Subsidence prediction of compacted subgrade soil (Kanto loam soil)using cyclic traffic loading experiments

Minagata, H., Yoshizaki, K. & Hagiwara, N.Tokyo Gas Co., Ltd., Pipeline Technology Center

Akagi, H.Professor of Civil Engineering, Faculty of Science and Engineering, Waseda University

Keywords: settlement, volcanic cohesive soil, partially saturated soil, repeated load, compaction, road base

ABSTRACT: During road excavations for laying gas distribution pipelines, the reuse of the excavated soil for thesubgrade backfill significantly reduces the environmental impact. However, whenKanto loam soil, which is a localsoil that is widely distributed in the Kanto region, Japan, is employed for the backfill material, the relationshipbetween the amount of the subsidence of the subgrade soil and the cyclic traffic loading must be investigated. Inthis study, cyclic loading tests on compacted Kanto loam soils are conducted to investigate the subsidencecharacteristics by using a confined compression apparatus. The subsidence prediction method for subgrade soil(Kanto loam soil) under cyclic traffic loading is proposed on the basis of the cyclic loading test results. Full-scaletraffic loading experiments are performed on roadbeds backfilled with Kanto loam soil. Consequently, goodcorrelation is obtained between the amount of the predicted subsidence and the measurement results.

1 INTRODUCTION

In Japan, large amounts of soil are dumped intolandfills. As a result, the capacity for landfills anddisposal sites is declining year-by-year, and shortagein capacity is expected in the future. Furthermore, thereare other issues such as the effect on the environmentfrom air pollution caused by CO2, NOx and otheremissions from vehicles transporting soil and sandand from the extraction of new materials.

Confrontedwith this situation, the JapaneseMinistryof Land, Infrastructure and Transport (JMLIT) adoptedthe ‘Construction Recycling Promotion Plan 2002’.In this plan, the ministry set a goal for the re-utilizationof 90% of the construction soil excavated for publicworks projects by 2010.

However, for relatively small-scale road excavationworks such as the laying of gas distribution pipes, theexcavated soil is not reused to backfill the roadbed,since the use of low-quality materials for the backfillcan result in the gradual subsidence of the road surfaceunder cyclic traffic loading. In particular, sincethe strength of Kanto loam soil, which is widelydistributed across the Kanto region, can significantlydecrease after being disturbed by excavation, it is notgenerally considered suitable for backfilling (ExpressHighway Research Foundation of Japan 1973).According to the soil quality classification standardsfor construction soil formulated by the Public WorksResearch Center (1997), Kanto loam soil that has beendisturbed by excavation is classified as a Type 3 orType 4 construction soil depending on its cone index,

which in either casemeans that its use in road (roadbed)banking requires some type of special installation orstabilization technique.

Monismith et al. (1975) and Dingqing et al. (1996)proposed a prediction method for the subsidence ofthe roadbed based on the results of cyclic loadingtriaxial compression tests. According to the results,the amount of road subsidence that occurs due tocyclic traffic loading not only depends on the qualityof the backfill material alone but can also varyaccording to the amount of cyclic traffic loadingafter the work is completed. As a result, even if theexcavated soil is Kanto loam, theremay be situations inwhich the soil quality and traffic loading conditionsallow its use as backfill material. The use ofjust-excavated Kanto loam for backfill can providebenefits such as reduction in the volume of excavatedsoil, environmental impact and construction costs.However, it has not yet been understood whatcombination of backfill material and traffic loadingcould avoid a major subsidence.

In this study, laboratory tests are conducted onKanto loam soil by using repeated applications ofstress that corresponded to actual traffic loading, andthen the amount of road surface subsidence dueto cyclic traffic loading is predicted on the basis ofthe experimental results. Furthermore, a full-scaleexperiment is conducted to investigate cyclic trafficloading on a road backfilled with Kanto loam soil.Then, the measurement results are compared with thepredicted subsidence values to verify the suitabilityof the subsidence prediction method.

Deformational Characteristics of Geomaterials, Burns, Mayne and Santamarina (eds) 327

2 COMPACTION CHARACTERISTICS OFKANTO LOAM SOIL

The roadbed soil conditions during the excavation andsubsequent backfilling of a road affect the subsidenceof a road surface. In order to investigate the roadbedsoil conditions when Kanto loam is used forbackfilling, compaction tests are conducted.

