Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE NINTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY
Edited by: B.H.V. Topping and M. Papadrakakis
Automated Baseline Correction, Fling and Displacement Estimates from the Chi-Chi Earthquake using the Wavelet Transform
A.A. Chanerley1 and N. Alexander2
1School of Computing & Technology, University of East London, United Kingdom
A.A. Chanerley, N. Alexander, "Automated Baseline Correction, Fling and Displacement Estimates from the Chi-Chi Earthquake using the Wavelet Transform", in B.H.V. Topping, M. Papadrakakis, (Editors), "Proceedings of the Ninth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 197, 2008. doi:10.4203/ccp.88.197
Keywords: correction, seismic, wavelet transform, integration, Chi-Chi, fling, decomposition, reconstruction.
This paper proposes a novel approach using a wavelet-based algorithm for the routine processing of seismic events, baseline correction and displacement. It shows that the wavelet transform extracts the acceleration 'fling' completely naturally from seismic events. It uses seismic records from the Chi-Chi event in order to extract the fling by applying wavelet filter banks for decomposition and reconstruction. The choice of wavelet is based on the shape of the long-period Fling and in this case the Haar Transform and Daubechies  and their associated filter banks seem well suited. These wavelets also give the depth of decomposition such that an optimal estimate of the fling and thus consequent displacement is obtained. Other wavelet transforms also give good results therefore the method is not transform specific. This method can automatically correct for baseline shifts in the velocity characteristics, to then obtain displacements. Moreover the process is routine and relatively straightforward to implement. The displacements are compared with GPS readings and the initial results are encouraging. In particular the acceleration and velocity 'fling' are manifest as the transform runs through the decomposition levels, which lends credibility to this routine method of seismic and baseline correction and displacement estimation. The point to make is that at a particular level of decomposition, the wavelet transform separates the long and short period acceleration sub-bands, it then becomes easier to perform a first and second integration separately on the short period and the long period in particular. This identifies acceleration and velocity 'fling' and velocity baseline offsets, which can be easily corrected.
purchase the full-text of this paper (price £20)