A system for estimating a rock property away from a well may include one or more hardware processors configured to access acquired three-dimensional (3D) seismic data that includes seismic traces from a 3D seismic survey of an area of interest. The system may also include a multi-head Convolutional Neural Network (CNN) model. The multi-head CNN model may include a plurality of kernels of various sizes for determining spatial and temporal relationships of the captured 3D seismic data at different resolutions. The multi-head CNN model may be trained to generate an estimated rock property value of a formation zone included in the area of interest, away from the well. The one or more hardware processors are further configured to update a drilling program for a production system based on the estimated rock property value. The drilling program may be executed on a computing device of the production system.

Systems and methods are disclosed that include generating reservoir property profiles corresponding to reservoir properties for pseudo wells based on reservoir data, generating seismic attributes for the pseudo wells, and training a machine learning model by comparing the reservoir property profiles against the seismic attributes. In this manner, the machine learning model may be used to predict reservoir properties for use with seismic exploration above a region of a subsurface that contains structural or stratigraphic features conducive to a presence, migration, or accumulation of hydrocarbons.

A method is described for seismic amplitude analysis including receiving a seismic dataset representative of a subsurface volume of interest wherein the seismic dataset includes an angle or angle stack dimension; select a plurality of sets of sub-cubes in the seismic dataset wherein each set of sub-cubes includes a plurality of the angles or the angle stacks; compute standard score statistics for each of the plurality of sub-cubes; identify amplitude variation with angle (AVA) anomalies based on the standard score statistics for each of the set of sub-cubes; classify the AVA anomalies to generate classified AVA anomalies; and displaying, on a user interface, the classified AVA anomalies as a graphical display. The method is executed by a computer system.

A method includes receiving a first transition probability matrix (TPM) of a subsurface region, wherein the TPM defines, for a given lithology at a current depth sample (or micro-layer), a probability of particular lithologies at a next depth sample (or micro-layer), receiving seismic data for the subsurface region, utilizing the first TPM and the seismic data to generate first pseudo wells, calculating a second TPM from the first pseudo wells, determining whether the second TPM is consistent with the first TPM, and utilizing the first pseudo wells to characterize a reservoir in the subsurface region when the second TPM is determined to be consistent with the first TPM.

Aspects of the subject technology relate to systems and methods for identifying values of mud and formation parameters based on measurements gathered by an electromagnetic imager tool through machine learning. One or more regression functions that model mud and formation parameters capable of being identified through an electromagnetic imager tool as a function of possible tool measurements of the electromagnetic imager tool can be generated using a known dataset associated with the electromagnetic imager tool. One or more tool measurements obtained by the electromagnetic imager tool operating to log a wellbore can be gathered. As follows, one or more values of the mud and formation parameters can be identified by applying the one or more regression functions to the one or more tool measurements.

Method for automated and quantitative assessment of multiple direct hydrocarbon indicators (“DHI's”) extracted from seismic data. DHI's are defined in a quantitative way (33), making possible a method of geophysical prospecting based on quantification of DHI anomalies. Instead of working in a particular spatial region of seismic data pre-defined as a hydrocarbon opportunity, the present invention works on entire data volumes derived from the measured seismic data (31), and identifies opportunities based on quantified DHI responses. In some embodiments, a series of algorithms utilizes the geophysical responses that cause DHI's to arise in seismic data to search entire data sets and identify hydrocarbon leads based on the presence of individual and/or combinations of DHI's (34).

A method for improved data-driven estimation of a stimulated reservoir volume may generate an optimized surface that encloses a set of data points including microseismic event data corresponding to a treatment of a subterranean formation. A Delaunay triangulation may be performed on the set of data points to generate a set of polytopes. A Voronoi polygon may be generated about each data point and used to obtain a local density measure that is locally and adaptively determined for each data point. Based on the local density measure, polytopes in the set of polytopes may be discriminated for inclusion in the optimized surface.

A method and apparatus for imaging seismic data includes obtaining an initial model of a subsurface formation, wherein the model includes a plurality of nodes that form at least part of a grid; an initial dip value for the nodes; and a set of origin coordinates for each of the nodes; performing bottom-up ray tracing for each node in the model, resulting in a set of arrival coordinates for each node; identifying a plurality of gathers from the seismic data; for each gather: calculating a set of midpoint coordinates; defining a midpoint vicinity surrounding the set of midpoint coordinates; identifying the nodes having arrival coordinates within the midpoint vicinity; and estimating a unique aperture for each of the gathers based on the respective origin coordinates; storing the estimated apertures in a table; and generating a subsurface volume or image with subsurface reflectors determined with apertures of the respective gathers.

An innovative apparatus and computer implemented methods to obtain values for a set of scalars corresponding to each force and displacement, which may be obtained from acoustical signals captured by sensors of a drill bit while drilling, in a material of known mechanical properties, such as a cement from casing the well, such that the application and use of the scalars in relation to measurements of the mechanics while drilling, such as the acceleration of the bit and motion of the bit captured by sensors such as accelerometers, allow for absolute values of mechanical rock properties to be obtained in rock formations, being drilled through, with otherwise unknown mechanical properties prior to drilling.

Simultaneous inversion of multi-vintage seismic data obtains seismic data for vintages and generates an initial earth model for each vintage. A cost function includes a data norm term having for at least one pair of vintages of seismic data a difference norm between a difference in obtained seismic data for the at least one pair of vintages and a difference in modeled seismic data for the at least one pair of vintages. The cost function also includes a model norm term for each pair of vintages selected from at least three vintages of seismic data. Each model norm term includes a difference norm between earth models for a given pair of vintages. A closure relationship is imposed on all earth models. The earth models are adjusted for the vintages to drive the cost function to a minimum and to produce updated earth models.