Artificial ground freezing (AGF) is a process used to strengthen soil and rock by freezing trapped pore water. Freezing is accomplished by pumping calcium chloride brine, chilled to approximately – 30˚C in ammonia refrigeration plants, through heat exchangers drilled into the ground. A knowledge gap exists in the field of AGF regarding the relationship between the performance of the refrigeration plants and the ground heat removal process. The coupling of these two aspects of AGF requires knowledge of the plant’s refrigeration capacity as a function of many factors; the most important of which is the temperature of the brine returning from the freeze pipes. However, refrigeration plant manufacturers do not provide sufficient information about the plant’s performance as a function of brine temperature. Typically, AGF plants are only rated at one operating point due to the impracticality in experimentally rating such large plants and the lack of any standard test methods. Refrigeration system models available in the existing literature do not emulate the compressor control system responsible for preventing compressor overloading. Therefore, the goal of this research is to develop a model that can predict the performance of an AGF refrigeration plant over a range of operating points, using plant specifications that are readily available in the documentation provided by the manufacturer of the plant. To fill the knowledge gap, a thermodynamic model is developed of an existing 1500 TR AGF plant at Cameco’s Cigar Lake mine. The Cigar Lake plant uses flooded shell-and-tube evaporators, two-stage economized twin screw compressors, and air cooled condensers packaged into five refrigeration modules. Each component in the system, including the evaporator, compressor, and condenser, is modeled individually, and then the individual models are combined to calculate the overall system capacity. The model emulates the behavior of the compressor’s slide valves, which are used to limit the plant capacity, limit suction pressure, control intermediate pressure, and control the discharge pressures in the system. In addition, the model accounts for the effects of the oil injection into the screw compressors, which cools the compressors and seals the spaces between the lobes of the compressor rotors. The model is validated using operating data from the Cigar Lake plant, which was collected over a period of eight months by plant operators. After calibration, the modeled plant capacities and the temperature of the brine leaving the refrigeration plant are found to be in agreement with the measured capacities and brine temperatures. The overall plant capacity results match measured capacities within ±14%, and the predicted brine temperatures match the measured values leaving the plant within ±5%. The modeled capacities match the measured capacities within the uncertainty in the measured data. The simulation of the Cigar Lake plant demonstrates that the performance of the plant is highly dependent upon the temperature of the brine returning to the plant. For example, a ±10% change in brine temperature causes a 22% overall change in the capacity of the refrigeration plant. The simulation also demonstrates that, even with the plant’s air cooled condensers, changes in the ambient temperature have little effect on the performance of the plant with the existing equipment. Furthermore, the results show that the selected suction pressure of the second compression stage, or intermediate pressure, affects the performance of the refrigeration plant. These findings lead to important plant performance optimization opportunities. An optimization study using the model demonstrates that, by selecting a lower intermediate temperature than what the existing literature suggests, an improvement in overall refrigeration plant capacity of 3% can be achieved. Additional simulations identify the brine tank, which allows for different brine flow rates to exist on the field and plant side of the tank, as an inefficient component in the system. The brine tank not only cools the brine returning from the field before it is pumped to the refrigeration modules but it allows heat to be transferred between the warm and cold brine. By eliminating the tank, plumbing all of the refrigeration modules in parallel, and installing appropriately sized evaporators, the capacity of the refrigeration plant can be increased by 17%. Further capacity gains can be realized by upgrading the evaporators to increase their capacity.

