Incorporating the Effect of Moisture Variation on Resilient Modulus

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Incorporating the Effect of Moisture Variation on Resilient Modulus for Unsaturated
Fine-Grained Subgrade Soils
36
Submitted to:
37
94th Transportation Research Board Annual Meeting
38
January 2015
39
Washington, D.C.
40
41
42
43
44
45
Murad Y. Abu-Farsakh, Ph.D., P.E. (Corresponding Author)
Research Professor
Louisiana Transportation Research Center
Louisiana State University
4101 Gourrier Avenue
Baton Rouge, LA 70808
E-mail: cefars@lsu.edu
Ayan Mehrotra
Former MS Student – Louisiana Transportation Research Center
Project Manager
Professional Service Industries, Inc.
Mandeville, LA 70471
E-mail: ayan.mehrotra@psiusa.com
Louay Mohammad
Professor
Louisiana Transportation Research Center
Louisiana State University
4101 Gourrier Avenue
Baton Rouge, LA 70808
And
Kevin Gaspard
Senior Pavement Engineer
Louisiana Department of Transportation and Development
Baton Rouge, LA 70808
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
Incorporating the Effect of Moisture Variation on Resilient Modulus for Unsaturated
Fine-Grained Subgrade Soils
ABSTRACT
Moisture content and the state of stress, which for unsaturated soil includes the effects of
matric suction, exert a significant effect on the resilient modulus (Mr) values of fine-grained
subgrade soils. Due to seasonal variation, the moisture content and consequently, matric
suction, vary periodically in subgrades. This study aims at investigating the relationship
between Mr and matric suction through conducting repeated load triaxial testing on finegrained soils to obtain Mr values and evaluating the Soil Water Characteristic Curves
(SWCC), which provides a relationship between suction and degree of saturation. The SWCC
curves were evaluated utilizing a combination of two techniques: axis-translation and chilled
mirror hygrometer, which allows for representation of the SWCC across the entire range of
saturation. It was found that the PI has a significant impact on the matric suction-water
content relationship such that the SWCC shifts to the left as the PI value of the soil decreases.
The test results indicate a significant relationship between Mr and matric suction. The results
were also used to analyze the relationship between Mr and moisture variation in terms of
gravimetric water content and degree of saturation, which showed that the degree of
saturation is superior to gravimetric water content in its ability to capture the effect of
moisture variation on Mr values. A modified constitutive Mr-matric suction model is
proposed with the ability to capture the effect of moisture variation on Mr while taking into
account the stress state of unsaturated soils. The proposed model demonstrated better
performance than the existing models in terms of reducing the number of regression constants
and providing a better fit to the measured Mr data.
71
KEYWORDS: resilient modulus, matric suction, unsaturated soil, moisture variation, subgrade
72
soil.
73
INTRODUCTION
74
75
76
77
78
79
80
81
82
83
84
85
86
87
Premature pavement failure is often associated with loss of support in the subgrade layer,
especially in regions with soft subgrade soils. Resilient modulus (Mr) quantifies the support
provided by different pavement layers including the subgrade layer. The Mr, which is a
measure of the stiffness of a material, is considered a fundamental property for pavement
design. It serves as a key input parameter for different pavement layers in the Mechanistic
Empirical Pavement Design Guide (1). It is up to the design engineer to select an appropriate
Mr value representative of the individual pavement layers. For the subgrade layer, the
selection of Mr presents a unique challenge since the subgrade lies within the Active Zone
that usually experiences periodic changes in moisture conditions. Previous studies (2) have
shown that the water content beneath the pavement subgrades is expected to vary accordingly
due to seasonal variations. Studies have also shown that the Mr values are significantly
impacted by changes in the moisture conditions of soils (3) (4). It is evident that the effect of
moisture fluctuation must be taken into account when selecting an appropriate design value
of Mr for subgrade soils.
88
89
90
91
92
Mr is a measure of the elastic behavior of geomaterials under cyclic loading, similar to
what is experienced by a pavement layer under vehicular loading. Its value is dependent upon
several factors such as stress state, density, soil type, and water content. The MEPDG
adopted the so called ‘Universal’ model, displayed in equation 1, proposed by Witczak and
Uzan (5), to evaluate Mr for geomaterials by taking into account the bulk and shear stresses.
1
𝜃 𝑘2
𝑀𝑟 = 𝑘1 𝑃𝑎 (𝑃 )
93
𝑎
𝑘3
𝜏
( 𝑃𝑜𝑐𝑡 + 1)
𝑎
(1)
94
where: Pa = atmospheric pressure; θ = bulk stress = σ1 + σ2 + σ3; τoct = octahedral shear stress
95
=
√2
3
(𝜎1 − 𝜎3 ) when σ2 = σ3; and k1, k2, k3 = model regression constants.
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
The selection of Mr design value must account for the effect of moisture content along
with the effect of stress state. The moisture content can be a difficult variable to account for
due to its tendency to fluctuate periodically. For fine-grained subgrade soils, it is essential to
account for the effect of moisture variation on Mr values since a slight increase in the
moisture content can have a significant negative impact on the Mr value (3) (6). The MEPDG
utilizes the Enhanced Integrated Climatic Model (EICM) to account for the changes in the Mr
value due to moisture fluctuation by incorporating an adjustment factor based on the variation
of degree of saturation (S), and applying it to the Mr value determined by using equation 1.
