人工神经网络在大气环境质量评价中的应用

针对大气环境中的数据监测与模式识别问题,应用人工神经网络理论,在自然环境大气腐蚀试验网站建立大气环境质量B-P网络评价模型。江津试验站成功地应用了该模型。通过1996-2000年大气环境质量评价结果表明,人工神经网络用于大气环境质量评价的方法是一种新方法,也是一种可行的方法。     

基于BN的锅炉超压爆炸事故原因推理与故障诊断

分析了锅炉爆炸的原因,讨论了事故树与贝叶斯网络的特点,编制了以"锅炉超压爆炸事故"为顶事件的事故树,将其转化为锅炉超压爆炸事故贝叶斯网络拓扑结构,利用贝叶斯网络的技术优势,针对锅炉超压爆炸事故贝叶斯网络拓扑结构分别进行了原因推理与诊断推理。实例应用表明,在原因推理中,根节点V_1(压力上升)发生时锅炉超压爆炸事故发生的可能性最高;在诊断推理中,当锅炉超压爆炸事故发生时,根节点V_1(压力上升)一定发生,其次为V_4(定压不准),V_(11)(未装疏水管)发生的可能性最小,为企业进行锅炉超压爆炸事故预测与故障诊断提供了理论依据。     

基于贝叶斯网络模型的船舶搁浅事故分析

根据海上交通安全事故统计数据显示,船舶搁浅事故是主要的海上交通安全事故之一。为了研究船舶搁浅事故的发生机理,收集了我国东部海域部分海事局辖区的船舶搁浅事故样本。首先从人为因素、环境因素和船舶因素三方面分析导致船舶搁浅事故的原因,确定网络节点。然后找出每起搁浅事故的事故链,并建立船舶搁浅事故贝叶斯网络模型。再对建立的贝叶斯网络模型进行仿真,选取21起船舶搁浅事故对建立的模型进行验证,确定模型的有效性。最后,利用HUGIN软件找出船舶搁浅事故各影响因素的致因概率,得到导致船舶搁浅事故的致因链。在人为因素方面,瞭望不当和不熟悉航道情况占据的比例最高;在船舶因素方面,舵机设备故障对船舶搁浅事故影响最大;在环境因素方面,风/浪/流对船舶搁浅事故影响最大。     

煤尘润湿接触角的BP神经网络估算方法研究

通过采集和测定35个矿区煤样品的化学组成、结构参数和润湿接触角,构建了以13个影响因子为输入参数和以接触角为输出目标的3层BP人工神经网络,并利用该模型估算煤尘润湿接触角。结果表明,隐含层节点数为19时,接触角估算值与实测值的决定系数R2=0.957,平均相对误差为4.59%,表明基于BP神经网络建立的煤尘润湿接触角估算模型具有很高的精度。     

Elman神经网络在矿井突水水源判别中的应用

矿井突水是矿建与生产过程中最具威胁的自然灾害之一,准确判别突水水源是防治水害的关键。选取6种离子的质量浓度作为突水水源的判别因素,将河南省焦作矿区不同水层的39组水化数据以2种样本设计方案进行Elman神经网络模型的构建与检验。以不同的35组水源样品作为训练样本,运用Matlab软件进行Elman神经网络训练,将所建立的判别模型应用于(相应的)4组待测样本的判别,并与DDA、FDA、Bayes三种判别方法的判别结果进行分析比较。2种方案应用结果表明:将具有非线性动态特征的Elman神经网络应用于突水水源判别,在结合相应的水文地质条件前提下,可以准确判断突水来源;矿井多年的开采促使地下各水层水质呈动态变化,Elman神经网络判别模型能够反映这种变化特性,对探寻地下水运移与演化具有一定的应用价值。     

旧工业建筑再生利用项目施工安全的评价及改进

针对旧工业建筑再生利用项目施工过程中存在诸多安全隐患的问题,首先,结合此类项目遗留工业垃圾冗杂、施工场地狭窄、部分构件拆换、大型机械使用受限的施工特点从4MTE角度构建施工安全评价指标体系。然后,在考虑指标赋权的主客观综合性与评价工作的未确知性后,建立了基于结构熵权法与未确知测度理论的旧工业建筑再生利用项目施工安全评价模型,并通过工程实例验证该模型,得出的评价结果与实际情况相符。最后,结合未确知测度单因素指标测度函数的特点,提出了施工安全管理改进策略的制定方法。     

