Skip to main content

Service Assessment and Failure of Earth Structures

Final Report Summary - SAFES (Service Assessment and Failure of Earth Structures)

The engineering behaviour of unsaturated soils is still an emerging area, in terms of the development of the underlying concepts, the ability to model the behaviour and the application of the concepts to field situations. Practical applications in geotechnical or geo-environmental engineering are still commonly designed using the conventional assumption that the soil is saturated, even when this is not the case. In the drier regions of Europe (e.g. the Mediterranean regions) the assumption of saturation will rarely be appropriate, and even in the wetter, more northerly regions, many applications involve unsaturated soils. The assumption of saturation can result in over-design of engineering structures, and cost savings could be possible if a proper assessment of the unsaturated behaviour was made. Neglecting the unsaturated aspects of behaviour can also lead to potential problems being overlooked (e.g. shrinkage and swelling) which may be costly to remedy at a later stage. A proper understanding of unsaturated soil behaviour is also essential in prediction and mitigation of landslide hazards. Conventional assumptions can lead to erroneous prediction of landslides. In particular, the effect of rainfall on slope stability cannot be modeled unless the unsaturated behaviour is incorporated. Rainfall-induced landslides often cause significant property damage and also loss of human life. Among the most recent and catastrophic events of this type there are the landslides, which occurred between the 5th and 6th of May 1998, in the area of Sarno, in southern Italy, causing the death of 167 people. Over the last decade, significant strides have been made in measurement, testing and numerical modeling techniques for unsaturated soils. There is now a strong requirement to validate these novel techniques, by making comparisons between tests results and model predictions. A further, more important, aspect of validation will involve comparing predictions with actual observations in the field. Without such activity, the geotechnical or geo-environmental industry will remain sceptical about the new techniques. The above issues are at the forefront of current geotechnical research and the present project aim has been to address some of them while laying the basis for future advances in the area. The project has taken place during an overall period of 15 months. It has included:

- in-situ testing and soil sampling
- laboratory tests on undisturbed field-compacted samples from the embankment
- Laboratory tests on laboratory-compacted samples
- Interpretation of test results.

The filed work has been undertaken at an experimental flood defense embankment close to the city of Mantova in Northern Italy. The embankment is 4 metres high, 194 metres long with a crest width of 4 metres and maximum slope 2:3. It is made of the same soils as the main flood defences embankment of the Po river. The soil has been compacted in-situ at the optimum water content and dry density as determined by standard Proctor tests. Samples of soil have been initially collected from the embankment fill for the determination of basic soil properties such as Atterberg limits, specific gravity, organic content and grain size distribution. Subsequently, a number of undisturbed samples have been retrieved at different locations across the full depth of the embankment. These samples will be used during an extensive laboratory programme of hydro-mechanical testing. In particular, tests are planned to be performed on artificial samples compacted in the laboratory at values of water content and dry density similar to those existing in the embankment at the time of construction, as well as on undisturbed samples retrieved from the embankment.
The main laboratory programme has investigated the influence of current and previous stress states (described in terms of suction, degree of saturation and applied load) on the mechanical response of the soil at both small and large strains. Two complementary classes of suction-controlled tests have been performed. In one set of tests the external stress has been varied along loading/unloading paths while suction has been maintained constant. Conversely, in the second set of tests, the samples have been subjected to cyclic variation of suction while the external stress is kept constant. Such stress paths have been investigated in both the RCTS and TC devices but the test objectives will depend on the particular device used. The tests in the RCTS device have been mainly focused on the determination of shear stiffness and damping at small strains whereas the tests in the TC device have explored deformation at large strains and in the deviatoric plane up to critical state. Laboratory activity carried out on field and laboratory compacted samples have generated two twin experimental data sets. The interpretation of results has been devoted to the highlight potential differences between engineering properties of field-compacted and laboratory-compacted samples in terms of compressibility, initial shear stiffness, damping, water retention and pre-failure behaviour in the deviatoric plane. This analysis has allowed a better knowledge on whether laboratory-compacted samples are representative of the engineering behaviour of earth fills compacted in-situ according to standard construction practice. The laboratory activity performed during the project has generated a wide-ranging set of high-quality experimental data from both in-situ dynamic tests and suction-controlled laboratory experiments. This has been a unique attempt to draw upon a variety of testing techniques with the aim of establishing a broad material database, which has the potential of becoming an experimental benchmark for the development and validation of future constitutive models beyond the proposed project. The data cover several aspects of soil behavior including crucial features for which very limited experimental evidence currently exists, such as the nonlinearity of the stress-strain relationship at very small deformations, the characterization of damping or the dependency of strength and volumetric behaviour on degree of saturation. An additional element of originality of the project is given by the comparison between the engineering properties of soils compacted in-situ according to conventional construction practice and in the laboratory according to the Standard Proctor Method. Current geotechnical practice, in fact, assumes that material properties are identical in both cases provided that the dry density and water content of the soil in-situ are the same as those of the samples compacted in the laboratory. This often implies that engineering characterization of earthworks is undertaken by testing samples compacted in the laboratory rather than undisturbed compacted samples retrieved from the field. The variability of construction procedures in the field and the limited control on the fill emplacement cast, however, serious doubts about such practice. Despite the importance of the above assumption, very little investigation has been carried out so far in this respect and the proposed project adds original experimental evidence to help to identify potential source of errors in current design of earth structures. During recent years, unsaturated soil mechanics has become a key area of geotechnical research underpinning important applications with a potential to improve safety and contribute to tangible cost savings in civil engineering design. Europe is currently leading the world in terms of developments in unsaturated soil mechanics. It is, however, vital that European research is maintained at the fore-front internationally as it is in danger of falling behind work in Asia, particularly Japan and China which are investing heavily in this area. In this project, particular attention has been paid to the dissemination of results. This has the potential of increasing prominence and visibility of European research in unsaturated soil mechanics and consolidating the image of a strong European lead in the field, which is crucial to sustain Europe’s position as one of the major destinations for specialized training in soil mechanics. In particular, sustained research visibility is essential to reinforce Europe’s ability to attract top extra-European graduates and to establish strong research links with the best research institutions worldwide. As an added value to the project, collaborative research links have been strengthened between the University of Glasgow and the University of Naples “Federico II”. The synergy between the two Universities has been particularly beneficial given the complementarities of their core research expertise in constitutive modeling and laboratory and in-situ testing respectively. The project has facilitated the funding applications for short academic visits from both institutions with the purpose of delivering seminars and discussing joint research activities. This has been the first step towards the establishment of stronger research links between these two institutions facilitated by matching research interests, which can lead to joint publications and exchanges of doctoral students. This has the potential to support European excellence and competitiveness facilitating transfer of specialized skills and collaboration between European research units.