基于深度学习的复杂作业场景下安全帽识别研究

为有效预防由于个人防护缺失所造成的事故,着力探究复杂作业情况下施工人员安全帽佩戴情况的智能化识别.提出在Faster R-CNN目标检测算法的基础上,针对小目标的安全帽识别问题通过增加锚点提升检测能力,为解决数据集中类别不平衡问题采用Focal loss替代原本的损失函数,为解决安全帽预测区域不匹配问题,引入ROI A...     

海洋钻井平台事故致因分析与评价指标研究

该文在广泛调研国内外已发生的海洋平台事故基础上,从人因失误、工艺设备、物料危险、管理缺陷和危险环境五个方面系统分析事故致因因素,提出海洋钻井平台事故致因评价框架,提出5个一级评价指标和24个二级评价指标,根据层次分析法,计算确定各指标的标准权重值.最后针对某在役海洋钻井平台,开展实例研究.为实际评价海洋钻井平台事故风险提供一种新方法,具有一定的实际应用价值.     

海上钻井平台安全风险预警模型应用研究

2010年美国墨西哥湾“深水地平线”钻井平台爆炸和2011年我国蓬莱钻井平台井涌事故,引发社会各界对海上钻井平台安全问题的关注.介绍了海洋钻井平台的组成结构,对海洋钻井平台的危险源进行了系统辨识,试图通过对海上钻井平台作业现场危险因素加以观察,诊断、分析警源、警情,警兆,警级,结合专家意见,从人员、设施设备、安全管理、工程地质、海上环境等几个方面建立海上钻井平台安全风险预警体系,采用改进的层次分析法,得出指标权重,构建海洋钻井平台安全风险预警模型,得到量化的预警结果.并通过对某钻井平台的实例分析,提出安全风险的预控对策,以期本模型对海上钻井平台的事故的发生起到一定的防范作用.     

自升式钻井平台就位安全性与覆盖能力要求分析*

为保证自升式钻井平台安全就位并覆盖海上生产平台井槽进行钻完井作业,研究自升式钻井平台就位及插桩作业过程中的风险及应对措施,提出考虑就位作业误差影响下的自升式钻井平台覆盖能力要求计算方法。结果表明:桩腿穿刺是自升式钻井平台就位安全的首要考虑因素;就位作业误差会明显提高对自升式钻井平台覆盖能力的要求;提出的方法可以准确计算出实际作业对自升式钻井平台横向和纵向覆盖能力的要求,对自升式钻井平台选型及安全作业具有一定的指导意义。     

Allocation of performance shaping factors in the risk assessment of an offshore installation

Offshore oil production is one of the most important human productive activities. There are many risks associated with the process of constructing a subsea well, pumping oil to the platform, and transporting it to refineries via underwater pipes or oil tankers. All actions performed by workers in those operations are influenced by specific working conditions, involving the use of complex systems. Contextual factors such as high noise, low and high temperatures and hazardous chemicals are considered to be contributors to unsafe human actions in accident analysis and also give a basis for assessing human factors in safety analysis. Some failure modes are particularly dangerous and can result in severe accidents and damage to humans, the environment and material assets. Fires and explosions on oil rigs are some of the most devastating types of offshore accidents and can result in long-term consequences. The most typical root causes related to accidents include equipment failure, human error, environmental factors, work organization, training and, communication, among others. The principal objective of this study is to propose a methodological framework to identify the factors that affect the performance of operators of an offshore unit for oil processing and treatment. In this phase, an ergonomics approach based on operators' work analysis is used as a supporting tool. After identification of factors that affect the performance of operators, a decision-making model based on AHP (analytic hierarchy process) is applied to rank and weight the principal performance shaping factors (PSFs) that influence safe operations. The next step involves the use of the SHELLO model to group the main PSFs in elements named software, hardware, environment, liveware and organization. In the last phase, a relevant accident that occurred aboard a floating production storage and offloading (FPSO) vessel is analyzed. The allocation process of the factors that affect the operator's performance in risk assessment was developed through fuzzy logic and the ISO 17776 standard.     

