Stochastic resonance systems have the advantages of converting noise energy into signal energy, and have great potential in the field of signal detection and extraction. Aiming at the problems of the performance of classical stochastic resonance system whose model is not perfect enough and the correlation coefficients between parameters is too large to be optimized by algorithm, then a novel model of the tristable potential stochastic resonance system is proposed. The output SNR formula of the model is derived and analyzed, and the influence of its parameters on the model is clarified. Compared with the piecewise linear model by numerical simulation, the correctness of the formula and the superiority of the model are verified. Finally, the model and the classical tristable model are applied to bearing fault detection in which the genetic algorithm is used to optimize the parameters of the two systems. The results show that the model has better detection effects, which prove that the model has a strong potential in the field of signal detection.

随机共振系统具有将噪声能量转换成信号能量的优点，并且在信号检测和提取领域具有很大的潜力。针对经典随机共振系统性能不够理想，参数之间的相关系数过大而无法通过算法优化的问题，提出了一种新型的三稳态电位随机共振系统模型。推导并分析了模型的输出信噪比公式，并阐明了其参数对模型的影响。通过数值模拟与分段线性模型比较，验证了公式的正确性和模型的优越性。最后，该模型和经典的三稳态模型被应用到轴承故障检测中，其中遗传算法被用来优化两个系统的参数。结果表明，该模型具有较好的检测效果，证明该模型在信号检测领域具有强大的潜力。