SPEED CONTROL OF PMSM DRIVE USING ANFIS BASED SPEED CONTROLLER

Abstract

Nowadays, PMSM drive systems are being employed in many industrial applications ranging from low to
medium power because of their simple structures, ease of maintenance and efficiency. However the non-linear behavior
which arises mainly from motor dynamics and load characteristics and the presence of uncertainties makes control an
extremely difficult task. So, the speed control technique should be adaptive and robust for better performance of the
drive. The conventional PID controllers are used to control speed of the drive. Various tuning techniques are employed
such as Z-N tuning, Cohen- Coon, manual tuning etc. of which Z-N tuning is the most widely used method. Since, PID
controllers work well under linear operating conditions and fails when non-linearity arises. Therefore, Artificial
Intelligence techniques are being implemented to achieve better performance either in isolation i.e. Fuzzy, ANN, GA,
PSO etc. or in hybrid as Artificial Neuro Fuzzy Inference System. This project work presents design and simulation of a
modern approach of speed control of PMSM using ANFIS based speed controller. Hence an artificial intelligence control
strategy namely ANFIS has been used as it requires only a reduced computation power while maintaining satisfactory
static and dynamic performance and a good insensitivity to perturbations and parameter uncertainties. A model of
PMSM drive is simulated under various dynamic conditions like starting, load application and removal, speed reversal.
The combined capability of Neuro Fuzzy Controller in handling uncertainties and learning process is proven to be
advantageous in modeling highly non-linear systems. The designed ANFIS based speed controller is presented to develop
a dynamic model of PMSM as it does not require the explicit mathematical model of the plant and automatically takes
care of changing conditions as well. Hence ANFIS based speed controller gives better performance of the drive with
linear as well as non-linear load as compared to conventionally used Z-N tuned PID speed controller.