Exploring spatio-temporal correlation and complexity of safety monitoring data by complex networks

Gao Yuyue, Li Rao, Zhou Cheng, Jiang Shuangnan

Abstract

With the development of monitoring technologies, collected data has become more massive, precise, and timely, which can be an excellent foundation for more comprehensive assessment. However, existing data-analyzing researches still concentrate mainly on spatial or temporal characteristics separately, neglecting the dependence inside. Actually, the spatio-temporal correlation has been widely studied in other areas. Based on that, this paper proposed an undirected and unweighted data-based complex network as a risk assessment tool to explore the spatio-temporal correlation in safety monitoring data. Eigenvectors containing both spatial and temporal characteristic values are the nodes and the degree of correlation determines whether edges exist between each pair of nodes. The good application of the model in a metro construction project verifies the existence and significance of the correlation. This work not only reveals the spatio-temporal correlation of construction characteristics but also provides a new perspective in safety assessment.

Keywords: Complex network, Spatio-temporal correlation, Settlement time series, Metro construction, Wavelet transform, Topological analysis

https://www.sciencedirect.com/science/article/pii/S0926580521005665?via=ihub