自适应随机共振形态学在液压泵振动信号特征提取中的应用

Application of Adaptive Stochastic Resonance Morphology in Hydraulic Pump Vibration Signal Feature Extraction

液压泵振动信号常湮没在强噪声背景中,为准确提取其特征频率,提出自适应随机共振形态学方法。首先采用以广义相关系数为目标函数的量子遗传算法对随机共振系统参数进行优化,再将优化后的参数代入随机共振系统对液压泵振动信号进行降噪预处理,最后利用形态学差值滤波器提取振动信号的特征频率。仿真实验和液压泵故障模拟实验结果表明,该方法能够准确地提取出振动信号的各种频率特征,优于其他特征提取方法。

In order to extract the frequency characteristic of hydraulic pump vibration signal which is in high noise background, adaptive stochastic resonance（ASR） was proposed. Firstly, quantum genetic algorithm（QGA） was used to optimize the stochastic resonance system＇s parameters, meanwhile the genetic correlation function （GCF） was used as the object function of QGA. Then, make the optimization parameters into the stochastic resonance system for hydraulic pump vibration signal pretreatment. At last, the morphologic filter was used to extract the characteristic frequencies from vibration signal. Through numerical simulations and hydraulic pump fault simulation tests, the proposed method can be used to clearly extract various characteristic frequencies of vibration signals and it is superior to the others.