Forward prediction of aeolian systems using fuzzy logic, constrained by data from recent and ancient analogues

Abstract

Aeolian sands are the main reservoir rock in some of the largest gas fields, such as the Shell-Exxon Groningen Field, operated by NAM. Although aeolian reservoirs have been studied for many years, there is still room for improvement in the predictive modeling of such reservoirs. A pilot project with this objective was initiated by SIEP B.V. in 1997, together with Heriot-Watt University in Edinburgh, UK and with Uppsala University, Sweden, to evaluate the factors influencing aeolian systems, and to formulate a forward model using ‘fuzzy logic’.

The project was initiated to develop a fuzzy system for generic modeling of aeolian architectures. The key aims were to be able to predict the type, amount and distribution of major facies in generic aeolian systems and specifically to model regional-scale architecture in the sub-surface. Fuzzy rules and sets, which defined the behaviour of aeolian systems, were constructed and used to modify the pre-existing fuzzy modeling software which had been designed for shallow and deep marine systems. The modeling procedure used input data appropriate to the Rotliegend climate, and was validated by comparing the resulting models, in terms of thickness and spatial distribution of facies types, to well data from the Upper Rotliegend interval of the Lauwerszee Trough area, NE Netherlands (Figures 1 & 2).