Introduction
The majority of Appalachian coal deposits lie in a multi-seam environment. Historically, the decision to extract a particular seam has been based on ease of access, ownership and economics rather than ground control considerations. As mining progresses and seams are being mined out, it is increasingly common to find working and abandoned mines in close vertical proximity to each other. Such contiguous operations can produce both positive and negative ground control situations in neighboring seams. These ground control problems, defined as interaction problems, are frequently reported and their severity can range from minor roof problems that may require no additional support to complete loss of a large section of coal reserves (Stempie, 1955; Peng & Chandra, 1980; Haycocks & Karmis, 1983; Suet ale 1984). As a consequence, interaction problems are of growing concern to the Appalachian underground coal industry. Although interaction problems in multi-seam mining have been recognized for many years and many investigations have been conducted to study the problems, the ground control principles involved in their development are only partially understood. Previous research has identified four interaction mechanisms: pillar load transfer, innerburden shearing, archirig effects and upper seam subsidence (Haycocks et al, 1982; Haycocks & Karmis, 1983). Pillar load transfer mechanisms and innerburden shearing have been successfully used to explain interaction phenomena associated with undermining operations, while archirig and subsidence have been used to assess problems during overmining. In addition, interaction-controlling factors such as cover, innerburden and seam thicknesses, extraction percentages on both seams, time delay between operations on two seams, and characteristics of innerburden rock have been identified and utilized in evaluation of interaction effects (Haycocks & Karmis, 1983; Webster et al,1984). As an example, Figure 1 shows the relation of innerburden layering and thickness to interaction on the lower seam. The combination of such theoretical work and the analysis of case study information has made it possible to determine, with a certain degree of confidence, the interaction damage in undermining conditions. The results of such an approach have been sufficiently successful and complete to be incorporated into a software package to assist field engineers in estimating potential interaction effects when undermining a previously mined upper seam (Grenoble et ale 1984; Grenoble & Haycocks, 1985; Grenoble et ale 1985).
Figure 1. Number of Innerbeds vs. Innerburden Thickness for Stable/Unstable Lower Seam Conditions. (after Haycocks & Karmis, 1982) (available in full paper)
Although the conventional finite element and photoelastic modeling techniques have been employed in conjunction with field case studies to study the interaction problems under a variety of conditions, the results and inferences from such studies are not always exhaustive (Peng & Chandra, 1980; Haycocks & Karmis, 1983; Suet ale 1984). Results from these investigations have proven successful in predicting trends or identifying possible problem locations but do not utilize all the case study data available and may have omitted some critical variables that affect interaction. Statistical analysis can overcome some of the disadvantages and utilize the numerous case studies available. Previously, the statistical analysis method has been used with only moderate success due to limited case study data available for the complexity of the interaction problems.
