An Improved Seismic Damage Assessment Model for Buildings Based on Parallel Algorithms

The efficiency of post-quake emergency management depends heavily on the speed of seismic damage assessment software, which depends on the evaluation algorithm essentially. Traditional seismic damage assessment models based on sequence algorithms usually need tens of minutes to get visualized results. More efficient computing models for seismic damage assessment are demanded urgently. General parallel evaluation algorithm based on GPU computing framework was investigated by the authors, and an improved seismic damage assessment model for buildings was proposed in this paper. We designed a three-level parallel scheme to accelerate the computing and visualization. At the top level, the earthquake area is divided into smaller regions, such as cities or counties. At the middle level, fundamental geographic data for each region is split into layers by features. At the bottom level, buildings are grouped into several types, which can be processed by GPU cores simultaneously. The developed model can estimate and analyze the seismic damage of buildings and the spatial distribution of victims in earthquake region more rapidly than traditional models.