Static Fatigue (Stress Rupture). These specimens were pressurized to a constant pressure and tested for various length exposures. The specimens were held at constant internal pressure until they burst due to static fatigue. End caps were used to close the samples and apply the pressure. The samples were free to move, blocked up on wooden supports. The pressure end load was resisted by the end caps and transferred into the sample, as in the internal pressure burst tests. The samples were placed in individual safety containment chambers, and each was monitored separately by reading and recording a gage at periodic intervals.
External Pressure Tests. These specimens were tested under pure external pressure or under a combined external pressure and tension. External pressure was applied to the specimens by placing them in a test chamber consisting of a large diameter steel pipe with o-ring seals and containment flanges which allowed the chamber to seal on the outside diameter of the steel connectors at each end of the CPR joint. In this manner, the entire composite portion was subjected to external pressure, with no pressure end load effects applied to the part itself. This resulted in a pure external pressure loading on the part. The pressure was increased until the part collapses. In the combined loading tests, a specified external pressure was applied, and then axial tension was increased to failure.
Cyclic Fatigue Testing. These tests were performed under a cyclic tension load range centered about some mean tension level. The testing was performed in a load frame with a sinusoidal tension load application. Testing rate was 1 Hz or less depending on sample stretch (cylinder stroke), and load level. Cyclic fatigue testing was performed with 500 psi constant internal pressurization.
Design Verification Testing
Seven full-scale diameter/short length CPR specimens were tested under conditions of pure internal pressure, external pressure, axial load, and combined axial load and pressure. Fig. 9 presents a comparison between the predicted failure envelope based on the finite element analyses and the test results. The predicted failures were based on the allowable strength values given in Table 2 for both carbon/epoxy and glass/epoxy laminates.
The external pressure collapse capability of the tube body was demonstrated in two tests to be greater than 3550 psi, which exceeds the project requirement of 3300 psi (1.5 times the hydrostatic pressure at 5000 ft depth). In an internal pressure test with pressure end load removed, the tube body ruptured at 11,161 psi, exceeding the required internal pressure capability of 9000 psi.
A joint was tested in axial tension, reaching 824,500 pounds at failure which is higher than the goal of 770,000 pounds. The failure occurred at the MCI. Also, the measured extension of the specimen under load confirmed that the composite tube body meets the 100 (msi)(in2) requirement for axial stiffness. However, testing under combined internal pressure and axial loading indicated that one of the specimens failed at an internal pressure (with end effects) of 7700 psi with the failure in the metal-to-composite interface. This pressure was less than the goal of 9000 psi. The strain gage data from this test indicated that a bending moment may have been applied to the specimen by the test fixture, so the test was repeated with a different setup. This second internal pressure burst achieved a value of 9160 psi. A test was performed where the joint was pressurized to 6126 psi and axial load was applied until failure. An MCI failure occurred when the applied axial load reached 225,000 pounds, about 25,000 pounds less than desired.
In addition to the above tests, three CPR specimens were pressure tested after being subjected to 1 kJ drop impacts representative of rough handling, with each joint being impacted in the tube body and over the MCI. The impact damage was inflicted at the mid-specimen by dropping the specimen from a horizontal position onto a 4-inch diameter steel bar. The damage was inflicted at the metal-to-composite interface by raising one end fitting so the specimen was at approximately a 45° angle to the floor and then dropped on a 4-inch diameter bar.
Figs. 10 and 11 show the extent of damage on the 0.D. and I.D. of the mid-specimen impact. X-Ray radiographs of the damage areas revealed some delamination between the carbon helical and E-glass hoop layers that was indicated by surface inspection. Testing of these units showed a reduction of 48% in the burst strength of the tube body and a reduction of 24% to 39% in axial tension capability.
Building on the experience gained in design, manufacturing and testing of the CPR joints, a redesign of the MCI was performed because the analysis indicated that joint performance could be improved significantly without increasing product cost. The improvements involved optimization of the trap shape and some additional reinforcements in the traps. This improved design was used to fabricate the 65 joints which were used for the following design characterization testing.
Design Characterization Testing
Characterization of the final CPR design was accomplished by testing 65 full-scale diameter/short length specimens that were fabricated with the proposed laminate and metal-to-composite end fitting. All short-term rupture and static fatigue specimens used the same stub Acme thread as was used for the design verification testing. All cyclic fatigue specimens used the Hydril MAC-II 12 3/4-inch, 94.20 lb/ft premium thread, since the Acme thread would have failed well before the composite being tested. These 65 specimens were tested under the load conditions shown in Table 3.
The objective of this series of tests was to generate sufficient data from which a correlation between analytical and empirical data could be established and statistical variability could be assessed.
