Professor Jennifer X Wen

Hydrogen is a key pillar in the global Net Zero strategy. Rapid scaling up of hydrogen production, transport, distribution and utilization is expected. This entails that hydrogen, which is traditionally an industrial gas, will come into proximity of populated urban areas and in some situations handled by the untrained public. To realize all their benefits, hydrogen and its technologies must be safely developed and deployed. The specific properties of hydrogen involving wide flammability range, low ignition energy and fast flame speed implies that any accidental release of hydrogen can be easily ignited. Comparing with conventional fuels, combustion systems fueled by hydrogen are also more prone to flame instability and abnormal combustion. This paper aims to provide a comprehensive review about combustion research related to hydrogen safety. It starts with a brief introduction which includes some overview about risk analysis, codes and standards. The core content covers ignition, fire, explosions and deflagration to detonation transition (DDT). Considering that DDT leads to detonation, and that detonation may also be induced directly under special circumstances, the subject of detonation is also included for completeness. The review covers laboratory, medium and large-scale experiments, as well as theoretical analysis and numerical simulation results. While highlights are provided at the end of each section, the paper closes with some concluding remarks highlighting the achievements and key knowledge gaps.

•The HRR of LIB cluster in confined space increased exponentially in the late stage.•The heating effect of cell ejections is the dominant reason for TRP.•The rupture of aluminum top delayed the TRP of LIB cluster in confined space.•Different mitigation measures were tested to guide safety design of LIB cluster. Thermal runaway (TR) propagation is one of the most crucial failure modes for lithium-ion batteries (LIB). However, previous studies mainly focused on TR propagation (TRP) in open space while the transportation and utilization of LIBs are mostly in confined spaces, for which significant knowledge gaps exist concerning TRP and related safety issues. In this study, a series of TRP tests were conducted in an enclosed LIB cluster. For comparison, tests were also conducted with an open cluster. The acceleration of the TRP and the exponential growth of the peak heat release rate were observed in the enclosed cluster. TRP in the open space was found to be not self-sustainable with some cells escaped TR. Different safety measures were applied to mitigate TRP including various spaces between the cells, barriers, cover plate, aluminum (Al) top and lower ambient temperature. The gaps between the cells were found to delay TRP occurrence from row 1 to row 2, but ineffective for the remaining 4 rows due to the preheating effects of the particles and flame. Jumping propagation of TR over rows was observed. For the Al top design, the hole created due to Al melting by fire was found to effectively slow down the propagation speed as the hot ejections were relieved from the hole. The cover plate was found to be most effective in delaying TRP from row 1 to row 2 as it blocked hot gases reaching the cell tops. However, it accelerated TRP in the remaining 4 rows. This was thought to be due to the thermal insulation effect. When the ambient temperature was lower than 17.4 °C, TR could only propagate in the first row even if there were no gaps between the cells in the confined cluster. Based on these results, recommendations are made on mitigation measures to help ensure safety during LIB transportation and utilization.

The statistical narrow band (SNB), correlated-k (CK), and weighted-sum-of-gray-gases (WSGG) models have been incorporated into a computational fluid dynamics (CFD) code. Their accuracy and computing times are evaluated for three pool fires scenarios. CFD_CK and CFD_SNB yield similar predictions, and the former is three times more CPU demanding than the latter. CFD_CK is unrealistic for practical fire applications. Temperature and velocity predictions with CFD_WSGG and CFD_SNB agree better in the persistent and plume regions, but significant discrepancies are found in the intermittent region. The overall predictions of the two approaches show reasonable agreement. The WSGG model should not be discarded for CFD fire simulations.

At least 24 Liquefied Natural Gas (LNG) rollover incidents have been reported since 1960. During rollover, because of the heat ingress through the tank walls, a stratified LNG may be suddenly homogenised while releasing massive amounts of vapour. The latter can result in an overpressure in the tank and significant amounts of potentially explosive LNG vapour being vented out. Both represent considerable hazards. The current study is aimed at developing and validating rolloverFoam, a dedicated solver for simulating rollover phenomena within the frame of the open-source CFD toolbox OpenFOAM. The code is based on the Navier-Stokes equations and integrates the Boussinesq approximation in the momentum equation and the modelling of the transport of the different hydrocarbons constituting LNG. The traditional Reynolds-Averaged Navier-Stokes approach is used for computational efficiency in large-scale applications. For turbulence modelling, the k−ε model with additional terms to incorporate buoyancy effects is used. The code has firstly been successfully validated by comparing its predictions to data obtained during a medium-scale LNG rollover experiment. The newly developed solver has also been applied to the La Spezia incident. The results have provided interesting insights into the phenomenon. •A CFD solver for simulating rollover accidents has been developed.•The standard k-ε turbulence model is extended to include buoyancy effects.•The code is validated with published medium-scale experimental data.•Predictions have also been conducted for the La Spezia rollover incident.

Numerical simulations have been conducted for a range of liquefied natural gas (LNG) pool fires on land and water using FireFOAM, the large eddy simulation (LES)-based fire simulation code within the framework of open source computational fluid dynamics (CFD) toolbox OpenFOAM®. The studied pool diameters range from 14-400 m with cross winds from 1.6-9.6 m/s. The code uses the extended eddy dissipation concept (EDC) and a newly developed soot model based on the laminar smoke point concept. For the low-temperature (111.65-200 K) thermodynamic data of natural gas, the nine-coefficient correlations in the NASA thermodynamic database are used. Comparison between the predictions and measurements were carried out for the first four cases where full-scale test data are available. For all cases, the variations of flame length, tilt angle, and surface emissive power with LNG pool diameters are analyzed. New nonlinear correlations for predicting length-to-diameter ratio and tilt angle are also proposed.

Industries across the world are making the transition to net-zero carbon emissions, as government policies and strategies are proposed to mitigate the impact of climate change on the planet. As a result, the use of hydrogen as an energy source is becoming an increasingly popular field of research, particularly in the aviation sector, where an alternative, green, renewable fuel to the traditional hydrocarbon fuels such as kerosene is essential. Hydrogen can be stored in multiple ways, including compressed gaseous hydrogen, cryo-compressed hydrogen and cryogenic liquid hydrogen. The infrastructure and storage of hydrogen will play a pivotal role in the realisation of large-scale conversion from traditional fuels, with safety being a key consideration. This paper provides a review on previous work undertaken to study the characterisation of both unignited and ignited hydrogen jets, which are fundamental phenomena for the utilisation of hydrogen. This includes work that focuses on the near-field flow structure, dispersion in the far-field, ignition and flame characteristics with multi-physics. The safety considerations are also included. The theoretical models and computational fluid dynamics (CFD) multiphase and reactive flow approaches are discussed. Then, an overview of previous experimental work is provided, before focusing the review on the existing computational results, with comparison to experiments. Upon completion of this review, it is highlighted that the complex near-field physics and flow phenomena are areas lacking in research. The near-field flow properties and characteristics are of significant importance with respect to the ignition and combustion of hydrogen.

Numerical simulations have been conducted for large scale hydrogen detonation by solving Euler equations with a single step reaction for the chemistry. A total variation diminishing numerical scheme is used for shock capturing. Predictions were firstly conducted with a small domain to ensure that the reaction scheme has been properly tuned to capture the correct detonation pressure and velocity. On this basis, simulations were conducted for the detonation tests carried out at Kurchatov Institute in Russia [1,4]. Comparison is made between the predictions and measurements. Further simulations were then conducted for a hypothetical hydrogen-air cloud in the open to assess the impulse as well as overpressure distributions. (c) 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

The extension of the laminar smoke point based approach to turbulent combustion using the partially stirred reactor (PaSR) concept proposed by Chen et al. (2014) has been further improved to overcome the limitation in the formulations of Chen et al. (2014) which assumed infinitely fast soot oxidation chemistry and constant soot formation characteristic time. In the PaSR approach, each computational cell is split into two zones: the reacting zone and the non-reacting zone. Soot formation and oxidation are assumed to take place at finite rates in the reacting zone and computed from the corresponding laminar rates and the mass fractions for soot formation and oxidation, which are evaluated in each computational cell from the characteristic time scales for turbulent mixing, soot formation and oxidation. Since soot would be produced in not only the fine structures but also surrounding fluids in the Eddy-Dissipation Concept (EDC) model, the average field parameters between the fine structure and surrounding fluid are employed instead of those Favre-averaged values in Chen et al.'s soot formation model. The newly extended model has been implemented in FireFOAM, a large eddy simulation (LES) based solver for fire simulation based on the open source CFD code OpenFOAM (R). Numerical simulations of a 30 cm diameter heptane and toluene pool fires tested by Klassen and Gore (1992) were performed for validation. The predicted soot volume fraction and temperature have achieved improved agreement with the experimental measurements in comparison with that of Chen et al. (2014), demonstrating the potential of the improved PaSR-based soot model for fire applications. (C) 2017 Elsevier Ltd. All rights reserved.

With the anticipated introduction of hydrogen fuel cell vehicles to the market, there is an increasing need to address the fire resistance of hydrogen cylinders for onboard storage. Sufficient fire resistance is essential to ensure safe evacuation in the event of car fire accidents. The authors have developed a Finite Element (FE) model for predicting the thermal response of composite hydrogen cylinders within the frame of the open source FE code Elmer. The model accounts for the decomposition of the polymer matrix and effects of volatile gas transport in the composite. Model comparison with experimental data has been conducted using a classical one-dimensional test case of polymer composite subjected to fire. The validated model was then used to analyze a type-4 hydrogen cylinder subjected to an engulfing external propane fire, mimicking a published cylinder fire experiment. The external flame is modelled and simulated using the open source code FireFOAM. A simplified failure criteria based on internal pressure increase is subsequently used to determine the cylinder fire resistance. (C) 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Explosions in homogeneous reactive mixtures have been widely studied both experimentally and numerically. However, in practice, combustible mixtures are usually inhomogeneous and subject to both vertical and horizontal concentration gradients. There is still very limited understanding of the explosion characteristics in such situations. The present study aims to investigate deflagration to detonation transition (DDT) in such mixtures. Two cases in a horizontal obstructed channel with 30% and 60% blockage ratios filled with hydrogen/air mixture with vertical concentration gradients are numerically studied. These cases were experimentally investigated by Boeck et al. (2015), and hence some measurements are available for model validation. A density-based solver within the OpenFOAM CFD toolbox is developed and used. To evaluate the convective fluxes contribution, the Harten–Lax–van Leer–Contact (HLLC) scheme is used for shock capturing. The compressible Navier–Stokes equations with a single step Arrhenius reaction are solved. The numerical results are in good qualitative and quantitative agreement with the experiments. The predictions show that the overpressure at the DDT transition stage is higher in the non-uniform mixtures than that in homogeneous mixtures under similar conditions. It is also found that increasing the blockage ratio from 30% to 60% resulted in faster flame propagation and lower propensity to DDT. The Baroclinic torque and the resulting Richtmyer–Meshkov (RM) instability are also analyzed in relation to flame acceleration and DDT. •Numerical studies have been conducted to investigate DDT of non-homogenous mixture.•The first localized explosion occurred near the bottom wall where the mixture is lean.•The increase in the BR was found to increase the flame acceleration.•The role of hydrodynamic instabilities in DDT phenomena have also been studied.

Numerical simulations have been carried out for large scale hydrogen explosions in a refuelling environment and in a model storage room. For the first scenario, a high pressure hydrogen jet released in a congested refuelling environment was ignited and the subsequent explosion analysed. The computational domain mimics the experimental set up for a vertical downwards release in a vehicle refuelling environment experimentally tested by Shirvill et al. [6]. For completeness of the analysis, an analytical model has also been developed to provide the transient pressure conditions at nozzle exit. The numerical study is based on the traditional computational fluid dynamics (CFD) techniques solving Reynolds averaged Navier-Stokes equations. The Pseudo diameter approach is used to bypass the shock-laden flow structure in the immediate vicinity of the nozzle. For combustion, the Turbulent Flame Closure (TFC) model is used while the shear stress transport (SST) model is used for turbulence. in the second scenario, premixed hydrogenair clouds with different hydrogen concentrations from 15% to 60% in volume were ignited in a model storage room. Analysis was carried out to derive the dependence of over-pressure on hydrogen concentrations for safety considerations. (C) 2009 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

This paper reports on the further development and validation of CO(2)FOAM, a dedicated computational fluid dynamics solver for the atmospheric dispersion of Carbon Dioxide (CO2) from accidental pipeline releases. The code has been developed within the framework of the open source CFD code OpenFOAM (R) (OpenCFD, 2014). Its earlier version used the homogeneous equilibrium method for fully compressible two-phase flow. Validation of the code against CO2 releases through vertical vent pipes and horizontal shock tubes was previously reported by Wen et al. (2013). In the present study, the homogeneous relaxation model has been implemented as it is more suited to account for the presence of solid CO2 within the releases. For validation, the enhanced CO(2)FOAM has been used to predict CO2 dispersion in a range of full scale tests within the dense phase CO2 PipeLine TRANSportation (COOLTRANS) research programme (Cooper, 2012) funded by National Grid. The test case used in the present study involved a puncture in a buried pipe. The experimental measurements were supplied to the authors after the predictions were completed and submitted to National Grid. Hence, the validation reported here is indeed 'blind'. The validated model has also been used to study the effect of a commercial building located downstream from the release location. (C) 2016 Elsevier Ltd. All rights reserved.

High-resolution direct numerical simulations are conducted for under-expanded cryogenic hydrogen gas jets to characterize the nearfield flow physics. The basic flow features and jet dynamics are analyzed in detail, revealing the existence of four stages during early jet development, namely, (a) initial penetration, (b) establishment of near-nozzle expansion, (c) formation of downstream compression, and (d) wave propagation. Complex acoustic waves are formed around the under-expanded jets. The jet expansion can also lead to conditions for local liquefaction from the pressurized cryogenic hydrogen gas release. A series of simulations are conducted with systematically varied nozzle pressure ratios and systematically changed exit diameters. The acoustic waves around the jets are found to waken with the decrease in the nozzle pressure ratio. The increase in the nozzle pressure ratio is found to accelerate hydrogen dispersion and widen the regions with hydrogen liquefaction potential. The increase in the nozzle exit diameter also widens the region with hydrogen liquefaction potential but slows down the evolution of the flow structures. (c) 2020 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

The present study aims to test the capability of our newly developed density-based solver, ExplosionFoam, for flame acceleration (FA) and deflagration-to-detonation transition (DDT) in mixtures with concentration gradients which is of important safety concern. The solver is based on the open source computational fluid dynamics (CFD) platform OpenFOAM® and uses the hydrogen-air single-step chemistry and the corresponding transport coefficients developed by the authors. Numerical simulations have been conducted for the experimental set up of Ettner et al. [7], which involves flame acceleration and DDT in both homogeneous hydrogen-air mixture as well as an inhomogeneous mixture with concentration gradients in an obstucted channel. The predictions demonstrate good quantitative agreement with the experimental measurements in flame tip position, speed and pressure profiles. Qualitatively, the numerical simulations have reproduced well the flame acceleration and DDT phenomena observed in the experiment. The results have revealed that in the computed cases, DDT is induced by the interaction of the precursor inert shock wave with the wall close to high hydrogen concentration rather than with the obstacle. Some vortex pairs appear ahead of the flame due to the interaction between the obstacles and the gas flow caused by combustion-induced expansion, but they soon disappear after the flame passes through them. Hydrogen cannot be completely consumed especially in the fuel rich region. This is of additional safety concern as the unburned hydrogen can be potentially re-ignited once more fresh air is available in an accidental scenario, resulting in subsequent explosions. •ExplosionFoam, a density-based solver, has been developed within the frame of open source computational fluid dynamics (CFD) platform OpenFOAM®.•The solver has been tested on the predictions of flame acceleration and deflagration-to-detonation transition in mixtures with concentration gradients.•The predictions demonstrate good quantitative agreement with the experimental measurements in flame tip position, speed and pressure profiles.•Qualitatively, the numerical simulations have reproduced well the flame acceleration and DDT phenomena observed in the experiment.

