| 103C2 |
|---|
| Compensation of Periodic Disturbance in Servo System by Using NN Based Estimator |
| AUTHORS |
| Hiromi Wada, Nobuhiro Kyura, Hiromu Gotanda (Kinki University) |
| ABSTRACT |
| In a mechatronic servo system activated by electric motors, there are many unknown periodic disturbances. A frequency of these disturbances is proportional to an angular frequency of the motor speed. A cogging-torque is a typical example of these disturbances. And this cogging-torque generates a fluctuation of rotational speed of motor. The fluctuation of a motor-speed reduces shape accuracy or surface smoothness of final result of products. But this fluctuation can not be described by theoretical equation based on motor dynamics. So we try to estimate the fluctuation by a neural network. An estimator of the fluctuation is constructed by a neural network and a control system model. The neural network is a three-layer type one. And control system parameters are known roughly, then these values are used to make the system model. Mechatronic servo system of 2nd order is studied in this paper. The neural network based estimator calculates a disturbance and its output is used to compensate the disturbance. We can reduces an average amplitude of the speed fluctuation to 1/3 comparing without compensation. A simulation result is shown in the paper. |
Back to Program
Time Table
Top of This Site
Annual Conference 2002