Geoinformatika 2012; 4(44) : 46-52

PREDICTION OF THE SHEAR WAVES VELOCITIES MODEL ACCORDING TO THE DATA OF GEOPHYSICAL RESEARCHES OF WELLS AND SEISIMIC-SURVEY USING NEURAL NETWORKS

Kh.B. Aghayev 

The properties of prediction of thin-layered two-dimensional model on velocities of shear waves are given. Prediction method of two- and three-dimensional models of velocities was developed on the basis of GSW (Geophysical Studying of Wells) data on pressure and shear waves, and seismic survey of 2D/3D on pressure waves. The method was based on a creation of medium physical properties models, conducting of cluster analysis and prediction of velocities using neural networks. The model of velocities of shear waves is predicted by “Teaching” neural networks on GSW data according to the results of seismic inversion. The method was tested on geophysical data of one of the South-Caspian Basin structures. The complicated character was revealed between petrophysical properties of medium on cluster analysis.  As a result of prediction the section on velocities is more differentiated on depth and profile than on empirical dependences.

Keywords:  Neural network, cluster, prediction, velocity of shear waves, seismic inversion, time section, elastic parameters, one- and two- dimensional models of medium.