Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE TWELFTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING
Edited by: B.H.V. Topping, L.F. Costa Neves and R.C. Barros
On Fling and Baseline Correction using Quadrature Mirror Filters
A.A. Chanerley1, N.A. Alexander2 and B. Halldorsson3
1University of East London, United Kingdom
A.A. Chanerley, N.A. Alexander, B. Halldorsson, "On Fling and Baseline Correction using Quadrature Mirror Filters", in B.H.V. Topping, L.F. Costa Neves, R.C. Barros, (Editors), "Proceedings of the Twelfth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Stirlingshire, UK, Paper 177, 2009. doi:10.4203/ccp.91.177
Keywords: fling, baseline, correction, wavelet, strong motion, filters, denoise, displacement, acceleration, Chi-Chi, Selfoss, frequency, integration, sub-band, filter banks.
This paper presents a new and alternative method of estimating displacements from seismic acceleration records. The method uses biorthogonal and, or debauchies wavelets in order to denoise the noisy accelerometer data. Wavelets use quadrature mirror filters, these are maximally flat filters and in the case of the biorthogonal filters have linear phase. These are useful properties and when used with a denoising scheme these filters are able to extract the long period fling from the record. The net result is that integration to velocity and to displacement is made easier. Furthermore, the method also enables automatic baseline correction, without having to manually locate suitable time points. The wavelet transform can be considered as a set of filter banks, which downsample by 2 and successively decompose (filter) the data into low frequency and a higher frequency sub-bands, each time halving the sub-band to lower freqeuncies. This enables the low-frequency fling sub-band to be isolated and integrated with relative ease. At the same time a de-noising procedure is applied, which removes the noise by applying a threshold scheme to the data. The net result is that the method extracts both the low and higher frequency sub-band, which are then easily integrated by automatically applying a small correction by finding a timing point at which instrument tilt and, or noise shifts the baseline from zero. The records used in this paper are from the strong motion Chi-Chi 1999 event of magnitude 7.3 and the 6.3 magnitude event at Selfoss in Iceland on May 29th 2008.
purchase the full-text of this paper (price £20)