流体激振下离心式压缩机动叶片在线监测与运行可靠性评估方法研究

针对流体激振下大型离心式压缩机动叶片疲劳失效亟待解决的紧迫问题，瞄准动叶片在线监测与运行可靠性评估方法问题，开展流体激振下的叶片疲劳失效规律、叶片运行过程的在线监测、微弱特征提取、参数表征与运行可靠性评估方法研究。从流体激振造成叶片疲劳失效的研究入手，结合流体动力学仿真、结构动力学分析、声疲劳仿真分析，并在压缩机模型级试验台进行验证，获取流体激振下气动载荷传递规律与叶片动载荷的作用机理，建立流体激振下叶片疲劳失效的表征方法。建立叶片运行过程中最佳的监测位置和监测参数，展开表征叶片状态的特征参数提取方法研究，研究叶片裂纹的早期预警方法；融合多个监测参数进行定量评估。在此基础上，应用粒子滤波方法进行基于状态空间预测算法实现，建立基于状态监测的叶片运行可靠性评估模型。此研究将促进动叶片在线监测与运行可靠性评估方法的发展，并可为燃气轮机、航空发动机、汽轮机等叶片安全可靠运行的问题解决提供借鉴。

Rotor blade fatigue failure based on flow induced vibration is a very urgent problem for large scale centrifugal compressor, which could lead to serious catastrphe. This research aims at rotor blade condition monitoring and operational reliability estimation problems to keep centrifugal compressor working in normal conditions. Blade fatigue failure behavior based on flow induced vibration, rotor blade oeprational condition monitoring, weak feature extraction, parameters charactersitics, quantitative evaluation, and operational reliability estimation are investigated in this research. Transfer behavior of flow induced vibration and action mechanism for dynamic load are determined based on blade fatigue failure as the reason of flow induced vibration together with fluid dynamics simulation analysis, structural dynamics simulation analysis, and acoustic fatigue simulation analysis. As well, experimental verification is also launched on the model stage test-rig. Blade fatigue failure behavior based on flow induced vibration characterization can be constructed. Characteritic parameters for blade working condition are investigated to determine the best monitoring position and monitoring parameters. As well, feature extraction method for blade fatigue failure is investigated in this research. Early warning method for blade crack are carried on. Quantitative evaluation based on multiple monitoring paramters fusion are investigated. Based on those results, state-space model is carried out for the operational reliability estimation based on particle filter method. Operational reliability estimation model is constructed according to blade condition monitoring information. This research will contribute to the development of blade reliability design theory and operational reliability estimation method. As well, this research will provide a reference to blade faitgue failure problem solvement, such as gas turbine, aero-engine, steam turbines, and so on.