Superfamilies of networks for analyzing the correlations of different flow fields

Uncovering the correlations of different flow fields, in terms of time series network analysis, is important and relevant to both engineers and businesses. We first systematically carry out experiments and record different groups of time series signals from the wind field and the gas concentration field simultaneously. Then, we employ the visibility graph method for mapping the experimental measurements into complex networks and calculate network motif frequency. Next, subgraph ratio profiles (SRPs) for different flow fields are exploited based on the motif analysis. Finally, the hierarchical cluster method is adopted to identify a more accurate number of clusters in the SRPs, and then assign each network to one superfamily. The results suggest that the time series of gas concentration, wind speed and wind direction belong to different superfamilies. Moreover, we find that compared with wind direction signals, the correlations between gas concentration signals and wind speed records are stronger. These interesting and significant findings suggest that the motif of a complex network can be a useful tool for revealing the correlations of different flow fields.