Prediction of Thin Sand body Reservoirs using Facies Model-Constrained Stochastic Optimization Inversion

Yungui Xu¹Chunyuan Bai^1*Ronghu Zhang²Guikang Chen³Lei Wu⁴Xuri Huang¹

• ¹School of Geoscience and Technology, Southwest Petroleum University, Chengdu, China
• ²Research Institute of Petroleum Exploration and Development, Beijing, China
• ³BGP Southwest Geophysical Company of CNPC, Chengdu, China
• ⁴China National Petroleum Corp Chuanqing Drilling Engineering Co Ltd, Chengdu, China

Abstract—Globally, subtle hydrocarbon reservoirs in petrolifer ous basins have always been challenging targets for exploration r esearch, with thin sandbodysand body reservoir prediction being a key focus in this field. Thin sandbodysand body reservoirs typic ally manifest as thin interbeds of sandstone and mudstone. Curre nt seismic inversion techniques struggle to accurately characteriz e the distribution patterns of such thin sandbodysand body reserv oirs in the subsurface, necessitating novel inversion methods. This study proposes a stochastic optimization inversion method for thi n sandbodiessand bodies based on facies-model constraints. Dyna mic forward modeling is conducted using typical thin sandbodysa nd body patterns to establish reasonable identification templates f or superimposed configurations. The seismic data are subjected t o steerable pyramid processing to achieve a multi-scale represent ation. By incorporating sedimentary facies analysis data, a facies-constrained volume is created. The acoustic parameter difference s between the thin sandbodysand body reservoirs and the mudsto ne are analyzed, followed by stochastic optimization inversion of t he sensitive parameters and interpretation of the inversion result s. The inversion results for beach-bar thin sandbodysand body re servoirs in the Yingmaili area of the Tarim Basin show strong con sistency with well-based sandbodysand body correlations and pla nar sedimentary facies distributions. Blind well validation demon strated prediction accuracies of 63% for sandbodiessand bodies t hicker than 4 m and 84% for those exceeding 10 m. In compariso n, conventional inversion methods achieve a prediction accuracy of only about 50% for thin sandbodiessand bodies ranging from 4 to 10 m in thickness, and approximately 77% for those exceeding 10 m. The new methodology demonstrates improved prediction a ccuracy for thin sandbodiessand bodies, thereby providing more reliable support for interpreting and evaluating the hydrocarbon potential of thin sandbodysand body reservoirs.This study achiev es high-accuracy prediction of thin sandbodiessand bodies and pr ovides a novel methodology for the detailed characterization of th in sandbodiessand bodies in beach-bar sedimentary basins world wide.