Laminar to turbulent transition is the process in which a smooth orderly laminar flow becomes turbulent, chaotic and unpredictable. A laminar boundary layer (BL) may become turbulent due to growing disturbances in the flow. If a disturbance is small its behavior is governed by linear stability theory (LST). Early works on LST were focussed on the growth of normal modes in incompressible or ideal gas compressible boundary layers, with the more recent inclusion of high-temperature and/or dense gas effects. Consideration for growth of perturbations other than modal started in 1970s, and while it has received considerable attention in ideal gas boundary layers, the study of non-modal growth in heavily stratified boundary layers remains limited.
In recent years, research in supercritical fluids and their applications has risen substantially, where supercritical CO2 (SCO2) stands out. It has been proposed as a working fluid for power generation turbines, a heat carrier fluid in geothermal, etc. A supercritical fluid near its pseudo-boiling temperature exhibits extremely large variations of physicochemical properties, leading to strongly stratified transcritical boundary layer flows, which may heavily influence its stability. Recently, transcritical boundary layers have been shown to be unstable to a novel mode, not found in ideal gas boundary layers, further motivating the study of the hydrodynamic behaviour of supercritical fluids.
This study investigates the linear stability of SCO2 boundary layers in the region of the Widom line in the phase diagram. Both heating and cooling are considered, in the sub-, super- and transcritical regimes. Regarding the amplification of normal modes, a special focus is given to 3D perturbations and the conditions for which a 3D perturbation is more amplified than a 2D perturbation. A moderate Mach number is found to be necessary in the further destabilization of 3D waves when compared to 2D waves. As indicated by previous works, the Tollmien-Schlichting (TS) mode is found to be preferentially 3D for moderate Mach number in the sub- and supercritical regimes. However, in the transcritical regime, the TS mode is found to be most amplified for stream-wise propagating waves. The novel mode II, found solely in transcritical flows, is preferentially 3D at sufficiently high Mach number, both in the viscous and inviscid regimes. Regarding non-modal stability, the energy growth of sub- and supercritical boundary layers is comparable to that of an ideal gas compressible boundary layer, with optimal energy growth driven by the lift-up mechanism. In the transcritical regimes, the lift-up mechanism also dominates, but energy growth is considerably larger due to the presence of strong thermodynamic gradients within the flow. This results in substantial energy growth associated with density and temperature streaks in the perturbations. When considering purely kinetic effects, the cooled wall case shows higher growth when compared to an ideal gas boundary layer, whereas the heated wall case shows a reduction of the kinetic energy growth. The cooled wall condition results in high mean vorticity far from the wall, leading to high energy growth, while the heated wall condition results in low mean vorticity far from the wall, leading to a reduced energy growth.

Supercritical CO₂ (SCO₂) is a promising alternative to traditional working fluids in heat pumps and power cycles due to its high density, thermal efficiency, and stability. These properties allow for the design of more compact and efficient equipment. However, accurately modeling supercritical heat transfer, especially near its pseudocritical point, is challenging, because of extreme variations in its thermophysical properties. As a result of these inaccuracies in modeling, large fault margins must be taken into account, leading to overdesign of equipment.

Current methods for estimating heat transfer in such systems include empirical measurements, computational fluid dynamics (CFD), and Nusselt correlations. Although measurements provide accurate data, it is not a scalable solution. CFD methods offer this scalability. However, CFD cannot be applied in complex scenarios such as modeling of SCO₂ due to the trade-off between computational cost and accuracy. Nusselt correlations have low computational demand and are a scalable solution but come with low accuracy. Therefore, Nusselt correlations are not suitable for the design of equipment.

To overcome these limitations, this research applies a new model to predict heat transfer to SCO₂ flowing upward in a heated vertical tube. Current prediction methods use artificial neural networks (ANN). This research applies a convolutional neural network (CNN) for its ability to capture spatial context from the heat transfer trajectory. This property enables CNN to capture global and local patterns important for accurate prediction in SCO₂ heat transfer. In addition, the multiple kernels per layer enables the model to extract a large range of features from the heat transfer trajectory allowing for higher accuracy. The developed CNN model has shown superior performance, achieving an R² score of 0.970 and an MSE of 1477, outperforming existing ANNs and Nusselt correlations.

Furthermore, a feature importance study is done, to identify the nondimensional parameters that are important for heat transfer prediction The feature importance study highlighted the importance of parameters such as the Reynolds number and the Froude number in improving the model predictions. In addition, the feature importance study led to the development and identification of a new nondimensional parameter, (q_wk)/(T d), as one of the most important features influencing heat transfer. Furthermore, the experiments show that normalization is vital to enhance model stability and performance, countering issues such as exploding or vanishing gradients.

The findings in this research suggest great potential for using machine learning models to design more effective and compact heat transfer equipment. Future work will focus on a feature importance study for normalized, dimensional and nondimensional parameters for improved predictions. In addition, more synthetic data should be generated in the heat transfer detoriation and heat transfer enhancement areas for better prediction of those parts. Finally, a study into the application of machine learning models for designing heat transfer equipment should be done to show the benefits of such models.

