CRACK DETERMINATION IN A 2-D SCALAR WAVE FIELD USING CONVOLUTIONAL NEURAL NETWORKS

DOI

Bibliographic Information

Other Title
  • 畳み込みニューラルネットワークを用いた2次元スカラー波動場におけるクラック決定解析

Search this article

Description

This paper considers the estimation of the location, angle, and length of a crack in a 2-D scalar wave field using convolutional neural networks (CNN). In a nondestructive evaluation especially for a crack determination, the analysis of scattered ultrasonic waves is important. In this paper, a CNN is trained using waveform data generated by the time domain boundary element method (TD-BEM) to construct a model for estimating the location, angle, and length of a crack. Inputting images generated using data obtained with two different measurement settings to the CNN, we carry out the determination of the crack. The experimental results confirm that the proposed method can predict the location, angle, and length of a crack with accuracy for unknown data.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390303932781669888
  • DOI
    10.60443/jascome.24.0_121
  • ISSN
    27593932
    13485245
  • Text Lang
    ja
  • Data Source
    • JaLC
  • Abstract License Flag
    Allowed

Report a problem

Back to top