Response Calculations of Semi-Submersible Column Exposed to Slamming Loads
Master thesis
Date
2016Metadata
Show full item recordCollections
- Institutt for marin teknikk [3526]
Abstract
Slamming is a highly non-linear phenomena arising from interaction between fluid and structure. A slamming load is characterized by high pressure peaks localized to a limited part of a structure, with short time duration. It is well known for causing problems related to structural integrity and fatigue, and is among the highest loads acting on offshore structures in rough environments.
Several uncertainties contributes to challenges when estimating the effect of slamming loads on offshore structures. Two areas of concern are method for calculation added mass and the effect of hydroelasticity. Today, there is no simple way to calculate loads from breaking waves, and the recommended practice is to perform model tests. A typical model test is carried out on a rigid model that does not deflect locally due to wave loads. This is not the case for a full-scale platform, and the challenge is to determine how to modify the loads from a model test to include the effects of added mass and hydroelasticity. The topics are investigated from a structural point of view, with basis in one of the columns on the semi-submersible production/drilling platform Sinaa. The main objective has been to study how added mass and hydroelasticity affect the structural response, by use of finite element analysis.
Hydroelasticity is the mutual interaction between fluid and structure. When the water pressure acts on the structure, the speed and shape of the structural deformation influence the pressure in the water, which again influence the structural deformation. There are mainly two methods for including the effect of hydroelasticity in a structural analysis. The loads obtained in a model test may be modified to account for the effects, or an integrated hydrodynamic and structural analysis may be performed. A coupled analysis will provide the most correct realization of the hydroelastic effects, but this method is time consuming and complex. If the extent of the hydroelastic effects are known, it is possible to modify the load measured from a rigid structure model test to include these effects, by multiplying the duration with a factor and divide the magnitude with the same factor. This approach corresponds with the typical consequences of hydroelasticity, which are lowered pressure peak and increased load duration.
When a breaking wave impact with a structure, energy is used to accelerate and deform the structure, as well as decelerate the fluid. The hydrodynamic forces that are in phase with the acceleration, are called added mass. Added mass is frequency dependent, and since a rigid structure is used in model tests, limited forces on the structure related to added mass are captured. The majority of theoretical methods used to estimate added mass, are based on oscillations of entire rigid bodies. For slamming, only a limited area experience vibrations that will lead to localized added mass forces. Calculations to estimate the added mass for the platform column studied, were performed in the numerical analysis software WAMIT. Two theoretical approaches were also used. The theoretical formula for added mass for a rigid surface, inserted with a circular piston to simulate the structural response, was proven to give a too high estimate. The formula defined in DNV-RP-C205 gave an estimate more similar to numerical calculations.
To study the effect of added mass and hydroelasticity on the structural response, finite element analyses were performed by use of the software Abaqus. The sensitivity studies show that both added mass and hydroelasticity influence the structural response. Up to a certain point, hydroelastic effects reduce the deformation significantly. In this range, dynamics effects are essential for the response, and it is necessary to perform dynamic FE-analyses. If the load duration increase above a limit (approximately 1.2 seconds), the structure behave similar to a static situation. The results show unrealistic high deformations in the static case, and the structure was completely destroyed. In terms of added mass, an increasing added mass leads to a decreasing structural response. It requires more energy to accelerate a higher mass and since the force on the structure is constant, the result of an increased mass is lower acceleration and response. Assuming that the numerical estimate for the added mass is correct, the structural deformation decreases by approximately 50\% when including added mass in the finite element analyses.
It appears that most of the uncertainties related to added mass and hydroelasticity in FE-analyses of slamming events, are connected to the estimation of these effects, and not how the effects are included in the FE-analyses. The comparison with a single degree of freedom system, confirms that the physics are simulated in a realistic way in the FE-analyses performed. There are uncertainties associated with the results of the analyses performed, but the trends clearly show that both added mass and hydroelasticity affect the structural response. A slamming analysis not including these effects, is likely to be very conservative.