A horizontally oriented jet fire could occur if the leaking liquefied natural gas (LNG) from the side surface of a pipe or storage tank was ignited. Previous work with LNG mostly focused on pool fires. In the present study, horizontally oriented LNG jet fires were studied through 10 open field full scale tests. The flames were visualized by both infrared and video cameras. The recorded flame shapes are compared and analysed. Peak temperatures and heat fluxes at various flow rates were measured and recorded. For relatively low reservoir pressure, a small amount of LNG was found to spray through the fire and rainout onto the ground, forming an LNG pool. A correlation was established to calculate the flame length from the mass flow rate. •Field tests were conducted using horizontally oriented nozzle for flashing LNG jet fires with release rates from 0.01 kg/s to 0.075 kg/s.•The peak temperature of the LNG jet fire stabilized at about 1100 °C.•In some cases, the released LNG did not completely vaporize in the flame zone but rained out, resulting in LNG pool fire.•Incident radiative heat fluxes were almost constant near the nozzle but increased with the release rate further away.•A new correlation was proposed to correlate the measured flame length with mass flow rate.

For radiative transfer in complex geometries, Sakami and his co-workers have developed a discrete ordinates method (DOM) exponential scheme for unstructured meshes which was mainly applied to graymedia. The present study investigates the application of the unstructured exponential scheme to a wider range of non-gray scenarios found in fire and combustion applications, with the goal to implement it in an in-house Computational Fluid Dynamics (CFD) code for fire simulations. The original unstructured gray exponential scheme is adapted to non-gray applications by employing a statistical narrow-band/correlated-k (SNB-CK) gasmodel and meshes generated using the authors' own mesh generator. Different non-gray scenarios involving spectral gas absorption by H(2)O and CO(2) are investigated and a comparative analysis is carried out between heat flux and radiative source terms predicted and literature data based on ray-tracing and Monte Carlo methods. The maximum discrepancies for total radiative heat flux do not typically exceed 5%.

Abstract only

A subgrid scale (SGS) model for partially premixed combustion has been implemented and applied to simulate the backdraft phenomena and its mitigation by watermist. The model is based on the coupling of independent approaches for non-premixed and premixed turbulent combustion. The "flame index" concept was used to separate the two different combustion regimes. This index describes the structure of the flame based on fuel and oxygen gradients. By using this approach, it is possible to implement individually the most suitable combustion models for each structure. In the current study, the Large Eddy Laminar Flamelet Model (LELFM) was used for non-premixed combustion and the flame surface density approach for premixed combustion. Simulations were conducted for the reduced scale backdraft tests of Weng and Fan. The predicted pressure-time curve is in good agreement with the measurement The predicted mass flow rate versus time has also captured the correct trend indicated by the measurement but quantitatively relatively larger discrepancies are found. For the simulation with watermist, a correlation for the laminar burning velocity of the methane-air-diluent-water vapor system was introduced following Stone and Clarke and Liaio et al.. In line with the experimental observation, the watermist was found to have mitigated the backdraft by reducing the prevailing laminar flame velocity, resulting in lower temperature and pressure distributions within the compartment However, in this particular case, the mist did not completely suppress the turbulent deflagration. Further study to optimize the mist injection time, speed, quantity, and direction is needed to achieve this goal.

In most earlier experimental investigations of condensation on low-fin tubes, vapor-side heat transfer coefficients have been found from overall (vapor-to-coolant) measurements using either predetermined coolant-side correlations or “Wilson plot” methods. When the outside resistance dominates, or is a significant proportion of the overall resistance, these procedures can give satisfactory accuracy. However, for externally enhanced tubes, and particularly with high-conductivity fluids such as water, significant uncertainties may be present. In order to provide reliable, high-accuracy data, to assist in the development of theoretical models, tests have been conducted using specially constructed plain and finned tubes fitted with thermocouples to measure the tube wall temperature, and hence the vapor-side heat transfer coefficient, directly. The paper describes the technique for manufacturing the tubes and gives results of systematic heat transfer measurements covering the effects of fin height, thickness, and spacing, tube diameter, and vapor velocity. The tests were carried out with steam, ethylene glycol, and R-113, with vertical vapor downflow. The heat flux was measured using an accurately calibrated 10-junction thermopile and paying particular attention to coolant mixing and isothermal immersion of thermocouple junctions. Care was taken to avoid errors due to the presence in the vapor of noncondensing gas and the occurrence of dropwise condensation. Smooth, consistent, and repeatable results were obtained in all cases. The data are presented in easily accessible form and are compared with the results of previous investigations, where indirect methods were used to determine the vapor-side data, and with theory.

The present study aims to develop a simplified mathematical model for the evolution of heating-induced thermal runaway (TR) of lithium-ion batteries (LIBs). This model only requires a minimum number of input parameters, and some of these unknown parameters can be obtained from accelerating rate calorimeter (ARC) tests and previous studies, removing the need for detailed measurements of heat flow of cell components by differential scanning calorimetry. The model was firstly verified by ARC tests for a commercial cylindrical 21700 cell for the prediction of the cell surface temperature evolution with time. It was further validated by uniform heating tests of 21700 cells conducted with flexible and nichrome-wire heaters, respectively. The validated model was finally used to investigate the critical ambient temperature that triggers battery TR. The predicted critical ambient temperature is between 127 °C and 128 °C. The model has been formulated as lumped 0D, axisymmetric 2D and full 3D to suit different heating and geometric arrangements and can be easily extended to predict the TR evolution of other LIBs with different geometric configurations and cathode materials. It can also be easily implemented into other computational fluid dynamics (CFD) code.

For small-scale pool fires, Vali et al. [1] showed a pair of vortices in the liquid pool. The first vortex appeared just close to the sidewall of the container, and the second one emerged slightly away from the first vortex. Large-eddy simulations of small methanol pool fires coupled with liquid fuel convective flow were conducted using an in-house version of FireFOAM to investigate the above phenomenon. In this study, a three-dimensional liquid phase model is newly developed. The model incorporates the effects of thermocapillary Marangoni convection, buoyancy, shear stress, and evaporation. For the gas phase, the combustion model is the extended eddy dissipation concept model coupled with the laminar combustion model. This combustion model uses the viscous diffusion rate to consider laminar-turbulent transition. The predictions were in reasonably good agreement with the measured local mass burning rate, flame height and distributions of liquid temperature. The error of the mass burning rate was within 4%. The present predictions captured a pair of vortices in line with Vali et al.'s experiment [1]. Their sizes increased with increasing the liquid temperature. The Reynolds analogy could explain the sensible reason behind this trend. Shear stress and thermocapillary force caused convection in the liquid pool, and this convection formed a pair of vortices. Thermocapillary force was due to the different distributions of convective and radiative heat transfer. Sensitivity test for sub-models for the liquid phase demonstrated that their effects on the mass burning rate were all less than 5.1%. Conversely, the simulation assuming zero gravity only in the liquid phase resulted in almost 64% reduction in the mass burning rate.

An approximate solution for the complete problem of attenuation of fire radiation by water mist is presented. This solution is based on simplified approaches for the spectral radiative properties of water droplets, the radiative transfer in the absorbing and scattering mist, and transient heat transfer taking into account partial evaporation of water mist. An analysis of the example problem makes it possible to recommend a decrease in the size of supplied water droplets with the distance from the irradiated surface of the mist layer. This can be achieved with the use of a multi-layered mist curtain. The advantage of this engineering solution is also confirmed by numerical calculations.

•Thermal management system (TMS) based on PCM-fin structure is proposed.•The electro-thermal model considers thermal contact resistance during heat transfer.•PCM-fin structure achieves superior thermal control compared with pure PCM.•Optimization design is discussed on the cooling performance of PCM-fin structure.•The suitable thermal performance is exhibited in continuous charge-discharge cycles. The safety, performance and durability of the Li-ion battery module are limited by the operating temperature especially in the hot temperature regions, hence the thermal management system is essential for battery module. In this paper a novel phase change material (PCM) and fin structure was proposed for the thermal management system of LiFePO4 battery module to reduce the maximum temperature and improve the temperature uniformity in high-temperature environment (40 °C). Carefully designed experiments were performed for model validation. The effects of PCM species, fin thickness, fin spacing and PCM thickness on the cooling performance of battery module were investigated numerically. The results showed that PCM-fin structure thermal management system with optimized design exhibited good thermal performance, keeping the maximum temperature of the battery surface under 51 °C at relatively high discharge rate of 3C. Moreover, by investigating the thermal behavior of PCM during discharge process and cycle test, it has been found that PCM-fin structure has the advantage of improving natural convection and heat conduction within the PCM structure, and as a result enhances heat dissipation efficiency and reduces failure risk in passive thermal management systems using PCMs.

The development and validation of CFD-DECOM, a pipeline depressurization model based on the arbitrary Lagrangian-Eulerian method (ALE) and the homogeneous equilibrium assumption is presented. In CFD-DECOM, the convection terms are separately solved from the other terms in a sub-cycled explicit manner using a sub-timestep that is only a fraction of the main computational timestep. This approach significantly simplifies the solution procedure and improves the computational efficiency. The model is validated against five release scenarios including one fast decompression of a rich gas pipeline and four slow blowdown cases of a liquefied petroleum gas pipeline. The predicted pressure, temperature and fluid inventory-time traces are found to be in good agreement with the measurements in all the cases. (C) 2014 Elsevier Ltd. All rights reserved.

The eddy dissipation concept (EDC) is extended to the large eddy simulation (LES) framework following the same logic of the turbulent energy cascade as originally proposed by Magnussen but taking into account the distinctive roles of the sub-grid scale turbulence. A series of structure levels are assumed to exist under the filter width “Δ” in the turbulent energy cascade which spans from the Kolmogorov to the integral scale. The total kinetic energy and its dissipation rate are expressed using the sub-grid scale (SGS) quantities. Assuming infinitely fast chemistry, the filtered reaction rate in the EDC is controlled by the turbulent mixing rate between the fine structures at Kolmogorov scales and the surrounding fluids. In order to extend the laminar smoke point soot model (SPSM) to LES, the partially stirred reactor (PaSR) concept is used to relate the filtered soot formation rate to the soot chemical time scale, which is assumed to be proportional to the laminar smoke point height (SPH) of the fuel. The turbulent mixing time scale for soot is computed as a geometric mean of the Kolmogorov and integral time scale. A new soot oxidation model is also developed by imitating the gas phase combustion within EDC. The newly extended EDC and SPSM are implemented in the open source FireFOAM solver and tested with two medium scale heptane and toluene pool fires with promising results. •The EDC is extended to the LES framework using the turbulent energy cascade.•The PaSR concept is used to extend the laminar-based SPSM to turbulent flames.•A new soot oxidation model is developed by imitating the EDC.•Promising results are achieved for heptane and toluene pool fire simulations.

A numerical approach is developed to simulate detonation propagation, attenuation, failure and re-initiation in hydrogen-air mixture. The aim is to study the condition under which detonations may fail or re-initiate in bifurcated tubes which is important for risk assessment in industrial accidents. A code is developed to solve compressible, multidi-mensional, transient, reactive Navier-Stokes equations. An Implicit Large Eddy Simulation approach is used to model the turbulence. The code is developed and tested to ensure both deflagrations (when detonation fails) and detonations are simulated correctly. The code can correctly predict the flame properties as well as detonation dynamic parameters. The detonation propagation predictions in bifurcated tubes are validated against the experimental work of Wang et al. [1,2] and found to be in good agreement with experimental observations. (C) 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Explosions in homogeneous reactive mixtures have been widely studied both experimentally and numerically. However, in accident scenarios, mixtures are usually inhomogeneous due to the localized nature of most fuel releases, buoyancy effects and the finite time between release and ignition. It is imperative to determine whether mixture in homogeneity can increase the explosion hazard beyond what is known for homogeneous mixtures. The present numerical investigation aims to study flame acceleration and transition to detonation in homogeneous and inhomogeneous hydrogen-air mixtures with two different average hydrogen concentrations in a horizontal rectangular channel. A density-based solver was implemented within the OpenFOAM CFD toolbox. The Harten-Lax-van Leer-Contact (HLLC) scheme was used for accurate shock capturing. A high resolution grid is provided by using adaptive mesh refinement, which leads to 30 grid points per half reaction length (HRL). In agreement with previous experimental results, it is found that transverse concentration gradients can either strengthen or weaken flame acceleration, depending on average hydrogen concentration and channel obstruction. Comparing experiments and simulations, the paper analyses flame speed and pressure histories, identifies locations of detonation onset, and interprets the effects of concentration gradients. (C) 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

In a compartment fire environment, the high temperature encountered could induce important stresses in glass panes, resulting into cracks and possible fallout of the glazing. The aim of the present work is to investigate thermal stress distributions in a glazing system for fire scenarios. A two dimensional glass thermal stress model to calculate the transient temperature and thermal stress distributions in a typical window glass under fire conditions was developed based on the Kong's work. The basic thermal conduction equation and thermal stress equation for glass were discretized by using the Galerkin method. A computer program based on the model was also developed. For validation purposes, simulations have been carried out using literature experimental data on glazing behavior in an enclosure fire. The glass surface temperature (exposed side) and thermal stress distributions in the glass pane were calculated. The simulation results of the transient temperature and thermal stress are overall in line with the experimental data reported in literature. The major principal thermal stress distribution in the glass at the time of first crack is consistent with the experimental crack patterns. The calculated maximum stress is located at the top edge of the glass pane, as the first crack recorded by experiments. The model does not predict second or later cracks. These results illustrate the relatively good predictions and usefulness of the developed simulation code.

采用预定的混合分数概率密度函数,将基于混合分数的离散反应模型拓展到火灾大涡模拟,并将其嵌入火灾模拟软件FireFOAM.单方程的亚网格湍流方程用于对经空间滤波的输运方程组进行封闭,辐射传热模型采用适用于光学厚火焰的P-1模型.通过模拟两种典型的火灾情景：池火和欠通风室火,对扩展的模型进行了验证.结果表明,模型可以给出合理的预测结果,适用于火灾数值预测.

