热障涂层厚度反射式脉冲太赫兹/高频涡流复合无损检测新方法研究

Nondestructive testing and evaluation of thermal barrier coating is urgently demanded for structural integrity of in-service advanced aero-engines. Thermal barrier coating consists of top coating (TC), thermal grown oxide (TGO), bonding coating (BC), and nickel-base superalloy. Each layer varies significantly in terms of thickness, electromagnetic and optical parameters. Therefore, a single nondestructive method is hardly capable of evaluating thermal barrier coating in an accurate and noncontact way. This projects aims to develop a novel method combing reflected terahertz pulses with high-frequency eddy current testing. According to bright specular reflections from metallic materials, the pulse-echo terahertz signals will be adopted to determine TC thickness considering the differences of refractive indexes between the air and TC. We will be concentrated on the physics that the amplitude of the first peak of a pulse-echo signal could be used for measurement of TC refractive index when TC thickness is unknown. The delay time between the adjacent peaks would serve as the feature for determination of TC thickness. TGO is extremely thin (up to 10 um) and has low conductivity (about 0.4 MS/m). Hereby, high-frequency eddy current testing will be proposed to infer the thicknesses of TGO and BC with model-based inversion, since the increase of excitation frequency enhances the resolution of eddy current testing. The key tasks are to reduce the discrepancy between the calculations and measurements, to select excitation frequencies, and to remove the effect of liftoff variations (the distance between a probe and a target) for accuracy enhancement. The Mento Carlo method will be employed to characterize the uncertainty for reliability evaluation. This project could be expected to generate a noncontact, accurate and high-resolution method for comprehensively nondestructive determination of thickness of each layer in thermal barrier coating.

热障涂层厚度无损检测与评估是先进航空发动机发展迫切需求的关键技术。热障涂层由陶瓷层（TC）、氧化物（TGO）、金属粘结层（BC）和合金基体组成，各层厚度、电磁和光学参数差异大，属于复杂多层结构，因而单一方法难以胜任TC、TGO和BC厚度的准确非接触检测。本项目提出基于脉冲太赫兹/高频涡流的复合检测新方法。依据空气和TC折射率的差异以及太赫兹对金属强反射特性，提出基于反射式脉冲太赫兹TC厚度检测方法，探寻反射信号首个峰值表征折射率的机理，利用相邻峰值时间差量化TC厚度。根据TGO极薄且电导率低的特点，依据涡流法频率增加可提高分辨率的特性，构建基于解析模型的TGO和BC厚度高频涡流反演算法，重点探究如何减小模型与实验偏差、选取激励频率和补偿提离效应以提高精度，并应用蒙特卡洛算法评价检测结果可靠性。项目有望形成一种全面无损检测热障涂层厚度的新方法，具有非接触、分辨率高、精度高等优点。

热障涂层具有耐高温、低导热、防腐蚀等优点被广泛应用于航空发动机。长期服役于恶劣且复杂的工作环境导致热障涂层内部结构容易出现异常，因此对其进行定期无损检测与评估至关重要。热障涂层属于复杂多层结构，由陶瓷层（TC）、氧化层(TGO)、金属粘结层（BC）和合金基体组成，单一检测手段难以完成对TC、TGO和BC厚度精确测量。为此，本项目提出了一种基于脉冲太赫兹/高频涡流热障涂层复合检测新方法。首先，构建了热障涂层脉冲太赫兹反射信号解析模型，设计了光学参数提取与厚度检测方法，建立了折射率、消光系数、TC厚度与时域信号特征的数学关系。其次，提出了基于机器学习与理论模型的TC缺陷分类算法，实现了多类缺陷自动识别，并结合太赫兹图像噪声产生机理建立了基于教与学优化算法的去噪手段，消除了噪声对缺陷量化影响。然后，构建了热障涂层高频涡流信号解析模型，提出了BC与TGO厚度反演算法，并利用蒙特卡洛对反演结果进行了不确定度评估。最后，研究了BC厚度与电导率变化对涡流信号的影响，设计了涡流解析模型校正方法，减小了仿真与实验信号的偏差，提高了厚度测量精度。本项目研究过程中发表学术论文8篇（全部标注本项目批准号61701500，第一标注5篇，第二标注2篇，第三标注1篇），其中SCI论文6篇，核心期刊2篇；参加国内会议和展览16人次；申请并授权发明专利2项；获取软件著作1项；培养硕士研究生9人（已毕业6人）；成功研发了高频涡流检测仪器，并销售1台；搭建了太赫兹检测仪器样机。项目成果为热障涂层各层厚度测量提供了理论依据，已形成一种非接触式、高分辨率、高精度的无损检测与评估新方法。