ing subgrades. Specimens of the base materials
special and class 4 special bases were nearly iden-
and subgrade were tested in a frozen saturated
tical to the methods of this study. Details of the
condition at three temperatures. The same base
testing procedures for dense graded stone are de-
material specimens were subsequently warmed to
scribed in Cole et al. (1986); the procedures for
above freezing and tested again at several mois-
Albany taxiway A subbase are in Cole et al. (1987).
ture contents created by drawing a suction at the
Both materials were molded into 15-cm (6-in.)
base of the specimens. To obtain unfrozen data for
diam. specimens in the laboratory from field
the subgrade materials, different specimens, re-
samples and then frozen from the top down at 2.5
ferred to as "never frozen," were molded to spe-
cm (1 in.) per day with open system freezing.
cific moisture conditions and tested at room tem-
Resilient modulus testing of the substitute materi-
perature.
als differed from the procedure used for the Mn/
The specimen size used for the subgrades and
ROAD materials in two ways: 1) the substitute
the class 3 subbase was 5.1 cm (2 in.) diameter and
materials were not saturated prior to freezing, and
12.7 cm (5 in.) long. The coarser class 6 material
2) the loading sequence was different (Tables 11
was tested using specimens measuring 15.2 cm (6
and 12). Neither of these variations is expected to
in.) in diameter and 39.4 cm (15.5 in.) in length.
cause significant differences in modulus values.
Specimens tested in the frozen/thawed condi-
Data analysis
tions were first molded at the specified moisture/
For each set of applied deviator and confining
density, which was usually at optimum as deter-
stresses, we recorded the resilient and permanent
mined from the compaction testing. To approxi-
axial and radial strains, and thus we can calculate
mate field conditions, subgrade specimens received
a compaction effort of 575 kJ/m3 (12,000 ft-lb/
a resilient modulus and a resilient Poisson's ratio.
ft3), while class 3 and class 6 special bases were
The following data were then tabulated in a spread-
compacted with an effort equal to 2,630 kJ/m3
sheet: confining stress, deviator stress, resilient
(55,00 ft-lb/ ft3) (Table 10). The specimens were
axial strain, resilient radial strain, density, and
moisture condition or temperature (Berg et al.
then frozen from the top downwards in a manner
1996).
similar to the procedure used in the frost suscepti-
Linear regression analysis was performed on
bility test (Chamberlain 1987).
the data from the resilient modulus testing of the
For the 1206 subgrade, the "never frozen" speci-
Mn/ROAD materials and the substitute materials
mens were prepared at 3 compactive efforts, with
for class 4 special and class 5 special. Although
moisture contents intended to be at optimum, 2%
regression analysis had been performed on the
above optimum, and 2% below optimum (Table 10).
substitute materials for the class 4 and class 5
Table 11a illustrates the sequence of stress con-
special subbases, new regression analyses were
ditions applied to the unfrozen specimens, whether
performed using our current set of governing pa-
thawed or never frozen. We applied only the stress
rameters. In the regression analysis, the resilient
combinations that would avoid excessive perma-
modulus was the dependent variable. For the fro-
nent strains in the specimens (5% decrease in axial
zen condition, a function of the unfrozen water
length), depending on the moisture condition and
content was the independent variable; for the
estimated available strength. The frozen specimens
thawed/never frozen condition, various forms and
were tested holding the confining pressure con-
stant at 69 kPa(10 lb/in.2) and varying the devia-
combinations of stress, density, and degree of satu-
ration were the independent variables.
tor stress (Table 11b). The cyclical deviator stress
The frozen and unfrozen data were analyzed
was applied at each test point until the residual
separately. However, the nonlinear form of the
axial strain remained a constant value, which oc-
equation used to model the resilient modulus was
curred at about 70100 applied cycles.
the same in both cases, as given by
Test procedures--
Mr = K1 (P)K2
(5)
previously tested materials
The methods used during the resilient modulus
where K1 and K2 are constants and P is a govern-
testing of the materials substituted for the class 5
ing parameter.
18
Previous Page
