There are several well-established empirical and mechanistic methods used to predict mine subsidence due to longwall mining. These methods can provide reliable predictions of vertical subsidence, tilt and the overall/macro curvature when properly calibrated to the local conditions.The prediction of horizontal movement and strain at the surface are much more difficult, as these parameters are sensitive to variations in the surface topography and the presence of near-surface geological structures, which can result in localised and elevated movements. Many existing predictive methods do not consider the potential for these irregular movements or are not capable of predicting such behaviour.Existing predictive methods often only provide a single predicted value of strain that represents the regular or general levels of strain. The irregular strains due to the effects of surface topography and near-surface geology can exceed the regular movements by factors of two to three times.Strain is also one of the most important parameters for assessing the potential for impacts on natural and built features on the surface. It is recognised in the industry that improved methods for the prediction of horizontal movement and strain are required to better assess the potential impacts on surface features.Predictive methods for strain have been developed through this research that consider the effects of the mining geometry, surface topography, near-surface lithology and the potential for irregular anomalous movement. These methods provide the range of potential strains, rather than a single predicted value, based on both the regular and irregular movements. The predicted distribution of strain provides better guidance on the magnitudes of the localised and elevated movements, which are often not considered in other existing methods.A large database of ground monitoring data from the Australian coalfields was available for this research. The methods have therefore been derived using an empirical approach supplemented with numerical modelling.Strain is predicted using a two-step process. Firstly, the net horizontal movements of the surface are predicted across each of the hogging and sagging curvature zones above the active longwall. The distributions of strain are then predicted within each of these zones based on these net horizontal movements. A predictive method has also been developed for shear deformation based on the parameter horizontal mid-ordinate deviation.The predictive equations for net horizontal movement, bay length difference and strain for the Southern Coalfield are provided in the appendices of this thesis. These methods can be applied to coalfields elsewhere in Australia and overseas by calibrating the coefficients for the local conditions using the available ground monitoring data.The proposed methods are an improvement on many existing empirical and mechanistic methods, as they consider the effects of surface topography, near-surface lithology and the potential for irregular anomalous movement. The mining-induced impacts on the natural and built features at the surface are often governed by the localised and elevated strains that develop from these effects.
展开▼
