Shib Sankar Ganguli
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Shib Sankar Ganguli has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
Prakash Kumar
6 shared publications- Souvik Sen
Department of Petroleum Engineering, Indian Institute of Technology (Indian School of Mines), Dhanbad, India
6 shared publications V. P. Dimri
CSIR-National Geophysical Research Institute
6 shared publicationsNimisha Vedanti
5 shared publicationsSouvik Sen
5 shared publicationsV.P. Dimri
CSIR Natl Geophys Res Inst, Uppal Rd, Hyderabad 500007, Telangana, India
5 shared publications
Latest Publications
- Open Access
- 1 November 2026
- Published by Wiley
A BAYESIAN PHYSICS-DRIVEN NEURAL NETWORK FOR ELASTIC LOG PREDICTION VIA CROSS-BASIN TRANSFER LEARNING
Driven by the demands of reservoir characterization, accurate elastic log prediction remains a critical research priority that bridges rock physical properties with seismic responses for hydrocarbon exploration. Elastic logs are conventionally estimated using rock physics models (RPMs) or, more recently, machine learning (ML) frameworks. However, these RPMs are constrained by idealized assumptions and fail to account for uncertainty, whereas existing ML approaches often lack physical constraints. To address this, Bayesian Neural Networks (BNNs) have proven successful in quantitative log prediction by simultaneously enhancing accuracy and quantifying model uncertainty. We introduce a novel BNN framework designed for the quantitative estimation of elastic properties from well logs. By integrating rock physics constraints via a physics-based loss term derived from synthetic velocity realizations, the proposed method ensures physically consistent predictions and calibrated uncertainty. Using variational inference (VI) governed by the Kullback-Leibler divergence objective function, the framework approximates the neural network’s posterior distribution to generate ensemble predictions that closely align with the target logs. To bridge the gap between velocity measurements and rock physical properties, the network incorporates a physics-dependent loss term derived from various combinations of velocity realizations. Furthermore, we employ cross-domain adaptation via a two-stage fine-tuning transfer learning strategy to predict seismic velocities in an unseen basin. By utilizing selective layer updates and regularization on a limited subset of target well data, the framework adapts the pre-trained BNN without sacrificing prior knowledge, ensuring physically meaningful predictions and physically consistent trends despite domain shift. When tested on field datasets from the Canning Basin (Western Australia) and the Sichuan Basin (China), our proposed approach outperformed the standard BNN, delivering higher correlation coefficients, lower prediction errors, and lower uncertainties. Remarkably, this framework offers a scalable, robust paradigm for subsurface characterization in complex, data-limited regimes by synthesizing physics-informed constraints with cross-domain transfer learning.
- 1 November 2026
- Published by Wiley
Open AccessA BAYESIAN PHYSICS-DRIVEN NEURAL NETWORK FOR ELASTIC LOG PREDICTION VIA CROSS-BASIN TRANSFER LEARNING
Driven by the demands of reservoir characterization, accurate elastic log prediction remains a critical research priority that bridges rock physical properties with seismic responses for hydrocarbon exploration. Elastic logs are conventionally estimated using rock physics models (RPMs) or, more recently, machine learning (ML) frameworks. However, these RPMs are constrained by idealized assumptions and fail to account for uncertainty, whereas existing ML approaches often lack physical constraints. To address this, Bayesian Neural Networks (BNNs) have proven successful in quantitative log prediction by simultaneously enhancing accuracy and quantifying model uncertainty. We introduce a novel BNN framework designed for the quantitative estimation of elastic properties from well logs. By integrating rock physics constraints via a physics-based loss term derived from synthetic velocity realizations, the proposed method ensures physically consistent predictions and calibrated uncertainty. Using variational inference (VI) governed by the Kullback-Leibler divergence objective function, the framework approximates the neural network’s posterior distribution to generate ensemble predictions that closely align with the target logs. To bridge the gap between velocity measurements and rock physical properties, the network incorporates a physics-dependent loss term derived from various combinations of velocity realizations. Furthermore, we employ cross-domain adaptation via a two-stage fine-tuning transfer learning strategy to predict seismic velocities in an unseen basin. By utilizing selective layer updates and regularization on a limited subset of target well data, the framework adapts the pre-trained BNN without sacrificing prior knowledge, ensuring physically meaningful predictions and physically consistent trends despite domain shift. When tested on field datasets from the Canning Basin (Western Australia) and the Sichuan Basin (China), our proposed approach outperformed the standard BNN, delivering higher correlation coefficients, lower prediction errors, and lower uncertainties. Remarkably, this framework offers a scalable, robust paradigm for subsurface characterization in complex, data-limited regimes by synthesizing physics-informed constraints with cross-domain transfer learning.
