Bridging Scales, from Atoms to Applications 跨越尺度，从原子到应用
About the Group 研究组简介
ZecYu Group for Multiscale Simulation is an independent, multi-institutional non-profit research collective operating at the convergence of domain-specific physical simulations and autonomous intelligence. We develop and apply computational methods — from first-principles and molecular dynamics to coarse-grained and mesoscale modeling — to address fundamental challenges in construction materials, composite interfaces, and functional nanomaterials.
泽川多尺度模拟研究组是一个跨机构的独立非营利性学术研究集体，致力于领域特定物理模拟与自主智能的融合。 我们开发并应用从第一性原理、分子动力学到粗粒化和介观尺度的计算方法，以解决建筑材料、 复合材料界面和功能纳米材料中的基础性挑战。
The traditional scientific method faces a bottleneck: the complexity of multi-physics, multi-scale material systems creates vast spaces too large for brute-force computation or human intuition alone. We believe the solution lies in AI for Science (AI4S). By integrating molecular dynamics, coarse-grained modeling, and advanced machine learning potentials into automated, agent-driven scientific workflows, we accelerate the discovery, optimization, and characterization of next-generation materials.
传统科学方法面临瓶颈：多物理场、多尺度材料系统的复杂性产生了仅靠暴力计算或人类直觉 无法企及的广阔数据空间。我们相信解决方案在于AI for Science (AI4S)—— 将分子动力学、粗粒化建模与先进的机器学习势函数集成到自主驱动的科学工作流中， 加速下一代材料的发现、优化和表征。
Our mission is built on three pillars:
What We Study 研究领域
From atomistic calcium-silicate-hydrate (C-S-H) to mesoscale cement matrix — developing coarse-grained models and remapping algorithms to predict transport properties, mechanical performance, and durability of cement-based composites.
从原子尺度的水化硅酸钙（C-S-H）到介观尺度的水泥基体，开发粗粒化模型和重映射算法， 预测水泥基复合材料的传输特性、力学性能和耐久性。
Molecular-level understanding of fiber-matrix interfaces — adhesion, creep, and environmental degradation in FRP, CFRP, and steel fiber-reinforced systems. Silane coupling and hydrophobic treatments for enhanced durability.
从分子层面理解纤维-基体界面——FRP、CFRP及钢纤维增强体系中的粘附、蠕变与 环境劣化机制。探索硅烷偶联和疏水处理对界面耐久性的增强。
Chitin/chitosan interfaces, cellulose nanocrystals, topological gels, and hydrogel mechanics — using molecular dynamics to probe structure-property relationships in biological and soft materials.
甲壳素/壳聚糖界面、纤维素纳米晶体、拓扑凝胶和水凝胶力学——利用分子动力学探索 生物与软材料中的结构-性能关系。
Carbon nanotube-reinforced composites, photoacoustic behavior of CNT arrays, nanocellulose dispersibility, and photocatalysis — functional nanomaterials explored through simulation and experiment.
碳纳米管增强复合材料、CNT阵列的光声行为、纳米纤维素分散性以及光催化—— 通过模拟与实验探索功能性纳米材料。
High-throughput molecular dynamics combined with machine learning for accelerated materials discovery. ML-based anomaly detection for structural health monitoring and predictive modeling of material performance.
高通量分子动力学与机器学习相结合，加速材料发现。基于机器学习的结构健康监测 异常检测方法，以及材料性能预测建模。
People 团队成员
Zechuan Yu 余泽川
Principal Investigator 课题组负责人
Zechuan Yu has over a decade of experience in molecular dynamics and multiscale simulation of materials. His research spans cementitious composites, chitin-based biomaterials, polymer interfaces, and data-driven materials modeling. He has authored more than 50 peer-reviewed publications in leading journals including Scientific Reports, ACS Nano, Cement and Concrete Research, and Composites Part B: Engineering.
