Modeling Matrix-related Heterogeneities in Carbonates with Non Process-based Geostatistical Methods: Real and Synthetic Examples
Philippe Ruelland¹, Florent Lallier¹, Henk Kombrink¹, and Sophie Prat²
¹Geoscience Research Centre, TOTAL E&P UK, Aberdeen, UK
²TOTAL S.A, E&P, CSTJF, Pau, France
Modeling carbonate reservoirs in 3D is a key challenge which is not so simple to solve. The first reason for this is related to the fact that every field is unique, and this is true for any field whether carbonate or silico-clastic. The second and main reason for this difficulty lies in understanding which heterogeneity is controlling flow dynamics in carbonate reservoirs. If matrix plays a dominant role in controlling the dynamic behavior of the reservoir, then the key aspect control is the complexity of the interaction between sedimentary facies and diagenesis.
3D modeling techniques that are available today enable to build structural models which are populated with facies and petrophysical parameters and, when necessary, with explicit or implicit fracture networks. The objective of a 3D model is to provide a set of geologically sound realizations for which a satisfying match between flow simulations and production data is obtained. For a given reservoir, 3D models may need to be different depending on the recovery mechanism, because the heterogeneity which will drive the flow may differ.
Thus, focusing on modeling key heterogeneity can infer simplifying or by-passing some of the other specificities of the matrix.
Recognizing the key heterogeneities is the first step of any reservoir geology study. This implies taking into account both static and dynamic information and building a picture of the carbonate system from well data and analogues. The analogues (present-day systems, outcrops or conceptual cartoons) often being asymmetrical do enhance extra complexity in the making of a 3D model. Depositional facies and diagenesis modeling should preferably be conducted in two separate steps because diagenesis comes as an overprint onto an existing facies distribution.
Another difficulty arises when considering high permeability streaks. The temptation is then to generate grids which consist of very fine layers (<50cm) to capture the highest frequency of the petrophysical data at log or plug scale, and to propagate it over a significant number of cells along this layer. But just how representative is this value when the cell dimensions in X and Y are 100 or 200m?
AAPG Search and Discovery Article #120034©2012 AAPG Hedberg Conference Fundamental Controls on Flow in Carbonates, Saint-Cyr Sur Mer, Provence, France, July 8-13, 2012