|
| 一种基于Lamb波的智能结构损伤识别方法 |
| An intelligent structural damage identification method based on Lamb wave |
| Received:May 30, 2021 Revised:June 07, 2021 |
| DOI：10.7520/1001-4888-21-134 |
| 中文关键词: Lamb波 卷积神经网络模型 数据增强技术 结构损伤识别 智能检测 |
| 英文关键词:Lamb wave convolutional neural network model data enhancement technology structural damage identification intelligent detection |
| 基金项目:国家自然科学重点国际合作项目(No.11520101001) |
|
| Hits: 632 |
| Download times: 108 |
| 中文摘要: |
| 常规基于超声Lamb波的损伤检测方法依赖于精确的结构物理模型、繁琐的信号处理技术以及受限于不同的成像算法，致使损伤成像的精度和效率难以兼顾。针对高端装备关键结构快速、精准的智能化损伤检测需求，本文构建以压电片激励、激光多普勒测振仪拾振的检测平台以研究卷积神经网络自动提取Lamb波含损伤特征信息的方法。对结构内损伤待检测区域进行合理分区，将对损伤的检测转化为基于卷积神经网络的图像分类任务。依据损伤所在位置、大小的变化建立数据库，并使用数据增强技术扩大数据库；构建基于卷积神经网络的智能结构损伤识别模型用以建立结构损伤区域与Lamb波信号特征的映射关系，进而实现对板结构损伤的快速智能识别。实验验证结果表明：将卷积神经网络运用于基于Lamb波的损伤识别方法中，在实现快速、智能化的损伤检测方面有潜在的工程应用前景。 |
| 英文摘要: |
| Conventional damage detection methods based on ultrasonic Lamb wave rely on accurate physical models, cumbersome signal processing techniques and limited by different imaging algorithms, which makes it difficult to balance the accuracy and efficiency of damage imaging. Aiming at the rapid and accurate intelligent damage detection requirements of key structures of high-end equipment, the method of automatically extracting Lamb wave damage features information based on the constructed convolution neural network by using the detection platform of piezoelectric excitation and Scanning Laser Doppler Vibrometer pickup. The area of damage to be detected in the structure is divided reasonably, and the damage detection is transformed into an image classification task based on convolution neural network. The database was established according to the change of the location and size of the damage, and the data enhancement technology is used to expand the database. The intelligent damage identification model based on convolution neural network is used to establish the mapping relationship between damage area and Lamb wave signal features, so as to realize the rapid and intelligent identification of damage. The experimental testing prove that the application of convolutional neural network to damage recognition based on Lamb waves has potential engineering application prospects in realizing fast and intelligent damage detection. |
| View Full Text Download reader |
| Close |
