Professional Documents
Culture Documents
CR97 06
CR97 06
Ice Strength as a
Function of Hydrostatic
Pressure and Temperature
Anatoly M. Fish and Yuri K. Zaretsky October 1997
τi = J 2
p 2
τ i = (c + bp) – (c + bp*) ––
CRREL REPORT
b1 M p*
τ max 1
τi, Shear Strength
φ1
φ2
τ max 2 T1
c1
c2
τ max 3 c φ3 T2
3
T3
Cover: Strength envelopes of ice at various temperatures: c, φ, p*, and T are the ice cohesion, friction angle, ice
melting pressure, and absolute temperature, respectively.
Ice Strength as a
Function of Hydrostatic
Pressure and Temperature
Anatoly M. Fish and Yuri K. Zaretsky October 1997
Prepared for
OFFICE OF THE CHIEF OF ENGINEERS
This report was prepared by Dr. Anatoly M. Fish, Research Civil Engineer, of the
Civil Engineering Research Division, Research and Engineering Directorate, U.S.
Army Cold Regions Research and Engineering Laboratory; and Dr. Yuri K. Zaretsky,
Director, Institute of Geomechanics and Hydrostructures, Moscow, Russia.
Funding of this research was provided by DA Project 4A762784AT42, Cold Regions
Engineering Technology, Work Package 201, Conventional Facilities in Cold Regions,
Work Unit CA-D13, Geotechnical Structures in Cold Regions.
The authors extend their appreciation to Dr. Devinder Sodhi, Dr. Patrick Black,
and Dr. Stephen Ketcham of CRREL for technical review of this report and their
valuable comments.
The contents of this report are not to be used for advertising or promotional
purposes. Citation of brand names does not constitute an official endorsement or
approval of the use of such commercial products.
ii
CONTENTS
Page
Preface ................................................................................................................... ii
Nomenclature ....................................................................................................... iv
Introduction .......................................................................................................... 1
Strain rate effect ................................................................................................... 3
Effect of hydrostatic pressure ............................................................................ 4
Temperature effect............................................................................................... 5
Ice melting pressure ....................................................................................... 5
Ice cohesion ...................................................................................................... 5
Angle of internal friction ............................................................................... 6
Temperature criteria of ice strength .................................................................. 6
Test data ............................................................................................................... 7
Parameter evaluation .......................................................................................... 8
Ice cohesion .......................................................................................................... 9
Angle of internal friction .................................................................................... 10
Strength predictions ............................................................................................ 11
Summary and conclusions ................................................................................. 13
Literature cited ..................................................................................................... 13
Abstract ............................................................................................................... 15
ILLUSTRATIONS
Figure
1. Strength envelopes of ice at various temperatures ............................. 2
2. Temperature diagrams of the strength parameters of ice ................. 6
3 Strength test data of ice under triaxial compression .......................... 7
4. Temperature dependencies of the strength parameters of ice .......... 8
5. Determination of parameters co and α .................................................. 9
6. Determination of parameters bo and β .................................................... 9
7. Angle of internal friction of ice as a function of temperature ........... 10
8. Strength of ice as a function of temperature and confining
pressure ................................................................................................ 11
9. Predicted and test magnitudes of ice strength under triaxial
compression at various temperatures .............................................. 12
10. Strength of ice in the range of high hydrostatic pressures at
ε̇ ≈ 5 × 10 −3 s −1 ....................................................................................... 12
TABLES
Table
1. Strength parameters of ice at various temperatures ........................... 8
2. Strength predictions of Labrador iceberg ice ...................................... 11
iii
NOMENCLATURE
iv
Ice Strength as a Function of
Hydrostatic Pressure and Temperature
ANATOLY M. FISH AND YURI K. ZARETSKY
τi = J2
p 2
τ i = (c + bp) – (c + bp*) ––
p*
b1 M
τ max 1
τ max = c + —
b σ
max
2
τ i , Shear Stress
φ1
φ2 p*2 c –1
τ max 2 T1 σmax = ––– — + p*
c1 2 b
c2
τ max 3 c φ3 T2 p* = p*(T); c = c(T)
3 b = b(T)
b(T) = tan φ(T)
T3
2
proportional to temperature. It has been shown the hydrostatic pressure is, the greater the viscos-
that the hydrostatic pressure being applied to ice ity coefficient, the lower the strain rate of ice, and
decreases its melting temperature and thus its the higher the ice strength will be. For p > σmax the
strength under triaxial compression. higher the hydrostatic pressure is, the smaller the
In the present report a different approach has viscosity coefficient, the higher the strain rate,
been undertaken. The authors considered the ice and the lower the ice strength will be. This is
strength as a function of two variables: the hydro- confirmed well by test data (Jones 1982, Jones and
static pressure and temperature, i.e., Chew 1983).
