otechnique 59, No. 7, 637641 [doi: 10.1680/geot.8.T.020] Chai, J.-C., Matsunaga, K., Sakai, A. & Hayashi, S. (2009).

Ge

TECHNICAL NOTE

Comparison of vacuum consolidation with surcharge load induced consolidation of a two-layer system
J. - C . C H A I * , K . M AT S U NAG A , A . S A K A I * a n d S . H AYA S H I * Laboratory consolidation tests for a two-layer soil system under oedometer conditions were conducted under both vacuum pressures and surcharge loads. The effects of the order of soil layers on the behaviour of vacuum consolidation of the two-layer soil system have been investigated by comparing vacuum pressure and surcharge load induced consolidations. Under one-way drainage conditions, for both a surcharge load and a vacuum pressure, the order of soil layers only inuences the rate of consolidation but not the nal settlement. When a layer with a relative lower value of hydraulic conductivity (k) is located immediately adjacent to the drainage boundary, the consolidation rate is slower. However, for vacuum pressure applied under two-way drainage conditions, the order of the soil layers not only inuences the rate of consolidation but also the magnitude of settlement.
KEYWORDS: consolidation; laboratory tests; settlement

Des tests de consolidation en laboratoire pour un sys` me de sol a ` deux couches, avec un oedome ` tre, sont te s en pre sence de pressions sous vide et de effectue surcharges. On examine les effets de lordre des couches de sol sur le comportement de la consolidation ` me a ` deux couches en comparant les sous vide du syste consolidations induites par pression sous vide et surcharge. Avec un drainage unidirectionnel, et aussi bien avec pression sous vide et surcharge, lordre des couches de sol ninue que la vitesse de consolidation, ` ve une et non pas le tassement nal. Lorsquon rele sentant une conductivite hydraulique (k) a ` couche pre rieure en un point adjacent a ` la valeur relative infe roule plus limite de drainage, la consolidation se de lentement. Toutefois, en cas dapplication de la pression sous vide avec un drainage bidirectionnel, lordre des couches de sol inue non seulement sur la vitesse de galement sur la magnitude du consolidation, mais e tassement.

INTRODUCTION Vacuum consolidation as a preloading method has received attention recently (Bergado et al., 1998; Tang & Shang, 2000; Tran et al., 2004). Chai et al. (2005a; 2005b) discussed the characteristics of vacuum consolidation using the results of a laboratory odometer test with a uniform soil layer under both vacuum pressures and surcharge loads. The response of a layered depositnatural deposits are often layeredunder a vacuum pressure has not yet been the subject of comprehensive experimental and theoretical investigation. Hence, in order effectively to design a vacuum consolidation system, there is a need to clarify the characteristics of vacuum consolidation of a two-layer soil system. A series of laboratory tests involving vacuum pressure and surcharge load induced consolidation under oedometer conditions were conducted to investigate the fundamental behaviour of vacuum consolidation of a two-layer soil system. The effects of drainage condition (one-way or two-way) and the relative hydraulic conductivity (k) of the soil layer immediately adjacent to the boundary where vacuum pressure is applied were studied with reference to the amount of settlement and the rate of the consolidation.

A BRIEF DISCUSSION ON THE CHARACTERISTICS OF VACUUM CONSOLIDATION Vacuum consolidation for a uniform soil layer Chai et al. (2005b) reported the laboratory oedometer test results for a uniform soil layer with both surcharge load and vacuum pressure. The conditions adopted can be found from Chai et al. (2005b), and the main ndings are as follows. (a) Settlements induced by vacuum pressure under one-way drainage conditions. For vacuum consolidation, if inward lateral displacement occurs, the settlement induced by a vacuum pressure will be less than that of a surcharge load with the same magnitude. Chai et al. (2005b) dened a stress ratio kr as follows kr vac vac v 90 (1)

Manuscript received 16 May 2008; revised manuscript accepted 14 November 2008. Published online ahead of print 17 March 2009. Discussion on this paper is welcomed by the editor. * Saga University, 1 Honjo Saga 840-8502, Japan Chuoh Consultants (formerly Saga University), 6-1-1 Arado, Chuoh-ku, Fukuoka, Japan

so that, if kr < k0 (k0 is at-rest horizontal earth pressure coefcient), there will be no lateral displacement and vice versa. Under oedometer conditions, Svac /Sl denotes the settlement ratio of the nal settlement induced by a vacuum pressure (Svac ) to that by a surcharge load with the same magnitude (Sl ). Fig. 1 (modied from Chai et al., 2005b) shows the relationship between the stress ratio kr and the settlement ratio Svac /Sl obtained from the oedometer test results. It can be seen that when kr . k0 , Svac /Sl increases almost linearly with decreasing kr value. When kr < k0 , Svac /Sl is very close to unity. (b) Effect of drainage boundary condition. In the case of two-way drainage, at the bottom of the sample the excess pore pressure is xed at zero and effectively no vacuum pressure can be applied. It is obvious therefore that vacuum consolidation involving two-way drainage should result in less settlement than one-way drainage. 637

