Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Mukerji, Tapan
ABSTRACT Exploration seismology uses reflected and refracted seismic waves, emitted from a controlled (active) source into the ground, and recorded by an array of seismic sensors (receivers) to image the subsurface geologic structures. These seismic images are the main resources for energy and resource exploration and scientific investigation of the crust and upper mantle. We survey recent advances in applications of machine-learning methods, more specifically deep neural networks (DNNs), in exploration seismology. We provide a technically oriented review of DNN applications for seismic data acquisition; data preprocessing tasks such as interpolation/extrapolation, denoising, first-break picking, velocity picking, and seismic migration; data processing tasks such as geologic and structural interpretations; and data modeling tasks such as the inference of subsurface structures and lithologic and petrophysical properties. DNNs have entered almost every sector of exploration seismology. They have outperformed many traditional algorithms for the automation of seismic data acquisition, data preprocessing, data processing, interpretations, and data modeling tasks. However, despite the impressive performances of DNN-based approaches, the out-of-distribution generalization and interpretability of these models remain challenging. To overcome these challenges, incorporating domain knowledge into the DNNs is a promising path and a focus of current deep-learning research in seismology.
- Geology > Geological Subdiscipline > Stratigraphy (1.00)
- Geology > Rock Type > Sedimentary Rock (0.93)
- Geology > Structural Geology > Tectonics > Plate Tectonics (0.46)
ABSTRACT Exploration seismology uses reflected and refracted seismic waves, emitted from a controlled (active) source into the ground, and recorded by an array of seismic sensors (receivers) to image the subsurface geologic structures. These seismic images are the main resources for energy and resource exploration and scientific investigation of the crust and upper mantle. We survey recent advances in applications of machine-learning methods, more specifically deep neural networks (DNNs), in exploration seismology. We provide a technically oriented review of DNN applications for seismic data acquisition; data preprocessing tasks such as interpolation/extrapolation, denoising, first-break picking, velocity picking, and seismic migration; data processing tasks such as geologic and structural interpretations; and data modeling tasks such as the inference of subsurface structures and lithologic and petrophysical properties. DNNs have entered almost every sector of exploration seismology. They have outperformed many traditional algorithms for the automation of seismic data acquisition, data preprocessing, data processing, interpretations, and data modeling tasks. However, despite the impressive performances of DNN-based approaches, the out-of-distribution generalization and interpretability of these models remain challenging. To overcome these challenges, incorporating domain knowledge into the DNNs is a promising path and a focus of current deep-learning research in seismology.
- Geology > Geological Subdiscipline > Stratigraphy (1.00)
- Geology > Rock Type > Sedimentary Rock (0.93)
- Geology > Structural Geology > Tectonics > Plate Tectonics (0.46)
Bayesian geophysical basin modeling with seismic kinematic metrics to quantify uncertainty for pore pressure prediction
Fonseca, Josue (Stanford University) | Pradhan, Anshuman (Stanford University, California Institute of Technology) | Mukerji, Tapan (Stanford University, Stanford University, Stanford University)
ABSTRACT Bayesian geophysical basin modeling (BGBM) methodology is an interdisciplinary workflow that incorporates data, geologic expertise, and physical processes through Bayesian inference in sedimentary basin models. Its application culminates in subsurface models and properties that integrate the geohistory of a basin, rock-physics definitions, well-log and drilling data, and seismic information. Monte Carlo basin modeling realizations are performed by sampling from prior probability distributions on facies parameters and basin boundary conditions. After the data assimilation, the accepted set of posterior subsurface models yields an uncertainty quantification of subsurface properties. This procedure is especially suitable for pore pressure prediction in a predrill stage. However, the high computational cost of seismic data assimilation decreases the practicality of the workflow. Therefore, we introduce and investigate seismic traveltime criteria as computationally faster proxies for analyzing the seismic data likelihood when using BGBM. Our surrogate schemes weigh the prior basin model results with the available seismic data with no need to perform expensive seismic depth-migration procedures for each Monte Carlo realization. Furthermore, we apply BGBM in a real field case from the Gulf of Mexico using a 2D section for pore pressure prediction considering different kinematic criteria. The workflow implementation with the novel seismic data assimilation proxies is compared with the complete computationally expensive benchmark approach, which uses a global analysis of the residual moveout in depth-migrated seismic image samples. Moreover, we validate and compare the outcomes for predicted pore pressure with mudweight data from a blind well. The fast proxy for analyzing the depth positioning of seismic horizons developed in this work yields similar uncertainty quantification results in pore pressure prediction compared with the computationally expensive benchmark. Our fast proxies make the BGBM methodology efficient and practical.
