skip to main content

DOE PAGESDOE PAGES

Title: Combining multiple lower-fidelity models for emulating complex model responses for CCS environmental risk assessment

Numerical modeling is essential to support natural resource management and environmental policy-making. In the context of CO 2 geological sequestration, these models are indispensible parts of risk assessment tools. However, because of increasing complexity, modern numerical models require a great computational effort, which in some cases may be infeasible. An increasingly popular approach to overcome computational limitations is the use of surrogate models. This paper presents a new surrogate modeling approach to reduce the computational cost of running a complex, high-fidelity model. The approach is based on the simplification the high-fidelity model into computationally efficient, lower-fidelity models and on linking them with a mathematical function (linking function) that addresses the discrepancies between outputs from models with different levels of fidelity. The resulting linking function model, which can be developed with small computational effort, can be efficiently used in numerical applications where multiple runs of the original high-fidelity model are required, such as for uncertainty quantification or sensitivity analysis. The proposed approach was then applied to the development of a reduced order model for the prediction of groundwater quality impacts from CO 2 and brine leakage for the National Risk Assessment Partnership (NRAP) project.
Authors:
 [1] ;  [1] ;  [1]
  1. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States). Energy Geoscience Division
Publication Date:
Grant/Contract Number:
AC02-05CH11231
Type:
Accepted Manuscript
Journal Name:
International Journal of Greenhouse Gas Control
Additional Journal Information:
Journal Volume: 46; Journal Issue: C; Journal ID: ISSN 1750-5836
Publisher:
Elsevier
Research Org:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
Sponsoring Org:
USDOE Office of Fossil Energy (FE)
Country of Publication:
United States
Language:
English
Subject:
54 ENVIRONMENTAL SCIENCES; 97 MATHEMATICS AND COMPUTING
OSTI Identifier:
1474924
Alternate Identifier(s):
OSTI ID: 1337579