However, in many cases, the subgrade soil layers can be under an unsaturated state condition
due to their shallow depths. In order to accurately describe the stress state behavior in
general, and the resilient modulus in specific for unsaturated soils, the effect of matric suction
must be incorporated in the design models. Several Mr - stress state constitutive models have
been proposed in the literature that incorporate the effect of matric suction when evaluating
Mr value for unsaturated subgrades (7) (8) (9) (10). Additionally, previous studies (11) (12)
have also shown the impact of matric suction, and consequently moisture variation, on the
shear modulus, which is a dynamic stiffness property similar to the Mr albeit at much smaller
strains than the Mr, of fine-grained soils.
113
114
115
116
117
118
119
120
121
122
123
124
125
The objective of this paper is to evaluate the existing Mr constitutive models and
develop/modify a model for incorporating the effect of moisture variation in estimating Mr
values for unsaturated subgrade soils. A comprehensive laboratory testing program was
performed to evaluate Mr values by performing Repeated Load Triaxial (RLT) tests on four
selected fine-grained soil types with varying plasticity indices (PI) to represent the range of
subgrade soils commonly found in the state of Louisiana. The tested soil specimens were
prepared at different moisture contents. The Soil Water Characteristic Curves (SWCC) were
also evaluated for the four soil types. Methods utilizing gravimetric water content, degree of
saturation (S), and matric suction to evaluate changes in Mr values due to moisture
fluctuation were analyzed. Based on the laboratory test results and external data, a modified
Mr-matric suction constitutive relationship, which incorporates the variation in matric suction
due to variation in moisture content for evaluating Mr of unsaturated subgrades soils, was
proposed and validated.
126
LABORATORY TESTING PROGRAM AND RESULTS
127
Resilient Modulus (Mr)
128
129
130
131
132
133
134
135
136
137
The selected physical properties for the four soil types tested in this study are provided in
Table 1. To evaluate the impact of moisture variation on Mr, the tested soil specimens were
prepared at different compacted moisture contents utilizing the Standard proctor effort. The
moisture contents of the tested specimens varied from OMC -6% to OMC+6% for A-7 soils
and from OMC -3% to OMC+3% for the other three soil types, where OMC represents the
optimum moisture content. Triplicate Mr specimens were prepared and tested at each target
moisture content in accordance with the AASHTO T-307-99 procedure. This implies testing
the subgrade Mr specimens at three confining pressures with five different deviatoric stresses
at each confining pressure, which leads to Mr values being evaluated at 15 stress states per
test. RLT tests were performed using the Material Testing System MTS 810 device with a
2
138
139
140
141
142
143
144
145
146
147
closed loop servo hydraulic loading system. Figure 1 displays the effect of compacted
moisture content on Mr values in terms of normalized water content (w/wopt) and normalized
Mr value (Mr/Mropt), where w is the compacted moisture content, wopt is the OMC, and Mropt
is the Mr at OMC. The decrease in Mr values with the increase in moisture content can be
attributed to weakening of the soil fabric as the moisture content increases. The Mr values in
Figure 1 and generally throughout this paper are presented at a bulk stress of 155 kPa and an
octahedral shear stress of 13 kPa, which is the recommended stress state for highway
subgrade soils per Strategic Highway Research Program Protocol (SHRP) P-46 (3).
TABLE 1 Selected Physical Properties and Classifications for the Soil Types Tested
Soil
Type
P-7
P-17
P-26
P-53
148
149
150
151
152
153
154
155
156
157
158
159
160
161
%
Plasticity
Passing
Index
No. 200
(PI)
Sieve
7
68.9
17
43.8
26
95.4
53
95.7
%
Silt
%
Clay
52
63
61
13
13
18
35
84
MDD* OMC*
AASHTO USCS
(N/m3) (%)
17
17.3
15.8
12.3
17
16
22
35
A-4
A-6
A-7-6
A-7-6
ML
CL
CL
CH
*Maximum Dry Density (MDD) and Optimum Moisture Content (OMC) Based on Standard Proctor
(ASTM D698)
FIGURE 1 Normalized Mr Values (Mr/Mropt) versus Normalized Water Content (w/wopt)
While there is a decrease in Mr value with increasing moisture content, this relationship
can also be considered inversely, i.e. an increase in Mr value as the moisture content
decreases. It can be explained that the soil becomes stiffer, since Mr is analogous to stiffness,
as the water content decreases. This increase in stiffness can best be appreciated by observing
the effect of deviatoric stress on Mr for tests conducted on the dry side of optimum versus
tests conducted on the wet side of optimum, especially for the A-7 soils. Figures 2(a) through
2(d) display the results of Mr tests performed at OMC - 6% and OMC +3% for soil P-26 and
OMC - 6% and OMC + 6% for soil P-53, respectively. From these figures, it is evident that
3
162
163
164
165
166
167
168
169
170
171
172
173
174
175
Mr decreases sharply with increasing cyclic stress when the soils are on the wet side of
optimum, indicating an overall decrease in the stiffness of the soil specimen. However, the
decrease in Mr with increasing cyclic stress for specimens on the dry side of optimum is not
as pronounced as on the wet side of optimum. As will be discussed later in this paper, this
decrease in stiffness on the wet side of optimum can be attributed to a decrease in matric
suction. A-7 soils tend to develop significant magnitudes of matric suction on the dry side of
optimum, therefore the decrease is matric suction and consequently decrease in stiffness is
drastic as the soils enter the wet side of optimum moisture content range. The results
presented in Figures 1 and 2 represent Mr specimens with varying compaction moisture
contents; therefore, the decrease in stiffness (i.e., Mr) could be attributed to both, a decrease
in the matric suction and a change in the soil structure since soil structure is affected by the
compaction moisture content, especially as moisture content increases from dry to wet of
optimum.
