Ocean Framework, C#, .NET, C++
Facies modeling is one of the pillars of geological reservoir modeling. It defines the spatial distribution of rock types and largely controls the dynamic behavior of the reservoir in the subsequent stages of simulation and history matching. Petrel natively offers a consolidated set of facies modeling algorithms, such as SIS (Sequential Indicator Simulation) and object-based methods, but certain geological contexts require more sophisticated techniques to represent contact relationships and transition rules between facies in a physically consistent way.
Pluri-Gaussian simulation is a geostatistical technique widely recognized for modeling facies distributions by truncating correlated Gaussian random fields, allowing explicit control over neighborhood relationships and contacts between lithofacies. It is particularly relevant in carbonate reservoirs and in depositional environments with well-defined transition rules, exactly the scenarios that are highly relevant to Petrobras' portfolio. The absence of this algorithm natively in Petrel forced the technical team to run it in external tools, with the resulting loss of fluidity and traceability in the modeling workflow.
DeepSoft developed Facies Modeling Extensions, a BR-Kalman module that extends Petrel's facies modeling capabilities by adding the Pluri-Gaussian simulation algorithm. The solution is built on top of the Ocean Framework, leveraging the platform's extension points to integrate the new algorithm into the standard Property Modeling workflows, so that users can apply it like any other native method: from the property modeling dialog, with access to the grids, regions, and well data already available in the project.
The implementation covers the definition of the underlying Gaussian random fields, the truncation rules that determine how the continuous fields are converted into discrete categories, and support for spatially variable proportions, allowing the expected distribution of each facies to follow geological trends such as vertical or lateral variations across the reservoir. The result is a facies distribution that simultaneously honors well data, the local proportions defined by the geologist, and the expected contact relationships.
As a BR-Kalman module, Facies Modeling Extensions integrates directly with the ensemble-based history matching workflow, allowing the Pluri-Gaussian parameterization to be updated by the data assimilation process rather than used only as a static input to the model.
Pluri-Gaussian simulation became available inside Petrel as an additional facies modeling method, without the need for external tools or intermediate exports.
Explicit control over contact relationships, transition rules, and spatially variable proportions produces models that are more consistent with the geological interpretation, especially in carbonate reservoirs.
The Pluri-Gaussian parameterization is compatible with BR-Kalman's ensemble workflow, enabling the calibration of facies models from dynamic data such as production, 4D seismic, and well logs.