Experimental robotic applications to monitor railway track defects and alignments

Abstract

Railways are critical infrastructure to many nations for servicing huge numbers of riders each day. Therefore, railway companies need to tackle wear and tear issues that lead to track defects. Failure to address them can result in ridership disturbances, economic losses, and even catastrophic loss of lives. This is especially important in cities such as Singapore where railways are necessary daily commute for most of its citizens. Railway track defects can occur due to a myriad of reasons, notably contact stresses or rolling contact fatigue. Other surface and internal defects include corrosion, inclusions, transverse fissures, and wheel burns. Static and dynamic loads also deteriorate the lifespan of existing rail tracks. Settlements and misalignment of tracks occur over time due to subgrade settlements, ageing factors, and sagging of materials. In Singapore’s context, alignment of the “third-rail” is also important as it provides power to the trains. Misalignment has been known to cause problems in the past with severe consequences. Thus, structural health monitoring (SHM) techniques and equipment to check for alignment are necessary to prevent such events from happening. Most of the present studies adopt signature/ signal techniques for non-destructive testing detection of static loads, dynamic loads, settlements and misalignment using techniques such as wave propagation techniques, and changes in electromagnetic field strength in the case of ACFM. These technologies while useful, are expensive and become huge sunk costs accrued by the railway company. It drives up expenditure and makes it harder for railway companies to continue running railways running at affordable costs to commuters. This project aims to develop a low-cost system that is capable of detecting railway defects and detect changes in alignments and settlement, adopting both robotic and traditional monitoring equipment. Thus, a comprehensive view of modern and existing techniques was studied to devise a cheaper way of railway monitoring using a robotic arm.