At present there is no simple, yet scientifically robust method for calculating insurance loss estimates due to a fire. Therefore building owners and insurers can not make suitably informed decisions when selecting fire protection measures or setting premiums as they have no way of defining the true risk they face. As a consequence this research aims to investigate a number of techniques in an effort to define one as appropriate for further research. Three different methods were explored and consist of risk based analysis, deterministic hand calculations and Computational Fluid Dynamics (CFD). Extensive literature reviews were conducted in each area and the final models were based on the outcomes of this research. Rack storage warehouses were chosen for analysis as they are currently topical within the fire engineering community and are a particular concern for insurers. The risk based analysis employed statistical techniques including event tree analysis and monte carlo simulation to calculate loss distributions and sensitivity analyses. The hand calculation method was based on equations presented in the literature and incorporated the use of a zone model (BRANZFire) to calculate deterministic loss estimates. The CFD model used was Fire Dynamics Simulator and full scale warehouse fires were modelled using this programme. It was concluded that Fire Dynamics Simulator is an inappropriate tool as the capability for providing loss estimates in a timely manner is currently beyond the model's capabilities. Of the two remaining methods the statistical risk based model was selected as the most appropriate for further investigation. The primary reasons for this decision were the ability to calculate loss distributions and conduct sensitivity analyses, as well as its versatility and user friendliness. Improved statistical data was defined as imperative for future development of the model.

This report describes an investigation into the prediction of the activation times of domestic ionisation and photoelectric smoke detectors within a two storey dwelling, the work undertaken being an extension to that previously presented by Brammer (2002). Three fire scenarios are considered, each having been a real test fire undertaken at the Building Research Establishment in Cardington. These three fire scenarios all involved the flaming combustion of an upholstered armchair within the lounge on the Ground floor. During the experiments various results were recorded, including temperatures, optical densities and smoke detector activation times. The fire scenarios where modelled using FDS, Version 5. Base parameters regarding the fuel load where defined to be 0.05kgsoot/kgfuel and 20MJ/kg. Consideration was also given to the effect varying the effective heat of combustion and defined soot yield would have on derived smoke detector activation times. Additional simulations where thus run considering soot yields of 0.04kgsoot/kgfuel and 0.10ksoot/kgfuel, and an effective heat of combustion of 25MJ/kg. Three prediction methods where applied to the results of the FDS simulations for derivation of the activation times of smoke detectors located throughout the house. These methods where the temperature correlation method, Heskestad’s method, and Cleary’s method. The temperature correlation method considered activation criterions of 4°C, 13°C and 20°C above ambient. The Heskestad and Cleary methods were found to derive comparable activation times for each detector location. None of the prediction algorithms where however found to predict activation times consistently comparable to the test data. Rather, it was determined that for an appropriate prediction method to be adopted for accurate assessment of a given fire scenario, consideration must be given to the: • type of detector being assessed; • location of the detector relative to the fire; • mode of combustion (i.e. flaming or smouldering); and the • growth rate of the fire.

Presently in ground vehicle industries, conducting durability tests with a high acceleration factor have become increasingly demanding for the less time and cost involvement. In the previous work, to accelerate the field test, the standard ‘test tailoring’ approach has been modified due to the requirement of high acceleration factors and the limitations of testing implementation. In this study, a computer-aided testing method is developed for the validation of this modified approach. Hence, a new test-piece has been designed by a conjugative approach involving the finite element technique and fatigue analysis. Afterwards, the accelerated durability loading profiles synthesized via the modified approach have been applied on the designed test-piece and the fatigue life has been simulated to verify the effectiveness of those loading profiles. Simulation results show that loading profiles with an acceleration factor up to 330 can be successfully generated with an accuracy of 95% by this modified approach.

