利用卫星遥感数据估算PM2.5浓度的应用研究进展

近年来,PM_(2.5)已成为中国大气污染的首要污染物,危害人体健康。为弥补地基监测站点在空间分布上的局限性,借助卫星遥感技术估算PM_(2.5)浓度已成为研究热点。文章总结了利用卫星估算PM_(2.5)浓度的各种研究方法,探讨了不同方法的优势和不足,指出不同方法对不同应用目的的选择性差异较大。提出,应针对不同应用目的选择相应的方法,从而取得满足各方面需求的研究成果,为未来PM_(2.5)浓度估算应用工作提供参考。     

深部煤层钻屑粒度随钻进深度分布规律

为了研究煤层钻屑粒度随钻进深度分布规律，选取具有冲击危险性的平煤八矿己15煤作为研究对象，采取钻屑量测试和孔口瓦斯浓度监测，通过筛分实验煤样并应用Rosin Rammler分布模型，探究了钻屑粒度和钻屑量大小随孔深变化关系，分析不同点钻屑粒度分布特征。结果表明:小于0.075 mm钻屑粒度分布随孔深变化与钻屑量变化规律相吻合；不同钻孔随深度变化分别对应不同Rosin Rammler分布函数，随着应力过渡越平缓，粒径分布宽度系数n值越小，煤体应力越大粒径相关系数D越大；不同范围钻屑粒度占比大小也会影响钻屑量大小；在钻屑量较大时，孔口瓦斯体积分数会出现增高现象。通过对钻屑粒度分布规律分析，更好地了解深部煤体应力分布，有助于冲击危险的预警。     

Cycle Aware: Piloting a module for novice drivers

Introduction: In low-cycling countries, motor-vehicle traffic and driver behavior are well known barriers to the uptake of bicycles, particularly for utility cycling. Lack of separation between cyclists and faster-moving traffic is one key issue, while attitudes of drivers toward and/or harassment of cyclists is another. Cyclist-related driver education has been recommended as a means to improve driver-cyclist interactions. Methods: The driver licensing process provides an opportunity for such education. The Cycle Aware module was developed to test and enhance novice drivers’ knowledge of interacting safely with cyclists. It was piloted across three Australian jurisdictions targeting both novice and experienced drivers. Participants were asked to complete the Cycle Aware module and an accompanying survey. A total of 134 novice and 97 experienced drivers completed the survey with 42 novice and 50 experienced drivers going on to complete the module. Results: Both groups of drivers scored equally well in the module but the very youngest and very oldest participants were more likely to have some incorrect responses. We did not find any relationship between correct module scores and attitudes toward cyclists. Survey results showed both novice and experienced drivers had somewhat positive attitudes toward cyclists. The two cohorts differed on several attitude questions. Sixty percent (60%) of novices compared to 30% of experienced drivers reported feeling concerned when sharing the road with cyclists, and novices were less likely to agree that cyclists had a right to use the roads. Conclusions and practical applications: The analysis suggests novices need to be better equipped to share roads confidently with cyclists and to recognize cyclists as legitimate traffic participants.     

基于视频分析的空管员违规行为识别方法*

为利用视频数据对空管员违规行为进行智能化分析,降低不安全事件发生率,提出2阶段的违规行为识别模型（AR-ResNeXt）,基于实地调研构建空管员视频数据集,利用最小化动态多实例学习损失函数和中心损失函数,获得违规行为检测的判别特征表示,结合异常回归网络和ResNeXt网络,完成对空管员违规行为的时序区间检测与动作分类。研究结果表明:AR-ResNeXt模型在自制数据集中,其帧级AUC达到82.9%,分类准确率达到87.8%,可准确识别空管员发生违规行为的时序区间并进行分类,研究结果可为保障空中交通安全奠定基础。     

深埋水平岩层隧道开挖稳定性分析及控制*

针对深埋高地应力水平岩层掌子面开挖稳定性及支护结构失效问题,以大峡谷隧道为工程背景,通过现场测试、室内试验、数值模拟等方法,探究深埋高地应力水平岩层失稳机理及控制措施。研究结果表明:坚硬岩体被节理面切割后,在高地应力作用下容易发生挤压破碎,破碎岩体遇水发生软化,导致掌子面发生大范围塌方,初支和超前支护失效;隧道开挖后岩层发生不均匀沉降,浅部岩层最先发生弯折破坏,层内块体错动滑移,继而向上方岩层发展,并伴随层间分离和层内裂隙发育,最终形成宏观破裂面;提出的台阶法、2 m开挖进尺、砼喷层、双层小导管、提高初支强度的整体优化控制措施,可有效提高现场支护效果。     

