Bearing fault detection with the use of acoustic signals recorded by a hand-held mobile phone

It is well know that bearings are the most often failing elements of electric motors. Percentage of bearing failures vs. other failures is even more significant for the case of low voltage motors. At the same time diagnostics of rolling element bearings constitutes a well-established part of the rotating machinery condition monitoring domain. In many cases, however, the cost of installing a high-end accelerometer-based bearing condition monitoring system, which is currently the most common approach in the industry, might be difficult to justify on non-critical machinery due to potentially long payback period of the investment. Therefore, still the very first diagnostic is often performed by human ear and the assessment of the nature of the emitted sound. Considering an example of electric motors, it is typically the abnormal or excessive sound that is first recognized by the operators as the indicator of faulty motor operating condition. This text investigates the possibility of performing condition monitoring of rolling element bearings based on acoustic signals recorded by a standard and easily accessible mobile phone. The main difficulty in using mobile phone-embedded microphone for rotating machinery diagnostic purposes is the fact that the frequency response of the mobile phone microphone is very poor (below 200 Hz) where typically the bearing fault frequencies lie. The results seem to indicate that given an appropriate signal processing approach is taken, it is possible to indicate the presence of faults in the bearings.