NEURAL NETWORKS operation speed ESTIMATION for identification of signals in defect detection

Authors

  • O.  Starodubtsev
  • V.  Khandetskyi

Keywords:

neural networks, defect detection, composite materials

Abstract

Using the computer simulation, we determined the degree of influence of structural parameters and learning methods of multilayer perceptrons with back propagation algorithm on their operation speed when identifying the defect detection signals for composite materials. The obtained results have a practical importance for real-time studies.

References

1. Khandetsky V., Antonyok I. Signal processing in defect detection using back-propogation neural networks. - NDT&E International, 35, 2002, p. 483-488.
2. Стародубцев О.Л. Дослідження нейронних мереж для ідентифікації сигналів вихорострумової дефектоскопії. - Системні технології, 1(108), 2017, с.с115-122.
3. Khandetsky V, Gerasimov V, Gnoevoy S. Research of probability characteristics in defect detection of composite materials using wavelet transform. - Int. J. Materials and Product Technology, vol.27, Nos.3/4, 2006, p. 210-220.
4. А.А. Рындин, В.П. Ульев. Исследование скорости обучения нейронных сетей. -Вестник Воронежского гос.техн.университета, Воронеж, 2012, с.7-8.
5. Battiti R. Accelerated backpropagation learning: two optimization methods.- Complex Systems, 1989. Vol.3. - p. 331-342.
6. Ф. Уоссермен. Нейрокомпьютерная техника: Теория и практика. Пер.с англ. - М.: Мир, 1992. - c. 26-54.
7. Haykin S. Neural Networks. A Comprehensive Foundation. Second edition. – New Jersey: Prentice Hall, 2008. – 1103 p.

Downloads

Published

2020-05-04