Active control of plane frames by compatible neural network

Abstract

Controlling the behavior of frame building is very common these days. This goal is achieved by changing the structural behaviors through applying forces to the frames. Recently, extensive studies have been carried out in the field of structural control related to the earthquakes. All studies conducted in this area can be divided into two groups. The first category is devoted to the control devices. Since accuracy and sensitivity of required equipment play an important role, some industries are trying to build better and more robust instruments. The key subject of the second group of researchers is developing new control algorithms. These approaches need some innovations. The purpose of this study is to minimize the structural response against earthquake utilizing two actuators. The purpose of this study is to minimize the structural response against earthquake utilizing two actuators. The relationship between the control forces of the actuators was so arranged that the first mode force becomes zero. In order to minimize the structural responses, the genetic algorithm was used. The controlling system, which is exploited in this paper, is a closed circle. In addition, the neural network was employed to predict the earth acceleration. The authors selected a kind of the neural network to have compatibility with earthquake acceleration variation. To achieve this, the number of the neurons in layers should be varied. The comprehensive experimental numerical results for a variety of earthquakes and structures indicated that the suggested method is very effective. However, the present study drawback is in decreasing the responses of tall frames.

Keywords

• structural control
• neural network
• variable neurons
• genetic algorithm
• predicting earthquake acceleration
• closed circle
• earthquake