Important structures will vibrate inelastically on the seismic demand of extreme level and then their dynamic time history analyses should be performed under specified base excitation For the purpose of seismic design design spectra are usually proposed where their forms are constituted based on many free-field records Hence ground accelerograms compatible with design spectra are generated only artificially No matter what artificial accelerograms are generated in the time or the frequency domain the results of convergence is reached as long as the maximum error compared with a target response spectrum is less than a tolerance without the consideration of PGA matching wave form maintenance and baseline drift prevention At first this study develops a new algorithm theoretically to generate artificial earthquakes The algorithm contains: (1) the use of a high-order Gaussian wavelet as the base function (2) adjusting the response displacement time history corresponding to a single period directly then obtaining the ground accelerogram by differential (3) adding several low-order Gaussian wavelets to satisfy some demands due to the finite duration of ground accelerograms and (4) suggesting the optimal demand setting Then the new algorithm are compared and verified through various simulation study and practical cases on generating artificial earthquakes Those comparisons include: (1) using smooth or piecewise-defined target response spectra general or near-fault site design spectra (2) covering 23 to 1000 various matching periods from an arithmetic or a geometric sequence (3) selecting different base functions and time histories under modification and (4) reducing tolerance of matching On the basis of theoretical formulation and practical verification the new algorithm which uses compound Gaussian wavelets to directly adjust the response deformation at one period at a time can achieve convergence in all cases Therefore our new algorithm of generating artificial earthquakes is worthy of practical application
Date of Award | 2020 |
---|
Original language | English |
---|
Supervisor | Li-Ling Hong (Supervisor) |
---|
A New Algorithm of Generating Ground Accelerograms by Compound Gaussian Wavelets
昱喬, 黃. (Author). 2020
Student thesis: Doctoral Thesis