A Novel Intelligent Fractional Order Cascade Control to Enhance Wind Energy Conversion in Wind Farms: A Practical Case Study

• Published in: IEEE transactions on energy conversion Vol. 40; no. 3; pp. 1736 - 1749
• Main Authors: Peykarporsan, Rasool, Oshnoei, Soroush, Fathollahi, Arman, Lie, Tek Tjing
• Format: Journal Article
• Language: English
• Published: IEEE 01.09.2025
• Subjects: deep deterministic policy gradient; Doubly fed induction generators; doubly-fed induction generator; four-degree-of-freedom controller; fractional-order cascade controller; Mathematical models; PI control; Power system stability; Renewable energy; Renewable energy sources; Rotors; Stators; Training; Uncertainty; wind farm; Wind turbines
• ISSN: 0885-8969; 1558-0059
• DOI: 10.1109/TEC.2025.3543144

As the world's demand for electricity is rising with a growing emphasis on environmental sustainability, the need for efficient renewable energy solutions becomes increasingly critical. Wind power, which comprises 26% of renewable resources, is essential in this transition. Nevertheless, the performance of wind farms (WFs) can be adversely affected by uncertainties in wind speed. In response to this challenge, we introduce a novel four-degree-of-freedom (4DoF)-based fractional-order cascade control approach for WFs based on doubly fed induction generators (DFIGs) to enhance the efficiency and robustness of wind energy conversion systems (WECSs). The presented control method leverages the flexibility and disturbance-reduction capabilities of fractional-order proportional-integral-derivative and tilt-integral-derivative controllers in a 4DoF framework called 4DoF-IHYB. Then, the 4DoF-IHYB controller is cascaded with a fractional-order tilt-derivative controller to mitigate the impact of input noises and disturbances. Furthermore, a deep deterministic policy gradient (DDPG) method is utilized to optimize the controller's parameters and improve the control system's efficiency in the face of uncertainties stemming from volatile environmental conditions. DDPG is an algorithm based on deep reinforcement learning that integrates the advantages of both deep learning and policy gradient methods. The proposed control technique's effectiveness is assessed in a case study of a prominent wind energy facility in New Zealand subject to various operating conditions. Moreover, the presented control method's efficiency is compared with control methods available in the literature. The simulation results disclose that the proposed control method provides much better dynamic stability for the practical case study than the other methods.