Acoustic analysis techniques for condition monitoring of roller bearings

Condition monitoring and maintenance of bearings for inservice machinery is a priority for railway operators. Degraded bearings can lead to inefficient operation, failures and their associated downtime. Vibration and acoustic emission analysis are well-known techniques for condition monitoring of rolling elements. However, these techniques usually require each bearing to be fitted directly with condition monitoring equipment. For systems with a large number of mobile assets, such as a railway network, this is an expensive solution that may be better addressed by a fixed (line side) measurement system capable of capturing data from a number of vehicles. The most commonly used device for checking the state of bearings from the line side is the hot axle box detector, which detects only severely damaged bearings through infrared axle temperature measurement. A wayside condition monitoring technique that can also detect early stage bearing faults is highly desirable, both to train operating companies and infrastructure managers. This paper considers the use of microphones and recording equipment to acoustically detect bearing failures at an early stage. This two-part investigation first shows the development of testing equipment and algorithms for acoustic condition monitoring of rolling bearings under ideal conditions, as well as initial results. Work on development of this research into a working line side system is then presented, including real world measurements of passing trains.