航空发动机叶片表面损伤与检测研究进展

Research Progress on Aeroengine Blade Surface Damage and Inspection

中文关键词:  叶片损伤  无损检测  机器视觉  深度学习  航空发动机

英文关键词:blade damage  nondestructive inspection  machine vision  deep learning  aeroengine

中文摘要:

      航空发动机叶片的工作环境极其恶劣，表面会出现各种类型的损伤。在损伤早期进行表面检测能够有效预防因损伤扩 展导致的叶片失效断裂。发动机叶片表面损伤的检测和评估主要由人工操作，严重依赖工作经验，但人工检测不仅效率低下，而 且检测结果容易受到人为因素的影响。为了高效、高精度地检测发动机叶片表面损伤，从叶片失效形式出发，综述了发动机叶片 在停放和运行2种状态下的损伤机理，并重点阐述了涡流检测、渗透检测等常用于叶片表面损伤检测的方法。总结了基于机器视 觉的检测技术，分析机器视觉检测面临数据集稀缺和单一性的挑战，认为收集大量数据并进一步完善评估标准是未来发动机叶片 表面损伤检测系统研究的重点方向。

英文摘要:

      The working environment of aeroengine blades is extremely harsh, leading to various types of surface damage. Surface inspection at the early stages of damage can effectively prevent blade failure and fracture caused by damage propagation. Currently, the inspection and evaluation of engine blade surface damage heavily rely on manual operation, which not only lacks efficiency but also suffers from susceptibility to human factors. In order to achieve efficient and accurate inspection of engine blade surface damage, from the forms of blade failure, the damage mechanisms under both non-operating and operating conditions are reviewed, with emphasis on methods com? monly used in blade surface damage inspection, such as eddy current and penetration. Additionally, this paper summarizes machine vision- based inspection technology while addressing the challenge posed by dataset scarcity and singularity in machine vision applications. It is believed that collecting extensive datasets and further enhancing evaluation criteria are key directions of future research on engine blade surface damage inspection systems.