176
(a)
(b)
177
178
179
180
181
(c)
(d)
FIGURE 2 Mr Values at Different Confining Pressures versus Cyclic Stress; a) P-26 at
OMC -6%; b) P-26 at OMC +3%; c) P-53 at OMC -6%; d) P-53 at OMC +6%
182
183
The SWCC describes the relationship between suction and water content for a given soil
type. The total suction consists of two components, osmotic suction and matric suction. The
Matric Suction
4
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
matric suction defines the decrease in the thermodynamic potential of soil pore-water due to
capillarity and short-range adsorption (13), while the osmotic suction is attributed to the
presence of dissolved salts in the pore water. The SWCC curves, for the four soil types in this
paper, were established by utilizing the axis-translation and chilled-mirror hygrometer
techniques. The axis-translation technique involves maintaining pore-water pressure (uw) at a
constant value (i.e., atmospheric pressure) while applying positive air-pressure (ua), thus
inducing a matric suction (ua – uw). The SWC-150, Fredlund device, manufactured by GCTS
Testing Systems Inc., was utilized to apply the axis-translation technique in this study. The
axis-translation technique is limited to applying a maximum matric suction of 1,500 kPa due
to the limitations of the high air-entry value (HAE) ceramic disk. To generate an accurate
SWCC, data points are needed near the residual saturation conditions (14), which for finegrained soils occur well above the 1,500 kPa matric suction. Therefore, the chilled-mirror
hygrometer technique was utilized to measure the data points above the 1,500 kPa matric
suction. The chilled-mirror hygrometer technique indirectly measures the total suction
through measuring the relative humidity (RH). The relationship between the total suction and
RH, as proposed by Fredlund and Rahardjo (15), is provided by equation 2.
𝜓𝑇 = −
200
𝑅𝑇
𝑉0 ∗𝜔𝑣
𝑈
ln(𝑈 𝑣 )
(2)
𝑣0
201
202
203
204
where: 𝜓𝑇 = total suction, R = universal gas constant (J/mol K), T = absolute temperature
(K), V0 = specific volume of water (m3/kg), ωv = molecular mass of water vapor (g/mol), Uv
= partial pressure of pore-water vapor (kPa), and Uv0 = saturation pressure of water (kPa). In
𝑈
equation (2), 𝑣 represents Relative Humidity (RH).
205
206
207
208
209
210
211
212
213
214
The WP4C Dewpoint Potentiameter manufactured by Decagon Devices, Inc. was
utilized to apply the chilled mirror hygrometer technique. The WP4C can measure matric
suction up to 450 MPa; however, its accuracy decreases when measuring matric suction
below 1 MPa. Therefore, data points below 1,500 kPa matric suction were obtained from the
Fredlund device. A technique similar to that utilized by Nam (16) was utilized to create and
test specimens via WP4C. Since the WP4C measures the total suction, specimens were
prepared with de-aired/distilled water to minimize the effect of osmotic suction, since
measuring the matric suction was the main objective of this study. A nonlinear least squares
optimization technique proposed by Fredlund and Xing (17) (described in equation 3) was
utilized to present the complete SWCC curves.
215
𝑈𝑣𝑜
𝜃𝑤 = [1 −
𝜓
)
𝜓𝑟
1,000,000
ln(1+
)
𝜓𝑟
ln(1+
]
𝜃𝑠
𝜓 𝑛
𝑎
𝑚
(3)
(𝑙𝑛(𝑒+( ) ))
216
217
where: θw is the volumetric water content; θs is the saturated water content; Ψ is the matric
suction; and Ψr, a, n, and m are fitting parameters.
218
219
220
221
222
223
224
225
226
227
Figure 3 provides the SWCC curves obtained for the four different soil types used in this
study. From this figure, pertinent information such as air-entry value (AEV), saturation water
content, and residual water content can be obtained. Generally, a shift to the left is seen in the
SWCC curves as the PI value of the soil decreases. This is expected since higher PI soils
generally have a larger clay fraction, which leads to a larger water holding capacity due to the
adsorption and surface charge properties of clay particles (13). McQueen and Miller (18)
presented the concept that different ‘regimes’ dominate the water holding for certain suction
ranges in a SWCC. Sandy and silty soils tend to derive most of their water holding capacity
from the ‘capillary’ regime; while for clay soils, due to their mineralogy, a significant portion
of the water holding capacity may come from the ‘adsorbed film’ and ‘tightly adsorbed’
5
228
229
230
regime, which are in the higher suction range. This helps explain the shift to the left, or
narrowing of the desaturation range, displayed by the SWCC curves with respect to the
decrease in PI value of the soil.