Delivered by ICEVirtualLibrary.com to: IP: 161.139.192.2 On: Wed, 03 Nov 2010 07:28:43

638
10

CHAI, MATSUNAGA, SAKAI AND HAYASHI

Vacuum applied Drained Layer-1 (k1) Layer-2 (k2) 08 Drained Flow direction (a) 07 k2 k1 (b) vac Vacuum applied Drained Layer-2 (k2) Layer-1 (k1) Drained vac

Stress ratio, kr vac/(vac v0)

09

Fig. 2. Illustration of vacuum pressure distribution in two-layer soil system: (a) layer-1layer 2; (b) layer 2layer 1

06

05 Assumed k0 value 04 06

07

08 09 Settlement ratio, Svac/Sl

10

11

Fig. 1. Stress ratio kr plotted against ration Svac /Sl (modied from Chai et al., 2005b)

Under a vacuum pressure, for both one-way and twoway drainage conditions, water is drained out of a soil sample only at the boundary where the vacuum pressure is applied. Therefore, for both the cases, the rate of consolidation should theoretically be the same.

General discussion on the characteristics of vacuum consolidation for two-layer soil system Generally, the effect of initial stress ( v 90 ) in a soil sample on the characteristics of consolidation of a two-layer soil system with a vacuum pressure should be similar to that of a uniform soil layer. However regarding the effect of drainage boundary condition, it is strongly affected by the order of soil layers relating to the boundary where the vacuum pressure is applied. (a) One-way drainage. The nal state will be a uniform vacuum pressure in the soil layers and no ow in the system. The order of soil layers will therefore not inuence the nal settlement. It will, however, inuence the rate of consolidation. Under the condition of no lateral displacement occurring, when a layer with a relative lower k value is located at the drainage boundary, the consolidation rate will be slower; this is the same for a surcharge load case (e.g. Pyrah, 1996). (b) Two-way drainage. Under a two-way drainage condition with a vacuum pressure, the nal state involves a steady water ow towards the boundary where the vacuum pressure is applied. In a two-layer soil system, to satisfy the ow continuity at steady state, the following equation must be hold i1 k v1 i2 k v2 (2) where i1 and i2 are the hydraulic gradients of layer-1 and layer-2, and kv 1 and kv 2 are the vertical hydraulic conductivities of layer-1 and layer-2, respectively. As can be seen from equation (2), a layer with a lower k value must have a higher i value in order to maintain the continuity of ow. Then, changing the order of the soil layer will not only inuence the rate of consolidation, but also the distribution of vacuum pressure in the system as illustrated in Fig. 2. The nal settlement will be a function of both the relative k values and the compression indexes of the soil layers (Cc ).

LABORATORY TEST PROGRAMME FOR TWO-LAYER SOIL SYSTEM The equipment used was a Maruto multiple oedometer apparatus (manufactured in Tokyo, Japan) (Chai et al., 2005b). The equipment has ve consolidation cells, which can be used either as individual consolidation cells or connected to form a 5-layer system. Each sample is 60 mm in diameter and typically 20 mm in thickness. In this study two cells were connected to form a two-layer soil system as illustrated in Fig. 3. For the case of an applied surcharge load, the same amount of load was applied at the tops of both layer-1 and layer-2. For the case of a vacuum pressure loading, however, the desired amount of vacuum pressure was only applied at the top of layer-1. The soil preparation procedure, initial stress and boundary conditions, and the magnitude of the load applied are the same as that of the tests for a uniform soil layer (Chai et al., 2005b). To make a direct comparison between a surcharge load and a vacuum pressureinduced consolidation for both one-way and two-way drainages, only cases where v 90 80 kPa were tested. As shown in Fig. 1, for v 90 80 kPa (kr 0.5), no or very minimal lateral displacement is expected for a vacuum pressure increment of 80 kPa. Two types of soil were used. One was reconstituted Ariake clay, and another was reconstituted mixed soil consisting of 50% Ariake clay and 50% sand (passing 2 mm sieve) by dry weight. The grain size distribution curves of the Ariake clay and the sand used are given in Fig. 4. Some of the physical and mechanical properties of the soils are listed in Table 1. The values of k and Cc were deduced from the standard oedometer test results. The k values listed represent average values in the normally consolidated region. At an average effective vertical stress of 100 kPa, the coefcient of consolidation (Cv ) of the mixed soil sample is about 4 times of that of the Ariake clay sample. The tests conducted are listed in Table 2. TEST RESULTS FOR TWO-LAYER SOIL SYSTEM One-way drainage The settlement/time curves under one-way drainage are compared in Figs 5 and 6 for the surcharge load and the vacuum pressure cases respectively. It can be seen that with v 90 80 kPa, the settlements induced by the vacuum pressure are almost the same as those induced by a surcharge load of the same magnitude. Also, the order of the soil layers did not have an inuence on the nal settlement. Both
Layer-1 Layer-2

