Analysis of the failure mechanism of inclined coal pillars is one of the complicated issues. The wide variability of dip angles of inclined coal pillars makes it more complex. The asymmetric stress distribution and the tendency of shearing along the bedding planes make the inclined coal pillars to behave differently from the flat coal pillars. There is a need for in-depth investigation of the failure mechanism for addressing the instability problems of the inclined coal pillars. Most of the literature quantifies only the magnitudes of the mean principal stresses by classical statistics. As the stress is a second-order tensor having six independent components, the classical statistics is not appropriate to calculate the mean and variability of the principal stresses at the onset of failure of the pillars. In this paper, a comprehensive analysis is done to understand the complex failure mechanism of the inclined coal pillar using numerical modelling as well as tensorial statistics and validated the results with field measurement data of failure cases. The failure mechanism is analysed by quantification of the characteristics of the inclined coal pillars by the principal-stress magnitude and its orientation, induced at the time of failure. Since the spatial variability of the magnitudes and orientations of the induced principal stresses exist within the inclined coal pillars, the mean induced principal stresses are used to quantify the stress states within it. The failure stress states within the coal pillars having different dip angles are generated by the calibrated elasto-plastic numerical modelling with the ubiquitous joint model. Several statistical parameters are calculated to quantify the stress-tensor variability and the correlation among the stress-tensor components. It is found that the correlation coefficients among the shear components increase significantly with the increase of the coal pillar dip angle. Therefore, the inclined coal pillars are highly susceptible to shear failure. The magnitudes, as well as orientations of the mean induced principal stresses within the coal pillars obtained through numerical modelling, are quantified by the tensorial as well as classical statistics. It is found that the magnitude of the mean major induced principal stress ((sigma) over bar (1), (sigma) over bar (2), and (sigma) over bar (3)) are used to quantify the stress states within it. The failure stress states within the coal pillars having different dip angles are generated by the calibrated elasto-plastic numerical modelling with the ubiquitous joint model. Several statistical parameters are calculated to quantify the stress-tensor variability and the correlation among the stress- tensor components. It is found that the correlation coefficients among the shear components increase significantly with the increase of the coal pillar dip angle. Therefore, the inclined coal pillars are highly susceptible to shear failure. The magnitudes, as well as orientations of the mean induced principal stresses within the coal pillars obtained through numerical modelling, are quantified by the tensorial as well as classical statistics. It is found that the magnitude of the mean major induced principal stress ((sigma) over bar (1)) at the time of failure, i.e. the strength of the pillar decreases with the increase of the dip angles. The validation of the results with the actual stress measurement data shows that all the failed pillar cases are correctly predicted by the tensorial statistical approach whereas the classical statistical approach does not effectively predict the actual failed condition of the pillars. The study would help to characterise the behaviour of the inclined pillars and address the instability issues for safe and efficient mining of inclined coal seams.