231
232
233
234
235
236
237
238
239
240
241
242
243
244
FIGURE 3 Predicted SWCC Curves for Each Soil Type with Measured Values
245
246
247
FIGURE 4 Graphical Procedure to Obtain Residual Water Content
Another advantage of obtaining SWCC curves spanning the entire range of saturation,
for each soil type, was having the ability to evaluate the residual condition. Residual water
content is generally defined as the water content where a large increase in suction causes a
relatively small decrease in water content. However, obtaining a value to represent the
residual water content can be difficult since it has no clear definition. Figure 4 illustrates the
graphical procedure utilized in this study to obtain a value to represent the residual water
content. Table 2 presents the residual water content values obtained for the four soil types
along with the air-entry values and the saturated water contents; the parameters that are
important in accurately describing a SWCC curve. It should be noted from Table 2 that, in
general, there is a decrease in the residual water content value with the decrease in PI value.
The physical meaning of the residual water content and its significance in the Mr-matric
suction relationship will be discussed later in this paper.
6
248
249
Table 2: Selected parameters obtained from the SWCC curves.
Soil
Saturated Water
Air-Entry Value
Type
Content
(kPa)
Residual Water Content
P-53
0.56
43
0.18
P-26
0.41
65
0.11
P-17
0.32
42
0.08
P-7
0.34
10
0.06
Note: Water content given is volumetric
250
251
252
ANALYSIS OF RESULTS
253
254
255
256
257
258
259
260
261
262
263
264
265
The gravimetric water content is a popular property in geotechnical engineering and
commonly utilized to create correlations with other soil properties since it is widely accepted
and easy to measure. The normalized Mr values obtained for the four soil types, compacted
at different moisture contents, tested in this study were first correlated with the variation of
gravimetric water content from optimum conditions. Figure 5 (a) displays an exponential
trend relationship for the (w – wopt) versus the normalized Mr value. The correlation is
considered adequate but not excellent with a coefficient of determination (R2) less than 70%.
As can be seen, the gravimetric water content, by itself, is not enough to accurately capture
the effect of moisture variation on Mr values. In this study, the degree of saturation parameter
was realized as a viable alternative to gravimetric water content to represent the moisture
conditions of the soil. The degree of saturation can be evaluated without extensive laboratory
testing and provides a better description of the soil state since it takes into account both the
effects of density and moisture content.
266
267
268
269
270
271
272
Utilizing Degree of Saturation to Predict Changes in Mr Due to Moisture Variation
(a)
(b)
FIGURE 5 a) Mr/Mropt Versus (w – wopt); Mropt = Mr Value at OMC; wopt = OMC; b)
Mr/Mropt versus (S - Sopt) (%) for the Four Soils Types Subjected to Mr Testing.
As mentioned earlier, EICM utilizes an adjustment factor, Fu, to adjust Mr at optimum
conditions due to changes in moisture content. The Fu factor is defined in equation 4 as:
𝑏 −𝑎
log 𝐹𝑢 = 𝑎 +
1+𝑒
−𝑏
ln(
)+𝑘𝑚 (𝑆 −𝑆𝑜𝑝𝑡 )
𝑎
(4)
7
273
274
275
276
where: Fu = Mr/Mropt, the ratio of Mr at a given condition to Mr at optimum condition, Mropt;,
a = minimum of log FU (-0.5934 for fine-grained); b = maximum of log FU (0.3979 for finegrained); km = regression parameter (6.1324 for fine-grained); (S-Sopt) = variation of degree
of saturation (S) from optimum condition (Sopt), expressed as a decimal.
277
278
279
280
The impact of moisture variation on Mr data from this study was also evaluated in terms
of changes in the degree of saturation. Figure 5 (b) displays the normalized Mr versus the
variation in degree of saturation (S – Sopt). A nonlinear regression analysis yields the best fit
line shown in Figure 5 (b) and the corresponding equation 5.
281
𝑀𝑟
𝑀𝑟𝑜𝑝𝑡
= −0.0009𝑥 2 − 0.0511𝑥 + 1
(5)
282
where: x = (S - Sopt) (%), The equation is valid for the range of −30 ≤ 𝑥 ≤ 10.
283
284
285
286
287
288
289
290
To assess the validity of equation 5, the measured Mr data from this study along with
data from Drumm et al. (3) was utilized for comparison. Figures 6 (a) and 6 (b) compare the
measured Mr values versus the predicted Mr values obtained utilizing equations 4 and 5,
respectively. Both equations provide an acceptable agreement with the measured data, while
equation 5 performs slightly better than equation 4 (R2 = 0.8 versus R2 = 0.75). Based on the
performance of equations 4 and 5 on the measured Mr data, it can be concluded that degree of
saturation can serve as a viable predictor variable for evaluating the changes in the Mr value
due to moisture fluctuation.
291
(a)
(b)
292
293
294
295
FIGURE 6 Measured versus Predicted Mr Values Utilizing; a) Equation 4, EICM
model; b) Equation 5
296
297
298
299
300
301
302
303
304
The key difference between saturated and unsaturated soils is the role played by the porewater pressure (PWP). While PWP is usually positive, with depth, in saturated soils, negative
PWP (or matric suction) exists in unsaturated soils due to the presence of air-water-soil
interface. The value of PWP is important in determining the effective stresses in soils. For
unsaturated soils, the effective stress can be evaluated by applying the model proposed by
Bishop (19), given in equation 6, which takes into account the effect of negative PWP (i.e.,
matric suction). Consequently, the matric suction contributes to ‘suction stress’ in unsaturated
soils (13). Suction stress represents the combined macroscopic effects of the air pressure
acting on particle surfaces, the water pressure acting on wetted portions of soil particles
Comparison of Data with Existing Mr – Matric Suction Constitutive Models
8
305
306
where a meniscus forms (due to difference between air pressure and water pressure), and the
surface tension at the air-water-soil interface (20).
