French practice for the design of embedded walls intensively relies on active and passive limit equilibrium analyses, in the tradition of the work by Coulomb, Caquot, Kérisel and Absi, as well as on the more recent method involving spring coefficients, the use of which has become almost systematic, in the continuity of the approaches initiated by Terzaghi and Ménard. This latter method has ever since been improved and refined by intensive geotechnical monitoring, and is of course substantiated, as needed, by the finite element method. The parameters needed to check the serviceability limit states are most often derived from pressuremeter tests, interpreted in accordance with the principles initially established by Ménard, but here again refined by the feedback of monitoring. The ultimate limit states are traditionally checked using global safety factors (approach 2 or now RFA as defined in EC7), the partial factors on the soil resistance parameters (approach 3 or now MFA) being used only for certain situations or types of structures. This tradition, built up on pragmatism, that excludes the unconditional use of a systematic approach and a single method, is based on the belief, every day strengthened by observation and measurement of the behavior of structures, that there is no unique model that would apply to all situations, and that, on the contrary, identification of the limits of existing models is essential. In this lecture, we propose to present the different aspects of this national practice by explaining not only the backgrounds, but also the limits, as well as the way they are taken into account, or at least they should be when this is not the case.

Soil nailing is a technique developed in France during the 1970s for the retainment of excavations. The nails are steel bars introduced in the soil and preventing the soil mass to fail. Such structures are often designed through classical slope stability analysis software, i.e. based on Limit Equilibrium. The tension in the nails is generally considered equal to the maximum tensile forces admissible in the reinforcement. Yet, the service loads in the reinforcements are generally smaller than the calculated ones, especially at the bottom of the excavation because of the construction phasing. This is critical for the design of the facing and Limit Equilibrium based software need to be adapted by considering the construction phasing to modulate the mobilization of reinforcements. Based on the study of soil nailed walls through real-scale experiments, centrifuge and numerical modelling, an improvement of the limit equilibrium classical design is proposed. The software used was PROSPER, developed by Laboratoire Central des Ponts et Chaussées in the 1990s. The particularity of PROSPER is to derive the reactions of the nail by imposing a displacement of the failing soil mass. This displacement is generally considered as homogenous along the failure surface. However, considering a high displacement for the top nails (fully mobilized) and a small one for the bottom nails (partially mobilized) provide a relevant distribution of soil nail tensile forces. A distribution is proposed for such displacement and this design approach has been tested on an experimental wall, providing an efficient and time-saving design of soil-nailed walls.

The basic finite element formulations of upper and lower bound limit analysis are first reviewed with an emphasis on arriving at a unified formulation applicable to arbitrary yield criteria. A number of extensions are then reviewed: elastoplasticity (perfect and hardening), dynamics, and consolidation. In all cases, it is emphasized how these follow as direct extensions from limit analysis and may be treated using the same solution algorithms.

In his Essai, Charles-Augustin Coulomb showed that the earth pressure on a retaining wall can be calculated maximising the force obtained by considering equilibrium and strength compatibility. This same approach is still of practical relevance in some design situations and, more importantly, its basic assumptions have paved the road for the numerous successive developments that lead, among other factors, to consideration of the changes in earth pressure induced by seismic forces. A modern approach to the seismic design of structures entails the evaluation of their performance under different shaking intensities, including seismic actions strong enough to activate the system resistance. This paper intends to show that the basic ingredients of the Coulomb’s analysis, that is, equilibrium and strength compatibility, can be used in a rational way not only to predict limit values of the earth pressure, but also to evaluate the performance of retaining structures subjected to moderate and severe seismic actions.

The seismic cone penetration test (SCPT) was developed in the early 1980s and significant developments have been made in its use and application. This paper provides a review of these developments with particular emphasis on its application to identify soil microstructure. The Ménard pre-bored pressuremeter test (PMT) is one of the most popular in situ tests in France and links between SCPT results and the PMT are suggested.

