Soil-foundation interaction: 10 years of scientific developments

(La Lettre Terrasol, June 2016)

Over the last few years we have carried out a number of scientific studies, in particular on the topic of soil-foundation interaction. These developments, all motivated by practical issues encountered in the geotechnical studies that we work on every day, have been applied through new calculation tools incorporated step by step into the Foxta software suite. Here is a brief summary.

The development of the Tasplaq module was a major step forward in the modelling of rafts and slabs, providing a practical calculation method breaking away from the “slab on springs” representation while avoiding the use of complete 3D finite-element processing. The method has demonstrated its effectiveness in a number of major projects (high-rise buildings at La Défense, EPR in the United Kingdom, etc.) and has benefited from several recent updates, such as the processing of a non-horizontal multilayer and the integration of unloading/ reloading effects.

The lessons learned from the ASIRI research project have been used to extend the Taspie+ module, which applies a simplified approach to process the complex interactions involved in the behaviour of soil reinforced by rigid inclusions. The power of this method lies in its ability to account for the effects of soil/inclusion relative rigidity in the modelling of load transfer mechanisms.

The latest update of the module incorporates edge effects in the case of a limited number of inclusions. The general application of the “t-z” and “p-y” models as implemented in the Taspie+ and Piecoef+ modules provides a satisfactory and rigorous solution to the issue of parasitic effects, taking into account the soil-foundation interaction (negative skin friction, lateral thrusts, swelling, etc.). Buckling is another parasitic aspect specific to the design of micropiles. The mathematical formulation of Piecoef+ provides a new approach for direct estimation of the critical buckling load, an approach that also evaluates the second-order effects in the case of a non-zero initial curvature.

The new Groupie+ module for pile groups enables easy processing of the 3D equilibrium of the pile/soil/foundation system in a single model, taking account of the non-linearities related to the soil behaviour. Development is continuing, with the aim of introducing the footing flexibility and processing pile/soil/pile interaction effects.

The recent development of geothermal foundations has motivated the development of a new module, Thermopie+, for evaluation of thermal-mechanical loads in a group of geothermal piles connected at the top by a structural element of any stiffness. The proposed approach accounts for the effects of pile/structure/pile interaction, unlike conventional modelling which is limited to an isolated pile with a single elastic connection at the top.

In parallel, the recent increased awareness of the challenges related to the seismic risk has encouraged us to work on additional analysis methods in order to take full account of the dynamic character of seismic loads. In this context, we are currently focusing on the following topics:

  • Piles under seismic conditions: a rigorous treatment of the seismic response of a pile foundation necessitates taking into account the (dynamic) effects of pile/soil/ pile interaction, which are ignored implicitly in the usual ‘pseudo-static’ analyses. This was the motivation for the initiation at the beginning of 2016 of a PhD thesis together with AP Consultants, Ecole Centrale de Nantes and IFSTTAR Nantes. This project is aiming to develop a “macro-element” able to represent the non-linear response of a pile group under seismic conditions, taking frequency-related effects into account.
  • Rigid inclusions under seismic conditions: going beyond the calculation of internal loads under inertial and kinematic effects, the seismic design of a foundation on rigid inclusions requires the evaluation of the irreversible displacements related to the “fuse” effect played by the granular layer. A “generalised Newmark model”, integrating the effects of soil-structure interaction, has been developed and subjected to practical evaluation as part of an on-going R&D partnership with Hyundai Engineering & Construction.

F. Cuira

Seismic design of retaining structures

Yield design calculation (La Lettre Terrasol, June 2016)

The kinematic method of the yield design calculation as implemented in Talren enables to assess the “dynamic” active/passive pressure diagrams to be considered for the seismic design of retaining structures. This goes well beyond the scope of validity of the usual analytical models such as the Mononobe-Okabe one extended to the case of cohesive soil (Okabe, 1924).

The results of a comparison made for a wall retaining homogeneous soil subjected to an earthquake characterised by kh = ah/g = 0.3 and kv = 0 are given in the table opposite: the Mononobe-Okabe model leads to results strictly identical to the results of a yield design calculation restricted to plane kinematics (θmax = 0), which corresponds to the assumption introduced implicitly in the Coulomb’s problem from which the Mononobe-Okabe model is derived. However, this assumption leads to over-optimistic results when the cohesion term becomes predominant, with more than 20% under-estimation of the dynamic active pressure compared with a more sophisticated model analysing spiral arc kinematics.

This observation is confirmed by the example shown in the figure, dealing with the dynamic active pressure behind a wall retaining heterogeneous soil. Application of the yield design calculation with spiral arcs mechanisms leads to a resultant dynamic active pressure of around 980 kN/m, 50% higher than the value given by a calculation restricted to plane kinematics (660 kN/m).

