Professional Documents
Culture Documents
1996
Japanese Geotechnical Society .
ABSTRACT
The use of vertical drains and surcharge to accelerate the consolidation of clayey soils is well established in soil im-
provement technology. For this method of soil treatment, it is important to monitor the progress of consolidation
where the average degree of consolidation is usually defined as the ratio of current settlement to the ultimate settle-
ment due to primary consolidation. An accurate estimate of the ultimate primary settlement is, therefore, required in
order to determine the state of consolidation, and the appropriate time for surcharge removal when the desired degree
of consolidation has been achieved.
Recently, Tan (1995) proposed a new hyperbolic method to identify the 60% and the 90% consolidation stages
from field settlement data due to one dimensional consolidation. Knowing the settlement at these points, it is possible
to determine the ultimate primary settlement. Once sufficient data has been recorded to identify the linear portion of
the hyperbolic plot between the 60% to 90% consolidation stages, the method can be employed. This method has
been validated by application to various well documented case histories in the published literature. In Japan,
however, Asaoka's method for estimating ultimate primary settlement from field observation has been widely used
and is well accepted by the geotechnical engineering community.
This research paper presents a comparison of two different observational methods for monitoring the progress of
consolidation to determine their similarity and differences, and their accuracy of predictions compared to actual obser-
vations. First, the theoretical basis of the two methods are examined. It was found that settlement data beyond the
60% consolidation stage are needed in both methods to make accurate predictions of ultimate primary settlement.
Next, both methods are applied to oedometer laboratory consolidation data for kaolin specimens in order to deter-
mine how prediction of ultimate primary settlement and coefficient of consolidation would compare in carefully con-
trolled laboratory experiments. Finally, both methods are applied to several well documented case histories of vertical
drain projects, and the predictions of ultimate primary settlement and in situ coefficient of consolidation are com-
pared. It was found that both methods gave good agreement with predictions, making them complementary tools for
use in monitoring consolidation in field applications.
Key words: Asaoka's method, clays, hyperbolic method, ultimate primary settlement, vertical drains (IGC: D4/E2)
i) Senior Lecturer, Department of Civil Engineering, National University of Singapore, 10 Kent Ridge Cresent, Singapore 0511.
ii) Lecturer, ditto.
Manuscript was received for review on April 3, 1995.
Written discussions on this paper should be submitted before April1, 1997 to the Japanese Geotechnical Society, Sugayama Bldg. 4 F, Kanda
Awaji-cho 2-23, Chiyoda-ku, Tokyo 101, Japan. Upon request the closing date may be extended one month.
31
This is an Open Access article under the CC-BY-NC-ND license.
ties are highly variable however, and even with the best 1.4
soil investigation program, the settlement predictions can
1.2 - Terzaghi Theory
be quite varied and inaccurate when compared to the ac-
tual observed settlement in the field. It is desirable there- 1.0
fore to use observational procedures, based upon which 0.8
the ultimate primary settlement can be estimated once ;f
J-? 0.6
sufficient settlement data has been recorded. The hyper-
bolic method which was proposed and refined by Tan 0.4
(1993, 1994, 1995) is one such observational method, and 0.2
the other established method is that proposed by Asaoka (a)
0.0
(1978), in which early field settlement data is used to 0 0.2 0.4 0.8 1.2
predict completion of primary settlement and the in situ
coefficient of consolidation. These procedures have a
good theoretical basis, and they have the distinct advan- 0.6~------------~--------------.
