Adaptive real-time fault diagnosis for track circuits

As an important component of the Train Control System, the railway track circuit is mainly used to realise the detection of the presence of trains and continuous information transmission between the ground-equipment and the on-board equipment. Generally, the track circuit has a lot of indoor and outdoor equipment and complex application environment, of which the fault detection can be quite difficult and any track circuit failure can cause significant disruption to rail services and be a safety risk. Therefore, it is of great importance and value to realise the fault identification and diagnosis of the track circuit. Current track circuits generally use a scheduled maintenance regime. This type of maintenance scheme is costly and time consuming since inspection has to be carried out on every track circuit periodically (typically every 6 weeks). However, urgent trackside maintenance carried out in the event of such failures is also costly, particularly when it has to be carried out during traffic hours. Recently, research into novel methods for increasing the operational dependability of industrial processes has become prevalent. The ability to detect certain incipient faults and/or provide diagnosis of failed track circuit, in a more 'intelligent' way, would have significant operational and economic advantages. In this paper, an adaptive realtime fault diagnosis system for railway track circuits was developed based on multi-section joint analysis and hierarchical diagnosis method. With the help of the real-time monitoring of track circuits data changes using condition monitoring equipment and the model-based diagnosis method, the system could realise adaptive real-time fault diagnosis for the track circuit in different environment. Based on the chain circuit model and the multi-section coupling equation of track circuits, a multi-section coupling mathematical model of the track circuit was established. The fault insertion simulation of the model was carried out to obtain the voltage and current data of the specific acquisition points when different carrier frequencies, different configurations and different fault locations were obtained. Based on the big data analysis and characteristic parameter extraction of the track circuit fault data, the fault hierarchical diagnosis method was established to realise the hierarchical location of the fault section, the 5 rough faulty areas and the 11 fine fault areas of the track circuit respectively. Based on the uniform transmission line theory and the impedance matching theory, the signal transmission mathematical model of railway digital signal cable was established and the cable fault characteristic parameters were deduced, to realise the accurate fault location method of cable fault position. Finally, to verify the track circuit model and to collect fault data, a lab-based track circuit test rig was constructed. According to the fault test results on the test rig, it is shown that the track circuit fault diagnosis system proposed by the paper can achieve the positioning accuracy of the fault section and the five rough faulty areas of the track circuit up to 100 percent, and the positioning accuracy of the 11 fine faulty areas up to 98 percent, and the fault positioning accuracy of the signal cable at the transmitting end is within plus or minus 500 meters, and the diagnosis time of all the fault is within 10 seconds for different carrier frequencies and different configurations.