Page 270 - Reservoir Geomechanics

251 Wellbore failure and stress determination in deviated wells

Of course, the interesting observation in this well is that the occurrence of tensile
◦
fractures abruptly ceased when the well reached a deviation of 35 (Figure 8.9a). In fact,
this is exactly what is expected for the stress field determined in Figure 8.10.As shown
in Figure 8.9b, near-vertical wells are expected to fail in tension at mud weights just a
◦
few MPa above the pore pressure, in contrast to wells deviated more than 35 which
require excess wellbore pressures over 9 MPa to initiate tensile failure. As the ECD
was approximately 6 MPa above the pore pressure in this well, there was sufficient mud
weight to induce tensile fractures in the near vertical section of the well, but insufficient
mud weight to do so in the more highly deviated sections.
This type of forward modeling is quite useful in putting constraints on the magnitude
and orientation of S Hmax when observations of wellbore failure are available in deviated
wells. As we often have knowledge of the vertical stress and least principal stress,
we can use iterative forward modeling to constrain values of S Hmax magnitude and
orientation that match the inclination of en echelon tensile failures with respect to the
wellbore axis, ω, and their position around the wellbore circumference. As was the case
with vertical wells, the absence of drilling-induced tensile fractures in a deviated well
allows us to put upper bounds on the magnitude of S Hmax .
Because the position of both tensile fractures and breakouts around a deviated well-
bore depends on the magnitude and orientation of all three principal stresses (as well as
the orientation of the wellbore), independent knowledge of S v and S hmin enables us to
constrain possible values of the orientation and magnitude S Hmax . This technique was
used by Zoback and Peska (1995)to model the position of breakouts around a deviated
well in the Gulf of Mexico to determine the magnitude and orientation of the maximum
horizontal stress. In this case the position of the breakouts was determined from multi-
arm caliper data, the magnitude of least principal stress was known from mini-frac data,
and the vertical stress was obtained from integration of density logs, leaving the mag-
nitude and orientation of S Hmax as the two unknowns. As illustrated in Figure 8.11a, an
iterative grid search technique was used to find the range of values of S Hmax magnitude
and orientation compatible with the observations cited above. The breakouts will only
occur at the position around the wellbore in which they were observed if the orientation
of S Hmax is at an aximuth of about 136 ± 8 . This corresponds to a direction of S hmin that
◦
is orthogonal to the strike of a nearby normal fault (Figure 8.11b), exactly as expected
from Coulomb faulting theory. The estimate of S Hmax obtained from this analysis ranges
between 39.5 and 43 MPa (Figure 8.11a). Predictions of wellbore stability based on
such values are consistent with drilling experience.
As a practical point, it turns out to be quite difficult to utilize the technique illustrated
in Figure 8.11 with highly deviated wells in places like the Gulf of Mexico where the
sediments are extremely weak. The problem with using multi-armed caliper data is that
because deviated wells are usually key-seated, the caliper arms usually get locked in
the key seats and will not detect breakouts. In a study of approximately 40 wells in the
South Eugene Island area of the Gulf of Mexico, Finkbeiner (1998)was only able to