A Multi-Stage NSGA-III Optimization Model for False Data Injection Attacks in Integrated Power-Hydrogen Cyber-Physical Systems

Abstract

The growing integration of renewable energy sources and hydrogen technologies into the power grid has introduced significant cyber-physical vulnerabilities, particularly to false data injection attacks (FDIA). This paper presents a novel three-stage FDIA optimization model targeting integrated power-hydrogen systems. The model leverages the interdependencies between power grids and hydrogen infrastructures, aiming to maximize system disruption while minimizing detection likelihood. Using a multi-objective optimization framework based on the non-dominated sorting genetic algorithm III (NSGA-III), the model identifies Pareto-optimal attack strategies by balancing conflicting objectives. The case study uses synthesized data to represent realistic conditions in the IEEE 118-bus system, incorporating hydrogen storage facilities and production infrastructure. Simulations demonstrate how coordinated attacks can exacerbate imbalances in power load and hydrogen storage, resulting in cascading failures across both systems. The results show that the power system's recovery time increases exponentially with the attack intensity, while the hydrogen system suffers from slower, more complex recovery dynamics. This work highlights the critical need for advanced cybersecurity frameworks to protect integrated power-hydrogen infrastructures from sophisticated cyber-physical threats, emphasizing the importance of monitoring and securing multiple layers of the system.