De-carbonizing aviation is necessary for a sustainable future, and using hydrogen in a fuel cell, that produces water, can greatly reduce greenhouse gas emissions. Achieving higher gravimetric energy density with hydrogen compared to conventional jet fuels, involves storing it in cryogenic liquid form. Along with it, its cryogenic temperature range not only enables its use as chemical energy storage but also as a potential heat sink. However, rapid vaporization, known as boil-off, limits the long-term storage of liquid hydrogen in fuel tanks, requiring regular venting due to self-pressurization over time. Additionally, hydrogen needs to be heated to the fuel cell’s operating temperature before being used as a reactant. Understanding these requirements, this thesis focuses on three areas: predicting the maximum boil-off rate of liquid hydrogen while charging the fuel tank using a MATLAB simulation and the boil-off rate during the flight journey using a validated software called BoilFAST to understand the feasibility of retrieving the boil-off for its integration with the fuel supply; designing a fuel supply system that integrates the boil-off gas with the vaporized liquid hydrogen supply line to the fuel cell system; and integrating the cryogenic energy of hydrogen with a ram air-cooled vapor compression refrigeration system (VCRS) based thermal management of fuel cell, with the intention of reducing its parasitic load and improving system compactness. Two methods were used for the thermal integration: VCRS involving fuel cooled heat exchangers that function as an intercooler and as a separate de-superheater before a ram air-cooled condenser; and VCRS with a separate single-phase, 52\% ethylene-glycol based serial cooling circuit with multiple fuel cooled heat exchangers (FCHX). This resulted in a significant reduction of parasitic load by 13.4% and 26% when integrated with the intercooler system and single-phase serial cooling system, respectively. The study also examined the expected additional component weight, considering the aviation sector's preference for lighter systems. The findings demonstrate that the holding time of the fuel for minimum 13 minutes after tank filling and before the start of the propulsion system unit can allow a controlled amount of boil-off gas to be integrated with the fuel supply. Utilizing cryogenic energy for thermal management can significantly enhance the system's coefficient of performance by 15.3% and 33.3% respectively. Future work should involve experiments to obtain actual boil-off rates at different ambient exposures of fuel tank, tests on sloshing effect due to turbulence during flight journeys, analysis of thermal stress effects in cryogenic heat exchangers due to high temperature gradients, and testing new compatible mixed refrigerants with improved thermal properties for optimum cryogenic heat exchange.

Steel manufacturing is a carbon intensive process, that is responsible for approximately 7% of the total global CO2 emissions. Therefore, TATA Steel IJmuiden aims to lower its carbon emissions. One way to bring down emissions, is to replace the existing blast furnace (BF) CO reduction process with the H2-based direct reduction of iron ore (DRI). The two most widely applied H2-DRI processes around the world are the low pressure MIDREX, abbreviated as MLP, and medium pressure HYL-Energiron, abbreviated as EMP. Since TATA Steel IJmuiden wants to study the switch from BF to H2-DRI steelmaking, it is relevant to gain insight into which gas phase reactions are dominant for both processes, into the direct reduction process itself and into the behaviour of the carburization reactions that improve the steel quality. In the gas phase reactions, it was seen that for the MLP process in situ reforming of natural gas can be a viable option before switching to a 100% H2 process. This may prove to be worthwhile in the early stages of H2-DRI steel production, when green H2 is still scarce and expensive. For EMP, internal reforming seems less of a possibility due to the high reaction rate for the reverse water gas-shift. When comparing both MLP and EMP, reaction rates are generally higher for EMP than for MLP and hence smaller reactor volumes are required for the EMP process to acquire the same amount of output. Direct reduction with H2 has a higher reaction rate than reduction with CO, while for the carburization reactions methane cracking was found to be the dominant reaction. Techno-economic scenarios for 100% H2-based DRI in which green H2 is imported are only feasible when H2-prices fall below €1.80/kg. Meanwhile, a scenario with an on-site electrolyzer powered by grey grid electricity only proves to be worthwhile for electricity prices lower than €20/MWh. The most promising techno-economic scenario, which includes an on-site electrolyzer and the construction of a wind farm just off the coast from the TATA Steel IJmuiden site, assumes an electricity price of €40/MWh.

