Yes / Safety and reliability are two important aspects of dependability that are needed to be rigorously evaluated throughout the development life-cycle of a system. Over the years, several methodologies have been developed for the analysis of failure behavior of systems. Fault tree analysis (FTA) is one of the well-established and widely used methods for safety and reliability engineering of systems. Fault tree, in its classical static form, is inadequate for modeling dynamic interactions between components and is unable to include temporal and statistical dependencies in the model. Several attempts have been made to alleviate the aforementioned limitations of static fault trees (SFT). Dynamic fault trees (DFT) were introduced to enhance the modeling power of its static counterpart. In DFT, the expressiveness of fault tree was improved by introducing new dynamic gates. While the introduction of the dynamic gates helps to overcome many limitations of SFT and allows to analyze a wide range of complex systems, it brings some overhead with it. One such overhead is that the existing combinatorial approaches used for qualitative and quantitative analysis of SFTs are no longer applicable to DFTs. This leads to several successful attempts for developing new approaches for DFT analysis. The methodologies used so far for DFT analysis include, but not limited to, algebraic solution, Markov models, Petri Nets, Bayesian Networks, and Monte Carlo simulation. To illustrate the usefulness of modeling capability of DFTs, many benchmark studies have been performed in different industries. Moreover, software tools are developed to aid in the DFT analysis process. Firstly, in this chapter, we provided a brief description of the DFT methodology. Secondly, this chapter reviews a number of prominent DFT analysis techniques such as Markov chains, Petri Nets, Bayesian networks, algebraic approach; and provides insight into their working mechanism, applicability, strengths, and challenges. These reviewed techniques covered both qualitative and quantitative analysis of DFTs. Thirdly, we discussed the emerging trends in machine learning based approaches to DFT analysis. Fourthly, the research performed for sensitivity analysis in DFTs has been reviewed. Finally, we provided some potential future research directions for DFT-based safety and reliability analysis.

Fault tree analysis, FTA, is one of the most commonly used techniques for safety system assessment. Over the past five years the Binary Decision Diagram (BDD) methodology has been introduced which significantly aids the analysis of the fault tree diagram. The approach has been shown to improve both the efficiency of determining the minimal cut sets of the fault tree, and also the accuracy of the calculation procedure used to quantifY the top event parameters. To utilise the BDD technique the fault tree structure needs to be converted into the BDD format. Converting the fault tree is relatively straightforward but requires the basic events of the tree to be placed in an ordering. The ordering of the basic events is -critical to the resulting size of the BDD, and ultimately affects the performance and benefits of this technique. There are a number of variable ordering heuristics in the literature, however the performance of each depends on the tree structure being analysed. These heuristic approaches do not always yield a minimal BDD structure for all trees, some approaches generate orderings that are better for some trees but worse for others. Within this thesis three pattern recognition approaches, that of machine learning classifier systems, multi-layer perceptron networks and radial basis function neural networks, have been investigated to try and select a variable ordering heuristic for a given fault tree from a set of alternatives. In addition a completely new heuristic based on component structural importance measures has been suggested with significant improvement in producing the smallest BDD over those methods currently in the literature.

Small Unmanned Aerial Vehicles (SUAVs) are rapidly being adopted in the National Airspace (NAS) but experience a much higher failure rate than traditional aircraft. These SUAVs are quickly becoming complex enough to investigate alternative methods of failure analysis. This thesis proposes a method of expanding on the Fault Tree Analysis (FTA) method to a Bayesian Belief Network (BBN) model. FTA is demonstrated to be a special case of BBN and BBN can allow for more complex interactions between nodes than is allowed by FTA. A model can be investigated to determine the components to which failure is most sensitive and allow for redundancies or mitigations against those failures. The introduced method is then applied to the Virginia Tech ESPAARO SUAV. / Master of Science

Continuous-time Markov chains (CTMC's) form a convenient mathematical framework for analyzing random systems across many different disciplines. A specific research problem that is often of interest is to try to predict maximum probability sequences of state transitions given initial or boundary conditions. This work shows how to solve this problem exactly through an efficient dynamic programming algorithm. We demonstrate our approach through two different applications - ranking mutational pathways of HIV virus based on their probabilities, and determining the most probable failure sequences in complex fault-tolerant engineering systems. Even though CTMC's have been used extensively to realistically model many types of complex processes, it is often a standard practice to eventually simplify the model in order to perform the state evolution analysis. As we show here, simplifying approaches can lead to inaccurate and often misleading solutions. Therefore we expect our algorithm to find a wide range of applications across different domains.

