In a significant leap towards shaping the future of energy, the U.S. Department of Energy (DOE) has established the Center for AI @PNNL at Washington State's Pacific Northwest National Laboratory (PNNL). Dedicated to advancing artificial intelligence (AI) applications, the center brings together hundreds of researchers to explore AI's potential across various domains, from smart grids to electric vehicles.
PNNL researchers are actively engaged in cutting-edge projects, ranging from developing AI tools ensuring a seamless power supply to homes and businesses to utilizing machine learning for optimizing chemical compounds in electric vehicle lithium batteries. The center's initiatives span training algorithms to detect theft in distribution circuits and modeling high-performance wind turbine designs, showcasing the versatility of AI in the energy sector.
A standout feature of the center's approach is its emphasis on physics-informed machine learning, utilizing advanced algorithms based on Earth models, embedded systems, and power grid components. This approach significantly reduces the number of samples required to train AI programs, making it particularly valuable for processing large datasets. Experiments in materials chemistry, for instance, involve the analysis of thousands of microscope images per second.
Furthermore, the Center for AI @PNNL leverages exascale supercomputing, enabling billions of computations per second. Collaborating with the Exascale Computing Project (ECP), PNNL has contributed to various areas, such as predicting the performance of multiphase energy conversion devices and aiding the development of small modular reactors. These advancements underscore the potential of AI in revolutionizing energy systems and addressing national security challenges related to the transition to wind and solar resources.
The center's researchers are also focused on energy resilience and security, training predictive AI models to safeguard the nation's energy infrastructure. Projects include generating synthetic data for machine learning to detect theft in distribution circuits, developing algorithms to optimize grid response to weather events, and using deep reinforcement learning for emergency control protocols in electric grids. The Center for AI @PNNL is poised to play a pivotal role in shaping the landscape of AI-driven innovations for sustainable and secure energy solutions.
Source: eepower.com