INFLUENCE OF DIFFERENT MODELING TECHNIQUES ON THE PERFORMANCE OF PARICÁ GLULAM BEAMS

Abstract

Despite advances in timber structure modeling, a gap remains in the comparative assessment of techniques applied to glued laminated timber (glulam) beams made from Amazonian species, aggravated by the scarcity of physical–mechanical data. This study evaluates the impact of different numerical modeling strategies on the structural performance of paricá glulam beams. Three approaches were considered: two layered-section models with orthotropic and isotropic behavior, and a unified orthotropic section. Geometry and properties were selected from the literature and complemented by normative correlations, with validation against four-point bending tests and a mesh sensitivity analysis. Among the discretizations evaluated, the beam with 10 mm elements presented the best balance between accuracy and computational time. Considering the values from the experimental reference, the layered and unified models satisfactorily reproduced the global responses, with errors of 0.75% and 0.76%, respectively, whereas the isotropic assumption artificially increased flexural stiffness and resulted in a 7.25% error. The stress distributions confirmed that orthotropy was crucial to capture gradients and local concentrations, while the unified section maintained good agreement with reduced processing time.