基于分段线性非饱和随机共振的机械早期故障诊断方法研究

发布时间：2018-01-16 03:06

本文关键词：基于分段线性非饱和随机共振的机械早期故障诊断方法研究　出处：《中国测试》2017年08期 　论文类型：期刊论文

【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。
[Abstract]:Aiming at the problem that the operation environment of large mechanical equipment is bad and it is difficult to extract the fault features. An unsaturated stochastic resonance method based on piecewise linear bistable state function model is proposed, in which the classical bistable state function is replaced by piecewise linearized potential function. Then the frequency shift scale transform is used to realize the small parameterization of mechanical equipment vibration signal, which satisfies the input condition of stochastic resonance system. Finally, the SNR of the output signal is taken as the objective function of genetic algorithm to optimize the parameters of unsaturated stochastic resonance system. The simulation results show that the proposed method can effectively extract the weak fault impact features submerged in the strong noise background. The results of simulation and experiment show that the proposed method is better than the classical bistable stochastic resonance method. It can not only obtain high output SNR, but also have higher amplitude at characteristic frequency. The advantages of the proposed method are not only to overcome the intrinsic output saturation problem of classical bistable stochastic resonance system. Moreover, genetic algorithm is used to realize the optimal matching between the system and the input signal.
【作者单位】：西安工程大学;
【分类号】：TH17
【正文快照】： 0引言随着科学技术的发展,机械设备朝着高、精、尖方向发展,致使设备结构愈来愈复杂,而且其恶劣的运行环境,导致故障特征遭受传递路径影响极其微弱,通常被淹没在强背景噪声中,因此机械设备故障难以提取与诊断。于是,早期故障特征提取一直是大型机械设备故障诊断中的关键难题[1

【相似文献】