Turbulent supercritical mixing. Selection of methods and tools

Resumo

Turbulent mixing is the ability of turbulent flows to effectively mix entrained fluids to a molecular scale. At high pressures, the fluid dynamics behave very differently than at reduced pressures lower than unity. Hence, turbulent mixing is also different. The aim of this thesis is to gain an understanding in turbulent mixing (macro- and micromixing) involving supercritical fluids, and its influence in some of the most important supercritical-based processes nowadays. To this end, first macromixing prediction is discussed. An assessment of turbulence models to discern which is the most suitable choice when dealing with supercritical fluids has been done. Two studies have been carried out and validated against experimental data available. First, an analysis of a turbulent submerged nitrogen jet, secondly a RTD (Residence Time Distribution) analysis carried out in a reactor working with CO2 at supercritical conditions. The results show that the Realizable k- epsilon model seems to be the most accurate choice for the simulation of high pressure systems, providing average deviations lower than 18% from experimental values. Additionally, a methodology based on the solution of a transport equation for the residence time has been employed to identify and locate backmixing zones inside the reactor. The Standard k - $\varepsilon$ model is employed for RTD prediction in the supercritical nanoparticle synthesis of \ce{TiO2} for the explanation of experimental results. In that process, the particles are generated by decomposition of an organometalic precursor in supercritical CO2. The kinetic of the process is slow, and the final product quality is macromixing dependent. Peng-Robinson equation of state with Huron-Vidal mixing rule has been used to predict density variations within fluid-dynamic equations. A pseudo-first order kinetic (r{TiO2} = k C_{D}, mol L-1 s-1 with k =0.0297 min-1) has been fitted to experimental data. Results show how the experimental particle size distribution has the same shape as the predicted RTD. Since the CFD predictions of the global flow have been validated, micromixing can be explored with some confidence. A first insight into micromixing has been achieved by studying its influence on the hydrothermal combustion processes. The combination of a micromixing model for liquid reactions along with the Eddy Dissipation Concept (EDC) allows for a good estimation method of the turbulent reaction rates. The method is validated for the combustion of methanol in supercritical water, predicting the experimental temperature profile with absolute average deviations below than 10\% for most of the cases tested. The influence of different parameters (inlet temperature, inlet mass fraction, etc.) has been explored. The main result shows that mixing in these types of systems is liquid-like, since Schmidt number is higher than one. The model proposed is compared with the original EDC model, showing a much better performance for the case studied. To continue the study of micromixing at supercritical conditions, a more complex model, based in the Direct Quadrature Method of Moments model (DQMoM-IEM model) is employed to quantify mixing efficiency in supercritical water hydrothermal reactors (SWHR). In fact, mixing plays a crucial role in determining the final particle size distribution and therefore in the final product quality in these types of systems. The performance of the model is investigated in three different scenarios, corresponding to situations with very different values of the Richardson number, and in different mixer configurations. Results showed how a global mixing time can be used to quantify mixing, and how turbulence can enhance the nanoparticle synthesis process, leading lower particle size and narrrower particle size distributions. These results confirm previous research in the topic. To conclude the study of micromixing at high pressures, the supercritical antisolvent process is studied thoroughly. This is a system that exhibits a strong coupling between mixing and precipitation. Controlling the particle formation in order to obtain small particles is a key issue in the Gas Antisolvent (GAS) and Supercritical Antisolvent (SAS) processes, and this is directly related to mixing in all scales. To describe this coupling, presumed Probability Density Function (PDF) methods has been employed: DQMoM-IEM model along with the quadrature method of moments are used to describe the behaviour of the particulate phase. Different nozzles were analyzed in order to get an insight into the mixing process associated to each geometry in tubular reactors, and its influence in the final particle size distribution. Results showed how mixing can be determinant in getting lower particle size, and how mixing in the microscale can be a significant parameter to account for in the design of precipitators. An auxiliary study of flammability limits estimation is performed. The lower flammability limit (LFL) of a fuel is the minimum composition in air over which a flame can propagate in a stable way, and defines the range of fuel concentrations for flame propagation to occur. The methodology proposed is an adaptation of the so-called analytical method to high pressures. For such purpose, an equation of estate has been added to the original model, and a different calculation path is proposed. Results show how the lower limit can be estimated with good accuracy for engineering calculations in the H2-O2 system.