- Open Access
- 25 June 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Solid Earth
- Vol. 131 (6)
Computing Effective Elastic Properties From Microtomographic Images: An Efficient 3D Digital Rock Workflow
Digital rock physics (DRP), using high‐resolution 3D microtomographic rock images, is widely employed to predict elastic and transport properties of subsurface formations. Accurate microstructural characterization is often hindered by limited image resolution and overlapping grayscale values of minerals. Consequently, DRP simulations are often restricted to samples with simple structures and mineralogies. We propose a practical workflow that leverages a voxel‐resolved finite element solver to rapidly predict effective elastic properties from segmented 3D digital rock images, incorporating mineral phases and their corresponding volume fractions. We selected Bentheimer and Doddington sandstones to assess the accuracy and performance of the workflow and to demonstrate its practical application. Including feldspar and clay in the solver, despite the grayscale similarities to quartz, significantly improved the predictions of elastic moduli and produced modulus‐porosity trends that closely matched laboratory ultrasonic measurements for Bentheimer sandstone. Uncertainty and sensitivity analyses evaluated how mineral volume fractions, input mineral moduli, grain contact size, and mineralogy, including aspect ratios and coordination numbers, impacted the predictions. The simulation results reveal that the dominant source of uncertainty is attributed to the level of sophistication of the modeling approach–whether a single‐mineral (quartz only) or multi‐mineral segmented model is applied. The quartz‐only model exhibited errors of up to 36% in bulk modulus and 22% in shear modulus relative to laboratory measurements. These errors were reduced to 22% and 9.9%, respectively, when a proper mineral‐weighted heterogeneous segmentation was employed. These results demonstrate that the workflow can be extended to estimate rock physical properties beyond elastic parameters.
- 25 June 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Solid Earth
- Vol. 131 (6)
Open AccessComputing Effective Elastic Properties From Microtomographic Images: An Efficient 3D Digital Rock Workflow
Digital rock physics (DRP), using high‐resolution 3D microtomographic rock images, is widely employed to predict elastic and transport properties of subsurface formations. Accurate microstructural characterization is often hindered by limited image resolution and overlapping grayscale values of minerals. Consequently, DRP simulations are often restricted to samples with simple structures and mineralogies. We propose a practical workflow that leverages a voxel‐resolved finite element solver to rapidly predict effective elastic properties from segmented 3D digital rock images, incorporating mineral phases and their corresponding volume fractions. We selected Bentheimer and Doddington sandstones to assess the accuracy and performance of the workflow and to demonstrate its practical application. Including feldspar and clay in the solver, despite the grayscale similarities to quartz, significantly improved the predictions of elastic moduli and produced modulus‐porosity trends that closely matched laboratory ultrasonic measurements for Bentheimer sandstone. Uncertainty and sensitivity analyses evaluated how mineral volume fractions, input mineral moduli, grain contact size, and mineralogy, including aspect ratios and coordination numbers, impacted the predictions. The simulation results reveal that the dominant source of uncertainty is attributed to the level of sophistication of the modeling approach–whether a single‐mineral (quartz only) or multi‐mineral segmented model is applied. The quartz‐only model exhibited errors of up to 36% in bulk modulus and 22% in shear modulus relative to laboratory measurements. These errors were reduced to 22% and 9.9%, respectively, when a proper mineral‐weighted heterogeneous segmentation was employed. These results demonstrate that the workflow can be extended to estimate rock physical properties beyond elastic parameters.
