Multi-criteria design of undersea arch-type structures using shape parameterization

Abstract

The increasing need for infrastructure to solve communication problems between territories motivates the development of new techniques to find the better solution, taking into account technical, economic, safety and serviceability criteria. One of the most used structural type to design submerged structures has been the arches, which design involves several variables to meet the aforementioned requirements.
The design of such structures was traditionally focused in their funicular or moment less shapes. Past works, shown that for intermediate depth ratios, the funicular shape of a submerged arch has a form between the ellipse and parabola conic curves.
Later, some authors have proposed to involve in the design of these structures other aspects as the enclosed airspace to improve their serviceability, in addition to the arches mechanical behavior. For this purpose, they proposed to parameterize the
curves that defined the shapes of the submerged arches and use metaheuristics to find the shape of the most appropriate solution for the submerged arches according with these criteria.
These processes traditionally also require the implementation of a finite element model and the corresponding parametric function to evaluate the mechanical behavior of these structures. Since this process may be greatly simplified by using machine learning, this opens a new path to be considered in the design of these types of structures.
In this work, the methodologies to design submerged arches type structures are presented. To this aim, three different parametric functions are considered: conics, elliptics and Bézier curves, and a proposal of surrogate model are presented.