Equation 2 can be rewritten in terms of maxi-
τ *i = τ *i ( p , T ) ; γ˙ i = Const. (1d) mum (peak) shear strength τ *i = τ i and presented
as a product of two independent functions: a yield
At a constant strain rate the strength of ice in a function τio(p) and a nondimensional function
multiaxial stress state is described by the para- Φ ( γ˙ i ) of the constant strain rate
bolic yield criterion with three parameters: the
cohesion, and the friction angle, which are differ- τ *i ( p , γ˙ i ) = τ io ( p) Φ ( γ˙ i ) . (2a)
ent nonlinear functions of temperature, and the
ice melting pressure. Then a strength criterion of Function Φ ( γ˙ i ) has been selected in its simplest
ice, which takes into account the combined effect form:
of all three variables in eq 1, the strain rate, hy-
drostatic pressure and temperature, is obtained γ˙ t 1/ n γ˙ i 1/ n
by combining eq 1b and 1d. Φ ( γ˙ i ) = i o = (3)
γ io γ˙ io
3
[(σ − σ )
and Gammon 1995) or remains unchanged (Jones 1
+ (σ 2 − σ 3 )
2 2
1982). Apparently, the magnitude of the instanta- τi = 1 2
6
neous strain rate depends on the type of ice, its
structure and other factors and varies between
10–1 and 10–3 s–1. Since the magnitude of ice
+ (σ 1 − σ 3 ) ]
2 1/ 2
strength is greatly affected by the strain rate, se- = octahedral shear stress (resultant)
lection of an adequate value of this referenced σmax = magnitude of the hydrostatic pres-
strain rate is extremely important for an accurate sure at which the shear strength
prediction of the ice strength at lower strain rates. reaches a maximum
In experimental studies of ice under triaxial b
(biaxial) compression, the radial strains are as- τ max = c + σ max (6)
2
sumed to be small and are usually ignored. In this
case eq 3 can be replaced by p *2
σ max =
c (7)
2 + p *
b
ε˙ t 1/ n ε˙ 1/ n
Φ ( γ˙ i ) = Φ (ε˙ ) = o = (3a) p* = ice melting pressure at which the
εo ε˙ o
shear strength of ice equals to zero.
where ε̇ = axial strain rate Subscript (o) at τio in eq 5 indicates that param-
ε˙ o , ε o = instantaneous (referenced) axial strain eters c and b are referred to the instantaneous
rate and strain respectively. condition when the applied strain rate is equal to
to = ε o /ε˙ o = given by eq 4. the instantaneous strain rate.
Note that the yield curve may also intersect the
Thus, the strain rate function varies in the limits hydrostatic axis in the domain of the negative
hydrostatic pressures (–p) at point h2 (not shown
1 ≥ Φ ( γ˙ i ) = Φ (ε˙ ) ≥ 0. in Fig. 1), the abscissa of which is equal to
The effect of the strain rate on the strength of ice −p *
h2 = .
