项目来源
国家重点研发计划(NKRD)
项目主持人
顾祥林
项目受资助机构
东南大学
项目编号
2022YFC3803004
立项年度
2022
立项时间
未公开
研究期限
未知 / 未知
项目级别
国家级
受资助金额
285.00万元
学科
城镇可持续发展关键技术与装备
学科代码
未公开
基金类别
“城镇可持续发展关键技术与装备”重点专项
关键词
城市工程系统 ; 多灾害耦合 ; 韧性评估 ; 可视化应用平台 ; 人工智能 ; Urban engineering systems ; Multi-hazard coupling ; Resilience assessment ; Visualization application platform ; Artificial intelligence
参与者
冯德成
参与机构
同济大学；东南大学土木工程学院
项目标书摘要：随着城市化进程的不断推进，确保城市基础设施群具有足够的安全性能以及抗灾韧性变得至关重要。为考虑多灾害及其耦合作用以有效展开城市韧性评价，本研究致力于研究多灾害耦合作用机理、城市基础设施单体的致灾机理、城市工程系统的级联失效机理，并探索多灾害耦合激励下的城市工程系统多尺度多分辨率区域韧性评估框架，力求搭建涵盖不同城市特征的工程系统多灾害韧性智能评估平台。现阶段已开发了人工智能驱动的易损性快速预测模型和物理增强的结构响应预测机器学习模型，构建了城市工程系统的多分辨率韧性评估方法，初步确定了城市工程系统的韧性评估平台架构。具体而言：首先，基于表格数据提出单体基础设施地震概率易损性快速评估模型，并基于时序数据提出了考虑不确定性的单体基础设施非线性时程响应预测方法。考虑到获取高保真数据的成本巨大，提出了多保真数据驱动的基础设施安全状态预测机器学习方法，可有效利用高保真数据精度高、低保真数据易获取的各自优势。其次，考虑多种因素构建了考虑时变效应的单体基础设施抗震韧性量化框架，确定了结构震后损失的主要来源。在区域层次，基于建筑倒塌瓦砾分布、道路交通系统通行效率与可达性、以及消防系统响应覆盖率与响应时间，建立了“建筑群—道路交通—消防”多系统灾害链功能韧性量化分析方法，基于数据驱动实现了区域桥梁群的全寿命周期地震损伤快速评估。最后，确定平台架构的核心为信息感知层、互联网络层、韧性评估层、平台应用层。
Application Abstract: With the continuous advancement of urbanization,it is crucial to ensure that urban infrastructures are equiped with sufficient safety performance and disaster resilience.To effectively carry out urban resilience evaluation considering multiple disasters and their coupling effects,the coupling mechanism of multiple disasters,the disaster-causing mechanism of individual urban infrastructure,and the cascading failure mechanism of urban engineering systems are researched.Moreover,an intelligent multi-hazard resilience assessment platform for engineering systems covering different urban characteristics will be built after exploring a multi-scale and multi-resolution regional resilience assessment framework for urban engineering systems under the stimulation of coupled multiple disasters.At the current stage,an artificial intelligence-driven vulnerability rapid prediction model and a physics-enhanced structural response prediction machine learning model have been developed.Meanwhile,a multi-resolution resilience assessment method for urban engineering systems has been constructed,and the architecture of resilience assessment platform for urban engineering systems has been initially determined.Specifically,a rapid assessment model for the probabilistic seismic vulnerability of individual infrastructure based on tabular data and a method for predicting the nonlinear time-history response of individual infrastructure considering uncertainty based on time-series data were proposed.Considering the huge burden of obtaining high-fidelity data,a machine learning method for predicting the safety status of infrastructure driven by multi-fidelity data was proposed.Taking into account multiple factors,a seismic resilience quantification framework for individual infrastructure considering time-varying effects was constructed,and the main sources of post-seismic structural loss was determined.At the regional level,a quantitative analysis method for the functional resilience of the"building group-road traffic-fire"multi-system disaster chain was established.Furthermore,a rapid assessment of seismic damage throughout the lifecycle of regional bridge groups based on data-driven methods was achieved.Finally,the information perception layer,the interconnected network layer,the resilience assessment layer,and the platform application layer were determined as core of the platform’s architecture.
项目受资助省
江苏省
