基于风谱特性和频带分离的大跨度空间结构风压数据压缩感知方法OA

Long-span spatial structures exhibit pronounced sensitivity to wind effects,making continuous wind load monitoring essential for structural safety.However,sustained high-frequency data acquisition considerably increases the energy demand of wireless sensing systems.To address this challenge,this study proposed a compressed sensing method for wind pressure measurements in long-span spatial structures,leveraging wind spectrum characteristics and a frequency-band separation framework.The method constructed a wind pressure decomposition model through band-wise separation and employs a spectrum-informed,customized dictionary to compactly represent the dominant energy components of the signals.The proposed approach was validated using measured wind pressure data from the Beijing Daxing International Airport terminal,and benchmarked against LASSO and Bayesian compressed sensing techniques.The results demonstrate that the proposed decomposition model effectively captures mean wind pressure as well as the low-and high-frequency components of fluctuating wind pressure.Even under high compression ratios(the ratio of original signal length to measured signal length),the proposed method significantly improves reconstruction accuracy while preserving the spectral characteristics of the original signals,demonstrating superior performance compared to the benchmark methods.