can be excluded from consideration by selecting b (8)
1+ c p *
the strength test data of ice corresponding to
Φ (ε˙ ) = 1 . Equation 5 can be considered an extended Von
Mises–Drucker-Prager yield criterion. At low
stress level p << σmax the third term in the right
EFFECT OF HYDROSTATIC PRESSURE side of eq 5 approaches zero and eq 5 transforms
into the Drucker-Prager (1952) (extended Von
When applied strain rate is equal to the instan- Mises) yield criterion:
taneous strain rate γ˙ i = γ˙ io or ε˙ = ε˙ o , in eq 2a and
3a, Φ ( γ˙ io ) = Φ (ε˙ o ) = 1, and the strength depen- τio = c + bp . (9)
dency of ice upon the hydrostatic pressure is de-
For frictionless materials (b = 0) eq 5 reduces to
scribed by the parabolic yield criterion (Fish 1991)
the Von Mises yield criterion
depicted in Figure 1,
τio = c . (10)
b
τ io ( p) = c + bp − p2 (5)
2σ max Thus, in a multiaxial stress state, the strength
of ice as well as its strength characteristics are
or
functions of only three parameters: p*, c and b
p 2 which all have a definite physical meaning and are
τ io ( p) = (c + bp) − (c + bp*) (5a)
p * easily determined from test data. Studies show
that all these parameters are functions of tem-
where c = ice cohesion on the octahedral plane perature, i.e.,
b = tan φ, where φ is the angle of internal
friction of ice on the octahedral plane p* = p * (T ), c = c(T ), b = b(T ) .
p = (σ 1 + σ 2 + σ 3 ) / 3 = hydrostatic pres- Consequently, the shear strength of ice as well as
sure (mean normal stress) the strength characteristics of ice are also func-
σ 1 , σ 2 , σ 3 = principal stresses, tions of temperature:
4
τ io = τ io (T ), τ max = τ max (T ), drostatic pressure p* at which the shear resistance
of ice equals zero, i.e.,
σ max = σ max (T ), h2 = h2 (T ).
Tm + θ = Tm − Ap * (13)
Thus, if a series of strength tests of ice is carried
out at a constant strain rate instead of one yield and
curve, one obtains a family of curves for various −θ
temperatures (Fig. 1). It should be emphasized p* = (14)
A
that parameters c, b and τmax are also strain-rate
dependent, while parameters σmax, p* and h2 are where θ = ice temperature (°C) and Tm = 273.1 K
independent of the strain rate (Fish 1992, 1993). is the ice melting temperature at the atmospheric
pressure. Note that the ice melting pressure is
unrelated to the grain size or the structure of ice.
TEMPERATURE EFFECT The magnitudes of ice melting pressures p* for
various temperatures are presented in Table 1.
Ice melting pressure
It is well known that the ice melting tempera- Ice cohesion
ture, as well as the melting temperature of other The ice cohesion defines the ice strength when
crystalline materials, is a function of the hydro- the hydrostatic pressure p = 0. Ice strength is a
static pressure. This pressure can be determined stochastic event, a culmination point of failure of
(Zaretsky and Fish 1996a) from the Clapeyron intermolecular bonds and growth of cracks. A cer-
equation, according to which a small change in tain number of these bonds in the unit volume of
the equilibrious melting temperature of a solid ice are formed during freezing of water at tem-
∆Tm attributable to a small change in the hydro- perature 0°C. Further temperature decrease brings
static pressure ∆p can be calculated from the rela- about formation of new bonds, attributable to
tionship freezing of the liquid phase on grain boundaries,
and a sharp increase of the ice strength (cohe-
V − Vs V − Vs sion). Thus, one may conclude that the ice cohe-
dTm = l dp = Tm l dp (11)
Sl − Ss Lm sion is a function of temperature and consists of a
sum of two components:
whereVl and Vs = unit mass volume of the ma-
terial in the liquid and in the c (T ) = co + c1 (T ) (15)
solid state, respectively
where co is the component of the ice cohesion
Sl and Ss = unit mass entropy of the liq-
brought about at the time of ice formation at 0°C
uid and of the solid state, re-
and c1(T) is the temperature-dependent compo-
spectively
nent of the cohesion brought about by freezing of
Lm = specific heat of melting of the
liquid phase at temperature below 0°C.
unit mass.
