Analysis on causes of chemical industry accident from 2015 to 2020 in Chinese mainland: A complex network theory approach

Chemical manufacturing is a long-process industry, where an end product may pass through numerous dangerous and complex steps. In such long chains of coordinated activity, accidents remain common. This study made loss-prevention recommendations for the chemical industry after conducting a review of accident reports and creating a complex network model. A human factor analysis and classification system (HFACS) was used to classify data from 109 investigation reports from the Chinese mainland (2015–2020). Levels Ⅱ and Ⅲ of the HFACS output were fed into a complex network model to generate a map of causes and chains of risk. It was shown that most accidents were directly or indirectly caused by human action, and human factors played a decisive role in occurrence, evolution, and resolution. The model used was visualized in Gephi, and the key cause nodes were identified by their topological characteristics. A modularity algorithm was used to derive the community structures and segment the network map. Crucial nodes in each community were compared with factors for each class in the HFACS model. It was also found that there was a biasing factor in the causal processes of explosive accidents and poisoning and asphyxiation accidents according to the associations classified by modularity. Risk abatement strategies were proposed for the crucial factors.     

Revising HFACS for SMEs in the chemical industry: HFACS-CSMEs

Chemical accidents have occurred frequently in recent years, and most have occurred in small and medium-sized enterprises (SMEs). SMEs in the chemical industry face greater challenges than large enterprises with regard to accident prevention. However, SMEs have been unable to effectively learn from accidents due to the limited resources. The accident causation model is an effective tool to help the analyst learn from accidents. As a systematic accident causation model, the causes classification in the human factors analysis and classification system (HFACS) can match the characteristics of SMEs, but the cause of chemical accidents can be ineffectively identified by HFACS. In this study, HFACS was revised for the SMEs in the chemical industry, mainly consisting of three parts. First, based on the definition of factors in the original HFACS, the extended HFACS framework was obtained, which include 78 manifestations with the characteristics of the chemical accidents. Second, 101 accidents occurring in a SME in the chemical industry from 2012 to 2016 were analyzed though the extended HFACS framework. Finally, a new model, known as the HFACS-CSMEs, was obtained by further revising the manifestations and causes classification according to the statistical results of the accident analysis. HFACS-CSMEs consists of 15 cause factors and 56 manifestations, which can effectually identify and distinguish the causes in chemical accidents. Moreover, the easy-to-understand and statistically acceptable features of HFACS-CSMEs can cater to the SMEs regarding accident analysis. HFACS-CSMEs solves the problem that HFACS cannot be directly applied to chemical accidents and provides new ideas about preventing accidents in SMEs in the chemical industry.     

高处坠落事故的人因失误与干预策略研究

为系统分析导致高处坠落人因事故的产生机理,通过统计152起建筑工程高处坠落事故的调查与分析报告,从组织影响、安全监管、不安全行为前提条件和不安全行为等4个层次,辨识影响高处坠落事故的人为失误因素,修订人为因素分析与分类系统框架（HFACS）。设计高处坠落人因失误调查问卷,开展一线高处作业人员问卷调查,建立高处坠落人因失误结构方程模型,对导致高处坠落事故的人为失误因素进行路径分析。结果表明:各潜在因素间均呈正相关,且高处坠落人因失误事故的关键路径为资源管理不到位→安全监督培训不充分→班组管理不良→操作违规。综合各因素间相关性,提出了针对性的预防高处坠落事故的人因干预策略。     

基于TEM模型的民航不安全事件关联规则挖掘与分析*

为了预防民航不安全事件的发生,应用机组威胁与差错管理（TEM）模型分析2014—2020年民航事故/征候的航空安全报告资料,提取事件里存在于民航运行风险中潜在的情况、威胁、机组差错等因素,通过改进的关联规则方法挖掘其中的关联关系,包括挖掘与事件严重程度有关的因素,找到TEM模型中的关键因素和影响航空器结束状态的致因因素,并进行关联网络图分析。研究结果表明:手动操纵/飞行控制差错、缺少/不足的飞行培训和安全管理、飞行员之间沟通差错与程序执行错误是造成事故/征候的显著因素;关联规则能够有效利用航空安全报告信息,通过定量的方法挖掘事故/征候的特征,找到影响民航不安全事件的强关联因素,为民航安全管理人员提供决策依据。     

