基于多层流模型的核动力装置可靠性分析及故障诊断方法研究

发布时间：2018-06-25 10:35

本文选题：核安全 + 风险评估　；参考：《哈尔滨工程大学》2013年博士论文

【摘要】：在确保核安全的基础上实现良好的经济性是核电长期可持续发展的动力,可靠性分析技术是保障核电厂安全性、可靠性和经济性的重要手段。可靠性分析技术不仅可以定量评估设备可靠性对核电厂安全性和可用性的贡献,用于优化定期测试和在役检查周期,制定最佳的维修策略,而且可以揭示设备的故障原因和对系统运行和安全的影响,用于改进设计提高系统的固有可靠性和安全性,或者辅助运行人员进行故障诊断、制定应急规程或防范措施以消除和缓解故障影响。论文来源于国家自然科学基金和“十二五”核能开发项目,目的是在改进多层流模型(MFM)建模理论的基础上,研究基于MFM的核电厂可靠性分析方法及其在故障诊断中的应用,论文的主要工作如下：针对传统的MFM只能描述系统静态特性的问题,论文首先通过引入逻辑门、逻辑属性、基本目标、时间点和信号等新概念,对MFM建模方法进行了补充和改进,改进后的MFM可以更清晰地描述系统和设备之间的可靠性逻辑关系,同时也可以描述系统和设备的阶段任务和时序关系。论文提出了利用MFM层次化流结构进行可靠性分析的基本方法和特征量,提出了两状态系统可靠性分析方法和结合决策表分析的多状态系统可靠性分析方法,提出了MFM生成GO-FLOW模型方法,可以分析有阶段任务的动态系统可靠性问题。为了进行可靠性定性分析,论文提出了基于MFM的故障模式及影响分析(FMEA)方法,遵循守恒原理沿着MFM的流结构依次分析功能故障的原因和影响,方法简便且不容易遗漏重要的失效模式。在此基础上,论文给出了两状态系统MFM生成故障树方法,通过将MFM功能元件依次生成故障树模型、选择顶事件后连接各元件故障树模型并断开逻辑环路,可以对导致顶事件发生的最小割集、元件的重要度和敏感度等可靠性指标进行定性分析。论文将MFM可靠性定性分析方法应用于核电厂故障诊断领域,提出了一种综合故障诊断技术,集成了三种故障诊断方法：警报分析法根据守恒原理分析功能故障状态之间的因果关系,通过排除结果性警报,在实现警报压缩功能同时,给出可能的故障集；最小割集法以警报分析法为基础,通过进一步将多状态MFM转化为故障树模型,可以求解系统当前异常状态的最小故障模式；针对警报限值的设定对警报分析法和最小割集法可能造成的影响,论文提出了基于Bayesian原理的不确定性推理方法,通过将具有不确定性因果关系的多状态MFM转化为故障树模型,可以定量给出导致系统故障状态的近似却又合理的解释,在操纵员利用深层次知识进行故障诊断时可以辅助决策。论文对MFM图形化建模平台、基于MFM的可靠性分析程序和基于MFM的核电厂故障诊断系统进行了设计和开发,可以通过简单的鼠标和键盘操作快速生成和修改系统MFM,通过一次分析可以评价多个系统目标和组态变化,并可处理多状态、时延、时序和阶段任务等问题,可以分析设备故障原因及其对系统可运行性和安全性的影响,综合了基于目标、基于规则、基于风险和基于专家经验的特点,采用统一的建模语言和建模平台,辅助操纵员全面了解系统运行要求,进行推理判断并深入了解系统设计和运行中的薄弱环节,具有良好的工程实践前景。
[Abstract]:Achieving good economy on the basis of ensuring nuclear safety is the motive force of long-term sustainable development of nuclear power. Reliability analysis technology is an important means to ensure the safety, reliability and economy of nuclear power plants. Reliability analysis technology can not only quantify the contribution of equipment reliability to the safety and availability of nuclear power plants, but also can be used to optimize the determination of nuclear power plant. The period test and the service inspection cycle, make the best maintenance strategy, and can reveal the cause of the failure of the equipment and the effect on the system operation and safety. It can improve the inherent reliability and safety of the design to improve the system, or assist the operator to diagnose the fault, make emergency procedures or precautions to eliminate and alleviate the fault. Ringing.
The thesis comes from the National Natural Science Foundation and the "12th Five-Year" nuclear energy development project. On the basis of improving the multi-layer flow model (MFM) modeling theory, the paper studies the reliability analysis method of nuclear power plant based on MFM and its application in fault diagnosis. The main work of this paper is as follows:
In view of the problem that the traditional MFM can only describe the static characteristics of the system, the paper first supplements and improves the MFM modeling method by introducing the new concepts such as logic gate, logical attribute, basic target, time point and signal. The improved MFM can describe the reliability logic relationship between the system and the equipment more clearly, and can also describe the reliability of the system. The phase task and time sequence relationship of the system and equipment.
The paper puts forward the basic method and characteristic quantity of reliability analysis by using MFM hierarchical flow structure. The reliability analysis method of two state system and the method of multi state system reliability analysis with decision table analysis are put forward, and the method of generating GO-FLOW model by MFM is put forward, which can be used to analyze the dynamic system reliability problem of the stage task.
In order to carry out the qualitative analysis of reliability, the paper puts forward the fault mode and influence analysis (FMEA) method based on MFM, and follows the conservation principle along the flow structure of MFM to analyze the cause and influence of the functional failure in turn. The method is simple and not easy to omit the important failure mode. On this basis, the paper gives the fault tree of the two state system MFM. Methods the fault tree model of MFM functional components is generated in turn, and the fault tree model of each element is connected after the top event, and the logic loop is disconnected. The reliability index, such as the minimum cut set which causes the top event, the importance and sensitivity of the element, can be analyzed qualitatively.
In this paper, the MFM reliability qualitative analysis method is applied to the field of nuclear power plant fault diagnosis. A comprehensive fault diagnosis technology is proposed, and three kinds of fault diagnosis methods are integrated. The alarm analysis method is used to analyze the causality between the functional failure states according to the conservation principle, and by eliminating the result alerts, the alarm compression function is also given. The minimum cut set method is based on the alarm analysis method. By further converting the multi state MFM into a fault tree model, the minimum fault mode of the current abnormal state of the system can be solved. The paper puts forward the principle of Bayesian based on the effect of the warning limit set on the alarm analysis method and the minimum cut set method. By converting the multi state MFM with uncertain causality into a fault tree model, the method of uncertain reasoning can give an approximate but reasonable explanation for the failure state of the system. It can assist the decision when the manipulator uses deep knowledge to diagnose the fault.
The paper designs and develops the MFM graphical modeling platform, the MFM based reliability analysis program and the MFM based nuclear power plant fault diagnosis system. It can quickly generate and modify the system MFM through a simple mouse and keyboard operation. By one analysis, multiple system targets and configuration changes can be evaluated, and multi state and delay can be processed. The problem of time series and stage tasks can analyze the cause of equipment failure and its influence on the system's runtime and security. It combines the characteristics of target, rule based, based on risk and expert experience, and uses a unified modeling language and modeling platform to assist the operator to fully understand the system operation requirements and to conduct reasoning and judgment. An in-depth understanding of the weaknesses in system design and operation has good prospects for engineering practice.
【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TL364

【相似文献】