GEUS is using Petrel™ for modelling reservoirs, incorporating data from a number of different scales, ranging from core data to seismic information. Modelling sessions and decisions about model size, gridding scheme and boundary conditions, are always guided by the ultimate purpose of the constructed model – whether it is for simple data display, volumetric calculations or dynamic flow simulation. The produced models are easily integrated with flow modelling in Eclipse™.
The three main issues for modelling are related to integration, data re-scaling, and uncertainty handling. The integration involves the combination of the structural framework and the property model, and how these are updated consistently. The data-rescaling is needed when data from different measurement methods are combined for a common volume support in the model grid. The uncertainty handling needs to be based on a combination of deterministic scenario-based model suites and stochastic methods for reflecting variability within each scenario. This SmartModel workflow combines a balance between the two model processes but still needs well-informed user decisions to avoid biased representation.
The relations between different types of data and data from different scales are studied as a preparation for inclusion in the model. If individual parameters are derived from different measurements on different scales, the consistency of the model input can be studied by using the scaling laws. As an example, porosity can be derived from core-plug analysis, well-log interpretation and from seismic inversion. The support volume for these data varies from cubic centimetres for the core plugs to hundreds of cubic metres for the seismic resolution in inversions.