自复位学校建筑抗震韧性区域评估用数字孪生模型

Digital Twin Model for Regional-Scale Seismic Resilience Assessment of School Buildings with Modular Retrofit Panel System

摘要:

在区域层面上对具有自复位模块结构的学校建筑进行抗震损伤和恢复力评估研究，建立了一个基于Python语言的从结构非线性分析到损伤程度可视化演示用的数字孪生模型。Python语言也是目前通用于机器学习模型训练的编程语言，方便在该数字孪生模型中引入人工智能模型来代替结构仿真计算，通过使用实时监测地震动和结构响应数据结合人工智能模型进行区域灾害响应和功能恢复快速评估。结构损伤和恢复力分析结果（如不同地震下修复成本、修复时间、不可修复的概率等指标）通过生成 shapefile在地理信息系统（GIS）软件中进行三维可视化，从而对采用自复位模块结构的学校建筑结构群在区域范围上进行抗震韧性定量评估。该模型的计算模块包括区域建筑结构清单生成、简化数值模型建立、地震响应非线性分析、结构响应参数生成、建筑结构易损部件定义、地震损失概率模型评估以及区域灾害损失的结果输出。这里的基于概率模型的学校建筑结构地震损失评估采用了FEMA P-58 方法，并使用了 Pelicun 软件包进行计算。以旧金山湾区近2 000栋学校建筑作为案例，对假定使用偏心支撑框架自复位结构作为抗震结构体系的学校建筑群进行了抗震韧性评估，研究了通过采用数字孪生模型对新型结构模块系统在学校建筑群区域灾害响应和功能恢复的影响。研究中使用了简化结构模型来缩短震后损伤和功能恢复仿真运算时间，同时提出了通过采用人工智能模型来实现强震后实时预测建筑群震后损失，并通过研究偏心支撑框架自复位结构体系的能量耗散比这一参数变化对学校建筑结构恢复力指标的影响，演示了采用不同的结构设计对建筑结构群在区域层面上的抗震损伤和恢复力的影响。

Abstract:

With the goal to develop a digital twin model with a seamless procedure for performing an intensity-based seismic resilience assessment of school buildings with self-centering modular bracing panel （SCMBP） systems on a regional scale， a computational framework comprised of sequential steps was built in the Python programming language by adopting multiple packages. The results of the analysis （e.g.， repair cost， repair time， probability of irreparability， etc.） were generated in different contexts such as graphs， tables， and multiple shapefiles containing the building footprints and resilience metrics such as repair time and repair cost at different seismic intensities that could be visualized three-dimensionally in geographical information system （GIS） software to present a more intelligible quantitative evaluation of the regional seismic loss of the building inventory with a retrofit modular bracing panel system. The steps consisted of generating the building inventory， generating simplified numerical models， response history analysis （RHA），obtaining engineering demand parameters （EDPs），estimating the quantity of the vulnerable components，probabilistic seismic loss assessments， and generating the building-specific and regional outputs. The probabilistic loss assessment was performed based on the component-level FEMA P-58 methodology by adopting the Pelicun package. As a case study， the regional seismic resilience assessment of buildings equipped with SCMBP systems was conducted by performing a study of nearly two thousand school buildings in the San Francisco Bay Area with such systems. A simplified structural model for simulating the SCMBP systems was adopted to reduce the computing time of regional-scale seismic resilience evaluation while exhibiting an identical story-shear hysteretic behavior. The effect of the key parameter of the energy dissipation ratio， β， of SCMBP systems on the resilience metrics of the school buildings was studied by performing a parametric study.