Abstract: In order to solve the problem that the vibration information in rotating machinery fault diagnosis was susceptible to noise pollution and it was difficult to extract the weak component of early fault, the commonly used weak signal extraction methods were studied, and a method of bearing weak fault diagnosis based on timedelay underdamped tristable stochastic resonance was proposed. Firstly, the tri-stable potential function and its shape variation characteristics were discussed, and the time-delay parameters, feedback parameters and damping parameters were introduced. The equivalent time-delay potential well function and signal-to-noise ratio were obtained by combining the Fokker-Planck equation, which proved the stochastic resonance phenomenon of the timedelay underdamped tri-stable system. Then, to improve the ability of weak fault information extraction, the output signal-to-noise ratio was used as the evaluation standard to determine the parameter optimization effect of the adaptive algorithm. Finally, the output performance of the timedelay underdamped tristable stochastic resonance method was evaluated through simulation analysis and experimental verification. The results show that the method performs better in filtering high frequency noise when the signal frequency is 85Hz and 162.7Hz, which is far better than that of the classical time-delay stochastic resonance method and the classical underdamped stochastic resonance method, and has higher applicability and reliability. This method can effectively detect weak fault signals.

Key words:  bearing weak fault diagnosis; time-delay underdamping; tri-stability; stochastic resonance; signal-to-noise ratio

LI Zhi-xing, WANG Guang-jin, BAO Hui-ru.Bearing weak fault diagnosis based on time-delay underdamped stochastic resonance[J].Journal of Mechanical & Electrical Engineering,2021,38（11）：1378-1386.