To improve the current design standards of the hydrogen composite cylinders, it is essential to understand the thermal response of the hydrogen composite cylinders subjected to fire impingement. In the present study, a fully coupled conjugate heat transfer model based on a multi-region and multi-physics approach is proposed for modelling the transient heat transfer behaviour of composite cylinders subjected to fire impingement. The fire scenario is modelled using the in-house version of FireFOAM, the large eddy simulation (LES) based fire solver within the frame of OpenFOAM. Three dimensional governing equations based on the finite volume method are written to model the heat transfer through the regions of composite laminate, liner and pressurized hydrogen, respectively. The governing equations are solved sequentially with temperature-dependent material properties and coupled interface boundary conditions. The proposed conjugate heat transfer model is validated against a bonfire test of a commercial Type-4 cylinder and its transient heat transfer behaviour is also studied. •Simple model with 4 parameters for predicting overpressure in vented explosions.•Two of these parameters only depend on fuel and are pre-tabulated.•Other two parameters are simple functions of enclosure geometry.•Predictions either more accurate or comparable with other models in literature.•Reasonably good predictions for realistic accidental scenarios.

Computational Fluid Dynamics (CFD) codes are widely used for gas dispersion studies on offshore installations. The majority of these codes use single-block Cartesian grids with the porosity/distributed-resistance (PDR) approach to model small geometric details. Computational cost of this approach is low since small-scale obstacles are not resolved on the computational mesh. However, there are some uncertainties regarding this approach, especially in terms of grid dependency and turbulence generated from complex objects. An alternative approach, which can be implemented in general-purpose CFD codes, is to use body-fitted grids for medium to large-scale objects whilst combining multiple small-scale obstacles in close proximity and using porous media models to represent blockage effects. This approach is validated in this study, by comparing numerical predictions with large-scale gas dispersion experiments carried out in DNV GL's Spadeadam test site. Gas concentrations and gas cloud volumes obtained from simulations are compared with measurements. These simulations are performed using the commercially available ANSYS CFX, which is a general-purpose CFD code. For comparison, further simulations are performed using CFX where small-scale objects are explicitly resolved. The aim of this work is to evaluate the accuracy and efficiency of these different geometry modelling approaches. (C) 2016 Elsevier Ltd. All rights reserved.

The manuscript investigated experimentally the influence of two parallel walls on the flame geometric parameters and radiative heat flux distributions along the wall from fires on rectangular burners. Four sets of burners with equal area and different aspect ratios were used. The burner aspect ratios, separation distance between two parallel walls and the fire heat release rates were systematically varied in the tests. Measurements were conducted for the flame height evolution and radiation hazards. The results were analysed to build the change trends of the vertical flame height and radiant heat flux with the change of the separation distance between two parallel walls. A normalized flame height equation incorporating the separation distance was proposed. The results also revealed that the radiant heat fluxes along a vertical target do not change monotonously. Comparison between the measurements and the radiant heat fluxes calculated by some published empirical models revealed relatively large discrepancies. A view factor based formula was hence proposed by assuming the flame shape as a triangular prism based on the probability flame contours for relatively larger burner aspect ratios (n ≥ 3) and found to correlate well with the measurements.

The effect of pressure boundary rupture rate on hydrogen spontaneous ignition has been numerically investigated. A mixture-averaged multi-component approach was used for accurate calculation of molecular transport. Spontaneous ignition and combustion chemistry were accounted for using a 21-step kinetic scheme. A 5th-order WENO scheme coupled with ultra fine meshes was employed to reduce false numerical diffusion. The study has demonstrated that the rupturing process of the initial pressure boundary has important influence on the spontaneous ignition of pressurized hydrogen release. When the pressure boundary rupture rate is below a certain threshold value, the predictions showed that there would be no spontaneous ignition. As the rupture rate increases, the shock-heated air temperature drops more quickly due to earlier flow expansion. Once the rupture rate is sufficiently high, spontaneous ignition can still occur. However, the initial flame width would be narrower compared to the sudden release case.

Dynamic predictions of the mass burning rate of pool fires under different burner conditions are essential to facilitate pool fire simulations without the need for artificially setting the inlet boundary conditions for the fuel surface. Such capability can remove the need for prescribed mass burning rates of pool fires in quantified assessment of the fire hazards. A fully coupled three-dimensional (3-D) model based on a multi-zone approach has been developed. In the gas-phase region, a compressible solver was employed. In the liquid-phase region, an incompressible solver with temperature-dependent thermophysical properties was utilized to directly solve fuel flow, accounting for the Marangoni effect, buoyancy effect, and incident radiation. In the solid-phase region, the 3-D heat transfer equation was resolved. The heat and mass transfer processes between different regions were simulated using conjugate heat transfer and an evaporation model based on "film theory". The proposed model has been validated through comparison with the 9 cm diameter methanol pool fire experiments. The predictions showed promising agreement with experimental measurements and empirical corrections, with the error in mass burn rate being within 3.1 %. Additionally, the predictions have captured a pair of vortices in sizes and directions closely resembling experimental observations. The sizes of the predicted vortices increased with the rising temperature at the base of the pool due to buoyancy and shear force. The analysis revealed that the wall effect not only leads to differences in the number of vortices and Marangoni velocity but also leads to a smaller mass burning rate in the burner with a high thermal conductivity than in the one with a poor thermal conductivity in the 9 cm diameter methanol pool fire. Neglecting the wall heat transfer would result in up to 18 % underprediction of the mass burning rate.

Computational Fluid Dynamics solvers are developed for explosion modelling and hazards analysis in Hydrogen air mixtures. The work is presented in two parts. These include firstly a numerical approach to simulate flame acceleration and deflagration to detonation transition (DDT) in hydrogen air mixture and the second part presents comparisons between two approaches to detonation modelling. The detonation models are coded and the predictions in identical scenarios are compared. The DDT model which is presented here solves fully compressible, multidimensional, transient, reactive Navier-Stokes equations with a chemical reaction mechanism for different stages of flame propagation and acceleration from a laminar flame to a highly turbulent flame and subsequent transition from deflagration to detonation. The model has been used to simulate flame acceleration (FA) and DDT in a 2-D symmetric rectangular channel with 0.04 m height and 1 m length which is filled with obstacles. Comparison has been made between the predictions using a 21-step detailed chemistry as well as a single step reaction mechanism. The effect of initial temperature on the run-up distances to DDT has also been investigated. In the second part, one detonation solver is developed based on the solution of the reactive Euler equations while the other solver has a simpler approach based on Chapman-Jouguet model and the programmed CJ burn method. Comparison has shown that the relatively simple CJ burn approach is unable to capture some very important features of detonation when there are obstacles present in the cloud. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

Abstract only

This paper aims to provide a comprehensive review of available empirical models for overpressures predictions of vented lean hydrogen explosions. Empirical models and standards are described briefly, with discussion on salient features of each model. Model predictions are then compared with the available experimental results on vented hydrogen explosions. First comparison is made for standards tests, with empty container and quiescent starting conditions. Comparisons are then made for realistic cases with obstacles and initial turbulent mixture. Recently, a large number of experiments are carried out with standard 20-foot container for the HySEA project. Results from these tests are also used for model comparison. Comments on accuracy of model predictions, their applicability and limitations are discussed. A new model for vented hydrogen explosion is proposed. This model is based on external cloud formation, and explosion. Available experimental measurements of flame speed and vortex ring formation are used in formulation of this model. All assumptions and modelling procedure are explained in detail. The main advantage of this model is that it does not have any tuning parameter and the same set of equations is used for all conditions. Predictions using this model show a reasonably good match with experimental results. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Numerical investigations have been conducted for flame acceleration and transition to detonation in a horizontal obstructed channel with 60 percent blockage ratio filled with hydrogen/air mixture. Both homogeneous and inhomogeneous hydrogen/air mixtures have been considered. The later has a vertical concentration gradient. The density-based solver within the OpenFOAM CFD toolbox developed by the present authors [1] is used. High-resolution grids are facilitated by using adaptive mesh refinement technique, which leads to 30 grid points per half-reaction length (HRL) in the finest region near the flame and shock fronts. The forward and backwards jets which represent Richtmyer-Meshkov (RM) instability, were found to impact on the shock front, resulting in the appearance of a secondary triple point on the initial Mach stem on the flame front. Moreover, since both the forward and backwards jet propagates in the shear layer, some small vortices can be found on the surface of the secondary shear layer, which represents the Kelvin-Helmholtz (KH) instability. Additionally, it has been found that the inhomogeneous (non-uniform) mixtures cause higher shock and flame velocities compared to the homogeneous mixtures concentration. Also, for both homogenous and inhomogeneous mixtures with 30% hydrogen concentration, the onset of detonation occurs within the obstructed channel section, but the homogeneous mixtures show slightly faster flame acceleration and earlier onset.

In a previous paper, we reported the development of CFD-DECOM, a Computational Fluid Dynamics (CFD) model based on the Arbitrary Lagrangian Eulerian (ALE) approach and the Homogeneous Equilibrium Method (HEM) for simulating multi-phase flows, to predict the transient flow following the rupture of pipelines conveying rich gas or pure carbon dioxide (CO2). The use of CFD allows the effect of pipe wall heat transfer and friction to be quantified. Here, the former is considered through the implementation of a conjugate heat transfer model while the two-phase pipe wall friction is computed using established correlations. The model was previously validated for rich gas and to a limited extent dense phase CO2 decompression against the available shock tube test data. This paper describes the extension of the model to the decompression of both gaseous and dense phase CO2 with impurities. The Peng-Robinson-Stryjek-Vera Equation Of State (EOS), which is capable of predicting the real gas thermodynamic behaviour of CO2 with impurities, has been implemented in addition to the Peng-Robinson and Span and Wagner EOSs. The liquid-vapour phase equilibrium of a multi-component fluid is determined by flash calculations. The predictions are compared with the measurements of some of the recent gaseous and dense phase CO2 shock tube tests commissioned by National Grid. The detailed comparison is presented showing reasonably good agreement with the experimental data. Further numerical study has also been carried out to investigate the effects of wall friction and heat transfer, different EOSs and impurities on the decompression behaviour.

Spontaneous ignition of pressurized hydrogen release through a tube into air is investigated using a modified version of the KIVA-3V CFD code. A mixture-averaged multi-component approach is used for accurate calculation of molecular transport. Autoignition and combustion chemistry is accounted for using a 21 step kinetic scheme. Ultra fine meshes are employed along with the Arbitrary Lagrangia–Eulerian (ALE) method to reduce false numerical diffusion. The study has demonstrated a possible mechanism for spontaneous ignition through molecular diffusion. In the simulated scenario, the tube provided additional time to achieve a combustible mixture at the hydrogen–air contact surface. When the tube was sufficiently long under certain release pressure, autoignition would initiate inside the tube at the contact surface due to mass and energy exchange between low temperature hydrogen and shock-heated air through molecular diffusion. Following further development of the hydrogen jet downstream, the contact surface became distorted. Turbulence plays an important role for hydrogen/air mixing in the immediate vicinity of this distorted contact surface and led the initial laminar flame to transit into a stable turbulent flame.

Abstract only

A numerical approach has been developed to simulate flame acceleration and deflagration to detonation transition in hydrogen-air mixture. Fully compressible, multidimensional, transient, reactive Navier Stokes equations are solved with a chemical reaction mechanism which is tuned to simulate different stages of flame propagation and acceleration from a laminar flame to a turbulent flame and subsequent transition from deflagration to detonation. Since the numerical approach must simulate both deflagrations and detonations correctly, it is initially tested to verify the accuracy of the predicted flame temperature and velocity as well as detonation pressure, velocity and cell size. The model is then used to simulate flame acceleration (FA) and transition from deflagration to detonation (DDT) in a 2-D rectangular channel with 0.08 m height and 2 m length which is filled with obstacles to reproduce the experimental results of Teodorczyk et al. The simulations are carried out using two different initial ignition strengths to investigate the effects and the results are evaluated against the observations and measurements of Teodorczyk et al. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

Numerical simulations of far-field carbon dioxide dispersion were conducted for a vertical vent release and a horizontal release from a shock tube. These scenarios had also been studied experimentally at field scale commissioned by National Grid. This work and the experiments both form part of the National Grid dense phase CO2 pipeLine TRANSportation (COOLTRANS) research programme. All tests involved releases of dense phase CO2 into an atmospheric flow. The dispersing plumes were subjected to transient wind conditions where both the direction and magnitude of the wind fluctuated with time. As part of the COOLTRANS research programme, the far-field dispersion simulations started from source terms derived from the near-field simulations conducted by the University of Leeds and outflow simulations conducted by University College London. The numerical model used for the far-field simulations is based on OPENFOAM, which is an object-oriented open source computational fluid dynamics toolbox. A dedicated solver CO(2)FOAM has been developed within the framework of OPENFOAM for simulating dispersion from dense phase CO2 releases. This has included the implementation of the homogeneous equilibrium method for fully compressible two-phase flow, treatment of the transient atmospheric boundary conditions and the time-varying inlet boundary conditions. The experimental measurements were supplied to the authors after the predictions were completed and submitted to National Grid. Hence, the validation reported here is indeed "blind." While further fine tuning of the model and validation is still underway, the relatively good agreement between the predictions and measurements in the present study has demonstrated the potential of CO(2)FOAM as an effective predictive tool for far-field CO2 dispersion in the context of pipeline transportation for carbon capture and storage.

Experimental investigations were conducted to characterise the impacts of crosswind and burner aspect ratio on the flame evolution characteristics and flame base drag length of gas diffusion flames on rectangular burners. The burners have the same surface area of approximately 100 cm(2). The tests to capture the flame base drag length were conducted three times for each condition with the differences between the original and repeated tests being less than 6%. The thermocouple readings were corrected for the effect of radiative and convective heat exchange with the surroundings. Overall, 84 independent test conditions were conducted on 4 different burner aspect ratios, 3 fuel supply rates and 7 crosswind conditions. The changing behaviour of the flame with different burner aspect ratios, heat release rates and crosswind speeds were carefully analysed. The appearance of "blue flames" in the upstream edge of the main diffusion flames just above the burner in relatively strong winds was analysed. Unlike the flame tilt angle and flame height which either increase (the former) or decrease (the later) monotonically with the increase of wind speed, the flame base drag length was found to increase with the wind speed firstly until a critical point and then decrease with further increase of the crosswind for a given heat release rate. This is thought to be due to the competing influence of thermal buoyancy and wind induced inertial forces. The transition point for the maximum flame base drag length with regard to crosswind was found to decrease with the increasing aspect ratio of the burner for a given heat release rate. A new physics-based correlation considering decay phase with the crosswinds was proposed for the flame base drag length incorporating all important physical factors including inertia force, fire induced thermal buoyancy, Froude number, dimensionless heat release rate and fuel/air density ratio. The proposed formulations were found to correlate well with the current measurements of gas burner fires as well as some published data in the literature for pool fires on the ground which were not used in their derivation. (C) 2018 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Spontaneous ignition of compressed hydrogen release through a length of tube with different internal geometries is numerically investigated using our previously developed model. Four types of internal geometries are considered: local contraction, local enlargement, abrupt contraction and abrupt enlargement. The presence of internal geometries was found to significantly increase the propensity to spontaneous ignition. Shock reflections from the surfaces of the internal geometries and the subsequent shock interactions further increase the temperature of the combustible mixture at the contact region. The presence of the internal geometry stimulates turbulence enhanced mixing between the shock-heated air and the escaping hydrogen, resulting in the formation of more flammable mixture. It was also found that forward-facing vertical planes are more likely to cause spontaneous ignition by producing the highest heating to the flammable mixture than backward-facing vertical planes. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

The present study reports on the development and validation of a finite element program, GLAZ-CRACK, for predicting crack initiation and propagation of glass in fire or under other thermal loadings. The model is based on three crack modes to calculate the stress intensity factors (SIFs) and strain energy release rates. The crack initiation is predicted from the stress distribution using either probabilistic or deterministic method. The crack growth can be predicted by one of the three criterions, which are SIFs based mixed-mode criterion, energy release rates based mixed-mode criterion and SIFs based maximum circumferential stress criterion. The crack spread rate and crack direction are calculated based on first principles of fracture mechanics. A moving crack tip mesh topology is proposed to locally refine the grid resolution in the tip region. Predictions for the SIFs of a central horizontal crack in a square plate, a central horizontal crack in a long plate and a single edge cracked plate under plane stress condition show good agreement with either the previous predictions of ANSYS or the theoretical values. Exploratory calculations of a single crack under thermal loading have shown that the crack initiation and crack propagation pattern agree with the experimental observations. (C) 2013 Elsevier Ltd. All rights reserved.

