Their project is dedicated to designing AI-driven tools capable of detecting cyberattacks and anomalies in modernized power grids—tools with mathematical safeguards, an area where traditional industry methods have repeatedly fallen short. “By integrating these advanced ML and AI techniques into existing grid monitoring and control systems,” Lavaei said, “our hope is that power operators will be able to detect and respond to anomalies in real time, reducing the risk of cascading failures and ensuring a more stable and resilient energy infrastructure.” Powerful Research As power systems evolve to become more efficient and sustainable, they are also becoming more data-centric, creating new opportunities for attackers to manipulate information and deceive operators. A successful cyberattack could trigger cascading failures across the grid, leading to widespread blackouts and economic paralysis. Lavaei and his team’s research directly addresses this vulnerability, developing deep learning techniques that not only detect anomalies but also pinpoint the most vulnerable sections of the U.S. power grid. The stakes could not be higher. If power grid operators had access to more sophisticated data analysis and learning tools, catastrophes like the Camp Fire and the Manhattan blackout could have been avoided. While these incidents were not cyber-induced, they illustrate the potentially disastrous consequences of grid failures. A reliable system for detecting anomalies and attacks could mean the difference between stability and widespread devastation. “This project will enable a safe data-driven operation of sustainable and resilient power systems,” Lavaei said. UCNI Impact The UCNI-funded research has already made significant strides. Lavaei and his team have successfully studied the U.S. grid at an unprecedented scale, an achievement that has the potential to transform how the nation safeguards its energy infrastructure. Their work has also fostered collaborations, including a partnership with Boston University, and has supported three Ph.D. students through the UCNI project. The impact of the research has led to the publication of eight papers, contributing critical knowledge to the field of cybersecurity in energy systems. “We have designed a deep learning technique to detect anomalies. We have also designed an ML technique that pinpoints what parts of the U.S. grid are vulnerable,” he said. “This is the first result in the literature that has successfully studied the grid at such a scale.” Recognizing the importance of this work, the Army Research Office awarded Lavaei and Sojoudi a $600,000 grant to continue their research on networks. The national security implications of this work are immense, as cyber threats to infrastructure continue to evolve. As the power grid becomes increasingly reliant on data-driven operations, the need for advanced cybersecurity measures grows more critical by the day. Through this UCNI-backed project, the U.S. is making significant strides toward a power grid that is not only efficient and sustainable but also fortified against the growing threats of cyberattacks and infrastructure failures—ensuring a more secure energy future. ◆
Stock image
ucnoyce.org 59
Made with FlippingBook Annual report maker