Site dependent and spatially varying response spectra

The goal of this study is to provide a stochastic method to investigate the effects of the randomness of soil properties due to their natural spatial variability on the response spectra spatial variation at sites with varying conditions. For this purpose, Monte Carlo Simulations are used to include the variability of both incident ground motion and soil parameters in the response spectra by mean of an appropriate coherency loss function and a site-dependent transfer function, respectively. The approach is built on the assumption of vertical propagation of SH type waves in soil strata with uncertain parameters. The response spectra are obtained by numerical integration of the governing equation of a single-degree-of-freedom(SDOF) system under non-stationary site-dependent and spatially varying ground motion accelerations simulated with non-uniform spectral densities and coherency loss functions. Numerical examples showed that randomness of soil properties significantly affects the amplitudes of the response spectra, indicating that as the heterogeneity induced by the randomness of the parameters of the medium increases, the spectral ordinates attenuate.     

Influence of spatial variability of soil friction angle on sheet pile walls’ structural behaviour

The uncertainty in terms of soil characterisation is studied to assess its effect on the structural behaviour of extended structures as sheet pile walls. A finite element model is used. This integrates a numerical model of the soil–structure interaction together with a stochastic model that allows characterising the soil variability. The model serves in propagating the variability and the system parameter uncertainties. Discussion is mainly focused on two points: (1) testing the sensitivity of the structural behaviour of a sheet pile wall to different geotechnical parameters and (2) assessing the influence of spatial variability of soil properties on the structural behaviour by identifying the most sensitive geotechnical parameter and the most significant correlation length values. The findings showed that in assessing the sheet pile wall’s structural behaviour, there are spatial variability parameters that cannot be considered negligible. In this study, soil friction angle is found to be an important parameter.     

Probabilistic comparative analysis between design rules of a shallow foundation safety

ABSTRACT

A fact that is generally overlooked in many geotechnical uncertainty analyses is that input data of the model may be correlated. While this correlation may influence the system response, epistemic uncertainties i.e. lack of knowledge of this correlation appears as a risk factor. This paper discusses how a negative correlation between cohesion (c’) and friction angle (Ø’) with their associated uncertainties can influence both the bearing resistance of a shallow strip foundation footing and the estimation of its safety. A probabilistic approach that considers both the negative correlation and the uncertainty is used in this work as a reference. This method is compared to Eurocode 7 variants that do not for the correlation. These variants, resistance and material factoring methods appears to be more or less conservative depending on the negative correlation degree between (c’–Ø), their associated uncertainties and soil configurations. Finally, the proposed probabilistic comparison shows that the material factoring method is more conservative than the resistance one.     

An identification procedure of soil profile characteristics from two free field acclerometer records

In this paper, an analytical, numerical and experimental approach for identifying soil profile characteristics by using system identification and free field records, is presented. First, a theoretical soil amplification function for two sites is defined and expressed in terms of the different parameters of the layers constituting the soil profiles (thickness, damping ratio, shear wave velocity and unit weight). Then, this function is smoothed with an analogous function obtained from experimental data by using the least squares minimization technique. The identification of the parameters is performed by solving, numerically, a non-linear optimisation problem. To demonstrate the numerical efficiency and the validity of this approach, two examples are treated. The first one consists in the identification of characteristics of a given uniform soil layer. The second example consists in the experimental validation of this approach with the data recorded within the Garner Valley Down Hole Array (GVDA). Finally, this approach is applied to identify, simultaneously, soil profile characteristics of sites from only a single soil acceleration record at free surface of each site. This procedure is utilised to identify soil profile characteristics of sites by using strong ground motions data recorded during the recent Boumerdes earthquake of May 21, 2003.     

Longitudinal variability of soils and structural response of sewer networks

A simple model which describes the soil–pipe interaction and accounts for the longitudinal soil variation has been developed. It is used for the analysis of the static response of a section of a buried sewer. A probabilistic analysis (Monte-Carlo method) enabling to quantify the influence of spatial variability of the geomechanical characteristics of the soil makes possible to study the parameters which can influence and drive the longitudinal response of a section of sewer. The system response is complex, soil–structure interaction depending on three different stiffnesses: soil stiffness, pipe components stiffness and joints stiffness. Various analyses have been performed to identify the parameters whose influence is the larger. A specific attention has been devoted to the fluctuation scale of the soil properties and to the stiffness of joints.     

Seismic response analysis of pipes by a probabilistic approach

Seismic response of buried pipes in longitudinal direction is studied. The effect of the variation of geotechnical properties of the surrounding soil on the stiffness, mass and damping of the soil is considered. The soil–structure interaction depends on pipe stiffness, joint stiffness, the variation of the soil stiffness and the soil mass and damping. Variations of the properties of the surrounding soil along the pipe are described by the random field theory. A numerical model is developed in order to simulate the effects of the variation of the soil on displacements, bending moments in the pipe and also to carry out a statistical analysis. The influence of different parameters regarding design and safety level of the pipe is conducted.