2.1 Soil Physical Properties

Table 1 and Figure 1 show the soil physical propertytest results (based on JGS 0051, JIS A 1202, JIS A1203, JIS A 1205, JIS A 1211 and JIS A 1228) and thegrain size accumulation curve (based on JIS A 1204),respectively, of Kanto loam soil used in the compactiontests. Musashino loam soil, which is known to have thetypical characteristics of Kanto loam soil, is shown tohave a grain density of 2.80–2.88 (g/cm3) and naturalwater content of approximately 100–120 (%), andthese values are similar to those of Kanto loam soil.In addition, when excavated, the cone penetrationindex for this soil is determined to be 334 (kPa);hence, it is classified as a Type 4 soil on the basis ofthe soil-type classification standards for constructionsoil regulated by the MLIT.

2.2 Compaction Test

2.2.1 Test MethodTable 2 shows the conditions required for thepreparation of test samples. To investigate theeffects of water content when compacting Kanto

loam, the water contents are adjusted to 90% (testcases: A1 to A4), 100% (B1 to B4), and 110% (C1toC4). The samples are compacted in three layers (eachlayer of approximately 40mm) in steel cylindricalmolds with inner diameters of 100mm and heightsof 127mm. To investigate the effects of compactionenergy, the number of compaction blows is varied foreach sample. In order to avoid even the slightestvariation in the characteristics due to thermal effects,thewater content in theKanto loam soil for the samplesis reduced at room temperature.

The compaction energy Ec for each sample iscalculated by the following Eq. (1):

Ec ¼ WR � H � NB � NLV ð1Þ

in which Ec:Compaction energy (kJ/m3),WR: Rammer

weight (kN),H:Rammer drop height (m),NB:Numberof compacting blows per layer, NL: Number of layersand V: Volume of compacted sample (m3).

After the completion of soil compaction, therespective dry density (based on JIS A 1225),saturated density and the relationship between thesevalues and the water content are obtained.Furthermore, an unconfined compression test (basedon JIS A 1216) is conducted for the sample preparedunder the same conditions for investigating therelationship among the unconfined compressivestrength, water content and compaction energy.

2.2.2 Test ResultsFigure 2 shows the relationship between compactionenergy and dry density. Comparing Cases A to C, it isdetermined that the lower the water content, the higheris the dry density after compaction. In addition,comparing the compaction energy at 110.4 (kJ/m3)and 882.9 (kJ/m3), it is shown that the lower the watercontent, the greater is the amount of increase in thedry density. These results show that Kanto loam soilwith low water content becomes more effectivelycompacted than that with high water content.

Figure 3 shows the relationship between thecompaction energy and the degree of saturation. Inall the Cases A to C, when the compaction energy is�286.9 (kJ/m3), the degree of saturation lies in therange 94%–97%.

Figure 4 shows the relationship between thecompaction energy and unconfined compressivestrength. It is assumed that the unconfinedcompressive strength increases with lower watercontent. In addition, in Cases A and B, the unconfinedFigure 1. Grain size distribution curve.

Table 1. Soil physical propertiesSoil type Sand mixed with volcanic co-

hesive soil Soil classification type VH2-S Soil grain density s(g/cm3) 2.832 Natural water content(%) 102.5 Plastic limit L(%) 85.3 Liquid limit p(%) 126.7 Plasticity index Ip 41.4 Design CBR(%) 0.7 Cone index (kPa) 334

Table 2. Sample preparation conditionsCase A1 A2 A3 A4 B1 B2 B3 B4 C1 C2 C3 C4 Sample Sand mixed with volcanic cohesive soil Water content w (%) 90.6 88.9 89.2 90.1 100.8 101.5 99.3 98.2 110.8 108.9 109.1 109.4Rammer weight W (kN) 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025Rammer drop height H (m) 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 Number of compaction blows per layer N 5 13 25 40 5 13 25 40 5 13 25 40 LNumber of layers 3 3 3 3 3 3 3 3 3 3 3 3 Compaction energy (kJ/m3) 110.4 286.9 551.8 882.9 110.4 286.9 551.8 882.9 110.4 286.9 551.8 882.9

328 © 2008 IOS Press, ISBN 978-1-58603-908-0

compressive strength is maximized at the compactionenergy of 286.9 (kJ/m3). At higher compactionenergies, the unconfined compressive strengthsteadily declines. These results are comparable withthose in the previous publication (The JapaneseGeotechnical Society 2000).

