Search

The previously developed models for fuel droplet heating and evaporation processes, mainly the Discrete Multi Component Model (DMCM), and Multi-Dimensional Quasi-Discrete Model (MDQDM) are studied, improved and generalised for a broad range of bio-fossil fuel blends so that the application areas are broadened with increased accuracy. The main distinctive features of these models are that they consider the impacts of species thermal conductivities and diffusivities within the droplets to account for the temperature gradient, transient diffusion of species and recirculation. The research carried out in this thesis is focused on four key aspects: (1) application of the previously developed models for a broad range of fossil fuels, biofuels and their blends including ethanol/gasoline, biodiesel/diesel, E85-diesel (E85 refers to 85% ethanol and 15% gasoline) and ethanol/biodiesel/diesel fuel blends; (2) formulation of fuel surrogates, using a new model referred to as ''Complex Fuel Surrogate Model (CFSM)'', and analysing their heating, evaporation and combustion characteristics; (3) modelling of fuel droplet heating and evaporation, using a modified version of the MDQDM with a new transient algorithm referred to as ''Transient Multi-Dimensional Quasi-Discrete Model (TMDQDM)''; and (4) providing a proof of concept with the implementation of the developed model into a commercial CFD code ANSYS-Fluent, for the three-dimensional modelling of complete combustion processes. A case study is made for the CFD modelling of gas-turbine engine using kerosene fuel surrogate. The non-ideal vapour-liquid equilibrium is accounted for, using the Universal Quasi-Chemical Functional-group Activity Coefficient (UNIFAC) model. A new approach to the formulation of fuel surrogates in application to gasoline, diesel, and their biofuel blends (including blends of biodiesel/diesel and ethanol/gasoline) is proposed. This new approach, described as a ''CFSM'', is based on a modified version of the MDQDM. The CFSM is aimed to reduce the full composition of fuel to a much smaller number of components based on their mass fractions, and to formulate fuel surrogates. A new algorithm for the auto-selection of Components/Quasi-Components in MDQDM is suggested and applied to the analysis of fuel droplet heating and evaporation. In contrast to the MDQDM, the new model takes into account the transient contributions of all groups of hydrocarbons, aiming for higher accuracy of the selection of quasi-components than that produced using the original MDQDM. Finally, a surrogate for kerosene is proposed using the CFSM. The model is implemented into ANSYS-Fluent via a user-defined function in order to provide the first full simulation of the combustion process. Detailed chemical mechanism is also implemented into ANSYS CHEMKIN for the combustion study.

Soot is carbonaceous particulate matter formed due to the incomplete combustion of hydrocarbon fuels. The prediction of soot emissions is crucial if next-generation combustion devices are to mitigate the deleterious effects of particulate matter on human health and the environment. Theories for the distribution of size and shape of soot particles in turbulent reacting flows are required, not only for accurate predictions related to flame characteristics but also to meet increasingly stringent regulations. Over the last decades, advances in the understanding of key processes controlling soot formation and oxidation have led to the development of models that can replicate soot emissions and size in laminar flames, sometimes even at a quantitative level. However, predictions in turbulent flames still lack behind due to uncertainties in the intricate coupling between kinetics, aerosol dynamics and turbulence, and the wide range of scales that have to be simulated. There is a clear need to explore the fundamentals of soot evolution in turbulent conditions and develop effective methodologies to predict the soot particle size distribution (PSD) accurately and rapidly. This thesis presents a step in this direction for flames in geometrical configurations of high relevance to aviation combustors. In the first part of the thesis, a comprehensive modelling strategy is proposed using a detailed physicochemical sectional soot model coupled with the Conditional Moment Closure (CMC) turbulent combustion model and Large Eddy Simulation (LES). This modelling approach allows for an elaborate description of the unsteady reacting field and explicitly accounts for transport, history and finite-rate chemistry effects on soot precursors and the PSD. The soot PSD evolution is investigated first in simplified configurations followed by detailed simulations of a canonical turbulent jet flame. The results are analysed to reveal the hierarchy of reaction pathways during soot formation and oxidation and demonstrate the effects of residence time, micromixing and differential diffusion of soot particles. These analyses are necessary to understand the sooting flame structure and act as preparatory investigations for the rest of the thesis. In the second part, a lab-scale swirl flame with addition of dilution air is simulated to explore soot PSD evolution in a Rich-Quench-Lean (RQL) burner configuration widely used in practice for emissions control. Results show a reasonably good agreement with experiments for the mean reaction zone and soot locations and their variations with different airflow provided in the burner primary and dilution regions. The predicted PSDs at the burner exit are fairly well captured for a high-dilution condition but show too few and too small particles for a dilution-free condition, which may be due to an over-prediction of the oxidation rates or the underlying assumptions for particle transport. The results are then used to indicate how dilution air modifies the soot PSD within the primary zone. The method is shown to reproduce the known sensitivity of soot and its precursors on history and scalar dissipation rate effects, a prerequisite for reliable predictions. As a result, it offers a framework for accurately capturing soot PSD in realistic combustion devices. In the third part of the thesis, a new approach based on Incompletely Stirred Reactor Network (ISRN) modelling is presented. The aim is to develop an emissions screening tool that can be utilised during the design phase of combustors. ISRN modelling simplifies calculations so that parametric studies with very complex chemistry and soot models can be performed, or a large number of geometries can be explored, all at a modest computational cost. The approach shares similarities with reactor network and compartmental modelling methods from chemical engineering but offers elaborate molecular mixing and transport treatment. It relies on a network of incompletely stirred reactors, which are inhomogeneous in terms of the flow and mixing fields but characterised by homogeneous conditional averages, with the conditioning performed on the mixture fraction. The ISRN approach is demonstrated on an ethylene model RQL combustor and a single sector lean-burn model combustor operating on Jet-A1 fuel in pilot-only mode, showing very good accuracy in reproducing the mean reaction zone as revealed by LES-CMC or experiments. It is then found that reasonable accuracy can be produced for soot emissions at a significantly reduced computational cost, further enabling the use of multiple chemical mechanisms and soot models and provide estimates of the soot PSD. Finally, a new framework for analysing turbulent non-premixed flames and history effects on soot evolution is presented. The framework is formulated based on the concept of conditional particle age, denoting the total time the mixture or particles have spent at a particular mixture fraction, and conditional thermal age, which allows for a quantification of time-temperature history. Without the need for strong modelling assumptions, governing equations for the two age types are derived that can be used both in the CMC or ISRN context. The approach is then demonstrated on a simple 1D configuration and an ISRN computation of a model RQL combustor. The findings suggest that the concept of conditional age has excellent potential for estimating particle surface reactivity and develop age-dependent closure for soot surface growth and oxidation.

The aim of this study was to understand spray break up behaviour of hydrous ethanol as a fuel for direct injection spark ignition engines. The motivation of using hydrous ethanol is mainly associated with the removal of the expensive distillation and dehydration processes required to produce anhydrous ethanol, that is commonly used in engines as a renewable additive to gasoline in varying percentages across the world. Anhydrous (E100) and hydrous ethanol used in the form of blends including E96W4 (96% ethanol and 4% water by volume), E94W6 (6% water), E90W10 (10% water), E85W15D (15% water) were sprayed by 6-hole injectors inside an optical injection chamber using 150 bar fuel pressure. Iso-Octane was also used as a reference fuel. The fuel temperature was measured at -15 °C, -10 °C, -5 °C, 20 °C, 50 °C, 90 °C and 110 °C at 1.0 bar and 0.5 bar gas pressures to simulate early homogeneous injection strategies for part-load and wide open throttle engine operation at fully warm and ultra-cold engine conditions. Shadowgraphy and side illuminated Mie scattering spray imaging techniques were used. Phase Doppler Anemometry (PDA) was also used to characterize droplet sizing and velocities. It was found that the increase in water content to ethanol in the range of fuel temperatures between 20-90 °C reduced plume tip penetration in ascending order of water content with E85W15D having the shortest penetration. All forms of anhydrous and hydrous ethanol displayed signs of plume merging and spray collapse at 110 °C, 0.5 bar due to flash boiling effects. In general, E96W4 the fuel closest to the azeotropic point, displayed larger outer spray envelope cone angles across all conditions. To study atomization in the absence of plume to plume interactions, high speed spray imaging techniques were used on an isolated single plume of the multihole injector, along with PDA measurements 25 mm downstream of the nozzle hole. At ambient conditions the differences between fuels were minimal but at 110 °C, 0.5 bar, there was an increase in the single plume's cone angle with the increase in water content. Iso-Octane consistently exhibited smaller Sauter Mean Diameter (SMD) droplet sizes than the alcohols. For all fuels, increasing the temperature and reducing the pressure decreased SMD values from levels close to 18 µm down to about 10 µm, with E96W4 showing the largest difference between temperatures. Increasing the temperature and reducing the pressure showed an inverse trend of faster mean droplet velocities. At cold conditions, the increase of water content led to narrowing of the plume and reduction in penetration. An increase in SMD was observed for all fuels as the temperature was reduced, with the highest SMD values in excess of 20 µm for E90W10 at -15 °C. For comparison, pure water injection was associated with an SMD of a similar value at 20 °C. This highlights that, in general, hydrous forms of ethanol up to 15% of water level studied here, are not overly detrimental to the atomization process. Future work in optical engines can highlight combustion characteristics to consider these fuels for widespread use.

Carbon dioxide (CO2) is a global threat, so continuing to improve technologies such as diesel engines that already emit low CO2 levels is vital to ensure the sustainable future of transportation. Internal combustion engines (ICEs) are powerful, but the depletion of fossil fuels and the detrimental effect of emissions on health, the environment, and engine efficiency, are driving innovations in combustion. An emission of concern is soot. Advancing the understanding of injection and combustion processes, especially of less-characterised transients, can help optimise combustion to reduce soot. The use of oxygenated fuels is also promising, but studies of long carbon-chain ones are still needed. The aim of this research was to address these two areas to contribute to improving ICEs, focusing on soot. Experiments were performed in a constant volume chamber, and data were collected using high-speed imaging techniques. The first part of this thesis explored the evolution of diesel transients under inert and reactive conditions. Early injection and end of injection (EOI) transients were investigated under inert conditions for nine pilot-dwell-main combinations. During early injection stages, the main spray's penetration was governed by collision and wake effects; increasing the dwell generally reduced collision. Changes in dwell also affected EOI expulsions after the pilot. For reactive conditions, soot was characterised for single injections after the EOI (aEOI) at ambient temperatures between 860 K and 1310 K. Near-nozzle soot aEOI formed faster as temperature increased. It was revealed that aEOI, up to 75% of soot formed recessively from the flame base and the rest formed progressively from the nozzle. The second part of this thesis identified that C8 oxygenates have similar thermo-physical properties to diesel, showing potential to be alternative fuels. Their combustion was studied neat and as drop-ins. Spatio-temporal flame data were obtained using two-colour pyrometry; the results showed that all fuels had similar flame temperatures, between 1700 K and 1800 K. The sooting tendency increased in the order of ester < alcohol ≤ ketone < diesel, with reductions relative to diesel between 83% and 71%. Dilution and oxygen effects had a larger influence on soot reduction than moiety-specific effects. The C8 oxygenates were also blended with dodecane to have a matching oxygen content of 2.5% and a cetane number of ~ 52. The sooting tendency increased in the order of ketone ≈ alcohol < ester < aldehyde < diesel, with reductions relative to diesel between 55% and 41%. Octanone and octanol had the highest soot reduction potential, but all oxygenates constrained high sooting regions to smaller areas and all reduced soot aEOI. The results showed that even as drop-ins, all oxygenates reduced soot significantly relative to diesel: they show potential to further drive the improvement of ICEs.