tage of estimating the ultimate primary settlement from -+-- Field settlement
the early settlement record of a drain installation to see if
the system is functioning as intended in the design. ,........,0.4
In order to evaluate their similarity and differences, the E
.g
theory and applications of the two methods are com- ~ T = Cvt/H 2
"0
pared in this paper. :;:::. 0.2 Cv= 1 m2 /yr
H=5n'l
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Theory of Hyperbolic Method
t (yr)
The hyperbolic method as proposed by Tan (1995) had
its origins 'in the rectangular hyperbola fitting method Fig. l(a), (b). Hyperbolic plots of Terzaghi theory and field settle-
proposed by Sridharan et al. (1981, 1987), which is a ment data
method for obtaining the coefficient of consolidation, Cv
from oedometer tests, by fitting laboratory settlement
data to the linear segment of the hyperbolic plot between
the U60 and U9o points of the theoretical Tv/ Uv vs Tv plot. and 90% consolidation stages for any consolidation set-
The basis of this method is shown in Fig. 1. The Terzaghi tlement data by noting the relationships that the slopes of
theory of consolidation gives a unique settlement-time lines radiating from the origin to these points make with
plot in terms of the average degree of consolidation the linear segment of the plots. Thus, the slopes S6o and
against the time factor, Uv vs Tv. When the settlement is S9o can be determined by:
plotted as Tv! Uv vs Tv as shown in Fig. 1(a), the feature a6o S;
of the plot is an initial concave segment up to U6o, fol- S6o=Si-=(1/0.6)- (2)
a; a;
lowed by a linear segment between U6o and U9o· This
straight line portion can be represented by the equation: (3)
Tv
u=aTv+/3 (1) where, S; is the slope of the initial linear segment of the
t I o vs t plot, and a; is the slope of the theoretical Tv/ Uv
which is the equation representing a rectangular hyperbo- vs Tv plot as shown in Fig. 1. In the earlier work by Tan
la, with a as the slope and /3 as the intercept of the hyper- (1995), the symbol a instead of a; was used. The latter
bolic plot. When lines radiating from the origin are symbol is preferred as the subscript "i" emphasized that
drawn to U6o and U9o, the slopes of these lines are a6o = the theoretical slope considered is the initial linear seg-
1/0.6 and a9o=1/0.9, respectively. The slope of the ment between the 60% and 90% consolidation stages. If
linear segment is determined from a least squares linear the slope of initial linear segment (S;) of the field hyper-
regression with coefficient r 2 greater than 0.9999, to be bolic plots between the 60% and 90% consolidation
ai=0.821, a unique value applicable only to Terzaghi's stages can be determined during field monitoring of settle-
theory. ments; S 60 and S9o can be calculated. By constructing
In field monitoring of the consolidation of soils, plots these radiating lines to intersect the first linear segment of
of settlement (o) versus time (t) are recorded. When a the field hyperbolic plot, the 60% and 90% settlement
hypothetical field settlement data is plotted in the form points can easily be identified. The ultimate primary set-
of t/ o vs t, as in Fig. 1(b), that is time/ settlement vs tlement can then be estimated as 06o/0.6 or o9o/0.9,
time, the same features as the theoretical plot are ob- where o6o and o9o are 60% and 90% of the ultimate prima-
served, namely an initial concave segment, followed by a ry settlement. Based on Eq. (2), the line radiating from
linear segment between the 60% and 90% consolidation the origin to the 60% settlement point in the Fig. 1(b) can
state. This feature can be expanded to identify the 60% also be expressed by:
TTv=
v,
~
f;- for 2 < 0 ... 2
Cvf
Tv= H (6a) 0.15 o Uso
Drain Influence
~ c. Ugo Diameter (D)
....... -- 1m
T,= ;~<::.0 ...2
0.10
U,= 1-:, exp ( -:'T,) for (6b)
-2m
---- 3m
0.05 d =0.05 m
ClayH 10m
The average degree of consolidation for radial consoli-
dation by Barron's theory is given by Hansbo (1981) as:
0.05 0.10 0.15 0.20 0.25
U•=l-exp(-:T•) for T.=~; (7a) Tv
Fig. 2. Hyperbolic plots of typical settlement with vertical drains
where
n2 3n 2 -1
p=-2-1ln (n)--4
2 (7b) 0.82 r-----------------------.
n- n
where n=drain spacing ratio, Dl d; D=diameter of an 0.80
~ . . . 0. . . . . . . . ---o------ ----------0 o
o H/D=2
H/D=5
equivalent soil cylinder influenced by each drain, which is ,~ a. HID= 10
;:? I
o H/D =·20
equal to 1.13s for a square pattern and 1.05s for a triangu- g?
cl
::::> 0.78
lar pattern (s=drain spacing) and d=diameter of sand
drain or equivalent diameter of prefabricated flexible ?