There is currently a resurgence of interest in nuclear propulsion within the maritime sector, which is reflected by the ambition, as mentioned in the Dutch maritime sector report "No Guts, No Hollands Glorie", to develop a standardized, modular nuclear reactor for ship integration within 10 years. Nuclear energy has the potential to reduce the maritime sector’s contribution to climate change, as it does not emit CO2 during operation. Additionally, in contrary to many other renewable energy sources, nuclear energy offers a high-energy-density power source capable of providing sufficient energy for longer periods of operation. Not only would this improve the strategic autonomy of the Royal Netherlands Navy, but it would also be a solution for the increasing energy demands of additional unmanned systems or advanced combat systems, like high power radars and rail guns.

Implementing nuclear propulsion in future (naval) vessels presents challenges, particularly regarding the dynamic power profile of ships during operation. Land-based nuclear reactors typically operate as stable power sources, which contrasts with the fluctuating power demands of ships, especially naval vessels. This research aims to find a solution for these dynamic power requirements, without the use of energy storage capabilities, like batteries, for peak shaving capabilities.

Based on the expected implementation of a small modular reactor (SMR) by 2034, the Future Air Defender and the Amphibious Transport Ship were selected as potential vessel types of interest for the Royal Netherlands Navy to implement an SMR. Additionally, the high-temperature gas-cooled reactor (HTGR) and the supercritical carbon dioxide (sCO2) recompression power conversion cycle were selected for the nuclear power plant. A dynamic model of the selected SMR and its energy conversion system has been developed to compare its ramp rate with those of conventional naval prime movers, such as diesel engines and gas turbines.

The simulation results indicate that the reactor dynamics alone are insufficient to meet common ramp rates of naval vessels, demonstrating that relying solely on reactor control is not a viable control strategy. However, the implementation of the turbine bypass valve, while operating the reactor at a constant load, provides dynamic power behaviour comparable with diesel engines and even gas turbines. Potential drawbacks include reduced cycle efficiency at part load, as well as significant pressure and temperature gradients within the heat exchangers. The unacceptable temperature increase at the reactor’s inlet was addressed by incorporating a dump cooler into the primary circuit. This research therefore concludes that an HTGR, in combination with an sCO2 cycle and a bypass valve, is capable to provide the dynamic power requirements of a naval vessel.

This report analyses the improvement potential of zeotropic refrigerant mixtures on the performance of vapour compression heat pumps integrated into industrial dryers. The report reviews the scope of total global energy consumption and the percentage of this attributable to industry to demarcate application area.

It proposes that by targeting industrial process heating, significant reductions to energy consumption and carbon emission by industry can be achieved. Heat pumps are introduced as the preferred system to accomplish this for temperatures up to 200 ◦C. A performance limitation is identified in the form of heat transfer between process streams that reject and receive heat at non-constant temperatures.

Zeotropic refrigerant mixtures are introduced in heat pumps to improve the interaction with non-isothermal heat sources and-sinks. The considered mixture compounds are limited to future-proof refrigerants and the mixtures are limited to binary ones. Humid air encountered in industrial drying is targeted as one such non-isothermal process stream with a large potential for waste heat recovery. The most prominent industries that use dryers are identified and evaluated as to provide representative operating conditions at which a heat pump integrated dryer would have to work. Heat pump cycles are described both on a cycle scale as well as per component and the most important design aspects and considerations are presented. The methods used to allow heat pumps to interact favourably with temperature glides are presented namely zeotropic mixtures and trans-critical heat pump cycles.

A modeling strategy is proposed to simulate heat pump integrated dryers in large numbers in order to investigate the effects of dryer inlet and outlet conditions as well as refrigerant mixture composition both in terms of constituents and mixing ratio. The total process flow diagram is given and divided into definable thermodynamic states for both the (humid) air as well as the refrigerant. Governing equations for each component are presented and the calculation method for the thermodynamic states is discussed in terms of known state variables and used thermodynamic libraries. Finally a novel heat pump cycle is proposed, motivated by a desire to decouple the glide matching in the evaporator from that in the condenser, that attempts to use the fractionation risk present in zeotropic mixtures as an advantage instead. The criteria placed on refrigerant are presented both within the context of the future-proof limitation, legislative limitation as well as those introduced by the process requirements.
The most promising refrigerant candidates from literature are presented for applications in the defined process conditions. Finally a selection is made of 13 refrigerants namely water, ammonia, (iso)butane,
(iso)pentane, methane, ethane, propane, CO2, propylene, ethylene and hexane. The thirteen refrigerants are combined into 78 refrigerant pairs and modelled using the described modeling strategy. The modelling reveals that using zeotropic mixtures does improve the COP for many refrigerants when compared to their pure cycle performance. For drying at 180 ◦C two heat pumps are proposed at different outlet temperatures. The cycles use 87.5%mol Isobutane mixed with Ethane and 87.5%mol NH3 mixed with Propane for a high temperature and lower temperature outlet respectively. COPs of 3.38 and 3.44 were calculated with PRs of 16.53 and 6.84 for the Isobutane-based and NH3-based cycles respectively. A lower drying temperature of 120 ◦C was explored and here CO2-based mixtures were identified as highly desirable refrigerants due to their non-toxic, non-flammable nature as well as low global warming and ozone depletion potential. The mixtures 87.5%mol CO2-Isopentane and 90%mol CO2-Isobutane were proposed, both feasible with single stage compression and possessing a COP of 3.96 and 4.02. Zeotropic mixtures improved the COP in at least 48 out of 78 possible refrigerant combinations when compared to the pure cycle COPs and allowed the dampening of flammability and toxicity where the pure refrigerant possesses those properties.