In the highly competitive automotive market the share of diesel engine is continuously increasing. They are extensively used in passenger cars as well as in long distance haulage sector vehicles to impart motive force. Regulations for diesel engine emissions as well as public concern for fuel economy have forced the research community to address the combustion and emission issues associated with the use of the diesel engines in general and during the transient operation mode in particular. Turbocharging of diesel engine is the most extensively used technology for the improvement of power density, emissions and enabling downsizing of engines without compromising power. However, turbocharged diesel engine suffer from turbo-lag which is a common phenomenon especially during the rapid transient conditions. Turbo lag causes engine performance deterioration and increased emissions during transient events. The studies on reducing the turbo lag and maintaining the desirable air fuel ratio are considered to be very important for making turbocharged engine compliant with the current environmental regulations. The thesis covers the important aspects of a turbocharged diesel engine with a specifc focus on the transient response of the diesel engine. In the first facet of this research, computer-based investigations using commercialengine simulation packages are performed to simulate the transient response of the system using different methods. Torque assistance of 0.16 Nm reduces the turbo lag by 3.6 sec for both compressor exit pressure and compressor speed. Optimum value of inertia reduction is found to be -10% which reduces the turbo lag by 2.9 sec for compressor exit pressure and by 0.6 sec for compressor speed. The effect of 2.5 bar air injection is found to reduce the turbo lag by 3.9 sec for both compressor exit pressure and compressor speed. A comparison is made for the assessment of relative improvement in transient response brought by the three methods and based on this comparison, air injection system is chosen that yields maximum benefit in the performance of the engine. The experimental results from the real CI engine is used to validate the simulation model. A good agreement is achieved between the simulation and experimental results. The effect of air injection on a heavy-duty turbocharged diesel engine under various operating conditions forms the second facet of this study. The turbocharger response parameters are analyzed under the effect of air injection for different transient operating conditions of speed and load transients. For speed transients, considering the energy imparted for air injection, 1 bar is the optimum injection pressure with turbo lag reduction per unit energy as 0.290, 0.392 and 0.555 per joule for compressor exit pressure, turbine inlet pressure and turbine inlet temperature (TIT) respectively. Faster recovery time is noted for 1 sec rapid acceleration than that of 2 sec with the application of air injection. For load transients, optimization of air injection is also performed for injection pressure and orifice diameter. Under constant load step, maximum improvement in turbo lag reduction is observed at 1000 {1400 rpm whereas its effect on maximum attainable value is at 1600 { 1800 rpm. For load magnitude variations, the optimum injection pressure of 3 bar at 15 mm orifice diameter brings more improvement for lighter load than for stronger load. Maximum improvement is noted for 50-70% load step. 3 bar air injection at 15mm orifice diameter is the optimum injection pressure which brings more improvement in terms of turbo lag reduction for faster load application. The effect of this optimized air injection is more beneficial for 1 sec load schedule than for 2 sec. During the transient operation of diesel engines, exhaust emissions are the imperative issue that needs to be addressed. In the third facet of the research exhaust emissions are analyzed to make sure that the applicationof air injection technique does not need compromise on emissions. The air injection technique is evaluated for exhaust emissions through simulation. 3 bar air injection at 10 mm orifice diameter is the optimum air injection for emissions under speed transient which satisfies Euro 6 standard for CO and HC emissions. This air injection reduces the concentration of CO and HC emissions by 5% and 0.4% respectively. Under load transient the optimum value of air injection is 1.2 bar at 10 mm orifice diameter which reduces the concentration of CO, HC and NOx emissions by 0.8%, 0.01% and 0.4% respectively. The study reveals that air injection technique while improving the transient response of the turbocharged diesel engine doesn't increase the emissions, highlighting the magnitude of contribution of the technique to the overall performance of the system. The novelty in the research is the compilation of these methods in a cohesive approach of modeling the transient response of turbocharged engine system.

The research involves a feasibility study of micro gas turbine (MGT) performance. The engine is based on off-the-shelve turbochargers provided by Cummins Turbo Tech Ltd (project sponsor). The intended applications of the MGT are: hybrid electric vehicles (HEV) and small portable power-plant (SPP). A market research was initially conducted to assess the scope for MGT. Following this, a preliminary parametric design point (DP) study, using an in-house modified code, has been performed on relevant Brayton cycles to choose the optimum DP in contrast to the market research. It was followed by a full off-DP (ODP) analysis. Actual component maps were used to model the output. Analysing the holistic DP and ODP parametric studies, an optimum cycle has been defined. A full thermo-economic analysis was conducted for this cycle to evaluate the applicability of the final MGT for HEV or SPP. Following this, a vehicle performance analysis has been conducted using the final MGT performance maps. Models capable of analysing HEV thermodynamic and kinematic performance were developed using a Simulink based analysis tool called QSS. The performance has been compared with other HEVs and standard vehicle output. A comparative analysis has been presented to assess the MGT based HEV performance. Following the above performance analyses, an experimental study has been pursued to investigate the effects of swirler, primary jets and various side-entry injection mechanisms on combustor flame-tube aerodynamics. Results show the influence of these factors on primary zone recirculation and exit plane pressure and axial velocity distribution. These factors are crucial for efficient combustion. A mathematical model for the recirculation zone has also been devised based on the empirical findings. Future scopes of the project, such as: further studies on flame tube aerodynamics of the sideentry combustor, MGT test rig setup, and heat-exchanger testing, have been described.

Combustion engines' exhaust emissions have impacted the environment with greenhouse gas emissions on a large scale and this will reach a global catastrophe limit in the coming decades. It has been an important issue of consideration for many years and substituting fossil fuels to decarbonise the environment has been of utmost importance. With the increase in knowledge and research to understand and control particulate matter emissions, the fundamental research still holds unanswered questions. The study carried out in the following thesis is primarily focussed on the particulate matter's inception, evolution, control and its characterisation from a compression ignition engine. The thesis proceeds with the initial study on the evolution and course of particulate matter inside the exhaust tailpipe using a zero-dimensional numerical model. The model aims to investigate the nucleation of water and sulphuric acid from the engine out, and its impact on the particulate matter as it is transported autonomously along a 3.5-metre exhaust pipe. The research is also concerned with explaining the effects of exhaust temperatures on particulate matter and gas emissions, by characterising them into size, mass, and concentrations at consecutive testing positions. The simulated and analysed data are used for a comparative analysis with empirical results acquired from similar exhaust temperature and particulate matter conditions that were confirmed as the assumptions were established. Further, the research is based on an empirical investigation of particulate matter evolution and the impact of the external cooling of an exhaust tailpipe. The cooling was produced using copper coil tube windings with a decreasing pitch along the length of the pipe and supplied with an ice water and antifreeze mixture solution. The external cooling of the exhaust tailpipe was an important parameter to study the effects of external cooling on the particulate matter flowing through the internal space of the tailpipe. The evolution of the particulates and the impact of the reduction in temperature gradient provided agreeable results. The objective was achieved in understanding and contributing to the knowledge of particulate behaviour inside the tailpipe under various engine operating conditions. In consideration of the previous studies mentioned above, it is critical to research the control of particulate matter and gas emissions at this stage. Hence, a diesel particulate filter is equipped as an exhaust after-treatment system for the abatement and oxidation of toxic gases and particulate matter. A catalyst is developed to be coated on the filter substrate with a novel nano-fibrous morphology using a rare-earth metal catalyst. The conceptualisation of the research was to investigate the morphological effects on particle trapping and the oxygenated catalytic effect on soot burn at low exhaust temperatures. Tests were performed at laboratory scale and test bench scale, where the filter substrate was coated with the catalyst; and the results acquired depicted an increased filtration efficiency consistent at 95-99%, and a high oxidation and continuous regeneration rate at reduced local exhaust temperatures, contributing to overall lower back pressure on the particulate filter and engine. Finally, the thesis provides details of the research findings and conclusions to provide a valuable contribution to the knowledge of exhaust emissions' characteristics and their control.

An on-going challenge with Gasoline engines is achieving rapid activation of the three-way catalyst during cold starts, in order to minimise pollutant emissions. Retarded combustion can help achieve rapid light-up of the three-way catalyst and can be facilitated by stratified charge using late injection. Injecting late in the compression stroke however, provides the fuel insufficient time for fuel entrainment, resulting in locally fuel rich diffusion combustion. Employing a split injection strategy can help tackle these issues. The effects of a split injection strategy on the spray characteristics and in-cylinder charge formation are investigated in the current study. Varying pulse width (PW) combinations, split ratios and dwell times are investigated, with pressures of up to 35MPa, using a state-of-the-art solenoid actuated high pressure gasoline injector. The experiments were performed in a constant volume spray chamber. The droplet velocities and sizes were measured using Phase Doppler Anemometry. Short and large PWs, in the range of 0.3ms to 1.5ms, were investigated. The results revealed that the highest injected quantity of fuel was measured with the shortest dwell time of 2ms, owing to increased interactions between the injection events, which led to larger drop sizes measured. The drop sizes from the short PW of 0.4ms were generally larger than 0.8ms PW, due to closely spaced opening and closing events of the solenoid valve. The high injection pressure had also resisted the timely closing of the Solenoid valve when short PWs operating in the ballistic zones were used. This led to larger overall duration of injection. The studies on the charge motion using split injections are performed inside an optical Gasoline engine using high-speed particle image velocimetry in the tumble and Omega-tumble planes, at a repetition rate of 10KHz. The engine's conditions were representative of low-load operations. The results revealed that a small split ratio of 25%-75%, with both injections in the intake stroke, was effective at generating a flow field with high turbulence levels close to the spark plug, when compared to 75%-25% split ratio. The injection coupled with inlet valve opening formed strong tumble charge motion, which was preserved throughout the compression stroke. This provides favourable conditions for fast flame propagation. The fuel injection timings which maximised interactions with the piston surface were detrimental for mixture formation due to heavy surface impingement. The findings from the study helped determine the optimum split injection properties.

In recent years, the exploration of new combustion technologies has accelerated due to new stringent emissions regulations and fuel economy requirements. Virtual engineering tools, that enable the screening of non-traditional hardware and engine calibration at the early stage of engine development, have become imperative to meet new emission regulations. In the current engine development process benchmarking and historical test data, are used to carry out simple 1-D engine system calculations and define the overall engine concept design. Later, to provide a definitive design ready for prototyping, more complex Computational Fluid Dynamics (CFD) calculations are coupled to 1-D engine system codes to optimise initial concept geometries and high-level calibrations. However, to provide meaningful results, 1-D engine system codes often use empirical based combustion models that require an initial input, called engine burn rate. Realistic engine burn rate responses, for the entire engine map and for different design concepts, are also required to provide 3D CFD codes with correct boundary conditions during the design optimisation phase. Thus, the engine burn rate of new combustion technologies, for which little experimental data is available, need to be initially assumed. To improve the predictive capabilities virtual engine development processes, the industry’s attention shifted towards Quasi-Dimensional (Q-D) combustion models capable of providing engine burn rate predictions. However, within the Q-D modelling framework, turbulence models, adding extra user-input variables, are required to capture the effect of different combustion chamber geometries on the engine combustion rate. Rigorous validation of Q-D turbulence models for different engine concepts and engine maps is needed to enable Q-D combustion models to predict the engine burn rate. Therefore, an alternative methodology characterised by limited dependency on previous test data is required to enhance the exploration of novel combustion strategies and geometric architectures. In this thesis, an alternative engine development process that uses a combination of a Q-D combustion Stochastic Reactor Model (SRM), a 1-D engine system model and noncombusting, “cold” CFD calculations, is proposed. The SRM code captures the combustion chemistry in a computationally efficient manner but does not capture in isolation geometric variables such as port and piston geometry. To account for that, the approach uses limited non-combusting CFD baseline calculations to characterise the engine in-cylinder flow of each screened engine concepts. A physics-based scaling factor response was developed and used to provide the SRM with the correct turbulence input, known as scalar mixing time (τSRM). The response was assessed against four different engine variants over a variety of engine operating conditions. The same response was used to predict the effect of different bore to stroke ratios (B/S) on the engine combustion rate and knock tolerance. Non-combusting CFD and 1-D engine system simulations have been carried out to investigate the effect of different engine variants and operating conditions on the in-cylinder turbulence. It was shown that τSRM of different operating conditions can be scaled to the intake flow velocity predicted by 1-D engine system analysis. This allows to predict the engine RoHR at the explored engine variants and operating conditions within the experimental standard deviation. The presented methodology showed augmented predictive capabilities and has potential to move the engine development towards a less hardware dependent approach for the exploration of new engine concepts.

