Quick Search

734 Glass smoke screens and structural safety

Report ID: 734

Published: Newsletter No 51 - July 2018

Report Overview

A reporter was asked to investigate the spontaneous failure of glass smoke control screens at a major retail store.

Report Content

A reporter was asked to investigate the spontaneous failure of glass smoke control screens at a major retail store. The screens were made of fire rated toughened glass suspended from the roof structure by aluminium cleats - 3 bolts per glass pane.

It became evident, based on examination of the remaining glass panes, that there were no gaskets or bushes to separate the glass panes from their connections and there was no tolerance within the bolt holes. In addition, bolts were loose and washers were missing, indicating they were working in shear rather than acting as tension clamps.

It was also discovered that a smoke extract vent was locked open adjacent to the failed panes. This permitted a differential temperature to arise between the ceiling void and occupied space adjacent to the failure. In this case the external air temperature was as low as -6oC.

The trigger for failure was thus differential expansion and contraction between the aluminium cleats and glass screen. However, the defective connections were the underlying cause. They were inherently vulnerable to a trigger mechanism - an accident waiting to happen. Had the building been occupied at the time of failure, a serious accident could have resulted.

In preparing remedial details, it has become apparent that standard glass smoke barriers are essentially designed for a specific fire rating, but there may be little consideration of structural factors. For example, they may not be designed for internal wind pressures and toughened glass is used irrespective of the height at which the barrier is installed. Toughened glass fails as a mixture of dice and heavy clumps of glass, which can cause serious injury, particularly as height increases.

Glass suppliers do not necessarily make the specifier (normally an architect) or client aware of the inherent risks that may be associated with their product - it is not even clear that they appreciate risks exist. It is the reporter's view that a risk assessment should always be carried out when designing overhead glass components.


Glass should never be directly supported by a material like steel as it does not yield and will cause concentrated stresses to form in the glass, causing it to fail. This is a well-known behaviour of glass so to hear of support details not following this principle is somewhat concerning. Normal practice would be to separate the glass from the supporting metalwork by nylon bushes, neoprene gaskets, poured resin or applied mastic with plastic spacer bars; appropriate to the situation.

As to the adoption of toughened glass at height, there is some merit in its use as its mode of failure is small pieces known as 'dice' once the tension and compression pre-stress becomes unbalanced within the glass pane causing it to shatter. The report's author is right to point out toughened glass can fail with clumps of 'dice' falling from height that can cause harm, however, this is not as dangerous as large shards of glass following the failure of annealed and heat strengthened glass.

The use of laminated glass is advisable for overhead glazing to reduce the risk of falling glass, however, the support detailing must be robust enough to restrain the glass in the first place following a failure of one of the plies within the laminated pane. For further guidance on this see the CIRIA Guidance on glazing at height (C632).

The report is not about ‘structural glass’ as such, but if there is any doubt, then it should be borne in mind that the design and installation of structural glass is a specialised business and appropriate competence and expertise is needed. The specifics of each location need to be considered including thermal and external loads and the consequences of failure understood and mitigated against. Guidance is given in the IStructE guidance on Structural use of glass in buildings (second edition).


CROSS depends on you for reports - if you have experienced a safety issue that you can share with CROSS, please Submit a CROSS Report

If you have any comments regarding this CROSS report, please Submit Feedback

View other CROSS reports published in Newsletter No. 51

Email Updates

How to Report

Online submission:
Submit by post: