【作者】 黎昌海；

【导师】 陆守香；

【作者基本信息】 中国科学技术大学 ， 安全技术及工程， 2010， 博士

【摘要】 腔室火灾是常见的建筑火灾场景,在建筑火灾安全中得到了广泛的关注和研究。前人针对有门窗等开口的建筑空间火灾进行了大量研究并建立了相关的火灾模型,然而很少研究涉及到没有对外开口或仅有顶棚开口的封闭空间火灾。本文针对以船舶机舱为代表的无开口或仅顶棚开口封闭空间火灾,在自行研制的封闭空间火灾实验系统中进行了池火实验研究,目的是为了揭示封闭空间池火特征及行为规律。首先研究了无开口条件下的池火行为,在体积较小的A舱(0.75m3)中进行了无开口条件下封闭空间池火燃烧行为特征研究,包括火焰高度特性、脉动特性、自熄灭特性、燃烧速率特性和气体温度分布特性等；随后在体积较大的B舱(17.55m3)中进行了游走火行为的探索研究。其次,在无开口封闭空间实验结果的基础上,基于氧气质量守恒建立了无开口封闭空间的熄灭时间预测模型,通过无量纲分析研究了无开口封闭空间池火自熄灭时间的影响因素。最后在A舱中进行了单顶棚开口条件下池火燃烧行为研究,重点研究开口大小对燃烧状态和熄灭时间的影响。具体工作包括：自行设计了封闭空间实验系统。用于本文池火实验的封闭空间有两个,体积较小的A舱内尺寸为1000mm(L)×1000mm(W)×750mm(H),体积较大的B舱内尺寸为3000mm(L)×3000mm(W)×1950mm(H), B舱与某机舱的几何比约为1：4。A舱顶棚一角可设正方形开口,最大开口边长为0.490m。实验采用正庚烷为燃料,在A舱中进行的实验油池直径有0.100m、0.141m、0.200m和0.300m四种,在B舱进行的实验油池直径有0.200m和0.300m两种。实验利用电子天平测量了燃烧过程中油池质量随时间的变化从而得到质量损失速率,利用K型热电偶阵列测量了舱内气体温度的分布,利用烟气分析仪测量了池火根部附近和空间上层气体组分的变化,利用数码摄像机记录了实验过程中的火焰图像。分析了封闭空间内火焰噪声特性和色彩空间特性,提出了用H、B和I分量联合分割火焰彩色图像的方法。对封闭空间内燃烧实验中捕获的典型图像噪声、图像在RGB色彩空间和HSI色彩空间各个分量上的特性进行的分析表明,虽然封闭空间火焰图像红色R、绿色G和亮度I分量图受与火焰颜色相似的淡红色噪声污染严重、火焰与背景之间的分界模糊,但噪声在色调H分量图和蓝色B分量图上平稳均一,H和B分量图受噪声污染较小,火焰与背景区分明显。联合使用色调H、蓝色B分量与亮度I共同分割火焰彩色图像,并将分割后的各分量图像进行融合处理,能有效减少噪声的干扰而取得较好的分割效果。研究了无开口封闭空间的池火行为特征。结果表明,火焰根部面积的变化频率与火焰高度的振荡频率相等,在此基础提出了采用火焰根部面积的变化频率求取火焰振荡频率的方法,并应用于具有顶棚射流的封闭空间池火火焰振荡频率的确定。燃烧过程中油池火焰振荡频率波动较小,火焰振荡频率小于开放空间经验公式预测值,火焰振荡频率与油池直径的关系拟合结果为f=1.33D-0.5,且满足无量纲关系式St＝0.26Fr-0.532。燃烧过程中的单位面积平均燃烧速率小于开放空间所得的实验结果,但也随着油池直径的增大而增大。燃烧因缺氧而发生自熄灭,不同直径油池的燃料的总消耗基本上相等,不随燃料池直径大小的变化而明显变化。火焰熄灭时池火根部附近氧浓度值在10.5%至15.3%之间,而空间剩余氧平均浓度大约为14.1%,熄灭时间与空间体积成正比而于油池大小呈反比并符合关系式tE=4.418V／D2。在燃烧过程的中后期,烟气分层不明显,在火焰熄灭时,空间内只有热烟气层和污染层。上部空间的热烟气层温度高于下部污染层温度,熄灭时污染层的温升随高度的减小基本上呈线性递减。并建立了无开口封闭空间熄灭时间预测模型。基于氧气质量守恒建立了无开口封闭空间的熄灭时间预测模型。定义了空间体积形状因子、无量纲火源体积和无量纲自熄灭时间,结果表明无量纲熄灭时间正比于封闭空间初始氧质量分数与剩余氧分数的差值以及燃料属性诸如燃烧热、燃烧反应化学当量比等,但反比于环境温度、空间体积因子、无量纲火源体积和综合燃烧系数。综合本文实验和NRL实验结果,综合燃烧系数的经验公式为研究了不同油池高度对池火行为的影响,探索了无开口封闭空间中池火的游走火行为。在B舱中进行了不同池火位置高度的庚烷池火实验,实验结果表明,无开口的封闭空间池火熄灭时间随着油池高度位置的增大呈先增大后减小的趋势,燃料沸腾后无开口封闭空间出现了明显的游走火现象,燃料沸腾产生大量可燃蒸气及卷吸到火焰区的氧气含量过低,导致空间内在燃料区以外较远的区域存在满足燃烧条件的蒸气浓度和氧气浓度是游走火出现的根本原因,油池位置高度越高,火焰游走距离越远,而游走火出现时间和持续时间与油池位置增加并非简单递增或递减。研究了单顶棚开口封闭空间的池火行为特征。结果表明,顶棚开口大小对封闭空间中的池火行为有决定性影响,按顶棚开口的大小对封闭空间池火行为的影响,封闭空间池火的燃烧模式可分为四种：密闭燃烧模式、近密闭燃烧模式、开口控制燃烧模式和通风控制燃烧模式；当燃料充足时,只有当顶棚开口大于一定值(临界开口大小)时,燃料才能因燃烧耗尽,按熄灭时油池内是否有燃料剩余可分为燃料剩余区和燃料耗尽区,两区之间的临界开口大小随油池的增大而增大。在燃料剩余区,开口较小时熄灭时间和燃料消耗率与无开口时差别不明显,而开口较大时熄灭时间和燃料消耗率随开口的增大而增大。在油池直径较小时,燃烧过程中燃料难以沸腾,熄灭时油池边缘氧浓度随开口的增大而增大；在油池直径较大时,当燃烧进行到一定时期燃料发生沸腾后,游走火现象明显,熄灭时油池边缘测到的氧浓度普遍较低,在本文实验中,游走火现象明显的顶棚开口条件下,油池边缘氧浓度在熄灭时刻最低可以达到10%左右。在燃料耗尽区,熄灭时间随开口的增大先急剧减小后趋于稳定。平均质量损失速率随顶棚开口大小的变化呈S形变化。开口很小时,平均质量损失速率接近于无开口时同直径池火,而开口很大时,平均质量损失速率接近于自由燃烧时同直径池火,在燃料剩余区和燃料耗尽区分界线附近,平均质量损失速率随开口增大而增大。在定义了无量纲时间(?)和ω的基础上,发现近密闭燃烧模式到开口控制模式的临界开口因子约为0.03,燃料剩余区和燃料耗尽区临界开口因子约为0.057。在燃料剩余区,不同直径池火的无量纲熄灭时间(?)随开口因子ω的变化可以近似采用经验公式(?)=1.097+ω951来描述。更多还原

【Abstract】 Compartment fires are widespread building fire scenarios, which are studied extensively for fire safety issues in building fire safety. A great deal of studies on compartment fires with doors or windows have been done by fire researchers and lots of models had been built, while few concern with the cases where there is no vent from the compartment to outside or is ceiling went only. Aiming at fires in closed compartment without vent or only ceiling vent such as ship machinery spaces, a series of tests dealing with pool fire behaviors were conducted under either no vent or ceiling vent conditions in the fire test system developed specially for closed compartment, to find out what characteristics and rules there are. First, fire behaviors under no vent condiction were investigated, the characteristics of flame height, flame pulsation, self-extinction, burning rates, gas temperature distribution were investigated in Chamber A (0.75m3) while ghosting behaviors were explored in Chamber B (17.55m3). Secondly, a model on self-extinction time of fires in closed compartment was developed based on oxygen conservation and experimental results of self-extinction time, factors affect the self-extinction behaviors were analyzed by nondimensionalization method. Finally, Fire behaviors under the single ceiling vent condictions focus on the extinction