Browse Prior Art Database

Pattern-Based Approach to Multiphase Flow Upscaling Using Distance-Based Modeling

IP.com Disclosure Number: IPCOM000200463D
Publication Date: 2010-Oct-14
Document File: 5 page(s) / 64K

Publishing Venue

The IP.com Prior Art Database

Abstract

We propose a pattern-based upscaling technique that preserves multiphase flow behavior on a single-reservoir-model basis. By utilizing a fast static algorithm, a subgrid-scale profile of displacement front is rapidly generated for each of simulation gridblocks based on fine-scale geological description. By applying distance-based modeling, the simulation gridblocks are grouped into clusters in accordance with the similarity of the pattern of displacement front profile in subgrid-scale. All simulation gridblocks belonging to the same cluster can share the same pseudo relative permeability because of the strong relationship between displacement front profile and pseudo function.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 1% of the total text.

Page 01 of 5

Pattern-Based Approach to Multiphase Flow Upscaling Using Distance- Based Modeling

Abstract

We propose a pattern-based upscaling technique that preserves multiphase flow behavior on a single-reservoir-model basis. By utilizing a fast static algorithm, a subgrid-scale profile of displacement front is rapidly generated for each of simulation gridblocks based on fine-scale geological description. By applying distance-based modeling, the simulation gridblocks are grouped into clusters in accordance with the similarity of the pattern of displacement front profile in subgrid-scale. All simulation gridblocks belonging to the same cluster can share the same pseudo relative permeability because of the strong relationship between displacement front profile and pseudo function.

Introduction

Methods for multiphase flow upscaling have been investigated and proposed by numerous researchers for decades. Most of such works are categorized as the generation of pseudo relative permeability (or effective relative permeability) using dynamic (i.e. flow-based) method and/or methods to assign pseudo functions to simulation gridblocks. However, technologies to generate and assign pseudo relative permeability have not yet reached maturity in both terms of robustness and practicality. The major limitation of existing dynamic pseudo methods is that they generate as many pseudo relative permeability curves as number of simulation gridblocks, which is not manageable in field applications. Several researchers have proposed methods for grouping pseudo relative permeability. Dupouy et al. (1998) applied cluster analysis and principal component analysis for grouping pseudo relative permeability. Their workflow is classified as "a posteriori grouping (Dupouy et al., 1998)", as opposed to "a priori grouping (Dupouy et al., 1998)", and first generates pseudo function for every coarse-scale gridblock by running multiphase flow simulation on fine-scale model, then, groups them into clusters. Christie and Clifford (1998) proposed a method of "a priori grouping" where simulation gridblocks are grouped first, and then, only one pseudo function is generated for each group. Their approach runs streamline simulation on fine-scale model and generates breakthrough curves (i.e. fractional flow vs. breakthrough time of individual streamlines) for each coarse-scale gridblock. The coarse-scale simulation gridblocks are grouped based on the similarity of the breakthrough curves using cluster analysis. Chen and Durlofsky (2007) proposed another "a priori grouping" method which groups simulation gridblocks based on mean and standard deviation of find-scale fluid velocity within the coarse-scale gridblocks by using k-mean clustering. The velocity is obtained by running single-phase flow simulation on fine-scale model. Their method is developed as "ensemble-level upscaling" which preserves multiphase flow behavior on ensemble-level basis for the purpose of ensemble-based performance...