Dr. Theses

Development of methods of processing and interpretation of data of seismic exploration, well logging using non-classical theory of deformations and neural networks

Aghayev Khanlar Boyukaga oglu

The methods and algorithms of their processing and interpretation are developed and improved taking into account the geodynamic state of the medium in order to increase the efficiency of seismic exploration and well logging. The "GEOPRESS" complex consisting of 45 programs is created on their basis, based on the geophysical package "CWP". The studying of the medium on some areas in the SCB on creation of models of petrophysical, geodynamic and elastic properties, geodynamic correction of models, prediction of models on NN, data processing of MSL, VSP, 2D and interpretation of all these materials is conducted with their use on well logging and seismic data.

The anisotropy related with the direction of packing of the rock grains is determined on velocities of shear waves. It is established that the elasticity moduli of the third order are more sensitive to changes of the elasticity of the rock than the second order. A new possibility of studying the porosity of the medium on seismic data is created. The thin layering of the medium, petrophysical variability, and the reliability of lithological division are quantitatively determined on cluster sections obtained by NN on well logging data. It is determined, that the relation between the petrophysical properties of rocks is more accurately characterized by clusters than by lithology.

The models are predicted on petrophysical and elastic properties of the medium on Shah-deniz, Nakhchivan and Umid structures. The nature of change of pressure, velocity, clayiness, porosity and density of rocks along seismic profiles are determined. It is shown that the revealed anomalies in the models, are related to oil and gas.