Mechanical Properties of Friable Sands From

Conventional Log Data

N. Stein, SPE-AlME, Mobil Research and Development Corp.

Introduction
Certain in-situ mechanical properties of friable sands slip by each other as the acoustic signals pass through.
must be measured to determine if such formations will Many of the rock properties could then be calculated
remain stable under different wellbore conditions. Sta- directly5,6 from such measurements. Grain slippage
bility problems that can then be resolved include esti- apparently occurs readily in many friable sands; the
mating (1) how fast oil or gas may be produced with- shear-wave signal is dampened severely, and it is dif-
out causing a sand problem,1,2 (2) if there is sufficient ficult to record reliable data. Techniques are developed
cementation between adjoining sand grains to avoid a in this paper that make it possible to substitute conven-
possible sand problem when water production occurs, 1,3 tional sonic-log data for full-wave log data for calculat-
and (3) the formation fracture pressure gradient. 4 ing the mechanical properties of friable sands.
Specific mechanical properties that enable such esti- Other methods have been reported l ,4,7 for using log
mates are sand strength, strength of cementation be- data to measure the mechanical properties of friable
tween grains, bulk compressibility, and Poisson's ratio. sands. None of these methods incorporated all the fol-
This paper presents techniques for calculating these lowing factors that can affect the mechanical properties:
properties. (1) overburden weight gradients at different depths, (2)
The means for determining the in-situ mechanical other earth loads caused by local geological conditions,
properties of friable sands is through the use of well-log (3) cementation, (4) fluid saturation, (5) fluid pressure,
data.' The accuracy of these properties determined from (6) different bulk compressibilities, and (7) the change
log data should be better than the properties determined in elastic modulus values l of friable sands at different
by direct strength tests with core samples. Weakly load conditions. These factors are considered in the cal-
~ cemented materials probably would disaggregate during culation technique reported here.
the core recovery operation, and no strength of cemen-
tation would be detected in direct strength measure- Modulus Properties of Friable Sands in a
ments. Relief of overburden load results in core expan- Test Area
sion and subsequent breaking of weak bonds. Modulus values have been used as a measure of the
Logs used include conventional sonic, density, induc- strength properties of friable sands. 1,2,7 It was found that
tion, and open-hole gamma ray. The conventional sonic a straight-line function exists between modulus values
log is used in preference to a full-wave sonic log for of friable sands and grain-to-grain vertical load (effec-
calculating the mechanical properties of friable sands. tive stress) at reservoir depths. A curved line would re-
The full-wave log can provide both compression and sult if depths are plotted instead of effective stresses;
shear-wave travel times if the formation grains do not this occurs because the overburden weight gradient in-

A new log interpretation method is presented for determining mechanical properties offriable
sands. Factors considered in the method include overburden weight, fluid pressure, and
additional earth loads caused by local geological conditions. Log data needed for the
interpretation include conventional sonic, density, induction, and open-hole gamma ray.

