Stress-Sensitive Reservoirs

Abstract

Technology Today Series articles are general, descriptive representations that summarize the state of the art in an area of technology by describing recent developments for readers who are not specialists in the topics discussed. Written by individuals recognized as experts in the area, these articles provide key references to more definitive work and present specific details only to illustrate the technology. Purpose: to inform the general readership of recent advances in various areas of petroleum engineering. Introduction Changes in reservoir fluids during production (fluid expansion, dissolution of gas, among others) have long been recognized, but reservoir strata themselves, except for compaction-drive reservoirs, typically have been considered to be static systems. However, a growing number of increasingly sophisticated measurements have demonstrated that some variations in reservoir deliverability are related to interactions between changing fluid pressures, reservoir stresses, and natural-fracture permeability during production and/or injection. Where appropriate tests have been conducted, the deliverability within many, if not most, reservoirs has been found to vary to some degree as a function of a natural-fracture conductivity that changes with changing stresses. This sensitivity to changing stresses is probably most pronounced in tight, overpressured, naturally fractured reservoirs where the elasticity of rock and large pressure changes can cause significant changes in fracture apertures. Fractures in such reservoirs may both dilate during injection and close during drawdown.1 The mechanical properties of the strata in some reservoirs are such that entirely new fracture sets can be created solely by production-related stress changes. The evolving model of stress-sensitive permeability in rock has met with local resistance.2 In fact, however, the recognition of the potential for rock to deform under stress led to the general belief, common several decades ago, that stresses at depth would close all fractures and that" fracture permeability" at depth was an oxymoron. This tenet has been modified and superseded with the recognition that pore pressure can counteract a significant percentage of in-situ confining stress. This produces an effective confining stress, while at the same time adding to the total stress acting on rock at depth and allowing fractures to remain open under anisotropic stress conditions. The idea that changes in stress can cause changes in the flow properties of a rock mass is not a radical departure from documented mechanical behavior of rock. Knowledge of the effect of stress changes in changing the permeability of many types of matrix rock (by changing pore-throat apertures) has existed since at least the early 1950's and led to the well-known necessity of measuring core plug permeabilities under restored-state confining-stress conditions. Changes in fracture aperture caused by variations in confining stress and/or pore pressures, and the related changes in theoretical and measured fracture permeability, are well-documented in the rock-mechanics literature.3 On a larger scale, the effects of stress changes caused by increased fluid pressures have been dramatically, if accidentally, demonstrated by the generation of earthquakes during injections of liquid waste at depth. Injection served to reduce the effective normal stresses across fault planes in these examples, allowing local shear stress to initiate slippage across those planes. On a geologic time scale, high fluid pressures are known to allow the dilation of cracks in rock and to change basic rock properties such as porosity, ductility, and strength. The mechanics of the relationship between overpressuring and the undercompaction of strata are also understood reasonably well. Mechanics of Stress Sensitivity Stress sensitivity is typically thought of in association with an increase or decrease in the apertures of natural fractures, and the related changes inconductivity within a reservoir. Although other types of stress sensitivity occur, as noted previously, this paper focuses on the causes and effects of changes in fracture apertures. Cook3 summarized the pre-1992 literature on the mechanics of fracture closure under increasing normal stress and noted that fractures that have noncongruent opposing faces, and which do not fit together perfectly(" unmated" or "mismatched" fractures), display a highly nonlinearstress-closure relationship (Fig. 1). Unmated surfaces are typical of shear fractures and mineralized fractures, probably the most common fracture types in reservoirs. Most fracture-closure strain occurs during the early stages of increasing stress; up to 70% of the potential fracture closure may occur during application of only 10% of normal stress required to reach an elastic, intact-rock stress/strain response. Fracture closure continues with increasing normal stress but at diminishing rates because of the enlarging areas of contact at the high points (" asperities") of the opposing, irregular fracture walls. Thus, early stages of pressure drawdown of a reservoir may produce a high percentage of the total potential fracture closure and related conductivity decrease. This is important because damage to fracture permeability during closure is commonly irreversible below a certain threshold of closure. Normal stress across a fracture is composed of tectonic stress, overburden stress, and pore pressure. The total stress in a system increases as porepressure increases, but the effective and differential stresses decrease concurrently. Likewise, as pore pressure in a reservoir system decreases (as during production), the differential and effective stresses become larger. Thus, the effective stress normal to a fracture increases during drawdown of reservoir pressures and is capable of narrowing fracture apertures in the system. The magnitudes of the increases in stress across a fracture vary for different lithologies by a (a poroelastic constant) and with the orientations of the fracture planes with respect to the principal stresses. Thus the amount of permeability reduction for a given reservoir pressure decrease is not a constant but must be predicted for each reservoir once the rock properties and relative orientations of the fractures and principal stresses are known.