Prilling is a spray solidification process of producing spherical beads, or prills, typically for fertilizers like urea and ammonium nitrate. The process involves molten material passing through perforated rotating containers, forming spiraling jets and breaking up into droplets due to a Rayleigh-Plateau instability by surface tension. These droplets undergo heat exchange with a cooling air stream and turn into solid prills. However, controlling prill size and shape remains challenging due to complex fluid dynamics, thermal effects, and non-Newtonian material properties. This research focuses on understanding jet breakup under the influence of thermal effects and body forces related to gravity and rotation (centrifugal and Coriolis forces). The overarching goal is to optimize the prilling process for a desired narrow prill size distribution by predicting and controlling the jet breakup process. A fundamental and holistic approach was taken based on a combination of experiments, stability analysis and various types of numerical simulations.
The study begins by exploring the dynamics of a spiraling liquid jet emitted from a rotating orifice. Using experiments, linear stability analysis and a nonlinear slender-jet model, the influence of the body forces from rotation and gravity on jet thinning and breakup is quantified. Imposed mechanical vibrations within the nozzle modulate the jet conditions at nozzle exit, enabling control over the jet breakup length and droplet size. The findings reveal that spiraling jets behave similarly to straight jets under gravity, where the body forces influence the thinning rate of the base jet and local base jet conditions control the growth rate of unstable jet perturbations. This insight opens pathways for more precise control mechanisms in various jet-based industrial applications....@en

The presence of particles in a flow can change the flow behaviour in an unpredictable manner, that leads us to multiphase flow. Multiphase flows are present in nature and in everyday life around us. The most common example of a multiphase flow, that one can think of, is the treatment of drinking water using liquid-solid fluidization in softening processes. Multiphase flows can also have very high relevance in various practical applications ranging from sedimentation of particles for process industries, transport of a dense suspension (slurry) through a pipe and dredging applications for land reclamation. Even though a lot of research has been done on liquid-solid fluidization in inertial and viscous dominated regimes, it remains far from being understood completely.
Experimental studies were conducted on the fluidization of monodisperse solid spheres within a rectangular channel. This research saw the development of an experimental setup, to study this phenomenon. The bulk solid volume fraction was studied in three different regimes. At a dilute regime (Φ ≤ 4%), moderate regime (Φ = 8 − 14%), and a dense regime (Φ ≥ 15%). To study the impact of inertia, three distinct Galileo numbers in the inertial regime were examined: 𝐺𝑎 = 147, 196, and 252. At these values, a singular settling/rising sphere presents markedly varied wake and path characteristics.
Throughout the study, the particle/fluid mass density ratio remained constant at 0.87. Results indicate that within the range of Φ = 4 − 18%, both the microstructure and dynamics of the suspension are heavily influenced by the bulk solid volume fraction. This aligns qualitatively with previous literature. Enhanced fluidization was noticed in the experiments. This is usually driven by the formation of clusters.
Particle clustering within these regimes is driven by various mechanisms: trapping of particles in wakes, a drafting–kissing–tumbling (DKT) instability causing vertically aligned particles to swiftly realign horizontally, and multiparticle interactions with collisions playing a minor role. On the macrostructure level, intrinsic convection was captured in the fluidization column. To further study these characteristics, instantaneous snapshots were illustrated for varying cases. Profiles for both mean velocity and root-mean-square (RMS) velocity were obtained. Wall effects emerged as pronounced, diminishing in importance with increased bulk solid volume fraction and augmented 𝐺𝑎 values. Intrinsic convection is observed in these experiments, and the velocity profiles provided an indication. Probability density functions (pdf) for normalized velocities highlighted evident particle clustering.

Thermal convection is a phenomenon seen in almost all facets of life, ranging from planetary convection to ocean currents and convection inside the Earth. The physics of thermal convection complicates when a porous wall layer is present. Flow over urban canopies, forest canopies or flow in underground aquifers are classic examples where thermal convection occurs in the presence of turbulent flow over a medium that may be described as a porous wall layer. The present work focuses on simulating pressure-driven turbulent flow over a simplified, ordered porous medium consisting of a regular array of cubes. The work further couples it with natural convection arising due to unstable stratification, to provide insight into its momentum and heat transfer characteristics. Direct numerical simulations (DNS) have been performed with a finite-difference Navier-Stokes solver to validate the model for buoyancy-driven convection and the classical Rayleigh-B´enard convection. Further, we extended the solver with a volume penalization Immersed Boundary Method (IBM) to model the ordered porous medium, which was validated against reference data. The bulk Reynolds in the overlying free channel region is fixed at 5500, the Prandtl number at 0.1, with an adiabatic boundary condition on the surface of the cubes. The bulk Richardson number, to cover different flow scenarios, from pure shear to purely buoyancy-driven flows, is also varied. The flow statistics show three distinct regimes spatially, (1) the porous region, (2) the turbulent channel and (3)the interface regime, where heat transfer timescales vary drastically between the porous region and the turbulent flow region. Looking at the regime transition from the perspective of changing Rib, the change starts at about bulk Richardson number Rib ≈ 0.1, with the flow switching from mostly buoyancy dominated to shear-dominated convection. The possible regimes as a function of the governing parameters are described, with three regimes formed using critical Rayleigh number limits for convection in free media and in porous media. The qualitative analysis of the thermal structures formed reveals, indeed, that the theorized critical limits are indeed seen through simulations. Lastly, length and velocity scalings are also suggested separately for shear-driven and buoyancy-driven regimes for different regions of the flow domain.

