Condition monitoring of pitting evolution using multiple sensing

Abstract

Pitting on surfaces is a type of rolling contact fatigue (RCF) occurs in rolling-sliding contacts operating under mixed or boundary lubrication conditions. The early detection of pitting is of great importance due to its potential detrimental effects on the performance and reliability of machinery components, such as rolling element bearings and gears. This research aims to investigate the responses of multiple sensors to the progression of pitting and achieve early detection of pitting initiation. Experiments were conducted on a TE74 twin-disc tribometer to investigate the behaviour of bearing steel discs. Mild wear and pitting fatigue were obtained with specimens of different roughness combinations. During testing, vibration, acoustic emission (AE) and electrostatic (ES) data were recorded, and post-test signal analysis was conducted in both the time domain and frequency domain. After testing, the worn surfaces were examined to determine the mechanisms responsible for specific features seen in the sensor data. The presence of pitting in the near-surface region was observed, and its development was effectively monitored using the employed sensing techniques. The stages of running-in, pitting initiation, and pitting formation were identified through the analysis of time-domain parameters and frequency spectrums. Vibration signal analysis exhibited a more prominent indication of pitting formation, whereas AE and ES methods demonstrated an ability to detect the onset of pitting at an earlier stage.