关于改进《生态保护》教学方法的探索与实践

目前教育改革的目标是如何在教学过程中最大限度地调动学生的学习积极性，让学生能动地、创造性地学习，而不是被动地接受知识灌输。根据《生态保护》教学的特点，提出多媒体辅助教学、设计情景、实例分析等方法，开展讨论式教学，实现对新时期高职高专学生的素质教育。     

安钢90 m2烧结机尾除尘工艺设计特点

介绍了安钢90 m2烧结机尾除尘工艺设计中的设备选型、系统配置、管网阻力平衡的特点及应注意的几个问题.     

网络攻击检测中的机器学习方法综述

在详细介绍网络攻击检测机器学习原理的基础上,对现有的各种方法进行了评述,并结合网络攻击检测系统的应用需求,阐述了网络攻击检测系统机器学习技术的发展方向。     

Mapping epistemic cultures and learning potential of participants in citizen science projects

The ever‐widening scope and range of global change and interconnected systemic risks arising from people–environment relationships (social‐ecological risks) appears to be increasing concern among, and involvement of, citizens in an increasingly diversified number of citizen science projects responding to these risks. We examined the relationship between epistemic cultures in citizen science projects and learning potential related to matters of concern. We then developed a typology of purposes and a citizen science epistemic‐cultures heuristic and mapped 56 projects in southern Africa using this framework. The purpose typology represents the range of knowledge‐production purposes, ranging from laboratory science to social learning, whereas the epistemic‐cultures typology is a relational representation of scientist and citizen participation and their approach to knowledge production. Results showed an iterative relationship between matters of fact and matters of concern across the projects; the nexus of citizens’ engagement in knowledge‐production activities varied. The knowledge‐production purposes informed and shaped the epistemic cultures of all the sampled citizen science projects, which in turn influenced the potential for learning within each project. Through a historical review of 3 phases in a long‐term river health‐monitoring project, we found that it is possible to evolve the learning curve of citizen science projects. This evolution involved the development of scientific water monitoring tools, the parallel development of pedagogic practices supporting monitoring activities, and situated engagement around matters of concern within social activism leading to learning‐led change. We conclude that such evolutionary processes serve to increase potential for learning and are necessary if citizen science is to contribute to wider restructuring of the epistemic culture of science under conditions of expanding social‐ecological risk.     

Stakeholder participation in building resilience to disasters in a changing climate

The resilience perspective has emerged as a plausible approach to confront the increasingly devastating impacts of disasters; and the challenges and uncertainty climate change poses through an expected rise in frequency and magnitude of hazards. Stakeholder participation is posited as pivotal for building resilience, and resilience is not passive; rather, stakeholders are actively involved in the process of building resilience. Who is involved and how they are involved are crucial aspects for developing resilience in practice. Nevertheless, there are few empirical studies available to inform theory or show how these issues are addressed. This study focuses on revealing how practitioners frame the issue of participation in relation to resilience, its relevance to a changing climate and how, in consequence, they construct practices. Using Hajer's [(1995). The politics of environmental discourse: Ecological modernization and the policy process. New York] ‘Social-interactive discourse theory’, in this interdisciplinary research, we study the frames and subsequent practices developed around a disaster management policy initiative in Australia: the Natural Disaster Resilience Program in Queensland. What emerges from the research findings as critical and requiring urgent attention is stakeholder and especially local government and community participation, and for this to become socially relevant, challenges such as meaningful communication and power structures need to be addressed. What is also critical is to move from experiential learning to social learning. Additionally, the results presented here offer empirical evidence on how broadening the pool of actors can be implemented, and the opportunities that this opens up for building resilience.     

