The Study & Application Of Seismic Multi-attribute Inversion

With regard to delicate oil-gas exploration ,the results of inversion play a more important role in the oil pool description and lateral in oil-gas prediction Seismic inversion integrates more acute lithology, drilling log data in vertical huge and intensive horizontal information continuity of seismic horizon.By means of inversion it reflect underground geological structured and reservior rock character. Expert can use the various rules of reservior in oil gas bearing basin to find out the profitable abundance zone of hydrocarbon to offer scientifical reference for explore and development.At present regular post-stack seismic inversion include recursive inversion , model inversion seismic which are impedance inversion and have their limitations.In of oil and gas explore because of limitation of the low-quation information and reservoir and non-reservoir often overlap between the impedence ,they result of the use simple impedance inversion to make accuracy prediction very difficult.We must adopt a new inversion method.Now seisimc attributes inversion get more and more people's attention .Seismic attributes contain abundant geological information .We can pick up hided lithology and various informations about the various reservior to establish many linear or non-linear relationship . Attributes inversions are bound to any wave and geological model and can invert any lithology parameters, which greatly improved the accuracy of the inversion .However a single seismic attribute inversion often have multiple solution and a low degree of reliability .With deep study of attribute we can build relation between multi-attribute and lithology ,use seismic attributes to analy sensitivity of reservoir parameters for the effectiveness of seismic attributes .By BP neural network to achieve reservior ,it greatly improve accuracyKonys oil field have buried deep gas reservoir, poor quality seismic data,reservior thin ,quick changs neural in horizon and so on ,using impedance inversion can not effectively distiguish ,so we use neural network technology for multi-attribute analysis to research reservior prediction ,which can resolve problem of low accuracy .However various seismic attribute, extraction ,analysis optimization best attributes were studied in seismic attribute,especially establish mathematical link between multi-attribut and reservior.By BP neural network inversion can effectively reflect regular of horizontal information of lithology changes to improve the accuracy of inversion.