Khalid, Junaid, Uddin, Muhammad Helal, Smailes, Michael, Liang, Jun  ORCID: https://orcid.org/0000-0001-7511-449X and Wang, Sheng
2026.
Physics informed adaptive frequency support with rotor energy-aware recovery for grid forming wind turbines.
IEEE Transactions on Sustainable Energy
10.1109/tste.2026.3737065
|
Preview |
PDF
- Accepted Post-Print Version
Available under License Creative Commons Attribution. Download (16MB) | Preview |
Abstract
High penetration of converter-interfaced wind power reduces system inertia and primary frequency response. This makes frequency nadir and rate-of-change-of-frequency (RoCoF) increasingly difficult to control. While kinetic-energy (KE)-based synthetic inertia from wind turbines can provide fast support without long-term energy loss. However, existing designs rely on fixed or heuristic gain scheduling and ad hoc rotor-speed recovery, which may underutilise available KE, cause unsafe rotor deceleration, or lead to secondary frequency dips. This paper proposes a physics-informed, constraint-embedded neural adaptive (PINA) framework for grid-forming wind turbines. The machine-side converter provides KE-based inertia and droop response, while the grid-side converter operates in virtual synchronous control mode to regulate the dc-link voltage and maintain grid synchronisation. An adaptive gain network (GainNet) schedules inertia and droop gains using local measurements of frequency deviation, RoCoF, rotor speed, and normalised rotor-KE headroom, while RecNet shapes energy-aware rotor-speed recovery. Both networks are trained with physics-informed objectives and deployed with hard rotor-energy, converter-power, and ramp-rate constraints. Numerical case studies evaluate the proposed framework against recent state-of-the-art frequency-support methods.
| Item Type: | Article |
|---|---|
| Date Type: | Published Online |
| Status: | In Press |
| Schools: | Schools > Engineering |
| Additional Information: | RRS policy applied |
| Publisher: | Institute of Electrical and Electronics Engineers |
| ISSN: | 1949-3029 |
| Date of First Compliant Deposit: | 28 September 2026 |
| Last Modified: | 28 Sep 2026 11:45 |
| URI: | https://orca.cardiff.ac.uk/id/eprint/189829 |
Actions (repository staff only)
 |
Edit Item |
Dimensions
Dimensions