Mining, process and energy industries suffer from billions of dollars of worldwide losses every year due to Vapour Cloud Explosions (VCE). Moreover, explosion accidents are often tragic and lead to a high number of severe injuries and fatalities. The VCE scenario is complex and controlled by various mechanisms. The interplay among them is still not entirely understood. Understanding all these intricate processes is of vital importance and requires detailed experimental diagnostics. Coupling accurate numerical simulations to well documented experiments can allow an elaborate description of these phenomena. This thesis focuses on explosions occurring on configurations that are either semi-confined or confined. In such configurations, the explosion is generally initiated by a mild ignition and a subsonic flame front emerges from the ignition source. An important feature of self-propagating flames lies in their intrinsically unstable nature. When they propagate in an environment with high levels of confinement and congestion, which is the case in most industrial sites, a Flame Acceleration (FA) process is often observed that can give rise to very fast flames, known for their destructive potential. In some cases, the FA process can create the appropriate conditions for the initiation of detonations, which corresponds to a rapid escalation of the explosion hazard. To reproduce the confinement and congestion conditions that one can find in industrial sites, the university of Munich TUM equipped a confined chamber with a series of obstacles and analysed the influence of repeated obstructions on the propagation of hydrogen/air deflagrations. This experimental study showed a strong influence of the mixture composition on the acceleration process. A Deflagration to Detonation Transition (DDT) has also been observed for a certain range of equivalence ratio. This configuration is therefore ideal to study the mechanisms of flame acceleration as well as the intricate DDT process. A numerical study of both scenarios is performed in this thesis: -First for a lean premixed hydrogen/air mixture, a strong flame acceleration is observed experimentally without DDT. The characteristic features of the explosion are well reproduced numerically using a Large Eddy Simulation (LES) approach. The crucial importance of confinement and repeated flame-obstacle interactions in producing very fast deflagrations is highlighted. -DDT is observed experimentally for a stoichiometric hydrogen/air mixture. This thesis focuses on the instants surrounding the DDT event, using Direct Numerical Simulations (DNS). Particular attention is drawn to the impact of the chemistry modelling on the detonation scenario. The failure of preventive measures is often observed in many explosion accidents. To avoid a rapid escalation of the explosion scenario, mitigative procedures must be triggered when a gas leak or an ignition is detected. Metal salts (like potassium bicarbonate and sodium bicarbonate) have received considerable attention recently because well-controlled experiments showed their high efficiency in inhibiting fires. The last part of the thesis focused on the mechanism of flame inhibition by sodium bicarbonate particles. First, criteria based on the particle sizes are established to characterize the inhibition efficiency of the particles. Second, two dimensional numerical simulations of a planar flame propagating in a stratified layer of very fine sodium bicarbonate particles showed that under certain conditions these powders can act as combustion enhancers. These results echo a number of experimental observations on the possible counter-effects of the inhibitors.