F. Cuira

K-Réa software: current developments and new features

(La Lettre Terrasol, June 2015)

We have recently developed a new calculation engine for the K-Réa software, scheduled
for integration into the commercially-available version of K-Réa in early-autumn 2015.
Based on an all-matrix formulation, the new engine considerably extends the options
offered by the K-Réa software for the design of retaining walls.
The new key features include:

  • Automated processing of bank and berm effects using the kinematic approach with the yield design method (see picture),
  • Direct processing without iterations of double-wall projects, without limitation of the number of linking anchors,
  • Generalisation of initial conditions, including the possibility to define a loading history preceding the construction of the retaining wall,
  • Upgrading of the anchoring conditions: local or surface anchors with capping of the anchoring reaction (elastic-plastic behaviour), generalisation of “creep” and “prestress” actions to all anchor types including linking anchors,
  • Automated processing of combinations of loads applied onto the wall and the soil (necessary for harbour projects, for instance),
  • Hydraulic gradient: possibility of defining a profile point by point (z, u) on each side of the wall,
  • Ultimate limit state design: integration of approaches 1 and 3 of Eurocode 7 (in addition to approach 2 already included in the current version), stability analysis of the anchoring block (Kranz) extended to double-wall projects, automatic checking of the stability of the bottom of the excavation.

F. Cuira

Earthquake engineering and soil-structure interaction

(La Lettre Terrasol, June 2014)

Measures must be taken to minimise the material and human consequences of the threat constituted by the seismic risk. Above all, this preventive approach requires appropriate design of structures taking fully into consideration the dynamic character of seismic loads.

Until recently, earthquakes were taken into account in the design of geotechnical structures mainly through simplified pseudo-static models enabling to join up with common design practices. Today, the introduction of the Eurocodes and the recent awareness of the issues related to the seismic risk lead to favour computer models which are capable of realistic processing of the dynamic behaviour of the soil interacting with the structure that it supports.

The development of these models, which demand a fresh and enlightened view by the engineer, is nevertheless hampered by a geotechnical corpus which is not familiar with seismic aspects and which by a wide margin prefers semi-empirical design based on the pressuremeter test data.

In this context, and thanks to the multidisciplinary training of its team, TERRASOL has developed advanced practice in design of earthquake-resistant structures. This requires effective control of soil-structure interaction effects going well beyond the radical simplification consisting in assuming the structure to be perfectly embedded at its base, a simplification sometimes leading to excessively unrealistic results.

In practice, this interaction may be analysed by representing the soil response using dynamic impedance functions (expressing the soil stiffness and damping according to the natural frequency content of the load) and, strictly speaking, by considering the seismic movement affected by the presence of the structure (kinematic interaction). This type of analysis, which can be performed easily using a SASSI or similar model, is regularly used for design of new or existing earthquake-resistant industrial facilities, in particular for EDF and AREVA.

F. Cuira and B. Simon

Thermal classification of soils

(La Lettre Terrasol, December 2013)

Taking the thermal properties of soils into account in projects for geothermal heat exchangers, burial of electrical cables, or roadways and hydrocarbon pipelines subjected to freeze/thaw cycles, is a major issue. TERRASOL, in partnership with BRGM (French Geological Survey) and the University of Paris VI, has conducted a study for RTE (French Transmission System Operator) aiming to develop a practical method for determination of the thermal conductivity of soils, based on the “GTR” classification from the French Road Earthworks Guide (French standard NF P11-300).

This method takes account of the saturation state of the soil using the determination of the particle-size distribution curves of the various categories starting from the single points defined in the GTR. In a second step, the relationship between the particle-size distribution and the characteristic water content of the soils is defined by a simple empirical model. Lastly, the thermal conductivity is estimated according to various models in the literature. The predictions obtained by these different thermal conductivity determination methods have been compared successfully with several data sets covering all soil categories. This tool has been used in combination with the Thermo-Hydro-Mechanical module recently introduced in PLAXIS 2D to study heat flows around some reference structures (an example is shown here, around a high-voltage cable burial trench).

P. Reiffsteck

Recent scientific developments within Foxta software

(La Lettre Terrasol, December 2013)

2013 has been marked for TERRASOL by sustained scientific activity which has resulted in notable upgrades of the calculation engines of several of our software products, in particular Foxta:

  • The Groupie+ program now has a new mathematical formulation processing the 3D equilibrium of the combination “foundation + piles + surrounding soil” simultaneously and in a single system: the formulation directly integrates the non-linearities related to the soil behaviour, thus avoiding the complexity of the present procedure which requires back and forth transfers between the three modules Taspie+, Piecoef+ and Groupie.
  • The Taspie++ program generalizes the elementary cell model by integrating the shear likely to develop at the cell periphery: this enables processing of the cases with a finite number of inclusions, a line of inclusions, and more generally the edge effects on the periphery of the reinforced area (the method developed was the subject of an oral communication at the 2013 Paris Conference).
  • Recasting of the initial formulation of the Tasplaq program now makes it possible to take into account a non-horizontal multilayer with a dip defined by layer and by direction.