基于HFACS的化工企业火灾爆炸事故人因分析*

为研究化工企业火灾爆炸事故的主要人为因素,以63起火灾爆炸事故案例为样本,构建人为因素分析与分类系统（HFACS）模型,进行火灾爆炸事故人为因素分类统计与分析,并利用卡方检验和比值比分析HFACS模型上下层级间的因果关系。结果表明:HFACS模型中上下层级人为因素间存在显著的因果关系,层级1中的“不良的组织氛围”及“组织过程漏洞”和层级2中“监督不充分”在HFACS模型中可以显著增大事故发生的可能性,且“资源管理漏洞”、“不良的组织氛围”、“组织过程漏洞”→“监督不充分”→“人员因素”→“违规”是引发事故的关键路径,并根据HFACS模型中引发事故的关键路径及其人为因素的主要表现形式,提出针对性的化工企业火灾爆炸事故预防措施。     

建筑安全事故人为因素分类研究

为系统研究导致建筑安全事故的人为因素及对策,首先,在文献分析和专家访谈的基础上结合建筑行业特征提出人为因素分类分析系统（HFACS）框架中应增加社会环境层,在修正框架层次和人为因素的基础上构建建筑安全事故人为因素分类分析系统（C-HFACS）框架;其次,对150起建筑安全事故进行案例分析,探讨C-HFACS框架中对事故影响显著的人为因素及其内在关联性,验证了构建C-HFACS框架的合理性;最后,得出政府监管等九个人为因素对事故和下层人为因素影响显著,并从政府、企业、现场和个体四个维度提出独立第三方“飞行式”巡检等有针对性的对策,以期为建筑安全事故分析和管理提供新的方法和工具。     

Developing the understanding of underlying causes of construction fatal accidents

A study of a small sample of construction fatal accidents was used to pilot a potential method of analysis for the UK Health and Safety Executive (HSE) Construction Division. The work was conducted in the context of a Governmental Inquiry into construction fatalities in 2009, but has been developed further since. The sample of 26 accidents (28 fatalities) was drawn from the 211 fatal accidents in the years ending 2006–2008, to be broadly illustrative of the range of accident characteristics. The accidents were analysed on the basis of available inspectorate reports and structured interviews with the investigating inspectors. A standard method of classification on four levels was developed, based on the Human Factors Analysis and Classification System (HFACS) classification of errors and task level factors, with additional categories covering the organisational and regulatory/market levels of the system. The results showed a concentration of underlying factors associated with inadequacies in planning and risk assessment, competence assurance, hardware design, purchase and installation, and contracting strategy. These findings were partially validated by comparing them with another sample of 50 accidents analysed earlier by the HSE. This paper describes the development and testing of the investigation and classification method and how it is being further developed since the initial study. It also provides a summary of the findings relating to underlying causes.     

基于贝叶斯网络的煤矿事故人因推理

煤矿事故的不可重现性决定了事故原因的调查具有很强的不确定性,如何通过事故发生后的相关信息提高事故深层次原因调查的准确性是非常重要的。将HFACS与贝叶斯网络相结合,以煤矿事故HFACS分析结果为样本,通过卡方检验和让步比分析建立人因的贝叶斯网络因果图,进一步利用最大似然估计算法确定了煤矿事故人因的贝叶斯网络参数。最后,以双柳煤业顶板事故的调查信息为证据推理导致煤矿事故发生的深层次原因,提高事故原因调查的准确性,从而验证模型的有效性。     

航空人为差错事故/事件分析(ECAR)模型研究

为深入研究航空人为差错事故/事件的影响因素,以人为差错相关理论为基础,对比分析几种典型的人为差错分析模型;通过借鉴ECCAIRS分析框架,并在基元事件分析(EEAM)逻辑和CCAR396部的分类方法基础上,构建航空人为差错事故/事件分析(ECAR)模型,它从事件层、描述层、原因层和组织因素与改进建议层,分析航空事故和不安全事件的人为差错。此外,还将组织因素概念引入该模型。     

