Development of realistic design fire time-temperature ：urves for the testing of cold-formed steel wall systems

Fire resistance rating of light gauge steel frame （LSF） wall systems is obtained from fire tests based on the standard fire time-temperature curve. However, fire severity has increased in modem buildings due to higher fuel loads as a result of modern furniture and light weight constructions that make use of thermoplastics materials, synthetic foams and fabrics. Some of these materials are high in calorific values and increase both the spread of fire growth and heat release rate, thus increasing the fire severity beyond that of the standard fire curve. Further, the standard fire curve does not include a decay phase that is present in natural fires. Despite the increasing usage of LSF walls, their behavior in real building fires is not fully understood. This paper presents the details of a research study aimed at developing realistic design fire curves for use in the fire tests of LSF walls. It includes a review of the characteristics of building fires, previously developed fire time-temperature curves, computer models and available parametric equations. The paper highlights that real building fire time-temperature curves depend on the fuel load representing the combustible building contents, ventilation openings and thermal properties of wall lining materials, and provides suitable values of many required parameters including fuel loads in residential buildings. Finally, realistic design fire time-temperature curves simulating the fire conditions in modem residential buildings are proposed for the testing of LSF walls.

Study on the Uncertainty of the Available Time Under Ship Fire Based on Monte Carlo Sampling Method

Available safety egress time under ship fire （SFAT） is critical to ship fire safety assessment, design and emergency rescue. Although it is available to determine SFAT by using fire models such as the two-zone fire model CFAST and the field model FDS, none of these models can address the uncertainties involved in the input parameters. To solve this problem, current study presents a framework of uncertainty analysis for SFAT. Firstly, a deterministic model estimating SFAT is built. The uncertainties of the input parameters are regarded as random variables with the given probability distribution functions. Subsequently, the deterministic SFAT model is employed to couple with a Monte Carlo sampling method to investigate the uncertainties of the SFAT. The Spearman＇s rank-order correlation coefficient （SRCC） is used to examine the sensitivity of each input uncertainty parameter on SFAT. To illustrate the proposed approach in detail, a case study is performed. Based on the proposed approach, probability density function and cumulative density function of SFAT are obtained. Furthermore, sensitivity analysis with regard to SFAT is also conducted. The results give a high-negative correlation of SFAT and the fire growth coefficient whereas the effect of other parameters is so weak that they can be neglected.

Effects of recovery time after fire and fire severity on stand structure and soil of larch forest in the Kanas National Nature Reserve, Northwest China

Forest recovery may be influenced by several factors, of which fire is the most critical.However, moderate-and long-term effects of fire on forest recovery are less researched in Northwest China.Thus, the effects of different forest recovery time after fire(1917(served as the control), 1974, 1983 and 1995) and fire severities(low, moderate and high) on larch(Larix sibirica Ledeb.) forest were investigated in the Kanas National Nature Reserve(KNNR), Northwest China in 2017.This paper analyzed post-fire changes in stand density, total basal area(TBA), litter mass, soil organic carbon(SOC) and soil nutrients(total nitrogen, total phosphorus and total potassium) with one-way analyses of variance.Results indicate that litter mass, TBA, SOC and soil nutrients increased with increasing recovery time after fire and decreasing fire severity, while the stand density showed an opposite response.The effects of fire disturbance on SOC and soil nutrients decreased with increasing soil depth.Moreover, we found that the time of more than 43 a is needed to recover the litter mass, TBA, SOC and soil nutrients to the pre-fire level.In conclusion, high-severity fire caused the greatest variations in stand structure and soil of larch forest, and low-severity fire was more advantageous for post-fire forest stand structure and soil recovery in the KNNR.Therefore, low-severity fire can be an efficient management mean through reducing the accumulation of forest floor fuel of post-fire forests in the KNNR, Northwest China.

Evaluation of deriving fire cycle of forested landscape based on time-since-fire distribution

Estimation of fire cycle has been conducted by using the negative exponential function as an approximation of time-since-fire distribution of a landscape assumed to be homogeneous with respect to fire spread processes. The authors imposed predefined fire cycles on a virtual landscape of 100 cell ×100 cell, and obtained a mosaic composing of patches with different stand ages (i.e. time since fire). Graphical and statistical methods (Van Wagner 1978; Reed et al. 1998) were employed to derive fire cycle from the virtual landscape. By comparing the predefined and the derived fire cycles, the two methods and tested the effects of sample size and hazard of burning (i.e., stand’s susceptibility to fire in relation to its stand age) were evaluated on fire cycle deviation. The simulation results indicated a minimum sample size of 10 times of the annual burnt area would be required for partitioning time-since-fire distribution into homogeneous epochs indicating temporal change in fire cycle. Statistically, there was significant difference among the imposed and the derived fire cycle, regardless of sample sizes with or without consideration of hazard of burning. Both methods underestimated the more recent fire cycle without significant difference between them. The results imply that deviation of fire cycle based on time-since-fire distribution warrants cautious interpretation, especially when a landscape is spatially partitioned into small units and temporal changes in fire cycle are involved.

Spatio-temporal changes of underground coal fires during 2008-2016 in Khanh Hoa coal field(North-east of Viet Nam) using Landsat time-series data

Underground coal fires are one of the most common and serious geohazards in most coal producing countries in the world. Monitoring their spatio-temporal changes plays an important role in controlling and preventing the effects of coal fires, and their environmental impact. In this study, the spatio-temporal changes of underground coal fires in Khanh Hoa coal field(North-East of Viet Nam) were analyzed using Landsat time-series data during the 2008-2016 period. Based on land surface temperatures retrieved from Landsat thermal data, underground coal fires related to thermal anomalies were identified using the MEDIAN+1.5×IQR(IQR: Interquartile range) threshold technique. The locations of underground coal fires were validated using a coal fire map produced by the field survey data and cross-validated using the daytime ASTER thermal infrared imagery. Based on the fires extracted from seven Landsat thermal imageries, the spatiotemporal changes of underground coal fire areas were analyzed. The results showed that the thermalanomalous zones have been correlated with known coal fires. Cross-validation of coal fires using ASTER TIR data showed a high consistency of 79.3%. The largest coal fire area of 184.6 hectares was detected in 2010, followed by 2014(181.1 hectares) and 2016(178.5 hectares). The smaller coal fire areas were extracted with areas of 133.6 and 152.5 hectares in 2011 and 2009 respectively. Underground coal fires were mainly detected in the northern and southern part, and tend to spread to north-west of the coal field.

基于网络开放数据的区域消防救援总体有效覆盖率评估

高效准确地掌握消防救援有效覆盖率是优化消防站点资源配置的前提。将判断是否有效覆盖的消防救援行程时间阈值定为240 s,建立了基于时间加权的区域消防救援总体有效覆盖率评估模型。基于网络开放数据收集了长沙的50座消防站作为消防救援供给点、5746家被抽查到的社会单位作为消防救援需求点,并调用网络地图应用程序编程接口基于实时路况来仿真消防救援行程时间。在连续7 d内设置197个评估场景,共获得了1131962个有效样本,进而得出:长沙消防救援总体有效覆盖率为21.22%,结果等级为“C”,消防救援水平一般,需加强消防站建设。这种基于网络开放数据的区域消防救援有效覆盖率评估,具备高效和准确的特点,可为优化消防站点资源配置、提升公共安全水平提供关键理论和方法支撑。

卷积与长短期记忆网络在火灾源强实时预测中的应用

针对火灾场景中火源位置及强度实时、准确识别困难的问题,利用卷积神经网络(Convolutional Neural Networks,CNN)与长短期记忆(Long Short-Term Memory,LSTM)网络的算法优势,构建一种火灾源强实时预测模型,该模型通过建筑内温度传感器接收的序列数据,实时、准确地预测火灾场景中的火源位置及强度信息。采用火灾动力学模拟软件(Fire Dynamics Simulator,FDS)模拟火灾场景,获得温度传感器实时接收的序列数据,建立火灾场景数据库,进行火灾场景数据分析并对火灾源强实时预测模型完成训练,通过实例验证该模型的准确性、及时性和鲁棒性。结果表明,该模型能够通过较短长度样本数据实时、准确预测火灾场景中火源位置及强度,预测准确率为99.18%,在温度传感器间隔损坏且损坏率不高于70%时,预测准确率仍可达到95.10%以上。

