机器学习在设计高性能锂电池正极材料与电解质中的应用

Application of Machine Learning in the Design of Cathode Materials and Electrolytes for High-Performance Lithium Batteries

随着大数据和人工智能的发展以及机器学习(ML)与化学学科领域的交叉,ML技术与电池领域的结合激发了更有前途的电池开发方法,尤其在电池材料设计、性能预测、结构优化等方面的应用愈加广泛。应用ML可以有效地加速电池材料的筛选进程并预测锂电池(LBs)的性能,从而推动LBs的发展。本文简要介绍了ML的基本思想及其在LBs领域中几种重要的ML算法,之后讨论了传统模拟计算方法与ML方法各自的误差表现及分析,借此来提高LBs专家对ML方法的理解。其次,重点介绍了ML在电池材料实际开发中的应用,包括正极材料、电解质、材料多尺度模拟及高通量实验(HTE)等方面,借此介绍ML方法在电池领域应用的思想和手段。最后,总结了ML方法在锂电池领域中的研究现状并展望了其应用前景。本综述旨在阐明ML在LBs开发中的应用,并为先进LBs的研究提供借鉴。

The rapid application of big data and artificial intelligence, and the deep intersection of machine learning (ML) and chemistry disciplines have inspired more promising development approaches for the integration of ML technology with battery materials, especially in the material design of battery, performance prediction, structure optimization, and so on. The application of ML can effectively accelerate the selection process of battery materials and predict the performance of lithium batteries (LBs), consequently driving the development of LBs. This review briefly introduces the basic idea of ML and several important ML algorithms in the field of LBs, then the error performance and analysis of the traditional simulation calculation method and ML method are discussed, thereby increasing understanding of ML methods by LBs experts. Secondly, the application of ML in the practical development of battery materials, including cathode materials, electrolytes, multi-scale simulation of materials and high-throughput experiments (HTE), is emphatically introduced to draw out the ideas and means of applying ML methods in the field of batteries. Finally, the recent works of ML in lithium batteries are summarized and their application prospects are foreseen. It is hoped that this review will shed light on the application of ML in the development of LBs and promote the development of advanced LBs.