Despite advances in transit vehicle safety technology, vehicle brakes remain a common cause of accidents, injuries, roadcalls, and mechanical breakdowns. In the United States there are over 480,000 truck and bus crashes annually with almost 30% of them involving brake failure as a factor. This project explored a new approach for improving transit safety based on detecting problems with transit vehicle brakes by analyzing sounds emitted during braking events. Acoustic theory indicates that as brake linings wear the resonant frequency of the shoe or pad increases, potentially enabling the monitoring of lining wear through acoustics. To test this approach passive acoustic sensors were placed roadside at the exit of a transit bus facility for 9 months. The sensors collected almost 10,000 recordings of a bus fleet of 160 vehicles braking over a variety of conditions.

It was found that the acoustic spectra of the braking sounds change as friction linings wear, in concurrence with theory, where resonant frequencies increased as the brake linings wore thinner. During the course of this study eight vehicles with drum brakes and six with disc brakes were observed to wear below minimum thickness specifications when the frequency reached a critical threshold and the linings had to be replaced.

Because the frequencies shift over time, the rate of wear can be tracked closely and used to predict when linings will need to be replaced. The use of this information in predictive maintenance could significantly reduce the maximum time a problem could go undetected between periodic inspections.  This data would help with scheduling planned maintenance and parts inventory, and also identify sudden changes such as breakages, contamination of components, and reduce unplanned downtime due to failed inspections.  The roadside drive-through passive nature of the acoustic sensor system can provide a cost effective method of tracking brake wear across an entire fleet.

The final report is available.