Sunteți pe pagina 1din 6

http://www.paper.edu.

cn
Transportation Research Record 1849
Paper No. 03-3116

47

Evaluation and Quality Control of


Dry-Jet-Mixed Clay Soil-Cement Columns by
Standard Penetration Test
Songyu Liu and Roman D. Hryciw
Dry jet mixing has been widely used since the 1980s for stabilization of
soft soil. The quality and strength of the dry-jet-mixed columns must be
evaluated to confirm the success of the stabilization. The standard penetration test (SPT) is shown to be a simple and effective method for this
task. The strength characteristics along the length of the column were
determined, and correlations between the SPT blow count and the
unconfined compressive strength were developed.

Dry jet mixing (DJM) is a soil improvement technique that pneumatically delivers powdered reagent into the ground and mixes it
with in situ soils to form a soil-cement column. The chemical reactions between soils and dry reagents such as cement powder or lime
increase the strength and reduce the compressibility of soft ground.
Because of its many advantages, DJM has attracted increasing attention compared with other in situ soil stabilization methods since its
development by Kjeld Paus in Sweden, in 1967. In the 1980s a large
increase in DJM usage occurred in the Nordic countries and in
Japan. Today the method is used worldwide, especially in Europe,
North America, and Asia (1, pp. 1525).
The DJM method was introduced to China through Japan in the
early 1980s. The first DJM equipment in China was developed in
1983 at the Wuhan Research Institute of Engineering Machinery of
China. The equipment was successful in soft soil improvement for
railway engineering. Although there were some early failures of
DJM column foundations for buildings during the 1980s (2), DJM
rapidly spread throughout China in the 1990s, especially for road
and railway embankment applications. Han et al. (3) presented a
detailed review of the state of the practice of deep soil mixing in
China.
The binder material for DJM in China is typically cement, but
occasionally a lime-cement mixture is used. The binder material
content is usually 10% to 15% of the soil weight. Column diameters
are normally 500 to 600 mm, and the maximum depth of improvement has been 15 m. The column pattern is generally triangular with
plan area ratios (Acolumns/Atotal) between 0.10 and 0.25. Columns are
constructed by mixing the soil with cement one to three times (the
mixing equipment goes all the way down and back up). However,
cement is only added during the initial mix. Small pilot studies that
include in situ column testing are typically conducted before production installation to determine the optimum cement contents for a
particular site and soil conditions. To ensure quality control, a load
S. Liu, Institute of Geotechnical Engineering, Southeast University, Nanjiang,
210096, Peoples Republic of China. R. D. Hryciw, Department of Civil and
Environmental Engineering, University of Michigan, 2340 GG Brown, Ann Arbor,
MI 48109-2125.

transducer measures the weight of cement applied per length of column. Mix uniformity is verified by in situ testing, as will subsequently
be discussed.
Many theoretical and empirical methods for calculating settlements and analyzing the stability of DJM-improved soils were
developed during the last decade (4, pp. 125153). Although the
modulus and strength of the soil-cement column material can be
measured by laboratory tests, there can be large differences between
the performance of samples prepared in the laboratory and field performance of full columns. Thus, a dynamic column penetrometer
was developed in Sweden and Finland by Halkola (5) for investigating column strength and integrity in situ. The cone penetration
test (CPT), vane penetrometer test, and other in situ tests are also
finding their way into practice for in situ testing of soil-cement
columns. However, these methods are not sufficient to evaluate soilcement columns (6, pp. 285294), because they can only be used in
the upper portion of each column where penetration is not met with
refusal. Furthermore, none of the above-mentioned tests provides a
sample of the DJM column for visual inspection.
In this paper, use of the standard penetration test (SPT) for estimating the strength and split-spoon sampling for visual inspection
of the DJM soil-cement column material is presented. The distribution of strength along the length of soil-cement columns is shown.
The relationships between SPT blow count and unconfined compressive strength are developed, and the engineering applications of
soil-cement columns are discussed.
SOIL-CEMENT STRENGTH ALONG
LENGTH OF COLUMNS
The shear strength or unconfined compressive strength of a soilcement column is a function of many factors, including soil type,
binder content, construction method, time, and the ambient environment, specifically temperature. To comprehensively investigate
soil stabilization by DJM soil-cement columns, SPTs were performed on cement columns in the LianYunGong section of the LianHuo Freeway in China. The cement columns were 0.5 m in diameter
and the lengths ranged from 8 to 13 m corresponding to the typical
thickness of an upper soft clay deposit. Table 1 summarizes the range
of properties of the clay encountered on the project. The amount of
cement used ranged from 55 to 75 kg/m depending on the soil water
contents. As water content increased, the percentage of cement was
increased.
Because of installation, the central 100-mm core of a soil-cement
column is not representative of the overall column. Therefore, SPTs
were performed at a distance of two-fifths of the column radius

