Thermally Induced Fracturing of Ula Water Injectors

Abstract

Summary This paper describes the impact that thermally induced fracturing (TIF) hashad on the North Sea Ula field injection wells, allowing higher thananticipated water injection rates to be achieved. This work also discusses howthermal stress reduced fracture propagation pressures by 2,000 psi and how a 3Dsimulation code developed to model TIF was used. Injection-water-qualityspecifications and techniques to optimize TIF are presented. presented. Introduction The performance of all the Ula field water-injection wells has been stronglyinfluenced by TIF, a phenomenon reported and modeled in several reservoirs. Thepractical consequences of TIF depend on several factors, which are discussed inthis paper and have resulted in different behaviors in individual Ula wells. After more than 8 months of successful injection with the first injector (WellA-03), a second injector (Well A-04) was started up in mid-Nov. 1988. Thiswell, however, achieved an injection rate of only 18,000 B/D at the maximumavailable injection pressure. This was not unexpected considering the poorerreservoir quality at this reservoir location than at Well A-03, but work was toidentify means of improving well performance. After several weeks of stableinjection, the performance of Well A-04 improved very rapidly. The injectionrate increased by 6,000 B/D and the injection pressure decreased by 70 psiwithin a 15-minute period. During a second, slower, increase in performance 3days later, the period. During a second, slower, increase in performance 3 dayslater, the injection rate built to 29,000 B/D. However, the injection pressureremained at the same level as a week before when the rate had been 18,000 B/D. Both events correlated with drops in injection-water temperature resulting fromoperational changes. Further examples are given for a variety of wells and areex-plained in terms of TIF. These follow a description of the Ula field, theinjection hardware, and an introduction to the theory and modeling of TIF. Finally, the impact of TIF on operations is discussed. Fig. 1 gives an overviewof each well's performance and shows the timing of events that are the mostsignificant to performance and shows the timing of events that are the mostsignificant to our understanding of Ula injection-well behavior. Ula Field Reservoir Description. The Ula field is an undersaturated oil reservoirlocated in Block 7/12 of the Norwegian sector of the North Sea, about 180 milessouthwest of Stavanger. The reservoir lies in laterally extensive Upper Jurassic shallow marine sandstones formed in a four-way-dip, closed domalstructure. It consists of several upward-fining sequences that are grouped intofive reservoir zones-Zones 1A, 1B, 2A, 2B, and 3A-from the top downward. Thesezones are characterized by rock properties resulting from perceived changes insea level. The individual zones correlate across the perceived changes in sealevel. The individual zones correlate across the field but thin away from thecrest. Horizontal and vertical pressure communication is good, althoughporosity and permeability both deteriorate downdip. Of the reserves of 435million STB, 115 million STB was produced from seven wells between fieldstartup in Oct. 1986 and the end of June 1990. Six water injectors have beendrilled to supplement the negligible aquifer support. All injectors werecompleted directly above the oil/water contact to maximize recovery whileavoiding the poorer-quality reservoir farther downdip. Fig. 2 shows the welllocations on the reservoir structure. The reservoir crest is about 10,825 ftbelow sea level, and the reservoir is hot, with an average temperature of 295degrees F. Reservoir pressure has fallen by about 3,800 psi from the initial7,110 psig. psig. Injection Hardware. All injection wells were deviated from asingle platform and completed with 7-in. tubing. The design specification forthe platform and completed with 7-in. tubing. The design specification for theseawater injection system was to deliver 120,000 B/D at a wellhead injectionpressure (WHIP) of 3,000 psig. The discharge pump characteristics allowedhigher manifold pressures at lower rates early in the field life when few wellshad been drilled. Two additional pumps were added 3 years after field startupto boost the pressure into two of the low-rate wells (Well A-05 and A-08)drilled in the poorer-quality reservoir. The response of Well A-08 to thehigher injection pressure is discussed later. Fig. 3 is a schematic of theinjection hardware. A side of the injection water was originally passed througha heat exchanger to cool produced oil before being remixed with the remainderof the injection water. Increasing the water temperature to 104 degrees Fimproved deoxygenation efficiency, and cooling of the process fluids aidedoptimization of gas/liquids recovery. The water leaving the heat exchanger waslater dumped overboard to lower the overall injection-water temperature to aid TIF. The water-quality specification originally required removal of 98% ofparticles greater than 2 m by fine filtration but was subsequently relaxed. TIF Theory. Rock at depth is compressed by high levels of stress in bothvertical and horizontal directions. When the pore pressure of the fluid in therock exceeds this compressive stress, a fracture forms. Rock toughness, whichmust be overcome when the fracture is first formed, can increase the fractureinitiation pressure by 500 to 1,000 psi above the minimum compressive stress. At Ula depth, vertical stresses exceed horizontal stresses, so the fractureforms in the vertical plane. However, cold-water-injection wells often arefractured at bottomhole injection pressures (BHIP's) lower than the originalminimum compressive horizontal stress because of a reduction in the compressivestress in the rock surrounding the wellbore caused by cooling. This process offracturing cooled rock at injection pressure less than the in-situ stress isknown as TIF. The minimum compressive horizontal stress in Ula, before coolingor pressure depletion, is about 10,500 psig. However, as demonstrated later, pressure depletion, is about 10,500 psig. However, as demonstrated later, fracturing occurs at pressures below the maximum BHIP of about 8,500 psigavailable from the injection wells. When rock around the wellbore is cooled byinjection water, it tends to contract, shrinking away from the hot rocksurrounding it. This sets up tensile stress in the cooled rock and reduces theinitial rock compressive stress. The thermoelastic stress reduction may reach11 psi/degrees F of cooling. It increases linearly with temperature change anddepends on the shape of the cooled region and the rock properties. Additionaltensile stresses resulting from poroelasticity, which occur in regions wherereservoir pressure is reduced, can aid fracture initiation. Horizontalcompressive stresses are rarely homogeneous and can be resolved into orthogonalminimum and maximum. The fracture will open against the direction of leastresistance, perpendicular to the direction of least compressive stresses. Therefore, it propagates parallel to the direction of maximum horizontalin-situ stress. Fig. 4 shows the fracture orientation and position relative tothe horizontal stress directions and injection front. Fracture initiation atthe wellbore is expected to occur at the point where stress reduction isgreatest. P. 384