Limit Analysis in 2012
If anyone needs evidence that the Offshore Oil & Gas business is seasonal, then perhaps the timing of this post on Limit Analysis is a good starting point. Here we are again in late spring, almost one year after my last bout of postitis and I find myself with enough time to revisit this blog.
Much of interest has occurred in the last year. From this seat one of most interesting occurrences is the gradual adoption of limit analysis software as a general analysis tool for geotechnical engineering problems.
In 2009 / 10 we found ourselves dealing with some rather complex problems that required a fast resolution: too fast for us to spend time playing with FEA and sufficiently complex to invalidate standard Limit Equilibrium checks. At this point we fell back on the upper bound plasticity approach where we guessed at kinematically admissible failure mechanisms, applied a weighty safety factor and walked away, slightly dissatisfied and seeking to do better “next time”.
With Advanced Geomechanics pulling AGSPANC off the market at about the same time, we were also in need of a replacement tool that could assist in the rapid sizing and load point optimisation of suction anchors, prior to detailed design.
It just so happened that I was in the South of England in early summer of 2010 with Andrew Ripley, catching up with various people with an intention of kicking off a technology development campaign. It was by good fortune that while visiting Byron Byrne and Chris Martin at the University of Oxford, Chris decided to do his party trick of demonstrating his Limit Analysis code to us. It was impressive, very impressive.
Andrew and I walked out of the meeting and I can recall being totally consumed by the possibilities that open up when having such a technique on hand. A week later, Andrew had found LimitState:Geo as the only commercially available code and we had obtained a test copy with which to do damage.
This was the start of our journey and apparently we weren’t alone.
Seeing is believing?
If you check out the LimitState:Geo website and linger on the “news” page, you’ll see a list of early adopters starting with RailCorp (in march 2010), most being focused on onshore problems. Following about 6 months of validation and testing, we obtained a license in December 2010. This happily coincided with a number of papers being presented at the ISFOG II conference in Perth in November 2010, giving us a nice warm feeling that we weren’t alone in the wilderness on this subject.
Since then, a series of other companies focused on offshore oil and gas projects have licensed the software (Boskalis, Fugro, IKM, Hyundai, GL Noble Denton and so forth) and I assume they are looking to leverage the benefits of this technique.
At Acergy, and subsequently Subsea 7 (following the merger of January, 2011) we found ourselves working rapidly to validate this new tool. We invested a significant amount of energy benchmarking against existing designs and analysis techniques, as well as published case studies, the NGI validation of Plaxis 3D for suction anchor design being just one example.
As part of this validation exercise, we ran a series of typical problems past Colin Smith and Co at LimitState, just to check we weren’t breaking their software and the limits of the DLO method. For most of the types of problems we encounter, it seems that the key assumptions relating to Limit Analysis (Tresca failure criterion, perfect plasticity and associated flow) are not a significant constraint.
It was satisfying to discover this, and in the process we learnt a good deal about the limitations of the DLO implementation and how to work our way around these, where it is feasible to do so. Naturally we remain cautious and check the validity of results and conclusions against various methods, but at this stage we are happy to use Limit Analysis as a day-to-day tool.
A Quick Sense Check
The key here is that we are practicing engineers, rather than researchers, and we believe we have a handle on how to use this tool safely and demonstrate that we are doing so, while gaining the benefits of the method. In general, I would say that it is more reliable than Limit Equilibrium hand checks for the complex soil layering and loading that we tend to consider. I’ve posted on this previously if you care to check it out.
A Touch of Hubris?
As practicing engineers, we found that limit analysis provides two major advantages: speed and insight. It is the latter reason that prompted my Danish colleague Christian Olsen to pen the immortal lines:
“To be quite honest, I can’t see how we did geotechnical engineering before we had the software. It has really allowed us to take the guesswork out of engineering non-standard designs.”
This opened the door to some rather entertaining conversations, both internally within Subsea 7 and externally. I bumped into Dave White from UWA at the SAFEBUCK JIP session in Wallingford in Autumn of 2011 and took a gentle ribbing and on this subject (at least I think it was gentle).
To be fair to Christian though, if this statement is viewed through the lens of our line of work it rings true. If you view it from the perspective of a Danish geotechnical engineer, shaped by the teachings of Brinch Hansen, it holds true – this software literally takes the guesswork out of plasticity analysis and is very powerful (in the right hands).
I’m going to take the time to write a few posts looking how Limit Analysis seems to be evolving as a tool, both for research and in engineering practice. I’ll mention a bit about how we are using it and how our workflow has been greatly improved. Hopefully you’ll find this interesting and keep checking back over the next month or two. It would be good to get some debate going, so feel free to comment below or via email if you have some experience or observations you wish to share.