Building specific catastrophe modeling for mid/high rise buildings supported by wind tunnel data

Abstract

Catastrophe modeling has been widely used to optimize portfolio management and facilitate public decision making for hazard mitigation. Among different natural perils, wind hazard imposed risk is one of the most significant. Three main aspects involved in catastrophe modeling includes hazard quantification, vulnerability assessment and monetary/economic impact prediction. Hazard quantification has been advanced in past decades due to advanced observational records, more powerful computational facilities and extensive research activities. Exposure data used for predicting monetary impact is also improved with more information collected. The estimation of building component failure includes significant uncertainties for mid/high rise buildings located in those coastal metropolitans, when generic models are applied. It is common for mid/high rise commercial buildings that geometry is unique and irregular and the surrounding building conditions are complex. Typically, wind tunnel tests are conducted to obtain accurate design pressures for cladding system and structural wind loads. Code calculated wind pressures are often used in modeling risk of building envelope breach when a generic model is considered. However, due to the complexity of the surrounding buildings and unique geometry of the building, code calculated pressures could introduce large uncertainties for specific buildings. The measured pressure distribution from the wind tunnel for specific building can be used to accurately quantify the risk of building envelope breach, which is one of the key failure modes for mid/high rise buildings. Structural failure is typically not being modeled for mid/high rise engineered buildings. However, there is still chances that the structural damage could occur at very high wind speed. The impact of considering the structural failure is investigated in this study. The developed vulnerability curve based on pressure calculated for generic model is compared to that developed by using building specific dataset.