As the global energy transition accelerates, transitioning away from fossil fuels has become critical to assist the mitigation of climate change. Hydrogen, with its dual role as both an energy carrier and fuel, holds great potential for this transition. However, its large-scale production, particularly for industries like chemicals and transportation, still relies heavily on fossil fuels, posing a challenge for sustainable development.

This research focuses on advancing green hydrogen production, specifically through alkaline water electrolysis, a technology associated with zero greenhouse gas emissions. A key aspect of this work is addressing a gap in large-scale electrolysis modeling, with an emphasis on a modular system design. Modularity, as opposed to traditional single-unit scaling, offers improved operational flexibility and safety. This approach is especially relevant when electrolysis systems are powered by renewable energy sources, another critical component of the energy transition.

This thesis presents an investigation into the performance optimization of a modular alkaline water electrolysis system, designed to handle fluctuating renewable energy inputs. A physics-based numerical model was developed in Python, to simulate a large-scale AWE system composed of multiple modular units, capturing critical parameters such as temperature evolution, gas purity, and energy losses. The model was built progressively, starting from the cell level, incorporating a thermal model for temperature development and a mass transfer model for gas purity estimation. These combined elements formed a robust tool for simulating and optimizing the performance of modular electrolysis systems.

After validating the model against existing numerical and experimental data at the cell level, it demonstrated a strong ability to accurately capture the behavior of a single-cell system across all modeled parameters, including cell potential, temperature, and gas impurities. The differences between the simulated values and experimental data were minimal, further confirming the model's accuracy and reliability. This validation provided confidence in the model's predictive capabilities and laid the foundation for extending it for a larger modular system.

Following this, a scaling analysis was conducted to evaluate the model's performance when applied to a modular system. The simulations were carried out under both steady and varying power inputs, reflecting realistic operational conditions, particularly when coupled with renewable energy sources. The results highlighted the model's capacity to predict temperature evolution and gas impurity levels in such a scaled system. These findings indicated that the model not only captured the thermal and mass transfer behavior but also provided valuable insights into the effect of system scaling on overall performance and safety.

The outcomes of this research demonstrate that the developed model could act as a valuable tool for optimizing the performance of modular alkaline water electrolysis systems. The model successfully predicts the thermal behavior, gas purity, and energy losses across a range of operational conditions, including fluctuating power inputs typical of renewable energy sources. By enabling the fine-tuning of operational parameters prior to system deployment, this model provides a significant advantage in designing safe, efficient, and scalable hydrogen production systems. Future work could extend the model's capabilities by incorporating additional factors such as degradation mechanisms and detailed component-level interactions, ensuring even more robust predictions over long-term operation.

The utilization of water electrolysis for hydrogen production is expected to increase significantly in the upcoming decades. Modular system design offers great advantages for emerging markets, whilst providing the necessary flexibility in capacity that intermittent renewable energy sources require. Literature research reveals little information regarding modular water electrolysis systems. However, their availability on the consumer market proves that there is demand for such systems. To facilitate the emerging hydrogen market, a modular engineering design approach is used to develop a containerized alkaline water electrolysis system. This containerized system achieves a high degree of modularity by utilizing multiple modular sections. Each section functions as a stand-alone hydrogen production station, containing all necessary components required for safe hydrogen production. These components include heat exchangers, gas-liquid separators, modular alkaline electrolyzers, and a novel component called flow panels. In conjunction with the Dutch company XINTC, the design of the modular sections and their components is presented. Significant capital expenditure reduction is achieved by operating at atmospheric pressure, allowing for cost-efficient material selection and their corresponding manufacturing methods; both of which are discussed. Thermal, hydrodynamic, and particle dynamic models from literature are presented and evaluated. The proposed hydrogen separator is expected to achieve a separation efficiency of over 99.8%. A thermal system analysis concludes that the system is capable of operating at maximum capacity for ambient temperatures of up to 35 C. Additionally, a hydrodynamic system analysis shows that the pressure drop inside the system is dominated by that of the heat exchanger, and is in conformity with the Pressure Equipment Directive.

