Basic Predictive Model of Railway Track Circuit Failure

Railway superstructure availability is a clear precondition for presenting railway services. As public demands does not accept any interrupt in the planned railway services, superstructure failure prediction will be very important in maintenance. Among the superstructure elements, the track circuits with direct influence on track availability, has the most complex behavior to predict its failures. Recent literature has not provide any failure time estimation for general conditions of equipment specifications, maintenance and utilizations. Most of the research activity has been devoted to evaluating signaling system in a limited space of models for operational conditions. The basic predictive model is founded on the estimation of probability distribution and reliability functions over large number of real events with the validation of basic probability theorems on estimation and standard statistical test. Tehran railway station track circuits are selected as a general track circuit for the case study, to estimate its cumulative distribution function of failures by analyzing ten years experience of operation in past. The estimation is fitted on the Weibull distribution function to obtain the shape, scale and threshold parameters of the model. Experimental results indicates that the track circuit, with the mean time between failures as 2.7 days, is foreseen to operate correctly for the next 1.7, 7.5 and 11 days after starting with the probability of 50%, 10% and 5%, respectively.