A water electrolyzer is an electrochemical device that produces hydrogen and oxygen from water via electricity. A sustainable way of producing hydrogen with renewable electricity, also known as green hydrogen, could enable its use in decarbonisation of sectors that depend on fossil fuels. However, to scale up the production of green hydrogen, larger water electrolyzers are required, which presents challenges. Water electrolyzers produce hydrogen in the form of bubbles, which are transported away through outlet channels. The geometry of the outlet channel is important for performance and safety. A channel that is too wide can result in unwanted ionic current leakage, reducing efficiency and increasing the risk of explosion. To reduce the leakage current, the electrical resistance in the channel can be increased to make it difficult for ions to flow through. However, a narrow channel can increase the hydraulic resistance, leading to clogging due to bubble build-up. This study focuses on the characterization of the slug flow regimes that eventually lead to clogging of the outlet channel in a novel electrolyzer design. To characterize the flow when blockage develops, electrochemical tests with industrially relevant 6M KOH were carried out at room temperature and ambient pressure in conjunction with high-speed video recording. The examined channel was 2 mm high, 3 mm wide, and 40 mm long. The experiments showed that clogging occurs for certain flow regimes. The hydraulic configuration caused clogging at the hydrogen side, which caused the liquid flow prefer to travel via the anode rather than the cathode, which was reflected in the temperature plots. Clogging did not occur when the superficial gas velocity was greater than 0.3 cm/s, which corresponds to a current density of 140 mA/cm2. The clogging was also not visible at a superficial liquid velocity greater than 0.25 cm/s. Below these values, clogging is expected to take place for the investigated channel dimensions. The bubble diameters in the bubble layer in front of the channel were measured with a microscopic camera. The mean bubble diameter was 0.4 mm, which is 80% smaller than the height of the channel through which they flow. However, exceptional bubble sizes that were 3 to 4 times the channel height, which ultimately led to clogging of the channel, were also present. The bubble layer thickness and velocity at increasing liquid flow rates and current densities were measured using a high-speed camera. At low current densities, the velocity and thickness of the bubble layer were not affected by the liquid flow rate. At higher current densities, a higher liquid flow rate appeared to reduce the thickness of the bubble layer, but not the velocity of the bubble layer. The bubble layer thickness as well as the velocity strongly depend on the applied current density (and thus the gas flow velocity). This knowledge might be, when the flow regimes are scaled to larger electrolyzer dimensions accordingly, essential for future electrolyzer manufacturers and operators for the operation and design of larger-scale electrolyzers. It will give insight into which flow regimes should be avoided in order to avoid clogging and thus explosion risks during operation.

In the non-dissipative regime, the potential energy is the difference between the strain energy of the deforming solid and the work done by the external forces. For configuration-dependent external forces, whose direction is perpendicular to the deformed shape, we obtain a simple formula for the strain energy release rate of peeled strips experiencing large deformations and prove rigorously that the same formula applies for external forces having fixed direction. We then apply Griffith's criterion for fracture to calculate critical loads for two cases: peeling produced by a uniform follower pressure distributed along the flexible strip and peeling produced by a localized follower shear force applied at the edge of the strip. We found that for these loads, the critical pressure for peeling follows approximately q_c∼ΓL⁻¹, where Γ is the solid–solid interface energy and L is the initial peeling length; for the shear force, the corresponding critical value instead follows Q_0c∼Γ, independently of the initial length. These formulas are, unexpectedly, independent of the bending stiffness EI of the strips and differ from the ones predicted for small deformations, i.e. q_c∝L⁻²EIΓ and Q_0c∝L⁻¹EIΓ. We apply our results to predict the critical hydrodynamic load necessary to exfoliate graphene sheets from graphite, a fluid–structure interaction problem where the load is of the follower type. We find that a follower load peeling model gives significantly improved predictions than fixed load peeling. For the same Γ, L and b, the critical hydrodynamic follower load is always lower than the one with fixed forces: approximately half for the case with uniform pressure, and one third for the case with shear force.