Urban microclimate significantly affects people’s experiences and activities in urban environments by a series of phenomena, among which urban flow is an important factor to be considered. Computational Fluid Dynamics (CFD) method has become a popular tool for studying urban airflow because of its low cost compared with experiment methods. However, flows over urban areas exhibit turbulence nature of being three-dimensional, unsteady, and multi-scale. Additionally, the large computational domain that should be covered and the inherent inhomogeneity of the urban structures make it challenging to do full-scale modelings. Large Eddy Simulation (LES), with the development of computing power, becomes a promising tool to study such flows.

In the campus of Delft University of Technology (TU Delft), a crossroad near the EWI building (the main building of Faculty of Electrical Engineering, Mathematics and Computer Science) is constantly complained for its strong wind. This research tackles such problem using LES, and takes TU Delft campus area itself as case study. The development of this research is composed of three stages.

In the first stage, Vreman eddy viscosity model is implemented into Canonical Navier-Stokes (CaNS), a massively-parallel Navier Stokes solver developed by Costa, (2018). Based on a structured three dimensional Cartesian grid, the subgrid scale (SGS) eddy viscosity model is inserted into the Navier Stokes equation by adding an extra diffusion term. The inserted diffusion term is discretized with second-order difference scheme along with interpolation of the velocity field due to the staggered grid arrangement. The implementation is validated with a turbulent channel flow with friction Reynolds number 𝑅𝑒𝜏 = 360. The good agreement is found and the discrepancy is small.

In the second stage, the solver employs a direct-forcing Immersed Boundary Method (IBM) and is further validated with the flow over periodic cube arrays. Signed-Distance Field (SDF), as a convenient tool, is generated and functions as read-in data for IBM. The IBM processes the effect of the boundary as an added force on the fluid points at the interface. The stair-step approach approximates the structure boundary with the cuboid cells faces. The results match well with the wind tunnel test data from Castro et al., (2006) and a previous LES study by Tomas et al., (2016).

In the last stage, the validated solver is applied to a scaled-down TU Delft campus model. The simulation setup is designed by considering the achievability of a possible future wind tunnel measurement. Three grids are used for a grid convergence analysis by comparing the total IBM force, mean velocity, and Reynolds stress at certain locations. The flow converges with the finest grid with grid number 𝑁𝑥 × 𝑁𝑦 × 𝑁𝑧 = 960 × 880 × 240. Around the EWI building, a high-speed region is found at the cross road location. Behind the building, a wake area is observed, and a clear shear layer is on the top of the building.

