Pressurised hydrogen storage vessel failure scenarios are analysed through numerical simulation to understand the influence of the surrounding geometry. The separation distance from a wall, failure diameter, storage pressure and wall curvature are varied in independent scenarios. Varying separation distance reveals a range within which ignition upon failure is possible due to a coupling of the shock-front compressive heating and the diffusion of the hydrogen plume. At reduced separation distances, there is insufficient time for the shock-wave to expand and provide sufficient shock-heating. When separation distance is increased, the separation between the shock-front and the stoichiometric mixture fraction, ‘lag’, is increased and concurrent heating and flammable conditions are absent. Variation to the release jet exit diameter suggests that smaller failures lack the required momentum to reach ignition temperatures upon failure. At smaller jet exit diameters, the specific enthalpy across the shock-wave is decreased due to the greater radial expansion. Increasing storage pressure does not alter the likelihood of ignition, instead, a minor reduction in the ignition delay after release is observed. Analyses reveal a non-linear relationship between separation distance and ignition, prompting an investigation to characterise the range within which ignition may occur, indicating that for a 5 mm jet exit diameter, separation distances greater than 4 mm, but less than 7 mm undergo ignition. Further analysis is conducted in batch reactors, revealing the relationship between impingement temperature, mixture fraction, residence time and ignition. These results provide valuable insight into the safe storage of pressurised hydrogen.

We report on a method to scale NOx emissions from a numerical study of a 7.5 MWth industrial-scale, non-premixed, swirl burner (register-type) used as a hot gas generator (HGG) for an iron ore pelletisation plant firing a range of blends containing hydrogen. The method considers pure natural gas (NG), hydrogen or their blends. A detailed 3D CFD model of the burner and HGG was developed in Ansys Fluent 2024 (version R1) incorporating a realisable k-ε turbulence model, detailed GRI 3.0 chemistry, a non-adiabatic steady diffusion flamelet model, and the Discrete Ordinates radiation model. NOx formation was modeled via both thermal and prompt pathways. The blend ratios were varied with both a constant thermal input, whilst also keeping one of the following two additional parameters constant relative to the NG baseline, namely: (i) constant inlet air mass flow rate, and (ii) constant equivalence ratio. The model was verified against plant-scale NG operation data, and laboratory-scale NG, H2 and NG/H2 flame experiments, demonstrating good agreement. Results show that while flame shape remains relatively insensitive to fuel type, H₂ addition significantly increases peak flame temperature and Emission Index of NOx (EINO) values. Local extinction zones near to the base of the burner were found to more likely as the fraction of NG is increased, leading to a reduction in NOx emissions. It was also found that blending with a constant air mass flow rate maintains the flue gas enthalpy and temperature near to the baseline levels, ensuring minimal disruption to downstream pelletising operations - unlike the constant equivalence ratio approach. A method of scaling NOx emissions was proposed, based on local strain rate, with all calculated data following a linear trend regardless of fuel type and firing strategy. Two linear NOx emissions regimes were also observed, namely high NOx and low NOx regimes. It also shows that retrofitting the current industrial burner for H2 co-firing (up to 90% H₂ by volume) is technically feasible with minimal process impact, provided that the air flow rate is appropriately managed. However, full H₂ conversion will necessitate burner redesign and refined firing strategies to meet NOx emission constraints.

Adaptation of hydrogen fuel for heating applications presents several challenges, particularly thermal NO x production and reduced radiative heat transfer. Oxy-steam combustion in the MILD (moderate or intense low-oxygen dilution) regime addresses both challenges by removing the nitrogen from the oxidizer and introducing steam as a diluent that can increase thermal radiation. However, both steam and pure oxygen are required to achieve these conditions. This study investigates the feasibility of using hydrogen peroxide (H2O2) as an oxygen and steam carrier for hydrogen-fuelled heating applications, as when heated, H2O2 decomposes into water and oxygen. Batch reactor, perfectly stirred reactor, and opposed flow laminar flamelet simulations are conducted over a range of representative conditions to investigate the combustion characteristics of this system. Reactor ignition is found to be enhanced with an increase in H2O2 mass fraction within the oxidizer mixture. However, even at a high H2O2 content, preheating does not allow significant decomposition of the H2O2 into an appropriate radical pool to promote ignition within furnace-relevant residence times, suggesting that external preheating or cracking may be required for practical applications. Fundamentally, H2O2-steam flames without precracking show unique flame structures when compared to the equilibrium/cracked mixtures in MILD conditions, similar to those observed in conventional combustion of H2O2. A double-peak structure of heat release is observed in the H2O2-steam cases, where one peak correlates with the exothermic decomposition of H2O2, and the second peak corresponds to the primary combustion process with the fuel. That finding suggests that there is a significant heat addition into the flow that comes from the decomposition process that cannot be neglected in the reactor analysis, and can contribute to low temperature ignition in these conditions.

