Transdimensional inversion of flow data with a cascaded reversible jump algorithm on a layer-cake model

@inproceedings{herrero_transdimensional_RM2023,
 abstract = {The evaluation of georesources involves to appropriately manage the level of detail needed in geomodels for subsurface porous flow and transport problems. In contrast to classical Bayesian modeling in which permeability values are inferred in a fixed reservoir geometry, we propose to use transdimensional Monte Carlo methods on a 2D layered reservoir model. These approaches are a way to solve the inverse problem with a suitable geological parameterization when the number of geometrical parameters (here the number of layers) is unknown. For simplicity, layers are horizontal and defined by thickness and interface depth information, and constant isotropic petrophysical values. Production data (pressure and fluid saturation) are used jointly with permeability log data in the Markov chain as a sequential acceptance criterion to reduce the computational cost: The likelihood probability of well log data is computed as a fast transdimensional regression problem, and flow simulations are performed only on candidate models accepted by this regression step to further reduce the uncertainty. Numerical flow simulations are solved on a mesh conformal to discontinuities, which is locally updated at each iteration. First inversion results from a prior permeability model compare a simple Metropolis sampler with a reversible jump sampler based on the joint use of flow and well log data. It demonstrates that the transdimensional algorithm can capture the main geological discontinuities and properly quantify the uncertainty. This suggests that this approach could be extended to more complex reservoir geometries. However, it also highlights the necessity to improve the algorithm convergence since the non-linearity of the forward flow simulation can aggravate the difficulty of recovering the geological structures accurately.},
 author = {Herrero, Julien and Caumon, Guillaume and Bodin, Thomas and Zakari, Mustapha and Giraud, Jeremie},
 booktitle = {2023 {RING} meeting},
 language = {en},
 pages = {29},
 publisher = {ASGA},
 title = {Transdimensional inversion of flow data with a cascaded reversible jump algorithm on a layer-cake model},
 year = {2023}
}