Dense suspensions can be found in various industrial and natural processes. A relatively new technique uses the principle of additive manufacturing to produce products from a wide variety of materials by printing with dense suspensions. To reach a high print quality the suspension rheology must be understood very well. Current knowledge about suspension rheology however lacks the capability of predicting the exact behaviour of a predefined suspension, especially for dense suspensions. Direct numerical simulation (DNS) in combination with a second order accurate immersed boundary method (IBM) (Breugem, 2012) can be a useful tool to research suspension rheology. However, currently no numerical results are known for suspensions close to the jamming limit produced with this method.
The current work validates the capability of the IBM to simulate dense suspensions by comparing produced results with existing numerical and experimental data. This is done by simulating a plane Couette flow for a range of particle volume fractions 𝜙=0.2−0.6, all of which are simulated with two friction coefficients (𝜇𝑐=0 and 0.39). Furthermore, the focus is on Stokes flow of neutrally buoyant non-colloidal suspensions with monodisperse spherical particles and the channel height was chosen equal to 13.5 particle diameters. DNS has been used to analyse the suspension rheology in terms of mean concentration profiles, velocity profiles, interactions in the microstructure and particle stress profiles. Steady-state concentration profiles of the simulated cases show a particle layering effect close to the confining walls. This layering effect alters the suspension rheology significantly and for that reason the wall regions are analysed separately from the core region. For both regions the microstructure, the relative viscosity and the normal particle stresses are analysed. The results agree well with existing numerical results (Gallier et al., 2016) (Yeo & Maxey, 2010a). Comparison with experimental work (Dbouk et al., 2013) (Zarraga et al., 2000) shows that results for the relative viscosities of the unlayered core regions are lower in general, reasons for this difference can be higher friction factors or the use of non-spherical particles in experiments. However, the same asymptotic trend is observed for increasing 𝜙. The maximum packing fraction 𝜙𝑚 was found by fitting the relative viscosities to the Marron & Pierce equation (Maron & Pierce, 1956), which gave 𝜙𝑚=0.69 for 𝜇𝑐=0 and 𝜙𝑚=0.635 for 𝜇𝑐=0.39. These results for 𝜙𝑚 are in good agreement with results from (Gallier et al., 2014).
In general the IBM turns out to be capable of reproducing existing numerical data. Furthermore, results have been obtained for suspensions closer to the jamming limit than known so far in numerical work on this particular flow regime. Besides that, the present results are obtained with an advanced soft-sphere collision model, including lubrication corrections for close approach of particles, which has been extensively validated with collision experiments in a previous study. Comparison of the results from this work with experimental data shows larger differences. The reason for this can be that the suspensions in experimental set-ups deviate from the idealized suspension in the simulations. Besides that experimental results differ significantly from each other, indicating that differences also exist between the experimental suspensions. These differences have to be clearly defined in order to make a valuable comparison. Therefore it is currently difficult to determine how accurately the IBM simulates suspension rheology.

A cylindrical column of fluid breaks into drops under the action of surface tension when other forces are relatively small. This phenomenon is called the capillary break-up. Since the introduction of the seminal theory of inertial-capillary jet instability by Rayleigh in the late 19th century, this phenomenon has seen a rich influx of academic research and industrial interests over the past few decades. The theory suggests that out of all the unstable wavelengths only the most unstable wavelength determines the drop size. But in reality, random background perturbations and non-linear pinch-off dynamics result in bi-modal drop size distribution. The two peaks in the bi-modal distribution correspond to the formation of main drops and satellite drops. The main drops are the desired product of the industry and the satellite drops are relatively smaller in size and are discarded by post-production processes like sieving. To reduce post-production processes and ensure repeatability, the industrial demand is to have a "narrow uni-modal" drop size distribution. Meaning, elimination of satellite drops and having uniform-sized main drops. Narrow distribution is achieved by employing finite-amplitude perturbation on the jet. This way the jet is not affected by the random background perturbation as the finite-amplitude perturbation dominates the break-up process. But this does not ensure satellite drop elimination. Uni-modality is achieved when the satellite drops merge with the main drops. Merging happens when the satellite and the main drop have different velocities and follow the same trajectory. The amplitude of the sinusoidal perturbation can be tuned to have this merging condition. The externally perturbed jet finds its application in a plethora of industrial processes such as powder production, combustion, micro-encapsulation, extreme ultraviolet rays (EUV) and, laser-plasma production (LPP), ink-jet printing, etc. The focus of this thesis is the production of crystalline fertilizer pellets from the capillary break-up of a spiralling jet of molten fertilizer. This process is called rotary prilling. Experimental studies in spiraling jets without external perturbation also shows bi-modal distribution of drops. The fine satellite drops produced in prilling process are a potential safety hazard as they can cause dust explosions which can be detrimental to air quality. Expensive filters and wet scrubbers are placed inside the prilling tower to remove the fine satellite drops. Hence uni-modal drop size distribution in prilling is highly beneficial for both economic and environmental reasons. As perturbed spiralling jets are not studied in detail in the existing literature, this thesis employs sinusoidal perturbations on spiraling water jets and aims to achieve uni-modal drop size distribution. While the straight jets have a nearly constant base flow, the spiraling jets have a base flow that accelerates due to the centrifugal and Coriolis forces in the downstream direction. As a result of mass conservation the accelerating base flow results in jet stretching in the downstream direction. The evolution of perturbations on such a stretched jet not only depends on surface tension and inertial forces of the local jet radius, but also on the wavelength that is excited initially. Hence the evolution of perturbations on straight jets with constant base flow is an eigen-value problem whereas, on a stretched jet it is an initial value problem. When the stretching rate is faster than the capillary growth rate damping of the surface perturbations will occur. For the process parameters investigated in this thesis, the capillary growth rate is faster resulting in negligible damping. The other effect of jet stretching is that the wavelengths are also stretched, that is the wavelength that is excited initially becomes longer in the downstream direction, thus the wavenumber decreases. The linear spatial instability theory for spiraling jets is experimentally validated in this thesis. For the set of non-dimensional numbers chosen in this thesis, uni-modal distribution of drops is achieved by tuning the frequency and amplitude of the sinusoidal perturbation. Uni-modality is observed when the non-dimensional wavenumber near break-up is approximately 0.7 and the amplitude is tuned accordingly. As the wavenumber decreases in downstream direction the wavenumber excited at the nozzle should be greater than 0.7.