2.2.3 Dynamic Compaction of Kanto Loam soilusing Mechanical Rammer

The optimum compaction method is experimentallyinvestigated by using a mechanical rammer forcompacting the Kanto loam soil.

The decrease in unconfined compressive strengthdue to the compaction that was described in theprevious section is assumed to occur as follows. Ifexcessive compaction energy is applied, the moistureabsorbed into the Kanto loam soil is converted into freewater; this weakens the Kanto loam soil in a mannersimilar to that when water is added. This phenomenonis one of the reasons for the difficulty in using Kantoloam soil for backfilling. Therefore, when Kanto loamsoil is used for backfilling roadbeds, the compactionenergy must be suitably adjusted.

The unconfined compressive strength increases tothe maximum value at a compaction energy ofapproximately 300 (kJ/m3), and this can be assumedto weaken with the application of higher compactionenergy. Therefore, if Kanto loam soil is compacted at acompaction energy of approximately 300 (kJ/m3), theoptimum backfilling conditions can be achieved.

Table 3 shows the specifications for the mechanicalrammer used in this study. According to George et al.(1993), when this rammer is used to compact sandysoil, the compacting energy per one rammer stroke is66.4–75.5 (J). In addition, as shown in Figure 5, thedefinition of one compaction round is the up-and-backpass of a single round of compaction across anexcavated trench. The actual time required for onecompaction round when Kanto loam soil is used forbackfilling an excavated trench of length 1.0m andwidth 0.4m and then compacting the soil surface, isapproximately 60 s. For each compaction round, 600rammer strokes with the mechanical rammer arerequired.

The compaction energy per unit volume absorbedby the roadbed soil can be calculated under theabovementioned conditions along with the bed layerthickness and the number of compaction rounds.Table 4 shows the calculation results for thecompaction energy. For a bed layer thickness of300mm, the compaction energy increases up to332.0–377.5 (kJ/m3) per compaction round or pertwo compaction rounds if the bed layer thicknessis 600mm; this brings the roadbed soil relativelyclose to the optimum backfilling condition. However,if the bed layer thickness is significantly large, thecompaction effect may not adequately reach the lowerlayers of the roadbed and only the surface layermay be excessively compacted. As a result, whena mechanical rammer is used for backfilling theKanto loam soil used in this test, it would be mostappropriate to compact the bed layer thickness to300mm with one compaction round.

Figure 4. Relationship between compaction energy andunconfined compression strength.

Table 3. Specifications of the mechanical rammer used in study

Engine Two-cycle engine Output (W) 3000 Strike number (times/s) 10 Weight (kg) 62 Strike plate Steel

Figure 5. Backfill compaction method.

Figure 3. Relationship between compaction energy and degree ofsaturation.

Table 4. Compaction energy of mechanical rammer

Exca-vation width

Layer thickness

Compaction number (times/layer)

Compaction energy

(mm)(mm) (kJ/m3)

1 332.0~377.52 664.0~755 300 (2 layers) 3 996.0~1132.51 166.0~188.8400

2 332.0~377.5600 (1 layer) 3 498.0~566.3

Figure 2. Relationship between compaction energy and drydensity.

Deformational Characteristics of Geomaterials, Burns, Mayne and Santamarina (eds) 329

3 CYCLIC LOADING TEST

A cyclic loading test is conducted to predict theamount of road surface subsidence due to cyclictraffic loading. In the past studies, the cyclic loadingtriaxial compression tests were conducted on soils toinvestigate the subsidence characteristics. However, itis thought that the lateral displacement is restrictedin the backfill of the relatively small-scale roadexcavation works.

In this study, vertical stress loading is repeatedlyapplied to the same Kanto loam soil samplesmentioned in the previous chapter with the lateraldisplacement is restricted. The relationships amongthe physical properties of the soil sample, magnitudeand number of cyclic loading and amount ofcumulative compressive strain in the sample areinvestigated.

