Rapid visual simulation of the progressive collapse of regular reinforced concrete frame structures based on machine learning and physics engine

Senna Wang,Xiaowei Cheng,Yi Li,Xiaoyan Song,Ruijie Guo,Haoyou Zhang,Zihan Liang
DOI: https://doi.org/10.1016/j.engstruct.2023.116129
IF: 5.582
2023-07-01
Engineering Structures
Abstract:Assessing the collapse region in the progressive collapse of buildings is one of the important issues in urban planning for disaster recovery. A rapid visual simulation method for the progressive collapse of regular reinforced concrete (RC) frame structures was proposed. First, a modeling solution using affine transformation and modification script was designed to rapidly create a geometric model. Second, a machine learning model of the internal collapse area and a horizontal collapse propagation algorithm were built to predict the internal collapse region. Third, a visual simulation method based on Voronoi tessellation and physics engine was proposed to represent the collapse process of the RC frame. Finally, the accuracy and efficiency of the proposed method were validated using the collapse event of the Murrah Federal Building and finite element models, and the effect of the mesh size of Voronoi tessellation and different initial failures on a collapse scenario was further discussed.
engineering, civil
What problem does this paper attempt to address?