Renovated controller designed by genetic algorithms
โ Scribed by Tzu-Kang Lin; Yi-Lun Chu; Kuo-Chun Chang; Chia-Yun Chang; Hua-Hsuan Kao
- Publisher
- John Wiley and Sons
- Year
- 2009
- Tongue
- English
- Weight
- 519 KB
- Volume
- 38
- Category
- Article
- ISSN
- 0098-8847
- DOI
- 10.1002/eqe.863
No coin nor oath required. For personal study only.
โฆ Synopsis
Abstract
A novel smart control system based on genetic algorithms (GAs) is proposed in this paper. The system is comprised of three parts: the fiber Bragg grating (FBG) sensorโbased sensing network for structural health monitoring, the GAโbased location optimizer for sensor arrangement, and the GAโbased controller for vibration mitigation under external excitation. To evaluate the performance of the proposed system, an eightโstory steel structure was designed specifically to represent a structure with large degrees of freedom. In total 16โFBG sensors were deployed on the structure to implement the concept of a reliable sensing network, and to allow the structure to be monitored precisely under any loading. The advantage of applying a large amount of information from the sensing system is proven theoretically by the GAโbased location optimizer. This result greatly supports the recent tendency of distributing sensors around the structure. Two intuitive GAโbased controllers are then proposed and demonstrated numerically. It is shown that the structure can be controlled more effectively by the proposed GAโstrain controller than by the GAโacceleration controller, which represents the traditional control method. A shaking table test was carried out to examine the entire system. Experimental verification has demonstrated the feasibility of using this system in practice. Copyright ยฉ 2008 John Wiley & Sons, Ltd.
๐ SIMILAR VOLUMES
as by rigorous 2.5D EM analysis for design verification, and good broadband performance is indeed obtained.
In this study we composed a recurrent neural network learning controller and applied it to the swinging up and stabilization problem of the inverted pendulum. A recurrent neural network was trained by a genetic algorithm which had an internal copy operator or inter-individual copy operator. An appro