Shinji Utsuki

Development of Measurement System of Seismic Wave Generated by the Excavation Blasting for Evaluating Geological Condition Around Tunnel Face

Mountain tunneling in Japan is commonly performed according to the new Austrian tunneling method (NATM), and preliminary investigation for understanding the geological conditions of the mountain are performed from the surface by borehole sampling and seismic exploration using the refraction method. Exploratory boring is performed in only select locations with specific terrain or geological features, whereas seismic exploration using the refraction method is generally performed over the area in which the tunnel will be excavated. The certainty of such seismic exploration is reduced according to depth, making it difficult to ascertain the details of the geological conditions. The results of preliminary investigations must therefore be constantly compared with the actual geological conditions as revealed during construction. Seismic velocity is a significant factor to determine the design of tunnel support. Evaluation of geological condition requires the seismic velocity a round tunnel face, so there is a need for seismic exploration techniques that do not affect the excavation work cycle of tunnels under construction. We have therefore developed the tunnel face tester (TFT), which is a seismic exploration system that uses excavation blasting as wave source. This paper describes the developed system, and the results of verification experiments performed at a tunnel construction site.

Development of a System for Automatic Evaluation of the Geological Conditions of Tunnel Faces Using Artificial Intelligence and Application to a Construction Site

When constructing rock structures such as dams and tunnels, it is important to first ascertain in detail the geological conditions at the planned site during the preliminary survey stage, and to implement the optimal design according to these conditions. During construction, after evaluating the geological conditions found at the drilling surface and tunnel face, it will sometimes be necessary to sequentially revise the design and construction plans. In this respect, various studies using artificial intelligence (AI) have been carried out in recent years, and substantial advancement and labor-saving has been achieved in various evaluations, especially in industries outside of construction. This paper presents the details of investigations for constructing a system to automatically evaluate the geological conditions of tunnel faces by the convolutional neural network method, which is an AI image recognition technology that has been put to practical use in industries outside of construction, using face observation records and elastic wave velocities measured at the corresponding points as training data.