Fault Detection of Planetary Gears Based on Signal Space Constellations.

Area of Science:

• Mechanical Engineering
• Vibration Analysis
• Condition Monitoring

Background:

• Planetary gearboxes are critical components in many industrial applications.
• Early detection of faults in gearboxes is essential for preventing catastrophic failures and minimizing downtime.
• Vibration analysis is a widely used technique for condition monitoring of rotating machinery.

Purpose of the Study:

• To propose a novel method for processing vibration signals from planetary gearboxes.
• To develop a technique for early detection of potential faults using vibration data.
• To validate the proposed method using experimental vibration signals.

Main Methods:

• A phenomenological model-based approach is employed.
• Healthy vibration signals are projected onto an orthonormal basis.
• Low-pass components representation and Gram-Schmidt's method are utilized to construct the basis.
• The processed signals are represented by scalars indicating the gear state.

Main Results:

• The proposed method effectively represents vibration signals by a set of scalars.
• These scalars provide diagnostic information about the gear's condition.
• The method demonstrated successful validation using laboratory test bench data.

Conclusions:

• The developed method offers a reliable way to process planetary gearbox vibration signals.
• It enables the diagnosis of gear health by comparing scalar values to a predefined range.
• This technique contributes to improved condition monitoring and fault detection in planetary gearboxes.