局部均值分解(Local Mean Decomposition,LMD)算法在对建筑物变形监测数据进行噪声抑制时存在端点效应及阈值设置困难问题,提出一种基于噪声辅助和奇异值分解(Singular Value Decomposition,SVD)改进的LMD噪声抑制方法,首先通过对原始...局部均值分解(Local Mean Decomposition,LMD)算法在对建筑物变形监测数据进行噪声抑制时存在端点效应及阈值设置困难问题,提出一种基于噪声辅助和奇异值分解(Singular Value Decomposition,SVD)改进的LMD噪声抑制方法,首先通过对原始变形数据加入受控高斯白噪声再进行LMD分解的方式解决传统LMD方法的端点效应问题,然后利用连续熵差值特征确定LMD分解的高频分量和低频分量分界点,其中低频分量为变形趋势信息,高频分量包含噪声分量和变形信息,最后利用SVD对高频分量进行噪声抑制,并将其与低频分量叠加得到噪声抑制后的建筑物变形信息。采用仿真数据和实际工程实例对所提方法的噪声抑制性能进行验证,结果表明相对于传统LMD方法,经验模态分解方法(Empirical Mode Decomposition,EMD),所提方法可以获得更好的噪声抑制性能,能够明显提升建筑物变形预测精度。展开更多
A set of basic deformation modes for hybrid stress finite elements are directly derived from the element displacement field. Subsequently, by employing the so-called united orthogonal conditions, a new orthogonalizati...A set of basic deformation modes for hybrid stress finite elements are directly derived from the element displacement field. Subsequently, by employing the so-called united orthogonal conditions, a new orthogonalization method is proposed. The result- ing orthogonal basic deformation modes exhibit simple and clear physical meanings. In addition, they do not involve any material parameters, and thus can be efficiently used to examine the element performance and serve as a unified tool to assess different hybrid elements. Thereafter, a convenient approach for the identification of spurious zero-energy modes is presented using the positive definiteness property of a flexibility matrix. More- over, based on the orthogonality relationship between the given initial stress modes and the orthogonal basic deformation modes, an alternative method of assumed stress modes to formulate a hybrid element free of spurious modes is discussed. It is found that the orthogonality of the basic deformation modes is the sufficient and necessary condition for the suppression of spurious zero-energy modes. Numerical examples of 2D 4-node quadrilateral elements and 3D 8-node hexahedral elements are illustrated in detail to demonstrate the efficiency of the proposed orthogonal basic deformation mode method.展开更多
文摘局部均值分解(Local Mean Decomposition,LMD)算法在对建筑物变形监测数据进行噪声抑制时存在端点效应及阈值设置困难问题,提出一种基于噪声辅助和奇异值分解(Singular Value Decomposition,SVD)改进的LMD噪声抑制方法,首先通过对原始变形数据加入受控高斯白噪声再进行LMD分解的方式解决传统LMD方法的端点效应问题,然后利用连续熵差值特征确定LMD分解的高频分量和低频分量分界点,其中低频分量为变形趋势信息,高频分量包含噪声分量和变形信息,最后利用SVD对高频分量进行噪声抑制,并将其与低频分量叠加得到噪声抑制后的建筑物变形信息。采用仿真数据和实际工程实例对所提方法的噪声抑制性能进行验证,结果表明相对于传统LMD方法,经验模态分解方法(Empirical Mode Decomposition,EMD),所提方法可以获得更好的噪声抑制性能,能够明显提升建筑物变形预测精度。
基金Project supported by the National Natural Science Foundation of China(No.10972188)the Fundamental Research Funds for the Central Universities of China(No.2010121073)the Scientific Program of Fujian Province of China(No.2007F3096)
文摘A set of basic deformation modes for hybrid stress finite elements are directly derived from the element displacement field. Subsequently, by employing the so-called united orthogonal conditions, a new orthogonalization method is proposed. The result- ing orthogonal basic deformation modes exhibit simple and clear physical meanings. In addition, they do not involve any material parameters, and thus can be efficiently used to examine the element performance and serve as a unified tool to assess different hybrid elements. Thereafter, a convenient approach for the identification of spurious zero-energy modes is presented using the positive definiteness property of a flexibility matrix. More- over, based on the orthogonality relationship between the given initial stress modes and the orthogonal basic deformation modes, an alternative method of assumed stress modes to formulate a hybrid element free of spurious modes is discussed. It is found that the orthogonality of the basic deformation modes is the sufficient and necessary condition for the suppression of spurious zero-energy modes. Numerical examples of 2D 4-node quadrilateral elements and 3D 8-node hexahedral elements are illustrated in detail to demonstrate the efficiency of the proposed orthogonal basic deformation mode method.
