Separation of the vibration-induced signal of oil debris for vibration monitoring

Oil debris sensors are designed for monitoring machine component conditions by detecting oil debris in the circulating oil lines. However, these sensors are not only sensitive to metallic particles, but are susceptible to machinery vibration as well. The vibration-induced signal has thus far been treated as interference and is accordingly removed to better reveal the particle signature. As the vibration signal also contains important information on machine health, which can be used to detect not only the machine component faults but also machine structural malfunctions, we propose a joint integral and wavelet transform approach to separate the vibration and particle signals to make the oil debris sensor multi-functional. The recovered vibration signal is then used to detect faults that cannot be revealed by examining oil debris content. Our experimental results have shown that the separated vibration signal is, in general, consistent with the vibration velocity and hence can be used as an auxiliary vibration monitoring tool.