飞机平尾夹芯蜂窝复合材料结构的超声智能检测技术

doi:10.19936/j.cnki.2096-8000.20251028.013

复合材料科学与工程 ›› 2025, Vol. 0 ›› Issue (10): 83-90.DOI: 10.19936/j.cnki.2096-8000.20251028.013

飞机平尾夹芯蜂窝复合材料结构的超声智能检测技术

涂思敏¹, 陈振华^1*, 章俊燕², 涂东坤², 徐云林², 卢超¹

1.南昌航空大学无损检测技术教育部重点实验室,南昌 330063;
2.江西洪都航空工业集团有限责任公司,南昌 330000

收稿日期:2024-09-20 出版日期:2025-10-28 发布日期:2025-12-02
通讯作者: 陈振华(1982—),男,博士,教授,硕士生导师,研究方向为超声波无损检测技术及系统,zhenhuachen@yeah.net。
作者简介:涂思敏(2000—),女,硕士研究生,研究方向为超声波无损检测技术及系统。
基金资助:
国家自然科学基金(12464059);江西省重点研发计划项目(20212BBE51006)

Ultrasonic intelligent testing technology for sandwich honeycomb composite structure of aircraft horizontal tail

TU Simin¹, CHEN Zhenhua^1*, ZHANG Junyan², TU Dongkun², XU Yunlin², LU Chao¹

1. Key Laboratory of Nondestructive Testing of Ministry of Education, Nanchang Hangkong University, Nanchang 330063, China;
2. Jiangxi Hongdu Aviation Industry Group, Nanchang 330000, China

Received:2024-09-20 Online:2025-10-28 Published:2025-12-02

摘要/Abstract

摘要： 飞机平尾的蜂窝复合材料结构尺寸大、材料结构复杂、质量要求高,喷水式超声聚焦成像检测技术可实现对蜂窝结构的成像检测;而大量检测图像的评价依赖于技术人员丰富的工程经验和高强度的工作,不可避免地会因主观因素的影响导致评价可靠性变差。由此,提出基于深度学习网络的飞机平尾蜂窝复合材料超声C扫描检测图像的智能识别技术。首先,通过喷水式超声聚焦检测方法采集飞机平尾C扫描检测图像,构建和扩充飞机平尾超声检测图像数据集;其次,基于检测图像对应的检测信号幅度分布,将检测图像按粘接完好性程度划分为三个目标区域类别;第三,构建Faster R-CNN网络并对其进行优化,形成蜂窝复合材料结构超声C扫描区域微小特征变化的智能识别网络;最后,通过试验方法测定智能识别模型的性能,验证其评价蜂窝结构超声C扫描图像的能力。研究结果表明:基于深度学习的智能模型对蜂窝复合材料分类识别的平均准确率均值达到88.2%,对粘接状态最差区域(三类区域)的识别平均准确率可达91.9%,能够用于分类统计蜂窝复合材料结构超声C扫描检测图像。

关键词: 蜂窝复合材料, Faster R-CNN, 喷水式超声聚焦检测, 深度学习, 智能识别

Abstract: The honeycomb composite structure of aircraft horizontal tail has large size, complex material structure and high quality requirements. The water-jet ultrasonic focusing imaging detection technology can realize the imaging detection of honeycomb structure. The evaluation of a large number of detected images depends on the rich engineering experience and high-intensity work of technicians, which inevitably leads to poor evaluation reliability due to the influence of subjective factors. Therefore, an intelligent recognition technology of ultrasonic C-scan detection image of honeycomb composite material of aircraft horizontal tail based on deep learning network is proposed. Firstly, the C-scan detection image of the horizontal tail of the aircraft is collected by the water-jet ultrasonic focusing detection method, and the data set of the ultrasonic detection image of the horizontal tail of the aircraft is constructed and expanded. Secondly, based on the amplitude distribution of the detection signal corresponding to the detection image, the detection image is divided into three target area categories according to the degree of bonding integrity. Thirdly, the Faster R-CNN network is constructed and optimized to form an intelligent recognition network for small feature changes in the ultrasonic C-scan area of honeycomb composite structures. Finally, the performance of the intelligent recognition model was measured by experimental methods to verify its ability to evaluate the ultrasonic C-scan images of honeycomb structures. The research results show that the average accuracy of the intelligent model based on deep learning for the classification and recognition of honeycomb composite materials reaches 88.2%, and the average accuracy of the worst bonding area (three class of region) can reach 91.9%, which can be used for classification and statistics of ultrasonic C-scan detection images of honeycomb composite structures.

Key words: honeycomb composites, Faster R-CNN, water-jet ultrasonic focus testing, deep learning, intelligent recognition

中图分类号:

TB332

涂思敏, 陈振华, 章俊燕, 涂东坤, 徐云林, 卢超. 飞机平尾夹芯蜂窝复合材料结构的超声智能检测技术[J]. 复合材料科学与工程, 2025, 0(10): 83-90.

