Improving Smart Grid security: Spectral and fractal analysis as tools for detecting cyberattacks

Improving Smart Grid security: Spectral and fractal analysis as tools for detecting cyberattacks

Blog Article

Objectives.Cyberattacks are major potential sources of disturbances in modern electrical networks (Smart Grid).However, distinguishing between the various kinds of harmonic distortions and malicious interventions can be challenging.The objective of this work is to develop an effective tool for detecting and quantifying the differences between harmonic and anomalous signals.This will permit the identification of cyberattacks associated with harmonic signal distortions to provide a Hats more accurate classification of patterns characteristic of malicious impacts.

Methods.A comparative analysis of various anomaly detection methods was conducted, including fractal analysis, multifractal analysis, Shannon entropy calculation, and power spectral density (PSD) analysis.Results.Harmonic distortions and anomalous signals caused by cyberattacks may share similar fractal and multifractal characteristics, making it harder to distinguish between them.The use of the Shannon entropy method does not fully capture the complexity and uncertainty of harmonic and anomalous signals.

To gain a deeper understanding of the nature of these signals, a comprehensive approach was applied, including analysis of their frequency characteristics and the use of other uncertainty assessment methods, such as multifractal analysis and PSD.Use of the ANTIFLAMX PSD method revealed significant differences in energy distribution between these signals, permitting a more accurate identification of cyberattacks.Conclusions.For the effective detection of cyberattacks associated with harmonic signal distortions in power systems, a comprehensive approach is required, including time series analysis, frequency analysis, and machine learning methods.This approach not only detects anomalies in signals but also provides their quantitative assessment to improve the accuracy of classifying malicious impacts.

The integration of these methods enhances the reliability and security of power systems, making them less vulnerable to cyberattacks.