DQ-PINN-Based Intelligent Reference Current Generation for LVRT Enhancement in Grid-Connected DFIG Systems
Contributors
Dr. Sajan CH
Keywords
Proceeding
Track
General Track
License
Copyright (c) 2026 Sustainable Global Societies Initiative

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Abstract
Grid codes require grid-connected doubly fed induction generator (DFIG) wind turbines to remain controllable during voltage sags without exceeding converter current or destabilizing the DC link. This work presents a physics-informed neural network formulated in synchronous dq coordinates (DQ-PINN) to generate adaptive current references during normal and low voltage ride through (LVRT) conditions. Measured three-phase variables are transformed into the dq frame, and the network estimates d-axis and q-axis commands under constraints derived from the DFIG electrical model, active- and reactive-power relations, DC-link energy balance, converter-current capability, and temporal smoothness. When a sag occurs, a supervisory LVRT layer prioritizes reactive-current injection and projects each command onto the feasible current region before it is applied to the conventional inner current loops. Frilled Lizard Optimization is used only to tune an auxiliary PI regulator for the DC link (FLO-PI). The proposed MATLAB/Simulink validation covers variable wind, balanced and unbalanced voltage sags, nonlinear-load disturbances, and post-fault recovery, with conventional dq-PI and data-driven ANN reference generators used as comparison methods.