Meeting the ever-growing energy demands of the world while not sacrificing energy security/environmental sustainability by relying on a single fuel source means that combustion systems must demonstrate fuel flexibility. Currently, the gas turbine burns a wide range of fuels and fuel combinations. However, there seems to be a limit to an exploration of its fuel flexibility. Whereas there are numerous investigations into multiphase fuel combustion in other internal combustion engines like the diesel engine, there is a dearth of such studies for the gas turbine. Consequently, this thesis investigates the simultaneous combustion of practical liquid and gaseous fuels in a 20 kW swirl-stabilised gas turbine relevant combustor. The investigation involved developing a dual-phase fuel injection system capable of handling diesel/methane, diesel/syngas, biodiesel/methane, biodiesel/syngas and blends of methanol/glycerol co-combusted with methane. The effect of partly replacing the liquid fuel with a gaseous type fuel on combustion characteristics like flammability limits, flame stability, flame structure and exhaust emissions were studied for the diesel and biodiesel blends. The gas substitution ratio was based on heat energy contribution in such a manner that a certain percentage of a desired heat output is contributed by the gaseous fuel and the balance by the liquid fuel. The nature of non-reacting flows in the system, including air flow and liquid fuel spray was also investigated using CFD while experimental measurements were supported using numerical chemical kinetics modelling. Flame extinction tests proved that as gas substitution ratio increases, flammability limits decrease owing to changing non-reacting and reacting flow dynamics. Intermediate combustion species chemiluminescence imaging was key to the investigations and was used in evaluating reaction zone characteristics and flame stability. These parameters as well as exhaust emissions were assessed as test conditions were varied. It was important, for the methanol/glycerol blends, to establish the feasibility of its combustion without retrofitting the burner used for the other blends in order to prove its practicality. Thereafter, the influence of methane addition on flame structure and stability was investigated.

Compression Ignition (CI) engines are state of the art power generating machines, and as such have found widespread use in a variety of implementations due to their high thermal efficiency, fuel efficiency, durability, reliability, and low carbon dioxide (CO2) emissions. These include power generation, marine propulsion, light and heavy duty road vehicles, and off-road applications. However, due to the nature of a CI engine’s combustion process, emissions of other harmful pollutants is increased, such as nitrogen oxides (NOx) and particulate matter (PM). Multiple injection strategies have been used to combat this rise in emissions; along with low temperature combustion (LTC) methodologies these have been able to reduce both NOx and PM emissions, and as such this thesis will focus on the modelling of the transient period of the diesel spray and its impact on the combustion of the fuel. Initially, this thesis will consider the transient period of diesel injection by focusing on the prediction of the early stage of the spray formation at the Engine Combustion Network’s (ECN) "Spray A" condition, comprising of a single-hole injection of ndodecane (diesel surrogate) fuel. This is achieved by comparing two different commercially available Computational Fluid Dynamics (CFD) codes and their predictions of the liquid and vapour lengths, initially with different computational set-ups before these set-ups are converged to being identical with each other. All simulations are undertaken under a Reynolds Averaged Navier Stokes (RANS) framework, in a well characterised domain for both CFD codes. This convergence of set-ups shows that the transient region of the spray is highly dependent on the break-up model, however comparison with experimental data showed a deficiency in the implementation of the break-up model within Star-CD. This was corrected with the inclusion of a novel break-up length criterion, with the corrected model showing good agreement with experimental data, with particular strengths in decoupling the liquid and vapour length predictions. Following the implementation of the novel break-up length criterion within Star- CD, the performance of this model at a combusting condition is tested. This study was performed under the same framework as previously, however with the implementation of a commonly used chemical mechanism for n-dodecane combustion. When the novel break-up length criterion is compared to the original baseline case within Star-CD the results match very well to each other, with predicted ignition delays, lift-off-lengths and combustion fields being closely aligned. An over-prediction in lift-off-length to experimental data is noted, however this is commonly seen for the mechanism used. Finally, by utilising the decoupling of the liquid and vapour penetrations offered through the novel break-up length criterion, the impact of the vaporising match on the combustion criteria detailed previously is investigated. A variety of cases are considered, with high, low and matched variations on both the liquid and vapour lengths compared against each other. The results from these tests show a strong effect of certain model constants on the combusting criteria, with break-up model constants especially having a large impact on the mixture fraction and temperature predictions. In contrast, the turbulence model constants often used when matching simulated tests to experimental results have very minimal impact on either the mixture fraction or temperature fields, with only the position of the combustion field changing, as expected. The effect of the combustion field position on the combustion temperatures is also considered, further reinforcing the break-up model constant’s impact on combustion prediction.

Within the last decades, the turbocharger has become a key part of the internal combustion engine and thus is no longer seen as optional. The turbocharger is especially crucial for heavy-duty engines for on-road truck applications as it addresses two main development goals simultaneously, which are typically oppose each other: lowest fuel consumption and meeting all emission regulations. To enable down-sizing, which is a common strategy to fulfil these two aims, the turbocharger must be well matched to the internal combustion engine and furthermore, must operate as efficiently as possible. The latter implies further optimisation of the turbocharger's components, e.g. the compressor, which supplies the engine with compressed air. As the compressor already achieves high efficiencies, further optimisation requires substantial efforts, a full understanding of the component and new development methods. To address the described challenges, a systematic methodology is developed and applied in this thesis to improve the efficiency of the baseline turbocharger compressor while achieving the same operating range. To also guarantee the mandatory, good matching between the engine and the turbocharger, the main operating points of the turbocharger compressor stage are defined for the real-world operation in a first step. A CFD setup is developed and validated extensively for the baseline geometry especially to analyse the flow field at these main operating points in detail. Two novel approaches are introduced and applied to the CFD results to determine the location and to further quantify the intensity of the most significant types of losses. Taking the resulting loss distribution into account, design parameters are defined which promise a high impact on specific kinds of losses and thus also on the stage's performance. In a next step, the selected design parameters are varied simultaneously applying a systematic Design of Experiments (DoE) plan. The resulting geometry variants are generated automatically using parameterised CAD models and they are evaluated via CFD simulation with an automated work flow. The optimisation is conducted based on mathematical surrogate models and results in design proposals promising improved compressor stage designs. Five of these design proposals are selected and their performance is validated against hot-gas test-rig measurements. Under real-world operation, the best performing compressor design is considered to achieve in average an approximately 1% higher efficiency. Additionally, the operating range is increased in average by around 10%. Furthermore, the best performing compressor design is compared to the baseline geometry in detail, also analysing the changes in the aerodynamic loss distribution leading to this improvement.

This research presents low engine-speed behaviour analysis of a two-stage inseries turbocharged air-path diesel engine. A variable geometry turbine (VGT) at the high-pressure stage and a high-pressure exhaust gas recirculation (EGR) path are included. A 1D computational fluid dynamic model of the air-path is simulated using Ricardo Wave and analysed for steady-state and transient behaviour. A mean value model of the air-path is constructed in MATLAB and validated against theWave model for model based control. First it is demonstrated that operating single-stage turbocharged diesel engines under steady load at low speeds is fuel efficient but the resulting reduction in power due to limited available air and torque will make the low-speed operation difficult during load transients. The engine's ability to track load transients is limited by emission constraints due to the rate of production values for smoke and nitrogen oxides (NOx). Then the proposed two-stage in-series air-path configuration is shown to meet performance expectations by extending the air-path operating range and reducing its response time at low engine-speeds. The configuration improves the typical part-load performance of regulated two stage in-series arrangements at low engine-speeds through closed loop adjustments to the turbine expansion ratios. Better EGR rates (NOx reduction) at low engine-speeds can be achieved while the engine transient response is maintained. The resulting interactions are complex and for optimal behaviour multivariable control with control objective switching during steady load and transient load conditions is proposed. The control proposals are validated in simulation using MATLAB and calling the Wave model.

The Intergovernmental Panel on Climate Change (IPCC) have conveyed in their fifth assessment report that anthropogenic emissions and endeavours are responsible for approximately 100 % of global warming since 1950 [1] and electricity and heat generation account for 41 % of the 32.8 billion tons of global CO2 emission from fossil fuel combustion in 2017 [2]. There is thus a sense of urgency to capture CO2 from large point sources of fossil fuel combustion to limit global temperature rise. Natural gas combustion is envisaged to play a fundamental role towards a zero-carbon economy as opposed to coal and oil due to its low carbon content. However, capturing CO2 from natural gas combustion, which emits about 6.7 billion tonnes of CO2 in 2017 is challenging as it, bestows a parasitic energy penalty on Natural Gas Combined Cycle (NGCC) power plants. This is due to the low partial pressure of CO2 in the flue gas of gas turbines, which necessitate that substantial reboiler heat duty is employed for solvent regeneration. To address the aforementioned impasse, pertinent experimental campaigns at the UKCCSRC- PACT National Core Facility were carried out to simulate Selective-Exhaust Gas Recirculation (S- EGR) under the influence of 40 wt(%) of Monoethanolamine (MEA). This was to enhance the driving force behind CO2 capture and to reduce the Specific Reboiler Duty (SRD), consequently counterweigh against the forfeit on the power plant’s productivity. Furthermore, the impact of varying Pressurized Hot Water (PHW) temperature at the inlet of reboiler was studied. The influence of oxidative degradation of the amine solvent at 15 vol(%) of O2 and 5 vol(%) of CO2 has been experimentally investigated. Results from these studies have demonstrated that Selective Exhaust Gas Recirculation (S-EGR) is favourable in reducing the solvent regeneration energy requirement by about 25 % at CO2 concentration of 6.6 vol(%) prior to flue gas introduction in the Post-combustion Carbon Capture (PCC) system. PHW temperature at 125 °C was identified to give the lowest SRD by 6 % against the baseline SRD. Detection of Dissolved Oxygen (DO) peaks was observed as water from the water-wash column was transferred to the absorber column which may have a possible impact on the oxidative degradation of the amine solvent in the PCC system. The concentration of the Iron in the amine solvent, which is a key indicator of the solvent decay increased by approximately 10 times from 3.68 to 36.20 mg/l over a course of 545 hours of experimental operation. Results and recommendations from these studies will potentially reduce the solvent regeneration energy requirement of the next generation PCC technologies and facilitate the global deployment of such technologies towards decarbonisation of the fossil fuel combustion industries and strengthening the efforts of limiting global temperature increase.

Downsized gasoline turbocharged direct injection (GTDI) engines deliver superior fuel economy by operating the engine at higher loads but become prone to knocking combustion at boosted operations, which requires the application of knock mitigation strategies, such as the retarded spark timing, with a negative impact on the engine performance and efficiency. Furthermore, the use of wide valve-overlaps to maximise positive scavenging by elevated intake pressure at low and medium engine speeds leads to greater tailpipe NOx emissions. In light of increased use of Real Driving Emissions (RDE) test where higher load operations are far more prominent, there is a strong need to further explore the approaches to improve engine efficiency and lower harmful emissions at knock limited operations. This project investigates the use of stratified flame ignition (SFI) combustion with exhaust gas recirculation (EGR) on a downsized GTDI engine. EGR dilution is added to control the knocking combustion to replace the traditional knock mitigation strategies. The subsequent combustion is further improved by stratified fuel injection and uprated ignition system. The valve-overlap duration is shortened to avoid the air short-circuiting and hence reduced tailpipe NOx emission through greater conversion efficiency of the 3-way catalyst, but with trade-off with lower volumetric efficiency and knock onset. The novelty of the study is identified as the combination and optimisation of these strategies to improve operational efficiency and reduce harmful tailpipe emissions at knock limited loads. The results showed that the EGR dilution lowered knock tendency and high energy (HE) ignition accelerated the combustion and recovered stability exclusively at boosted operations. Split injection strategy showed fuel consumption benefit at very limited cases, and most cases led to the loss of efficiency from slower less efficient combustion. Through computational fluid dynamics (CFD) analysis, this was found to be caused by the unfavourable mixture preparation of the multi-hole injectors due to high spray penetration and insufficient mixture preparation time. However, the combined use of EGR dilution and reduced valve-overlap improved mixture preparation due to increased charge temperatures and induced turbulence. Indicated specific fuel consumption (ISFC) improvements of 4.8% and 5.2% were achieved at 13.7bar and 16.4bar Net_IMEP at 2000rpm, respectively. The reduction of valve-overlap also improved the combustion efficiency and reduced emissions of tailpipe NOx and particulates due to eliminated short-circuit air and enhanced turbulence for faster mixture preparation. Hence, a synergy between valve-overlap reduction, split injection, and EGR dilution was found, and the proposed strategy successfully lowered fuel consumption and harmful emissions from this combined synergy effect.

