Fault Diagnosis Method for Axial Piston Pump Slipper Wear Based on Symmetric Dot Pattern and Multi-Channel Densely...

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Fault diagnosis in axial piston pumps is critical for hydraulic system reliability.
• Slipper wear presents a diagnostic challenge due to subtle fault variations.
• Improving diagnosis accuracy enhances operational reliability and offers insights for similar rotating machinery issues.

Purpose of the Study:

• To propose an effective fault diagnosis method for axial piston pump slipper wear.
• To enhance the accuracy and reliability of diagnosing different forms and degrees of slipper wear.
• To provide a novel approach for fault diagnosis in rotating machinery.

Main Methods:

• Symmetric Dot Pattern (SDP) transformation to convert vibration signals into images, fusing triaxial data.
• Multi-channel densely connected convolutional networks (DenseNet) for feature extraction.
• Improvements to the inception module, incorporating CBAM and DropBlock for enhanced performance.

Main Results:

• The proposed SDP and DenseNet method achieved higher diagnostic indices compared to traditional convolutional neural networks.
• The method demonstrated effectiveness in fusing triaxial vibration signals and enriching signal features.
• Experimental results validated the superiority of the proposed approach on slipper wear fault datasets.

Conclusions:

• The developed method offers a superior and effective solution for diagnosing slipper wear in axial piston pumps.
• The approach provides a robust framework for fault diagnosis in complex rotating machinery.
• This work contributes to improving the reliability of hydraulic systems through advanced fault detection techniques.