These developments, already made available to our engineers in-house, will be incorporated into the next commercial version of Foxta.

F. Cuira

Uncertainties and risk analysis in geotechnical engineering

(La Lettre Terrasol, June 2013)

Taking uncertainties into account is a constant concern in geotechnical engineering. Everyone involved in the profession seeks to reduce these uncertainties and their effects, by proposing for example:

  • Actions aimed at reducing uncertainties concerning geotechnical data and anticipating more effectively the means to react to these uncertainties. These include for example two recent evolutions relating to standards: the publishing by French Tunneling Association of recommendation GT32 on the characterisation of uncertainties and geological, hydrogeological and geotechnical risks, and the current updating of French Standard NF P 94-500 which defines the geotechnical missions and minimum geotechnical investigations to be carried out in the frame of each project. Terrasol took part in the working groups in both these cases;
  • Actions enabling these uncertainties to be taken into account more effectively in the structures design; this is the issue which we have chosen to develop here, in connection with the Eurocodes in particular.

Since 2009, Terrasol has undertaken research and development work focusing on risk analysis techniques and their applications in the design of geotechnical structures. This is in answer firstly to high demand from project owners faced with ageing infrastructures and a relatively undeveloped body of technical expertise in this area in France, and secondly to requests from construction companies for optimisation of projects in increasingly complex contexts, such as concessions for example.

These techniques are based on a probabilistic approach derived from the reliability theory and are used to define the safety level with respect to a failure mechanism given in the form of a probability of failure or collapse of the structure. Thus, in this approach, the geotechnical properties are random variables described by a probability distribution with a mean value and a standard deviation. Compared to conventional design by a deterministic or semi-probabilistic approach, the use of reliability engineering techniques enables to ensure a structure is safe when the geotechnical properties present significant scatter, or to optimise its design when the scatter is limited.

This work resulted in the development of a risk analysis tool based on the approximate method known as the “Response Surface Method” (RSM) which has proved to be effective in
its applications on geotechnical structures. This tool will be progressively implemented in Terrasol’s calculation software in the form of a “reliability engineering design” wizard. This wizard will evaluate the probability of exceeding a limit state predefined by the user (bearing capacity, sliding, settlement, thrust, etc). The probabilities thus obtained are to be compared with the target probability values defined by Eurocode 0 for ultimate or service limit states.

The method which has been developed is also intended to clarify the application of the Eurocodes when numerical methods such as the finite element method or the finite difference method are used for the structures design.

F. Cuira & B. Simon

3D FEM modelling of geotechnical structures

(La Lettre Terrasol, June 2013)

More and more often, large-scale projects require complex stress-strain modelling for a better understanding of soil-structure interaction.
For a number of years, TERRASOL has been carrying out 3D finite element modelling for structures as varied as retaining structures (the Odeon Tower in Monaco), foundations of complex buildings (the Louis Vuitton Foundation for Creation, the new High Court in Paris), underground works (the extension of Paris metro line 14), foundations of civil engineering structures (the Third Bosphorus Bridge), etc.

Let’s consider 2 recent cases:

  • The new High Court in Paris in the Batignolles area (a 160 m high building): the aim of the 3D modelling was to understand the behaviour and interaction of the deep foundations of the three building cores with the complex Paris geology. This model made it possible to better estimate the settlements and the load transfer within the piled-raft foundation system.
  • The dismantling chamber for the metro line 14 TBM : this underground excavation is located below old buildings in the 8th Paris district. The 3D model was to prove that the compact shape of the structure and the chosen construction methods helped to limit settlements.

Through these numerous cases, we have developed our know-how, and determined the limits of such modelling. We are thereby able to propose relevant 3D approaches in accordance with specific issues and design progress of each project.

A. Beaussier, JP Janin and P. Reiffsteck

Geotechnical engineering and rehabilitation of structures

(La Lettre Terrasol, December 2012)

Rehabilitation projects for constructions, buildings or civil engineering structures include a geotechnical engineering component which must simultaneously seek to meet imperative technical requirements and provide an optimum answer to Sustainable Development needs. This means aiming to make maximum re-use of existing foundations to limit demolition and reconstruction work and minimise consumption of new materials.

