An adaptive stochastic resonance method based on grey wolf optimizer algorithm and its application to machinery fault...

Area of Science:

• Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Stochastic resonance (SR) is a valuable technique for enhancing signal detection in machinery fault diagnosis.
• Conventional SR methods struggle with parameter optimization, limiting their effectiveness.
• System parameter sensitivity hinders satisfactory analysis results in traditional SR applications.

Purpose of the Study:

• To develop an adaptive stochastic resonance (SR) method for machinery fault diagnosis.
• To improve the performance and accuracy of SR in identifying faults.
• To overcome the limitations of fixed-parameter SR methods.

Main Methods:

• An adaptive SR method is proposed, utilizing the grey wolf optimizer (GWO) algorithm.
• The GWO algorithm optimizes SR system parameters based on a redefined signal-to-noise ratio (SNR).
• Optimized parameters enable adaptive SR output matching the input signal characteristics.

Main Results:

• The proposed GWO-based adaptive SR method demonstrated superior accuracy compared to fixed-parameter SR, GA-SR, and fast kurtogram.
• Validation was performed on simulated signals, a rolling element bearing test bench, and gear fault diagnosis.
• The method successfully achieved adaptive SR output tailored to input signals.

Conclusions:

• The adaptive SR method based on GWO significantly enhances machinery fault diagnosis accuracy.
• This approach offers a robust solution for overcoming parameter sensitivity issues in SR.
• The method shows considerable practical value and potential for engineering applications.