Conditioning surface-based geological models to well data using artificial neural networks

Titus, Zainab; Heaney, Claire E.; Jacquemyn, Carl; Salinas, P.; Jackson, Matthew D.; Pain, Christopher C.

doi:10.1007/s10596-021-10088-5

Cited by 10 publications

(2 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SBMs consider that the different geological features (layers, architectural elements, facies, etc.) can be separated by surfaces (Pyrcz et al, 2015;Titus et al, 2021;Jo et al, 2020). These methods also integrate the notion of time during which geological objects are deposited.…”

Section: Introductionmentioning

confidence: 99%

A novel surface-based approach to represent aquifer heterogeneity in sedimentary formations

Schorpp,

Straubhaar,

Renard

2024

Preprint

View full text Add to dashboard Cite

Sedimentary formations that compose most aquifers are difficult to model as a result of the nature of their deposition. Their formation generally involves multiple processes (alluvial, glacial, lacustrine, etc.) that contribute to the complex organization of these deposits. Representative models can be obtained using process-based or rule-based methods. However, such methods have several drawbacks: complicated parametrization, large computing time, and challenging, if not impossible, conditioning. To address these problems, we propose a new hierarchical surface-based algorithm, named EROSim. First, a predefined number of stochastic surfaces are simulated in a given order (from older to younger). These surfaces are simulated independently but interact with each other through erosion rules. Each surface is either an erosive or a deposition surface. The deposition surfaces represent the boundaries of depositional events, whereas the erosive surfaces can remove parts of the previously simulated deposits. Finally, these surfaces delimit sedimentary regions that are filled with facies. The approach is quite simple, general, flexible, and can be conditioned to borehole data. The applicability of the method is illustrated using data from fluvio-glacial sedimentary deposits observed in the Bümberg quarry in Switzerland.

show abstract

Section: Introductionmentioning

confidence: 99%

A novel surface-based approach to represent aquifer heterogeneity in sedimentary formations

Schorpp,

Straubhaar,

Renard

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Further, a refill of initial model realizations containing missing zones is carried out using the trained model. Titus, Heaney, Jacquemyn, Salinas, Jackson and Pain (2022) used ANNs to condition a surface-based geological model (SBGM), constructed with a parametric non-uniform rational B-spline (NURBS) approach to well data. ANNs were applied in the following way: (i) to map input parameters of SGBM to types of facies in the vicinity of well locations in the framework of the forward modeling step and (ii) to obtain the optimized set of input parameters of SBGM using a back-propagation method, so that the constructed SBGM complies with the types of facies obtained in well measurements.…”

Section: Introductionmentioning

confidence: 99%

Combined mechanistic and machine learning method for construction of oil reservoir permeability map consistent with well test measurements

Kanin¹,

Garipova²,

Boronin³

et al. 2023

Preprint

View full text Add to dashboard Cite

We propose a new method for construction of the absolute permeability map consistent with the interpreted results of well logging and well test measurements in oil reservoirs. Nadaraya-Watson kernel regression is used to approximate two-dimensional spatial distribution of the rock permeability. Parameters of the kernel regression are tuned by solving the optimization problem in which, for each well placed in an oil reservoir, we minimize the difference between the actual and predicted values of (i) absolute permeability at the well location (results of interpretation of well logging); (ii) absolute integral permeability of the domain around the well and (iii) skin factor (results of interpretation of well tests). Optimization task (inverse problem) is solved via multiple solutions to forward problems, in which we estimate the integral permeability of reservoir surrounding a well and the skin factor by the surrogate model. The last one is developed using an artificial neural network trained on the physics-based synthetic dataset generated using the procedure comprising the numerical simulation of bottomhole pressure decline curve in reservoir simulator followed by its interpretation using a semi-analytical reservoir model. The developed method for reservoir permeability map construction is applied to the available reservoir model (Egg Model) with highly heterogeneous permeability distribution due to the presence of highly-permeable channels. We showed that the constructed permeability map is hydrodynamically similar to the original one. Numerical simulations of production in the reservoir with constructed and original permeability maps are quantitatively similar in terms of the pore pressure and fluid saturations distribution at the end of the simulation period. Moreover, we obtained an good match between the obtained results of numerical simulations in terms of the flow rates and total volumes of produced oil, water and injected water.

show abstract

Reconstrucción probabilística mediante aprendizaje automático del sistema acuífero de la cuenca del Po (Italia)

et al. 2023

View full text Add to dashboard Cite

A machine-learning-based methodology is proposed to delineate the spatial distribution of geomaterials across a large-scale three-dimensional subsurface system. The study area spans the entire Po River Basin in northern Italy. As uncertainty quantification is critical for subsurface characterization, the methodology is specifically designed to provide a quantitative evaluation of prediction uncertainty at each location of the reconstructed domain. The analysis is grounded on a unique dataset that encompasses lithostratigraphic data obtained from diverse sources of information. A hyperparameter selection technique based on a stratified cross-validation procedure is employed to improve model prediction performance. The quality of the results is assessed through validation against pointwise information and available hydrogeological cross-sections. The large-scale patterns identified are in line with the main features highlighted by typical hydrogeological surveys. Reconstruction of prediction uncertainty is consistent with the spatial distribution of available data and model accuracy estimates. It enables one to identify regions where availability of new information could assist in the constraining of uncertainty. The comprehensive dataset provided in this study, complemented by the model-based reconstruction of the subsurface system and the assessment of the associated uncertainty, is relevant from a water resources management and protection perspective. As such, it can be readily employed in the context of groundwater availability and quality studies aimed at identifying the main dynamics and patterns associated with the action of climate drivers in large-scale aquifer systems of the kind here analyzed, while fully embedding model and parametric uncertainties that are tied to the scale of investigation.

show abstract

Conditioning surface-based geological models to well data using artificial neural networks

Cited by 10 publications

References 32 publications

A novel surface-based approach to represent aquifer heterogeneity in sedimentary formations

A novel surface-based approach to represent aquifer heterogeneity in sedimentary formations

Combined mechanistic and machine learning method for construction of oil reservoir permeability map consistent with well test measurements

Reconstrucción probabilística mediante aprendizaje automático del sistema acuífero de la cuenca del Po (Italia)

Contact Info

Product

Resources

About