Insight of thermal behaviour of lithium-ion batteries under various operating conditions is crucial for the development of battery management system (BMS). Although battery thermal behaviour has been studied by published models, the reported modelling normally addresses either normal operation or thermal runaway condition. A comprehensive electro-thermal model which can capture heat generation, voltage and current variation during the whole process from normal cycling to thermal runaway should be of benefit for BMS by evaluating critical factors influencing potential transition to thermal runaway and investigating the evolution process under different cooling and environment conditions. In this study, such a three-dimensional model has been developed within the frame of open source computational fluid dynamics (CFD) code OpenFOAM to study the electrical and thermal behaviour of lithium-ion batteries (LIBs). The equations governing the electric conduction are coupled with heat transfer and energy balance within the cell. Published and new laboratory data for LiNi0.33Co0.33Mn0.33O2/Li1.33T1.67O4 (LNCMO/LTO) cells from normal cycling to thermal runaway have been used to provide input parameters as well as model validation. The model has well captured the evolution process of a cell from normal cycling to abnormal behaviour until thermal runaway and achieved reasonably good agreement with the measurements. The validated model has then been used to conduct parametric studies of this particular type of LIB by evaluating the effects of discharging current rates, airflow quantities, ambient temperatures and thickness of airflow channel on the response of the cell. Faster function losses, earlier thermal runaway and higher extreme temperatures were found when cells were discharged under higher current rates. The airflow with specific velocity was found to provide effective mitigation against over-heating when the ambient temperature was below 370 K but less effective when the ambient temperature was higher than the critical value of 425 K. The thickness of airflow channel was also found to have critical influence on the cell tolerance to elevated temperatures. These parametric studies demonstrate that the model can be used to predict potential LIB transition to thermal runaway under various conditions and aid BMS.

Growth in demand for Liquefied Natural Gas (LNG) has increased calls for further research and development on LNG production and safer methods for its transportation. This paper presents the implementation of numerical models for dispersion of evaporated LNG in the open atmosphere. The developed model incorporates in its formulation LNG spill and pool formation into a source model. It is then coupled with a Computational Fluid Dynamics (CFD) approach in OpenFOAM for dispersion calculations. Atmospheric conditions such as average wind speed and direction were used to resolve wind boundary layers. The model also accounts for the humidity effect and its influence on air-density and buoyancy change. Verifications have been conducted using the experimental results from Maplin Sands series of tests by comparing the maximum evaporated gas concentration in every arc in relation to the release point. The results show good agreements between the model's predictions and experiments.

This paper describes the application of a fire field model combined with detailed chemistry to the simulation of sooting propane jet fires in a 135 m 3 compartment. The purpose of the research is to investigate the behavior of under-ventilated jet fires, and the formation of toxic products. The kinetic formation of major species (CO, C 2H 2) and trace species (soot) has been studied using, a detailed reaction mechanism, for propane. The concept of strained laminar diffusion flamelet was adopted to model the main hydrocarbon combustion and the formation of soot. The turbulence-chemistry, interaction followed the conserved scalar and assumed, probability density function (PDF) description of turbulent diffusion flames. The soot formation was modeled by the two-equation approach. Instantaneous temperatures given by the flamelet were modified to account for radiative heat loss. The predictions for velocity, CO, soot, and other main species are given in the paper, and comparison is made between predictions and measurements on a 1,5-MW fire test case that showed the important effect of the entrainment on the formation of toxic species in the enclosure. The predicted result of soot at the vent location is in agreement with the experimental data. The general trend, of temperature distributions have been correctly predicted, but the neglecting of convective heat transfer from the compartment walls and the simplified treatment for radiative heat loss from the jet fire has resulted in some discrepancies on the predictions of temperatures, particularly at the vent location.

•The inhibition of N2 and CO2 on methane/hydrogen/air premixed flame was investigated.•CO2 has better inhibiting performance than N2 at corresponding conditions.•The inhibiting mechanisms were revealed from thermal and kinetic aspects. Hydrogen enriched natural gas (HNG) is a promising alternative fuel. But the blended fuel will inevitably have different ignition and combustion characteristics as compared to natural gas. The extent of the resulting difference depends on the percentage of hydrogen addition. It may affect the compatibility of combustion systems and have safety implications. The present study was aimed at enhancing the safety of HNG through inhibition by inert gases. Laboratory tests were conducted for methane/hydrogen/air premixed flame propagating in a closed channel with either nitrogen (N2) or carbon dioxide (CO2) as the inhibitor. Mixtures with different hydrogen volumetric fractions in fuel, including 0%, 20%, 50% or 80% were investigated. The flame shape evolution and the overpressure in the channel were recorded by high-speed Schlieren photography and pressure sensor, respectively. The flame shape was found to change in various ways depending on the inhibitor and hydrogen content. The pressure wave had remarkable impacts on flame and pressure dynamics. The effect of buoyancy on the flame deformation was observed and discussed. Both N2 and CO2 were found to have considerable inhibiting effect on the flame speed and overpressure build-up in the channel while the inhibiting effect of CO2 was stronger. The inhibition mechanisms of either N2 or CO2 were revealed from thermal and kinetic aspects.

A modelling strategy has been developed for consequence analysis of medium and large scale gaseous detonation. The model is based on the solution of Euler equations with one-step chemistry. The van Leer flux limited method which is a total variation diminishing scheme is used for shock capturing. Preliminary calculations were firstly conducted for small domains with fine grids which resolve the wave, relatively coarse grids which have less than 10 grids across the wave and coarse grids in which the minimum grid size is larger than the wave thickness to ensure that the reaction scheme has been properly tuned to capture the correct detonation pressure, temperature and velocity in the resolutions used in the different cases. The model was firstly tested against a medium scale detonation test in a shock tube with U-bends. Reasonably good agreement is achieved on detonation pressure and mean shock wave velocities at different measuring segments of the tube. Following the validation, the detonation of a hypothetical planar propane-air cloud is simulated. The predictions uncovered some interesting features of such large scale detonation phenomena which are of significance in the safety context, especially for accidental investigations. The findings from the present analysis are in line with the forensic evidence on damages in some historic accidents and challenges previous analysis of a major accident in which forensic evidence suggested localised detonation but was considered as the consequence of fire storms by the investigation team.

The computational analysis of downward motion and evaporation of water droplets used to suppress a typical transient pool fire shows local regions of a high volume fraction of relatively small droplets. These droplets are comparable in size with the infrared wavelength in the range of intense flame radiation. The estimated scattering of the radiation by these droplets is considerable throughout the entire spectrum except for a narrow region in the vicinity of the main absorption peak of water where the anomalous refraction takes place. The calculations of infrared radiation field in the model pool fire indicate the strong effect of scattering which can be observed experimentally to validate the fire computational model. (C) 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license.

•Critical breakage condition of coated, insulated and laminated glass was determined.•Insulated and laminated glass can survive much longer than the single ones in a fire.•Laminated glass demonstrates good capability to prevent new vent formed in fires.•Heat transfer mechanism of three different glasses was revealed and compared. To make constructions more artistic, various new kinds of glazing are increasingly employed in building envelopes. However, when subjected to a fire, these glass façades may easily break and fall out, significantly accelerating the development of enclosure fire. Thus, it is necessary to investigate and compare their different fire performance and breakage mechanisms. In this work, a total of ten tests, including single coated, insulated and laminated glazing, were heated by a 500×500mm2 pool fire. Breakage time, glass surface and air temperature, incident heat flux and crack initiation and propagation were obtained. The critical conditions of three different kinds of glazing were determined. It was established that the insulated and laminated glass can survive longer than the single glass. The air gap and fire side glass pane was found to play a key role for the thermal resistance of ambient side pane in the insulated glazing. Although both panes of the laminated glazing broke, it could be held together by the layer of gel, effectively avoiding the formation of a new vent. Numerical simulations were performed to investigate the heat transfer process through the glazing panels and the temperatures in the glazing were predicted well. Suggestions for glass fire resistance design are proposed.

Numerical simulations have been carried out for spontaneous ignition in the sudden release of pressurized hydrogen into air. A mixture-averaged multi-component approach was used for accurate calculation of molecular transport. Spontaneous ignition and combustion chemistry were accounted for using a 21-step kinetic scheme. To reduce false numerical diffusion, extremely fine meshes were used along with the arbitrary Lagrangian–Eulerian (ALE) method in which convective terms are solved separately from the other terms. Spontaneous ignition of pressurized hydrogen release was previously observed in laboratory tests and suspected as a possible cause of some accidents. The present numerical study has successfully captured this phenomenon and demonstrated a possible mechanism for spontaneous ignition due to molecular diffusion between the shock-heated air and the expanding hydrogen. The role of turbulence in the mixing at the region of the distorted hydrogen–air contact surface and the potential development from the initial laminar flame to a turbulent jet flame has also been discussed.

Unplanned and uncontrolled releases of hazardous materials could result in major casualties and environment impact. Natural gas (sour) containing significant amounts of hydrogen sulfide is toxic, flammable and corrosive. Incidents like 2003 Kaixian blowout ("12.23 disaster") and 2015 Saskatchewan in Oil & Gas industry highlight the significance of conducting appropriate technical (safety) risk assessments and exercising effective risk management. It is estimated that for 90% of accidental releases of sour gas during pipeline transfer results toxic impacts as typical releases will not get immediately ignited. Lack of adequate information of the hazards and the lack of realistic estimate of the hydrogen sulfide toxic exposure zone are the main challenges in addressing the risk to public from sour natural gas. The challenge risk analysts come across is the lack of guidance on appropriate tool and methodology to estimate the toxic impact zone following an accidental loss of containment. Dispersion following accidental release of high pressure and high flow rate sour gas in complex terrain should take account of multicomponent thermodynamics, terrain effect and the phase transitions. For selecting processing sites, pipeline routes etc., stakeholders require convincing results addressing the uncertainties. Simple correlation like Gaussian model alone is not considered as suitable and appropriate. This paper is based on the academic research conducted to overcome the uncertainties in sour gas dispersion modelling. The focus of this research is on the dispersion following an accidental release from sour natural gas pipeline. The expansion following release and the initial air entrainment will be estimated to determine a range of cloud behaviour. Based on the sensitivity analysis, this paper provides guidance on the natural gas composition and the source term characteristics to define and select the appropriate dispersion phenomenon. The results and analysis will minimize the knowledge gap/uncertainty with the consequence calculations by identifying the key assumptions and parameters that should be put through sensitivity analysis.

Experimental investigations have been conducted in hydrogen-oxygen mixtures with equivalence ratio of 1.5 at cryogenic temperature (77 K) and initial pressure (P0) from 0.2 to 0.5 atm. The pressure sensors and optical fibers are used to measure the overpressure and flame velocity, respectively. The precursor shock wave is formed by the combination of the 1st and 2nd shock waves. When flame catches up with the precursor shock wave, detonation occurs. Strong flame acceleration was observed in all tested conditions, which can be well characterized and predicted by the Zel'dovich number and expansion ratio. The stuttering and galloping modes were observed at 0.5 and 0.3 atm, respectively. With the decrease of the initial pressure to 0.20–0.25 atm, detonation could not occur, only the deflagration mode was observed. The stability of the mixture can be indicated by the parameter χ, and the improvement of stability of mixture will shorten the initial pressure range where the galloping mode occurs. The heat loss effect on detonation limit has subsequently been examined for the present experiments as well as those with equivalence ratio of 2.6 reported in our previous study [Shen, X. et al., Proceedings of the Combustion Institute 2022]. For equivalence ratio of 2.6, the heat loss is found to have dominant effect on the appearance of detonation limit. Its influence can be quantitatively characterized by critical ratio of heat loss to heat release. However, for equivalence ratio of 1.5, the heat loss effect is found to have only relatively small effect on detonation limit. •The combustion of hydrogen-oxygen mixture at cryogenic temperature of 77 K is investigated.•The 1st and 2nd shock waves would merge to form a stronger precursor shock wave.•Strong flame acceleration could be predicted by expansion ratio and Zel'dovich number.•The stuttering mode and the galloping mode of detonation are both observed.•The effect of heat loss on the detonation limit is examined.

Heptane pool fires with different mass burning rates ranging from 0.00123kg/s (corresponding to steady burning stage in experiments) to 0.0247kg/s (corresponding to fuel boiling burning stage) on a hollow square pan are numercally studied in this paper. The fireFoam, the large eddy simulation (LES)–based fire simulation code within OpenFOAM®, was employed with the modified Eddy Dissipation Concept combustion model[13] and PaSR based soot model[14]. The predictions achieved good qualitative and quantitative agreement with the experimental data[9]. Further analysis was conducted on the fire behaviour and in particular the fire merging phenomena. The dependency betwwen the key parameters such as flame height, centerline temperature and soot distributions on the mass burning rate was also discussed.

A laboratory experimental work is carried out to investigate the attenuation ability of water sprays subjected to thermal radiation. The objective is to analyze the key parameters involved in the mitigation properties of this fire protection technique. The spectral transmittances of two types of sprayers, TG03 and TG05, are measured with a Fourier infrared spectrometer under various conditions. The wavelength range varies from 1.5 to 12 μm. The influence on the transmittance of both the flow rate and the pressure ranging from 1 to 7 bars, as well as the effect of the number of spray nozzles are considered. The results clearly show the advantage of small drops with high concentration. An investigation on the multi-ramp curtain configuration also provides valuable information on the mitigation behavior of the whole spray. Key guidelines are provided for fire protection engineering.

