During my internship at Volkswagen, I was able to work on a really cool ML & optimization problem: how can we streamline the materials design process for an EV battery? Traditionally, running electrochemical and physics-based simulators, such as COMSOL, for a given set of battery designs can take weeks & even months. But what if this process could be accelerated to a time-scale of minutes? I looked at different machine learning techniques (Gaussian Process Regression) for creating a model that was highly accurate in its predictions. I took this model one step further to develop a method for “searching” for battery material parameters that can produce the most “optimal” performance using Bayesian Optimization. I was able to create a framework that could screen through millions of battery design candidates in only milliseconds per prediction. My former colleagues and I are in the process of writing a publication detailing the work that we did, but our abstract has already been accepted.
Zhang et al. (2020) demonstrates "materials design".
Abstract Title: AI-based High-throughput Screening Framework for Battery Materials Design
Abstract: Identification of the optimal combination of parameters in materials design requires large effort since many possible candidates have to be evaluated by experiments or simulations. Creating synthetic materials already decreases the effort significantly, but one still needs to focus on a reduced selection of material combinations. We propose a data-driven high-throughput screening framework for materials design that allows materials screening of millions of candidates in only milliseconds per prediction. We demonstrate the feasibility
of our approach on the design of battery cathode materials combining synthetic microstructure generation and electrochemical modeling considering also battery cell properties. We apply simple Machine Learning models on averaged microstructure properties for materials screening and complex Deep Learning models on the 3D microstructures to enable generative materials design. We use model uncertainty to efficiently create new simulation data samples for incremental model improvement. The most promising candidates selected by materials screening are then validated with simulations.
Authors: Melanie Senn, Nasim Souly, Alina Negoita, Prateek Agrawal, Christian Tae, Vedran Glavas, Julian Wegener, Kai Gerstner, Alex Alekseyenko
of our approach on the design of battery cathode materials combining synthetic microstructure generation and electrochemical modeling considering also battery cell properties. We apply simple Machine Learning models on averaged microstructure properties for materials screening and complex Deep Learning models on the 3D microstructures to enable generative materials design. We use model uncertainty to efficiently create new simulation data samples for incremental model improvement. The most promising candidates selected by materials screening are then validated with simulations.
Authors: Melanie Senn, Nasim Souly, Alina Negoita, Prateek Agrawal, Christian Tae, Vedran Glavas, Julian Wegener, Kai Gerstner, Alex Alekseyenko
Abstract was accepted to 6th World Congress on Integrated Computational Materials Engineering (ICME 2021).