This study investigates the factors influencing the colour and visibility of hydrogen flames in domestic settings, where flame visibility is an important safety control, particularly in cooktop applications. Controlled laboratory experiments and field measurements were conducted to isolate key variables. Results show hydrogen flames are predominantly orange due to sodium emissions, primarily originating from sea salt aerosols in ambient air. This sodium-driven luminosity was found to be correlated to flame brightness, but variable depending on environmental conditions, particularly wind direction. Winds that had passed over the ocean increased flame visibility, while filtered air reduced luminosity, rendering flames nearly invisible under typical lighting. Blue and red emissions were visible only under low-light conditions. Comparative tests confirmed that hydrogen flames are less visible than natural gas flames, even under favourable conditions. The findings demonstrate that hydrogen flame visibility is inconsistent and cannot be reliably used as a safety indicator in domestics applications.

Energy storage systems are paramount in enabling progress towards a zero-emission future. The structural integrity of a latent heat thermal energy storage system for transport of refrigerated food has been investigated for the first time. Fatigue failure is a critical issue for tubes made from aluminium alloys containing phase change material at low temperatures, due to random vibration loads induced by trucks moving on rough roads. Random vibration fatigue life analysis was conducted for tubes with nominal diameters of 50 mm and 32 mm, and thicknesses of either 1.6 mm and 3 mm, using ANSYS Mechanical. The results of the modal analyses showed that the natural frequencies of the tube-PCM systems are higher than 136 Hz. This is outside the frequency domain induced by driving on rough roads, which is between 1 and 100 Hz. Tube-PCM behaves as one solid body when filled with PCM and it is bonded to tube wall resulting in balanced deformation and stress on both sides of a middle support. However, tube filled partly with PCM bonded to tube wall showed an unbalanced distribution of stress concentrated to the side with confined PCM. During cyclic melting and solidification processes, contact between tube and PCM changes where PCM can slide or form a gap. This reduces the maximum stress while still unbalanced between the two sides of tube with and without confined PCM. Tube-PCM with larger ratio of diameter to thickness (Di/t) are more sensitive to contact type, for instance tube diameter of 50 mm and 1.6 thickness compared with tube 32 mm and 3 mm thickness with no sensitivity. The minimum fatigue life was found 2.6 million hours for tube-PCM size 50 mm with 1.6 mm tube thickness. This was assuming a full cycle with minimum to maximum stress ratio equal to one. The lifetime reduced to 2.8 million hours under random vibration cycles with stress ratio equal to 0.5. These results indicate that aluminium tube sizes of 50 and 32 mm with a thickness of 1.6 mm are satisfactory for an expected life of 20 years. Aluminium is well-suited for thermal energy storage in refrigeration due to its low density, excellent corrosion resistance, and high thermal conductivity. The optimal size and thickness depend on the expected lifetime and specific design criteria.

This study explores the application of Generative Adversarial Networks (GANs) in combustion science, utilizing a flame image dataset. By comparing the produced images with the original dataset, we qualitatively analyze the discriminator and generator loss to assess the performance of the GAN. The results show improvements in the discriminator's ability to distinguish between real and generated images, as well as improvements in the generator's ability to add missing details, leading to the generation of images that are more realistic. Fundamental properties of flames are well captured in the resulting images, despite the absence or distortion of minor details. The study advances the fields of AI, image processing, and combustion science by highlighting possible uses in the creation of synthetic images and data augmentation. Overall, our qualitative analysis enriches comprehension of combustion science.

