基于ai智能图像的工控火电现场爆炸风险评估

对工控火电现场爆炸风险进行评估时，若采用人工识别现场图片信息的方法，容易导致现场图片特征信息采集不准确，存在评估耗时长、评估效率低和评估结果不准确的问题。针对该问题，提出1种基于ai智能图像的工控火电现场爆炸风险评估方法；通过自适应融合方法提取工控火电现场ai智能图像中的特征信息，根据特征信息对工控火电现场ai智能图像进行识别；结合层次分析法和问卷调查法，选取工控火电现场爆炸风险评估指标；在图像识别结果的基础上，通过风险等级集合、评估指标权重，建立工控火电现场爆炸风险评估模型，并与另外2种工控火电现场爆炸风险评估方法进行了对比。研究结果表明:所建方法能够缩短评估时间，且评估效率较高、评估结果准确率较高。     

A novel approach based on non-parametric resampling with interval analysis for large engineering project risks

The issue of risk assessment has been always the matter of debate in large engineering projects (LEPs). The assessment is an indispensable means for the projects to accomplish their objectives. It is firmly accepted that LEPs are particularly subject to more potential risks than other business activities because of their complexity, uncertainty and ambiguity. These characteristics are often conducive to small sample sizes of the gathered risk data in practice. Consequently, traditional statistical techniques cannot contribute significantly to analyze the risk data. The non-parametric resampling technique, namely bootstrap, has been used subsequently to solve numerous complicated problems and evaluate the accuracy of a parameter estimator in situations where commonly used techniques are not valid. It is also more natural, applicable and simple to estimate the risk data in an interval form under decision-making process by considering the concept of safety by professional experts in LEPs. Hence, in this paper, a new approach based on bootstrap technique with the interval analysis is presented in the context of the project risk assessment. The proposed approach not only plays an important role in reducing risk data and saving time but also is more economical. A real case study is conducted to illustrate the applicability of the approach. Finally, the comparison results indicate that the proposed approach outperforms the traditional technique in terms of the accuracy and efficiency.     

基于支持向量机的化工工艺安全评价模型构建及优化研究*

为解决现阶段基于风险分级的安全评价方法仍存在着高维数据处理不当、评价智能化程度不高等问题,创建支持向量机的安全评价模型,利用核函数解决安全评价因子分类问题,粒子群算法（PSO）寻找最适合模型的正则项C,进一步提升安全评价模型的正确率,形成适用高维数据的化工工艺安全评价方法。研究结果表明:该模型与经典支持向量机模型和BP神经网络评价模型相比具有更高的正确率,研究结果对借用机器学习来创新安全评价理论及工程应用具有现实意义及理论价值。     

Development of novel combustion risk index for flammable liquids based on unsupervised clustering algorithms

Current liquid flammability classification mainly relies on flash point and its risk is largely dependent on consequence and probability. However, combustions of liquefied marine fuels have their uniqueness, leading to a less consistent with the common classification. This work aims at classifying flammable liquids in compression ignition engines for further safety evaluation. Besides liquid flammability characteristics, flame propagation and aerosol formulation are considered. Two unsupervised machine learning clustering algorithms, k-means and spectral clustering, are applied to the collected liquid compounds database. To consider both cluster cohesion and separation, the global mean silhouette value is used to find the optimal number of clusters and to evaluate the clustering performance. The results show that the spectral clustering outperforms k-means on classifying the risk ratings for all proposed models, while the clustering accuracy of the optimal model has been doubled by employing spectral clustering algorithm. Moreover, principal component analysis and star coordinate diagrams are presented to visualize high dimensional data to 2-D graphs. Finally, the overall liquid safety performance is evaluated by a novel combustion risk index via the weight values determined by the information entropy approach. This index can be used to explore inherently safer fuels in the process industries.     

