You are on page 1of 12

CONTROL OF INDUCTION

MOTORS BY ARTIFICIAL
NEURAL NETWORKS
OBJECTIVE

 The objective of this project is to control the speed of an


induction motor using neural networks.
 The variable motor speed and current is obtained
experimentally.
 Based on the obtained data artificial neural network is
trained using back propagation algorithm .
 The neural network model of the induction motor is obtained
using MATLAB software.
 Neural network model predictive controller is designed for
the induction motor. The performance of motor is to be
compared with open loop and closed loop responses.
FUNCTIONAL BLOCK DIAGRAM
NEURAL NETWORK PREDICTIVE
CONTROLLER
• The predictive controller predicts future plant performance
• The controller then calculates the control input that will
optimize plant performance.
BACK PROPAGATION ALGORITHM

Steps involved in training process

 Assemble the training data


 Create the network object
 Train the network
 Simulate the network response to new inputs
FORMING NETWORK OBJECT

• The function “newff” forms the network object.


• The function newff requires four inputs ,

 Minimum and maximum values of the input vector


 Sizes of each layer
 Transfer function to be used in each layer
 Training function to be used
EXAMPLE

net=newff([-1 2; 0 5],[3,1],{'tansig','purelin'},'traingdm')
 The default performance function for feed forward network is
mean square error( the average squared error between the
network outputs a and the target outputs t.)
TRAINING RESULTS
Performance is 7.61696e-005, Goal is 1e-005
1
10

0
10
T raining-B lue G oal-B lac k

-1
10

-2
10

-3
10

-4
10

-5
10

-6
10
0 50 100 150 200 250 300
300 Epochs
Performance is 1.67776e-005, Goal is 1e-005
1
10

0
10

-1
Training-Blue Goal-Black

10

-2
10

-3
10

-4
10

-5
10

-6
10
0 500 1000 1500 2000 2500
2500 Epochs
MODEL OUTPUT
TRANSFER FUNCTION OF THE MOTOR
THANK YOU

You might also like