Verification of a real-time ensemble-based method for updating earth model based on GAN

Fossum, Kristian; Alyaev, Sergey; Tveranger, Jan; Elsheikh, Ahmed H.

doi:10.1016/j.jocs.2022.101876

Cited by 8 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many previous studies have shown that, as an approximate Bayesian method, EnRML can provide posterior estimates of lower order moments (e.g., mean and variance) similar to those produced by the most accurate MCMC. 40,63 The method is applicable to the cases where only the lower order moments are important (e.g., the mean and the uncertainty of deflections), but not to the cases where the higher order moments and the exact probability are critical.…”

Section: Discussionmentioning

confidence: 99%

“…The golden standard of Bayesian inference (i.e., MCMC sampling) is not adopted because the MCMC sampling‐based Bayesian updating is computationally intractable when a numerical model of excavation is used as the forward model. Many previous studies have shown that, as an approximate Bayesian method, EnRML can provide posterior estimates of lower order moments (e.g., mean and variance) similar to those produced by the most accurate MCMC 40,63 . The method is applicable to the cases where only the lower order moments are important (e.g., the mean and the uncertainty of deflections), but not to the cases where the higher order moments and the exact probability are critical. The unknown parameters to be updated ( θ ) should be selected as the soil parameters that significantly affect the quantity of interest (i.e., wall deflection).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A bi‐fidelity inverse analysis method for deep excavations considering three‐dimensional effects

Tao,

Pan,

Sun

et al. 2024

Num Anal Meth Geomechanics

View full text Add to dashboard Cite

Inverse analysis methods are commonly used in braced excavations for improved deformation predictions. This paper proposes a bi‐fidelity ensemble randomized maximum likelihood (BF‐EnRML) method for efficient inverse analyses of deep excavations considering the three‐dimensional effects. The bi‐fidelity (BF) model is developed by the low‐fidelity model (i.e., two‐dimensional finite element model, 2D FEM) and the high‐fidelity model (i.e., 3D FEM) for a balance between efficiency and accuracy. A large number of 2D FEMs are first used to explore the relationship between soil parameters and wall deflections. A few 3D FEMs are then performed to calibrate the discrepancy between 2D‐3D deflections caused by the inability of 2D FEM to consider the three‐dimensional effects. The constructed BF model serves as the forward model in inverse analysis. The soil parameters are updated by incorporating the monitoring data based on EnRML and further used to predict wall deflections in later stages. A hypothetical excavation and a real project are studied to evaluate the performance of the proposed method. The results show that the BF model can provide wall deflection predictions close to those calculated from 3D FEM while using a computational cost of 2D FEM. The BF‐EnRML method can efficiently update the soil parameters and improve the wall deflection predictions. Moreover, factors affecting the accuracy of the BF model are studied, including the number of required 3D FEMs, the distance from the evaluated wall section to the excavation corner, the number of data points along the wall depth, and the number of excavation stages.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

A bi‐fidelity inverse analysis method for deep excavations considering three‐dimensional effects

Tao,

Pan,

Sun

et al. 2024

Num Anal Meth Geomechanics

View full text Add to dashboard Cite

show abstract

“…The user then reviews and refines the generated horizon shapes and repeats this process for each horizon of interest. More recent versions of these algorithms leverage various comprehensive features within the 3D seismic volume, see [20,[22][23][24] and incorporate advanced data processing approaches like deep learning, see [25][26][27][28][29][30].…”

Section: Horizon Auto-trackingmentioning

confidence: 99%

Real-Time Automated Geosteering Interpretation Combining Log Interpretation and 3D Horizon Tracking

D’Angelo,

Zhao,

Zhang

et al. 2024

Geosciences

View full text Add to dashboard Cite

Existing methods for estimating formation boundaries from well-log data only analyze the formation along the wellbore, failing to capture changes in the 3D formation structure around it. This paper presents a method for real-time 3D formation boundary interpretation using readily available well logs and seismic image data. In the proposed workflow, the mean formation boundary is estimated as a curve following the well path. 3D surfaces are then fitted through this boundary curve, aligning with the slopes and features in the seismic image data. The proposed method is tested on both synthetic and field datasets and illustrates the capabilities of accurate boundary estimation near the well path and precise representation of boundary shape changes further away from the well trajectory. With this fully automated geological interpretation workflow, human bias and interpretation uncertainty can be minimized. Subsurface conditions can be continually updated while drilling to optimize drilling decisions and further automate the geosteering process.

show abstract

“…This approach successfully addressed uncertainty and optimized the geologic model of a sandstone greenstone belt. More recently, Fossum et al (2022) use the ensemble randomized maximum likelihood to update the subsurface uncertainty in earth models and simulate electromagnetic logs generated with generative adversarial networks and a forward deep neural network, respectively.…”

Section: Introductionmentioning

confidence: 99%

Quantitative interpretation of geoelectric inverted data with a robust probabilistic approach

et al. 2023

View full text Add to dashboard Cite

The non-uniqueness of the solution of the geophysical inverse problem can lead to misinterpretation while characterizing the subsurface. To tackle this situation, ground-truth information from excavations and wells can be used to improve, calibrate and to interpret inverted models. We refer to quantitative interpretation as the decision analysis based on probability theory, which is focused on solving a classification problem. First, we present a probabilistic approach to classify different types of materials or categories observed in borehole logs using multiple data sources: inverted 2D electrical resistivity tomography (ERT) and induced polarization (IP) data, and the positions (x, z) of these boreholes. Then, using Bayes rule and permanence of ratios, we compute joint conditional probabilities of each category, given all data sources in the whole inverted model domain. We validate this approach with synthetic data modeled for a complex anthropogenic-geologic scenario and using real data from an old landfill. Afterwards, we assess the performance of the probabilistic approach for classification and compare it with the machine learning algorithm of multi-layer perceptron (MLP). Additionally, we analyze the effect that the different data sources and the number of boreholes (and its distribution) have on both approaches with the synthetic case. Our results show that the MLP performance is better for delineating the different categories where the lateral contrasts in the synthetic resistivity model are small. Nevertheless, the classification obtained with the probabilistic approach using the real data seems to provide a more geologically realistic distribution. We conclude that the probabilistic approach is robust for classifying categories when high spatial heterogeneity is expected and when ground-truth data is limited or not sparsely distributed. Finally this approach can be easily extended to integrate multiple geophysical methods and does not require the optimization of hyperparameters as for MLP.

show abstract

Verification of a real-time ensemble-based method for updating earth model based on GAN

Cited by 8 publications

References 23 publications

A bi‐fidelity inverse analysis method for deep excavations considering three‐dimensional effects

A bi‐fidelity inverse analysis method for deep excavations considering three‐dimensional effects

Real-Time Automated Geosteering Interpretation Combining Log Interpretation and 3D Horizon Tracking

Quantitative interpretation of geoelectric inverted data with a robust probabilistic approach

Contact Info

Product

Resources

About