0
drain, for which Hansbo (1981) recommended that 8_0.76
.2
d=2(b+t)l n. (.J)
tr
The average degree for combined vertical and radial 0.74
consolidation can be obtained from Carillo's theorem ch/cv =1 ch/cv = 3 ch/cv = 9
(1942):
10 20 30 40 50 60 70 80 90 100 110
(8)
Drain Spacing Ratio (n =D/d)
Using Eqs. (6) to (8), the theoretical settlement-time Fig. 3• Plotofextasafunctionofn=D/d,H/Dandch/cvforvertical
for any vertical drain problem can be computed and plot- drains
straight line part can be observed beyond the 60% con- and for radial drainage only, in situ ch can be estimated
solidation stage. With Ot.i and Si determined, 860 and Sgo is by:
computed from Eqs. (2) and (3), and the two radial lines
lnPt
from the origin drawn with these slopes to intercept the (14)
first linear segment of the field hyperbolic plot at the 60% .Lit
and extended to the 90% consolidation points. Thus o6o To evaluate what the appropriate range of settlement
and Ogo can be estimated, and the ultimate primary settle- data would be for least squares linear regression to deter-
ment can be computed from Eq. (5). mine Po and Pt in the Asaoka plot, that would give ac-
Since this procedure also identifies the time for 60% curate estimates of end of primary settlement in Eq. (11),
and 90% consolidation (t6o and tgo) for the combined the Terzaghi theory settlement is used for time incre-
flow consolidation in the vertical drain system, it is possi- ments .LI Tv of 0.01, 0.025 and 0.05 using Eqs. (6a), (b).
ble to estimate the gross average in situ field consolida- These are plotted in the Asaoka form of Un-t vs Un in
tion coefficient ch, assuming that the value of Cv from Fig. 4. Least squares linear regression is used on the plots
laboratory oedometer tests on high quality undisturbed
field samples is known. For the vertical drain case, Uv
will rarely exceed 50%. Therefore, with t= t6o and t= tgo 1.2~----------------------------~
for the combined flow known, Uv can be calculated from Asaoka's plots for
Eq. (6a). Using Carillo's theorem of Eq. (8), Uh at the Terzaghi's Uv data
time t6o or tgo can be calculated, since U=0.6 or 0.9. Rear-
ranging Eq. (7a), the values of ch corresponding to t6o or
tgo are given as:
-f..lD 2 ln (1- Uh)
(9)
8t
ATV
Theory of Asaoka's Method --0.01
In 1978, Asaoka proposed an "observational proce-
dure" in which early settlement data can be used to -+- 0.025
predict the ultimate primary settlement and in situ ..... 0.05
coefficient of consolidation for one dimensional consoli-
dation. The method is becoming increasingly popular be-
cause of its simplicity and lack of required for detailed 0.2 0.4 0.6 0.8 1 1.2
sampling and laboratory testing to determine soil proper-
ties or for the monitoring of field pore pressure behavior. un-1
The Asaoka method is based on the fact that one dimen- Fig. 4. Asaoka's plots of Terzaghi's Uv data for various time incre-
sional consolidation settlements oo, o~, o2, etc. at times 0, ments
.LIt, 2.LI t, etc., can be expressed as a first order approxima-
tion by:
Table 1. Influence of data range used for regression, on Asaoka's
(10)
method
which represents a straight line in a On- 1 vs on known as Time increment Regression
Po Pt Cv Uutt or U100
the Asaoka's plot, where Po is the intercept and Pt is the and data range coefficient =Po/(1-Pt)
