Session H16: Computational Fluid Dynamics: Applications I

Minimizing Downstream Heat Transfer for Turbomachinery Design via Bayesian Optimization

When a turbine engine is manufactured, an inter-platform gap is created between the combustor annulus and the turbine annulus due to them having to be manufactured as separate parts.  This gap has to be supplied with leakage air to prevent ingestion of the extremely hot combustion gases into the interior of the engine. The two parameters that affect the effectiveness of this cooling system are the blowing ratio of the air inlet and the radius of curvature of the outlet of the gap. This work utilizes Bayesian optimization to find the ideal configuration of these parameters such that downstream cooling is maximized. Here, Bayesian optimization is a sequential design strategy that utilizes Bayes Theorem to direct the search for a global maximum or minimum of a black box function, and  results for every iteration were obtained by large-eddy simulation.   

Similarity theory for wind and buoyancy combined natural ventilation using CFD simulations

Natural ventilation has gained popularity in response to the increasing need for a sustainable and healthy built environment, but the design of a naturally ventilated building remains challenging due to the high variability in a building's operating conditions, i.e., wind and buoyancy. In the current study, we propose leveraging similarity theory to characterize the operating conditions with one non-dimensional number, the Richardson number, such that we can efficiently deal with this variability in computational fluid dynamics (CFD) models. We first validate high-fidelity large-eddy simulations (LES), which precisely resolve the flow and temperature fields, against full-scale ventilation rate measurements using the tracer concentration decay technique. Then, we verify the existence of an empirical relationship that quantifies the natural ventilation flow rate as a function of the Richardson number under two different scenarios: daytime with indoor temperature stratification and nighttime with uniform indoor temperature. The proposed relationship will enable us to characterize the ventilation rates under a wide range of operating conditions based on a limited number of simulations with different Richardson numbers.   

Vortex induced vibration of cylinders with an attached elastic plate

We investigated the effect of attachment of an elastic plate of high aspect ratio (AR= [10, 20]) to leeward side of cylinder undergoing vortex induced vibration at low Reynolds number of 150 using a flow solver based on the curvilinear immersed boundary method coupled with a finite element solver by the partitioned approach. Lock-in was observed when the non-dimensional natural frequency of the elastic plate, the non-dimensional frequency or the inverse of reduced velocity of the cylinder undergoing vibration, and the non-dimensional vortex shedding frequency from the cylinder were close to each other. The elastic plate at resonance was observed to be attaining very high modes of deformation leading to vortex-to-vortex interaction with cylinder. The transverse displacement of the cylinder with an attached elastic plate of length twice the diameter of the cylinder was found to be nearly twice of an isolated cylinder. Further, the lock-in regime in the presence of attached elastic plate was observed to be narrowed when compared to that of an isolated cylinder.   

Computational investigation of low-Reynolds-number flow vortices in coiled tubes

Coiled tubes are deployed in polymerisation, production of nanoparticles, extraction, and retrofits to heat-exchanger systems, where plug flow is a desirable characteristic. In this study, we superimpose oscillatory fluid motion at a low flow rate (with a Reynolds number, Re = 50) to characterise the vortex structures affecting plug flow. We apply a transient computational fluid dynamics analysis with a scalar transport technique, and tracer injection into the water working fluid. Oscillation amplitudes are varied for a set of coil geometries in order to understand the effect of vortices on residence time distribution and tracer mixing concentration. For optimal oscillation amplitudes, enhanced plug-flow performance is achieved due to the formation of Dean vortex structures. Existing processes with coiled tubes can be improved by easily implementing an oscillating motion.    

Effect of enrichments on the CH4 yield and CO2 conversion in a Sabatier process modeled in a microchannel reactor.

Sustainable energy systems for the future are gaining interest in terms of reducing carbon emissions and efficiently utilizing CO₂. The Sabatier reaction is one which has the potential to convert the atmospheric CO₂ into methane (fuels), and has attracted attention of the researchers because of its applications in the power to gas technology and space exploration missions. The Sabatier process, which is a catalytic reduction of CO₂ into methane, is highly reversible and exothermic in nature, and requires a higher than ambient temperature for occurrence. It is usually accompanied by a reverse water gas shift (RWGS) reaction, which results in an oxidizer-fuel (O/F) ratio of 3.5:1, as opposed to 2:1 with the Sabatier process alone. In the current study, a CFD model is developed to describe the reaction-coupled transport phenomena and effect of enrichments on the conversion of CO₂ and methane yield inside a microchannel reactor at different temperatures and pressures to determine the optimum operating conditions. The use of a microchannel reactor helps in providing a better thermal and mass transfer performance, while providing savings in space requirements. In addition, H₂ and CO₂ enrichment studies conducted here, deliver significant gains in terms of methane yield and CO₂ conversion, and also present very interesting effects on the reaction rates and associated products. In summary, a detailed set of parametric CFD studies of the Sabatier reaction help in identifying the optimal reactant ratios in order to achieve the highest efficiency   

Thermal phase change and bacterial inactivation in a superheated steam application using CFD simulations

