Deng Changyu, Kim Andrew, Lu Wei
Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109, USA.
Department of Materials Science and Engineering, University of Michigan, Ann Arbor, MI 48109, USA.
Patterns (N Y). 2022 Jun 20;3(7):100531. doi: 10.1016/j.patter.2022.100531. eCollection 2022 Jul 8.
Battery optimization is challenging due to the huge cost and time required to evaluate different configurations in experiments or simulations. Optimizing the cycling performance is especially costly since battery cycling is extremely time consuming. Here, we introduce an optimization framework building on recent advances in machine learning, which optimizes battery parameters efficiently to significantly reduce the total cycling time. It consists of a pruner and a sampler. The pruner, using the Asynchronous Successive Halving Algorithm and Hyperband, stops unpromising cycling batteries to save the budget for further exploration. The sampler, using Tree of Parzen Estimators, predicts the next promising configurations based on query history. The framework can deal with categorical, discrete, and continuous parameters and can run in an asynchronously parallel way to allow multiple simultaneous cycling cells. We demonstrated the performance by a parameter-fitting problem for calendar aging. Our framework can foster both simulations and experiments in the battery field.
由于在实验或模拟中评估不同配置需要巨大的成本和时间，电池优化具有挑战性。优化循环性能尤其昂贵，因为电池循环极其耗时。在此，我们引入了一个基于机器学习最新进展构建的优化框架，该框架能有效优化电池参数，显著减少总循环时间。它由一个剪枝器和一个采样器组成。剪枝器使用异步连续减半算法和超参数优化算法，停止没有前景的循环电池，以节省进一步探索的预算。采样器使用帕曾估计树，根据查询历史预测下一个有前景的配置。该框架可以处理分类、离散和连续参数，并且可以以异步并行方式运行，以允许多个电池同时循环。我们通过一个日历老化的参数拟合问题展示了其性能。我们的框架可以促进电池领域的模拟和实验。