slope of the. line. When the ultimate primary settlement used to obtain rz
has been reached, On= On-1 = Outt· Hence the end of prima- Ps
ry or ultimate primary settlement Outt is given by: (1) (2) (3) (4) (5) (6)
LlTv=O.Ol
(11) Uo to U3o 0.095 0.712 0.971 13.76 0.33
U3o to U60 0.030 0.966 0.999 1.42 0.87
The number of time increments for 90% consolidation, u60 to U90 0.024 0.976 1 1 1
ho, is given as: LlTv=0.025
U0 to U30 0.170 0.562 0.968 9.34 0.39
U3o to U6o 0.071 0.919 0.999 1.37 0.88
(12)
U6o to U90 0.060 0.940 1 1 1
to determine Po and Pt for the data range of U0 to U3 0 , U3o used with each load applied for 24 hours. The settlement
to U6o, and U6o to U9o for the three time increments. data was recorded electronically at every 1Os interval, in
These are shown in Table 1, together with the regression order to permit Asaoka's method to be applied for vari-
coefficients, and the computed Cv using Eq. (13), and com- ous time increments of Ll t= 10, 20, 30, 40 and 50s. The
puted ~utt using Eq. (11). The exact values for both Cv and typical Asaoka plot for the load step of 2 to 4 kPa at the
~utt is 1. The results showed that using early settlement five time increments is shown in Fig. 5. Using a least
data in the Asaoka plot, say, from O% to 30% consolida- squares linear regression on the plots for data between
tion, would give very low estimates of ~utt and gross over- the 60% and 90% consolidation stages, the values of Po
estimates of Cv, with improving estimates for larger time and Pt were obtained with regression coefficient greater
increments. Using the data range from 30% to 60% con- than 0.999, and the corresponding values ofho, Cv and ~utt
solidation showed much better estimates, ~utt however is computed from Eqs. (11) to (13), with the results tabu-
still underestimated by about 10%, with Cv overestimated lated in Table 2. Although the values of Po and Pt are
by about 30%. Only regression with settlement data be- different for the different time increments used, the
yond 60% consolidation, determine the exact values for values of Cv and ~utt are almost the same, varying between
both ~utt and Cv. 0.68 and 0.72 mm2 Is, and 0.983 and 0.988 mm, respec-
This conclusion was supported by the work of Bergado tively. It would seem therefore that for normal primary
et al. (1991), who found that applying Asaoka's method consolidation settlement as seen in kaolin, the choice of
to their field trial with vertical drains in Bangkok, indi- time increment does not really matter, as long as the ho
cated that the accurate prediction of in situ coefficient of fall within the range between 10 to 90.
consolidation was only obtained when Pt was determined Using the kaolin consolidation test data, the values of
using data beyond 60% consolidation. Both the hyperbol- Cv and ~utt were estimated by the conventional, hyperbolic
ic and Asaoka's methods therefore require data beyond
60% consolidation for accurate prediction of settlement
and in situ coefficient of consolidation. 1.2 .
Asaoka's plots for
Kaolin specimen consolidatio
COMPARISON USING LABORATORY ( load from 2 to 4 kPa )
OEDOMETER DATA
Edil et al. (1991) reported that the accuracy of the At(s)
Asaoka's observational method is strongly dependent on .......
