基于灰色BP网络的火灾事故预测

将BP网络模型与灰色系统预测方法相结合,利用人工神经网络对GM(1,1)模型进行残差修正,建立了火灾事故预测模型.该模型兼具灰色系统与BP网络的优点,不但能体现火灾复杂的灰色系统行为,又能根据每次学习误差的不同调整学习速率,加快收敛速度.与单一的GM(1,1)模型相比,该模型预测结果具有更高的精度.     

基于元胞自动机的城市区域火蔓延概率模型探讨

考虑建筑结构物震后可能发生的次生火灾的不确定性,给出解决问题的基本思路,构建了基于元胞自动机的城市地震次生火灾蔓延概率模型,指出元胞着火的概率与建筑物的特性和灾害条件有关,包括建筑材料、外墙是否有开口、与着火元胞的距离、地震对建筑的破坏程度以及气象条件等因素。模型通过设置没有开口的外墙以及道路和空地来区分单体建筑和建筑之间的火蔓延。该模型量化了影响火蔓延概率的因素,能快速实现城市区域火蔓延过程的动态模拟,特别适宜用于城市宏观角度上的火蔓延模拟。模拟结果再现了火灾过程,这不仅提高了人们对城市区域火蔓延危害的认识,而且为城市规划设计和消防扑救措施提供了有益的理论依据。     

水幕在隧道中灭火的数值模拟研究

利用CFD软件中的专用FDS软件对水幕在隧道产生火灾进行灭火模拟研究,模拟水幕对隧道中火灾产生的烟气流动速度、烟气浓度以及温度等特性参数的影响,模拟得到水幕能有效地减慢地下隧道中烟雾的扩散和降低烟雾的温度,为地下工事中的人员逃离提供有力的保证,为加强地下工事的防御提供了理论参考。     

电缆隧道火灾数值仿真及分析

电缆隧道灭火以及人员疏散的关键,在于对灾变条件下隧道火灾参数变化的正确预测,特别是火焰、烟气蔓延范围,烟气浓度变化以及有毒气体的扩散范围等参数的预测。为了获得电缆隧道火灾参数,应用美国国家标准和技术研究院(NIST)开发的FDS(Fire Dynam ics Simulator)软件,建立电缆隧道模型,对隧道火灾进行全尺寸模拟,通过对模拟实验数据处理和分析,给出电缆隧道火灾时烟气浓度和氧气浓度,纵向温度的变化规律,火焰蔓延情况以及高温烟气在隧道中水平蔓延规律,为有效救援和紧急疏散以及消防决策提供一定依据。     

公共聚集场所防火设计中的火灾场景设计研究

在对公共聚集场所性能化防火设计的基础上,介绍火灾场景设计的内容和方法,通过对火灾荷载和火灾热释放速率的计算,得到火灾发展的变化情况和热释放速率曲线,进而构建适合公共聚集场所等大空间建筑的火灾设定模型。并应用该计算方法和模型对某超市进行火灾场景设定来确定该火灾场景的最大热释放速率,并对其火灾蔓延可能性进行讨论,为火灾危害性评估以及火灾探测和扑救系统的有效性评估提供依据。     

消防工程专业本科课程体系的结构研究

在分析讨论国内外消防工程专业教育现状与课程体系研究进展的基础上,针对消防工程专业所涉及的学科领域,运用系统论的观点和方法,分析消防工程专业本科课程体系的结构层次特点。从课程体系的结构与功能的相互关系出发,按照培养目标、知识范围和应用领域的要求,构建出"课程体系→课程群→主干课程群→主干课程"的消防工程本科课程体系的结构模式,初步给出了课程群、主干课程群的结构与组成以及主干课程群的基本内容和教学要求。     

基于统计和信度理论的火灾保险费率分级方法

通过对建筑火灾保险费率厘定的调查、研究,确定了关于建筑物火灾保险费率的主要影响因素;详细介绍基于统计理论和信度理论的火灾保险费率分级方法,并以美国马萨诸塞州东海岸地区为例,给出其实际的应用算例;结合实际中的问题,分析了基于统计理论和信度理论的火灾保险费率分级方法的优缺点,为我国建筑火灾保险费率厘定的研究提供了依据和参考。     

电缆隧道火灾有效灭火技术试验研究

针对工业企业的电缆隧道的灭火技术进行试验研究,探讨电缆隧道火灾的有效灭火安全措施并对不同灭火措施的灭火效果进行分析。利用火灾动力学模拟程序软件对电缆隧道火灾进行模拟,研究火灾中喷淋系统启用前后各时间点测点温度以及各时段内的降温速率,并与试验结果作对比分析,论证了灭火实验结果有效性,确定了电缆隧道火灾中无机隔板和防火门的安全防火时效,以及4种常用灭火剂在各时段的灭火效果。     

The Impacts of Human Visitation on Mussel Bed Communities Along the California Coast: Are Regulatory Marine Reserves Effective in Protecting These Communities?

Rocky intertidal habitats frequently are used by humans for recreational, educational, and subsistence-harvesting purposes, with intertidal populations damaged by visitation activities such as extraction, trampling, and handling. California Marine Managed Areas, particularly regulatory marine reserves (MRs), were established to provide legal protection and enhancement of coastal resources and include prohibitions on harvesting intertidal populations. However, the effectiveness of MRs is unclear as enforcement of no-take laws is weak and no regulations protect intertidal species from other detrimental visitor impacts such as trampling. The goal of this study was two-fold: (1) to determine impacts from human visitation on California mussel populations (Mytilus californianus) and mussel bed community diversity; and (2) to investigate the effectiveness of regulatory MRs in reducing visitor impacts on these populations. Surveys of mussel populations and bed-associated diversity were compared: (1) at sites subjected to either high or low levels of human use, and (2) at sites either unprotected or with regulatory protection banning collecting. At sites subjected to higher levels of human visitation, mussel populations were significantly lower than low-use sites. Comparisons of mussel populations inside and outside of regulatory MRs revealed no consistent pattern suggesting that California no-take regulatory reserves may have limited effectiveness in protecting mussel communities. In areas where many people visit intertidal habitats for purposes other than collecting, many organisms will be affected by trampling, turning of rocks, and handling. In these cases, effective protection of rocky intertidal communities requires an approach that goes beyond the singular focus on collecting to reduce the full suite of impacts.     

Spatio-temporal analysis of fire events in India: implications for environmental conservation

Information on fires in different geographic regions of India is relatively scarce. This study quantifies spatial and temporal patterns in fire occurrences covering different states and districts in India. Two important scientific questions are answered in this study: (1) how are the fire events distributed across different geographical regions? (2) are there any specific districts where fire events clustered across space and time? To address these questions, Along Track Scanning Radiometer (ATSR) derived satellite fire counts from 1997–2006 were used and the datasets were analysed using spatial scan statistic. Spatial scan statistic provides a test statistic for most likely ‘hotspot’ spatial clusters, based on the likelihood ratio test and Monte Carlo simulation. Results from geographical analysis based on state boundaries suggested Maharastra state had the highest number of fires followed by Madhya Pradesh, Chattisgarh, Orissa, etc., during the 10-year period. Among the several districts, the spatial scan statistic identified the most likely cluster of fire events in Dausa, Karauli, Sawai Madhopur, Bharatpur and Alwar in addition to several other secondary clusters, with high statistical significance. These results are based on a large sample of cases, and they provide convincing evidence of spatial clustering of fire events in the Indian region. Results relating to hotspot areas of fire risk can guide policy makers towards the best management strategies for avoiding damages to forests, human life and personal property in the ‘hotspot’ districts.