Abstract: A computational stratigraphy model may be run for M mini-steps to simulate changes in a subsurface representation across M mini-steps (from 0-th subsurface representation to M-th subsurface representation), with a mini-step corresponding to a mini-time duration. The subsurface representation after individual steps may be characterized by a set of computational stratigraphy model variables. Some or all of the computational stratigraphy model variables from running of the computational stratigraphy model may be provided as input to a machine learning model. The machine learning model may predict changes to the subsurface representation over a step corresponding to a time duration longer than the mini-time duration and output a predicted subsurface representation. The subsurface representation may be updated based on the predicted subsurface representation outputted by the machine learning model.

Abstract: Data in physical space may be converted to layer space before performing modeling to generate one or more subsurface representations. Computational stratigraphy model representations that define subsurface configurations as a function of depth in the physical space may be converted to the layer space so that the subsurface configurations are defined as a function of layers. Conditioning information that defines conditioning characteristics as the function of depth in the physical space may be converted to the layer space so that the conditioning characteristics are defined as the function of layers. Modeling may be performed in the layer space to generate subsurface representations within layer space, and the subsurface representations may be converted into the physical space.

Abstract: Data in physical space may be converted to layer space before performing modeling to generate one or more subsurface representations. Computational stratigraphy model representations that define subsurface configurations as a function of depth in the physical space may be converted to the layer space so that the subsurface configurations are defined as a function of layers. Conditioning information that defines conditioning characteristics as the function of depth in the physical space may be converted to the layer space so that the conditioning characteristics are defined as the function of layers. Modeling may be performed in the layer space to generate subsurface representations within layer space, and the subsurface representations may be converted into the physical space.

Abstract: Data in physical space may be converted to layer space before performing modeling to generate one or more subsurface representations. Computational stratigraphy model representations that define subsurface configurations as a function of depth in the physical space may be converted to the layer space so that the subsurface configurations are defined as a function of layers. Conditioning information that defines conditioning characteristics as the function of depth in the physical space may be converted to the layer space so that the conditioning characteristics are defined as the function of layers. Modeling may be performed in the layer space to generate subsurface representations within layer space, and the subsurface representations may be converted into the physical space.

Abstract: A computer implemented method for identifying reservoir facies in a subsurface region includes obtaining a set of seismic data points of both petrophysical and geophysical parameters relating to the subsurface region, identifying one or more correlated clusters of petrophysical parameters, generating, from the one or more correlated clusters of petrophysical parameters, one or more corresponding multi-dimensional clusters of seismic data points, storing, in a facies database, a multi-dimensional cluster center point for at least one multi-dimensional clusters, and recursively splitting the multi-dimensional clusters into distinct sub-clusters of seismic data points corresponding to facies types.

Abstract: A computational stratigraphy model may be run for M mini-steps to simulate changes in a subsurface representation across M mini-steps (from 0-th subsurface representation to M-th subsurface representation), with a mini-step corresponding to a mini-time duration. The subsurface representation after individual steps may be characterized by a set of computational stratigraphy model variables. Some or all of the computational stratigraphy model variables from running of the computational stratigraphy model may be provided as input to a machine learning model. The machine learning model may predict changes to the subsurface representation over a step corresponding to a time duration longer than the mini-time duration and output a predicted subsurface representation. The subsurface representation may be updated based on the predicted subsurface representation outputted by the machine learning model.

Abstract: A computer implemented method for identifying reservoir facies in a subsurface region includes obtaining a set of seismic data points of both petrophysical and geophysical parameters relating to the subsurface region, identifying one or more correlated clusters of petrophysical parameters, generating, from the one or more correlated clusters of petrophysical parameters, one or more corresponding multi-dimensional clusters of seismic data points, storing, in a facies database, a multi-dimensional cluster center point for at least one multi-dimensional clusters, and recursively splitting the multi-dimensional clusters into distinct sub-clusters of seismic data points corresponding to facies types.

Abstract: An enhanced multi-point statistical (MPS) simulation is disclosed. A multiple-grid simulation approach is used which has been modified from a conventional MPS approach to decrease the size of a data search template, saving a significant amount of memory and cpu-time during the simulation. Features used to decrease the size of the data search template include: (1) using intermediary sub-grids in the multiple-grid simulation approach, and (2) selecting a data template that is preferentially constituted by previously simulated nodes. The combination of these features allows saving a significant amount of memory and cpu-time over previous MPS algorithms, yet ensures that large-scale training structures are captured and exported to the simulation exercise.