Green-Internet of Fuzzy Things for Irregularity Detection in the Advanced Metering Infrastructure

Advanced Metering Infrastructure (AMI) relies on smart meters to collect and monitor electricity consumption data from residential and industrial entities in the context of smart cities. These smart meters facilitate data collection and monitoring and enable the detection of non-technical losses through consumption pattern analysis. This study introduces a Green-Internet of Fuzzy Things (GIFT) solution to identify anomalies within smart metering data, a critical component for enhancing AMI reliability and efficiency. By leveraging the evolutionary fuzzy system, the GIFT solution delivers multi-class anom-aly classification with enhanced interpretability and robustness against data vagueness and uncertainty. After validation using a real-world smart metering dataset, the proposed system achieves an average Area Under the Curve (AUC) of 0.9496 across six anomaly types, outperforming existing methods with an accuracy of 89.89%, precision of 79.26%, recall of 74.71%, F1-score of 0.7443, and a low false positive rate (FPR) of 0.0632. Additionally, the GIFT solution demonstrates environmental sustain-ability by measuring CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> eq emissions, maintaining lower emissions than other methods across multi-ple scenarios. This work significantly advances the field of IoT-based energy management by addressing critical gaps in anomaly detection mechanisms while promoting environmental sustainability.