Developing a novel technique for testing glass strength

Abstract

The use of glass in structural applications has become more and more prevalent, with applications like modern cruise ships incorporating large glass façades into their architecture. With this trend ongoing, the concept of structural optimization is naturally extended to reach glass elements, as part of a global shift towards lightweight, more sustainable design. The foundation of an optimized glass structural element is an efficient method for strength determination. Glass is a brittle material; that is, its strength is primarily driven by the flaws existing in its surface rather than intrinsic bond strength between atoms. This makes characterizing its strength a challenging task. For this reason, glass strength should be tested so that a uniform biaxial stress field is created within the test area to eliminate the influence of the orientation of surface flaws, as flaws are most critical when they are normal to the direction of the maximum stress. The four-point-bending test (4pb) and the coaxial double ring (CDR) test are two of the most common standardised testing methods for glass strength. However, these tests do not present efficient methods for determining glass strength, as the former includes glass edges, where flaws are more critical, within the region of maximum stress, while the latter only considers small specimens. For large specimens that undergo large deflections where the effect of geometric nonlinearity is significant, the standardized approach for the CDR test not only becomes overly sophisticated, but it also fails to maintain an equibiaxial stress state. This thesis aims to propose a novel technique for testing glass strength, utilizing the CDR method, that is simpler and more efficient than the current common testing methods. 21 identical fully tempered glass specimens of dimensions 1000×1000×6 mm were tested using the proposed technique, utilizing tools like Digital Image Correlation (DIC) and strain gauges for data acquisition. The results showed that the proposed test setup is capable of reproducing the same loading and boundary conditions for the majority of the tested specimens, while maintaining a somewhat uniform stress field under large deflections. Overall, the proposed test setup showed efficacy and potentially presents a simpler and reliable alternative to the standardised testing methods currently in use.