신경회로망을 이용한 유도전동기의 센서리스 속도제어에 관한 연구

Abstract

Recently, with the development of vector control technology, induction motors has been used more in the industrial variable speed drive system.

Generally, induction motor controller requires rotor speed sensors for commutation and current control, but they increase cost and size of the motor.

So in these days, various researches including speed sensorless vector control are reported and some have been put to practical use.

Many sensorless control algorithms use mathematical model based upon analysis techniques which have been used for the steady and transient states of induction motor.

However, their control performances are greatly influenced by the parameters and load deviations. Futhermore, many difficulties occurred in starting and low-speed range.

In this paper a new speed estimation method using neural networks is proposed.

The neural network structure was again found by trial and error and it was found that the 8-16-1 neural network has given correct results for the instantaneous rotor speed.

The eight inputs to the neural network are monitored values of the stator voltages and stator currents. The neural network contains a singles hidden layer with 16 nodes, and the activation functions used in the hidden layer are tansigmoid. The output layer contains a single node, which outputs the rotor speed.

A feedback signal is necessary for only training. Supervised learning methods, through which the neural network is trained to learn the input/output pattern presented, are typically used.

The back-propagation technique is used to adjust the neural network weights during training. The rotor speed is calculated by weights and eight inputs to the neural network.

Also, the proposed methods have advantages such as the independency on machine parameter, the insensitivity to the load condition, and the stability in the low speed operation.

The results of simulation and experiment indicate good response characteristics even in the low speed range and in the parameter variation.