It is clearly demonstrated that future-proof binary zeotropic mixtures increase the performance of VCHP-integrated dryers across all relevant temperature ranges, by as much as 21.47%. The highest COPs are found when zeotropic refrigerant mixtures are used together with trans-critical operation. It is concluded that zeotropic mixtures improve the COP of vapour compression heat pump integrated dryers and should be utilised for non-isothermal processes like drying.

Amidst the rise in fossil fuel consumption and the global energy crisis, there is a growing demand for clean, environmentally friendly, and sustainable energy solutions in industrial processes. One promising approach is to substitute conventional working fluids in power cycles with supercritical fluids. Over recent decades, substantial research effort has been made in investigating supercritical fluids. Among these, supercritical carbon dioxide (sCO2) has emerged as a practical and alternative solution. Its low critical pressure and temperature, high thermal efficiency, and operational flexibility have garnered widespread interest across various energy applications, particularly in heat exchangers and gas-powered cycles. However, the complex flow dynamics and heat transfer characteristics of sCO2 near its critical point have become the subject of intense research. When sCO2 flows through a heated channel, strong density gradients can generate dominant buoyancy forces that can significantly affect the supercritical fluid structure, mixing and transport properties. Understanding buoyancy-affected flows is highly crucial as it can lead to flow stratification and heat transfer deterioration or enhancement in the channel.

The role of buoyancy forces in flow stratification is quite substantial and can stabilize or destabilize the stratified structure by inducing or dampening instabilities. While numerous computational simulations have been performed to understand the mechanism of buoyancy-affected stratification in ideal fluids, whilst, a notable research gap exists in the literature on supercritical fluid stratified flow. Therefore, the present study aims to investigate the influence of buoyancy on the sCO2 flow stratification in a channel, considering heating from the top and bottom walls. In this context, a Direct Numerical Simulation (DNS) is conducted using the open-source CFD package ”OpenFOAM” to simulate pressure-driven sCO2 channel flow under constant wall heat flux. A buoyant pimple foam solver was adapted to simulate transient supercritical flow. To gain good accuracy in the thermophysical properties, a custom library was prepared to interpolate supercritical fluid properties at simulation run-time. To assess the influence of buoyancy in the heated channel flow of sCO2, a developing flow profile is initiated at a constant pressure of 80 bar. Stratification is achieved by imposing a heat flux at the wall boundary spanning in the heating range of 5 − 15 kW /m2, resulting in density and temperature variation across the fluid. By varying the heat flux at the wall boundary, we analyze the resulting variation in the flow field (temperature, velocity, and pressure distribution), the dynamics of heat transfer, and the influence of buoyancy by the non-dimensional parameter Richardson number in the stratification of supercritical carbon dioxide. The results show that the effect of heating on the developing boundary layer in sCO2 channel flow is substantial. With increasing heat flux, the flow is accelerated near the heated wall while decelerating in the bulk. A strong non-linear variation in the temperature and density distribution is observed in the wall-normal direction. Moreover, as the heat flux increases, the wall shear stress decreases due to strong property variation, while the Richardson number and Reynolds number increases, and the heat transfer coefficient decreases.

In recent years, a considerable amount of research has been directed towards making energy generation more efficient to combat global warming. To aid this goal, the use of supercritical fluids (SCFs) is gaining a lot of traction. SCFs have only one phase and experience a sharp variation of thermophysical properties when heated sufficiently. These facts can be exploited to gain several advantages over conventionally used sub-critical fluids. However, the strong property variation in heated SCFs also complicates the flow Physics considerably if the wall heating is strong enough. Further, the flows in such settings are often turbulent and spatially developing. For these flows, a strong variation of properties can lead to a modulation of turbulence, which the conventional turbulence models can not predict well. This thesis is an attempt to better understand how spatially developing heated supercritical turbulent flows behave and use this understanding to improve turbulence model predictions.

In this thesis, we investigate two supercritical heated developing turbulent flows from Nemati et al. (2015). One case is oriented horizontally and the other is oriented vertically. The latter has an additional effect of buoyancy resulting from its vertical orientation and strong density variation near the wall. We analyze how the strong property variation modulates the turbulence in both cases. Then, we assess if Semi-Local Scaling (SLS) and Apparent Reynolds Number (ARN) theories can characterize the modulated turbulence. Further, we propose a methodology to make use of ARN by itself, and in combination with SLS to improve turbulence model predictions.

The results indicate that ARN theory provides a robust way to sensitize conventional turbulence models to the additional effects arising in supercritical heated developing turbulent flows due to property variation. Additional research in turbulent heat flux modeling is deemed necessary to improve the model performance further.

