基于数理统计方法的锂电池电解液电导率优化设计

Statistics method-based optimization of electrolyte conductivity of lithium-ion battery

离子电导率是评估锂离子电池电解液性能的重要特征参数,直接影响着电池的低温、倍率等性能,在电解液的设计中极具指导价值。传统的电解液研发模式主要基于经验和实验的“试错法”,存在变量多、实验成本高、开发周期长等问题。针对以上问题,本工作提出了一种结合空间填充混料设计与高斯过程回归的电导率优化设计方法,以包括环状碳酸酯(EC)及不同种类线性碳酸酯、羧酸酯的电解液溶剂组成作为模型的输入,电导率作为模型的输出,并运用最大似然估计求解超参数;通过后续实验验证了模型的有效性,并可预测满足任意电导率要求的电解液溶剂配方。

Ionic conductivity is an important parameter in the evaluation of lithium-ion battery electrolyte performance.Ionic conductivity affects the low temperature and rate capability of the battery and provides guiding principles for electrolyte design.Traditional research and development methodologies are primarily based on trial and error,which involves many variables.This results in high experimental costs and a long discovery cycle.To solve the above issues,a conductivity optimization design method that combines a space filling mixture design and Gaussian process regression is proposed in this paper.According to the formulation parameters of the electrolyte,including different types of cyclic carbonate(ethylene carbonate),linear carbonates,and carboxylic acids as the model's input,the ionic conductivity is output by the model,and the maximum likelihood estimation is employed to solve the super parameters.The effectiveness and precision of the proposed model were verified in subsequent experiments,and we found that an electrolyte solvent recipe that satisfies any conductivity requirements can be predicted.