The Gullfaks Field is a large oilfield in the Norwegian Sector of the North Sea. The field is compartmentalized by a dense and complex fault pattern, and most of the oil is contained in poorly consolidated but high-quality reservoir sands. Production started in 1986, and waterflooding is the main recovery method. Water-alternating-gas (WAG) injection -or supplementary injection of gas in existing water injection wells -has been identified as a possible method for increasing and accelerating oil recovery from Gullfaks. To verify the potential of this technique, a WAG pilot was initiated in 1991. The target area for the gas was primarily one fault block where major water breakthrough had occurred. In this paper, relevant field observations are presented, and the effects of the WAG pilot injection on flow performance and recoveries are discussed on the basis of the field observations and a detailed, history-matched 3D simulation model for the pilot area. In the present case, the gas migrates rapidly towards the top of the formation and accumulates in secondary gas caps. Improved immiscible displacement of attic oil is the main benefit of gas injection in this case.

This paper presents parts of the work performed to develop and qualify a polymer-assisted surfactant flooding (PASF) system for economical use in the Gullfaks Field, Norway. The paper addresses experimental work done in the laboratory, numerical simulation of PASF, and the evaluation of the potential for PASF in full-field scale.The experimental part comprises core-flooding experiments at different temperatures, pressures, and gas/oil ratios (GOR) to optimize the PASF system for the Gullfaks Brent formation conditions. The surfactant in the PASF system is a branched sulphonate (5,000 ppm) and xanthan (500 ppm). The surfactant-polymer slug is followed by a slug of xanthan (500 ppm) for mobility control. No cosolvent is used. In coreflood experiments more than 70% of the waterflood residual oil was recovered.Using reservoir simulation, a suitable pilot area was found in the Brent reservoir. Additional results from simulations were the amount of chemicals, the time needed for the pilot test, and additional oil recovery.Much effort was put into estimating the full field PASF potential. First, the areas of the field where PASF possibly could be used were selected. Key factors were existing and planned well locations, production data, and long-term production forecasts. Then, the amount of chemicals needed and the expected technical efficiency for each area were calculated. To verify these calculations, an area of the field containing two possible injection wells and three producers was selected for a simulation study. This area was considered the most promising area for PASF.The main conclusion from this work is that, with the present crude-oil price and chemical costs, the PASF process is not economically attractive for use in the Gullfaks field, mainly because the residual oil was considerably lower than believed at project start.

No abstract

The surfactant test iriitiated in a Gu!Ifaks Well in January 1991 1 . was nterrupted before surfactant injection due to operational problems iri the well. Based on this expenence, a new well was identified, and a single well surfactant injection test was planned and conducted dunng the first haif year of 1992. The results, and a discussion of borh the results and the experiences, are presented for the two well operatons.The main objective of the test was to evaluate the efficiency of the surfactant to mobilize oil at reservoir conditions, by measuring the residual oil saturation after vater flooding, and after ihe surfactant was injected.The residual oil saturation was measurcd by the SWCT method. applyirig ethyl forrnate as the reactive tracer. A ota1 of six SWCT tests were performed during the operations. In both welis, extensive data collection programs were executed. This included measuremenr.s of fluid saturations (GST Iogs), fluid inflow and outflow performances (PLT 1025,) as well as pH, ionic composition, and gas content in the back produced fluid. These data were integrated in the evaluaiion of the SWCT tests, and proved to be crucial in understanding arid interpreting reservoir processes dunng testing. A chemical flood simulator was used to model the tracer profiles from the SWCT tests.Special attention was paid to the reduction ofpH in back produced water due to hydrolysis of ethyl formate, and the reduced gas content in back produced water due to stripping of light components of the residual oil by the injected water.The residual oil saturations interpreted from these tests.showed that between 40 and 70% of the remaining oil after water flooding was mobilized by the surfactant.

The development program for processes to reduce watercut during production ofthe lower Brent reservoir unit ofthe Gullfaks field culminated by a large scale injection of alkaline Na-silicate gel in a lower Brent production well durmg summer 1993. The operation was prepared and designed through an extensive laboratoiy program as well as small scale onshore field tests.The injection sequence consisted of a preflush of fresh water added low èoncentratkns of KC1 or Na-silicate. The purpose ofthe preflush was to condition reservoir parameters to improve injectivity ofthe Na-silicate gel. A volume ofapproximately 5500 m3 4-6% Na-silicate gel was mjected and displaced into the reservoir with one tubing volume fresh water. In total 9700 m 3 fluid was injected. An extensive sampling and anaJysis program of production fluids was worked out and results will be presented.Analysis of ion composition of the produced water allowed to determine the fractions of sea water, formation water and fresh water used as solvent for the injected fluids. Aftertreatment, the fraction of fresh water was initially about 50%, rapidly raising to 90% and higher. The injected volume was backproduced during a period of about five months. The reduction of permeability of the treated volume was also observed asan increased draw-down m the well. The resuits show that permeability is reduced significanlJy in the treated volume, and that an efficient water diversion has been achieved. The water-cut was initially at the same level as before treatment, but decreased within three months to 72-74%. The water-cut stayed at this level for another three months before rising to 77%. The treatment gave initially an increase in sand production and the well had temporarily to be produced at reduced rates. After this initial period, the total fluid rate could be increased to about 2000 Sm3 /d, the same level as before treatment, implying that the oil rate has been significantly higher most of the period after treatment than before.Based on history matching ofthe reservoir section with the Eclipse black oil simulator, predictions of well performance after treatment were made. The bistory matching included an active use of well influx data, including results from four production logs. The result indicated that the oil saturation in the drainage voluine ofthe well was rather low at the time oftreatment. Predictions with the history matched model reconstructed the water-cut development, and an estimate of incremental oil production can be done by comparing with predictions ofwell productivity without gel treatment. Using the fresh water as tracer, the perineability reduction factor in the treated volume is calculated to be approximately hundred.The conclusion is that a further testing and application of the technology iri North Sea reservoirs seems to be very promising for the pupose of improving oil recovery and well productivity.