可靠性工程中的重要度分析
发布时间：2018-08-13 10:57
【摘要】：重要度分析是进行系统可靠性分析的关键环节之一，用于定量分析系统中组件对系统影响的重要程度。它是融合了灵敏度、风险性、危害度和重要性等多类知识的前沿和热点研究领域之一，同时也是一种确定系统薄弱环节和提高系统可靠度的有力工具，对提高系统可靠性、安全性和进行故障诊断具有积极的意义。当前，对于重要度的分析尚缺乏全面深入的研究，仅仅作为可靠性工程中的一个概念，而没有独立全面的研究。这种局限于单一性的重要性研究可能会导致对系统的分析不够深入。本文尝试以数控冲床模具子系统和供油子系统的可靠性分析为研究对象，以重要度分析为主线，结合二元决策图BDD、多元决策图MDD、逻辑微分学以及马尔可夫随机过程等方法，研究二态系统和多状态系统组件的重要度，以期为进行系统可靠性评定、寿命预测和故障诊断与预防提供新思路和新方法，对丰富和完善现有的可靠性理论和方法起到一定的积极作用。本文主要研究内容如下： (1)提出基于二元决策图(BDD)的二态系统重要度分析方法。采用BDD来表示二态系统的结构函数，以布尔定律为依据，推理出BDD的算法规则，并在此基础上分析基于BDD的概率路径搜索以及可靠性的计算方法。利用递归原理实现由故障树(FT)向BDD的转化，由BDD代替FT来进行二态系统的组件重要度分析，解决当前FT法所面临的计算量大、结果不精确以及组合爆炸问题，为重要度的计算提供了一种高效精确的方法。 (2)提出基于多元决策图(MDD)的多状态系统重要度分析方法。对上述所建立的BDD模型进行拓展，，研究能表示多状态系统结构函数的多元决策图(MDD)，并结合逻辑微分学中的直接逻辑偏导数(DPLD)，对多状态系统的可靠性框图(RBD)进行模型转换。将MDD与DPLD理论引入到多状态组件重要度分析中，解决传统上行法和下行法等路集分析法的不足，使模型建立所耗费的时间短、计算直观简便。 (3)提出基于性能水平的多状态系统可靠性分析方法。考虑到组件的工作性能需求可能会因季节气候或者载荷的变化而不同，提出采用性能水平来分析组件对系统的重要程度。探讨组件性能随机性与马尔可夫离散随机过程之间的关系，用一个给定的性能水平对原始系统进行划分以建立起新的子系统，分析在给定性能水平下的组件重要度。表明在不同的性能水平下组件状态的重要性也不同，为系统检测及维修提供了一个更贴近实际工况的参考依据。
[Abstract]:Importance analysis is one of the key links in system reliability analysis, which is used to quantitatively analyze the importance of components in the system. It is one of the leading and hot research fields which combines sensitivity, risk, hazard and importance, and it is also a powerful tool to determine the weak links of the system and to improve the reliability of the system. Safety and fault diagnosis have positive significance. At present, the importance of the analysis of the lack of comprehensive and in-depth research, only as a concept of reliability engineering, and no independent comprehensive research. This study of the importance of singularity may lead to a lack of in-depth analysis of the system. This paper attempts to take the reliability analysis of the die subsystem and oil supply subsystem of NC punching machine as the research object, take the importance analysis as the main line, combine the binary decision diagram BDDD, the multivariate decision diagram MDD, the logic differentiator and the Markov stochastic process, etc. The importance of two-state and multi-state system components is studied in order to provide new ideas and methods for system reliability evaluation, life prediction, fault diagnosis and prevention, To enrich and improve the existing reliability theory and methods play a positive role. The main contents of this paper are as follows: (1) A two-state system importance analysis method based on binary decision graph (BDD) is proposed. The structure function of two-state system is represented by BDD. Based on Boolean law, the algorithm rules of BDD are deduced. On this basis, the probabilistic path search based on BDD and the calculation method of reliability are analyzed. The transformation from fault tree (FT) to BDD is realized by recursion principle, and the component importance analysis of two-state system is carried out by BDD instead of FT, which solves the problems of large computation, inaccurate results and combined explosion in the current FT method. It provides an efficient and accurate method for the calculation of importance degree. (2) the importance analysis method of multi-state system based on multivariate decision graph (MDD) is proposed. In this paper, the BDD model is extended to study the multivariate decision graph (MDD), which can express the structural function of the multistate system, and the model transformation of the reliability block diagram (RBD) of the multistate system with the direct logical partial derivative (DPLD), in the logic differentiator. The theory of MDD and DPLD is introduced into the importance analysis of multi-state components to solve the shortcomings of traditional uplink method and downlink method, so that the time spent on modeling is short. (3) A multi-state system reliability analysis method based on performance level is proposed. Considering that the performance requirements of components may vary according to the seasonal climate or load, it is proposed that the performance level be used to analyze the importance of components to the system. The relationship between component performance randomness and Markov discrete stochastic process is discussed. The original system is partitioned with a given performance level to establish a new subsystem, and the component importance at a given performance level is analyzed. It is shown that the importance of component state is different at different performance levels, which provides a reference basis for system detection and maintenance.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB114.3