JULY, 1976 757

creases with depth. Thus, a plot of modulus values vs Means for calculating overburden weight gradients from
effective stress removes the influence of the variations density log data were reported earlier. 8
of overburden weight gradient, and the straight-line Effective vertical stress provided by the overburden
function is a property of the sands alone. Discussion on weight may be determined from Eq. l.
the development of this straight-line function follows.
(Tevo b = D (gob - gf) ....................... (1)
Sonic and density log data were available for a sub-
stantial number of different friable sands at depths vary-
The log data were then used to calculate elastic modulus
ing from 1,000 to 8,000 ft. The wells penetrating these
values 1 from Eq. 2.
sands are located within a rectangular area of about
2,500 sq miles. Wells included in this study were logged 1.34 X 10 10
Ec = .. , ........... , ....... (2)
with an ll-lb/gal mud in the hole; different mud weights t2
could affect sonic log values (see Appendix). Sands that
Straight lines averaging the data for liquid-filled
were saturated with liquid at a pressure gradient of about
sands in each well were drawn. These lines were found
0.44 psi/ft were considered in the first correlations; then
to be the same for all the normally pressured intervals
the study was extended to include sands containing gas at
of wells in the test area. A line for sands containing gas
the same pressure gradient.
was developed similarly. Plots of these averaging lines
Overburden weight gradients were computed from
are drawn in Fig. 2. The lines are parallel and the angle
density log data; these gradients are plotted in Fig. I.
'" between the straight lines and the horizontal axis is
common to both sand systems. The straight line for
liquid-saturated sands intercepts the modulus axis at
0.74 X 10 6 psi, and the line for sands containing gas
intercepts at 0.44 X 10 6 psi; therefore, the difference in
" ,, modulus to be considered when comparing mechani-
,, cal properties of oil and gas sands appears to be 0.3 x
,, 106 psi.
The common angle, "', for each of these straight-line
\
\ , \
modulus plots indicates that all the different friable
sands at depths ranging from 1,000 to 8,000 ft over the
2,500-sq mile test area follow a given pattern. It ap-
\ pears reasonable to assume this angle'" applies to fri-
\
x \ able sands in other areas also because the pattern was
\ demonstrated for the many sands in the test area.
3
\
:0:::
\ Strength Plots
l-
e.. \
u.J
\ The above test data are used in determining mechanical
o
\ properties of friable sands by considering the modulus
10 t- \ plot to be a mathematically remapped Mohr-envelope 3
\ strength plot. Discussion gf the common properties of
\
these plots follows.
12 t-
Mohr-Envelope Strength Plot
A material will remain stable at any set of grain-to-grain
14L-----~ ______L __ _ _ _ ~ ____ ~ ______ ~~ load conditions below the Mohr-envelope strength plot 3
0.80 0.84 0.88 0.92 0.96 1.00 drawn in Fig. 3. When load conditions reach any point
along the Mohr envelope, the strength of the material is
OVERBURDEN WEI GHT GRADI ENT W), psi/It
overcome and part of the material can slip, or "fail."
Fig. 1 - Overburden weight gradients in test area.
In Fig. 3, the envelope is drawn as a straight line hav-
".
ing the angle <p with the horizontal axis; a straight line
"
, / / ...... /
is a reasonable approximation of the Mohr envelope for
...... / / / stress conditions normally found in reservoirs .
". "./ ' " Fig. 3 illustrates that there can be two contributions
///....... /
Liquid Filled ". /,'" to the strength of the material: (1) friction between sand
sands ". grains must be overcome to cause failure, and (2) cohe-
"
""
" . ' - - Gas Sands
sion or cementation between adjoining sand grains must
". " be broken to cause failure. The contribution of friction
".
".
".
to the strength may be expressed as the tangent <p. The
".
".
". contribution of cementation to the strength is the value
,0/ /
".
of strength at (T e = O.
".
Slope of Plots tan 0/ 0.000615 X 106
". 0 0

Modulus Strength Plot

"0/
oL-__ ~ __-L__ ~ ____ ~ __ ~ __ ~ __ ~ ____ ~
The modulus plot for our test area (Fig. 2) has contribu-
o tions to the combined elastic modulus value, E c ,
3
NORMAL EFFECTIVE STRESS I <r I, psi X 10- analogous to the Mohr-envelope strength plot; however,
Fig. 2 - Mod ulus plots for friable sands in test area. there is one added factor contributing to the value of
758 JOURNAL OF PETROLEUM TECHNOLOGY
Ee. This factor is the bulk modulus, E/j, or the recip-
rocal of bulk compressibility. The E/j values depend on
the compressibilities of both formation solids and fluids; 'v;
0-
therefore, it should be expected that the E/j value for a
given sand containing oil should be different from the I:
uo
E/j value for the same sand containing gas. C<:

The combined modulus may be expressed 9 as 3

~
4 uo
'"
Ec ="3 E.• + E/j ........................ (3)
I
~
<.:>
V>
::::>
«
c:
.~
is
C<:
u
~ 0 c:
V> ~

Shear modulus of a material subjected to a given total V>

V>
uo
load, E s , is defined!O as the ratio of shear load to lateral C<:
~

deformation; and Es serves as a direct measure of the V>

C<:
strength of the formation solids alone.! ;::s
I
Let us examine all the features of the modulus plot as ~

illustrated in Fig. 4. There are three factors contributing Cohesion or Cementation

to the combined modulus value: friction, cohesion or Contribution to Strength

cementation, and bulk modulus.

