Gaussian process regression for response spectrum
Abstract/Contents
- Abstract
- Design spectra are an important tool to determine peak structural response to seismic loads. A common approach to obtain a design spectrum is to approximate the mean response spectrum to a set of ground motion records by linear segments. The idea behind such method is to obtain a curve that is smoother than the response spectrum and, therefore, can be a better generalization to new ground motions that might occur. We propose the use of Gaussian Process regression as an alternative method to obtain such curves. Once model parameters have been defined, Gaussian Process regression is simple to be implemented to any ground motion record and can be a more natural approximation to the mean response spectrum. In addition, regression results take into account the uncertainties on the process and allow for prediction confidence regions to be readily obtained. The model proposed was able to generate smooth curves that can be appropriate to be used as design spectra.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Rosa dos Santos Cordeiro, Leticia |
|---|---|
| Associated with | Stanford University, Civil & Environmental Engineering Department. |
| Advisor | Law, K. H. (Kincho H.) |
| Thesis advisor | Law, K. H. (Kincho H.) |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Leticia Rosa dos Santos Cordeiro. |
|---|---|
| Note | Submitted to the Department of Civil and Environmental Engineering. |
| Thesis | Thesis (Engineering)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by Leticia Rosa dos Santos Cordeiro
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...