STG Symposium "Selected Topics on Strength, Vibrations and Noise"

Structural optimization of a large cruise ship section

Carl Rupp, Technische Universität Hamburg

In July 2023 the IMO aggravated the strategy for the reduction of greenhouse gas emissions from ships. This leads to a strategy which calls out a net-zero greenhouse gas emissions in international shipping until 2050. This challenges all fields in maritime industry to reduce the overall emissions. A common opportunity to reduce the energy needed to move a ship is to increase its efficiency. This is also highly relevant for alternative fuel concepts. The contribution of structural design to this field can be the creation of lightweight structures with optimized designs reducing the power demand. This paper presents a method reducing the weight of a large ship section. The optimization is combining the structural response of a local FEM model of a ship structure with the particle swarm optimization.

Therefore, the structure is build parametric so that the employed particle swarm optimization (PSO) can alter the FE model. The focus is here especially on the change of the plate thicknesses, while using thicknesses, that are commercially available. The solution space is of significant size as we have 78 parameters, which can have 10 different values (thickness range) there are 10 to the power of 78 structure variations possible. As boundary condition a global bending case is applied and is transferred to the local FE model. The deformations are evaluated in sections as linear functions from the global model and interpolated to the local one. It is assumed, that the applied static load case is relevant for the dimension of each plate. Solution constraints are the stress limits of the DNV class rules. Moreover the influence of component stresses in longitudinal and transversal directions in comparison to equivalent stresses is investigated. The process is illustrated on a large cruise ship section that is optimized reducing its weight. Expected are weight savings of 5 to 10 percent of the original weight. The time consuming process of building a parametric model and performing the optimization is considered.