Results
Arnab Dhara discusses his paper, "Physics-guided deep autoencoder to overcome the need for a starting model in full-waveform inversion," in the June issue of The Leading Edge. Arnab proposes employing deep learning as a regularization in full-waveform inversion. He explains why physics-based solutions with machine learning are challenging to develop, how he made it possible to train the network without known answers, and why he tested his approach with the Marmousi and SEAM models. Arnab also shares why this research took over 20 years to build on the initial idea and how he used full-waveform inversion without a starting model. This is a cutting-edge conversation that may represent the future of FWI.
- Information Technology > Artificial Intelligence > Representation & Reasoning (0.70)
- Information Technology > Artificial Intelligence > Machine Learning > Neural Networks > Deep Learning (0.32)
We present a new alternative for the joint inversion of well logs to predict the volumetric and zone parameters in hydrocarbon reservoirs. Porosity, water saturation, shale content, kerogen and matrix volumes are simultaneously estimated with the tool response function constants with a hyperparameter estimation assisted inversion of the total and spectral natural gamma-ray intensity, neutron porosity and resistivity logs. We treat the zone parameters, i.e., the physical properties of rock matrix constituents, shale, kerogen, and pore-fluids, as well as some textural parameters, as hyperparameters and estimate them in a meta-heuristic inversion procedure for the entire processing interval. The selection of inversion unknowns is based on parameter sensitivity tests, which show the automated estimation of several zone parameters is favorable and their possible range can also be specified in advance. In the outer loop of the inversion procedure, we use a real-coded genetic algorithm for the prediction of zone parameters, while we update the volumetric parameters in the inner loop in addition to the fixed values of zone parameters estimated in the previous step. We apply a linearized inversion process in the inner loop, which allows for the quick prediction of volumetric parameters along with their estimation errors from point to point along a borehole. Derived parameters such as hydrocarbon saturation and total organic content show good agreement with core laboratory data. The significance of the inversion method is in that zone parameters are extracted directly from wireline logs, which both improves the solution of the forward problem and reduces the cost of core sampling and laboratory measurements. In a field study, we demonstrate the feasibility of the inversion method using real well logs collected from a Miocene tight gas formation situated in the Derecske Trough, Pannonian Basin, East Hungary.
- Geology > Geological Subdiscipline (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock > Mudrock > Shale (0.71)
- Europe > Slovakia > Pannonian Basin (0.99)
- Europe > Serbia > Pannonian Basin (0.99)
- Europe > Romania > Pannonian Basin (0.99)
- (9 more...)
- Information Technology > Artificial Intelligence > Representation & Reasoning (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Evolutionary Systems (1.00)
During the last few years, there has been an increasing focus on automation solutions for drilling, with many of these solutions already in the market. Artificial Intelligence (AI) techniques are usually part of these solutions. During this SPE Live, we focus on AI systems that are targeting higher levels of automation, some reaching all the way to autonomy. One requirement for such systems to be taken into use is to be trustworthy. But what does this mean and how to achieve it? At the same time, drilling involves multiple providers, which reflects on the data and information flow, hence adding a new level of complexity for any automation solution.
Adaptive laterally constrained inversion of time-domain electromagnetic data using Hierarchical Bayes
Li, Hai (Chinese Academy of Sciences, Chinese Academy of Sciences) | Di, Qingyun (Chinese Academy of Sciences, Chinese Academy of Sciences) | Li, Keying (Chinese Academy of Sciences, Chinese Academy of Sciences, University of Chinese Academy of Sciences)
Laterally constrained inversion (LCI) of time-domain electromagnetic (TEM) data is effective in recovering quasi-layered models, particularly in sedimentary environments. By incorporating lateral constraints, LCI enhances the stability of the inverse problem and improves the resolution of stratified interfaces. However, a limitation of the LCI is the recovery of laterally smooth transitions, even in regions unsupported by the available datasets. Therefore, we have developed an adaptive LCI scheme within a Bayesian framework. Our approach introduces user-defined constraints through a multivariate Gaussian prior, where the variances serve as hyperparameters in a Hierarchical Bayes algorithm. By simultaneously sampling the model parameters and hyperparameters, our scheme allows for varying constraints throughout the model space, selectively preserving lateral constraints that align with the available datasets. We demonstrated the effectiveness of our adaptive LCI scheme through a synthetic example. The inversion results showcase the self-adaptive nature of the strength of constraints, yielding models with smooth lateral transitions while accurately retaining sharp lateral interfaces. An application to field TEM data collected in Laizhou, China, supports the findings from the synthetic example. The adaptive LCI scheme successfully images quasi-layered environments and formations with well-defined lateral interfaces. Moreover, the Bayesian inversion provides a measure of uncertainty, allowing for a comprehensive illustration of the confidence in the inversion results.
