Localization, both in the form of compaction bands and of shear bands, poses special challenges to finite element simulations, as the results tend to vary with the used discretization. As a rule, so called non simple continua in the sense of Noll are used to resolve the problem. In this way an internal length is introduced, which controls the size of the domain in which strain localizes. Such an approach requires an element length that is significantly smaller than the internal length, making it thus unsuitable for practical applications in the large scale. An alternative was suggested by Pietruszczak and Mróz, as far as shear bands are concerned. In this approach the shear band thickness is assumed to be a material parameter, with an element size that is significantly larger. In the present work this approach is extended to compaction bands and the discretizations required for the two approaches are compared.
Within the CRC461 project in 2004 field investigations concerning the near surface layers in Bucharest were performed. These tests include Seismic Cone Penetration Tests (SCPTU) and seismic refraction measurements for shallow depths. SCPTU is used to obtain a detailed distribution of the shear wave velocities and in situ state parameters of soils. The results are used for site response analysis with linear and nonlinear wave-propagation models. A 1-D wave propagation model is used to study the influence of the non-linear effects in the site response analysis. This numerical approach is based on the hypoplastic constitutive laws which can describe the non-linear plastic behaviour of soil during dynamic loading. It is also able to simulate the liquefaction of saturated granular materials, as well as the diffusion of pore water excess during earthquakes. A comparison of both models subjected to input signals of different magnitudes is presented
Visco-hypoplastic constitutive models reproduce properly the non-linear soil dynamic behavior, carrying effective stress analysis out. That is, viscohypoplastic models can simulate pore pressure build up, stiffness degradation, permanent deformation, non-linear viscous effects (time and rate dependence) and some non-linear effects observed in the wave propagation in undrained soft soils. In this work dynamic equations are derived and analyzed using this model. The results of numerical simulations of shear wave propagation in Bogotà soft soils as well as the influence of initial and boundary conditions are presented.
Some analytical methods have been proposed for the computation of the stability of plane open trenches supported by bentonite slurry. Nevertheless, there is no validated analytical method for the case of non-plane trenches which are required as stiff watertight corners in excavation pits. Simplified approaches replace the geometry of the corner (non-plane) trench by an equivalent plane trench. In this work it is shown, that some of these approaches are not adequate, as they oversimplify the geometry of the sliding body in the limit state and may overestimate the factor of safety against failure. Using a 3D finite element model it was possible to identify the geometry of the sliding body by simulating the stepwise excavation process of a corner trench in normal consolidated clay and its failure due to the reduction of bentonite support. The influence of the geometry of the trench, drainage conditions, material properties and anisotropy of the soil has been considered. In contrast to the simplified sliding body with the shape of a wedge on which analytical methods are based, the sliding body observed in the simulations has the shape of a quarter of a vertical cylinder, intersected by an inclined plane at its bottom.
Paraelasticity (PE) is a hysteretic, fully reversible model with rate-independent damping and with variable secant stiffness. PE has been proposed in the companion paper (Niemunis et al. in Acta Geotech, 2011) for the 3D case. Here, paraelasticity is extended for demanding application in geomechanics by implementation of a reversible dilatancy–contractancy. Moreover, an implicit FE implementation for Abaqus and calculation of a boundary value problem are presented.
Apart from time-driven creep or relaxation, most viscoplastic models (without plastic and viscous strain separation) generate no or a very limited accumulation of strain or stress due to cyclic loading. Such pseudo-relaxation (or pseudo-creep) is either absent or dwindles too fast with increasing OCR. For example, the accumulation of the pore water pressure and eventual liquefaction due to cyclic loading cannot be adequately reproduced. The proposed combination of a viscous model and a hypoplastic model can circumvent this problem. The novel visco-hypoplasticity model presented in the paper is based on an anisotropic preconsolidation surface. It can distinguish between the undrained strength upon triaxial vertical loading and horizontal loading. The strain-induced anisotropy is described using a second-order structure tensor. The implicit time integration with the consistent Jacobian matrix is presented. For the tensorial manipulation including numerous Fréchet derivatives, a special package has been developed within the algebra program MATHEMATICA (registered trade mark of Wolfram Research Inc.). The results can be conveniently coded using a special FORTRAN 90 module for tensorial operations. Simulations of element tests from biaxial apparatus and FE calculations are also shown.
A computer aided calibration, benchmarking and testing of constitutive models is presented. A large collection of test data for Karlsruhe sand is used together with the neohypoplastic model as an example. The essential feature of the procedure is an automatic evaluation of the discrepancies between a test and its simulation. They can be minimized by modifications of material constants or used as a benchmark for different models. Apart from the curve-fitting part one may check whether a constitutive model violates the Second Law. For this purpose several specially designed stress loops are tried out. A shake down of all state variables is established. The neohypoplasticity and the barotropic hypoelasticity are used as examples in the thermodynamic tests. In conclusion several modification to neohypoplastic models are presented.
