Prediction of Performance of Miscible-Gas Pilots

Abstract

Summary A model was developed to predict field performance of miscible-gas injectionprojects in oilwet reservoirs. The model represents a vaporization and/orliquid extraction multiple-contact-miscible (MCM) process by astabilized-contact process by a stabilized-contact miscible process. A criticalcomponent of the model is the inclusion of a solvent relative permeability(SRP), different from the oil relative permeability, in the presence of a finaloil saturation. An example match of a misciblegas injection pilot is shown with the newmodel. Satisfactory matches of both production and injection wellperformancedata with model predictions were obtained. The pilot predictions were obtained. The pilot reservoir description was not changed to match miscible-gas floodperformance. The new model was also used performance. The new model was alsoused to match field-performance data of other pilots using the same procedure. Although not shown in this procedure. Although not shown in this paper, matchesof similar quality paper, matches of similar quality were obtained for all theother pilots. Introduction Accurate performance predictions of a miscible-gas injection project requirea reservoir simulator validated with misciblegas performance and a realisticreservoir description. Validation of a simulator that incorporates physics fora new process requires matching of actual process performance. Pilot testsprovide valuable data on the performance. Pilot tests provide valuable data onthe process performance for such validation. We process performance for suchvalidation. We conducted several miscible-gas pilots for performance data thatare the basis for the performance data that are the basis for the modeldiscussed here. This paper describes a new first contactmiscible (FCM) SRP model to predictperformance of a water-alternating-gas (WAG) performance of awater-alternating-gas (WAG) injection process in oil-wet reservoirs. Thetechnique uses a compositional model designed to operate in the modifiedblack-oil mode. An SRP (miscible gas) curve is used to approximate the mobilityof the MCM process. The SRP curve has a physical basis for vaporization- andextraction physical basis for vaporization- and extraction type MCM processesand is based on Schneider and Owens data. Wellington and Vinegar and Christiesrecently indicated that the presence of mobile water during simultaneousinjection of water and CO2 stabilizes viscous fingering compared withcontinuous CO2 injection. The new model implicitly assumes that the reservoirmixing, although significant because of dispersion, is not dominated by viscousfingering. Miscible-Gas Pilots The miscible-gas pilots, all oil-in-the-tank pilots, were normal five-spotpatterns. pilots, were normal five-spot patterns. Fig. 1 is an exampleoil-in-the-tank pilot configuration referred to as Pilot A. The pilotinjectors, Wells 21 through 26 in Fig. 1, pilot injectors, Wells 21 through 26in Fig. 1, injected the miscible gas alternately with water. The pilotproducers, Wells 19 and 20, were influenced only by the fluids injected in thepilot injectors. In general, the same statement cannot be made for invertedpatterns in which a single well injects patterns in which a single well injectsmiscible gas to surrounding producing wells. For inverted patterns, thesurrounding producers may experience the miscible-gas producers may experiencethe miscible-gas flood from the central injector and may respond simultaneouslyto a conventional oil recovery process from an area outside the pilot. In suchcases, it is much more difficult pilot. In such cases, it is much moredifficult to quantify the recovery caused solely by miscible-gas injection. The oil-in-the-tank pilots also have an advantage over loggingobservation-well pilots alone. Some of the oil-in-the-tank pilots, includingPilot A, had logging observation and fluid sampling wells to aid in theinterpretation of process performance. These wells usually are placed nearmiscible-gas injectors in a region where the reservoir pressure is above theminimum miscibility pressure is above the minimum miscibility pressure (MMP). Observation alone can pressure (MMP). Observation alone can account for neitherareal sweep nor the performance in an area where reservoir performance in anarea where reservoir pressure falls below the MMP. On the other pressure fallsbelow the MMP. On the other hand, oil-in-the-tank pilots can have areas nearthe producer where reservoir pressure may fall below the MMP. Thus, oil-inthe-tank pilots yield performance over a wide range of pressures and provide abetter data source for validating a miscible-gas simulator. Pilot Reservoir Descriptions Pilot Reservoir Descriptions The pilots weredrilled in areas of their respective fields not affected by waterflooding orwater influx. Extensive logs were obtained for all the wells, and some wellswere cored to aid in the reservoir descriptions. Waterfloods were subsequentlycompleted in each pilot to provide performance data. During this period, extensive pressure-transient data were also obtained. Initial reservoirdescriptions of the pilots were developed through geological interpretation, permeability/porosity ratio (klo) plots, and core, permeability/porosity ratio(klo) plots, and core, log, and pressure-transient data. The pilot waterfloodperformance as well as the primary and waterflood performance of the primaryand waterflood performance of the offset wells were history matched to yield afinal realistic reservoir description of each pilot. Knowledge of a realisticreservoir pilot. Knowledge of a realistic reservoir description is necessary toseparate the effects of the reservoir heterogeneities from those of themiscible-gas process variables. JPT P. 1564