Data-driven Subcomponent Design and Engineering

Developing new and effective energy storage technologies is a complex challenge that is difficult to tackle through trial-and-error experimentation alone. However, computational simulation, in conjunction with data science, has emerged as a complementary tool to accelerate the development process and enhance understanding. In the first part of this talk, we will present various case studies, such as a multi-faceted high-throughput screening method, a query-based method utilizing material databases, and a descriptor-based materials design. In the second half of the talk, we will discuss the multi-scale design and product integration efforts at Rivian Automotive. Our approach encompasses the smallest scale (e.g., materials), subcomponents (e.g., electrodes), and entire systems (e.g., battery packs, electric vehicles). While there is a strong focus on developing better "predictive" models that require less experimental input, there are still several ongoing challenges in utilizing the data collected from R&D, pilot testing, and the automated factory floor. Data-driven subcomponent engineering represents one of the key methods that can help accelerate the EV transition underway at Rivian.