A synthetic data generation apparatus according to an embodiment includes a generator for generating synthetic data from an input value, a first discriminator learned to distinguish between actual data and the synthetic data, a second discriminator learned to distinguish between the actual data and the synthetic data while satisfying differential privacy, and a third discriminator learned to distinguish between first synthetic data which is output from the generator learned by the first discriminator and second synthetic data which is output from the generator learned by the second discriminator.

A reflection full waveform inversion method updates separately a density model and a velocity model of a surveyed subsurface formation. The method includes generating a model-based dataset corresponding to the seismic dataset using a velocity model and a density model to calculate an objective function measuring the difference between the seismic dataset and the model-based dataset. A high-wavenumber component of the objective function's gradient is used to update the density model of the surveyed subsurface formation. The model-based dataset is then regenerated using the velocity model and the updated density model, to calculate an updated objective function. The velocity model of the surveyed subsurface formation is then updated using a low-wavenumber component of the updated objective function's gradient. A structural image of the subsurface formation is generated using the updated velocity model.

A computer-implemented method for augmented inversion and uncertainty quantification for characterizing geophysical bodies is disclosed. The method includes machine-learning-augmented inversion that also facilitates the characterization of uncertainties in geophysical bodies. The method may further estimate wavelets without a well-log calibration, thereby enabling a pre-discovery exploration phase when well log data is unavailable. The machine learning component incorporates a priori knowledge about the subsurface and physics, such as distributions of expected rock types and rock properties, geological structures, and wavelets, through learning from examples. The methodology also allows for conditioning the characterization with the information extracted a priori about the geobodies, such as probabilities of rock types, using other analysis tools. Thus, the conditioning strategy may make the inversion more robust even when a priori distributions are not well balanced. Using the method, a scenario testing workflow may evaluate different candidate subsurface models, facilitating the management of uncertainty in decision-making processes.

A method, and a system for implementing the method, are disclosed wherein coordinates of survey region are used to locate small pieces of a seismic wave model, usually defined by their nodes (or vertices) and contain information about physical properties, such as liquid or solid, density, velocity that seismic waves propagates in it; and connects them to the appropriate source and receiver sensor. In particular, the method and system disclosed, generates a multi-layer mapping of the survey region by decomposing the survey region into cubes containing small pieces of seismic wave models (the elements), as well as source and receiver location. Those cubes are then indexed depending upon their location and the elements, sources and receivers are assigned to a particular cube thereby creating a multi-layer relationship between the survey region map, the cube map, the elements map, as well as the source and receiver locations.

Methods and systems are disclosed for updating a seismic velocity model of a subterranean region of interest. The method includes receiving an observed seismic dataset and a seismic velocity model, and generating a simulated seismic dataset based on the seismic velocity model and the geometry of the observed seismic dataset, wherein each dataset is composed of a plurality of seismic traces. The method further includes determining a transformed observed seismic dataset and a transformed simulated seismic dataset by determining the instantaneous frequency of at least one member of the plurality of observed seismic traces; and at least one member of the plurality of simulated seismic traces. The method still further includes forming an objective function based on the transformed observed seismic dataset and the transformed simulated seismic dataset and determining an updated seismic velocity model based on an extremum of the objective function.

Systems and methods include a computer-implemented method for characterizing low-permeability reservoirs by using numerical models. A numerical model modeling production of a well is prepared using reservoir data and well data. The numerical model is updated, including adjusting numerical model properties, until results of performing a quality assurance/quality control check indicate that the numerical model is within acceptable limits. Pressure derivatives are extracted from a transient test to create a functional numerical model. Simulations are run on the functional numerical model and reservoir features and properties are adjusted until acceptable results are achieved on: 1) a pressure match between pressures modeled in the functional numerical model and transient pressures of the well, and 2) a log-log plot derivative match between a pressure derivative of the functional numerical model and a pressure derivative of the transient pressures of the well. A simulation output that is based on the simulations is provided.

Seismic exploration of an underground formation uses seismic excitations to probe the formation's properties such as reflectivity that can be imaged using reverse time migration. Using an equal area spherical binning at reflection points improves and simplifies RTM imaging together with adaptability to the data acquisition geometry, while overcoming drawbacks of conventional cylindrical binning.

The present disclosure provides a method and system for stabilizing a Poynting vector of a seismic wavefield. The method includes: adjusting an amplitude of a time derivative of the seismic wavefield, and computing a Poynting vector of the adjusted time derivative of the seismic wavefield to obtain a first Poynting vector, where a difference between the amplitude of the first Poynting vector and the amplitude of a second Poynting vector is within a set range, and the second Poynting vector belongs to the seismic wavefield; and conducting operation on the second Poynting vector and the first Poynting vector to obtain a final Poynting vector of the seismic wavefield. The present disclosure addresses instability of Poynting vectors computation.

A surface wave prospecting method for jointly extracting Rayleigh wave frequency dispersion characteristics in a seismoelectric field. A surface wave prospecting method includes following steps of: acquiring jointly acquired data, where the jointly acquired data includes seismic wave data and electric field data; carrying out jointly imaging processing on jointly acquired data to obtain a superposed frequency dispersion spectrum; carrying out extraction processing on superposed frequency dispersion spectrum to obtain a frequency dispersion curve, outperforming inversion processing on frequency dispersion curve to obtain a stratum structure profile. As seismic wave data and electric field data are adopted to carry out combined imaging processing to obtain superposed frequency dispersion spectrum, multi-mode frequency dispersion curve is extracted, multiplicity of solutions of inversion is greatly reduced during inversion, precision and stability of surface wave prospecting are greatly improved.

A method may include obtaining seismic data regarding a geological region of interest. The method may further include obtaining a property model regarding the geological region of interest. The method may further include determining an adjoint migration operator based on the property model. The method may further include updating the property model using the seismic data and a conjugate gradient solver in a least-squares reverse time migration to produce a first updated property model. The conjugate gradient solver is based on the adjoint migration operator. The method may further include updating the first updated property model using a threshold shrinkage function to produce a second updated property model. The threshold shrinkage function comprises a sign function and a maximum function that are applied to the first updated property model. The method may further include generating a seismic image of the geological region of interest using the second updated property model.