- Geology > Mineral (0.93)
- Geology > Sedimentary Geology > Depositional Environment (0.34)
- Oceania > Australia > Western Australia > North West Shelf > Carnarvon Basin > Exmouth Plateau > WA-1-R > Scarborough Field (0.99)
- Europe > Norway (0.91)
- Reservoir Description and Dynamics > Reservoir Characterization > Exploration, development, structural geology (1.00)
- Reservoir Description and Dynamics > Reservoir Simulation > Evaluation of uncertainties (0.93)
- Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (0.79)
- (2 more...)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Uncertainty > Bayesian Inference (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Learning Graphical Models > Directed Networks > Bayesian Learning (1.00)
An integrated approach for sewage diversion: Case of Huayuan mine, Hunan Province, China
Kouadio, Kouao Laurent (Central South University, Hunan Key Laboratory of Nonferrous Resources and Geological Hazards Exploration, Universit Flix Houphout-Boigny) | Liu, Jianxin (Central South University, Hunan Key Laboratory of Nonferrous Resources and Geological Hazards Exploration) | Liu, Wenxiang (Central South University, Guangdong Geological Bureau) | Liu, Rong (Central South University, Hunan Key Laboratory of Nonferrous Resources and Geological Hazards Exploration) | Boukhalfa, Zakaria (Centre de Recherche en Astronomie)
Environment protection is a core priority of many governments in this century. Most environmental problems have diverse causes: emission of greenhouse gases from fossil fuels, resource depletion, or intense mining activities such as the Huayuan manganese mine. The positioning of mining factories and water treatment stations impacts the surrounding groundwater reservoir. As the mine expands, the environmental impact also increases and the previous plan based on monitoring wastewater leakage has become inappropriate. Therefore, to solve this issue, a new study is required to understand the lateral resistivity distribution underground and to define a new station location for water treatment and divert the sewage to that station. In this study, the audio-frequency magnetotelluric method was used. Surveys of two long lines that cross the mining area to its boundaries were carried out. Data was robustly processed and inverted. Based on the inverted models in addition to geological information, drilling inspections, and solid waste distributions map, the integrated interpretation proposed two sites on the top of impermeable layers which constitute a buffer point between the unsafe (high concentration of pollutants) and the safe zones in the northwestern part of the mine. From the resistivity distribution combined with the water quality analysis, a relationship between fault structures reveals an interconnected conductive zone in the southeastern part. Being, the main channels for water circulating underground, these conductive zones delineate the main groundwater reservoir with a clastic aquifer layer. However, close to factories, water from faults contains solid wastes thereby making the groundwater in that zone non-potable, unlike the safety zone located in the northwestern part. To conclude, this workflow could become a field guide to improve the environment of mines and the deployment of hydrogeological drilling in a safe area.
- North America > United States (1.00)
- Asia > China > Hunan Province (0.40)
- Geology > Mineral (1.00)
- Geology > Structural Geology > Fault (0.93)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.66)
- (2 more...)
- Materials > Metals & Mining (1.00)
- Energy > Oil & Gas > Upstream (1.00)
- Water & Waste Management > Water Management > Lifecycle > Treatment (0.54)
- Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Exploration, development, structural geology (1.00)
- Health, Safety, Environment & Sustainability > Environment > Water use, produced water discharge and disposal (0.88)
- Information Technology > Data Science (1.00)
- Information Technology > Artificial Intelligence > Representation & Reasoning (1.00)
Multichannel deconvolution with a high-frequency structural regularization
Wang, Pengfei (China University of Petroleum) | Zhao, Dongfeng (China University of Petroleum) | Niu, Yue (National Engineering Research Center for Oil and Gas Exploration Computer Software) | Li, Guofa (China University of Petroleum) | Gu, Weiwei (China University of Petroleum)
The resolution of seismic data determines the ability to characterize stratigraphic features from observed seismic record. Sparse spike inversion (SSI) as an important processing method can effectively improve the band-limited property of the seismic data. However, the approch ignores the spatial information along seismic traces, which causes the unreliability of the reconstructed high-resolution data. In this article, we have developed a high-frequency structure constrained multichannel deconvolution (HFSC-MD) to alleviate this issue. This method allows the cost function to incorporate high-frequency spatial information in the form of prediction-error filter (PEF), to regularize the components of the result beyond the original frequency. The PEF also called high-frequency structural characterization operator (HFRSC operator), is estimated from the mapping relationship of low and high-frequency components. We adopt the alternating direction method of multipliers (ADMM) to solve the cost function in HFSC-MD. Synthetic and field data demonstrate that the proposed method recovers more reliable high-resolution data, and enriches the reflective structures.
