Spatiotemporal correlation analysis of the dynamic response of supertall buildings under ambient wind conditions

The structural health monitoring systems of supertall buildings are typical big data systems with abundant and high-dimensional data. Correlation analysis is a vital technique for data mining and analysis. This study aims to reveal the spatial and temporal correlation of 4 years of structural health monitoring data from the 632-m-high Shanghai Tower under normal wind and typhoon conditions. In this paper, the root mean square (RMS) value of the acceleration response is used to characterize the structural vibration intensity. The maximal information coefficient (MIC) and Pearson product-moment correlation coefficient (PPMCC) are adopted to measure the spatial correlations along Shanghai Tower under different wind conditions. The temporal correlation of structural vibration is investigated with kernel density estimation and the density-based spatial clustering of applications with noise (DBSCAN) algorithm. The results show that the structural vibrations of Shanghai Tower in different spatial positions are correlated. The correlation of the structural vibrations of the same floor is at a high level under both normal wind and typhoons, while that of different floors rises with increasing mean wind speed. Furthermore, the MIC is more stable than the PPMCC in measuring this spatial correlation. Based on temporal correlation analysis, a circadian rhythm tendency of the structural vibration intensity of Shanghai Tower on the daily scale is identified, and it is verified that the structural response is temporally correlated. These findings can provide a structural health state estimation index based on the correlation coefficient and provide a basis for the screening of modeling data.