EMD包络线拟合算法改进及在泡沫尺寸趋势提取中的应用

发布时间：2018-01-06 07:00

本文关键词：EMD包络线拟合算法改进及在泡沫尺寸趋势提取中的应用　出处：《江南大学学报(自然科学版)》2015年06期 　论文类型：期刊论文

【摘要】：包络线拟合是EMD算法中的关键步骤,针对现有三次样条插值(CSI)法拟合包络线存在插值点冗余、端点发散、拟合松散等问题,提出了改进的EMD包络线提取算法。将插值点由传统的极值点改为特殊点,并对特殊点分段处理,删除冗余点;利用加权最小二乘多项式拟合对边界的特殊点进行三次拟合,求出端点值作为边界处的插值点抑制端点效应;用带约束的保形三次埃尔米特插值法替代三次样条插值方法,保证每个插值子区间的单调性使得拟合更紧密。定义包络线与信号的方差和以及曲线长度协同作为评判标准,判定所提算法拟合效果良好,在浮选泡沫尺寸趋势提取中的应用也验证了其有效性。
[Abstract]:Envelope fitting is a key step in EMD algorithm. For the existing cubic spline interpolation, there are some problems such as redundant interpolation points, divergence of endpoints, loose fitting and so on. An improved EMD envelope extraction algorithm is proposed. The interpolation points are changed from the traditional extremum points to special points, and the special points are segmented to delete redundant points. The special points of the boundary are fitted by the weighted least square polynomial, and the endpoint value is obtained as the interpolation point at the boundary to suppress the endpoint effect. The conformal cubic Hermitian interpolation method with constraints is used to replace the cubic spline interpolation method. Ensure the monotonicity of each interpolation subinterval to make fitting closer. Define the variance sum of envelope and signal as well as the length of curve as the evaluation criteria, and determine that the proposed algorithm fits well. Its effectiveness is also verified by its application in the trend extraction of flotation foam size.
【作者单位】：中南大学信息科学与工程学院;
【基金】：国家自然科学基金项目(61273187);国家自然科学基金创新研究群体科学基金项目(61304126,61321003) 国家“863”项目(2014AA041803)
【分类号】：TN911.7
【正文快照】： 经验模态分解法(Empirical Mode Decomposi-tion,EMD)在非平稳及非线性信号的处理及分析上具有明显优势,被广泛应用于海洋内波探测[1]及温度预测[2]、大气时间序列预测[3]、天体观测[4]、地震记录[5]、机械故障诊断[6]等方面。包络线拟合是EMD算法中非常关键的一步。目前,常用

【参考文献】