突发环境事件风险源识别与监控技术创新进展——（I）环境风险源识别技术与应用

环境风险防控与环境风险源管理是我国当前环境管理的重要工作之一。为了防控重大环境污染事件、实现环境风险源有序管理、科学决策,开展了重大环境污染事件风险源识别监控与管理技术的研发创新。借鉴国外环境风险源管理经验,遵循环境风险源“分类、分级、分区”管理原则,研发了环境风险源分类、识别与定量分级方法,实现了典型化工园区、特大城市及全国尺度的环境风险源分级,初步提出了全国石化化工行业重点风险企业名单。开发了环境风险源综合管理系统,以期为重点风险源监管、环境风险防控提供技术支持。     

基于层次分析法的有毒作业场所职业危害程度分级指标研究

我国目前职业危害形势依然严峻,急、慢性职业中毒案例时有发生。对可能引起急、慢性职业中毒的有毒作业场所,按照其职业危害严重程度实施分级、分重点监管,成为我国职业危害监管工作需要解决的迫切需求。实施有毒作业场所分级监管的关键,是确定合理的分级指标,从而将有毒作业场所区分开。本文应用层次分析法,对有毒作业场所的各项评价指标进行分析,并最终确定毒物浓度超标倍数作为最终的分级指标,从而为制定有毒作业场所分级标准,开展有毒作业场所分级监管提供科学的依据。     

基于BERT-BILSTM-CRF模型的电力行业事故文本智能分析*

为解决电力行业事故报告文本较长、语义复杂,难以进行有效文本识别问题,提出1种以BERT作为底层的预训练模型,并设计1种双重注意力机制编码器,结合BILSTM-CRF深度挖掘事故文本语义特征,从而实现文本智能分析。首先构建电力词典,通过对BERT预训练,进行BIO标注,然后引入BILSTM-CRF模型实现对文本标签智能分类,最后将该模型与现行其他4种深度学习模型进行对比。研究结果表明:该模型智能识别精确率、召回率及F1值(查准率)均达到约97%,较其他4种模型中效果最好的模型分别提高0.02,0.03,0.02。研究结果可为电力行业事故报告文本分析提供1种新思路。     

油气长输管道作业分级定量风险评价方法研究及应用

为客观掌握油气管道生产运营岗位主要作业风险,实现油气长输管道企业作业风险管理,提出一种作业分级定量风险评价方法。首先,辨识油气管道系统主要危险作业类型和引起作业事故的危险有害因素;然后由作业分级法确定事故发生的可能性与严重性;最后建立基于作业分级的风险矩阵,得到各作业活动的风险级别。将该方法应用于某管道公司,以该公司生产区内部高处作业为例,对作业过程中的危险有害因素进行识别与评价,得出高处作业过程中各步骤的风险程度。结果表明该评价方法具有一定的可操作性,可反映各类危险作业过程中的风险程度,为后期有针对性的风险控制提供指导,有助于公司的风险管理。     

供水管网震后流量监测点的动态分级优化布局研究

为了提高管网地震监测点布局的准确性和合理性,基于管网微观水力计算模型和动态分级法,提出供水管网震后流量监测点的动态分级优化布局模型。首先,利用管网微观水力计算模型计算管段流量的影响系数,构建管段的影响系数矩阵,并利用信息熵确定管段权重;其次,标准化处理影响系数矩阵,通过聚类迭代提出供水管网地震流量监测点优化布局的动态分级方法,对供水管网震后流量监测点进行优化布置分级评定;最后,根据工程实例进行方法实践,结果表明:供水管网中的管线分类较为科学合理,地震监测点在供水管网上分布也比较均匀,而且该模型在一定程度上消除了人为因素的影响,保障了震时管网的监控效果和日常建设的合理性。     

基于随机森林的突发事件分级模型研究

突发事件的分级是快速响应和有效应对的基础,为了解决目前突发事件分级宽泛、主观性强和动态适应性弱等问题,将多分类器集成引入突发事件的分级过程中,借助随机森林组合分类器,针对突发事件特征,构建突发事件分级的随机森林模型,形成突发事件分级过程,进而对事前的应急预案制定、事中的应急资源配置和应急决策提供有力支撑;最后,以2014-2016年洪涝灾害突发事件为例检验构建的模型和过程。研究结果表明:与支持向量机方法相比,通过检验随机森林组合分类器分类结果准确率达到97.56%,在突发事件分级的应用上是可行的,进而可为突发事件的快速响应和应急决策提供信息支撑和参考依据。     

关于危险化学品重大危险源分级的研究

在重大危险源死亡半径分级法的基础上,引入经济密度和人员密度,并给出不同情况下相应的计算方法。然后以＂最大危险原则＂和＂概率求和原则＂为财产损失和人员伤亡计算的原则,以《重大危险源分级标准（征求意见稿）》中的分级标准为标准,对危险化学品重大危险源进行分级,这样进行危险源分级时,可将周边环境中可能波及到的财产损失和人员伤亡考虑进去,使得危险化学品重大危险源分级考虑的因素更加全面,分级的结果和实际情况更加吻合。本文的研究结果对危险化学品重大危险源分级具有一定的参考价值。     

建筑安全风险因素分级排序的风险矩阵法

为减少风险分级排序时产生的风险结,在引入施工安全事故频率的基础上,采用单元格细化和Borda序值法改进原有的风险矩阵法。改进后的方法使得风险等级数量增加,风险结数量减少。将该方法应用于脚手架工程的风险因素的分级与排序,风险等级分为6级,且实现了12个因素的明确排序。     

The Aging of America's Reservoirs: In‐Reservoir and Downstream Physical Changes and Habitat Implications

Reservoirs are important for various purposes including flood control, water supply, power generation, and recreation. The aging of America's reservoirs and progressive loss of water storage capacity resulting from ongoing sedimentation, coupled with increasing societal needs, will cause the social, economic, environmental, and political importance of reservoirs to continually increase. The short‐ and medium‐term (<50 years) environmental consequences of reservoir construction and operation are well known and include an altered flow regime, lost connectivity (longitudinal, floodplain), an altered sediment regime, substrate compositional change, and downstream channel degradation. In general, reservoir‐related changes have had adverse consequences for the natural ecosystem. Longer term (>50 years) environmental changes as reservoirs enter “old” age are less understood. Additional research is needed to help guide the future management of aging reservoir systems and support the difficult decisions that will have to be made. Important research directions include assessment of climate change effects on aging and determination of ecosystem response to ongoing aging and various management actions that may be taken with the intent of minimizing or reversing the physical effects of aging.     

Improving the Urban Stream Restoration Effort: Identifying Critical Form and Processes Relationships

Stream restoration projects are often based on morphological form or stream type and, as a result, there needs to be a clear tie established between form and function of the stream. An examination of the literature identifies numerous relationships in naturally forming streams that link morphologic form and stream processes. Urban stream restoration designs often work around infrastructure and incorporate bank stabilization and grade control structures. Because of these imposed constraints and highly altered hydrologic and sediment discharge regimens, the design of urban channel projects is rather unclear. In this paper, we examine the state of the art in relationships between form and processes, the strengths and weaknesses of these existing relationships, and the current lack of understanding in applying these relationships in the urban environment. In particular, we identify relationships that are critical to urban stream restoration projects and provide recommendations for future research into how this information can be used to improve urban stream restoration design. It is also suggested that improving the success of urban restoration projects requires further investigation into incorporating process-based methodologies, which can potentially reduce ambiguity in the design and the necessity of using an abundant amount of in-stream structures.