Research On Neural Network Control Of Structural Control

China is the conutry which is destroied enormously in the earthquake. The civil engineering structure vibration control is a positive effective method and technology to ensure the structure aseismatic and against natural disasters. It may effectively reduce the damage accumulates of the structure under dynamic action the vehicles, wind and earthquake. The development of structure intelligence control research can make the uncertain system control to be possible. The nerve network controller is taught to automatically learning system characteristic, thus adapting the system along with the time characteristic variation, and achieving the system optimum control. The unique from functions of Nerve Nentwork, such as self-study, self- organization, association memory and parallel processing, are good to enhance the structure resistance ability, reduce the loss in the earthquake.Firstly, the Nerve Network control is applied to the tall building with annex construction structure earthquake responded control, the structural motion equation and the state space form equation of the earthquake-excited the tall building and annex's structural model are analyzed, the earthquake response model of single MR damper structure and the MR damper's mathematical model are studied and the method of how to apply nerve network control principle to the tall building and annex's structural control as well as how to design the controller are discussed.Secondly, the improved BP Nerve Network study algorithm is researched in the dissertation, the Nerve Network training sample is discussed, and the improvement BP Nerve Network training algorithm using the MATLAB language is realized. This algorithm enhanced the network training speed and the precision.Finally, the simulation models of the independent tower, the independent podium and the MR damper are given. The simulation program of the structure under no control and Nerve Network control based on the improved BP study algorithm is designed using Matlab/Simulink Tools. The comparison of the response under no control and Nerve Network control with different earthquake-excited is made.The simulation results show that the Nerve Network control is a simply and effective control arithmetic. It can greatly reduce earthquake response of the tall building with annex and mitigation effect is very available.