@en

This work describes a modelling approach to SARS-CoV-2 dispersion based on experiments. The main goal is the development of an application integrated in Ansys Fluent to enable computational fluid dynamics (CFD) users to set up, in a relatively short time, complex simulations of virion-laden droplet dispersion for calculating the probability of SARS-CoV-2 infection in real life scenarios. The software application, referred to as TU Delft COVID-app, includes the modelling of human expiratory activities, unsteady and turbulent convection, droplet evaporation and thermal coupling. Data describing human expiratory activities have been obtained from selected studies involving measurements of the expelled droplets and the air flow during coughing, sneezing and breathing. Particle Image Velocimetry (PIV) measurements of the transient air flow expelled by a person while reciting a speech have been conducted with and without a surgical mask. The instantaneous velocity fields from PIV are used to determine the velocity flow rates used in the numerical simulations, while the average velocity fields are used for validation. Furthermore, the effect of surgical masks and N95 respirators on particle filtration and the probability of SARS-CoV-2 infection from a dose-response model have also been implemented in the application. Finally, the work includes a case-study of SARS-CoV-2 infection risk analysis during a conversation across a dining/meeting table that demonstrates the capability of the newly developed application.

@en

In liquid-based material processing, hydrodynamic forces are known to produce severe bending deformations of two-dimensional (2D) materials such as graphene. The non-linear rotational and deformation dynamics of these atomically thin sheets is extremely sensitive to hydrodynamic particle-particle interactions. To investigate this problem, we developed a computational model of the flow dynamics of elastic sheets suspended in a linear shear flow, solving the full fluid-solid coupling problem in the two-dimensional, slender-body, Stokes flow regime. Both single and pairs of sheets in close proximity are analyzed. Despite the model being two-dimensional, the critical non-dimensional shear rate yielding single-particle buckling is comparable in order of magnitude to that reported for fully three-dimensional, disk-like sheets. For pairs of interacting sheets, hydrodynamic interactions lead either to parallel sliding or bending, depending on the value of an elasto-viscous number based on particle length. For sufficiently low bending rigidity or large shear rates, large deformations of initially stacked sheets lead to sheet reattachment after separation, unlike for the rigid case. A peeling-like dynamics where lubrication provides a viscous bonding force is observed for sheet pairs when one of the two sheets is more rigid than the other. Practical implications for graphene processing and exfoliation are discussed.

@en

A magnetorheological fluid is a type of smart fluid that can change its rheological properties when a magnetic field is applied to the fluid. The fluid is a suspension of magnetizable particles in a non-magnetic carrier fluid. When a magnetic field is applied to the fluid, the particles will form structures along the magnetic field lines. These structures resist flow, thereby increasing the viscosity of the fluid. In a non-uniform magnetic field that is generated by a permanent magnet, the particles separate from the fluid and aggregate on the surface of the magnet. This phenomenon is of use in hydrodynamic bearings, where textures are used to increase the load capacity of the bearing. Replacing the fluid in the bearing by a magnetorheological fluid and placing permanent magnets at the desired texture locations, results in particles separating from the fluid, aggregating on the permanent magnets and forming textures. These textures are of a self-healing nature, because any particle that is sheared off, returns into the fluid and is replaced by another particle. Currently, not much is known about the formation of these textures and the influence of their formation on the rheological behavior of the magnetorheological fluid. Therefore, several discrete element models are constructed, that can simulate a magnetorheological fluid in non-uniform field. These models use basic physical laws to determine the dynamics of the particles. A single core model is constructed to simulate the behavior of the magnetorheological fluid in small domains using two different methods for the magnetic interaction forces between the particles. A parallel model is made to simulate the magnetorheological fluid in larger domains. Finally, a model that is used to simulate particulate flows, is modified to research whether two-way coupling is required to determine the steady state behavior of the magnetorheological fluid. The models show that the particles do not separate from the fluid due to the attractive force of the magnet by itself, but rather by a combination of the attractive force and the deformation of the structures when a flow is applied. Furthermore, the rheological behavior of the fluid can still be approximated using the standard viscoplastic models. Finally, the modified particulate flow model showed that two-way coupling is not required to determine the steady state behavior of the magnetorheological fluid.