http://www.paper.edu.cn

48

Transportation Research Record 1849

Paper No. 03- 3116

TABLE 1

The SPT blow counts and the unconfined compressive strengths


along the lengths of all the cement columns at different curing times
are shown in Figures 1 and 2, respectively. The data were collected
from 43 soil-cement columns (with slightly different soils and different cement contents), and thus data scatter is significant. Considerably
greater uniformity is observed in single DJM columns as shown by
the N- and qu-values in Table 2. Other columns showed similar uniformity. Despite the scatter in the composite data shown in Figures 1
and 2, there is a clear increase in strength observed with time. These
observations agree with test results from Norway (7, pp. 295301).

Soil Properties in LianYunGong Section of LianHuo Freeway


15.4 kN/m3 to 17.5 kN/m3

Unit Weight
Void Ratio

1.0 to 2.1

Water Content

46.5% to 85.0%

Plasticity Index

16 to 38

Liquid Limit

40 to 65

Clay Fraction (<2 microns)

50% to 60%

Coefficient of Consolidation

0.3 x 10-3 cm2/s to 3.7 x 10-3 cm2/s

Unified Soil Classification

CH and CL with occasional organics

RELATIONSHIPS BETWEEN N AND qu


The relationships between N and qu are shown in Figure 3. Linear
regression yields the following empirical equations (R2-values are
shown in Figure 3):

(0.10 m) from the center. Disturbed samples from the split-spoon


sampler were visually inspected for cement mix uniformity. In addition, undisturbed 88-mm samples were taken every 1.5 m along the
length of the columns by dry coring. The cores were protected by a
tube and sealed with wax for transport to the laboratory, where
unconfined compressive strength tests were performed. The cores
were taken from just above the SPT depths. Thus, each reported
unconfined compressive strength (qu) corresponds to an SPT blow
count (N ) at approximately the same location.

qu ( kPa ) = 6.8 N + 20

(7 days)

(1a)

qu ( kPa ) = 8.0 N + 150

(14 days)

(1b)

qu ( kPa ) = 10.0 N + 270

(28 days)

(1c)

Standard Penetration Test N-Value

Depth (m)

10

15

20

25

30

10

15

20

25

30

10

15

20

25

30

10 15 20 25 30

10

10

10

10

10

12

FIGURE 1

12

12

12

(a)

(b)

10 15 20 25 30

12

(c)

(d)

(e)

Variations of SPT blow count along columns: at (a) 7 days, (b) 14 days, (c) 28 days, (d) 60 days, and (e) 90 days.
Unconfined Compressive Strength qu (MPa)

Depth (m)

0.5

0.5

1.5

0.5

1.5

0.5

1.5

10

10

10

10

10

12

12

(a)
FIGURE 2

(c)

1.5

12

12

12

(b)

0.5

(d)

(e)

Variations of unconfined compressive strength along columns at (a) 7 days, (b) 14 days, (c) 28 days, (d) 60 days, and (e) 90 days.