Tube connection

Fig. 3. Two-layer system

Delivered by ICEVirtualLibrary.com to: IP: 161.139.192.2 On: Wed, 03 Nov 2010 07:28:43

COMPARISON OF VACUUM CONSOLIDATION WITH SURCHARGE LOAD INDUCED CONSOLIDATION

100 80 60 40 20 0 0001 20 01 00

639

Percent finer by weight

05

Ariake clay Sand

Settlement: mm

10

15

MC, one-way (S-2) CM, one-way (S-1) Two-way (S-3) 1 10 100 Elapsed time: min 1000 10 000

001

01 1 Diameter: mm

10

100

Fig. 4. Grain size distribution curves

Fig. 5. Settlement/time curves (two-layer, surcharge load)

00

gures show, however, that the order of the soil layers did have an obvious effect on the rate of consolidation. The case where the mixed soil sample was located at the drainage boundary (V-2 and S-2 in Table 2) had a faster consolidation rate. For the vacuum consolidation case, the time for 50% of the nal settlement reached (ts50 ) for C+M case was about 37 min and for M+C case was about 28.5 min, and these times will be compared later with those for the two-way drainage cases. The measured excess pore pressures at the bottoms of layer-1 (u1 ) and layer-2 (u2 ) are given in Figs 7 and 8 for the surcharge load and the vacuum pressure cases respectively. It can be seen that u1 dissipated much faster in the M+C case (S-2 and V-2) than in the C+M case (S-1 and V-1). For the surcharge load cases, at the interface of layer-1 and layer-2, the time for u1 to reduce to 40 kPa (half of the applied surcharge load) was 9 min for the M+C case (S-2) and 62 min for the C+M case (S-1). For the case of vacuum pressure, the time for u1 to reach 40 kPa was about 5 min and 35 min for the M+C (V-2) and the C+M (V-2) cases, respectively. Obviously, the vacuum pressure propagated faster than the dissipation of the surcharge load induced excess pore pressure at the early stages of consolidation. The reason considered is the possibility of very small inward
Table 1. Physical properties of the soil samples Soil particles: % Clay: ,5 m Ariake clay Mixed soil 31.0 Silt 67.8 Sand 1.2 13.9 18.1 Unit weight, t : kN/m3

Settlement: mm

05

10 CM, one-way (V-1) MC, one-way (V-2) CM, two-way (V-3) MC, two-way (V-4) 1 10 100 Elapsed time: min 1000 10 000

15

20 01

Fig. 6. Settlement/time curves (two-layer with one-way drainage, vacuum pressure)

lateral displacement occurring in the upper soil sample (layer-1) under the vacuum pressure, which could form micro-gaps between the conning ring and the soil sample causing an apparent increase in the hydraulic conductivity of the sample. In layer-1 and layer-2, the average hydraulic gradients (i) can be estimated as j u0 u1 j= h and

Liquid limit, Wl: %

Plasticity limit, Wp: % 57.5

Void ratio, e0

Compression Hydraulic index, Cc conductivity, k: 109 m/s 0.88 0.21 1.44 3.13

116.6

3.63 1.11

Table 2. Summary of the tests conducted No. Sample condition Soils V-1 V-2 V-3 V-4 S-1 S-2 S-3
y

Drainage condition Type One-way Two-way One-way Two-way

Consolidation pressure Magnitude: kPa 80

Remarks

v 90 : kPa 80 80 80 80 80 80 80

C+M y M+C C+M M+C C+M M+C C+M

Vacuum pressure

Two-layer

Surcharge load

80

9 is initial vertical effective stress. v0 C+M: the rst layer is the Ariake clay and the second layer is the mixed soil; M+C: reverse the order of the soil layers.