Superheated steam is a promising technique used in applications to kill bacteria and sanitize dishes. To analyze the different phases of the cleaning process, OpenFOAM simulations were performed in an idealized 3D applications geometry with a nozzle and a plate at a temperature of 180 °C and a pressure of 10 bar. The interThermalPhaseChangeFoam solver [1] was used to calculate the heat and mass transfer between phases. The K-omega-SST turbulence model was used to capture the turbulent flow conditions. Bacterial inactivation was modeled using first-order kinetics in combination with the Arrhenius equation.  An unstructured tetrahedral mesh of the dishwasher geometry was created using ICEM CFD v.17.1.  Due to the high flow velocity of the steam and the strong steam shocks, which resulted in an increase in temperature, a higher steam condensation rate, and a lower bacterial concentration, the bacteria were initially deactivated at locations near the perimeter of the plate.  Most bacteria on the plate surface were deactivated in a short time of about 10 s.  This work forms the basis for future development and optimization of next generation superheated steam applications.

[1] M. Nabil, A. S. Rattner. SoftwareX, 5:216-226, 2016.   

Flow Analysis for the Development of the Coanda Vortex Ring Thruster (CoVoRT)

 The vortex ring thruster (VRT) is a propulsion system mimicking the propulsive technique of cephalopods, such as squids and jellyfish. VRT's acknowledged advantages include low noise generation, delicate attitude control, and precise maneuverability. However, due to the low propulsion efficiency, its application is limited to low-speed micro-scale vessels only. In the present study, VRT enhanced by the Coanda effect, dubbed CoVoRT, is proposed, aiming for propulsion efficiency improvement. The flows around the typical VRT and CoVoRT were analyzed using Computational Fluid Dynamics (CFD) simulations. For computational validation, the CFD results for the flow around VRT were compared with the previous simulations and experimental results. Then the propulsion performance of VRT and CoVoRT was compared in terms of the propulsion performance evaluation index (PPEI), and the vortex ring formation process. The results showed that CoVoRT's PPEI is approximately 50% higher than that of VRT. Also, the vortex ring formation by CoVoRT was more stable compared to that by VRT.   

Marangoni vs Weber (and Reynolds): Complex multiphase flows in the presence of surfactants

Surface active agents (surfactants) are present either as chemical additives or as contaminants in flows involving interfaces. These molecules adsorb at these interfaces and lower the interfacial tension. Interfacial gradients of their concentration lead to interfacial tension gradients, which, in turn, induce Marangoni stresses. The presence of these Marangoni stresses has a profound effect on the interfacial dynamics, which we illustrate numerically through several exemplar flows that include turbulent jets, falling films, bursting bubbles, and retracting filaments. These flows are characterised by Reynolds, Weber, and Marangoni parameters that reflect the delicate interplay amongst inertial, viscous, and capillary forces, as well as Marangoni stresses. The physico-chemical effects related to surfactant elasticity, diffusivity, solubility, and sorption kinetics are also taken into account.  We use an interface-tracking/level-set approach to capture accurately the interfacial dynamics which are punctuated by topological transitions in certain cases. We highlight the mechanisms underlying the surfactant-induced effects, which are often related to interfacial rigification and associated qualitative changes to the emergent flow structures. Open problems which constitute departure points for future research are also highlighted.    

Effect of impeller rotation on oil-water emulsion formation in stirred vessels

The mixing of two immiscible liquids, oil and water, by a pitched blade turbine (PBT) in a cylindrical vessel with various impeller rotating frequencies (f = 5,6,7,8,9,10 Hz) is studied. The vessel is filled with oil in the upper half and water in the lower half, and the PBT is immersed in the water phase with a small clearance from the vessel bottom. A hybrid front-tracking/level-set method is employed to capture the complex flow and interfacial dynamics within such a system. The rotation of the impeller introduces a primary vortex that can be observed in idealised rotational flows as well as in several secondary vortical structures resembling Kelvin-Helmholtz, vortex breakdown, blade tip-vortices, and end-wall corner vortices. These vortical structures lead to the lifting of the water phase around the vessel wall, while the oil in the centre is pulled downwards resulting in a deforming interface of a helical shape with 4 rotating curtains. As the rotation frequency increases, the interface deformation is accelerated, and ligaments are produced which subsequently break up into small droplets. Beyond a critical rotation rate, the system dynamics become extremely complex where the creation of ligaments, and the breakage and coalescence of drops occur simultaneously. The complex dynamics underlying such a mixing system are discussed in detail.    

Effect of surfactants on immiscible liquid-liquid dispersion formation in a stirred vessel

In this study, we model the mixing of two immiscible fluids, oil and water, in the presence of surfactant, in a stirred vessel equipped with a pitched blade impeller. Fully-three dimensional direct numerical simulations are employed in order to capture the complex interfacial deformation, rupture, and droplet coalescence, the change in topology, as well as the surfactant distribution along the interface. We demonstrate the development of a large vortex that reaches the impeller blades which is then followed by interfacial breakup into a multitude of droplets. The addition of surfactants modifies the local value of the surface tension along the interface and also induces Marangoni stress. We show how these surfactant-induced stresses alter the interfacial dynamics and give rise to profound changes in the drop size distribution depending on the choice of physico-chemical parameters such as the surfactant elasticity.