148 / 2024-03-04 04:09:24

A novel method for measuring rock bolt load through 3D laser scanning and deep learning

rock bolt load, measurement, laser scanning, deep learning

Estimating the load on rock bolts is crucial for ensuring the stability and safety of underground constructions. Traditional methods for measuring this load are often cumbersome and slow, mainly due to the need for strain gauge devices that must be installed on the rock bolts. This study introduces an innovative technique for measuring rock bolt load, leveraging the relationship between the bolt's axial load and the deformation of its bearing plate. By utilizing 3D laser scanning for data collection and a specially developed convolutional neural network (CNN) model for data analysis, this method offers a fast and non-destructive means of external monitoring. The efficacy of this approach was confirmed through 27 laboratory experiments, which demonstrated its ability to accurately estimate the axial forces on rock bolts, with an average error margin of ±5 kN. This highlights the technique's applicability in real-world underground engineering scenarios. The research advances intelligent system development in rock mechanics and engineering and bears significant implications for industries like mining, tunneling, and other forms of underground construction.