基于广义相关系数自适应随机共振的液压泵振动信号预处理方法

Hydraulic pump vibration signal pretreatment based on adaptive stochastic resonance with a general correlation function

针对液压泵故障振动信号信噪比低,故障特征难以提取的问题,对液压泵振动信号预处理方法进行研究。针对现有自适应随机共振优化算法及其目标函数存在的问题,将量子遗传算法(Quantum Genetic Algorithm,QGA)引入自适应随机共振中,提出一种改进的自适应随机共振的信号预处理方法。该方法以广义相关系数为目标函数,采用QGA算法对随机共振系统的结构参数进行优化,从而实现对信号的降噪预处理。仿真及实验结果表明,该方法能够有效提取强噪声背景下的液压泵振动信号频率特征,是液压泵故障特征提取及故障诊断中信号预处理的有效方法,可进一步发展至实际工程应用。

Aiming at the problem that the signal-to-noise ratio（SNR）of vibration signals of hydraulic pumps is low and that fault features are difficult to be extracted,vibration signal preprocessing methods of hydraulic pump were studied. An improved adaptive stochastic resonance（ASR）pretreatment method was proposed based on quantum genetic algorithm （QGA）.The method proposed in this paper used a general correlation function（GCF）as the object function,and QGA was the algorithm used to optimize the parameters of stochastic resonance systems to realize the pretreatment of vibration signals.Both simulation and experiments indicate that the proposed method can be used to extract the frequency character of hydraulic pump vibration signal from strong background noise,and the pretreatment is effective in fault character extraction and diagnosis of hydraulic pump vibration signals,and it can be developed to practical application in future research.