Flow-Based Model Assisted Characterization of Sub-Seismic Karst

Abstract

Evaluating uncertainty in reservoir performance related to karst is a fundamental challenge for field development planning and optimizing recovery. Deciding which karst features matter-to flow and how to capture their effect for flow simulation is a critical, but uncertain step in developing a reservoir characterization and modeling strategy. Permeability multipliers, that are commonly required to match reservoir production history, are potentially one manifestation of not explicitly including sufficient karst features in carbonate reservoir simulation models. Karst characterization is compromised by the resolution and relative abundance of different measurement methods. Many sub-seismic karst features fall in a “data gap” because they have dimensions and networks of pores that are not, or only partially, resolved in subsurface datasets. This generates debate about the abundance, distribution and connectivity of under-sampled karst features with the objective of deciding if they matter for flow. A specific challenge is how to capture the flow significance of karst features that are smaller than a full-field scale geological model cell. High resolution geologic models termed element models, populated with distinct geological configurations of sub-seismic karst constrained from concepts and outcrop analogs, are used to address this karst characterization and modeling challenge. Flow-based averaging was used to calculate the effective reservoir properties at the scale of a full field geologic model cell. Karst was populated using two different facies modeling methods: 1) Object Based Modeling and 2) Multiple Point Statistics. Results for two karst examples presented demonstrate that the presence of certain types of karst, even < 5% by volume in a geological model cell, can increase the vertical permeability by an order of magnitude or more relative to a matrix only case. In contrast the effect of karst on horizontal permeability is less significant. Permeability anisotropy (kv:kh) is therefore increased. The proportion of karst to matrix is a property common to both the element and full field model and can be used to populate effective properties at the field scale. “Model Assisted Characterization” using high-resolution element models, that capture distinct styles of karst, in combination with flow-based averaging optimizes the characterization workflow and provides a richer understanding of irreducible uncertainty that expedites geologic model design.

AAPG Datapages/Search and Discovery Article #90259 ©2016 AAPG Annual Convention and Exhibition, Calgary, Alberta, Canada, June 19-22, 2016