Performance Degradation Prediction Using LSTM with Optimized Parameters.

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Predictive maintenance of mechanical equipment relies on accurate condition monitoring.
• Degradation prediction of components like rolling bearings is crucial for operational reliability.
• Existing methods for degradation modeling have limitations in predictive accuracy.

Purpose of the Study:

• To develop an advanced method for predicting the degradation of mechanical components, specifically rolling bearings.
• To enhance the accuracy of degradation prediction using a novel LSTM-based approach.
• To optimize the LSTM model parameters for improved predictive performance.

Main Methods:

• Feature extraction from bearing vibration signals using the kernel joint approximate diagonalization of eigen-matrices (KJADE) method.
• Development of performance degradation indicators based on between-class and within-class scatter (SS).
• Optimization of LSTM model parameters using an improved particle swarm optimization (IPSO) algorithm.

Main Results:

• The proposed LSTM-IPSO method effectively identifies degradation trends in rolling bearing performance.
• Experimental results demonstrate superior predictive accuracy compared to Extreme Learning Machine (ELM) and Support Vector Regression (SVR).
• The fused features and optimized LSTM model contribute to enhanced degradation prediction capabilities.

Conclusions:

• The novel LSTM-IPSO method offers a significant improvement in predicting mechanical component degradation.
• This approach provides a more accurate and reliable tool for condition monitoring and predictive maintenance.
• The study highlights the potential of advanced AI algorithms in mechanical system health management.