Investigations on Intelligent Control Algorithms for the Design of Speed Controller in Variable Frequency Induction Motor Drives

Abstract

Variable Frequency Induction Motor Drives (VFIMDs) are the vertebrae of the industrial world today. With the advent of advanced speed control techniques for high power VFIMDs, it is imperative to study their dynamic behavior under different operating conditions like starting, braking, load dynamics and field weakening, to which these drives are frequently subjected to. To analyze the dynamic response of VFIMDs, their speed control is normally achieved by deploying Field Oriented Control (FOC) and Direct Torque Control (DTC) schemes. An attempt has been made in this thesis to obtain an improved performance of the drive, in terms of reduced torque ripple and reduced transient response specifications such as rise time, peak overshoot, peak time and settling time, using new/modified controllers. In order to overcome the shortcoming of three level torque controller, a model of five-level torque controller has been developed in this research work. The thesis therefore, is directed towards investigating the various AI techniques in order to properly design and tune the speed controller for the direct vector controlled and for the five-level torque controller based DTC induction motor drives. The four popular AI techniques: viz. Fuzzy Logic (FL), Artificial Neural Network (ANN), Genetic Algorithm (GA) and Adaptive Network Based Fuzzy Inference System (ANFIS) have been considered for the above purpose. Their performances have been compared with that of the conventional Zeigler-Nichols (Z-N) method tuned PI speed controller of VFIMDs. In the course of the research work, the two AI techniques (FL and ANFIS) have been used for designing the speed controller. The performance of VFIMDs has been analyzed by replacing the existing PI controller by these AI controllers. Further, the two AI techniques (ANN and GA) have been used for tuning the existing PI controller. The comparisons of the proposed techniques with the conventional one, and those among themselves facilitate need based selection of VFIMDs for various applications.