Detection of Gear Damage in Motion - Highspeed Recording and Machine Learning for Damage Progression Analysis and Condition Monitoring (HighspeedGearVision)

• Applicant: Professor Dr.-Ing. Karsten Stahl
• Subject Area: Engineering Design, Machine Elements, Product Development
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 532790251

Gear transmissions are a fundamental part of many drive systems. An unexpected failure therefore usually comes with high risks for people, machines and costs. For this reason, precise knowledge of possible damage, its occurrence and progression, is important. The information serves both in the design for optimal dimensioning, as well as in the operation for condition monitoring by correlation with sensor signals. However, today's models and analyses lack data with high-resolution and continuous damage information, especially at rapid damage progressions. This means that damage is often only recorded quantitatively at the beginning and end, sometimes with a few interruptions, of gear tests or field runs. Continuous damage information over the entire service life, so-called run-to-failure data, is difficult to obtain and is therefore not available for research. The present project is intended to research an innovative approach for continuously recording precisely this temporal and local damage progression of surface damage during operation. Due to the sometimes very high speeds, a high-speed image recording technology will be integrated into the test rig and measurement data acquisition and quantifiable damage information will be continuously derived by means of automatic image recognition and machine learning algorithms. On the one hand, this makes it possible to gain fundamental knowledge about damage occurrence and its progression, which is important for detailing and validating mechanisms of action and damage models. On the other hand, necessary damage information for the research of precise condition monitoring algorithms can be derived.

• DFG Programme: Research Grants