航空发动机结构可靠性多元建模及仿真理论与方法研究

Traditional fatigue design theory cannot uniformly deal with low cycle, high cycle and super high cycle fatigue mechanisms as well as multiple site damage and multiple part correlation, traditional reliability model fails to completely reflect the effect of complicated random load history, strength degradation and structure geometry..Focused on the fatigue reliability design and evaluation of the mechanical components and system of aeronautical engine, the present proposal researches the way to characterize the uncertainty of complex load history hierarchically to support time-dependent fatigue reliability modeling, investigates the interaction and probabilistic competition among low cycle fatigue, high cycle fatigue and super high cycle fatigue, studies the size effect of wide spread damage part and the stochastic behavior of strength degradation under complex load history, systematically deals with complicated engine part fatigue problem involving multi-site damage, multiple failure mechanisms and multiple failure modes within the framework of probability theory, and builds universal fatigue damage evolution model, fatigue life prediction model and time-dependent reliability model which can take into account the different fatigue mechanisms completely and uniformly..Proceeding from system engineering point of view, multi-level statistics principle and conditional independence concept, based on the comprehensive description to the interaction among different failure mechanisms as well as the probabilistic competition relationship between different failure modes, innovates multi-mode fatigue reliability modeling methodology, constructs multi-variate reliability model and simulation approach for complex mechanical parts concerning multiple failure modes, accurately expresses the time-dependent issues (i.e. the statistical risk effect of load history, material strength degradation), spatial issues (load uncertainty in macro level and part geometry), as well as the emergency (statistical dependence among different failure sites and/or different failure mechanisms) of the reliability of aero engine parts and system.

传统疲劳理论缺乏统一处理多机理耦合、多部位损伤及多构件关联的方法，传统可靠性模型未能全面反映载荷历史、强度退化、结构特征等要素。针对航空发动机及其关键零部件疲劳可靠性设计与评估，研究载荷历程不确定性分层表征方式、低周-高周-超高周疲劳三种失效机理之间的相互作用和概率竞争关系、广布损伤结构的尺度效应、复杂载荷历程下的强度退化统计规律等，在概率框架下统一处理多部位、多机理疲劳问题，实现不同失效机理疲劳损伤演化和寿命预测模型的统一及可靠性模型的统一。基于系统论思想，借助多层统计分析等数学工具和随机事件之间的条件独立原理，立足于对失效机理和失效模式的科学表述，创新系统可靠性建模途径，建立统一表达多影响因素、多失效模式复杂零部件及系统的多元可靠性模型，发展仿真技术，精确反映可靠性的时间属性（载荷统计风险效应、部件性能退化）、空间属性（载荷宏观不确定性、部件结构特征）和关联属性（失效相关性）。

机械系统及其零部件的可靠性是航空发动机极其重要的质量指标。对于航空发动机疲劳可靠性评估问题，传统疲劳可靠性理论缺乏统一处理多机理并、多部位损伤及多构件关联等问题的方法。传统可靠性模型不能全面反映载荷不确定性、材料强度退化、结构多部位损伤效应、零部件之间的失效相关性等要素。针对航空发动机的随机载荷环境及其关键零部件的多轴应力状态，发展、完善了疲劳可靠性设计与评估方法。重点研究并提出了载荷历程不确定性的结构性表征方式、关键零部件低周-高周-超高周疲劳三种失效机理之间的相互作用和概率竞争关系模型、广布损伤结构的尺度效应、复杂载荷历程下的强度退化统计规律、多轴应力状态下的寿命预测方法等，在概率框架下统一处理多部位、多机理疲劳问题，实现了不同失效机理的疲劳损伤演化和寿命预测模型的统一及可靠性模型的统一。基于系统论思想，借助多层统计分析等数学工具和随机事件之间的条件独立特性及全概率计算原理，立足于对失效机理和失效模式的科学表述，建立了统一表达多影响因素、多失效模式复杂零部件及系统的多元可靠性模型。同时，进行了详细的动力学分析、发展了高效可靠性仿真技术，精确反映可靠性的时间属性（载荷统计风险效应、部件性能退化）、空间属性（载荷宏观不确定性、部件结构特征）和关联属性（失效相关性）。.该项目发展、完善了航空发动机载荷历程及其不确定性表征方法，为可靠性评估创建了良好开端；采用的系统层可靠性建模方法，避免了传统建模程式处理零部件失效相关性的困难，有普遍应用价值；项目提出的材料强度退化规律测试原理及数据拟合方法，综合了早期失效信息，弥补了过往方法的不足；建立的齿轮传动系统可靠性模型，更新了从可靠性分析角度对系统类型的认识；提出的基于虚拟/经验大数据的超小样本右截尾寿命数据条件下的可靠性评估方法，显著提高了Weibull分布的应用价值。