Building Rigorous Pillar Datasets Using Visual Observations and Ground Penetrating Radar to Calibrate Pillar Failure Models

Abstract Empirical and numerical studies of pillar strength rely on the availability of high-quality pillar databases. These databases often consist mostly of visual observations and subjective assessments of pillar failure. Typically, no objective measure is available to verify that the pillars had indeed failed. This study explored the use of ground-penetrating radar to collect data on pillar condition. Some agreement was obtained between the radar data and borehole camera observations for pillars with minor spalling. The ground-penetrating radar equipment is, however, difficult to use if the pillars are subjected to extensive spalling, and care should be exercised when interpreting the data. As a secondary objective, the data collected were used to determine the strength of the hard rock UG2 (Upper Group 2 chromitite) pillars in the Bushveld Complex of South Africa. A database of 78 intact and failed pillars was compiled. The data indicated that a power-law strength equation with exponents similar to those of the Hedley and Grant equation and a $$K$$ K value of 95 MPa provides a good approximation of pillar strength. For this mining area, an improved calibration of a limit equilibrium model in a boundary element code was also possible. A key finding of the study is that even with advanced technology, such as ground-penetrating radar, it is difficult to determine whether pillars retain an intact core when extensive spalling at the pillar edges is present. This unknown factor affects the quality of the pillar databases and the accuracy of the derived equations.