Correlating and Averaging Connate Water Saturation Data

Abstract

Abstract The use of simple hyperbolic relationships in correlating connate water saturation data is shown to have wide application in a variety of reservoirs. This can be of considerable practical significance to the reservoir engineer because volumetrically weighted average saturations for any combination of core samples, reservoir strat a or lease areas can be obtained directly from such correlations with a knowledge of average porosity alone. Also, in systems represented by a simple equilateral hyperbola the porosity- saturation productis constant and hydrocarbon volumes are conveniently derived from the difference between the pore volume and a constant fraction of the bulk volume. Potential applications and limitations of the technique have been explored fora number of sand and carbonate reservoirs, and several examples are described. Introduction Of fundamental importance in all oil and gas reserve estimates is the concept of connate water saturation. In fact, as it is one of the most basic parameters, like pay thickness or porosity, connate water must be measured orestimated in any and all types of reservoir analyses. Because water saturations may range from as little as 1 or 2 per cent up to 100 percent of the pore space within the same reservoir, it is common practice to condense the data into asingle "average" value related to some other average reservoir characteristic such as porosity or permeability. Although such an average may be applicable tothe reservoir as a whole, it can, in some cases, lead to appreciable errors ifapplied to individual wells or areas within the pool. Even when the entire poolis considered, there may be considerable difficulty in obtaining the correct average. The use of permeability as a correlating parameter has long been popular (1,2, 3), and in some reservoirs it may be the only reasonable choice. Difficulties, however, can be encountered in applying such relationships toobtain true average saturations. Porosity-saturation correlations have beengiven somewhat less attention in the literature, although they have been widely used in Western Canada and in some areas of the United States. There are inherently fewer problems in establishing representative average values from porosity relationships, as the differences between arithmetic and geometric means, medians and modes, and the effects of sample orientation, need not be considered.