Influencing diversity beliefs through a personal testimonial,promotion-focused approach

Although researchers have highlighted the importance of diversity beliefs (i.e., team members' perceived value of diversity) for the elaboration of information in teams, little attention has been paid to whether and how diversity beliefs can be shaped. Drawing on theory and research on team diversity beliefs, we propose that diversity beliefs are more effectively influenced by interventions using a promotion (compared with a prevention) focus toward diversity and personal testimonial (compared with factual) knowledge. Results from an experiment conducted with 175 teams revealed that both a promotion focus and personal testimonial knowledge independently contributed to more positive diversity beliefs and consequently increased team elaboration of task-relevant information as well as integration of different perspectives. Our results reveal key factors that can influence diversity beliefs and underscore the pivotal role of diversity beliefs in improving the extent to which team members elaborate information and integrate diverse perspectives.     

Machine learning based quantitative consequence prediction models for toxic dispersion casualty

Incidental release of toxic chemicals can pose extreme danger to life in the vicinity. Therefore, it is crucial for emergency responders, plant operators, and safety professionals to have a fast and accurate prediction to evaluate possible toxic dispersion life-threatening consequences. In this work, a toxic chemical dispersion casualty database that contains 450 leak scenarios of 18 toxic chemicals is constructed to develop a machine learning based quantitative property-consequence relationship (QPCR) model to estimate the affected area caused by toxic chemical release within a certain death rate. The results show that the developed QPCR model can predict the toxic dispersion casualty range with root mean square error of maximum distance, minimum distance, and maximum width less than 0.2, 0.4, and 0.3, which indicates that the constructed model has satisfying accuracy in predicting toxic dispersion ranges under different lethal consequences. The model can be further expanded to accommodate more toxic chemicals and leaking scenarios.     

Good acting requires a good cast: A meso‐level model of deep acting in work teams

This study proposes and tests a meso‐level model of deep acting in work teams that draws on emotional contagion theory to explain how shared means of complying with display rules can arise in work teams. We argue that the presence of influential deep actors can lead to greater convergence (lower dispersion) on individual deep acting in the team. That is, team members behave more similarly. When a team has greater convergence, deep acting by individual members should be related to lower emotional exhaustion and higher job satisfaction and in‐role performance. In a sample of mature work teams, these hypotheses received general support. Our findings suggested that team‐level deep acting effects can foster benefits for team members (lower emotional exhaustion and higher satisfaction) and organizations (higher job performance). Copyright © 2014 John Wiley & Sons, Ltd.     

基于SA-ELM的爆破震动速度预测

针对爆破震动速度与其影响因素之间的复杂非线性,结合模拟退火算法(SA)的全局寻优性,提出了一种新的SA-ELM算法.以矿山周边建筑物爆破震动实测数据作为训练样本,选取总药量、最大段药量、测点与爆破点距离、地面震动特性、建筑物震动特性等8个影响因素作为输入变量,建立了爆破震动速度预测的SA-ELM模型.模型训练值和预测值与实测值的均方误差(MSE)分别为0.20和3.26,平均相对误差控制在5％以内,显示出该模型具有良好的训练精度和泛化能力.对比传统ELM模型,SA-ELM模型不但提高了精度和泛化能力,而且降低了隐层节点数变化对训练结果的影响,提高了模型的稳定性.     

Adaptive co-management and network learning in the Room for the River programme

Adaptive co-management and learning are paramount for integrated flood risk management. Relevant literature focuses on adaptation at the level of physical and societal systems. The level of projects and programmes is largely overlooked, but they comprise interventions that adapt our physical systems and they provide opportunities for learning to contribute to transitions of societal systems. This paper aims to increase understanding on how learning takes place and can be stimulated within a programme. The mixed-method case study of Room for the River, a €2.3 billion programme for flood risk management, shows that a programme can be organised using various governance arrangements to stimulate learning and be a means for adaptive co-management to deliver upon environmental objectives.