3.1 Test Conditions

3.1.1 Preparation of Soil SamplesTable 5 shows the conditions for the preparation of thesample. The compaction test described in the previouschapter demonstrates that the unconfined compressivestrength of compacted Kanto loam soil is dependenton the compaction energy and water content. Inaddition, for the maximum unconfined compressivestrength, the degree of saturation is approximately95%. Consequently, the water contents in the soilsample are adjusted to 90% (Case A), 100% (CaseB) and 110% (Case C). The Kanto loam soil withadjusted amounts of water content is compacted in asteel cylindrical mold (inner diameter of 100mm,height of 127mm, and with the inner surface coatedwith silicon grease to reduce friction) so that it achievesa sufficiently compacted state with a degree ofsaturation of 95%.

3.1.2 Method of LoadingFigure 6 shows an overview of the device used in thecyclic loading test. Porous stones were mounted onupper and lower surfaces of the compacted samples sothat the pore water and gases could be drained offfrom the samples. Vertical stress was repeatedlyapplied to the samples by using an electropneumatictransducer with a capacity of 980 kPa (Compressionside: J813-004, manufactured by Moog Japan Inc.;Extension side: T6000-06U, manufactured byFairchild Corporation).

Figure 7 shows the vertical stress loading sequencein the test. First, the vertical stress is monotonicallyincreased upto sminþ (Ds/2), and then the stress is

repeatedly loaded and unloaded up to 100,000 timesin a sinusoidal waveform at a frequency of 1Hz forgenerating the minimum stress smin and the maximumstress smax values.

The minimum stress smin is a static soil overburdenpressure of 9.2 kPa at a depth of 0.462m correspondingto the average value of the vertical soil pressure on atypical pavement.

The maximum stress values of smax are 23.6 kPa,62.5 kPa and 101.3 kPa, which are the maximumvertical stress magnitudes on the roadbed due to arear wheel load of 39.3 kN.

A load cell (Capacity: 1000N, TCLZ-1000Nmanufactured by Tokyo Sokki Kenkyujo Co., Ltd.)is used to measure the magnitude of the acting load.In addition, an axial displacement gauge (Capacity:30mm, DDP-30A manufactured by Tokyo SokkiKenkyujo) is employed to measure the verticaldisplacement Dl from the initial position at the topof the sample when the cyclic loading stress issminþ (Ds/2). As shown below in Eq. (2), thecumulative compressive strain e is defined as thevalue of the measured displacement Dl divided bythe initial height L of the sample.

e ¼ DlL

ð2Þ

3.2 Compressive Strain of Compacted Sample

Figures 8, 9 and 10 show the relationships between thenumber of cyclic loadings and the cumulativecompressive strain of the sample. The compressivestrain increases sharply during the initial stage ofloading; however, the amount of increase becomes

Table 5. Backfill compaction conditions

Case A B C Soil grain density (g/cm3) 2.832 Degree of saturation Sr(%) 95.0 Dry density ρ d(g/cm3) 0.767 0.707 0.660 Water content ω (%) 90.3 100.8 110.4

Figure 6. The device used for cyclic loading.

Figure 7. Cyclic loading method.

330 © 2008 IOS Press, ISBN 978-1-58603-908-0

extremely small when the number of loadings reaches10,000. Subsequently, it becomes almost constant.This implies that the magnitude of compressivestrain beyond 10,000 loadings depends on themaximum value of the loading stress.

The relationship between the compressive strain for100,000 cyclic loadings and the maximum value ofstress smax is shown in Figure 11. This result showsthat the compressive strain is almost proportional to themaximum vertical stress value smax.

4 PREDICTION OF SUBSIDENCE

Based on the results of the cyclic loading test describedin Chapter 3, the amount of road surface subsidence

due to cyclic loadings is predicted. Furthermore, afull-scale traffic loading experiment is conducted forroads backfilled with Kanto loam soil. The full-scaleexperimental results are compared with the predictedvalues.

