Machine Learning Enabled Power Quality Enhancement in FACTS Devices: A Systematic Review

Power quality (PQ) degradation due to nonlinear loads, distributed energy resources (DERs), and renewable energy variability presents major challenges to grid stability and reliability. Flexible AC Transmission Systems (FACTS), such as STATCOM, SVC, and UPFC, offer dynamic voltage regulation and reactive power compensation but often rely on traditional control strategies that lack adaptive capability. The integration of machine learning (ML) has shown promise for improved classification, prediction, and adaptive control of PQ disturbances. This systematic review synthesizes recent research on ML and AI techniques applied to PQ enhancement specifically within the context of FACTS devices. It critically analyzes methodologies, compares performance metrics, identifies gaps in literature, and proposes a roadmap for advancing research and practical implementation. Results indicate that ML enables superior detection and adaptive responses in PQ tasks, but challenges remain in real world validation, dataset standardization, and controller integration. This study provides a comprehensive foundation for future work aimed at resilient, scalable, and explainable machine learning–based PQ control frameworks for FACTS devices.