Mitigating sub-synchronous oscillation using intelligent damping control of DFIG based on improved TD3 algorithm with...

Area of Science:

• Power Systems Engineering
• Renewable Energy Integration
• Control Systems

Background:

• Sub-synchronous oscillations (SSO) in doubly-fed induction generators (DFIGs) pose a significant threat to power grid stability.
• Conventional damping controllers struggle with the dynamic nature of modern power systems.

Purpose of the Study:

• To develop an Intelligent Sub-Synchronous Damping Controller (I-SSDC) for DFIGs to effectively mitigate SSO.
• To enhance controller adaptability and interpretability using deep reinforcement learning and integrated knowledge.

Main Methods:

• Formulation of an I-SSDC framework using an improved Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm with Softmax.
• Construction of a surrogate model via Weighted Linear Regression and regularization to identify key SSO influencing factors.
• Integration of domain knowledge into agent training to improve exploration efficiency.

Main Results:

• The proposed I-SSDC demonstrates effective suppression of SSO across various operating conditions.
• The surrogate model aids in optimizing the controller's output signal selection for improved decision-making.
• Enhanced environmental adaptability and interpretability of the intelligent controller were achieved.

Conclusions:

• The I-SSDC effectively addresses the limitations of conventional controllers for SSO mitigation in DFIG-based wind turbines.
• The integration of deep reinforcement learning and knowledge provides a robust solution for maintaining power grid stability.
• The developed approach offers improved performance and adaptability in dynamic power system environments.