Fault Diagnosis System of Electromagnetic Valve Using Neural Network Filter
-
- HAYASHI Shoji
- Nuclear Power and Energy Safety Engineering, Graduate School of Engineering, Fukui University
-
- ODAKA Tomohiro
- Graduate School of Engineering, Fukui University
-
- KUROIWA Jousuke
- Graduate School of Engineering, Fukui University
-
- OGURA Hisakazu
- Graduate School of Engineering, Fukui University
Bibliographic Information
- Other Title
-
- ニューラルネットワークフィルタを用いた空気圧電磁弁の音響故障診断
- ニューラル ネットワーク フィルタ オ モチイタ クウキアツ デンジベン ノ オンキョウ コショウ シンダン
Search this article
Abstract
This paper is concerned with the gas leakage fault detection of electromagnetic valve using a neural network filter. In modern plants, the ability to detect and identify gas leakage faults is becoming increasingly important. The main difficulty in detecting gas leakage faults by sound signals lies in the fact that the practical plants are usually very noisy. To solve this difficulty, a neural network filter is used to eliminate background noise and raise the signal noise ratio of the sound signal. The background noise is assumed as a dynamic system, and an accurate mathematical model of the dynamic system can be established using a neural network filter. The predicted error between predicted values and practical ones constitutes the output of the filter. If the predicted error is zero, then there is no leakage. If the predicted error is greater than a certain value, then there is a leakage fault. Through application to practical pneumatic systems, it is verified that the neural network filter was effective in gas leakage detection.<br>
Journal
-
- Transactions of the Atomic Energy Society of Japan
-
Transactions of the Atomic Energy Society of Japan 7 (3), 186-193, 2008
Atomic Energy Society of Japan
- Tweet
Keywords
Details 詳細情報について
-
- CRID
- 1390282680163951488
-
- NII Article ID
- 10021307502
-
- NII Book ID
- AA11643165
-
- ISSN
- 21862931
- 13472879
-
- NDL BIB ID
- 9629307
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- Crossref
- CiNii Articles
-
- Abstract License Flag
- Disallowed