摘要
利用金属磁记忆法对煤矿系统中的钢绳芯输送带进行早期故障诊断时,发现信号中包含的噪声对诊断结果影响较大。针对此问题,提出了基于总体平均经验模态分解和改进小波半软阈值的降噪算法。首先利用总体平均经验模态分解得到若干个本征模函数,经过相关性分析后提取本征模函数的有效分量,用人工蜂群优化算法改进阈值函数,再分别对有效分量进行改进的小波半软阈值函数降噪处理,最后将信号重构。经过降噪处理后的磁记忆信号能较好地保留信号中的有用信息。仿真实验结果表明,该算法可实现噪声环境下的钢绳芯输送带应力集中区特征的有效提取,从而实现早期故障诊断。
When using metal magnetic memory method to diagnose the early faults of steel‐cord belt ,we found that the noise that comes along in the signal ,to a great extent ,affects the re‐sults of the diagnosis .Aiming at overcoming the deficiency ,we proposed a noise decomposition algorithm based on ensemble empirical mode decomposition （EEMD ） and improved semi‐soft wavelet threshold .The algorithm first decomposes the empirical mode to obtain several intrinsic mode functions ,and extracts valid components through Intrinsic Mode Function （IMF）correlation analysis .For each of the components ,the improved semi‐soft wavelet threshold function is used to reduce the noise and reconstruct the signal ,thus enhancing the ability of metal magnetic memory signals after the noise reduction to keep useful information ,which is collected from original signal in an effective manner .The results of simulation experiment show that the algorithm can effec‐tively extract thecharacteristics from steel‐cord belt stress concentration in noisy environments and thereby enable early faults to be diagnosed .
出处
《太原理工大学学报》
CAS
北大核心
2015年第5期592-597,共6页
Journal of Taiyuan University of Technology
基金
山西省特色学科资助项目:煤矿运输系统物联网安全监控关键技术开发与应用(晋教财[2012]145号)
山西省高等学校留学回国人员科研项目(晋教外[2011]63号)
关键词
总体平均经验模态分解
人工蜂群算法
小波半软阈值
金属磁记忆法
广义交叉验证
相关性
分析
ensemble empirical mode decomposition （EEMD）
artificial bee colony（ABC）
semi-soft wavelet threshold
metal magnetic memory method
generalized cross validation（GCV）
correlation analysis