behaviors were investigated. The work has been done includes the following.Fire test system for closed compartment was developed. Two closed compartments with different sizes are used. The smaller one with the inner dimension of 1000mm(L)×1000mm(W)×750mm(H) was named as Chamber A, the larger one with the inner dimension of 3000mm(L)×3000mm(W)×1950mm(H) was named as Chamber B. Chamber B was built with 1:4 scaled to the machinery space in a certain ship.Tests under no vent condition were performed in both Chamber A and Chamber B, while tests under ceiling vent only conditions were performed in Chamber A. The variable opening with the largest side size of 0.490m was located in one of the coners of the ceiling wall. Fire sources are n-heptanes pans with the diameters of 0.100m,0.141m, and 0.200m and 0.300m in Chamber A, as well as 0.200m and 0.300m in Chamber B. The mass loss rates (MLR) of fuel were got by monitoring the mass changes of fuel pan by an electronic balance. Oxygen concentrations nearby the fuel pans and at the upper layer during the combustion were measured by smoke analyzers. Gas temperatures were measured by several TC arrays K type sensors. The flame images of the combustion process were recorded by video cameras to investigate the flame charatistic and extinction behavior.A method on segment flame color images by joint utilization of hue, blue and intensity was proposed based on the analysis on the characteristics of noises, components in color space and HIS space of flame in closed compartment.The analysis on the characteristics of noises, components in color space and HIS space shows that the red-component, green-component and the intensity-component gray images are seriously polluted and that the boundary lines between the flame and background are blurred by light red noises, which color is similar to that of flame. But in the much less polluted hue-component and blue-component gray images, the noises are almost uniform and distinguishing from the flame apparently. The segment flame color images by joint utilization of hue, blue and intensity and fuse the components could reduce the disturbance of various noises efficiently and have a good performance.Characteristic of pool fire behaviors in closed compartment without vent were investigated. The results show that the pulsation frequency of flame bottom area equals to that of flame height, a method using flame bottom area was proposed to determine flame pulsation frequency and applied to pool fires in closed compartment with ceiling jet flames. The results show that the variation of flame pulsation frequency is small, and flame pulsation frequency is smaller than that predicted by empirical equations developed for burnings in free atmosphere while it fits well with the scaling relationship st=0.26Fr-0.532, the expression of f=1.33D-0.5 is appropriate in describing the relationship between the flame pulsation frequency and pool diameter. The average burning rates were found to be lower than those in the burning at free atmosphere and increase as the increase of pool sizes, but the total fuel consumed of each tests are almost the same, do not change with pool sizes.The behaviors of self-extinction take place due to oxygen starvation. The fire self-extinction happened when local oxygen mole fraction in the vicinity of the flame descended to a level of 10.7%-15.3%. The mean remaining oxygen mole fraction when the fire self-extinguished is about 14.1%.It is found that the fire self-extinction time is propotion to the compartment volume but inversely to the pool size and fit well with the expresseion of tE= 4.418V/D2. In the middle and later of burning process, the stratification behavior of smoke is not clear. Only hot smoke layer and polluted layer remain when extinction takes place.The temperatures in upper layer (hot smoke layer) is much higher than those in lower layer (polluted layer). The temperature decreases linearly with height in the polluted layer.Model on prediction of self-extinction time was developed. Based on oxygen mass conservation, a prediction model