Subsidence Profile Functions Derived From Mechanistic Rock Mass Models - A tentative assessment of practical applicability

Daemen, Jack J. K.
Organization: Society for Mining, Metallurgy & Exploration
Pages: 18
Publication Date: Jan 1, 1982
A wide variety of subsidence prediction methods exist. They can be classified into empirical methods and mechanistic models. The empirical methods include profile functions and influence functions (Brzuner, 1973a,b; Kratzsch, 1974; Berry, 1978; Hood et al., 1981), the two empirical approaches of most practical value, if only be- cause they are readily programmed on calculators or computers. The profile functions have the advantage that they allow for ready determination of the necessary empirical parameters from field observations. They have the disadvantage of being restricted to subsidence calculations at a very limited number of points or lines above very simple (rectangular) mine geometries. This reduces greatly their practical applicability for damage predictions, which usually require estimates of strains, slopes or curvatures at structures that have arbitrary position and orientation with respect to planned mine excavations. Influence functions can be used to calculate subsidence at any point near an excavation of any in- seam outline. The main difficulty associated with empirical influence functions is the lack of uniqueness in the relation between subsidence and its cause, the convergence distribution over a mined excavation and its immediate surroundings. A secondary practical difficulty is the need for fairly arbitrary adjustments near the excavation edges, usually in the form of compensation zones (e.g. Brsuner, 1973; Hood et al., 1981). Determination of the appropriate empirical constants in influence functions is facilitated greatly when a direct mathematical connection exists between influence and profile function (several examples are given by Brauner, 1973 and by Hood et al., 1981), so that field data can be used to determine the parameters through backfitting of the profile functions, which is considerably simpler and more direct than backfitting of influence functions.
Full Article Download:
(7797 kb)