The drainage time of liquid-filled PET bottles can be greatly reduced by pre-rotating the bottle with a certain angular velocity. Depending on the magnitude of the angular velocity, up to four different flow regimes can be distinguished during the emptying of a bottle: bubble regime, transition regime, vortex
regime and the swirl regime. At zero or low pre-rotation, the flow is in the so-called bubble regime in which the intermittent downward liquid flow is accompanied by an upward motion of irregularly shaped air bubbles into the bottle. At sufficiently high pre-rotation, the picture is completely different. After
initial transient behaviour in which the flow is first in the bubble regime, the flow undergoes transition towards a so-called vortex regime. This is characterized by regular downward motion of liquid along the bottle wall in a free-surface vortex and accompanying upward motion of air through the nozzle core. Finally, close to the end of the drainage process another transition towards a so-called swirl regime takes place, in which the last bit of liquid swirls around in the bottle before being slowed down sufficiently to exit the bottle through the nozzle opening. Dimensional analysis indicates that for a specific bottle geometry the non-dimensional total drainage time td/td,0, where td,0 is a characteristic drainage time scale for stationary low-viscosity fluids in the bubble regime, depends primarily on 3 non-dimensional numbers: (1) the rotation number, Π (2) the Morton number, Mo, and (3) the E¨otvos number, Eo. The former represents the characteristic ratio of centrifugal to hydrostatic forces inside the liquid phase. The objective of this study is to determine the relationship between the non-dimensional drainage time and Π and gain insight in the influence of E¨otvos and Morton onto this. To this purpose, a parametric CFD study of a model PET bottle has been conducted and the results have been compared with previous preliminary experiments performed in our group. The numerical study is divided into three categories, each discussing the influence of one of the dimensionless parameters. The influence of the Π-number was studied by altering the initial rotational velocity of the bottle. It was found that at some critical Π-number, the drainage time was minimal. A further increase beyond the critical Π-number resulted in a longer drainage time, due to the stronger centrifugal force acting on the liquid layer. It was found through an analytical solution and verified with the numerical results that the rate at which the Π-number grows in relation to the dimensionless drainage time is to the power 3 for the laminar case and to the power 3/2 for the turbulent case. Below the critical value, the flow is expected to remain in the so-called bubble regime, also increasing the drainage time. Furthermore, the onset of the vortex regime was expected to occur at a constant value of Π. This was also verified with the numerical results and a strong correlation was found between the onset and the corresponding local Π-number. The influence of the Morton number is also discussed, where an increase in the Morton number was realized with an increase in viscosity. It was found that the drainage time got substantially reduced with an increase in viscosity. The effects of viscous dissipation resulted in a bigger effective area and higher axial downward flow velocity. However, increasing the Morton number excessively would result in the flow regime remaining in the bubble regime. Therefore, the critical Πnumber shifts depending on the different fluid compositions. When varying the E¨otvos number, by both changing the viscosity and the surface tension, it was found that the latter had negligible effect on the generation of the air-core or the drainage time.

Sedimentation of particle suspensions is a relevant process to a variety of applications, such as the reclamation of land and chemical engineering systems. Direct numerical simulations of the settling of a single, two and multiple solid spheres under gravity in a square duct have been performed. The simulations involve 4 different Galileo numbers, i.e. Ga=144, 178, 200 and 250 for a fixed solid-to-fluid-density ratio Γ=1.5 and solid-volume fractions ranging from Φ=0.005 to Φ=0.3. It is shown that for these Ga the particle motions of a single particle are steady vertical, steady oblique, oblique oscillating and chaotic. In the simulations for a single particle it has been found that the observed fluctuations of the vertical particle velocity are due to particle-wall interactions. Drafting-kissing-tumbling (DKT), which is a key mechanism in the formation of particle clusters, has been observed in the two particles simulations and show that DKT initiates when the particles are almost aligned vertically and move in the same direction in the horizontal plane. DKT could occur even when the particles are several particle diameters apart.In the simulations of multiple particles only the steady vertical, steady oblique and oblique oscillating regimes have been considered and the particles are initially randomly distributed in the computational domain. It is shown for these regimes that for solid-volume fractions Φ≤0.01 an increase of the settling velocity of the suspensions compared to the value of a single particle is observed for all regimes, while for higher solid-volume fractions a hindered settling effect occurs. The statistics of a spatial parameter related to the angle of a particle and its nearest neighbour show that the particles align vertically more than a set of random distributed particles for solid-volume fractions Φ≤0.01. This indicates the formation of particle clusters. These findings are supported by multiple instances of DKT observed in the animations of the simulations. Finally, it has been observed for all regimes that the vertical particle velocity fluctuations relative to the settling velocity of the suspension increase roughly with Φ^(1/3).