Water electrolysis is a popular energy storage technique used in tandem with many renewable sources to convert the generated energy into storeable hydrogen. The efficiency of water electrolyzers is greatly affected by overpotential losses. Bubble evolution is a unique characteristic of flows in water electrolysis. The evolution of bubbles alter electrokinetics close to the electrodes and vary ionic mass transport. The localized flow features close to a bubble are often attributed to the variation in the current density and subsequently, the electrolyzer efficiency. For the purpose of modelling flows close to a bubble better, a Lagrangian method for the simulation of passive tracers is developed and programmed to be coupled to the flow field of Bluebottle, an open-source particulate multiphase flow solver that uses the Physalis algorithm.
The dynamics of the tracers are modelled using a simplified Langevin equation. In the present work, the migration flux is omitted and priority is given to convection and diffusion with the objective of establishing a foundation for the simulation of ionic mass transport. Brownian motion is described using a random displacement term. The coupling with the flow field is achieved using trilinear interpolation. The domain boundaries in regards to tracer dynamics are modelled as either a rigid wall pair or as a periodic boundary pair. Specular reflection is programmed for the former ensuring elastic collision of a tracer with the domain boundary. For the latter, the tracer position is altered so as to place the tracer in the opposite side of the domain in the axis of intrusion. Particles are assumed to be non-penetrative and hence, specular reflection is implemented at the surface of each particle. Since the tracer module is coupled one-way with the flow field and executed after a Bluebottle time-step, a subroutine is developed to push the tracers out of a particle radially if a tracer is located inside a particle after a Bluebottle time-step. To ensure particle interaction is ensured across periodic boundaries, a subroutine is developed that places the particle in an apparent location that enables particle-tracer interaction. 
The module execution time is found to be linearly proportional to the number of particles and the number of tracers and consumes roughly 10% of a Bluebottle iteration execution time in nominal tests. The module is tested to ensure the Brownian displacement term obeys diffusion statistics and also to ensure that the trilinear interpolation works as intended. The numerically enforced no-penetration boundary at particle surfaces is also tested and observed to prevent intrusion of tracers. 
The tracer module is then used to stochastically simulate mass transport across a particulate suspension in a stagnant and a sheared flow field. The Sherwood number Sh determined from the tracer module is found to agree well with the expected experimental and numerical results of Wang et al. (2009). The tracer module is also compared to a scalar field solver of Bluebottle. The tracer module is observed to capture features of the flow field quite well. However, the transient tracer positions upon conversion to a transient continuous concentration field exhibits noise due to the discrete nature of the tracers. Hence, transient comparisons with a continuous field in terms of absolute magnitude requires a large number of tracers.
Recommendations for improvement of the code is provided. The present work is intended to be followed up with the addition of migration flux to the equations of motion for the tracers through the solution of an additional equation for velocity of the tracer using a force equivalence of Coulomb's law and Stokes' drag law. Future challenges that will be encountered in the development of an accurate ionic mass transport solver is briefly discussed.

The decarbonization of the transport sector is a major challenge for the transition towards a sustainable economy. Amongst the different measures to lower the impact of road transportation on the environment is the use of biofuels, especially for heavy duty vehicles. For biofuels to be affordable and well integrated, improvements in the production process in terms of energy performance and economic feasibility need to be made.

