摘要
针对大型机械设备运行环境恶劣、故障特征难以提取的问题,提出一种基于分段线性双稳态势函数模型的非饱和随机共振方法。该方法通过分段线性化的势函数代替经典的双稳态势函数,然后利用频移尺度变换实现机械设备振动信号的小参数化,使其满足随机共振系统的输入条件;最后,将系统输出信号的信噪比作为遗传算法的目标函数,优化非饱和随机共振系统参数,实现机械设备早期故障特征的增强与提取。仿真分析表明该方法可以有效地提取淹没在强噪声背景下的微弱故障冲击特征,而且轴承实验证明提出的方法能够有效增强与提取故障特征频率。仿真与实验结果进一步表明提出的方法优于经典双稳态随机共振方法,不仅能够获得高的输出信噪比,而且在特征频率处具有更高的幅值。以上优点归功于提出方法不仅克服经典双稳态随机共振系统的内在输出饱和问题,而且利用遗传算法实现系统与输入信号之间的最佳匹配。
An unsaturated stochastic resonance method based on piecewise linear bistable potentials function model is proposed aiming at the problem that fault signature of large-scale mechanical equipment is hard to be extracted in hostile operation environment. Initially, the method replaces classical bistable potentials function with piecewise bistable potentials function. Second, frequency- shifted scale transform was used to transfer the vibration signals of machinery into small parameters for satisfying the input conditions of stochastic resonance. Finally, the system parameters of unsaturated stochastic resonance were optimized by using SNR of system output signal as objective function of genetic algorithms to strengthen and extract incipient fault of mechanical equipment. Simulation analysis shows that the proposed method was able to effectively extract weak fault impact characteristics overwhelmed by strong noise background, and the bearing test proves that it can be used to strengthen and extract the fault characteristics frequency. In addition, the simulation and test results further indicate that the proposed method is superior to classical bistable potentialsstochastic resonance method and it can not only obtain higher input signal-to-noise ratio but also possess higher amplitudes at the fault characteristic frequencies of bearings. The proposed method can not only overcome the inherent output saturation problem of classical bistable potentials stochastic resonance system but also achieve moreover can achieve the optimal matching between the system and input signals based on genetic algorithm.
出处
《中国测试》
北大核心
2017年第8期106-112,共7页
China Measurement & Test
关键词
非饱和随机共振
故障特征提取
信噪比
早期故障诊断
unsaturated stochastic resonance
fault characteristic extraction
signal-to-noise ratio
incipient fault diagnosis
