The fracture stress of float glass

Abstract

A fracture test [1] which uses concentrically loaded square plates supported near their corners has been used to measure the fracture stress of float glass. The plates were 102mm square and 5.98mm thick. The maximum displacement at fracture was less than 0.4mm. Under these circumstances it has been shown that use of a linear finite element solution for the stress distribution and the plate deflections is justified. The glass plates had greater edge damage than had the alumina plates tested in an earlier investigation. In order to secure an adequate proportion of failures in the central plate region, it was necessary to move the supports inwards towards the centre of the plate. This reduced the ratio of the maximum edge stress to the maximum stress in the plate. Batches of plates were tested with loading circle diameters of 7.5 and 25mm, to measure volume effects, with the side of the plate that had been in contact with the liquid tin in tension. Median ranking was used in the statistical analysis and edge failures were treated as suspensions, it being assumed that the minimum fracture stress of the central region of the edge-fractured plates was the plate centre stress at the fracture load. The Weibull modulus was determined by a linear regression in which extreme members of the population were given reduced weighting using the relationship of Faucher and Tyson [3]. The average fracture stresses were 147.2 and 107.3 N mm−2 for the 7.5 and 25 mm loading circles, respectively, and the Weibull moduli were 4.49 and 5.44. These data are shown to agree well with Weibull statistics. Tests using a 7.5 mm diameter loading circle on plates with the non-tin side in tension gave a significantly higher average fracture stress of 242.1 N mm−2, confirming the fact that the non-tin side has a higher strength.