Presentations

Abstract

This research analyses the effectiveness of a vision-based condition monitoring (VCM) framework designed to assess gear health in gearboxes. The framework addresses the limitations of traditional condition monitoring (CM) techniques, such as vibration, oil, and temperature monitoring, which often miss key defect details and are affected by external factors such as weather conditions. In this study, the effectiveness of VCM is investigated by integrating a camera system into a back-to-back gear tester for helical gears. The system captures images at regular intervals during testing, synchronising them with accelerometer data to allow comparison with vibration monitoring. The VCM uses an auxiliary system to manage challenges like lubricant splash, ensuring high-quality image capture even during gearbox operation. The analysis of the VCM focuses on three levels: qualitative (detecting different damage types), quantitative (measuring damage metrics), and comparative (assessing benefits over vibration monitoring). The results show that VCM can detect gear damage earlier than traditional vibration-based methods. It can distinguish the type of gear damage and provide information on the location and size of the damage. Furthermore, it is less influenced by external factors such as changes in temperature, wind speed and wind direction. Finally, in some cases, it can offer insights into the root cause behind the occurring damage. This shows the potential of a VCM to provide an alternative for the real-time monitoring of gears, which is faster, more detailed, and more robust than vibration monitoring.