𝜎 ′ = (𝜎 − 𝑢𝑎 ) + 𝜒(𝑢𝑎 − 𝑢𝑤 )
307
(6)
308
309
310
where: σ’ = effective stress; σ = total stress; χ = Bishop’s parameter, representing the
contribution of matric suction to effective stress; ua = pore-air pressure; and uw = pore-water
pressure.
311
312
313
314
315
316
317
318
It is well-known that effective stresses control the strength, stiffness and deformation
characteristics of soils. Also, the effective stress for unsaturated soils is highly dependent
upon the matric suction, as evidenced in equation 6. Therefore, a sound theoretical
framework for Mr, which represents a deformation characteristic of soils, should incorporate
the matric suction. The matric suction varies with water content as evidenced by the SWCCs.
Therefore, incorporating the matric suction in a Mr model will indirectly capture the effect of
moisture variation on Mr. The Mr-matric suction models proposed by Gupta et al. (8) and
Liang (9) are given in equations 7 and 8, respectively.
𝜃 𝑘2
𝑀𝑟 = (𝑘1 𝑃𝑎 (𝑃 )
319
320
𝑎
𝑘3
𝜏
( 𝑃𝑜𝑐𝑡 + 1) ) + 𝛼(𝑢𝑎 − 𝑢𝑤 )𝛽
𝑎
(7)
where: α and β = fittings constants.
𝜃+𝜒𝜓𝑚 𝑘2
𝑀𝑟 = 𝑘1 𝑃𝑎 (
321
(𝑈𝑎 −𝑈𝑤 )𝑏 .55
𝜏
( 𝑃𝑜𝑐𝑡 + 1)
𝑎
𝑘3
(8)
322
Where: 𝜒𝑤 = (
323
324
325
326
327
328
329
330
The Mr-matric suction relationship displays a non-linear trend. Gutpa et al. (8) proposed
that the Mr-matric suction relationship resembles a power function. If Mr values are
considered independent of stress state (i.e., both bulk stress and octahedral shear stress are
constant) across different moisture contents, then during the regression analysis the terms k1,
k2 and k3 in equation 7 become unnecessary. Only the matric suction (ua – uw) varies if the
stress state is constant, and equation 7 can be reduced to the form given in equation 9 for the
regression analysis. Equation 9 represents a simple power function relationship between Mr
and matric suction for the special case of constant stress state.
331
𝑈𝑎 −𝑈𝑤
)
𝑃𝑎
)
; (ua – uw)b = air-entry pressure; and ua – uw = matric suction.
𝑀𝑟 = 𝛼(𝑢𝑎 − 𝑢𝑤 )𝛽
(9)
332
333
334
335
336
337
Figure 7 presents the plots of Mr values obtained at the SHRP P-46 stress state condition
(θ = 155kPa, τoct = 13 kPa) but at different moisture contents, for the four soil types, versus
the matric suction in a semi-logarithmic scale. The data points for each soil type are fitted
with the best fit power regression function. The data in Figure 7 displays an excellent fit to
the power function with high R2 values, which demonstrates that the Mr-matric relationship
follows a non-linear trend.
338
339
340
341
342
343
344
345
346
The ability of equations 7 and 8 to predict Mr across different stress states was evaluated
by performing non-linear regression analysis on the laboratory measured Mr data obtained
from this study. Table 3, presented later in the text, provides the regression constants
obtained for each soil type and for each model along with the coefficient of determination
(R2) values. Based on R2 values, the reader can realize that equation 7 has stronger agreement
with the measured data than equation 8, while also possessing the ability to capture the effect
of moisture variation. However, equation 7 treats the matric suction as an additional term,
which leads to an increased number of fitting parameters, as compared to equation 8 and the
‘Universal’ model.
9
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
FIGURE 7 Measured Mr Values versus Matric Suction for the Four Soil Types Tested
The Liang (9) model (equation 8) has a sound approach since it incorporates the effect of
matric suction within the bulk stress term. Similar to bulk stress, the matric suction has a
stiffening effect on the soil specimen. However, based on R2 values that are presented in
Table 3, equation 8 only provided a marginal fit to the measured Mr data, especially for soils
P-7 and P-17. The definition of the χ parameter, which represents the contribution of matric
suction to effective stress, utilized by Liang (9) was introduced by Khalili and Khabbaz (21).
It represents a linear relationship in a log-log scale, which was obtained from the shear
strength tests on unsaturated soils. The applicability of this relationship to Mr tests may be
questionable since Mr tests are conducted under dynamic loading, while shear strength tests
are conducted under static loading conditions.
368
Proposed Modified Mr-Matric Suction Constitutive Relationship
369
370
371
372
373
374
The results of the laboratory testing program were used to develop a model to estimate Mr
that takes into account the stress state of unsaturated soils while also capturing the effect of
moisture variation. Fredlund et al. (22) proposed a linear function, presented in equation 10,
which incorporated the matric suction in predicting the shear strength of unsaturated soils.
Equation 10 has two components, the saturated shear strength of the soil and the contribution
of matric suction to shear strength.