- Geophysics > Seismic Surveying > Seismic Processing (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (0.34)
- Information Technology > Artificial Intelligence > Representation & Reasoning (0.54)
- Information Technology > Artificial Intelligence > Machine Learning (0.46)
Elastic Full Waveform Inversion (EFWI) is a process used to estimate subsurface properties by fitting seismic data while satisfying wave propagation physics. The problem is formulated as a least squares data fitting minimization problem with two sets of constraints: PDE constraints governing elastic wave propagation and physical model constraints implementing prior information. The Alternating Direction Method of Multipliers (ADMM) is used to solve the problem, resulting in iterative algorithm with well-conditioned subproblems. Although wavefield reconstruction is the most challenging part of the iteration, sparse linear algebra techniques can be used for moderate-sized problems and frequency domain formulations. The Hessian matrix is blocky with diagonal blocks, making model updates fast. Gradient ascent is used to update Lagrange multipliers by summing PDE violations. Various numerical examples are used to investigate algorithmic components, including model parameterizations, physical model constraints, the role of the Hessian matrix in suppressing interparameter cross-talk, computational efficiency with the source sketching method, and the effect of noise and near surface effects.
- Geophysics > Seismic Surveying > Seismic Processing (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
During the last few years, there has been an increasing focus on automation solutions for drilling, with many of these solutions already in the market. Artificial Intelligence (AI) techniques are usually part of these solutions. During this SPE Live, we focus on AI systems that are targeting higher levels of automation, some reaching all the way to autonomy. One requirement for such systems to be taken into use is to be trustworthy. But what does this mean and how to achieve it? At the same time, drilling involves multiple providers, which reflects on the data and information flow, hence adding a new level of complexity for any automation solution.
- Information Technology > Artificial Intelligence > Robots (0.85)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Agents (0.49)
ABSTRACT Prismatic reflections in seismic data carry abundant information about subsurface steeply dipping structures, such as salt flanks or near-vertical faults, playing an important role in delineating these structures when effectively used. Conventional linear least-squares reverse time migration (L-LSRTM) fails to use prismatic waves due to the first-order Born approximation, resulting in a blurry image of steep interfaces. We develop a nonlinear LSRTM (NL-LSRTM) method to take advantage of prismatic waves for the detailed characterization of subsurface steeply dipping structures. Compared with current least-squares migration methods of prismatic waves, our NL-LSRTM is nonlinear and thus avoids the challenging extraction of prismatic waves or the prior knowledge of L-LSRTM results. The gradient of NL-LSRTM consists of the primary and prismatic imaging terms, which can accurately project observed primary and prismatic waves into the image domain for the simultaneous depiction of near-horizontal and near-vertical structures. However, we find that the full Hessian-based Newton normal equation has two similar terms, which prompts us to make further comparison between the Newton normal equation and our NL-LSRTM. We determine that the Newton normal equation is problematic when applied to the migration problem because the primary reflections in the seismic records will be incorrectly projected into the image along the prismatic wavepath, resulting in an artifact-contaminated image. In contrast, the nonlinear data-fitting process included in the NL-LSRTM contributes to balancing the amplitudes of primary and prismatic imaging results, thus making NL-LSRTM produce superior images compared with the Newton normal equation. Several numerical tests validate the applicability and robustness of NL-LSRTM for the delineation of steeply dipping structures and illustrate that the imaging results are much better than the conventional L-LSRTM.
- Geophysics > Seismic Surveying > Seismic Processing > Seismic Migration (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (0.46)