Evacuation, escape, and rescue experiences from offshore accidents including the Deepwater Horizon

When a major hazard occurs on an installation, evacuation, escape, and rescue (EER) operations play a vital role in safeguarding the lives of personnel. There have been several major offshore accidents where most of the crew has been killed during EER operations. The major hazards and EER operations can be divided into three categories; depending on the hazard, time pressure and the risk influencing factors (RIFs). The RIFs are categorized into human elements, the installation and hazards. A step by step evacuation sequence is illustrated. The escape and evacuation sequence from the Deepwater Horizon offshore drilling platform is reviewed based on testimonies from the survivors. Although no casualties were reported as a result of the EER operations from the Deepwater Horizon, the number of survivors offers a limited insight into the level of success of the EER operations. Several technical and non-technical improvements are suggested to improve EER operations. There is need for a comprehensive analysis of the systems used for the rescue of personnel at sea, life rafts and lifeboats in the Gulf of Mexico.     

基于人脸信息的信息安全技术研究现状及应用前景展望

讨论了国内基于人脸谱信息的信息安全技术研究总体现状 ,包括人脸谱信息的检测、定位、特征提取及主要识别方法和算法 ,并对各种方法的优点和问题进行了评述 ,最后对基于人脸谱信息的信息安全技术的应用前景进行了分析和预测。     

基于FLACS的海上钻探平台井喷爆炸事故后果模拟分析*

为研究海上钻探平台井喷燃爆事故后果,运用FLACS软件对某深海钻探平台井喷爆炸事故进行模拟,研究在不同事故场景下气云爆炸发展过程及平台荷载分布规律,讨论井喷速率、风向、点火位置等对爆炸超压的影响。研究结果表明:随泄漏速率增加,爆炸强度和爆炸范围均增大,爆炸严重程度不仅与井喷速率密切相关,同时也受平台结构影响;点火位置会对爆炸超压产生影响,在可燃气体与空气混合气体比例为化学理论当量比处点燃气体,生活区承受的爆炸超压最大;在设施及建构筑物分布较为密集、拥塞度较高的地方产生的爆炸超压更大。研究结果可为可为平台的阻隔防爆性能设计与应急响应提供指导。     

复杂背景环境下基于SSD_MobileNet深度学习模型的火焰图像识别研究

针对现有的火灾火焰图像识别方法在光照和红花等类似火焰干扰的复杂环境下存在错检和漏检的问题, 提出一种基于SSD_MobileNet的复杂环境火焰区域标记方法。首先,将深度卷积神经网络SSD300的基础卷积网络VGG16替换为MobileNet网络,应用深度可分离卷积,降低网络参数,进而构建一种火焰图像检测的SSD_MobileNet模型;然后,迁移第一次训练模型所有的卷积层参数,初始化新的待训练模型;最后,加入新的数据样本用于削弱光照、红花等干扰对象的影响。通过与SSD300、以及深度学习的目标检测算法Faster R-CNN和YOLOv3-tiny对比,实验结果表明,构建的火焰检测和火焰区域标记SSD_MobileNet模型的综合性能优于Faster R-CNN和YOLOv3-tiny模型,更适用于实时火焰检测领域。     

基于VR技术的机场应急救援虚拟演练平台

为解决机场实地应急演练中存在的成本消耗高、不安全、扩展性差等问题,研发基于VR技术的机场应急救援虚拟演练平台。平台采用模块化设计方案,并通过定义平台操作流程支持虚拟演练操作。采取分细节度的场景构建方法,运用细节层次模型以3层不同细节度模型方案将虚拟灾害场景划分为3类,同时利用Unity3D粒子系统控制机制实现火灾及灭火粒子特效的模拟,运用双顶层本体模型构建机场应急预案本体模型（AERPO）,并利用XML Spy建立存储模型,且支持基于典型场景模型编辑形成整个灾害场景。结果表明:平台模块设置科学、场景效果合理,有助于提升演练人员应急处置能力。     

支持向量机法在煤与瓦斯突出分析中的应用研究

通过分析采煤工作面煤与瓦斯涌出量与地质构造指标的对应关系,应用支持向量机(SVM)方法对煤与瓦斯涌出类型及涌出量进行分析。建立两类突出识别的SVM模型、多类型突出识别的H-SVMs模型以及预测瓦斯涌出量的支持向量回归模型。研究结果表明:SVM方法能够很好地对煤与瓦斯突出模式进行识别,所建立的采煤工作面瓦斯涌出量预测模型的精度高于应用BP神经网络预测精度;SVM理论基础严谨,决策函数结构简单,泛化能力强,并且决策函数中的法向量W可以反映突出模式识别的地质结构指标的权重。     

Does the potential safety risk affect whether particular construction hazards are recognized or not?

