Evaluation of Dual-Injection Methods for Multiple-Zone Steamflooding

Abstract

Summary For multiple-oil-zone reservoirs, steamflooding two zones at the same timeis advantageous. Three methods for dual injection through a common wellborewere tested at Texaco's Kern River field in Bakersfield, CA. Because dualinjection results in an excessive amount of heat transfer between injectionstreams, it was imperative to predict steam quality or water temperature at thesandface for each injection stream. To evaluate the merits of each methodadequately, a computer model was developed to predict downhole temperatures, steam qualities, and pressures for each predict downhole temperatures, steamqualities, and pressures for each injection stream. Field tests of the threeinjection configurations provided measurements necessary for validation of themodel. This paper provided measurements necessary for validation of the model. This paper describes the computer model, the field test, and the results fromthe three dual-injection scenarios. Introduction For fields having heavy oil in multiple noncommunicating zones, dual steaminjection is practical for several reasons. Recompleting an existing injecteris more economical than drilling a new one. Also, with dual injection, floodsat different maturities can be operated simultaneously, and separate injectionprovide the most control over rates and steam qualities to each zone. Threetypes of dual-injection completions have been used at Texaco's Kern Riverfield: bare-tubing, concentric-tubing, and parallel tubing completions. Insulated tubing is not included in parallel tubing completions. Insulatedtubing is not included in this investigation. The bare-tubing completionconsists of a thermal packer placed within the casing between two sets ofperforations on a string of uninsulated tubing. Steam is injected through thetubing to the lower set of perforations. Steam is injected simultaneously intothe casing/tubing annulus and flows to the upper set of perforations. Theconcentric completion is similar to the bare-tubing completion except that twodifferent diameter tubings are used instead of one. The space between the twotubing strings is packed off at the surface and at the thermal packer toprovide some insulating effect. The parallel completion consists of twointegral joint tubing strings and two thermal packers within a common casing. The first packer is a stab-in, wireline-set packer set between the two sets ofperforations. The first string of tubing, which provides injectionperforations. The first string of tubing, which provides injection to the lowerset of perforations, is run and stabbed into the bottom packer. Within thefirst tubing string, a hydraulic set dual-tubing packer. Within the firsttubing string, a hydraulic set dual-tubing packer is installed. This packer isset above the top set of perforations. packer is installed. This packer is setabove the top set of perforations. The second tubing string is stabbed into thetop packer and run beside the first string and is used to provide steam to theupper set of perforations. The void space between the tubing walls and thecasing wall provides some insulation. To evaluate the relative merits of thesethree methods better, a computer model was developed to predict heat losses andpressure profiles. Models that deal with dual steam injection are not readilyprofiles. Models that deal with dual steam injection are not readily available. General-purpose wellbore models that predict dual-injection heat losses oftendo not give adequate treatment to the two-phase flow aspects of the pressurepredictions or to the heat transfer during boiling and condensation. Tovalidate the computer model and to compare the three injection completionsbetter, a field test was conducted. The field test consisted of pressure andtemperature surveys of the three completion types. This paper describes thefield test, the use of the computer model to quantify the heat transfer, andthe relative merit of the three types of injection completions. Field Test Three main objectives were considered in the design of the Kern River fieldtest:to better the understanding of boiling heat transfer when hot wateris injected in one stream and steam in the other,to test the effectivenessof concentric tubing strings for reducing heat transfer, andto determinemodel accuracy in predicting pressure drop for a wide range of injection steamqualities. predicting pressure drop for a wide range of injection steamqualities. Surface measurements recorded during the field test included thetemperature, pressure, and flow rate of each injection stream. Differentialpressure across an orifice plate was also measured to provide an accurate meansof calculating steam quality at the surface. Downhole measurements of pressurewere made at intervals ranging from 10 to 50 ft. Temperature was measured atthe same intervals as pressure but was not run for all tests because theinformation is redundant for saturated steam. The generators were placed closeto the wellsite to eliminate excessive heat loss in surface flowlines. Turbinemeters were use to measure feedwater rates to each generator. Fig. 1 shows thelayout of the surface equipment around the wellhead. Surface temperatures andpressures were recorded at the generator discharge and near the wellhead. Orifice-plate meter runs were placed in each flowline near the wellhead torecord pressures and differential pressures through the orifice platescontinuously. Steam quality was calculated with correlations developed by Marriott and James for flow through sharp-edged orifices. Downhole pressure andtemperature measurements were made with surface-readout recording loggingtools. The pressure tool was a capillary tubing pressure tool with a 3/4-in. hollow chamber run on the bottom of a string of 1/8-in. capillary tubing. Thetemperature tool was a 3/4-in.-diameter thermocouple tool run on an electricline. Pressures were measured simultaneously in each injection stream. Temperature runs, made separately from the pressure runs, were also measuredsimultaneously in the two injection streams. Twelve tests were made atinjection rates from 205 to 445 B/D cold water equivalent (CWE). Steamqualities varied from 0% to 78 %. Injection pressures ranged from atmosphericto 445 psia. Only a few of the tests are discussed here.Parallel-Tubing Tests Parallel-Tubing Tests Parallel-tubing tests wereperformed at San Joaquin Well 550, a Parallel-tubing tests were performed at San Joaquin Well 550, a dual injecter completed with 5 1/2-in. casing to atotal depth (TD) of 1,138 ft. Fig. 2 shows the completion for this well. Forcomparison, all simulation runs were made at a common depth of 900 ft becauseall test wells were completed to at least this depth. Figs. 3a and 3b showresults for the parallel-tubing tests from both field test and simulation. Pressure drops for all parallel runs at high rates and high steam qualitiesresulted in pressure drops of 250 to 300 psi. Downhole steam qualities were notmeasured, so the match psi. Downhole steam qualities were not measured, so thematch of the pressure profile was used to determine the adequacy of theheat-loss calculations. Pressure drop increases with higher injectionqualities, as can be seen by the difference in the slope of the pressureprofiles in these runs. The fact that the simulated pressures are close to themeasured pressures for the various pressures are close to the measuredpressures for the various qualities, ranging from 4% to 78%, indicates that theheat loss, and therefore the steam-quality estimates, are accurate. Thecomputer model was used to quantify the wellbore heat losses and to determinethe heat content of the injection streams at the sandface (Table 1). Steamquality changes from top to bottom were all less than 5 quality points.