@article{Molkenthin:2015:10.1785/0120140294,
author = {Molkenthin, C and Scherbaum, F and Griewank, A and Kuehn, N and Stafford, PJ and Leovey, H},
doi = {10.1785/0120140294},
journal = {Bulletin of the Seismological Society of America},
pages = {1810--1822},
title = {Sensitivity of Probabilistic Seismic Hazard Obtained by Algorithmic Differentiation: A Feasibility Study},
url = {http://dx.doi.org/10.1785/0120140294},
volume = {105},
year = {2015}
}

Download

TY  - JOUR
AB  - Probabilistic seismic-hazard analysis (PSHA) is the current tool of thetrade used to estimate the future seismic demands at a site of interest. A modern PSHArepresents a complex framework that combines different models with numerous inputs.It is important to understand and assess the impact of these inputs on the modeloutput in a quantitative way. Sensitivity analysis is a valuable tool for quantifyingchanges of a model output as inputs are perturbed, identifying critical input parameters,and obtaining insight about the model behavior. Differential sensitivity analysisrelies on calculating first-order partial derivatives of the model output with respect toits inputs; however, obtaining the derivatives of complex models can be challenging.In this study, we show how differential sensitivity analysis of a complex frameworksuch as PSHA can be carried out using algorithmic/automatic differentiation(AD). AD has already been successfully applied for sensitivity analyses in variousdomains such as oceanography and aerodynamics. First, we demonstrate the feasibilityof the AD methodology by comparing AD-derived sensitivities with analyticallyderived sensitivities for a basic case of PSHA using a simple ground-motion predictionequation. Second, we derive sensitivities via AD for a more complex PSHA studyusing a stochastic simulation approach for the prediction of ground motions. The presentedapproach is general enough to accommodate more advanced PSHA studies ofgreater complexity.
AU  - Molkenthin,C
AU  - Scherbaum,F
AU  - Griewank,A
AU  - Kuehn,N
AU  - Stafford,PJ
AU  - Leovey,H
DO  - 10.1785/0120140294
EP  - 1822
PY  - 2015///
SN  - 0037-1106
SP  - 1810
TI  - Sensitivity of Probabilistic Seismic Hazard Obtained by Algorithmic Differentiation: A Feasibility Study
T2  - Bulletin of the Seismological Society of America
UR  - http://dx.doi.org/10.1785/0120140294
UR  - http://hdl.handle.net/10044/1/29750
VL  - 105
ER  - 

Download