Presently in ground vehicle industries, conducting durability tests with a high acceleration factor have become increasingly demanding for the less time and cost involvement. In the previous work, to accelerate the field test, the standard ‘test tailoring’ approach has been modified due to the requirement of high acceleration factors and the limitations of testing implementation. In this study, a computer-aided testing method is developed for the validation of this modified approach. Hence, a new test-piece has been designed by a conjugative approach involving the finite element technique and fatigue analysis. Afterwards, the accelerated durability loading profiles synthesized via the modified approach have been applied on the designed test-piece and the fatigue life has been simulated to verify the effectiveness of those loading profiles. Simulation results show that loading profiles with an acceleration factor up to 330 can be successfully generated with an accuracy of 95% by this modified approach.

Magíster en Ciencias de la Ingeniería, Mención Mecánica / El presente trabajo tiene por objeto caracterizar la energía liberada por una llama de difusión dentro de un túnel modelo. El trabajo se enmarca dentro del proyecto FONDECYT 1085015, que apunta a determinar la efectividad de cortinas de aire doble jet - doble flujo (DJ-DF) operando como un sistema de protección contra incendios en túneles. El estudio se llevó a cabo en una instalación experimental que representa un túnel vial a escala 1:34, al interior del cual se genera una celda de confinamiento mediante dos cortinas DJ-DF. Con el fin de representar la potencia liberada y la dinámica de un incendio se procedió a diseñar, construir y montar un quemador estándar de llama de difusión, equipado con un sistema de refrigeración, un circuito de alimentación de combustible y un mecanismo de encendido remoto. El comportamiento de la llama se caracterizó de la siguiente manera. En primer lugar, se efectó una simulación numérica con el programa Fire Dynamics Simulator (FDS) para predecir la radiación incidente a lo largo de un tramo del túnel. Luego se midió experimentalmente la radiación en dichos puntos bajo las mismas condiciones del modelo. Al contrastar estos resultados se valida el modelo de radiación y combustión empleado por FDS, por cuanto se logra predecir efectivamente los datos experimentales. Sin embargo, una desviación en la estimación numérica de la producción de hollín se traduce en una leve discordancia de la predicción computacional. El confinamiento del calor generado por la llama se abordó como sigue. Se obtuvo experimentalmente la altura de la llama para diversas tasas de liberación de calor en el marco de dos escenarios, uno en que la llama se desarrolla sin perturbaciones, y otro en que la llama se encuentra dentro de la celda de confinamiento. En el primer caso los resultados experimentales se ajustan a las correlaciones publicadas al respecto, dando cuenta que la metodología empleada es adecuada y certera, aunque pierde validez para bajas liberaciones de calor y cuando la llama toca el techo propagándose por éste. Luego, al comparar ambos escenarios se observa una disminución en la altura de la llama para altas liberaciones de energía, y la situación contraria para bajas liberaciones de energía. Así se concluye que la configuración DJ-DF cumple su objetivo para llamas cuya altura comparte el mismo orden de magnitud que la altura del túnel. Para llamas más débiles, las cortinas tienden a favorecer su desarrollo, aunque esta conclusión está sujeta a reducir la incertidumbre en la medición de la altura de llama.