Hydrogen's competitiveness as a sustainable energy carrier depends greatly on its transportation and storage costs. Liquefying hydrogen for these purposes offers benefits like purity, versatility, and higher density. However, current industrial hydrogen liquefaction processes face efficiency and cost challenges, with a second-law efficiency below 25% and costs between 2.5-3.0 US$/kg_LH2, and they are mainly limited to small-scale. Scaling up liquefaction plants is a potential solution to reduce costs, but existing models often overlook the economic and technical viability of the conceptual plants.

Most reported liquefaction costs rely on "ballpark" estimation, hindering process economic comparisons between various processes. Industry-sponsored studies often use confidential data, limiting accessibility. Thus, this thesis aims to develop a comprehensive framework for modelling large-scale liquefaction processes and assessing their feasibility, focusing on the large-scale high-pressure hydrogen Claude-cycle liquefier concept.

The technical analysis focuses on the preliminary designs of main equipment, such as compressors, turboexpanders, and heat exchangers, ensuring compatibility with existing technology limitations. Aspen Process Economic Analyzer (APEA) is employed to estimate the capital and operating expenditure. At 0.1 €/kWh electricity price, the techno-economic assessment of the 125 tonnes per day (TPD) Claude-cycle concept yields specific liquefaction costs (SLC) of 1.55 €/kg_LH2. Sensitivity analysis shows electricity price has significant influence. Applying the established design approach, incorporating high-speed centrifugal compressors could reduce the SLC by 5.42% and potentially more.

Scaling up to 250 and 500 TPD shows potential SLC reductions of 7.80% and 9.45%, respectively, at electricity price of 0.1 €/kWh. Future projections suggest a potential 12.6% SLC reduction as the Claude-cycle liquefiers matures. In an ideal scenario, with incentives and low-cost renewable electricity, costs could range from 0.87 to 1.09 €/kg_LH2.

Finally, a cost-scaling curve for hydrogen liquefaction plants are estimated based on the cost results from this study. The curve is comparable to existing cost curves reported by industrial and government joint research projects, validating the methodologies developed in this thesis. Moreover, an experience curve is also predicted using cost data from this study and data found in the literature. The curve indicates a 17% price decrease in hydrogen liquefiers with each doubling of global liquefaction capacity.

Variable green energy sources were sought to supply the growing demand for energy without impacting the environment. Wind and solar energy are at the forefront of the energy transition. However, their intermittent nature poses a challenge that necessitates the development of energy storage technologies.
Several technologies were studied over the years including pumped hydro energy storage, compressed air energy storage, electrochemical energy storage, and pumped thermal energy storage.
Pumped thermal energy storage provides a mean to store excess electrical energy in the form of heat by employing a heat pump cycle during the charging process and a heat engine cycle during the discharging process. There are multiple variations of pumped thermal energy storage cycles of which Rankine and transcritical cycles. Unlike Brayton cycles, Rankine and transcritical cycles typically operate at low-temperature levels < 200o𝐶 which facilitate their
integration with low-temperature waste heat in order to boost their round trip efficiencies even beyond 100%. Therefore, the focus of this thesis was to model and optimize Rankine and transcritical thermally integrated small scale pumped thermal energy storage.
Critical to the storage system’s efficacy are the performances of its key components: the compressor and expander. Geometric models were developed and validated for these components, facilitating the determination of their isentropic efficiencies and their variation with pressure ratio and rotation speed.
Additionally, a small-scale pumped thermal energy storage system model was developed to study the system performance. The choice of working fluid and operating parameters were guided by both a simplified optimization scheme and a multi-objective approach utilizing the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II). This yielded 𝑅13𝐼1 as the optimum working fluid with source temperature 𝑇𝑠𝑜𝑢𝑟𝑐𝑒 = 80o𝐶 and storage temperature 𝑇𝑠𝑡𝑜𝑟𝑎𝑔𝑒 =140o𝐶. This configuration led to a transcritical heat pump charging cycle and a subcritical heat engine discharging cycle employing pressurized water as the storage medium.
Consequently, the compressor and expander geometric models were integrated into the pumped thermal energy storage system model working with 𝑅13𝐼1 at the source and storage
temperature mentioned earlier which yielded a round trip efficiency of 𝜂𝑟𝑡 = 93%, energy density 𝜌𝑒𝑛 = 12𝑘𝑊ℎ𝑟/𝑚3, and exergy efficiency 𝜓𝑒𝑥 = 29%.
Finally, a case study was demonstrated with the integrated cycle employed in a Dutch household solar system where dual electric and thermal panels are employed for the electrical and heat energy input in the pumped thermal energy storage. The system was developed for a 1𝑘𝑊 electrical input from the solar system to be stored for further use during off peak hours.

