|Ph.D Student||Pinkert Shmuel|
|Subject||Solution of Steady State Plastic Flow Problems in|
|Department||Department of Civil and Environmental Engineering||Supervisor||Professor Assaf Klar|
There are several geotechnical problems for which the formulation of large deformations is vital for their solution. Among these problems are in-situ penetration tests, used for determining the soil strength and other mechanical problems such as subsea pipelines buckling.
In this thesis, a new numerical approach is developed to solve such problems efficiently. The method is based on numerical solution of the plastic flow problem (i.e. rigid plastic material), where the spatial distribution of strength is determined by converting time changes into spatial distributions using the governing equation of steady state flow. Rather than building streamlines for back integration of disturbances, the method treats the domain as continuous with the associated field equations. For this purpose, the method employs an upstream weighting technique for determination of information flow within the domain. The execution order of the calculation is based on topological ordering, which leads to complexity of O(N), compared with O(N1.5) for the streamline methods, which significantly reduces the calculation effort.
Using the suggested method, the resistance factors for in-situ T-bar and Ball penetrometers are obtained under various soil conditions. These include the rate effect of the soil, strain softening and anisotropy, all of which affect the shear strength of the soil. In the thesis, expressions and procedures which relate between the overall penetration resistance and the fundamental soil properties are developed, specifically,  a relation for extracting the fundamental soil viscous parameter from penetration under different velocities,  a general procedure for obtaining the softening effect of a soil element from the global degradation observed by cycling penetration procedure. The analytical approach is compared to previously published field test results from five different sites around the world. Investigation of the different terms in the analytical solution reveals why normalization of the field tests results by the undrained shear strength of the field vane has the lowest scatter.
Some deficiencies in the commonly used strain softening model are noted, and a new degradation model is suggested, by investigation of the relation between degradation at the element scale and at the global resistance scale. The newly suggested model appears to fit well the field observation. General expressions for the resistance factors of the T-bar and Ball penetrometers are finally suggested.
In addition to penetration problem, which constitutes the focal point of this thesis, the developed method is applied for the solution of subsea soil-pipe interaction during thermal buckling .