Statistical Characteristics of Extreme Wind Duration Using Surface Observations in the United States

Le résumé fourni par la source

ABSTRACT The current prescriptive wind design heavily depends on the statistical analysis of wind speed and direction, yet the duration of wind events has received limited attention due partially to its insignificant effects on linear, elastic structural response. However, the recent trend towards the performance-based wind design (PBWD), by allowing for limited inelastic behavior of structural components, poses a concern on low-cycle fatigue (along with cyclic deterioration of strength and stiffness) that depends on wind duration (above a high threshold speed). In addition, the operation of smart infrastructure systems heavily relies on wind duration (above a relatively low threshold speed). Following the methodology adopted in the wind speed map development of ASCE 7, this study investigates the wind duration statistics of extreme thunderstorm and synoptic events based on surface observations. Specifically, the non-hurricane non-tornadic wind events obtained from the automated surface observing systems (ASOS) stations across the United States are classified into the thunderstorm and synoptic categories, and then the wind records from each category are preprocessed to make them suitable for duration statistical analysis. The analysis first briefly explores the geographical characteristics of mean wind duration in the United States. Then, a comprehensive investigation of wind duration statistics for a selected region is conducted, including wind duration dependence on speed threshold, wind duration probability distribution, extreme value distribution of wind duration, evolutionary and ramp-up wind speed characteristics, and the relationship among wind duration, speed and direction. The obtained statistical characterization reveals that synoptic events exhibit substantially longer above-threshold persistence than thunderstorms. Also, exceedance durations exhibit strong threshold dependence. Specifically, the conditional median durations at the 50-year wind speed with a threshold of 70 mph are approximately 1.5 minutes and 3.5 minutes for thunderstorms and synoptic events, respectively. However, these values increase to approximately 3 hours for thunderstorms and 31 hours for synoptic events with a threshold of 29 mph. While the current 1-hour duration assumption in PBWD focuses on the duration of peak intensity, the present findings provide a complementary perspective by capturing event duration above multiple thresholds. The framework developed here can be applied at any threshold relevant to cumulative effects such as rational design of loading protocols (for quantifying low-cycle fatigue) and effective operation of smart infrastructure systems.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.