http://www.paper.edu.cn

Liu and Hryciw

Paper No. 03- 3116

TABLE 2

N and q u for Three Typical DJM Columns

Column 23-12

Column 19-16

Column 14-20

qu
(kPa)

qu
(kPa)

qu
(kPa)

1.5

17

455

16

507

23

585

4.0

16

520

22

640

21

620

6.0

15

486

18

523

18

495

8.0

14

430

17

540

16

508

10.0

14

452

20

621

19

465

Depth
(m)

0.6

0.2

qu (MPa)

qu (MPa)

qu= 6.8N+ 20

R2=0.62

qu= 8N + 150

R2=0.52

0.4

0.1
0.2

0
0

10

15

20
SPT-N

10

(a)

20

25

30
35
SPT-N

(b)

1
qu = 8.3N + 365

qu = 10N + 270

0.8

0.8

qu (MPa)

R =0.37
qu (MPa)

15

0.6
0.4
0.2

R2=0.14

0.6
0.4
0.2

0
10

15

20

25

30

35

10

15

20

SPT-N

25

30

(c)

(d)

qu (MPa)

1
qu = 6N+ 445

0.8
0.6
0.4

R2=0.13

0.2
0
10

35

SPT-N

15

20

25

30
35
SPT-N

(e)
FIGURE 3 Relationships between SPT-N and q u at (a) 7 days, (b) 14 days, (c) 28 days, (d ) 60 days, and
(e) 90 days.

49

http://www.paper.edu.cn

50

Transportation Research Record 1849

Paper No. 03- 3116

qu ( kPa ) = 8.3 N + 365

(60 days)

(1d )

700

qu ( kPa ) = 6.0 N + 445

(90 days)

(1e)

600

qu ( kPa ) = AN + quo

500

q u (kPa)

As expected, the unconfined compressive strengths increase with


SPT blow count at all times. Furthermore, the slopes of the lines corresponding to different ages are almost identical, as observed in the
comparison of regression lines in Figure 4.
Equations 1a through 1e have the following general form:

100

where A is the slope of the linear regression, and quo is the unconfined compressive strength extrapolated to N = 0 at each age, t (in
days). Figure 5 illustrates the functional relationship, quo, which can
be fit very well by the following equation:
(3)

Since the lines in Figure 5 are virtually parallel, an average slope


value of A = 8 can be assumed. Finally, combining Equations 3 and
2 obtains
qu ( kPa ) = 8 N + 162 ln t (days) 286

( 4)

In the foregoing equations N is the SPT blow count modified


from the raw blow count (Ni) according to Chinese practice (8).
This practice requires multiplying Ni by the following amounts for
the rod lengths specified: 1.00 for 3 m, 0.92 for 6 m, 0.86 for 9 m,
0.81 for 12 m, 0.77 for 15 m, and 0.73 for 18 m. Although in China
this procedure is referred to as a rod-length correction, this is not
the same rod-length correction used in the United States. In fact, the
Chinese correction actually appears to be a correction for overburden pressure. The SPTs were performed using a 63.5-kg automatic
trip hammer.
Equation 4 has two practical uses. First, it can predict the unconfined compressive strength of soil-cement column material at any
time t based on an SPT performed on the column at t. Second, it
can also be used to conservatively estimate the strength at a future
time t based on an SPT performed at an earlier time. The conser-

300
200

(2)

quo ( kPa ) = 162 ln t (days) 286

400

0
0

20

40

60

80

100

Time (days)
FIGURE 5 Variation of unconfined compressive strength
with time.

vatism stems from the fact that N actually increases with time, and
therefore the first term in Equation 4 should gradually increase.
Nevertheless, the second term in Equation 4 appears to account for
most of the observed time effects, especially if the SPT is performed after 14 days. If the SPT is performed at only 7 days, the
authors recommend doubling the value of N for estimation of qu
beyond 28 days.
Equation 4 is very practical for finding the right design parameters,
such as optimum cement contents for a given soil based on 7- or
14-day tests on pilot soil-cement columns.

