| |
A NEURAL ROTOR FLUX OBSERVER FOR INDUCTION MOTOR CONTROL
| [View]
[Download]
|
| Author(s) |
A. Alessandri; S. Dagnino; M. Marchesoni; M. Mazzucchelli; P. Saccani |
| Abstract |
Rotor flux observers can provide an useful mean for achieving field oriented control of induction machine. Estimating the state of a nonlinear system, through a noisy measurement channel, is still a challenge and the aim of this paper is to describe how multilayer feedforward neural networks can be trained successfully in the case of rotor flux
estimation. The method presented exploits off-line training on a large set of learning patterns. Both steady-state and transients conditions over full operative speed range have been considered.
The basic idea of this approach is to constrain the optimal state estimator to take the structure of a feedforward neural network. This strategy involve a potential leak of accuracy,
nevertheless approximating features of neural nets allow to achieve excellent results. The rotor flux observer bas been trained and tested by means of numerical simulations. Computer
results establish the effectiveness and the robustness of the suggested method. |
| Download |
| Filename: | Unnamed file |
| Filesize: | 413.6 KB |
|
| Type |
Members Only |
| Date |
Last modified 2018-04-10 by System |
| |
 |