Skip to main content
Fig. 9 | Advanced Modeling and Simulation in Engineering Sciences

Fig. 9

From: Solving multiphysics-based inverse problems with learned surrogates and constraints

Fig. 9

Permeability inversions from CO2 saturations sampled at three well locations at 6 early snapshots. The well locations are denoted by the red vertical lines. a Unconstrained inversion result based on PDE solves. b Same as a but now with FNO surrogate approximation. c Constrained inversion result based on PDE solves. d Same as c but now with FNO surrogate approximation. e–h The error of the permeability inversion results in a–d compared to the unseen ground truth shown in Fig. 6a. i The FNO prediction errors as a function of the number of iterations for b and d

Back to article page