of self-extinction time of the pool fire in closed chambers was developed. By defining the concepts of the chamber shape factor and the dimensionless fire volume, the dimensionless fire self-extinction time is proportional to the difference between the initial and remaining oxygen mass fraction, fuel properties such as heat of combustion and stoichiometric ratio etc., but inverse to the ambient temperature, chamber shape factor, the dimensionless fire volume and the integrated combustion coefficient. The prediction has a good agreement with the experimental results. The model also reveals a good prediction to the results of NRL’s tests. An experimental formula of Xo= 0.094 is also proposed to estimate the integrated combustion coefficient.Fire behaviors under different pool height in position were investigated, as well as ghosting fire behaviors in closed compartment without vent were explored. A series of fire tests with different height of fire source location were conducted in Chamber B without vents. The results showed that extinction time increase before descreasing as the pool height in position. Ghosting fire occurred in all tests tests when fuels boiled. Excessive combustible gases produced by boiling but low oxygen entrained to flame zone result in redundant combustible gases flow to somewhere far from the fuel zone, mixing with oxygen and reach ignition condition, leading to the ghosting fire. Travel distances of ghosting fire could be longer as the pool position raised, while time of ghosting fires begin and time of ghosting fires last do not simply increase or decrease.Characteristic of pool fire behaviors in closed compartment with sinlge ceiling vent were investigated. The results show that the size of the ceiling vent has remarkable influence on the fire behaviors in closed compartment. Four regimes of burning states named completely closed regime, nearly closed regime, vent size controll regime and ventilated-controll regime were distinguished. Only when ceiling vent is larger than a critical size can fuel consumed completely if there is enough fuel. The "fuel remaining stage" and "fuel exhaustion stage" could be distinguished according to whether there is fuel left at extinction. In "fuel remaining stage", Extinction time and total fuel consumed are similar to those in closed compartment without vent when vent size is small but increasing with vent size if vent size is large.Extinction time reduced as distance to the criticality between the fuel remaining stage and fuel exhaustion stage. Extinction time and average burning rates are similar to those in closed compartment without vent when the ceiling vent size is remarkable small or similar to those in free atmosphere when ceiling vent size is large enough.It’s hard to reach boiling states when pool is small and oxygen concentration increase as vent size. While it’s easy to reach boiling state for large pool and ghosting flames are easy seem. Oxygen concentration at extinction is found to be rather low when ghosting fires occur even to 10% in our tests. Extinction time increases before sharply dereasing as vent size in "fuel exhaustion stage". The average mass loss rates changing like a "S" as vent size increases. The average burning rates are similar to those in closed compartment without vent when the ceiling vent size is remarkable small or similar to those in free atmosphere when ceiling vent size is large enough. The average mass loss rates increase as vent size when near the criticality between the fuel remaining stage and fuel exhaustion stage.By defining the dimensionless ceiling vent sizeωas vent size comparting to characteristic length of fire and distance to vent to pool,ωabout to be 0.03 for the criticality beween nearly closed regime and vent size control regime, whileω=0.057 is the criticality between the fuel remaining stage and fuel exhaustion stage. The dimensionless extinction time (?), defined as self-extinction time comparing to that in closed compartment without vent, couble be describled using an experimental formula (?)=1.097+ω9.51。更多还原