A multi-scale U-Net machine learning (ML) model is developed to assess its validity as a surrogate for non-intrusive flame temperature measurement in jet-in-hot-coflow (JHC) flames. Inputs to the model are simultaneous hydroxyl (OH) and formaldehyde (CH2O) planar laser-induced fluorescence (PLIF) measurements, with target temperature fields derived from Rayleigh scattering measurements. Coflow oxygen (O2) concentration, jet Reynolds number, coflow temperature, and fuel inputs were considered in the dataset, resulting in 33 unique flame conditions, and ∼17,000 training images. The ML model reconstructs the instantaneous temperature images to within an absolute error of 7 ± 10% of the measured values for the initial natural gas/ethylene model. Transfer learning was employed to accelerate training across conditions of different fuels, with mean absolute reconstruction error of 7 ± 9% for models of ethanol and dimethyl ether. A fourth model including data from all four fuels was trained using transfer learning, with 6 ± 9% error. Multi-layer perceptron (MLP) classification models were applied to the deepest layers of the four-fuel model to identify whether physical and chemical flame features were being encoded. Experimentally verifiable features were found to be encoded within the latent space of the model, demonstrating that it is not solely an image regression tool, but is capable of predicting flame features including the fuel type, initial coflow temperature, downstream axial location, and coflow O2 concentration. These findings indicate that the model is not only well suited for temperature predictions, but can provide scientific insight as well.

Advanced combustion concepts, such as moderate or intense low-oxygen dilution (MILD) combustion, offer a reduction in NOx emissions and increased thermal efficiency. MILD is characterised by low-oxygen, high-temperature conditions, where finite-rate chemistry effects are significant. Modelling this regime using reduced or single-step chemistry remains a challenge in part due to the finite-rate chemistry. This work proposes a generalised method for assigning Arrhenius coefficients of a single-step reaction using the outputs of a detailed mechanism. Assigning the typical chemical conservation equation to a progress variable, activation energy and pre-exponential factor for an Arrhenius kinetics global reaction are determined for hydrogen and methane across a number of conditions, with the proposed method extended to n-heptane as well. The temperature exponent in the modified Arrhenius equation is determined by minimising the ignition delay error between detailed and single-step simulations. Functional forms of each coefficient are calculated from a multivariate regression, dependent on initial temperature, pressure, and oxidant mole fraction. The predicted mechanisms are compared against the detailed kinetics in closed homogeneous batch reactors, with comparisons for hydrogen extended to both laminar opposed flow diffusion flames, and computational fluid dynamics (CFD) simulations. Ignition delay and equilibrium temperature are both well predicted for all three fuels in the batch reactors. Notably, the negative temperature coefficient behaviour of n-heptane is successfully recreated with the single-step mechanism. Temperature and heat release of hydrogen flames are well captured in both opposed flow laminar flames, and in turbulent CFD simulations. The computational time was also significantly reduced through the single-step mechanisms, resulting in ∼100 times reduction in compute time for CFD simulations. The function form of the Arrhenius coefficients shows promise for extension outside of the ranges and fuels analysed herein, and presents interesting phenomena for exploring how initial reactant temperature and pressure influence the effective activation energy of an oxidation process.

This work reports an application of physics-informed machine learning models on reconstructing key parameters of acoustically forced, time-varying laminar sooting flames, highlighting the potential of the machine learning methods as a complementary tool to conventional laser diagnostics. First, a physics-informed neural networks (PINNs) model was developed to reconstruct the fields of velocity and temperature in the region where is inaccessible with laser-based diagnosing methods due to soot scattering. The PINNs model was trained using experimental data from planar laser diagnostics and constrained with the momentum and energy conservations. The model shows effective capability of fulfilling the velocity and temperature fields. Second, an Autoencoder (AE)-based Deep Operator Network (DeepONet), also as a physics-informed model, was developed to predict the planar distribution of soot volume fraction in the flames. The AE-DeepONet framework was trained using planar images of temperature and hydroxyl radical (OH) with a hybrid way by combining physics-informed and data-driven approaches. The AE-DeepONet model outperforms the conventional data-driven-only machine learning models. The results show that, constrained by physical laws, machine learning based models can properly predict soot distribution, velocity and temperature in unsteady laminar flames, shedding light on the physics-informed machine learning methods as a complement to laser diagnostics.

