A statistical learning approach to model the uncertainties in reservoir quality for the assessment of CO2 storage performance in the lower Permian Rotliegend Group in the Mid North Sea High Area

Abstract

It has been identified that the Rotliegend sandstone reservoir in the Mid North Sea High region, in the UK Quadrants 27-29, has a large-scale CO 2 storage potential of national importance. In this paper, the authors develop a reservoir model using extensive datasets available from seismic interpretations and core analysis. An advanced statistical learning approach was applied to characterise the uncertainties in the spatial distribution of reservoir quality. The model was used to assess the CO 2 injection performance and the preliminary results obtained thusfar indicate promise in the available storage capacities.