D. B. JOHNSON, D.D. BALDWIN, J.R. LONG
Fig. 12 presents a comparison between the failure envelope as predicted by the finite element analyses and test results under both axial and combined internal pressure and axial loads. The combined load specimens were tested under conditions of internal pressure with end effect. The failure under teansile load occurred as predicted in the MCI (Fig. 13) at an average of 942 kips. As shown in Fig. 12, the failure envelope predicts that there is an equal probability that the failure of the CPR under the condition of pressure test with end effect (i.e. combined pressure and axial load) may occur in the MCI (tensile failure, Fig. 14) or in the tube body (burst failure. Fig. 15). The pressure testing of six specimens resulted in three failures in the MCI and three in the tube body, at an average of 11,635 psi. These results clearly demonstrate the accuracy of the analytical predictions. The small variations in the burst and ultimate tensile strength results illustrate the suitability and cost effectiveness of the fabrication and quality control procedures.
In order to establish industry confidence in the long-term behavior of the CPR joints under load, several joints were tested under conditions of static fatigue (stress rupture). An objective of these tests is to establish the time dependent allowable strength values for the composite laminates at different failure probabilities. Fig. 16 presents the results of the static fatigue tests, along with the static fatigue curves calculated using probability functions established by Robinson for the stress rupture performance of composite pressure vessels (Ref. 8). The slope of the regression line of the test results shows a strength loss of about 2.5% per decade.
Fig. 17 presents the results of the cyclic fatigue tests, along with the cyclic fatigue or S-N curves. The slope of the fatigue curve depends on the material system (i.e., carbon or glass) and the failure mode ( i.e., tensile, compressive or shear). For failure dominated by shear strength, as is the case for the traps, published data indicates that the value of n for carbon fiber composites is between 18.73 and 23.83. To ensure conservatism, a value of 18.73 was used. The second parameter is derived by considering the fatigue strength corresponding to 1/4 cycle. This fatigue strength value is equal to the static strength of the joints at the service temperature. The accuracy of the static strength predictions has been confirmed as shown in Fig. 18. For the mean fatigue curve, the value selected is the mean or average strength at temperature based on static ultimate test data. For the design fatigue curve, the value used is the A-basis allowable strength at temperature based on the statistical distribution of the static ultimate test data. Using the above approach, the mean S-N fatigue curves for 70｡? and 170｡? service are as shown in Fig. 17, along with design curves based on A-basis allowables. Shown also on Fig. 17 are the results of fatigue tests of twelve CPR joints at three different stress ranges. The accuracy of this approach is demonstrated by the test results.
Finally, Fig. 18 shows the average strength envelope for the CPR. These results show good correlation with predictions made by FEA. with failures occurring at predicted stress levels consistent with composite material strengths developed by CEAC. Clearly, this testing demonstrates that the tube body and final MCI have adequate safely margin with respect to the expected service loads for the CPR.
Full-length Joint Fabrication
Upon completion of the testing of the 65 pre-production prototype joints, fabrication and testing of three full-length joints began. The fabrication process was identical to that described earlier. No problems were encountered during the fabrication (Figs. 19 and 20). The quality control plan developed during pre-production fabrication was closely followed and is proposed that this would be adopted for certification. A full-length CPR joint is shown in Fig. 21.
Conclusions
The CPR project is a joint-industry development of a new product for offshore application, under the auspices of the NIST ATP. During the project, a total of 80 prototypes have been fabricated and tested ─ three 7” subscale joints, 12 design verification test joints, and 65 pre-production prototype joints. Performance envelope, static and cyclic fatigue curves have been generated, showing compliance with the Functional Specification and Performance Criteria document developed by the end-users. At this point, the CPR design has been shown to meet the cost, weight and performance goals of the project.
It should be noted, however, that some CPR prototypes remain in testing, and additional prototypes are being fabricated to put into testing, with the expectation of obtaining even longer-term static and cyclic test data than that presented here.
Acknowledgments
The authors would like to express their gratitude to the NIST ATP program office and the sponsoring companies (Amoco, Conoco, Shell, Brown & Root, Hexcel, Hydril, Lincoln Composites, Stress Engineering, and CEAC/UofH) for their technical and financial support. The authors would like to thank Carol Schutte and Felix Wu of NIST for providing clear direction while encouraging innovation. Special recognition is due to Su-Su Wang, Akira Miyase, Xiaohua Lu, Metin Karayaka and Zhong-Qing Gong of the CEAC/UofH for their excellent work on this project. The authors would also like to acknowledge the efforts of Bob Burden (formerly of Hydril) in the design and analysis of the 12 3/4-inch MAC-II connection.
And finally, the authors also express their gratitude to the following representatives in this program, each of whom has contributed to the group's understanding and whose input is reflected in this program: Hin Chiu, Bill Cole, and Bernie Stahl of Amoco; Fikry Botros. Ashok Kumar and Mamdouh Salama of Conoco; Chris Howell, F. Joseph Fischer and K. Him Lo of Shell; Mohammed Abdullah of Hexcel; and, Reza Rashedi of Brown & Root.
