Download files
Abstract
Soft soils are often associated with low shear strength and high water content and many of them show large time dependent deformations when structures like road embankments are constructed over them. The generation of excess pwp and settlement continues for a long time after the end of construction. The excessive ground deformations in such soils, causes severe damage to pavements and or related structures and has been a great concern to the geotechnical engineers in designing and maintaining these structures. Despite that, due to rapid growth of infrastructure and transpiration development and environmental considerations, the necessity of constructing road embankments and other structures on such soils is common. To facilitate design, construction and maintenance, predicting the multiple behavirour characteristics of these structures accurately is a necessity and is often a challenge. This is partly due to the complexities in the constitutive models and partly due to some other modeling aspects in the analysis. To address this challenge, a new, relatively simple, elastic viscoplastic (EVP) model has been developed in this thesis. This model is then tested against well instrumented field cases. Some other important aspects of modeling the behaviour of PVD improved soft soils and structures constructed over them have also been dealt with. To describe the behaviour of soft soils, a number of EVP models have been proposed in the literature. Many of them use a constant creep coefficient to incorporate time dependency in their mathematical formulations. However, according to many researchers, the plot of void ratio or vertical strain against log (time) may not be a straight line and thus, the creep coefficient may not be a constant. The use of a constant coefficient in the calculation may yield misleading results. Based on long duration (more than 3 months) creep tests, a new EVP model has been proposed in this thesis which allows the non linear variation of the creep coefficient. The new model also incorporates a new yield surface function developed based on effective stress path of undrained triaxial test results. The performance of the model is then tested against long duration field monitoring data from a well instrumented case history (i.e., Leneghans embankment) both in one and two dimensional analyses. The predictions obtained from the newly developed model have been compared with the predictions from another EVP model and an elastoplastic model. The results obtained using the new model are significantly better than those obtained from this analyses with elastoplastic and the other EVP model. Another important issue investigated in this thesis is the effect of using matched permeability on the predicted performances in a plane strain analysis. Hird s (1992) matching procedure is often used to estimate the equivalent permeability for foundation soils improved with PVDs in a plane strain analysis. Good predictions in terms of settlement can be achieved using this technique. However, the predicted excess pwp can be significantly different than an equivalent axisymmetric analysis. A procedure is developed in this thesis, to correct the predicted excess pwp at the plane strain unitcell boundary. The results obtained with the correction procedure are much closer to the field measured values than the originally predicted ones. Later in the thesis, the effect of post construction soil movement on embedded structures such as abutment piles has been investigated. Bridge abutments are often constructed first for joining roadways and they are commonly supported on piles. These structures and consequently their pile foundations are often subjected to forces exerted by the lateral movements of soils on one side of them. So far, there has been limited research on such embedded structures subjected to lateral loading in a field scale. In this thesis, the behaviour of two abutment piles subjected to the lateral movements of the foundation soil due to the construction of Leneghans embankment have been modeled using a combination of two and three dimensional analyses. These two piles were installed on the northern berm section of Leneghans embankment about one and a half years after the construction of the embankment. They were instrumented with inclinometers at the center and were monitored for about four months. Excellent predictions for the lateral bending behaviour of the piles were achieved.