An Automated Data Fusion-Based Gear Faults Classification Framework in Rotating Machines

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Previous studies on frequency domain fusion for rotating machine fault diagnosis were limited to rotor faults.
• Manual diagnosis approaches are impractical for multi-component machines with high-frequency data acquisition.
• Existing methods lacked proficiency in classifying non-rotor component faults, such as those in gearboxes.

Purpose of the Study:

• To develop an automated framework for enhanced fault diagnosis in industrial rotating machines.
• To extend the capabilities of frequency domain fusion techniques to include gearbox fault classification.
• To improve the robustness and efficiency of condition monitoring systems.

Main Methods:

• Feature generation using coherent composite spectrum (CCS).
• Data dimensionality reduction via principal component analysis (PCA).
• Faults classification employing artificial neural networks (ANN).

Main Results:

• The automated framework successfully classified rotor-related faults.
• The approach demonstrated effective and automatic classification of gearbox faults using experimental data.
• Visualized decision maps were generated, simplifying fault characterization.

Conclusions:

• The proposed automated framework enhances the diagnosis of rotating machinery faults, including those in gearboxes.
• This approach minimizes diagnostic downtime and reduces the need for specialized expertise.
• The study improves the robustness of condition monitoring for industrial rotating machines.