A fuzzy c-means algorithm guided by attribute correlations and its application in the big data analysis of tunnel boring machine

Abstract

Tunnel boring machine (TBM) is a complex engineering system used for tunnel construction, and its design is mainly based on knowledge from previous projects. With the development of measurement techniques, massive operation data have been recorded and partitioning these data can provide useful references to the designers and help in the design of TBM. In this paper, a fuzzy c-means algorithm guided by attribute correlations, named attribute correlation-guided fuzzy c-means algorithm (ACFCM), is proposed to accomplish this work. The proposed algorithm is based on fuzzy c-means algorithm (FCM) and involves a new objective function in which the attribute correlation is described by the linear model. A synthetic dataset is used to evaluate the performance of the ACFCM algorithm, which demonstrates its higher effectiveness and advantages compared with conventional FCM. The ACFCM algorithm is applied to cluster the TBM operation data from a tunnel in China, and the load and penetration rate of the TBM are predicted based on the clustering results. The results indicate that the ACFCM algorithm can not only provide competitive clustering results but also significantly increase prediction accuracy. This work also addresses the applicability and potential of data clustering in the design and analysis of other complex engineering systems similar to TBMs.