复杂系统动态故障树分析的新方法及其应用研究

发布时间：2018-06-05 06:15

本文选题：系统可靠性分析 + 动态故障树分析　；参考：《电子科技大学》2013年博士论文

【摘要】：随着现代工程系统的大型化、复杂化以及高新技术的引入，系统可靠性已经成为制约复杂系统发展的关键所在。可靠性分析技术作为实施系统可靠性工程的关键基础技术，目前正面临着复杂系统所带来的若干技术难点和应用挑战。针对复杂系统的可靠性分析技术已经成为可靠性工程领域的研究热点及难点问题之一。系统可靠性分析常规方法主要包括：可靠性框图法、故障模式影响及危害性分析法、故障树分析法、Petri网方法以及蒙特卡洛数值仿真方法等。常规方法通常不考虑系统的动态失效特性，且多数建立在零部件故障相互独立和故障数据完备的基础之上。在实际复杂工程系统中，零部件失效之间通常并不是相互独立的，往往存在着多种复杂的关联关系和动态特性，比如部件失效的顺序关系。另一方面，由于成本、时间、管理和人因等多方面的原因导致零部件失效数据存在模糊不确定性。目前，在考虑动态失效特性的故障树分析方面已取得了一定成果。然而，在同时考虑模糊不确定性以及动态失效特性等情况下的故障树分析方面的研究工作还很缺乏，以致用常规方法分析所得结果与实际情况不符甚至相差甚远。因此，迫切需要开展考虑零部件动态失效特性和模糊不确定性的系统可靠性分析方法的研究。针对上述问题，本文主要开展了以下研究工作：（1）基于模糊马尔科夫模型的动态故障树分析方法。马尔科夫模型方法是一种状态空间分析方法，用该模型能够准确地描述失效分布与维修分布都服从指数分布的系统的失效及维修过程。本文在基于马尔科夫模型的基础上，考虑了零部件失效信息的模糊不确定性，研究了在模糊失效率下的动态故障树分析方法。通过建立系统的动态故障树模型，并运用三角模糊数来描述零部件和系统的失效率，通过已经得到的动态故障树模型建立系统失效过程的模糊马尔科夫模型。运用模糊理论中扩展原理的思想和Laplace-Stieltjes变换求解该模型，得到系统在给定时刻下的模糊失效概率和给定隶属度下的模糊可靠度曲线。最后应用该模糊马尔科夫模型对某数控加工中心液压系统进行可靠性建模与分析。研究结果表明，该方法能够有效地对具有动态失效特性和模糊不确定性的系统进行可靠性建模及定量评估。（2）基于离散时间贝叶斯网络的动态故障树可靠性评估模型。研究了基于贝叶斯网络和动态故障树的系统可靠性建模和评估方法。通过把系统失效的动态故障树模型转化为贝叶斯网络模型，，并运用贝叶斯网络的拓扑结构来表达系统中部件失效之间的逻辑关系。针对基于马尔科夫模型的动态故障树求解方法中存在的状态爆炸问题，借助贝叶斯网络的条件独立性来降低模型求解的复杂度。在此基础上，建立了静态和动态故障树中各种逻辑门的条件概率分布的公式，以实现对系统失效过程及其动态特性进行建模和分析。以卫星太阳翼驱动机构为对象，建立了动态故障树模型和相应的贝叶斯网络模型，并运用联合树推理算法对该模型进行了双向概率推理。实例分析结果表明：该方法能够有效地解决具有动态失效特性的复杂系统的可靠性分析和评估问题。（3）模糊数据下基于连续时间贝叶斯网络的动态故障树分析方法。研究了考虑模糊不确定性的基于连续时间贝叶斯网络的系统可靠性建模与分析方法。基于连续时间贝叶斯网络模型的方法能够直接得到系统的可靠度和失效概率的解析表达式。本文用三角模糊数描述零部件的失效率，并用其来构造零部件的模糊边缘失效密度函数及模糊失效分布函数。用单位阶跃函数和冲激函数来构造贝叶斯网络中非根节点失效事件的条件概率密度函数和分布函数。在此基础上，推导了在模糊失效率下的几种典型的故障树逻辑门输出事件发生的模糊边缘失效密度函数和模糊失效分布函数的表达式。最后，运用算例验证了该方法的正确性和有效性，并通过对大型矿用挖掘机电气系统整流回馈子系统的建模与分析阐述了该方法在实际工程系统中的应用。（4）考虑共因失效的动态故障树分析方法。运用故障树分析方法对具有共因失效的系统进行了可靠性分析。阐述了当前共因失效研究中的一些经典模型和建模方法，运用显式建模方法与平方根模型对某动车组追尾事故进行了故障树分析。分别计算了考虑共因失效和假设部件失效独立两种情况下的系统失效概率。结果表明：不考虑共因失效因素的影响会对可靠性分析结果带来较大的误差，说明了共因失效对于交通工具这种重要设施的安全性影响非常重大，同时也表明了考虑共因失效的动态故障树分析方法可为列车安全性及可靠性评估提供基础。同时，本文还提出了各种备份条件下考虑共因失效的动态故障树及贝叶斯网络可靠性建模及评估方法。建立了考虑共因失效条件下，确定贝叶斯网络中各种备件门输出事件对应节点的条件概率分布表的方法。通过算例验证了该方法的有效性，并通过与蒙特卡洛数值仿真方法对比，验证表明该方法的计算精度能够满足实际要求。
[Abstract]:With the large-scale, complex and high technology of modern engineering system, system reliability has become the key to restricting the development of complex systems. As the key basic technology for implementing system reliability engineering, the reliability analysis technology is facing some technical difficulties and application challenges brought by the complex system. The reliability analysis technology of complex systems has become one of the hot and difficult problems in the field of reliability engineering. The conventional methods of system reliability analysis mainly include: reliability block diagram, failure mode influence and hazard analysis, fault tree analysis, Petri network method and Monte Carlo numerical simulation method. Generally, the dynamic failure characteristics of the system are not considered, and most of them are based on the independence of the parts and the complete fault data. In the actual complex engineering system, the failure of the parts is usually not independent, and there are many complex correlation and dynamic characteristics, such as the sequence relation of the failure of the components. On the other hand, due to many factors such as cost, time, management and human cause, there are fuzzy uncertainties in the failure data of parts. At present, some achievements have been achieved in the fault tree analysis considering dynamic failure characteristics. However, the fault tree analysis under the condition of fuzzy uncertainty and dynamic failure characteristics is considered simultaneously. The research work is still very short, so that the results obtained by the conventional method are not quite different from the actual situation. Therefore, it is urgent to carry out the study of the system reliability analysis method considering the dynamic failure characteristics and the fuzzy uncertainty of the parts.
In view of the above problems, the following research work has been carried out in this paper.
(1) the dynamic fault tree analysis method based on the fuzzy Markov model. The Markov model method is a state space analysis method, which can accurately describe the failure and maintenance process of the system whose failure distribution and maintenance distribution are all subject to exponential distribution. Based on the Markov model, this paper considers the zero part. The fuzzy uncertainty of the failure information is studied. The dynamic fault tree analysis method under the fuzzy failure rate is studied. By establishing the dynamic fault tree model of the system and using the triangular fuzzy number to describe the failure rate of the parts and systems, the fuzzy Markov model of the system failure process is established by the dynamic fault tree model obtained. Using the idea of the extended principle in the fuzzy theory and the Laplace-Stieltjes transformation to solve the model, the fuzzy failure probability and the fuzzy reliability curve under given membership are obtained. Finally, the fuzzy Markoff model is applied to the reliability modeling and analysis of the hydraulic system of a CNC machining center. It is shown that this method can effectively model and evaluate the reliability of systems with dynamic failure characteristics and fuzzy uncertainties.