Similar to the Mohr envelope, the friction contribu- I ncreasi ng

tion to Ee may be expressed as the tangent t/J; for the LOAD

sands in the test area described above, the tangent t/J = (EFFECTIVE STRESS W P I, psi 1
0.000615 X 10 6 • The cementation contribution plus the Fig. 3 - Mohr-envelope strength plot.
bulk modulus contribution to Ee is the value of Ec on
the modulus plot at (J eV = 0, or
[Ee[ (Tpv = 0 = Ec - 0.000615 X 10 6 (JeV

'"
c:
'v;
= 11 Es 1(Tpv = 0 + Eb ......... (4) ro
l'"
u
c:

These two latter contributions may be separated

by evaluating E b , then subtracting this result from u
[Ee [(Tpv = 0 to yield [4/3 Es [(Tpv = 0, the cementation
contribution.
The correlation in Fig. 5 was develored, to estimate
separately the bulk modulus from [Ee (Tpv = o. Data
points in the figure include values for the liquid-filled 8z
friable sands of our test area and other formation mate- "'
2:
o
rials reported in the literature.!O These other formation u

materials include gneiss, granite, gabbro, tuff, calcari-

ous sands, and breccia mosaic in granite matrix. The
correlation may be used directly to determine Eb values
for liquid-filled friable sands. It also may be used for 0 - - - - - - - - - - - - - - - - · Increasing
friable sands containing gas if the following assumption
is made: the 0.3 X 10 6 psi shift in Ec values between EFFECTIVE STRESS ( (T 1 psi
liquid- and gas-containing sands of our test area (see Fig. 4 - Modulus plot.
Fig. 2) applies to other friable sands. This shift is ap-
parently caused by a change in Eb when gas is present;
the friction contribution to Ec remains the same for
either system.
The procedures for calculating Eb are given below:
o Based on Published Data 10
• Based on our Test Area Data
For Formations Containing No Gas:
I. Calculate [Ec [(Tpv = 0 using Eq. 4.
2. Read Eb from Fig. 5.
For Formations Containing Gas:
1. Calculate [Ec[(Tpv = 0 using Eq. 4.
2. Add 0.3 X 10 6 psi to the value of [Ec[(Tev = 0 ob-
tained in Step 1. This eliminates the effects of gas.
3. Read Eb from Fig. 5.
4. Subtract 0.3 X 10 6 psi from the value of Eb ob-
tained in Step 3. This reintroduces the effects of gas and
provides the value of Eb for the conditions existing in COMBINED MODULUS AT (J = 0 qEclQ'"~ol, pSI X 10-
6

the formation. Fig. 5 - Estimating the bulk modulus.

