Modelling tremie concrete placement in deep foundations.

Lead Research Organisation: University of Cambridge
Department Name: Engineering


Recent construction problems have been observed in the UK and abroad caused by tremie concrete placement in deep foundations. This project involves a collaboration with Arup on the development of numerical models of concrete placement for deep foundations using the material point method (MPM). The aim is to improve understanding of concrete flow during placement and around obstacles such as the reinforcing cage. This will facilitate the prevention of construction issues in deep foundations which will be of great value to the construction industry. Modelling other problematic phenomena such as static segregation and bleed will also be investigated. The research work will focus on developing the MPM in close collaboration with the Arup numerical team in London. Comparisons will also be made with other models such as LS DYNA and fluid dynamic models. It will also be very important to undertake laboratory tests and field trials to calibrate the models. The experimental component of the research project will be carried out in collaboration with Laing O'Rourke's testing laboratories and foundation contractor Expanded Piling for the field trial. This CASE project will form part of a larger European project, which also includes collaboration with other universities such as Delft, Munich and LP et C and industry partners European Federation of Foundation Contractors (EFFC) and the Deep Foundations Institute (DFI)


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/P510440/1 01/10/2016 30/09/2022
1769998 Studentship EP/P510440/1 01/10/2016 31/01/2021 Christopher Wilkes
Description Results of MPM simulations of the Slump-flow test using dynamic rheological
properties of real concretes found a good level of agreement between simulated and physical
results. The static results, however, were not in agreement. With the introduction of
thixotropy after a rest period, the simulated concrete became stiffer, resulting in reduced
Slump-flow measurements. This was in contrast to the experimental evidence but in line
with the expected behaviour. It was concluded that some process in the physical test was
obscuring the true nature of the concrete behaviour, most likely related to a reduction in
friction coefficient between the concrete and the steel base-plate of the Slump-flow test.

As with the Slump-flow test, physical L-box results for concrete resting for 240 s in
showed minimal changes, yet the flowability of the virtual concrete was severely
impacted. The difference in behaviour was explored further, with the same conclusion arising
from the L-box results as was discussed in regard to the Slump-flow results: A reduction
in boundary friction is likely to be occurring. It is hypothesised that this could be a result
of one of two main possibilities: Bleeding of concrete during the rest period was creating a
lubricating effect between the concrete and steel base, or a change in the availability of fines
(or an increase in plastic viscosity) due to thixotropy was reducing the friction coefficient
between steel and concrete.

CFD simulations of the Slump-flow test also performed well when using dynamic properties
of experimental and virtual concretes. Both MPM and CFD simulations showed the same
change in concrete flow behaviour when reinforcement inserts were added to the simulations.
This is consistent with what was observed in the physical case. In general, therefore, it seems
that both CFD and MPM are representing the bulk flow behaviour of Tremie Concrete.

CFD will likely struggle to replicate the ability of MPM to represent time-dependent concrete
behaviour, however, it is computationally more efficient and there are a multitude of software
packages available to suit many different use-case scenarios. Conversely, it is clear that
volumetric locking presents an issue to stiffer concretes that must be accounted for in MPM,
limiting its ability to simulate a range of concretes. Though it is worth noting that all
continuous methods may struggle to simulate high yield stress concretes.

The objective of the thesis was to use numerical modelling to define what conditions
encourage the occurrence of defects in cast-in-place foundations. This has been achieved by
numerically modelling concrete testing devices and the Gravity Flow Box.

A new classification system of flow behaviours that relies on a scale of flow restriction was
presented, where restriction ranges from high to low. High levels of restriction will lead to a
large flow-front differential that encourages the enveloping of interface material, arrow-head
secondary flows which may breach the flow front more regularly, and reduced flow rates
that may prevent the adequate flow of concrete. Conversely, low levels of restriction may
encourage the formation of a rising plug of old concrete atop the pile, necessitating the use
of superplasicisers to maintain good workability. Contributors to flow restriction identified in
the thesis are the use of reinforcement cages, high levels of concrete thixotropy and the use
of a support fluid. Thus, greater clarity on the conditions during the flow of concrete within
a deep foundation that can lead to defective foundations has been achieved.
Additionally, prior to this study, it was difficult to explain why empirical concrete testing was
unable to detect thixotropic changes in concrete. However, now there is a clear investigatory
path lain out for future researchers to follow, starting with an analysis of the frictional
boundary conditions.
Exploitation Route Continued development of the model, along with full-scale (100m+) models will push the boundaries of the knowledge. Also, detailed isolated models would help refine some of the understanding.
Sectors Construction

Description Performance of polymer support fluids for piling and diaphragm walls
Amount £27,800 (GBP)
Funding ID 2109009 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 09/2018 
End 09/2022
Description Academic partner for: Guide to Tremie Concrete for Deep Foundations 
Organisation European Federation of Foundation Contractors
Country United Kingdom 
Sector Learned Society 
PI Contribution Attended multiple meetings in Europe to partake in academic discussions on the best practice for producing deep foundations.
Collaborator Contribution Multiple options for numerical models were discussed during the meetings and the development of the guide. The numerical method we use was discussed in detail but unfortunately was not ready in time for current version publication.
Impact Publication of wide-reaching best practice guide.
Start Year 2016
Title CB-GEO HPC Code 
Description The CB-GEO High-performance computing code is an application that can run numerical models using the material point method (MPM). The code is parallelised for both serial and distributed applications. 
Type Of Technology Webtool/Application 
Year Produced 2018 
Open Source License? Yes  
Impact One of the first (to the best of my knowledge) open source fully parallelised MPM codes using a modular design scheme for rapid alteration and modification by the user. 
Description Lunch-Time talk's within Arup Geotechnical team 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Gave multiple hour-long talks within ARUP to update on research progress, CFD project, and opinions on concrete technology.
Year(s) Of Engagement Activity 2019,2020