Since the physical nature of co and c1 is the
Since at melting the volume Vl < Vs , dTm < 0 , i.e., same, and taking into account that at temperature
T = Tm, c1 = 0, the temperature dependency of
the equilibrious melting temperature of ice Tm
component c1 can be presented in the form
decreases as well,
c 1 (T ) = co α (1 − T / Tm ) (16)
dTm = − Adp . (12)
Thus, for ice when p = 0, Tm = 273.1 K, Vl = 10–3 where α is a parameter. Combining eq 15 and 16
m3 kg–1, Vs = 1.09 × 10–3 m3 kg–1, Lm = 3.336 × 105
J kg–1, and parameter A for ice at temperature c (T ) = co [1 + α (1 − T / Tm )] . (17)
θ = 0°C is equal to
Equation 17 establishes a linear dependency of
A = 0.074 K/MPa. the ice cohesion upon temperature. Test data show
that such a relationship is valid in the domain of
The magnitude of parameter A coincides with relatively low temperatures below –20°C. In the
that calculated by Barnes et al. (1971). It is not range of moderate temperature, this relation-
difficult at this point to calculate the critical hy- ship becomes nonlinear and somewhat better
5
b (T ) = bo + b1 (T ) (19)
c* a.
where bo = tan φo and φo is the angle of internal
friction of ice brought about at the time of ice
c, Cohesion formation at 0°C; b1(T) = tan φ1 (T), and φ1 (T) is
µ
α(1 – T/T ) the angle of internal friction attributable to the
c(T) = c 0 e m
b*
b. [
b (T ) = bo 1 + β (1 − T / Tm ) . ] (21)
b, Friction Parameter
b (T ) = bo e β (1− T / Tm )
ν
(22)
or
b0 ν
b (θ) = bo e β 1 θ (22a)
T = 0 ( −273.1°C) ; b* = bo e β
described by
µ The temperature diagram of parameter b(T) is
c (T ) = co e α(1− T / Tm ) (18)
given in Figure 2b. Note that at very low tem-
or
peratures certain adjustments of eq 15 through 22
c (θ) = co e α|θ|/ Tm (18a) will probably be required.
6
23 should be combined with eq 2a through 4. TEST DATA
Then eq 1 can be written as
The validity of the temperature dependencies
τ *i ( p , ε˙ , T ) = τ io ( p , T ) Φ (ε˙ ) of the strength parameters of ice presented above
1/ n was verified using test data of Gagnon and
b(T ) ε˙ (24) Gammon (1995). The triaxial (σ2 = σ3) compres-
= c(T ) + b(T ) p − p2
2σ max (T ) ε˙ o sion tests were carried out using cylindrical speci-
mens of Labrador iceberg ice, 9.58 cm in diameter
In the range of the hydrostatic pressures p << and 26 cm long. Mean grain diameter d ≈ 8.1 mm.
σmax, eq 24 takes the form The test temperature varied between –1°C and
1/ n
–16°C. The tests were conducted at constant axial
ε˙ strain rates between 10–2 and 10–5 s–1. Test data
τ *i ( p , ε˙ , T ) = [c (T ) + b (T )p] (25)
ε˙ o corresponding to the strain rate ε̇ o ≈ 5 × 10 −3 s −1
were selected in the following analysis to evalu-
when p = 0 (pure shear) ate the strength parameters of ice. This strain rate
1/ n was defined by the authors of this report as the
ε˙ instantaneous strain rate because the test strength
τ *i (ε˙ , T ) = c (ε˙ , T ) = c (T ) (26)
ε˙ o magnitudes at this rate reached their maximum
Accordingly values. The tests were carried out at four different
confining pressures: 1.38, 3.45, 6.89 and 13.79 MPa.
1/ n The test data replotted by the authors of this re-
ε˙
b (ε˙ , T ) = b(T )
ε˙ o
(25a) port in terms of the shear strength τ i = (σ 1 − σ 3 ) / 3
vs. the hydrostatic pressure p = (σ 1 + σ 2 + σ 3 ) / 3
where parameters c(T) and b(T) are correspond- are presented in Figure 3. Note that each point in
ing to the strain rate ε˙ = ε˙ o . Figure 3 represents the mean of five tests.
If the radial strain rates are taken into account In Figure 3 for comparison tests, data are pre-
function Φ(ε˙ ) in eq 24, 25 and 26 should be re- sented on triaxial compression of freshwater poly-
placed by function Φ( γ˙ ) . crystalline ice obtained by Jones (1982). The ran-
18
σσ21 == σ33
.