This thesis is concerned with experimental and computational studies on pulse detonation engine (PDE) that has been envisioned as a new concept engine. These engines use the high pressure generated by detonation wave for propulsion. The cycle efficiency of PDE is either higher in comparison to conventional jet engines or at least has similar high performance with much greater simplicity in terms of components. The first part of the work consists of an experimental study of the performance of PDE under choked flame and partial fill conditions. Detonations used in classical PDEs create conditions of Mach numbers of 4-6 and choked flames create conditions in which flame achieves Mach numbers near-half of detonation wave. While classical concepts on PDE's utilize deflagration-to-detonation transition and are more intensively studied, the working of PDE under choked regime has received inadequate attention in the literature and much remains to be explored. Most of the earlier studies claim transition to detonation as success in the working of the PDE and non-transition as failure. After exploring both these regimes, the current work brings out that impulse obtained from the wave traveling near the choked flame velocity conditions is comparable to detonation regime. This is consistent with the understanding from the literature that CJ detonation may not be the optimum condition for maximum specific impulse. The present study examines the details of working of PDE close to the choked regime for different experimental conditions, in comparison with other aspects of PDEs. The study also examines transmission of fast flames from small diameter pipe into larger ducts. This approach in the smaller pipe for flame acceleration also leading to decrease in the time and length of transition process. The second part of the study aims at elucidating the features of deflagration-to-detonation transition with direct numerical simulation (DNS) accounting for and the choice of full chemistry and DNS is based on two features: (a) the induction time estimation at the conditions of varying high pressure and temperature behind the shock can only be obtained through the use of full chemistry, and (b) the complex effects of fine scale of turbulence that have sometimes been argued to influence the acceleration phase in the DDT cannot be captured otherwise. Turbulence in the early stages causes flame wrinkling and helps flame acceleration process. The study of flame propagation showed that the wrinkling of flame has major effect on the final transition phase as flame accelerates through the channel. Further, flame becomes corrugated prior to transition. This feature was investigated using non-uniform initial conditions. Under these conditions the pressure waves emanating from corrugated flame interact with the shock moving ahead and transition occurs in between the flame and the forward propagating shock wave. The primary contributions of this thesis are: (a) Elucidating the phenomenology of choked flames, demonstrating that under partial fill conditions, the specific impulse can be superior to detonations and hence, allowing for the possibility of choked flames as a more appropriate choice for propulsive purposes instead of full detonations, (b) The use of smaller tube to enhance the flame acceleration and transition to detonation. The comparison with earlier experiments clearly shows the enhancements achieved using this method, and (c) The importance of the interaction between pressure waves emanating from the flame front with the shock wave which leads to formation of hot spots finally transitioning to detonation wave.

<p>In this thesis, gas explosions inside pipes are considered. Laboratory experiments and numerical simulations are the basis of the thesis. The target of the work was to develop numerical models that could predict accidental gas explosions inside pipes.</p><p>Experiments were performed in circular steel pipes, with an inner diameter of 22.3 mm, and a plexiglass pipe, with an inner diameter of 40 mm. Propane, acetylene and hydrogen at various equivalence ratios in air were used. Pressure was recorded by Kistler pressure transducers and flame propagation was captured by photodiodes, a SLR camera and a high-speed camera. The experiments showed that acoustic oscillations would occur in the pipes, and that the frequencies of these oscillations are determined by the pipe length. Several inversions of the flame front can occur during the flame propagation in a pipe. These inversions are appearing due to quenching of the flame front at the pipe wall and due to interactions of the flame front with the longitudinal pressure waves in the pipe. Transition to detonation was achieved in acetylene-air mixtures in a 5 m steel pipe with 4 small obstructions.</p><p>Simulations of the flame propagation in smooth pipes were performed with an 1D MATLAB version of the Random Choice Method (RCMLAB). Methods for estimation of quasi 1D burning velocities and of pipe outlet conditions from experimental pressure data were implemented into this code. The simulated pressure waves and flame propagation were compared to the experimental results and there are good agreements between the results.</p><p>Simulations were also performed with the commercial CFD code FLACS. They indicated that to properly handle the longitudinal pressure oscillations in pipes, at least 7 grid cells in each direction of the pipe cross-section and a Courant number of maximum 1 should be used. It was shown that the current combustion model in FLACS gave too high flame speeds initially for gas explosions in a pipe with an inner width of 40 mm.</p>

In this thesis, gas explosions inside pipes are considered. Laboratory experiments and numerical simulations are the basis of the thesis. The target of the work was to develop numerical models that could predict accidental gas explosions inside pipes. Experiments were performed in circular steel pipes, with an inner diameter of 22.3 mm, and a plexiglass pipe, with an inner diameter of 40 mm. Propane, acetylene and hydrogen at various equivalence ratios in air were used. Pressure was recorded by Kistler pressure transducers and flame propagation was captured by photodiodes, a SLR camera and a high-speed camera. The experiments showed that acoustic oscillations would occur in the pipes, and that the frequencies of these oscillations are determined by the pipe length. Several inversions of the flame front can occur during the flame propagation in a pipe. These inversions are appearing due to quenching of the flame front at the pipe wall and due to interactions of the flame front with the longitudinal pressure waves in the pipe. Transition to detonation was achieved in acetylene-air mixtures in a 5 m steel pipe with 4 small obstructions. Simulations of the flame propagation in smooth pipes were performed with an 1D MATLAB version of the Random Choice Method (RCMLAB). Methods for estimation of quasi 1D burning velocities and of pipe outlet conditions from experimental pressure data were implemented into this code. The simulated pressure waves and flame propagation were compared to the experimental results and there are good agreements between the results. Simulations were also performed with the commercial CFD code FLACS. They indicated that to properly handle the longitudinal pressure oscillations in pipes, at least 7 grid cells in each direction of the pipe cross-section and a Courant number of maximum 1 should be used. It was shown that the current combustion model in FLACS gave too high flame speeds initially for gas explosions in a pipe with an inner width of 40 mm.

