| |
NEURAL NETWORK CURRENT CONTROL OF VS-PWM INVERTERS
| [View]
[Download]
|
| Author(s) |
M. P. Kazmierkowski; D. L. Sobczuk; M. A. Dzieniakowski |
| Abstract |
This paper presents the application of Neural Network (NN) for control of three phase PWM inverter output currents. Two systems are presented. In the first a single layer perceptron with precalculated weights is applied. With modified sampling technique the NN controller always chooses adjacent active vectors or zero vector. Moreover, this simple one layer NN controller is very robust and can work correctly without information in one of three phases. In the second part the feedforward multilayer based NN controller trained off line by a model PWM pattern generated by optimal current control algorithm is presented. It is shown that three layers feedforward NN is able to regulate PWM inverter output current very well without on-line calculations needed in the case of the optimal controller. The performances of the proposed NN controllers are illustrated by simulation and experimental results. |
| Download |
| Filename: | Unnamed file |
| Filesize: | 355.1 KB |
|
| Type |
Members Only |
| Date |
Last modified 2018-05-16 by System |
| |
 |