财产(哲学)
网格
计算机科学
地质学
大地测量学
认识论
哲学
标识
DOI:10.3997/2214-4609.202335037
摘要
Summary Over the past forty years, the petroleum industry has consistently employed the same subsurface property modeling workflow. In this workflow, a geomodeller constructs a structural framework based on interpreted stratigraphic horizons and faults from seismic data. Subsequently, a three-dimensional grid is generated from this framework. Well log and seismic data are then blocked, or upscaled, to match the grid's cell resolution. Geostatistical parameters, such as variogram models, are inferred from the upscaled data, and utilized to simulate representative facies indices or average petrophysical property values in all grid cells. The main reason for this grid-based approach is that the resulting property models are primarily used for volumetric calculations and flow simulations. An alternative approach, devoid of grids, consists in simulating property values at point locations using the original volume support from the well log data. This approach, detailed in this paper, offers distinct advantages, including a more efficient and flexible geomodeling experience, multi-scale modeling that can be shared across exploration and development teams, and a multi-resolution simulation capacity that enables fit-for-purpose modeling. Additionally, this paper outlines the computation of volumetrics from point-based models and introduces a novel static-to-dynamic modeling workflow that eliminates the need for flow simulation grid upscaling.
科研通智能强力驱动
Strongly Powered by AbleSci AI