关键词: 多点地质统计学, 地质建模, 变差函数, 训练图像, 现代沉积和野外露头, 精细油藏描述

Abstract: Focusing on the significance of geological modeling in fine reservoir description, this study introduces the present situation
of multi-point geological statistics modeling and its differences with traditional geological modeling methods. Taking a steam flooding
area in West depression of Liaohe Basin as an example, we analyzed the basics of multipoint geostatistics modeling, the establishment
of the training images, the advantages of multipoint geostatistics modeling compared with the traditional geological modeling. It is
pointed out that geostatistics is better than other traditional modeling methods in the reservoir prediction of cross-wells. Based on the
literature review and the authors investigations, this paper discusses the existing problems and future research directions of multi-point
geostatistics modeling. The future research in geological statistics modeling mainly includes the geological genesis analysis in order to
gain training images, and the algorithm to improve, perfect, and expand the application field of multiple-point geostatistics, etc.

Key words: Multiple-point geostatistics, geological modeling, variation function, training images, modern sedimentary and field outcrops, fine reservoir description