CROSS Safety Alert
Effects of scale
This Alert has been prompted by concerns raised about some particularly large structures and is based on evidence made known to SCOSS. However, the disclosures also prompted wider thinking about safety issues related to ‘scale’, so historical events are also discussed.
Anecdotal evidence suggests that many designers are unfamiliar with scale effects thus increasing the risk that safety and commercial aspects may be overlooked. Validation of our current design rules has largely been based on years of satisfactory building performance.
The recommendation from SCOSS is that a far better understanding of large structure performance is needed. Large structures being defined as those whose dimensions are such that effects that are not well understood or where secondary effects pose challenges.
Note: this Safety Alert was published by SCOSS (Standing Committee on Structural Safety). Since March 2021, SCOSS is integrated under the name Collaborative Reporting for Safer Structures (CROSS).
Who should read this?
Principal Designers, civil and structural engineers, Principal Contractors, steelwork fabricators, building control engineers, HSE construction inspectors.
Submit a report
Your report will make a difference. It will help to create positive change and improve safety.
Our secure and confidential safety reporting system gives professionals the opportunity to share their experiences to help others.
I applaud the 'Effects of Scale' article and would say that it reflects concerns that I have felt for many years and which, largely, have fallen on deaf ears.
I would though like to mention my belief that there is another issue, not mentioned, that should come under this heading. Designers, often more accustomed to smaller scale structures, are routinely designing structures with at least one and often multiple levels of transfer structures. This is often driven by architecture that arises from mixed use developments and, commonly, car parking modules conflicting with residential or commercial modules. The result is often something that would fall a long way short of any exhortations about robustness, collapse resistance and direct load paths. If you have a failure in a transfer structure, the spread of collapse could be catastrophic, so the issue is serious.
Designers sometimes attempt to address the issue by applying prescriptive and highly simplistic tying rules that pay lip service to the problem, at best. The potential for disaster is compounded still more by designers who manipulate, or stretch to breaking point, the disproportional collapse categorisation and strive to obtain a client-favourable interpretation. The guidance is clear but is not being implemented, I fear.
Therefore, much clearer and perhaps a more explicit and robust statement on disproportionate collapse should be made and the issues that flow from that spelled out in no uncertain terms. It is my view that the rules in the AD's (number of stories, basements, mansards) are misused and we would be better off saying that the rules should apply to all structures. Structures of modest height, say 4-5 storeys, can still be quite large and extensive and, especially if non-robust structural forms exist in them, may present a potential for great loss.
Thank you once again for an excellent and appropriate SCOSS Alert.
I feel it should be highlighted, however, that the deflection effects on the long spans you illustrate are in fact even more sensitive to length than you advise. Whilst it is correct that deflection is proportional to (span)^3, in the cases you highlight, the main loading (especially the construction loads) are all UDLs. As the overall load applied by a UDL is itself a function of the span, then for UDLs, deflection is in fact proportional to (span)^4.
The significance of this is often overlooked and is compounded by the presentation of many standard formulae, for example, the Steel Designers’ Manual: Simply-supported UDL deflection, say, is given as 5WL^3 / (384EI), where W = wL (w being the UDL). Although obvious to experienced engineers, and for those with the time to review such things, the significance of this can often be overlooked in time-pressed design houses, leading to the examples you describe in your Alert.
Your Alert on the Effects of Scale is excellent and extraordinarily comprehensive. Congratulations to the author(s). Other associated topics might include fatigue-induced cracking and destruction of marine structures from wave-loading; cyclical flexure of thick members or tensile loading of hangers under vehicular loading; fatigue-induced cracking of testing equipment (fully loaded for 20m cycles); all of which I have encountered. How far can RC design for shear be extrapolated? I am sure I am not alone in having designed slabs over 3.5m thick for large span tunnel roofs, largely based on rules derived from test on 0.35m thick specimens or shallower. Have we stacked up serious problems for posterity?
The issue of concrete ponding on long spans was considered in detail in my paper Floor slabs, lasers and levels in Concrete September 2011. Unfortunately, despite support from some of the main steel sheeting manufacturers, I have been unable to persuade anyone to address the issue properly in specifications. In practice it is only the good sense of some specialist contractors plus a fair degree of luck that has prevented major disasters