Output details

Sparse calibration of subsurface flow models using nonlinear orthogonal matching pursuit and an iterative stochastic ensemble method

This paper introduces, for the first time, a subsurface model calibration method combining ensemble techniques and signal processing techniques using sparse regularization. The proposed method enables the reconstruction of the unknown subsurface permeability fields using very limited number of field observations to reduce the cost of collecting observation. Finding the unknown permeability fields, in what is called model calibration, sharpens the model predictions and assist operational decision-making. This paper is a result of a multi-disciplinary, international collaboration (UT-Austin, KAUST). This work was funded through the KAUST UT-Austin Academic Excellence Alliance Program (UT-Austin office of sponsored projects OSP-200702891, $3M).