| The roof of coal seam is usually dominated by sandstone aquifer with large porosity and developed fissure,which is one of the main water sources of mine water inrush.Therefore,it is of great significance to predict the physical parameters of water bearing in sandstone of coal seam roof for coal mine safety mining.Based on geological,well logging and seismic data,and using seismic petrophysics as a bridge,this thesis conducts analysis of elastic parameters and AVO forward for sandstone in coal seam roof,and studies the sensitivity of elastic parameter attributes and AVO attributes of sandstone to porosity.The Deep Feedforward Neural Network(DFNN)method was used to predict the sandstone porosity of coal seam roof by seismic multi-attribute fusion.In this thesis,the relationship between sandstone porosity and elastic parameters of sandstone is analyzed by using the equivalent model of rock physics and Gassmann equation.When sandstone porosity increases,elastic parameters such as the volume modulus,density,Young’s modulus,Lamé coefficient,p-wave velocity and s-wave velocity of sandstone decrease,while elastic parameters such as Poisson’s ratio and p-s wave velocity ratio of sandstone increase,and vice versa.The forward AVO modeling of the physical model of coal seam roof sandstone was carried out to analyze the influence of dominant frequency of wavelet,pore fluid type and water saturation on the reflection coefficient.The AVO numerical simulation of sandstone with different porosity showed that with the increase of porosity,the AVO attribute G value and(PG)/2 value were most sensitive to the change of porosity.On this basis,the elastic parameters and AVO attributes of the sandstone layer of the coal seam roof are inversed,and the porosity calculated by logging of the sandstone layer at drilling(sandstone layer porosity)is intersected with its seismic attributes,and the attributes sensitive to porosity prediction are selected to predict the porosity.However,since the porosity prediction accuracy of a single seismic attribute is not high,in order to further improve the accuracy of prediction,sandstone porosity is taken as the target,the DFNN method is adopted to carry out seismic multi-attribute fusion.The correlation coefficient between the prediction result of multi-attribute fusion(apparent porosity)and the porosity of the sandstone layer at the borehole reaches 0.92.By comparing the porosity prediction results of sandstone layers of different heights in eighth coal roof with the data of drilling and releasing water in coal seam roof,it is found that the area with large porosity of sandstone layer has greater water richness.It is proved that the prediction method of sandstone porosity by seismic multi-attribute fusion based on DFNN method is correct,which enriches the evaluation method of water content of sandstone in coal seam roof.There are 54 figures,17 tables and 120 references in this thesis. |
