Power Quality Solutions for Rail Transport using AI-based Unified Power Quality Conditioners

Abstract

This research proposes an AI-controlled Unified Power Quality Conditioner (AI-UPQC) to enhance power quality in railway power supply systems. The AI-UPQC utilizes artificial neural networks (ANNs) to generate optimal reference signals for controlling the series and shunt active power filters (APFs). Simulation analysis in a typical 25 kV, 50 Hz traction power supply network demonstrates the effectiveness of the AI-UPQC in maintaining balanced supply voltage and mitigating current harmonics under nonideal operating conditions. The AI-based control strategy outperforms the conventional PI controller in tackling nonlinearity and parameter variations, resulting in superior harmonic mitigation, resonance damping, and dynamic performance. The AI-UPQC significantly reduces voltage and current total harmonic distortion (THD) compared to the uncompensated case and the PI-UPQC. Economic analysis reveals substantial cost savings from reduced equipment maintenance, avoided penalties, and improved energy efficiency. The proposed data-driven AI-UPQC system offers a promising solution to the power quality challenges faced by modern electrified railway transportation networks. Future research directions include advanced machine learning algorithms, real-world testing, scalability, integration with renewable energy sources, and comprehensive economic analysis.