ABSTRACT
The efficient operation of Tunnel Boring Machines (TBMs) is essential for the successful completion of mechanized tunnelling projects within budget and on schedule. It is essential to increase productivity by optimizing the tunnel boring process and reducing downtime. There are a number of reasons for downtime in tunnelling projects, including delays in material supply and unexpected maintenance issues with the TBM or secondary processes. This study analyses TBM downtime by investigating fault codes based on duration and frequency, in order to identify dependencies between them. By analyzing TBM sensor data, the study explores correlations between sensor readings and failures with the goal of early detection of downtime using machine learning techniques. Anomaly detection shows a correlation between machine failures and outliers in sensor data, supporting the use of these algorithms for proactive maintenance. This approach demonstrates the potential of artificial intelligence in optimizing mechanized tunnelling projects.