Gas turbines in combined cycle are the cleanest and most efficient form of large-scale thermal power generation, and their use is expected to increase in the future. The most arduous conditions within a turbine are faced by the first stage nozzle guide vanes (NGV). These are located immediately after the combustion chamber and face the highest thermal loads, as well a a complex flow field influenced by combustor conditions. For this thesis, experiments and computational methods are developed and used together to investigate steady and unsteady heat transfer and aerodynamics in the nozzle guide vanes It is beneficial to reduce the heat transfer to components in order to increase efficiency and maximise component life. Improvements were developed to experimental methods which could be carried out in a scale experimental rig, which operated at realistic Reynolds number and Mach number. Thin film gauges were developed for heat transfer measurements at higher sensitivity and spatial resolution than has been possible in the past. The use of a large numbers of unsteady pressure sensors has been pioneered successfully. Improved experimental techniques are applied to optimising heat transfer and aerodynamics resulting from realistic features which are often overlooked in simplified studies. Several combustor chamber geometries are compared, finding that a combustor design that minimised length scales of turbulence structures reduces NGV heat transfer coefficients significantly. This shows the benefits which can be achieved by designing combustor and turbine together in a holistic manner. The unsteady effects of large-scale flow structures generated in the combustion chamber were observed in detail by high resolution unsteady measurements and LES. Instantaneous flow phenomena which occur are measured experimentally, and are explained physically with the support of LES. Instantaneous heat transfer events are found to be sufficient to cause potentially damaging temperature fluctuations to thermal barrier coatings, despite being disguised in time-averaged measurements. The impact of the mid-passage gap between vane platforms on heat transfer and aerodynamic efficiency is measured. An analytical model is developed, which can be used to assess the importance of such gaps in the general case. It is found that great improvements are possible if gap sizes are reduced below a critical threshold. The optimum distribution of cooling flows required to prevent ingress can be calculated by means of a one-dimensional equation, with no need for iterative optimisation. Ingress into circumferential gaps is also investigated, and found to cause damaging local heat transfer. A measurement campaign investigating surface texturing effects on heat transfer is also presented.

Thermoacoustic instabilities that arise in the context of lean premixed combustion are a major challenge in the development of gas turbines and aero engines. This thesis contributes to the research on modal analysis of thermoacoustic instabilities by applying linear stability analysis and sensitivity analysis of longitudinal thermoacoustic modes within a low order network model framework. The first part of the thesis is dedicated to the better understanding of different types of modes, particularly the interplay between classical acoustic modes and intrinsic thermoacoustic (ITA) modes in a simple thermoacoustic system. Anticorrelated modal sensitivities are found to arise due to a pairwise interplay between acoustic and ITA modes. The magnitude of the sensitivities increases as the interplay between the modes grows stronger. The results show a global behaviour of the modes linked to the presence of exceptional points in the spectrum. Non-normal behaviour and its consequences are also investigated in this setup. The use of sensitivity information is shown to be capable of quantitatively assess the degree of coupling and decoupling of the thermoacoustic modes \francol{from} the acoustic components of a simplified geometry. The investigation shows the effect of an area expansion on the modes of a thermoacoustic tube. The last part of the thesis proposes an optimisation procedure based on the sensitivity information to identify the optimal volume and placement of dampers, which are used to damp thermoacoustic oscillations.

Despite the improvement in fuel economy and engine performance, gasoline direct injection (GDI) engines have been relatively recently identified as a prominent source of ultra-fine particulate matter (PM) which mainly consists of soot. Adverse health impacts of these airborne particles call for stringent legislations to monitor both the number and mass emissions. Set against this background, an integrated computational fluid dynamics (CFD)-chemical kinetic modelling approach is developed in order to acquire an in-depth understanding of PM processes within GDI engine configurations. Given the nature of spark-ignited engines, GDI combustion is characterised by flame propagation under high-temperature conditions. To represent gasoline fuels, reduced mechanisms of toluene reference fuel (TRF), coupled with the chemistry of polycyclic aromatic hydrocarbon, are evaluated using ignition delay (ID) time, laminar flame speed and speciation data as validation targets. The optimal TRF mechanism predicts the ID times of stoichiometric gasoline-air mixture at 55 bar with the highest deviation of 13% throughout 900-1160 K. The effect of high pressures at 20 bar and 25 bar on the flame speeds of a binary surrogate is also captured accurately (< 14% difference). In addition, the concentrations of PAH from flame burners and reactors are computed within the same order of magnitude as the experimental measurements. The selected TRF mechanism is subsequently incorporated to CFD modelling of the GDI engine. A complete set of CFD sub-models is formulated to describe the whole panoply of in-cylinder events numerically, covering turbulence, spray, fuel impingement, liquid fuel film, ignition, combustion, flame propagation and emissions. The dynamic multi-zone partitioning method is introduced within the Detailed Chemistry model to expedite the calculations. Along with the sectional method for soot modelling, the resulting computational time only increases by 13% as compared to the modelling approach that relies on tabulated chemistry. Based on the reference case at 2300 rpm and 90 N m, in-cylinder pressures are replicated with the peak value predicted within a 1% margin. Computed number density of soot differs from the tailpipe measurement by 8% while its mass density is under-predicted by a factor of two. The modelled particle size distribution function captures the decreasing trend in particle number as the size becomes larger within the range of 10-100 nm. The integrated approach is extended across different speed-load points whereby engine speed is varied from 1600 rpm to 3000 rpm while engine load is changed by altering the torque from 60 N m to 120 N m. The effects of speed and load are manifested in terms of in-cylinder flow conditions and mixing time, thus affecting the mixture distribution at spark ignition timing. A conceptual model describing the ultra-fine PM processes in the wall-guided GDI engine under the homogeneous stoichiometric mode is developed. The dominant PM formation mechanisms are the presence of fuel-rich regions and remaining liquid fuel droplets at the onset of spark event. The mixture inhomogeneity is attributed to film stripping and evaporation from the liquid film deposited due to spray impingement. Piston wetting is the most severe, accounting for up to 76% of the maximum film mass. Pyrene and acetylene contribute towards the increase in soot particle number and mass, respectively while post-flame oxidation becomes effective at high-temperature regions with large concentrations of hydroxyl radicals. Based on the reference case, three engine operating parameters are varied in isolation to examine their effects on PM emissions. In the order of decreasing sensitivity, they are ranked as fuel injection timing, spark ignition timing and rate of exhaust gas recirculation. Overall, the detailed fundamental understanding of PM processes obtained from this study is beneficial in strategy optimisation for PM mitigation in GDI engines.

The gas turbine has been widely used for mechanical drive, electric power generation and jet propulsion. Challenged by strict CO2 emission regulations and competition between manufacturers, designers push the boundaries for an ever more efficient gas turbine. The cycle efficiency of the gas turbine is a crucial parameter that drives the engine performance and fuel burn. Improved cycle efficiency can be achieved by raising the turbine entry temperature, as well as the pressure ratio across the compressor. Today’s gas turbine engines use bypass cooling air from the high-pressure compressor, to cool down the seal disc cavities and extend the life cycle of critical components in the stage. The cooling air (secondary system) can account for 25% of the total bypass air, approximately 5% of the total bypass air is used to seal the disc cavities. To further enhance the benefits of the sealing flow (cooling air), rim seals are fitted at the periphery of both, stationary and rotating discs. Rim seals help to further reduce the ingestion of hot mainstream gases (combustion chamber gases) that can cause damage to the discs and blade roots. The cause of hot gas ingestion is principally due to the circumferential pressure asymmetry in the mainstream flow, furthermore, the mixing between the sealing flow and mainstream gas, results in a deterioration of aerodynamic performance. The first part of this thesis concerns experiments using a 1.5-stage turbine rig, from which the flow physics associated with ingestion could be studied. This thesis presents experimental results using the turbine test rig with wheel-spaces, upstream and downstream of a rotor disc. Ingress and egress were quantified using a CO2 concentration probe. The probe measurements have identified an outer region in the wheel-space and showed a flow structure consistent with Batchelor-type flow. This is the first time asymmetric variations of concentration have been shown to penetrate through the seal clearance and the outer portion of the wheel-space. For a given flow coefficient in the annulus, the concentration profiles were invariant with rotational Reynolds number. The measurements reveal that egress flow provides a film-cooling benefit, on the vane and rotor platforms. The second part of this thesis investigates the effects of ingress through a double radial rim seal. The effect of the vanes and blades on ingress was investigated by a series of carefully controlled experiments: firstly, the position of the vane relative to the rim seal was varied; secondly, the effect of the rotor blades was isolated using a disc with and without blades. Measurements of steady pressure in the annulus show a strong influence of the vane position. The relationship between sealing effectiveness and purge flow rate, exhibited a pronounced inflexion for intermediate levels of purge; an inflexion did not occur for the experiments with a bladeless rotor. Shifting the vane closer to the rim seal, and therefore the blade, caused a local increase in ingress, at the inflexion region; again, this effect was not observed in the bladeless experiments. Unsteady pressure measurements revealed the existence of large-scale flow structures (flow instabilities) which depended weakly on the vane position but strongly on the sealing flow rate. Unsteady pressure was measured with and without the blades on the rotor disc. In all cases, the flow structures rotated close to the disc speed. The third part of this thesis involves, experiments and computations of flow through a gas turbine chute seal. The study investigates ingress and the phenomena of flow instabilities. The aim of this study is to investigate the steady and unsteady flow features, near the rim seal mixing plane, for a geometrically scaled stage with a chute seal. The work presented here forms part of a future partnership with the KTH Royal Institute of Technology, to study the effect of engine scaling on ingress. Experiments and computations for; pressure, swirl and sealing effectiveness, were done at an engine representative turbulent flow (lT) condition. Under this condition, an engine representative wheel-space flow structure exists. CFD results suggest that flow instabilities influence ingress, instabilities caused by a shear interaction between the egress and mainstream. This supports the hypothesis of flow instabilities being driven by shear gradients. The final part of this thesis investigates re-ingestion. A mixture of upstream egress flow and mainstream, gets re-ingested in the downstream wheel-space. Re-ingestion was experimentally measured and evaluated using a model developed by Prof. Mike Owen. To the author’s knowledge, it is the first time this methodology has been applied to quantify re-ingested fluid in downstream wheel-space. The results showed that the upstream egress does not influence the fluid dynamics, downstream of the rotor blades. The experiments were conducted at incompressible flow conditions. Probe concentration measurements demonstrated that, an interaction occurs between the re-ingested fluid and the downstream egress flow, at the rim seal mixing region. Re-ingestion was evaluated for a range of sealing flow rates. It was shown that the mass fraction of re-ingestion increases with increasing downstream seeded sealing flow rate.

Porous multi-wall cooling schemes such as double-wall cooling combined with effusion cooling offer a practical approximation to transpiration cooling which in turn present a potential for high cooling effectiveness. Most of the existing double-wall effusion cooling studies have been on flat plate geometries. There are varying external static pressure and secondary flow features in flow over an aerofoil which have an influence on overall cooling performance. Thus, there is need to study double-wall effusion cooling applied to an aerofoil. The main aim of this research was to extend the double-wall effusion-cooling technology research, that has long been undertaken on flat plates, onto a gas turbine blade. In the present study, both numerical simulations and experiments were undertaken to study double-wall effusion-cooled (DWEC) turbine-representative aerofoils. The aerofoils were built from double-wall block elements that have been validated by another author. Both low porosity and high porosity circular and diamond pedestal designs were considered. A novel decoupled numerical analysis tool for preliminary cooling performance analysis of DWEC aerofoils was first developed. In this analysis method, a modified flat plate correlation from the literature was used to represent the two-dimensional distribution of film cooling effectiveness. The internal heat transfer coefficient was calculated from a validated conjugate analysis of a wall element representing an element of the aerofoil wall and the conduction through the blade solved using a finite element code in commercial CFD solver. The developed decoupled numerical analysis method was validated using results from fully coupled conjugate heat transfer (CHT) simulations. In addition, high-speed experimental tests at engine representative Mach and Reynold numbers flow conditions were carried out to study film cooling effectiveness over the full surface of three circular and six diamond pedestal DWEC blade designs using pressure sensitive paint. All the blades were tested within a range of representative modern engine coolant mass flow rate to mainstream mass flow rate ratios; 0.5% to 5.5%. The novel simplified numerical analysis method offered good performance approximation particularly on the suction surface of the aerofoil. In addition, compared to CHT, the novel simplified numerical analysis method reduced computational time by approximately 50 times and therefore, computationally efficient for use during preliminary design and optimization stages. High effective porosity designs exhibited better film cooling effectiveness, than the low effective porosity counterparts, but this came at an expense of internal cooling efficiency. CFD results compared well with the experiments and were able to capture similar film effectiveness trends on both the pressure surface and the suction surface, however, inability of Reynolds-Averaged Navier-Stokes equations (RANS) models to correctly predict diffusion resulted in an overprediction of film cooling effectiveness around the vicinity of film cooling holes and an overprediction of film superposition on the suction surface. This work has contributed knowledge of the DWEC aerofoils performance including overall cooling effectiveness predictions, internal cooling effectiveness predictions and film cooling effectiveness performance measurements. There is still much work to be done (including investigation into aerodynamic losses, weight and stresses associated with this cooling technology) to realise a practical double-wall effusion-cooled blade. The present author has included recommendations for future work.