La congélation artificielle des terrains est une technique d'imperméabilisation et de renforcement des terrains régulièrement employée dans le génie civil et l'industrie minière. Dans un objectif de prédiction fiable de l'évolution de la congélation dans le milieu poreux, cette recherche propose deux nouveaux modèles numériques permettant la simulation du problème global de la congélation artificielle des terrains. Un premier modèle a pour objectif la représentation des mécanismes couplés thermo-hydrauliques associés à la congélation du matériau tandis qu'un deuxième modèle se concentre sur l'estimation des échanges de chaleur entre un puits de congélation et le terrain environnant. Le modèle thermo-hydraulique, en plus d'être cohérent sur le plan thermodynamique, a été vérifié à la fois par rapport à des solutions analytiques et par rapport à des résultats expérimentaux obtenus à grande échelle en conditions d'écoulements importants. Le modèle puits-terrain adopte une approche innovante par rapport à la bibliographie. Il permet de déterminer les conditions aux limites des modèles de congélation des terrains, difficiles à connaître en pratique, et d'optimiser les conditions opératoires du système grâce à des temps de simulation limités. De par les hypothèses considérées, leur fiabilité et leur praticité d'utilisation, ces deux modèles sont particulièrement adaptés à des sites industriels comme celui de la mine d'uranium de Cigar Lake (Canada) qui présente deux contraintes majeures : la présence potentielle d'écoulements importants et la forte hétérogénéité des terrains à congeler. Dans de tels contextes, des applications d'utilisation conjointe des deux modèles ou non sont présentées par rapport à des cas simples et au cas industriel de Cigar Lake. Ils peuvent ainsi être employés pour prédire l'évolution de la congélation dans le terrain en tenant compte des interactions thermo-hydrauliques, pour optimiser le système de congélation, ou encore pour évaluer l'impact sur la progression des zones congelées de conditions géologiques, hydrogéologiques et opératoires particulières. / Artificial ground freezing is a ground sealing and reinforcement technique regularly used in civil and mining engineering. In order to reliably predict the freezing evolution in the porous medium, this research offers two new numerical models allowing the simulation of the global problem of artificial ground freezing. A first model aims at representing the thermo-hydraulic coupled mechanisms associated with the material freezing while a second model focuses on the estimation of heat transfers between a freeze pipe and the surrounding ground. The thermo-hydraulic model, in addition to being thermodynamically consistent, has been verified both with respect to analytical solutions and large- scale experimental results obtained under conditions of high water flow velocity. The pipe-ground model adopts an innovative approach compared with literature. It allows to determine the boundary conditions of the ground freezing models, not readily available in practice, and to optimize the operating conditions of the system thanks to limited simulation times. By the considered assumptions, their reliability and their practicality, these two models are particularly well adapted to industrial sites like the uranium mine Cigar Lake (Canada) which presents two major constraints: the potential presence of high seepage-flow velocities and the strong ground heterogeneity. In these contexts, applications of the two models, jointly used or not, are presented with respect to simple cases and to the industrial case of Cigar Lake. They can be employed to predict the freezing evolution in the ground considering the thermo-hydraulic interactions, to optimize the freezing system, or to evaluate the impact of specific geological, hydrogeological and operating conditions on the freezing progress.

Artificial ground freezing as a method to temporarily stabilize and create hydraulic sealing in urban as well as in rural areas has been used in a number of Swedish construction projects, particularly during the last decade. One problem with the freezing of soil and rock is that fine-grained clayey types of soils have showed a tendency to under certain circumstances, during the thawing process, create a pore water overpressure and to consolidate, despite a change in the external loading conditions. In certain cases, this condition can be a desired effect as the soil mass after a freeze- and thaw cycle acquires overconsolidated properties. The main objectives of this study are, to describe and review the knowledge and current state of practice of artificial ground freezing, to increase the understanding about the conceptual behaviour for prognosis of the vertical deformation concerning artificial ground freezing and to compare and discuss results from laboratory and field studies concerning vertical deformation during thawing process for Bothnia soil. The field studies and the laboratory tests in this research study have been performed with soil from the freezing of the Bothnia Line in the vicinity of Stranneberget. The Bothnia Line is the railway link between Nyland, north of Kramfors, and Umeå. This thesis relates to a part of the Bothnia Line. It deals with the behaviour of soil during thawing by means of temporary stabilization and hydraulic sealing of fine-grained soil through artificial freezing using brine as the cooling agent. However, the reason behind the problem consists of the final deformations due to the thawing process. The general conclusions of this study are; the Bothnia soil water content decreased in mean approximately 14 % after a freeze-thaw cycle, which approximately corresponds to; wth = 0.8w – 1.5 the decrease of the water content has no correlation to the depth below ground surface, in contrast, there is a strong correlation between the undisturbed soil water content and the magnitude of the decrease in water content the soil liquid limit decreases after a freeze-thaw cycle, simultaneously as the relative share of clay and fine silt grains decreases while the relative share of more coarse grains increases the coarser and denser soil created after a freeze-thaw cycle obtains an increased preconsolidation pressure and an increased undrained shear strength. / QC 20100721