Reservoir Rock Type Classification and Variation of Reservoir Quality in the Arab Formation, Dukhan Field, Qatar

Abstract

Abstract This paper illustrates the importance of classification and identification of rock types within integrated reservoir characterization studies. The goal is not to present a new classification scheme, but to draw attention to the benefits of a classification that provides the basis for populating a geo-cellular model with key attributes (porosity, permeability, saturation). Rock type, as derived from an integration of core examination, routine and special core analysis and wireline logs, and conditioned with well tests, facilitates geologically reasonable property prediction. The study concludes that a single classification scheme is impractical for all aspects of geological modeling. This work stems from a comprehensive reservoir description study of the Jurassic Arab C and D reservoirs of the Dukhan field. Core descriptions, routine core analysis, and visual examination of log data form the basis for recognition of an initial classification internally named ResFacs (Reservoir Facies). Because ResFacs originate from core, volume fractions and distribution are readily derivable from a cycle-based, sequence stratigraphic interpretation. Mineralogical and depositional trends of ResFacs form an initial conditioning parameter for models. This core-based classification is also used in all routine and special core analysis (SCAL) to segregate rocks of differing reservoir quality. It forms the common variable between all other classifications. SCAL data reveals reservoir quality trends that are not resolvable from core description and routine analyses alone. Most important were trends in fine-grained, muddy limestones and very coarse-grained lime grainstones. MICP data and saturation height modeling reveals a small volume of low porosity lime mudstones with such fine pore systems that they are unlikely to have been filled with hydrocarbons. At the other end of the spectrum, SCAL data and petrography reveal a class of coarse grainstones with point of contact cements that exhibit extremely high permeability. These data are used to modify the ResFac classification for use in calibration of well logs. Log-derived rock types, referred to as Petrophysical Reservoir Facies (PRFs) have been established to increase input data density for property distribution in geological models. As with many oilfields, historical SCAL and routine core analyses of variable quality are available. In order to make use of these data, samples were classified within the ResFac scheme. In addition, 100 representative samples were selected for further analysis (MICP, NMR, petrology, and porosity-permeability measurement at ambient and overburden conditions). This dataset provides a means of resolving pore systems by rock type and allows incorporation of historical data analyzed under variable laboratory conditions. The ResFac classification scheme is impractical for numerical modeling. Accordingly, ResFacs were consolidated based on texture, mineralogy, and rock type with unique saturation height functions. The final input into flow simulation models is termed Reservoir Rock Types (RRTs). These RRTs are used for property prediction, and calibration occurred with static data (e.g., initial water saturation from logs) and dynamic data (e.g., well test permeability-thickness and profiles from production logging tools). Rigorous rock description and comparison to measured static and dynamic data ensure that models based on RRTs describe fluid distribution and fluid flow capability.