Abstract:

This essay predicts the lithology parameters using seismic attribute, by which to obtain the wave impedance body, the apparent resistivity body and the porosity of body from the inversion of velocity, density, resistivity and porosity of longitudinal wave selectively using multi-attribute linear regression and probabilistic neural network technology. In the pragmatic examples, it is consistent to the knowledge of rock physics that to predict the area with comparatively higher aquosity and meanwhile in line with the aquifer defined by the theory of the mining geological. It’s no doubt significant for the prevention of coal mine water inrush accident by the analysis of aquosity to the coal seam roof and floor.

Key words: seismic attributes technique, multiattribute linear regression, probability neural network, coal roof and floor, property of water-bearing