Simulation And System Identification Of Seismic Resistant Structure Base On Neural Network

Identification of a structure system is inevitable requires of structure control and diagnose. It is also an important subject of disaster fortification and mitigation engineering. Cosmically parallelism, getting redundant, getting fault-tolerant, essential no linearity, study adaptive capacity and so on are characteristic of artificial neural network(ANN), so it become the tool for system identification naturally. After studies concepts of ANN, this text have probed the simulation of seismic resistant structure base on ANN, including the collecting of samples, the choosing of performance function, the construction of NN's topology, the choice of training methods and so on. And it works out the programs for the simulation of seismic resistant structure base on Matlab, and analyses some calculating examples. From the examples it summarizes several methods and laws. It brings forward the methods for identification of the dynamical parameter of linear structure base on NN and analyses some calculating examples, too. Finally, it concludes that it is feasible that NN is applied to the simulation of seismic resistant structure. This technique is characteristic of high precision, fast, fault-tolerant, none limiting by the structure model etc. And the identification of the dynamical parameter of linear structure base on NN is feasible, too.