This thesis by articles addresses the role played by Chapman-Jouguet (CJ) deflagrations in deflagration to detonation transition (DDT) events. By definition, CJ deflagrations are flames propagating with a sonic flow in the burned gases, and are theoretically the fastest subsonic combustion waves able to propagate steadily, predicted using conservation of mass, momentum and energy. DDT is difficult to describe, as many complex phenomena and their interaction take place, including flame instabilities, turbulent combustion, and combustion in compressible medium, among others. Recent experiments and numerical simulations however showed that, prior to transition to detonations, deflagrations plateau at the CJ regime before rapid acceleration. In the present thesis, multiple aspects of the last stages of DDT are studied, and are each presented in published articles or articles in preparation. The two articles presented in Chapter 2 focus on experiments performed on the transition of a shock-flame complex to a detonation downstream of a single obstacle, in a stoichiometric propane-oxygen mixture at low pressure, mimicking the common configuration found at the last stages of DDT in experiments and numerical simulations performed in a channel filled with obstacles. The relative large size of the obstacle and the low gas initial pressure permitted to visualize the details of the initiation of the detonation around the obstacle. Transition to detonation was found to occur in a similar fashion for variously shaped obstacles, after flame acceleration due to the interaction with reflected shocks. This acceleration process was found to occur rapidly in the case where the incident flame propagated with a burning rate close to the Chapman-Jouguet value. The third article presented in Chapter 3 describes a model aimed to predict the properties of shocks followed by a CJ deflagration, in experimental configurations where the burned gases can be vented. The formulation is similar to the double discontinuity problem adapted from the work of Chue (1993), extended to cases where the burned gases are not confined by a rear wall anymore, but can be vented through an opening of known dimensions. The properties of the shock / CJ-deflagration complex could then be predicted and compared to flame measurements done prior the initiation of detonations, obtained on a selection of large scale DDT experiments. The good agreement suggests that DDT occurs when deflagrations reach the CJ regime, corroborating with observations done in shock tubes. The article presented in Chapter 4 is aimed to present a consistent method for calculating the structure of flames propagating at arbitrary burning velocities, from the low-Mach case (isobaric) up to the CJ deflagration regime. The method uses a dynamical system approach to calculate the steady wave structure, described by ordinary differential equations. A stability analysis near the burned and unburned gases permitted to develop a numerical shooting technique, which was used to obtain the flame structure and burning rate eigenvalue. Chapter 5 is a numerical study of the deflagration to detonation transition problem in one-dimension. By linearly increasing the burning rate eigenvalue to increase the flame burning velocity, the flame first reached the CJ condition. Subsequent increase in the burning rate leads to the self-organization of the flame into a CJ deflagration - shock complex. This self-organization triggers a pulsating gasdynamic instability leading to the transition of the flame to detonation.

