Abstract:
This thesis presents design and simulation of Artificial Neural Network based speed estimation of
induction motor using integer order (IOPI) and fractional order controller (FOPI) for control of
IM drives. Induction motors are the most widely used motors in industrial application control
systems. Vector control methods require speed sensor which is mounted on the shaft, but their
installation makes the drive system bulky, unreliable, expensive and even installing them might
not be feasible in areas where motor drives in hostile environment or high-speed drives. Induction
motor drives without direct speed sensors have the features of low cost, high reliability, better
noise immunity and less maintenance requirements therefore, problems due to mechanical
position sensor of motor during speed control are alleviated by designing an estimation technique
for estimating an instantaneous speed of the rotor. In this thesis, the proposed estimation method
is based on ANN to obtain the the rotor speed signal. The ANN is used as estimator, trained by
Levenberg- Marquardit algorithm. The data for training the ANN estimator are obtained from the
fractional order controller of the induction motor simulations when the motor drive is working in
closed loop at various values of speeds and loads for speed observation. Stator voltages and
currents are used as an input of ANN and rotor speed as an output of ANN. The complete drive
system is modelled using MATLAB®2020a. This thesis also presents design of fractional order
PI controller, integer order PI controller for indirect vector-controlled induction motor. The error
of simulation result between actual and estimated speed have been less than 0.28% for transient
response, 0.21% for speed tracking. The parameters of the two controllers were genetically
optimized using square of error as a fitness function. The performance of fractional order
controller was compared with integer order PI controllers using MATLAB simulation results
with different operating conditions. It was observed from the simulation results that by using
FOPI and IOPI controllers, for the reference speed of 100 rad/sec, the fractional order controller
has shorter settling time of 0.32 sec and has less steady state error of 0.016 radsec in comparison
with the conventional integer order PI controller which has settling time of 0.64 sec and steady
state error of 0.023 radsec. FOPI controller showed better performance than IOPI controller for
IM drive, this is because of FOPI controller has one additional parameter for tuning which is
integration order.
viii
Keywords: Induction Motor, Neural Network, Speed Control, Field Oriented Control.