In a society increasingly more influenced by technology and electricity, electrical and computer cables will play a more vital role in humans’ everyday life. With an increasing number of cables being introduced into society, the risk of fires caused by or involving cables will increase and become a more common danger to property and human lives. The fire properties of cables are tested according to Standard EN 50399 where vertically mounted cables are exposed to a burner for 20 minutes. The present work consists of running simulations imitating the conditions of Standard EN 50399 for testing cables using a Computational Fluid Dynamics program called Fire Dynamic Simulator (FDS). The general idea was to test the material in small-scale and running simulations to verify how well simulated values corresponded to values from actual testing, providing a potential less costly method of predicting the correct Euroclass in the development phase of new cables. During a visit at RISE in Borås, material for testing and a script previously used for testing a module of EN 50399 in FDS were obtained from previous work. The FDS script was later altered by adjusting the meshes inside the model in an effort to reduce simulation time. This was done by prioritizing smaller grid cells in high activity areas and using large grid cells in low activity areas. To verify the function of the model on the current version 7.5.0 of the FDS software, simulations were run empty without modelling the cables. To validate the FDS-model, temperatures were measured at four heights using a resemble of plate thermometers and the results were compared to older temperature measurements from an actual experiment using plate thermometers in the apparatus used at RISE when testing in the EN 50399 apparatus. To obtain the material data necessary for FDS, the material used as cable sheeting (surrounding the conductive metal core) molded into thin square plates were tested using a cone calorimeter at Luleå University of Technology. Two tests were conducted at irradiance levels of 50 and 25 kW/m2 where heat release rate was measured. Thenceforth followed 14 repeated tests at varying irradiance levels with the sole purpose of measuring time to ignition. In total 16 experiments were conducted, of which ten resulted in ignition, four of which did not ignite after exposure for 20 minutes and two which were interrupted due to swelling of the sample. After testing in the cone calorimeter, a critical irradiance level and ignition temperature of the material were verified using a theory presented by Janssens (1991). Two ramps – a controlled way of determining the materials heat release over time in FDS – were created based on the two tests at different irradiance levels. Using these new parameters simulations recreating the scenario for testing according to EN 50399 were run using FDS. Three simulations were run, testing different ramps and different implementations of the cables. The results proved it difficult to achieve the same heat release rate for cables simulated using FDS as heat released rate measured at experiments. With the simulations results at hand, in combination with uncertainties regarding material data it became clear the material had proven more difficult then anticipated. A possible reason for the big gap in heat release rate between simulations and experimental values could be considered to be the high ignition temperature given as material input for the cable in FDS.

The fact that a building is covered by a well-arranged fire protection is today a basic precondition for the safety of a building. Good and organized fire protection is also an essential part of a good safety culture. For this reason, it is common for buildings to be equipped with a higher fire protection than what is required by current regulations. In buildings intended for hotel services, sprinklers are usually installed even though it is not required. Fans in operation is a protection method for the ventilation that falls within the framework of analytical dimensioning, where an analysis is required to verify the formation. As the name implies, fans in the operating solution are based on the fans continuing to run in the event of a fire, the protection mechanism is that the pressure in the supply air duct must be overcome by the fire pressure before the fire gases can spread. As a calculation basis for the verification, the fire pressure (when sprinklers are not used) amounts to 1500 Pa and the fire gases temperature in the early stage of the fire to 350 ͦ C according to Boverkets “allmänna råd” (BFS 2013: 2) on analytical dimensioning of buildings fire protection (BBRAD3). When using sprinklers, these values ​​are expected to be much lower, which should also change the conditions for fire protection design. Together with the fans in operation method, self-actuating