华东交通大学
土木建筑学院 · 全国重点实验室 · 岩土工程(智能建造方向)硕博连读
- 长期围绕深基坑、隧道与工程智能监测开展研究,方向覆盖结构安全感知、智能分析与辅助决策。
- 导师:徐长节教授、丁海滨副教授;依托实验室开展工程 AI、监测预警与数字建造相关研究。
- 持续以英文文献、国际会议报告与跨学科合作方式拓展研究视野,具备良好的英文科研交流能力。
岩土工程(智能建造方向)硕博连读,博士三年级;导师为徐长节教授、丁海滨副教授;当前以博士毕业、代表性成果产出与工程化项目推进为主线。
既能做理论推导、AI 建模与论文写作,也能完成系统架构设计、云端平台开发、数据接入与工程现场协同,并在多角色协作中承担资源整合、任务拆解、团队推进与方向把控。
持续以博客与公众号记录智能建造与岩土信息化探索,同时通过 GitHub 与工具化实践沉淀个人作品和研发方法。
具备跨专业团队组织与持续运转经验,能够围绕目标进行方向拆解、资源对接、进度把控与成员协同;曾组织 20+ 人团队,推动国省级项目、竞赛与研发任务并行落地。
土木建筑学院 · 全国重点实验室 · 岩土工程(智能建造方向)硕博连读
大土木实验班 · 本科
注:作者名后的 * 号表示通讯作者。
基于最小势能原理推导半解析解,面向不对称开挖与附加载荷场景建立围护体系整体变形评估方法,并结合工程监测数据验证。
将改进遗传算法与多目标优化结合,面向复杂不对称荷载工况实现自动化逆向设计,文中给出寻优效率约提升 30%、变形控制精度提升约 25% 的验证结果。
以 IFC 语义拓展为核心,把基坑围护结构、地质信息、自动化力学分析与 BIM 建模打通,实现三维模型直生与设计流程的统一表达。
针对 BIM/IFC、仿真模型、规范与施工图之间的语义割裂问题,提出多 LOD 统一表达与 Mapper 链,实现几何、材料、结构、阶段与监测信息的可执行映射。
围绕卫星 InSAR、无人机与地面仪器融合,构建 SAG 监测框架与七步实施流程,梳理规范接纳度、数据融合策略及 AI 预测方法。
围绕“建项—接入—监测—判定—处置—复盘”闭环设计软件体系,融合 BIM/IFC、Plaxis、优化算法、多源监测与 AI 预测,实现可追溯的安全—经济—环保一体化决策。
提出 PCRS 框架,将多任务 U-Net、渗流物理模型与可学习软修正机制耦合,在现场校准场景中将物理残差降低约 77%,并嵌入 NSGA-II 进行设计优化。
面向基坑支护与厂房框架等耦合结构,先预测位移场再反推连接件受力,结合质控与现场实时校准,将计算速度提升至毫秒级,适配实时监测场景。
构建“双目视觉 + 语义分割 + 几何映射”流程,在隧道与外墙场景中实现对细小裂缝位置与宽度的稳定量测,弱化对传统立体匹配关键点的依赖。
形成“定位—分割—量化”级联流水线,结合注意力增强检测与迁移学习分割,实现裂缝宽度统计效率提升约 5 倍,支撑隧道巡检自动化。
Year-3 PhD candidate at the National Key Laboratory for Safety and Resilience of Mountain Civil Engineering, East China Jiaotong University. My work sits at the intersection of AI, geotechnical engineering, digital construction, and engineering software products, covering automated design of excavation support systems, monitoring & early-warning platforms, BIM/IFC-to-simulation workflows, and physics-informed AI surrogates, with additional experience in resource integration, team coordination, and technical direction setting.
Combined MSc–PhD student in geotechnical engineering, advised by Prof. Changjie Xu and Assoc. Prof. Haibin Ding; currently focused on PhD completion, representative research outputs, and engineering-oriented delivery.
I can move from theory derivation and AI modeling to system architecture, data pipelines, product interfaces, and on-site deployment, while supporting resource coordination, task decomposition, team execution, and direction control.
I continuously document smart-construction practices through blogs and public writing, and preserve tool-based workflows through code and reusable engineering assets.
I have hands-on experience organizing cross-disciplinary teams, aligning resources with objectives, decomposing workstreams, tracking progress, and coordinating people; this includes leading a 20+ member studio across projects, competitions, and engineering tasks.
School of Civil Engineering and Architecture · National Key Laboratory · Combined MSc–PhD in Geotechnical Engineering (Smart Construction)
BEng · Grand Civil Experimental Class
Note: * denotes corresponding author.
Derived a semi-analytical solution from the minimum potential energy principle to evaluate global retaining-system deformation under asymmetric excavation and surcharge conditions, with field-data validation.
Combined an improved genetic algorithm with multi-objective optimization to automate inverse design for complex asymmetric loading scenarios, targeting safer and more economical excavation support schemes.
Expanded IFC semantics to cover retaining structures and geological entities, then connected BIM models with mechanics simulation for direct 3D generation and streamlined design automation.
Proposed a multi-LOD information framework and executable mapper chain that links geometry, materials, structures, phases, monitoring, and project semantics across BIM and simulation ecosystems.
Built a monitoring review and framework that combines InSAR, UAV sensing, and high-precision ground instruments, mapping standards, fusion strategies, and AI prediction methods into an operational workflow.
Designed an end-to-end software loop spanning project setup, data access, monitoring, rule evaluation, response, and review, while integrating BIM/IFC, simulation, optimization, sensing, and AI forecasting.
Developed the PCRS framework by coupling multi-task U-Net, seepage physics, and a learnable soft-correction mechanism, enabling design optimization after strong residual reduction in calibrated field settings.
Predicted displacement fields first and then recovered connector forces for coupled systems such as retaining structures and industrial frames, enabling fast and stable safety inference with on-site calibration.
Designed a full pipeline using stereo vision, semantic segmentation, and geometric mapping to measure tiny cracks robustly in tunnels and facades without relying heavily on conventional stereo keypoint matching.
Built a cascade from detection to segmentation to quantitative measurement, improving efficiency for tunnel lining inspection and supporting scalable automated maintenance workflows.