The effect of size and stress state on the strength of architectural glass. Experiments versus theory

The multi-scale approach in the statistical evaluation of the macroscopic strength of annealed glass for architectural applications indicates a close correlation with the population of surface microcracks present in the glass sheets. The theory predicts a dependence of the strength capacity on the size of the glass element and on the state of stress acting on it. To investigate this effect, an experimental campaign has been conducted on samples of three different sizes, tested in a ring-on-ring apparatus. Data were re-normalized according to statistics of the Weibull type and re-scaled to account for the phenomenon of static fatigue. For small- and medium-size specimens, the results are in agreement with the prediction of Weibull statistical theory, according to which the larger the specimen, the greater the probability of finding cracks of critical size and, consequently, the lower the macroscopic strength on a statistical basis. On the other hand, quite surprisingly, large samples show an opposite trend. Since the state of stress is not uniform and equibiaxial in the large specimens, on the contrary to what happens in the small and medium size ones, this finding suggest a revision of classical approach based on linear elastic fracture mechanics and its correlation with the weakest-link-in-the-chain rationale in the evaluation of the mechanical properties of brittle materials such as glass.