Confidence Booster: Learn How to Incorporate Contemporary Design with Structural Glass

By Richard Green

Glass is one of the strongest materials in compression known to man. It is underutilized as a structural material. It is, however, brittle. Follow sound principles and combine glass with modern interlayers, and owners, architects and engineers can launch into advanced glass design confidently.

What is ‘Structural Glass?’

In jurisdictions governed by the International Building Code (IBC), glass design has two recognized design standards: ASTM E 1300 for determining the load resistance of glass in buildings and ASTM E 2751 for design and performance of supported laminated glass walkways. ASTM E 1300 has as its scope uniform lateral loads of windows and includes no redundancy or post-damage requirements. E 2751 includes the principles of retention,
redundancy, and residual capacity at full design loads in a post-damage condition, i.e. the critical ply broken. For our purposes, consider structural glass to mean glass that has consequences should it fail completely. Just as importantly, consider non-structural to be systems that have limited consequences. While the cases for windows and floors are well-defined, a wide range of glass applications fall between these two extremes that do not have guidance from codes or standards. Here are some principles for empowering responsible, economic and beautiful glass design.

What is Risk?

Structural engineers generally consider risk to be the probability of failure. In ductile systems, where the failure mechanisms are well-controlled, the probability of failure is a good approximation of the total risk. However, risk engineers define risk as the probability of failure multiplied by the consequence. This is reflected in the two standards above where vertical windows are supported on four sides, and not in a hurricane region and have no specific requirements should there be glass breakage. In contrast, a floor panel, which has a direct correlation between collapse and human injury, requires a highly robust solution.

Responsible design considers the benefit to society by balancing the need for safe glass that cannot fall down against the cost of providing that assurance. Many applications have excellent track records without additional retention, while others are beyond reproach as requiring robustness. However, for systems such as glass fin walls, the need for robustness can depend on size, use and occupancy.

When designing glass, it is essential to consider the circumstances of its use, the level of occupancy and the proximity of the occupants to the glass. Glass design can be classified into four different risk categories, similar to the building risk categories, but glass risk is governed only by the occupancy in the immediate proximity of the glass.

Glass Risk Category 1 is for systems that have low occupancy or no robustness requirements;
Glass Risk Category 2 is for typical applications not in categories 1, 3 or 4: requirements vary with the application;
Glass Risk Category 3 is for applications in which high occupancy creates risk, or there is a design objective to avoid economic impact or inconvenience if a unit is partially damaged: units are generally retained;
Glass Risk Category 4 is glazing designated as essential to the facility’s performance: units have residual capacity.

The requirements within each category also consider the orientation of the glass and any live load requirements either on or under the glass.

While there are many ways of preventing glass from falling, lamination is probably the most common (IBC also recognizes catch-nets, and Australian Standard AS-1288 recognizes safety films).

It is generally recognized that glass does not fail due to design load events, with glass meeting design strength. Rather, non-design events, such as rigid body impact, installation damage, and rare but inherent weaknesses, are responsible for the failures. Suppose we accept that damage/failure to a ply within a laminated unit does not occur due to design loads. It also follows, in that case, that we do not need to design the remaining system for the full design load capacity. This idea of conditional probability is already embedded in ASCE-7 where companion loads are reduced in magnitude when combined with another primary load. ASCE-7 does have extraordinary load combinations for rare events such as fire-, blast-, and vehicular-impact. However, glass has all of these and additional inherent flaws, including but not limited to nickel sulfide, which may reduce capacity to one ply.

How Safe is Fully Tempered Safety Glass?

Traditionally, we have thought of fully tempered glass as a safety glass that breaks into small, harmless pieces. Fully tempered glass is created by a quenching process where tension in the core of the glass creates compression on the outer skins. When a fully tempered piece of glass fractures, the crack wavefront propagates through the tensile part of the glass but not necessarily through the zone that was previously in compression. Consequently, as the unit fails, there can be a skin on each side of the fully tempered glass that remains unfractured, holding the failed glass together in clumps, which only break into small pieces after hitting something solid. If that something solid could be a person, the height of fall should be taken into consideration.

Unlike most materials in which adding thickness not only reduces the probability of failure to design loads, it also reduces the risk because the failure mechanism is controlled in a ductile manner. For glass, the opposite can be true. When glass fails, the crack propagates entirely through the ply, causing complete failure no matter how thick the ply is or how low the probability of failure is under design loads. In this respect, glass is a unique
construction material. Adding thickness may reduce the probability of failure to design loads but increases the risk due to non-design loads in monolithic glass. This is because the fracture runs full depth; the unit still fails in its entirety, and the resulting collapse is just bigger, heavier glass. It is more important to control the failure mechanism against non-design loads than to design for excessively low failure probabilities when considering design loads. Increasing glass thickness without robustness does not reduce the overall risk.

Making Smart Design Choices

Just like glass floors, structural glass design must consider both the undamaged and post-damaged states. Monolithic glass has catastrophic failure mechanisms. Laminated glass has significantly reduced capacity once one ply is damaged. Considering both the reduced capacity and the reduced loading resulting from conditional probability, responsible design can be achieved while minimizing the cost impact of redundancy and robustness.

Glass can be strong and robust, and we can deal with it being brittle: technology exists for a beautiful, exuberant design. Considering conditional probability allows the creation of efficient designs that follow good design practices, ensure safety, and minimize cost impact.

A new era of transparency in architecture is already here. If we establish sound design principles, owners, architects and engineers can launch into it with confidence.

The science and the art have outpaced current standards and codes. There is a full spectrum of designs between “non-structural” windows and “structural” floors. Until a guide is published addressing robustness, architects and engineers need to choose and specify how the glass should perform, and specify design parameters for structural glass assemblies on a project-by-project basis.

Richard Green, P.E., CPEng, is the founding principal of Green Facades LLC, a façade consultancy and design engineering practice headquartered in Seattle. Green has participated in the writing of glass standards in the United States, Australia, and Europe, and represents the United States as an expert for the strength of glass at International Standards Organization.

To view the laid-in version of this article in our digital edition, CLICK HERE.