Torque Ripple Minimization for A Bearing-Less Synchronous Motor Using Neural Network Based Control Technique.

Reliable torque production in bearing-less synchronous motors (BSMs) is often degraded by torque ripple, which can increase vibration, acoustic noise, and speed fluctuation, especially at low-speed operation. This study proposes a compact torque-ripple assessment framework for two BSM prototypes (37 kW/400 V and 11 kW/690 V) using a DSP-based platform with torque and speed sensing, supported by MATLAB/Simulink. The method identifies dominant ripple sources (cogging, harmonic distortion, and reluctance torque due to d–q interaction), summarizes ripple severity using a torque ripple factor (TRF), and applies a lightweight feed-forward neural network to predict electromagnetic torque from d–q currents, rotor position, and rotor speed. Network parameters were optimized by minimizing an MSE-based cost function. Results show that cogging torque is periodic, with pulsations repeating every 30° rotor movement and becoming most pronounced below 500 rpm. Harmonic analysis reveals five significant components with a dominant 3rd harmonic (0.2 Nm). Total torque varies between 8.2 Nm and 11.8 Nm, while reluctance torque contributes about 15% of the ripple. Frequency-domain results indicate dominant ripple bands at 50 Hz, 100 Hz, and 150 Hz, with the 100 Hz component approximately 2.5× stronger than the fundamental. The neural model converged within 40 epochs, reducing MSE from 0.52 to 0.08, and predicted torque within ±0.5 Nm over the 10–12.5 Nm range. The proposed low-complexity framework effectively characterizes ripple contributors and enables reliable torque estimation. It is recommended that future mitigation efforts prioritize dominant low-order harmonics, particularly around the 100 Hz band.