An investigation is carried out to assess the two-flux model for evaluating the attenuation ability of a water spray curtain in fire protection. Transmittances calculated with this model are compared with the “exact” discrete ordinates solutions for a range of water curtains under practical conditions. The results show the unsuitability of the two-flux method under a collimated incidence boundary condition, even if some improvements could be expected with very small droplets. Whereas the diffuse incidence type provides relatively better results, it is more reliable for transmittance calculations. [S0022-1481(00)00101-8]

A sub-grid scale model for partially premixed combustion has been implemented into an existing LES code and applied to simulate turbulent deflagration in backdraft and its mitigation by watermist in a scaled-compartment with different vent geometries. The model is based on the coupling of independent approaches for non-premixed and premixed turbulent combustion, while the flame index concept was used to separate the two different combustion regimes. Simulations were conducted for the reduced scale tests of Weng and Fan ( 2002 ). Reasonable agreements have been obtained for species concentrations, total mass outflow and inflow rates, maximum pressure, and likelihood of the occurrence of the fireball outside the container. The numerical study has highlighted the mitigation effect of watermist by reducing the laminar burning velocity of the mixture and the influence of the end opening geometries on the occurrence of backdraft.

•Two-step method is applied to radiation calculation of developing flames.•Focusing of evaporating water droplets in local areas of the flame is studied.•A strong infrared scattering by small water droplets is analyzed.•The use of the infrared scattering in flame observations is discussed. A procedure based on two-step method is suggested to simplify time-consuming spectral radiative transfer calculations in open flames containing scattering particles. At the first step of the problem solution, the P1 approximation is used to calculate the divergence of radiative flux, and it is sufficient to determine the flame parameters. The second step of solution is necessary to obtain the radiation field outside the flame, and this can be made independently using the ray-tracing procedure and the transport source function determined at the first step. Such a splitting of the complete problem results in much simpler algorithm than those used traditionally. It has been proved in previous papers that the combined two-step method is sufficiently accurate in diverse engineering applications. At the same time, the computational time decreases in about two orders of magnitude as compared with direct methods. An axisymmetric pool fire at the initial stage of fire suppression by a water spray is considered as the case problem. It is shown that evaporating small water droplets characterised by a strong scattering of infrared radiation are mainly located in regions near the upper front of the flame and one can observe the scattered radiation. This effect can be used in probe experiments for partial validation of transient Computational Fluid Dynamics (CFD) simulations.

To satisfy the needs of large-scale hydrogen combustion and explosion simulation, a method is presented to establish single-step chemistry model and transport model for fuel-air mixture. If the reaction formula for hydrogen-air mixture is H 2 +0.5O 2 →H 2 O, the reaction rate model is ω̇ = 1.13×10 15 [H 2 ][O 2 ]exp(−46.37 T 0 / T ) mol (cm 3 s) −1 , and the transport coefficient model is µ= K/C P = ρD =7.0×10 −5 T 0.7 g (cm s) −1 . By using current models and the reference model to simulate steady Zeldovich-von Neumann-Doering (ZND) wave and free-propagating laminar flame, it is found that the results are well agreeable. Additionally, deflagration-to-detonation transition in an obstructed channel was also simulated. The numerical results are also well consistent with the experimental results. These provide a reasonable proof for current method and new models.

A variation of the Laminar Flamelet Decomposition (LFD) method for the Conditional Source Term (CSE) model developed by Bushe and Steiner (Phys Fluids 15:1564–1575, 2003 ) is implemented into an existing LES code. In this approach, the set of basis functions, on which the decomposition is based, is reduced using the mixture fraction dissipation rate as external parameter for the selection. It was found that reducing the basis improves and stabilises the inversion, resulting in reasonably accurate approximation for the average conditional quantities. Some modifications have been introduced to improve the inversion process by reducing the number of flamelets. This modification is found to help stabilize the inversion and keep the dimension of the linear system small. The model is used to simulate the turbulent non-premixed piloted SANDIA Flame D. Reasonably good predictions for conditional and unconditional average variables were found for different planes and at centreline of the flow field. However, an over prediction of the consumption rate in the near field of the flame is found, which may be partially attributed to the use of the Steady Laminar Flamelets (SLF) as functions for the decomposition and the use of a constant boundary condition for the species mass fractions in solving the flamelets. The present simulation of a turbulent reacting jet is the first test of the LFD approach in a realistic scenario using only the temperature field to calculate the inversion. The model is found to be computationally inexpensive.

The present study reports a modular phenomenological model for predicting peak pressure in vented explosions. Modelling assumptions are explained in detail and model components are validated against experimental and computational results. A basic version of this model is reported in our earlier paper (Sinha et al., 2019). Previous experimental and modelling efforts on vented explosion have primarily focussed on idealized condition of empty container with uniformly mixed fuel. However, in real accidents, there are often obstacles in flame path, and a leaked fuel may not get enough time to mix uniformly. These realistic accidental scenarios are accounted for in this extended model. First the model components are assessed using available experimental results. Comparison of flame arrival time and flame propagation inside the enclosure are made, which demonstrate the ability of the model to capture flame propagation accurately. Suggestions are also made for vent panel installation to reduce peak overpressure in accidental explosions. Predictions for external cloud radius and pressure generated by external explosion are found to be in close agreement with the experimentally measured values. The model is further simplified, and a final equation is proposed which depends on two fuel related parameters and two geometric parameters. Fuel dependent parameters are pre-tabulated, and geometric parameters are easy to compute. Procedure to calculate pressure generated by external explosion and internal pressure are outlined in detail. Experimental results available in literature are used to evaluate model predictive capabilities. The model, in principle should be applicable for any gaseous fuel. However, the focus of the present investigation is to assess it for hydrogen explosions. Experimental repeatability is also discussed, and role of wall deflection is highlighted. In parallel to the modelling effort, a dedicated in-house CFD solver HyFOAM is developed utilizing OpenFOAM platform. The HyFOAM predictions are validated against experimental results from the recently published test data involving hydrogen explosion in a 20-foot ISO container (Skjold et al., 2017, 2018, 2019). Moreover, as experimental investigations are expensive and require significant testing and safety infrastructure, a limited number of scenarios can be tested experimentally. In addition to the experimental results, few more cases are simulated using HyFOAM. Phenomenological model results are compared with the CFD results, and a reasonably good match is observed.

Modelling the atomization process in fire sprinklers has remained a challenge mainly due to the complexity of sprinkler geometry. A review of existing fire sprinkler spray modelling approaches, including film flow and sheet tracking models, showed that they mainly assumed a constant sheet velocity and linear attenuation of the sheet thickness before its disintegration. In the present study, a liquid sheet trajectory sub-model based on the solution of stream-wise conservation equations has been used to predict both sheet thickness and velocity as it radially expands. This will also help to investigate the extent to which a change in the release angle can affect the sheet characteristics. The analysis carried out shows that the proposed approach improves the predictions of mean droplet diameter and initial droplet speed. A semi-empirical approach is further introduced in the study by using experimental volume fraction measurements to characterize sprinkler sprays in the near field. For a given direction predictions have been conducted for droplet volume median diameter, water volume flux and droplet average velocity at different elevation and azimuthal locations. A reasonably good agreement is found for the near field measurements. (C) 2014 Elsevier Ltd. All rights reserved.

In this paper the blast resistance of cracked steel structures repaired with fibre-reinforced polymer (FRP) composite patch are investigated. The switch box which has been subjected to blast loading is chosen to study. The steel material is modelled using isotropic hardening model, pertaining to Von Mises yield condition with isotropic strain hardening, and strain rate-dependent dynamic yield stress based on Cowper and Symonds model. Three different cracked structures are chosen to investigate their capability in dissipating the blast loading. To improve the blast resistance, the cracked steel structures are stiffened using carbon fibre-reinforced polymer (CFRP) composite patches. The repaired patches reduce the stress field around the crack as the stress is transferred from the cracked zone to them. This situation prevents the crack from growing and extends the service life of the steel structure. It will be shown that CFRP repairing can significantly increase the blast resistance of cracked steel structures. (C) 2010 Elsevier Ltd. All rights reserved.

A series of experiments were carried out in a closed tube at cryogenic temperature (77 K) for hydrogen-oxygen mixtures. Flame propagation speed and overpressure were measured by optical fibers and pressure sensors, respectively. The first and second shock waves were captured in the cryogenic experiments, although the shock waves always precede the flames in all cases indicating the absence of stable detonation. However, strong flame acceleration was observed for all situations, which is consistent with the prediction by expansion ratio and Zeldovich number. Besides, the tube diameter and length are also critical for flame acceleration to supersonic. All the flames in this work accelerate drastically reaching the C-J deflagration state. But at 0.4 atm, only fast flame is formed, while at higher initial pressures, the flame further accelerates to a galloping mode manifesting a near-limit detonation, which could be indicated by the stability parameter χ.

Classical modes of one-dimensional (1D) detonation characterized by a simplified reaction model are reproduced by using a real chemical kinetics for the H-2-O-2 system with argon dilution. As Ar dilution is varied, the bifurcation points of pulsating instability are identified and a formed bifurcation diagram is compared with that obtained by the one-step reaction model. Eventually, the numerical results demonstrate that, for real detonations with detailed chemistry, the criterion of Ng et al. works well on prediction of the 1D detonation instability. Furthermore, the detonability limits are found respectively at low and high Ar dilutions. Above the high Ar dilution limit, detonations decays to the minimum level where long autoignition time and small heat release rate make reestablishment impossible for both 1D and 2D simulations. However, below the low Ar dilution limit, a 1D detonation cannot be sustained due to high instability, while the corresponding cellular detonation can propagate sustainably due to the role of transverse instability.

Abstract only

The aim of the study is to predict the thermal and stress behavior of a framed glass subjected to typical fire conditions, and the initial glass fracture time and locations using a probabilistic approach as an alternative to Pagni's deterministic criterion. Thermal stresses in glass have been little researched. The probabilistic approach has the advantage of taking into account some uncertainties such as the edge conditions. The model employed is based on stress and conduction heat transfer models, a spectral discrete ordinates radiation model, and a failure probability model. Some results of its verification and applications are reported here.

Radiative heat transfer plays a major role in the analysis of glazing behavior in fires, but its rigorous modeling has received little attention. In the present study, a spectral radiative heat transfer model, based on the discrete ordinates method (DOM), is developed and employed to analyze heat transfer and the transient temperature distribution in a glazing structure subjected to fire heat flux. Comparisons are made between model predictions and literature experimental data; acceptable agreements are found. The study also investigates the influence of the glass properties and geometry on the temperature and time to breakage.

This paper demonstrates experimental and numerical study on spontaneous ignition of H-2-N-2 mixtures during high-pressure release into air through the tubes of various diameters and lengths. The mixtures included 5% and 10% (vol.) N-2 addition to hydrogen being at initial pressure in range of 4.3-15.9 MPa. As a point of reference pure hydrogen release experiments were performed with use of the same experimental stand, experimental procedure and extension tubes. The results showed that N-2 addition may increase the initial pressure necessary to self-ignite the mixture as much as 2.12 or 2.85 - times for 5% and 10% N-2 addition, respectively. Additionally, simulations were performed with use of Cantera code (0-D) based on the ideal shock tube assumption and with the modified KIVA3V code (2-D) to establish the main factors responsible for ignition and sustained combustion during the release. (C) 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Abstract only

Spontaneous ignition of a pressurized hydrogen release has important implications in the risk assessment of hydrogen installations and design of safety measures. In real accident scenarios, an obstacle may be present close to the release point. Relatively little is known about the effect of such an obstacle on the salient features of highly under-expanded hydrogen jets and its spontaneous ignition. In the present study, the effect of a thin flat obstacle on the spontaneous ignition of a direct pressurized hydrogen release is investigated using a 5th-order WENO scheme and detailed chemistry. The numerical study has revealed that, for the conditions studied, the presence of the obstacle plays an important role in quenching the flame following spontaneous ignition for the release conditions considered. (c) 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

A fully coupled fluid–solid approach has been developed within FireFOAM 2.2.x, a large eddy simulation (LES) based fire simulation solver within the OpenFOAM® toolbox. Due consideration has been given to couple the radiative heat transfer and soot treatment with pyrolysis calculations. Combustion is modeled using the newly extended eddy dissipation concept (EDC) for the LES published by the authors’ group. Soot formation and oxidation are handled by the published extension of the laminar smoke point concept to turbulent fires using the partially stirred reactor (PaSR) concept also from the authors’ group. The gases radiation properties are evaluated using the established weighted sum of grey gas model while soot absorption coefficient is calculated using a single Planck-mean absorption coefficient. The effect of in-depth radiation is treated with the relatively simple Beer's law and the solid surface regression length is calculated from the local pyrolysis rate. Systematic validation studies have been conducted with several published experiments including simple pyrolysis test without the gaseous region, small scale wall fires and large scale flame spread. The predictions are in very good agreement with the relevant experimental data, demonstrating that the present modeling approach can be used to predict upward flame spread over PMMA with reasonable accuracy. Further parametric studies have also been conducted to demonstrate the effectiveness of the present modifications to capture the underlying physics. The detailed field predictions for vortex structures and flame volume including laminar–turbulent transition have also been analysed to uncover further insight of the unsteady flame spread phenomena. Potentially, the model can be used to aid further fundamental studies of the flame spread phenomena such as investigating the effects of width, inclination angles and side walls on flame spread as well as the predictions of flame spread in practical applications.

Large-eddy simulations (LES) of two plane impinging jets have been conducted. Predictions were first conducted for a natural impinging jet and found to be in good agreement with the experimental data of Yoshida et al. The validated code was then used to study the vortical structures of a forced impinging jet which had been experimentally investigated by Sakakibara et al. The numerical results show that the predictions have clearly captured the spanwise rollers, successive ribs, cross ribs and wall ribs observed by Sakakibara et al. They also show the predicted average convection velocity to be in good agreement with the measured value. Overall, the present study demonstrates the potential of LES simulations as a reliable tool to optimize the performance of engineering systems involving the use of forced impinging jets by regulating cross ribs through the inlet perturbations. (C) 2013 Elsevier Masson SAS. All rights reserved.

Flame quenching by fine mesh is one of the oldest known methods for mitigating flame propagation. Sir Humphry Davy pioneered the study of flame-wall quenching while developing the mining safety lamp. Installing a quenching mesh around the equipment is an effective preventive technique against hazardous flame propagation. Quenching of flame at the wall is due to the coupled thermo physical process involving heat transfer, flame stretch and preferential diffusion. While laminar flame-wall quenching has been extensively studied both theoretically and numerically, the reported studies of turbulent flame-wall interactions are limited to some Direct Numerical Simulations (DNS), which have subsequently led to improvement of models for predicting flame characteristics in the vicinity of the wall. Large Eddy Simulation (LES) of turbulent flows is considered as a powerful tool to predict the occurrence of instabilities due to heat release, hydrodynamic flow fields and acoustic waves. It provides a better description of turbulent-combustion interaction than the classical Reynolds Averaged Navier Stokes techniques (RANS). In the present study, the single mesh quenching of turbulent flame deflagration is investigated in stoichiometric methane-air mixture, which is equi-diffusive with unit Lewis number. It is well known that flame stretch and preferential diffusion has negligible influence in flame quenching for equi-diffusive flames. Therefore a unity Lewis number flamelet formulation can be used for simulating the turbulent combustion process within the premixed flamelet regime. In the present study, LES predictions are performed using the OpenFOAM CFD toolbox solver. The Coherent Flame Model (CFM) in the LES context as proposed by Richard et al. (2007) is implemented for modeling flame deflagration. During the flame/wall interactions, enthalpy loss through the wall affects the flamelet speed, flamelet annihilation and flame propagation; and the decrease in turbulence scales near the wall affects turbulent diffusion and flame strain. These flame-wall interactions are accounted for through extending the closures proposed by Bruneaux, Poinsot, and Ferziger (1997) for the CFM-RANS model to the LES context. Preliminary testing has demonstrated good potential of the modified CFM to capture the quenching effect of the wall. (C) 2012 Elsevier Ltd. All rights reserved.

