Abstract

The growing exposure of Industrial Control Systems (ICS), especially smart grids, to the internet and the proliferation of wireless networks exacerbate the cybersecurity risks to these systems. Meanwhile, developments in using cutting-edge technology for managing and controlling industrial systems enhance remote troubleshooting and operational flexibility; nevertheless, they also increase vulnerability to cyberattacks. A testing environment is required to simulate and assess the impact of cyberattacks on smart grids. This work introduces a security framework that offers intrusion detection and prevention functionalities at both the host and network levels, increasing the trust in threat detection approaches. Utilizing machine learning and artificial intelligence to enhance the detection of threats and assaults efficiently. Machine learning and blockchain are used to fulfill the security objectives of the suggested architecture. Extreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP) are used to analyze the architecture and identify the most significant elements of industrial data, while blockchain technology, coupled with smart contracts, is utilized to document and secure transactions between Internet of Things (IoT) devices. Moreover, the proposed design provides an explanatory environment to interpret the output of the machine learning algorithms, showing the significance of each feature element of the system. The proposed methodology has outstanding effectiveness in detecting cyberattacks, as shown by precision, recall, F1-Score, accuracy, and other performance metrics.