Topology optimization for stability problems of submerged structures using the TOBS method.

• Ano de defesa: 2021
• Autor(a) principal: Mendes, Eduardo Aguiar
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Binary variables; Buckling constraints; Carregamento de pressão; Otimização topológica; Pressure loading; Restrição de flambagem; Topology optimization; Variáveis binárias
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3151/tde-22022022-114244/

Structural topology optimization is increasingly used across academia and industry because of the great design freedom it offers and due to the rising computational power availability. Typical Topology Optimization (TO) problems seek stiffness maximization for volume-constrained structures via density-based methods, which may generate solutions with poor stability performance, e.g. prone to buckling. A valid alternative is to include the buckling parameter as a constraint in order to obtain final designs that fulfill this criterion. In this context, binary methods - which generates clear [0,1] designs - emerge as an effective approach to solve multiphysics problems, wherein precise definition of the structural boundary is essential. A challenging TO application that benefits from this class of methods are submerged structures, e.g. offshore industry components, which are subject to design-dependent loads and might present stability issues. This loading type imposes a constant change on fluid loading location, direction and magnitude, which is not trivial for optimization procedures. In this scenario, the aim of this work it to investigate the binary nature of the TOBS method by solving topology optimization problems that consider buckling constraints and design-dependent loads, characteristic of submerged structural systems. The proposed topology optimization problem has not been explored in the literature. The linear buckling implementation is verified through analytical methods, and a benchmark optimization problem for buckling-constrained formulation is solved for efficiency analysis. Numerical examples of pressure-loaded structures are optimized and investigated regarding the stability parameter effect when compared to classic compliance minimization solutions. Further discussions are held concerning the common issues associated with the buckling eigenproblem, as well as the main parameters adopted in the TOBS method. The proposed binary framework presented promising results by obtaining final solutions with significant improvement in buckling resistance and minimal stiffness loss when compared to the compliance designs. Computational time studies showed that the buckling sensitivities are the bottleneck of the optimization process and, thus, alternative techniques should be investigated.