Adaptive Sparse Regression for Data-Driven Modeling and Control of Grid-Connected DFIG Wind Turbines

This article aims to introduce a novel technique, adaptive sparse regression (ASR), for data-driven sparse model identification of doubly-fed induction generator (DFIG) wind turbine systems. Unlike other data-driven techniques, the proposed approach does not rely on large amounts of data for training and is interpretable. In addition, the proposed ASR method achieves higher accuracy in matching true system dynamics compared with conventional sparse regression techniques. This is accomplished by using different sparsity-promoting hyperparameters for different state variables. Using this algorithm, open-loop and closed-loop models of DFIG wind turbine systems are derived from measurements and utilized for control design. The effectiveness of the proposed data-driven approach for data-driven modeling of DFIGs is validated in various case studies, and its application in grid impedance estimation is evaluated, highlighting its capabilities compared with other data-driven techniques. Finally, the proposed approach is validated through real-time simulations.