Summary
Field measurements of fluid pressure inside hydraulic fractures have shown
rapid pressure declines along the fracture length. The consequence of this
pressure profile is rapidly tapering fracture width. This means that a
disproportionate volume of fluid and proppant injected inside hydraulic
fractures remains near the wellbore, thus creating excessive near-wellbore and
substantially less far-field fracture conductivity. This explains why history
matching of oil well production figures yields much lower effective fracture
lengths than when the same exercise is performed for gas wells, as oil wells
require higher fracture flow capacity because of their higher permeability.
The rapid tapering of the fracture width also restricts the movement of the
proppant inside the fracture, causing its accumulation near the wellbore. As
the treatment progresses, and if sufficient proppant volume has been injected
inside the fracture, the near-wellbore segment of the fracture can begin to
fill with proppant, thus reducing the open width available for further movement
of the fluid. Essentially, accumulation of proppant near the wellbore reduces
the fracture width available for fluid flow, which then results in higher
frictional pressure losses inside the fracture, further skewing the pressure
distribution and eventually leading to screenout.
Introduction
Over the last two decades, the industry has been gradually recognizing the
complexity of the theory and practice of hydraulic fracturing. Simple theories
of tensile fracturing (Haimson and Fairhurst 1967) evolved successively into
concepts of near-wellbore tortuosity (Aud et al. 1994), of the presence of
branches and shear fractures (Weijers et al. 2000), and of off-balance growth
(Daneshy 2003). The hydraulic fracture is now viewed as extending under a
mixture of tensile and shear forces and containing numerous branches, with the
extension occurring randomly around the fracture tip and highly influenced by
local inhomogeneities and planes of weakness. Randomly distributed proppant
packs forming behind narrow shear fractures are now offered as alternatives to
simple viscosity-dominated proppant transport models (Daneshy 2005). The
natural consequence of these developments is the recognition that our simple
models of pressure distribution inside the fracture also need revision and
re-evaluation. This paper extends the new fracturing concepts into analysis of
fluid pressure variations inside the fracture and reviews their impact on
fracture shape, fracture conductivity, analysis of post-frac treatment pressure
analysis, and the causes of screenout, as well as the other major application
of hydraulic fracturing, namely in-situ stress measurement.
Fluid pressure variations observed during a fracturing treatment have always
been very complex and unpredictable. Most hydraulic fracturing theories assume
the fluid pressure to be relatively constant inside the fracture, except for
the small region near its tip. Under this condition, tensile fracturing models
coupled with energy balance equations predict that fluid pressures tens or a
few hundred psi in excess of the least in-situ principal stress will be
sufficient to extend a hydraulic fracture. These same models predict the
extension pressure to be relatively flat or slightly decreasing. Yet actual
hydraulic fracturing pressures sometimes vary by several thousand psi. One
explanation offered for such pressure variations has been fracture containment
(Nolte and Smith 1979). The hypothesis offered is that fluid pressure in a
vertically contained fracture would increase during the treatment and that
constant or decreasing pressure means the fracture is experiencing excessive
vertical growth. However, in spite of its simplicity and appeal, this theory
cannot explain how the enormous energy contained in the pressurized fluid is
consumed inside the fracture, nor how the formation can resist the very large
stresses created by these pressure increases.
© 2007. Society of Petroleum Engineers
View full textPDF
(
655 KB
)
History
- Original manuscript received:
14 July 2005
- Meeting paper published:
9 October 2005
- Revised manuscript received:
20 March 2006
- Manuscript approved:
20 March 2006
- Version of record:
20 February 2007
View Cart