A practical dialogue protocol for sustainability science to contribute to regional resources management: its implementation in Réunion

Over the last decade, adaptive co‐management has been recommended as a policy framework to address complex and uncertain resources management issues. Implementing this theoretical management concept requires the integration of multidisciplinary research and local knowledge. Yet practical protocols to link science, policymaking and societies have yet to be developed. We designed a protocol to produce legitimate, credible and relevant solutions to a regional resources management issue. This is a two‐component protocol. A stakeholder grid categorizes stakeholder representatives in three distinct specialized dialogue arenas: institution representatives, technical experts and local end‐users. An iterative co‐design process then builds on these arenas to assess the institutional legitimacy, technical credibility and empirical relevance dimensions of a common solution initiated by an initial plausible promise. We tested this framework in Réunion to address organic waste management issues at the regional level. The plausible solution explored was the introduction of a recycling industry involved in collecting organic waste and producing and selling organic fertilizers tailored for local crop systems. The protocol application outcomes were consolidated and documented scenarios accepted by all, with knowledge exchange and the broad spread of a stabilized expectation contributing to private initiatives and public policy change.     

A framework for sustainable supply chain in emerging economies

Sustainable supply chains (SSC) are an essential topic in emerging economies for achieving sustainable development. Because of their rapid economic growth, emerging economies are experiencing both sides of this development: the well‐being enhancement and the environmental impact. However, decision makers and practitioners do not have robust models for managing supply networks that consider the sustainability complexities. This study aims to provide a better understanding of SSC in emerging economies and proposes a generic framework to improve the decision‐making process. Based on a review of the literature, a decision framework for sustainability is proposed. A green sustainable decision concept is also discussed in the SSC literature. This study has implications for scholars, decision makers, and practitioners interested in advancing the SSC field based on an extension of the framework.     

Road Erosion and Delivery Index (READI): A Model for Evaluating Unpaved Road Erosion and Stream Sediment Delivery

The Road Erosion and Delivery Index (READI) is a new geographic information system–based model to assess erosion and delivery of water and sediment from unpaved road networks to streams. READI quantifies the effectiveness of existing road surfacing and drain placements in reducing road sediment delivery and guides upgrades to optimize future reductions. Roads are draped on a digital elevation model and parsed into hydrologically distinct segments. Segments are further divided by engineered drainage structures. For each segment, a kinematic wave approximation generates runoff hydrographs for specified storms, with discharge directly to streams at road–stream crossings and onto overland‐flow plumes at other discharge points. Plumes are attenuated by soil infiltration, which limits their length, with delivery occurring if plumes intersect streams. Sediment production and sediment delivery can be calculated as a relative dimensionless index. READI predicts only a small proportion of new drains and new surfacing results in the majority of sediment delivery reductions. The model illustrates how the spatial relationships between road and stream networks, controlled by topography and network geometries, influence patterns of road–stream connectivity. READI was applied in seven northern California basins. The model was also applied in a recent burn area to examine how reduced hillslope infiltration can result in increased hydrologic connectivity and sediment delivery.     

京津冀化工场地地下水污染修复治理对策研究

京津冀化工场地地下水污染问题突出,严重威胁当地饮水安全和可持续发展,亟待开展修复治理.针对京津冀化工场地地下水污染现状,分析了化工场地地下水污染修复面临的挑战,提出了分区分级的修复治理对策.结果表明:①针对可能存在NAPL(非水相液体污染物)的高风险污染源区,采取高强度修复措施,以实现污染物总量的快速削减;②针对中度污染区,采取单位能耗强度更低的长效修复措施,降低修复成本和二次污染风险;③针对低风险的轻度污染区,采取风险管控措施.结合对典型化工场地地下水污染修复技术的分析,提出的分区分级修复治理对策具有以下特点:①多技术耦合,形成互补效应,可提高修复效率;②节约修复成本,降低二次环境影响;③体现基于风险的原则,避免过度修复.