Ring-on-ring tests and load capacity of cladding glass

Abstract

Although ring-on-ring test results have been used in the past to obtain information on the strength of glass, no methodology has so far been developed in the literature explicitly relating such results to the load capacity of cladding glass. The main purpose of this report is to propose such a methodology. The proposed methodology makes use of recent advances in the modeling of the fracture mechanics behavior of glass and the calculation of stresses in plates exhibiting geometric nonlinearity. Evidence is presented which strongly suggests that the probability distribution of the load capacity of cladding glass panels whose failure is due to surface flaws can be estimated reliably on the basis of results of ring-on-ring tests used in conjunction with (a) numerical methods for the analysis of stresses in plates, and (b) information on the elastic and fracture mechanics behavior of glass currently available or that can be obtained routinely. Two interesting findings are noted. First, owing to the way in which results of ring-on-ring tests are utilized, the relatively large variabilities typical of fracture mechanics parameters, as well as the uncertainties with respect to the shapes of surface flaws, have a minor effect on the estimation of load capacities. Second, two-parameter Weibull distributions, previously used in the literature to model the strength of glass and the load capacity of cladding panels, are not consistent with experimental results. On the other hand, three-parameter Weibull distributions model the observed glass behavior credibly.

A numerical method for analysis of fracture statistics of glass and simulations of a double ring bending test

“…Virtual glass specimens were tested until failure in double ring bending and the results were compared with data from the experiment conducted by Simiu et al (1984). The analysis was carried out using the software ABAQUS/CAE (2013) and MATLAB (2016).…”

“…The Weibull shape parameter value that was estimated based on the double ring bending experiment carried out by Simiu et al (1984), i.e. m = 3.8, might indicate a high dispersion for the experimental data because the value is quite low.…”

“…The data report is complete with values for the fracture stress even when the failure originated from outside the loading ring. Because Simiu et al (1984) reported the fracture locations it is possible to make comparisons with the simulated failure origin data. The data is challenging to model.…”

“…Some experimenters discard any observation associated with a failure originating from outside the loading ring radius. Simiu et al (1984) carried out experiments on 56 small square glass plates in double ring bending. The plates had the nominal dimensions 179 × 179 × 6 mm 3 .…”

“…Using a formula for a flat circular plate of constant thickness, Simiu et al (1984) calculated the in-plane MPTS for each fractured specimen. The stress was calculated at the centre of the plate, even when the fracture origin was not located within the loading ring radius.…”