基于超椭球贝叶斯网络的高原寒旱区动车组可靠性分析方法研究——以兰新客专动车组为例

The static reliability analysis of Electric Multiple Unit(EMU) components, as the total manifestation of the internal state and the reflection of internal coupling and components mutual influence of EMU, which has been deeply studied by many scholars. But until now the comprehensive results have not been found yet for the integrated research how to quantitatively assess the risk of system performance changes and losses in different external operating environments. Lanzhou-Xinjiang passenger dedicated line, as the first high-speed railway in northwest China with special operating conditions,The EMUs have complex working conditions and functional hierarchy structures, Therefore, features, such as the ambiguity of fault status and failure rate of the EMUs’ components, the uncertainty of the logic between components and the dynamic sequential, are found in the formation of faults. The project is planning to concentrate on the methods of reliability modeling and risk status assessment EMU based on the results of the reliability analysis and failure analysis of the EMU's components under the background of integrating knowledge of different fault data and operating environment factors. Aiming at the small sample characteristics of the EMU operation data and its variety of knowledge types and the problems with various uncertainties, the data fusion method based on D-S fusion is adopted to improve the information utilization rate. In view of the following characters existing in the operation process of EMU including complex polymorphism, dynamic timing characteristics and uncertain characteristics, the reliability modeling and analysis method based on Bayesian network and evidence theory is studied combined with the fault interval probability which is subject to hyper-ellipsoidal model and Bayesian networks to improve the accuracy of reliability analysis, and an associated model is set up with different operating environment factors to quantitative assessment of the degree of risk exposure and find weak links of the system, and simulation calculation and experimental verification were carried out. It is expected that the research results will not only provide concrete modeling and analysis methods for the reliability analysis of EMU, but also provide reference for the adaptability optimization of EMU on Lanzhou-Xinjiang passenger dedicated line.

动车组部件静态可靠性分析作为动车组部件内部性态及其耦合关系的体现，研究已比较深入。但在不同运营环境中如何对动车组部件性能变化和失效可能性进行量化评估，尚未见到综合性成果。兰新客运专途经高原高寒干旱地区，动车组工作条件、功能层次结构复杂，在故障演化过程中存在部件故障状态、故障率的模糊性，故障逻辑关系的不确定性及动态时序性特点。本项目拟在融合不同故障知识与运营环境因素背景下，融入团队动车组部件可靠性分析方法及失效分析等成果，对动车组部件可靠性建模与风险状态评估方法进行研究。针对采集的兰新客专动车组运维数据及其具备的多源异构、小样本特性，利用D-S证据理论融合专家领域知识与运营样本数据，获取动车组部件底事件故障概率数据，为可靠性分析及计算打下基础；针对动车组运营过程中的多故障状态特点及动态特性，并考虑故障逻辑的不确定性，构建动车组多态可靠性模型与动态可靠性模型，并将超椭球模型约束的故障区间概率与贝叶斯网络相结合提升分析的精确度；针对高原寒旱区环境条件进行动车组故障的运营环境因素关联建模，对遭受风险程度进行定量评估，找到系统薄弱环节，进行仿真计算与实验验证，为动车组属地适应性优化提供参考。

兰新客专是高原寒旱区修建的首条高速铁路，探究动车组可靠性建模分析方法可为该类区域高速铁路安全运营积累宝贵经验。动车组在故障演化过程中存在部件故障状态、部件故障率的模糊性、故障逻辑关系的不确定性及动态时序性特点，课题对基于多源数据融合的动车组部件故障率获取及处理方法、动车组复杂多态可靠性建模分析方法、动车组复杂动态可靠性建模分析方法及环境因素关联建模方法进行了研究。首先，针对兰新客运专线动车组运维数据存在小样本特性、多源异构、不确定性及不均衡等问题，提出基于D-S证据理论的动车组多源数据融合处理方法获取动车组部件的故障率。根据故障等级、故障频次等属性获取动车组部件的主要故障及故障率参数；通过基于信心指数的专家调查法获取部件主要故障及重要程度；根据D-S证据理论实现主观专家领域知识信息与客观数据样本信息的结合，保证动车组关键部件故障率数据的可用性，为动车组部件可靠性分析奠定基础；其次，针对动车组由于工作条件多样及内部结构复杂而呈现出的多故障状态和多失效模式以及故障机理模糊、故障关系不确定等问题，利用T-S模糊故障树进行建模，并利用模糊数表述故障的多种状态，分别构建动车组系统复杂多态、动态贝叶斯网络可靠性分析模型，将T-S模糊多态故障树模型映射成为模糊贝叶斯网络，引入超椭球模型约束故障区间概率，提升可靠性分析结果精确度，最终确定影响动车组部件可靠性的薄弱环节与高风险事件，并通过建模仿真验证算法的可行性和有效性；最后，对高原寒旱区环境条件进行动车组故障的运营环境因素关联建模，通过引入环境因子以及模拟环境因素冲击下的退化模型对动车组部件可靠性建模，得到环境因素影响下的部件可靠度规律。以上成果为动车组安全运营及属地适应性优化提供参考。

内容获取失败，请点击重试