--> Application of Geostatistical Inversion in the Thin Sand Body Prediction: A Case Study in Yangqian 19 Area, Nanxiang Basin
[First Hit]

AAPG/SEG International Conference & Exhibition

Datapages, Inc.Print this page

Application of Geostatistical Inversion in the Thin Sand Body Prediction: A Case Study in Yangqian 19 Area, Nanxiang Basin

Abstract

Abstract

Yangqian 19 area, located in Nanxiang Basin, is characterized by rapid facies change, strong heterogeneity, and small thickness of single sand body. As the earthquake-band of the conventional deterministic inversion is limited, the vertical resolution of the seismic inversion body is very low, and hence it is often difficult to identify the thin sand body. The geostatistical inversion method based on stochastic modeling technique can effectively integrate geological, logging and seismic data, which can greatly improve the seismic vertical resolution, and identify the thin sand layers more accurately. Thus, the geostatistical inversion method is used to predict the sand body spatial distribution in this research.

First, a correlation analysis is carried out between lithology (from core data) and well logging data, and resistivity log shows the best correlation with lithology. Then the wave impedance data body is acquired by using constrained sparse spike inversion, which is used to get the horizontal variogram, while the vertical variogram is acquired based on the well data. After a correlation analysis between resistivity and wave impedance at the well point, resistivity logs are discretized to lithological data from 30 wells which are used for the inversion. Next, regarding resistivity value as principal variable, whilst the wave impedance data body as covariate, MCMC method is used to inverse principal variable and several equal probability realizations of resistivity invertomer are obtained finally. Then by calculating root mean square (RMS) of all probabilities, the data body of RMS is the final resistivity invertomer, and then transformed into the lithology invertomer, so as to forecast sand body distribution. Results of inversion are consistent with well data by Previous HitverificationNext Hit wells and vertical resolution is significantly improved. In addition, forward Previous HitmodelTop is also established for further evaluating inversion results. By comparison, synthetic seismogram and original seismic trace is highly matched. In conclusion, stochastic seismic inversion has good effect on this research.