A recent study published in Engineering has shed light on a significant cybersecurity risk facing smart grids as they become more complex with the increasing integration of distributed power supplies. The research, conducted by Zengji Liu, Mengge Liu, Qi Wang, and Yi Tang, focuses on false data injection attacks (FDIAs) targeting data-driven algorithms in smart grids.
As smart grids expand with the addition of distributed power sources like battery energy storage systems and photovoltaic installations, data-driven algorithms have become crucial for managing the complex power systems. However, these algorithms are vulnerable to cyberattacks. FDIA is a type of disruption that can compromise system operations by hijacking or tampering with data.
The researchers introduced a novel black-box FDIA method. Unlike traditional approaches that manipulate data in communication networks, this new method directly injects false data at the measurement modules of distributed power supplies. It uses a generative adversarial network (GAN) to generate stealthy attack vectors. One of the key advantages of this method is that it requires no detailed knowledge of the target system, making it a practical threat in real-world scenarios.
The study also proposed an estimation method for controller and filter parameters of distributed power supplies, which simplifies the execution of the black-box attack. A false data injection technique was developed to introduce the attack vector into the power system through the measurement module of distributed power supplies. To enhance the efficiency of the attack and minimize errors, a piecewise process for generating the attack vector was described.
To validate the effectiveness of the proposed attack method, the researchers conducted a case study on the New England 39-bus system. They targeted the transient stability prediction (TSP) approach based on a deep convolutional neural network. The results showed that the attack could significantly reduce the prediction accuracy of the TSP model. For example, the prediction accuracy dropped from 98.75% to 56.00% after the attack.
The attack method was also tested on different neural network architectures and various IEEE bus systems. The findings indicated that the attack vector could deceive different targets, and the attack was effective across different system scales. Larger systems, such as the IEEE 118-bus and 145-bus systems, were more affected, highlighting the need for robust defense mechanisms in smart grids.
This research serves as a wake-up call for the smart grid industry. As smart grids continue to evolve, it is essential to develop effective security measures to protect data-driven algorithms from FDIA threats. Future studies may focus on developing countermeasures to defend against such attacks and strengthening the security of smart grid systems.
The paper "False Data Injection Attacks on Data-Driven Algorithms in Smart Grids Utilizing Distributed Power Supplies," authored by Zengji Liu, Mengge Liu, Qi Wang, Yi Tang. Full text of the open access paper: https://doi.org/10.1016/j.eng.2024.11.025
