The paper studies the application of neural network in chaotic vibration isolation system. Because the dynamical characteristics of the isolation system change with working state of the isolated object, the control system must have the ability to adjust the system parameters online to obtain the designated target. During the controlling process, chaotic vibration signal identification is the first step. New technology, however, is needed to identify the category of the signals because the traditional methods such as Lyapunov exponent and correlation dimension are time-consuming. The incompact wavelet neural network is trained to classify different signals effectively in this paper, and it is beneficial to adjust the nonlinear isolation system parameters timely. Hence the system works in a chaotic state, and the linear spectrum in the waterborne noise can be reduced.

This content is only available via PDF.
You do not currently have access to this content.