Evaluation of rock muck using image analysis and its application in the TBM tunneling

Abstract

The rock muck is the direct product of a rock-cutter interaction in the tunnel boring machine (TBM) tunnelling. It indicates both the TBM performance and rock mass quality. To investigate the relationships between the rock muck and the TBM operational parameters/rock mass parameters, the most important work is how to characterize the morphological properties of rock blocks reliably. Traditionally, only the block size is characterized using the sieving method, which is time consuming and not practical in the TBM tunnelling. Image analysis can serve as a quick and convenient 2D approach for analyzing particle size and shape. In the present study, three projection methods were used to characterize the morphological properties of rock blocks, including a randomly single projection, the maximum-area and minimum-area projections (selected from 1000 projections) and three orthogonal projections. These morphological properties were obtained using a computer algorithm written in the MATLAB and the relationships between the 3D and 2D parameters were tried to develop. These developed correlations were applied to a TBM-excavated tunnel project to investigate the potential relationships between the rock muck and the TBM operational parameters/rock mass quality. The results show that 3D morphological properties of rock blocks can be estimated using the corresponding 2D morphological properties measured on a randomly single projection despite this two-dimensional method strictly failing to describe all particle attributes. The proposed method was found to be useful for estimating the 3D size and shape of rock blocks. The rock blocks trended to be larger in size and more elongated in shape when the TBM worked at a low FPI or the rock mass was fractured, vice versa.