余泽川在材料分子动力学与多尺度模拟领域拥有超过十年的研究经验。其研究涵盖水泥基复合材料、 甲壳素生物材料、高分子界面以及数据驱动材料建模。在包括Scientific Reports、 ACS Nano、Cement and Concrete Research和Composites Part B: Engineering 等期刊上发表同行评审论文逾50篇。
External Advisory Fellows 外聘顾问研究员
Ao Zhou 周傲
External Advisory Fellow 外聘顾问研究员
Guangdong Provincial Key Laboratory of Intelligent and Resilient Structures for Civil Engineering 广东省智能与韧性结构重点实验室（土木工程）
Huali Hao 郝华丽
External Advisory Fellow 外聘顾问研究员
School of Civil Engineering, Wuhan University 武汉大学土木建筑工程学院
Renyuan Qin 秦任远
External Advisory Fellow 外聘顾问研究员
School of Environment and Civil Engineering, Dongguan University of Technology 东莞理工学院生态环境与建筑工程学院
Wei Jian 简玮
External Advisory Fellow 外聘顾问研究员
Ningbo University 宁波大学
Research Associates 研究助理
Yize Li李逸泽
Research Associate 研究助理
Molecular dynamics simulation for civil engineering materials, with emphasis on interfacial degradation and durability under environmental exposure.
土木工程材料的分子动力学模拟，重点关注环境暴露下的界面退化与耐久性问题。
Jingbo Zhuo卓靖博
Research Associate 研究助理
Multi-scale molecular dynamics of cementitious composites, bridging atomistic C-S-H models to mesoscale structural predictions.
水泥基复合材料的多尺度分子动力学模拟，连接原子尺度C-S-H模型与介观结构预测。
Kai Wang王凯
Research Associate 研究助理
Machine learning potential development for cementitious materials, enabling high-accuracy simulations at reduced computational cost.
水泥基材料的机器学习势函数开发，以降低计算成本实现高精度模拟。
Xuelian Sun孙雪莲
Research Associate 研究助理
Machine learning potential development for metal-polymer interfaces, targeting adhesion and corrosion-resistant coating design.
金属-聚合物界面的机器学习势函数开发，面向粘附与耐腐蚀涂层设计。
Kexuan Li李柯璇
Research Associate 研究助理
Durability and interfacial behavior of fiber-reinforced composites in marine and alkaline environments.
纤维增强复合材料在海洋与碱性环境中的耐久性及界面行为研究。
Chunlong Liu刘春龙
Research Associate 研究助理
Graph neural network development for molecular dynamics, with applications to material property prediction and force field augmentation.
面向分子动力学的图神经网络开发，应用于材料性能预测与力场增强。
Zhenyu Wang王振宇
Research Associate 研究助理
Machine learning assisted molecular dynamics simulation, integrating data-driven models with classical MD for accelerated materials discovery.
机器学习辅助分子动力学模拟，结合数据驱动模型与传统MD以加速材料发现。
Caisheng Wang王才盛
Research Associate 研究助理
Generative models for engineering applications, focused on inverse design of composite materials.
面向工程应用的生成模型研究，专注于复合材料的逆向设计。
Junzhen He何俊臻
Research Associate 研究助理
Artificial intelligence for scientific computing, applying deep learning methods to accelerate multiscale simulations.
面向科学计算的人工智能研究，应用深度学习方法加速多尺度模拟。
Yujun Yang杨喻钧
Research Associate 研究助理
Machine learning potential development for organic-inorganic interfaces, enabling predictive modeling of hybrid material systems.
有机-无机界面的机器学习势函数开发，实现杂化材料系统的预测性建模。
Research Output 研究成果
Multiscale Simulation of Cementitious Materials 水泥基材料多尺度模拟
Interfacial Mechanics in Composites 复合材料界面力学
Biomolecular & Polymer Materials 生物分子与高分子材料
Nanoscale Functional Materials 纳米功能材料
Data-Driven & ML Methods for Materials 数据驱动与材料机器学习
More than 50 peer-reviewed publications since 2014. 自2014年以来发表同行评审论文逾50篇。
Get in Touch 联系我们
Email: 电子邮箱： imiku@zecyu.com
This email domain matches our group website domain (zecyu.com). 邮箱域名与本课题组网站域名一致（zecyu.com）。