A method for training a backpropagation-enabled regression process is used for predicting values of an attribute of subsurface data. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A predicted value of the attribute has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.

Aspects of the technology described herein identify geologic features within seismic data using modern computer analysis. An initial step is the development of training data for the machine classifier. The training data comprises an image of seismic data paired with a label identifying points of interest that the classifier should identify within raw data. Once the training data is generated, a classifier can be trained to identify areas of interest in unlabeled seismic images. The classifier can take the form of a deep neural network, such as a U-net. Aspects of the technology described herein utilize a deep neural network architecture that is optimized to detect broad and flat features in seismic images that may go undetected by typical neural networks in use. The architecture can include a group of layers that perform aspect ratio compression and simultaneous comparison of images across multiple aspect ratio scales.

The present disclosure relates to systems, methods, and non-transitory computer-readable media for dynamically utilizing offset drill-well data generated within a threshold geographic area to determine formation-top trends and identify formation-top depths at a subject drill-well site. To do so, in some embodiments, the disclosed systems estimate a variogram for observed formation-top depths of a subset of offset drill-wells, and, in turn, map a predicted response from the estimated variogram. For example, using weighted combinations (e.g., with Kriging weights) of the formation-top depths of the subset of offset drill-wells, the disclosed systems can map a continuous surface of a formation and identify a top-depth thereof. Moreover, the disclosed system can do so for multiple formations at the subject drill-well site, and (in real-time in response to a user input) provide for display at a client device, the associated formation-top depths, various predicted drilling events and/or predicted drilling metrics.

A system can receive seismic data that can correlate to a subterranean formation. The system can derive a set of seismic attributes from the seismic data. The seismic attributes can include discontinuity-along-dip. The system can determine parameterized results by analyzing the seismic data and the seismic attributes using a deep learning neural network. The deep learning neural network can include a dilation module. The system can determine one or more fault probabilities of the subterranean formation using the parameterized results. The system can output the fault probabilities for use in a hydrocarbon exploration operation.

Methods, computing systems, and computer-readable media for interpreting seismic data, of which the method includes receiving seismic data representing a subterranean volume, and determining a feature-likelihood attribute of at least a portion of a section of the seismic data. The feature-likelihood attribute comprises a value for elements of the section, the value being based on a likelihood that the element represents part of a subterranean feature. The method also includes identifying contours of the subterranean feature based in part on the feature-likelihood attribute of the section, and determining a polygonal line that approximates the subterranean feature.

A lithology prediction that uses a geological age model as an input to a machine learning model. The geological age model is capable of separating and recoding different seismic packages derived from the horizon interpretation. Once the machine learning model has been trained, a validation may be performed to determine the quality of the machine learning model. The quality may be improved by refining the training of the machine learning model. The lithology prediction generated by the machine learning model that utilizes the geological age model provides an improved lithology prediction that more accurately reflects the subterranean formation of an area of interest.

In contrast to existing methods wherein derived horizons are interpreted in isolation, the disclosure provides a process that does not interpret patches themselves but determines the relationships between patches, in order to associate and link patches to derive a holistic geological interpretation. Predefined patches, such as from a pre-interpreted suite, are received as inputs to determine the relationships and derive an interpretation for a complete volume. In one aspect the disclosure provides an automated method of generating a geological age model for a subterranean area. In one example, the automated method includes: (1) abstracting seismic data of a subsurface into a limited number of patches, (2) abstracting the patches by defining patch-links between the patches, and (3) generating a geological age model of the subsurface by solving for the relative geological age of each of the patches using the patch-links.

A system and method for acoustically profiling a wellbore while drilling, and which identifies depths in the wellbore where the wellbore diameter is enlarged or has highly fractured sidewalls. These depths are identified based on monitoring either travel time or signal strength of acoustic signals that propagate axially in the wellbore. Correlating wellbore depth to travel time of a signal traveling downhole inside of a drill string and uphole outside of the drill string yields an average signal velocity in the wellbore. Depths having a lower average signal velocity indicate where the wellbore diameter is enlarged or has highly fractured sidewalls. These depths are also identified by generating separate acoustic signals inside and outside of the drill string, comparing signal strengths of signals reflected from the wellbore bottom, and identifying the depths based on where there is an offset in the strengths of the reflected signals.

The present invention belongs to the field of treatment for data identification and recording carriers, and specifically relates to a method and system for evaluating the filling characteristics of a deep paleokarst reservoir through well-to-seismic integration, which aims to solve the problems that by adopting the existing petroleum exploration technology, the reservoir with fast lateral change cannot be predicted, and the development characteristics of a carbonate cave type reservoir in a large-scale complex basin cannot be identified. The method comprises: acquiring data of standardized logging curves; obtaining a high-precision 3D seismic amplitude data body by mixed-phase wavelet estimation and maximum posteriori deconvolution and enhancing diffusion filtering. According to the method and the system, the effect of identifying the development characteristics of the carbonate karst cave type reservoir in the large-scale complex basin can be achieved, and the characterization precision is improved.

Geologic modeling methods and systems disclosed herein employ an improved simulation gridding technique that optimizes simulation efficiency by balancing the computational burdens associated with remeshing against the performance benefits of doing so. One method embodiment includes: (a) obtaining a geologic model representing a subsurface region as a mesh of cells, at least some of the cells in the mesh having one or more interfaces representing boundaries of subsurface structures including at least one fracture; (b) determining a fracture extension to the at least one fracture; (c) evaluating whether the fracture extension is collocated with, or is proximate to, an existing cell interface, and using the existing cell interface if appropriate or creating a new cell interface if not; and (d) outputting the updated version of the geologic model.