Consequently, this thesis aims at the techno-economic assessment of the integration of pyrolysis oil gasification with electrolysis and syngas catalytic upgrading. Oxygen, being a co-product of electrolysis, is used as an oxidizer for entrained flow gasification of biomass derived pyrolysis oil. Hydrogen from the electrolyzer is used to enhance the production of biofuel. The process was designed to have a high-temperature solid oxide electrolysis cell (SOEC) and the biofuel was chosen to be compressed natural gas (bio-CNG). The modeling of the process was performed on Aspen Plus software and calculations of parameters for the gasifier and electrolyzer were done using Fortran and Matlab respectively.
As aimed by the integration of electrolysis, the hydrogen produced is added to the syngas to obtain a feed ratio of 3, required for an optimum methane production. The increase in the feed ratio at the level of the gasifier is mainly limited by the oxygen equivalence ratio (OER), considering the extent of combustion reactions and the operating temperature. The amount of steam inputted was found to increase the molar ratio of H2 to CO without significant change in the feed ratio due to the increase in the concentration of CO2. However, steam was still necessary for adjusting the temperature of the gasifier to values that are in accordance with experimental ones in literature. The obtained final product stream of bio-CNG at 250 bar and 15℃ has a purity of 98.63 mol% and 99.15 wt%.

Moreover, by applying heat integration to the process, the resulting surplus heat was 10.71 MW, which was assumed to be used to generate steam that is inputted to the SOEC for producing hydrogen as an additional product. The process was found to have an overall efficiency of 73.8% (LHV basis) and 78.2% (HHV basis). Heat integration increased the process efficiency by 15% (LHV basis) and 23% (HHV basis), by removing to a great extent the need for external heat and by having hydrogen as an energy output added to CNG.
As a complement to the energy performance, the economic feasibility of the process was assessed. The total capital investment was found to be 889.5 Euros/kWinput, the net present value was 6.0342 million Euros and the rate of return on investment 30.02%.

Biochar for horticultural and agricultural applications using high temperature torrefaction technology Pradeep Ravi Supervisors: Prof Dr. D.J.E.M Roekaerts, ir.Bart de Vries, Dr. Luis Cutz, Dr.Lorenzo Botto & Dr.Ralph Lindeboom Biomass currently accounts for less than 10 percentage of the world’s renewable energy production. Currently the major global sustainability issue stems from the sourcing of virgin wood chips from dense forests for pellet production. An alternative is to use residual biomass from agriculture or forestry, which is produced in large volumes, to produce different products that range from biofuels to chemicals via thermochemical conversion technologies. Among thermochemical technologies, torrefaction is a promising route to produce solid biofuel known as biochar. With an increasing potential for biomass production coupled with an increased scrutiny on the use of biomass as a green fuel, the need for alternative clean applications for the biochar is critical. The aim of this study is to investigate new and novel agricultural residues or other waste streams to produce biochar using high temperature (350 °C) torrefaction technology. The obtained biochar is evaluated experimentally to determine the best feedstocks out of the ones that are selected from a performance and cost point of view for horticultural applications. . This research aims to provide a clear and useful analytical tool which will benefit the scientific community to select suitable biomass materials based on material properties and end applications. The efficacy for the various torrefied biomass feedstocks on the soil and its stability are tested. Overall, about 50 different biomass feedstocks were identified and evaluated based on past performances from literature. The top 10 best performing feedstocks were sourced and subjected to various physical and chemical characterization tests with a specific focus on soil remediation. The selected materials were torrefied in a fixed bed pilot reactor Scanning Electron Microscopy with Energy Dispersive X-Ray Spectroscopy (SEM-EDS), Brunauer–Emmett–Teller (BET) and pH measurements. Ultimately the feedstocks were scored and ranked from best to worst performing biochar for soil remediation and sequestration-based applications. The results of this project indicate potential for biochar production from woody, grassy and other processed materials that could help to remove the dependence on evergreen forests and wood chips. The system proposed in this work could also yield negative emissions since the feedstocks are residual flows and the biochar is going to be used in the soil.

Rigid platelike particles displaying interfacial slip can attain a constant orientation in a shear flow when the slip length is sufficiently large. But actual thin particles such as single-layer graphene are flexible and prone to bending deformations when exposed to shear stress. To study the effect of bending deformation on the dynamics of flexible platelike particles with large interfacial slip in a shear flow, we develop a two-dimensional (2D) fluid-structure interaction model. Our model is based on coupling the Euler-Bernoulli beam equation with a boundary integral method to solve the hydrodynamic stress at the particle surface. Emphasis is placed on resolving accurately the stress distribution at the edges of the particle. We find that (i) a stable alignment occurs even for relatively flexible particles and that (ii) edges effects on the shape of the plate are important for values of the length-to-thickness aspect ratio as large as 100. Our results are particularly relevant in view of recent research on the hydrodynamics of suspended flexible sheets made of 2D nanomaterials.

