Prognostic for hydraulic pump based upon DCT-composite spectrum and the modified echo state network.

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Prognostic is crucial for effective condition-based maintenance (CBM).
• Accurate prediction of equipment degradation enhances operational reliability and reduces downtime.
• Existing prognostic methods may lack the precision required for complex machinery like hydraulic pumps.

Purpose of the Study:

• To propose a novel prognostic method for hydraulic pumps to improve prediction performance.
• To introduce a DCT-composite spectrum (DCS) fusion algorithm for multi-channel vibration signals.
• To develop a modified Echo State Network (ESN) model for enhanced prognostic accuracy.

Main Methods:

• Fusion of multi-channel vibration signals using the proposed DCT-composite spectrum (DCS) algorithm.
• Extraction of DCS composite spectrum entropy as a key feature for degradation.
• Establishment of a modified Echo State Network (ESN) model with updated reservoir and redefined neighboring matrix elements.

Main Results:

• The proposed DCS fusion algorithm effectively integrates multi-channel vibration data.
• The extracted DCS composite spectrum entropy serves as a sensitive indicator of hydraulic pump degradation.
• The modified ESN model demonstrated improved prediction accuracy in hydraulic pump degradation experiments.

Conclusions:

• The novel prognostic method, utilizing DCS fusion and a modified ESN, is feasible for hydraulic pump health monitoring.
• The approach offers significant potential for enhancing condition-based maintenance (CBM) strategies.
• This research contributes to more reliable and predictive maintenance of critical industrial equipment.