Delivered by ICEVirtualLibrary.com to: IP: 161.139.192.2 On: Wed, 03 Nov 2010 07:28:43

640
80

CHAI, MATSUNAGA, SAKAI AND HAYASHI

u1 (S-1) u2 (S-1) u1 (S-2) u2 (S-2) u0 u1 u2

60

40

20

0 101

100

101 102 Elapsed time: min

103

Fig. 7. Measured excess pore pressure variations (surcharge load and one-way drainage)

0 u1 (V-1) u2 (V-1) u1 (V-2) u2 (V-2) u0 u1 u2

Vacuum pressure: kPa

20

40

60

80 101

100

101

102

103

Elapsed time: min

Fig. 8. Measured vacuum pressure (vacuum pressure and oneway drainage)

j u2 u1 j= h (h is the thickness of the layer, and u0 is 0 for the surcharge load case and 80 kPa for the vacuum pressure case). It can also be observed that during the whole consolidation process, for the C+M case, the difference between u2 and u1 is very small, which implies that there was a higher hydraulic gradient (i) in the clay sample than that in the mixed soil sample. For the M+C case with a surcharge load, when the elapsed time was longer than about 60 min, ju2 u1 j became larger than j u0 u1 j (Fig. 7), that is, the i values became larger in the clay sample (layer-2) than in the mixed soil sample, and for the vacuum pressure case when the elapsed time is longer than about 30 min, j u2 u1 j became larger than j u0 u1 j (Fig. 8), the i values in the clay sample were larger or almost equal to those in the mixed soil sample. Two-way drainage Under a surcharge load with two-way drainage, the order of the soil layers should have no inuence on both the magnitude of settlement and the rate of consolidation. Only one test (S-3 in Table 2) was therefore conducted, which represents both the C+M and the M+C cases. The settlement/time curve for this case is also given in Fig. 5. The rate of consolidation is much faster, and the nal settlement is almost the same when compared with that of the oneway drainage. Under vacuum pressure with two-way drainage, however, the order of the soil layers not only inuenced the rate of consolidation but also the magnitude of settlement. The settlement/time curves are included in Fig. 6. The M+C case (V-4) had a faster consolidation rate and

a larger nal settlement. The nal settlement was 0.80 mm for the C+M case (V-3) and 0.94 mm for the M+C case (V4), and the values of ts50 were 43.3 min and 28.5 min respectively. The values of ts50 are comparable with the corresponding values of a two-layer soil system under oneway drainage conditions as discussed previously (37.0 min and 28.5 min accordingly), and it supports the statement that under vacuum pressure, the rates of consolidation for one-way and two-way drainage should be the same if the coefcient of consolidation of soil sample is a constant during the consolidation. The measured u1 variations are given in Fig. 9. It can be seen that M+C case had a much higher j u1 j values than that of C+M case. At the elapsed time of 48 h (2880 min), the measured u1 values were 19.0 for the C+M case (V-3) and 45.6 kPa for the M+C case (V-4) (52.0 kPa at time t 60 min and increased to 45.6 kPa at t 2880 min). Using equation (2), the ratios of hydraulic conductivity of the mixed soil sample (km ) to that of the clay sample (kc ) (km /kc ) are about 3.2 for the C+M case and 1.9 for the M+C case (using u1 of 52.0 kPa). All these numbers are comparable with the ratio of 2.2 obtained using the k values given in Table 1. Theoretically, if kc and km are constants, the km /kc ratios for both cases should be the same. There are two reasons to explain the difference in km /kc ratios. First, the k value of a soil relates to its void ratio (e), and a reduction in void ratio will result in a reduction of k value (e.g. Taylor, 1948). Compressions of the clay samples for both the C+M and the M+C cases are compared in Fig. 10(a). This gure shows that the clay sample had more compression (more void ratio reduction) in the C+M case (V-3) than that in the M+C case (V-4), and therefore a smaller kc value for the C+M case. Further, the compression of the mixed soil sample in the C+M case was less than that in the M+C case (Fig. 10(b)), and therefore a larger km value for the C+M case. Consequently, the C+M case had a larger km /kc value. Second, for the M+C case, there was a possibility of leakage of vacuum pressure for t . 60 min owing to the possible vacuum pressure induced inward lateral displacement of the samples, which tends to reduce the km /kc ratio. The mixed soil sample (M) may have a larger internal friction angle (9) than the Ariake clay sample (C), which possibly implies a smaller at-rest earth pressure coefcient, ko . The condition for inward lateral displacement to occur can be written as follows (Chai et al., 2005b) vac . k o v 9o 1 ko (3)