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.93)
- Geophysics > Seismic Surveying > Seismic Processing > Seismic Migration (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling (1.00)
- Geophysics > Seismic Surveying > Seismic Interpretation (1.00)
Exploration seismology uses reflected and refracted seismic waves, emitted from a controlled (active) source into the ground, and recorded by an array of seismic sensors (receivers) to image the subsurface geological structures. These seismic images are the main resources for energy and resource exploration and scientific investigation of the crust and upper mantle. The exponential growth of data volumes from seismic surveys using modern instruments drives the need for algorithmic developments to process large-volume high-dimensional data. This is where artificial intelligence and modern machine learning techniques could play a key role. We survey recent advances in applications of machine learning (ML) methods, more specifically deep neural networks (DNNs), in the exploration seismology. We provide comprehensive and technically oriented review of DNNs applications for seismic data acquisition, data pre-processing tasks such as interpolation/extrapolation, denoising, first break picking, velocity picking, seismic migration, data processing tasks such as geological and structural interpretations, and data modeling tasks such as the inference of sub-surface structures and lithologic and petrophysical properties. Our goal is to document the technological trends, progress, and challenges for DNN approaches in the exploration seismology. This could benefit other sectors of the seismological community such as earthquake and engineering seismology by sharing how progress was achieved. data pre-processing tasks such as interpolation/extrapolation, denoising, first break picking, velocity picking, seismic migration, data processing tasks such as geological and structural interpretations, and data modeling tasks such as the inference of sub-surface structures and lithologic and petrophysical properties. Our goal is to document the technological trends, progress, and challenges for DNN approaches in the exploration seismology. This could benefit other sectors of the seismological com-#xD;munity such as earthquake and engineering seismology by sharing how progress was achieved.
- Geology > Structural Geology > Tectonics > Plate Tectonics > Earthquake (1.00)
- Geology > Rock Type > Sedimentary Rock (1.00)
- Geology > Geological Subdiscipline (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
- Geophysics > Seismic Surveying > Passive Seismic Surveying > Earthquake Seismology (1.00)
- Geophysics > Seismic Surveying > Seismic Interpretation > Seismic Reservoir Characterization > Amplitude vs Offset (AVO) (0.67)
Banana hole cave systems can be reservoirs to be targeted, or near-surface drilling hazards to be avoided. Modeling the spatial uncertainty of these caves is critical for developing strategies for well placement, or resource production or storage. First, a spatial statistical analysis is performed on observed banana holes. This analysis is done to better understand the morphology of banana holes and to create a reference of statistical metrics. Secondly, deep generative models are explored for banana hole simulation. The performance of the models are evaluated by comparing their realizations to observed caves.