- Open Access
- 7 February 2026
- Published by Wiley
Computing Effective Elastic Properties from Microtomographic Images: An Efficient 3D Digital Rock Workflow
Digital rock physics (DRP), using high-resolution 3D microtomographic rock images, is widely employed to predict elastic and transport properties of subsurface formations. Accurate microstructural characterization is often hindered by limited image resolution and overlapping grayscale values of minerals. Consequently, DRP simulations are often restricted to samples with simple structures and mineralogies. We propose a practical workflow that leverages a voxel-resolved finite element solver to rapidly predict effective elastic properties from segmented 3D digital rock images, incorporating mineral phases and their corresponding volume fractions. We selected Bentheimer and Doddington sandstones to assess the accuracy and performance of the workflow and to demonstrate its practical application. Including feldspar and clay in the solver, despite the grayscale similarities to quartz, significantly improved the predicted elastic moduli and produced modulus-porosity trends that closely matched laboratory ultrasonic measurements for Bentheimer sandstone. Uncertainty and sensitivity analyses evaluated how mineral volume fractions, input mineral moduli, grain contact size, and mineralogy, including aspect ratios and coordination numbers, impacted predictions. The simulation results reveal that the dominant source of uncertainty is attributed to the level of sophistication of the modeling approach–whether a single-mineral (quartz only) or multi-mineral segmented model is applied. The quartz-only model exhibited errors of up to 36% in bulk modulus and 22% in shear modulus relative to laboratory measurements. These errors were reduced to 22% and 9.9% respectively, when a proper mineral-weighted heterogeneous segmentation was employed. These results demonstrate that the workflow can be extended to estimate rock physical properties beyond elastic parameters.
- 7 February 2026
- Published by Wiley
Open AccessComputing Effective Elastic Properties from Microtomographic Images: An Efficient 3D Digital Rock Workflow
Digital rock physics (DRP), using high-resolution 3D microtomographic rock images, is widely employed to predict elastic and transport properties of subsurface formations. Accurate microstructural characterization is often hindered by limited image resolution and overlapping grayscale values of minerals. Consequently, DRP simulations are often restricted to samples with simple structures and mineralogies. We propose a practical workflow that leverages a voxel-resolved finite element solver to rapidly predict effective elastic properties from segmented 3D digital rock images, incorporating mineral phases and their corresponding volume fractions. We selected Bentheimer and Doddington sandstones to assess the accuracy and performance of the workflow and to demonstrate its practical application. Including feldspar and clay in the solver, despite the grayscale similarities to quartz, significantly improved the predicted elastic moduli and produced modulus-porosity trends that closely matched laboratory ultrasonic measurements for Bentheimer sandstone. Uncertainty and sensitivity analyses evaluated how mineral volume fractions, input mineral moduli, grain contact size, and mineralogy, including aspect ratios and coordination numbers, impacted predictions. The simulation results reveal that the dominant source of uncertainty is attributed to the level of sophistication of the modeling approach–whether a single-mineral (quartz only) or multi-mineral segmented model is applied. The quartz-only model exhibited errors of up to 36% in bulk modulus and 22% in shear modulus relative to laboratory measurements. These errors were reduced to 22% and 9.9% respectively, when a proper mineral-weighted heterogeneous segmentation was employed. These results demonstrate that the workflow can be extended to estimate rock physical properties beyond elastic parameters.
- 27 February 2024
- Published by Society of Exploration Geophysicists in Interpretation
- Vol. 12 (2) , SB1-SB8
Constraining maximum horizontal stress using wellbore breakouts — A case study from the Ordovician tight reservoir of the northeastern Oued Mya Basin, Algeria
In this study, we interpret the maximum horizontal stress (SHmax) azimuth from the breakout positions of the wellbore and attempt to constrain the SHmax gradient based on the interpreted breakout width. A cumulative of 110 m of breakouts are deciphered within the Ordovician Hamra Quartzite interval of the Oued Mya Basin from a 138 m acoustic image log. These breakouts are ranked as “A-Quality” following the World Stress Map ranking guidelines. We infer a mean SHmax orientation of N28°E ± 8°. Following the frictional faulting mechanism and stress polygon approach, measurement of the minimum horizontal stress (Shmin) from minifrac tests and observations of the compressive failures from the acoustic image log provide strong constraints on the SHmax magnitude in the reservoir interval in the absence of core-measured rock strength. Interpreted breakout widths exhibit a range between 32.6° and 90.81°, which indicates a SHmax range of 24.4–34.7 MPa/km. The average breakout width of 62.58° translates to a narrower SHmax gradient range, varying between 27.2 and 31.2 MPa/km. The relative magnitudes of the principal stresses indicate a strong strike-slip tectonic stress state. Considering all the uncertainties, we infer a SHmax/Shmin ratio of 1.41–1.81 within the Ordovician interval.
