Reservoir Description by Interference Testing of the Clayton Field

Abstract

Summary This paper describes the design, implementation, and analysis of the Claytonfield interference test. The test was performed to gather reservoir informationthat was used with geologic and petrophysical data to assess the magnitude ofinterwell communication and areal and vertical reservoir heterogeneities and todevelop a detailed reservoir description. Interference-test data were evaluatedwith conventional semilog and type-curve-matching techniques. A reservoirsimulator was used to improve the test interpretation, to includerate-variation effects on the observation-well responses, and to develop acomplete reservoir description. The simulator match with interference-test dataused the best geologic and petrophysical information so that the finalreservoir description would honor the best of all available data. Thus, dataidentified by these disciplines were honored during the history match eventhough the match of the test data was insensitive to some of the values. Introduction The Clayton field is located in northeast Michigan in Arenac County (see Ref. 1 for map). Structurally, the field is a large anticline elongated on anorthwest/southeast trend. The northernmost boundary of the field is delineatedby a fault that follows the same trend. The major productive formations in thefield are the Upper and Massive Prairie du Chien. Reservoir property variationsbetween each of these horizons are quite property variations between each ofthese horizons are quite significant. As a result, multilayering is not only apossibility but is expected. The field produces a rich (151 -bbl/MMscf) gascondensate with a dewpoint pressure about 350 psi below discovery pressure. The complex nature of the produced fluid led to a desire to develop areservoir description as early in the life of the field as possible. Evaluationof the extent of both vertical and areal heterogeneity was needed to develop aneconomic depletion plan because, clearly, a depletion plan that includedcycling would be detrimentally affected if high-conductivity layers or regionswere found later. Poor vertical or areal sweep, together with the initialinvestment required for injection of a cycling gas, would preclude anydepletion plan based on pressure maintenance. pressure maintenance. Interference and pulse testing have been used extensively in the petroleum andgroundwater-management industries to determine the extent of communicationbetween producing wells and to derive specific descriptions of areal andvertical property variations that exist in a field. Kamal and Hegemannpresented general overviews of interference- and pulse-testing research done bythe petroleum industry. Review of their studies shows that extensive petroleumindustry. Review of their studies shows that extensive work has been done onthe analysis of interference and pulse-test responses in single-layer systems. The effects of skin, storage, vertical fractures, and areal permeabilityheterogeneity on the interference-test responses of such systems has beenconsidered. The influence of other reservoir heterogeneities (e.g., nearbyfaults, compressible regions, or anisotropic trends) on estimated permeabilityalso has been studied and the resulting theories permeability also has beenstudied and the resulting theories applied to field tests. None of these studies, however, give results applicable to the multilayerreservoir encountered in the Clayton field. Most studies in the petroleumliterature concerned with multilayer systems focus on naturally fracturedformations and/or the concept of vertical interference testing to determine theextent of interlayer communication. Of the four papers in the petroleumliterature that deal directly with the effects of layering on interference andpulse data, two (Refs. 17 and 18) are shown by Onur and Reynolds to be based ontechnically incorrect equations, though the calculations in Ref. 18 arereasonably accurate. Woods and Onur and Reynolds give methods that use combinedrate and pressure data to obtain accurate estimates of layer properties. Without rate data, Onur and Reynolds concluded that a history-matching processwould be the only recourse for data analysis. Kamal advocates this same ideafor multilayer systems. A number of field studies published recently have takenthis history- matching approach, using analytical and/or numerical simulatorsto match the measured data. In this study, we use a multiwell, numerical, gas-reservoir simulator tomatch the measured interference response at six observation wellssimultaneously. Preliminary estimates of average interwell transmissibilitiesare computed with conventional semi- log and/or type-curve-matching methods. Final values of interwell parameters are obtained through trial-and-errormanipulation of parameters are obtained through trial-and-error manipulation ofthese parameters until model and observed pressure responses match. This methodis particularly attractive for this work because one of the major goals was toobtain a preliminary reservoir description that could be used in acompositional simulator to yield a more realistic evaluation of how thereservoir should be depleted. Geologic Setting The Clayton field produces from an Ordovician formation commonly referred toas the Prairie du Chien. JPT p. 524