A Weak Signal Detection Method Based on Stochastic Resonances and Its Application to the Fault Diagnosis of Critical Mechanical Components

Abstract

This chapter presents studies on the enhanced detection of characteristic signals from critical mechanical components such as bearings by the nonlinear effect of stochastic resonance (SR). In the past decades, classical stochastic resonance (CSR) method has been extensively studied to enhance the fault detection of these critical mechanical components such as bearings and gears. Based on CSR theories, the main content of this chapter includes two parts. The first is aiming at identifying the component characteristic frequencies in the spectra, SR normalized scale transform is proposed based on parameter-tuning bistable SR model, which leading to a new method via averaged stochastic resonance (ASR) to enhance the result of incipient fault detection. Then, rather than achieving the improvement of the signal-to-noise ratio (SNR) by increasing the noise intensity, a new approach is developed based on adding a harmonic excitation with a frequency based on the system’s Melnikov scale factor to the system while the noise is left unchanged. The effectiveness of this method is confirmed and replicated by numerical simulations. Combined with the strategy of the scale transform, the method can be used to detect weak periodic signal with arbitrary frequency buried in the heavy noise. In addition, the chapter also presents the case study of applying these methods for the enhancement of fault characteristic signals in detecting incipient faults of roller bearings.