This paper summarizes the results of a subsidence monitoring program performed by the authors under a grant from Bureau of Mines, U.S. Department of Interior (Contract No. J0328010) and provides a comparative analysis of the subsidence data collected with three popular subsidence prediction models which have been used in the region. The subsidence monitoring program was conducted over a room-and- pillar panel at Westmoreland Coal Company's Ferrell Mine located in south-central West Virginia, approximately 73 kilometers (45 miles) south of Charleston, WV, the State Capitol. The Cedar Grove Coal seam ranging in height from 1.4 to 1.6 meters (55 to 62 inches) was being mined. The thickness of overburden strata in the study area ranged from 245 to 303 meters (810 to 1000 feet) of which more than 40% consists of sandstone layers 1.5 meters (5 feet) thick or greater. Surface subsidence was monitored by installing 43 surface monuments in four (4) rows over the No. 3 panel (Fig. l-A). The distance between the monuments in the critical areas was 15 meters (50 feet), whereas, all of the remaining monuments were spaced at 30 meters (100 foot) intervals.
2. MEASURED SUBSIDENCE
The transverse subsidence profile over the No. 2 and No. 3 panels along the centermost monument row is shown in Figure 3 for a survey taken one month after the mining was completed. The maximum subsidence measured was 0.45 meters (1.48 feet) which gives a subsidence factor of 0.30. It was located on the surface over the center of the barrier pillar which separated the No. 2 and No. 3 panels (see Fig. l-B). The maximum angle of draw observed for several subsidence profiles in this study is 24 degrees. The subsidence factor is relatively low and is believed to be so because of two reasons. First, during pillar mining, the percentage of extraction was only about 75 percent, and as a result, the gob area was partially filled with small coal stumps. This unmined coal provided part of the filling material instead of the caved immediate roof rock. Thus, a lesser caving height resulted and the maximum subsidence on the surface was reduced accordingly. Second, the amount of stronger rock (sandstone) layers greater than 1.5 meters (5 feet) in thickness) comprising the overburden is approximately 40 percent. These stronger rock layers underwent a smaller downward deflection after pillar mining which also contributed significantly to the lower subsidence readings at the surface.
3. SUBSIDENCE PREDICTION METHODS
Currently there are more than 30 subsidence prediction models available (Hall, 1980). In general, these prediction models rely on mathematical models of deformation mechanisms in the overburden or on empirical methods. A complete discussion of both approaches can be found elsewhere (Brauner, 1973, Hall 1980). This paper outlines the popular prediction methods available for the Appalachian Coalfields and compares the measured subsidence at the Ferrell Mine with the predicted subsidence from two empirical methods and one finite element method. The two empirical methods are the National Coal Board (NCB 1975) and the Hyperbolic Profile Function Methods (Kohli 1984). The third method is a Finite Element Method (Kohli 1984) which calculates with the aid of the computer.