基于叶尖定时技术的航空发动机轮盘裂纹识别方法

Disk break will be generated if the crack of aero-engine disk spread to a certain extent, thus it is important for the safety operation of aero-engine to identify the crack on-line before the disk is broken. The disk crack of aero-engine has little impact for the shaft system vibration response. The vibration characteristic signals are difficult to be monitored. However, the crack of disk will influence the time of group blades’ tip near crack reaches the reference position and shows the law of advance or lag. The tip-timing principle and equipment of aero-engine disk’s crack are about to be utilized for the identification of disk crack on-line. Research content: the damage propagation laws of disk’s crack will be studied, the typical position and the expansion path of disk crack will be revealed; the finite element reduction model of the disk-blade structure with cracks will be established, the vibration characteristics of the disk-blade structure containing cracks in the disk will be obtained, the effects of crack parameters on the vibration characteristics of the blade will be revealed; The characteristics of the tip-timing signal of the cracked disk will be studied, the method of noise reduction and feature extraction of the tip-timing signal will be proposed, the characteristic information of the tip-timing signal of the crack in the disk will be extracted; the crack disk identification model based on the characteristics of the tip-timing signal will be established, a accurately disk crack identification method based on the tip timing vibration measurement technology will be formed. The project provides a theoretical method for on-line identification of aero-engine disk cracks, it is with great significance for the safe operation and health maintenance of aero-engine.

航空发动机轮盘裂纹扩展至一定程度将引发轮盘破裂，轮盘破裂前在线识别出裂纹对航空发动机安全运行至关重要。轮盘裂纹对轴系振动响应影响微弱,振动特征信号难以监测，但会引起裂纹附近叶片组各叶尖到达参考位置的时间呈现超前或滞后的规律，应用航空发动机叶片裂纹的叶尖定时识别原理与设备，研究轮盘裂纹在线识别方法。研究内容：研究轮盘裂纹扩展机理，揭示轮盘裂纹的典型位置和扩展路径；建立含裂纹轮盘-叶片结构有限元减缩模型，获取含轮盘裂纹的轮盘-叶片结构振动响应特性，揭示裂纹参数对叶片振动特性的影响规律。研究轮盘裂纹叶尖定时信号特性，提出叶尖定时信号降噪与特征提取方法，提取轮盘裂纹叶尖定时信号的特征信息。建立基于叶尖定时信号特征的轮盘裂纹识别模型，形成基于叶尖定时测振技术的轮盘裂纹识别方法，实现轮盘裂纹的准确识别。项目为实现航发动机轮盘裂纹机载在线识别提供理论方法，对于航空发动机的安全运行与健康维护具有重要意义。

航空发动机轮盘裂纹扩展至一定程度将引发轮盘破裂，轮盘破裂前在线识别出裂纹对航空发动机安全运行至关重要。轮盘裂纹对轴系振动响应影响微弱，振动特征信号难以监测，但会引起裂纹附近叶片组各叶尖到达参考位置的时间呈现超前或滞后的规律。项目应用航空发动机叶尖定时识别原理与设备研究轮盘裂纹在线识别方法，主要开展了如下研究：航空发动机转子-轮盘-叶片系统动力学特性，分析了轮盘-叶片-转轴整体受力环境；研究了轮盘外缘裂纹扩展机理，考虑了初始裂纹位于两叶片间的相对位置对裂纹扩展路径的影响，得到不同位置处的径向裂纹扩展路径；研究了初始裂纹的初始角度对裂纹扩展路径的影响，同等载荷下，裂纹路径主要与裂纹尖端所在的位置相关，与裂纹初始角度几乎不相关；研究了非贯穿式裂纹长深比对裂纹扩展路径的影响，同等载荷下，非贯穿式裂纹将逐渐演变为贯穿裂纹，继续扩展后，裂纹扩展路径主要与裂纹初始位置相关，最终得到了一定载荷下轮盘外缘裂纹扩展路径，为后续裂纹植入与含裂纹轮盘的动力学分析提供支撑；研究了含裂纹轮盘的固有特性与谐响应特性，轮盘裂纹将引起叶盘的模态局部化现象，导致叶盘各叶片的共振频率分离，共振频率区变宽，裂纹越靠近叶根，影响更明显；研究了轮盘裂纹导致的叶尖定时信号响应规律，研究了叶尖信号处理方法，对比分析了深度信念网络、支持向量机、BP神经网络三种方法，其中基于DBN方法能够更有效地对轮盘裂纹故障特征进行识别。项目研究获得的轮盘外缘裂纹扩展路径、含外缘裂纹的轮盘-叶片结构的固有特性、响应特性、叶尖定时信号响应特性、裂纹诊断方法将为航空发动机轮盘裂纹在线诊断提供基础。