backflow protectors is often used, which completely restricts the flow in the supply air duct when the fire pressure becomes too high and the flow reverses, but these are relatively costly. The thesis mainly aims to investigate whether deviations from the fire technical installation of backflow protection can be made in buildings where sprinklers have been installed. This in the hope of finding more cost-effective solutions and at the same time achieve satisfactory fire protection. To get an answer to if the deviation can be made, the computer programs Fire Dynamics Simulator (FDS), Program Flow System (PFS) and material from an existing project have been used. As a complement, a small literature study has also been done to gain a basic understanding of the underlying theory and to substantiate proposed solutions. Various CFD (Computational Fluid Dynamics) programs are often used in many engineering professions. The basis for the programs is that they are based on flow calculations based on the Navier-Stokes equations. FDS is a CFD program adapted for simulations of fire scenarios. PFS is a ventilation program where flows can be calculated, both during normal circumstances and in the happening of fire. In FDS, a hotel room was built based on the existing project and requirements from BBRAD3. Based on data given in the FDS runs, the fire gas spread via the ventilation system was calculated using PFS. The result shows that a possible design of the ventilation technical fire protection without backflow damper is not possible to achieve a satisfactory fire protection. This is because the spread of fire gases to the adjacent fire compartment exceeded the requirement of 1% of its volume. The results show, however, that there are several different acceptable ventilation technical solutions in the form of an increased capacity of the supply air fan in the event of a fire or by installing a flow adjustment damper on the supply air duct. When applying the flow adjustment damper as a fire technical solution, the result has shown that with a 200 Pa pressure relief on the damper, the fire gas spread is less than the 1% requirement. When the pressure in the supply air ducts becomes higher, the possibility of fire gases spreading becomes more difficult. As the estimated fire gas spread was below the requirements, this solution is also considered to be applicable in buildings that are slightly more airtight than the building studied. A disadvantage, however, is that more powerful fans for the supply air will have to be used, which can increase operating costs and energy use. The other proposed fire technical solution was to increase the capacity of the supply air fan from normal operating function to 350 Pa and 380 Pa pressure effect in the event of a fire. This function is intended to be connected to the building's fire alarm. In the event of a fire, the fire alarm should be activated at an initial stage so that the supply air fan can reach the selected capacity within 44 seconds. Since at 44 seconds the normal pressure drop of 50 Pa in the supply air duct is overcome, which means that fire gas spread from the fire compartment to adjacent fire compartments is initiated. The application of an existing or external supply air fan whose capacity only increases if it is activated by a fire alarm is the report's main solution for maintaining satisfactory fire protection when substituting a backflow damper. This is because the solution is considered to be the most cost-effective. / Att en byggnad omfattas av ett väl anordnat brandskydd är idag en grundläggande förutsättning för en byggnads säkerhet. Ett bra och organiserat brandskydd är även en väsentlig del av en god säkerhetskultur. Av denna anledning är det vanligt att byggnader utrustas med ett högre brandtekniskt brandskydd än vad som är kravställt enligt gällande regelverk. I byggnader som avses för hotellverksamhet finns oftast sprinkler installerat, trots att det inte är kravställt. Fläktar i drift är en ventilationstekniks skyddsmetod som går inom ramen för s.k. analytisk dimensionering, där en analys krävs för att verifiera utformningen. Som man hör på namnet bygger fläktar i driftlösningen på att fläktarna fortsätter att gå vid händelse av brand, skyddsmekanismen är att trycket i tilluftskanalen måste övervinnas av brandtrycket innan spridning kan ske. Som beräkningsgrund vid verifieringen uppgår brandtrycket vanligtvis (när sprinkler ej nyttjas) till 1500 Pa och brandgastemperaturen i branden tidiga stadie till 350 ͦC enligt Boverkets allmänna råd (BFS 2013:2) om analytisk dimensionering av byggnaders brandskydd (BBRAD3). Vid användandet av sprinkler förväntas dessa värden bli mycket lägre, vilket även bör ändra förutsättningarna till brandskyddet utformning. Tillsammans med fläktar i drift nyttjas ofta backströmningsskydd som helt stryper flödet i tilluftskanalen (där risken för brandgasspridning är) när brandtrycket blir för högt och flödet vänder, dessa är dock relativt kostnadsdrivande. Examensarbetet syftar främst på att undersöka om avsteg från den brandtekniska installationen av backströmningskydd kan göras i byggnader där sprinkler har installerats. Detta i hopp om att hitta mer kostnadseffektiva lösningar och samtidigt uppnå ett tillfredställande brandskydd. För att få svar på avsteget kan göras har datorprogrammen Fire Dynamics Simulator (FDS), Program Flow System (PFS) samt material från ett befintligt projekt nyttjats. Som komplement har även en mindre litteraturstudie gjorts för att få grundförståelse för den bakomliggande teorin samt att styrka föreslagna lösningar. Olika CFD-program (Computional Fluid Dynamics) nyttjas ofta inom många ingenjörsyrken. Grunden för programmen är att de bygger på flödesberäkningar som utgår ifrån Navier-Stokes ekvationer. FDS är ett CFD-program anpassat för simuleringar av brandscenarion. PFS är ett ventilationsprogram där flöden kan beräknas, både vid normal- och branddrift. I FDS byggdes ett hotellrum upp utifrån det befintliga projektet samt krav från BBRAD3. Utifrån data givet i FDS-körningarna beräknades brandgasspridningen via ventilationssystemet med hjälp av PFS. Resultatet påvisar att en eventuell utformning av det ventilationstekniska brandskyddet utan backströmningspjäll ej är möjligt för att uppnå ett tillfredställande brandskydd. Detta med anledning av att brandspridningen till intilliggande brandcell översteg kravet på 1% av dess volym. Resultatet visar dock att det finns flera olika godtagbara ventilationstekniska lösningar i form av en ökad kapacitet på tilluftsfläkten vid händelse av brand eller genom en installation av ett injusteringspjäll på tilluftskanalen. Vid tillämpning av injusteringspjäll som brandteknisk lösning har resultatet påvisat att vid en 200 Pa tryckavlastning på spjället understiger brandgasspridningen 1%-kravet. När trycket i tilluftskanalerna blir högre, försvåras möjligheten för brandgaser att sprida sig. Då den beräknade brandgasspridningen understeg kravställningen, anses denna lösning även kunna tillämpas i byggnader som är något tätare än den aktuella byggnaden. En nackdel är dock att kraftigare fläktar för tilluften kommer behöva nyttjas vilket kan öka driftkostnaderna och energianvändningen. Den andra föreslagna brandtekniska lösningen var att öka kapacitet på tilluftsfläkten från normal driftsfunktion till 350 Pa respektive 380 Pa tryckverkan i det fall då brand uppstått, denna funktion är tänkt att vara kopplad till byggnadens brandlarm. Vid händelse av brand bör brandlarmet aktiveras i ett inledande skede så att tilluftsfläkten kan nå utvald kapacitet inom 44 sekunder. Detta med anledning av att vid denna tidpunkt övervinns det normala tryckfallet på 50 Pa i tilluftskanalerna, vilket innebär att brandgasspridning från brandrummet mot intilliggande brandceller inleds. Tillämpningen av en befintlig eller extern tilluftsfläkt vars kapacitet endast ökar i det fall då den aktiveras av ett brandlarm bedöms utgöra rapportens främsta brandtekniska lösning för att upprätthålla ett tillfredställande brandskydd vid substitution av ett backströmningspjäll. Detta då lösningen anses vara den mest kostnadseffektiva samtidigt som det innebär en smidigare tillämpning.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2018-04-04T12:31:51Z No. of bitstreams: 1 Rafael Gialdi Salvagni_.pdf: 4078552 bytes, checksum: 3046bf3bf4c04b0ca0ff91981f73727e (MD5) / Made available in DSpace on 2018-04-04T12:31:51Z (GMT). No. of bitstreams: 1 Rafael Gialdi Salvagni_.pdf: 4078552 bytes, checksum: 3046bf3bf4c04b0ca0ff91981f73727e (MD5) Previous issue date: 2017-04-25 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / PROSUP - Programa de Suporte à Pós-Gradução de Instituições de Ensino Particulares / Este trabalho apresenta o estudo numérico-experimental da dispersão de poluentes e da geometria da chama de uma poça de combustível. Foi realizada A modelagem da combustão em uma poça, com a dispersão da pluma em função do vento incidente, com o objetivo de caracterizar o fenômeno. Foi utilizado um túnel de vento em escala laboratorial para executar a análise experimental de poças de diesel (S-500) e biodiesel (B-100), em um tanque cilíndrico com tamanho reduzido de Ø110 x 57,4 mm. Além disso, foi utilizado o programa FDS para análise e comparação dos dados em escala de mesma grandeza da bancada experimental. A influência da velocidade do vento sobre a geometria da chama – inclinação, altura e comprimento – foi analisada, bem como outras questões pertinentes à estrutura da chama, como temperatura adimensional da chama, da pluma e temperatura periférica. Por último, uma análise da taxa de queima mássica foi feita para complementar os dados experimentais e obter mais informações sobre o comportamento