The long-distance maritime transport sector plays one of the most relevant roles in the transport sector. Therefore, this project designs a modular solid hydrogen solution, studying the assumptions made in the literature and including new simplifications that reduce the computational costs of its simulation. Moreover, this project compares the viability of this solution with the different prototypes found on the market.

In this work, the added value of machine learning (ML) molecular force fields (FF) for the community of molecular simulations is showcased by successfully calculating transport properties of aqueous potassium hydroxide (KOH (aq)). Classical FFs use relatively simple interatomic potentials to simulate the nano scale. These simulations can predict macroscopic properties, such as density, heat of evaporation, viscosity, and self-diffusivity of the modeled materials. However, these FFs struggle to model materials in which more complicated interactions are relevant for the macroscopic behavior. Examples of such interactions are three-body interactions and chemical reactions. Quantum scale simulation methods are able to compute properties of materials in which these challenging interactions occur, although these methods are limited in length and time scales that can be modeled with realistic computational costs. Transport properties, such as viscosity, self-diffusivity and electric conductivity need these larger length and time scales to be determined accurately. ML can be used for a multi scale approach, bridging the gap between the quantum and the nano scale by training coefficients of general interatomic potentials. This provides the possibility of reaching the time and length scales of traditional molecular simulations with the accuracy of quantum mechanic models. KOH (aq) is selected to highlight the prospects of these multi scale techniques, as the self-diffusion of OH- in this electrolyte is dominated by proton transfer reactions, which has not been modeled successfully with classical FFs.

Results of structure properties produced with ab initio molecular dynamics (AIMD, at quantum scale) simulations are compared with machine learning molecular dynamics (MLMD, at multi scale) simulations. There are no significant differences in the calculated shortest typical atomic distances and coordination numbers for both KOH (aq) and pure water systems. The determined transport properties are in the same order of magnitude as experimental results, although the calculated viscosity is overestimated and the self-diffusion of H2O and K+ are underestimated. This is because the system is simulated at a higher than experimental density and hydrogen bonding is overestimated with the selected quantum mechanics model. The proton transfer reactions are captured in the MLMD simulations, calculating the enhanced self-diffusion of OH- to be (6±2)e-9 m squared per second, which matches experimental results at infinite dilution.

In this thesis, the first part will be a review of the literature that includes as much of the commercially available information on carbon capture and mainly, Direct Air Capture (DAC). What are the difficulties surrounding capturing CO2 from gas mixtures and what are the possible solutions that have been investigated? What type of negative emission technologies have been proposed, what are the important parameters for the design of carbon removal technologies and what are the biggest obstacles that need to be overcome? Gathering sufficient data and comparing the different finding will be of importance, given that DAC technologies are still in developing stages and information is limited. The focus of this thesis will be on chemically amine based solid sorbents, as this can be argued that this is the most promising technology for a viable DAC system.

The modelling part of the thesis will be focused on the design of and development of a DAC model in the open source software ’Python’. The amine based sorbent that is investigated and used for the creation of the model is Lewatit VP OC 1065. Using experimental data gathered from the literature together with feasible assumptions, a model will be built to recreate the the whole DAC process and analyse the system. The main focus will be on acquiring a flexible model for both the adsorption and desorption parts of the DAC process to make further investigation of the system parameters possible. This model will be used after this thesis for further development and, for instance, analysing different possible sorbents.