Bayesian Geophysical Basin Modeling with Seismic Kinematics Metrics to Quantify Uncertainty for Pore Pressure Prediction
Fonseca, Josue (Stanford University) | Pradhan, Anshuman (Stanford University, California Institute of Technology) | Mukerji, Tapan (Stanford University, Stanford University, Stanford University)
Bayesian geophysical basin modeling (BGBM) methodology is an interdisciplinary workflow that incorporates data, geological expertise, and physical processes through Bayesian inference in sedimentary basin models. Its application culminates in subsurface models and properties that integrate the geo-history of a basin, rock physics definitions, well log and drilling data, and seismic information. Monte Carlo basin modeling realizations are performed by sampling from prior probability distributions on facies parameters and basin boundary conditions. After data assimilation, the accepted set of posterior subsurface models yields uncertainty quantification of subsurface properties. This procedure is especially suitable for pore pressure prediction in a predrill stage. However, the high computational cost of seismic data assimilation decreases the practicality of the workflow. Therefore, we introduce and investigate seismic traveltimes criteria as computationally faster proxies for analyzing the seismic data likelihood when employing BGBM. The proposed surrogate schemes weigh the prior basin model results with the available seismic data with no need to perform expensive seismic depth-migration procedures for each Monte Carlo realization. Furthermore, we apply BGBM in a real field case from the Gulf of Mexico using a 2D section for pore pressure prediction considering different kinematics criteria. The workflow implementation with the novel seismic data assimilation proxies is compared with the complete computationally expensive benchmark approach, which utilizes a global analysis of the residual moveout in depth-migrated seismic image samples. Moreover, we validate and compare the outcomes for predicted pore pressure with mud-weight data from a blind well. The fast proxy of analyzing the depth-positioning of seismic horizons proposed in this work yields similar uncertainty quantification results in pore pressure prediction compared to the computationally expensive benchmark. The proposed fast proxies make the BGBM methodology efficient and practical.
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.93)
- Geophysics > Seismic Surveying > Seismic Processing > Seismic Migration (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling (1.00)
- Geophysics > Seismic Surveying > Seismic Interpretation (1.00)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Uncertainty > Bayesian Inference (0.86)
- Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (0.67)
- Information Technology > Artificial Intelligence > Machine Learning > Learning Graphical Models > Directed Networks > Bayesian Learning (0.48)
Abstract Petrophysical inversion is an important aspect of reservoir modeling. However, due to the lack of a unique and straightforward relationship between seismic traces and rock properties, predicting petrophysical properties directly from seismic data is a complex task. Many studies have attempted to identify the direct end-to-end link using supervised machine learning techniques, but they face challenges such as lack of a large petrophysical training data set or estimates that may not conform with physics or depositional history of the rocks. We present a rock- and wave-physics-informed neural network (RW-PINN) model that can estimate porosity directly from seismic image traces with no wells or with a limited number of wells and with predictions that are consistent with rock physics and geologic knowledge of deposition. The RW-PINN takes advantage of auto-differentiation to compute the gradients across the rock- and wave-physics models. As an example, we use the uncemented-sand rock-physics model and normal-incidence wave physics to guide the learning of the RW-PINN to eventually get good estimates of porosities from normal-incidence seismic traces and limited well data. Training the RW-PINN with few wells (weakly supervised scenario) helps in tackling the problem of nonuniqueness as different porosity logs can give similar seismic traces. We use a weighted normalized root mean square error loss function to train the weakly supervised network and demonstrate the impact of different weights on porosity predictions. The RW-PINN's estimated porosities and seismic traces are compared to predictions from a completely supervised model, which gives slightly better porosity estimates but matches the seismic traces poorly and requires a large amount of labeled training data. We demonstrate the complete workflow for executing petrophysical inversion of seismic data using self-supervised or weakly supervised RW-PINNs.
- Europe > Norway (0.46)
- North America > United States > California (0.28)
- Europe > United Kingdom (0.28)
- Geophysics > Seismic Surveying > Seismic Processing (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
- Europe > Norway > Norwegian Sea > Halten Terrace > Block 6406/2 > Lavrans Field (0.99)
- Europe > Norway > Norwegian Sea > Halten Terrace > Block 6406/2-1 > Lavrans Field (0.99)
ABSTRACT The prediction of petrophysical properties, such as porosity and rock-fluid volumes, from partially stacked seismic data typically requires a rock-physics model that often is lithology dependent and difficult to calibrate. We adopt canonical correlation analysis (CCA) to infer the underlying relation between petrophysical properties and elastic attributes estimated from seismic data. We develop a two-step inversion approach: first, we predict elastic properties from partially stacked seismic data using a Bayesian linear inverse method based on an amplitude-variation-with-offset (AVO) linearization in terms of fluid, rigidity, and density factors, and then we predict petrophysical properties from the estimated AVO attributes using CCA. The novelty of our approach is the application of CCA to the fluid and rigidity factors, which avoids the calibration of an explicit rock-physics model by automatically deriving a linear relation in the lower dimensional space of the canonical variables. The parameterization of the linearization in terms of fluid, rigidity, and density factors maximizes the correlation with respect to the petrophysical properties of interest. Furthermore, the probabilistic approach is extended to the petrophysical inversion using Bayesian linear theory and the posterior distribution of petrophysical properties conditioned by seismic data is computed by combining the probability distributions obtained from seismic and petrophysical inversion to propagate the uncertainty from the seismic to the petrophysical domain. The inversion is validated on a synthetic case that finds high accuracy of our formulation. A case study with synthetic and real partially stacked seismic data also is presented and compared to a traditional inversion with an explicit rock-physics model.