da chama. Os dados experimentais foram aplicados nas correlações semi-empíricas de comprimento e ângulo de chama, a fim de comparar seu comportamento com a previsão de outros autores. Foi observado nos experimentos que, com a mudança da geometria da chama, ocorre a mudança da posição e consequentemente da dispersão da pluma. O ângulo da chama mostrou variação diretamente proporcional à velocidade do escoamento. A variação da altura e comprimento de chama foi inversamente proporcional à velocidade do escoamento. Os comportamentos do ângulo e da altura concordaram com a literatura, mas o comprimento apresentou diferenças. A temperatura adimensional da chama aumentou com o aumento da velocidade, sendo a do biodiesel 49% superior à do diesel. A temperatura adimensional da pluma apresentou um decremento com o aumento da velocidade de escoamento, atingindo, para o diesel, temperaturas cerca de 110% menores em relação ao biodiesel. A mesma tendência ocorreu com a temperatura periférica, que reduziu conforme aumentou a distância de medição da poça, havendo diferença de 20,3% entre os dois combustíveis. A taxa de queima mássica foi verificada e se observou que foi regida por diferentes mecanismos de trocas térmicas e flutuabilidade, o que provocou comportamentos diferentes para o diesel e o biodiesel, sendo as taxas de queima do diesel maiores, em geral. Os dados experimentais obtidos foram comparados com os resultados da análise numérica realizada no FDS. Obteve-se boas aproximações para o ângulo, comprimento, altura e temperaturas da pluma e periférica. As temperaturas da chama mostraram tendência diferente em relação aos dados experimentais. As análises de dispersão de poluentes mostraram uma tendência de redução abrupta das concentrações com o aumento da distância em velocidades menores e uma redução mais suave e constante para as maiores velocidades nos ensaios experimentais, apresentando, no entanto, uma grande diferença em relação aos valores numéricos, embora com tendências semelhantes, para ambos os combustíveis. / This work presents a numerical-experimental study of pollutants dispersion and flame geometry in a pool fire. Combustion modeling in a pool fire with plume dispersion as a function of the incident wind is carried out with the objective of phenomenon characterization. A laboratory-scale wind tunnel is used to perform the experimental analysis of diesel (S-500) and biodiesel (B-100) pool fire in a cylindrical tank with a reduced size of Ø110 x 57.4 mm. In addition, the FDS software was used to analyze and compare the results, using a model in a scale of the same magnitude of the experimental setup. The wind speed influence on the flame geometry – tilt angle, height and length – was analyzed as well as other questions related to the structure of the flame, such as dimensionless flame and plume temperature and outer layer temperature. Finally, an analysis of the mass burning rate was done to complement the experimental data and to obtain more information about the flame behavior. The data obtained were applied in the semi-empirical correlations of flame length and tilt angle to compare their behavior with the prediction of other authors. It Was observed that the change of flame geometry induces a change of plume position and dispersion. The behavior of the flame geometry was observed; the angle changes proportionally to the air flow speed. The variation in flame height and length was inversely proportional to air flow speed. The angle tilt and height agreed with the literature, but the length presented differences. The temperature of the flame increased with increasing of air flow speed being the values for biodiesel 49% higher than for diesel. The plume temperature presented a decrease with the increase of air flow speed, temperatures for diesel were about 110% smaller than for biodiesel. The same trend occurred with the measured outer layer temperature that reduced as the pool fire measurement distance increased, with a 20.3% difference between the fuels. The mass burning rate was governed by different mechanisms of heat feedback and buoyancy, which caused different behaviors for diesel and biodiesel, with diesel mass burning rates being higher in general. The experimental data obtained were compared with the results of the numerical analysis performed in the FDS software, and thus the numerical validation was done. The simulated results for tilt angle length, height and temperature of the plume and the outer layer agree well with experimental ones. The flame temperatures show an inverse trend in relation to experimental data. Pollutant dispersion analyzes showed a trend of abrupt reduction of concentration with increasing distance at lower air flow speed and a smoother and steady reduction at higher speeds, yet presenting a large discrepancy in relation to the numerical values for both fuels.