Stratified turbulent flows abound in environmental and industrial settings. Examples are atmospheric boundary layer flows, the transport of nutrients and organisms and the mixing of heat and salinity in the oceans, fluid flow in heat exchangers, and the transport of reactants and products in chemical reactions. These examples and many others consider stratified wall-bounded turbulence, in which the creation of turbulence by mechanical processes contends with its dissipation due to buoyancy effects. These flows are said to be stably stratified as these are inherently stable flows and are averse to mixing. The buoyancy effects alter the structure of the flow, and consequently the dynamics of mass, heat, and momentum transport. As density fluctuations become more severe, the Oberbeck-Boussinesq approximation becomes inaccurate and the resulting dynamics are not correctly predicted. In the current work, we developed and validated a numerical solver for direct numerical simulations (DNS) of turbulent flows featuring strong property variations. More precisely, we solve the Navier-Stokes equations in the limit of vanishing Mach number (so-called low-Mach number limit), with the fluid density given by the ideal gas law, and the dynamic viscosity and thermal conductivity also expressed as functions of temperature.
Our numerical solver is used to study stably-stratified turbulent channel flow under non-Oberbeck-Boussinesq conditions. The simulations are carried out at friction Reynolds number of  180, Prandtl number of 0.71, and friction Richardson number of the O(10). These non-dimensional numbers are the governing parameters and are defined based on the prescribed pressure drop and properties of the fluid at the reference temperature. Stratification is achieved by imposing constant temperature boundary conditions, with a high upper-to-lower wall temperature ratio (larger than 2), resulting in strong density variations in the flow. We will vary the temperature ratios and adjust gravity to maintain a similar Richardson number between cases, thereby isolating the effects of strong property variations in the flow dynamics. We will analyze the dynamics of heat and momentum transport under strong stratification for these conditions, also in light of DNS data of the same system under the Oberbeck-Boussinesq regime.

The rise in temperature attributed to human CO2 emissions and the escalating energy needs of society necessitates the development of clean energy production. Solar and wind energy, both renewable sources, have emerged as cost-effective alternatives to conventional fossil fuel systems. They now account for a substantial portion of the world's electricity generation (IEA, 2022). However, their intermittent nature poses a challenge to their reliability. To overcome this, the implementation of grid-scale energy storage systems is crucial. Such systems can store excess energy produced during peak periods and release it during low-generation or high-demand periods, ensuring a stable and dependable power supply to the grid.

Pumped Thermal Energy Storage (PTES) is one such type of promising grid-scale storage solution based on the concept of storing electricity in the form of heat. These systems are not reliant on rare earth metals, are not restricted by geographical location, and are relatively economical over their lifetime. They employ a heat pump cycle for charging and a heat engine cycle during times of discharge. Often, in the thermodynamic modelling of PTES systems, a fixed value of turbomachinery efficiency is assumed. This approach holds well for the first estimate of performance, but for better accuracy and further analysis, meanline models could be used to arrive at the efficiency value and preliminary geometric design. Hence, this work presents a method for developing meanline models for centrifugal compressors and radial inflow turbines. Modelling techniques and guidelines from the literature are noted and presented here. The accuracy of these models is dependent mainly on the loss models used. Using suitable models selected from the literature, a fair agreement was found between the meanline model's prediction and experimental data from open literature, validating the methodology.

An essential function of energy storage is to provide load flexibility, meaning its charging and discharging cycles must adjust to match the net load curve. As a result, the turbomachinery would need to operate under off-design conditions to meet these demands. Therefore, this report introduces an approach to extend the PTES model by Radi (2023) for off-design operation based on turbomachine performance.

In response to climate change, the marine sector is increasingly focused on the energy transition. New marine power plants operating more efficiently and on cleaner fuels are receiving significantly more attention. While the transition to different power plants and fuels results in reduced fossil fuel consumption and emissions, waste heat recovery systems can aid in obtaining both as well. Current as well as future marine power plants do not convert all of the energy contained in the fuel into useful power, with significant amounts of energy being wasted as heat. Waste heat recovery technologies can be applied to recover some of this energy and generate additional power, effectively boosting total system efficiency, with the resulting benefits of improved fuel economy and reduced specific emissions.
This study first provides an extensive summary of all the different marine power plants, fuels, and waste heat recovery technologies, of both the present and future. One type of power plant that could be an advantageous alternative for future marine propulsion is the solid oxide fuel cell (SOFC) due to its high efficiency and potential to run on clean fuels. SOFCs generate electricity from chemical energy using a high-temperature electrochemical process, resulting in high-temperature waste heat being expelled. Therefore, the potential for a waste heat recovery technology to boost system efficiency is high, and it is chosen to investigate several power cycles for the waste heat recovery of a case study vessel powered by a 2 MW SOFC system. The aim of this study is to develop and execute an approach to evaluate these waste heat recovery technologies regarding their efficiency, size, and associated cost.
While numerous power cycles for waste heat recovery are in existence, a selection of potentially suitable systems with a number of different setups is made to investigate further. Thermodynamic models are created for the selected power cycles to determine and compare their theoretical efficiencies. Subsequently, the size of the heat exchangers are calculated for the evaluation of system size, considering the size of other components such as turbomachinery as well. Two types of heat exchangers are considered in this study: the compact and innovative printed circuit heat exchanger (PCHE) and the classic but commonly large shell and tube heat exchanger (STHE). Finally, the investigated systems are subjected to an economic analysis based on the cost associated with the main components, with again considerations being made regarding excluded additional components.
The results indicate that various configurations of the (transcritical) Rankine cycle operating on steam and CO2, as well as the (supercritical) Brayton cycle operating on CO2 and air, present with significant theoretical efficiencies ranging from 41 to 52% and electrical power outputs ranging from approximately 530 kWe to over 670 kWe. From the evaluation of the system size it is concluded that the smallest systems are those operated on CO2 equipped with PCHEs, while the largest are the air Brayton cycles equipped with STHEs. The economic analysis revealed that the systems with the lowest costs are the configurations of the (transcritical) Rankine cycle operating on steam, as well as certain air Brayton cycles equipped with PCHEs. The systems with the highest cost are found to be the air Brayton cycles equipped with STHEs, due to the significant sizes of the required heat
exchangers. In general, it is concluded that no system outperforms the others simultaneously across all three investigated aspects of efficiency, size, and cost, and trade-offs will be required when selecting a waste heat recovery technology for the presented case study vessel. Nonetheless, a detailed process-oriented approach has been developed and executed to allow various waste heat recovery solutions to be compared, and it is proven that a significant amount of power can be produced from recovered waste heat. The results from this study can be directly consulted by ship owners and designers considering the application of waste heat recovery to an SOFC powered vessel. Furthermore, the developed approach can also be applied to waste heat recovery in other industries and power generation systems, as it provides a step-by-step guide on relevant calculations and considerations.

