Nonplanar Fracture Propagation From a

Horizontal Wellbore: Experimental Study

Summary were confined in a triaxial loading vessel and the principal stresses ap-
This paper presents experimental results related to hydraulic frac- plied were 3,000-psi vertical, 2,500-psi maximum horizontal, and
turing of a horizontal well, specifically the nonplanar fracture ge- 1,400-psi minimum horizontal stresses. No pore fluid was present.
ometries resulting from fracture initiation and propagation. Experi- Fracturing fluid used in all tests was 30-weight motor oil with ap-
ments were designed to investigate nonplanar fracture geometries. parent viscosities of 580, 360, and 14 cp at 74, 83, and 195°F, respec-
This paper discusses how these nonplanar fractures can be responsi- tively. All experiments were conducted at room temperature with
ble for premature screenout and excessive treatment pressure when injection rate of 30 cm3/min. Identical rock type, rock properties,
a horizontal well is hydraulically fractured. Reasons for unsuccess- loading conditions, fracturing fluid properties, injection rate, and
ful hydraulic fracturing treatments of a horizontal well are presented fracturing treatment were used throughout the course of testing. The
and recommendations to ensure clear communication channels be- only variable was the wellbore orientation.
tween the wellbore and the fracture are given.
Nonplanar Fracture Geometry
Introduction Fig. 2 shows the pressure-time record for the experiment where
Several researchers have investigated critical mechanisms related to q+0°. This is a typical example of the treatment pressure recorded
during these experiments. Two experiments of each orientation
fracturing horizontal wells. In the first experimental and analytical
angle, q, were conducted to ensure data reliability and repeatability.
work on fracture reorientation and shear failure mechanisms, Dane-
Any fracture plane that does not follow the conventional single pla-
shy1 showed that inclined hydraulic fractures exhibit shear as well as
nar fracture geometry will be referred to, in this paper, as a nonplanar
tensile failure as they propagate away from the wellbore toward the
fracture. Fig. 3 depicts a typical single planar fracture created along
unaltered in-situ stress field. He showed experimentally that shear
the wellbore that is called “longitudinal fracture.” In this test, the well-
failure occurs during the reorientation process, bringing about frac-
bore is in the direction of maximum horizontal stress (q+0°). The
ture “steps” along the nonplanar fracture surface. Veeken et al.2 per-
created fracture was single and smooth, with full communication with
formed experimental and theoretical studies and observed three types
the wellbore, not only through the open hole section but along the
of limited communication between the fracture and wellbore: (1) lim-
cased portion as well. Nonplanar fractures will be categorized as mul-
ited number of communicating perforations, (2) reduced fracture
tiple-parallel fractures, reoriented fractures, and T-shaped fractures.
width near the wellbore, and (3) tortuous (or crooked) fracture.
In his experimental work, El Rabaa3 showed that multiple fractures Multiple Parallel Fractures. Multiple fractures may be created in
were created when the perforated interval was greater than four well- a deviated or horizontal well from the same interval, resulting in re-
bore diameters and the deviation angle is less than 75° (a deviation duced fracture width near the wellbore. This can be the cause of high
angle is referenced, in his work, from the minimum horizontal stress). treating pressures and/or screenouts.
In the first analysis of microfrac and minifrac tests in horizontal Fig. 4 shows a perforated horizontal wellbore oriented at a 45° angle
wells, Soliman4 showed that fracture reorientation may affect the from the direction of maximum horizontal stress (q+45°). Multiple
analysis of these tests. Hallam et al.5 concluded that surface rough- fractures were created and each perforation acted as a fracture initiation
ness of fractures initiated from deviated wellbores as observed by source. Multiple exposures in Fig. 4 demonstrate the progress of frac-
other researchers are caused by “starter fractures,” which are indi- ture propagation at and away from the wellbore. The number of propa-
vidual fracture planes initiated from separate perforations. gating fractures decreased as they extended away from near the well-
Kim and Abass6 showed that for wells with high deviation angles, bore area. Ultimately, these fracture strands would become one fracture
a pair of mutually perpendicular fractures was created and was in- propagating perpendicular to the minimum horizontal stress.
jection-rate dependent. Fig. 5 shows an openhole horizontal wellbore with q+45°. Three
fractures were created from both sides of the wellbore. Fig. 6 shows an
Experimental Study oblique cut across the wellbore where two fractures were encountered.
A series of laboratory experiments were conducted to promote re- As it is known, fracture width development is caused by loading and
search in the area of fracturing horizontal wells. Rock samples used displacing rock formation by application of fluid pressure. When two
in these experiments were rectangular blocks of hydrostone (gyp- fractures closely propagate from the same interval, as shown in Fig. 6,
sum cement) with dimensions of 6 6 10 in. These blocks were they will share the same rock material in developing width. Therefore,
cast from mixing water and hydrostone with a weight ratio of instead of having a wide fracture, two narrow fractures result.
32/100, respectively. The physical and mechanical properties are as
follows: porosity+26.5%; permeability (N2)+3.9 md; grain densi- Reoriented Fractures. If the wellbore is inclined from the direction
ty+2.32 gm/cc; bulk density+1.71 gm/cc; Young’s modu- of maximum horizontal stress, an induced fracture may initiate
lus+2.07 106 psi; Poisson’s ratio+0.21; uniaxial compressive along the wellbore and, as it propagates away from the near wellbore
strength+8,032 psi; and tensile strength (Brazilian)+807.6 psi. effect toward the unaltered in-situ state of stress, it will reorient itself
A wellbore was cast in the center of the block perpendicular to the to be perpendicular to the minimum in-situ stress. Because of this
sample axis along the 10-in. side (Fig. 1). The wellbore was cast reorientation mechanism, Daneshy1 postulated that shear as well as
with different orientation angles such that it makes an angle, q, from tensile failure modes may occur, causing wall waviness or rough-
the maximum horizontal stress. A series of angles was considered: ness along the fracture surface and within the reorientation plane.
q+0°, 15°, 30°, 34°, 45°, 60°, 67.5°, and 90°. One sample with a He demonstrated that the existence of steps on the fracture surface
vertical hole was fractured to provide reference data for a fractured is a proof of shear failure. In our work, these steps or roughness were
vertical hole under similar triaxial loading conditions. All samples noticed in all samples where the wellbore is cast in an angle such that
0tqt90. These steps may show that a single rough fracture is
Copyright 1996 Society of Petroleum Engineers
created; however, a closer examination suggests that each step is a
potential fracture propagating away from the wellbore.
Original SPE manuscript received for review Nov. 23, 1992. Revised manuscript received
Nov. 15, 1995. Paper peer approved Jan. 29, 1996. Paper (SPE 24823) first presented at the
Fig. 7 shows results of Sample HO60 (this code stands for horizon-
1992 SPE Annual Technical Conference and Exhibition held in Washington, DC, Oct. 4–7. tal wellbore with openhole section, oriented 60° from the direction of

SPE Production & Facilities, August 1996 133