Environmental pollution has been a major point of interest in recent years, and aerospace gas turbine combustion systems must address increasingly more stringent emissions requirements. One potential way of addressing these requirements is the introduction of lean burn combustion technology. However, lean burn fuel injectors are much larger, relative to traditional rich burn injectors. This leads to increased aerodynamic interactions with upstream and downstream components and a highly non-uniform feed to the injector. The impact of this on the combustion process is currently not well understood. Importantly, these effects are not accounted for in typical test facilities used for injector development, as these are generally plenum fed. Hence, the main aim of this work is to study the impact of these aerodynamic interactions on the combustion process in order to obtain the true embedded performance of a lean burn injector and hence improve future low emission fuel injector designs. Using computational fluid dynamics (CFD), validated by particle image velocimetry (PIV), a modified inlet was designed which could be retrofitted into a single sector plenum fed reacting flow facility as typically used in injector development programme. This modified inlet was designed to reproduce the key aerodynamic features associated with the diffuser-injector interaction. Additionally, a new effect associated with the diffuser-injector interaction was identified, which was a mass flow redistribution between the injector passages that clearly has the potential to alter local air-fuel-ratios (AFR). The impact of these aerodynamic interactions on the combustion process was experimentally evaluated using a single sector test facility with both a plenum inlet and the new modified inlet design. Images of the flame and emissions of NOx, CO and UHC all highlighted that inclusion of the modified inlet, and the associated aerodynamic interactions, had a large impact on the combusting performance of the injector. Changes in local AFR distribution, caused by the mass flow redistribution effect, were directly linked to some of the measured changes. However, further data was required to link the diffuser-injector interaction to the remaining measured differences. To allow for a more detailed analysis of the chemical processes a 1-D chemical reacting network model was used. A methodology was developed to build the network based on a predicted flow field thus accounting for realistic residence times and spatial variation in mass flow and temperature. Results from the model showed similar trends to the experimental data and thus a sensitivity analysis was performed. This identified several possible changes to the flow field that influence the emissions in a similar trend to that observed in the experimental data. These include the amount of flow recirculating between the pilot and mains flames and the residence time in the mains flame. Overall, the work has shown the importance of representative inlet conditions on the combusting performance of low emissions lean burn fuel injectors. A methodology for generating these conditions in single-sector test facilities used in injector development has also been developed.

The compression ignition engine will continue to be a significant factor in the transportation sector due to its unmatched efficiency and robustness. As emissions legislation becomes more stringent, pressure on the automotive industry will continue to increase as both government and consumers demand cleaner technologies. The detailed understanding of in-cylinder physics is paramount to improving efficiency and reducing engine-out emissions. Consequently, numerical analysis is playing a much more pivotal role in the engine design phase with the requirement to predict complex combustion events and pollutant formation driving the need for ever better models. In this investigation, current numerical tools are used to elucidate the effects of spray targetting and piston bowl geometry on combustion evolution and pollutant formation. Closed cycle computational fluid dynamics simulations are performed on a sector mesh at various load points using the 3 Zones Extended Coherent Flame Model coupled with adaptive mesh refinement. The computational fluid dynamics model is validated experimentally at the baseline conditions at each test point after-which, parametric sweeps of bowl geometry, exhaust gas recirculation rate and nozzle tip protrusion are conducted. Results indicate that appropriately pairing fuel injection strategy and piston geometry is essential in reducing engine-out emissions. In addition, a novel chemical source term closure based on Conditional Moment Closure (CMC) is developed to simulate diesel combustion. Conditional Source-term Estimation (CSE) uses the conditional averages in evaluating the mean chemical source term. However, unlike Conditional Moment Closure where transport equations are solved for the conditional averages. CSE approximates the conditional averages through inversion of an integral. Previous studies have shown CSE is capable of accurately simulating non-premixed flames of light hydrocarbon fuels. In this study, CSE is extended to simulate spray flame combustion by coupling the CSE combustion model with Flamelet Generated Manifolds chemistry reduction methodology. The CSE-FGM model is applied to the Engine Combustion Network n-Dodecane Spray.A -- in a Large Eddy Simulation turbulence modelling framework. The model is successfully able to predict ignition delay and flame lift-off length with good agreement to experimental measurements. Additionally, the CSE-FGM model is validated further within a RANS framework to predict ignition delay and flame lift-off length over a wide range of ambient temperature and oxygen concentration conditions. The CSE-FGM model is successfully able to predict experimental trends and sensitivity with respect to ambient conditions.

This study presents an experimental investigation of rotating flows related to gas turbine blade cooling. It focuses on measuring the development of turbulent flow and heat transfer inside a cooling passage which consists of two straight square-sectioned ducts connected with a square-ended bend, using porous metal foam blocks as turbulent promoters. The cooling passages are either stationary, or in orthogonal rotation. This study aims to improve the understanding of the effects of using porous metal foam on flow and thermal development in rotating cooling passages and exploring the possibility of using porous media for gas turbine blade cooling. Particle image velocimetry (PIV) and Thermochromic liquid crystals (TLC) experiments have been performed to measure the flow and heat transfer characteristics, respectively. A set of 12 static pressure taps around the passage provide detailed picture of the pressure variation along both sides of the rotating passage. Aluminium porous foam blocks of an aspect ratio of 1.5 with 0.93 porosity and pore density per cm of 2, were attached to two opposite walls of the straight sections normal to the bend in a staggered manner. The ratio of blocks spacing to the duct's hydraulic diameter (D) is 1; whereas the block's height ratio (h/D) is 0.6. Water is used as the working fluid at Reynolds numbers of 16,000, 26,000 and 36,000 and rotation numbers of 0.32 and 0.64. The resulting tests have generated original knowledge and information which advances our understanding of the effects of porous metallic foams on the flow and thermal development in rotating cooling passages and also provides detailed data for CFD validation. The results generated by the PIV method show the serpentine manner of the flow both upstream and downstream of the bend region due to the presence of the porous blocks. Within the bend, in contrast to the case with a smooth upstream section, a single vortex dominates the flow. While rotation does not change the overall flow character, it does force more fluid through the blocks on the trailing (pressure) side of the duct. For both stationary and rotating conditions, the upstream section is long enough for the flow to become periodic over successive rib intervals, which produces very attractive data for CFD validation. The Nusselt number distribution is significantly affected by the presence of the porous blocks and the heat transfer levels are improved by about 17%. The local Nusselt number distribution shows that the influence of rotation is negligible compared to the effect of the porous blocks.

Diesel engines are widely known for their high thermal efficiency, high torque, reliability, durability, fuel economy, and low carbon dioxide emissions in various industrial applications, such as power generation, mass transportation and off-road applications. However, the combination of a stratified air-fuel mixture and a non-premixed flame gives rise to nitrogen oxides (NOx) and particulate matter (PM) emissions for diesel combustion. Studies have shown that diesel engines can achieve very low NOx and PM emissions at high efficiency through the use of a range of low-temperature combustion (LTC) strategies and this thesis seeks to investigate the turbulent mixing influence on spray atomization and combustion processes encountered in compression ignition diesel engines under low temperature conditions. Recent studies indicate that end-of-injection (EOI) processes may support ignition recession back to the injector nozzle, thereby helping to reduce emissions. The first part of the thesis contributes to the physical understanding of this EOI phenomenon, combustion recession, using computational fluid dynamics studies at LTC conditions. Simulations are performed on a single-hole injection of n-dodecane under a range of the Engine Combustion Network's (ECN) "Spray A" conditions. The primary objective of this part is to assess the ability of a Flamelet Generated Manifold (FGM) combustion model in predicting and characterizing the combustion recession. All simulations are performed under the Reynolds-Averaged Navier-Stokes (RANS) framework in a grid-converged Lagrangian spray scenario. The simulation of combustion recession is qualitatively validated against experimental data from the literature, and the efficacy of each model in predicting combustion recession is evaluated. Overall, it was found that the FGM model was able to capture the combustion recession phenomenon well --- showing particular strength in predicting distinct auto-ignition events in the near nozzle region. Following the validation of the FGM model in predicting combustion recession, the importance of the chosen chemical mechanism in predicting diesel fuel spray combustion is also investigated. Studies were again performed under the RANS framework using the Flamelet Generated Manifold (FGM) model with four different chemical mechanisms for n-dodecane that are commonly used in the engine simulation communities - including recently developed reduced chemistry mechanisms. The flamelet database for each of the chemical mechanism is generated using two distinct methods: 0D homogeneous reactor (HR) ignition flamelets and 1D igniting counterflow diffusion (ICDF) flamelets. The effect of different tabulation approaches is investigated first following the discussion of the impact of chemical mechanisms on the prediction of combustion recession. Further discussions include an evaluation of the performance of the chemical mechanisms in predicting the most relevant reacting spray characteristics compared to the ECN experimental database: ignition delay time (IDT), flame lift-off length (LOL) and the flame reactive region. Results show that the choice of both the tabulation method and chemical mechanism plays a significant role in initial flame stabilization and end of injection (EOI) transient processes. In general, both tabulation techniques were able to qualitatively capture the flame characteristics before EOI; however, ICDF tabulation is better suited for the FGM approach in order to capture the combustion recession. Furthermore, the chemical mechanisms studied indicate that mechanisms with stronger low temperature chemistry predictions are more likely to promote combustion recession under an FGM framework. In addition, a novel combustion modelling approach is proposed here to further study the transient effects of diesel spray. Conditional Source-term Estimation (CSE) is a combustion model which invokes the Conditional Moment Closure (CMC) hypothesis to provide an approximation of the mean chemical source term in an averaged transport equation. Unlike CMC, where transport equations are solved for conditional moments, CSE recovers these conditional moments through the solution of an inverse problem. Integral equations are inverted for the conditional moments, by assuming spatial homogeneity in the conditional averages where Tikhonov regularization is applied. Previous CSE studies have shown that the model is able to predict the flame characteristics successfully for both premixed and non-premixed combustion modes. However, most of these investigations were based on methane flames. This study will be the first CSE application to a complex hydrocarbon fuel, n-dodecane, under the Engine Combustion Network's (ECN) "Spray A" conditions. Detailed chemistry is included in tabulated form using the Flamelet Generated Manifold (FGM) methodology. The predictions of this study include both the Favre averaged conditional mass fraction of reactive species and temperature. The results are compared with available experimental data and previous numerical results. Both RANS and LES simulations are performed under the same condition. The objectives of this part of the thesis are (i) assessment of the application of CSE on igniting diesel spray (ii) comparison of the CSE numerical results with available experimental results and previous numerical simulations. Overall, the combination of a chemical mechanism that has been tuned to predict "Spray A" conditions with the CSE-FGM model is able to successfully predict autoignition delay time and lift-off length of n-dodecane spray within the scatter of the experimental data. CSE-FGM offers a feasible tool for detailed combustion analysis of diesel spray flames. Both RANS and LES can give reasonably good global predictions of the flame. The LES approach is more data-rich, given the opportunity to explore more local and unsteady phenomenon present in a transient diesel jet.

