Restrain end-effect in empirical mode decomposition by mirror extension and radial basis function neural network prediction

	Restrain end-effect in empirical mode decomposition by mirror extension and radial basis function neural network prediction
	Han, Jianping1 ; Qian, Jiongn 2
	2010
会议名称	12th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures
会议录名称	Proceedings of the 12th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures
卷号	1
页码	375-381
会议日期	April 23, 2010 - April 25, 2010
会议地点	Yantai, China
出版者	CI-Premier Pte Ltd
摘要	Hilbert-Huang Transform (HHT) is a new powerful signal processing approach, especially for nonlinear and non-stationary signal. HHT consists of Empirical Mode Decomposition (EMD) and Hilbert Transform (HT) and EMD is the crucial step. However, there is a troublesome end-effect issue to apply spline interpolation to get the upper and lower envelopes. Based on discussion of cause and influence of end-effect, mirror extension and prediction via radial basis function (RBF) neural network are investigated to restrain end-effect in EMD. Then a simulated signal and a recorded acceleration signal from shaking table test of a 12-stroey reinforced concrete frame model are preprocessed by these methods and decomposed by EMD respectively. The results indicate that the proposed methods can restrain end-effect to some extent, but still can not completely eliminate end-effect in EMD. In addition, for complex signal such as acceleration record of a structure, only extending signal by RBF neural network is not effective to restrain end-effect in EMD. But applying mirror extension after predicting the filtered signal using RBF neural network can restrain end-effect in EMD effectively.
关键词	Forecasting Interpolation Mathematical transformations Mirrors Radial basis function networks Reinforced concrete Acceleration signals Empirical Mode Decomposition End effects Hilbert Huang transforms Nonlinear and non-stationary signals Radial basis function neural networks Reinforced concrete frames Spline interpolation
收录类别	EI
语种	英语
EI入藏号	20173604122913
EI主题词	Signal processing
来源库	Compendex
分类代码	412 Concrete - 716.1 Information Theory and Signal Processing - 741.3 Optical Devices and Systems - 921.3 Mathematical Transformations - 921.6 Numerical Methods
文献类型	会议论文
条目标识符	https://ir.lut.edu.cn/handle/2XXMBERH/116836
专题	教务处（创新创业学院）
作者单位	1.Lanzhou University of Technology, China; 2.Planning and Construction Bureau of Anji County, China
第一作者单位	兰州理工大学
推荐引用方式 GB/T 7714	Han, Jianping,Qian, Jiongn. Restrain end-effect in empirical mode decomposition by mirror extension and radial basis function neural network prediction[C]:CI-Premier Pte Ltd,2010:375-381.