Detection of Electrical Transmission Line Overload with Blockchain using Isolation Forest with Recurrent Neural Network

Nowadays, energy has emerged as a dynamic support of modern society, with electricity playing a crucial role in powering our daily lives. As the world's dependence on electricity continues to grow, the significance of a consistent and reliable supply cannot be exaggerated. Power companies face the daunting task of meeting this escalating demand, while also ensuring the delivery of high-quality energy services to consumers, where even a momentary disruption can have severe consequences for businesses, communities, and individuals alike. To overcome these limitations, an Isolation Forest with Recurrent Neural Network (IF-RNN) for detecting overload in electrical transmission lines using Power Demand Benchmarking Electric Power Load (PDB-EPL) dataset. For preprocessing, a Singular Value Decomposition (SVD) is employed for reducing noise and irrelevant features, for improving generalization, and dimensionality. Then, Blockchain is considered for feature extraction that performs decentralization, fault tolerance, immutability, auditability, and transparency. Eventually, IF-RNN is utilized as classifier for anomaly detection, high-dimensional data handling, and scalability. The results conclusively demonstrate the efficacy of the proposed IF-RNN interms of balanced accuracy, and accuracy when compared to the existing method like Feed-forward Neural Network (FNN).