Dissertations

The objective of this thesis was to develop and apply thermal and thermo-mechanical analyses to evaluate the structural and energy efficiency of a Geothermal Energy Pile Foundation System in cold regions. Such systems were used to re-harvest the buildings energy loss through their below-grade enclosures for providing heating and cooling energy demand. To investigate the energy performance of the below-grade envelope of a building in cold regions, a thermal analysis was carried out for an institutional building, the Stanley-Pauley Engineering Building (SPEB) located in the campus of the University of Manitoba. Knowing the amount of the annual heat dissipation from the sub-grade enclosure of the building to the ground, soil temperature increase was calculated. To efficiently harness the leaked heat from the basement, a geothermal energy system was proposed to be integrated to the foundation of the SPEB and the energy efficiency of such a system was assessed. In addition to the energy efficiency of the proposed system, the thermo-mechanical response of the proposed thermo-active foundation to the applied thermal and mechanical loads was also evaluated. Results showed that 8% of the annual energy consumption of the SPEB in terms of space heating was leaked into the ground. This energy loss increased the temperature of the soil underneath the building. Using the geothermal energy foundation system, the lost energy was aimed to be re-harvested. Results showed that the thermoactive foundation system could supply 4-15% of the building heating demand during Nov-Apr and 7-41% of the building cooling demand during May-Oct. It should be noted that application of such foundation system necessitated larger factor of safety effective on the allowable load.

Harvesting shallow geothermal energy by means of energy piles coupled with ground source heat pump systems for heating and cooling buildings has increased in recent years. However, the structural integrity of such systems subjected to thermo-mechanical loads or heating-cooling cycles should be studied. Therefore, a comprehensive understanding of their structural and geotechnical performances is vital for successful applications. This thesis aims to investigate the responses of concrete energy piles subjected to thermal and thermo-mechanical loads using fully coupled thermo-hydro-mechanical (THM) finite element analyses. The axisymmetric models were carried out for two case studies of full-scale energy pile tests. Two hypothetical energy piles in Winnipeg were also analyzed to study their performances by considering local geological and climatic conditions. In general, it was found that the THM numerical models could capture considerably well the behavior of energy piles during cooling and heating cycles in comparison with the field data published in the literature. The thermo-mechanical loads did have significant effects on pile responses. From sensitivity analyses, it was found that there were considerable effects of the thermal expansion of concrete and soil stiffness on the thermo-mechanical pile responses. The pile head restrained conditions also affected the behavior of energy piles with stronger effects in the upper part of the pile near the pile head. From the simulations of the energy piles in Winnipeg, settlements of the pile head kept on increasing with increasing numbers of thermal cycles (ratcheting settlement phenomena). It was also found that the ultimate geotechnical pile capacities generally increased when the pile was heated but reduced when cooled.

Energy performance of basement structure of Stanley Pauley Engineering Building of the University of Manitoba is studied. A total of eighteen soil samples were collected at different locations and depths based on the geological profile of the ground. Thermal properties of the collected soil samples were obtained in the Geotechnical Engineering Lab using a KD2-Pro device by Decagon. Heat losses are predicted for the below-grade envelope by considering thermal properties of soil and building materials. For that purpose, a numerical model was created based on construction drawings using COMSOL Multiphysics software. The applied approach includes the calculation of heat loss due to heat transfer through conductive mechanism, also it considers phase change of water during freezing/thawing cycles. Different alternatives for heat preservation are suggested and compared in terms of economic effectiveness.

The objective of this thesis is to apply thermal analysis, coupled thermo-hydraulic analysis, and fully coupled thermo-hydro-mechanical analysis developed in this research to different engineering designs in cold regions. The applications vary from Geothermal Snow-Melting System Design to Structural and Thermal Integrity of Buried Infrastructure in cold regions. A feasibility study of geothermal energy pile based snow melting systemwas performed for six major cities in Canada. The coefficient of performance (COP) of heat pump is derived and the number of geothermal piles are determined for each city based on its specific local geological condition and heating demand.Also, the structural and thermal integrity of buried infrastructure were studied by performing an optimum design of rigid plastic foam insulations to protect buried utilities against frost damage and reduce the excavation cost. Furthermore, the effect of frost heave on the structural integrity of pavement structure and culverts are assessed.

The purpose of this dissertation is to develop a boundary element method (BEM) for multiphase porous media. Nowadays, the application of the BEM for solving problems of unsaturated porous media is still limited, because no fundamental solution exists in the published literature, neither in the frequency nor time domain. This fact rises from the complexity of the coupled partial differential equations governing the behaviour of such media. The developments of the BEM for the unsaturated soils carried out during this thesis are based on the thermo-hydro-mechanical (THHM) and hydro-mechanical (HHM) models presented in the first part of this dissertation. These phenomenological models are presented based on the experimental observations and with respect to the poromechanics theory within the framework of the suction-based mathematical model presented by Gatmiri (1997) and Gatmiri et al. (1998). After having presented the THHM and HHM models, for the first time, one establishes the boundary integral equations (BIE) and the associated fundamental solutions for the unsaturated porous media subjected to quasi-static loading for both isothermal (2D in the Laplace transform domain) and non-isothermal (2D and 3D in Laplace transform and time domains) cases. Also, the boundary integral equations as well as the fundamental solutions (2D and 3D in the Laplace transform domain) are obtained for the fully coupled dynamic model of unsaturated soils.In the next step, the boundary element formulations (BEM) based on the convolution quadrature method (CQM) regarding the saturated and unsaturated porous media subjected to isothermal quasi-static and dynamic loadings are implemented via the computer code HYBRID. Having integrated the BEM formulations for the wave propagation, as well as the consolidation problems in the saturated and unsaturated porous media, it seems that now the first boundary element code is obtained that can model the various problems in dry, saturated and unsaturated soils. Once the code is verified and validated, parametric studies on seismic site effects are carried out. The aim is to achieve a simple criterion directly usable by engineers, combining the topographical and geological characteristics of the soil, to predict the amplification of acceleration response spectra in sedimentary as well as hollow valleys.

