| The mine water inrush accident is easy to happen under the disturbance of coal mining.The roof water damage is one of the main water hazards in coal mining in China.So predicting and evaluating water content of sandstone roof of coal seam is of great practical significance for rational deployment of coal roadway and working face,and for realizing safe and efficient production of mining area.In no.81 mining area of Xinhu Coal Mine,Huaibei mining area,the main coal seam is 8 8 coal seam which over 700 m deep and covered with multi-layer medium-thick sandstone.The roof of the coal seam is sandstone with poor cementation and good permeability.The water-bearing sandstone is weak in strength,so it may become a water-rich layer and a good permeability layer.This paper introduces the theoretical basis of AVO forward and inversion and several theoretical models applicable to sandstone,and analyses the relationship between P-wave and S-wave velocities of sandstone with pressures by P-wave and S-wave velocity tests on roof sandstone samples.I used the test results and the dual time difference log to establish a rock physical model of the sandstone,and apply the Gassmann fluid replacement equation to forward calculate the p-and s-wave velocities and corresponding AVO properties of sandstone with different porosity and pore fluids to analyze the relationship between AVO properties and different porosity and pore fluid and search for AVO properties sensitive to water-rich sandstone.Based on the existing geological and logging data,the thesis analyze the characteristics of aquiclude and aquifers in no.8 coal roof of Xinhu Coal Mine.The logging curve was normalized and the porosity curve was reconstructed using gamma logging.After the geological drilling,logging calibration,synchronous inversion of prestack AVO were carried out to extract various seismic attributes such as density,Poisson’s ratio and velocity of P-wave and S-waves.In order to reduce the multiplicity of a single property,the thesis take apparent resistivity of sandstone as desired output,and use the PNN probabilistic neural network to coalesce attributes.The result of attributes fusion is compared with the actual drilling results,which confirms the reliability of research on aquifer prediction method of coal seam roof sandstone based on seismic and logging data,which provides reference for safe production of coal mine working face.There are 64 figures,7 tables and 99 references in this thesis. |