A large eddy simulation of the turbulent syngas non-premixed jet flame of Sandia ETH/Zurich B is conducted using an in-house version of FireFOAM, a fire simulation solver within OpenFOAM. Combustion is modeled using (i) the newly extended eddy dissipation concept for the large eddy simulation published by the authors’ group, (ii) the 74-step CO–H2–O2 mechanism and (iii) the relatively simple tabulated chemistry approach. The effects of the nonunity Lewis number and thermal diffusion (=Soret diffusion) are considered in the mass fraction and enthalpy equations. Systematic validation and model sensitivity studies have been conducted against published experiments of the turbulent syngas diffusion flame from the international workshop on measurement & computation of turbulent flames (TNF workshop). The predictions were in very good agreement with the relevant experimental data. The axial position of peak H2O moved toward the nozzle direction owing to the different diffusion coefficients of H2. In the radial direction, the effect of thermal diffusion was observed at x/d < 20, whereas those of the nonunity Lewis number and difference in chemistry were found at x/d < 40. After considering the effects of the nonunity Lewis number and thermal diffusion, as well as the detailed reaction mechanisms, the results were slightly better than those obtained under previous numerical conditions.

Large eddy simulation has been applied to the prediction of a small pool fire experimentally tested by Venkatesh et al. The Smagorinsky's eddy viscosity model is used for subgrid-scale (SGS) turbulence closure and a modified laminar flamelet model (MLFM) based on the Cook and Riley approach is used for SGS combustion modeling. The predictions have captured the unique characteristics of small pool fires such as flame anchoring and double flame, as shown in previous theoretical analysis and experiments. The existence of the premixed zone of fuel and air, which is thought to be the reason for flame anchoring, is evidenced by the low gradient of temperature and mixture fraction near the burner rim. The experimentally observed double flame can be seen from both the predicted temperature contour and velocity vectors. For the mean temperature field where experimental data are available for quantitative comparison, the predictions with the MLFM are found to be in very good agreement with the data. In line with general expectations, the study also reveals that the predicted small pool fire is nearly axially symmetric. Comparison of the predictions with the two different SGS combustion models have highlighted the importance of SGS combustion modeling in capturing the fine details of such small pool fires. Considerable discrepancies have been found in the predictions of the velocity and temperature fields, and the predictions of the mixture fraction model indicate a slightly higher and more centered premixed flame near the burner rim.

Hydrogen process equipment are often housed in 20-foot or 40-foot container either be at refueling stations or at the portable standalone power generation units. Shipping Container provide an easy to install, cost effective, all weather protective containment. Hydrogen has unique physical properties, it can quickly form an ignitable cloud for any accidental release or leakages in air, due to its wide flammability limits. Identifying the hazards associated with these kind of container applications are very crucial for design and safe operation of the container hydrogen installations. Recently both numerical studies and experiment have been performed to ascertain the level of hazards and its possible mitigation methods for hydrogen applications. This paper presents the numerical modelling and the simulations performed using the HyFOAM CFD solver for vented deflagrations processes. HyFOAM solver is developed in-house using the open source CFD toolkit OpenFOAM libraries. The turbulent flame deflagrations are modelled using the flame wrinkling combustion model. This combustion model is further improved to account for flame instabilities dominant role in vented lean hydrogen-air mixtures deflagrations. The 20-foot ISO containers of dimensions 20' x 8' x 8'.6 '' filled with homogeneous mixture of hydrogen-air at different concentration, with and without model obstacles are considered for numerical simulations. The numerical predictions are first validated against the recent experiments carried out by Gexcon as part of the HySEA project supported by the Fuel Cells and Hydrogen 2 Joint Undertaking (FCH 2 JU) under the Horizon 2020 Framework Programme for Research and Innovation. The effects of congestion within the containers on the generated overpressures are investigated. The preliminary CFD predictions indicated that the container walls deflections are having considerable effect on the trends of generated overpressures, especially the peak negative pressure generated within the container is overestimated. Hence to account for the container wall deflections, the fluid structure interactions (FSI) are also included in the numerical modelling. The final numerical predictions are presented with and without the FSI. The FSI modelling considerably improved the numerical prediction and resulted in better match of overpressure trends with the experimental results. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Abstract only

Fire and deflagration are extreme manifestation of thermal runaway (TR) of Li-ion cells, and they are characterized for fully charged LiNiCoAlO2 (LNCA) 18650 cells in this investigation. The cells are over-heated using a cone calorimeter under different incident heat fluxes. When the cells are exposed to the incident heat flux larger than 35 kW m−2, both fire and deflagration present. The pressure valve opens when the temperature of the cell is higher than 132 °C. The fire occurs with the valve opening when the concentration of the venting vapour in the air is higher than the lower flammability limit. The deflagration happens after the cell temperature arrives about 200 °C, and is mainly arising from the cathode decomposition, the combustion of solvents and the anode relevant thermal reactions. The extreme temperatures of the cell and the flame during deflagration are over than 820 and 1035 °C, respectively. The production of COx, mass loss, heat release rate (HRR) are quantitative identified, and are found increase as the increasing incident heat flux. Based on revised oxygen consumption method, the HRR and liberated heat during the fire and deflagration for the cells are up to 11.8 ± 0.05 kW and 163.1 ± 1.5 kJ, respectively. •Critical incident heat flux that activates deflagration or/and fire is found.•Duration, evolution and key parameters of fire and deflagration are characterized.•Underlying reactions for fire and deflagration occurrence are determined.•Hazards like temperatures of cell and flame and mass of ejected gas are quantified.•Revised oxygen consumption method is developed and the heat release is specified.

Modelling and simulating the rapid pressure drop inside nozzles is a significant challenge because of the complexity of the multiple associated phenomena. In the present study, FlashFOAM a compressible solver for calculating the phase change within various nozzle geometries undergoing rapid pressure drops has been developed in the frame of the open source Computational Fluid Dynamics (CFD) code OpenFOAM. FlashFOAM accounts for the inter-phase heat transfer with the Homogeneous Relaxation Model (HRM). The work describes the development of a pressure equation within a different formulation than in other studies. The surface forces due to liquid-gas interfacial instabilities are modelled here in a novel coupling of HRM with the volume of fluid method giving rise to a conservative method for modelling primary atomisation. This new pressure equation is validated with published experimental measurements. A validation series dedicated to long nozzles is included for the first time. Novel additional tests for the flow characteristics and vapour generation in cryogenic liquid cases are included showing that the solver can be employed to gain some new insights into the physics of the flow regimes of sudden depressurising cryogenic liquids. The dependency of the geometry of the nozzles, pressure and subcooled degree on the vapour generation has been analysed including the effect of turbulence on the nozzle flow avoiding the laminar flow scenarios of previous validation studies. The validation study has demonstrated that FlashFOAM can be used to simulate flash boiling scenarios accurately and predict the properties of flash atomisation. (C) 2018 Elsevier Ltd. All rights reserved.

The behaviour of FRP strengthened flexural elements is a subject of this experimental study on the effect of elevated temperatures after heating and cooling. Six groups of CFRP strengthened simply supported small scale beams were loaded prior to the heating with a point load of I RN and heated uniformly in loaded condition to a specified temperature of 50 degrees C, 100 degrees C, 150 degrees C, 200 degrees C, 250 degrees C or 300 degrees C. Once a uniform temperature was established the samples were left to cool to room temperature and tested in four point bending. Results from the deflection of the small beams, development of strain in the laminate and failure modes are presented and discussed. (c) 2014 Elsevier Ltd. All rights reserved.

The eddy dissipation concept (EDC) is extended to the large eddy simulation (LES) framework following the same logic of the turbulent energy cascade as originally proposed by Magnussen but taking into account the distinctive roles of the sub-grid scale turbulence. A series of structure levels are assumed to exist under the filter width “Δ” in the turbulent energy cascade which spans from the Kolmogorov to the integral scale. The total kinetic energy and its dissipation rate are expressed using the sub-grid scale (SGS) quantities. Assuming infinitely fast chemistry, the filtered reaction rate in the EDC is controlled by the turbulent mixing rate between the fine structures at Kolmogorov scales and the surrounding fluids. The newly extended EDC was implemented in the open source FireFOAM solver, and large eddy simulation of a 30.5cm diameter methanol pool fire was performed using this solver. Reasonable agreement is achieved by comparing the predicted heat release rate, radiative fraction, velocity and its fluctuation, temperature and its fluctuation, turbulent heat flux, SGS and total dissipation rate, SGS and total kinetic energy, time scales, and length scales with the corresponding experimental data.

An experimental study of hydrogen/air premixed flame propagation in a closed rectangular channel with the inhibitions (N-2 or CO2) was conducted to investigate the inhibiting effect of N-2 and CO2 on the flame properties during its propagation. Both Schlieren system and the pressure sensor were used to capture the evolution of flame shape and pressure changes in the channel. It was found that both N-2 and CO2 have considerable inhibiting effect on hydrogen/air premixed flames. Compared with N-2, CO2 has more prominent inhibition, which has been interpreted from thermal and kinetic standpoints. In all the flames, the classic tulip shape was observed. With different inhibitor concentration, the flame demonstrated three types of deformation after the classic tulip inversion. A simple theoretical analysis has also been conducted to indicate that the pressure wave generated upon the first flame-wall contact can affect the flame deformation depending on its meeting moment with the flame front. Most importantly, the meeting moment is always behind the start of tulip inversion, which suggests the non-dominant role of pressure wave on this featured phenomenon. (C) 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

•The heat generation and gas production of four main thermal-chemical reactions are detected.•The fire-impingement takes an unordinary thermal runaway propagation for battery module.•There is a “smoldering period” before the explosion of lithium ion battery module.•Semenov and Frank-Kamenetskii models are used to analysis and predict the onset of runaway. Insight of the thermal characteristics and potential flame spread over lithium-ion battery (LIB) modules is important for designing battery thermal management system and fire protection measures. Such thermal characteristics and potential flame spread are also dependent on the different anode and cathode materials as well as the electrolyte. In the present study, thermal behavior and flame propagation over seven 50Ah Li(Ni1/3Mn1/3Co1/3)O2/Li4Ti5O12 large format LIBs arranged in rhombus and parallel layouts were investigated by directly heating one of the battery units. Such batteries have already been used commercially for energy storage while relatively little is known about its safety features in connection with potential runaway caused fire and explosion hazards. It was found in the present heating tests that fire-impingement resulted in elevated temperatures in the immediate vicinity of the LIBs that were in the range of between 200°C and 900°C. Such temperature aggravated thermal runaway (TR) propagation, resulting in rapid temperature rise within the battery module and even explosions after 20min of “smoldering period”. The thermal runaway and subsequent fire and explosion observed in the heating test was attributed to the violent reduction of the cathode material which coexisted with the electrolyte when the temperature exceeded 260°C. Separate laboratory tests, which measured the heat and gases generation from samples of the anode and cathode materials using C80 calorimeter, provided insight of the physical-chemistry processes inside the battery when the temperature reaches between 30°C and 300°C. The self-accelerating decomposition temperature of the cell, regarded as the critical temperature to trigger TR propagation, was calculated as 126.1 and 139.2°C using the classical Semenov and Frank-Kamenetskii models and the measurements of the calorimeter with the samples. These are consistent with the measured values in the heating tests in which TR propagated. The events leading to the explosions in the test for the rhombus layout was further analyzed and two possible explanations were postulated and analyzed based on either internal catalytic reactions or Boiling Liquid Expansion Vapor Explosion (BLEVE).

Abstract only

The standard 20 ft ISO containers are studied both experimentally and numerically with model obstacles to ascertain the peak overpressures generated in case of an accidental fast deflagrations. Apart from overpressure its often important to know the maximum deflections of the enclosures to examine the structural integrity. The container walls are not rigid; they not only contribute through acoustic response to the pressure waves but also through structural resonance response to the generated overpressures, to capture these effects its necessary to do the Fluid Structure Interaction (FSI) or in simple terms the coupled Computational Fluid Dynamics (CFD) and Finite element analysis (FE) simulations. Alternatively, the structural response can be coupled to CFD simulation in either one-way or two-way interactions to reduce the associated computation efforts. Numerical simulations have been conducted with a pseudo two-way coupled CFD & FSI, to aid our understanding of the combination of various factor contributing for the generation of overpressures using an in-house solver developed based on open source CFD code OpenFOAM named as HyFOAM. The CFD solver solves the compressible Navier-stokes equations along with the spring-mass-damper system's single degree of freedom motion equation for coupled fluid structure interactions. The turbulent flame deflagrations are modelled using the flame area combustion model, the necessary modification are incorporated to governing equations to include the dominant flame instabilities present during the vented deflagration process. For lean hydrogen mixtures the Lewis number effects are important hence suitable modelling improvements are added to the combustion model. The numerical results are validated against the recent experiments conducted at Gexcon, Norway as part of the HySEA project supported by the Fuel Cells and Hydrogen 2 Joint Undertaking (FCHJU) under the Horizon 2020 Framework Programme for Research and Innovation. Numerical predictions of CFD and FSI coupled, are assessed against the experimental results to study the contributing factors affecting the generated overpressure with in the ISO containers in view to overall improve the numerical predictions.

Flash boiling is the rapid phase change of a pressurised fluid that emerges to ambient conditions below its vapour pressure. Flashing of a flowing liquid through an orifice or a nozzle can occur either inside or outside the nozzle depending on the local pressure and geometry. Vapour generation during flashing leads to interfacial interactions that eventually influence the jet. Empirical models in the literature for simulating the inter-phase heat transfer employ many simplifying assumptions, which limits their applicability. Typical models, usually derived from cavitation, fail to describe the physics of heat and mass transfer, making them unreliable for flashing. The Homogeneous Relaxation Model (HRM) is a reliable model able to capture heat transfer under these conditions accounting for the non-equilibrium vapour generation. This approach uses a relaxation term in the transport equation for the vapour. On the basis of the generic compressible flow solver within the open source computational fluid dynamics (CFD) code OpenFOAM, the HRM has been implemented to create a dedicated new solver HRMSonicELSAFoam. An algorithm that links the standard pressure–velocity coupling algorithm to the HRM is used. In this method, a pressure equation is derived which employs the continuity equation including compressibility effects. A relaxation term has been defined such that the instantaneous quality would relax to the equilibrium value over a given timescale. Although it is possible to consider this timescale constant, it is calculated via an empirical correlation in the present study. Validations have been carried out by simulating two-phase flows through sharp-edged orifices. The relatively good agreement achieved has demonstrated that the solver accurately calculates the pressure and vapour mass fraction. This demonstrates the potential of HRMSonicELSAFoam for flash boiling simulations and predicting the properties of the subsequent flash atomisation.