4.1 Prediction Method

4.1.1 Subsidence MechanismFigure 12 shows the mechanism assumed for theprediction of subsidence and the road model (a Type35 asphalt-concrete mix). The amount of cumulativecompressive strain due to cyclic traffic loading on thebackfilled roadbed (roadbed thickness of 600mm) isassumed to be the reason for the road surfacesubsidence. Since the compressive strain within thepaving portion (paving thickness of 350mm) isexpected to be extremely small as compared to thatinduced at the roadbed portion, the paving portion isassumed to be a completely rigid body.

4.1.2 Vertical Stress Due to Traffic LoadingThe vertical stresssz on the paving and roadbed, whenthe vehicles pass over the road surface is calculatedas follows. By utilizing the relationship between thewheel loading and the tire air pressure p, Foster andAhlvin (1954) analytically determined the verticalstress sz due to the wheel load by taking intoaccount the fact that the tire air pressure p is evenlydistributed over the contact area of a circle with radiusa, as shown in Figure 13. The diagram in Figure 14shows the contact radius a, depth z and vertical stresssz expressed in terms of the percentage of the tire airpressure p and distance r from the center of the tire.The vertical stress sz distributions at the wheel center(r¼ 0) due to a passenger vehicle (total weight 9.8 kN,rear wheel load of 3.9 kN) and a truck vehicle (total

Figure 8. Relationship between the number of cyclic loading andcompressive strain (Case A).

Figure 9. Relationship between the number of cyclic loading andcompressive strain (Case B).

Figure 10. Relationship between the number of cyclic loadingand compressive strain (Case C).

Figure 11. Relationship between the maximum vertical loadingstress and compressive strain.

Figure 12. Stress distribution due to wheel loading.

Figure 13. Contact pressure due to wheel loading.

Deformational Characteristics of Geomaterials, Burns, Mayne and Santamarina (eds) 331

weight 39.2 kN, rear wheel load of 15.7 kN) areobtained, as shown in Figure 15.

4.1.3 Prediction of Road Surface SubsidenceThe subsidence of the road surface that was subjectedto cyclic traffic loading and whose roadbed wasbackfilled with the Kanto loam soil used in theexperiments in Chapter 2 and Chapter 3 waspredicted. In the cyclic loading test described inChapter 3, the compressive strain decreasedsignificantly after the number of cyclic loadingsexceeded 10,000. It is assumed that the roadbed hadbeen sufficiently subjected to cyclic loadings when thenumber of cyclic loadings reached 100,000. The roadsurface subsidence under various conditions for watercontents in the Kanto loam soil is obtained on the basisof the relationship between the maximum value of theloaded vertical stress shown in Figure 11 and thevertical stress and soil depth due to rear wheelloading shown in Figure 15, as follows.

The magnitude of road surface subsidence d (mm)can be calculated by the integration of the compressivestrain generated by cyclic traffic loading over subgradethickness values ranging from 350mm to 950mm, asshown by the following Eq. (3).

d ¼ð950340

eðzÞdz ð3Þ

in which e(z): compressive strain due to cyclic trafficloading at the depth z; z: depth from the surface (mm).

The relationship between the maximum verticalstress smax(z) and the compressive strain e(z)obtained from the cyclic loading test described inChapter 3 is shown in Figure 11. The linearapproximate equations between e(z) and smax(z) are

obtained by using the least squares method for eachwater content, as follows:

ew¼90:3%ðzÞ ¼ 0:0189� smaxðzÞ ð4Þ

ew¼100:8%ðzÞ ¼ 0:0582� smaxðzÞ ð5Þ

ew¼110:4%ðzÞ ¼ 0:1161� smaxðzÞ ð6Þ

where smax(z): Maximum vertical stress (kPa)generated at depth z due to traffic loading.

The road surface subsidence prediction results areshown in Table 6. The roads traversed by passengervehicles and truck vehicles experience road surfacesubsidence up to 14mm, even when the watercontent of the Kanto loam soil is at relatively highlevels of 110%.

In actual road construction, a relatively simplepavement is constructed immediately after theroadbed is backfilled and temporarily recovered.After vehicles are allowed to pass for a specificperiod, the final paving is reconstructed andrecovered. According to the cyclic loading testresults, the surface subsidence occurs during theinitial stages of cyclic traffic loading and thenremains at a fixed level. It is believed that (1) evenduring the temporary recovery stage, a compressivestrain from the backfilled material accumulates in theroadbed due to cyclic traffic loading and a certaindegree of subsidence occurs on the road surface and(2) after the main recovery, the amount of compressivestrain that accumulates in the roadbed due to cyclictraffic loading is extremely small. This means that theroad surface subsidence after the main recovery isfurther smaller than the amount of subsidence predicted.