JULY, 1976 759

Problems When Using the Modulus Plot Correction for Mud Weight
There are two additional problems that must be solved The second problem with using the modulus plot for
before the modulus plot developed for friable sands in friable sands occurs when log data are taken with mud
our test area can be used to determine mechanical prop- in the borehole having a density other than II Ib/gal.
erties of friable sands in general. The first problem is We have observed cases where the use of different mud
to know the vertical effective stress acting on the sand; weights can alter the travel time recorded on the sonic
the second is to know how to deal with log data taken log (sonic and density data are used in Eq. 2 to calcu-
in wells where the mud weight is different from the late modulus values). There is a relaxed state of effec-
II-Ib/gal mud of the test area. Our solutions to these tive stress in the sand adjoining the wellbore,11 and the
problems are given below. recorded travel time on the sonic log reflects the path
followed by the recorded acoustic signals in this stress
Effective Stress field. Different mud weights will change this stress
The effective vertical stress, (TeV, acting on a formation field.
is made up of the weight of the overburden and reser- The modulus plot for friable sands was developed
voir pressure in accord with Eq. I and additional earth from log data taken with II-Ib/gal muds in the bore-
loads, (TE. An example of an additional earth load is a holes with corresponding stress fields at different depths.
salt dome that provides pressure from some lower The slope of this modulus plot is used in calculations
depth. Then, of mechanical properties of different friable sands.
Therefore, sonic log data used to calculate modulus
(Tev=(Tevob + (TE . . . . . . . . . . . . . . . . . . . . . . (5)
values of these different sands should be on the basis
There is a way to evaluate the additional earth load of the same stress fields that would exist if an 1 I -Ibl
using resistivity log data for normally pressured shales. gal mud were in the wellbore.
Straight lines may be drawn through plots of the Travel times from sonic log data for friable sands
logarithm of the shale resistivity, R sh , vs the effective corrected to an I I-Ib/gal mud-weight basis can be calcu-
vertical stress for normally pressured shales; modifying lated using Eq. 8. The derivation of Eq. 8 is given in
the form of Eq. I, the Appendix; this derivation is based on the developed
modulus plot for friable sands.
I(T el b In = D(gob - Igfl n) .................... (6)
t = tOB - 2.295 X 10- 8
The slope of each straight line from such a plot, Ll In R/
I
Ll (T evob In,
can be used in the correlation plotted in (tOB)3 (TeV (I - Fwv II I) ] ............ (8)
Fig. 6 to obtain a measure of additional earth loading, [
PbD
(TEll (T ev I n· Then,
ob

Use of the above correction should be applicable to log

data taken with the borehole-compensated sonic tool;
where that tool was used to log wells in the test area.

is obtained from Fig. 6, Mechanical Properties

I(T evob In With the information presented above, it is now possi-
and ble to estimate the mechanical properties of friable
I(T evob I n is obtained using Eq. 6. sands from log data. These calculations are illustrated
below using data for a shaly sand at a depth of I 1,154 ft
The data points plotted in Fig. 6 were obtained for in Well B. Mud weight while logging this well was 17
shales in three geographically distant areas. Ib/gal; the sonic log value at the sand depth was 100
microsec/ft; the density log value was 2.2 gm/cc. No
gas was in this formation.
·From the Slope of the ReSlStlvlty Plot for Shales In
whJCh the Hydraulic Gradient IS Normal for the Area. Modulus Data
Slope "~In Rsh 16.1 (T ~~ I n where I (J" ~~ In' Effective Stress Determine Vertical Effective Stress,
Contributed by Overbur~n Weight, when the Hydraulic Gradient IS Normal for the Area.
ob
•• O"E • Effectl.ve Stress Caused by Additional Earth Load over that
Contributed by Overburden Weight Alone
(Tev=(Tev + (TE (Eq. 5):
0.4r-------'------''------------,
0.2
First,
o AREA 1
.6. AREA 2
o AREA 3
-0.2

~-O.4
where
i5';:
~-O.6 D= 1 1,154 ft (sand depth)
b -0.8
gob = 0.956 psi/ft (from density log data 8 )
-1.0

gf= 0.86 psi/ft (from field test).

-1.4 '--_.l..-_-'--_-'--_--'--_--'--_...L_....1...._-'-_-L-I
-30 -25 -20 -15 -10 -5 0 10 15 Then,
6in R:.h 16.10'"~~ I n' ohm meters/psI X 10
4

(T el b = (II, 154) (0.956 - 0.86)

Fig. 6 - Estimating' effective stress" caused by additional
earth load. = 1,070.8 psi.
760 JOURNAL OF PETROLEUM TECHNOLOGY
Second, Determine Combined Modulus,
10
E = 1.34 X 10 PbD (E q .2)'
e t2 .