ε1=≈55××1010–3–s3–1s – 1
Strength (MPa)
(MPa)
–6 Polycrystalline Ice
(data from Jones,
σ11 ––σσ33)/)/√3,
1982)
–1
σ3 Trajectory
6
= ((σ
–
√3
ττii =
0 15 30 45
p, Hydrostatic Pressure (MPa)
Figure 3. Strength test data of ice under triaxial compression. Data from Gagnon
and Gammon (1995) and Jones (1982).
7
domly oriented laboratory-made ice Table 1. Strength parameters of ice at various temperatures*.
samples were 20 mm in diameter and 60
θ°C c (MPa) b ~ φo σmax (MPa) τmax (MPa) p* (MPa)
mm long with mean grain diameter less
than 1 mm. The test data selected for –40 12.93† 0.241 14 245.87 43.52 540.54
comparison refer to the axial constant –16 9.43 0.125 7 80.15 14.44 216.22
–11.8 8.14 0.106 6 53.81 10.99 159.46
strain rate ε˙ = 5.4 × 10 −3 s −1 . The test tem-
–11 7.91 0.103 6 49.0 10.43 148.65
perature was –11.8 ± 0.9°C. The confin- –6 6.63 0.08 4°30′ 20.05 7.43 81.08
ing pressure (σ2 = σ3) varied between ~5 –1 5.56 0.053 3 (4.79)** (5.69)** (36.85)**
and ~ 85 MPa. 0 5.37†† 0.04†† 2††
In Figure 10 (below), results of triaxial * Data in this table for strain rate ε˙ ≈ 5 × 10 −3 s −1 .
constant strain rate compression tests † Calculated by eq 17.
α = 9.61 ; β = 4.69.
PARAMETER EVALUATION
Temperature variations of ice strength in a mul- Using these parameters, we calculated the magni-
tiaxial stress state were shown to be defined by tudes of the ice cohesion c(θ), the friction param-
three temperature-dependent parameters: the ice eter b(θ) and the friction angle φ(θ) by eq 18a and
melting pressure p(T), ice cohesion c(T) and fric- 22a (Table 1). The magnitudes of σmax(θ) and
tion parameter b(T). The magnitudes of
the ice melting pressure p(T) were cal- 16 16 × 10 –2
culated by eq 14 for the test tempera-
tures and are included in Table 1. The
b (θ)
magnitudes of p* for θ = –11.8°C was
found to be in agreement with p* = 123.7
MPa obtained earlier (Fish 1991) from 12 12
the analysis of Jones’ data and with that
calculated by an empirical equation of c (θ)
b, Friction Parameter
c, Cohesion (MPa)
8
10 al. (1991) reported that for θ = –10°C and for the
sliding velocity of 10–3 m s–1, µc ≈ 0.13. In Table 1
we found that for –10°C, b = 0.098 that is in agree-
8 ment with the test data.
Comparing parameter bo with the kinetic dry
α (Tm – T)/Tm friction coefficient µd of ice on ice is particularly
c = c 0e
6 interesting. Since the latter is temperature-depen-
dent, the comparison should be made for a low
ln c
ICE COHESION
ln b
9
sis of ice strength was carried out using test data range between –1° and –16°C. These data were
of Rist and Murrell (1994). This polycrystalline ice extrapolated to calculate the friction angle φ and
had a grain size of 1.7 mm. The comparison was the ice strength for temperature –40°C (Fig. 7) and
performed for temperature –20°C and the strain compared (see Fig. 10 below) with test data of
rate range between ~10–3 s–1 and ~ 10–5 s–1. The Rist and Murrell (1994) obtained for this tempera-
results of the analysis confirmed the above con- ture and for the grain size of 1.7 mm.
clusion. At the same time our studies indicate One can see that despite the considerable dif-
that the ice cohesion magnitude is strongly de- ference (almost five times) in the grain sizes of
pendent on the ice structure. these two types of ice and the difference in the
test temperatures, the predicted angle of internal
friction correlates well with the test data. We find
ANGLE OF INTERNAL FRICTION the friction angle magnitudes are unaffected by
variations of the grain size of ice.