The release of high-pressure hydrogen is studied to develop the understanding of the potential for ignition and flame acceleration. Numerical simulations of the sudden release of 12.5 MPa hydrogen through a 7.5 mm orifice into a semi-enclosed concave domain are reported and analysed. An obstructing cylindrical bluff-body is positioned downstream of the hydrogen jet. Ignition was observed to be initiated through shock-shock intersection, regardless of obstruction. In addition to this, ignition due to bluff-body stagnation was observed. Intersection between the hydrogen jet, the shock-wave and the cylindrical bluff-body resulted in a localised heating at the stagnation point, forming a hot-spot which transitioned into a deflagration. The presence of an introduced bluff-body impingement promotes mixing, resulting in the deflagration undergoing significant flame acceleration and exceeding the local speed of sound. The absence of the impinging cylindrical bluff-body results in the shock-wave reaching the concave reflecting wall unobstructed, and the subsequent shock reflections from this surface self-intersect, providing the energy required for autoignition and consequent deflagration.

Hydrogen-fueled aircraft are a promising innovation for a sustainable future in aviation. While hydrogen aircraft design has been widely studied, research on airport requirements for new infrastructure associated with hydrogen-fueled aircraft and its integration with existing facilities is scarce. This study analyzes the current body of knowledge and identifies the planning challenges which need to be overcome to enable the operation of hydrogen flights at airports. An investigation of the preparation of seven major international airports for hydrogen-powered flights finds that, although there is commitment, airports are not currently prepared for hydrogen-based flights. Major adjustments are required across airport sites, covering land use plans, airside development, utility infrastructure development, and safety, security, and training. Developments are also required across the wider aviation industry, including equipment updates, such as for refueling and ground support, and supportive policy and regulations for hydrogen-powered aircraft. The next 5–10 years is identified from the review as a critical time period for airports, given that the first commercial hydrogen-powered flight is likely to depart in 2026 and that the next generation of short-range hydrogen-powered aircraft is predicted to enter service between 2030 and 2035.

Turbulent flames in a confined and pressurised jet-in-hot-coflow combustor, under a range of pressures and coflow conditions, are investigated in this study. The investigation involves both experimental and numerical analyses, with the central focus on extending the use of an atmospheric-pressure computational fluid dynamics (CFD) model to describe hot and low-oxygen combustion at elevated pressures. A series of CFD simulations are compared against experimental images, in terms of both flame structure and chemiluminescence behaviour. The imaging of OH* and CH* is performed experimentally for flames with Reynolds numbers of 10,000 and 15,000, issuing into coflows with O2 concentrations ranging from 3%–9%. The experimental results show a reduction in OH* and CH* intensity with increasing pressure under the various coflow conditions, while the CFD results display some key differences in the trends with pressure and O2. The computational analysis is complemented by a series of one-dimensional laminar flame simulations at a range of pressures and oxidant O2 concentrations. These simulations enable the changes in chemical kinetics with pressure and O2 concentration to be studied in greater detail, and several key differences in comparison with the CFD results are observed. Ultimately, the results highlight the importance of accurately predicting both the flow-field and finite-rate chemistry to reproduce the trends observed under hot and low-oxygen combustion conditions at elevated pressures.

The elevated temperature of hydrogen combustion increases the formation of thermal NOx. Moderate or intense low oxygen dilution (MILD) combustion is known to reduce NOx emissions and increase thermal efficiency. Pressure is often also used for increasing thermal efficiency. The impact that pressure has on fluid dynamics and chemical kinetics is especially relevant in MILD combustion conditions. Hydrogen jet flames issuing into a hot and vitiated coflow were imaged using OH* chemiluminescence at different pressures (1-7 bar) and oxygen levels (3-9% by vol.). Laminar flame simulations complemented the experiments. The observed mean radial OH* width increased with increased pressure, but only at O2 content less than 9%, suggesting that pressure has greater influence on kinetics when oxygen is reduced. The integrated OH* signal strength remained constant at 3% coflow O2, despite an apparent increase in flame width, suggesting a spatial broadening of the flame with pressure. Numerical results indicate that at 3-6% O2, conditions for MILD combustion of H2 are met across a wide range of strains and pressures, supporting the experimental observations for 3% O2.& COPY; 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY license (http://creativecommons.org/ licenses/by/4.0/).