城市应急物流需求区域确定及等级划分方法

为提高应对突发事件的物资保障能力,提出基于联合国救灾署(UNDRO)灾害风险评估模型的应急物流需求区域风险预评估流程。采用应急物流需求总量,分类别、分事件需求量,需求密度,需求比例等指标进行物流需求测度,研究突发事件发生概率与需求扰动水平、需求密度、需求的自我供给水平相结合的3种需求区域等级划分方法。实例研究表明:区域4、3、1作为城市商业中心,人口密度高及消防隐患多,应急物流需求等级高于其他区域;城市外围区域9,由于工业基础较好,抗风险能力及自我修复能力高,需求等级低。作为城市的有机组成部分,各区域地质灾害事故的发生概率相差不大,城市管理者应关注火灾事故及安全生产事故对应急物流需求的重大影响。     

基于无人机的火灾检测系统设计与实现

针对传统森林火灾检测手段响应速度慢、效率低、误报率高等问题,设计了无人机搭载的由云台和相机组成的图像采集平台,通过火灾智能识别技术,实时识别监测火灾的发生,并达到了自动抵近侦察及实时态势感知的效果。在火灾智能识别算法方面,提出了improved-YOLOv3算法,通过在特征交互阶段增加yolo层,加强了网络对特征的融合度,从而增加了网络的检测能力。通过与性能相似的网络进行对比测试,验证了改进算法的有效性。测试结果表明,提出的算法检测准确率高、漏检率低、推理速度快,能够适用于实际火灾现场监测。     

1种基于视频的油田危险区域入侵检测智能综合识别技术研究

为解决油田日常生产作业中缺乏危险作业区域的等级划分与自动识别方式以及缺乏人员踏入危险区域的识别方法.提出1种基于视频智能综合识别技术的全天油田危险区域入侵检测算法,该算法首先结合油田危险因素对油田危险区域进行危险等级的划分与危险区域的识别;然后,针对光照条件良好的白天场景,在训练数据集中融合油田作业区视频数据和公开行人...     

基于ICEEMDAN和MC-CNN的矿山声发射信号识别分类方法

为精准识别地下矿山声发射事件,采用基于改进的完全集合经验模态分解模型(ICEEMDAN)和多通道卷积神经网络(MC-CNN)模型对声发射信号进行处理后得到分量图,根据各通道输入分量峭度值赋予不同权重,并利用卷积神经网络对输入数据进行训练,最终采用五折交叉实验方法验证该分类识别方法的可行性及有效性.结果表明:基于ICEE...     

Risk analysis in Natech events: State of the art

Natech events (Natural Hazard Triggering Technological Disasters) are industrial accidents caused by natural events such as hurricanes, floods, earthquakes, tsunamis, and so on. In recent decades, the probability of these events occurring has increased, activating the interest of researchers in the study of new methods of risk analysis to prevent and mitigate possible damage to people, the environment, and processing facilities. On the other hand, the concept of multi-hazard is summarized in the combination of two or more threat factors manifested in isolated, simultaneous manner, or by chain reaction, to produce a trigger event of a disaster, where hazardous events can be one or more natural events. Considering that, it is essential to know the progress in risk analysis for Natech events, to identify the gaps for future research. Therefore, in this paper, a systematic review of the Natech events literature with single and multi-hazard approaches was developed. The review was conducted by searching the Science Direct, Web of Science, and Scopus databases for scientific documents. Subsequently, the words Natech and Multi-hazard were taken as keywords, and 208 results were obtained. Then, some management documents were consulted in international organizations to compare academic literature and industrial risk management. In conclusion, the risk analysis methods revised are specific to a particular hazard and apply mainly to earthquakes, floods, and lightning. Regarding a multi-hazard approach, the methods focus on risk mitigation in urban areas without taking into account Natech risk. In the case of industrial risk assessment, some methodologies were found that briefly consider Natech risk in risk assessment processes in industry.     

井下高温作业场所危害程度分级标准研究

对我国热害分级相关标准进行了简要总结分析,表明我国尚未制定综合考虑温湿度条件和体力劳动强度从而对井下高温作业场所的热害程度实施有效分级的相关标准.由井下现场管理人员、企业劳资(人事)管理人员、研究人员等3类共30名相关专家对井下62类主要工种劳动强度进行主观分级,通过对其结果的统计分析,研究制定了井下高温作业场所不同工种最高体力劳动强度级别建议表.通过定性和定量分析,确定了井下作业场所不同劳动强度级别下 WBGT指数限值,将结果与南非井下热害分级标准限值进行了对比.最终制订了基于井下高温作业场所最高体力劳动强度级别和WBGT指数的井下高温作业场所危害程度分级标准,提出了标准的宣传贯彻建议及分级监管策略.     

