Prediction of TLP Responses: Model Tests vs. Analysis

Abstract This paper presents an overview of prediction of TLP responses: model tests vs. analysis, sponsored by DeepStar Phase V program. ABB and Marintek were invited to carry out the task, which was intended to provide an overall assessment of the current capabilities of the industry in predicting TLP responses and highlight areas of uncertainties and sensitivities. The paper summarizes key results of TLP responses in 6,000 ft water depth of Gluf of Mexico field. Non-linear coupled dynamic analyses were employed indenpently by both ABB and Marintek to model the system in a consistent and accurate manner. The measured hull, tendon and riser configurations, as well as the measured wave elevations, wind loads and mean current velocity profile were applied. The overall correlations between model tests and analyses demonstrated the industry has the analytical capability in predicting the TLP responses. However, analytical tools are not perfect and physical model test is still an important design tool to verify the analyses. Introduction The DeepStar Program sponsored a series of tasks to evaluate the current industry capability in predicting the responses of deepwater theme structures (FPSO, TLP and SPAR). In its Phase IV program, model tests of the FPSO, TLP and SPAR were conducted, and in the Phase V program, engineering companies as well as test basins were invited to evaluate the correlations between the tests and analyses. The TLP physical model tests were carried out in MARIN wave basin tank in January 2001 in water depth of 6,000 feet under Gulf of Mexico hurricane and loop-current conditions. The details of test setups, wave, wind and current generations and cablibrations were described in paper OTC 16582 (ref. 1). In deep or ultra deepwater, the TLP platform tends to interact more pronouncedly to its tendons and risers. The dynamic interactions among platform, tendons and risers cannot be evaluated accurately and consistently by using the conventional uncoupled analysis tools, where the platform, tendons and risers are treated separately. Therefore, analytical capability of fully coupled dynamic analyses was required to complete the project. Both ABB and Marintek performed fully coupled dynamic analyses indenpenly by their own softwares. The purpose of this excerise is to investigate whether the existing numerical tool applying the measured data could reproduce the measured results and identify the gaps for further study. This paper presents an overview of both ABB's and Marintek's work. It includs the following aspects of the study:Model test setup and environmental criteriaThe state-of-art analytical tools available to the IndustryKey results of comparisons - tests vs analysesSensitivities and uncertainties in predicting TLP responsesAssessment of current industry capabilitiesAreas of future efforts Model Test Setup TLP hull, tendon and riser configurations, tendon numbering, both tendon and riser locations, tendon porch elevation and riser top elevations were illustrated in Figure 1.Hull ParameterHull configuration, measured hull weight, tendon top tensions and riser top tensions are given in Table 1.Tendon ParameterTendon configuration, measured tendon dry weight and wet weight are summarized in Table 2.