TERRASOL is currently committed to these goals:

  • both in a research project entitled Rufex (Re-Use of EXisting Foundations) as part of the competitive cluster Advancity, aiming to develop the use of Soil Mixing with a retractable tool (Solétanche-Bachy process) for reinforcement of railway platforms without removal of the track, and of existing foundations of buildings;
  • and in various building renovation projects in the Paris region: the former Calberson warehouses on Boulevard MacDonald (for conversion into offices and flats), the East sector of Jussieu university campus, and a former postal sorting centre in Pantin redeveloped as a Data Center.

In such contexts, the geotechnical problems are different from those encountered in new projects and lead to the development of very different study methodologies:

  • to re-use existing foundations, it is necessary to know them well and therefore, because the «as-built» drawings are not always available, to implement means of investigation and inspection of the geometry of the existing foundations, the quality of the concrete, etc;
  • in cases when the renovation leads to an increase in loads, it is necessary either to reinforce the existing foundations or to create new ones, which raises load redistribution questions involving all phases of the construction work, from demolition through to reconstruction. The modelling is often complex and must take into account this whole loading history;
  • furthermore, even when there is no significant change in the loads, the application of existing regulations could lead paradoxically to the reinforcement of foundations which have served their purpose very well for decades, on the grounds that they are no longer in line with regulations;
  • lastly, whether it is necessary to reinforce the existing foundations or to create new ones, the often restricted spaces oblige us to choose unusual and innovative construction techniques, all of this being done while minimising the demolition of existing structures.

We applied this methodology particularly to the renovation of the Calberson warehouses, the structure of which was originally designed to withstand higher loads than those of the future construction. However, the foundations were designed to the “bare minimum”, and the discovery of an anomalous layer under the base of the piles led to reinforcement of the existing foundations by Jet Grouting at the tip of the piles, thus enabling them to work at an optimum level.

The creation of new foundations on micropiles was then the second significant step in the project. At this stage, a study of the interaction between these new “flexible” foundations and the stiffer existing piles was conducted to evaluate the load transfers as accurately as possible. Lastly, in order to check the correct behaviour of the construction, which combined new and old structures, it was decided to monitor the behaviour of the building over time in accordance with the principles of the observational method.

Renovation projects thus involve a desire to incorporate an element of the past in the buildings of the future. And such projects oblige us to review and rethink our engineering know-how.

A. Guilloux

Foxta v3 software - Buckling calculations

(La Lettre Terrasol, June 2012)

A “direct” buckling calculation has been integrated into the Piecoef+ module of the Foxta v3 software. It has already proven itself efficient in the design of a large number of projects including foundations on flexible inclusions and vertical or inclined micropiles, for which the buckling needed to be taken into consideration.

The method used is based on the matrix formulation employed in Piecoef+ and overcomes the limits of standard analytical approaches by providing a direct estimation of the critical buckling load for complex configurations. It also makes it possible to evaluate second order effects in the case of a defect in shape or of a lateral load concomitant to an axial load.

F. Cuira

Terrasol contribution to ASIRI research project

(La Lettre Terrasol, December 2011)

2011 saw the successful completion of the national ASIRI Project devoted to soil reinforcement using rigid inclusions. It was launched in 2005 after various exploratory studies, which had been conducted as of 1999 with the support of the IREX and the RGCU. Terrasol was involved in this collaborative research project for over 11 years, with Bruno Simon in charge of its coordination and technical management from the very outset.

Strong commitment on the part of all 41 partners allowed the project to keep within the initial budget (2.4 million euros) while meeting the designated objectives: better understanding of behaviour mechanisms and drafting Recommendations for the design, execution and monitoring of these works. The 360-page document published at the end of 2011 by Presses des Ponts is the fruit of this work. Events held in Lyon, Paris, Toulouse and Nantes to present these Recommendations during the last quarter highlighted the interest this project has generated in the construction world by attracting over 400 people overall.

Terrasol’s investment in this project has been considerable, its efforts equating to more than the annual workload of a full-time engineer since 2005. In addition to the project’s scientific management and the coordination of experimental and numerical actions, numerous studies were performed in-house to assess the data from test sites, take part in benchmarking exercises, as well as develop and assess new calculation methods suited to this technique. Part of this work served to update the Foxta Taspie+ and Piecoef+ calculation modules. Several articles were also written to present simplified but effective methods, such as that used to calculate footings over rigid inclusions.

This project provides an exemplary illustration of what national projects can bring to civil engineering research as a whole, and geotechnical engineering in particular. The field of foundations is one very much open to innovation insofar as it is able to benefit from experimentation and players with complementary skills.

B. Simon