Excess pore pressure: kPa

Assuming 328 for the mixed soil sample, and k o 1 sin 9, with 9vo 80 kPa, equation (3) yields a
0 V-3 (CM)

Vacuum pressure: kPa

20

40

V-4 (MC)

60

Vacuum pressure between 2-layers (u1) 100 101 102 Elapsed time: min 103

80 101

Fig. 9. Measure vacuum pressure between two layers (two-layer with two-way drainage, vacuum pressure)

Delivered by ICEVirtualLibrary.com to: IP: 161.139.192.2 On: Wed, 03 Nov 2010 07:28:43

COMPARISON OF VACUUM CONSOLIDATION WITH SURCHARGE LOAD INDUCED CONSOLIDATION

00

641

02

04

06 Compression of clay samples Vacuum pressure 2-layer, 2-way drainage 01 1 10 100 Elapsed time: min (a)

08

CM MC 1000

10

00

02

Compression: mm

(a) One-way drainage. For a two-layer soil system with one-way drainage tested with both surcharge load and vacuum pressure, the order of soil layers only inuences the rate of consolidation but not the nal settlement. When a layer with a relative lower value of hydraulic conductivity (k) is located at the drainage boundary, the consolidation rate is slower. (b) Two-way drainage. It is well known that for two-way drainage with a surcharge load, the order of soil layers has no inuence on both the rate of consolidation and the nal settlement. However, for two-way drainage with a vacuum pressure (vacuum pressure applied at one face only), the order of the soil layers not only inuences the rate but also the magnitude of the consolidation settlement. This is because for two-way drainage with a vacuum pressure, the nal state is a steady water ow towards the boundary where the vacuum pressure is applied. To satisfy the ow continuity condition, the vacuum pressure distribution in the soil layers is not only a function of relative values of the hydraulic conductivity, but also a function of the order of the soil layers.

Compression: mm

04

06 Compression of mixed samples Vacuum pressure 2-layer, 2-way drainage 01 1 10 100 Elapsed time: min (b) CM MC 1000

08

10

ACKNOWLEDGEMENTS This research has been partially founded by the programme of Grants-in-Aid for Scientic Research, Japan Society for Promotion of Science (JSPS) under grant No. 18560488. Sincere thanks are extended to Professor J. P. Carter at the University of Newcastle, Australia, for his valuable comments/suggestions during the preparation of this technical note.

Fig. 10. Comparison of the compressions of soil samples in a two-layer soil system under vacuum pressure and two-way drainage condition: (a) compressions of the clay samples; (b) compressions of the mixed samples.

REFERENCES
Bergado, D. T., Chai, J.-C., Miura, N. & Balasubramaniam, A. S. (1998). PVD improvement of soft Bangkok clay with combined vacuum and reduced sand embankment preloading. Geotech. Engng 29, No. 1, 95121. Chai, J.-C., Hayashi, S. & Carter, P. J. (2005a). Characteristics of vacuum consolidation. Proc. 16th ICSMGE, Osaka 3, 1167 1170. Chai, J.-C, Carter, J. P. & Hayashi, S. (2005b). Ground deformation induced by vacuum consolidation. J. Geotech. Geoenviron. Engng, ASCE 31, No. 12, 15521561. Pyrah, I. C. (1996). One-dimensional consolidation of layered soils. Geotechnique 46, No. 3, 555560. Tang, M. & Shang, J. Q. (2000). Vacuum preloading consolidation of Yaogiang Airport runway. Geotechique 50, No. 6, 613623. Taylor, D. W. (1948). Fundamentals of soil mechanics. New York: John Wiley. Tran, T. A., Mitachi, T. & Yamazoe, N. (2004). 2D nite element analysis of soft ground improvement by vacuum-embankment preloading, Proc. 44th Annual Conference of Japanese Geotechnical Society, Hokkaido Branch, 127132.

no lateral displacement vac value of less than about 65 kPa. Therefore for an applied vacuum pressure of 80 kPa, it was highly possible that some kind of inward lateral displacement occurred. The direct evidence is the slight reduction of the vacuum pressure at the interface of the soil layers after t . 60 min (Fig. 9). CONCLUSIONS Laboratory consolidation tests under oedometer conditions were conducted with both vacuum pressures and surcharge loads for a two-layer soil system. The effect of the order of soil layers on the characteristics of vacuum consolidation has been investigated. Based on the results of the laboratory tests and theoretical interpretations, the following conclusions can be drawn.

Delivered by ICEVirtualLibrary.com to: IP: 161.139.192.2 On: Wed, 03 Nov 2010 07:28:43