基于Transformer的矿井内因火灾时间序列预测方法

传统的基于机器学习的矿井内因火灾预测方法尽管具备一定的预测能力,然而在处理复杂的多变量数据时不能有效捕捉数据间的全局依赖关系,导致预测精度较低。针对上述问题,提出了一种基于Transformer的矿井内因火灾时间序列预测方法。首先,采用Hampel滤波器和拉格朗日插值法对数据进行异常值检测和缺失值填补。然后,利用Transformer的自注意力机制对时间序列数据进行特征提取及趋势预测。最后,通过调节滑动窗口的大小与步长,在不同的时间步长和预测长度下对模型进行不同时间维度的训练。结合气体分析法将矿井火灾产生的标志性气体(CO,O_(2),N_(2),CO_(2),C_(2)H_(2),C_(2)H4,C_(2)H_(6))作为模型输入变量,其中CO作为模型输出的目标变量,O_(2),N_(2),CO_(2),C_(2)H_(2),C_(2)H4,C_(2)H_(6)作为模型输入的协变量。选取陕煤集团柠条塔煤矿S1206回风隅角火灾预警的束管数据进行实验验证,结果表明:①对CO进行单变量预测和多变量预测,多变量预测相比单变量预测有着更高的预测精度,说明多变量预测能通过捕捉序列间的相关性提高模型的预测精度。②当时间步长固定时,基于Transformer的矿井内因火灾预测模型的预测精度随着预测长度的增加而下降。当预测长度固定时,模型的预测精度随时间步长增加而提高。③Transformer算法的预测精度较长短时记忆(LSTM)算法和循环神经网络(RNN)算法分别提高了7.1%~12.6%和20.9%~24.9%。

基于OSG引擎的森林火势蔓延仿真系统关键技术研究

针对森林火势蔓延趋势的仿真推演需求,以及森林消防队伍的行动路线选择和实时路径生成的需要,提出了一种融合数字高程信息(digital elevation model,DEM)的扩展晶格模型,将地形的植被、受灾属性与地理高程信息进行结合,通过实时运算得到结果数据并转换为可渲染的资源形式,利用OSG三维渲染引擎将其呈现到最终用户层面。借助这一数据结构以及对应的关键算法技术的研究实现,完成了一套较为完整的森林火势蔓延趋势仿真和灭火演练系统。该系统具有灵活性高,仿真效率和效果较好,支持各种并行计算方法的特点,可以在相关行业和系统中广泛推广和使用。

广西林火发生动态及气象影响因素研究

了解森林火灾的时空分布格局,定量分析其统计学规律,能够为该地区的火源管理与森林火灾科学预防提供基础支撑。根据广西1990—2020年的森林火灾数据、气象数据和基础地理数据,分析广西大尺度下的森林火灾时空分布趋势,同时构建零膨胀负二项回归模型,探讨气象因素对森林火灾的影响。结果表明,在年际变化上,广西林火数量呈“M”形波动,2005年以后森林火灾波动较大,季节上多发于冬季和春季,每年10月至次年4月累计发生13 249起,占比80%;森林火灾的发生具有明显的地域聚集性,主要发生在海拔0~400 m和斜坡为25°以下的平原低山地区;气象因子对森林火灾的发生具有重要影响,其中降雨量和平均风速对广西森林火灾的影响最为显著。未来广西应加大防火期,尤其是重要节假日前后的野外火源管控力度与应急处置能力,同时完善火险天气预警,以应对极端天气带来的森林火灾隐患。

