摘要
为提高船舶液舱液位测量精度,研究了液位测量信号的特性,提出基于小波去噪和经验模态分解(EMD)的水位信号处理新方法。用平移不变小波去噪算法对液位信号进行预处理,消除异常事件对EMD方法的影响,用EMD分解信号,按照给定的置信度去除高频固有模态(IMF)分量,提取低频IMF及趋势项进行重构,得到的无干扰成分即为液位真实信号。理论及实验研究表明,只有在有效消除异常干扰的情况下才能获得可靠的IMF分量,采用该方法提取有异常干扰的液水位信号真实趋势是有效可行的。
To improve the accuracy of ship hold liquid level, the nonlinear characteristics of level signals are exploited. A novel method for analyzing the water level signals based on wavelet de-noises technique and Empirical Mode Decomposition（EMD） has been developed. Wavelet transformation with translation invariance is adopted to eliminate the effect of abnormal signals, and the new signal is decomposed into Intrinsic Mode Functions（IMF） by EMD. After removing high frequencies IMFs according to the given confidence degree, the low useful frequencies IMFs are extracted to reconstruct the true level signal. The results show the proposed method is reasonable and suitable to nonlinear and non-stationary water level signal.
出处
《计算机工程》
CAS
CSCD
北大核心
2008年第18期13-16,共4页
Computer Engineering
基金
2006辽宁省博士科研启动基金资助项目(20061065)
国家高新优船舶关键技术开发专项基金资助项目"货油装卸系统自动化和智能化研究"
关键词
液位信号
小波去噪
平移不变
经验模态分解
liquid level signal
wavelet de-noise
translation invariance
Empirical Mode Decomposition（EMD）
