论文标题
基于LI的电池材料数据库的高通量计算：化学过程 - 培训关系
High-Throughput Computation of Li-based Battery Material Databases: Chemistry-Processing-Property Relationships
论文作者
论文摘要
在本文中，我们将处理纳入基于LI的尖晶石电池材料的数据驱动属性建模中。当将lime2O4形式的尖晶石化合物与我一起作为金属或金属时，有125,000个可能的组合假设最多三个金属元素。在这项工作中，我们专注于预测财产的能力，并通过将处理纳入建模来将可能的组合数量增加到200万。由于处理与属性之间的非线性关系以及确保适当的灵敏度，因此引入了一种随着变化处理而跟踪变化的新方法。这项工作为指导下一代电池实验提供了宝贵的工具，从而大大缩小了无限的设计空间。
In this paper, we incorporate processing into the data driven property modeling of Li-based spinel battery materials. When considering spinel compounds of the form LiMe2O4 with Me as a metal or metals, there are 125,000 possible combinations assuming a maximum of three metallic elements. In this work, we focus on capacity for our predicted property, and increase the number of possible combinations to two million by incorporating processing into the modeling. Due to the non-linear relationship between processing and property, as well as ensuring proper sensitivity, a new approach which tracks the change with changing processing is introduced. This work provides an invaluable tool for guiding the next generation of battery experiments, providing a significant narrowing of the infinite design space.