Chaque année en France, les feux de compartiments donnent lieu à plus de 88 000 interventions impliquant plus de 15 000 personnes dont plusieurs centaines de décès et de blessés graves. Aujourd’hui, lors de ces feux, les prises de décision et les délais d’intervention des équipes de secours sont principalement basés sur des décisions humaines, fruits de l'expérience. Une connaissance parfaite de la situation, de son évolution dans le temps et des dangers qui peuvent apparaître est impossible. C’est en partie la cause majeure des mauvais chiffres répertoriés ci-dessus. En effet, les processus physicochimiques qui régissent les feux de compartiments et les situations collatérales extrêmement réactives et dangereuses sont complexes. La transition entre un feu localisé et un feu généralisé peut prendre plusieurs formes. L'un des vecteurs les plus importants dans la propagation de la combustion pour les feux de compartiments sont les fumées, du fait de leur température élevée (souvent supérieure à 600°C) et des quantités importantes d’énergie, sous forme de chaleur, qu'elles contiennent. Malgré leur extrême dangerosité, les fumées restent importantes à étudier car elles véhiculent de précieuses informations, notamment sur l’apparition de phénomènes thermiques redoutés par les pompiers. Afin de mener cette étude, une cellule expérimentale composé de deux containers maritimes a été installée sur le site de formation incendie des sapeurs-pompiers de Seine-Maritime. Cette plateforme va permettre, grâce à un brûleur alimenté en propane, de produire des fumées chaudes dans uneconfiguration dite « feu réel ». Elles seront analysées en partie grâce à une technique de mesure non intrusive, la PIV. Les mesures par PIV grands champs seront comparées à des simulations LES de l’expérience (FDS). La double compétence (numérique / expérimentale) est essentielle dans ce type d’étude où les données expérimentales souffrent d’un manque de résolution (spatiale et temporelle) mais pour autant représentent des informations nécessaires à la validation des codes. / Each year in France, compartment fires result in more than 88,000 interventions involving more than 15,000 people, including several hundred deaths and serious injuries. Today, during fire compartments, the decision-making of the rescue teams is mainly based on human decisions, as a result of the accumulated experience. However, a perfect knowledge of the situation, its evolution over time and the dangers that may appear is impossible. Therefore, studying the fumes is of major interest. Indeed, smoke remains important to study because it conveys valuable information, especially on the appearance of thermal phenomena feared by firemen. In order to carry out this study, an experimental cell made up of two maritime containers was installed on the site of the Seine-Maritime fire brigade fire training. This platform will produce hot smokes in a configuration called "real fire" thanks to a propane burner. In this study, the smoke dynamics in a large scale experimental setup is analyzed using a non-intrusive measurement technique such as PIV (Particle Image Velocity). All the performed measurements are compared with LES (Large Eddy Simulation) simulations of the experiment using Fire Dynamics Simulator (FDS). The double expertise (numerical / experimental) is essential in this type of study where the experimental data suffer from a lack of resolution (spatial and temporal) but nevertheless represents an important source of information necessary for the validation of the codes.