L’opto-pyrotechnie (amorçage de la détonation par système optique) est l’une des innovations les plus prometteuses en termes de fiabilité, de sécurité et de performances pour les futurs lanceurs spatiaux. Le but de la thèse est d’étudier et de modéliser le premier des deux étages d’un Détonateur Opto-Pyrotechnique, constitué d’un explosif confiné dans une chambre de combustion fermée où se déroulent les premières phases d’une Transition Déflagration-Détonation. L’amorçage par laser de l’explosif puis la combustion en chambre isochore sont traités par le code EFAE, lequel est couplé au logiciel LS-DYNA qui simule la déformation et la rupture du disque de fermeture de la chambre, puis la propulsion du projectile résultant vers le second étage. En parallèle, diverses techniques expérimentales (adsorption de gaz, vélocimétrie hétérodyne, microscopie) ont mis en valeur plusieurs procédés physiques, ce qui a permis de tester le couplage entre EFAE et LS-DYNA, puis de déterminer et de hiérarchiser les paramètres affectant les critères industriels. / Opto-pyrotechnics (ignition of detonation by optical systems) is one of the most promising innovations to improve reliability, safety and performances on future space launchers. This thesis aims at studying and modeling the first stage from a two-stage opto-pyrotechnic detonator that consists of a condensed explosive confined in a closed combustion chamber, in which the beginning of a Deflagration-to-Detonation Transition occurs. The laser ignition of the explosive and its isochoric combustion are modeled by the EFAE code. This code is coupled with LS-DYNA software to deal with the deformation and the rupture of the metallic disk that closes the combustion chamber, and then with the subsequent propulsion of the projectile to the second stage. In parallel, various experimental technics (gas adsorption, photonic Doppler velocimetry, microscopy) have underlined several physical processes that allow first to test the coupling between EFAE and LS-DYNA, then to determine and classify influent parameters that affect the industrial specifications.

Les incidents d'explosions intervenant sur les sites industriels sont souvent accompagnés de dégâts matériels et humains importants. Les dégâts varient d’une explosion à une autre, suggérant l’existence de mécanismes capables d’aggraver le scénario d’explosion. Réduire les risques d'explosion nécessite une compréhension fine des différents mécanismes mis en jeu. Avec l’augmentation considérable de la puissance de calcul, la simulation numérique est devenu une approche incontournable pour l’étude et la compréhension de ces scénarios. Cette thèse se focalise sur les explosions de gaz initiées par un noyau de flamme subsonique. Lorsque la flamme se propage dans un environnement offrant un haut niveau de confinement et d’obstruction, ce qui est souvent le cas des sites industriels, une forte accélération de la flamme est généralement observée, accompagnée d’une augmentation de la pression. Dans certains cas, l’accélération de la flamme peut conduire à l’initiation d’une onde de détonation. Ce scénario coïncide avec une augmentation brutale de la surpression et donc une aggravation des dégâts observés. Pour reproduire des conditions de confinement et d’obstruction représentatives des sites industriels, l’université de Munich TUM a équipé une chambre confinée de 5.4m de long d’une série d’obstacles et analysé l’impact de ces obstructions sur la propagation de déflagrations hydrogène/air. Cette étude expérimentale a montré une forte influence de la richesse du mélange sur l’accélération de la flamme. Une transition à la détonation est notamment observée pour une certaine gamme de richesse. Cette configuration est donc idéale pour étudier les mécanismes d’accélération de flamme ainsi que les conditions qui peuvent mener à l’initiation de détonations. Une étude numérique des deux scénarios a été menée mêlant simulations directes (DNS) et simulations aux grandes échelles (LES):-Pour un mélange d’hydrogène/air pauvre, une forte accélération de la flamme est observée expérimentalement sans transition à la détonation. Les grandeurs caractéristiques de l’explosion ont été reproduites avec des simulations aux grandes échelles (LES). Plusieurs mécanismes d’accélération de flamme ont été identifiés et attribués au haut niveau de confinement et de congestion dans la chambre. Le couplage de ces mécanismes explique les grandes vitesses de propagation observées. -Pour un mélange stoechiométrique, une transition à la détonation est observée. Cette thèse s’est focalisée sur les instants précédant l’initiation de la détonation afin de caractériser les conditions nécessaires pouvant mener à cet événement soudain, en se basant sur une approche de simulation directe (DNS). Une attention particulière a été portée à l’influence du schéma