基于人工鱼群算法的轴承故障随机共振自适应检测方法

针对传统的自适应随机共振以单个参数为优化对象忽略参数间交互作用的不足及采用遗传算法优化参数在种群数量增加时算法收敛速度明显减缓的缺陷，提出基于人工鱼群算法的自适应随机共振新方法。该方法利用人工鱼群算法对初值、参数设定容许范围较大、具备并行处理能力及人工鱼个体数目增加时鱼群算法收敛速度能提高的特性，自适应实现与输入信号最佳匹配的随机共振系统。仿真数据与轴承滚动体故障数据分析表明，基于该算法的自适应随机共振方法可有效实现微弱特征检测与早期故障诊断。

Bearing fault detection method with adaptive stochastic resonancebased on artificial fish swarm algorithm

Based on analyzing the disadvantages of traditional adaptive stochastic resonance, for example, optimizing only one parameter which ignores the interaction between parameters and the convergence speed of genetic algorithm slowing down with the increasing population, a new adaptive stochastic resonance method is proposed. With a wide range of initial value and parameter setting, the proposed method adaptively realizes the optimal stochastic resonance system matching input signals, utilizing the ability of parallel processing and the characteristics that the algorithm has faster convergence speed with the increasing of the number of artificial fish. The analysis of the simulation data and the bearing fault data shows that the new adaptive stochastic resonance method effectively realizes the weak signal detection and early fault diagnosis.