An automatic history-matching technique has been developed for analyzing pressure buildup data affected by wellbore phase segregation. This automated type-curve analysis method uses a Gauss-Newton least-squares procedure to match observed field pressures and pressures simulated with Fair's1 phase segregation model. From this analysis we determine not only estimates of wellbore skin factor and formation permeability but also indications of the magnitude of the phase segregation effects.
The basis of conventional semilog analysis techniques for buildup tests is the presence and correct identification of a middle-time line whose slope is related to formation permeability. However, wellbore phase segregation may delay or completely suppress the development of this line, so semilog analysis methods often cannot be used. Fair1 has developed type curves for analyzing buildup data distorted by wellbore phase segregation, but unique solutions are difficult to obtain from manual type-curve matching. Since our automatic history-matching technique does not rely upon the presence of a middle-time line, this new method becomes especially useful when conventional semilog analysis techniques cannot be applied. In addition, comparisons of results from our automatic history-matching and Fair's1 manual type-curve fitting suggest that more reliable solutions can be achieved with this new automated procedure.
We present analyses of pressure buildup data taken from the petroleum literature and a previously unpublished field test. Comparisons of the results from our automatic history-matching, Fair's1 manual type-curve fitting, and the Horner2 semilog plotting method illustrate the applicability of this new method for analyzing field buildup tests distorted by wellbore phase segregation.