Research On Alglorithm Of Structural Control Based On Neural Network And Fuzzy Logic

The algorithm of structural control is an important factor of active vibration control,and the kernel part of intelligent structure system. The paper aims at the integration of BP neural network and fuzzy logic to seek a new vibration control algorithm for adapting to nonlinear structure and large flexible structure of which the model is difficult to obtain.Firstly,the basic principle and the recent development of structural control are introduced in this paper,and are stressed on structural control algorithm. Secondly,the basic theory and algorithm of BP neural network and fuzzy control are presented,and to characteristics of BP neural network and fuzzy logic are discussed respectfully. In this paper a new active control algorithm,the neural network and fuzzy algorithm is proposed by making use of their superiority to the full.In this algorithm the input values are displacement and velocity,and control force is output value. They are determinate by mechanical analyses of structure vibration and fuzzy control. The fuzzy control rule database is built by using sample data to train BP neural network,the control force is calculated by exploiting fuzzy logic in terms of the fuzzy rule database which has been built,and the vibration is decreased by bringing to bear the control force. The advantages of neural network are leamable;seasoning no known and non-determination system,and the superiority of fuzzy control is fuzzy reasoning that imitates person brain. The integration of them enables the algorithm possess the advantages of self-learning,robust and handling nonlinear easily.Lastly,to show the effectiveness of the algorithm,the simulation study for vibration-suppression of cantilever-beam is carried out. The results show that the algorithm has remarkable effect on suppressing structure vibrations,excited under the periodic and random excitations. By reason of its no-dependency on structure,the algorithm has importance for nonlinear structure and large flexible structure of which the model is difficult to obtain.