difficult,high-cost and the normal use of the track circuit is affected.So it is necessary to use soft sensing to research.Based on soft sensing technology and the relevant data obtained during train operation,auxiliary variables such as shunt current curve and wayside data of the train are taken as the main research object.Considered the characteristics of the track circuit in our country,the influence of basic parameters for onboard and wayside data is analyzed deeply.And the soft sensing result of basic parameters in the track circuit can be measured through the genetic algorithm and D-S evidence theory.The main work of this paper is as follows:(1)The two terminal pair network model of track circuit was structured by transmission-line theory based on the basic structure and working principle of track circuit.The front and back shunt current curve and wayside data was acquired by simulation.(2)Based on the linear relationship between voltage envelope of the cab signal and track current signal,the influence relationship of basic parameters and monitoring data was acquired based on analysis.(3)Because of the difficulties of measuring shunt resistance directly,the soft sensing method based on onboard data and wayside data.In this method,with the shunt resistance chosen as the decision variable,with the minimum difference between real signal and simulate signal as the objective to form the fitness function,by genetic operations such as selection,crossover and mutation,the optimal value of each decision variable in the current state was obtained.(4)The onboard data and wayside data was combined according to D-S evidence theory to measure ballast resistance.In this method,the final results was obtained based on constructing the membership function for the multi-interval of ballast resistance,calculating the basic trust probability assignment and making decision by the D-S rules.The innovations of the thesis are as follows:(1)An analysis method which add back shunt current to onboard information was proposed.The onboard information and wayside information are combined to analyze,which can make up for the defects of past way such as inaccuracy and instability caused by single data sources.(2)A soft sensing method for shunt resistance based on multiple population genetic algorithm was proposed.(3)A soft sensing method for ballast resistance base on D-S evidence theory which is a multi-information fusion technology was proposed.