Numerical modeling of glued laminated timber beams without finger joints: Identifying load-bearing capacity and analyzing failure mechanisms

Abstract

Developing simulation approaches to predict the structural behavior of glued laminated timber (GLT) beams under loading is essential for formulating efficient design concepts, particularly for large-span structures. While experimental data for GLT beams according to European standards are limited to depths of about 1 m with rather small sample sizes, a substantial knowledge gap exists for beams already used in practice with depths up to 3 m. For such large beams, conflicting size effects have been reported by simulation studies. Numerical modeling of GLT beams is challenging, e.g., the quasi-brittle nature of wood requires modeling progressive failure mechanisms and nonlinear behavior during loading. This study examines the influence of three global failure criteria on bending strength, modulus of elasticity (MOE), and damage states of simulated GLT beams without finger joints. The examination is based on data from 11 800 simulations covering seven beam sizes up to 3.3 m and two strength classes. The simulations incorporated discrete vertical and horizontal cracking and plastic deformations. The findings reveal that the choice of failure criterion significantly influences the predicted bending strength, while the MOE remains practically unchanged. A load-drop criterion most effectively captured the peak loads, highlighting the role of progressive damage accumulation. The reported damage states provide insights into the failure processes modeled in the simulations. These results underscore the importance of selecting appropriate failure criteria and mechanisms in numerical models to predict the performance of large-scale GLT beams.