Enhancement of the signals collected by oil debris sensors

Oil condition data is a major source of information for machine condition monitoring. It contains information about the metallic particle content and thus reflects the level of wear and fatigue-induced damage in the mechanical system. Oil debris sensor is a popular measurement device used to collect oil condition data. This sensor generates an output signature with the passage of a metallic particle through the oil return lines. Analysis of the measured data leads to an estimate of the size and number of metallic particles present in the lubricating oil and consequently health state of the mechanical system. However, the signal measured through the oil debris sensor is severely tainted by various noises, e.g., the background noise present as well as the interferences caused by the vibrations of the structure where the sensor is mounted. These interferences affect the performance of the health assessment unit considerably. This will inevitably cause misleading maintenance decisions and hence premature machine failure as well as lost productivity. As such, this paper focuses on the enhancement of the signals acquired from oildebris sensors. This is achieved by a two stage de-noising scheme. In the first stage the adaptive line enhancement (ALE) technique is applied to remove the vibration related interferences. Following this step, the partly purified signal is further enhanced using the wavelet decomposition based denoising method to remove the background noise mainly caused by the wiring and measurement system flaws. The proposed approach has been validated using both simulated and experimental data.