From the above studies an important conclu- Thus, eq 27 can be used for calculations of
sion can also be made regarding the physical na- parameter b(T) in the above equations and for
ture of the angle of internal friction of ice. While prediction of the ice strength over a wide spec-
the temperature dependency of the friction pa- trum of temperatures between 0° and –40°C. It
rameter b(θ) = tan φ(θ) is nonlinear (Fig. 4), the should be remembered that the ice friction angle
test data from Table 1 plotted in Figure 7 suggest is a function of the strain rate as well. It decreases
that the friction angle is a simplest linear function rapidly with the strain rate decrease. Fish (1991,
of temperature: 1993) showed that at low strain rates ~10-7 s–1 and
below, ice at –10°C can be considered as an ide-
φ(θ) = φo + ω|θ| (27) ally cohesive (φ = 0) material, the strength of which
is defined by eq 10 and 26. At high temperatures
where φo = 2°50′, ω = 16.4′ degree–1 andθ= (Tm the internal friction angle is small and so is its
– T ) is ice temperature (°C). It should be remem- effect on the ice strength. However, at low tem-
bered that the magnitudes of the parameters in eq peratures or at high strain rates and high confin-
27 refer to a strain rate ε˙ ~ 10 −3 s −1 . ing pressures, the effect of the internal friction
Note that the magnitudes of the internal fric- angle on the strength of ice can be considerable.
tion angles φ in Table 1 and Figure 7 were calcu- Since parameters c, b and p* were found to be
lated based mainly upon test data on the Labra- unaffected by variations of the grain size. Table 1
dor iceberg ice (d = 8.1 mm) at the temperature data can be used to predict strength of various
24°
.
Data from ε ≈ 10 –3 s –1
Gagnon and Gammon (1995)
Jones (1982)
φ , Angle of Internal Friction
lated
apo
Extr
ω
8°
φο
10
types of iceberg ice and freshwater randomly ori- Table 2. Strength predictions of Labrador iceberg ice.
ented polycrystalline ice in a multiaxial stress state.
Confining Strain
θ pressure rate ε̇ σ1 − σ3 (MPa)
(°C) (MPa) (s–1) Measured Predicted Ice type
STRENGTH PREDICTIONS
–11 1.38 4.8210–5 3.50 4.36 Labrador ice
–11 6.89 5.96 10–5 4.75 4.93 Labrador ice
Let us verify the accuracy of strength predic-
tion by comparing the computed and the test re-
sults of ice strength as functions of temperature,
strain rate and hydrostatic pressure. ε˙ 1/ n
σ 1 − σ 3 = 3 (c + bp) . (25a)
For low hydrostatic pressures p << σmax and ε˙ o
the strain rate ε˙ = ε˙ o , ice strength can be calcu-
lated by eq 9 using data of Table 1 and the hydro- In Table 2 the results of such calculations are pre-
static pressure magnitudes p from Figure 3, corre- sented for n = 4, ε̇ o = 5 × 10 −3 s −1 , two different
sponding to confining pressures σ3 = 1.38 MPa confining pressures (assuming that p = σ3), and
and 6.89 MPa. The results of such calculations for two different strain rates together with the test
the temperature range between –1° and –16°C are results from Table 2 of Gagnon and Gammon
presented in Figure 8 together with the test data (1995).
from Gagnon and Gammon (1995). One can con- From Table 2 it follows that predicted values of
clude that the predicted ice strength values are in ice strength practically coincide with the test data.
agreement with the test data. Unfortunately only two tests of this ice type were
The data in Table 1 can also be used to evaluate available for comparison.
the ice strength in those cases when the strain rate In Figure 9 predicted dependencies of the ice
ε˙ < ε˙ o . For p << σmax the ice strength is calculated strength are presented calculated by eq 5 using
by eq 25. To simplify the comparison with the test values of hydrostatic pressures and the
published data, the latter can be rewritten in terms strength parameters from Table 1. One can see
of deviator stress: that in the moderate temperature range between
–6° and –16°C the calculated values of the ice
16 strength are in good agreement with test data
of Gagnon and Gammon (1995). An exception
Labrador Iceberg Ice is the test data for –1°C for which the magni-
———————————
14
Data from Gagnon and Gammon (1995)
tude of the ice melting pressure p* = 13.51
MPa calculated by eq 14 turned out to be
(MPa)
Shear Strength (MPa)
11
18
σ21 = σ3
.
