Scholarlink

Introduction <p>Non-destructive testing (NDT) has been employed since the early 1900s for maintenance and quality control, typically requiring partial or full disassembly of components. In contrast, Structural Health Monitoring (SHM), introduced in the 1980s, enables continuous, in-situ monitoring using low-cost sensors. However, inspecting additively manufactured (AM) parts remains challenging due to their complex geometries.</p> Objectives <p>This study investigates the integration of NDT and SHM techniques to estimate applied loads and detect damage in AM parts, aiming to enhance inspection and monitoring capabilities.</p> Methods <p>Five stainless steel specimens with varying defects were examined using Ultrasonic Contact Transducer (UCT) and Electromagnetic Acoustic Transducer (EMAT) probes from the NDT system, while SHM sensors simultaneously recorded the structural response. The collected data were processed using Fast Fourier Transform (FFT) and Continuous Wavelet Transform (CWT) methods to train and evaluate deep learning models, including Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks.</p> Results <p>A 2D CNN trained on CWT scalograms achieved 100% classification accuracy. LSTM and 1D CNN models also performed well, with accuracies ranging from 90% to 100%.</p> Conclusion <p>This study demonstrates the feasibility of integrating NDT and SHM systems for real-time load estimation and structural health assessment of AM parts. The hybrid NDT-SHM approach offers a promising pathway for improving the precision and efficiency of inspection and monitoring processes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Inspection of Additively Manufactured Structures Using Combined Nondestructive Testing (NDT) and Structural Health Monitoring (SHM) Systems

  • 期刊论文

  • 发表日期：

    2025年6月11日

查看全文
  • Romaine Byfield,
  • Ahmed Shabaka,
  • Milton Molina Vargas,
  • Ibrahim Tansel

摘要

Introduction

Non-destructive testing (NDT) has been employed since the early 1900s for maintenance and quality control, typically requiring partial or full disassembly of components. In contrast, Structural Health Monitoring (SHM), introduced in the 1980s, enables continuous, in-situ monitoring using low-cost sensors. However, inspecting additively manufactured (AM) parts remains challenging due to their complex geometries.

Objectives

This study investigates the integration of NDT and SHM techniques to estimate applied loads and detect damage in AM parts, aiming to enhance inspection and monitoring capabilities.

Methods

Five stainless steel specimens with varying defects were examined using Ultrasonic Contact Transducer (UCT) and Electromagnetic Acoustic Transducer (EMAT) probes from the NDT system, while SHM sensors simultaneously recorded the structural response. The collected data were processed using Fast Fourier Transform (FFT) and Continuous Wavelet Transform (CWT) methods to train and evaluate deep learning models, including Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks.

Results

A 2D CNN trained on CWT scalograms achieved 100% classification accuracy. LSTM and 1D CNN models also performed well, with accuracies ranging from 90% to 100%.

Conclusion

This study demonstrates the feasibility of integrating NDT and SHM systems for real-time load estimation and structural health assessment of AM parts. The hybrid NDT-SHM approach offers a promising pathway for improving the precision and efficiency of inspection and monitoring processes.