摘要
针对强噪声背景下的液压泵故障特征提取问题,提出一种自适应级联单势阱随机共振的特征提取方法。首先验证广义相关系数可作为自适应随机共振优化算法的目标函数,然后采用量子遗传算法优化单势阱随机共振系统的结构参数,再将所得的自适应单势阱随机共振系统进行级联。该方法只需调节每一级随机共振的一个系统结构参数,优化速度快,且采用级联方式能更准确地提取液压泵故障信号的低频成分。数值仿真分析表明:该方法可有效地提取淹没在强噪声背景下的多频信号;实际测试结果证明其能有效地检测液压泵故障信号的特征频率,为液压泵故障预测和诊断奠定基础。
An adaptive cascaded single-potential well stochastic resonance method（ACSPSR） has been proposed to extract hydraulic pump fault characteristics in strong noise backgrounds. This paper first verified that general correlation function could be used as the fitness function of stochastic resonance optimization algorithm and then used quantum genetic algorithm（QGA） to optimize the parameters of single-potential well stochastic resonance（SPSR）. The last step was to cascade the SPSR. The proposed method only requires the optimization of a systematic structural parameter at each cascade of stochastic resonance. The speed of optimization is fast and by using the cascaded stochastic resonance,the low-frequency components of hydraulic pump fault signals can be more accurately extracted. Simulation data indicates that the method can effectively extract multi-frequency signals in strong noise backgrounds. Practical test results show that the ACSPSR can effectively detect the characteristic frequency of hydraulic pump fault signals, thus laying a good foundation for pump fault prediction and diagnosis.
出处
《中国测试》
CAS
北大核心
2016年第5期107-112,共6页
China Measurement & Test
基金
国家自然科学基金项目(51275524)
关键词
单势阱随机共振
级联
广义相关系数
量子遗传算法
液压泵故障特征信号
sing le-well potential stochastic resonance
cascaded
general correlation function
quantum genetic algorithm
hydraulic pump fault characteristic signal
