An optimization framework for the design of LMUs used in offshore platform float-over installation

Abstract

Leg mating units (LMU), consisting of steels and hyperelastic materials, are utilized to protect the structures during offshore platform float-over installation. Initial LMU design usually does not satisfy required performance since it is geometrically adjusted to fit given requirements based on previous design. To meet design criteria, design optimization can be applied to the initial design. However, a direct implementation of an accurate local search algorithm may cause heavy computations with the risk of falling into a local optimum. It can be also challenging to find an accurate optimum with efficient global search algorithms due to the characteristics of LMU problems. A small perturbation in nonlinear material can have a significant impact on LMU performance. Therefore, in this research, to find an optimum LMU design efficiently and accurately, an optimization framework that combines global and local search algorithms is implemented. In the framework, an efficient meta-model of optimal prognosis based evolutionary algorithm is first applied. The optimal design obtained from the first step, premature but near the optimum, is set as a new starting point for the sequential optimization where accurate polynomial based local adaptive response surface method is utilized. The robustness of the proposed framework is shown by the development of an LMU suitable for float-over installation of a 20,000-ton topside over an eight-legged fixed offshore jacket platform.