133 制御入力の大きさと位相の異常に対応可能な知的振動制御システム

概要

論文の詳細を見る
With progress of technology in recent years, gigantism and complication such as high-rise buildings, nuclear reactors and so on have brought about new problems. Particularly, the safety and the reliability for damages in abnormal situations have become more important. Intelligent control systems which can judge whether the situation is normal or abnormal at real time and cope with these situations suitably are demanded. In this study. Cubic Neural Network (CNN) is adopted, which consists of the controllers possessing cubically some leve s of information abstracting. In addition to the usual quantitative control, the qualitative control is used for the abnormal situations. And, by selecting a suitable controller, CNN can cope with the abnormal situation. In order to confirm the effectiveness of this system, the structural vibration control problems with sensory failure and tural vibration control problem with and elasto-plastic response is dealt with. As a result of simulations, it was demonstrated that CNN can cope with unexpected abnormal situations which are not considered in learning.
一般社団法人日本機械学会の論文
2000-09-01