Abstract:The expansion in technology and attainability of a large number of sensors has led to a huge amount of real-time streaming data. The real-time data in the electrical distribution system is collected through distribution-level phasor measurement units referred to as $\mu$PMU which report high-resolution phasor measurements comprising various event signatures which provide situational awareness and enable a level of visibility into the distribution system. These events are infrequent, unschedule, and uncertain; it is a challenge to scrutinize, detect and predict the occurrence of such events. For electrical distribution systems, it is challenging to explicitly identify evolution functions that describe the complex, non-linear, and non-stationary signature patterns of events. In this paper, we seek to address this problem by developing a physics dynamics-based approach to detect anomalies in the $\mu$PMU streaming data and simultaneously predict the events using governing equations. We propose a data-driven approach based on the Hankel alternative view of the Koopman (HAVOK) operator, called DynamoPMU, to analyze the underlying dynamics of the distribution system by representing them in a linear intrinsic space. The key technical idea is that the proposed method separates out the linear dynamical behaviour pattern and intermittent forcing (anomalous events) in sequential data which turns out to be very useful for anomaly detection and simultaneous data prediction. We demonstrate the efficacy of our proposed framework through analysis of real $\mu$PMU data taken from the LBNL distribution grid. DynamoPMU is suitable for real-time event detection as well as prediction in an unsupervised way and adapts to varying statistics.

Dimensionality Increment of PMU Data for Anomaly Detection in Low Observability Power Systems

Early Anomaly Detection in Power Systems Based on Random Matrix Theory

Sequential Feature Selection for Power System Event Classification Utilizing Wide-Area PMU Data

Online Anomaly Detection Using Machine Learning and HPC for Power System Synchrophasor Measurements

A Deep Learning Approach to Anomaly Sequence Detection for High-Resolution Monitoring of Power Systems

IRFLMDNN: hybrid model for PMU data anomaly detection and re-filling with improved random forest and Levenberg Marquardt algorithm optimized dynamic neural network

Using PMU Data for Anomaly Detection and Localization

Pattern Mining and Anomaly Detection Based on Power System Synchrophasor Measurements

DynamoPMU: A Physics Informed Anomaly Detection and Prediction Methodology using non-linear dynamics from $μ$PMU Measurement Data

Optimal PMU Placement Approach for Power Systems Considering Non-Gaussian Measurement Noise Statistics

Hybrid Clustering‐based Bad Data Detection of PMU Measurements

Early Anomaly Detection and Location in Distribution Network: A Data-Driven Approach

Robust Fast PMU Measurement Recovery Enhanced by Randomized Singular Value and Sequential Tucker Decomposition

Learning from Time-Synchronized Power System Measurements with Rare Anomalous Event Records to Detect PMU Data Manipulation Attacks

Early Anomaly Detection and Localisation in Distribution Network: a Data‐driven Approach

Anomaly Detection Using Optimally-Placed Micro-PMU Sensors in Distribution Grids

A low-quality PMU data identification method with dynamic criteria based on spatial–temporal correlations and random matrices

An Unsupervised Anomaly Detection Approach Using Energy-Based Spatiotemporal Graphical Modeling

Correlation-based Bad Data Detection of PMU Measurements for Improving Network Structure with High Penetration of Renewable Energy

Framework of Random Matrix Theory for Power System Data Mining in a Non-Gaussian Environment.

Distributed Framework for Detecting PMU Data Manipulation Attacks with Deep Autoencoders