Double-layer platform of high performance computing for damage prediction of rc buildings subject to earthquake and subsequent tsunami in a target area

Abstract

For tsunami scenarios, evaluation of tsunami load acting each building depends on surrounding circumstance. Therefore, modeling of all buildings in a target area is important for damage prediction from earthquake and subsequent tsunami. In this study, sequential earthquake and tsunami simulation was developed to predict structural damage from earthquake and subsequent tsunami by means of the application in Integrated Earthquake Simulation (IES). Since IES can simulate only earthquake scenarios with bean-column frame models, IES was modified for input tsunami load acting on a proposed wall-frame model in order to simulate tsunami scenarios using predicted data of tsunami inundation depth. A target area in Kochi city was selected to simulate an earthquake and tsunami scenario because this area has many public buildings and is important for economic activities. A double-layer platform of high performance computing was proposed to simulate this earthquake and tsunami scenario with parallel processing on CPUs and GPUs. The results of sequential earthquake and tsunami simulation show that three-story RC buildings had a significant risk that maximum drift ratio could occur during sequential tsunami response. However, maximum drift ratio from sequential tsunami response was still less than 0.3% in which structural damage didn't occur obviously. For the worst case scenario that tsunami inundation depth was double, structural damage from sequential tsunami response was much more serious than that of the normal-case scenario in which maximum drift ratio was less than 5% for a four-story RC building. In addition, it was found that low-rise buildings (three- to seven-story) had a significant risk that maximum drift ratio was higher than 1% during sequential tsunami response. The results of sequential earthquake and tsunami simulation can be used to construct further prevention measures.