Ice storms are one of the most devastating natural hazards which have the potential to inflict significant damage to the built environment. The multi-hazard nature of ice events complicates the analysis of their induced risk due to their inherent nonlinear interactions. In addition, the concurrent and interacting hazards are often responsible for several aerodynamical/dynamical instabilities such as the galloping mechanism. Moreover, the existing risk approaches are usually not suited for large-scale risk evaluation over extended geographical regions due to the involved high-computational costs. Therefore, in this study, a novel data-driven multi-scale performance-based ice engineering (PBIE) framework is developed to support the design of new structures subjected to ice storms or the rehabilitation of existing ones. In addition, the proposed PBIE is capable of rapidly estimating the real-time risk over an extended region due to an ice event. Specifically, it leverages the superior capabilities of state-of-the-art data-driven techniques (e.g., machine learning) to efficiently generate the corresponding risk maps and identify the high-risk areas. The proposed PBIE framework is applied to a simplified example in which the galloping-induced risk on iced conductors, in terms of the galloping amplitude, is evaluated for both local and regional scales. The resulting PBIE framework can be readily applied for design or retrofitting purposes or integrated within an emergency response management system to inform preventive actions that can mitigate the ice storm-induced damages and save lives.
Read full abstract