Consistency In Glass Design

Glass structural elements have become increasingly common to the point of ubiquity; however, there currently is no universally recognized and codified Glass Standard in the United States. As a result, while there have been several efforts at developing design guides for structural glass (ISE Structural Use of Glass in Buildings, ASTM Standards, GANA Glazing manual etc.) each has somewhat different design methodologies and concomitant assumptions. With the variety of design references, there is a potential for accidental confusion of valid application of these methods.

Recently, the NCSEA Engineering Structural Glass Design Guide was published with equations for glass strength that are predicated on an “allowable” glass stress in combination with a series of adjustment factors. The resulting method is attractive in its apparent simplicity. However, sole reliance on the single largest maximum principal tensile stress (SLMPTS) may not always be representative of actual performance as random surface flaws often precipitate failure at lower stresses and different locations from the SLMPTS.

This paper analyzes the design examples from the NCSEA guide using the glass failure prediction model from ASTM E1300 to determine the probability of failure for each of the examples using various designs. Comparisons of the results show the importance of accurately modeling actual material behavior to 1) minimize material usage through efficient designs and 2) produce designs with consistently reliable levels of risk. Results show that maximal sustainability of glass (via minimal material usage) correlates to consistent probability of failure metrics. The average plate thickness found using allowable stress design methods is 16.4% greater than those calculated with strictly probabilistic-based design.

While the GFPM is potentially more computationally involved than closed-form equations in other references, the results indicate that the equations from the NCSEA guide may produce conservative designs, but they do not result in consistent probabilities of failure.