基于改进MobileNetV3的选煤厂实时火灾识别方法研究

选煤厂的早期火灾识别可以极大保障人员和财产的安全,然而现有的火灾识别方法往往存在误报率高、实时性差等问题,难以满足实际需求。为了解决选煤厂实时火灾识别效果不理想的问题,基于计算机视觉领域中网络结构、轻量级网络和注意力机制的研究现状,针对目前选煤厂等工业生产环境中火灾识别方法存在识别延迟、推理速度较慢、难以进行实时识别等问题,提出将注意力模块CBSAM(Convolutional Block Second-order Attention Module)集成到Mobile-NetV3网络中来增强模型的表示能力,解决选煤厂的实时火灾识别问题。为提升特征表示能力,CBSAM模块在平均池化和最大池化特征的基础上引入二阶池化特征,通过通道注意力模块生成通道注意力图,然后通过空间注意力模块生成空间注意力图。为验证该模型在选煤厂实时火灾识别中的有效性,数据集采用开源数据集Fire Dunning Dataset和选煤厂现场收集的数据,通过消融实验验证CBSAM模块的有效性,通过对比实验验证CBSAM模块与改进MobileNetV3网络的泛化性与有效性。结果表明:通过引入CBSAM模块与MobileNetV3网络相结合,实现了轻量级网络与高效特征提取的有机结合,提高了网络的分类效果;CBSAM-MobileNetV3收集到的选煤厂火灾识别数据集具有良好的识别准确率。在MobileNetV3网络的基础上通过集成CBSAM模块减少了冗余信息的干扰,网络分类效果明显优于其他轻量级网络架构,能够有效地提升火灾识别的准确率,可以应用于选煤厂火灾识别任务。

节能砌块隐形密框复合墙体双面受火后滞回特性研究

为了评价节能砌块隐形密框复合墙体(以下简称复合墙体)受火后的滞回特性,制作了6个缩尺比例为1∶2的试件,并对试件进行了双面受火试验和受火后平面内循环往复加载试验,研究了轴压比、受火时间对墙体骨架曲线及滞回特性的影响。基于试验结果和现有恢复力模型理论,考虑轴压比和受火时间影响,提出了适用于受火后墙体试件的退化四线型恢复力模型。相关计算分析结果表明:计算得到的骨架曲线与试验情况以及计算得到的恢复力模型与墙体滞回曲线吻合程度均较好,表明提出的恢复力模型可较好地反映轴压比和受火时间对其滞回特性的影响,可为该墙体的结构弹塑性分析及工程应用提供参考。

流变助剂对膨胀型钢结构防火涂料性能的影响

以溶剂型丙烯酸树脂及氯化石蜡为基料,多聚磷酸铵、季戊四醇和三聚氰胺为膨胀体系,通过不同温度下炭层质量保持率、炭层强度、膨胀倍数、炭层形貌、耐火时间等考察了不同类型的流变助剂对防火涂料性能的影响。结果表明:流变助剂对膨胀型钢结构防火涂料性能的影响非常大。聚酰胺蜡、水合硅酸镁类流变助剂对体系的炭层强度、膨胀倍数、耐火性能等基本无影响;改性水合硅酸镁可以提高防火涂料的耐火时间和炭层强度;亲水气相二氧化硅会导致炭层膨胀倍数和耐火时间增加,同时会降低炭层强度,导致炭层易脱落;疏水二氧化硅、改性蒙脱石类流变助剂会抑制膨胀发泡、提高炭层强度,同时会大幅度降低耐火时间。分析探讨了不同类型流变助剂对防火涂料的性能影响的机理。不同流变助剂在高温膨胀发泡过程中主要影响体系的熔融黏度和膨胀倍数,从而影响防火涂料的综合性能。

