A model for probable maximum loss in hurricanes

The quantitative prediction of wind damage to structures has emerged in recent times as a major issue confronting construction and insurance industry personnel as well as emergency management planners. While a majority ofthe structures in the coastal areas exposed to hurricanes are non-engineered damage-prone buildings, these same buildings are insured against wind damage. Reliable insurance underwriting for structures, efficient wind damage mitigation programs, and post-storm recovery strategies are all dependent upon good quantitative damage prediction.

A new approach using the concept of building damage bands is proposed for predicting the probable maximum damage degree to individual buildings or groups of buildings for any given hurricane scenario. The damage prediction model employs an objective weighting technique driven by component cost factors, conditional failure probabilities, and location parameters to obtain upper and lower hurricane damage thresholds. Component failure probabilities were obtained via a quasi fault tree analysis of a multiple fault tree scheme in which the damage of individual building envelope components, including the building interior, serve as the Top events. Important phenomena such as damage propagation and common cause failures which complicate the wind damage process were considered in the present model.