375
𝜏𝑓 = 𝑐 ′ + (𝜎𝑛 − 𝑢𝑎 ) tan 𝜙′ + (𝑢𝑎 − 𝑢𝑤 ) tan 𝜙 𝑏
(10)
376
377
378
379
where: τf = shear strength of unsaturated soil; c = effective cohesion of saturated soil; ϕ =
effective angle of shearing resistance for saturated soil; ϕb = angle of shearing resistance with
respect to matric suction; (σn – ua) = net normal stress on the plane of failure; and (ua – uw) =
matric suction on plane of failure.
380
381
382
383
384
385
However, a subsequent study by Gan et al. (23) showed that the shear strength- matric
suction relationship for unsaturated soils follows a non-linear trend. The non-linearity in the
Mr-matric suction relationship was noted earlier in this paper. The non-linearity in the matric
suction-shear strength relationship can be attributed to the variation in the contribution of
matric suction to effective stress. Vanapalli (24) argued that the contribution of matric
suction to shear strength is related to the area of water menisci in contact with soil particles,
10
386
387
388
389
390
391
392
393
and as water content decreases the area of contact decreases as well. Initially, the soil is
saturated and the area of water in contact with soil particles is continuous. As suction
increases to values above the air-entry value, air begins to enter the soil pores, and
consequently, the area of water in contact with the soil particles is reduced. This process
continues till the residual saturation condition is achieved. At this point, the area of water in
contact with the soil particles is discontinuous and insignificant. This concept was furthered
by Lu and Likos (20) by presenting a non-linear relationship between suction stress, defined
as intergranular stress in unsaturated soils, and matric suction.
394
395
396
397
398
399
400
401
402
Vanapalli et al. (25) proposed a shear strength model for unsaturated soils, presented in
equation 11, which accounts for the variation in the contribution of matric suction to shear
strength by relating the area of water in contact with the soil particles to the normalized water
content. The normalized water content utilized by Vanapalli et al. (25) accounts for the
residual water content when the area of water in contact with the soil particles is
insignificant. Suction forces act at the particle contacts, and the strength of the force depends
on the size of the liquid contact area between the soil particles (26). It can be inferred that the
increase in matric suction at the residual stage does not significantly contribute to the
increase in the shear strength, and similarly the Mr value.
𝜏 = [𝑐 ′ + (𝜎𝑛 − 𝑢𝑎 ) tan 𝛷′] + (𝑢𝑎 − 𝑢𝑤 )[(𝛩𝑘 )(𝑡𝑎𝑛𝛷′)]
403
(11)
𝜃−𝜃
404
where: 𝛩 = 𝜃 −𝜃𝑟 ; Θ = normalized water content; θ = water content; θr = water content at
405
406
residual condition; θs = water content at saturated condition; and k = fitting parameter to
obtain better agreement amongst measured and predicted values.
407
408
409
410
411
412
413
414
415
The model proposed in this paper, displayed in equation 12, represents a modified
version of the ‘Universal’ and Liang (9) Mr constitutive models. The proposed model also has
distinct similarities with the small strain shear modulus (Go) – matric suction relationship
proposed by Sawangsuriya et al. (12). The Sawangsuriya et al. (12) Go-matric suction
relationship utilized principles from Vanapalli (24) and Vanapalli et al. (25) that are also
included in equation 12 by incorporating the impact of normalized water content on the
contribution of matric suction to the Mr value. It should be noted here that evaluating the
normalized water content requires data from a SWCC curve, which has been evaluated over
the entire range of saturation.
416
𝑠
𝑟
𝑀𝑟 = 𝑘1 𝑃𝑎 (
(𝜃+𝛩𝑘 𝛹)
𝑃𝑎
𝑘2
)
𝑘3
𝜏
( 𝑃𝑜𝑐𝑡)
𝑎
(12)
417
418
where: k = 1/n; n is obtained from the Fredlund and Xing (17) SWCC fitting model, which
represents the rate of change of matric suction due to changes in water content.
419
420
421
422
Equation 12 also establishes an explicit link between the SWCC and the Mr-matric
suction relationship by utilizing the exponent k, which is evaluated based on the parameter n.
The parameter n is obtained from equation 13, the 3-parameter version of the Fredlund and
Xing (17) model.
423
𝜃𝑤 =
𝜃𝑠
𝜓 𝑛
𝑎
𝑚
(13)
(𝑙𝑛(𝑒+( ) ))
424
425
426
427
428
The n parameter depends on the ‘slope’ of the SWCC, and it implicitly takes into
account the soil type. Leong and Rahardjo (14) demonstrated the effect of the n parameter on
the shape of the SWCC curve in the desaturation zone. The n parameter generally captures
the rate of change of suction with respect to water content in the desaturation zone of the
SWCC. As mentioned earlier, different ‘regimes,’ which are dependent on soil type, can be
11
429
430
431
used to determine the desaturation zone of a SWCC.
432
433
434
435
436
437
438
439
440
441
The proposed model was first evaluated utilizing the laboratory Mr data from this study by
performing non-linear regression analysis. Table 3 presents the regression constants obtained
from the regression analysis utilizing the proposed model in equation 12, in addition to
equation 7 and 8. Figure 8 displays the comparison between the laboratory measured Mr
values and the predicted Mr values obtained utilizing equation 12 for the four soil types. The
figure clearly demonstrates strong agreement between the measured and predicted Mr values.
