Innovative mechanism design for data mining and enhanced gear misalignment detection via vibration analysis

Amirnia, Amirhossein; Mohammadi, Somaye; Behzad, Mehdi

doi:10.22064/tava.2025.2070794.1274

Innovative mechanism design for data mining and enhanced gear misalignment detection via vibration analysis

Document Type : H. Ahmadian Prize

Authors

Amirhossein Amirnia

Somaye Mohammadi

Mehdi Behzad

Condition Monitoring Center, Mechanical Engineering Department, Sharif University of Technology, Tehran, Iran

10.22064/tava.2025.2070794.1274

Abstract

With the rapid and continuous advancement of industry and considering the vital role of gearboxes in various machines and industrial systems, condition monitoring and maintenance of these systems are of great importance. One of the most common faults in industrial gearboxes is the occurrence of misalignment between meshing gears. The presence of misalignment provides favorable conditions for the development of gear and bearing defects. Given the widespread use of helical gears in most industrial gearboxes, a deeper investigation of their behavior under misalignment conditions is required. In this study, to bring simulations closer to real-world and industrial cases, experiments are conducted on an industrial gearbox operating under gear misalignment. A dedicated mechanism has been designed and fabricated to impose controlled misalignment. After data acquisition and extraction of vibration signals, a total of nine features is calculated and analyzed. The results reveal that among the extracted features, Energy Ratio and Kurtosis exhibited the highest percentage variations relative to the aligned condition. Furthermore, detailed analysis shows that these features demonstrated more significant increases in the horizontal direction and at measurement points closer to the meshing location of the misaligned gears compared to the aligned state.

Highlights

Innovative mechanism enables precise gear misalignment in an industrial gearbox.
Nine vibration features extracted using the Time Synchronous Averaging technique.
Kurtosis and Energy Ratio show the highest sensitivity to misalignment.
Feature sensitivity increases at measurement points closer to gear meshing.
Approach supports early, accurate detection of gear misalignment in real machines.

Keywords

Condition monitoring

Fault diagnosis

Gear misalignment

Vibration analysis

Feature extraction

Subjects