Toward the automation of mechanized tunneling “exploring the use of big data analytics for ground forecast in TBM tunnels”

Automation of construction machines has grown rapidly in recent years as a response to the need to increase productivity, increase construction safety, decrease costs, and overcome the lack of availability of qualified labor. However, tunnel automation still lags. Automation of mechanized tunneling is essential to the future of tunneling construction, as the current practice still relies heavily on the experience and the human judgment of the machine operator to steer the TBM, which could lead to undesirable events. The primary motivation for this review paper stems from the statement: the success of tunneling automation relies on precise ground prediction. Even small inaccuracies can have significant implications, and the current reliance on human experience and judgment presents limitations and risks. With the abundance of machine data now available from TBMs, and the advancements in data analytics and machine learning (ML), many models have been proposed. These models have the potential to revolutionize tunneling by providing better decision-making support, including geology forecasts and anomaly detection. However, despite the numerous research studies and advantages of these models, they are not widely implemented in real-world scenarios. Thus, this review paper aims to address this issue by focusing on ground prediction models for TBM tunnels, providing a comprehensive overview of the current state-of-the-art, illuminating the existing practices, and highlighting the limitations that hinder these models from being the catalysts of tunneling automation. By emphasizing these challenges, the paper seeks to not just critique but also guide, providing recommendations for future research that promise to bridge the gaps and potentially usher in an era of fully automated TBMs.