New understanding of turbulent jet flames of natural gas, and blends of natural gas with ethylene (C2H4), issuing into a range of preheated coflowing oxidisers with reduced oxygen (O2) concentrations are reported. Comparisons are made for coflow O2 concentrations of 3 %, 4 %, 5 %, 6 %, 9 % and 11 % and coflow temperatures of 1250 K, 1315 K and 1385 K. Instantaneous and simultaneously planar imaging measurements of temperature, hydroxyl radicals (OH) and formaldehyde (CH2O) were taken at eight locations ranging from 9 mm to 75 mm downstream of the jet exit plane. The experimental data is supplemented with chemical reaction modelling. It is shown that increasing the proportion of C2H4 in the fuel supply resulted in an increase in the flame front temperature. At 3 % O2, CH2O signal did not very between the three fuels containing C2H4, whilst at 9 %, the signal increased with increasing C2H4. OH number density was highest for a blend of 66 % NG and 33 % C2H4. The fuel composition made little effect on the jet spread rate. Reaction modelling found evidence of changes in the reaction process, both in the suppression of reactions with different fuels, and in the suppression of reactions, and the locations they occur, with reducing O2 concentrations. Reaction modelling also found a significant decrease in ignition delay when C2H4 was present in the fuel mixture, and an increase in ignition delay as all four fuels transition to MILD combustion.

New understanding of turbulent ethylene (C2H4) jet flames, issuing into a range of preheated coflowing oxidisers with reduced oxygen (O2) concentrations are reported. These conditions emulate moderate or intense low-oxygen dilution (MILD) combustion. To resolve previously reported non-monotonic trends, comparisons are made for coflow O2 concentrations of 3 %, 4 %, 5 %, 6 %, 9 % and 11 % and coflow temperatures of 1250 K, 1315 K and 1385 K. Instantaneous and simultaneously planar imaging measurements of temperature, hydroxyl radicals (OH) and formaldehyde (CH2O) were taken at eight downstream locations ranging from 9 mm to 75 mm. The new data reveal non-monotonic lift-off trends in the OH and CH2O formation heights at all three temperatures. Compared with extant measurements, the higher resolution and greater control of coflow composition provide a detailed exploration of the transitional behaviour that occurs when MILD combustion conditions are achieved. The results experimentally demonstrate non-monotonic variation in lift-off height, and show that the same behaviour is present in both the OH and CH2O formation heights, and occurs at all three temperatures, which is due to the movement of the location of the stoichiometric and most reactive mixture fractions to the coflow side of the jet shear layer. The results also show that increasing the oxidiser temperature does not have a significant effect on the OH number density for a given temperature but does result in a significant decrease in lift-off height and CH2O concentration.

This work investigates the effects of ambient and injection parameters on the ignition and combustion characteristics of hydrogen (H-2)- methane (CH4) jet (50% H-2 by volume, with the remaining CH4) in simulated direct-injection, compression-ignition conditions. Parameter variations include ambient gas temperature (1060-1200 K), ambient oxygen (O-2) concentration (10-21 vol %), and injection pressure (10-20 MPa reservoir pressure). The results show that the ignition delay of the H-2-CH4 jet decreases with increasing ambient temperature. In most cases, the ignition initiates from a localized kernel before spreading across the jet volume downstream. The lower ambient O-2 cases display a more voluminous ignition sequence. The results also show that the jet flame recesses upstream to attach or stabilize close to the nozzle but becomes increasingly lifted with lower ambient temperature and O-2 conditions. The flame autoignition process displays increased variation at the lowest tested ambient temperature condition in this work, which affects the ensuing flame evolution and heat release profile.

