Last year in 2024 we where lucky enough to win and carry out a £120,000 Innovate UK grant, collaborating with Astute in Swansea. To help Richter Technologies not just provide the fastest design packages, but numerically superior ones. This is a summary run through of our journey. if you would like to know more detail about any particular part of the process, perhaps leave a comment below and we can create another blog post.
Our proposal was that Richter Technologies had been successful creating several modules to generate common Temporary Works (TW) design packages on the fly for our clients. While we had managed to capture and improve on key metrics which our clients where thrilled with, like increasing turn around time dramatically and reducing overall cost. We wanted to go a stage further and utilise our digital workflow to also give the most accurate results.
Our most popular design which is often required short notice at the beginning of jobs or with flexibility to last minute changes of plant etc, is working platforms or piling mats. These common items can be a significant cost to site and minimising material cost is also a high priority to the people who comes to us. Our original package used the Temporary Works Forum method alone. However we were aware that there is a huge number of different industry recognised methods available to calculating the depth of these compacted granular mats and that there is a variation between methods to which is optimal in different ground and loading conditions. This leads to there being no clear cut superior option for all scenarios. We identified 13 different methods utilised across the UK and Europe and 34 peer reviewed academic papers for bearing pressures spanning over the last century. These were brought together by Richter's own Valdemar Mejia, who had taken interest in a similar topic in their Masters project. Valdemar prepared an in depth 64 page report taking a look into each method, its context in industry, how the methods may of evolved into one another. Commenting on their strengths and weaknesses. Also detailed in the report was which numerical values we typically derive as practical engineers who work on these things daily. This gave the researchers at Astute some great guidance to avoid including parameters which may not be readily available to us.
After this initial review of the current state of the art, Dr Nia Fry at Astute began digesting this information through collaboration we were able to derive through parametric studies that BRE470, EC7, TWf and Tensar's Triax methods would provide both a good range of results in different conditions and give us good reinforced options. While covering that extreme cases could yield extreme platform depths which were unnecessary. This enabled us to build all of these methodologies into our generation process giving the user the option of which methodology they would like to calculate with. The conclusions enabled us to provide a better service in the short term.
However, while we could compare like to like within these methods, going the extra step would be to measure them against reality. While we know these methods are all industry accepted, and have been followed countless times and yielded no collapses. it is difficult to say which curve of results actually follows realistic conditions.
Astute being connected to Swansea has some of the best computational analysis tools and minds at their disposal and they began to take the research a step further. They developed the above analysis run, utilising a Mohr-Coulomb plasticity model method. Which yielded realistic plastic deformation. We collaborated on reducing the range of possible variables of the problem as best we could, giving realistic ranges of values we commonly use in calculations for these problems, Phi, Cu, Youngs Modulus, Thickness of platform etc. Simplifying the sample size of the issue. However to analyse every possible outcome generates a huge requirement on resources, even with our reduced variable ranges. therefore a subspace of 500 combinations of parameters are randomly generated. An analysis of each is completed on their super computer and after post processing of the results, we were able to begin to map out the distribution and compare it to the other key methods. It would be possible with enough computing power and further optimisation to the analysis, to analyse every possible permutation of the space and provide one large data set of results. meaning that every possible answer of every possible combination could eventually be reduced to a data base look up table, given the appropriate inputs. With the deadline looming however, we had to draw a close to the research and conclude our work. While an analysis of the entire space was not viable at this time, it will not be long before computational time costs reduce further to a stage where this approach will be very possible.
The whole process was incredibly interesting, working with some great minds. Our service has upgraded given the guidance on which methods would be best to add first. We now complete 5 different numerical methods on clients inputs simultaneously and enable our engineers to deliver rapidly and efficiently. without sacrificing risk mitigation. we calculated in the average size of platform with the average difference we find between the methods in height, we will be able to save our clients more than the cost of the grant in material savings within the first year of this new development being deployed.
We are looking to continue the work and collaboration with Astute and have taken steps to sponsoring a PhD student down in Swansea. Its exciting to push the edge of what's possible and look forward to exploring more.
Comments