An Experimental Study of Casing Performance Under Thermal Cycling Conditions

Abstract

Summary. We investigated the behavior of casing pipe body and connections under simulated thermal recovery conditions. The study, performed in a new computer-controlled thermal-well simulator, examined the thermal stress behavior and leak resistance of pipe and connections at temperatures up to 354 degrees C [670 degrees F] under severe loading conditions similar to those encountered in thermal wells. We also studied the biaxial collapse resistance of the casing under the large axial tension that would exist after the cooling period in a steam-stimulation process. Results indicate that in a steam-injection process in which the casing is maintained at its maximum temperature for a period of time, stress relaxation can occur in a constrained production casing, resulting in the development of excessive tensile stress during the cooling period. Premium connections with metal-to-metal seals maintained gas leak resistance at temperatures up to 354 degrees C [670 degrees F], whereas the connections without metal seals did not. The biaxial collapse resistance of casing under large axial tension depended on the stress/strain characteristics of the material. Hot-rolled Grade K55 casing with work-hardening characteristics showed a biaxial collapse resistance higher than that predicted by the von Mises yield criterion. From results of this study, we concluded that heavyweight Grade K55 casing with premium connections is a good candidate for thermal well completions. Introduction Cyclic steam stimulation can impose severe stresses on well casing. In this process, steam is injected into the formation through the tubing or casing for a period of a few days to several weeks. The steam is then allowed to soak for a short period, after which the well is placed on production. Thus, the casing is subjected to a period of heating during the steam-injection phase, a short period of nearly constant temperature during the soak period, and a cooling period during subsequent well production. When casing is heated or cooled (as in the cyclic steam stimulation operation), one of two things will happen: the casing will either expand or contract if allowed to do so. If the casing is fixed at both ends, as is often the case when the casing is cemented to the surface, then compressive and tensile stresses are generated in the casing. These stresses may be large enough to exceed the pipe yield strength or the coupling joint strength, resulting in casing failure. Injection temperatures up to 343 degrees C [650 degrees F] and injection pressures up to 20.7 MPa [3,000 psi] have been reported in steam-injection wells. Willhite and Dietrich presented a conceptual model of the thermal load history of the casing during cyclic steam stimulation. During steam injection, the casing is heated and compressive stress is created in the casing in proportion to the temperature change. If the temperature is high enough, the yield strength of the casing material will be exceeded and the casing will become plastically deformed. Therefore, during the steam-injection phase of stimulation, the casing may fail as a result of plastic deformation and the connection may fail as a result of excessive compressive load. When the casing is cooled, the tensile stress generated may be high enough to cause tensile failure of the pipe or the connection. When the casing is cooled to its original temperature (before steam injection), a permanent residual tensile stress well be left in the casing. In addition to creating the potential for tensile failure, this residual tensile stress causes the casing to be more susceptible to biaxial collapse failure. Thermal-well operators are well aware of the dangers posed to casing by high temperatures and often take measures during well completions to minimize the risk of casing failure. One Step usually taken is the use of high-grade casing-such as P105, P110, S95, and S0095and API buttress thread coupling BTC) for thermal wells. The problem facing the casing designer is that very few experimental data are available on the amount of thermal stress to be expected and the sealing performance of the various connection types at steam-injection conditions on which to base casing design. The objectives of this study were (1) to measure the thermal load histories of various casing grades under simulated cyclic stem-stimulation conditions, (2) to test the leak resistance of API connections and the new premium connections at simulated steam-stimulation conditions, (3) to measure the biaxial collapse resistance of the casing materials under tensile loads similar to those generated in the casing during cyclic steam-stimulation, and (4) to propose a new casing design approach for thermal wells based on the results of our experimental measurements. We measured the thermal load behavior of casing pipes, high-temperature leak resistance of various connections, and the biaxial collapse resistance of the casing at simulated cyclic steam-stimulation conditions with a computer-aided thermal-well simulator and a biaxial-collapse tester. This paper presents the results of our study, which show that thick-walled Grade K55 casing equipped with a premium connection is a good candidate for high-temperature, steam-stimulation wells. Experimental Equipment and Procedure Thermal-Well Simulation Test. The thermal-well simulator was designed to subject a full-size pipe and connection to cyclic steam-injection conditions of thermal load and gas pressure. Fig. 1 is a schematic of the thermal-well simulator. Fig. 2 is a photograph of the simulator with a test sample installed. The simulator consists of (1) a 1900-mm [6.2-ft] -long electric heater, (2) hydraulic power cylinders capable of applying 9808 kN [2,205,000 lbf] of load, and (3) an integrated measurement and control system. Table 1 summarizes the simulator's specifications. A computer controls the operation of the thermal simulator. Feedback control systems control the temperature, thermal displacement, and thermal load. To control the temperature, six thermocouples are attached to the test sample. The measured and programmed temperatures are compared every second by the programmed temperatures are compared every second by the computer, and appropriate signals are transmitted to the temperature regulator to eliminate the difference between the two temperatures. To control the thermal displacement, two clamp rings are mounted 500 mm [19.7 in.] apart on the test sample. The relative movement of the rings is measured by magnetic scales located outside the heater and connected to the rings by heat-resistant steel rods. If a thermal elongation is about to occur, the computer permits a servo valve to apply enough compressive load on the test sample to keep its length constant. On the other hand, if the sample is about to contract, the computer applies sufficient tensile load to maintain a constant length. To test the sealing performance of the connection, high-pressure nitrogen gas was applied to the connection. Leaked gas was vented into a water vessel, where it was detected by photoelectric bubble detectors. SPEDE P. 156