Multi-phase flows are ubiquitous in nature and in everyday life surrounding us, impacting us in almost all possible ways. The presence of particles in a flow can change the flow behaviour in an unpredictable manner. The simplest example of a particle-laden flow, that one can think of, is the settling of a single sphere under gravity in a quiescent fluid. This seemingly simple problem has very high relevance in various practical applications ranging from sedimentation of particles for water treatment, process industries, transport of a dense suspension(slurry) through a pipe and even in land reclamation. The settling/ascension of a single sphere, even after having been subject to extensive study for more than a century, remains far from being understood completely. The path and the wake of a falling/rising sphere in a quiescent fluid may be subject to various instabilities depending upon two dimensionless quantities which are sufficient to characterize the motion. One being the Galileo number (Ga), which is the ratio of the net gravity force to the viscous force and the second one being the mass density ratio, which is the ratio of the density of the solid to the density of the fluid. Depending upon Ga and mass density ratio, the sphere can take up various regimes of motion such as vertical, oblique, zigzagging, helical to name a few. This is mainly due to wake instabilities that trigger such path instabilities. Based on Ga and mass density ratio, various regime maps have been proposed in literature. There have been several disagreements regarding the characterization of such paths taken by the sphere. This is due to the strong solid-fluid coupling and the inherent complexity due to triggering of the instabilities in such cases, which is far from being trivial to model numerically and also to test experimentally. The disagreements between different numerical works and different experimental works make the problem hard pressing and tempting to study. Moreover, the settling behaviour of a single sphere can also aid in understanding the collective effects displayed in the settling of dilute suspensions. The goal of the present study is to shed light on the confusion/disagreements in literature until now and characterize various path instabilities. A detailed experimental investigation is conducted to cover the parameter space (regime map) by employment of over 250 different combinations of Ga and mass density ratio to cover as many regimes of motions as possible within the given time framework. The motion of a sphere is tracked in time using high-speed cameras and corresponding path/regime of motion, higher-order statistics like velocity and physical characteristics such as the Strouhal number/ drag coefficient has been computed. The results validate well for some simple regimes of motion for which results from the previous studies perfectly agree with each other. With the confidence obtained after the validation, the current work attempts to draw points of consensus and disagreements with these earlier works for other more controversial regimes. Some regimes, which had only been observed using numerical simulations, have been observed experimentally for the first time. Also, intriguing bi-stable regimes (coexistence of two regimes) have been observed. Moreover, attempt is also made to characterize the suppression of the high-frequency oscillations with increase in the sphere inertia. An update of the regime maps is proposed with the results obtained from the experiments conducted. The results obtained will also serve as an excellent tool for validation of new numerical models, using which the Ga-mass density ratio parameter space can be covered in great detail. Recommendations for future work are given.

Salt intrusion through sea locks causes the mixing of fresh and salt water due to a gravity current. This process may occur gradually over tens of kilometers and affects the required quality of the inland water. A bubble screen alongside the locks is one of the available mitigating measures for this undesired phenomenon. However, current numerical simulations on bubble screens to mitigate salt intrusion are relatively scarce, and the validation of such models has not yet been thoroughly checked. The objective of this thesis is to study the performance of bubble screens for mitigating the salt intrusion using an Euler-Euler Computational Fluid Dynamics (CFD) model in Fluent 17.2. The work was conducted at TU Delft in collaboration with Deltares, a hydraulics institute in Delft in The Netherlands. First, a simulation of a gravity current is conducted to study the mixing of fresh and salt water in the absence of a bubble screen. The mass transport equation for the salt concentration, together with a linear approximation of the concentration-density relation, is involved into the governing equations. The results are validated with empirical predictions and experiments at Deltares of gravity currents to study the buoyancy effects of different turbulence submodels. Second, simulations of bubble screens in the fresh-fresh water system are performed and validated with particle image velocimetry (PIV) measurements conducted at Deltares to investigate the bubble screen dynamics. Larger flow circulations are generated with increasing air flow rates, while a constant surface current thickness of around 0.3 times the water depth is found for all cases. Lastly, bubble screens simulations for six different Froude air numbers $Fr_{air}$ in the fresh-salt water system are conducted and validated with the dye measurements. The dimensionless number $Fr_{air}$ serves as a ratio of the kinetic energy of the rising bubble plume to the potential energy of the gravity current. The salt transmission factor, defined as a ratio of the salt intrusion with mitigating measures to that without any measures, shows that $Fr_{air}$ in the range of 0.93 - 1.08 is most efficient for mitigating salt intrusion. At a lower $Fr_{air}$, the salt water generally intrudes at the bottom area as a salt tongue, whereas the salt intrudes through the surface current for larger $Fr_{air}$ owing to a higher rate of liquid entrainment into the bubble screen.