Introduction: Evidence from the global construction industry suggests that an unacceptable number of safety hazards remain unrecognized in construction workplaces. Unfortunately, there isn’t a sufficient understanding of why particular safety hazards remain unrecognized. Such an understanding is important to address the issue of poor hazard recognition and develop remedial interventions. A recent exploratory effort provided anecdotal evidence that workers often fail to recognize safety hazards that are expected to impose relatively lower levels of safety risk. In other words, the research demonstrated that the underlying risk imposed by a safety hazard can affect whether a hazard will be recognized or not. Method: The presented research focused on empirically testing this preliminary finding. More specifically, the study tested the proposition that Construction workers are more likely to recognize safety hazards that impose higher levels of safety risk than those that impose relatively lower levels of safety risk. The research goals were accomplished through a number of steps. First, a set of 16 construction case images depicting a variety of construction operations that included a number of known safety hazards was presented to a panel of four construction safety experts. The experts were tasked with examining each of the known safety hazards and providing a rating of the relative safety risk that the individual hazards impose. Having obtained an estimate of the underlying safety risk, a hazard recognition activity was administered to 287 workers recruited from 57 construction workplaces in the United States. The hazard recognition activity involved the examination of a random sample of two construction case images that were previously examined by the expert panel and reporting relevant safety hazards. Results: The results of the study provided support for the proposition that workers are more likely to recognize hazards that impose relatively higher levels of safety risk. Practical Applications: The findings of the study can be leveraged to improve existing hazard recognition methods and develop more robust interventions to address the issue of poor hazard recognition levels.     

附属管线对隔水管涡激振动影响的数值模拟研究

涡激振动现象会引起海上钻井平台隔水管发生结构失稳和疲劳破坏。针对附属管线安装位置对隔水管的涡激振动影响开展了三维数值模拟,采用CFD大涡模拟计算方法,对不同来流角度α、直径比d/D和间距比G/D情况进行数值模拟,得到隔水管所受的升力和阻力变化情况并对其进行频谱分析。模拟结果表明:隔水管所受的升力和阻力对来流角度的变化很敏感,阻力系数CD和升力系数均方根CrmsL随来流角α的增大而减小,并在α值为30~45°范围内降至最低;用来表征隔水管后方漩涡脱落频率的St数随α的增大略有增大,但总体稳定保持在较低值范围。根据模拟结果,附属管线在直径比d/D值应在0.2~0.3范围内,间距比G/D值在0.3~0.4范围内的布置方式对隔水管涡激振动的抑制效果最为理想。     

基于系统动力学的海洋平台安全脆弱性分析

海洋平台承受着风、浪、流等复杂环境载荷，安全系统在保障平台安全稳定方面的作用极为突出。为深入探讨影响海洋平台安全脆弱性的机理，针对海洋平台作业环境的极端性和复杂性，建立了海洋平台的安全脆弱性系统动力学分析模型，采用模糊层次分析法确定影响安全系统中各层因素的权重系数，并设定参数，考虑子系统不同的安全投入方案，进行了海洋平台脆弱性系统分析。研究结果表明:海洋平台设备系统对于暴露度和敏感度的影响程度最高，管理系统对海洋平台安全系统的适应度起主导作用；当安全投入总值一定，对各系统的投入比例不同时，为有效降低海洋平台脆弱性，应合理分配有限投入资金。     

A safety barrier-based accident model for offshore drilling blowouts

Blowout is one of the most serious accidents in the offshore oil and gas industry. Accident records show that most of the offshore blowouts have occurred in the drilling phase. Efficient measures to prevent, mitigate, and control offshore drilling blowouts are important for the entire offshore oil and gas industry. This article proposes a new barrier-based accident model for drilling blowouts. The model is based on the three-level well control theory, and primary and secondary well control barriers and an extra well monitoring barrier are established between the reservoir and the blowout event. The three barriers are illustrated in a graphical model that is similar to the well-known Swiss cheese model. Five additional barriers are established to mitigate and control the blowout accident, and event tree analysis is used to analyze the possible consequence chains. Based on statistical data and literature reviews, failures of each barrier are presented. These failures can be used as guidance for offshore drilling operators to become aware of the vulnerabilities of the safety barrier system, and to assess the risk related to these barriers. The Macondo accident is used as a case study to show how the new model can be used to understand the development of the events leading to the accident. The model can also be used as an aid to prevent future blowouts or to stop the escalation of events.     

