Testing and constitutive modelling of saturated interfaces in dynamic soil-structure interaction.

Cyclic direct and simple-shear experiments are conducted on remolded clay-steel interfaces under undrained conditions and a constant normal stress. Also, a constitutive model for the stress-strain-pore-pressure behavior of cohesive soil interfaces under dynamic loading is developed using the new unified disturbed state concept (DSC). The model is based on elasto-plasticity theory as defined by the hierarchical single surface (HISS) plasticity model. The proposed model is calibrated with laboratory tests and is shown to capture the complex strain-softening, degradation, and pore pressure behavior observed in cyclic loading of piles in saturated clay. Understanding of the mechanical behavior of saturated interfaces between structural and geologic materials and joints in rock, subjected to cyclic loading is important for safe and improved analysis and design of many geotechnical structures. Appropriate testing is vital for the determination of parameters in constitutive models to characterize the mechanical response in terms of stress-strain and failure behavior. A unique laboratory testing device for investigating the dynamic loading effects at the interfaces and joints of materials is described. This new device known as the cyclic multi-degree-of-freedom device with pore fluid pressure effects (CYMDOF-P), can automatically load and test various combinations of material interfaces with dry or saturated conditions and in a direct-shear or simple-shear mode. Based on field observations of instrumented piles it is proposed that there is a thin interface zone of clay between the moving pile and clay mass in which significant shear deformations and generation of pore pressure occur. To explore this behavior, a test program using Gulf of Mexico marine clay is carried out with the CYMDOF-P device. Important behavioral aspects are identified and incorporated as part of the new disturbed-state interface model. Laboratory test results are used for the determination of parameters for the model, and for verification of the model. The model predictions, in general, were found to provide satisfactory correlation with the observations. The procedure to find material parameters for the interface model is described. The model is simple enough to be easily implemented in numerical techniques such as the finite element method, and this implementation is briefly discussed