| 唐元晖,宋岩,何晓磊,张志军,张春晖,余立新,王晓琳,林亚凯.化学通报,2026,89(6):677-687,676. |
| 机器学习驱动膜材料设计与优化的研究进展 |
| Research Progress on Machine Learning-Driven Design and Optimization of Membrane Materials |
| 投稿时间：2026-02-23 修订日期：2026-04-02 |
| DOI： |
| 中文关键词: 机器学习 膜材料 材料信息 膜制备 性能优化 |
| 英文关键词:machine learning, membrane materials, materials informatics, membrane fabrication, performance optimization |
| 基金项目:中央高校基本科研业务费专项资金(2022YQHH04,2022YJSHH14,2023ZKPYHH04)、北京市自然科学基金项目(2232018,2252036)、国家自然科学基金面上项目(22378225)和北京市高层次创新创业人才支持计划科技新星计划项目(202504841001)资助 |
|
| 摘要点击次数: 20 |
| 全文下载次数: 0 |
| 中文摘要: |
| 作为膜分离技术的核心,膜材料的设计、制备和性能优化是膜领域的重要研究方向。然而,不同类型分离膜在制备过程中普遍面临原料来源广泛、制备方法多样、工艺参数复杂以及成膜机理尚未完全阐明等问题。尽管现有模型与经验方法在一定程度上能够为膜材料设计提供指导,但受限于影响因素维度高、耦合关系复杂,仍难以实现对新型膜材料的高效开发及膜性能的精准调控。近年来,机器学习(ML)技术在算法与工程应用层面取得了显著进展,并逐步引入膜材料研究领域,在气体分离膜、纳滤/反渗透膜等体系的材料筛选、性能预测和工艺参数优化中获得了成功应用,为膜材料的智能化设计与高效开发提供了新的技术路径。本文系统综述了ML驱动的膜材料设计与优化研究进展：首先阐述了基于ML的研究流程,包括数据挖掘、特征工程及常用算法的优势和局限；随后,重点评述了ML在不同类型膜材料设计与性能优化中的典型应用与最新进展；最后,对ML技术在膜材料领域的发展趋势与面临的关键挑战进行了展望。 |
| 英文摘要: |
| As the core of membrane separation technology, the design, fabrication, and performance optimization of membrane materials have become key research directions in the membrane field. However, different types of separation membranes generally face common challenges during fabrication, including diverse raw material sources, multiple preparation methods, complex processing parameters, and not yet fully understood membrane formation mechanisms. Although existing theoretical models and empirical approaches can provide certain guidance for membrane material design, the high dimensionality of influencing factors and the complexity of their coupled relationships still make it difficult to achieve the efficient development of novel membrane materials and precise control of membrane performance. In recent years, Machine Learning (ML) has found broad and successful applications in material screening, performance prediction, and process parameter optimization for gas separation membranes as well as nanofiltration/reverse osmosis membranes, highlighting its unique advantages and enabling new strategies for the intelligent design and efficient optimization of membrane materials. This review systematically summarizes recent progress in ML-driven membrane material design and optimization. First, it outlines the general ML-based research workflow, including data mining, feature engineering, and the advantages and limitations of commonly used algorithms. It then highlights representative applications and the latest advances of ML in the design and performance optimization of various types of membrane materials. Finally, the future development trends and key challenges of ML technologies in the membrane materials field are discussed. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
京公网安备11010802047317号