### Refine

#### Document Type

- Conference Proceeding (8)
- Article (4)

#### Institute

#### Keywords

- Angewandte Mathematik (9)
- Computerunterstütztes Verfahren (7)
- Angewandte Informatik (5)
- Strukturmechanik (4)
- Architektur <Informatik> (3)
- Building Information Modeling (3)
- Data, information and knowledge modeling in civil engineering; Function theoretic methods and PDE in engineering sciences; Mathematical methods for (robotics and) computer vision; Numerical modeling in engineering; Optimization in engineering applications (3)
- Stochastik (3)
- CAD (2)
- Bentonit (1)

In this paper we evaluate 2D models for soil-water characteristic curve (SWCC), that incorporate the hysteretic nature of the relationship between volumetric water content θ and suction ψ. The models are based on nonlinear least squares estimation of the experimental data for sand. To estimate the dependent variable θ the proposed models include two independent variables, suction and sensors reading position (depth d in the column test). The variable d represents not only the position where suction and water content are measured but also the initial suction distribution before each of the hydraulic loading test phases. Due to this the proposed 2D regression models acquire the advantage that they: (a) can be applied for prediction of θ for any position along the column and (b) give the functional form for the scanning curves.

In this paper we evaluate 2D models for soil-water characteristic curve (SWCC), that incorporate the hysteretic nature of the relationship between volumetric water content Θ and suction Ψ. The models are based on nonlinear least squares estimation of the experimental data for sand. To estimate the dependent variable Θ the proposed models include two independent variables, suction and sensors reading position (depth d in the column test). The variable d represents not only the position where suction and water content are measured but also the initial suction distribution before each of the hydraulic loading test phases. Due to this the proposed 2D regression models acquire the advantage that they: (a) can be applied for prediction of Θ for any position along the column and (b) give the functional form for the scanning curves.

Physicochemical forces are responsible for the swelling pressure development in saturated bentonites. In this paper, the swelling pressures of several compacted bentonite specimens for a range of dry density of 1.10–1.73 Mg/m3 were measured experimentally. The clay used was a divalent-rich Ca-Mg-bentonite with 12% exchangeable Na+ ions. The theoretical swelling pressure–dry density relationship for the bentonite was determined from the Gouy-Chapman diffuse double-layer theory. A comparison of experimental and theoretical results showed that the experimental swelling pressures are either smaller or greater than their theoretical counterparts within different dry density ranges. It is shown that for dry density of the clay less than about 1.55 Mg/m3, a possible dissociation of ions from the surface of the clay platelets contributed to the diffuse double-layer repulsion. At higher dry densities, the adsorptive forces due to surface and ion hydration dominated the swelling pressures of the clay. A comparison of the modified diffuse double-layer theory equations proposed in the literature to determine the swelling pressures of compacted bentonites and the experimental results for the clay in this study showed that the agreement between the calculated and experimental swelling pressure results is very good for dry densities less than 1.55 Mg/m3, whereas at higher dry densities the use of the equations was found to be limited.

This study contributes to the identification of coupled THM constitutive model parameters via back analysis against information-rich experiments. A sampling based back analysis approach is proposed comprising both the model parameter identification and the assessment of the reliability of identified model parameters. The results obtained in the context of buffer elements indicate that sensitive parameter estimates generally obey the normal distribution. According to the sensitivity of the parameters and the probability distribution of the samples we can provide confidence intervals for the estimated parameters and thus allow a qualitative estimation on the identified parameters which are in future work used as inputs for prognosis computations of buffer elements. These elements play e.g. an important role in the design of nuclear waste repositories.

Sand-bentonite mixtures are well recognized as buffer and sealing material in nuclear waste repository constructions. The behaviour of compacted sand-bentonite mixture needs to be well understood in order to guarantee the safety and the efficiency of the barrier construction. This paper presents numerical simulations of swelling test and coupled thermo-hydro-mechanical (THM) test on compacted sand-bentonite mixture in order to reveal the influence of the temperature and hydraulic gradients on the distribution of temperature, mechanical stress and water content in such materials. Sensitivity analysis is carried out to identify the parameters which influence the most the response of the numerical model. Results of back analysis of the model parameters are reported and critically assessed.

Due to the complex interactions between the ground, the driving machine, the lining tube and the built environment, the accurate assignment of in-situ system parameters for numerical simulation in mechanized tunneling is always subject to tremendous difficulties. However, the more accurate these parameters are, the more applicable the responses gained from computations will be. In particular, if the entire length of the tunnel lining is examined, then, the appropriate selection of various kinds of ground parameters is accountable for the success of a tunnel project and, more importantly, will prevent potential casualties. In this context, methods of system identification for the adaptation of numerical simulation of ground models are presented. Hereby, both deterministic and probabilistic approaches are considered for typical scenarios representing notable variations or changes in the ground model.

In this paper, we present an empirical approach for objective and quantitative benchmarking of optimization algorithms with respect to characteristics induced by the forward calculation. Due to the professional background of the authors, this benchmarking strategy is illustrated on a selection of search methods in regard to expected characteristics of geotechnical parameter back calculation problems. Starting from brief introduction into the approach employed, a strategy for optimization algorithm benchmarking is introduced. The benchmarking utilizes statistical tests carried out on well-known test functions superposed with perturbations, both chosen to mimic objective function topologies found for geotechnical objective function topologies. Here, the moved axis parallel hyper-ellipsoid test function and the generalized Ackley test function in conjunction with an adjustable quantity of objective function topology roughness and fraction of failing forward calculations is analyzed. In total, results for 5 optimization algorithms are presented, compared and discussed.

Performing parameter identification prior to numerical simulation is an essential task in geotechnical engineering. However, it has to be kept in mind that the accuracy of the obtained parameter is closely related to the chosen experimental setup, such as the number of sensors as well as their location. A well considered position of sensors can increase the quality of the measurement and to reduce the number of monitoring points. This Paper illustrates this concept by means of a loading device that is used to identify the stiffness and permeability of soft clays. With an initial setup of the measurement devices the pore water pressure and the vertical displacements are recorded and used to identify the afore mentioned parameters. Starting from these identified parameters, the optimal measurement setup is investigated with a method based on global sensitivity analysis. This method shows an optimal sensor location assuming three sensors for each measured quantity, and the results are discussed.