水利工程高危作业人因评价体系中的认知图理论研究

为加强水利工程施工安全管理和预防事故发生,通过建立认知地图对高危作业中的人为因素进行了分析。在修订的人为因素分析与分类系统(HFACS)对高危作业人为因素进行识别的基础上,通过引入认知地图理论,建立了人为因素的因果认知地图并进行头节点、尾节点和中心性分析;提出认知广度和认知深度概念,通过数理统计各人为因素出现的频数和采用G1法求其各因素权重,建立二维平面认知地图,对事故的关键影响因素进行了识别和分析。结果表明:教育培训、作业环境和人员健康状况是导致高危作业事故最核心因素,该方法能有效对水利工程人为因素进行分析研究,具有一定适用性。     

Analysis of train derailments and collisions to identify leading causes of loss incidents in rail transport of dangerous goods in Canada

The Canadian railway industry has improved safety performance in the last decade as measured by freight loss incidents per billion gross ton-miles. Further improvements in safety performance require a deeper analysis of the leading causes to identify weaknesses in implementing safety systems. In this paper, we classify the causes of railway loss incidents using a Safety Management System (SMS) framework to identify system weaknesses. The role of human factors is further analyzed through the Human Factors Analysis and Classification System (HFACS) approach. For this, we utilized data from 42 main track derailments and collisions involving the transport of dangerous goods in Canada between 2007 and 2018, which have been investigated by the Transportation Safety Board of Canada in detail. Associations between adjacent sub-categories of the HFACS framework are analyzed to identify any interdependency that exists between active and latent errors using a Chi-square test and Kruskal's lambda analysis. Furthermore, we implement the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method and the Analytical Network Process (ANP) to identify causal relationships between different sub-categories of the HFACS framework and calculate the weighted influence of each sub-category on main track derailments and collisions. Finally, a comparison is made between this work and others', which have analyzed human factors in the railway industry. There is good agreement between the results of these studies that highlight the importance of supervisory and organizational factors in the prevention of railway loss incidents. Based on these findings, we make recommendations to reduce railway loss incidents.     

基于HERA-JANUS模型的空管人误认知分析

空管人误分类分析是空管人误研究的基础。为了对管制员人误进行系统的分类研究,结合空管业务知识和认知心理学理论,对欧洲航空安全局和美国联邦航空局合作开发的HERA-JANUS模型的工作原理和流程进行较详细地分析。运用该方法模型,对我国一起空管不安全事件案例进行分析后得到3个由管制员所产生的人误差错,并对这3个人误差错分别从人误类型、人误认知、相关因素3方面进行详尽的分析研究,最后得出该不安全事件的21项人误结果。结果表明,HERA-JANUS模型能较全面地从深层次分析管制员的人误,其分类形式也便于开展空管人误统计。     

基于HFACS的空中相撞事故分析及建议

空中相撞事故往往是由诸多人为差错相互叠加、耦合和作用而导致的,要找出事故的真正诱因,防止类似事故再次发生,难度非常大。为了有效地分析和定位人为差错,以更好地服务于防相撞的管理与决策,提出一种基于人为因素分析分类系统（HFACS）的空中相撞事故分析方法,它按照从显性差错到隐性差错的思路来分析事故的诱因,最终找出组织因素对事故的影响。并利用HFACS对巴西卡欣布上空发生的一起空中相撞事故进行了系统分析。案例分析结果表明,该方法不仅能够找出导致空中相撞事故的人为差错,解释事故发生的原因和过程,而且能够据此提供防止相撞事故发生的安全建议。     

CREAM-based communication error analysis method (CEAM) for nuclear power plant operators’ communication

Communication error has been considered a primary cause of many incidents and accidents in the nuclear industry. In order to prevent these accidents, a method for the analysis of such communication errors is proposed here. This paper presents a qualitative and a quantitative method to analyze communication errors. The qualitative method focuses on finding a root cause of the communication error and predicting the type of communication error which could happen in nuclear power plants. We develop context conditions and antecedent-consequent links of influential factors related to communication errors. The quantitative analysis method focuses on estimating the probability of communication errors. To accomplish the quantification of communication errors, the Cognitive Speaking Process (CSP) is defined and a method to estimate the weighting factors and the probability is suggested. Finally, case studies conducted to validate the applicability of the proposed methods are detailed. From the results, we can foresee the effects of given plant conditions on communication errors and reduce the error occurrences.     