ε =≈ 5 × 10–3
– 3ss–1– 1
(MPa)
Strength(MPa)
– 16°C Labrador Ice
(data from
Shear Strength
12 Gagnon and
Gammon, 1995)
1995)
– 11.8
/ 3,Shear
– 11 Polycrystalline Ice
(data from Jones,
1982)
σ33))/√3,
–6
Predicted
(σ11 –– σ
6
i =(σ
τiτ=
–1
0 15 30 45
p, Hydrostatic Pressure (MPa)
Figure 9. Predicted and test magnitudes of ice strength under triaxial compres-
sion at various temperatures. Data from Gagnon and Gammon (1995), and Jones
(1982).
24
Polycrystalline Ice
18 – 40°C – 11 .8 Jones (1982)
ShearStrength
– 16
– 11 Gagnon and
Labrador Ice Gammon (1995)
–6
3,Shear
–1
12
Predicted
σ11––σσ33)/)/√3,
τi == ((σ
6 – 16°C
– 11
–6
p* = 81.08 p* = 148.65 p* = 216.22 MPa
–1
0 100 200 300
p , Hydrostatic Pressure (MPa )
12
eral trend of predicted values of ice strength as Fish, A.M. (1992) Three-dimensional visco-plastic
functions of the hydrostatic pressure is well con- flow model of polycrystalline ice. In Proceed-
firmed by Jones’ test data. ings, 3rd International Conference on Ice Technol-
ogy, Massachusetts Institute of Technology, Cam-
bridge, Massachusetts, 11–13 August (T.K.S. Murphy,
SUMMARY AND CONCLUSIONS W.M. Sackinger, and P. Wadhams, Eds.) South-
ampton, U.K.: Computational Mechanics Publ.,
1. A temperature model has been developed p. 193–207.
that describes the ice strength in a multiaxial stress Fish, A.M. (1993) Combined creep and yield model
state over a wide spectrum of negative (subfreez- of ice under multiaxial stress. Journal of Offshore
ing) temperatures. and Polar Engineering, 12(3): 130–139.
2. The strength dependency of ice on the hy- Gagnon, R.E., and P.H. Gammon (1995) Triaxial
drostatic pressure and temperature is well de- experiments on iceberg and glacier ice. Journal of
scribed by eq 5a, which is a parabolic yield crite- Glaciology, 41(139): 538–540.
rion: Hallam, S., and J.P. Nadreau (1988) Failure maps
2 for ice. In Proceedings, 9th International Conference
p
τ io ( p , T ) = (c + bp) − (c + bp*) (5a) on Port and Ocean Engineering under Arctic Condi-
p * tions (POAC ‘87) 17–22 August, Fairbanks, Alaska
with only three fundamental physically well- (W.M. Sackinger, and M.O. Jeffries, Eds.) Geophysi-
founded parameters: the ice cohesion c(T), angle cal Institute of Alaska–Fairbanks, vol. 3, p. 45–55.
of internal friction φ(T) or b(T), and ice melting Hausler, F.U. (1983) Comparison between differ-
pressure p*(T). They all have a definite physical ent yield functions for saline ice. Annals of Glaciol-
meaning and are functions of temperature. ogy, 4: 105–109.
3. The temperature model was developed Jones, D.E., F.E. Kennedy, and E.M. Schulson
based upon data on over 100 triaxial compression (1991) The kinetic friction of saline ice against
tests of iceberg ice and laboratory-made fresh- itself at low sliding velocities. Annals of Glaciology,
water randomly oriented polycrystalline ice with 15: 242–246.
a grain diameter between ~ 1 mm and ~ 8 mm in the Jones, S.J. (1978) Triaxial testing of polycrystal-
temperature range between –1° and –40°C at the line ice. In Proceedings, 3rd International Conference
range of strain rates between 10–3 s–1 and 10–5 s–1. on Permafrost, Edmonton, Alberta, Canada. vol. 1,
4. Further studies are required of the tempera- p. 670–674.
ture dependencies of the ice strength parameters Jones, S.J. (1982) The confined compressive
for various types of ice, particularly the ice melt- strength of polycrystalline ice. Journal of Glaciol-
ing pressure over a wider spectrum of tempera- ogy, 28(98): 171–177.
tures, strain rates, and confining pressures. Jones, S.J., and H.A.M. Chew (1983) Creep of ice
as a function of hydrostatic pressure. Journal of
Physical Chemistry, 87(21): 4064–4066.