This work focuses on creating a recipe for parametrizing flow around assemblies of non-spherical particles. A multi-relaxation time lattice Boltzmann method (MRT-LBM) is used to simulate the flow. The research focuses on 3 different developments. First, different boundary conditions available in the literature for LBM are tested to identify the best for the flow problem. The second part of the thesis focuses on developing more widely applicable scaling laws for drag and lift of various isolated non-spherical particles. In the third part, a recipe to describe hydrodynamic forces on assemblies of axisymmetric, non-spherical particles is proposed. With the described parameters, drag, lift and torque correlations are proposed accordingly. This research is funded by the European Research Council under its consolidator grant scheme, contract no. 615096 (NonSphereFlow).@en

Manganese nodules have been lying on the deep sea soil for centuries before they became known. The first ones to discover the nodules were Murray and Renard, during their scientific cruise between 1872 and 1876 (Murray and Renard, 1876). The Clarion-Clipperton Zone, located between Mexico and Hawaii, appears to have a virtually inexhaustible amount of mineral resources. The manganese nodules consist mainly of man- ganese, but because they also contain metals such as copper and cobalt, the nodules become more and more interesting for our modern society. Another name for the manganese nodules is polymetallic nodules, be- cause of the content of many other metals.
By means of a vertical hydraulic transport system the manganese nodules will be transported upwards from the seabed. Given that the nodules are located at a depth of about 5000 meter, it is plausible that they come into contact with the pipe wall during transport. When a nodule interacts with the wall, abrasive wear can occur. Abrasive wear, or sliding wear, causes the nodule to degrade, or decrease in mass, through the sliding contact with the pipe wall.
The main research question of this Master Thesis is formed as follows:
What is the contribution of abrasive wear to the degradation of polymetallic nodules, in a vertical transport system?
The contribution of abrasive wear is tested on the basis of the equation of Archard (1953). In the MTI laboratory in Kinderdijk various manganese nodules are tested in a specially designed experimental setup. The polymetallic nodules available for this Master Thesis originate from the Clarion-Clipperton Zone (CCZ). The first set of experiments is focused on single nodule degradation. The nodules are weighted before and after a test run, hereafter the nodule mass after a test run is divided by the initial nodule mass, resulting in the normalized remaining mass. Each test run consists of a specific sliding length. With the sliding length and normalized remains mass, a prediction can be made of the material specific wear rate factor, noted as k [m2/N]. The experiments resulted into a value of k = 4.16·10−9 [m2/N]. With this k-factor, in the worst case, 97% of the original nodule mass is remaining after 1000 meters of sliding contact.
Subsequently, a batch of nodules is tested in the experimental setup, with the expectation that the con- tribution of abrasive wear increases to the total nodule degradation, due to the presence of particle-particle interaction. Recently, Blue Mining has conducted trials in Freiberg, Germany, in an abandoned mine shaft. A vertical pipe of almost 130 meters has been placed inside, and the effect of abrasion on a batch of nodules has been observed. During the batch experiments in Kinderdijk, where the batch of nodules was in a horizontal position, it is seen that the contribution of abrasion is greater, compared to the tests in Freiberg.
Using the open-source program OpenFOAM, a vertical turbulent pipe flow is modeled. The pipe has simi- lar dimensions to the pipe of the experimental setup. The simulation was done using a Large Eddy Simulation (LES). By means of a one-way coupling, the data of the flow field is linked to the movement of the simulated particle. Subsequently, point particles were released at various points at the entrance of the pipe. The first series of point particles consists of a diameter of one-third of the pipe diameter, corresponding to the com- mon size of a manganese nodule. The collisions between the point particle and the pipe wall are solved with a Discrete Element Method (DEM), which makes use of a (non-linear) spring-damper system.
To test the sensitivity of the model, point particles of one-sixth and one-tenth of the pipe diameter are also released. The number of collisions per total number of simulations are compared with each other. After the sensitivity analysis it turned out that in 15% of the simulations there was contact with the pipe wall, regardless of the diameter of the particle. When this 15% is compared to the experimental results using single nodules, there is a weight loss of around 0.5%. This holds a weight loss of only a single - to a few - gram(s) for the average polymetallic nodule.
It can be concluded, according to the experimental trials with polymetallic nodules from the Clarion- Clipperton Zone and by means of a one-way coupled CFD-DEM model, that the contribution of abrasive wear results into a (maximum) remaining mass of 97%. However, it is advised to extend the computer model with a two-way (or multi-way) coupling between the liquid and the solids. It is also recommended to conduct vertical abrasive wear tests with nodules, in order to validate the model and gain more insight in degradation behaviour.

