In this talk, I will present and discuss our recently developed methodologies based on evolutionary algorithms and deep neural networks, at interplay with atomistic modeling, to efficiently explore the enormous range of chemical compositions and structures offered by the organic-materials realm. The primary aim is to achieve fundamental understanding of electrochemistry at the atomic level and then to accelerate the discovery of organic battery materials. More specifically, this study provides an assessment of the redox activity of organic materials, even for molecules with complex geometries, creating a framework that can aid the designing of novel organic electrodes.

Machine-Learning and AI-based strategies are key tools to capture the real-world dynamic behavior of batteries and perform predictive analytics. Using dynamic nonlinear model predictive control strategies in the past, we have developed optimal methods of battery charging by modeling and minimizing various battery degradation phenomena, while controlling internal temperatures inside the batteries, using physics-informed Machine Learning/AI models. In collaboration with NREL, we have shown improvement in battery life in excess of 100%.

With the increasing incidence of fire and catastrophic failure of electric vehicles, the necessity of an advanced battery management system and safety framework became crucial. The talk will focus on the importance of intelligent state estimation and control in lithium-ion battery management systems for transportation electrification applications. The current issues and challenges will be also discussed to highlight the scope of future research and development.