基于数据驱动的民机系统可靠性分析与预测研究

项目名称： 基于数据驱动的民机系统可靠性分析与预测研究

项目编号： No.U1233115

项目类型： 联合基金项目

立项/批准年度： 2013

项目学科： 电子学与信息系统

项目作者： 王华伟

作者单位： 南京航空航天大学

项目金额： 36万元

中文摘要： 针对民机系统具有的状态非线性及多失效模式的特征，提出基于数据驱动的可靠性分析与预测技术。分析和处理用于支撑民机系统可靠性分析与预测的复杂数据，建立数据与民机系统状态、失效模式和可靠性之间的关系，解决数据可利用及提升利用效率和效果的问题；采用模型集成的方法构建基于竞争失效的民机系统复合可靠性分析模型，通过数据定量分析退化失效和竞争失效的作用机制；研究基于复杂数据及复合模型的不确定性及灵敏度分析方法，提高民机系统可靠性分析与预测的准确性；分别建立基于性能退化及突发失效的剩余寿命预测模型；结合数据分析多失效模式竞争失效的变化规律，基于竞争失效的复合剩余寿命预测模型及风险决策机制，预测民机系统剩余寿命。通过在民机航电系统和发动机系统中的应用，验证方法的有效性。本项目对于丰富和发展可靠性工程理论和方法，推动民机系统可靠性工程管理科学化具有重要意义。

中文关键词： 可靠性分析；可靠性预测；数据融合；数据驱动；竞争失效

英文摘要： The reliability analysis and prediction on the data driven approach is proposed to describe the nonlinear condition and multiple failure mode of the civil aircraft system. Analyzing and processing data methods are used to build the data relationship with condition, failure mode and reliability, which improve the data utilization effectiveness and efficiency. Integration method is proposed to build composite model on reliability analysis of competing failure for civil aircraft system. The interaction mechanism between degradation failure and catastrophic failure is obtained by data driven approach. Uncertainty and sensitivity methods are proposed for complex data and composite reliability model, which improve the accuracy of reliability analysis and prediction. The remaining useful life prediction models are built for degradation failure and catastrophic failure , respectively. By analyzing the interaction mechanism variation of multiple failure mode, the remaining useful life method is proposed on reliability prediction model and risk decision. The approach is demonstrated effectiveness by avionics system and engine system of civil aircraft. The project has the important meaning for enriching and developing reliability engineering theory and methods, promoting commercial system scientific management on reliabili

英文关键词： reliability analysis；reliability prediction；data fusion；data driven；competing failure