Using the respiratory system as a route to deliver medication has many advantages, such as the large surface area of the respiratory system and the ease of absorption via the lung tissue. Nebulizer technology uses the respiratory system to administer a wide variety of medication. Nebulizer technology creates microscale droplets that can be inhaled by the patient. Different types of nebulizer technology exists, and the vibrating-mesh nebulizer is one of them. A mesh with over a thousand nozzles, each with a diameter of 4.3 micrometer, vibrates against the liquid reservoir with a frequency of 128 kHz. This technology has already been proven to work, except for the use of one particular liquid, pure water.
The relation between the concentration of salt in a solution and the performance of the vibrating-mesh nebulizer was researched by Beck-Broichsitter et al. (2014). They found that a significant drop in performance takes place when the salt concentration of the solution decreases from 1 mM sodium-chloride to 0.01 mM sodium-chloride.
This research first investigated the theory on the properties of these solutions, and how these properties change when the concentration decreases from 1 mM sodium-chloride to 0.01 mM sodium-chloride. It was found that the hydrodynamic properties of the fluid do not change significantly over this range, but the electrostatic and -kinetic properties do change significantly over this range. This led to the hypothesis that the electrostatic and -kinetic effects have an influence on the formation of droplets and therefore the performance of the nebulizer. Three different hypotheses were contrived on how these effects influence the formation of droplets. The first hypothesis was that charged induced wetting causes the liquid from the nozzle to spread out over the surface at the end of the nozzle, instead of forming a droplet. The second hypothesis was that due to electrostatic forces in the neck of the droplet formation process, the neck is prevented to decrease in diameter, no pinch-off takes place and therefore no droplet is formed. The third hypothesis was that a droplet was formed, taking with it an electrostatic charge. The next droplet is then prevented from exiting the nozzle, by the electrostatic force the first droplet exerts on it. These hypotheses where based on engineering intuition, rather than experimental data.
Two different types of experiments were conducted to test the hypotheses. First the mass flow rate of the nebulizer was measured using different electrolyte solutions. Second, images where recorded, with various imaging speeds and magnifications, of the nebulizer while it was in working.
The results of the first type of experiments indicated that there is indeed a relation between the electrostatic and -kinetic properties of the fluid and the performance of the nebulizer. The results from the second type of experiments showed that the first hypothesis, the charged induced wetting hypothesis, was false. The results did not prove the other hypotheses to be false or true, but rather led to a new hypothesis. The new hypothesis was that the increased viscosity or decreased diameter of the nozzle, due to the electric double layer, bring down the mass flow through the nozzles or causes a portion of the nozzles to be inactive. As a prove of concept, an experiment was conducted where the viscosity of the liquid was decreased by increasing the temperature. This experiment gave the indication that the new hypothesis can be true.

The freeze plug is a key safety component of the molten salt fast reactor (MSFR), one of six next-generation nuclear reactor technologies being developed under the Generation IV International Forum (GIF). It should be designed to melt if an accident occurs, allowing the MSFR to drain before it incurs structural damage.
Two freeze plug concepts have been considered in recent years, in which the plug is melted either through the decay heat produced in the core, or through heat generated by special heating rings and stored in steel blocks adjacent to the freeze plug. Variations consisting of both a single freeze plug, and multiple smaller plugs contained in a metal plate, have been proposed. This work seeks to evaluate the feasibility of these designs and study how parameters such as the sub-cooling of the plug affect melting times. Additionally, an alternative, wedge-shaped freeze plug design is proposed for increased reliability. 
Simulations performed in COMSOL showed that the decay heat plug melts within 600 s only if placed within 0.01 m of the mixed core flow. Because such a placement makes the plug vulnerable to temperature and velocity fluctuations in the core during regular operation of the reactor, this design is considered unfeasible and is not recommended for further study. On the other hand, melting times under 600 s were possible with the heating ring design for a range of sub-cooling amounts and plug configurations, suggesting that this
design is promising. A thin frozen layer was shown to form on top of the metal grate in the multi-plug configurations, preventing heat transfer through the top of the plate. Although the melting behavior of this layer warrants further investigation, its insulating effect was found to generally cause the single-plug designs to melt faster than the multi-plug designs.

