Geração de modelos de bielas e tirantes para estruturas de concreto armado por otimização topológica discreta via método ground structure

Abstract

The Strut-and-Tie Model (STM) is widely used for the design of reinforced concrete structures in D-regions, where the assumption of linear strain distribution is no longer valid. Traditionally, the definition of a STM relies heavily on the engineer’s experience and subjective interpretation of force flow, which may lead to conservative or inefficient solutions. In this context, this dissertation investigates the use of discrete topology optimization, based on the Ground Structure Method, as an auxiliary tool for the automated generation of strut-and-tie models. The proposed approach employs a bilinear material model that distinguishes between tensile and compressive behavior, combined with discrete and geometric filtering techniques to obtain simplified and directly interpretable topologies. Five numerical examples were analyzed: a simply supported beam used as a reference problem, a continuous beam and three structures representative of D-regions — a coupled shear wall, a deep beam with an opening, and a dapped beam. The results show that the method successfully identifies the fundamental load transfer mechanisms, producing physically consistent strut-and-tie models that are in good agreement with solutions reported in the literature. The application of separate filters for tension and compression elements revealed that the sensitivity of the optimal topology depends on the structural nature of the problem, allowing the exploration of alternative layouts without compromising structural feasibility. The study concludes that discrete topology optimization via the Ground Structure Method is a promising tool to support the design of strut-and-tie models in reinforced concrete structures.