1,720,979 research outputs found
Flow instabilities and reversals in non-uniformly thermocapillary driven melt pool
No full textWith transient LES and DNS simulations, we investigate flow in melt pools driven by thermocapillary forces. The developing pool is at first axisymmetric as are the boundary conditions, but flow instabilities arise that lead to 3D oscillatory flow patterns. At higher laser powers a sign-change in the surface tension temperature coefficient occurs, resulting in a flow reversal in the pool and thus two counter-rotating vortices, which exhibit similar though more complex flow instabilities
Marangoni driven free surface flows in liquid weld pools
No full textExtending the weldability of novel materials, and improving the weld quality by tailoring weld microstructures are key factors to obtain the welding techniques demanded in the modern manufacturing industries. This can be done, for example, by feeding chemical elements from a consumable wire into the weld pool during welding. The mixing of chemical components in the weld pool and the resulting post-solidification weld microstructures are influenced by weld pool hydrodynamics. Weld pool hydrodynamics is known to be primarily driven by Marangoni forces acting at the free liquid surface, i.e by tangential gradients in surface tension along the liquid surface due to pronounced lateral gradients in temperature and surface active element concentration. In this research, we develop a Computational Fluid Dynamics model to study steel weld pool hydrodynamics during conduction mode laser spot welding. It is concluded that free surface deformations and instabilities have a strong impact on the fluid flow and heat transfer in weld pools, and should therefore be accounted for in weld pool simulations. With increasing the surface active element concentration and laser power, the weld pool flow becomes highly unstable and can no longer be accurately modeled with a flat surface assumption. More accurate predictions of weld pool physics can be made if the free surface, solidification stage, and three-dimensionality are taken into account. This reduces the need for the use of unphysical parameter fittings widely reported in literature.Multi Scale PhysicsApplied Science
Flow, heat and mass transfer through CBRN protective clothing
No full textChemical EngineeringApplied Science
Numerical simulations of rarefied gas flows in thin film processes
No full textMany processes exist in which a thin film is deposited from the gas phase, e.g. Chemical Vapor Deposition (CVD). These processes are operated at ever decreasing reactor operating pressures and with ever decreasing wafer feature dimensions, reaching into the rarefied flow regime. As numerical simulation tools are frequently used to design and improve reactors, there is a need for numerical simulation tools capable of modeling rarefied internal gas flows. To fulfill this need, the Direct Simulation Monte Carlo (DSMC) method was implemented in the 2D STARS and 3D X-Stream codes. The DSMC method was developed for high speed rarefied flow simulations in space applications, but is also suitable for low speed rarefied gas flows as e.g. found inside deposition reactors. After the implementation, validation of both codes was performed on a wide variety of problems, including a thin film deposition experiment that was designed and performed especially for the validation of the codes. The numerical results of the developed codes were very satisfactory in all cases, leading to the conclusion that the DSMC method in general, and the STARS and X-Stream codes specifically, are a very promising tool for use in internal rarefied gas flows, including thin film deposition processes. Finally, the developed codes were used for calculations on three applications in the thin film industry, namely the heat transfer in a stagnation flow CVD reactor, thin film deposition through a shadow mask and the reaction rate of a gas at a surface with a reactive sticking coefficient.Applied Science
Breakup of confined droplets in microfluidics
No full textSegmented-flow microreactors have emerged as an attractive tool for fine chemical synthesis and (bio)chemical analysis, owing to their high heat and mass transfer rate, low axial dispersion, as well as rapid mixing. A key challenge for the use of segmented-flow microreactors in large-scale processing is their low throughput. This can be overcome by applying the concept of numbering-up in which several microreactors are placed and operated in parallel. A challenging aspect of this approach is to distribute segmented flows over those parallel microreactors with a high uniformity in the size and the speed of the fluid compartments. In this thesis, we propose to use a bubble-splitting distributor where a single stream of fluid compartments is recursively split into smaller ones via a series of T-junctions. We first investigate the fundamental physics of the breakup of droplets in a single T-junction using CFD simulations. Being able to explain the mechanism of the droplet breakup leads us to a more applied question, how to design a bubble/droplet distributor and what are the optimum operating conditions. We present theoretical and experimental analyses of the uniformity of the distribution of bubbles/droplets using the proposed distributor and provide guidelines to operate it for different flow conditions. The thesis ends with a discussion on how some of our main findings can be generalized and opportunities for future research.Chemical EngineeringApplied Science
Multi-Scale Modeling of Gas Flows with Continuum-Rarefied Transitions Application to Expanding Gas Jets in Thin Film Deposition Processes
No full textApplied Science
Electromagnetic control of oscillating flows in a cavity
