The implementation and use of ad-hoc approaches and solutions, combined with advanced simulation tools, allowed facing critical aspects during the development of the reservoir model of an extra heavy oil giant field. The main reservoirs consist of fluvio-deltaic stacked sandstones bearing an oil characterized by low API gravity (7-12°API) and high viscosity (2000-10000 cP). The definition of the so-called ‘Flushed zone’, a low salinity water zone above the regional aquifer, is one of the main challenges of the modelling activities, involving static and dynamic features.
The ‘Flushed zone’ is not straightforward detected by means of conventional log data set, so the combined interpretation with special logs (NMR and ADT) allowed reducing the uncertainties in its definition. The closeness of these fresh waters to the oil column negatively affects both the oil/water mobility ratio and the water-cut at the wells, so a proper modelling of the Flushed Zone is essential to reproduce a reliable well behaviour. The fine geological grid was upscaled in a coarse simulation grid; transmissibility upscaling was evaluated as the most appropriate solution in order to preserve the fine scale results.
The presence of a no-flat contact and non-equilibrium hydraulic system resulted in a non-conventional initialization of the dynamic model, in which the reservoir initial conditions were directly included by means of the enumeration procedure.
Due to the areal extension of the field, the coarse grid still had more than four millions of cells, so the use of a new generation simulator was also tested to assess time reduction and performance reliability. The validated dynamic model was finally used as a reference for the evaluation of possible development scenarios. A large number of development wells, grouped in clusters, were required for the field exploitation. The implementation of an automated ad-hoc workflow, that combines geological information (such as permeability and porosity) and hydrocarbon column maps, revealed to be a useful support/tool for identifying the best levels for each development cluster, reducing modelling time.
