A Method for Large-Span Roof Wind-Load Spectrum Construction: Integrating Meteorological Measured Data With Wind Field Template

Large-span metal roofs are widely applied in large-scale public buildings such as transportation hubs and stadiums due to their high space utilization efficiency and excellent architectural adaptability. However, existing studies have focused on the fatigue performance of local components, lacking a global analysis of the synergistic mechanism between full-scale structural systems and real-world environmental loads. To address these issues, this article proposes a method for large-span roof wind-load spectra construction by integrating meteorological measured data with a wind-field template. Taking the Daxing Airport roof as a case study, a wind-direction-wind-speed joint probability distribution model was developed using the Gumbel copula. Since standalone wind-direction-wind-speed probability models are insufficient for fatigue damage assessment, we establish a wind-load spectrum template to complete the mapping relationship between wind speed and direction. Forty load cases were defined by varying wind speed and direction, with initial wind-load distribution templates derived from finite element analysis (FEA) results and field measurements. Subsequently, static analysis and variable-amplitude cyclic loading simulations were performed on a finite element model of the metal roof panel to obtain its S-N curve. Miner's linear damage accumulation model was then introduced to map cyclic wind-load spectra into equivalent constant-amplitude stress ranges, calculate cumulative damage degrees at each node, and establish damage variable-maintenance stage mapping relationships. Using the northwest roof of Daxing Airport as a validation case, grid nodes were divided according to their geometric dimensions. By integrating cyclic load spectra and S-N curves, damage accumulation was computed point-by-point to generate fatigue life distribution maps.