Battery health is readily diagnosed in lab settings but can be difficult to measure during energy storage system operation, as common lab diagnostic tests require long times or expensive test equipment to perform. NREL researchers use physics-based models and machine learningto enable rapid, scalable diagnostic tests. Given that batteries degrade with use and storage, predictive models of battery lifetime must consider the variety of electrochemical, thermal, and mechanical degradation modes, such as temperature, operating windows,. With validated models of battery performance and lifetime, battery controls or energy storage system designs can be optimized for revenue,. Predicting Battery Capacity From Impedance at Varying Temperature and State-of-Charge using Machine-Learning, Cell Reports Physical Science (2022) Machine-Learning.