【作者】 王丹华；

【导师】 潘金贵；

【作者基本信息】 南京大学 ， 计算机应用技术， 2010， 博士

【摘要】 随着过程复杂性以及人们对过程安全性要求的逐渐提高,采用各种安全性分析技术来提高过程尤其是高风险过程(如医疗过程等)的安全性成为十分迫切的需求。故障树分析技术(Fault Tree Analysis,简称FTA)与故障模式及影响分析技术(Failure Mode and Effect Analysis,简称FMEA)是目前被广泛应用的两种互补的安全性分析技术。因此,本文提出将此两种技术应用于过程的安全性分析及改进,并且针对两种技术的手工分析方式中存在的耗时且易出错的缺点,分别对它们展开了自动化方法的研究。研究工作的具体内容包括：构建Little-JIL过程模型的方法、基于Little-JIL过程模型进行FMEA技术的自动化方法、基于Little-JIL过程模型进行FTA技术的自动化方法、以及基于FMEA与FTA两种安全性分析技术相结合的辅助过程安全性改进的方法。主要研究工作和贡献包括以下五个方面：1.提出一个基于Little-JIL过程模型的安全性分析及技术改进框架：在用Little-JIL语言对过程建立模型的基础上,采用各种静态分析技术对过程模型进行分析,并最终结合分析结果对过程提出改进措施。2.提出一种FMEA技术的自动化方法：在用Little-JIL语言为过程建模的基础上,实现对过程模型进行故障模式及影响分析的自动化,自动构建Little-JIL过程模型相应的影响树,通过对影响树的分析捕获过程中潜在的安全隐患。3.提出一种FTA技术的自动化方法：在用Little-JIL语言对过程进行建模的基础上,提出构建以给定故障事件为顶事件的故障树的自动化算法,并且能够自动计算故障树的最小割集,针对最小割集中的单因素故障事件提出在过程模型中增加一致性校验步骤的方法,提高过程安全性。4.针对FTA技术的自动化方法中的规模问题与循环缺陷分别提出优化策略：针对规模问题,提出的步骤抽象、步骤删除及单因素事件删除三种优化策略能够删除故障树中大量冗余结点,并保证优化前后故障树完全等价；针对循环缺陷,检测并删除故障树中所有循环,仅删除故障树分析结果中阶数很高的最小割集,提高故障树分析效率,节省时间空间开销。5.提出FMEA技术与FTA技术相结合的辅助过程安全性改进的方法：对于过程中的任意故障事件,采用故障模式及影响分析技术自底向上分析其在过程中形成的影响链,同时采用故障树分析技术自顶向下分析导致其发生的原因,为其构建相应故障树,结合两者分析结果提出过程安全性改进措施。最后,本文将提出的框架和自动化方法应用于医疗机构的两个实际过程分析。实践结果表明,本文所提出的方法可系统地分析过程中存在的漏洞与缺陷,辅助过程安全性措施的提出。本文的研究成果对于过程安全性分析和改进的理论与应用研究具有良好的参考价值和借鉴意义。更多还原

【Abstract】 Fault Tree Analysis (FTA) and Failure Mode and Effect Analysis (FMEA), two of the most widely used safety analysis techniques, can be combined together to analyze and evaluate safety problems in a process in an attempt to reduce the occurrence of severe hazards or their consequences. This paper focuses on applying these two techniques to process analysis, and finding out defects or weaknesses in processes, and finally leading to improved processes with higher safety. FMEA and FTA techniques are usually manually performed by skilled experts, which may be time-consuming, error-prone, and tedious when applied to complex processes. How to automatically perform FTA and FMEA techniques to analyze processes is the main objective of this paper. We also do research on finding an effective way to combine these two complementary techniques to analyze processes. The details of the above problems discussed in this paper are as follows,1. Framework of automatic safety analysis techniques for process analysis and improvement. We propose a framework that adopts several static analysis techniques to analysis weaknesses in processes after modeling them in Little-JIL language.2. Approach and tool support of an automatic FMEA technique. We propose an approach that automatically derives effect trees from Little-JIL processes.3. Approach and tool support of an automatic FTA technique. We propose an approach that automatically generates fault trees from Little-JIL processes, as well as computes MCSs of each fault tree.4. Optimization methods of addressing scaling and looping issues. To address the scaling issue, two optimizations are applied before the fault tree is derived, Step Abstract and Step Removal. Another optimization, Single-Cause Event Elimination, is applied after the fault tree is derived. To solve the looping issue, we propose an algorithm that can detect and remove all loops in the fault tree in a simple way.5. Approach of combining FTA and FMEA techniques. Both the failure mode and the effect in our automatic FMEA approach can be used as top-events in our automatic FTA approach to automatically generate fault trees for them, which realizes the combination of FMEA and FTA techniques. That is to say, given a failure mode, both effects (bottom-up) and causes (top-down) can be generated for it, and results can be combined to improve processes.Through applying the proposed framework and automatic techniques in analyzing two real-world medical processes which are adopted by Bayestate medical center in USA, it is proved that our approach can be used to systematically detect safety issues in processes. Subsequent improvement measures will be made to processes, which may lead to improved safety. Our research has directive significance for future work on safety analysis and improvement of processes both theoretically and in practice.更多还原