CROSS Safety Report
Unconservative design of flat slab due to software modelling issues
This report is over 2 years old
A reporter discusses how a design/modelling problem caused an under-designed reinforced concrete (RC) slab to be constructed.
Key Learning Outcomes
For civil and structural design engineers:
If there is uncertainty with design outputs from a design software it is good practice to carry out hand calculation checks to verify the outputs
Ensure that those using specialist software programs are suitably trained and competent to do so
It is good practice to have in-house checking of designs carried out by a competent and experienced engineer. Particularly for critical elements such as transfer slabs.
Find out more about the Full Report
Our secure and confidential safety reporting system gives professionals the opportunity to share their experiences to help others. If you would like to know more, please visit the reporting to CROSS-UK page.
A reporter's organisation recently came across a design/modelling problem which gave highly unconservative analysis results. This caused an under-designed reinforced concrete (RC) slab to be constructed within a large domestic property.
Correct modelling of blockwork walls
A loadbearing blockwork wall, supported on a transfer slab, was mistakenly modelled as a concrete shell element within a 3D finite element (FE) package. A more realistic approach is to model such walls as a series of pin ended columns.
A loadbearing blockwork wall, supported on a transfer slab, was mistakenly modelled as a concrete shell element within a 3D finite element (FE) package. A more realistic approach is to model such walls as a series of pin ended columns
When the transfer slab was exported to a 2D FE package for reinforcement and deflection checks, the 3D concrete wall element was converted to a line element of equivalent stiffness and incorporated within the 2D FE analysis.
The result of this was that the transfer slab was artificially stiffened by the line element, which was effectively acting as a very stiff beam with a depth equivalent to the height of the wall over. As such, both the long-term deflection prediction and the reinforcement demand was significantly underestimated.
Insufficient reinforcement in transfer slab
The already constructed slab was found to have around 50% of the necessary ultimate limit state design reinforcement and was about to receive a 75mm screed. Once the modelling error was discovered following observed excessive cracking to the supported masonry wall, temporary propping was installed.
Strengthening works required on site
A permanent strengthening solution was developed by way of a heavy steel transfer beam installed below the wall. Ceilings had to be removed and services diverted to achieve this.
To avoid such an error, when creating or checking a 3D FE model, it needs to be ensured that any loadbearing masonry wall that is transferred onto a slab below, or that is not vertically continuous down to foundation, is modelled as a series of individual pin ended columns. This ensures that they act in the vertical loadbearing direction only, and thus cannot act as a deep beam.
Wall shell elements within a 3D FE model should only be used where a vertically continuous RC concrete wall is proposed, as otherwise they can artificially stiffen the structure by acting as deep beams. The design checker should also ensure that they see an extruded and annotated view of the 2D model, in order to verify that the structure has been modelled correctly.
Expert Panel Comments
Find out more about the Expert Panels
Expert Panels comment on the reports we receive. They use their experience to help you understand what can be learned from the reports. If you would like to know more, please visit the CROSS-UK Expert Panels page.
There has been much disquiet expressed in engineering circles about the improper use of (or over reliance on) computer modelling with potential for results to be divorced from reality. This report is a classic illustration of the kind of problems that might arise. Safety demands that all model outputs are subjected to a simplified sanity check which appears not to have happened.
Safety demands that all model outputs are subjected to a simplified sanity check which appears not to have happened
Beyond that, the description of this model suggests an inappropriate level of refinement for the essentially simple task of designing an RC slab supporting a wall. If, however the slab in question is complex with, for example, significant openings, then accurate modelling is all the more important.
There were a number of opportunities to discover this mistake. For example, as the wall was in the model, a very quick review of the stresses in the wall would have highlighted that they were inappropriate for a masonry wall. This highlights the need to check the whole model during the design not just the element of immediate interest.
Modelling precast concrete planks
Similar errors can occur when concrete slabs are constructed from precast planks but modelled as a solid diaphragm leading to an underestimate in the loading to the supporting beams; a check of bending in the slab perpendicular to the span would have highlighted this. It is disturbing that such a slab can be detailed and constructed with only 50% of the required rebar without anyone in the office or on site thinking it looked odd.
Additional reports on design modelling failures
Questions regarding the sufficiency of computer modelling, the adequacy of peer review, and the role of the structural engineer in the field have been raised recently in the FIU bridge collapse. This will be the topic of an upcoming CROSS-US report. In the UK, the Standing Committee on Structural Safety (SCOSS) and CROSS have had a long-standing policy of endorsing third party checks for key structures. The rationale is to assure public safety. In 2016, SCOSS published a paper Reflective thinking which looked at over-reliance on computer modelling and posed a set of questions for the designer:
Is the model capable of satisfying the requirements? (the validation question)
Is the model the most appropriate in the context?
Has the software been validated and verified?
Has the model been correctly implemented? (the verification question)
This is an early indication of the value of CROSS sharing experiences between countries by linking the Hartford collapse, the FIU bridge collapse, this report, and other CROSS reports on computer modelling failures. There is an overriding need in the construction industry to have sufficient checking by suitably qualified and experienced persons to uncover such serious errors.
Share your knowledge
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.
No feedback has yet been published for this page.