基于风险修正的城市安全风险评估方法及应用*

为推动实现城市安全风险的系统化、信息化管理，基于我国城市安全风险管理的要求与特点，总结城市安全风险评估的分层分类原则、突出固有风险原则和合理修正原则等基本原则；提出“点位-行业-区域”逐层展开的城市安全风险评估程序，明确各层级的评估要点；提出城市安全风险评估方法，分别制定点位、行业风险修正规则，结合风险参数与情景构建开展基于风险矩阵法的点位风险评估和行业风险评估，采用加权计算法由行业风险评估结果叠加评估区域风险。结果表明:该方法能够从不同角度有效评估城市的风险状态，为不同层级的城市安全管理者明确风险管控重点提供决策依据。     

Risk assessment of chemical process considering dynamic probability of near misses based on Bayesian theory and event tree analysis

With the development of modern automatic control systems, chemical accidents are of low frequency in most chemical plants, but once an accident happens, it often causes serious consequences. Near-misses are the precursor of accidents. As the process progresses, near misses caused by abnormal fluctuation of process variables may eventually lead to accidents. However, variables that may lead to serious consequences in the production process cannot update the risk in the life cycle of the process by traditional risk assessment methods, which do not pay enough attention to the near misses. Therefore, this paper proposed a new method based on Bayesian theory to dynamically update the probability of key variables associated with process failure risk and obtain the risk change of the near-misses. This article outlines the proposed approach and uses a chemical process of styrene production to demonstrate the application. In this chemical process, the key variables include flow rate, liquid level, pressure and temperature. In order to study the dynamic risk of the chemical process with consideration of near misses, according to the accumulated data of process variables, firstly the abnormal probability of the variables and the failure rate of safety systems associated with the variables were updated with time based on Bayesian theory. On the basis of the dynamic probability of key process variables, an event tree of possible consequences caused by variable anomalies was established. From the logical relationship of the event tree, the probability of different consequences can be obtained. The results show that the proposed risk assessment method based on Bayesian theory can overcome the shortcomings of traditional analysis methods. It shows the dynamic characteristics of the probability of different near misses, and achieves the dynamic risk analysis of chemical process accidents.     

Job level risk assessment using task level ACGIH hand activity level TLV scores: a pilot study.

Existing upper extremity musculoskeletal disorder analytical tools are primarily intended for single or mono-task jobs. However, many jobs contain more than 1 task and some include job rotation. This case/control study investigates methods of modifying an existing tool, the American Conference of Governmental Industrial Hygienists (ACGIH) Hand Activity Level (HAL) Threshold Limit Value (TLV), to assess the upper extremity risk of multi-task jobs. Various methods of combining the task differences and ratios into a job level assessment were explored. Two methods returned significant odds ratios, (p < .05) of 18.0 (95% CI 1.8-172) and 12.0 (95% CI 1.2-120). These results indicate that a modified ACGIH HAL TLV may provide insight into the work-related risk of multi-task jobs. Further research is needed to optimize this process.     

基于深度学习的输煤皮带故障视觉检测方法研究

为了更好地检测皮带跑偏、撕裂和异物干扰等严重影响皮带安全运行的故障状态,围绕相关问题产生的原因及检测方法开展深入研究,通过对纵/横向裂缝、异物的检测分析、实验,提高基于视觉的检测精度。提出基于Canny边缘检测算法的皮带跑偏检测算法;基于深度学习的横向与纵向撕裂检测,尤其对于裂缝与纵向纹理区分不明显情况,提出一种红光透射的判别方式;基于最小距离分类算法将识别异物转换为分类问题,利用机器学习的方法对样本进行训练并建立无异物阈值,通过提取特征,最后利用最小距离分类算法得到有无异物的结果。研究结果表明:提出的视觉检测系统可以实时高效地检测出输煤皮带常见的3种故障,可进一步保障运输系统安全运行。     