Turbulent Jet Ignition (TJI) is an advanced ignition process where ultra-lean mixtures can ignite in standard gasoline spark ignition engine. In this research, a TJI unit by Mahle Powertrain USA was adopted and studied in a bespoke single-cylinder engine with optical acess. The TJI device features a very small pre-chamber that is connected to the main chamber by multiple small orifices and can be separately fuelled by a direct fuel injector. The spark plug shifts from the main chamber to the pre-chamber to ignite the pre-chamber mixture. A new cylinder head was designed and manufactured to accommodate the TJI unit and optical windows on the top and sides of the cylinder head block. A new direct inejector (DI) fuel supply system was set up for direct fuel injection in the pre-chamber. A new engine control and a data system were commissioned and used for engine experiments and heat release analysis. High-speed combustion imaging and spectroscopic techniques were developed to study the ignition and combustion in the main chamber through high-speed cameras and spectrographic equipment. Thermodynamic studies on TJI combustion in a single-cylinder engine demonstrate the ability of TJI to extend the lean-burn limit of gasoline operation at different engine speeds and loads. Similar effects are also observed with engine operations fuelled with ethanol and wet-ethanol. TJI exerts the greatest effect in extending the lean-burn limit of ethanol fuel and leads to near-zero NOx emissions near the lean-burn limit. In addition, the TJI ethanol engine operation has higher thermal efficiency as well as lower HC and CO emissions than the gasoline operation. Spectroscopic results reveal that ethanol combustion produces higher chemiluminescent emissions than gasoline during the normal spark ignition combustion in the main chamber. The OH spectral peak at 310 nm is the highest throughout the ignition and combustion, followed by CH emission at 430 nm and HCO at 330 nm. Their intensities peak before the maximum heat release rates measured by the in-cylinder pressure. Emission spectra produced by the pre-chamber ignition are stronger than the normal spark ignition in the main chamber. The highest emission intensities are observed with the fuelled pre-chamber ignition even with leaner air–fuel mixture in the main chamber. As pre-chamber fuel is increased, the pre-chamber pressure rises faster to a higher peak value, producing greater pressure differential between the pre-chamber and main chamber and faster turbulent jets of partially burned products at higher temperature. The increase in the pre-chamber pressure causes the jets to travel deeper into the main chamber and enlarges the ignition sites. In addition, the ignition delay of the main chamber combustion is shortened due to the higher temperature of turbulent jets, as indicated by the stonger emission spectra. The turbulent ignition jets of ethanol are characterised with greater momentum than gasoline due to the faster combustion speed of ethanol and higher energy input. When the pre-chamber spark timing is advanced, the OH and CH emission intensities increase due to higher pressure and temperature in the pre-chamber, causing the pre-chamber products to travel deeper to ignite most of the main chamber charge. In comparison, the pre-chamber fuel injection timing has minimal effect. Finally, the spectroscopic investigation at different air–fuel ratios with fuelled pre-chamber ignition shows that the peaks of OH, CH and HCO drop towards the lean-burn limits for both fuels. The intensity of the emission spectra is dependent on the ignition type, fuel properties and air–fuel ratios.

While Fossil-fuel-fired gas turbines remain the most reliable approach of power production, strict regulations and Acts have been imposed to limit NOx and carbon emissions. Innovative techniques have become resorts for the power generation industry to overcome such a low level of tolerance. The emerging concept of CO2-Argon-Steam Oxy-Fuel (CARSOXY) power generation has theoretically proven to increase gas turbine cycle efficiency whilst eliminating NOx emissions. Nevertheless, facilitating a higher level of technology maturity of CARSOXY gas turbines is essential to promote this technique to the industry within economically feasible scenarios while considering technical aspects of CARSOXY combustion. This thesis covers multidisciplinary aspects to facilitate further studies on CARSOXY, the performance of CARSOXY gas turbines under variable operation conditions and cycle arrangements, the production of CARSOXY, the techno-economic sustainability of CARSOXY and flame characterization. This will aid to bring CARSOXY to more mature status. A parametric study for several CARSOXY gas turbine cycles has been conducted to identify the ultimate working conditions for each cycle with respect to cycle efficiency. A cycle has been suggested for each range of working conditions. Further increase in CARSOXY cycle efficiency is promised using a newly suggested CARSOXY blend. CARSOXY gas turbines face the technical and economic challenges of conventional engineering practices for argon and carbon dioxide productions. Therefore, this thesis proposes a novel approach of continuously providing a gas turbine with the required molar fractions of CARSOXY blend. The elegance of this approach appears as an opportunity to use it in sites where ammonia is produced whilst proving its techno-economic sustainability. Finally, this thesis experimentally assesses CARSOXY in comparison to a CH4/air flame. OH Chemiluminescence integrated with Planar-Induced Fluorescence imaging techniques have been utilized to study flame stability, and flame geometry over a range of operation conditions. Results from this thesis provide a baseline investigation of CARSOXY gas turbines to be adopted by developers and manufacturers in the future.

Multi-level or hierarchical models present various features for dealing with data grouped at several levels. The majority of applications of hierarchical models use clustered data that is static in nature and collected over a long period of time. The purpose of this study is investigating hierarchical models for application with highly dynamic systems. Steady-state data are conventionally employed for engine torque mapping purposes. The data takes much time to collect and the dynamics of the system are routinely ignored. This valuable information could be used for better control of the system. In this study, an innovative transient spark-sweep approach is developed for collecting dynamic torque data more efficiently. The means of data collection implies a structure for which a multi-level model is best suited. A multi-model augmented D-optimal design is created, and the experimental data collected. Spark excitation is applied at speed/load points using Amplitude Modulated Pseudo Random Signal (AMPRS), and the torque response over the operating space is thus obtained. Conditional first-order linearization is used within the identification process for determining the hierarchical model parameters. The level-1 Nonlinear Auto Regressive eXogenous (NARX) models are separately determined using an Iterative Generalized Least Square (IGLS) method and the results are employed for initialisation of the covariance matrix and the model level-2 parameters. A novel gradient optimiser was established to facilitate the dynamic hierarchical model identification. Additionally, the uncertainty associated with model selection was mitigated using a multi-model approach. The model identified is evaluated and compared with experimental dynamic and steady-state data. It shows behaviour, both dynamic and steady state, providing prediction over a wider extrapolated spark range than conventional approaches. The new approach is eight time faster than current state-of-the-art approaches.

Lean premixed combustion is a promising strategy for the next generation of gas turbines, which is characterised by low pollutant emissions and high combustion efficiency. However, flame quenching and combustion instability arising from this technique could increase operating cost and decrease operating efficiency. The fundamentals behind these problems are not yet clarified. Therefore, it is essential to study and understand fundamental combustion phenomena of lean premixed flames in conditions relevant to practical combustion devices, which will promote the development of turbulent flame models and help to get full advantages of this technique. With the availability of increasingly powerful supercomputers, direct numerical simulation (DNS) of turbulent reacting flow has become feasible and affordable. This thesis investigates lean premixed turbulent H2/air flames at high Karlovitz (Ka) numbers under elevated pressures by using DNS with multi-step chemistry. The effects of the Karlovitz number, pressure, equivalence ratio and integral length scale on flame structures and chemical pathways are examined qualitatively and quantitatively. It is found that the relative probability of positive curvature to negative curvature is insensitive to Ka but sensitive to pressure and integral length scale (lt). On flame fronts, the local heat release rates in regions with high-positive curvatures are higher than those in regions with high-negative curvatures when conditioned on the same H2 consumption rate, whereas this phenomenon is getting weaker with decreasing Ka and increasing pressure. As pressure increases, the flame speed and thickness (δL) decreases, and the reaction zone moves to regions with higher values of progress variable. Moreover, the thickness of the inner layer conditioned on the laminar flame thickness becomes smaller under elevated pressures, which results in a lower probability of finding high curvatures in the high-pressure flames with a fixed Ka. Under conditions relevant to gas turbines, the heat release rate and scaled reaction zone thickness (δf /δL) increase with increasing equivalence ratio. However, flames demonstrate similar topological structures of flame fronts when Ka is fixed. Trenches of local equivalence ratio (φL) with small gradients are observed in concave structures outside the reaction zone, while φL plateaus with large gradients are observed in convex structures inside the reaction zone. When the integral length scale is smaller than the thickness of the corresponding laminar flame, turbulence is unable to stretch and interrupt the reaction zone and the flame presents laminar flamelet characteristics. However, the distributions of curvature and tangential strain rate are comparable with those in the same Ka flames with lt/δL ≥ 1.0. It is also found that keeping constant lt/δL ratio and Ka could isolate the effects of pressure on flame front structures. The turbulent flame with unity lt/δL ratio could capture the main features of heat release as those in flames with higher lt/δL ratios. Considering the chemical process, pressure could significantly modify the chemical pathways in both laminar and turbulent flames, and the effects are more significant than those of the Karlovitz number and integral length scale. Due to the combined effects of radical fractions and reaction rate constants, the local heat release is changed in different temperature windows when the mixture equivalence ratio varies.

This thesis investigates stability, control, and state estimation of free-piston engines (FPEs). Emphasis is placed on FPE electric power generators, currently targeted at potential application areas including electric vehicle range extension, efficient power sources in the field, and in combined heat and power systems. A general group of FPE configurations is considered; ranging from a single piston configuration through to an opposed piston configuration, which in each case may make use of either a bounce chamber, a mechanical spring, or second combustion chamber as the rebound device. To assist in verifying the theoretical results, one configuration type is physically modelled to create numerical simulation capability. The modelling includes representative descriptions of the combustion processes, the electrical machine, and the system dynamics. On stability, the thesis starts from first principles to newly develop a framework that relates key FPE physical parameters. Formal definitions of stability and instability are provided, and general technical statements on the stability of piston oscillations are proposed and verified. On control, the thesis newly applies model-based control theory to three control problems; namely, the control of compression ratio, engine start, and mitigation against abnormal combustion events such as misfire. Optimality and robustness are of key interest in addressing the control problems although other control approaches are investigated. On state estimation, the thesis newly develops a robust finite-time converging observer for FPE dynamics. The developed observer's effectiveness is mathematically proven and verified for the estimation of in-cylinder pressure and piston speed. To partly verify the findings in experiment, the thesis describes the design and creation of a two-stroke gasoline FPE in hardware. The hardware rig is used to provide measurement data for model validation, observer-based state estimation, and also to discuss FPE stability. In model validation, a parameter identification scheme is proposed in form of a general optimisation problem. In state estimation, the mean observer error is found to be very small; 0.05 bar and 0.1 m/s for the estimation of in-cylinder pressure and piston speed respectively.

The goal of producing more efficient internal combustion engines has led to the use of more advanced technologies and operating strategies. As a result, Computational Fluid Dynamics (CFD) is now a standard tool that is used to optimise the design of new combustion systems. Validation of these numerical simulations is essential, especially for the prediction of in- cylinder flow, which is known to have a profound effect on mixture preparation and combustion. Therefore, high quality, in-cylinder flow measurements are required to aid the validation process and the design of new combustion systems. This work outlines the development of an optical engine test facility and the installation of a high-speed Particle Image Velocimetry (PIV) system that provides in-cylinder flow measurements with high spatial and temporal resolution over a large range of crank angles. The system is capable of producing measurements at a maximum rate of 3.7 kHz at full resolution and a total of 6000 frame pairs in a single experiment. Additional improvements to the system for specific application to IC engines have been outlined, including the use of variable inter-frame delay and methods of scatter minimisation. To aid the validation of CFD simulations, metrics known as the Weighted Relevance Index (WRI) and the Weighted Magnitude Index (WMI), have been developed to quantify differences between flow fields in terms of both alignment and magnitude. These have been combined to produce a third metric, the Combined Magnitude and Relevance Index (CMRI), that produces a single value that rates the similarity of two flow fields. The application of these metrics has been demonstrated by investigating the differences between velocity measurements and CFD RANS simulations in the central tumble plane for three test conditions. The metrics were able to determine regions of the flow field that were significantly different between simulation and experiment, which would not be highlighted by conventional metrics. Flow field measurements were also made during the induction and compression strokes of firing cycles to investigate the effect of in-cylinder flow structures on cycle-by-cycle variations in combustion. The cycles were separated into subsets conditioned on burn rate, as indicated by in-cylinder pressure measurements. The WRI and WMI were then used to compare the conditionally averaged flow fields of fast and slow burning cycles. From this analysis, it was possible to determine regions of the flow that have a significant effect on the rate of combustion. Simultaneous combustion imaging showed that the flames for slow burning cycles tended to grow asymmetrically towards the exhaust valves. In contrast, the flames of fast burning cycles were convected away from the spark plug and grew rapidly in all directions.

