Multiple structural modifications to dendritic ecological networks produce simple responses

Structural modifications to landscapes affect the ability of organisms to access different habitat patches. There exist, however, very few general methods by which to relate modifications to expectations of effects, and even fewer that enable understanding of how multiple modifications may interact. In the absence of any guiding principles, ecologists have assumed that interactions will result in complex landscape-scale effects. One way of understanding such effects is through rendering a landscape as a graph or network, among the simplest of which are dendritic networks typified by stream systems. Yet even for stream networks, there are no known general principles concerning the nature of interactions between multiple modifications. We developed a model to describe the ability of fish to access and use different habitat patches within dendritic networks. We used mathematical and numerical analyses of the model to investigate how the habitat value of a network is affected by changes in connectivity and habitat quality, and then to examine interactions between multiple modifications. Rather than showing complex interactions, our analytic and simulation-based results show that the combined effect of multiple modifications approximately equals the sum of individually predicted effects. Dendritic networks thus appear to respond far more simply to multiple modifications than has previously been assumed. These results have implications for stream management planning, and offer a firm foundation from which to better understand population processes within dendritic networks.     

基于人工神经网络法的绿色建筑评价

以实际建筑物为例,介绍了用层次分析法建立绿色建筑评价模型的过程,并分别用层次分析法和人工神经网络法对实际建筑物进行了评价。评价结果显示,人工神经网络法与层次分析法相对误差不到0.5%,表明人工神经网络法作为一种客观科学的评价方法,应用于绿色建筑的评价,能有效降低主观因素带来的影响,会使结果更具有客观性。     

Quantitative risk analysis of urban natural gas pipeline networks using geographical information systems

This paper presents a novel quantitative risk analysis process for urban natural gas pipeline networks using geographical information systems (GIS). The process incorporates an assessment of failure rates of integrated pipeline networks, a quantitative analysis model of accident consequences, and assessments of individual and societal risks. Firstly, the failure rates of the pipeline network are calculated using empirical formulas influenced by parameters such as external interference, corrosion, construction defects, and ground movements. Secondly, the impacts of accidents due to gas leakage, diffusion, fires, and explosions are analyzed by calculating the area influenced by poisoning, burns, and deaths. Lastly, based on the previous analyses, individual risks and social risks are calculated. The application of GIS technology helps strengthen the quantitative risk analysis (QRA) model and allows construction of a QRA system for urban gas pipeline networks that can aid pipeline management staff in demarcating high risk areas requiring more frequent inspections.     

Social capital networks for achieving sustainable development

There have been few empirical studies investigating whether and how social capital (SC) relates to better achievement of sustainability goals and, if so, how. This research investigates the roles three types of SC, namely, bonding, bridging, and bracing SC, play in achieving sustainable development (SD), using a case study of the Japanese region to explore the process of SC accumulation leading to collective action. The research question as to whether SC accumulation makes a difference in the progress towards sustainability is addressed qualitatively and quantitatively using a case study, and network and regression analysis; in particular the impact and functions of bracing SC are closely investigated. The study concludes that SC accumulation can indeed make a difference in achieving sustainability and that bracing SC plays an essential role in expediting the processes of goal sharing and resource flow by connecting various networks across sectors and scales, thereby making collective action possible. These findings suggest that creating an environment in which the generation of all three types of SC is encouraged may help local governments to achieve their desired policy goals for SD.     

Climate change in mountain regions: how local communities adapt to extreme events

This paper examines how local communities adapt to climate change and how governance structures can foster or undermine adaptive capacity. Climate change policies, in general, and disaster risk management in mountain regions, in particular, are characterised by their multi-level and multi-sectoral nature during formulation and implementation. The involvement of numerous state and non-state actors at local to national levels produces a variety of networks of interaction and communication. The paper argues that the structure of these relational patterns is critical for understanding adaptive capacity. It thus proposes an expanded concept of adaptive capacity that incorporates (horizontal and vertical) actor integration and communication flow between these actors. The paper further advocates the use of formal social network analysis to assess these relational patterns. Preliminary results from research on adaptation to climate change in a Swiss mountain region vulnerable to floods and other natural hazards illustrate the conceptual and empirical significance of the main arguments.     

大型公园道路网络关键节点脆弱性评价*

为更加科学有效地辨识景区道路网络中的客流关键节点,以节点脆弱性为度量指标,提出1种基于FIM模型的关键节点脆弱性评价方法。以某大型公园为例,首先通过ArcGIS软件将该公园的道路网络信息转换成可编译的文本信息,经过Java枚举可行路径,然后利用嵌入FIM算法的Lingo进行优化,得出网络节点的重要度。最后综合节点容量、流量、吸引度等信息加权得出网络节点的脆弱性。结果表明:该方法能够有效地对道路网络节点进行脆弱性度量并排序,景区可据此制定相应的客流导向方案以提高景区资源的利用率,提高游览的安全性。     

基于多智能体建模方法的人群应急疏散模型构建研究——以腾龙芳烃（漳州）有限公司“4·6”爆炸着火重大事故疏散为例*

为从更微观角度分析人群疏散过程中疏散行为及路网设计对疏散效果的影响,基于腾龙芳烃（漳州）有限公司“4·6”爆炸着火重大事故,构建多智能体人群应急疏散模型,模拟人群中个体群组、惯性、就近、从众、信息传播的行为决策及相互交互影响。结果表明:群组行为会严重影响应急疏散效果,在疏散路网两端避难所附近会出现明显拥堵现象,在疏散路网设计和避难所选择时,应尽可能避免出现极端汇流路段或节点;在应急培训中,应告知群众减少群组行为。研究结果可为人群应急疏散提供借鉴。     

基于改进的YOLOv4安全帽佩戴检测研究

为了加强建筑工人佩戴安全帽情况的检测,防止安全事故的发生,提出1种改进的轻量级YOLOv4安全帽佩戴检测算法,用于运行在移动设备端,降低现场部署的条件;制作1个8 000幅图像的数据集,用于训练和评估安全帽检测算法;为了评估改进的YOLOv4的性能,从5个不同建筑工地采集到600张施工人员图像和60条施工视频作为验证集;根据建筑工地不同的视觉条件对图像进行分类,用于验证本文算法在不同外界环境下的性能。结果表明:改进后的模型检测速度是YOLOv4的3.4倍,可用于实时检测施工人员在不同施工现场条件下是否佩戴安全帽的情况,有利于提高安全检查和监督水平。     

水夹点技术用于石化行业水网络优化研究

运用水夹点技术,对某石化企业进行用水网络优化研究。在流量恒定和流量改变两种条件下,对某石化企业13个用水操作原有用水网络进行初始设计和优化改进。结果表明,在流量恒定和流量改变条件下,通过用水网络初始设计节约新鲜水用量分别为17.17%和13.72%,对初始设计用水网络的优化设计可简化用水网络,回用管路数分别由14条和16条减少为11条和8条,优化后的用水网络分别节约新鲜水用量17.25%和14.81%。水夹点技术在工业用水节约和废水减量方面具有实用价值和较好的应用前景。     

煤矿主提胶带运输系统远程自动化集中控制的工程实践

从煤矿生产作业环境的安全性出发,结合机械化、自动化技术的发展趋势,提出煤矿主提胶带运输系统远程自动化集中控制的技术方案,并以成功实例验证了方案的可操作性,节能效果显著。