the choice of time increments used. They suggested that
--10s
for reasonable predictions in clays, the chosen time incre-
E
ment should give a hs (number of time steps to reach the
E
.......... ---- 20s
95% consolidation stage) value between 10 and 30. They -+- 30s
also determined that the method produces bilinear plots
for peat and its applicability to peat settlement is highly -...-- 40s
questionable. This may be expected as settlement in peat
is largely a secondary compression phenomena at con- -@-50s
stant effective stress, unrelated to the consolidation proc-
ess. To check this dependence on time increment, a 0.2 0.4 0.6 0.8 1 1.2
laboratory oedometer test was conducted on kaolin with
a liquid limit=55, plasticity index=25, and initial void l>n-1 (mm)
ratio=0.8. The specimen was cut to 75 mm diameter, Fig. 5. Asaoka's plots of kaolin consolidation data for various time
and 25 mm height, and a load increment of unity was increments
Table 2. Influence of time increment Lf t, on Asaoka's method applied to kaolin test data
ln (1- U90 ) t90=j9oL1 t duu=f3o/(1-f3t)
L1 t (s) Po (mm) Pt j90 (s) cv (mm2 /s) (mm)
ln (/3 1)
(1) (2) (3) (4) (5) (6) (7)
10 0.026 0.973 86 860 0.682 0.986
20 0.051 0.948 43 868 0.697 0.988
30 0.078 0.921 28 841 0.721 0.983
40 0.102 0.897 21 851 0.712 0.986
50 0.124 0.874 17 862 0.702 0.988
Test data for kaolin specimen in oedometer 75 mm diameter by 25 mm height. LL=55; PI=25; Initial void ratio=0.8, Load step from 2 kPa to 4
kPa. Measured Ouu=l.Omm; Cv from Casagrande's t50 =0.513 mm 2 /s; cv from Taylor's t90 =0.613 mm2 /s
60% and 90% consolidation stages superimposed. From 0 400 800 1200 1600 2000 2400
these plots, si, t6o, and t9o is determined and the values of Time, t (s)
Cv and Outt computed from Tv6o=0.287, Tv9o=0.848, and
ouu=0.821/ Si. The Asaoka plots for a time increment of sooo~----------------------------~
Kaolin d=75mm H=24.3mm
30s for the same settlements data are shown in Fig. 8. A
Load 4kPa to 8kPa
least squares linear regression using data between the
60% and 90% consolidation stages gives Po and fit values
with a regression coefficient greater than 0.999. Based on 'E
these regression analyses, the values of cv and Outt were E
~ 4000
computed using Eqs. (13) and (11), and all the results are co
summarized in Table 3. Using the results indicated in Ta- :;;;;
0. 4 .--K-a_o_li_n_d_=_7_5_m_m_________________~L~o-a~d~S~te-p~
e o +4-SkPa
4000
.S.o.2
«:t
5i -0.4 'E 3ooo
E .§
.5! -0.6 .!.
~ ~ 2000
(/) -0.8
-1 (a) 1000
-1.2 ..__________ _________.________ _ _ j ---~.
(c)
10 100 1000 10000
Time, t (s) 400 800 1200 1600 2000 2400
Time, t (s)
0.4~---------------------~
Kaolin d=75mm 4000~--------------------------~
Initial ht = 25.5mm Kaolin d=75mm H=23.5mm
Load 16kPa to 32kPa
Load Step
-11- 2·4kPa
+4-8kPa
*8-16kPa
·1 --16-32kPa
-1.2 L......o........J..................._...........o...L.........._..~..........~~.........L..............o...L.................L................I.....L..o-......J (d)
0 10 20 30 40 50 60 70 80 90 1 00
Square Root Time (s) 300 600 900 1200
Time, t (s)
Fig. 6(a), (b). Kaolin oedometer settlements plotted at both
logarithmic and square root time scales Fig. 7(a)-(d). Hyperbolic plots of kaolin oedometer settlements
0
for time increment of 30s 0 0.5 1.5 2
1 6. 1u Observed primary settlement (mm)
2
(i) Gradient
0.8 ~ • Taylor too 1.09
..-.. g A Hyperbolic ~o 0.93
U'l
E 0.6 "0
0 • Hyperbolic t 00 0.95
E
..._.. Load Step ~ • Asaoka
E
c • 2 to 4 kPa :G 1
<-0 0.4 :5
0
-z&- 4 to 8 kPa >-
.a
"0
0.2 -'V- 8 to 16 kPa ~
E
iw (b)
J 0
0.2 0.4 0.6 0.8 1 1.2 0 1 2
Table 3. Determination of cv and duu for kaolin consolidation test using conventional, hyperbolic and Asaoka's methods
Load step Average ht. Casagrande's Taylor's Cvfrom t 50 Cvfrom t 90 Observed Outt
(kPa) (mm) tso (s) t 90 (s) (mm2 /s) (mm2 /s) (mm)
(1) (2) (3) (4) (5) (6) (7)
Clay 2H=10 m
tq_ 0~------------------------------~
c
(I) -40 cv=0.1, ch=0.3 m 2 /yr
Orebro (Geodrain at 1.1 m traingle)
s -theory -20
(I)
e(.)
