Numerical analysis and engineering application of large parameter stochastic resonance

Under the condition of large parameters, it is difficult in signal processing to use the small parameter stochastic resonance (SR) approach to detect a weak signal submerged in strong noise from the response power spectrum of a bistable system. We develop a new method, the re-scaling frequency stochastic resonance (RFSR) to solve this technical issue. In practical applications, the RFSR method requires that the ratio of sampling frequency to the signal frequency be equal to or larger than 50. The input and the output signal-to-noise ratios of the bistable system demonstrate the effectiveness of the RFSR method. Finally, two practical cases, the monitoring and diagnosis of mechanical faults and the vibration analysis of metal cutting show that the RFSR approach is suitable for detecting an early fault and extracting weak signals from strong noise. Thus, the RFSR technique has potential applications in the engineering signal processing.