材料科学
贝叶斯优化
阳极
概率逻辑
可扩展性
光热治疗
计算机科学
亚氧化物
退火(玻璃)
纳米技术
吞吐量
介孔材料
激光器
储能
模拟退火
硅
热的
过程(计算)
工艺优化
电极
电子设备和系统的热管理
纳米结构
化学过程
想象
贝叶斯网络
工艺工程
高效能源利用
光电子学
作者
Chaeyoung Park,Yeongje Lee,Sunho Jeong,Eungkyu Lee
摘要
ABSTRACT Silicon suboxide (SiO x ) has attracted significant attention as a promising anode material for next‐generation lithium‐ion batteries due to its high capacity and improved structural stability. However, precise control over stoichiometry and pore structure remains challenging under conventional thermal annealing. Herein, we introduce an empirical‐aided active learning (EAAL) approach to optimize high‐throughput laser‐induced photothermal annealing of high‐performance SiO x anodes. By synergistically combining probabilistic machine learning with empirical domain knowledge, the EAAL framework efficiently explores complex processing parameter spaces using a limited number of experiments. Remarkably, the EAAL‐optimized anode achieves electrochemical performance comparable to that of an empirically optimized counterpart while significantly reducing the number of laser irradiation passes from five to two, thereby improving practical production throughput in cumulative laser‐processing. Furthermore, the optimal conditions are identified using only 26 experimental data points, significantly improving process efficiency. This study demonstrates a data‐efficient strategy for process optimization by synergistically integrating machine learning with empirical domain knowledge and provides a scalable pathway for the advanced manufacturing of high‐performance energy storage materials.
科研通智能强力驱动
Strongly Powered by AbleSci AI