Reconstruction of Metal Roofs Wind Pressure Using POD-LSF Algorithm with Sparse Data

Metal roof envelope systems are critical components of large-scale structures, relying on sensors mounted on their surfaces to reconstruct wind pressure for essential fatigue monitoring. Due to the extensive structural coverage and large surface area involved, sensor deployment is often sparse, necessitating effective methods for reconstructing wind pressure fields from limited data. This paper proposes a Proper Orthogonal Decomposition-Least Squares Fitting (POD-LSF) algorithm for wind pressure reconstruction based on discrete measurement data. The method first employs POD in an offline phase to reduce the dimensionality of the wind pressure field data, significantly decreasing computational demands and enhancing reconstruction speed. Subsequently in the online phase, the least squares fitting technique (LSF) is applied to integrate information from the discrete measurement points, thereby improving the accuracy and stability of the reconstructed wind pressure distribution. This approach enables rapid and precise reconstruction of the wind pressure field even under conditions of sparse sensor placement. The proposed algorithm has been validated on a typical sloped roof structure, achieving an average relative error of 8.95% in wind pressure reconstruction and effectively capturing the overall wind pressure distribution across the roof surface.