No full textIn continuous steel casting, liquid steel flows turbulently through a submerged nozzle into a thin, vertical mould. In the mould the liquid steel is cooled, such that it solidifies and plate steel is formed. On top of the liquid steel in the mould, a slag layer is present and due to the turbulent behavior of the flow, particles and droplets from the slag layer can get entrained into the bulk flow. This leads to inclusions in the final product, which is unwanted. The flow in the mould needs to be stabilized, such that entrainment effects no longer play a role. For this, electromagnets are generally installed next to the mould. The flow of the electrically conductive liquid steel through the magnetic field induces an electrical current, from which an induced Lorentz force emerges, which acts as a so-called electromagnetic brake. This dissertation presents an experimental study on flow dynamics, heat transfer, and electromagnetic interaction in a thin slab continuous casting mould. To mimic the continuous casting process, a glass model of the mould was fabricated and (salt) water was used as modeling fluid, such that particle image velocimetry measurements could be performed. The flow of both single and bifurcated jets is studied. The jets issuing into the thin cavity and the induced flow in the cavity exhibit a self-sustained oscillating behavior with a frequency that grows linearly with the jet velocity. It was found that the self-sustained oscillations exist due to an imbalance between the inertial forces in the recirculation zones alongside the jets and the pressure force due to a low pressure zone in these recirculation zones. The low pressure zone in the center of the recirculation zones can exist due to the (semi) two dimensionality of the flow. When a thicker cavity is employed, the self-sustained jet oscillations vanish due to a less structured, and more three dimensional, flow pattern. Next, the influence of the self sustained jet oscillations on heat transfer at the wall is studied for flow from a bifurcated nozzle. A constant high inlet flow temperature is applied, in combination with cooling of one of the broad walls. Measurements of the temperature at the cooled wall are performed using thermochromic liquid crystals (TLC’s) attached to the cooled wall. The self-sustained jet oscillations show an imprint on the TLC’s. At the point where the shear layers of the jet reach the wall a hot spot is formed, and in the center of the recirculation zone alongside the jet a cold spot is found. The cold spot moves with the jet oscillation, leading to a non-uniform and time-dependent temperature distribution at the cooled wall. Measurements of the temperature drop of the liquid over the cavity have been performed and the average heat transfer coefficient h was found to scale with Re^0.8. Subsequently, the self-sustained jet oscillations are influenced by means of an applied electromagnetic force. This is done by applying an electrical current through a saline solution across the width of the cavity, in conjunction with a permanent magnetic field perpendicular to the electrical current. The combination of the electrical and magnetic field with a liquid with high electrical conductivity (as compared to tap water) leads to a permanent and local Lorentz force. This Lorentz force can be applied such that the jet oscillations are either suppressed or enhanced. In the oscillation suppressing configuration, the flow due to the Lorentz force prohibits the recirculation zones from forming, and so no low pressure zones emerge alongside the jet. Above a critical forcing strength, this suppresses the self-sustained jet oscillations completely. In the opposite, oscillation enhancing, configuration, the flow due to the Lorentz force increases the formation of the recirculation zones and hence the oscillation frequency increases. We finalize this thesis by discussing how the experimental results from this work can be used in the design and optimization of actual steel casters, and in the validation of numerical models to be used for that same purpose. Shortcomings of the experimental methods are discussed, as well as the appropriate scaling of physical parameters.Chemical EngineeringApplied Science
On the Interaction of Capillary Shapes with Solid Surfaces
No full textControl over the interaction of droplets with solid surfaces is commonplace in nature. Famous examples are the water-shedding capabilities of the lotus leaf and the water-harvesting skin of certain types of beetles. To date, this type of control remains a challenge in engineering applications. Consider, for example, droplets on the windscreen of a car that need to be wiped off because they unwantedly stick to the glass, or the dripping of droplets from fog harvesting devices, where one actually wants the droplets to stick. In this thesis, we address the interaction of droplets with solid surfaces that underlies the sticking behavior. In particular, we investigate how droplets stick to chemically and physically heterogeneous surfaces, and what shape droplets take when squeezed by various confining geometries. The results of these investigations can be used to predict the force needed to set a droplet on a given surface into motion, and to calculate the volume of droplets in confining geometries. An important tool in our studies is the principle of energy minimization, by which we determine the (local) equilibrium shape of a droplet under the influence of surface tension, gravity and solid boundaries. It turns out that the interaction of droplets with solid surfaces can largely be captured in terms of two aspects of the system. The first is the contact line of the droplet, the line where the fluid-fluid interface meets the solid surface. The second is the contact angle distribution around this line, where the contact angle is defined as the angle at which the fluid-fluid interface meets the solid surface. A core question in our work is how to predict the shape and sticking force of a droplet on a realistic, thus physically or chemically heterogeneous, inclined surface. The answer to this question constitutes an important step towards a model for the sticking force. We resolve this question by studying the model system of a droplet on an incline, using 2D analytical and 3D numerical energy minimization. An important conclusion from our work is that the maximum sticking force of the droplet is determined by local constraints on repositioning of the contact line, originating from a range of possible contact angles within which the droplet can deform while the contact line remains static. The possible, i.e. accessible, shapes of the droplet during the inclination process are then also determined by constraints on repositioning, which causes the initial shape and deformation history of the droplet to influence the shape of the droplet at roll off and thereby the maximum sticking force. Accounting for inaccessible shapes and droplet shape history hence results in accurate predictions of the sticking force, in contrast to currently available