• A Robust and Explainable Data-...

Treffer: A Robust and Explainable Data-Driven Anomaly Detection Approach For Power Electronics

Title:
A Robust and Explainable Data-Driven Anomaly Detection Approach For Power Electronics
Authors:
A. Beattie, Pavol Mulinka, orcid:0000-0002-9394-, S. Sahoo, I. T. Christou, Charalampos Kalalas, D. Gutierrez-Rojas, P. H. J. Nardelli
Publisher Information:
Zenodo
Publication Year:
2022
Collection:
Zenodo
Subject Terms:
Anomaly detection, Classification (of information), Data visualization, Embedded systems, Fault detection, Internet of things, Iterative methods, Learning algorithms, Machine learning, Uncertainty analysis, Cybe-physical systems, Detection approach, Fault classification, Industrial internet of thing, matrix, Power-electronics, Profile algorithms, Uncertainty, Cyber Physical System
Document Type:
Konferenz conference object
Language:
unknown
Relation:
https://zenodo.org/records/7505941; oai:zenodo.org:7505941; https://doi.org/10.1109/SmartGridComm52983.2022.9961002
DOI:
10.1109/SmartGridComm52983.2022.9961002
Availability:
https://doi.org/10.1109/SmartGridComm52983.2022.9961002
https://zenodo.org/records/7505941
Rights:
Creative Commons Attribution 4.0 International ; cc-by-4.0 ; https://creativecommons.org/licenses/by/4.0/legalcode
Accession Number:
edsbas.CEBA9107
Database:
BASE

Weitere Informationen

Timely and accurate detection of anomalies in power electronics is becoming increasingly critical for maintaining complex production systems. Robust and explainable strategies help decrease system downtime and preempt or mitigate infrastructure cyberattacks. This work begins by explaining the types of uncertainty present in current datasets and machine learning algorithm outputs. Three techniques for combating these uncertainties are then introduced and analyzed. We further present two anomaly detection and classification approaches, namely the Matrix Profile algorithm and anomaly transformer, which are applied in the context of a power electronic converter dataset. Specifically, the Matrix Profile algorithm is shown to be well suited as a generalizable approach for detecting real-time anomalies in streaming time-series data. The STUMPY python library implementation of the iterative Matrix Profile is used for the creation of the detector. A series of custom filters is created and added to the detector to tune its sensitivity, recall, and detection accuracy. Our numerical results show that, with simple parameter tuning, the detector provides high accuracy and performance in a variety of fault scenarios. © 2022 IEEE. ; This work is partly supported by the FIREMAN project CHIST-ERA-17-BDSI-003 funded by the Spanish National Foundation (PCI2019-103780), the Academy of Finland (AoF; n.326270) and the Greek General Secretariat of Research and Technology; also by (1) AoF via EnergyNet fellowship n.321265/n.328869/n.352654 and X-SDEN project n.349965, and (2) Baltic-Nordic Energy Research programme via Next-uGrid project n.117766. This work has been also funded by the "Ministerio de Asuntos Económicos y Transformación Digital" and the European Union-NextGenerationEU in the frameworks of the "Plan de Recuperación, Transformación y Resiliencia" and of the "Mecanismo de Recuperación y Resiliencia" under references TSI-063000-2021-39/40/41.