where First, correct the sonic log data to 11-lb/gal mud-

weight basis.

t = tOB - 2.295 X 10- 8

where
I gfl n = 0.465 psi/ft (nonnal fluid pressure
gradient in the area).
Then, where
I(T eVo
b
In = (11,154) (0.956 - 0.465) tOB = 100 microsec/ft (sonic log data)
= 5,476.6 psi, (TeV= 742.2 psi (calculated above)

and F wv = 17 lb/gal (given mud weight)

PbD = 2.2 gm/cc (density log data).
0.060.
I(T evo I nb Then,
t = 100 - 2.295 X 10- 8
(From Fig. 6, A In R/ A I(T evo b In = 6.64 X 10- 4 ohm
(100)3 (742.2) (1 - 17111)J
m/psi. A In R/ A I(T evo b In is the slope of the resistivity [
plot for normally pressured shales in the well - see 2.2
Fig. 7. Shale resistivity points shown were read at = 104.2 microsec/ft.
depths where the open-hole gamma ray log values were
high.) Finally,
10
Then, E - 1.34 X 10 (2.2)
e- (104.2)2
(TE = - (0.060) (5,476.6)
= 2.714 X 106 psi.
328.6 psi.
Compressibility = I/Eb:
Finally,
Eb may be read from Fig. 5 at I Ee IO'eV = 0 :
(TeV= 1,070.8 - 328.6
= 742.2 psi. IEelO'er = 0 = Ee - (TeV tan'" (Eq. 4)
= 2.714 X 106

3
- (742.2) (0.000615 x 106)
DEPTH. ft X 10-
1O~__-T____~_____6r-____T8__~~1~0____~12 = 2.257 X 106 psi.
8 Therefore,
Eb = 1.79 X 106 psi (Fig. 5),
Trend for Norma I
4 Hydraulic Gradient and compressibility = l/Eb = 0.56 X 10- 6 psi-I.
Measure Cementation Strength, IEs IO'eV = 0 :
l>ln Rsh 6.64 X 10-4 J From a form of Eq. 3, applicable to the 0
~ 2 l>i crOb i = ohm meters/psi '-0-, (T eV

~
E
-§
[
eV n
"..- _ _ _ ,
\
0
{,
\
"\0\
condition,

IEs IO'eV = 0 = ~ (I Ee IO'eV = 0 - Eb)

>-' 1. 0 (a ° , \ ° \
~ 0.8 ~° 0/
I--~./o \. oOB \
= 1 [(2.257 x 106) - (1.79 X 106)]
~ 0.6
V>
~
Zone of Rapid
Changes in
/ \0 ° oj 4
"" Overpressured "_0./
0.4 Water Salinity
Sha les = 0.35 X 10 6 psi.
and Resu Iti ng
Shale Properties
Measure Sand Strength In Situ, Es:
From Eq. 3,
0.2

Es = ~ (Ee - Eb)
0.1 L..____J.....____..J....____...l-____-L-____......L..____.......

a =1 [(2.714 x 106) - (1.79 x 106)]

NORMAL EFFECTIVE STRESS (I cr ~~ In). psi X 10-
3 4
Fig. 7 - Normal resistivity plot for shales in Well B. = 0.69 X 106 psi.
JULY, 1976 761
Determine Poisson's Ratio:
Poisson's ratio, JJ-, may be expressed in terms of the Acknowledgments
following modulus values. 12 The author wishes to thank Mobil Research and De-
velopment Corp. for pennission to publish this paper.
JJ- = 3Eb - 2Es The author also wishes to thank E. R. Parish for his aid
6Eb + 2Es in developing the data for the test area.

