Development of a roof-logging system by rockbolt drilling

Mechanical data from a rockbolt drilling machine, such as torque, thrust, revolution and stroke, are useful for understanding the geological and geomechanical conditions around roadways and hence selecting the optimum bolt type and density. Laboratory experiments showed that the level of torque or thrust reflects differences in rock type, with separation of strata and cracks appearing as a specific pattern. Algorithms of neural networks based on adaptive resonance theory and back-propagation were used to extract the pattern and estimate the location of discontinuities. For field testing in an underground coal mine, a prototype data acquisition system was attached to a portable drilling machine. Several sets of logging data from one locality were combined to construct a two-dimensional model of the strata, from which the slope angles of discontinuity faces could be estimated