桥梁振动控制系统的GA-BP神经网络非参数辨识方法研究

Non-parameter identification of the bridge vibration control system based on GA-BP networks

文章提出了一种基于GA-BP神经网络的地震激励下桥梁振动控制系统的非参数辨识方法,该方法避免了BP算法易陷入局部极小值、训练速度慢、误差函数必须可导及受网络结构限制等缺陷,克服了传统系统辨识要对结构进行预先假定、容错能力差、无法辨识系统非线性特性等问题,考虑到结构-控制装置的相互作用,可以真实反映桥梁结构动力特性。

A new structural system identification method based on GA-BP networks is presented. The new system identification method helps to avoid some disadvantages of the BP algorithm, which is easy to fall into the local minimum and limited by network structure, has a slow training speed and needs derivable error functions,and the new identification method also overcomes the drawbacks of the traditional system identification methods,which need presupposition of the structure, have poor fault-tolerant ability and are unable to identify the system＇s nonlinear characteristic. The interaction between the structure and the actuators is considered, and the true dynamic performance of the bridge is reflected.