基于混合不确定性信息处理的复杂分层系统健康评估

项目名称： 基于混合不确定性信息处理的复杂分层系统健康评估

项目编号： No.U1330130

项目类型： 联合基金项目

立项/批准年度： 2014

项目学科： 数理科学和化学

项目作者： 黄洪钟

作者单位： 电子科技大学

项目金额： 88万元

中文摘要： 以装备中的复杂分层系统健康评估为应用背景，考虑装备功能的复杂性、结构关系的层次性和应用环境的可变性，结合不确定性量化模型和测度论，研究同时计及随机和认知不确定性的非概率度量方法和健康评估非概率模型。同时，针对不确定性综合处理过程中信息不完备和计算量过大的问题，研究兼顾精度与效率的高精度代理模型和融合多源信息的高效算法，进行一定预测置信度下的定量推断。继而，基于层次分析思想，研究复杂分层系统层次模型的描述方法、不确定性在层次模型中从底层向顶层传播的处理方法。然后，以混合不确定性表征、处理和传播方法为基础，选取同时具有混合不确定性和分层特点的典型复杂分层系统进行实例验证，最终建立一套可以综合处理混合不确定性、能够融合包括多源、多层和异构不确定性信息在内的多元信息，且具有较高适应性的健康评估方法，为提高装备的健康评估水平提供支撑。

中文关键词： 分层系统；混合不确定性；信息融合；健康评估；代理模型

英文摘要： The health assessment problems of critical application systems in equipments are investigated, in consideration of the complexity of system functions, hierarchy of structural relationship and changeability of operation environments, by means of uncertainty quantification models and approaches, with the aim of non-probabilistic uncertainty measurements and non-probabilistic modeling under both aleatory uncertainty and epistemic uncertainty. Referring to inaccurate and incomplete information, as well as excessive calculation in the process of non-probabilistic assessment, this project investigates an assessment framework of high-efficiency surrogate model, along with an efficient information-fusion algorithm, with the result of quantitative inference under a specific confidence. Afterwards, such methodologies as hierarchical model description and from-bottom-to-top uncertainty propagation for complex key systems are presented, on the basis of analytic hierarchy process. Moreover, based on the aforementioned work, this project particularly provides an application of a typical key system under hybrid uncertainty, with the properties of hierarchical structure and insufficient information. Those efforts are devoted to a comprehensive health assessment, applying multiplex information fusion with multi-source, multi-lev

英文关键词： Hierarchical system；Hybrid uncertainty；Information fusion；Health assessment；Surrogate model