= 3(1.79 X 10 6 ) - 2(0.69 x 106 ) References

I. Stein, N. and Hilchie, D. W.: "Estimating the Maximum Pro-
6(1.79 X 10 6 ) + 2(0.69 X 106 ) ducing Rate Possible From Friable Sandstones Without Using
Sand Control," J. Pet. Tech. (Sept. 1972) 1157-1160; Trans.;
= 0.33 ....................... (9) AIME,253.
2. Stein, N., Odeh, A. S., and Jones, L. G.: "Estimating
Conclusions Maximum Sand-Free Production Rates From Friable Sands for
Different Well Completion Geometries," J. Pet. Tech. (Oct.
Conventional density, sonic, induction, and open-hole 1974) 1156-1158; Trans., AIME, 257.
gamma ray log data can be used to determine measures 3. Hall, C. D., Jr., and Harrisberger, W. H.: "Stability of Sand
of the mechanical properties of friable sands. These Arches: A Key to Sand Control," 1. Pet. Tech. (July 1970)
properties include compressibility, strength of cementa- 821-829.
4. Anderson, R. A., Ingram, D. S., and Zanier, A. M.: "Fracture
tion, in-situ strength, and Poisson's ratio. Such property Pressure Gradient Determinations From Well Logs," J. Pet.
measures can be useful to sand control efforts. The ap- Tech. (Nov 1973) 1259-1268.
proach is to consider that the strength of friable sands 5. Mason, W. P.: Physical Acoustics and the Properties of Solids,
will increase with increasing grain-to-grain loads at a D. Van Nostrand Co., Inc., Princeton, N. J. (1958).
6. Love, A. E. H.: A Treatise on the Mathematical Theory of Elas-
fixed rate empirically detennined. Factors considered in ticity, Dover Publications, Inc., New York (1944).
computing loads include overburden weight, fluid pres- 7. Tixier, M. P., Loveless, G. W., and Anderson, R. A.: "Estima-
sure, and additional earth loads caused by geological tion of Formation Strength From the Mechanical Properties
conditions. No previously reported method for deter- Log," J. Pet. Tech. (March 1975) 283-293.
mining sand properties from log data consider all these 8. Eaton, B. A.: "Fracture Gradient Prediction and Its Application
in Oilfield Operations," J. Pet. Tech. (Oct. 1969) 1353-1360;
factors. By following the above approach, it also be- Trans., AIME, 246.
comes possible to correct conventional sonic log data 9. Dobrin, M. B.: Introduction to Geophysical Prospecting, 1st ed.,
for sands to values that would have been measured if an McGraw-Hill Book Co., Inc., New York (1952).
Il-Ib/gal mud had been in the well. 10. Myung, J. I. and Helander, D. P.: "Correlation of Elastic Mod-
uli Dynamically Measured by In-Situ and Laboratory Tech-
Nomenclature niques," The Log Analyst (Nov.-Dec. 1972) 22-33.
11. Westergaard, A. M.: "Plastic State of Stress Around a Deep
D = depth, ft Well," Journal, Boston Society of Civil Engineers (Jan. 1940)
Eb = dynamic bulk modulus (reciprocal of 387-391.
bulk compressibility), psi 12. Jaeger, J. C. and Cook, N .. G. W.: Fundamentals of Rock
Mechanics, Halsted Press, New York (1969).
Ec = dynamic combined modulus, psi
Es = dynamic shear modulus, psi
F wv = weight of the mud while logging, Ib/gal Appendix
gf = fluid pressure gradient, psi/ft Derivation of Sonic Log Correction for Mud Weight
I gfl n = fluid pressure gradient normal to the Use of different mud weights, F wv , at a log depth D
area, psi/ft can alter the travel time of the recorded acoustic signal
gob = overburden weight gradient, psilft obtained with the sonic logging tool used in the test
t = travel time of an acoustic signal obtained area. It is necessary to convert such travel times to the
from sonic log data corrected to the Il-lb/gal mud-weight basis. The modulus infonnation
basis of having an 11-Ib/ gal mud in the for friable sands was developed using log data taken
wellbore while logging, microsec/ft with an II-Ib/gal mud in the hole. It is desirable to use
tOB = travel time of an acoustic signal obtained that developed infonnation for calculating mechanical
from the sonic log, microsec/ft properties of all friable sands; therefore, sonic log data
PbD = bulk density from the density log, gm/cc must be on a common basis.
JJ- = Poisson's ratio The change in travel time when logging with mud
a e = effective stress, psi weight F wv in the hole is caused by the difference in the
a eV = effective vertical stress, psi stress field at depth D from that provided by an II-Ib/
a el b = effective vertical stress provided by the gal mud. But at some depth D + llD, the pressure
overburden weight, psi exerted by an II-Ib/gal mud is equal to the pressure
Ia evob In = effective vertical earth stress provided by exerted by the F wv-Ib/gal mud at depth D; llD is posi-
the overburden weight if formations tive if F wv > 11 lb/gal, or is negative if F wv < 11
are normally pressured, psi Ib/gal. If the vertical effective stress in the sand, aev, is
aE = effective vertical stress provided by earth assumed to be the same in both cases, the sonic travel
loads other than overburden weight, times and the corresponding combined modulus values
pSI should be the same.
am = Ib/gal density equivalent to Fwv weight Correction of the II-Ib/gal-based modulus from the D
mud + llD depth to the D depth is used to obtain what the
a l l = Ib/gal density equivalent to Il-Ib/gal travel time would have been at depth D if an II-Ib/gal
weight mud mud had been in the well bore . The development of the
762 JOURNAL OF PETROLEUM TECHNOLOGY
approach used to correct the travel time is given below. But,
The equation for calculating the combined modulus
from log datal taken in our test areas with 11-lb/gal
am = Fwv
mud in the wellbore may be expressed as a 11 11
10 therefore
Ec = 1.34 X 10 PbD • • • • • • • • • • • • • • • • • (A-l)
(t OB )2
6.D = D ( 1 - Fl~v) . . ................. (A-5)
Differentiating with respect to taB and solving for dEc,
10 As stated in the first part of this section, it is assumed
dEc = - 2.68 X 10 PbD dtoB , .......... (A-2) that the vertical effective stress on the formation sand in
(taB )3
the imaginary 11-lb/gal system at D + 6.D is the same
From the modulus plot for friable sands in Fig. 2, the as in the real system at depth D. This stress is termed
slope of the straight line may be expressed as a eV; the stress gradient in psi/ft would be a evlD. Then,