The drive to increase the thermal efficiency and specific power output of the gas turbine results in increasing turbine temperatures which necessitate active cooling of the turbine blades. High performance cooling systems are required which achieve the necessary blade life requirements whilst minimising coolant consumption. This thesis explores one potential turbine cooling system advancement termed the double-wall, effusion cooled system which is envisaged to find application in areas where high cooling performance is required. Single wall effusion cooling performance is initially investigated given its importance in the double-wall system. An experimental investigation is performed and used to validate a computational model. A two-dimensional superposition method for predicting effusion cooling performance is presented and validated based upon the experimental and computational data. The superposition method was used as part of a computationally light, decoupled conjugate method which was developed to predict the cooling performance of double-wall geometries. This decoupled method was used to assess several developed double-wall geometries. Those geometries which exhibited preferable cooling characteristic were manufactured for testing in a novel experimental facility, designed and commissioned as part of the thesis work. The facility incorporates several features to improve the quality of data obtained and matches several relevant non-dimensional parameters. The facility was used to obtain both overall effectiveness and film effectiveness results for five flat plate, double-wall geometries. This data was then used to infer the convective cooling performance of each geometry at varying coolant flow rates, with preferable cooling features identified. The results indicate the high cooling performance achieved in double-wall, effusion cooled geometries. Two computational methods were used to replicate the experimental setup utilising boundary conditions obtained from the facility. The first was a fully conjugate solver, and the second was a modified form of the developed decoupled method. The results from both sets of simulations compared favourably with the experimental data. The decoupled conjugate method was particularly encouraging given the vast reduction in simulation time required when compared to the fully conjugate solver. Thermomechanical analyses were performed to identify how various geometric features influence the stress field developed under thermal load, which has direct implication on blade lifespan. The work was successful in indicating certain double-wall configurations which could reduce the magnitude of the stresses.

This project aims to improve the combustion performance of a compression ignition engine using three novel fuel pre-treatments, the employment of renewable fuels, nano additive modified fuels and supercritical (SC) fuel combustion, from the perspective of spray characteristics and engine performance. In this project, HVO and GTL are selected as the renewable fuels, whilst CeO2 nanopowder and CNT are the nano additives. A CVV system is fabricated to investigate macroscopic spray characteristics of test fuels at various conditions. A 2D CFD model coupled with the DoE method is developed to correlate experimental conditions to macroscopic spray characteristics. A Cummins ISB4.5 diesel engine test rig is employed to obtain the in-cylinder behaviour and pollutant emissions. A 3D CFD model is built to study the advantages of SC fuel combustion. GTL shows the smallest spray tip penetration during both the injection and post-injection periods, whilst DF has the largest penetration, but the average cone angles are almost the same. Nano additives have no impact on the average cone angle and spray tip penetration, except that CNT can increase the spray tip penetration slightly in the post-injection period. Empirical models are formulated and indicates different impacts of each experimental condition during injection and post-injection. HVO and GTL have lower fuel consumption and NOx, HC and PN emissions than DF. CeO2 nanopowder can significantly reduce NOx, HC and PN emissions, whilst CO can only be reduced in a certain engine load and speed range. CNT lowers down all emissions when blending with most test fuels except GTL. Compared with conventional spray combustion, SC fuel combustion illustrates significantly higher in-cylinder peak pressure and thus improved engine output power. Moreover, the fuel concentration and temperature field during the SC combustion are more evenly distributed, which enables more sufficient combustion and reduction of NOx and soot generation.

The correct management of air delivery to the combustion chamber is vital to the economic and clean operation of modern internal combustion engines. However, estimation of air mass trapped in the cylinders prior to combustion in these engines proved to be challenging and yet is fundamental to the engine control process.If such an engine is boosted and equipped with an exhaust after-treatment device, the result is many degrees of control freedom compounded with highly nonlinear behaviour. Control solutions require embedded models and on-line optimisation in order to manage the often conflicting objectives of fuel economy and low exhaust emissions. The work reported in this thesis addresses the particular issue of trapped air mass estimation in a heavy-duty engine using artificial neural networks (ANN).

Stationary gas turbines manufacturers and operators are under constant scrutiny to both reduce environmentally harmful emissions and obtain efficient combustion. Numerical simulations have become an integral part of the development and optimisation of gas turbine combustors. In this thesis work, the gas turbine combustion process is analysed in two parts, a study on air-fuel mixing and turbulent combustion. For computational fluid dynamic analysis work the open-source CFD code OpenFOAM and STAR-CCM+ are used. A fuel jet injected to cross-flowing air flow is simplified air-fuel mixing arrangement, and this problem is analysed numerically in the first part of the thesis using both Reynolds Averaged Navier Stokes (RANS) method and Large Eddy Simulation (LES) methods. Several turbulence models are compared against experimental data in this work, and the complex turbulent vortex structures their effect on mixing field prediction is observed. Furthermore, the numerical methods are extended to study twin jets in cross-flow interaction which is relevant in predicting air-fuel mixing with arrays of fuel injection nozzles. LES methods showed good results by resolving the complex turbulent structures, and the interaction of two jets is also visualised. In this work, all three turbulent combustion regimes non-premixed, premixed, partially premixed are modelled using different combustion models. Hydrogen blended fuels have drawn particular interest recently due to enhanced flame stabilisation, reduced CO2 emissions, and is an alternative method to store energy from renewable energy sources. Therefore, the well known Sydney swirl flame which uses CH4: H2 blended fuel mixture is modelled using the steady laminar flamelet model. This flame has been found challenging to model numerically by previous researchers, and in this work, this problem has been addressed with improved combustion modelling approach with tabulated chemistry. Recognizing that the current and future gas turbine combustors operate on a mixed combustion regime during its full operational cycle, combustion simulations of premixed/partially premixed flames are also performed in this thesis work. Dynamical artificially thickened flame model is implemented in OpenFOAM and validated using propagating and stationary premixed flames. Flamelet Generated Manifold (FGM) methods are used in the modelling of turbulent stratified flames which is a relatively new field of under investigation, and both experimental and numerical analysis is required to understand the physics. The recent experiments of the Cambridge stratified burner are studied using the FGM method in this thesis work, and good agreement is obtained for mixing field and temperature field predictions.

Combustion noise is central to several efforts to curb aircraft emissions. Indeed, acoustic waves originating in the combustor are a major contributor to aircraft noise. Moreover, they can act as a trigger for thermoacoustic instabilities, the consequences of which may range from decreased efficiency to outright failure. Modern engines designed to lower NOx emissions are particularly susceptible to this phenomenon. Unsteady combustion generates acoustic waves — direct noise — as well as convected flow disturbances, such as entropic, vortical or compositional inhomogeneities. These disturbances generate additional acoustic waves — indirect noise — if they are accelerated. The main objectives of this thesis are to examine the validity of current theoretical models for indirect noise, and to propose new ones where needed. First, a one-dimensional theoretical framework for the direct and indirect noise produced in a reflective environment is presented. The direct noise produced by the addition of mass, momentum and energy to a flow is determined analytically. A model for the entropic and compositional noise generated at a compact nozzle is then derived, accounting for nozzles with non-uniform entropy. Finally, the effect of reverberation (i.e. repeated acoustic reflections) is determined analytically. This enables direct and indirect acoustic sources to be identified and separated within experimental data, while eliminating the effect of acoustic reflections. The framework is applied to a model experiment — the Cambridge Wave Generator — in which direct, entropic and compositional noise are generated. Direct and indirect noise models are validated using experimental measurements of the sound field resulting from air injection and extraction, heat addition and helium injection. For the first time, direct, entropic and compositional noise are clearly identified in the experimental data, and shown to be in line with theoretical predictions. The results provide the first experimental demonstration of the compositional noise mechanism, and show that isentropic nozzle models are inadequate in predicting the indirect noise generated at nozzles with substantial losses.

Engine manufacturers are constantly seeking avenues to build cleaner and more ef cient engines to meet ever increasing stringent emission legislations. This requires a closer under- standing of the in-cylinder physical and chemical processes, which can be obtained either through experiments or simulations. The advent of computational hardware, methodologies and modelling approaches in recent times make computational uid dynamics (CFD) an important and cost-effective tool for gathering required insights on the in-cylinder ow, combustion and their interactions. Traditional Reynolds-Averaged Navier-Stokes (RANS) methods and emerging Large Eddy Simulation (LES) techniques are being used as a reli- able mathematical framework tools for the prediction of turbulent ow in such conditions. Nonetheless, the combustion submodels commonly used in combustion calculations are developed using insights and results obtained for atmospheric conditions. However, The combustion characteristics and its interaction with turbulence at Internal combustion (IC) engine conditions with, high pressure and temperatures can be quite different from those in conventional conditions and are yet to be investigated in detail. The objective here is to apply FlaRe (Flamelets revised for physical consistencies) model for IC engines conditions and assess its performance. This model was developed in earlier studies for continuous combustion systems. It is well accepted that the laminar burning velocity, SL, is an essential parameter to determine the fuel burn rate and consequently the power output and ef ciency of IC engines. Also, it is involved in almost all of the sophisticated turbulent combustion models for premixed and partially premixed charges. The burning velocities of these mixtures at temperatures of 850 ≤ T ≤ 950 decrease with pressure up to about 3 MPa as it is well known, but it starts to increase beyond this pressure. This contrasting behaviour observed for the rst time is explained and it is related to the role of pressure dependent reaction for iso-octane and involving OH and the in uence of this radical on the fuel consumption rate. The results iv seem to suggest that the overall order of the combustion reaction for iso-octane and gasoline mixture with air is larger than 2 at pressures higher than 3 MPa. The FlaRe combustion is used to simulate premixed combustion inside a spark-ignition engine. The predictive capabilities of the proposed approach and sensitivity of the model to various parameters have been studied. FlaRe approach includes a parameter βc representing the effects of ame curvature on the burning rate. Since the reactant temperature and pressure inside the cylinder are continually varying with time, the mutual in uence of ame curvature and thermo-chemical activities may be stronger in IC engines and thus this parameter is less likely to be constant. The sensitivity of engine simulation results to this parameter is investigated for a range of engine speed and load conditions. The results indicate some sensitivity and so a careful calibration of this parameter is required for URANS calculation which can be avoided using dynamic evaluations for LES. The predicted pressure variations show fair agreement with those obtained using the level-set approach. DNS data of a hydrogen air turbulent premixed ame in a rectangular constant volume vessel has been analysed to see the effect of higher pressure and temperature on the curvature parameter βc. Since the reactant temperature and pressure inside the cylinder are continually varying with time, the mutual in uence of ame curvature and thermo-chemical activities are expected to be stronger in IC engines and thus the parameter βc may not be constant. To shed more light on this, two time steps from the DNS data has been analysed using dynamic βc procedure. The results show that the effect of higher pressure and temperature need to be considered and taken into account while evaluating βc. When combustion takes place inside a closed vessel as in an IC engine the compression of the un-burnt gases by the propagating ame causes the pressure to rise. In the nal part of this thesis, the FlaRe combustion model is implemented in a commercial computational uid dynamics (CFD) code, STAR-CD, in the LES framework to study swirling combustion inside a closed vessel. Different values of βc has been tested and the need for dynamic evaluation is observed.

High combustion temperatures are required in gas-turbine engines to achieve high cycle efficiencies. With increasing temperature, however, the life span of the turbine components are reduced. The ability to accurately predict engine component temperature as a function of combustion temperature is required to strike this balance correctly. An experimental heat transfer measurement technique is developed in this thesis, which builds on a large body of existing literature. The technique enables a detailed quantification of turbine heat transfer on test rigs which closely represent gas-turbine engine configurations. Fundamental improvements are made to existing methods, in the definition of the ‘semi- infinite limit’ for transient measurement techniques, in Infra-red camera calibration, and in thermal effusivity measurement. The improvements were developed from first principles, verified experimentally, and have been used on a world leading heat transfer rig (the FACTOR combustor-turbine interaction rig, run on the NG-Turb facility at DLR Göttingen). It was found that optimisation of a number of measurement parameters was required to minimise the measurement uncertainty. It is shown that the optimum measurement parameters are dependant, and sensitive to the specific configuration of the test rig. An experimental procedure was developed and tested, which has been ‘tuned’ for measurements on the FACTOR test rig. Despite the challenging measurement environment on the FACTOR rig, it was found that state-of-the-art heat transfer measurement uncertainties of approximately 5%, could nevertheless still be achieved, by using the new methods. General principles and rules are established which can be used to guide the design of future heat transfer measurements, with the aim of minimising measurement uncertainty.