0.08
0.25
-.::
~
t.() 0.06
'E o.2
';;;::;
~ 0.15
.§..
0.04
~ 0.1
0.02
(b)
0.05
0
0 2 4 6 8 10 (b)
Time, t (yr) 0
0 5 10 15 20 25
Time, t (month)
Asaoka's plotsforSka-Edeby V1 120 ..--------------------------.,
{18 em sand drain 1.5m trian e) Asaoka's plots of Orebro G11
80 for time increment of 0.25 (Geodrain at 1.1m triangle}
1 oo for time increment of 0.04
.-..
E
(,)
..........
60
-0
E
..........
80
60
t-0 40 c
"0
40
20
(c) (c)
/3o:6.61 __.,
a~--~----~----~---~---~ a~--~--~----~-~---~--~
0 20 40 60 80 1 00 0 20 40 60 80 100 120
a Drain spacing, T means triangular spacing (D= l.05s), S means square spacing (D= 1.13s)
b Clay thickness, S means single top drainage, D means double drainage
Table 5. Determination of a1, S1, t560 , t590 for hyperbolic method, and P0 , P1 for Asaoka's method
Hyperbolic method Asaoka's method
Site Drains Lit pl
(1) (2)
Slope Slope t56o t59o Po
Ol; s~ (em) (em) (yr) (em)
(3) (4) (5) (6) (7) (8) (9)
Sweden-SkaEdeby
V03 Sand 0.777 0.013 34.6 52.6 0.25 4.77 0.919
V08 Sand 0.758 0.016 29.8 44.4 0.125 7.86 0.842
V13 Sand 0.767 0.01 46.3 68.9 0.25 12.11 0.839
Ali Sand 0.767 0.013 34.4 51.8 0.1 5.40 0.910
AIV No drain 0.824 0.02 25.2 37.8 0.1 1.37 0.967
Sweden-Orebro
011 Geodrain 0.763 0.008 57.1 89.5 0.042 6.61 0.937
014 Geodrain 0.766 0.008 58 87 0.042 5.18 0.948
Sll Sand 0.76 0.007 64.6 96.8 0.021 4.08 0.963
S14 Sand 0.762 0.007 62.4 93.7 0.042 5.94 0.943
Singapore
Changi Sand 0.748 0.008 58.5 87.7 0.014 4.60 0.955
CTE-SP2 OV drain 0.79 0.006 81.2 123 0.042 16.07 0.892
CTE-SP3 OV drain 0.785 0.005 93.7 142 0.042 26.94 0.826
Japan-Kobe
S6 Sand 0.78 0.011 42.8 64.2 0.014 8.32 0.889
For regression between t560 and t590 to obtain Po and P~> all the regression coefficients r 2> 0.995
Table 6. Comparison of ultimate primary settlement estimates using hyperbolic and Asaoka methods
Sweden-SkaEdeby
V03 Sand 59 76 59 58 58 59
V08 Sand 50 64 49 50 49 50
V13 Sand 75 100 76 77 77 75
All Sand 61 58 59 57 58 60
AIV No drain 43 51 42 42 42 41
Sweden-Orebro
011 Oeodrain 100 125 96 95 95 105
014 Oeodrain 100 128 97 97 97 100
Sll Sand 110 140 106 108 108 112
S14 Sand 110 138 105 104 104 105
Singapore
Changi Sand 100 130 98 97 97 102
CTE-SP2 OV drain 145 174 138 135 136 149
CTE-SP3 OV drain 155 201 158 156 158 154
Japan-Kobe
S6 Sand 74 92 72 71 71 75
Table 7. Comparison of consolidation coefficient estimates using hyperbolic and Asaoka methods
Sweden-SkaEdeby
V03 Sand 0.45 2.45 0.38 6.54 0.38 0.41
AIV No drain Cv= 1.10 6.51 Cv= 1.10 19.1 Cv= 1.10 Cv= 1.40
Sweden-Orebro
011 Oeodrain 0.5 0.62 0.5 1.49 0.53 0.55
014 Oeodrain 0.6 0.82 0.66 2.19 0.64 0.82
Japan-Kobe
S6 Sand 3 0.14 3.2 0.3 3.50 4.5
and 11(a}, from which the most likely ch applicable is de- Similarly, a comparison of estimated in situ ch by the
termined. The settlement observed for the field cases are hyperbolic (using Eq. (9)) and Asaoka (using Eq. (14))
shown as hyperbolic and Asaoka plots in Figs. 10(b}, methods with observed values is made in Table 7, and
ll(b), 10(c) and ll(c). From such plots, the values of Si, plotted in Fig. 12(b). It is apparent that the hyperbolic