Un effet de site se manifeste par la modification de la réponse d’un sol soumis à une sollicitation sismique par rapport à un cas dit de référence. Les codes de dimensionnement parasismiques actuels reposent sur des calculs issus de modèles unidimensionnels. Cette méthode permet de mesurer l’influence de la nature et de l’épaisseur de la couche sédimentaire sur la propagation verticale des ondes de volume. Cependant, ces résultats ne concordent pas avec les estimations fournies par des modélisations bidimensionnelles ou tridimensionnelles. Lorsque le contraste d’impédance entre le sédiment et le substratum ou la profondeur de la vallée alluviale augmente, la résonance verticale 1D des ondes de volume et la propagation latérale des ondes de surface tendent à intervenir simultanément (Bard et Bouchon, 1985). En bureau d’étude, on utilise les courbes de ‘Spectre de réponse’. Mais malheureusement, les règlements d’ingénierie ne comportent pas de solution complète tenant compte de l’influence des conditions topographiques et géologiques sur le mouvement sismique et sur le contenu fréquentiel. Ce travail vise à caractériser et à quantifier les effets de site dans des configurations bidimensionnelles, dans le domaine spectral. Il s’agit de prendre en compte les influences combinées de la topographie et de la géologie sur la réponse sismique de vallées alluviales. L’étude paramétrique est réalisée grâce au programme de simulation numérique HYBRID, mis au point par Gatmiri et al. (Gatmiri et Kamalian, 2002 ; Gatmiri et Dehghan, 2005). Il s’agit d’une technique de modélisation hybride combinant des éléments finis (champ proche) et des éléments frontières (champ lointain). La sollicitation sismique est une onde de Ricker de type SV à propagation verticale. Les propriétés mécaniques des sols sont fixées. Le site de référence est une station éloignée de l’épicentre, qui ne subit pas d’effet de site (effets topographiques et géologiques). Dans la littérature, nous avons trouvé que les déplacements horizontaux sont plus importants que les déplacements verticaux. Par conséquent notre travail va se concentrer sur xu. Dans le premier chapitre, nous expliquerons comment les résultats des simulations effectuées avec HYBRID doivent être transformés pour être exprimés dans le domaine spectral. La réponse unidimensionnelle, couramment utilisée dans les règlements parasismiques, sera également étudiée. Nous nous pencherons particulièrement sue l’EUROCODE 8. Comme on le verra, l’effet topographique n’est pas considéré dans les règlements parasismiques actuels. Mais les observations montrent l’existence de variations dans l’amplitude et le contenu fréquentiel de la réponse du sol le long des pentes des collines. Pour comprendre le problème de l’effet topographique, les spectres de réponse en accélération de différentes vallées vides seront étudiés dans le chapitre 2. Les courbes seront regroupées sur une figure unique, qui caractérisera l’effet topographique de manière quantitative et qualitative dans le domaine spectral. Toutes les configurations géométriques seront représentées. Dans le chapitre 3, nous chercherons à évaluer l’influence de la bidimensionnalité dans la réponse du site. Les réponses en accélération de vallées pleines seront comparées aux réponses de colonnes de sol unidimensionnelles. La hauteur de la colonne 1D de référence est choisie égale à l’épaisseur de la couche sédimentaire à l’aplomb du point d’observation considéré dans la vallée pleine. Pour modéliser l’effet topographique dans les différentes vallées vides pour obtenir les valeurs d’amplification spectrale sans utiliser les méthodes numériques comme HYBRID, on utilise une méthode de modélisation statistique (la méthode des moindres carrés, (least squares, LS)). Dans le chapitre 4, nous définirons des coefficients statistiques indépendants permettant de calculer l’amplification spectrale de l’accélération par une méthode prédictive. Un ingénieur civil a besoin de savoir identifier l’effet de site prépondérant en un point de la surface d’une configuration donnée. C’est pourquoi dans le chapitre 5 (conclusion), nous tenterons de définir un critère de prépondérance. Ce critère doit permettre de choisir un coefficient de sécurité adapté. Selon le cas étudié, nous verrons qu’il faut fonder le calcul du coefficient de sécurité : sur une analyse unidimensionnelle (effet géologique dominant), ou sur les courbes spectrales caractéristiques de l’effet topographique, présentées dans le chapitre 2 (effet topographique dominant).