Ekofisk Reservoir Voidage and Seabed Subsidence

Abstract

Summary Field data describing the time lag between Ekofisk subsidence and reservoirvoidage are given. A method to discriminate between real subsidence variationsand random-data errors and a procedure to derive a contemporary relationshipbetween reservoir voidage and seabed subsidence are presented. At Ekofisk, mostsubsidence lags reservoir voidage by 2 to 3 months. Introduction Phillips Petroleum Co. Norway operates the Phillips Petroleum Co. Norwayoperates the Ekofisk field, which is located roughly in the middle of the NorthSea. Several platforms, placed near the center of this field, platforms, placednear the center of this field, are connected by bridges. This group ofplatforms is referred to as the Complex. The platforms is referred to as theComplex. The field was discovered in 1969 and has produced oil and gas since1971. produced oil and gas since 1971. The Ekofisk field chalk reservoir islocated about 3000 m [9,843 ft] below the seafloor. As oil and gas areextracted, these high-porosity chalk layers are compacting. Part of thiscompaction is transferred to the Part of this compaction is transferred to theseabed and results in the observed subsidence. Subsidence of the platforms atthe Ekofisk Complex was noticed in late 1984. Ekofisk subsidence is not unique; other hydrocarbon-producing, reservoirs, such as Wilmington, Long Beach, and Bolivar Coast reservoirs. also show surfacesubsidence. To my knowledge, however, no published studies with a developedtime-lag relationship between surface subsidence and voidage are available. Different measurement systems were used to establish that a total subsidence>4 m [is greater than 13 ft] has occurred since 1971. From March 1985 toApril 1988, 35 measurements were taken with the NAVSTAR (navigation system withtime and ranging) global positioning system (GPS). Relative height differencesbetween, three benchmark platforms (Tor 2/4E, Edda 2/7C, and platforms (Tor2/4E, Edda 2/7C, and Albuskjell 2/4F) and the subsiding Hotel Platform 2/4Hwere measured at monthly Platform 2/4H were measured at monthly intervals. Fig.1 shows the geographical location of these platforms. We used thesemeasurements to derive our best estimate of subsidence progression for the 15thof each month for which GPS data are available. These procedures are describedin Ref. 2. Computed elevations, called "averaged interpolatedelevations" in Table 1, are used as a basis for further calculations. Thesevalues represent the history of relative elevation, changes between the HotelPlatform 2/4H and the Albuskjell Platform 2/4F. Well-defined and highly accurate data are required to relate short-termchanges in average monthly reservoir-voidage rates to additional monthlysubsidence. The subsidence history data in Table 1 were not sufficientlyaccurate to reveal the effects of changes in the voidage rates on theprogression of subsidence. The uncertainty in these progression of subsidence. The uncertainty in these data without further processing is perhaps 20 to 30 mm[0.8 to 1.2 in.] (see Ref. 2), an uncertainty too large for our purpose. Wedecided to process these elevation data mathematically to reduce theuncertainty of the interpreted Hotel Platform 2/4H elevation history as much aspossible using simple, recognized, and mathematical procedures; (2) tocalculate the lag time between procedures; (2) to calculate the lag timebetween total net reservoir voidage and the response of surface subsidence; and(3) to compute an estimate of the relationship of reservoir voidage andconsequent subsidence effects. Data Interpretation The subsidence rate is expected to decrease with time. Over a few years theobserved progression of subsidence would be progression of subsidence would besomewhat parabolic in nature, with a horizontal axis for the parabola. Thisseems reasonable because the subsidence history depicted in Fig. 2 does notshow a perfectly linear time/ subsidence relation but a slightly curved lineinstead. We decided to use this assumption to obtain a slightly better estimateof the seabed-subsidence history than previously reported. Data Used Table 1 gives the history of Hotel Platform 2/4H subsidence data used inthis report. Data are spaced by regular intervals, and their derivation isdescribed in Ref. 2. We reduced random errors in the adjusted Hotel Platform 2/4H elevationsfurther with numerical low-pass filtering, which required that we specify acutoff frequency in the filter. We could define only an arbitrary cutofffrequency, however, because we do not yet deal with cyclic data or data with awell-known frequency content. We had a 28-day production interruption for the Ekofisk jackup operationfrom Aug. 10 to Sept. 8, 1987. JPT, November 1990 P. 1434