A method for a physics-driven deep learning-based inversion coupled to fluid flow simulators may include obtaining measured data for a subsurface region, obtaining prior subsurface data for the subsurface region, and obtaining a physics-driven standard regularized joint inversion for at least two model parameters. The method may further include obtaining a case-based deep learning inversion characterized by a contracting path and an expansive path. The method may further include forming the physics-driven deep learning inversion with the physics-driven standard regularized joint inversion, the case-based deep learning inversion, and a coupling operator based on a penalty function. The method may further include forming a feedback loop between the physics-driven standard regularized joint inversion and the case-based deep learning inversion for re-training the case-based deep learning inversion. The method may further include generating an inversion solution for reservoir monitoring.

A method, apparatus, and program product may evaluate a field by receiving a dataset including well measurements collected from a plurality of wells in a field, generating a synthetic dataset from the received dataset by computing a plurality of synthetic samples from the received dataset using a self-organized may (SOM), and propagating one or more models generated from the synthetic dataset to the plurality of wells.

A system, method and program product for implementing a machine learning platform that processes a data map having feature and operational information. A system is disclosed that includes an interpretable machine learning model that generates a function in response to an inputted data map, wherein the data map includes feature data and operational data over a region of interest, and wherein the function relates a set of predictive variables to one or more response variables; an integration/interpolation system that generates the data map from a set of disparate data sources; and an analysis system that evaluates the function to predict outcomes at unique points in the region of interest.

A method can include selecting a type of geophysical data; selecting a type of algorithm; generating synthetic geophysical data based at least in part on the algorithm; training a deep learning framework based at least in part on the synthetic geophysical data to generate a trained deep learning framework; receiving acquired geophysical data for a geologic environment; implementing the trained deep learning framework to generate interpretation results for the acquired geophysical data; and outputting the interpretation results.

Computer-implemented methods, media, and systems for characterization and geomodeling of three-dimensional (3D) vugular pore system (VPS) in carbonate reservoir are disclosed. One example method includes determining an occurrence of a VPS in multiple layers of a carbonate reservoir based on data collected from multiple wells in the carbonate reservoir. A spatial distribution of multiple VPS intensity classes of the VPS is determined using at least one of well log data, borehole image log data, production log data, or seismic acoustic impedance data from the multiple wells. A 3D VPS intensity distribution model of the VPS is constructed using the spatial distribution of the multiple VPS intensity classes of the VPS. The 3D VPS intensity distribution model is provided for at least one of reservoir volumetric estimation, reservoir history matching, or reservoir quality prediction of the carbonate reservoir.

A computer-implemented method may include obtaining well logs data pertaining to a well of interest. The method may further include training a physics-constrained machine learning (PCML) model using the obtained well logs data as inputs. The method may further include outputting one or more sonic logs and mechanical properties of interest determined by using the trained PCML model and the obtained well logs data for the well of interest. The method may further include updating the determined sonic logs and mechanical properties of interest based on a breakout model and field breakout data for the well of interest. The method may further include outputting the final sonic logs for the well of interest. The method may further include determining one or more mechanical properties for well planning based on the final sonic logs for the well of interest.

Method, apparatus, and computer program product are provided for retrieving analogues using topological knowledge representation (TKR). In some embodiments, a TKR input query is built and/or validated using a domain-specific knowledge base (KB). A search database containing candidate analogues and corresponding pre-built TKRs is then searched to retrieve at least one analogue of the TKR input query using statistical analysis. In some embodiments, a system may build the TKR input query based on a seismic dataset. For example, the system may receive a seismic dataset, segment the seismic dataset and classify each region using a computer vision (CV) database and the KB, and build the TKR input query based on the segmented and classified seismic dataset. In some embodiments, the TKR input query may be input and/or edited by a user. For example, the TKR input query may be input and/or edited by the user and validated using the KB.

Storing and retrieving subsurface rock physical property models for well drilling by generating estimated subsurface rock physical property models based on a respective set of seismic data, determining a stratigraphic classification for each model based on seismic interpretation data; storing the models associated classification in a database, determining characteristics of the seismic data used to generate the models, selecting from the database a subset of the models based on a similarity of the stratigraphic classification, determining measures of similarity between the characteristics of the seismic data used to generate each of subset of the models and a new set of seismic data and retrieving an estimated subsurface rock physical property model from the subset models based on the similarity measures.

A system can include a back-end framework that validates a reservoir model and a production network model using field data to generate a validated integrated model based at least in part on the validated reservoir model and the validated production network model; and a front-end web portal application that receives integrated model simulation data from execution of a simulation using the validated integrated model.

Techniques to match a signature in seismic data with a seismic attribute space. A method includes automatically selecting a first plurality of seismic attributes corresponding to seismic data as first selected seismic attributes, combining the first selected seismic attributes into a first realization of attributes, performing a first cluster analysis on the first realization of attributes to generate a first clustered volume, selecting a region of interest (ROI) in the seismic data, projecting the ROI onto the first clustered volume to generate a first signature, determining a first level of correlation between the ROI and the first signature, and determining whether the first level of correlation between the ROI and the first signature exceeds a predetermined threshold and outputting a first correlation volume corresponding to the first signature when the first level of correlation between the ROI and the first signature exceeds the predetermined threshold.