LITERATURE CITED Nadreau, J.P., and B. Michel (1986) Yield and
failure envelope for ice under multiaxial com-
Barnes, P.D., F.R.S. Tabor, and J.C.F. Walker (1971) pressive stresses. Cold Regions Science and Technol-
The friction and creep of polycrystalline ice. ogy, 12(1): 75–82.
Proceedings, Royal Society of London, Ser. A. 347: Nadreau, J.P., A.M. Nawwar, and Y.S. Wang (1991)
493–512. Triaxial testing of freshwater ice at low confining
Casassa, C., H. Narita, and N. Maeno (1991) Shear pressures. Transactions of ASME, 113: 260–265.
cell experiments of snow and ice friction. Journal Reinicke, K.M., and T.D. Ralston (1977) Plastic
of Applied Physics, 69(6): 3745–3756. limit analysis with an anisotropic, parabolic yield
Drucker, D.C., and W. Prager (1952) Soil mechan- function. International Journal of Rock Mechanics
ics and plastic analysis of limit design. Quarterly and Mining Sciences, 14(3): 147–154.
Journal of Applied Mechanics, 10: 157–165. Richter-Menge, J.A., G.F.N. Cox, N. Perron, G.
Fish, A.M. (1991) Creep and yield model of ice Durell, and H.W. Bosworth (1986) Triaxial test-
under combined stress. USA Cold Regions Re- ing of first-year sea ice. USA Cold Regions Re-
search and Engineering Laboratory, Special Re- search and Engineering Laboratory, CRREL Re-
port 91-31. port 86-16.
13
Rist, M.A., and S.A.F. Murrell (1994) Ice triaxial Weiss, J., and E.M. Schulson (1995) The failure
deformation and fracture. Journal of Glaciology, of freshwater granular ice under compressive
40(135): 305–318. loading. Acta Metallurgica Material, 43(6): 2303–
Sayles, F.H. (1974) Triaxial constant strain rate 2315.
tests and triaxial creep tests on frozen Ottawa Zaretsky, Yu.K., and A.M. Fish (1996a) Model
sand. USA Cold Regions Research and Engi- of viscoplastic deformation of frozen and un-
neering Laboratory, Technical Report 23. frozen soils and ice. In Proceedings, 6th Interna-
Smith, M.B. (1974) A parabolic yield condition tional Offshore and Polar Engineering Conference,
for anisotropic rocks and soils. Ph.D. thesis, Rice 26–31 May 1996, Los Angeles, vol. 2, p. 291–296.
University, Houston, Texas. Zaretsky, Yu.K., and A.M. Fish (1996b) Effect
Timco, G.W ., and R.M.W . Frederking (1986) of temperature on the strength and viscosity of
Confined compression tests: Outlining the fail- ice. Journal of Soil Mechanics and Foundation En-
ure envelope of columnar sea ice. Cold Regions gineering (translated from Russian), vol. 33(2):
Science and Technology, 12: 13–28. 46–52.
14
Form Approved
REPORT DOCUMENTATION PAGE OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestion for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington,
VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
October 1997
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
A temperature model has been developed that describes the ice strength in a multiaxial stress state over a wide
spectrum of negative temperatures. The model takes into account the anomalous behavior of ice under high
hydrostatic pressure, when its strength reaches a maximum, and then gradually decreases with the pressure
increase. It has been shown that strength of ice under high hydrostatic pressure is described by a parabolic yield
criterion with only three fundamental parameters, ice cohesion, internal friction angle, and ice melting pressure,
which all have a definite physical meaning and are functions of temperature. The model has been verified using
test data on the strength of iceberg ice and laboratory-made polycrystalline freshwater ice under triaxial
compression at strain rates between 10–3 and 10–5 s–1 over the temperature range between –1° and –40°C.