Feature Extraction of Lubricating Oil Debris Signal Based on Segmentation Entropy with an Adaptive Threshold.

Area of Science:

• Tribology
• Mechanical Engineering
• Signal Processing

Background:

• Ferromagnetic debris in lubricating oil indicates mechanical equipment wear and predicts remaining useful life.
• Detection signals often contain noise, distorting weak debris features and challenging identification.

Purpose of the Study:

• To propose an effective debris signature extraction method for improved wear particle detection.
• To enhance the accuracy of debris signature identification and quantitative estimation in lubricating oil.

Main Methods:

• Developed a debris signature extraction method based on segmentation entropy with an adaptive threshold.
• Investigated five identification indicators to improve detection accuracy.

Main Results:

• The proposed algorithm effectively identifies wear particles in lubricating oil.
• The method successfully preserves crucial debris signatures amidst noise.

Conclusions:

• The segmentation entropy-based method offers a robust solution for ferromagnetic debris analysis.
• This approach enhances the reliability of condition monitoring and predictive maintenance for mechanical systems.