Identification of induction motor at standstill using artificial neural network

A Bechouche, H Sediki, DO Abdeslam… - IECON 2010-36th …, 2010 - ieeexplore.ieee.org
IECON 2010-36th Annual Conference on IEEE Industrial Electronics …, 2010ieeexplore.ieee.org
This work deals with electric parameters identification of the induction motor at standstill
using Adaline (ADAptive LInear NEuron) which is a type of ANN (artificial neural network).
We will show that, at standstill, the motor can be expressed by a second order differential
equation linking the current to the voltage. This second order transfer function can be
approximated by two first order differential equations: one valid at the low frequencies and
the other at the high frequencies. This decomposition in two subsystems, slow and fast …
This work deals with electric parameters identification of the induction motor at standstill using Adaline (ADAptive LInear NEuron) which is a type of ANN (artificial neural network). We will show that, at standstill, the motor can be expressed by a second order differential equation linking the current to the voltage. This second order transfer function can be approximated by two first order differential equations: one valid at the low frequencies and the other at the high frequencies. This decomposition in two subsystems, slow and fast, reduce much constraints related to the experiment. The identification, by ANN, of the coefficients corresponding to the two differential equations, enables us to go back easily to the electric parameters of the motor. The use of Adaline as identifier is not fortuitous; it is possible to interpret physically the weights, which is not done generally with the multi-layer neural network. In our case these weights merge with the constant parameters of the discretized equations.
ieeexplore.ieee.org
以上显示的是最相近的搜索结果。 查看全部搜索结果