• cyclic triaxial test
• resilient modul
• soil-cement

Engineering properties of soils exhibit, in nature, a wide variety. Often local soils are not adequate to meet the prescribed requirements for a construction project. Soils can be improved by adding cement or lime and an amount of water that allows obtaining the maximum degree of compaction of the soil-cement mixtures. This solution allows a significant increase of both stiffness and strength. The use of this technique is especially effective in road or high-speed railroad platforms with high traffic loads and a significant number of passages of vehicles.
The current work concerns mainly the study of the response of soil-cement, to be used in road and railroad platform, to cyclic load; the purpose is to contribute to the development of more rational design of these structures, with stringent criteria project. The results of cyclic triaxial tests performed in the Geotechnical Laboratory of FEUP (Faculdade de Engenharia da Universidade do Porto ) over soil-cement samples moulded with silty sand from granite residual soil form the background material for this work.
Several samples with different percentage of cement and void ratio were moulded, saturated and consolidated at different consolidation pressures. Then they were cycled in distinct load amplitudes, confining pressures and drainage conditions. The tests were carried out following the methodology of the standard EN 13286-7:2004 for cyclic tests on unbound granular materials in order to analyze the behaviour of these mixtures in terms of the resilient stiffness and plastic deformation.
The moulding conditions concerning the percentage of cement, curing water content and void ratio, were selected by means of an index parameter that is the ratio of porosity (η) to the volumetric cement content (Civ), expressed as a percentage of the cement volume regarding the total volume. This index, introduced by Consoli et al. (2007) as an attempt to normalize the results of uniaxial compression tests, is found to be a good index to classify the cement mixture behaviour even under cyclic loading.
Two different triaxial tests procedures (procedure A and procedure B) were carried out. The procedures essentially differ for the number of applied loading cycles. The development of the resulting permanent deformation, which accumulates with the repeated loading, is compared with the types of responses predicted by the shakedown concept and the classification given by the standard EN 13286-7:2004 (developed for unbounded materials), with the aim to establish the relevance of this standard for cemented materials. It seems that the standard is not very appropriate for these cemented mixtures.
Analyzing the results obtained applying the procedure with the lower number of loading cycles (procedure A), it can be seen that:
- deviatoric stress affects the resilient modulus, which decreases as deviatoric stress increases; - the soil-cement degrades as the number of cycles increases, in fact the resilient modulus decreases;
- drainage conditions affect the resilient behaviour of the material, being the resilient modulus always lower in drained tests;
- small permanent deformations occur in soil-cement samples, insomuch as, based on the difference (εp(5000)- εp(3000)) between the permanent deformation at 5000 and at 3000 load cycles, the material always falls in the range A – Plastic shakedown (stable deformation behaviour). The classification is the one suggested by the standard EN 13286-7:2004.

The main result obtained from tests carried out following procedure B is that applying more load cycles plastic deformations occurred even at high frequencies such as 1Hz and low loads. The permanent deformation increased with the number of cycles reaching values that fall in range B (plastic creep). The analysis of the hysteresis loops corroborates this result.

In addition, the thesis reports the results of ultrasonic wave measurements and uniaxial compression tests, showing the influence of the curing time on the strength and the stiffness.