Remote condition monitoring of phase sensitive AC track circuits

Abstract

On today's railways the story is the same the world over - demands for increased reliability at reduced cost. Leaving aside the trains themselves, the only way that this can be achieved is by improving and expanding Fixed Infrastructure Asset Monitoring (FIAM). In the UK the two biggest contributors to train delays are points and track circuits. Point condition monitoring has received much attention over the past ten years but track circuit monitoring has tended to be the poor relation. This is due in part to the wide variety of types of track circuit in use, with a broad distinction between DC, AC, high voltage impulse and audio frequency types but with many variations within those categories. Not only do these require differing monitoring techniques but also bring with them the added complication of requiring different asset models for the central monitoring system. Any monitoring system must also be proven to introduce no risk to track circuit operation. In the UK, DC track circuits are widely used in areas where traction is provided from 25kV overhead electrification. A DC track circuit can readily be monitored non-invasively by placing a DC current transducer over the wire to the track circuit relay and monitoring the signal with a logger. Any 50 Hz traction supply component is filtered out and data is then sent to a central server where the resultant profile can be analysed. The clear and occupied signal levels are affected by ballast condition and the transitions from clear to occupied and vice versa give valuable information about railhead contamination, e.g. from leaf fall particularly in the UK. Where DC traction is used, another track circuit commonly encountered in the UK is the 50Hz AC type. These are also found in mainland Europe and throughout the world; often operating at other frequencies e.g. 75 or 83.3Hz. Attempts have been made to monitor AC track circuits using a similar technique to DC, by placing an AC current transducer over the wire carrying the 'control' current to the vane type relay, but the results have generally been disappointing. This is primarily due to the effects of the traction current superimposed on the control current, which has no effect at all on the vane relay but may be interpreted by the same analysis used for DC track circuits as indicative of a problem, in particular with regard to the transitions between the clear and occupied states. Because the effects vary between trains and with train direction on reversible tracks it is not possible to apply a fixed algorithm to remove these traction effects This article describes a novel technique developed by CDSRail to overcome this problem by using correlated signal retrieval within the logger to remove the traction effects, leaving a clean profile capable of analysis. Because the resultant profiles closely resemble those from DC track circuits, an added benefit is that it allows the same asset model to be used for both DC and AC 50 Hz track circuits. The development is approved by Network Rail in the UK and is the subject of a patent at the UKPO. The monitoring unit is suitable for both single rail and double rail capacitor-fed AC track circuits and is available in two forms, either integral with a CDSRail logger or standalone. The latter outputs a 4-20 mA current which can be accepted by loggers from other suppliers. Results of trials on the UK rail infrastructure have proved valid and further benefit will result from continuous monitoring of the phase angle between local and control currents. (7 pages)