Dagens samhälle är beroende av el, allt från att få rent vatten i kranen till att kunna tanka bilen kräver el. Vattenkraftverk står för nästan 50% av dagens elproduktion i Sverige vilket medför att vattenkraft är en väsentlig verksamhet för att få samhället att fungera. Ur ett miljöperspektiv anses vattenkraft vara mer miljövänlig än el som produceras på fossila bränslen. I och med den stora rollen vattenkraft spelar i dagens samhälle är det viktigt att förebygga eventuella olyckor och produktionsstopp, till exempel bränder. Bränder har tidigare förekommit på vattenkraftverk runt om i världen. Colorado är ett exempel där det började brinna i det brandfarliga materialet som arbetades med i en tunnel på stationen. Detta resulterade i en brand där flera arbetare blev instängda i en tunnel under marken, samtliga omkom då det endast fanns en utrymningsväg. I vattenkraftstationer som är belägna under marken är det vanligt förekommande att det endast finns en utrymningsväg. Olyckan påvisar den komplexitet som finns vid dimensionering av vattenkraftstationer under marken. Syftet och målet med detta projekt har varit att identifiera brandrisker på två vattenkraftverk och finna passande åtgärder för att minimera och förebygga dessa. För att uppnå projektets mål har bland annat en litteraturstudie genomförts för att ge en förståelse över problem som kan uppstå. En riskidentifiering genom platsbesök har också genomförts för att se hur det ser ut i verkligheten på de två vattenkraftstationerna. Det har även gjorts en datasimulering för att ge en uppfattning av rökspridningen, denna datasimulering har genomförts med hjälp av Fire Dynamics Simulator (FDS). Litteraturstudien påvisade, genom statistik från MSB:s databas IDA, att transformatorbränder är en av de största riskerna i anläggningar under marken. Vid platsbesöket på Messaure, ett vattenkraftverk, fanns det endast lokaltransformatorer placerade under marken. Dessa lokaltransformatorer var av typen torrtransformator vilket medför en mindre brandrisk. Tidigare fanns även övriga transformatorer under mark men dessa är flyttade ovan mark för att minimera brandrisken. Till skillnad från Messaure fanns Akkats, ett vattenkraftverk, transformator placerad under mark i en egen transformatorhall vilket utgör en egen brandcell. Entrén till denna fanns i direkt anslutning till tillfartstunneln som också används för utrymning. För att undersöka hur mycket en brand i transformatorhallen skulle påverka utrymningen valdes detta utrymme för en FDS-analys. Den slutsatsen som kan dras är att skulle dörren stå öppen skulle en brand i transformatorn ha en stor negativ inverkan på utrymningen. Rökutvecklingen skedde snabbt och röken strömmade tidigt ut i tillfartstunneln vilket medförde en påverkan på all utrymning. Vanligtvis ska dörren till transformatorhallen vara stängd men den kan lätt glömmas öppen i och med att det inte finns en dörrstängare. En kostnadseffektiv förbättring vore att installera en dörrstängare, det skulle kunna rädda liv. / Today's society is dependent on electricity, from getting clean water in the faucet to being able to put fuel in the car electricity is required. Hydroelectric power plants account for almost half of today's production of electricity, which means that hydropower is a vital factor for the community to function. From an environmental perspective, hydropower is considered more environmentally friendly than electricity produced from fossil fuels. Since hydroelectric power plants play a big role in society makes it very important to prevent possible accidents and hold-ups in production, such as a fire. Fires have occurred at hydropower plants around the world. Colorado is an example where the fire started in the flammable material that was being used in a tunnel at the hydropower plant. This resulted in a fire where several workers were trapped in a tunnel underground, all workers died since there was only one escape route. In hydropower stations located underground, it is common that there is only one escape route. The accident demonstrates the complexity that exists in the dimensioning of hydropower stations underground. The purpose and goal of this project has been to identify fire risks at two hydropower plants and how to minimize and prevent them. In order to achieve the project's goals, a literature study has been conducted to provide an understanding for the problems that may arise. A risk investigation was also conducted at the hydropower plants to see what they look like in the reality. A computer simulation has also been made to give an idea of the smoke spread. This computer simulation has been performed using the Fire Dynamics Simulator (FDS). The literature study showed, through statistics from MSB's database IDA, that transformer fires are one of the biggest risks in underground facilities. During the visit to Messaure, a hydroelectric power plant, only the local transformers were placed below the ground. These local transformers were of a dry-transformer type, which results in a smaller fire hazard. Previously, there were also other transformers underground, but these are moved above ground to minimize the fire hazard. Unlike Messaure, Akkat's, a hydroelectric power plant, transformer was placed underground in a separate transformer hall, which is a separate fire cell. The entrance to this was directly adjacent to the access tunnel which is also used for evacuation. To investigate how much a fire in the transformer hall would affect the evacuation, this space was chosen for an FDS analysis. The conclusion that can be drawn is that if the door should be left open, a fire in the transformer would have a major negative impact on the evacuation. Smoke quickly streamed into the access tunnel, which has an effect on all evacuation. The door should always be closed . However, there is no automatic door closer, which can result in the door being left open. A cost-effective improvement would be to install a door closer, which could save lives.