This thesis presents an experimental investigation of turbulent explosions resulting from hydrogen enriched fuel blends in an obstructed flame tube facility. The influence of mixture reactivity and thermal radiation induced ignition in turbulent explosions were studied followed by a quantification of the flow field development. Fuel lean binary H2/CO, H2/CH4 and ternary H2/CH4/CO mixtures were first studied in a two-obstacle configuration to assess the impact of mixture reactivity on explosion overpressures and flame speeds. The mixture reactivity was varied by introducing different H2 substitutions at equivalence ratios of 0.80, 0.60 and 0.40. The results highlight significant differences in explosion behaviour between the two blending components, with CO mixtures providing substantially higher over- pressures than the corresponding CH4 blends. The results suggest that methane has a mitigating effect up to comparatively high hydrogen blending fractions and that synergistic effects between fuel components need to be taken into account. A new scaling parameter (β) is proposed that successfully linearises the peak explosion overpressure between different fuel blends in response to the hydrogen concentration. A scaling based on acoustic theory shows good agreement with experimental data and a simple method for estimating the overpressure change caused by variations in the mixture reactivity in a fixed geometry is also evaluated. The impact of thermal radiation induced ignition was explored in fuel lean H2/CH4/Air mixtures with a continuous wave laser operating in the near infrared as the radiation source and acetylene black particles as the radiation target due to their relationship with soot emissions. Influences of ignition location and ignition delay time were studied. The results show that the ignition kernels caused by irradiated particles can be successfully entrained into the main flow and/or re- circulation zones formed around obstacles and cause multipoint explosions. The resulting relationship between fuel consumption ahead of the advancing flame and the evolution of the strength of the explosion was shown to be complex and typically lead to increased explosion durations with reduced peak pressures. The complex effect of forward radiation induced ignition on the pressure development stems from the interactions of the two explosion kernels resulting in two sharp pressure rises separated by a quasi-stable stage with a duration depending on the radiation induced ignition time. The total impulse was estimated by integrating the instantaneous pressure over time and the results show different characteristics for different configurations. The mixture reactivity also affects turbulent explosions indirectly via turbulence- chemistry interactions in the critical recirculation zones behind the obstacles. The flow field development was therefore quantified in a two obstacle configuration using high-speed (10 kHz) particle image velocimetry (PIV), time-series PIV and Mie scattering in H2/CO/Air and H2/CH4/Air mixtures with H2 substitution levels of 50%, 80% and 100% for a fixed stoichiometry of 0.60. The time-resolved evolution of the recirculation zone behind the second obstacle was successfully captured with the explosion over-pressure and flame propagation speed also measured. Data is presented for the mean horizontal (u) and vertical (v) velocity components at 24 spatial locations for each mixture along with the translational velocities of the shear driven recirculating eddies. It is shown that, despite large differences in flow velocities and over-pressures, the impact of the mixture reactivity on the temporal evolution of the flow field evolution can be approximately normalised using a dimensionless time scale.

The spark ignition engine is one of the most widely used power sources for vehicles. Due to the global warming, the fuel economy is encouraged to contribute in designing spark ignition engines. Therefore, one promising design is a strongly boosted downsized spark ignition engine. The present work aims to investigate combustion from two perspectives, namely, bio-derived fuels performed as additives in a strongly charged engine, and flame propagation at high pressure. N-butanol and 2,5-dimethylfuran performed as additives in a strongly charged engine were investigated across a range of spark timings at an engine speed of 750 RPM under stoichiometric conditions. Knock characteristics of those fuels were studied, and reverse thermodynamic analysis was applied to derive the in-cylinder charge state and flame speed. The results show that 2,5-dimethylfuran preformed as additive for current unleaded gasoline and its toluene reference fuel provides good anti-knock performance in strongly charged engine. N-butanol have great improvement as an additive on anti-knock performance for toluene reference fuel. Turbulent flow in the engine was studied by using two dimensional particle image velocimetry. Meanwhile, the effects of thermal expansion on turbulent flow which is located in front of the flame was also investigated. Burning velocity and flame speed of commercial unleaded gasoline and iso-octane under turbulent intensity of 0.75 m/s and 1.20 m/s were investigated at temperature of 650 K, pressure of 30 bar. The results suggest that several burning velocity correlations for predicting turbulent burning velocity not match the experimental outcomes. Furthermore, the thermal expansion may alter the turbulent intensity which is located in front of the flame. Flame shape and wrinkles were studied based on the captured laser tomographic images. The power spectral density of the wrinkles shows that the flame wrinkling is possibly related to the turbulent energy cascade. However, due to the limited camera resolutions, the effect of flame instability on wrinkling spectrum can not be observed clearly.

After undergoing an extensive study about engine air mass flow measurement approaches as well as engine modelling for air mass flow prediction, a major problem found to exist is that engineers have still not found a suitable technique to accurately measure the air mass flow entering the cylinder of an internal combustion engine. The engine air mass flow is the most important parameter needed during engine development so the fuel control can be accurately calibrated and as a result increase performance and reduce emission output of an engine. The current methods used to determine the air mass flow lead to inaccuracies due to the large amount of mathematical assumptions and also sensor errors and as a result the mapping and calibration process of a new engine family takes approximately 2 years due to extensive modelling and testing required overcoming the above drawbacks. To improve this, the distributed lumped modelling technique (D-L) of the inlet manifold was chosen, where the intake system is separated into very small sections which are distributed continuously throughout the volume of the intake until entering the cylinder. This technique is validated against a CFD model of the engine's intake system and real engine data as well as a 1D engine model.

High altitude ice particles can accrete inside the core compression system of turbofan engines in cruise and descent. This can lead to severe in-flight events including blade damage, surge and flameout. This thesis describes the development and validation of a new comprehensive computational model to aid prediction of ice crystal icing in turbofan compressors. The Ice Crystal Icing ComputationaL Environment (ICICLE) delivers a step change in modelling of the phenomenon compared to the first generation of models in the open literature. Modelling of this multi-faceted problem is broadly divided into three strands: first, modelling of the ice particles in flight; second their interactions with solid surfaces; and third the thermodynamics of ice accretion. To aid development of models and provide validation data, three different experiments were also undertaken. Treatment of particle size and shape distribution is considered first, and a particle trajectory model based on Lagrangian tracking is presented. A Nusselt number correlation for non-spherical particles is used to develop a phase change model for the particle in flight, incorporating sublimation, evaporation and melting. The model is then validated against measured particle melt data in an ice crystal facility. A model for the change in enthalpy and humidity of the airflow as a result of the particle phase change is proposed. Existing icing codes do not attempt to model these affects, but evidence from engine encounters with ice crystals indicate that they are significant. It was assessed that experimentation was required to develop modelling capability in three areas: particle sticking, erosion and heat transfer. Two experimental campaigns were performed at the ice crystal wind tunnels of the National Research Council of Canada (NRC) using simple geometries (an inclined flat plate and a cone). Data was presented for the first time on heat transfer from a warm substrate under ice crystal conditions, and a method to predict the change in particle melt during surface impacts was proposed. New semi-empirical models were developed for sticking and erosion, with a substantially wider range of applicability than achieved in previous studies. A new thermodynamic ice crystal accretion model was developed. A literature model for supercooled water icing was adapted to ice crystal and mixed phase conditions, and to substrates either above or below freezing. In the former case, an entirely novel three-layer accretion model was developed, which is a substantial advancement in modelling ice crystal growth on initially warm engine surfaces. Finally, the complete model is validated against experimental accretions on the case of a compressor stator test article, also tested at the NRC. Agreement is seen generally to be good, with the transient behaviour of growth rates well predicted, typically within 20% of experimental measurements. It is shown that a substantial improvement in prediction accuracy may be attained by updating the fluid domain at discrete time points. This accounts for the influence of the growing accretion on the flowfield. The successful application of a quantitative code to a more complex, engine-realistic geometry is a significant step forward for the literature, as existing ice crystal codes have only been validated against simpler geometries.

Modern compression ignition engines offer higher thermal efficiency compared to gasoline engines, thus offering superior fuel consumption performance and lower CO2 emissions, a major greenhouse gas. With future legislation pushing the automotive manufacturers for even lower fleet average CO2 emissions the compression ignition engine can assist in achieving these goals, however further research is required to extend their efficiency. Understanding where the chemical energy of the fuel is transferred during the combustion process and from that identifying strategies that can assist in converting part of these energy flow terms into useful piston work can help enhance the engine's efficiency. This work looked into the study of these energy flows by using a first law analysis approach and by developing the necessary instrumentation and methods that allow for the more accurate measurement of the various energy flows, subsequently increasing the accuracy of the first law analysis. Two thermal studies were carried out on a single cylinder diesel engine. The first study investigated the effect of different high-pressure EGR strategies on engine efficiency and emissions, in an attempt to reduce the negative effects of EGR application on soot emissions under two load/speed conditions. The second study compared the effects of different piston material on engine efficiency under two speed/load conditions. A baseline aluminium design, was compared against an alloy steel piston which, due to its lower thermal conductivity, was shown to provide lower heat transfer losses during combustion thus increasing efficiency. The results of the thermal studies showed that ~40% of fuel energy was transferred to the exhaust. Therefore, being able to accurately measure the exhaust temperature can offer significant insights to engine designers. The exhaust event is highly unsteady and the exhaust temperature is typically measured using a 3 mm sheathed thermocouple which, due to its thermal mass, cannot capture this transient event. Instead, a time-average measurement is only possible. This can result in an under prediction of the exhaust enthalpy since the measured temperature is lower than that of the flow field of interest due to measurement errors. A lumped capacitance model was developed in order to better understand the behaviour of thermocouple sensors under an unsteady flow environment. The sensors are subject to both dynamic errors, due to their thermal inertia, and conduction and radiation errors due to temperature gradients between the sensor and the surrounding environment. Understanding how different size thermocouples react under unsteady flow conditions has the potential to improve the measurement process and increase the accuracy of the measured exhaust temperature. A temperature reconstruction method was developed which can correct both the dynamic and conduction errors that are prevalent during the engine cycle, thus approaching the true exhaust gas temperature. This technique requires the use of thermocouples with different thermal masses, thus resulting in a different response under the same flow conditions. This reconstructed temperature then allowed the estimation of exhaust enthalpy on a mass-average basis improving the accuracy of the first law analysis.

Emissions from internal combustion engines are a major contributor to anthropogenic climate change. In order to decrease the amount of emissions, car manufacturers are investing in increasing the efficiency of spark ignition engines. Means for this include downsizing and turbocharging which come with an exacerbated risk of abnormal and harmful combustion phenomena, notably autoignition, knock and pre-ignition and thus pose a limit to the efficiency of the engines. Abnormal combustion depends on the engine geometry, the operating conditions and the fuel. Industrial standard classification systems are outlined to be insufficient, misleading or non-existent for modern engines and fuels. This thesis aims to improve the understanding of the abnormal combustion phenomena through an experimental project which can be utilised for improved classification systems. The vast majority of the experiments were conducted on a variable compression ratio engine which was fitted with modern control, measurement and data acquisition equipment to resemble an industrially-used test engine. In a first study, methods of finding the ideal engine operating point were investigated. Knock was induced in the engine, and knock indicators and limitations of knock are discussed here. Enhanced humidity was passed into the heated air-inlet stream by means of a custom-built humidifying unit. Results showed that both the power output of the engine and the severity of knock were reduced with increased humidity. This was explained by the exclusion of combustible air. A fuel-vaporization unit allowed for experiments with fully vaporized fuel. It could be shown that this had an adverse effect on knock as the cooling effect of the enthalpy of vaporization was removed. A second study employed a temperature-controlled glow plug to induce surface pre-ignition. A range of analysis techniques were tested and discussed which ranged from flame ionization detection to several in-cylinder pressure based methods. A cycle-by-cycle analysis with a maximum pressure method revealed an unexpected trend of surface pre-ignition tendency in sweeps of stoichiometry and fuels, with slightly weak of stoichiometric mixtures being the most susceptible to pre-ignition. Enhanced humidity had a negligible effect on surface pre-ignition under real world conditions. A third study concerned itself with the analysis of knock-induced heat flux, which is both a major cause for damage to the engine and trigger for surface pre-ignition. A heat flux probe was fitted to the engine and results linking heat flux to knock could be obtained on cycle-by-cycle basis and cycleaveraged basis. A linear trend between heat flux and knock intensity was found.