Induction motor speed drive improvement using fuzzy IP-self-tuning controller. A real time implementation

Area of Science:

• Electrical Engineering
• Control Systems
• Machine Learning

Background:

• Induction machine speed control is crucial in industrial applications.
• Conventional Proportional-Integral (PI) self-tuning controllers require tuning two gains, increasing complexity.
• Existing methods often face challenges in dynamic performance and disturbance rejection.

Purpose of the Study:

• To propose and evaluate a novel Integral-Proportional (IP)-self-tuning controller optimized by a fuzzy logic adjustor for enhanced induction machine speed control.
• To simplify the tuning process by requiring adjustment of only one gain, unlike traditional PI-self-tuning controllers.
• To compare the performance of the proposed controller against classical IP and fuzzy-like-PI controllers.

Main Methods:

• Development of an IP-self-tuning controller integrated with a fuzzy logic-based gain adjustment mechanism.
• Simulation studies to initially assess controller behavior.
• Experimental validation using a 1.5KW induction machine implemented on a DSPace 1104 real-time control system.
• Comparative analysis based on computation time, speed tracking accuracy, and disturbance rejection capabilities.

Main Results:

• The proposed IP-self-tuning controller, tuned by a fuzzy adjustor, demonstrated competitive performance.
• Experimental results showed comparable or improved tracking performance and disturbance rejection over classical IP and fuzzy-like-PI controllers.
• The controller's ability to tune a single gain was confirmed, simplifying the control strategy.

Conclusions:

• The fuzzy-tuned IP-self-tuning controller offers an effective and simplified approach to induction machine speed control.
• This method presents a viable alternative to conventional PI-self-tuning controllers, particularly in scenarios demanding efficient gain tuning.
• The controller's performance in terms of speed tracking and disturbance rejection is robust, validated through simulation and experimental testing.