Continuous-time Markov chains (CTMC's) form a convenient mathematical framework for analyzing random systems across many different disciplines. A specific research problem that is often of interest is to try to predict maximum probability sequences of state transitions given initial or boundary conditions. This work shows how to solve this problem exactly through an efficient dynamic programming algorithm. We demonstrate our approach through two different applications - ranking mutational pathways of HIV virus based on their probabilities, and determining the most probable failure sequences in complex fault-tolerant engineering systems. Even though CTMC's have been used extensively to realistically model many types of complex processes, it is often a standard practice to eventually simplify the model in order to perform the state evolution analysis. As we show here, simplifying approaches can lead to inaccurate and often misleading solutions. Therefore we expect our algorithm to find a wide range of applications across different domains.

In 2004, the NASA Astronaut Office produced a memo regarding the safety of next generation launch vehicles. The memo requested that these vehicles have a probability of loss of crew of at most 1 in 1000 flights, which represents nearly an order of magnitude decrease from current vehicles. The goal of LOC of 1 in 1000 flights has since been adopted by the launch vehicle design community as a requirement for the safety of future vehicles. This research addresses the gap between current vehicles and future goals by improving the capture of vehicle architecture effects on reliability and safety. Vehicle architecture pertains to the physical description of the vehicle itself, which includes manned or unmanned, number of stages, number of engines per stage, engine cycle types, redundancy, etc. During the operations phase of the vehicle life-cycle it is clear that each of these parameters will have an inherent effect on the reliability and safety of the vehicle. However, the vehicle architecture is typically determined during the early conceptual design phase when a baseline vehicle is selected. Unless a great amount of money and effort is spent, the architecture will remain relatively constant from conceptual design through operations. Due to the fact that the vehicle architecture is essentially “locked-in” during early design, it is expected that much of the vehicle's reliability potential will also be locked-in. This observation leads to the conclusion that improvement of vehicle reliability and safety in the area of vehicle architecture must be completed during early design. Evaluation of the effects of different architecture decisions must be performed prior to baseline selection, which helps to identify a vehicle that is most likely to meet the reliability and safety requirements when it reaches operations. Although methods exist for evaluating reliability and safety during early design, weaknesses exist when trying to evaluate all architecture effects simultaneously. The goal of this research was therefore to formulate and implement a method that is capable of quantitatively evaluating vehicle architecture effects on reliability and safety during early conceptual design. The ConcepTual Reliability Growth Approach for CompariSon of Launch Vehicle ArchiTectures (CONTRAST) was developed to meet this goal. Using the strengths of existing techniques a hybrid approach was developed, which utilizes a reliability growth projection to evaluate the vehicles. The growth models are first applied at the subsystem level and then a vehicle level projection is generated using a simple system level fault tree. This approach allows for the capture of all trades of interest at the subsystem level as well as many possible trades at the assembly level. The CONTRAST method is first tested on an example problem, which compares the method output to actual data from the Space Transportation System (STS). This example problem illustrates the ability of the CONTRAST method to capture reliability growth trends seen during vehicle operations. It also serves as a validation for the development of the reliability growth model assumptions for future applications of the method. The final chapter of the thesis applies the CONTRAST method to a relevant launch vehicle, the Space Launch System (SLS), which is currently under development. Within the application problem, the output of the method is first used to check that the primary research objective has been met. Next, the output is compared to a state-of-the-art tool in order to demonstrate the ability of the CONTRAST method to alleviate one of the primary consequences of using existing techniques. The final section within this chapter presents an analysis of the booster and upper stage block upgrade options for the SLS vehicle. A study of the upgrade options was carried out because the CONTRAST method is uniquely suited to look at the effects of such strategies. The results from the study of SLS block upgrades give interesting observations regarding the desired development order and upgrade strategy. Ultimately this application problem demonstrates the merits of applying the CONTRAST method during early design. This approach provides the designer with more information in regard to the expected reliability of the vehicle, which will ultimately enable the selection of a vehicle baseline that is most likely to meet the future requirements.

