Hybrid AI-based Anomaly Detection Model using Phasor Measurement Unit Data

• DOI: 10.48550/arxiv.2209.12665
• Publication Date: 2022-01-01
• Authors: Yuval Abraham Regev; Henrik Vassdal; Ugur Halden; Ferhat Özgür Çatak; Ümit Cali
• Keywords: Units of measurement; Phasor measurement unit; Cyber-physical system; Anomaly (physics)

Over the last few decades, extensive use of information and communication technologies has been main driver digitalization power systems. Proper secure monitoring critical grid infrastructure became an integral part modern system. Using phasor measurement units (PMUs) to surveil system is one that have a promising future. Increased frequency measurements smarter methods for data handling can improve ability reliably operate grids. The increased cyber-physical interaction offers both benefits drawbacks, where drawbacks comes in form anomalies data. be caused by physical faults on grid, as well disturbances, errors, cyber attacks layer. This paper aims develop hybrid AI-based model based various such Long Short Term Memory (LSTM), Convolutional Neural Network (CNN) other relevant algorithms anomaly detection unit dataset used within this research was acquired University Texas, which consists real from measurements. In addition data, false injected produce analyzed. impacts mitigating prevent kind are discussed.