摘要
针对强噪声环境中故障声发射信号提取的难题,在最小熵解卷积MED(Minimum Entropy Deconvolution)和最大峭度解卷积MCKD(Maximum correlated kurtosis deconvolution)的基础上,提出一种基于双向循环重构滤波准则的改进MCKD降噪方法。双向重构准则能够充分利用数据中的冲击成分,具有较强的抗冲击幅值变化干扰能力,改进MCKD方法通过滤波器迭代提取有用信号达到降噪目的。通过仿真信号和轴承故障声发射实验信号的研究表明:改进MCKD方法能有效提取原始信号中的有用故障信息,其降噪性能优于MED方法和MCKD方法。
Aiming at the problem of extracting a useful acoustic emission signal of fault, an improved MCKD method was presented by two-way loop restructuring rule based on MED and MCKD method. The two-way loop restructuring rule fully used impact component of original data, and has strong anti-amplitude changing capacity. The improved MCKD method obtained de-noising signal by iterative updating filter coefficients. The result of simulation and testing data showed that the improved MCKD can effectively extract the useful fault, and the noise reduction performance was better than MED and MCKD method.
出处
《机械设计与研究》
CSCD
北大核心
2015年第1期70-73,77,共5页
Machine Design And Research
基金
国家自然科学基金资助项目(50775219)
军队科研资助项目:[2011]107
关键词
声发射
循环重构
最小熵解卷积
降噪
故障诊断
acoustic emission
loop restructuring
minimum entropy deconvolution
denoising
fault diagnosis
