A Sage-Husa Prediction Algorithm-Based Approach for Correcting the Hall Sensor Position in DC Brushless Motors

Area of Science:

• Electrical Engineering
• Control Systems
• Robotics

Background:

• Accurate rotor position is critical for brushless DC motor (BLDCM) control.
• Deviations cause performance issues like current/torque fluctuations, noise, and reduced efficiency.
• Hall sensors are commonly used but susceptible to installation offsets and signal delays.

Purpose of the Study:

• To develop a pre-calibration method to eliminate Hall sensor installation offsets and signal conditioning delays.
• To propose a self-adaptive position information prediction algorithm for refining rotor position accuracy.
• To enhance the overall control performance and efficiency of BLDCMs.

Main Methods:

• Estimation of Hall sensor installation offsets using the minimum deviation principle.
• Implementation of a pre-calibration method addressing magnet pole offset, signal conditioning delay, and armature response.
• Application of a Sage-Husa adaptive algorithm for filtering residual rotor position deviations.
• Experimental validation on a hydrogen circulation pump motor.

Main Results:

• Mean Square Error (MSE) of speed fluctuations reduced by 92.0%.
• Average phase current decreased by 62.8%, with significant reductions in motor vibration.
• The Sage-Husa algorithm improved speed stability, reduced overshoot by 56.0%, and enhanced commutation timing accuracy by 14.7% compared to traditional KF.
• Overall system efficiency was significantly improved.

Conclusions:

• The proposed pre-calibration and adaptive prediction methods effectively eliminate rotor position errors in BLDCMs.
• These techniques lead to substantial improvements in motor performance, efficiency, and stability.
• The Sage-Husa method demonstrates superior disturbance rejection and commutation prediction capabilities for BLDCM control.