@en

Solid particles can adhere to fluid interfaces, modifying interfacial properties such as the surface tension and the surface elasticity. We here describe a new simulation method, the Fast Interface Particle Interaction (FIPI) method, capable of simulating on commodity hardware up to 100k particles interacting with fluid interfaces of complex morphology. The method is based on resolving interfacial and hydrodynamic length scales larger than the particle, and model particle-level phenomena with physically consistent algebraic models. For its ability to simulate large numbers of potentially very small particles, FIPI represents a useful tool for investigating phenomena of particle adsorption/desorption and particle-induced surface tension modification in problems of froth flotation, drop/foam stabilisation, Bijelformation and other multiphase systems. Program summary: Program Title: FIPI Licensing provisions: GPLv2 Programming language: C++ Nature of problem: Simulation of particle-laden fluid interfaces (3-phase flows) Solution method: Spectral method Running time: Up to dozens of hours on a single processor@en

We present a micromechanical analysis of flow-induced peeling of a layered 2D material suspended in a liquid, for the first time accounting for realistic hydrodynamic loads. In our model, fluid forces trigger a fracture of the inter-layer interface by lifting a flexible “flap” of nanomaterial from the surface of a suspended microparticle. We show that the so far ignored dependence of the hydrodynamic load on the wedge angle produces a transition in the curve relating the critical fluid shear rate for peeling to the non-dimensional adhesion energy. For intermediate values of the non-dimensional adhesion energy, the critical shear rate saturates, yielding critical shear rate values that are drastically smaller than those predicted by a constant load assumption. Our results highlight the importance of accounting for realistic hydrodynamic loads in fracture mechanics models of liquid-phase exfoliation.@en

Liquid-phase exfoliation, the use of a sheared liquid to delaminate graphite into few-layer graphene, is a promising technique for the large-scale production of graphene. However, the microscale and nanoscale fluid-structure processes controlling the exfoliation are not fully understood. Here, we perform non-equilibrium molecular dynamics simulations of a defect-free graphite nanoplatelet suspended in a shear flow and measure the critical shear rate γ̇ c needed for the exfoliation to occur. We compare γ̇ c for different solvents, including water and N-methyl-pyrrolidone, and nanoplatelets of different lengths. Using a theoretical model based on a balance between the work done by viscous shearing forces and the change in interfacial energies upon layer sliding, we are able to predict the critical shear rates γ̇ c measured in simulations. We find that an accurate prediction of the exfoliation of short graphite nanoplatelets is possible only if both hydrodynamic slip and the fluid forces on the graphene edges are considered and if an accurate value of the solid–liquid surface energy is used. The commonly used “geometric-mean” approximation for the solid–liquid energy leads to grossly incorrect predictions.@en

In liquid-phase exfoliation for the production of 2D nanomaterials fluid forces are used to gently overcome adhesive interlayer forces, leading to single- or few-layer 2D nanomaterials. Predicting accurately the critical fluid shear rate for exfoliation is a crucial challenge. By combining notions of fluid mechanics and fracture mechanics, we analyze a mathematical model of exfoliation, focusing on the π − peel regime in which bending forces are much smaller than the applied hydrodynamic forces. We find that in this regime the shear rate is approximately proportional to the adhesion energy, independent of the bending rigidity of the exfoliated sheet, and inversely proportional to the size a of a (assumed pre-existing) material flaw. The model appears to give values comparable to those obtained in wet ball milling, but to overestimate the shear rate values reported for turbulent exfoliation (by rotor mixing or microfluidization). We suggest that for turbulent exfoliation a “cleavage model” may be more appropriate, as it gives a stronger dependence on a and smaller critical shear rates.

@en