Transdimensional reservoir data inversion for a 2D geological layered model with horizon slopes and horizontal permeability trends

@inproceedings{herrero:hal-05157109,
 abstract = {We introduce a transdimensional inversion framework to quantify stratigraphic uncertainties in reservoir layered models from well data. The goal is to adaptively discover the suitable number of subsurface model parameters. For this, we build upon an existing transdimensional sampler which considers a horizontal layered parameterization and an uncertain number of layers. Our approach integrates the dip of horizons to represent thickness variations, and a horizontal permeability gradient to reflect possible lateral petrophysical trends in geological strata. The number n_L of geological layers in the model is made variable and defines the model dimensionality, so that model parameters to infer correspond to (1) the number of layers n_L, (2) the average permeability for each layer, (3) the horizontal gradients of permeability in each layer, (4) n_L-1 interface depths and (5) interface slope angles. The inverse problem is set in a Bayesian framework and prior distributions are defined for each of these parameters. We use a reversible jump Markov chain Monte Carlo algorithm to determine both the optimal number of layers and their respective properties. We apply this approach to a simple case aimed at reconstructing a two-dimensional continuous field using two synthetic well logs of permeability, each located on opposite sides of the domain. This shows the ability of the method to retrieve a posterior density close to that of the reference solution, and demonstrates the potential application to real borehole data. Overall, the proposed methodology is a way to integrate geological parameters in a unified modeling framework, and can be extended to seismic inversion, flow data inversion (e.g., for well test interpretation), or other types of geophysical data.},
 address = {Ponta Delgada (A{\c c}ores), Portugal},
 author = {Herrero, Julien and Caumon, Guillaume and Bodin, Thomas},
 booktitle = {{International Geostatistics Congress 2024, A Springer book series Quantitative Geology and Geostatistics}},
 editor = {Leonardo Azevedo et al.},
 hal_id = {hal-05157109},
 hal_version = {v1},
 keywords = {Transdimensional inversion ; Uncertainty quantification ; Well correlation},
 month = {September},
 title = {{Transdimensional reservoir data inversion for a 2D geological layered model with horizon slopes and horizontal permeability trends}},
 url = {https://hal.univ-lorraine.fr/hal-05157109},
 volume = {2025},
 year = {2024}
}