Akademik Veri Yönetim Sistemi
Arabian Journal for Science and Engineering, cilt.50, 2025 (SCI-Expanded)
The efficient utilization of energy resources is crucial for the economic and social development of any country. Given that agricultural irrigation constitutes a significant portion of annual energy consumption, effective monitoring and billing mechanisms are essential. This study proposes an innovative system for detecting unauthorized electricity usage in agricultural irrigation systems, particularly focusing on irrigation motors directly connected to the distribution line. The system leverages a wireless IoT network and machine learning to identify electricity theft by analyzing current patterns from the feeder measurement point. The proposed approach overcomes the limitations of traditional impedance-based methods, which often fail due to the dynamic load profile and branching in low-voltage distribution lines. The Optimized Boosting Ensemble (OBE) algorithm demonstrated superior performance, achieving high validation accuracy and low error rates. When applied to simulation data, the algorithm accurately identified illegal load locations with minimal error. Field testing with real measurement data further confirmed the system’s effectiveness, showing successful localization with low average and maximum error rates. This study represents a pioneering application of IoT and machine learning for theft localization in agricultural settings, validated through both simulations and field measurements. The findings underscore the system’s potential to enhance energy efficiency in rural grid management by providing an adaptable solution for accurately identifying unauthorized loads.