Experiencing Biodiversity as a Bridge over the Science–Society Communication Gap

Drawing on the idea that biodiversity is simply the diversity of living things, and that everyone knows what diversity and living things mean, most conservation professionals eschew the need to explain the many complex ways in which biodiversity is understood in science. On many biodiversity‐related issues, this lack of clarity leads to a communication gap between science and the general public, including decision makers who must design and implement biodiversity policies. Closing this communication gap is pivotal to the ability of science to inform sound environmental decision making. To address this communication gap, we propose a surrogate of biodiversity for communication purposes that captures the scientific definition of biodiversity yet can be understood by nonscientists; that is, biodiversity as a learning experience. The prerequisites of this or any other biodiversity communication surrogate are that it should have transdisciplinary relevance; not be measurable; be accessible to a wide audience; be usable to translate biodiversity issues; and understandably encompass biodiversity concepts. Biodiversity as a learning experience satisfies these prerequisites and is philosophically robust. More importantly, it can effectively contribute to closing the communication gap between biodiversity science and society at large. Experimentando la Biodiversidad como un Puente sobre el Vacío entre la Comunicación Ciencia‐Sociedad     

基于机器学习和情景分析的我国焦化行业企业用地污染预测

针对焦化行业企业用地缺乏时序连续监测数据而无法预测其污染趋势的问题,从企业特征、企业管理水平、污染物特征和自然地理要素等4个方面选取13个影响企业用地污染的指标,识别焦化行业企业用地污染主控因子,在此基础上构建基于机器学习的焦化行业企业用地污染预测模型,并在不同情境下,对2025年和2030年焦化行业企业用地污染状况进行预测.结果表明,生产经营活动时间、建厂时间、企业环境监管记录、土壤黏粒和年均风速是焦化行业企业用地污染的主控因子;相对于支持向量机模型、BP神经网络模型和决策树模型,逻辑斯蒂模型预测价值高、性能指标稳健,其预测精度受试者工作曲线面积为0.91,模型准确率和召回率分别为84%和88%.在乐观情境下,2025年和2030年焦化行业高概率污染地块数量分别为1599块和1695块;在悲观情境下,2025年和2030年焦化行业高概率污染地块数量分别为1671块和1715块.研究结果可为焦化行业企业用地的修复治理和生态环境的宏观决策提供科学依据.     

基于应力分解法的整机加速因子预计方法研究

提出了计算整机加速因子的新方法.将整机的失效率通过定量分析方法按照部件、失效模式和应力的顺序逐级下分,从而得到各敏感应力的失效率,再通过加速模型计算各失效模式在加速应力下的失效率.最终可以得到整机在加速应力下的失效率,进而计算出整机加速因子.以典型电子产品为例,预计了弹上计算机在温度、湿度、振动综合环境应力下的加速因子为153.5.     

一种应用于不均衡小样本集的石化设备故障智能诊断方法

针对实际生产中,石化设备故障呈现出“不均衡小样本”特性,致使传统诊断模型适用性不足,准确率低下这一问题,基于一类支持向量机(One-Class SVM)和遗传算法优化的支持向量机(GA-SVM),提出一种分步诊断策略。利用One-Class SVM构建超球诊断模型,可在没有任何故障先例的前提下实现异常状态辨识和未知故障甄别,提升算法对不均衡监测数据(正常样本数量远超故障样本数量)的适用性;利用GA-SVM,针对小样本故障数据集构建并优化智能诊断模型,判别故障模式,降低算法对大量有标签故障样本的依赖。故障诊断实验结果表明,该方法能够在不均衡小样本场景中显著降低漏报率、误报率和误诊率,对实验数据和工程数据的诊断准确率分别达99%和100%。     

融合时空特征的PCA-PSO-SVM臭氧(O3)预测方法研究

针对目前臭氧预测方法未能考虑臭氧污染的区域性和在时间周期内的强自相关性的问题,提出一种融合时空特征的PCA-PSO-SVM臭氧组合预测模型.利用小波分析和系统聚类提取臭氧时间序列波动特征和站点空间分布相似性特征,并通过主成分分析和粒子群算法优化的支持向量机组合模型（PCA-PSO-SVM）对臭氧日最大8h平均浓度进行预测,以2016~2018年杭州市大气污染物观测数据和气象数据进行实验验证.结果表明：融合时空特征的PCA-PSO-SVM模型预测精度有较大提升,与未融合时空特征的PCA-PSO-SVM模型相比,精度提升19%.气象因素中温度对臭氧预测效果影响最大,在气象预报数据存在一定误差的情况下,提出的模型仍得到较高精度的预测效果,具备较好的鲁棒性.