基于热-力耦合的悬索桥结构抗火性能研究

利用火灾动力学软件FDS和有限元软件ABAQUS,开展基于热-力耦合原理的悬索桥结构抗火性能数值分析,探讨不同火灾工况(火源功率和环境风速)下悬索桥温度场及力学响应的变化规律。结果表明:不同火源工况的温升曲线变化趋势基本一致,但其温度上升速率及最高温度并不相同。随着火源横向、纵向距离增加,构件温度逐渐降低。随着风速增大,显著提高桥梁结构耐火时间,最大提高可达50%。随着火源功率增大,其耐火时间显著降低。当环境风速为2 m/s时,耐火时间下降约33.33%。

全国森林火灾形势及其时空聚集性特征分析

为进一步研究全国森林火灾发生的时间规律与空间聚集性特征,基于全国2004—2022年度、2015—2021年月度的森林火灾发生次数及2017年31个省(自治区、直辖市)水文地理条件数据,利用Prophet时间序列模型与模糊聚类分析法对森林火灾特征进行分析。结果表明:Prophet时间序列模型在森林火灾预测领域具有良好的适用性;时间规律方面,2004—2022年森林火灾次数呈先下降后上升再下降趋势,且季节因素影响突出,节假日影响因素不显著;空间聚集性方面,31个省(自治区、直辖市)可大致分为7个类别,其分类基本与地域特点有关。研究结果可为挖掘森林火灾时空分布及聚集性特征提供参考。

不利天气条件下机场应急救援能力评估研究

为评估机场在有风、低能见度等不利天气条件下的应急救援能力,以灭火时间和伤亡人数为评估指标,基于火灾面积和人员安全分析方法,利用仿真软件AnyLogic构建机场应急救援能力评估模型。结果表明:3种能见度条件下造成的重伤、死亡人数均高于最佳能见度条件;在低能见度环境下救援会有泡沫灭火剂施放中断和灭火失败的风险;机场消防响应时间控制在2min内对降低事故后果具有重要作用;3种风速条件下造成的重伤、死亡人数均高于无风条件;灭火时间随风速呈非线性变化。该仿真模型可快速推演出不同风速、能见度下的灭火时间和伤亡人数,仿真结果可为机场提供辅助决策支持。

基于RTMP的森林火灾预警监控系统设计

高温是森林火灾发生的主要原因之一,对林区进行分布式温度监测是提早发现火灾隐患的有效途径。设计了一种使用Linux系统和实时传输协议进行温度等环境参数检测的森林火灾预警监控系统,系统终端节点集成温湿度、位置和视频采集功能,利用TCP和RTMP协议实现多个分布式终端节点数据和视频流的实时传输,在监控端实时监测多个地点的环境数据,排查潜在的火灾隐患。通过3个节点的测试,系统可以实现终端节点数据、视频汇聚,且使用FFmpeg进行视频拉流比传统VLC拉流速度快约10秒。

化工园区多种消防救援队伍联勤联训联防机制建设分析

化工产业作为“工业中的工业”,是国民经济发展的重要基础和支持产业,面向化工园区“更清洁、更环保、更安全、碳中和”的发展愿景,安全高速的发展给消防工作带来了新的挑战。目前,各地方负责园区处置突发灾情的消防救援队伍有城市消防救援站和企业专职消防队,通过研究柳州市鹿寨化工园区依托这两种形式队伍建立起来的区域联防联动组织的运行情况,就如何建立联勤联训联防机制进行探讨。

多点位监控视频分析在火灾过程重构中的应用

以一起大型仓库火灾事故调查为例,介绍了多点位监控视频分析在火灾过程重构中的应用。通过点位初筛筛选出包含有效信息的视频,通过相对时间校准消除了各点位的时间误差,然后根据各点位视频画面综合分析确定了起火时间,根据各点位光影交叉定位和相互印证,确定了起火部位和火势发展情况,从而重构了该起火灾的起火过程。在此基础上,分析了火灾调查工作中提取多点位监控视频的注意事项,总结了多点位监控视频分析中的一些原则和方法。

智能网联在消防实时监测与预警中的应用

阐述智能化消防监控系统的应用案例,对系统性能进行评估与优化,包括硬件升级、代码优化、数据库优化、负载均衡、横向扩展、纵向扩展、缓存策略、并发控制、定期维护和监控、日志分析。