The main objective of this thesis was to explore the capabilities of neural networks in terms of representing governing differential equations, primarily in the purview of fluid/aero dynamic flows. The governing differential equations were accommodated within the loss functions for training the neural networks, thereby making them 'physics-informed'. Subsequently, this idea of physics-informed neural networks (PINNs) was extended to parameterized geometries generated with the help of commercial auto-encoders developed by the UK based company Monolith AI pvt. ltd. because neural networks have the capability to learn the desired PDEs over variable/parameterized geometries without the need to recompute the solution for every minor change in the input geometry, which proves out to be a huge advantage over classical numerical techniques. The advantages, limitations and scope for further research in the field of physics-informed deep learning have been discussed in the contents of the underlying thesis report.

To face the ongoing climate crisis, the Dutch government adopted a number of policies and measures to limit its greenhouse gas emissions. It is estimated that 10% of Dutch annual emissions are caused by the built environment, a large portion of which come from two sources: space heating and domestic hot water.
The use of heat pumps is said to meaningfully contribute to the electrification of domestic heat supply. However, heat pumps face intermittency issues during unfavourable climate conditions when heating is most needed, thus limiting their potential of being standalone heating systems.
A hybrid heat pump - meaning a single heat pump operating with multiple renewable heat sources - could efficiently provide heat all year round with maximised performance. However, literature on the subject is sparse and non-systematic. Therefore, conclusions on the system performance and its variability cannot be drawn.
Thus, the aim of this project was to create a modular numerical model of a hybrid heat pump system to analyse its performance. This was achieved by individually modelling and experimentally validating each component.
The chosen heat sources were wastewater and air. The numerical model showed satisfactory correlation with experimental results for each component, particularly the condenser of the heat pump, air source and wastewater storage bag. Improvements could be made in the evaporator of the heat pump, boiler and spiral heater.
After analysis of its yearly performance, it may be concluded that the modelled hybrid heat pump is not currently a worthwhile investment for a typical Dutch home. However, with a higher capacity heat pump, alongside other component and control optimisations, the system can be an all-electric alternative to many gas-heated homes.

Due to the global energy crisis, there is a growing need to make industrial processes more efficient and clean. One such method is to use effective and renewable working fluids in power and refrigeration cycles. Research in this field during the past decade has placed supercritical CO2 (SCO2) in the forefront due to its non-ideal properties and abundance in nature but, to fully understand their heat transfer properties, insight into boundary layer flow and hydrodynamic instabilities are crucial. Experimental setups such as closed loop wind tunnel contraction facilities are needed which provide a laminar steady flow to take boundary layer measurements. However, using a contraction causes various problems such as formation of secondary flows like boundary layer separation, Gortler vortices, cross-flows and non-uniformity’s. In this study a novel 1D contraction shape with curvature on only one side is pro- posed. This will completely eliminate the risk of Gortler vortices on the bottom plate and make it a suitable candidate for boundary layer experiments. Nevertheless, the risk of the other secondary flows and Gortler vortices on the top wall disturbing this laminar flow remains. Hence, the present research aims to optimise this 1D contraction to reduce the risk of such secondary flows and analyse the growth of Gortler vortices on the top wall using numerical simulations.

Steady state laminar simulations were performed to optimise the design of the 1D wind tunnel contraction. Optimum contraction and settling chamber lengths were found by identifying the risk of flow separation and boundary condition effects respectively. For the optimisation of the contraction wall shape, a family of transformed fifth order polynomial curves with varying inflection point distances were selected. To identify the curve with least risk of boundary layer separation and lowest flow non-uniformity, a multi- objective optimisation procedure was implemented. The procedure found that a curve with inflection point of 101mm downstream of contraction inlet gives the best performance.

