Hypertuning Random Forest for Enhancing Cyber Security in Industrial Power Electronics

Industrial Power Electronics (IPE) systems are increasingly integrated with communication technology, creating new vulnerabilities for cyber-attacks. In most cases, traditional security measures are insufficient to deal with the complexity and evolution of attacks that target critical infrastructure. This paper presents a machine learning-based framework for detecting cyberattacks in IPE systems, leveraging the Random Forest algorithm to classify anomalous network traffic. Using the NSL-KDD dataset, which encompasses multiple attack types such as Denial of Service (DoS), Probe, Remote to Local (R2L), and User to Root (U2R) attacks, we achieved a detection accuracy of 99.61%. Our approach highlights the potential of Random Forest models in effectively identifying intrusion attempts and enhancing the resilience of industrial systems against cyber threats. The high accuracy underscores the effectiveness of feature-rich datasets and the importance of machine learning in modern industrial cybersecurity solutions.