Cyber-Secure Operation of Battery Energy Storage Systems Providing Grid Services

In the domain of renewable energy, Battery Energy Storage Systems (BESSs) have emerged as pivotal agents accelerating the transition toward renewable power sources. Utilizing internet-of-things (IoT) technologies increases the robustness of BESS and BESS-enabled technologies such as virtual power plants (VPP). VPP allows to combine the resources of multiple units to provide a service leveraging the resources of each VPP component. Simultaneously, sharing security-critical data through communication channels among multiple parties opens up a vulnerability towards cyberattacks against BESSs. As the modern energy grid relies more and more on energy storage systems, increasing the resilience of BESS towards cyberattacks becomes critical for system stability. This Ph.D. study investigates security issues that could affect a BESS utilized to support the grid. The special focus is placed on BESS services that have a variable operational pattern and are crucial for keeping the grid stable as they are more prone to cyberattacks.
To address the vulnerability of BESS, the study explores ways to design secure BESSs and develops a new method to detect and mitigate cyberattacks against BESSs for a BESS as a part of VPP. The novel algorithm can detect the BESS-disabling attack while ensuring undisturbed system operation. The algorithm predicts how the BESS will behave in the future while utilizing the capabilities of the VPP, which brings together multiple BESSs to confirm the presence of a cyberattack. To validate the cyberattack detector, the BESS-disabling attack that is capable of shutting down a BESS with limited knowledge about the system was designed.
To predict BESS behaviour, historical data of a realistic and real BESS providing frequency regulation was recorded. Six machine-learning methods are chosen and compared by using real data from a BESS assisting the grid and data from Denmark's PowerLabDK, where a BESS is used for frequency regulation. To predict BESS behaviour and identify any unusual patterns that might indicate an attack, the study uses a powerful technique called Adaptive Boosting, which helps forecast the BESS's state of charge (SOC) accurately.
Furthermore, a technique called Moving Target Defence (MTD) is implemented. It uses aggregated resources of VPP to verify the presence of cyberattack. Applying hidden MTD increases the sensitivity of the algorithm without compromising the number of false alarms. MTD adjusts the VPP's operation setpoint based on cyberattack signals and evaluates how the system responds.
To counter more sophisticated attacks, the defense mechanism with hidden MTD is enhanced, adding an extra layer of protection. The adjustments made to the setpoint of each BESS in a VPP are limited to the measurement and estimation errors, making it difficult for attackers to detect the changes introduced by the cyber defense algorithm.
To demonstrate the effectiveness of the latter algorithm, this is tested using real data and simulations of various cyberattacks that could disable BESSs. The results are promising, as the system successfully detects and neutralizes all simulated attacks, ensuring the resilient operation of the BESSs.
In conclusion, with the increasing reliance on renewable energy and smart grid technologies, ensuring the security of BESSs is vital for a sustainable and resilient energy future. The research sheds light on potential cyber threats to BESSs within VPPs and presents a robust defense algorithm to protect these critical energy storage systems. In summary, this work has contributed to a safer and more secure energy future.