Context: The DataPower project has developed a software-defined power converter (TRL 9), but its online data acquisition platform—critical for intelligent control and smart maintenance—remains at TRL 2-3. The internship aims to advance this platform within the ANR-funded project ANR-21-CE05-0011, which explores power converters as both energy processors and data sources for machine learning.
Objective: Develop and integrate an online data acquisition platform to enable closed-loop intelligent control and smart maintenance, bridging model-based and data-driven approaches.
Key Tasks:
- Understand the existing power hardware, control firmware, and Python/MATLAB-based modules for smart maintenance and generic control.
- Implement a model-driven control database and define open-loop excitation procedures for data collection.
- Validate the closed-loop interaction: data acquisition → smart maintenance (parameter identification, anomaly detection, fault diagnosis, life estimation) → control system update.
Methodology: Use data-driven methods like Virtual Reference Feedback Tuning (VRFT) to bypass traditional modeling, relying instead on experimental input-output data for rapid controller synthesis.
Deliverables:
- A functional prototype of the online data acquisition platform integrated with the power converter.
- A case study demonstrating the intelligent control and smart maintenance loop.
- Documentation of excitation protocols, data requirements, preprocessing, and control database strategy.
Skills Required:
- Background in control systems and power electronics
- Proficiency in Python and MATLAB/Simulink
- Familiarity with embedded systems is a plus.
Contact: Luiz Villa and Pauline Kergus
