Source: Dan Hawkes
Inverse modeling involves repeated evaluations of forward models,
which can be computationally prohibitive for large numerical models. To
reduce the overall computational burden of these simulations, ESD’s
George Pau, Yingqi Zhang, and Stefan Finsterle recently studied the use
of reduced order models (ROMs) as numerical surrogates. Their paper
examines an input–output relational approach based on Gaussian process
regression (GPR). Among other findings, they observe that ROMs are more
accurate than the linear lookup tables with the same number of
high-fidelity simulations. They demonstrate how the use of a GPR-based
ROM in two many-query applications—uncertainty quantification and global
sensitivity analysis—significantly reduces the total computational
effort.
To read further, go to: http://link.springer.com/article/10.1007%2Fs10596-013-9349-z#page-1
Additional related publication: http://www.sciencedirect.com/science/article/pii/S0098300413002355
Citation: Pau, G.S.H., Y. Zhang, and S. Finsterle (2013), Reduced
order models for many-query subsurface flow applications. Computational
Geosciences, DOI 10.1007/s10596-013-9349-z, 2013.
Funding: FE, NRAP, Office of Sequestration, Hydrogen, and Clean Coal Fuels