Determination of human error probabilities for offshore platform musters

The focus of this work is on prediction of human error probabilities during the process of emergency musters on offshore oil and gas production platforms. Due to a lack of human error databases, and in particular human error data for offshore platform musters, an expert judgment technique, the Success Likelihood Index Methodology (SLIM), was adopted as a vehicle to predict human error probabilities. Three muster scenarios of varying severity (man overboard, gas release, and fire and explosion) were studied in detail. A panel of 24 judges active in the offshore oil and gas industry provided data for both the weighting and rating of six performance shaping factors. These data were subsequently processed by means of SLIM to calculate the probability of success for 18 muster actions ranging from point of muster initiator to the final actions in the temporary safe refuge (TSR). The six performance shaping factors considered in this work were stress, complexity, training, experience, event factors and atmospheric factors.     

基于SAPSO-BP网络模型的海洋平台落物碰撞损伤分析

为了确保海洋平台作业过程中落物坠落对甲板撞击后的结构安全,对坠落立管撞击海洋平台甲板的过程进行了非线性仿真分析。运用ANSYS/LS-DYNA建立有限元模型,对坠落立管不同撞击角度撞击海洋平台甲板进行模拟计算,得到甲板在不同工况下的损伤情况。运用1种自适应性变异的粒子群优化算法SAPSO与BP神经网络结合［1］,对海洋平台进行落物碰撞损伤分析。研究结果表明:落物与垂直方向偏离5°～10°为坠落时最危险的工况。SAPSO-BP提高了BP神经网络的拟合能力,减小了拟合误差,提高了拟合精度,验证了SAPSO-BP网络模型的实用性和可靠性。综合考虑制定适用于海洋平台落物安全的工作程序和平台及设备的防护措施,为海洋平台作业中落物风险评估和海洋平台作业安全保障提供参考。     

海洋平台密集作业空间内H_2S扩散分析

针对含硫油气田开发生产中,硫化氢(H_2S)在振动筛房等密集作业空间内泄漏扩散造成的潜在人员毒害问题,从确定H_2S扩散规律、建立事故控制措施的角度展开分析.根据海洋平台空间环境特点,采取计算流体动力学(CFD)建立用于分析H_2S扩散过程的理论模型和数值模拟方案.就某海洋平台振动筛房在舱门关闭、打开和房内安装换气机3种条件下的H_2S扩散过程进行分析,确定H_2S浓度变化和分布规律.通过对比不同舱室条件下的H_2S影响程度,推荐在平台生产过程中封闭可能出现H_2S泄漏的舱室并安装换气机.进一步分析了换气机的工作效率,建立了简化数学模型以快速预测换气机工作时间,并与CFD结果进行了验证.     

基于RBI技术的海上平台延期服役静设备定量化安全评估技术

针对海上延期服役平台静设备的安全风险评估与控制问题,应用RBI风险评估技术的评估理论和方法对海上延期服役平台静设备的定量化安全管理方法进行了实践和探索,并选择某延期服役平台进行风险分析相关结论的工程验证。其具体过程是运用定量RBI技术,首先根据工艺流程和介质特点对海上平台静设备分别进行了物流回路和损伤回路的划分,并以此为基础对延期服役平台静设备的安全风险进行定量评估,进而得出各静设备单元的失效可能性、失效后果及风险等级,最后通过得出的损伤模式、损伤机理及对应的腐蚀速率,从而制定针对性的检测策略。将利用这些检验策略所检测出的真实情况与分析结论进行对比,最终验证了RBI技术在延期服役平台静设备风险评估方面具有很强的适用性。     

海洋延寿平台随机冰激响应与损伤评估

直立结构上冰力作用是国际海洋石油工程界研究的一个热点,尽管很多学者建立了各种冰力破碎模型,但是这些理论模型实现对冰与直立结构交互作用过程的量化模拟还存有较多困难.针对胜利油田大量海上平台超期服役现状,选择渤海现场实测的随机冰力时程曲线,考虑腐蚀减薄、桩基冲刷、海生物附着等服役损伤,建立海洋延寿平台结构有限元模型,研究随机冰力作用下海洋延寿平台构件的时程响应,采用疲劳等效应力幅来描述随机冰激疲劳应力过程产生的疲劳效应,最后基于Miner线性累积疲劳损伤模型采用S-N曲线法评估平台构件的疲劳损伤,该方法为进行可适合工程应用的渤海延寿平台疲劳评估提供了思路.