Abstract: Estimation of Sand/Shale Ratio in Interbedded Formations Using Neural Network Processing of Multiple Seismic Attributes
KATZ, SIMON A., USC; GEORGE V. CHILINGAR, USC, VLADIMIR BENENSON, Academy of Natural Sciences, Russia, LEONID G. KIRYUKHIN, N SC; Tatiana Smolenchuk; NSC
The problem of integral characterization of sandstone reservoir rocks in terrigenous geologic sections is of complex nature especially when individual sand layers are I to 10 m in thickness. Thin-bedded sand-shale oil-gas-bearing formations are encountered in many parts of the world including: Gulf of Mexico, Western Siberia, Western Kazakhstan, Eastern China, and Sakhalin Island of Russia.
New method for integral evaluation of interbedded formations was developed and tested on synthetic seismic data. It is based on the processing of multiple seismic attributes by biradial-basis neural networks (BRB NN) which may be trained on recorded and synthetic seismic data. This approach is especially efficient when the studied formation is a combination of thin layers having alternating acoustic impedances and with thicknesses of individual layers beyond the resolution of seismic interpretation.
AAPG Search and Discovery Article #90937©1998 AAPG Annual Convention and Exhibition, Salt Lake City, Utah