Environmental Engineering Reference
In-Depth Information
variable is the same for each successive particle fraction,
but it would appear that
n
p
should be a function of particle
diameter.
40
30
5.12.12 Data Sets Analyzed in Testing M.D. Fredlund
(2000) Model
A sample data set of 188 soils was selected to test the pro-
posed model. The soils were from Sillers (1997), Rawls and
Barkensiek, et al., (1985), and Williams (1992). All soils
selected had a measured grain-size distribution and a mea-
sured SWCC.
The reliability of the M.D. Fredlund (2000) pedo-transfer
function was evaluated using cross-validation (Hjorth,
1994). The reliability of a pedo-transfer function was
assessed by (1) drawing a random sub sample from the
data set, (2) developing a pedo-transfer function for the sub
sample, and (3) testing the accuracy of the pedo-transfer
function against the data left after subsampling. Data sets
were first selected based on the availability of grain-size
distribution curves and volume-mass properties (i.e., void
ratio, dry density, and specific gravity). The database was
split into two parts with one part used to train the proposed
M.D. Fredlund (2000) pedo-transfer function and the other
part used to test the pedo-transfer function.
20
Proposed PTF
10
Experimental
drying curve
0
10,000 100,000
10
6
0.1
1
10
100
1000
Soil suction, kPa
Figure 5.105
Comparison of experimental and predicted SWCC
for sandy silt (Data after Schuh et al., 1991).
although all estimated SWCCs appear to be quite reasonable.
Results tend to be somewhat sensitive to the assumed packing
porosity
n
p
. More research is required to better understand the
influence of the packing porosity.
The M.D. Fredlund (2000) pedo-transfer function pro-
vided a reasonable estimation of the SWCC for a wide range
of textural classes. There are several groups of soils for
which it is particularly difficult to estimate the SWCC. These
general categories of soils include (i) soils having a large
amount of clay-sized particles, (ii) soils that contain large
amounts of coarse-size particles mixed with fines, (iii) soils
that exhibit bimodal behavior such as sand-bentonite mix-
tures, and (iv) man-made soils such as mine tailings and
waste rock.
The assumed minimum particle size has an influence on
the estimation of the SWCC. If the minimum particle size is
set too small, there will be an overabundance of clay-size
particles that will influence the estimation of the SWCC. If
5.12.13 Estimation of SWCC Using M.D. Fredlund
(2000) model
Comparisons of experimental and predicted results are shown
in Figs. 5.104-5.106. Ten optimal estimations of the SWCC
were performed using the M.D. Fredlund (2000) method and
the results can be seen in Fig. 5.107. The estimation of a
SWCC from grain-size distribution was attempted for all soil
types. There was greater difficulty in estimating the SWCC
for clay, till, and well-graded soils than for silt and sands,
30
Proposed PTF
20
Experimental drying curve
10
Log PDF
0
10
6
0.1
1
10
100
1000
10,000
100,000
Soil suction, kPa
Figure 5.104
Comparison of experimental and predicted SWCC and logarithmic probability
density curve (log PDF) for clay soil (Data from Russam, 1958).
Search WWH ::
Custom Search