Based on the R2 values presented in Table 3 and the results displayed in Figure 8, the reader
can realize that the proposed model presented in equation 12 performs well in predicting the
Mr value of unsaturated soils by incorporating the matric suction, and hence being able to
account for the effect of moisture variation on the Mr value through matric suction.
442
443
444
445
446
FIGURE 8 Comparison between the Laboratory Measured Mr Values and Mr Predicted
Values Obtained utilizing Equation 12 for the Four Soil Types
Validation of the Proposed Model
TABLE 3 Results of Regression Analysis Performed on Laboratory Data Utilizing
Equations 7, 8 and 12.
P-7
P-17
P-26
P-53
P-7
Equation 7
P-17
P-26
P-53
P-7
Equation 8
P-17
P-26
P-53
Equation 12
k1
52.0
71.7
150.9
225.4
345.5
123.2
165.5
347.1
205.0
58.5
169.6
368.1
k2
1.73
3.17
2.57
1.19
0.76
1.79
1.23
0.8
0.90
1.30
0.90
0.45
k3
-0.88
-7.71
-15.05
-6.06
-1.07
-2.27
-1.49
-1.58
-0.81
-1.2
-1
-1.1
α
968.8
119.4
1398.6
2472.7
β
0.46
0.86
0.36
0.24
n
0.73
1.05
0.76
0.59
k
1.37
0.95
1.32
1.69
RMSE
2
R
553.1
692.4
1090.2
484.1
881.2
1756.6
1449.1
786.6
772.2
579.4
1111.8
694.2
0.92
0.95
0.88
0.94
0.64
0.69
0.81
0.85
0.83
0.97
0.89
0.84
12
447
448
449
450
451
452
453
454
455
456
To further validate the proposed model in equation 12, additional Mr data was
collected from Gupta et al. (8) and Liang (9) studies. A non-linear regression analysis was
performed on the external collected Mr data, obtained at different moisture contents, to
evaluate the performance of equation 12. Figures 9 (a) and 9 (b) display the comparison
between the measured Mr values, from Gupta et al. (8) and Liang (9), respectively, versus the
predicted Mr values, obtained by performing regression analysis utilizing equation 12. The
figures clearly demonstrate excellent agreement between the measured and predicted Mr
values with high R2 values. It is very clear that the proposed model in equation 12 performs
very well in predicting the Mr values for unsaturated specimens prepared at different moisture
contents and tested at different stress states as demonstrated in Figures 8, and 9a, and 9b.
457
458
459
460
461
(a)
(b)
FIGURE 9 Measured Mr versus Predicted Mr Values Obtained Utilizing Equation 12:
a) Measured Mr Data from Gupta et al. (2007); b) Measured Mr Data from Liang (2008)
462
463
464
465
466
467
468
469
470
471
472
473
474
A laboratory testing program was conducted on four soil types to assess the impact of
moisture variation on the Mr values of unsaturated fine-grained subgrade soils. RLT tests
were conducted on laboratory specimens compacted at various moisture contents to measure
the Mr values at various stress states. The SWCC curves for the four soil types were
established utilizing the axis-translation and chilled mirror hygrometer techniques, which
allowed for the evaluation of the SWCC curves from saturation to residual saturation
conditions. The SWCC curves demonstrate that PI has a significant impact on the matric
suction-water content relationship such that the SWCC shifts to the left as the PI value of
the soil decreases. Analysis of the laboratory test results demonstrated that the effect of
moisture variation on Mr values can be captured by applying different relationships; Mrgravimetric water content, Mr-degree of saturation, and Mr-matric suction. It was
demonstrated that the degree of saturation is superior to gravimetric water content in its
ability to capture the effect of moisture variation on the Mr values.
475
476
477
478
479
480
481
482
The existence of a non-linear relationship between the Mr and matric suction was
emphasized in this study. The data obtained from the laboratory testing program was applied
to existing Mr-matric suction constitutive models (8) (9), which showed that the existing
models provided adequate fit to the laboratory measured Mr data. However, due to certain
concerns in the existing models (Gupta et al. (8) model needs five regression constants while
the Liang (9) model provided a marginal fit to the laboratory measured data for lower PI
soils), a modified Mr-matric suction constitutive model was proposed to evaluate Mr for
unsaturated soils. The proposed Mr model accounts for the non-linear contribution of matric
CONCLUSION
13
483
484
485
486
487
488
489
suction to the Mr values by utilizing the normalized water content. Also, the proposed model
implicitly includes the effect of soil type through incorporating the n parameter. The
proposed model was validated utilizing laboratory data obtained from this study, and data
collected from external sources available in literature. The proposed model demonstrated its
ability to accurately capture the variation in moisture conditions and the effect of stress state
on the Mr values for unsaturated soils.
490
ACKNOWLEDGEMENTS
491
492
493
494
495
496
497
498
This research is funded by the Louisiana Transportation Research Center (LTRC Project No.
12-2P) and Louisiana Department of Transportation and Development (State SIO No.
30000425). The authors would like to express their thanks to Mark Morvant, Zhongjie
Zhang, and Gavin Gautreau at LTRC for providing valuable help and support in this study.
The authors appreciate LTRC for allowing utilization of their Geotechnical Research
laboratory and equipment.
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
1. National Cooperative Highway Research Program (NCHRP). (2004). “Development of
the 2002 guide for the design of new and rehabilitated pavement structures.” Project
No. 1-37A, NCHRP Transportation Research Board, Washington, D.C.