models that allow all shapes and are thus bound to yield incorrect predictions. Based on the insight that local constraints control the global shape of the droplet we developed an approach to analytically predict the deformation of 3D droplets upon inclination and the maximum sticking force of these droplets. This prediction can be used for a number of engineering applications where the sticking force of droplets is an important factor. Examples include the design and positioning of solar panels and windscreens, as well as more industrial applications such as condensors and chemical reactors with small flow channels. The analytical model for 3D droplets builds on the findings on the role of local constraints by explicitly modelling the deformation process of the contact line, in contrast to the current state-of-the-art where contact line shapes are either assumed or predicted without regard for the constraints. To keep the model analytically tractable, we parameterize the deformation of the droplet in terms of the contact line shape and the contact angle distribution around the contact line. In combination with a few physically sound assumptions, based on our earlier findings, this model accurately predicts the deformation and maximum sticking force of a droplet, as compared with experiments and numerical simulations. Unlike currently available models, this model also accounts for the history of the droplet in the prediction of the maximum sticking force. Apart from droplets sticking to surfaces we also use energy minimization to determine the shape of bubbles and droplets in straight microchannels of various cross-sectional geometries. With this, we develop a model to calculate the volume of a droplet on the basis of only the channel geometry and the droplet length as observed from a top view. This determination of volume is for example important to accurately determine the concentration of gaseous species or nutrients in studies of mass transfer rates or growth of cells. The model is based on a physical description of the central part of the droplet, where we calculate the shape of a cross-sectional slice of the droplet based on energy minimization. The volume calculation is completed with two caps on the droplet that smoothly fit onto the central part and mimick the physically realistic shape. We compare our analytical model with 3D numerical energy minimization calculations and find excellent agreement for a large range of droplet lengths and channel geometries, thus validating the model for use in quantitative research. The findings of this research can, with small adaptations, be applied in related systems, such as pinned gas bubbles or droplets in a shear flow. We discuss the necessary adaptations and what open questions remain in the field of droplet pinning, specifically related to cases where assumptions made in our work do not apply. This results in a series of recommendations for future research on the role of viscous and inertial effects and heterogeneities with sizes on the order of the droplet.Chemical EngineeringApplied Science
Formation and Transport of Bubbles in Microfluidic Systems
No full textPrecise manipulation of minute volumes of fluids is at the heart of microfluidics and opens up an exciting route to miniaturize processes in the areas of chemistry, biology and medicine. An attractive way to transport small fluid samples through the channels of microfluidic devices is by enclosing these samples inside containers in the form of microbubbles or microdroplets. The work described in this thesis is motivated by the need to reliably form such microbubbles and microdroplets and robustly transport them through networks of microchannels at high throughput. One of the central questions addressed in this thesis is how bubbles form at T-junctions under conditions typical to bubble-based microfluidic systems. The basic question how large bubbles grow is resolved by studying the fundamentals of bubble formation and pinch-off using experimental techniques, such as micro Particle Image Velocimetry. Another key question addressed in this thesis is how bubbles are transported in confined geometries. We hereby considered transport of bubbles through simple microchannel networks as found in for instance micro reactors and laboratories-on-chips, as well as through more complex microchannel networks comparable to for instance the pore structure in micro packed beds or oil reservoirs.Multi-Scale Physics / Chemical EngineeringApplied Science
Efficient numerical methods for the instationary solution of laminar reacting gas flow problems
No full textIn production processes of micro-electronics, optical and mechanical coatings and solar cells, high-purity materials in the form of a powder or a thin film are significant importance. The deposition of thin films on irregularly shaped surfaces can be done by chemical vapor deposition (CVD). Computer simulations are widely used to design CVD reactors and to optimize the process itself. The core of a computer simulation is a mathematical model of the gasflow and all chemical processes within the CVD reactor. A lot of (commercial) computer software has been written for CVD simulation. The emphasis has always been on modeling and validation. For the 'old' CVD processes this approach was sufficient. However, with the deposited films getting thinner and thinner, process times are reduced and transient times become more important. Further, the technology is moving towards inherent transient CVD processes (such as atomic layer deposition) making instationary simulations indispensable. Computing the time-dependent solution of the underlying mathematical equations is hard, because the involved chemistry makes these equations hard to solve. Commercial CFD software packages have often great problems to compute these solutions. Solutions computed by various codes have been reported to differ a lot and the computational times needed to find solutions are generally excessive. In this thesis a rigorous mathematical approach has been applied to these problems, with the aim to reduce computational times for simulations of CVD and related applications. The numerical techniques proposed in this thesis enables us to perform instationary, multi-dimensional gas flow simulations with multi-species, multi-reaction CVD chemistry in a computationally efficient way.Electrical Engineering, Mathematics and Computer Scienc
- …