Numerical modelling of the cyclic deformation of soil can be performed with the use of two types of constitutive models: incremental plasticity models and so-called explicit cyclic models. The present study deals with the application of a model of the second type—the high-cycle accumulation model for sand—to the analysis of vibration-induced stress changes in saturated soil. The application of the model requires the concurrent solution of two coupled boundary value problems: a dynamic problem for the determination of strain amplitude and a quasi-static problem for the calculation of stress evolution. The coupling of the two problems has been implemented in a two-dimensional axisymmetric formulation with the finite-element program Abaqus. The model is applied to the analysis of stress changes near a concrete wall caused by a vibrating pile. The numerical results show that a large-amplitude vibration, e.g. during the installation of vibratory driven piles, may substantially reduce both the effective and the total stresses in front of the wall. The effective stress may be reduced to zero resulting in soil liquefaction.
Die Standsicherheit von nicht-ebenen Schlitzwand-Ecklamellen in weichem, normalkonsolidiertem bindigen Boden und die Setzungen von benachbarten Fundamenten infolge des Herstellungsvorgangs einer Schlitzwand wurden mit Hilfe der FE-Methode numerisch untersucht. Um das K0−anisotrope Verhalten des Bodens besser simulieren zu können, wurde eine Erweiterung der Visko-Hypoplastizität entwickelt und in einer FE-kompatiblen Subroutine implizit implementiert. Die neue anisotrope Visko-Hypoplastizität ist in der Lage, die Ratenabhängigkeit, das Kriechen, die Relaxation, die Anisotropie, und die Akkumulation des Porenwasserdruckes infolge undrainierter Deformationszyklen wiederzugeben. Bezüglich der Standsicherheit wurde der unter Suspension gestützte Aushub des Eckschlitzes und das Verhalten des Schlitzes bis zum Bruchzustand simuliert. Der dadurch entstandene Bruchkörper hat die Form eines vertikalen Viertelzylinders, der von seiner Unterseite her von einer geneigten Ebene begrenzt wird. Im Vergleich mit den in Deutschland üblichen verwendeten analytischen Verfahren zur Berechnung der Standsicherheit, die normalerweise von einem Erdkeilmodell ausgehen, ist der Mechanismus für den Eckschlitz mit dem Viertelzylinder kritischer. Dieses Ergebnis spricht für die Entwicklung eines analytischen Verfahrens, das auf diesem Mechanismus basiert. Die Simulationen der Gebrauchstauglichkeit von benachbarten Fundamenten infolge der Herstellung von Eckschlitzwänden zeigen, dass die Setzungen stark von der Abwicklungslänge des Eckschlitzes abhängen und wenig vom Verhältnis der Schenkellängen L1 /L2 . Zusätzlich nehmen die Setzungen mit der Gebrauchslast des Fundamentes zu. Für die gleiche Gebrauchslast nimmt die Setzung mit dem Abstand des Fundamentes zum Schlitz stark ab. Sowohl die Standsicherheit des offenen Schlitzes als auch die Setzungen des benachbarten Fundamentes hängen stark von der Zeit ab. Ein schnellerer Aushub des Schlitzes und eine schnellere Betonierung der Wand führen zu größeren Standsicherheiten und kleineren Setzungen des Fundamentes, als eine langsamere Vorgehensweise.
Einige Böden neigen aufgrund ihrer tonmineralogischen Zusammensetzung bei Kontakt mit Wasser zur Volumenänderung, welche sich phänomenologisch entweder als Expansion oder durch eine Quelldruckentwicklung äußert. In der Baupraxis müssen sowohl die Quellphänomene als auch das hydromechanisch gekoppelte Materialverhalten allgemein z.B. bei der Bemessung von Tunnelbauwerken oder Abdichtungen berücksichtigt werden. Existierende empirische Modelle sind für komplexe Randbedingungen oft unzureichend, doch die Anwendung höherwertiger mechanischer Stoffmodelle stellt aufgrund komplexer Modellformulierungen und meist aufwendiger Bestimmung der benötigten Materialparameter eine große Hürde dar. In diesem Aufsatz wird ein von den Autor:innen kürzlich vorgestelltes neuartiges hypoplastisches Stoffmodell für expansive Böden durch die Anwendung auf weitere experimentelle Studien validiert. Zur Erleichterung der Anwendung des Stoffmodells in der Praxis wird eine vereinfachte Parameterkalibrierung erläutert, die keine aufwendige experimentelle Untersuchung erfordert.