HFACS中不安全监管因素与事故致因24 Model中不安全动作因素的对比分析

为了使HFACS这种事故分析方法能够在我国各种类型事故中得到应用,研究了HFACS的不安全监管因素中各指标的定义和分类,并将其与事故致因24 Model中的不安全动作因素进行对比,得出两者的对应关系和各自特点。结果表明:HFACS中多数不安全监管因素属于24 Model中的不安全动作,这些不安全动作的发出者都是监管者;24 Model中不安全动作的发出者既可以是事故的直接引发者,也可以是监管者。从监管的范畴来看,HFACS中不安全监管因素的范围小于24 Model中的监管范围,只包括组织内部的监管。     

Cognitive failures, driving errors and driving accidents

INTRODUCTION: The impact of a driver's cognitive capability on traffic safety has not been adequately studied. This study examined the relationship between cognitive failures, driving errors and accident data. METHOD: Professional drivers from Iran (160 males, ages 18-65) participated in this study. The cognitive failures questionnaire (CFQ) and the driver error questionnaire were administered. The participants were also asked other questions about personal driving information. A principal component analysis with varimax rotation was performed to determine the factor structure of the CFQ. Poisson regression models were developed to predict driving errors and accidents from total CFQ scores and the extracted factors. RESULTS: Total CFQ scores were associated with driving error rates, but not with accidents. However, the 2 extracted factors suggested an increased effect on accidents and were strongly associated with driving errors. DISCUSSION: Although the CFQ was not able to predict driving accidents, it could be used to identify drivers susceptible to driving errors. Further development of a driving-oriented cognitive failure scale is recommended to help identify error prone drivers. Such a scale may be beneficial to licensing authorities or for developing driver selection and training procedures for organizations.     

Origin of equipment design and operation errors

The paper discusses the origin of chemical process equipment accidents by analyzing past accident cases available in the Failure Knowledge Database (FKD). The design and operation errors of the process equipment that caused the accidents were analyzed together with their time of occurrence. It was found that design errors contributed to 79% of accidents while the rest were only due to human and organizational errors in the operation stage and external factors. The most common types of errors were related to layout, organizational errors in the operation stage, considerations of reactivity and incompatibility, and wrongly selected process conditions (each approx. 13% of total accident contributors). On average there were about 2 design errors per accident. The timing of the errors was quite evenly distributed between various lifecycle stages. Nearly half (47%) of the errors were made in process design-oriented stages, one fourth (26%) in detailed engineering, and one fifth (20%) in operation. In addition, the most frequent design and operation errors for each equipment type were identified. A points-to-look-for list was created for each equipment type, showing also the typical time of occurrence of the error. The knowledge of type and timing of design errors can be utilized in design to focus the hazard analysis in each stage on the most error-prone features of design.     

航空维修差错分析及其管理

航空维修差错是诱发或直接导致飞行事故最重要的原因之一 ,对维修差错进行分类和分析有助于航空安全。笔者在分析航空维修环境变化的基础上 ,基于Reason模型构建了维修差错分类与诱因分析的框架 ,并结合机务维修的实际情况 ,对框架所包括的不安全行为、不安全行为的先兆、不安全的管理及组织因素进行了初步编码。文章还简要论述了维修差错的管理技术 ,指出借助框架编制详细的差错分类与分析编码系统是发展的方向。     

自愿报告信息分析模型（CRIAM）研究

为深入分析挖掘自愿报告信息中蕴含的人为因素,在借鉴SHEL模型、REASON模型、HFACS模型、ECCAIRS模型的基础上,结合自愿报告信息内容的特点,采用上行和下行分类并行的思想构建了自愿报告信息分析模型,并运用该模型对我国航空安全自愿报告信息进行分析和对比。统计结果显示:执行不力、经验及情景意识缺乏、违规、运行计划或组织过程、压力与应激所占比重较大,对此,笔者分别从个体和组织两方面为有效实施风险管理提供了改进建议。该研究成果有助于获取重要的人为因素信息、识别出可能存在的安全隐患,为更好地改善航空安全提供了一种分析方法和思路。