Non-destructive investigation of anisotropic materials using air-coupled ultrasound

使用空气耦合超声对各向异性材料进行无损研究

• 批准号: 518269513
• 负责人: Dr. Olga Popovych, Ph.D.
• 金额: --
• 依托单位: Centrum Baustoffe und Materialprüfung (cbm)
• 依托单位国家: 德国
• 项目类别: WBP Position
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/518269513?language=en
• 关键词: Non; destructive; investigation; anisotropic; materials

This project aims to develop database for automation of a broadband, single-sided, contact-free, ultrasound-based non-destructive-testing method for defect characterization. Preliminary own experiments have shown, that laser-based ultrasound (US) generation combined with an innovative laser-based, air-coupled, high-fidelity optical microphone for detection performs well in transmission mode testing. This approach is based on the Local Acoustic Resonance Spectroscopy (LARS) technique and a derivative of the known Air-Coupled Ultrasonic Testing (ACUT) using piezoelectric transducers.First results from single-sided ACUT measurements using a piezoelectric transducer for excitation and an optical microphone for detection in combination with Machine-Learning using a Convolutional Neural Network were promising. Expanding LARS from transmission to single-sided measurement mode and implementing machine learning for data analysis should yield highly interesting results with new possible applications in the industry. For basic materials research the new technique enables for the investigation of components with complex geometry and significant heterogeneity. The objective of this proposal, for which funding is requested, is a deeper analysis of LARS data and proceeding of database of training material that will be used for ML algorithms and further automation of the control process.

本项目旨在开发宽带、单面、无接触、基于超声的无损检测方法的自动化数据库，用于缺陷表征。初步实验表明，基于激光的超声波（US）生成与创新的基于激光的、空气耦合的、高保真光学麦克风相结合，用于检测，在传输模式测试中表现良好。这种方法是基于局部声共振谱（LARS）技术和已知的空气耦合超声检测（ACUT）的衍生物，使用压电transducers.First结果从单侧ACUT测量使用压电换能器的激励和检测结合机器学习使用卷积神经网络的光学麦克风是有前途的。将LARS从传输模式扩展到单侧测量模式并实施机器学习进行数据分析，应该会产生非常有趣的结果，并在行业中产生新的可能应用。对于基础材料研究，新技术可以用于研究具有复杂几何形状和显著异质性的组件。该提案的目的是对LARS数据进行更深入的分析，并对将用于ML算法和控制过程进一步自动化的培训材料数据库进行处理。