(2) based on the dynamic fault tree reliability evaluation model of discrete time Bias network, the system reliability modeling and evaluation method based on the Bias network and dynamic fault tree is studied. The dynamic fault tree model of the system failure is transformed into the Bias network model, and the topology of Bias network is used to express the system. According to the problem of state explosion in the dynamic fault tree solving method based on Markoff model, the condition independence of the Bayesian network is used to reduce the complexity of the model solution. On this basis, a formula for the conditional probability distribution of various logic gates in the static and dynamic fault tree is established. The system failure process and its dynamic characteristics are modeled and analyzed. A dynamic fault tree model and a corresponding Bayesian network model are established based on the satellite solar wing driving mechanism, and the joint tree reasoning algorithm is used to carry out two-way probabilistic reasoning. The example analysis results show that the method can be effectively solved. Reliability analysis and evaluation of complex systems with dynamic failure characteristics.
(3) the method of dynamic fault tree analysis based on continuous time Bayesian networks under fuzzy data. The method of system reliability modeling and analysis based on continuous time Bayesian network based on fuzzy uncertainty is studied. The method of continuous time Bayesian network model can directly obtain the analysis of reliability and failure probability of the system. In this paper, a triangular fuzzy number is used to describe the failure rate of parts and components, and to construct the fuzzy edge failure density function and fuzzy failure distribution function of parts and components. The conditional probability density function and distribution function of the failure events of the non root nodes in Bayesian networks are constructed by using the unit step function and impulse function. The expression of fuzzy edge failure density function and fuzzy failure distribution function of several typical fault tree logic gate output events under the fuzzy failure rate is derived. Finally, the correctness and effectiveness of the method are verified by an example, and the modeling and classification of the rectifier feedback subsystem for the electrical system of a large mining excavator are established. The application of the method in practical engineering system is expounded.
(4) the dynamic fault tree analysis method of common cause failure is considered. The reliability analysis of the system with common cause failure is carried out by the fault tree analysis method. Some classical models and modeling methods in the current common cause failure study are expounded. The fault tree of the rear end accident of a certain EMU is carried out by using the explicit modeling method and the square root model. The failure probability of the system is calculated in two cases, including the common cause failure and the hypothesis component failure. The results show that the influence of the common cause failure factor will bring the greater error to the reliability analysis results, which shows that the common cause failure has a great influence on the safety of the important facilities such as the vehicle. The dynamic fault tree analysis method considering common cause failure can provide a basis for the evaluation of train safety and reliability. At the same time, the dynamic fault tree and Bayesian network reliability modeling and evaluation method considering common cause failure under various backup conditions are proposed. The Bayesian network is established under the condition of common cause failure. The validity of the method is verified by a numerical example, and the results show that the accuracy of the method can meet the actual requirements by comparing with the Monte Carlo numerical simulation method.
【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TH165.3;TP18

【参考文献】