“The more I learn, the more I realise that I don’t know” is a quote attributed to Einstein. After a career designing foundations, especially piles, I realise more and more how little we really understand.
For instance, we still do not know how to properly design piles in London Clay, probably the most tested and researched material in the world. We started with Skempton’s 1959 paper and still argue about what the appropriate alpha is – a debate not settled even by the recently published and much disparaged LDSA document.
What is cohesion as critical stare theory tells us it does not exist? I remember a Peter Wroth lecture where as an aside he “designed” a pile showing that the friction would be about half of a function and that function represented shear strength (I wish I could reproduce that proof!). This “half” is, of course, what the industry knows as the alpha value, but would like to think can be significantly increased.
We cling to this approach. However, when you determine “shear strengths” by either banging a pointed bar in the ground or testing a highly strained sample squeezed out of a tube hammered into the ground we only have correlation, not understanding. Whatever happened to effective stress design?
While we argue about the details of design in London Clay it is as nothing compared with our ignorance in Chalk. The British Geotechnical Association’s Chalk 2018 conference in September showed progress in the understanding of the material and initiatives from Imperial College have hugely improved the design of offshore driven steel piles. As John Burland stated, a similar exercise is needed for onshore bored piles.
There has been much talk of establishing a database of test results to allow research. However, the combination of contractors’ commercial secrecy and their reluctance to carry out preliminary load tests means there is little or no data being collected.
The current design methods are based on a small sample since of those tested very few result in failure. For example, at the Reading Oracle (Illingworth and Chantler 1999) there were 13 preliminary tests on piles designed and installed by the same methodology, yet only two moved sufficiently to analyse and one of those was found to have construction issues; both piles were in areas of deeper putty chalk whereas those in more sound material did not move significantly. Of the few tests reported at the recent Chalk conference, it is notable that only Osterberg cell testing gave useful information in the structured chalks and the results considerably exceeded conventional design expectation.
In structureless chalk we are determining design methods from an unrepresentative sample, probably the lowest decile, and it is possible that movement only occurred in some of these because the pile was incorrectly constructed. Are the prescribed factors appropriate when you are designing so near the bottom of the probability curve? We then apply the same methods to structured chalk – surely this is a rock and should be treated as such?
End bearing capacities are shown to exceed our current approaches considerably but how do we use them? Applying higher factors because we are worried about settlement demonstrates our lack of faith in consistent construction and reluctance to design to serviceability.
There is evidence that the time taken to construct a bored pile is a significant factor in the development of shaft friction. Take too long, even with drilling fluids, and the friction is reduced. However, what is the improvement of performance with age? We impose rapid, brutal load increases on a test pile yet it takes probably a year to load a contract pile.
The piling industry needs to rise to this challenge, testing shorter piles outside their comfort zone to failure in a single cycle and share the results. Structural engineers need to ensure a geotechnical specialist advises on the ground investigation and manage their clients better. Then we may have data to interpret.
- John Chantler is technical manager for JRL Civil Engineering