Danny Smyl

Three-Dimensional Electrical Impedance Tomography to Monitor Unsaturated Moisture Ingress in Cement-Based Materials

The development of tools to monitor unsaturated moisture flow in cement-based material is of great importance, as most degradation processes in cement-based materials take place in the presence of moisture. In this paper, the feasibility of electrical impedance tomography (EIT) to monitor three-dimensional (3D) moisture flow in mortar containing fine aggregates is investigated. In the experiments, EIT measurements are taken during moisture ingress in mortar, using electrodes attached on the outer surface of specimens. For EIT, the so-called difference imaging scheme is adopted to reconstruct the change of the 3D electrical conductivity distribution within a specimen caused by the ingress of water into mortar. To study the ability of EIT to detect differences in the rate of ingress, the experiment is performed using plain water and with water containing a viscosity-modifying agent yielding a slower flow rate. To corroborate EIT, X-ray computed tomography (CT) and simulations of unsaturated moisture flow are carried out. While X-ray CT shows contrast with respect to background only in highly saturated regions, EIT shows the conductivity change also in the regions of low degree of saturation. The results of EIT compare well with simulations of unsaturated moisture flow. Moreover, the EIT reconstructions show a clear difference between the cases of water without and with the viscosity-modifying agent and demonstrate the ability of EIT to distinguish between different flow rates.

Optimizing elastic properties of structural elements considering multiple loading conditions and displacement criteria

Significant research effort has been devoted to topology optimization (TO) of two- and three-dimensional structural elements subject to various design and loading criteria. While the field of TO has been tremendously successful over the years, literature focusing on the optimization of spatially varying elastic material properties in structures subject to multiple loading states is scarce. In this article, we contribute to the state of the art in material optimization by proposing a numerical regime for optimizing the distribution of the elastic modulus in structural elements subject to multiple loading conditions and design displacement criteria. Such displacement criteria (target displacement fields prescribed by the designer) may result from factors related to structural codes, occupant comfort, proximity of adjacent structures, etc. In this work, we utilize an inverse problem based framework for optimizing the elastic modulus distribution considering N target displacements and imposed forces. This approach is formulated in a straight-forward manner such that it may be applied in a broad suite of design problems with unique geometries, loading conditions, and displacement criteria. To test the approach, a suite of optimization problems are solved to demonstrate solutions considering N = 2 for different geometries and boundary conditions.

Relating unsaturated electrical and hydraulic conductivity of cement-based materials

ABSTRACT Unsaturated hydraulic (K) and electrical (σb) conductivity are often considered durability indicators of cement-based materials. However, K is difficult to measure experimentally. This is due to the large pressure requirements at low degrees of saturation resulting from the fine pore-size distribution of cement-based materials. As a result, the commonly-used analytical models, requiring calibration of K from experimental data, are often inaccurate at low degrees of saturation. On the other hand, measuring σb is rather straight forward. Descriptions of the relationship between σb and K may therefore be particularly valuable when K is required. In this work, we use experimental data from previous works to determine the feasibility of models employing a van Genuchten-Mualem based framework to predict K and σb – expressions for diffusivity D are also provided. We then develop analytical expressions relating K and σb using these models. It is then shown that K = K(σb) and σb = σb(K) may be determined when either parameter is fully described. Lastly, we propose a simplified model and discuss the roles of pore-size distribution, saturation, pore connectivity and tortuosity in characterizing the relationship between K and σb.