cinétique sur ce scénario. Comme constaté dans bon nombre d’incidents industriels, les mesures préventives peuvent échouer. Le cas échéant, des procédures visant à contrôler l’impact des explosions doivent être utilisées pour éviter une catastrophe de grande ampleur. L’utilisation d’inhibiteurs chimiques est une technique qui a déjà fait ses preuves contre les feus. Elle consiste à injecter des poudres capables de réagir chimiquement avec la flamme et de réduire son taux de dégagement de chaleur. L’étude de l’interaction de ces particules solides avec la flamme correspond au deuxième volet de cette thèse. Un modèle simplifié de décomposition de ces particules solides (HetMIS) a été développé dans un contexte LES. Deux aspects ont été explorés : 1) l’interaction unidimensionnel flamme/particule a permis d’établir un critère, basé sur la taille des particules, caractérisant l’efficacité des poudres dans le processus d’inhibition; 2) l’effet de la distribution spatial des particules sur la propagation de la flamme est analysé dans le but d’apporter une explication à certains résultats expérimentaux révélant un effet opposé des inhibiteurs dans certaines conditions. / Mining, process and energy industries suffer from billions of dollars of worldwide losses every year due to Vapour Cloud Explosions (VCE). Moreover, explosion accidents are often tragic and lead to a high number of severe injuries and fatalities. The VCE scenario is complex and controlled by various mechanisms. The interplay among them is still not entirely understood. Understanding all these intricate processes is of vital importance and requires detailed experimental diagnostics. Coupling accurate numerical simulations to well documented experiments can allow an elaborate description of these phenomena. This thesis focuses on explosions occurring on configurations that are either semi-confined or confined. In such configurations, the explosion is generally initiated by a mild ignition and a subsonic flame front emerges from the ignition source. An important feature of self-propagating flames lies in their intrinsically unstable nature. When they propagate in an environment with high levels of confinement and congestion, which is the case in most industrial sites, a Flame Acceleration (FA) process is often observed that can give rise to very fast flames, known for their destructive potential. In some cases, the FA process can create the appropriate conditions for the initiation of detonations, which corresponds to a rapid escalation of the explosion hazard. To reproduce the confinement and congestion conditions that one can find in industrial sites, the university of Munich TUM equipped a confined chamber with a series of obstacles and analysed the influence of repeated obstructions on the propagation of hydrogen/air deflagrations. This experimental study showed a strong influence of the mixture composition on the acceleration process. A Deflagration to Detonation Transition (DDT) has also been observed for a certain range of equivalence ratio. This configuration is therefore ideal to study the mechanisms of flame acceleration as well as the intricate DDT process. A numerical study of both scenarios is performed in this thesis: -First for a lean premixed hydrogen/air mixture, a strong flame acceleration is observed experimentally without DDT. The characteristic features of the explosion are well reproduced numerically using a Large Eddy Simulation (LES) approach. The crucial importance of confinement and repeated flame-obstacle interactions in producing very fast deflagrations is highlighted. -DDT is observed experimentally for a stoichiometric hydrogen/air mixture. This thesis focuses on the instants surrounding the DDT event, using Direct Numerical Simulations (DNS). Particular attention is drawn to the impact of the chemistry modelling on the detonation scenario. The failure of preventive measures is often observed in many explosion accidents. To avoid a rapid escalation of the explosion scenario, mitigative procedures must be triggered when a gas leak or an ignition is detected. Metal salts (like potassium bicarbonate and sodium bicarbonate) have received considerable attention recently because well-controlled experiments showed their high efficiency in inhibiting fires. The last part of the thesis focused on the mechanism of flame inhibition by sodium bicarbonate particles. First, criteria based on the particle sizes are established to characterize the inhibition efficiency of the particles. Second, two dimensional numerical simulations of a planar flame propagating in a stratified layer of very fine sodium bicarbonate particles showed that under certain conditions these powders can act as combustion enhancers. These results echo a number of experimental observations on the possible counter-effects of the inhibitors.