Numerical and experimental study of erosion in fluidized bed heat exchangers

The aim of this dissertation is to address the issue of erosion in fluidized beds under concentrated solar power with fluidized beds (CSP-FB) conditions, i.e. using candidate granular materials under different conditions of fluidization velocity and temperature. From previous research conducted by the Energy Efficiency and Thermal Systems (EnerSys) research group, silica sand and silicon carbide (SiC) were selected as potential candidate powders for CSP-FB, each one showing its own strengths and weaknesses. Nonetheless, their erosive power has not been comprehensively studied. Therefore, this is the main topic this doctoral thesis focuses on. A comprehensive literature review on erosion modelling in fluidized beds was done at the first stage of this doctoral studies. This review included experimental and numerical works, as well as the erosion models that were previously published. The results of the review revealed that, despite being a known problem, experimental erosion studies are difficult to conduct because of the long time it takes to wear a real device, and the influence of temperature on the mechanical properties of the materials, which may distort the evaluation of the erosive power of the bed. In addition, there are multiple variables affecting the process, such as the fluidization velocity, pressure or geometry, among others. Moreover, modelling erosion is also challenging because of being the result of the interaction of a multi-phase flow (fluid and solids) and a solid material. On the other hand, classic erosion models were developed for lean flows (particle contents rounding 1-2%) at high velocity (in the order of 50-150 m/s) for aeronautic engine blades, hydraulic machines and piping. These conditions are different from those that are likely to be observed in fluidized beds, where particle concentration is high (over 10% in volume, up to 60%) and velocities are in the order of 1 m/s. In the second stage, erosion was predicted using several erosion models available in the literature that were implemented using the hydrodynamical data obtained with numerical MP-PIC simulations. Moreover, a new erosion model based on the necessary energy to grind a material was developed. The interest in MP-PIC technique lays on the fact that MP-PIC simulations are faster and less computational resources demanding than CFD-DEM, whose application is limited to small-scale. Conversely to CFD-DEM, MP-PIC has been successfully applied to predict the fluid dynamics, thermodynamics and chemical behaviour of industrial-scale fluidized systems. The predictions were compared to previous experimental results and other models available in the literature to estimate erosion in fluidized beds, finding that the new model is able to predict erosion reducing the deviation by a 9%, approximately. The last stage of this doctoral studies included experimental studies performed under conditions resembling those expected in CSP-FB plants using silica sand and silicon carbide powders. The choice of silica sand was motivated by its availability (low price) and the fact that it can be easily fluidized using low flow rates ought to its low density. Silicon carbide is more expensive and dense than silica sand, but its superior optical and thermodynamic properties may balance its high price. The results revealed that the angular shape and high hardness of silicon carbide make it about a hundred times more abrasive than silica sand under similar fluidization conditions.