A METHOD FOR REDUCING MODEL ERROR WHEN ESTIMATING RELATIVE PERMEABILITIES FROM DISPLACEMENT DATA

Sarma, H.K. and Bentsen, R.G., 1989. A method for reducing model error when estimating relative permeabilities from displacement data. J. Pet. Sci. Eng., 2: 331-347. The importance of reliable effective permeability data in reservoir engineering studies is well recognized in the petroleum industry. Often the effective permeability correlation for a particular rock-fluid system is established on the basis of laboratory displacement studies. Moreover, the external-drive technique is frequently used to obtain such effective permeability data. When using the external-drive method, it is common to use functional forms to fit the experimental data. The selection of such functional forms requires some care, if the introduction of model error into the analysis is to be avoided. This study demonstrates how improved functional forms for smoothing cumulative-oil and pressure-drop histories can be obtained by choosing regression equations which are consistent not only with the integrals of the two differential equations which describe the displacement process, but also with various known physical conditions which can be imposed on the displacement process. Moreover, it is shown that when these improved functional forms are used to fit the experimental data, the resulting reduction in model error can effect, in some cases, an improvement in the accuracy of the effective permeability curves estimated using the external-drive method. Relative permeability curves obtained using this approach are presented also.