機械学習を用いたAEモニタリングにおけるノイズ除去手法の開発

DOI Web Site

書誌事項

タイトル別名
  • Development of Noise Signals Rejection Method in AE Monitoring Using Machine Learning

抄録

<p>The Acoustic Emission (AE) method is used for detecting and preventing serious damages in structures because it can detect small cracks of the material and small deformation before breaking. However, AE signals detected during the test were not only from damage in materials but also noise signals so a method for classification is needed. In this study, our aim is to develop a method to classify noise and AE signals automatically, during the corrosion accelerated test. The corrosion accelerated test was carried out for about 30 days and about 20 days data from the beginning were used for teacher data, and the 10 days data from the last were classified. A neural network, which is a type of machine learning, was used for classification. Classification was performed by combining thirteen kinds of features including general AE parameters as feature quantities for classification. Other result, recognition rate was more than 90%.</p>

収録刊行物

  • 年次大会

    年次大会 2018 (0), J0410202-, 2018

    一般社団法人 日本機械学会

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