2. Phan, T.H., Cushman, M., White, D.J., Jahren, C.T., Schaefer, V.R., Sharma, R.S.,
2008. Case Study of Seasonal Variation in the Subgrade and Subbase Layers of
Highway US 20.
3. Drumm, E.C., Reeves, J.S., Madgett, M.R., Trolinger, W.D., 1997. Subgrade Resilient
Modulus Correction for Saturation Effects. Journal of Geotechnical and
Geoenvironmental Engineering 123, 663–670. doi:10.1061/(ASCE)10900241(1997)123:7(663)
4. Khoury, N.N., Zaman, M.M., 2004. Correlation between resilient modulus, moisture
variation, and soil suction for subgrade soils. Transportation Research Record: Journal
of the Transportation Research Board 1874, 99–107.
5. Witczak, M. W. and Uzan, J. (1988). The universal airport pavement design system.
Report I of V: Granular material characterization. University of Maryland, Department
of Civil Engineering, MD.
6. Nazzal, M.D., Mohammad, L.N., 2010. Estimation of resilient modulus of subgrade
soils for design of pavement structures. Journal of Materials in Civil Engineering 22,
726–734.
7. Yang, S., Huang, W., Tai, Y., 2005. Variation of Resilient Modulus with Soil Suction
for Compacted Subgrade Soils. Transportation Research Record: Journal of the
Transportation Research Board. No. 1913. pp. 99-106
8. Gupta, S., Ranaivoson, A., Edil, T., Benson, C., Sawangsuriya, A., 2007. Pavement
design using unsaturated soil technology.
9. Liang, R., Rabab’ah, S., Khasawneh, M., 2008. Predicting Moisture-Dependent
Resilient Modulus of Cohesive Soils Using Soil Suction Concept. Journal of
Transportation Engineering 134, 34–40. doi:10.1061/(ASCE)0733947X(2008)134:1(34)
10. Cary, C.E., Zapata, C.E., 2011. Resilient Modulus for Unsaturated Unbound
Materials. Road Materials and Pavement Design 12, 615–638.
doi:10.1080/14680629.2011.9695263
REFERENCES
14
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
11. Yang, S., Lin, H., Kung, J., Liao, J., 2008. Shear Wave Velocity and Suction of
Unsaturated Soil Using Bender Element and Filter Paper Method. Journal of
GeoEngineering, Vol. 3, No. 2, pp. 67-74
12. Sawansuriya, A., Edil, T.B., Bosscher, P.J., 2009. Modulus-Suction-Moisture
Relationship for Compacted Soils in Postcompaction State. Journal of Geotechnical
and Geoenvironmental Engineering, Vol. 135, No. 10.
13. Lu, N., Likos, W.J., 2004. Unsaturated soil mechanics. J. Wiley.
14. Leong, E., Rahardjo, H., 1997. Review of Soil-Water Characteristic Curve Equations.
Journal of Geotechnical and Geoenvironmental Engineering 123, 1106–1117.
doi:10.1061/(ASCE)1090-0241(1997)123:12(1106)
15. Fredlund, D.G., Rahardjo, H., 1993. Soil Mechanics for Unsaturated Soils. John
Wiley & Sons.
16. Nam, S., Gutierrez, M., Diplas, P., Petrie, J., Wayllace A., Lu, N., Munos, J.J., 2009.
Comparison of testing techniques and models for establishing the SWCC of riverbank
soils. Engineering Geology. 110. 1-10. doi:10.1016/j.enggeo.2009.09.003
17. Fredlund, D.G., Xing, A., 1994. Equations for the soil-water characteristic curve. Can.
Geotech. J. 31, 521–532. doi:10.1139/t94-061
18. McQueen, I.S., Miller, R.F., 1974. Approximating soil moisture characteristics from
limited data: Empirical evidence and tentative model. Water Resour. Res. 10, 521–
527. doi:10.1029/WR010i003p00521
19. Bishop, A. W. (1959). “The principle of effective stress.” Tek. Ukeblad, 106(39),
859–863.
20. Lu, N., Likos, W.J., 2006. Suction Stress Characteristic Curve for Unsaturated Soil.
Journal of Geotechnical and Geoenvironmental Engineering. Vol 132. No. 2.
21. Khalili, N., Khabbaz, M.H., 1998. A Unique Relationship of Chi for the
Determination of the Shear Strength of Unsaturated Soils. Geotechnique 48.
22. Fredlund, D.G., Morgenstern, N.R., Widger, R.A., 1978. The shear strength of
unsaturated soils. Can. Geotech. J. 15, 313–321. doi:10.1139/t78-029
23. Gan, J.K.-M., Fredlund, D.G., 1996. Shear strength characteristics of two saprolitic
soils. Can. Geotech. J. 33, 595–609. doi:10.1139/t96-085-307
24. Vanapalli, S.K., 1994. Simple test procedures and their interpretation in evaluating
the shear strength of an unsaturated soil [microform].
25. Vanapalli, S.K., Fredlund, D.G., Pufahl, D.E., Clifton, A.W., 1996. Model for the
prediction of shear strength with respect to soil suction. Can. Geotech. J. 33, 379–392.
doi:10.1139/t96-060
26. Mitchell, J. K.,, and Kenʼichi S., 2005. Fundamentals of Soil Behavior. Hoboken, NJ:
John Wiley & Sons.
15