To analyse the effect of Gortler vortices on outlet uniformity, unsteady laminar simulations were performed on the contraction by forcing sinusoidal perturbations of different wavenumbers. This study found that perturbations form symmetrical steady steam-wise Gortler vortices in the contraction. The most unstable wavenumber was found to be 𝜆=83.33 𝑚^-1 which formed secondary vortices of high vorticity. However it was also seen that for all wavenumbers, the vortices lose energy as they exit the contraction and do not affect the bottom boundary layer. Further, the effects of a side wall on the formation of Gortler vortices were also investigated. The results showed that the Gortler vortex closest to the wall is absorbed into high vorticity corner vortices, while those close to the centreline develop into steady Gortler vortices. No effect on the centreline velocity profile was seen due to the vortices making the 1D contraction suitable for boundary layer experiments. Finally, it was also seen that the 1D contraction shape produces asymmetrical Gortler vortices due to alternative perturbation methods such as random inlet perturbations.

The Intergovernmental Panel on Climate Change (IPCC) believes that multiple solutions must be deployed simultaneously to reduce the emission of greenhouse gases into the atmosphere. Carbon, Capture & Storage (CCS) is an unavoidable technique within this portfolio as an intermediate solution. CCS requires transport of CO₂ through pipeline systems and into wells. There are still large uncertainties on the thermodynamics of the CO₂ in flow through valves. The CO₂ flow undergoes significant changes in pressure, temperature and phase distribution when it passes this control valve. Therefore, the behaviour of the CO₂ flow flowing through a control valve is examined in this study.

The simulation of CO₂ in a 3D valve including phase transitions is complex. Furthermore, few validation experiments are available. As a first step, more simple nozzles are simulated. In these devices, the same processes occur and validation data is available. These simulations are validated with experimental data by Nakagawa et al. to examine the accuracy. Three types of models (isenthalpic, Euler and Enhanced Mass Transfer (EMT)) are implemented in increasing levels of complexity to investigate the differences between these models and to consider when complexity is needed or simplifications are valid.

The validation cases show experimental pressure data of high-pressure CO₂ flow through converging-diverging nozzles with phase transitions. The results showed that the EMT model matched the experimental data best. A substantial similarity was achieved regarding the pressure data. The mass transfer mechanism, however, needed adjustments in its coefficients to match the experimental data. Finally, after finding the right values, the EMT shows the best technique for modelling flashing or cavitation.

In short, the overall consequences of the transition in a valve are substantial and must be considered. The behaviour of the high-pressure CO₂ flow is heavily influenced by flowing through a valve. Substantial amounts of vapour are formed, but only after the throat. This is the same for choking condition, which is achieved in the diverging section of the nozzle. The large expected drop in temperature due to pressure reduction has also been noted.

Although a high degree of similarity between the results of the model with experimental data is obtained, there is room for improvement regarding the model. A flaw was discovered in the handling of the thermodynamic properties of the fluid near critical points. Also, the surface tension has not been considered, but might have a substantial influence. Next steps in the research are 2D and 3D simulations of actual valves, but require experimental validation data.

The continued development of improved algorithms and architecture for numerical simulations is at the core of increased computational performance and, therefore, the ability to perform more complex and precise numerical simulations in less time in areas such as Computational Fluid Dynamics. Employing faster algorithms on more efficient processing units, such as Graphics Processing Units (GPUs), can reduce not only the time spend per simulation but also the energy required to perform these computations as well. This will be of significant benefit to different areas of research and engineering and through improvement achieved in these to society at large as well.

As simulations grow in dimensions and accuracy the wall clock time is bound to increase substantially, reaching days and potentially months depending on the parameters and geometry of the chosen simulation. The performance of different finite difference solvers with different degrees of optimization on different types of compute hardware was investigated and the achieved speedups assessed. Specifically, one serial CPU-based solver was presented as a baseline, which was then transitioned to a GPU-based solver. This in turn was then optimized further with regards to improved memory redundancy. To make comparisons fairer all of the solvers used the same temporal and spatial discretization techniques. Further, a benchmarking scenario was proposed to be used for the different solvers across used hardware, including the relevant geometry, gird, and initial and boundary conditions.

The speedups between the different solvers were observed and contextualized with regard to the effort that went into implementing the solvers and the capability and cost of the used hardware. The speedups at different problem sizes were investigated with the aim to establish how the performance gain from parallelizing and optimizing solvers scales with the chosen number of grid points and therefore the computational load.

Very significant speedups were achieved between the regular CPU solver and its GPU implementation, clearly showing the possible performance gains when moving from a serial to a parallel implementation running on an accelerator. The speedups between the two GPU-based solvers were more modest but still significant when considering the possible time spend on one simulation, depending on the chosen number of grid points.

Finally, an additional performance analysis was performed on two Navier-Stokes, one of which was the optimized version of the other, to investigate whether the performance increases were in line with prior findings and what magnitude of a reduction of wall clock time was possible for a state-of-the-art finite difference Navier-Stokes solver.