本文编号：2180768
[Abstract]:Importance analysis is one of the key links in system reliability analysis, which is used to quantitatively analyze the importance of components in the system. It is one of the leading and hot research fields which combines sensitivity, risk, hazard and importance, and it is also a powerful tool to determine the weak links of the system and to improve the reliability of the system. Safety and fault diagnosis have positive significance. At present, the importance of the analysis of the lack of comprehensive and in-depth research, only as a concept of reliability engineering, and no independent comprehensive research. This study of the importance of singularity may lead to a lack of in-depth analysis of the system. This paper attempts to take the reliability analysis of the die subsystem and oil supply subsystem of NC punching machine as the research object, take the importance analysis as the main line, combine the binary decision diagram BDDD, the multivariate decision diagram MDD, the logic differentiator and the Markov stochastic process, etc. The importance of two-state and multi-state system components is studied in order to provide new ideas and methods for system reliability evaluation, life prediction, fault diagnosis and prevention, To enrich and improve the existing reliability theory and methods play a positive role. The main contents of this paper are as follows: (1) A two-state system importance analysis method based on binary decision graph (BDD) is proposed. The structure function of two-state system is represented by BDD. Based on Boolean law, the algorithm rules of BDD are deduced. On this basis, the probabilistic path search based on BDD and the calculation method of reliability are analyzed. The transformation from fault tree (FT) to BDD is realized by recursion principle, and the component importance analysis of two-state system is carried out by BDD instead of FT, which solves the problems of large computation, inaccurate results and combined explosion in the current FT method. It provides an efficient and accurate method for the calculation of importance degree. (2) the importance analysis method of multi-state system based on multivariate decision graph (MDD) is proposed. In this paper, the BDD model is extended to study the multivariate decision graph (MDD), which can express the structural function of the multistate system, and the model transformation of the reliability block diagram (RBD) of the multistate system with the direct logical partial derivative (DPLD), in the logic differentiator. The theory of MDD and DPLD is introduced into the importance analysis of multi-state components to solve the shortcomings of traditional uplink method and downlink method, so that the time spent on modeling is short. (3) A multi-state system reliability analysis method based on performance level is proposed. Considering that the performance requirements of components may vary according to the seasonal climate or load, it is proposed that the performance level be used to analyze the importance of components to the system. The relationship between component performance randomness and Markov discrete stochastic process is discussed. The original system is partitioned with a given performance level to establish a new subsystem, and the component importance at a given performance level is analyzed. It is shown that the importance of component state is different at different performance levels, which provides a reference basis for system detection and maintenance.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB114.3
【参考文献】
相关期刊论文 前9条
1 田宏,陈宝智,吴穹,高永庭;多态系统可靠性及元素的不确定性重要度[J];东北大学学报;2000年06期
2 毕卫星;郭成宇;;提高重要度的兼容性算法——联合关键重要度[J];大连交通大学学报;2010年05期
3 孙红梅;高齐圣;朴营国;;关于故障树分析中几种典型重要度的研究[J];电子产品可靠性与环境试验;2007年02期
4 王海涛;吴宜灿;李亚洲;胡丽琴;FDS团队;;核电站实时风险管理系统部件重要度计算方法研究[J];核科学与工程;2008年01期
5 姚成玉;张荧驿;陈东宁;王旭峰;;T-S模糊重要度分析方法研究[J];机械工程学报;2011年12期
6 史定华;单元的重要度及其计算[J];科学通报;1984年06期
7 杨林娟;沈士明;;基于粗糙集理论的故障树重要度分析[J];南京工业大学学报(自然科学版);2007年01期
8 张英芝;郑珊;申桂香;郑锐;谷东伟;牛序磊;;采用重要度和模糊推理的数控刀架危害性分析[J];吉林大学学报(工学版);2012年05期
9 古莹奎;邱光琦;储茜;;多状态串并联系统工作性能及其状态概率分析[J];江西理工大学学报;2013年03期
本文编号：2180768
本文链接：https://www.wllwen.com/guanlilunwen/gongchengguanli/2180768.html