dEc = tan 1/1 = 0.000615 X 106 , 6.a eV = 6.D ( a~v) ...................... (A-6)
&rev
or Combining Eqs. A-5 and A-6,

Fl~v)
6
dEc = 0.000615 x 10 daev ............. (A-3)
6.a ev =a ev ( 1- ................. (A-7)
Substitute Eq. A-3 into Eq. A-2 to obtain
_-2.295x 1O- 8 (t oB )3da e v Finally, substituting Eq. A-7 into Eq. A-4, the correc-
dtOB - tion to travel time observed is given by
6.tOB = -2.295 X 10- 8
or
(tOB)3 aev (1- F wv !11)] ......... (A-8)
6.tOB = - 2.295 X 10- 8 (tOB)3 6.a eV ...... (A-4) [
PbD
PbD
And the travel time converted to the 11-lb/gal mud
Now it is necessary to find the 6.a eV value in Eq. A-4
weight is taB + 6. taB, or
corresponding to the 6.D distance discussed at the be-
ginning of this section for the conditions when a dif- t = taB - 2.295 X 10- 8
ferent mud weight is in the hole. Pressures exerted by
(toB)3aev (1 - Fwvlll)] ......... (A-9)
the two mud columns of different weights are equal at [
two depths in accord with the following equation: PbD JPT

amD = a ll (D + 6.D),
Original manuscript received in Society of Petroleum Engineers office July 28,
or 1975. Paper accepted for publication Feb. 13, 1976. Revised manuscript received
March 15, 1976. Paper (SPE 5500) was first presented at the SPE·AIME 50th
am Annual Fall Technical Conference and Exhibition, held in Dallas, Sept. 28-0ct. 1,
6.D=D(I- ). 1975. © Copyright 1976 American Institute of Mining, Metallurgical, and Pe·
all troleum Engineers, Inc.

