A Sensitivity Study of Three-Phase Relative Permeability and Capillary Pressure Models
MetadataVis full innførsel
During the exploitation of the hydrocarbon reservoirs the large amount of hydrocarbon is swept my pressure sources. This source can be either natural or artificial. Natural sweep might occur when the reservoir has strong aquifer or gas cap that, however when these sources are unable to provide enough pressure the Enhanced Oil Recovery (EOR) methods are used. One of these methods which have been used widely in North Sea, Water alternating gas injection (WAG) is studies in this thesis. WAG injection is implemented by injection of water and gas alternating each other after certain period of time. WAG injection usually result in lower SOR than its two-phase counterpart displacement. Accomplishment of this method involves the three phase flow through the reservoir. For the purpose of understanding and control of this flow several Three-Phase Relative Permeability and Capillary Pressure models have been introduced such as Default (Cheshire’s)11, Stone14, Stone25 and ODD3P9 model. Sensitivity of these models has been studied in this project. Provided original two-dimensional model with eight components the base models for each three phase relative permeability correlations have been developed. Since the traditional (Default, Stone1 and Stone2) models use the saturation functions in residual water saturation, original saturation data for ODD3P model has been modified for these models. Mentioned models have common parameters as well as different ones to be studied. The parameters that are the same or identical for all the three phase models have been used for comparison of these models with one another. Some of the main parameters examined in this project include different use of saturation functions, various hysteresis models due to wetting phase, endpoint saturations, different ways of capillary pressure input and range of values that Stone exponent can have. Furthermore, the effect of equilibrium gas and dry gas injection has been examined. Simulation studies show that the endpoint saturation values are of high importance on saturation profiles. Therefore, these values should be paid more attention since they influence recovery severely. Although gas-oil data for ODD3P model is different from the ones used in traditional models, applying the same range of endpoint values would make it possible to compare these models. Moreover, different capillary control methods, various stone exponent values to Stone’s first model and hysteresis options were also applied. Studies indicate that selection of different hysteresis models can produce noticeably different results. However, outcome does not depend on the selection of imbibition or drainage curve for wetting phase. Significantly different results are obtained in the simulation if hysteresis is applied or not.