Laminar hydrogen flame propagation in a channel with a perforated plate is investigated using 20 reactive Navies-Stokes simulations. The effect of the perforated plate on flame propagation is treated with a porous media model. A one step chemistry model is used for the combustion of the stoichiometric H-2-air mixture. Numerical simulations show that the perforated plate has considerable effect on the flame propagation in the region downstream from the perforated plate and marginal effect on the upstream region. It is found to squeeze the flame front and result in a ring of unburned gas pocket around the flame neck. The resulting abrupt change in flow directions leads to the formation of some vortices. Downstream of the perforated plate, a wrinkled "M"-shape flame is observed with "W' shape flame speed evolution, which lastly turns back to a convex curved flame front. Parametric studies have also been carried out on the inertial resistance factor, porosity, perforated plate length and its location to investigate their effects on flame evolution. Overall, for parameter range studied, the perforated plate has an effect of reducing the flame speed downstream of it. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

Gas explosions in homogeneous reactive mixtures have been widely studied both experimentally and numerically. However, in practice and industrial applications, combustible mixtures are usually inhomogeneous and subject to vertical concentration gradients. Limited studies have been conducted in such context which resulted in limited understanding of the explosion characteristics in such situations. The present numerical investigation aims to study the dynamics of Deflagration to Detonation Transition (DDT) in inhomogeneous hydrogen/air mixtures and examine the effects of obstacle blockage ratio in DDT. VCEFoam, a reactive density-based solver recently assembled by the authors within the frame of OpenFOAM CFD toolbox has been used. VCEFoam uses the Harten-Lax-van Leer-Contact (HLLC) scheme fr the convective fluxes contribution and shock capturing. The solver has been verified by comparing its predictions with the analytical solutions of two classical test cases. For validation, the experimental data of Boeck et al. (1) is used. The experiments were conducted in a rectangular channel the three different blockage ratios and hydrogen concentrations. Comparison is presented between the predicted and measured flame tip velocities. The shaded contours of the predicted temperature and numerical Schlieren (magnitude of density gradient) will be analysed to examine the effects of the blockage ratio on flame acceleration and DDT.

The paper is focused on comparative computational modelling of the attenuation of fire radiation by water mists of pure water or sea water. The use of sea water in fire protection could be a more convenient and practical choice in coastal areas, on offshore installations or transported ships. The spectral absorption and scattering properties of both water droplets and salt particles formed by evaporation of sea water droplets are considered. A combined heat transfer problem is based on a combination of the spectral radiative transfer in a mist curtain, the kinetics of water evaporation, and convective heat transfer along the curtain. The developed computational model is used to analyze the radiative heating and evaporation of droplets of pure water and more complex multi-phase processes in droplets of sea water at all stages of the process. The numerical results for the case problem indicate sufficiently good shielding quality of a sea-water mist curtain. The suggested approach is expected to be useful for important engineering applications in fire protection. •A model for attenuation of fire radiation by water mists is modified.•An approach for behaviour of pure and sea water droplets is suggested.•Heat transfer problem for heated and evaporated water mist is solved.•The shielding properties of a sea-water mist curtain are analyzed.•Computational results for the case problem are presented.

Water spray screening is a mitigation technique to reduce thermal radiation from flames to acceptable levels required by safety regulations. The current literature on water screening is limited to hydrocarbon flames. The main goal of the present study is to bridge this knowledge gap by presenting a methodology to predict the water spray screening performance of hydrogen flames using their actual emission spectra. A wide range of hydrogen flames and water screen scenarios are investigated. The results show that although hydrogen flames' emissivities are relatively lower than hydrocarbons', the radiated heat could pose safety risks and water screening is an effective mitigating method. The study further analyses the total transmissivity of water screens calculated from (i) the actual hydrogen spectrum and (ii) a blackbody spectrum. For optically thick spray screens, the blackbody spectrum widely used for hydrocarbon flames screening, could yield unacceptable overestimations of the total transmissivity. Copyright (C) 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

Evaporating liquid cascades resulting from gasoline and liquefied natural gas tanks overfilling or rupture of elevated pipes create a source of flammable vapor cloud. Such phenomena were responsible for the formation of the large fuel vapor cloud, the ignition of which resulted in the large scale explosion, in Buncefield [Buncefield Major Incident Investigation Board, Explosion Mechanism Advisory Group Report, 2007] on December 11, 2005 at the Hertfordshire Oil Storage Terminal, an oil storage facility located by Hemel Hempstead in Hertfordshire, England. Despite its significance, there lacks adequate models treating the underlying physics of this phenomenon. The present study numerically analyses fuel cascades which are considered as a droplet-laden system. Consideration is given to vapor production inside the cascade due to droplets evaporation and breakup. The solver used here is a modification of the sprayFoam solver which is present in the open source computational fluid dynamics (CFD) toolbox OpenFOAM (R) [OpenFOAM 2.3.0, Available at http://www.openfoam.com]. The fuel droplets evaporate during their motion and create a cloud of flammable vapor. In order to capture the characteristics of the hazardous phenomena, the CFD model needs to address the underlying physics with adequate submodels. In the present study, the multiphase flow is simulated with a combined Eulerian-Lagrangian approach. The governing equations of the gas phase represent the conservation equations of mass, momentum, and energy including the source terms arising from the interaction with the droplets. The Reynolds Averaged Navier-Stokes simulation approach was used for its computational efficiency. The Large-Eddy Simulation would be more robust in handling the interaction of the droplets with the flow but it would require more computational resource. The particulate phase is simulated through a Lagrangian deterministic or stochastic tracking models to provide particle trajectories and particle concentration. Particular emphasis is given to the effect of impingement of droplets to account for the effect of splashing in the impact region. The study involves developing robust and accurate modeling approaches for the instabilities and aerodynamic breakup in the cascade which contribute to the formation of the cloud, air entrainment, and fuel impingement on deflector plates. Suitable submodels have been implemented in OpenFOAM (R) to facilitate the study. The predictions are compared with the experimental measurements and CFD predictions previously conducted by Atkinson and Coldrick [Research Report 908, 2012] from the Health and Safety Laboratory, an agency of the Health and Safety Executive (HSE). The present predictions are found to better capture the interaction between the droplets and the gas phase. Improved agreement with the experimental measurements in the gasoline fuel cascades has also been achieved. (C) 2015 American Institute of Chemical Engineers

The lateral ceiling temperature distribution induced by wall-attached fires with various burner aspect ratios in underground space have been experimentally investigated with a specially designed test rig of 1:8 scale. A series of tests were conducted with rectangular gas burners with different aspect ratios set close to the side wall, different fire heat release rates and fire source-ceiling height. The ceiling temperature profile and lateral ceiling flame extension length under the ceiling were recorded and analyzed in detail. The results show that the lateral ceiling temperature distribution induced by rectangular burners with different aspect ratios is quite different. The temperature for a given horizontal location in the lateral direction below the ceiling increases with the increase of the burner aspect ratio. A new correlation is proposed with modified characteristic diameter. Its predictions are in reasonably good agreement with the present measurements as well the previous measurements of the others which were not used in its derivation.

A solution for the complete problem of attenuation of fire radiation by water mist is presented. This solution is based on simplified approaches for the spectral radiative properties of water droplets, the radiative transfer in the absorbing and scattering mist, and transient heat transfer taking into account partial evaporation of water mist. A computational study of the conventional model problem indicates the role of the main parameters and enables one to formulate some recommendations to optimize possible engineering solutions. The method developed is also applied to more realistic case study of a real fire. It is suggested to decrease the size of supplied water droplets with the distance from the irradiated surface of the mist layer. The advantage of this engineering solution is confirmed by numerical calculations. Potential possibility of microwave monitoring of water mist parameters is analyzed on the basis of Mie theory calculations. (C) 2016 Elsevier Ltd. All rights reserved.

This paper summarises the results from a blind-prediction benchmark study for models used for estimating the consequences of vented hydrogen deflagrations, as well as for users of such models. The work was part of the HySEA project that received funding from the Fuel Cells and Hydrogen Joint Undertaking (FCH JU) under grant agreement no. 671461. The first blind-prediction benchmark exercise in the HySEA project focused on vented explosions with homogeneous hydrogen-air mixtures in 20-foot ISO containers. The scenarios selected for the second blind-prediction study focused on vented deflagrations in inhomogeneous hydrogen-air mixtures resulting from continuous stratification of hydrogen during vertical jet releases inside 20-foot ISO containers. The deflagrations were vented through commercial vent panels located on the roof of the containers. The test program included two configurations and four experiments, i.e. two repeated tests for each scenario. The paper compares experimental results and model predictions and discusses the implications of the findings for safety related to hydrogen applications. Several modellers predicted the stratification of hydrogen inside the container during the release phase with reasonable accuracy. However, there is significant spread in the model predictions, especially for the maximum reduced explosion pressure, and including predictions from different modellers using the same model system. The results from the blind-prediction benchmark studies performed as part of the HySEA project constitute a strong incentive for developers of consequence models to improve their models, implement automated procedures for scenario definition and grid generation, and update training and guidelines for users of the models. [Display omitted] •Blind-prediction benchmark study for consequence models.•Vented hydrogen deflagration experiments in 20-foor ISO containers.•Model predictions by computational fluid dynamics (CFD) tools.•Model predictions by empirical or semi-empirical engineering models.

The effect of turbulence models on CFD predictions of gas and liquid pool fires is investigated. A buoyancy-modified four-equation (4EQ) model and two modified versions of the standard two-equation (2EQ) k-epsilon model are employed for simulating three pool fire scenarios. The 4EQ model includes all the important source terms in the turbulent heat flux expressions and emphasizes the anisotropy of turbulence due to buoyancy effects by adding an algebraic term to the eddy-viscosity expression of Reynolds stresses. Predicted results show that the 4EQ model predicts far larger values of buoyancy production of turbulent kinetic energy than the 2EQ models, and consequently achieves better agreement with experimental temperature and velocity data.

•A novel Coupled PCM-liquid cooling system (CPLS) for Li-ion battery pack.•Design of CPLS is optimized for the thermal performance of battery pack.•Effectiveness of CPLS is verified by the designed experiments.•Controlling strategy for the velocity and inlet temperature of coolant.•Improved cooling performance and energy-saving at different ambient temperatures. In order to improve the working performance of the lithium-ion battery pack in continuous operation under different ambient temperatures, a coupled composite phase change material and liquid cooling thermal management system is proposed. The simulation for this system under a cycle that a 3C rate discharging and then a 0.5C charging was conducted, as well as comparison tests concerning factors such as cell-to-cell spacing, cell-to-tube distance, channel number and coolant velocity. Simulation results showed that the coupled system with suitable design exhibited good thermal performance at an ambient temperature of 30 °C, which kept the maximum surface temperature and the temperature difference of the battery pack at 41.1 °C and 4 °C at the end of 3C discharge. Then, the latent heat of phase change material was also recovered by the liquid cooling during the 0.5C charge. Specially designed experiments have also been conducted to verify the effectiveness and practicability of the proposed coupled system. Based on this system, a liquid cooling strategy was proposed for controlling the velocity and inlet temperature of coolant by monitoring the temperature of the phase change material and environment. This further improved the thermal performance of the battery pack during cycling at different ambient temperatures and significantly reduced the unnecessary power consumption of liquid cooling during this process.

The paper presents the latest experimental data in a continuing investigation into condensation on horizontal, integral-fin tubes. The effect of radiused fillets at the roots of rectangular-section fins has been investigated. Theory suggested that, for a high conductivity tube, fillets should enhance the heat transfer by preventing retention of condensate in the corners at the base of the fins on the upper "unflooded" part of the tube. Three condensing fluids were used, namely, steam, ethylene glycol and R-113. Vapour-side, heat-transfer coefficients were obtained from accurately measured overall heat-transfer coefficients using a modified Wilson plot. Radiused fillets at the fin root were found to increase the vapour-side, heat-transfer coefficient, the enhancing effect increasing with the surface tension of the condensate.

The authors presented a basic mathematical model for estimating peak overpressure attained in vented explosions of hydrogen in a previous study (Sinha et al. [1]). The model focussed on idealized cases of hydrogen, and was not applicable for realistic accidental scenarios like presence of obstacles, initial turbulent mixture, etc. In the present study, the underlying framework of the model is reformulated to overcome these limitations. The flame shape computations are simplified. A more accurate and simpler formulation for venting is also introduced. Further, by using simplifying assumptions and algebraic manipulations, the detailed model consisting of several equations is reduced to a single equation with only four parameters. Two of these parameters depend only on fuel properties and a standard table provided in the Appendix can be used. Therefore, to compute the overpressure, only the two parameters based on enclosure geometry need to be evaluated. This greatly simplifies the model and calculation effort. Also, since the focus of previous investigation was hydrogen, properties of hydrocarbon fuels, which are much more widely used, were not accounted for. The present model also accounts for thermophysical properties of hydrocarbons and provides table for fuel parameters to be used in the final equation for propane and methane. The model is also improved by addition of different sub-models to account for various realistic accidental scenarios. Moreover, no adjustable parameters are used; the same equation is used for all conditions and all gases. Predictions from this simplified model are compared with experimentally measured values of overpressure for hydrogen and hydrocarbons and found to be in good agreement. First the results from experiments focussing on idealized conditions of uniformly mixed fuel in an empty enclosure under quiescent conditions are considered. Further the model applicability is also tested for realistic conditions of accidental explosion consisting of obstacles inside the enclosure, non-uniform fuel distribution, initial turbulent mixture, etc. For all the cases tested, the new simple model is found to produce reasonably good predictions. (C) 2019 The Authors. Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC.

With the global move towards performance based fire design, fire safety assessment in and around buildings becomes increasingly important. However, key knowledge gaps still exist concerning the behavior of fire swirling, which may be generated if one or more accidental fires are in the passage of the vortices behind an adjacent tall building. The present study is focused on the experimental investigations of the burning behavior of two pool fires behind 1/50 scaled tall buildings with heights varying from 0.565 to 1.165 m in a cross-wind. The objective is to gain insight of the effect of the distance between the two fires ( D2), the distance between the fires and the building ( D1), wind speed ( V), and the height of the scaled building ( H) on the burning behavior. Important conclusions have been drawn about the influence of D1 and D2 on the fuel mass loss rate, the influence of D1 on fire swirling, the influence of D2 on the possible merging of the two fires and the effect of wind speed on the mass loss rate. The results suggested the existence of a critical velocity for the cross-wind on the initiation of fire swirling and an approximate value was identified for the conditions in the tests. The investigations also covered the effect of height of the scaled building on the fuel mass loss rate and the occurrence of fire swirling. This relationship was found to be also dependent on the wind speed. Analysis of the results has led to some important recommendations to enhance the fire protection of tall buildings.