智能管道时代的管道风险评价技术展望

当前管道风险评价工作普遍使用半定量方法,通过对该方法的应用现状进行总结分析,得出目前半定量风险评价方法存在的问题,包括录入数据准确性不足、评价结果准确性不高,时效性不强等。通过分析物联网、智能移动终端、云技术、大数据等技术开始在管道的应用前景,展望了智能管道时代管道风险评价发展方向:实现管道风险评价实时动态评价,提高风险评价结果的准确性,实现风险的预测预警等,为未来管道风险评价技术发展提供研究思路。     

A Classification System for Characterization of Physical and Non-Physical Work Factors

A comprehensive evaluation of work-related perfomance factors is a prerequisite to developing integrated and long-term solutions to workplace performance improvement. This paper describes a work-factor classification system that categorizes the entire domain of workplace factors impacting performance. A questionnaire-based instrument was developed to implement this classification system in industry. Fifty jobs were evaluated in 4 different service and manufacturing companies using the proposed questionnaire-based instrument. The reliability coefficients obtained from the analyzed jobs were considered good (.589 to .862). In general, the physical work factors resulted in higher reliability coefficients (.847 to .862) than non-physical work factors (.589 to .768). performance work-factor system classification     

我国危险化学品产业转移及安全风险分析*

为深入分析我国国内（不含港澳台,下同）危险化学品产业转移及安全风险,通过调研产业转移现状,统计典型地区产业政策、转移企业数量、转移时间及安全生产形势,分析产业转移带来的安全风险。研究结果表明:危险化学品产业转移呈现明显地域趋势,从以江苏、浙江为主的东部沿海地区向中西部等地区转移;转移时间集中在2个阶段,分别是:2002—2008年、2017—2020年;转移集中在精细化工、炼焦、煤化工等产业,医药等高端产业未发生转移;危险化学品产业转移安全风险主要表现在:“重经济,轻安全”现象突出、转移项目安全风险高、承接地基础薄弱、人才储备不足等方面;针对呈现的产业转移安全风险,提出相关对策措施及建议。研究结果可为我国危险化学品产业转移安全风险管控政策措施制定提供相关数据与理论支撑。     

Drunk driving detection based on classification of multivariate time series

IntroductionThis paper addresses the problem of detecting drunk driving based on classification of multivariate time series.MethodsFirst, driving performance measures were collected from a test in a driving simulator located in the Traffic Research Center, Beijing University of Technology. Lateral position and steering angle were used to detect drunk driving. Second, multivariate time series analysis was performed to extract the features. A piecewise linear representation was used to represent multivariate time series. A bottom-up algorithm was then employed to separate multivariate time series. The slope and time interval of each segment were extracted as the features for classification. Third, a support vector machine classifier was used to classify driver's state into two classes (normal or drunk) according to the extracted features.ResultsThe proposed approach achieved an accuracy of 80.0%.Conclusions and practical applicationsDrunk driving detection based on the analysis of multivariate time series is feasible and effective. The approach has implications for drunk driving detection.     

Emission factor estimation for oil and gas facilities

Fugitive emission rate quantification in an oil and gas facility is an important step of risk management. There are several studies conducted by the United States Environmental Protection Agency (USEPA) and American Petroleum Institute (API) proposing methods of estimating emission rates and factors. Four major approaches of estimating these emissions, in the order of their accuracy, are: average emission factor approach, screening ranges emission factor approach, USEPA correlation equation approach, and unit-specific correlation equation approach. The focus of this study is to optimize the USEPA correlation equations to estimate the emission rate of different units in an oil and gas facility. In the developed methodology, the data available from USEPA (1995) is used to develop new sets of equations. A comparison between USEPA correlation equations and the proposed equations is performed to define the optimum sets of equations. It is observed that for pumps, flanges, open-ended lines, and others, the proposed developed equations provide a better estimation of emission rate, whereas for other sources, USEPA equations supply the better estimate of emission rate.     

基于HMM和DT的无人机异常检测方法

为了实时检测无人机异常状态,提出基于隐马尔可夫模型(Hidden Markov Model,HMM)和决策树(Decision Tree,DT)的无人机异常检测方法(HMMDT).首先根据异常致因将无人机异常分为干扰异常和硬件异常;然后结合HMM和DT建立无人机异常检测模型,定义无人机异常度衡量异常状态的严重程度,确定...