TU Simin, CHEN Zhenhua, ZHANG Junyan, TU Dongkun, XU Yunlin, LU Chao. Ultrasonic intelligent testing technology for sandwich honeycomb composite structure of aircraft horizontal tail[J]. COMPOSITES SCIENCE AND ENGINEERING, 2025, 0(10): 83-90.

参考文献

[1] KUMAR N C, VASANTHANATHAN A. State-of-the-art review on honeycomb sandwich composite structures with an emphasis on filler materials[J]. Polymer Composites, 2021, 42(10): 5011-5020.
[2] 吴斌, 宋协宏, 高杰, 等. 厚胶层复合材料弱粘接结构的超声反射特性[J]. 声学学报, 2024, 49(1): 137-145.
[3] 檀桢, 王明泉, 刘康驰, 等. 蜂窝夹芯结构脱黏的空气耦合超声检测技术研究[J]. 压电与声光, 2021, 43(6): 799-804.
[4] MIAO H, FAN J. Debonding failure analysis of honeycomb sandwich composite radome[J]. Journal of Physics: Conference Series, 2021, 1865(3): 032068.
[5] 陈韩青, 徐志远, 屈仲毅, 等. 蜂窝夹层结构无损检测方法研究综述[J]. 材料导报, 2024, 38(10): 296-310.
[6] RUSAKOV D Y, CHERNUSHIN V A, KURYATIN A A. Features of the research of the structure and assessment of the quality of composite materials by means of non-destructive testing techniques[J]. Journal of Physics: Conference Series, 2021, 1758(1): 012033.
[7] 郝威, 李明, 王珏, 等. 复合材料芳纶蜂窝夹层结构件特殊缺陷检测[J]. 航空学报, 2022, 43(7): 438-446.
[8] 杨涵, 李斯涵, 舒江鹏, 等. 基于阵列超声和特征融合神经网络的钢筋混凝土结构内部裂缝检测[J]. 建筑结构学报, 2024, 45(7): 89-99.
[9] 余泽禹, 袁洪强, 卫小龙, 等. 基于深度学习的管道焊缝超声检测缺陷识别方法[J]. 科学技术与工程, 2022, 22(30): 13288-13292.
[10] 胡文博, 邱实, 许馨月, 等. 基于深度学习的钢轨伤损超声检测与分类[J]. 铁道学报, 2021, 43(4): 108-116.
[11] KHUMAIDI A, YUNIARNO E M, PURNOMO M H. Welding defect classification based on convolution neural network(CNN) and Gaussian kernel[C]//2017 International Seminar on Intelligent Technology and Its Applications (ISITIA). Surabaya: IEEE, 2017: 261-265.
[12] 李容, 冯侃, 闫静, 等. 基于YOLOv5s的复合材料层合板层裂损伤局部稳态波场识别研究[J]. 固体力学学报, 2023, 44(5): 672-678.
[13] 薛阳, 张舒翔, 贾巍, 等. 基于改进Faster RCNN的电缆外护套破损检测[J]. 电子测量技术, 2023, 46(15): 158-164.
[14] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas: IEEE, 2016: 770-778.
[15] 彭伊娟, 王振超, 张秋菊. 改进的Faster RCNN算法在聚乙烯管接头内部缺陷检测中的应用[J]. 应用声学, 2023, 42(5): 984-992.
[16] 蒋占军, 吴佰靖, 马龙, 等. 多尺度特征和极化自注意力的Faster-RCNN水漂垃圾识别[J]. 计算机应用, 2024, 44(3): 938-944.
[17] 罗仁泽, 唐祥, 余泓, 等. 基于改进Faster RCNN的射线图像焊缝缺陷检测方法[J]. 电子测量技术, 2023, 46(22): 160-168.
[18] 刘应桃, 郭世伟, 付孟新, 等. 基于改进Faster RCNN的轮对踏面缺陷检测[J]. 电子测量技术, 2023, 46(12): 34-41.
[19] 包妮沙, 韩子松, 于嘉欣, 等. 基于改进Faster-RCNN的露天煤矿开采区遥感识别方法[J]. 东北大学学报(自然科学版), 2023, 44(12): 1759-1768.

飞机平尾夹芯蜂窝复合材料结构的超声智能检测技术

Ultrasonic intelligent testing technology for sandwich honeycomb composite structure of aircraft horizontal tail

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

编辑推荐

Metrics

本文评价

[1]	李磊磊, 王明泉, 赵付宝, 朱焕宇, 丰晓钰, 谢绍鹏. 基于Mask R-CNN的复合材料夹杂缺陷自动检测研究[J]. 复合材料科学与工程, 2024, 0(1): 83-88.
[2]	张睿, 文传博. WTB-Net:基于ShuffleNet V2的轻量级风机叶片表面缺陷识别算法[J]. 复合材料科学与工程, 2023, 0(8): 66-71.
[3]	白国栋, 童小燕, 姚磊江. 基于深度学习的复合材料铺层优化方法[J]. 复合材料科学与工程, 2020, 0(7): 68-73.