A method and apparatus for automated seismic interpretation (ASI), including: obtaining trained models comprising a geologic scenario from a model repository, wherein the trained models comprise executable code; obtaining test data comprising geophysical data for a subsurface region; and performing an inference on the test data with the trained models to generate a feature probability map representative of subsurface features. A method and apparatus for machine learning, including: an ASI model; a training dataset comprising seismic images and a plurality of data portions; a plurality of memory locations, each comprising a replication of the ASI model and a different data portion of the training dataset; a plurality of data augmentation modules, each identified with one of the plurality of memory locations; a training module configured to receive output from the plurality of data augmentation modules; and a model repository configured to receive updated models from the training module.

A subterranean earth formation is evaluated by running a process with a logging tool residing in a borehole in the earth formation to collect shallow measurements of a property of the formation and deep measurements of the property of the formation. An inversion is performed on the shallow measurements to produce a group of possible formation models that fit the shallow measurements. A machine-learning algorithm is applied to estimate the shallow formation structure, using the group of possible formation models that fit the shallow measurements, to produce a shallow formation structure. An inversion is performed on the deep measurements to produce a group of possible formation models that fit the deep measurements. The shallow formation structure is expanded using the group of possible formation models that fit the deep measurements to produce a deep formation structure.

Systems and methods of determining an equivalent black oil model are disclosed. In one embodiment, a method of determining an equivalent black oil model of a reservoir includes generating three-dimensional PVT properties using a compositional model, calculating original fluid in place and reservoir performance characteristics from the three-dimensional PVT properties, and converting the three-dimensional PVT properties to a two-dimensional PVT table. The method further includes, until an equivalency metric is satisfied, generating one or more grouped PVT property tables, initializing a black oil model with the one or more grouped PVT property tables, calculating estimated fluid in place and estimated reservoir performance characteristics using the black oil model, and comparing the estimated fluid in place and the estimated reservoir performance characteristics with the original fluid in place and the reservoir performance characteristics to determine whether the equivalency metric is satisfied.

Embodiments herein relate to identifying occurrence of a trigger condition related to an existing physical well. Embodiments further relate to simulating, based on identification of the occurrence of the trigger condition, a plurality of computer-simulated ancillary wells in a vicinity of the existing physical well. Embodiments further relate to determining one or more simulated parameters related to respective ones of the plurality of computer-simulated ancillary wells. Embodiments further relate to determining, based on the one or more simulated parameters, a parameter of a sidetrack well that is to be related to the existing physical well. Embodiments further relate to outputting an indication of the parameter of the sidetrack well. Other embodiments may be described or claimed.

Provided are embodiments for hydrocarbon reservoir development that include the following: identifying proposed well locations within a reservoir boundary, for each location, developing a well plan by: (a) identifying layers of the reservoir located below the proposed location; (b) iteratively assessing the layers (from deepest to shallowest) to identify a deepest “suitable” layer that is not dry, congested, or unsuitable for gas production; and (c) performing the following for the identified layer and the location: (i) determining a borehole configuration for the location; (ii) determining a completion type for the location; and (iii) determining a stimulation treatment for the location, where a well plan for the location (e.g., for use in developing the reservoir) is generated that specifies some or all of a well location, the target layer, a borehole configuration, a completion type, and a stimulation treatment that corresponds to those determined for the proposed well location.

Disclosed is a method of conditioning one or more parametric models. The method comprises obtaining a plurality of candidate parametric models, each describing a sequence of domains characterising a subsurface region and determining whether each sequence of domains described by one or more of said candidate parametric models is a valid sequence of domains. For each valid sequence of domains, each candidate parametric model describing that valid sequence of domains (or a subset of these models) is conditioned simultaneously, for example by using an Ensemble Kalman Filter or artificial neural network.

A method for performing a fracturing operation in a subterranean formation of a field. The method includes obtaining, during the fracturing operation, distributed optical fiber data from a downhole sensor of a treatment well in the subterranean formation, and determining, based on the distributed optical fiber data, an active perforation location from a number of pre-determined perforation locations of the treatment well. The active perforation location is a location of fluid flow into the subterranean formation during the fracturing operation. The method further includes generating, based at least on the active perforation location, a fracturing model for the subterranean formation, and performing, based on the fracturing model, modeling of the fracturing operation to generate a modeling result.

Systems and methods include a method for generating a real-time reservoir porosity log. Historical gas-porosity data is received from previously-drilled and logged wells. The historical gas-porosity data identifies relationships between gas measurements obtained during drilling and reservoir porosity determined after drilling. A gas-porosity model is trained using machine learning and the historical gas-porosity data. Real-time gas measurements are obtained during drilling of a new well. A real-time reservoir porosity log is generated for the new well using the gas-porosity model and real-time gas measurements.

This disclosure presents processes and systems for generating an image of a subterranean formation from seismic data recorded in a seismic survey of the subterranean formation. The seismic data is contaminated with low frequency noise in a low frequency band. Processes and systems reconstruct seismic data in the low frequency band of the seismic data to obtain low frequency reconstructed seismic data that is free of the low frequency noise. The low frequency reconstructed seismic data is used to construct a velocity model of the subterranean formation. The velocity model and the low frequency reconstructed seismic data are used to generate an image of the subterranean formation that reveals structures of the subterranean formation without contamination from the low frequency noise.

A system for determining pressure in a hydraulic fracturing system for a well includes a processing module executing code and configured to receive a plurality of input parameters. The processing module can predict either a bottomhole pressure, based on statistical predictions and physics-based predictions, or a surface pressure based on the predicted bottomhole pressure.