A simplified, isothermal model of the wedge-shaped plug was simulated using the enthalpy-porosity approach to account for convection. To model the sinking of the wedge, an extended Darcy term approach was developed based on an analytical solution which was validated experimentally, with good agreement. This model shows that melting of the wedge is unsteady, and that melting times depend linearly on the wedge angle and sub-cooling. Unfortunately, melting times of the wedge plug could not be estimated with realistic, non-isothermal, time-dependent boundary conditions. For future study, a customizable numerical solver such as OpenFOAM is recommended, which would allow the sinking of the solid phase to be modeled more robustly through an immersed boundary method.

Oblique impact of droplet onto a deep liquid pool is frequently observed in nature. As for engineering applications, the subsequent phenomenon of oblique droplet impact is of great significance for a proper design of the vanes in the extreme ultraviolet (EUV) chamber of ASML to prevent contamination. During the impact, the pool surface deforms and a cavity is generated, which later on collapses into a jet that shoots upward. Meanwhile, a liquid crown will be visible above the surface. When this crown becomes unstable, splashing occurs. We study these phenomena numerically by using fully three dimensional simulation in Basilisk C, an open source CFD package, coupled with an adaptive Cartesian grid and volume-of-fluid (VOF) method. The simulation is validated by both qualitative and quantitative comparison with the available experimental results and theory. In both simulation and experiment, three types of impact phenomena are identified: smooth coalescence of droplet with the pool, splashing in the impact direction only, and splashing in all directions. We quantify these three regimes by accounting for the Weber number and the droplet impingement angle, and compare the results with the experimental data and scaling argument proposed in the previous literature. We also study cavity dynamics and shape. The growth of the cavity angle is captured by our simulation. In analogy to the perpendicular drop impact, we give a model for the cavity expansion, from which we obtain a temporal function for the cavity depth which is in agreement with the simulation. The magnitude of the cavity depth and the displacement match with the available experimental data. Besides, our simulation captures the scaling dependence of both the depth and the displacement on the Weber number. We further point out the connection between the descent of the crown and the initiation the cavity collapse. The three-dimensional details about the crown behavior and cavity formation from our simulation provide an insight of the phenomena which could cause contamination in the EUV chamber, and offer an opportunity for a further research on this subject.

Heat transfer is important in many applications. For instance, due to the decrease in size of electronics, it
becomes more necessary to have efficient and smaller cooling systems. In order to increase the effect of the
cooling liquids used, it might be interesting to add extra solid particles with a high conductivity, to possibly
increase the effective heat transfer fromthe wall to the liquid. These particles can be separated in 2 categories:
point particles, with a very small size compared to the flow phenomena, and finite sized particles, which due
to their larger size are able to significantly influence the fluid flow. In this thesis, the finite sized particles
and their effect on the effective conductivity have been analyzed by using a CFD code. The main focus has
been on the effect of the mechanical and thermal Stokes numbers, which give an indication about the time
required for the particles to react to changes in surrounding flow and temperature compared to the relevant
flow time scale.
To investigate the effect of large particles on the effective conductivity of a fluid a numerical method to
solve heat transfer inside a fluid and between fluid and particles has been implemented. This method, based
on an immersed boundary method combined with DNS, is able to solve both isolated particles and extremely
high conductivity particles. To solve the heat transfer for finite conductivities, a volume of fluid method has
also been implemented. These methods have been verified by comparing the simulation results to known
results for single sphere heat transfer and conservation of energy.
With this code, the influence of the thermal and mechanical Stokes numbers have been analyzed for
laminar Couette flow. In order to gain a better understanding of the underlying heat transfer mechanics, it
has been assumed that the natural convection is negligible and the density ratio between the particle and the
fluid is taken to be equal to one (no effects of gravity). From this it appeared that the effective conductivity
of a suspension can be split in 3 components: the non-moving conductivity, an enhancement due to fluid
convection induced by the particles and an increase in heat transfer due to particle convective heat transfer.
The non-moving conductivity is only dependent on the conductivity of the particles and the fluid, and
on the particle concentration. It stays close to constant independent of Stokes numbers. In contrast, the
fluid convection appeared to scale significantly with the mechanical Stokes number and with the particle
concentration. This appeared to be due to the increase in particle inertia resulting in more movement in
wall normal direction and as a result moving more fluid in wall normal direction. The particle convection
appeared to not only scale with the thermal Stokes number and the particle concentration, but also with the
mechanical Stokes number. This increase was caused by the particles being able to absorb and release more
thermal energy for higher thermal Stokes numbers, and thus transport more heat from the hot wall to the
cold wall.
Finally, the resulting effective conductivity and effective viscosity of the suspension were compared and
it was shown that it is possible to enhance the heat transfer more than the viscosity, but only by either introducing
a small amount of highly conductive particles, or by introducing well-conducting particles with very
low mechanical Stokes numbers. It appeared to not be possible to increase the heat conductivity more than
the viscosity for particles with equal or lower conductivity compared to the fluid.