利用结构化神经网络识别振动系统非线性特性

本文提出了振动系统非线性特性识别的结构化神经网络方法。与传统的前馈神经网络不同的是，该法把系统分为线性和非线性两部分，学习得到的神经网络可以单独识别出系统非线性模型，而不是线性与非线性综合在一起的模型。本文将其应用于振动系统非线性特性的识别。实例表明该方法是可行的。

Identification of Nonlinear Characteristics ofVibration System Using Structural Neural Network

A structural neural network model for the identification of nonlinear characteristics of vibration sactern is presented in this paper. The model divides a complex syStem intO linear and nonlinear parts, whichis different from the typical feed-forward neural network. Neural network is used to learn nonlinear characteristics of the system only. The trained neural network can identify nonlinear model of the system independently(not the hybrid model of linear and nonlinear models). The method is used in the paper tO identify nonlinear characteristics of vibration syStems. Results show that the method is feasible.