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.46)
ABSTRACT The physics that describes the seismic response of an interval of saturated porous rocks with known petrophysical properties is relatively well understood and includes rock physics, petrophysics, and wave propagation models. The main goal of seismic reservoir characterization is to predict the rock and fluid properties given a set of seismic measurements by combining geophysical models and mathematical methods. This modeling challenge is generally formulated as an inverse problem. The most common geophysical inverse problem is the seismic (or elastic) inversion, i.e.,ย the estimation of elastic properties, such as seismic velocities or impedances, from seismic amplitudes and traveltimes. The estimation of petrophysical properties, such as porosity, lithology, and fluid saturations, also can be formulated as an inverse problem and is generally referred to as rock-physics (or petrophysical) inversion. Several deterministic and probabilistic methods can be applied to solve seismic inversion problems. Deterministic algorithms predict a single solution, which is a โbestโ estimate or the most likely value of the model variables of interest. In probabilistic algorithms, on the other hand, the solution is the probability distribution of the model variables of interest, which can be expressed as a conditional probability density function or a set of model realizations conditioned on the data. The probabilistic approach provides a quantification of the uncertainty of the solution in addition to the most likely model. Our goal is to define the terminology, present an overview of probabilistic seismic and rock-physics inversion methods for the estimation of petrophysical properties, demonstrate the fundamental concepts with illustrative examples, and discuss the recent research developments.
- Europe (1.00)
- Asia (1.00)
- South America (0.67)
- North America > United States > California (0.28)
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.67)
- South America > Argentina > Patagonia > Golfo San Jorge Basin (0.99)
- Europe > United Kingdom > Atlantic Margin > West of Shetland > Faroe-Shetland Basin > Judd Basin > Block 204/25 > Greater Schiehallion Field > Schiehallion Field (0.99)
- Europe > United Kingdom > Atlantic Margin > West of Shetland > Faroe-Shetland Basin > Judd Basin > Block 204/20 > Greater Schiehallion Field > Schiehallion Field (0.99)
- (7 more...)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Uncertainty > Bayesian Inference (1.00)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Optimization (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (1.00)
- (3 more...)
Abstract Rock physics plays an essential role in geophysical reservoir characterization. It aims to build a bridge between geophysical measurements and in-situ rock and fluid properties. With the advancement of microscopic imaging and computer science, rock physics is transitioning to the digital age. This is referred to as digital rock physics (DRP). DRP provides a nondestructive and efficient way to determine physical rock properties directly from digital images. Over the last decades, it has become a routine tool in reservoir characterization by complementing or replacing expensive and time-consuming laboratory measurements. With the emergence of deep learning, DRP has advanced significantly from image processing to physical simulation. This paper presents an application of deep learning in multiscale fusion of digital rock images. It aims to overcome the trade-off between image resolution and field of view (FoV) by integrating imaging data from multiple sources including (1) 3D microcomputed tomography images at micronscale with a large FoV and (2) 2D scanning electron microscopy images at nanoscale with a small FoV. The reconstructed image integrates information of microstructures at different scales and helps characterize heterogeneous porous rocks more accurately. It is helpful to improve the prediction accuracy of effective rock properties and to have deeper insight into physical processes at pore scale. Data fusion based on deep learning would unlock new pathways for geophysical characterization of porous rocks, with broad implications for various subsurface applications such as groundwater transport, enhanced oil recovery, and geologic carbon sequestration.