Improving gearbox fault detection and diagnosis by integrating innovative gear tooth contact analysis with neural network training​

Tasks

Expected results:

1. Development of new methods for analyzing tooth contact for different types of gears.

2. Implementation of a new TCA method to determine effective modification of the tooth surface in various cases of displacement.

The implementation of the new TCA method is expected to result in optimized modification of the tooth surface, improving the contact characteristics in various displacement cases. The results, including quantitative assessment of contact pressure, transmission error and contact path, are expected to demonstrate the versatility and reliability of the method.

3. Training a neural network to detect faults.

In the context of gear contact misalignment, the expected results include the development of a robust model capable of efficiently detecting and classifying misalignments in gear systems. The neural network is expected to demonstrate increased sensitivity to changes indicative of contact misalignment, which will be a valuable tool for proactively detecting faults in gear assemblies. Performance metrics including accuracy and precision will validate the network’s ability to recognize misalignment patterns, contributing to improved reliability and maintenance of gear systems.

4. Integration of the new TCA method with neural networks.

The integration of the innovative tooth contact analysis method with neural network learning is expected to significantly enhance the ability to detect gearbox faults. Using the new TCA datasets as the basis for machine learning models is expected to improve their ability to classify gear faults. This convergence of the tooth contact analysis method with machine learning is intended to improve the reliability and accuracy of gear fault diagnosis.

– during the entire period of the project implementation, successful results will be published in at least 2 (two) articles indexed in SCIE Web of Science Q1-Q3 by impact factor or with a CiteScore percentile in Scopus of at least 65 (sixty-five);