Continuous-time Markov chains (CTMC's) form a convenient mathematical framework for analyzing random systems across many different disciplines. A specific research problem that is often of interest is to try to predict maximum probability sequences of state transitions given initial or boundary conditions. This work shows how to solve this problem exactly through an efficient dynamic programming algorithm. We demonstrate our approach through two different applications - ranking mutational pathways of HIV virus based on their probabilities, and determining the most probable failure sequences in complex fault-tolerant engineering systems. Even though CTMC's have been used extensively to realistically model many types of complex processes, it is often a standard practice to eventually simplify the model in order to perform the state evolution analysis. As we show here, simplifying approaches can lead to inaccurate and often misleading solutions. Therefore we expect our algorithm to find a wide range of applications across different domains.

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第17885号 / 工博第3794号 / 新制||工||1580(附属図書館) / 30705 / 京都大学大学院工学研究科航空宇宙工学専攻 / (主査)教授 藤本 健治, 教授 泉田 啓, 教授 椹木 哲夫 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Real-world processes often undergo improvements to meet certain goals, such as coping with changed requirements, eliminating defects, improving the quality of the products, and reducing costs. Identifying and evaluating the defects or errors in the process, identifying the causes of such defects, and validating proposed improvements all require careful analysis of the process.Human-intensive processes, where human contributions require considerable domain expertise and have a significant impact on the success or failure of the overall mission, are of particular concern because they can be extremely complex and may be used in critical, including life-critical, situations. To date, the analysis support for such processes is very limited. If done at all, it is usually performed manually and can be extremely time-consuming, costly and error-prone.There has been considerable success lately in using static analysis techniques to analyze hardware systems, software systems, and manufacturing processes. This thesis explores how such analysis techniques can be automated and employed to effectively analyze life-critical, human-intensive processes. In this thesis, we investigated two static analysis techniques: Finite-State Verification (FSV) and Fault Tree Analysis (FTA). We proposed a process analysis framework that is capable of performing both FSV and FTA on rigorously defined processes. Although evaluated for processes specified in the Little-JIL process definition language, this is a general framework independent of the process definition language. For FSV, we developed a translation-based approach that is able to take advantage of existing FSV tools. The process definition and property to be evaluated are translated into the input model and property representation accepted by the selected FSV tool. Then the FSV tool is executed to verify the model against the property representation. For FTA, we developed a template-based approach to automatically derive fault trees from the process definition. In addition to showing the feasibility of applying these two techniques to processes, much effort has been put on improving the scalability and the usability of the framework so that it can be easily used to analyze complex real-world processes. To scale the analysis, we investigated several optimizations that are able to dramatically reduce the translated models for FSV tools and speed up the verification. We also developed several optimizations for the fault tree derivation to make the generated fault tree much more compact and easier to understand and analyze. To improve the usability, we provided several approaches that make analysis results easier to understand. We evaluated this framework based on the Little-JIL process definition language and employed it to analyze two real-world, human-intensive processes: an in-patient blood transfusion process and a chemotherapy process. The results show that the framework can be used effectively to detect defects in such real-world, human-intensive processes.

Yes / Flaring is a combustion process commonly used in the oil and gas industry to dispose flammable waste gases. Flare flameout occurs when these gases escape unburnt from the flare tip causing the discharge of flammable and/or toxic vapor clouds. The toxic gases released during this process have the potential to initiate safety hazards and cause serious harm to the ecosystem and human health. Flare flameout could be caused by environmental conditions, equipment failure, and human error. However, to better understand the causes of flare flameout, a rigorous analysis of the behavior of flare systems under failure conditions is required. In this article, we used fault tree analysis (FTA) and the dynamic Bayesian network (DBN) to assess the reliability of flare systems. In this study, we analyzed 40 different combinations of basic events that can cause flare flameout to determine the event with the highest impact on system failure. In the quantitative analysis, we use both constant and time-dependent failure rates of system components. The results show that combining these two approaches allows for robust probabilistic reasoning on flare system reliability, which can help improving the safety and asset integrity of process facilities. The proposed DBN model constitutes a significant step to improve the safety and reliability of flare systems in the oil and gas industry.