Acid-gas capture systems are used to remove acid-gasses from waste gas streams from combustion or other chemical processes. A commonly used solvent is an aqueous solution with the primary amine monoethanolamine (MEA). Removing the acid gasses typically involves heating the solvent to approximately 378-383K in the stripping reactor. Using methyldiethanolamine (MDEA) instead of MEA can reduce the heating energy consumption of the stripper due to its lower reaction heat. The design of new reactors using aqueous MDEA solvents requires more data describing the properties of this solvent. Literature reporting thermophysical properties of aqueous MDEA solvents and transport properties of acid gasses in these solvents is exceedingly scarce. This work fills that information gap. Molecular dynamics were employed to compute density, viscosity and diffusivity of MDEA, CO2 and H2S in aqueous MDEA ranging in 0-50 wt% MDEA and 288-323K. The simulations were conducted using fully atomistic force fields for all species. The charges of MDEA were computed using Gaussian09 and scaled to achieve optimal agreement with experimental density and viscosity data. It has become clear that N-C-C-O dihedral in MDEA is crucial to reproduce experimental data of the viscosity and the diffusivity of MDEA. Two dihedrals were tested to achieve the best results. The resulting computed density, viscosity and diffusivity of MDEA are in good agreement with experimental data. The mixing rules between MDEA and CO2 were adjusted to increase accuracy of the prediction of diffusivity of CO2. The results are in good agreement with experimental data at 0-10wt% MDEA or 288K. The deviations become larger with higher wt% MDEA or higher temperatures.

The power plant, steel and petrochemical industries are large sources of carbon emissions worldwide. For meeting the climate goals of 2030 and 2050, efforts are underway towards pioneering novel technologies. Gas separation by supported liquid membranes is an attractive option for its energy-efficient, continuous and low cost of operation. The separation of gases takes via a solution diffusion mechanism where the gases are separated based on their
selectivity. Krytox oil is a high performance perfluoropolyether polymeric lubricant, originally aimed for its application to supersonic transport aircraft. The lubricant is known for its chemical and thermal stability leading to a longer usable life. The oil has shown an affinity for carbon dioxide (CO2) gas, a weak Lewis acid, owing to the fluorine and oxygen atoms in the polymeric oil which act as Lewis base and is thus envisioned for use in supported liquid membranes for gas separation.
Molecular dynamics simulations serve as a powerful tool of study, overcoming the shortcomings of experimental methods such as the difficulty of measurements at elevated temperatures, pressures or handling dangerous chemicals. This project covers the study of transport properties of Krytox oil namely the oil viscosity and diffusivity of CO2 in the oil for varying conditions of temperature, pressure and polymer chain length using equilibrium molecular dynamics simulations. The suitability of various atomistic force field models for this particular study has been tested, proceeding with the Universal force field (UFF) as the model of choice for studying the oil properties. To shorten the simulation time and study long time scales, coarse-grained simulations have been carried out using state-of-the-art MARTINI force field. In addition to transport properties, Henry’s constant for the solubility of CO2 in Krytox oil
has been predicted via alchemical free energy calculations by molecular dynamics simulations.

The current industrial application of carbon capture utilization and storage (CCUS) is limited due to technological drawbacks such as high energy demand and environmental pollution. Ionic liquids (ILs) and deep eutectic solvents (DESs) are considered promising alternative solvents for the capture of carbon dioxide (CO2). DESs are often characterized by high viscosities, which hinders industrial application. This problem might be solved by mixing the DES with an organic solvent. This study aims to assess the DESs choline chloride-ethylene glycol (ethaline) and choline chloride-urea (reline) mixed with methanol and propylene carbonate (PC) for their suitability as a medium for the combined capture and electrochemical conversion of CO2. Molecular dynamics (MD) simulations are performed to obtain the densities, the viscosities, the self-diffusivities, the ionic conductivities and insight into the molecular interactions of these mixtures. Independent MD simulations are performed of these mixtures with low concentrations of the solutes CO2, oxalic acid and formic acid. Complementary studies within the Bio-cel project are conducted to characterize the solubility and electrochemical reaction of CO2 and the techno-economics.
The viscosities of the mixtures monotonically decrease for an increase of mole fraction of organic solvent, which is benign for the application of CCUS. The self-diffusivities of all constituents increase monotonically for an increase of mole fraction of organic solvent. The ionic conductivity is calculated based on the ion self-diffusivities. Ionic conductivity optima are found at a mole fraction of DES of approximately 0.6 for ethaline-PC and approximately 0.2 for ethaline-methanol and reline-methanol. For higher mole fractions of organic solvent, the ionic conductivity decreases due to a depletion of ions. Radial distribution functions (RDFs) are used to analyse the intermolecular interactions. RDF peaks between chloride-choline and chloride-ethylene glycol show an increase for an increasing mole fraction of organic solvent, which was unexpected. The numbers of hydrogen bonds decrease for addition of methanol to pure deep eutectic solvent. For addition of propylene carbonate, this decrease is less pronounced. The depletion of hydrogen bonds at low mole fractions of deep eutectic solvent is in correspondence with the decrease in viscosity and increase in self-diffusivities. The results indicate that, for the studied properties, deep eutectic solvents mixed with organic solvents are more favourable than pure deep eutectic solvents for the absorption and electrochemical conversion of CO2.