The paper puts in perspective the influence of Charles-Augustin de Coulomb on the research and practice of modern geotechnical engineering. In his “Essai” (1773), Coulomb presented clear propositions on friction and cohesion on a sliding plane, giving birth to today’s soil mechanics. Coulomb’s theory started with the suggestion that the strength on a slip plane must involve a combination of friction and cohesion. Coulomb used calculus to obtain active earth pressure and show how friction and cohesion together affect lateral earth pressure. Coulomb was a gifted experimental scientist and wrote, when discussing friction, cohesion and interlocking that “only experiment can help us to decide the reality of the different causes” [for resistance]. With his fundamental equation [±t = c + µσ = c + tanϕ σ], Coulomb brought his theory to design. Design guidelines and standards are written in terms of cohesion and friction, and Coulomb’s fundamental slip plane model still dominates today’s soil mechanics. Coulomb’s insight led to further developments by Terzaghi, Hvorslev, Skempton, and the ensuing work at Imperial College and Cambridge University in the UK, at Harvard and at MIT (Lambe, Whitman, and Ladd) in the USA. Coulomb’s work, originally for masonry, has enabled the geotechnical profession to map soil behaviour, and continues to inspire new generations of researchers to look into improved and novel models and interpretations of the shear strength of soil and rock.

The time-dependent response of rock salt has been mainly investigated using confined creep tests covering a differential stress range between 5 and 20 MPa. In recent years, efforts have been made to investigate the range 0.1–4.5 MPa, using dead-end drifts in underground mines to take advantage of the very stable ambient conditions (temperature, relative humidity). Up to now, the combination of experimental data in the two ranges is difficult because of the use of different salt facies, sample preparation methods, test temperatures and experimental conditions (e.g., confined vs. unconfined tests, scale of measurements). In this work, we conduct two long-term creep tests on two natural salt samples from the same origin and prepared using the same protocol. The thermo-mechanical loading path is the same for the two tests, with only a small difference in the lateral load. One test is performed in a remote drift in a salt mine and the other test is performed in a climatic chamber in the laboratory. The comparison of the results is consistent, allowing to investigate a large range of deviatoric stresses by combining results in the two facilities. The final goal of this approach is to reduce stress extrapolation by investigating the whole deviatoric range that is relevant for underground operations. Next steps include investigating in more detail the effect of intergranular fluids, testing different temperatures and performing confined tests in the mine.

Observations of the geological deformation of the Earth’s upper crust show both brittle behavior (faults) and viscous behavior (folds, shear zones). This paper explains the crucial role of pressure solution creep and sealing processes in these contrasting behaviors and in their evolutions over time. A description of natural deformation by pressure solution shows that the pressure solution creep process can accommodate large ductile deformation without any faults. This process can also accommodate near-stable ductile deformation through the coupling of pressure solution and fracturing. Even if pressure solution creep cannot accommodate the tectonic loading rate and earthquakes consequently occur, the post-seismic evolution is largely controlled by pressure solution processes such as post-seismic creep and fault healing and sealing. Some key experiments are presented that allow evaluating the thermodynamics and kinetics of these processes. Various models are then presented that could help engineers integrate pressure solution creep and sealing processes into predictions of the long-term behavior of rock deformation in underground storage and geo-energy facilities.

A newly built versatile device for mode I fracture toughness measurement is presented. To test this apparatus, measurements have been conducted on two crystalline rocks, the Aue Granite and the Äspö Diorite, and two chalks from the Paris basin, Obourg and Ciply chalks. The fracture toughness KIC can be measured with two different testing procedures, the Semi-Circular Bend (SCB) and the Straight Edge Cracked Round Bar Bend (SECRBB) methods, both known for the easiness of the notched sample preparation. For the SCB tests, ultrasonic sensors were mounted at the sample surface to monitor changes in P-wave velocity and record acoustic emission activity. Our results are in reasonably good agreement with published data on the same rocks. The SECRBB test provides values of the fracture toughness 37% higher compared to the SCB test for the Obourg chalk. This discrepancy may be explained by a sample size effect. The fracture toughness of water saturated chalks is strongly reduced compared to that of the dry chalks by almost 50%. This shows that fracture toughness is a valuable parameter to assess the importance of water weakening in porous rocks.