- 27 February 2024
- Published by Society of Exploration Geophysicists in Interpretation
- Vol. 12 (2) , SB1-SB8
Constraining maximum horizontal stress using wellbore breakouts — A case study from the Ordovician tight reservoir of the northeastern Oued Mya Basin, Algeria
In this study, we interpret the maximum horizontal stress (SHmax) azimuth from the breakout positions of the wellbore and attempt to constrain the SHmax gradient based on the interpreted breakout width. A cumulative of 110 m of breakouts are deciphered within the Ordovician Hamra Quartzite interval of the Oued Mya Basin from a 138 m acoustic image log. These breakouts are ranked as “A-Quality” following the World Stress Map ranking guidelines. We infer a mean SHmax orientation of N28°E ± 8°. Following the frictional faulting mechanism and stress polygon approach, measurement of the minimum horizontal stress (Shmin) from minifrac tests and observations of the compressive failures from the acoustic image log provide strong constraints on the SHmax magnitude in the reservoir interval in the absence of core-measured rock strength. Interpreted breakout widths exhibit a range between 32.6° and 90.81°, which indicates a SHmax range of 24.4–34.7 MPa/km. The average breakout width of 62.58° translates to a narrower SHmax gradient range, varying between 27.2 and 31.2 MPa/km. The relative magnitudes of the principal stresses indicate a strong strike-slip tectonic stress state. Considering all the uncertainties, we infer a SHmax/Shmin ratio of 1.41–1.81 within the Ordovician interval.
- 8 December 2023
- Published by American Chemical Society (ACS) in Energy & Fuels
- Vol. 37 (24) , 19575-19589
Characterization and Feature Ranking of Well Log Variables Using Data-Driven Algorithms for Total Organic Carbon Estimation of Organic-Rich Shales
- 8 December 2023
- Published by American Chemical Society (ACS) in Energy & Fuels
- Vol. 37 (24) , 19575-19589
Characterization and Feature Ranking of Well Log Variables Using Data-Driven Algorithms for Total Organic Carbon Estimation of Organic-Rich Shales
- Open Access
- 1 September 2023
- Published by KEAI Publishing in Energy Geoscience
- Vol. 5 (2)
Reservoir heterogeneity analysis using multi-directional textural attributes from deep learning-based enhanced acoustic impedance inversion: A study from Poseidon, NW shelf Australia
- 1 September 2023
- Published by KEAI Publishing in Energy Geoscience
- Vol. 5 (2)
Open AccessReservoir heterogeneity analysis using multi-directional textural attributes from deep learning-based enhanced acoustic impedance inversion: A study from Poseidon, NW shelf Australia
- 14 July 2023
- Published by Springer Nature
- p. 678-688
Inversion Theory in Geoscience
- Open Access
- 10 August 2022
- Published by Elsevier in Geosystems and Geoenvironment
- Vol. 2 (1)
Assessment of reservoir stress state and its implications for Paleozoic tight oil reservoir development in the Oued Mya Basin, northeastern Algerian Sahara
- 10 August 2022
- Published by Elsevier in Geosystems and Geoenvironment
- Vol. 2 (1)
Open AccessAssessment of reservoir stress state and its implications for Paleozoic tight oil reservoir development in the Oued Mya Basin, northeastern Algerian Sahara
- 8 August 2022
- Published by Elsevier in Tectonophysics
- Vol. 839
Variable source depth beneath the Indian Ocean geoid low area: Insights from L1 and L2 norm-based scaling power spectrum inversion
- 8 August 2022
- Published by Elsevier in Tectonophysics
- Vol. 839
Variable source depth beneath the Indian Ocean geoid low area: Insights from L1 and L2 norm-based scaling power spectrum inversion