Schlumberger Technology Corporation

Westerngeco L.l.p.

Schlumberger Canada Limited

Services Petroliers Schlumberger

Prad Research and Development Limited

Geoquest Systems B.v.

Logined B.v.

Victor Aarre

Nader Salman

Michael Hermann Nickel

Hilde Grude Borgos

Sergio Courtade

Patrick Smith

Christophe Laurent

Donatella Astratti

Reidar Devold Midtun

Edo Hoekstra

Bradley Clark Wallet

Jarl Eirik Tronerud

A method for processing an image includes receiving an input image including a signal and noise, and generating a filtered image based on the input image by removing at least a portion of the noise from the input image. A portion of the signal is also removed from the input image. The method further includes generating a residual image based on the input image. The residual image comprises the at least a portion of the noise and the portion of the signal that are removed from the input image to generate the filtered image. The method also includes identifying at least some of the portion of the signal that is in the residual image, and inserting the at least some of the portion of the signal identified in the residual image into the filtered image.

Residual signal detection for noise attenuation

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.

Geophysical deep learning

A method can include receiving n-dimensional data where n is equal at least three; analyzing a plurality of two-dimensional slices of the n-dimensional data to determine characteristic information with respect to a plurality of feature indexes for a feature in the n-dimensional data; and, based at least in part on the characteristic information, associating the feature with one of the feature indexes.

Feature index-based feature detection

A method, apparatus, and program product analyze time-series data such as seismic data collected from a subsurface formation by splitting a time-series data set such as an individual seismic trace into a plurality of spectral components, each having an associated frequency, determining an instantaneous frequency for each spectral component, determining a frequency difference for each spectral component based at least in part on the associated and instantaneous frequencies therefor, and determining a tuning parameter based at least in part on the determined frequency difference of each spectral component. Doing so enables, for example, thin-bed structures in the subsurface formation to be identified, and in some instances, thicknesses of such structures to be determined.

Thin bed tuning frequency and thickness estimation

A method can include receiving data sets where each of the data sets corresponds to one of a plurality of individual emitter-detector arrangements; calculating a multi-dimensional similarity metric for one of the data sets; and, based at least in part on the multi-dimensional similarity metric, assessing the one data set.

Seismic survey analysis

One or more computer-readable media include computer-executable instructions to instruct a computing system to, for a given time, solve the elastic wave equation in tensorial form for a geological reservoir model subject to stated conditions; and adjust a grid associated with the geological reservoir model based on solution of the elastic wave equation in tensorial form for the given time. Various other apparatuses, systems, methods, etc., are also disclosed.

Reservoir grid conversion for scaling strain

A method can include receiving seismic data that has an associated bandwidth; for a number of frequency bands, for a number of frequency bands, iteratively filtering and adjusting the seismic data by applying band-pass filters to extract information associated with each of the frequency bands where the adjusting the seismic data includes, after each iteration, subtracting extracted information from the seismic data prior to a subsequent iteration; balancing the extracted information to generate spectrally balanced seismic data; and outputting the spectrally balanced seismic data.

Seismic spectral balancing

A method can include receiving data that includes signal data and coherent noise data where the signal data includes signal data that corresponds to a multidimensional physical structure; generating filtered data by filtering at least a portion of the data to attenuate at least a portion of the coherent noise data by applying a multidimensional geometric coherent noise model defined by at least one geometric parameter; and assessing a portion of the signal data in the generated filtered data to characterize the multidimensional physical structure.

Noise attenuation

A method can include providing seismic data values for a subsurface region that includes a reflector; determining a gradient magnitude value based on at least a portion of the seismic data values; normalizing the gradient magnitude value using a nonlinear normalization equation that includes a gradient magnitude variable divided by a normalization variable raised to a power that depends on an adjustable parameter; and outputting the normalized gradient magnitude value. Various other apparatuses, systems, methods, etc., are also disclosed.

Normalization seismic attribute

A method can include receiving a request for a dip value for a location within a sub-surface volume; responsive to the request, locating complementary dip values for locations within the sub-surface volume where the location for the requested dip value lies within bounds defined by the locations for the complementary dips values; and interpolating a value for the requested dip value based on the complementary dip values. Various other apparatuses, systems, methods, etc., are also disclosed.

Dip seismic attributes

Implementations of discontinuity detection are described. In one possible embodiment, information associated with a computer-generated model is accessed. A location associated with a feature in the model can be reviewed for possible discontinuity properties by comparing an orientation associated with the location to an orientation of a true normal to the feature associated with a specific point nearby the location. In one possible aspect, if the orientation associated with the location and the true normal are within a given range of one another, the location can be flagged as a possible discontinuity in the feature.

Discontinuity detection

Systems and methods perform consistent dip estimation for seismic imaging. An example system applies global consistency constraints during iterative volume dip estimation of a seismic volume to improve upon conventional dip estimation methods. With each iteration, the system applies single and joint dip constraints, checking local dip estimates for reciprocity, causality, consistency, and vertical and lateral continuity. At discontinuities in the seismic volume, local divergences are marked with a quality attribute. Upon convergence of the volume dip estimation, the volume may be rendered in 3D, including the discontinuities. In performing volume dip estimation, the system can also provide automatic horizon interpretation and automatic fault detection.

Consistent dip estimation for seismic imaging

A method for processing seismic data. The method may include splitting the seismic data into multiple datasets according to one or more offsets; determining a first shift amount in three or more dimensions of the seismic data between a dataset having a first offset and a dataset having a second offset, determining a second shift amount in the three or more dimensions between the dataset having the second offset and a dataset having a third offset, determining a cumulative shift amount based on a shift of the first shift amount and the second shift amount and determining a corrected dataset based on the dataset having the third offset and the cumulative shift amount.

3D residual binning and flatness error correction

Implementations of various technologies for a method for processing seismic data. In one implementation, the method includes (a) selecting a first trace from a first seismic data set and a second trace from a second seismic data set; (b) extracting one or more features of the same types from the first trace and the second trace; (c) matching the extracted features from the first trace with the extracted features from the second trace; and (d) calculating for a displacement field using one or more of the matching features of the first trace and the second trace.

Displacement field calculation

A method of processing seismic data comprises identifying the value of a first parameter associated with an event in a first set of seismic data. The value of a second parameter associated with a corresponding event in a second sent of seismic data is then obtained using at least one look-up table. The first parameter may be PP travel time with the first data set being a raw PP data set, and the second parameter may be PS travel time with the second data set being a raw PS data set or reflection depth. The invention makes it possible to identify pairs of corresponding PP and PS events in raw data traces. The look-up table(s) are obtained using an assumed model for the velocity of propagation of acoustic energy within the earth. The results of the method may be used in quality control, or to correct the velocity model.

Growing community of inventors

Stavanger, Norway

Victor Aarre