JULY, 1976 763

Discussion SPE 6426
I would like to offer a few comments on Stein's paper,
"Mechanical Properties of Friable Sands From Conven-
tional Log Data" (July JPT, Pages 757-763).
1. The most important criticism bears on his Fig. 5
and its use in the calculation sequence. It is easily shown
that

and a quick check shows the slope of the straight-line

correlation to be associated with v == 0.3.
This is not surprising since Fig. 6 is based on the data
of Myung and Helander (and is similar to their Figs. 12
NOVEMBER. 1976 1303

and 13), which reveals a bias on close examination: the

mean value of Poisson's ratio for the 15 samples of their
Table 3 (three-dimensional velocity log data) is 0.3 with a
standard deviation of 0.03.
The use of Fig. 5 to determine Eb is, therefore, equiv-
alent to the arbitrary choice of Iv I0"=0 0.3.
The application of a correction for depth (Eq. 4) pro-
duces a shift in the straight line in Fig. 5 so that it does not
pass through the origin (whereas the uncorrected results Author's Reply SP E 6427
correlate perfectly with the theory). This would at least
show that such a correction is hazardous at shallow 1. I agree that the range of Poisson's ratio in the data
depths, if not that Eb is not independent of depth. included in Myung and Helander's Table 3 is limited,
2. The separation of moduli in Fig. 4 leading to the with the field-measured values ranging from 0.241 to
determination of the "cementation contribution" could 0.338. Most of their samples were taken at shallow
be discussed endlessly. What is very misleading, how- depths; however, some sample depths were more than
ever, is the use of names like "cementation strength," 5,000 ft. The Poisson-ratio values will include the influ-
• '5 trength plot," "sand strength," etc., applied to ence of stresses imposed at the depth of burial. Therefore,
moduli. a direct comparison of the Poisson ratios in Table 3
To my knowledge, there is no way of using moduli as a should not be made unless the data for all samples are
measure of strength; and no correlation has ever been corrected to the same stress levels. I have not as yet
found between elastic properties and failure properties. performed this work.
The main proposition in this paper is, therefore, simply The combined modulus values in Fig. 5 are already
an empirical method for determining in-situ effective corrected for depth effects. The bulk modulus values are
stress and therefrom correcting elastic properties for those listed in Myung and Helander's Table 3. Eq. 4 in
depth, with an assumption on the value at surface of my paper is directed toward removing friction effects
Poisson's ratio. provided by burial on combined modulus values for any
3. Some less important comments are in order about friable sand (see Fig. 4), not for correcting the existing
the modulus plot. Fig. 5. It is necessary to remove friction effects to obtain
It should be kept in mind that Fig. 2 has different scales a general correlation useful for determining the bulk
on the two axes, Ec being in 106 psi and a in 103 psi. modulus at the formation depth. Fig. 4 illustrates the
What appears to be the angle I/J has a real measure of 89.9° combined modulus behavior of the material at the sample
and it is rather spurious to speak of ,. angles" in abnormal depth.
coordinate systems. 2. Data tabulated by Wuerker* indicate a trend exists
Fig. 4 is even more misleading since both scales are for increased modulus values corresponding to increased
simply shown as pounds per square inch. strength values. Also, I have found that when the calcu-
The assumption that I/J applies to other friable sands is lated shear-modulus contribution provided by cementa-
obvious from its value. However, there is no more reason tion is above a minimum value, sand production will not
for tan I/J to be a constant than there is for tan cp (for the occur.
Mohr envelope) to be one. 3. The angle I/J was an average value determined for
Y. Kyvellos, SPE,AIME many friable sands in a rather large area. I believe it is
Compagnie Francaise des Petroles reasonable to apply this average value to other friable
Paris sands until more precise values can be derived.

1304 JOURNAL OF PETROLEUM TECHNOLOGY