1 March 2015 in Géotechnique
In this paper, quantitative observations of dry granular flows subject to varying magnitude and direction of Coriolis acceleration are presented in order to assess the robustness of the Savage–Hutter shallow water approximation of granular flow for modelling landslides. Nine tests carried out at 70° and 45° slope inclinations are described for cases where no Coriolis acceleration was applied (i.e. tests performed at 1 g ), and where the Coriolis acceleration acted into or out of the slope (centrifuge tests). Given that the magnitude of the Coriolis term is velocity dependent, each point within a model landslide on the centrifuge experienced a unique time and space-varying magnitude of Coriolis acceleration, altering the flow displacement with time and final deposit characteristics. The resultant behaviour is compared to a depth-averaged flow model based on a frictional rheology. A single set of model parameters was found to be appropriate for all test conditions and directions/magnitudes of Coriolis acceleration, illustrating the robustness of the frictional depth-averaged approach to capture the mobility of dry granular flows, albeit as a ‘Class C' type of prediction. It is noted that the empirically derived basal interface friction was found to be lower than the static interface friction determined by conventional testing, suggesting that new methods are needed for the a priori determination of suitable rheological parameters for high-speed flows. In this paper, quantitative observations of dry granular flows subject to varying magnitude and direction of Coriolis acceleration are presented in order to assess the robustness of the Savage–Hutter shallow water approximation of granular flow for modelling landslides. Nine tests carried out at 70° and 45° slope inclinations are described for cases where no Coriolis acceleration was applied (i.e. tests performed at 1 g ), and where the Coriolis acceleration acted into or out of the slope (centrifuge tests). Given that the magnitude of the Coriolis term is velocity dependent, each point within a model landslide on the centrifuge experienced a unique time and space-varying magnitude of Coriolis acceleration, altering the flow displacement with time and final deposit characteristics. The resultant behaviour is compared to a depth-averaged flow model based on a frictional rheology. A single set of model parameters was found to be appropriate for all test conditions and directions/magnitudes of Coriolis acceleration, illustrating the robustness of the frictional depth-averaged approach to capture the mobility of dry granular flows, albeit as a ‘Class C' type of prediction. It is noted that the empirically derived basal interface friction was found to be lower than the static interface friction determined by conventional testing, suggesting that new methods are needed for the a priori determination of suitable rheological parameters for high-speed flows.
1 January 2020 in Soil Dynamics and Earthquake Engineering
1 May 2020 in Engineering Structures
FGFernando Gómez-Martínez
1 August 2021 in Journal of Geotechnical and Geoenvironmental Engineering
This paper presents a novel method for performing equivalent linear analysis that allows for a variation of the stiffness and equivalent viscous damping properties throughout the duration of shaking. The ability to change the dynamic properties throughout the analysis comes from the conversion of the input ground motion into the time-frequency domain using the Stockwell transform whereby different transfer functions can be applied at different times before converting back to the time domain. The equivalent linear Stockwell analysis (ELSA) method provides a fully decoupled approach to modeling the dynamic site response of liquefiable soil deposits that can account for changes in properties due to strain effects and the build-up of excess pore pressure. The simplicity of the method means that only a limited number of dynamic soil properties are required, the same as those used in an equivalent linear analysis, as well as an estimation of the build-up of excess pore pressure and the relative density of the soil. While there are drawbacks of decoupling the estimation of the build-up of excess pore pressure from the dynamic response, this approach means that the effects of liquefaction on ground shaking can be independently assessed using different models for estimating pore pressure. Furthermore, the influence of liquefaction mitigation interventions, or the presence of a building, can present significant modeling challenges for fully-coupled or loosely coupled approaches, whereas they can easily be assessed using simple decoupled tools. Changing the dynamic properties throughout time using the ELSA method provides a rational way to correct equivalent linear analyses for the known drawback of using the same properties throughout the whole duration. Validation studies are presented of the ELSA method against field downhole recordings from the Wildlife Array and fully-coupled nonlinear effective stress analyses.
5 October 2020 in Earthquake Engineering & Structural Dynamics
1 March 2023 in Journal of Geotechnical and Geoenvironmental Engineering
The damage caused by seismic shaking and liquefaction-induced permanent ground deformation has conventionally been assessed as two separate problems often by different engineers. However, the two problems are inherently linked, since ground shaking causes liquefaction, and liquefaction-induced soil softening affects ground shaking. Modelling both problems within a single numerical model is complex for both the engineer and the software, and most finite element software only have the capabilities to address one of them. To improve the estimates of the seismic performance of buildings on liquefiable soil, a new sub-structuring approach is proposed called the macro-mechanism approach. This approach allows the soil-liquefaction-foundation-structure interaction to be considered in a series of sub-models accounting for the major nonlinear mechanisms of the system at a macro level. The proposed approach was implemented in the open-source finite element software OpenSees and then applied to a case study of a building where significant liquefaction- and shaking-induced damage was observed after the 1999 Mw 7.4 Kocaeli Earthquake. The case study building was also simulated using two different commercial software programs, the finite difference software FLAC, and the finite element software PLAXIS, by two different research teams. A comparison between the results from the macro-mechanism approach compared to full numerical models shows that the macro-mechanism approach can capture the extent of the foundation deformation and provide more realistic estimates of the building damage than full approaches since the FLAC and PLAXIS models consider elastic elements for the building.