A Preliminary Study of Model Predictive Control of Permanent Magnet Assisted Maglev System


AYHAN A. B., ERKAN K., BOZKURT A. F.

8th Global Power, Energy and Communication Conference, GPECOM 2026, Naples, İtalya, 3 - 05 Haziran 2026, ss.160-165, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/gpecom70462.2026.11578643
  • Basıldığı Şehir: Naples
  • Basıldığı Ülke: İtalya
  • Sayfa Sayıları: ss.160-165
  • Anahtar Kelimeler: 3DOF control, magnetic levitation, model predictive control
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

Zero-friction operation and low noise characteristics make magnetic levitation (Maglev) systems highly desirable for advanced motion control applications. However, controlling these platforms remains challenging due to their open-loop instability and strong nonlinear dynamics. While conventional PID controllers often struggle to satisfy strict performance criteria under such conditions, Model Predictive Control (MPC) offers a systematic way to handle system constraints. This study presents a decoupled MPC architecture to stabilize and track trajectories for a 3-Degree of Freedom (DOF) Maglev system. Instead of MIMO, the vertical (z), roll (α), and pitch (β) axes are governed by three independent MPC loops, called centralized control approach. A 4 × 4 control allocation matrix then translates these virtual control efforts into physical electromagnet currents, effectively minimizing cross-coupling. The internal prediction models for the controllers rely on a linearized state-space representation derived at a 5.5 mm nominal air gap. MATLAB/Simulink simulations validate the proposed architecture, showing distinct improvements in reference tracking and disturbance rejection when compared to a baseline classical I-PD controller.