| 202D4 |
|---|
| Parameter Estimation of Nonlinear Vibration System by use of Volterra Kernel |
| AUTHORS |
| Hiroshi HARADA, Hiroshi KASHIWAGI, Naoto SARUWATARI (Kumamoto University) |
| ABSTRACT |
| In this research, we identify a nonlinear mechanical system by using Volterra kernels. Most of an actual control system can be considered to be a nonlinear system. Therefore, the identification of the nonlinear system is an important problem in the field of control theory. In general, many classes of nonlinear systems can be represented by Volterra kernels. Many researches have been performed concerning the method of measuring the Volterra series. In a previous research, we obtained the Volterra kernels, g1 and g2 of the nonlinear vibration system by using M-sequence system signal and the cross-correlation function. From the results of computer simulation, it was shown that the maximum values of 1st and 2nd order Volterra kernels are proportional to the damping ratio of the mechanical system. From this proportionality relation, a nonlinear parameter was derived. This paper is theoretically presented by using approximation of the Kernel Invariance Method (KIM). In order to confirm this rerational expression, the derived relational expression and the relational expression which used the approximation of KIM were compared. As a result, it is understood that it is in the proportionality relation as well as the derived relational expression in the approximation of KIM. We estimate the nonlinear system that have nonlinear external force. A nonlinear parameter is estimated based on the proposed identification method. As a result, it has been understood that the estimated nonlinear parameter is almost corresponding to the theoretical value. |
Back to Program
Time Table
Top of This Site
Annual Conference 2002