Capacity development

A new study published in ‘Engineering’ highlights a growing cybersecurity threat to smart grids as they become more complex due to increased integration of distributed energy sources.

The research, conducted by Zengji Liu, Mengge Liu, Qi Wang, and Yi Tang, focuses on a sophisticated form of cyberattack known as a false data injection attack (FDIA) that targets data-driven algorithms used in smart grid operations.

As modern power systems adopt technologies like battery storage and solar panels, they rely more heavily on algorithms to manage energy distribution and grid stability. However, these algorithms can be exploited.

The study introduces a novel black-box FDIA method that injects false data directly at the measurement modules of distributed power supplies, using generative adversarial networks (GANs) to produce stealthy attack vectors.

What makes this method particularly dangerous is that it doesn’t require detailed knowledge of the grid’s internal workings, making it more practical and harder to detect in real-world scenarios.

The researchers also proposed an approach to estimate controller and filter parameters in distributed energy systems, making it easier to launch these attacks.

To test the method, the team simulated attacks on the New England 39-bus system, specifically targeting a deep learning model used for transient stability prediction. Results showed a dramatic drop in accuracy—from 98.75% to 56%—after the attack.

The attack also proved effective across multiple neural network models and on larger grid systems, such as IEEE’s 118-bus and 145-bus networks.

These findings underscore the urgent need for better cybersecurity defenses in the evolving smart grid landscape. As systems grow more complex and reliant on AI-driven management, developing robust protection against FDIA threats will be critical.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!