API - REPORT 26
Stage II - Time History Analyses Earthquake Analytical Studies for Offshore Structures
|Publication Date:||1 October 1981|
The overall objective of this Earthquake Analytical Studies for Offshore Structures project is to provide a basis for evaluating and improving current API RP-2A design guidelines for earthquake resistance, with particular emphasis on platform capacity and ductility requirements.
In the project, an example steel template-type drilling platform was simulated mathematically and analyzed through computer program INTRA, a recently-developed nonlinear finite element program. The structure's response to earthquake ground motion was calculated by several different methods consistent with current API RP-2A recommendations. The baseline analyses done in Stage I of the project comprised a typical strength requirement response spectrum analysis and an equivalent static pushover analysis to check ductility. In Stage II, the analytical model was subjected to three-dimensional ground motion histories from three real earthquakes, scaled first to roughly correspond to the API strength requirement and then doubled to roughly correspond to the ductility requirement.
In all of the strength requirement analyses, all of the superstructure and pile elements remained within their proportional limits, thus satisfying the API strength requirement. However, the responses calculated in each analysis differed substantially. The variations in some key structural forces, displacements and baseline strain energies were as high as 78, 42, and 142 percent respectively. These rather large variations were attributed to a combination of several factors:
1. Differences in the intensities of the input shaking, even though the records were nominally scaled to match the API design spectrum.
2. Differences in the characteristics of ground shaking, which led to differences in the distribution of forces in the structure.
3. Direct combination of time history responses versus SRSS combination of the spatial and modal responses in the response spectrum analysis.
In the ductility analyses, the analytical model included a detailed nonlinear inelastic model of the supporting soil. It appears that the soil limited the amount of earthquake force transmitted into the superstructure in the time history analyses, and further, that the soil accumulated substantial energy if the duration of shaking was sufficiently long. Of course this load limiting effect was not displayed in the static ductility check, where "equivalent" static loads were applied directly to the superstructure. The structure showed approximately 50 percent more mudline shear capacity in one of the time histories than was displayed in the static ductility check; this may have been due to soil damping. Limited damage (inelastic action) was displayed by the superstructure and piles in all but one of the ductility analyses, but the platform remained stable in all cases at energies at least six times the baseline strength requirement energy computed in the response spectrum analysis. However, in one time history, doubling the input ground accelerations was insufficient to demonstrate four times the baseline energy if the corresponding strength requirement time history was chosen as the baseline case.
It is concluded that the example analyses provide a sound data base for evaluating and improving current design guidelines. The wide variations in computed responses for different ground motion inputs and analytical methods emphasize the need for consistency between analytical procedures used in strength and ductility analyses. Furthermore, the observed importance of the soil in absorbing energy warrants efforts to standardize procedures for modeling soils.