Beruflich Dokumente
Kultur Dokumente
2014
Journal of Biodiversity and Environmental Sciences (JBES)
ISSN: 2220-6663 (Print) 2222-3045 (Online)
Vol. 5, No. 5, p. 197-204, 2014
http://www.innspub.net
RESEARCH PAPER
OPEN ACCESS
Abstract
Construction of buildings and different structures leads to soil consolidation, hence, causes soil settlement. Soil
settlement depends on numerous factors, for instance, pressure deformation, depletion of pore water, and so
forth. One way to calculate soil settlement, is utilizing compression index, which is obtained through
consolidation test. Obtaining this index through consolidation test is too time-consuming; thus, researchers have
attempted to relate compression index to soil physical parameters such as plasticity limit, liquid limit, void ratio,
and relative density, which all could be simply measured; therefore, there is great deal of empirical relations in
this regard. In this study, the correlation coefficients between the physical characteristics of fine soil and
compression index were investigated using the Artificial Neural Network (ANN). A few but common empirical
equations describing the relationship of compression index with other soil properties were evaluated along with
the developed ANN model in this study. The results have indicated that among the considered empirical
relations, the Rendon-Herrero formula performed better in calculating the compression index. By comparison,
the ANN calculates the compression index more accurately and with less error than the Rendon-Herrero formula.
*Corresponding
where,
e0
CC 0.009(wL 10)
designing
criterion
based
on
the
(3)
estimated
1 1 e0
CC
2 GS
2.4
(4)
CC 0.2243wL .GS
(5)
each
region,
in
addition,
research
and
CC
n0
371.747 4.27n0
where,
(6)
ratio.The
intact clay:
CC 0.007(wL 10)
(1)
following
relationship
CC 0.0681e1.405e0
where,
CC
wL (%)
was
presented
an
important
(7)
index
is
generally
CC 0.15(e0 0.35)
(2)
permeability and consolidation characteristics of Namontmorillonite swelling clay soil was investigated by
layer that hosts inputs and data is the input layer. The
a f(
pi w j,i b j )
i 1
In this formula
There are plenty of research studies, which are
recently published in the literature showing the
application of ANN method for simulation of different
phenomena in soil and water. For example, Erzin et
al., (2009) modeled the hydraulic conductivity of
compacted fine grained soils using ANN. Soil solution
(8)
w j ,i indicates
b j indicates
pi shows
the
is a
1
f ( z)
1 exp( z )
(9)
on
compression
index,
in
other
words,
the
CC f (e0 , wi , LL, PI , Gs )
(10)
wi
measured
data
for
prediction
of
is the primary
errors
was
obtained
for
training,
evaluation
predicted
and
comparison
of
the
CCm
CCp
with
the
corresponding
accurately
characteristic.
and
predict
this
very
complex
soil
Pattern
Training
Verification
Test
following formula:
CCmi CCpi
% E i 1
100
(11)
CCmi
RMSE
%E
R2
0.03
0.03
0.02
14.44
16.00
10.93
1.00
0.97
1.03
0.67
0.55
0.70
i 1
RMSE
model.
(CCmi CCpi )
i 1
(12)
Model
RMSE
%E
R2
Rendon-Herrero
0.04
Ahadiyan
0.02
0.06
Nishida
Nagaraj and Murty
0.12
0.05
0.11
Skempton
0.07
ANN(present study)
0.02
empirical models.
gravity,
classified
in
three
categories
20 % error
The
models
like
Nishida
equation
generally
Rendon-Herrero
Ahadiyan
Park and Koumoto
Nishida
Nagaraj and Murty
Terzaghi and Peck
Skempton
ANN (present study)
18.26
37.23
67.16
0.00
16.80
68.62
43.08
16.05
70.07
58.39
29.92
0.00
51.09
27.00
36.49
73.00
11.67
4.38
2.92
100.00
32.11
4.38
20.43
10.59
References
Conclusions
(In Persian).
the
most
important
parameters
and
soil
montmorillonite
swelling
clay.
Journal
2012.
of
Di
Matteo
L,
Bigotti
F,
Ricco
R.
2011.
46(8), 955-968.
index
equation.
Journal
of
geotechnical
and
index
equation.
Journal
of
the
Geotechnical
amended
soil-Bentonite
backfills.
Journal
of
102(1-4), 205-209
pp. 1448-