ABSTRACT: Yielding pillars are gaining increasing acceptance as a technique for improving ground control in coal mines. The field studies described in this paper provide insights into the failure mechanics of yielding pillars that may prove helpful in improving yielding pillar design. Analysis includes the derivation of expressions for the stress distribution within a yielding pillar implied by several well-known empirical pillar design formulas. The field studies were conducted in three narrow longwall pillars at two mines. The pillars were instrumented with vibrating wire stress meters, and, in one case, a sonic extensometer. As the longwall faces passed the pillars, measurements were made of vertical stress changes, changes in confining stress, and horizontal pillar deformation. The measurements, in particular the vertical stress profile measured in one pillar as it yielded, are compared with the predictions of several pillar strength models. The models include the analytical one developed by Wilson as well as those derived from the empirical pillar strength formulas. The paper concludes that the stress gradients derived from the empirical formulas provide good fits to the measurements, and that the analytical models used for yield pillar design would benefit from more reliable determinations of in situ material properties.
1 INTRODUCTION
Recent years have seen a renewed interest in the use of yielding coal pillars, particularly in connection with longwall mining. Yield pillars have been proposed as a means of improving longwall tailgate stability (Martin et al., 1985), eliminating pillar bumps (Harvey and King, 1985), controlling floor heave (Carr et al., 1985), and reducing stress-related roof falls (Serata et al., 1986). Mine design with yielding pillars can also increase overall extraction and reduce development time. Because longwall mining is being conducted at increasingly greater depth, the use of yielding pillars can be expected to increase, making the design of yielding pillars an important area of rock mechanics research. Yielding pillars present a unique design problem because failure is to be expected rather than avoided. Unfortunately, relatively little field testing has been performed to characterize the failure mechanics of yielding pillars. The goal of this paper is to evaluate several current pillar design methods using field measurements obtained from yielding coal pillars.
2 PILLAR DESIGN METHODS
Two approaches have been used for coal pillar design in the past. One traditional method has been to use empirical formulas, which were derived through curve-fitting to compressive strength data from coal specimens of various shapes and sizes. Empirical formulas typically incorporate a "size effect" through the concept of the "in situ coal strength," and a "shape effect" through a "width-to-height ratio" (Bieniawski, 1984). Examples of empirical pillar strength formulas are the Obert-Duvall/Wang formula (1), the Holland-Gaddy/Hustrulid- Swanson formula (2) and the Bieniawski formula (3). (mathematical equation)(available in full paper)
Empirical pillar strength formulas have one important advantage in that they have been validated through large-scale tests and many years of actual use. Their disadvantage is that they are not based on any kind of stress analysis or mechanical model of pillar behavior, so it can be difficult to apply them to new problems such as yielding pillar design.
