Abstract
The strength of window glass is sometimes assessed using ring-on-ring tests that are conducted by placing a glass plate on a circular support ring and then applying loads through a smaller concentric loading ring. Ring-on-ring test results clearly demonstrate that the spatial variability of glass strengths has a very significant effect on the effective strength of a glass plate. The article shows how the effects of the spatial variability of glass strength in ring-on-ring tests can be modelled using a probabilistic glass failure prediction model (based on a Weibull model of glass strengths). In particular, it is shown that the model can be used to predict the proportions of failures that occur inside and outside the loading ring, and to predict the influence that spatial variability has on the test strength statistics. It is shown that the probability distributions of test strengths are identical for failures originating inside or outside the loading ring (or combined), even if the local strength distributions inside and outside the loading ring are not the same, provided the Weibull shape parameter is constant. If the local strength distributions inside and outside the loading ring are not the same, the only indication is given by the proportions of failures inside and outside the ring. The theoretical results are shown to agree with some ring-on-ring test results for 4 mm annealed glass. Also, some apparently anomalous test results for 6 mm annealed glass are presented, and a possible explanation is given, consistent with the theoretical model.