Search

Prediction is a vague concept that is why we need to conceptualize it specifically for underground deformation time-series data. For this impending issue, this paper employs an advanced deep learning model Bi-LSTM-AM to address it. The results show its applicability for practical engineering. The proposed model is compared with other basic deep learning models including long short-term memory (LSTM), Bi-LSTM, gated recurrent units (GRU), and temporal convolutional networks (TCN). These models cover the most common three forms of deep learning for time-series prediction: recurrent neural networks (RNN) and convolutional neural networks (CNN). This research is supposed to benefit the underground deformation time-series prediction.

Structure-from-Motion Multi-View Stereo (SfM-MVS) photogrammetry is a viable method to digitize underground spaces for inspection, documentation, or remote mapping. However, the conventional image acquisition process can be laborious and time-consuming. Previous studies confirmed that the acquisition time can be reduced when using a 360-degree camera to capture the images. This paper demonstrates a method for rapid photogrammetric reconstruction of tunnels using a 360-degree camera. The method is demonstrated in a field test executed in a tunnel section of the Underground Research Laboratory of Aalto University in Espoo, Finland. A 10 m-long tunnel section with exposed rock was photographed using the 360-degree camera from 27 locations and a 3D model was reconstructed using SfM-MVS photogrammetry. The resulting model was then compared with a reference laser scan and a more conventional digital single-lens reflex (DSLR) camera-based model. Image acquisition with a 360-degree camera was 3× faster than with a conventional DSLR camera and the workflow was easier and less prone to errors. The 360-degree camera-based model achieved a 0.0046 m distance accuracy error compared to the reference laser scan. In addition, the orientation of discontinuities was measured remotely from the 3D model and the digitally obtained values matched the manual compass measurements of the sub-vertical fracture sets, with an average error of 2–5°.

Rock discontinuities play an important role in the behavior of rock masses and have a high impact on their mechanical and hydrological properties, such as strength and permeability. The surfaces roughness and physical aperture of rock joints are vital characteristics in joint shear strength and fluid flow properties. This study presents a method to digitally measure the physical aperture of a rock fracture digitized using photogrammetry. A 50 cm × 50 cm rock sample of Kuru grey granite with a thoroughgoing fracture was digitized. The data was collected using a high-resolution digital camera and four low-cost cameras. The aperture and surface roughness were measured, and the influence of the camera type and 3D model rasterization on the measurement results was quantified. The results showed that low-cost cameras and smartphones can be used for generating 3D models for accurate measurement of physical aperture and roughness of rock fractures. However, the selection of appropriate rasterization grid interval plays a key role in accurate estimations. For measuring the physical aperture from the photogrammetric 3D models, reducing rasterization grid interval results in less scattered measurement results and a small rasterization grid interval of 0.1 mm is recommended. For roughness measurements, increasing the grid interval results in smaller measurement errors, and therefore a larger rasterization grid interval of 0.5 mm is recommended for high-resolution smartphones and 1 mm for other low-cost cameras.

Funding Information: This work was made possible thanks to the funding provided by the State Nuclear Waste Management Fund (VYR) and the support of the Finnish Ministry of Economic Affairs and Employment on the Finnish Research Programme on Nuclear Waste Management KYT2018 and KYT2022 of the Nuclear Energy Act (990/1987) in the research projects Mechanical Properties of Rock Joints (KARMO) and Fluid flow in fractured hard rock mass (RAKKA). Funding Information: Funding: This work was made possible thanks to the funding provided by the State Nuclear Waste Management Fund (VYR) and the support of the Finnish Ministry of Economic Affairs and Employ‐ ment on the Finnish Research Programme on Nuclear Waste Management KYT2018 and KYT2022 of the Nuclear Energy Act (990/1987) in the research projects Mechanical Properties of Rock Joints (KARMO) and Fluid flow in fractured hard rock mass (RAKKA). Publisher Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. An accurate understanding of jointed rock mass behavior is important in many applications ranging from deep geological disposal of nuclear waste, to deep mining, and to urban geoengineering projects. The roughness of rock fractures and the matching of the fracture surfaces are the key contributors to the shear strength of rock fractures. In this research, push shear tests with three normal stress levels of 3.6, 6.0, and 8.5 kPa were conducted on two granite samples with artificially induced well-matching tensile fractures with sizes of 500 mm x 250 mm and 1000 mm x 500 mm. The large sample reached on average a -60% weaker peak shear stress than the medium-sized sample, and a strong negative scale effect was observed in the peak shear strength. The roughness of the surfaces was measured using a profilometer and photogrammetry. The scale-corrected profilometer-based method (joint roughness coefficient, JRC) underestimates the peak friction angle for the medium-sized slabs by -27% for the medium sample and -9% for the large sample. The photogrammetry-based (Z'(2)) method produces an estimate with -7% (medium) and + 12% (large) errors. The photogrammetry-based Z'(2) is an objective method that consistently produces usable estimates for the JRC and peak friction angle.