Experimental Validation: SINDy-Based Dynamic Identification of GFM/GFL Converters from PMU Data

This project develops a data-driven method to identify the internal dynamics of grid-forming (GFM) and grid-following (GFL) converters in renewable-rich power systems using only output-side PMU measurements. Small, continuous perturbations (e.g., 0.01 p.u. in active/reactive power setpoints) are injected during normal operation to safely excite the system, and SINDy is applied to the recorded voltage, current, and frequency data to obtain reduced-order models that capture the converters’ voltage–frequency behavior. From these models, we estimate key control parameters, such as PLL PI gains, P-f and Q-V droop coefficients, and virtual inertia and damping, and then design robust controllers that improve stability under high RES penetration. The full workflow has been implemented and validated in Simulink at McGill University, and the next phase will experimentally validate the approach using the double Power Hardware-in-the-Loop setup at Karlsruhe Institute of Technology (KIT). The project strengthens McGill’s capability in data-driven converter modeling and control and leverages KIT’s advanced laboratory infrastructure, fostering a long-term research collaboration between the two institutions.

Faculty Supervisor:

Xiaozhe Wang

Student:

Partner:

Karlsruher Institut für Technologie

Discipline:

Engineering

Sector:

Education

University:

McGill University

Program: