A Hybrid Intelligence Framework for Real-Time Anomaly Detection in Smart Grid Transactions

Abstract

The increasing digitization of smart grids has introduced new vulnerabilities, particularly in fraudulent and anomalous energy transactions that threaten both market stability and cybersecurity. This study addresses the urgent need for secure and explainable anomaly detection mechanisms in energy markets. The purpose of the research is to develop a hybrid intelligence framework that integrates machine learning–based anomaly detection with blockchain’s immutable ledger to provide real-time, trustworthy insights into transaction integrity. The proposed framework employs a combination of deep neural networks and explainable gradient-boosting models for anomaly classification, enhanced with synthetic oversampling to address class imbalance. Evaluation was conducted using standard performance metrics, including precision, recall, F1-score, detection latency, and interpretability assessment. Results demonstrate that the hybrid framework achieved a significant improvement in detection accuracy and false-positive reduction compared to conventional ML-only baselines. Moreover, the blockchain integration ensured tamper-proof record-keeping, while the explainable AI component enhanced stakeholder trust by offering transparent reasoning for flagged anomalies. The findings highlight the potential of combining AI-driven intelligence with blockchain consensus for advancing fraud detection in smart grid environments.