| Abstract: |
Power quality enhancement in distribution systems has become critical with increasing integration of renewable energy sources and nonlinear loads. This research investigates advanced intelligent control methods including Artificial Neural Networks, Adaptive Neuro-Fuzzy Inference Systems, and optimization algorithms for mitigating power quality issues in Indian distribution networks. The study employs a comprehensive methodology examining voltage sag, swell, harmonics, and reactive power compensation through FACTS devices controlled by intelligent algorithms. Five case studies with real-world data demonstrate significant improvements in total harmonic distortion reduction from 15.10% to 0.90%, voltage profile enhancement by 8.5%, and power loss reduction of 12.3% compared to conventional controllers. Deep Reinforcement Learning integrated with PI controllers shows 20.50% performance improvement in renewable energy integration scenarios. Statistical analysis validates that PSO-optimized ANFIS controllers outperform traditional methods with 95% confidence levels. The findings establish that hybrid intelligent control strategies combining machine learning with optimization techniques offer superior adaptability and robustness for modern distribution system challenges. Implementation recommendations suggest prioritizing ANFIS-based controllers for nonlinear load compensation and DRL-based systems for renewable integration applications in Indian power distribution networks. |