Numerical investigations have been conducted on the effect of the internal geometry of a local contraction on the spontaneous ignition of pressurized hydrogen release through a length of tube using a 5th-order WENO scheme. A mixture-averaged multi-component approach was used for accurate calculation of molecular transport. The auto-ignition and combustion chemistry were accounted for using a 21-step kinetic scheme. It is found that a local contraction can significantly facilitate the occurrence of spontaneous ignition by producing elevated flammable mixture and enhancing turbulent mixing from shock formation, reflection and interaction. The first ignition kernel is observed upstream the contraction. It then quickly propagates along the contact interface and transits to a partially premixed flame due to the enhanced turbulent mixing. The partially premixed flames are highly distorted and overlapped with each other. Flame thickening is observed, which is due to the merge of thin flames. The numerical predictions suggested that sustained flames could develop for release pressure as low as 25 bar. For the release pressure of 18 bar, spontaneous ignition was predicted but the flame was soon quenched. To some extent this finding is consistent with Dryer et al.'s experimental observation in that the minimum release pressure required to induce a spontaneous ignition for the release through a tube with internal geometries is only 20.4 bar. Copyright (C) 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

In order to characterize fire merging, pool fires on hollow trays with varying side lengths were burned under quasi-quiescent condition and in a wind tunnel with the wind speed ranging from 0 m/s to 7.5 m/s. Burning rate and flame images were recorded in the whole combustion process. The results show that even though the pool surface area was kept identical for hollow trays of different sizes, the measured burning rates and fire evolutions were found to be significantly different. Besides the five stages identified by previous studies, an extra stage, fire merging, was observed. Fire merging appeared possibly at any of the first four stages and moreover resulted in 50-100% increases of the fire burning rates and heights in the present tests. The tests in wind tunnel suggested that, as the wind speed ranges from 0 m/s to 2 m/s, the burning rates decrease. However with further increase of the wind speed from 2 m/s to 7.5 m/s, the burning rate was found to increase for smaller hollow trays while it remains almost constant for larger hollow trays. Two empirical correlations are presented to predict critical burning rate of fire merging on the hollow tray. The predictions were found to be in reasonably good agreement with the measurements. (C) 2015 Elsevier B.V. All rights reserved.

A sub-grid scale model for partially premixed combustion has been adapted and applied to simulate the backdraft phenomena. A fast deflagration or backdraft is produced when into a hot, fuel-rich compartment an inflow of fresh air is Suddenly allowed through all opening. It is essentially a violent combustion process involving both premixed and non-premixed regimes. The present model is based oil the coupling of independent approaches for non-premixed and premixed turbulent combustion. The 'flame index' concept was used to separate the two different combustion regimes. This index describes the structure of the flame based on fuel and oxygen gradients. Due to the lack of detailed experimental measurements, the results were largely analysed qualitatively. The predictions have provided valuable insight into the backdraft phenomenon suggesting that the development of backdraft can be divided into five phases, i.e. initial condition, free "spherical propagation, "plane" front propagation, stretching of the flame front through the opening and fireball outside the container. Quantitatively, the experimentally measured and predicted lapse of time between the maximum over- and under-pressure at the opening of the container is found to be in reasonably good agreement. (c) 2007 Elsevier Ltd. All rights reserved.

The issue of spontaneous ignition of highly pressurized hydrogen release is of important safety concern, e.g. in the assessment of risk and design of safety measures. This paper reports on recent numerical investigation of this phenomenon through releases via a length Of tube. This mimics a potential accidental scenario involving release through instrument line. The implicit large eddy Simulation (ILES) approach was used with the 5th-order weighted essentially non-oscillatory (WENO) scheme. A mixture-averaged multi-component approach was used for accurate calculation of molecular transport. The thin flame was resolved with fine grid resolution and the autoignition and combustion chemistry were accounted for using a 21-step kinetic scheme. The numerical study revealed that the finite rupture process of the initial pressure boundary plays ail important role in the spontaneous ignition. The rupture process induces significant turbulent mixing at the contact region via shock reflections and interactions. The predicted leading shock velocity inside the tube increases during the early stages of the release and then stabilizes at a nearly constant value which is higher than that predicted by one-dimensional analysis. The air behind the leading shock is shock-heated and mixes with the released hydrogen in the contact region. Ignition is firstly initiated inside the tube and then a partially premixed flame is developed. Significant amount of shock-heated air and well developed partially premixed flames are two major factors providing potential energy to overcome the strong under-expansion and flow divergence following spouting from the tube. Parametric studies were also conducted to investigate the effect of rupture time, release pressure, tube length and diameter on the likelihood of spontaneous ignition. It was found that a slower rupture time and a lower release pressure will lead to increases in ignition delay time and hence reduces the likelihood of spontaneous ignition. If the tube length is smaller than a certain value, even though ignition Could take place inside the tube, the flame is unlikely to be sufficiently strong to overcome under-expansion and flow divergence after spouting from the tube and hence is likely to be quenched. (C) 2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

In a compartment fire, the breakage and possible fallout of a window glass has a significant impact on the fire dynamics. The thermal breakage of glass depends on various parameters such as glass type, edge shading, edges conditions and constraints on the glass. The purpose of the present study is to investigate some of the key parameters affecting the thermal breakage of window glass in fire conditions using a recently developed and validated computer tool. Fallout is not within the scope of this study. Different boundary conditions of the glass pane (unconstrained and constrained) subjected to fire radiant heat are investigated. The analysis shows that to prevent glass thermal breakage, it is important to provide enough spacing between the frame and glass pane to accommodate the thermal expansion, and constraints on the glass structure should be avoided. The zones where the glass is likely to crack first are shown. The study also quantifies the effects of glass edge conditions on its thermal breakage in fire conditions; such analysis has not been reported in the literature due to its complexity and the statistical nature of edge flaws. The results show that an ordinary float glass mostly used in windows, with the "as-cut" edge condition would break later and is stronger than a ground edge or polished edge glass for the scenarios investigated. The study demonstrates how a predictive tool could be employed for a better understanding of thermal breakage of window glass in fires and for design guidance. (C) 2012 Elsevier Ltd. All rights reserved.

•Flash-atomisation is investigated studying both the mechanical and thermodynamic effects that break-up the liquid jet.•A unified Eulerian methodology is proposed for studying flashing jets simulating both the internal flow and the spray dynamics.•The Eulerian-Lagrangian-Spray-Atomisation (ELSA) model is extended for simulating superheated jets.•The atomisation model is coupled with a pressure equation which takes into account the non-equilibrium state of the fluid and surface tension.•The model is used to predict various spray characteristics such as the Sauter mean diameter, the jet velocity and the spray angle. The physics of the atomisation of flash boiling jets is known to be extremely complex with interactions of different mechanisms at microscopic and macroscopic level. Early studies describe both the mechanical and thermodynamic effects focusing on the influence of the initial pressure and temperature on the spray characteristics. The resulting flashing jet usually emerges to the low-pressure region with a high velocity and fragments to large blobs and ligaments which break up to droplets due to both mechanical and thermodynamic effects. This present study describes a numerical approach for simulating the atomisation of flashing liquids suitable for both primary atomisation and secondary break-up using the Eulerian-Lagrangian-Spray-Atomisation model coupled with a pressure equation for the metastable jet. The proposed approach aims at describing the atomisation of superheated jets and the impact of bubble nucleation at different stages and regimes inside the channel the liquid emerges from. The changes in the regime outside the nozzle are discussed for various cases of flashing liquids providing insights for the interactions of the mechanisms that contribute to the liquid fragmentation and the spray characteristics such as the droplet size and velocity and the spray angle.

A statistical narrow band (SNB) radiation gas model has been incorporated into a Reynolds-averaged Navier-Stoke (RANS) based computational fluid dynamics (CFD) code. The coupled CFD code is used in this work to simulate two upward flame spread scenarios: the first one concerns flame spread over a vertical PMMA wall, while the other represents flame spread along vertical corner walls. For comparison purpose, the weighted-sum-of-greygases (WSGG) model was also used. Simulation results indicate that the CFD model is capable of predicting reasonably the flame spread phenomenon over simple materials such as PMMA. Compared to the WSGG approach, the SNB model generally yields more accurate results thanks to a better predicted gas flow field which in turn improves the prediction of the radiative heat fluxes imposed on solid surfaces and consequently that of the flame spread rate and heat release rate.

Safety studies for hydrogen retail stations involve identification of possible accidental scenarios, modelling of consequences and measures to mitigate associated hazards with it. Accidental release of hydrogen during its handling and storage can lead to formation of ignitable mixture in a very short time. Ignition of such a mixture can lead to generation of overpressure affecting structure and people. Understanding of the possible overpressures generated is critical in designing the system safe from explosion hazards. In the present study, the worst-case scenario where high-pressure hydrogen storage cylinders are enveloped by a premixed hydrogen-air cloud is numerically simulated. The computational domain mimics the setup for premixed hydrogen cloud in a mock hydrogen cylinder storage congestion environment experimentally studied by Shirvill et al. [1]. Large Eddy Simulations (LES) are performed using OpenFOAM CFD toolbox solver. The Flame Surface Wrinkling Model in LES context is used for modelling deflagrations [2]. Numerical simulation results are compared against experiments. Simulations are able to predict experimental flame arrival and overpressure reasonably well. The effects of ignition location, congestion and confining walls on the turbulent deflagrations in particular on explosion overpressure are discussed. It was concluded that explosion overpressure increases with increase in confinement. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

•9 mm smallest cell size ensures grid independence, resolves ∼70% of RMS temperature.•Comparable accuracy and CPU times between nongrey WSGG and box model.•Newer oxyfuel WSGG models are slower and not more accurate for fire applications.•TRI modeling required to avoid relative radiant power loss of ∼40% at pool surface.•Pool surface radiation very sensitive to subgrid temperature fluctuation modeling. Non-grey radiation modelling of gas-phase combustion products is performed during runtime of large eddy simulations (LES) of 30 cm, 20 kW methanol pool fires, based on the experiments of Klassen and Gore [4] and Weckman and Strong [38]. FireFOAM, the turbulent flame solver part of open source CFD platform OpenFOAM®, was modified to include a new array of gas radiation models. Two grey and three non-grey implementations of the weighted-sum-of-grey-gases (WSGG) are compared in terms of both accuracy and CPU efficiency, along with a 'box' model based on the exponential wide band model (EWB) but specially optimised for fire scenarios. Turbulence-radiation interactions (TRI) are taken into account for the self-correlation of temperature in the emission term of the radiative transfer equation (RTE). Non-grey WSGG models consistently performed better than their grey counterparts, but the two newer WSGG correlations based on up-to-date spectral databases did not perform noticeably better or worse than the older WSGG model, which is a departure from other studies in oxy-fuel conditions. The work also showed that TRI is very important for the accurate prediction of the pool surface radiant feedback and the total radiant output. Some recommendations are made for the fire and radiation community.

Abstract only

A three-dimensional (3-D) code which can predict the complete performance of high efficiency boilers is described in this paper. The model is in two parts. The combustion chamber is modelled in the first part and the heat exchanger in the second. The thermal efficiency is calculated from the predicted results. The computational model was based on the two-dimensional (2-D) TEAM code which has been extended into 3-D and extensively modified by the present authors. In the code, the gaseous combustion was modelled by Simple Chemical Reaction System concept; the Radiation process was handled by the ‘Discrete Transfer’ method. The modelling results of the combustion chamber were used as input for the calculation of the consequent finned tube heat exchanger. The heat exchanger calculation included conduction, convection and radiation. Predictions were performed for a low Reynolds number boiler and compared with existing data. Good agreement was obtained.

A possible consequence of pressurized hydrogen release is an under-expanded jet fire. Knowledge of the flame length, radiative heat flux as well as the effects of variations in ground reflectance is important for safety assessment. The present study applies an open source CFD code FireFOAM to study the radiation characteristics of hydrogen and hydrogen/methane jet fires. For combustion, the eddy dissipation concept for multicomponent fuels recently developed by the authors in the large eddy simulation (LES) framework is used. The radiative heat is computed with the finite volume discrete ordinates model in conjunction with the weighted sum of grey gas model for the absorption/emission coefficient. The pseudo-diameter approach is used in which the corresponding parameters are calculated using the formulations of Birch et al. [24] with the thermodynamic properties corrected by the Able-Noble equation of state. The predicted flame length and radiant fraction are in good agreement with the measurements of Schefer et al. [2], Studer et al. [3] and Ekoto et al. [6]. In order to account for the effects of variation in ground surface reflectance, the emissivity of hydrogen flames was modified following Ekoto et al. [6]. Four cases with different ground reflectance are computed. The predictions show that the ground surface reflectance only has minor effect on the surface emissive power of the smaller hydrogen jet fire of Ekoto et al. [6]. The radiant fractions fluctuate from 0.168 to 0.176 close to the suggested value of 0.16 by Ekoto et al. [6] based on the analysis of their measurements. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

In this paper, we present some experimental and analytical model results of two-component fuel mixtures of methane, propane and hydrogen. Experimentally obtained turbulent burning velocity S T for outwardly propagating spherical lean turbulent premixed flames is examined with an algebraic flame surface wrinkling reaction model using 1) mean local burning velocity, and 2) the critical chemical time scale from the leading edge model by Zel'dovich and Frank-Kamenetskii. Based on the latter approach, the time scale that characterizes the effects of preferential diffusion phenomenon in critically curved spherical flames is incorporated into the reaction model. For this, a proposed simple linear model is used for estimating the effective Lewis number of the two-component fuel (CH 4–H 2 and C 3H 8–H 2)/Air mixtures. In general, both approaches are effective ways in achieving qualitatively consistent S T trends for both mixtures. However, in the second approach, model predictions show large S T deviation especially at high turbulence. This may be attributed to the use of approximate values of activation temperature and for the use of the effective Lewis number of both mixtures based on the simple linear model.

Safety issues raised by thermal runaway (TR) are the main obstacle hindering the booming of lithium-ion batteries. A comprehensive model can potentially help improve understanding of the TR mechanisms and assist the battery pack design. However, previous models generally neglected the particle ejection, which is integral to predicting TR. In this study, a multi-scale model for the multiphase process of battery venting has been proposed, covering the entire chain of chemical reactions and physical transformation during TR. A lumped model in battery scale unveiled the interplay of thermal abuse progression and pressure accumulation. The computational fluid dynamics coupled with the discrete phase model was adopted to simulate both generated gases and ejected particles. The newly developed model was checked quantitatively by experimental measurements for battery temperature, jet velocity and mass evolution under thermal abuse. Simulation results highlight two violent ejections of particles and gases with inverted conical contours, consistent with visualization by laser technique in the experiment. The electrolyte vapours are found to dominate the gas release before TR, while the generated reaction gases become the major release after the burst of chain reactions. The developed model fulfils the TR prediction including particle ejection, which can provide new references for the thermal safety design of battery packs.