From these results, it can be concluded that thesubsidence calculated in this study is within anacceptable range for practical applications and thatthe backfilling using Kanto loam soil in this researchis fully feasible for the roads, where the traffic loadinglevel is within 39.2 kN.

4.2 Verification of the Subsidence PredictionMethod

4.2.1 Full-scale Test for Traffic LoadingAs shown in Figure 16, two excavated trenches arebackfilled with Kanto loam soil and paved using theType 35 asphalt-concrete mixture, which is mostcommonly used on the residential streets in Japan.Table 7 shows the full-scale test implementationconditions and soil physical properties for the Kanto

Figure 15. Relationship between the vertical stresss z due to rearwheel loading and depth z.

Table 6. Predicted results for the road surface subsidenceWater content (%)

90.3 100.8 110.4 Subsidence (mm) in cyclic loading in 100,000 times Loading conditions

Passenger vehicle (9.8 kN) 0.6 1.9 3.8 Truck vehicle (39.2 kN) 2.3 7.1 14.1

Figure 14. Relationship between vertical stress s z and z/a(Foster and Ahlvin 1954).

332 © 2008 IOS Press, ISBN 978-1-58603-908-0

loam soil used in backfilling (based on JGS 0051, JISA1202, JIS A 1203, JIS A 1205, JIS A 1211 and JIS A1228). In addition, Figure 17 shows the grain sizedistribution curve (based on JISA1204). The thicknessof each backfill material layer is 300mm, and eachlayer is subjected to one round of compaction.

Vehicles pass over the center of the temporaryrecovery, as shown in Figure 16. Initially, 600 passesof a passenger vehicle of 9.8 kN are conducted andthe road surface settlement is observed until no furtherincrease in subsidence can be observed. Then, the samenumber of passes of a 39.2-kN truck vehicle over thetemporary recovery locations is performed.

4.2.2 Test ResultsFigure 18 shows the relationship between the numberof passing vehicles and the amount of road surfacesubsidence observed from the full-scale traffic loadingtest. From the cyclic loading test performed in thelaboratory, the full-scale test results confirmed a largeamount of subsidence in the initial stages of loading.Although subsidence up to approximately 2mm isobserved due to the initial irregularities in theadjacent road surfaces after backfilling, the roadsurface settles monotonically with increase in thevolume of traffic passage.

A difference of approximately 1mm is observedbetween the observed surface settlement values attwo locations (No.1 and No.2). It can be assumedthat this difference results from the differences incompaction time or paving construction precisionor measurement errors. Although the measuredsubsidence value increased by approximately 1mmand then decreased by 1mm for approximately800–900 vehicle passes, it was concluded that thisvalue must be a measurement error and therefore itwas excluded from the experimental data.

4.2.3 Comparison of Predicted SubsidenceValue with Measurement Results

The relationship between the amount of cyclic trafficloading and compressive strain (with water content of100%) shown in Figure 9 is used to investigate therelationship among the number of passing vehicles, themaximum vertical stress value and the compressivestrain. These relationships and the maximum verticalstress value for each roadbed soil depth obtained fromFigure 15 are combined to predict the road surfacesubsidence in a full-scale test.

Figure 18 compares the predicted subsidencevalue with the measurement value. Comparing themeasurement value from the full-scale test withthe predicted subsidence value, subsidence ofapproximately 2mm in the full-scale test due to theinitial irregularities in the adjacent road surfaces isobserved. Therefore, a difference of approximately1–2mm between the predicted value and measuredvalue is observed. It is confirmed that the subsidenceincreases during the initial loading stage in everycase, and thereafter tends to increase graduallywith the number of passing vehicles. These resultsshow that the proposed method can be used to predictsubsidence due to cyclic traffic loading for roadsurfaces backfilled with Kanto loam soil within areasonable range.

