The Burgan Sands constitute the major reservoir in the giant Greater Burgan field. The Upper Burgan Sands contain significant volumes of oil, several billions of barrels, and have generally poorer and more variable reservoir quality and poorer continuity than the bulk of the Burgan sequence. Secondary water flooding is currently under consideration.
This paper describes the challenges in developing part field models that are appropriate for waterflood simulation studies and that are, as far as is practicable, conditioned to the available dynamic data. The general approach to choosing "type areas" to represent typical portions of the field will be described. The process of developing geo-cellular models and conditioning them to dynamic data will then be illustrated.
The existing geological and simulation models are considered too coarse to provide a proper basis for modelling the Upper Burgan. The extent to which this resolution has allowed geological models to be conditioned to dynamic data has been limited.
A rock typing exercise integrated available core, log, and production data. Speed zones and flow barriers were identified. Calculated flow capacities and productivity indices generally matched field data well. Pressure breaks in the RFT profiles correlated well with vertical flow barriers. In addition, Permeability anisotropy ratios (Kv/Kh) were developed from detailed RCA and probe permeameter data acquired in key wells. Subsequently, MicroModels at whole core scale were developed and simulated to generate representative Kv/Kh ratios by Facies.
Based on sedimentology studies and rock typing work, and the existing structural model, detailed geo-cellular models were produced. Dynamic models were then developed and conditioned to selected dynamic data. The approach used to condition these models to pressure transient data, local water movement as indicated by a detailed water encroachment survey, and to the pressure breaks seen on RFTs is described. This process both confirmed the plausibility of the geo-model and reduced uncertainty in permeability anisotropy.
The field was segmented into equi-uniform polygons, on which waterflood patterns were evaluated, number of wells and throughputs determined, and the volumes of water to be handled estimated.
A series of numerical simulation models were developed, through variations on the reservoir quality, reservoir connectivity, oil type, and development options. Generated water-cut vs. recovery factor profiles were utilized in a hybrid approach combining analytic and numerical evaluations to determine production profiles and potential waterflood recovery factors over time for each of the polygons, and the whole field.
The work demonstrates a methodology by which relatively quick but comprehensive and robust evaluation of the potential value of a waterflood development could be made. It allows for sensitivity assessments and a degree of optimization.