Real-time forecast of hurricane-induced damage risk to envelope systems of engineered buildings through metamodeling

Abstract

The growing intensity and frequency of extreme wind events, together with the continued economic and population growth of coastal areas, has created an urgent need for methodologies that can inform emergency responders and managers of the increasing risk to the infrastructure of these areas. To this end, this work outlines a real-time risk forecast framework for hurricane-induced damage to building envelope systems of engineered buildings. Damage is quantified through a recently introduced multi-demand and coupled progressive fragility model, with a full range of uncertainty in structural properties, capacities, and wind stochasticity. To enable real-time assessment, an efficient Kriging metamodel is introduced to capture the damage statistics conditioned on intensity measures. From official real-time hurricane advisories, site-specific intensity measures are forecast based on a parametric wind field model while considering the uncertainty in, among others, the hurricane track, pressure deficit, and filling model. Damage risk is predicted through propagating uncertainty by Monte Carlo simulation through the Kriging metamodels calibrated to forecast intensity measures. For illustration, the real time damage risk of the envelope system of a 45-story building located in Miami, Fl, was estimated for hurricanes Matthew (2016), Irma (2017), and Dorian (2019). The efficiency and accuracy of the proposed framework is demonstrated.