o6o, and c590 were determined for the hyperbolic method, method estimates agree with the observed values within
and the values of Po and P1 were determined for the an error of 5%, whereas the Asaoka's method overesti-
Asaoka method. These are summarized in Table 5. mates the value of ch by about 10%. This is consistent
Using the results presented in Table 5 and the theories with the results determined from the laboratory kaolin
discussed earlier, the ultimate primary settlement were studies.
computed for both the hyperbolic (using Eq. (5)) and
Asaoka (using Eq. (11)) methods and is shown in Table
6. In order to compare the various estimates, these are SUMMARY AND CONCLUSIONS
plotted against the observed primary settlement in Fig. The theoretical basis for the hyperbolic and Asaoka ob-
12(a}. From this plot, it can be clearly shown that the servational methods for monitoring settlements due to
hyperbolic method proposed by Tan (1995) underesti- one dimensional consolidation with and without vertical
mates the primary settlement by less than 3%, whereas drains were described. It was determined that both the
the Asaoka method overestimates by less than 1%. The hyperbolic and Asaoka's methods require settlement
conventional hyperbolic method which was the inverse of data beyond the 60% consolidation stage in order to pro-
the hyperbolic slope would grossly overestimate primary vide accurate estimates of the ultimate primary consolida-
settlement by about 25%, which is unacceptable. tion settlement and the in situ consolidation coefficients.
One distinct advantage of the hyperbolic over the
Asaoka's method is the ease with which the hyperbolic
250~------------------------------~ method permits the identification of the different consoli-
'E dation stages because of the linearity between the 60%
.& (a) and 90% consolidation points. Although there have been
~ 200 • reports of the accuracy of Asaoka's method being strong-
0
.c ly dependent on the choice of time increment used, it was
Q)
E shown that for normal one dimensional consolidation set-
a; 150
tlement, no apparent influence of the choice of time incre-
.c
0 Gradient ment were observed using data from consolidation tests
>
.c o 1/S1 1.26 on kaolin specimens. Both the hyperbolic and Asaoka
"C 100
• 01..1s. o.97 methods gave very good predictions of ultimate primary
~ settlements for both laboratory oedometer tests on kao-
E v 1} 60 /0.6 0.97
m 50
• a,o/0.9 0.98
lin and field settlement for clays with vertical drains. The
agreement between the predicted and observed primary
• Asaoka 1.01 settlements are well within an error of 5%. Both the
0~----~----~----~------~----~ hyperbolic and Asaoka's methods can also be used to esti-
0 50 100 150 200 250
mate both laboratory and in situ consolidation
llu•u Observed primary settlement (em) coefficients. For this application, the error between
predicted and observed values would be larger compared
-~ (b)
to settlements, about 10%, with the hyperbolic method
underestimating and the Asaoka's method overestimat-
ing the consolidation coefficients. Based on this study, it
-(/) 4
"C
seems that both the hyperbolic and Asaoka methods can
-...
