Probability Distributions in the Glass Failure Prediction Model

  • Samir Blanchet Texas Tech University
  • H. Scott Norville Texas Tech University
  • Stephen M. Morse Michigan Tech University


Glass, a brittle material, fractures under tensile stress acting over a time duration. Lateral loads, such as wind, acting on a simply supported rectangular glass lite, put one surface of the lite primarily into tension. ASTM E 1300 defines load resistance of glass as the uniform lateral loading acting over a duration of 3 seconds that is associated with a probability of breakage of 8 lites per 1000 at the first occurrence of the loading. To determine load resistance, the underlying window glass failure prediction model facilitates determination of a probability distribution of 3 second equivalent failure loads, P3. The glass failure prediction model is based on a Weibull distribution, and most people believe the distribution of P3 is, in fact, a Weibull distribution. However, the authors contend that this is not the case. This paper provides an explanation of the glass failure prediction model, its basis, and a discussion of the method for determining surface flaw parameters with an example. The authors demonstrate the distribution of the equivalent failure loads does not follow a Weibull distribution, and they will elucidate the relationship between the distribution of P3 and the Weibull distribution.

How to Cite
BLANCHET, Samir; NORVILLE, H. Scott; MORSE, Stephen M.. Probability Distributions in the Glass Failure Prediction Model. Challenging Glass Conference Proceedings, [S.l.], v. 6, p. 681-684, may 2018. ISSN 2589-8019. Available at: <>. Date accessed: 27 sep. 2020. doi: