Data based complex network modeling and analysis of shield tunneling performance in metro construction

Shield tunneling performance depends mainly on changes in geological conditions and machine working status. Understanding its characteristics is the key to operating and controlling shield machine during the metro construction. Despite the large set of shield tunneling data in having been a big challenge in interpreting the underlying meaning, a systematical view of the shield tunneling performance has not yet been identified. In this study, a methodology for the modeling and analysis of shield tunneling performance network is proposed which aims at integrating the high dimensional data mining and the complex network approaches for shield performance evaluation. It is tested by analyzing the heterogeneous data of shield tunneling performance acquired from in the first Yangtze river crossing metro tunnel project in China. Each segment ring tunneling cycle in the construction were considered to be nodes of the network mapped while edges are determined by nodes having the similarity greater than an optimal threshold value. The construct network exhibits high clustering coefficient combined with comparatively short path lengths, which demonstrates a small world topology feature. Communities in the performance network with different size based on the complex network are detected, which provides the vital decision information for geological conditions identification and shield tunneling performance risk evaluation.