In flames, turbulence can either limit or enhance combustion efficiency by means of strain and mixing. The interactions between turbulent motions and chemistry are crucial to the behaviour of combustion processes. In particular, it is essential to correctly capture non-equilibrium phenomena such as localised ignition and extinction to faithfully predict pollutant formation. Reactor-based combustion models - such as the Eddy Dissipation Concept (EDC) or Partially Stirred Reactor (PaSR) - may account for turbulence-chemistry interactions at an affordable computational cost by calculating combustion rates relying upon canonical reactors of small fluid size and timescale. The models may include multiscale mixing, detailed chemical kinetic schemes and high-fidelity multispecies diffusion treatments. Although originally derived for conventional, highly turbulent combustion, numerous recent efforts have sought to generalise beyond simple empirical correlations using more sophisticated relationships. More recent models incorporate the estimation of scales based on local variables such as turbulent Reynolds and Damkohler numbers, phenomenological descriptions of turbulence based on fractal theory or specific events such as extinction. These modifications significantly broaden the effective range of operating conditions and combustion regimes these models can be applied to, as in the particular case of Moderate or Intense Low-oxygen Dilution (MILD) combustion. MILD combustion is renown for its ability to deliver appealing features such as abated pollutant emissions, elevated thermal efficiency and fuel flexibility. This review describes the development and current state-of-the-art in finite -rate, reactor-based combustion approaches. Recently investigated model improvements and adaptations will be discussed, with specific focus on the MILD combustion regime. Finally, to bridge the gap between laboratory -scale canonical burners and industrial combustion systems, the current directions and the future outlook for development are discussed.

The structure and stabilisation mechanisms of ethanol and n-heptane spray flames are investigated in this study. The burner configuration involves a dilute spray of dispersed droplets which is produced and transported via a carrier gas stream of air to the reaction zone, where the flames are stabilised by a hot coflow of combustion products. A range of coflow conditions were implemented for the different flame cases, allowing the effects of the coflow oxygen (O2) concentration and temperature to be examined independently. The resulting flames were analysed using three simultaneous laser diagnostic techniques, enabling the combined planar imaging of the hydroxyl (OH) and formaldehyde (CH2O) radicals, along with the location of droplets. For both fuel types, a noticeable shift in stabilisation behaviour was observed with a variation in the coflow O2 concentration from 11% to 3%, while the coflow temperature was not seen to have a significant impact. These flames also show an interesting departure from the typical behaviour observed for gaseous and prevaporised flames in a similar configuration, particularly for coflow conditions that are typically associated with the transition to the mild combustion regime.

The configuration of a recently installed CO2 heat pump system with a cooling capacity of 33.5 kW and heating capacity of 35.8 kW used on a dairy farm is presented. This system has been studied in order to determine the optimal performance when used for combined water heating and refrigeration. Furthermore, the effect of the interstage pressure on the CO2 heat pump system has been investigated. This required the development of a thermodynamic cycle based on the configuration of the actual constructed system and data acquired from monitoring. A compressor model with a compression ratio over three was developed and validated against experimental data with an acceptable agreement. The numerical model of the counter-flow heat exchanger between water and sCO2 has been developed assuming that the water outlet temperature is variable (just over the set water temperature), which is more consistent with actual practice. In the desktop study, the optimal COPs for heating, cooling and combined performance have been simulated considering the interstage pressure, discharge pressure and gas cooler outlet temperature. It was found that the optimal combined COPs generally decrease from 3.84 to 3.68 with the interstage pressure from 29.65 bar to 53.60 bar. Through optimisation of discharge pressure, it was found that the optimal combined COP of 4.27 occurs at the same pressure where the optimal heating COP of 2.63 is achieved (at a discharge pressure of 94 bar), while the optimal cooling COP of 1.66 occurs when the discharge pressure is around 88 bar. A further optimisation study was conducted based on a fixed cooling capacity, and it was found that the optimal cooling COP of 1.78 can be achieved when the discharge pressure is 80 bar, while the optimal heating and combined COPs are still achieved at the same discharge pressure of 93 bar (with the heating COP of 2.59 and combined COP of 4.18). Based on the simulation results, it can be summarised that the overall system thermal performance designed for heating and refrigeration is dominated by the heating performance.