Recently there has been an upsurge of technical papers discussing the specification of roof support in tailgates in both Australia and the United States. Examples include:
* ALTS – tailgate serviceability in Australia is empirically related to pillar size and the nature of the immediate roof.
* SPOT – an American software program that designs roof support in tailgates and suggests that support density can be decreased if some deformation is allowed.
* Cable support should be designed based on the suspension of the dead weight of a detached block.
* Tailgate roofs fail as a result of increasing horizontal stresses, and support density can be designed using the concepts of buckling Euler beams.
* Cable trusses in tailgates perform well.
Some of these concepts are contradictory, and certainly cannot all be used simultaneously. The purpose of this article is to present one possible interpretation of the conditions that develop in tailgates in the hope that it generates a more critical examination of this important issue. The article builds on the formative work done by Prof Kaiser and his colleagues in the underground metalliferous mines of Canada.
In particular it develops the idea that roof instability can develop in zones of stress reductions, both as a result of the reduction in the shear resistance of vertical joints and from the loss of bolt/cable anchorage.
Stresses in tailgates
Before the roof is exposed to tailgate stress conditions, it has experienced a range of stress conditions. On development there was a reduction of vertical stress in the immediate roof to zero and an increase in horizontal stress of approximately 20%. Bolts and cables are installed after most of these stress changes develop. There was relatively small increase in the vertical stress above the pillars.
At the maingate corner, the horizontal stresses in the roof increase by 20% to 120% depending on the orientation of the roadways with respect to the major principal horizontal stress. The vertical stresses on the pillar increase by about 100% and are distributed in a bulb shape with the sides of the bulb contained within a 25o angle.
Due to the elastic Poisson’s ratio effect, this vertical stress increase is associated with an increase in horizontal stress. It is important to note that this effect develops above the pillar and not within the roadway roof. At the rib-line the increase in horizontal stress due to the Poisson’s ratio effect is in the order of about 3% of the vertical stress increase. This then quickly drops to zero a small distance from the rib.
Behind the face line, the travel road experiences a major reduction in horizontal stress as a result of shielding from the newly formed goaf. By analogy with stress relief development headings, it is estimated that these stress reductions are in excess of 50% of the insitu values. The vertical stress on the chain pillar increases to approximately three times the insitu value. These stress conditions remain in the return roadway ahead of the face on the tailgate side.
The next change in stress develops at the tailgate/face corner. Here the vertical stresses increase to values of four to five times the in-situ stresses.
Drawing from the Canadian work, the other mechanism acting relates to the behaviour of the chain pillar. If the pillar compresses significantly, then the induced rotation of the tailgate roof will result in a reduction of the horizontal stress. The level of stress reduction is a function of the roof rock modulus, with a greater stress reduction occurring for higher modulus (stronger) rocks.
To achieve a significant reduction in the horizontal stresses, the pillar compression needs to be in the order of 150 mm or more. This is only possible if the pillar yields.
A simple longwall layout can be used to quantify the stress changes. The longwall is at 250m depth, with a 210m wide extracted void, roadways are aligned parallel to the major principal horizontal stress in a seam 2.5m thick.
The recently developed ALTS methodology is used to estimate the vertical stresses acting around the longwall panel. Two chain pillar sizes are used – 22m and 25m, with the 22m pillar specified to yield at the tailgate corner. The chart shows that the horizontal stresses increase at the maingate to about 14MPa and then decrease. In the tailgate the horizontal stresses in the immediate roof are about 6MPa for the 25m wide pillar, and –3MPa (tension) for the 22m pillar.
Once it is recognised that there are stress reductions in tailgates, then there is a need to critically review what appears to be an unstated assumption by roof support designers – that roof falls indicate high horizontal stresses. If there are stress reductions, then roof instability cannot be related to the onset of high horizontal stresses. The use of numerical methods that presume high horizontal stresses and the use of Euler concepts need to be reassessed for tailgates.
Rock masses can fail in a zone of low stresses because their jointing and bedding means that they have no tensile strength. It is suggested that tailgate failures are related to the onset of tensile stress conditions interacting with vertical jointing in the immediate roof.
The relationship between CMRR, tailgate stability, and pillar design can be explained by the general trend for thinner bedded materials (low CMRR) to have closer spaced jointing. Such rocks are particularly susceptible to failure along vertical joints especially when the roadways are aligned parallel to the joints.
For rocks with a higher CMRR, the joint spacing may be such that the face line and the yielding pillar can support individual joint-bounded blocks. This will be particularly the case if the roadways are not aligned parallel to one of the joint sets. There may be situations where a cantilever may form such that localised compressive stresses are developed on the underside of the joint block.
In zones of tensile stresses, suspension of detached blocks should be the design approach. The challenge is then to identify the height of the potential fall. Recalling the comments on induced horizontal stresses above the pillar, elastic theory would suggest that there is some horizontal confinement across vertical joints within an approximately 25o angle from the pillar rib. This means that the height of the unconfined zone is approximately equal to the roadway width.
An adequately conservative design would be obtained if the detached block is presumed to have vertical sides. An important complicating factor is the presence of jointing, such that the stress bulb may be vertically sided. In this case, the height of the detached block should be derived from knowledge of the location of possible thick layers with wide joint spacings.