5 CONCLUSION

In this study, cyclic loading tests on compacted Kantoloam soils are conducted to investigate the subsidencecharacteristics by using a confined compressionapparatus, and then the experimental results wereemployed to predict the amount of road surfacesubsidence due to cyclic traffic loading. A full-scaleexperiment was conducted to investigate the

Figure 16. Full-scale test of traffic loading.

Table 7. Backfill conditions of full scale testNo.1 No.2

bed grade thickness (m) 0.3 0.3 Method of com-paction Compaction (round) 1 1

Name Sand mixed with vol-canic cohesive soil Soil grain density (g/cm3) 2.802 Plastic limit (%) 88.2 Liquid limit (%) 111.2

Soil physical prop-erties

Design CBR (%) 0.8 Wet density (g/cm3) 1.368 1.378 Dry density (g/cm3) 0.679 0.695 Water content (%) 101.6 98.4 Void ratio e 3.13 3.03

Compacted soil condition

Cone index (kPa) 521 747

Figure 17. Grain size distribution curve.

Figure 18. Relationship between the number of vehicles passingand road surface subsidence.

Deformational Characteristics of Geomaterials, Burns, Mayne and Santamarina (eds) 333

subsidence due to cyclic traffic loading on a roadsurface, which was backfilled with Kanto loam soil.

The values measured in the full-scale experimentwere comparedwith the predicted subsidence values toverify the suitability of the subsidence predictionmethod. Consequently, good correlation is obtainedbetween the amount of the predicted subsidence andthe measurement results. The proposed method forpredicting the road surface with the compacted Kantoloam soil can be used to predict the subsidence due tocyclic traffic loading for road surfaces backfilled withKanto loam within a reasonable range.

ACKNOWLEDGMENTS

For this study, valuable advice was provided by Mr.Tomoaki Yokoyama of Tokyo Gas Co., Ltd. Theauthors are also indebted to Mr. Yasuyuki Takahashiof Kanpai Co., Ltd., and Mr. Mitsuo Hayashi of KisoJiban Consultants Co., Ltd., for their cooperation inconducting the experiments.

REFERENCES

Dingqing, Li and Ernest, T.S. (1996). “Cumulative PlasticDeformation For Fine-Grained Subgrade Soils”, Journal ofGeotechnical Engineering, ASCE.

Express Highway Research Foundation of Japan (1973) “CivilEngineering for Kanto Loam–Soil Quality and DesignExecution”, Kyoritsu Shuppan, pp. 38–39, pp. 82–91.

Foster, C.R. and Ahlvin, R.G. (1954). “Stress and DeflectionsInduced by a Uniform Circular Load”, HRB Proc.

George, M. Fliz, and Thomas, L. Brandon (1993). “CompactorForce and Energy Measurements”, Geotechnical TestingJournal, ASTM.

Kura, K. and Miki, H. (1991). “Some Considerations regardingUse of Improved Soil in Roadbed Backfilling”, CivilEngineering Technology Materials 33–10.

Minagata, H., Yoshizaki, K., Hagiwara, N., Ishikawa, M. andSakamoto, K. (2006). “Soil Suppression in Guided PipeWork”, 2006 City Gas Symposium Papers, The Japan GasAssociation.

Ministry of Land, Infrastructure and Transport (2002).“Construction Recycling Promotion Plan 2000”, p. 6.

Ministry of Land, Infrastructure and Transport (2003). “ActionPlan for Effective Utilization of Construction Soil, Etc.”,Study Committee for Effective Utilization of ConstructionSoil, Etc., Appendix-2, pp. 1–2.

Ministry of Land, Infrastructure and Transport (2005). “CurrentState of Construction Byproducts”, the Ministry of Land,Infrastructure and Transport Recycling HP (http://www.mlit.go.jp/sogoseisaku/region/recycle/fukusanbutsu/genjo/index.htm), pp. 13.

Monismith, C.U., Ogawa, S. and Freeme, C.R. (1975).“Permanent Deformation Characteristics of Sub-grade Soilsdue to Repeated Loading”, Transportation Research Board.

Public Works Research Center (1997). Technical Manual forUtilization of Construction Soil, p. 24.

The Japanese Geotechnical Society (2000). “Soil TestingMethodand Analysis”, First Revised Edition, p.757.

334 © 2008 IOS Press, ISBN 978-1-58603-908-0