Investigation into the mechanism of strength and failure in squat coal pillars in South Africa
Date
2016-03-17
Authors
Mathey, Markus
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The mechanism of strength and failure in squat coal pillars with large width-to-height ratio has piqued the interest of researchers and mining operators for many years. In South Africa, the view has been adopted that pillars with width-to-height ratios greater than 5 have the ability to increase their peak load-bearing capacity exponentially and therefore obey a distinctly different law than the more slender pillars.
This view has not always been shared in the international mining community, for good reasons: a literature review of evidence and indications on squat pillar behaviour is not conclusive. A competing philosophy on strength and failure in squat pillars is that such structures did not exhibit a peak strength, but rather performed in a ductile or strain-hardening manner. If this holds true, they were able to increase their load-bearing ca-pacity beyond the point of yield with increasing deformation.
Hence, the research presented in this thesis is designed to investigate the mechanism of strength and failure in squat coal pillars in greater detail.
A statistical survey of the mechanical properties of coal in South Africa was conducted with the intention to evaluate coalfield-, seam-, or site-specific patterns, which may de-mand individual treatment in the further analysis of pillar performance. Such differences are, however, not discernible from a laboratory specimen point-of-view. The average uniaxial compressive strength of coal seams is practically constant in South Africa and there are indications that this is also the case for individual mining sites. Only the triaxial compressive strength of coal specimens from different seams shows individual trends, but further statistical evidence is required to substantiate these. Nevertheless the survey yields important insights into the adverse influence of natural discontinuities on the strength of coal, which is observed to diminish in a sufficiently large triaxial confining environment. It is postulated that this confining environment can be generated in pillars with large width-to-height ratios, which ultimately means that the presence of natural discontinuities does not have a major influence on the strength of squat pillars.
A comparative laboratory study into the relationship between strength, failure, and the width-to-height ratios of model pillars of different materials reveals the following insight:
(1) The residual strength of model pillars after crushing increases progressively with in-creasing width-to-height ratio.
(2) The relationship between the peak pillar strength and the width-to-height ratio is bi-linear, whereas the second branch of strength increase is steeper than the first one. This behaviour is termed the squat effect.
(3) At sufficiently large width-to-height ratios, the peak strength trend ends in a brittle-ductile transition.
Bi-linearity is an entirely newly observed phenomenon and has been identified for hard and soft rock materials, as well as for a soft composite material, but notably not for coal.
The described characteristics (1 – 3) are qualitatively reproduced in a numerical model-ling investigation, which yields the following insights as to the mechanism behind the trends:
(Ad 1) The rate of progressive strength increase in the residual domain is controlled by the residual material properties. The higher the residual material properties, the more rapid the strength increase, and the lower the critical width-to-height ratio for brittle-duc-tile transition in pillars.
(Ad 2) Bi-linearity in the peak strength trend is caused by a change of failure mechanism in pillars from a critical width-to-height ratio onwards.
In the first and slower ascending branch of the bi-linear relationship, typically up to a width-to-height ratio of 5, the pillar’s maximum load-bearing capacity is reached when the first shear band develops in the pillar sidewall. The pillar cannot recover from this initial failure, and fractures soon propagate further into the core while the load-bearing capacity gradually decreases with further deformation.
In the second and faster ascending branch of the bi-linear relationship, pillars do re-cover from the first shear band in their extremities. The pillars then regain and increase their load-bearing capacity until core failure occurs.
The critical width-to-height ratio at which the squat effect occurs is dependent on the residual material properties. The higher the residual properties, the lower the critical width-to-height ratio. Also, the rate of strength increase in the second branch of bi-line-arity depends significantly on the competence of the residual material properties.
(Ad 3) Brittle-ductile transition appears to be only a practical concept for the interpretation of the stress-strain behaviour of pillars. Every model pillar is seen to ultimately show a certain amount of strength drop at the point when its core fails, even if 50 % strain or more is required at very large width-to-height ratios. However, because such large strains are unlikely to occur for coal pillars in-situ, the concept of brittle ductile transition may remain valid for practical purposes.
The new insight into the outstanding importance of residual material properties for the performance of squat pillars stimulates the demand for further research, in particular for rock pillars, for which the characteristics (1 – 3) have been observed in physical tests.
Coal model pillars, however, evidently behave different from rock: All physical tests in the laboratory demonstrate unambiguously that the strength versus width-to-height rela-tionship in coal follows one single, continuous trend from w/h= 1 up to 9 or 11. A squat effect is not discernible in this range for coal. This suggests that, at the current state of knowledge, coal pillar strength in South African mines is most suitably addressed by extrapolating established empirical strength equations into the squat range to at least w/h= 10. The extrapolation is further validated by experience made with squat coal pillars in mines in the United States.
Description
A thesis submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy
Johannesburg, 2015