2020 Rail Transit Session
Characterizing the Effect of Rail Hardness on Corrugation Formation, Grinding Cycles, and Noise
British Columbia Rapid Transit
In 2018, TransLink commissioned a SkyTrain Noise Study in response to noise concerns raised by residents. One output of the Noise Study was recommendations for next steps, including investigation of the feasibility and effectiveness of specifying harder rail steel to improve long term rail surface condition and reduce corrugation growth rates. This presentation describes the harder rail steel investigation with the objective of quantifying the noise benefits and costs for future SkyTrain rail replacements and other projects.
Historically SkyTrain has required AREMA Standard rail as a minimum. At the time of construction of the original Expo Line, this corresponded to rail hardness around 260 HB. Over time, AREMA standards have changed with newer areas of the network utilizing harder rail steels. This has enabled a comparison of maintenance requirements and long-term noise emissions from track sections with a wide range of rail hardness, up to 370 HB.
The analysis showed that recently ground track results in a consistent noise baseline regardless of the rail hardness, age, brand, or location. The influence of rail hardness on noise was identified by examining the magnitude of increase in noise over time between grinding intervals. The softest rail typically exhibits an 8-10 dB increase in noise level in the duration between grinding interventions. In many cases this significant increase in sound level occurred over a short time period, in areas that are ground as often as every three months. Areas with harder rail steel required grinding less frequently, and also exhibited a smaller range of variation in noise levels in the intervals between grinding, as low as a 5 dB noise increase over 2 years. The presentation covers the study methods, results and conclusions, in addition to a cost-benefit analysis and discussion of potential adverse effects of harder rail steels.