The global warming and increase in energy use demands the industry to come up with ideas to lower the CO₂ production and save valuable energy. Organic Rankine cycles have a great potential as low grade heat can be converted into electrical energy. In the last ten years, research in this particular field has increased together with the amount of installed systems. Triogen is a company which does a lot of research on how to increase the power output of their Organic Rankine Cycle systems. Their product involves a radial inflow cantilever turbine connected to a single shaft to the generator and the main pump which makes the system quite unique. Because of the fluctuating heat input the system scarcely runs on its design condition so research has to be done to ensure the turbine efficiency is as high as possible to extract as much power out of the heat source as possible. This study investigates whether the pump strategy can be modified in order to increase the system power output for part load situation. By means of elaborate turbine computational fluid dynamics simulations, an optimal rotational speed op n = 475 hz is found for all turbine inlet pressures. A simple cycle model is made in Excel in order to compute the increased power output, given the optimal rotational speed. The most cost effective and straight forward way to implement the fixed rotational speed in the system, a throttle valve is chosen. The amount of extra power gained using the optimal rotational speed is approximately 7 kW, which is an increase of 8 %. For evaporator heat inputs higher than 900 kW_th, the operation with throttle valve is no longer beneficial and the operation on design condition is favoured.

Experiments to characterise vortices in hyperbolic shaped funnels are being conducted at the Water Application Centre (WAC) in Wetsus. These have demonstrated their higher gas transfer rates in comparison to the conventional aeration systems presently in use. Depending on the imposed flow conditions, different regimes of vortices are formed, among which the Twisted vortical structure is observed to have the highest gas dissolution rates. This has probed several questions on the physical mechanisms responsible on both micro-and macroscopic scales. The present research aims to numerically analyse the flow field organisation in these vortices to reason the observed high gas transfer rates using Computational Fluid Dynamics (CFD).  Transient simulations were performed on a three-dimensional radially structured hexahedral mesh. Multi-phase modelling was done using the Euler-Euler approach-based Volume-of-Fluid (VOF) method, while the turbulent flow was modelled using Shear Stress Transport (SST) based on k - ω equations with curvature correction. The choice of boundary conditions and their location is crucial for forming a stable vortex in the hyperbolic funnels. The position of the air-water interface from experimental results was used to validate the obtained numerical results. Two regimes of the vortex, namely the Twisted and Straight vortical structures, were evaluated for their gas transfer capabilities in terms of Hydraulic Retention Time (HRT), interfacial area and mixing in the bulk.  Instabilities arise in the secondary flow field of these vortical structures analogous to the Taylor-vortices that develop in the well known Taylor-Couette flow systems. In hyperbolic funnels, these instabilities aid in advecting the bulk of liquid to the air-water interfacial region and also enhance mixing within the bulk of water. The former enhances the gas transfer rates while the latter promotes uniform mixing. The strength of these instabilities is qualitatively analysed in terms of average vorticity per unit mass of water. This is found to be higher in the Twisted regime in comparison to other regimes. This is augmented by high air-water interfacial area making this regime possess superior gas transfer rates.  Although the gas transfer rates are high, water exiting the funnel is undersaturated at the given operating conditions. In order to further enhance the amount of gas dissolved few possibilities are qualitatively discussed at the end of this study.     

Over the past decades, engineers have focused on a common goal - to reduce emissions by making industrial processes more efficient and by utilizing renewable energy sources. One possibility to reach this goal can be achieved by employing processes with non-ideal fluids, such as fluids at supercritical conditions in refrigeration, heat pump cycles and power cycles.
Most of the flows in industry are turbulent and hence the need to study turbulence in non-ideal fluids arose. Turbulence in non-ideal fluids is extremely challenging since there are many complex effects at play. One such effect is caused by variations in properties, which also occurs in compressible flows, flows with high concentration gradients, or flows in heat exchangers.

This thesis presents a review of the existing theory on semi-local scaling for variable property flows with the aim to take it one step further and apply it to turbulent heat flux modeling. Two types of variable property cases are analysed in this thesis; (1) low-Mach number flows with uniform pseudo-heating sources, (2) high-Mach number flows with non-uniform viscous heating. For the former, a Direct Numerical Simulation (DNS) data-base, already available at TU Delft, has been post-processed with the main goal to investigate if the semi-local theory can also be applied to thermal turbulence and its modeling. For the latter, additional DNS simulations of high-Mach number channel flows have been performed to investigate how the viscous heating and its correlations can be accounted for in the semi-local scaling framework.

Using a 2-equation heat flux model, we find that modeling thermal turbulence in semi-local scales considerably improves the results for low-Mach number flows. However, for high-Mach number flows, additional unknown (closure) terms arise due to fluctuations in the viscous heating source. A model for the source term in the enthalpy variance equation is successfully proposed. In addition, the DNS study of two high-Mach number flows with constant semi-local Reynolds number profiles shed light on the importance of a newly defined parameter (modified Eckert number) and also unveils one of the most important conditions in which semi-local theory can be compromised, e.g. extreme density gradients.