Nitrogen oxides (NOx) are significant sources of air pollution. Nitrogen oxides like Nitric oxide (NO) and Nitrogen dioxide (NO2) are mainly responsible for the acid rain and smog. Nitrous oxide (N2O), also known as the laughing gas, is the major greenhouse gas that is responsible for the ozone layer's damage in the troposphere. According to the Environmental Protection Agency (EPA) report, one pound of N2O is 300 times more potent greenhouse gas than one pound of CO2. The significant emitters of Nitrogen oxides (NOx) are automobiles, agricultural sources, thermal power plants, and chemical processes like Nitric acid production plants, paint manufacturing, etc.   This study mainly focuses on the tail gas emitted from the Nitric acid production facility. The tail gas emitted during the HNO3 production consists of almost 2% of O2, 200-400 ppm of NO2, and NO, whereas 800 ppm of N2O.  As N2O is the most emitted gas from the Nitric acid production facility, it is followed by NO2 and NO, so it is essential to reduce these pollutants from the tail gas. Selective catalytic reduction (SCR) is a well-known technique currently involved in reducing NOx via the adsorption process from the Nitric acid production facility. But the costs involved in these methods are quite high. Nanoporous materials like zeolite exhibit uniform pore size and high thermal stability are said to be the promising adsorbents of NOx. The availability of a large number of zeolites makes it impossible to identify the proper zeolite for NOx adsorption experimentally. In such situations, molecular simulations are a powerful tool that can help identify the perfect zeolite. The time and cost involved in the process of molecular simulations are very low.  In this work, Monte Carlo simulations involving reaction ensemble are implemented to obtain the equilibrium composition of NOx components at desired operating conditions in the Brick molecular simulation package. This is followed by Grand Canonical Monte Carlo simulations (GCMC) and Reactive Grand Canonical Monte Carlo simulations (RXMC-GCMC) for pure and quaternary NOx gas mixture adsorption in five different zeolites (FAU, FER, MOR, MFI, and TON) using simulation package RASPA. The composition results from the reaction ensemble are validated with the composition results obtained using the Gibbs minimization technique in the MATLAB model, and the results are in good agreement. The quaternary gas mixture adsorption results in five different frameworks from RXMC-GCMC simulations are then validated in Ideal adsorbed solution theory in the Python model, and the results are in good agreement at the given operating conditions.

The energy mix of the future is likely to feature hydrogen due to its versatility. For effective use in energy storage, hydrogen has to be compressed. Conventional electrolysis of water and subsequent mechanical compression of hydrogen is an energy intensive process. A one-step conversion and compression can take place in a Electrochemical Hydrogen Compressor (EHC). This has the disadvantage that the resulting stream of high pressure hydrogen contains water. In order to assess the capability of hydrophilic zeolites as a means of selectively adsorbing this water, a in silico screening study of 6 zeolites was performed. To this end, a force field was constructed to allow for the simulation of high pressure hydrogen dehydration using Monte Carlo methods. The validity of this force field was evaluated by replicating simulation studies of adsorption of water/ hydrogen on zeolite frameworks with the presence of extra-framework cations. At the system pressure of 875 bar, prediction of the fugacity coefficient by means of the Peng-Robinson Equation of State (PR-EOS) yields inaccurate results. Therefore, these are calculated in the CFCNPT ensemble. It is demonstrated that the Ideal Adsorbed Solution Theory (IAST) is not suited to predict binary adsorption isotherms in this system. As a result of the screening study, it is found that the selectivity of water over hydrogen is almost linearly correlated to the amount of Al atoms in the zeolite framework. Furthermore, it is observed that topologies which feature a low amount of void space outperform those where significant void spaces are present. This can be attributed to the fact that the interactions of the zeolite with water are stronger than those with hydrogen in the limit of a low Si/Al ratio (Si/Al=1). It is theorized that water adsorbs preferentially at the surface of the zeolite, and competitive adsorption by hydrogen can only take place in sufficiently large void spaces.