ENGINEERING APPLICATIONS
Foundations
For foundations, the bearing capacity of the soil-cement column system must be evaluated. Broms (9, pp. 177184) evaluated shortterm capacity based on cavity expansion theory and indicated that
when u (column) = 30 and h < 100 to 150 kPa,
p = qu + 3 h

700

120

( 5a )

and when u (column) = 0 and h > 100 to 150 kPa,

qu (kPa)

600

p = qu + h

500

90 d

400

60 d

300

28 d

200

14 d

where
p = bearing capacity of the cement column,
h = horizontal stress on the column, and
u = undrained angle of internal friction.

100

However, for shallow failures of foundations, qu is much larger


than h and therefore it can conservatively be assumed that

7d
0
0

(5b)

10

20

30

40

SPT-N
FIGURE 4 Comparison of best-fit relationships between SPT-N
and q u at 7, 14, 28, 60, and 90 days (d  days).

p qu

(6)

Liu et al. (10, pp. 153158) found that boring and sampling of the
soil-cement column causes a considerable decrease in the uncon-

http://www.paper.edu.cn

Liu and Hryciw

Paper No. 03- 3116

fined compressive strength of recovered specimens compared with


those cured in the laboratory. The strengths of the laboratory-cured
specimens reflect the in situ strengths. Thus, the strength predicted
by Equation 4 can be increased by a factor, k, to account for the
strength loss due to sampling:

Slopes and Embankments

p = k (8 N + 162 ln t 286)

sp = a p + (1 a) s

For slopes or embankments stabilized by DJM soil-cement columns,


the global shear resistance sp corresponds to the weighted average
shear strength of columns and unstabilized soils:

( 7)

p = 0.5qu

CONCLUSIONS

where

A large number of SPTs and laboratory unconfined compressive


strength tests have been collected on cement columns formed by
DJM for soft ground improvement. The strength characteristics
along the lengths of the cement columns were studied. The following
conclusions can be drawn:

sp = bearing capacity of composite ground,


s = bearing capacity of unstabilized soils, and
a = ratio of column area to total area (0.10 to 0.25).
Considering the relatively uniform observed N-values along the
lengths of cement columns, the average SPT blow count can be used
in Equation 7. However, should N-values be significantly lower in
the upper portion of the column, more deference should be given to
these values.
Table 3 shows a comparison of bearing capacities of single columns
determined by SPT and the static load test (SLT), where the SPT was
performed following the SLT for the same column. There appears to
be excellent agreement between the load test results and predictions by
Equation 7.
The settlement of DJM soil-cement columns can be predicted in
several ways. The Youngs modulus of the column, E, could be estimated using empirical relations between modulus and unconfined
compressive strength such as E/qu = 250, where qu is computed by
Equation 4. Alternatively, since SPT N-values are available, direct
correlations between E and SPT-N could be employed to compute
E for the soil-cement column.

Column
No.

16-13
10-11
11-19
13-9

(10)

where qu can be obtained by Equation 4 times the index of disturbance,


k = 2.5.

(8)

TABLE 3

( 9)

where p is the shear strength of the cement column and s is the


shear strength of the unstabilized soils.
Because the permeability of the cement column is very low, the
undrained shear strength can be used for the stability analysis (9,
pp. 177184). Therefore,

where k is an index of disturbance equal to the ratio of undisturbed (laboratory or DJM column) strength to disturbed strength.
For the soils in this study, k was found to be 2.5 (10, pp. 153
158). This value can conceivably be extended to other CH and CL
soils. For different soil types, other values of k would have to be
established.
Finally, since loading is distributed over both the column and the
surrounding soil, the bearing capacity depends on contributions of
both the column and the unstabilized soil. Thus, for engineering
applications, it may be estimated by a weighted volume average of
the column capacity and the soil bearing capacity:
sp = a p + (1 a) s