0 be applied effectively to monitor consolidation with verti-
.c:
Q) 3 cal drains, provided that the bulk of the settlement occur-
E
ring is due mainly to one dimensional primary consolida-
-
Q)
.c tion, and not secondary compression at constant effective
0
2 stress as in the case of the settlement of peat.
>
.c Gradien
"C e Hyperbolic t 60 0.98
Q)
a; 1 A Hyperbolic t11o 1.05 REFERENCES
E
~ 1.09
1) Asaoka, A. (1978): "Observational procedure of settlement predic-
w tions," Soils and Foundations, Vol. 18, No. 4, pp. 87-101.
0~----J-----~----~~----~----~ 2) Barron, R. A. (1948): "Consolidation of fine-grained soils by drain
0 1 2 3 4 5 wells," Trans ASCE, Vol. 113, pp. 718-754.
Observed ch (m 2 /yr) 3) Bergado, D., Asakami, H., Alfaro, M. C. and Balasubramaniam,
A. S. (1991): "Smear effects of vertical drains on soft Bangkok
Fig. 12(a), (b). Comparison of Ouu and Cv by hyperbolic and Asaoka clay," Journal of Geotechnical Engineering, ASCE, Vol. 117, No.
methods with observed values for vertical drain case histories 10, pp. 1509-1530.
4) Carrillo, N. (1942): "Simple two and three dimensional cases in the Testing Journal, Vol. 4, No. 4, pp. 161-168.
theory of consolidation of soils," Journal of Mathematics and 9) Sridharan, A., Murthy, N. S. and Prahash, K. (1987): "Rectangu-
Physics, Vol. 21, No. 1, Pt. 1, pp. 1-5. lar hyperbola method of consolidation analysis," Geotechnique,
5) Edil, T. B., Fox, P. J. and Lan, L. T. (1991): "Observational proce- Vol. 37, No. 3, pp. 335-368.
dure for settlement of peat," GEO-COAST'91, 3-6 September, 10) Tan, S. A. (1993): "Ultimate settlement by hyperbolic plots for
Yokohama, 2/4, pp. 165-170. clays with vertical drains," Journal of Geotechnical Engineering,
6) Hansbo, S. (1981): "Consolidation of fine-grained soils by prefabri- ASCE, Vol. 119, No. 5, pp. 950-956.
cated drains," Paper 12/22: Proceedings of the lOth ICSMFE, 11) Tan, S. A. (1994): "Hyperbolic method for settlements in clays
Stockholm, Sweden, Vol. 3, pp. 677-682. with vertical drains," Canadian Geotechnical Journal, Vol. 31, pp.
7) Magnan, J.P. and Deroy, J. M. (1980): "Analyse graphique des 125-131.
courbes de consolidation oedometrique," Bull. Laison Lab. Ponts 12) Tan, S. A. (1995): "Validation of hyperbolic method for settle-
et Chausses, Paris, France, (in French), Vol. 9, pp. 53-56. ments in clays with vertical drains," Soils and Foundations, Vol.
8) Sridharan, A. and Sreepada Rao, A. (1981): "Rectangular hyperbo- 35, No. 1, pp. 101-113.
la fitting method for one-dimensional consolidation," Geotech.