Phantom is an ongoing research project under OHR by Aditya V Kini and Ashwin Shankar,which is focused on soft sensing applications in aerospace
I have attempted to implement Field Oriented Control (FOC) of an Internal Permanent Magnet Synchronous Motor (IPMSM).
The objective is to achieve precise speed and current control, while generating datasets for advanced AI-driven motor switching optimization in aerospace applications.
- Implement Field Oriented Control (FOC) for an IPMSM.
- Design control loops for speed and current regulation.
- Generate and log training datasets for:
- TSAM (Time-Series Anomaly Model) – for fault detection & discharge monitoring.
- MRTCN (Multi-Resolution Temporal Convolutional Network) – for switching sequence optimization.
- Develop a scalable framework for motor drive health monitoring and predictive control.
- ✅ FOC model successfully developed and simulated.
- ✅ Operating under fixed-load and no partial discharge conditions.
- ⚙️ PSIM model under development — using Space Vector Modulation (SVM) for PMSM drive control.
- Still yet to decide the relevant parameters of the rotor system that can be fed to the TSAM/MRTCN Model
- 📊 Dataset generation underway for ML model training (TSAM & MRTCN).
| Module | Function | Implementation Tool |
|---|---|---|
| Motor Control | Field Oriented Control (Speed + Current) | MATLAB / Simulink |
| Inverter Logic | Space Vector Modulation | PSIM |
| Data Generation | Current, Voltage, Torque Waveforms | MATLAB |
| AI Models | TSAM & MRTCN (Motor Switching Optimization) | Python (PyTorch / TensorFlow) |
Field Oriented Control (FOC) separates the motor’s stator currents into two orthogonal components:
i_d→ controls the magnetic fluxi_q→ controls the torque
This allows independent control of torque and flux — similar to a DC motor — enabling:
- Smooth and fast torque response
- High efficiency
- Low torque ripple
The control loop includes:
- Clarke & Park Transformations
- Current Controller (PI Loop)
- Speed Controller (Outer Loop)
- Inverse Park Transformation
- PWM Generation using SVM
- Stable operation under fixed load.
- Smooth torque and flux control in FOC loop.
- Minimal ripple in electromagnetic torque waveform.
- Space Vector PWM under tuning for harmonic optimization.
- Incorporate partial discharge (PD) fault conditions.
- Integrate real-time adaptive switching using ML (TSAM + MRTCN).
- Validate system on aerospace-grade PMSM hardware.
- Add DC-link voltage stabilization and regenerative braking features.
MATLAB / Simulink • PSIM • Python • PyTorch • Control Systems • Power Electronics • AI for Motor Drives