This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 202775, “Application of Stratigraphic Forward Modeling to Carbonate-Reservoir Characterization: A New Paradigm From the Albion Research and Development Project,” by Jean Borgomano, Aix-Marseille University, and Gérard Massonnat and Cyprien Lanteaume, TotalEnergies, et al. The paper has not been peer reviewed.
Improving carbonate-reservoir prediction, field development, and production forecasts, especially in zones lacking data, requires novel reservoir-modeling approaches, including process-based methods. Classical geostatistic modeling methods alone cannot match this challenge, particularly if subtle stratigraphic architectures or sedimentary and diagenetic geometries not directly identified as properties with well data control the reservoir heterogeneity. Stratigraphic forward-modeling approaches can provide pertinent information to carbonate-reservoir characterization. The complete paper describes a modeling package tested and calibrated with high-resolution stratigraphic outcrop models. It allows valid prediction of carbonate facies associations mimicking the spatial distribution mapped along the Urgonian platform transects.
Classical carbonate-reservoir characterization protocols rely mainly on 3D geostatistical models based on well data, allowing the realization of 3D numerical grids of reservoir properties. These geostatistic property models are supported by deterministic geological interpretations such as stratigraphic well correlations that are commonly based on sequenced stratigraphic concepts and carbonate sedimentological interpretations. The stratigraphic framework obtained from these deterministic interpretations has a critical effect on further static and dynamic reservoir models because it constrains the spatial stationarity of the geostatistic property simulations or imposes discrete flow units or barriers. These deterministic carbonate sequence stratigraphic and associated sedimentological interpretations, however, introduce significant biases, uncertainties, and imprecisions in reservoir models and furthermore are not validated by process-based modeling approaches as one should expect from any scientific protocol. This lack of validation represents a fundamental scientific gap in classical reservoir-characterization work flows that is generally avoided in other scientific domains such as physics by iterations combining experimentation and process-based models to verify deterministic interpretations and hypothesis. The paradox is that this virtuous scientific method is applied at the ultimate stage of the reservoir flow modeling with the classical “flow history matching,” implying the following strong hypothesis (Fig. 1a): If the dynamic model obtained from the upscaled static model matches the dynamic history and the flow records of the studied field and carbonate reservoir, then the geological model, including the deterministic stratigraphic and sedimentary interpretations, is validated. Reservoir flow and dynamic behavior certainly are controlled by initial geological conditions, but those are not dependent on flow processes. According to fundamental scientific principles, geological interpretations and deterministic models must be validated by geological process-based models.
To fill this scientific gap in the presented carbonate-reservoir characterization approach, the authors introduce process-based stratigraphical and sedimentological models that are calibrated on pertinent, well-studied outcrop analogs.