The main objective of this study was to develop a quick model for an oil free twin screw compressor. The computational time of the model was to be kept as low as possible so that the model could be used for quick estimations and optimizations of the process. This was done by using a simple geometry focusing on the volume curve and port sizes, that can easily be changed by scaling the process as needed. The whole model was done in Matlab, with thermodynamic properties of the ammonia water mixture implemented by Rattner and Garimella. The model was validated by comparing general trends seen in oil free twin screw compressors using different working fluids.
The quick computational time allows for many different features to be examined in a short time to get a general idea of how the compressor would react to them. While using models with more detailed geometries one test can take hours compared to minutes with the model of this study. From the model search it was concluded that for this level of model a finite volume method would be the best option for this study. Various methods were explored to solve the governing equations, counting a few integrated solvers in Matlab and some numerical procedures for ODEs. The method that was chosen as the best for this system was the Euler method. All methods had in common was that the step size needed to be very small, for smaller geometries they needed to be even smaller in some sections of the compressor. To account for that the model was adjusted to accept variable steps. The variable steps allowed for smaller steps around the beginning and end of compression and thus lowering computational time. The main output of the model is the efficiency of the compressor and the Coefficient Of Performance (COP) of the cycle.
The model was tested for various operating conditions to determine how it, and the compressor, would respond. The main result is that the vapor quality should be kept rather low, around 40%, to achieve the highest efficiencies. The performance is also better with higher rotational speeds, however higher speeds also mean more mas flow through the whole system and increased mechanical losses. This means that up to a certain point the rotational speed can be increased until it starts decreasing the efficiency. The ideal ammonia concentration is around 33-35wt%, at higher or lower values the coefficient of performance decreases.
As some preliminary work for future experiments the experimental cycle was set up using pumps and heat exchangers to simulate the compressor. This included verifying the calibrations of sensors, confirming that correct data was read from sensor to LabView and getting the system leakproof. This cycle was also implemented in Matlab to be tested. From those tests the conclusion is that after going through the separator and the vapor and liquid streams being mixed back together the temperature decreases around 60 degrees usually, with a small increase in vapor quality around 2-3%.

For the design and optimization of different processes and technologies in the chemical and petrochemical industry, the knowledge of the accurate vapor-liquid phase equilibrium of hydrocarbons and their binary mixtures is fundamental. The main focus of this thesis is the binary mixtures of methane with various n-alkanes and the binary mixture of methane and toluene, at temperatures ranging from 400 to 650 K and pressures ranging from 2 MPa to 50 MPa. Despite the currently increased importance of these asymmetric binary mixtures of methane and long n-alkanes due to Enhanced Oil Recovery technologies and depletion of old, easily accessible and highly profitable hydrocarbon reservoirs, the available vapor-liquid equilibrium (VLE) data from experiments are scarce or unknown.Currently, the most common practice for volumetric and phase behavior calculations in the industry is based on different cubic, such as Peng-Robinson, SRK or on higher order equations of state like PC-SAFT . However, the predicted data by equations of state are not accurate enough, due to the lack of experimental data. This is more pronounced for high temperature and pressure conditions or in the vicinity of the critical point. This thesis aims to produce vapor-liquid phase equilibrium data for binary mixtures of methane with various long n-alkanes by performing Monte Carlo molecular simulations in the Gibbs ensemble with TraPPE force field. The simulation results are used to validate the applied TraPPE force field by comparing its results to available experimental data. At extrapolated conditions, the new data are compared to PC-SAFT predictions in order to assess the performance of the CBMC technique and to highlight the deviations between the CBMC and PC-SAFT results. Additionally, this new data could be used to adjust the parameters of the applied PC-SAFT equation of state to achieve better predictions at extrapolated conditions.