51

1. The SPT is an economic and effective method to evaluate


the strength and quality of stabilized soil-cement columns. The
test is highly recommended as the quality control method for
DJM.
2. The strength parameters along the cement columns are generally homogeneous, with a logarithmic increase in strength with time.
3. Unique empirical relationships between SPT blow count N
and unconfined compressive strength qu of the cement column were
determined for various curing times.
4. The SPT results can be used for determining the bearing
capacity both at the time of testing and by procedures established
herein, in the future. This conclusion makes the SPT particularly
useful for analysis of pilot test results.
5. The shear resistance and modulus for stability analysis and
settlement of DJM can likewise be estimated.

Comparison of Bearing Capacities Determined with SPT and SLT


Time After
Installation
(days)

Column
Length
(m)

Column
Diameter
(mm)

Average
N

63

13

500

45

12

67
46

Bearing Capacity (kN)


By
SPT

By
SLT

Error

13

240

240

500

16.4

227

270

15%

13

500

15

252

270

7%

13

500

14

219

240

8%

http://www.paper.edu.cn

52

Paper No. 03- 3116

Transportation Research Record 1849

ACKNOWLEDGMENTS

4. Broms, B. B. Design of Lime, Lime/Cement, and Cement Columns. In


Proc., International Conference on Dry Mix Methods for Deep Soil
Stabilization, Balkema, Rotterdam, 1999.
5. Halkola, H. In Situ Investigations of Deep Stabilized Soil. In Proc., 8th
European Conference on Soil Mechanics and Foundation Engineering,
Balkema, Rotterdam, 1983.
6. Halkola, H. Quality Control for Dry Mix Methods for Deep Soil Stabilization. In Proc., International Conference on Dry Mix Methods for
Deep Soil Stabilization, Balkema, Rotterdam, 1999.
7. Braaten, A., R. Aaboe, and F. Oset. Development of In Situ Control
Methods for Lime/Cement Columns. In Proc., International Conference
on Dry Mix Methods for Deep Soil Stabilization, Balkema, Rotterdam,
1999.
8. Ministry of Construction, China. National Code of Geotechnical Investigation (in Chinese). Chinese Construction Industrial Press, Beijing, 1995.
9. Broms, B. B. Progressive Failure of Lime, Lime/Cement, and Cement
Columns. In Proc., International Conference on Dry Mix Methods for
Deep Soil Stabilization, Balkema, Rotterdam, 1999.
10. Liu, S. Y., M. L. Shi, Y. Xu, Y. J. Du, and H. Zhu. New Approaches to
Evaluate Deep Mixing Piles in Soft Ground Improvement. In Proc.,
International Symposium on Lowland Technology, Saga, Japan, 1998.

Sponsorship from the Jiangsu Provincial Highway Administration


for this study is gratefully acknowledged. The field and laboratory
efforts of L. C. Miao, R. M. Li, S. Li, and D. W. Zhang and others at
Southeast University were essential to the successful completion of
the research.

REFERENCES
1. Bruce, D. A., M. E. C. Bruce, and A. F. Dimillio. Dry Mix Methods:
A Brief Overview of International Practice. In Proc., International
Conference on Dry Mix Methods for Deep Soil Stabilization, Balkema,
Rotterdam, 1999.
2. Gong, X. N. Composite Foundation (in Chinese). Zhejiang University
Press, Hangzhou, China, 1992.
3. Han, J., H.-T. Zhou, and F. Ye. State-of-Practice Review of Deep Soil
Mixing Techniques in China. In Transportation Research Record: Journal of the Transportation Research Board, No. 1808, TRB, National
Research Council, Washington, D.C., 2002, pp. 4957.

Publication of this paper sponsored by Committee on Transportation Earthworks.

S-ar putea să vă placă și