Study on the Mechanism of Piping and the Effect of Suspended Cutoff in Dike Foundation in Dike Foundation YAO Qiuling China Institute of Water Resources.

Study on the Mechanism of Piping and Study on the Mechanism of Piping and the Effect of Suspended Cutoffthe Effect of Suspended Cutoff

in Dike Foundationin Dike Foundation

YAO QiulingYAO Qiuling

China Institute of Water Resources and Hydropower Research China Institute of Water Resources and Hydropower Research (IWHR) (IWHR)

May 6, 2008, TorontoMay 6, 2008, Toronto

11 IntroductionIntroduction

2 Tests on Piping in the Single-stratum Foundation 2 Tests on Piping in the Single-stratum Foundation

3 Tests on Piping in the Two-stratum Foundation3 Tests on Piping in the Two-stratum Foundation

4 Tests on Piping in the Three-stratum Foundation 4 Tests on Piping in the Three-stratum Foundation

5 Tests on Seepage Control Effects of Suspended 5 Tests on Seepage Control Effects of Suspended

Cutoffs in the Two-stratum FoundationCutoffs in the Two-stratum Foundation

6 Conclusions6 Conclusions

ContentsContents

1 Introduction

Kinds of Kinds of risksrisks

PipingPipingRiverbank Riverbank collapsecollapse

LeakageLeakage ChinkChink DropDrop SloughingSloughingWave Wave

erosionerosionCulvert Culvert

gategateOthersOthers totaltotal

TotalTotal

risksrisks20252025 330330 27632763 27952795 106106 615615 316316 220220 235235 94059405

PercentagePercentage 21.5%21.5% 3.5%3.5% 29.4%29.4% 29.7%29.7% 1.1%1.1% 6.5%6.5% 3.4%3.4% 2.3%2.3% 2.5%2.5% 100%100%

TotalTotal

severe riskssevere risks366366 5656 4040 130130 66 5656 99 2020 1515 698698

PercentagePercentage 52.4%52.4% 8.0%8.0% 5.7%5.7% 18.6%18.6% 0.9%0.9% 8.0%8.0% 1.3%1.3% 2.9%2.9% 2.1%2.1% 100%100%

1.1 Statistics of Risks in Yangtze River Main Dike during the ’98

Flood

Because the piping mechanisms are not well understood, fights are taken wherever piping occurred. It costs plenty of manpower and materials to check the possible piping occurrences and to fight against them.

1.2 Problems in Countermeasures of Dike Foundation Piping :

In some cases, the calculated widths of landside seepage berms according to the existing design criterion of China are too large to be adopted in dike stabilization design. Thus, designers have to determine them through their individual experiences.

Suspended cutoffs are used to prevent piping in dike projects. However, it is not strictly coincident with the seepage control theories in dike projects. Further studies are needed to explain the action mechanism and to work out the design criteria of suspended cutoffs to prove their effectiveness.

1.2 Problems in Countermeasures of Dike Foundation Piping :

single-stratum dike foundation two-stratum dike foundation three-stratum dike foundation

1.3 Three typical types of dike foundation

2 Tests on Piping in the Single-stratum Dike Foundation

One perspex sheet is used to cover the sand sample as the dike body.

Samples are made by depositing sand in water in thin layers.

2.1 Setup of Models2.1 Setup of Models

2.2 Test Method2.2 Test Method

During the tests, the hydraulic head is increased During the tests, the hydraulic head is increased step by step from step by step from aa lower original lower original oneone. . When the When the seepage deformation seepage deformation stopsstops, and the seepage , and the seepage discharge and piezometer head keep invariabledischarge and piezometer head keep invariable under certain hydraulic headunder certain hydraulic head, the hydraulic head , the hydraulic head can be raised can be raised forfor the next step the next step..

2.3 Several Stages (1)2.3 Several Stages (1)

Some small holes, where several grains are taken up and fall down continuously, are observed at landside near the levee toe. No grains are taken away from the holes.

Some narrow and shallow piping channels are observed along the interface of the dike body and the foundation. They look like meandering rivers. Sand grains in the foundation are taken away along the channels by flow and brought out from the holes. Along the increase of the hydraulic head step by step, piping channels propagate toward the river side and stop intermediately. If the hydraulic head is lower than the critical one, the channels will stop finally.

2.3 Several Stages (2)

If the hydraulic head exceeds the critical value, the piping channels will propagate to riverside consistently. Plenty of grains are brought out by the muddy flow. The seepage discharge increases suddenly. Scoured by high flow, piping channels will enlarge continuously. Meanwhile, one main piping channel propagates toward riverside fast and connects to the river water finally that will cause dike failure.

2.3 Several Stages (3)2.3 Several Stages (3)

2.4 Relation Curve between Seepage Discharge 2.4 Relation Curve between Seepage Discharge and Average Horizontal Gradientand Average Horizontal Gradient

stage stage 11

stage stage 22

stagestage 33

Piping initiates near the dike toe at landside firstly, then Piping initiates near the dike toe at landside firstly, then propagates horizontally toward riverside along the propagates horizontally toward riverside along the interface of the dike body and the foundation acted by the interface of the dike body and the foundation acted by the horizontal seepage force.horizontal seepage force.

2.5 From the phenomenon mentioned above and 2.5 From the phenomenon mentioned above and the experimental results, it can be inferred the experimental results, it can be inferred that: that:

If the hydraulic head is lower than the critical value, the If the hydraulic head is lower than the critical value, the piping channel length is limited, and piping will cease and piping channel length is limited, and piping will cease and the seepage deformation will stop ultimately. Otherwise, if the seepage deformation will stop ultimately. Otherwise, if the hydraulic head exceeds the critical value, the piping the hydraulic head exceeds the critical value, the piping channel will propagate toward the river side and connect channel will propagate toward the river side and connect to the river water finally, that will cause the dike failure.to the river water finally, that will cause the dike failure.

For the single-stratum dike foundation, the critical average For the single-stratum dike foundation, the critical average horizontal seepage gradient is about 0.278 in these tests.horizontal seepage gradient is about 0.278 in these tests.

3 Tests on Piping in the Two-stratum Dike Foundation3 Tests on Piping in the Two-stratum Dike Foundation

A perspex sheet with a length of 2.25m fully covers the sand layer, A perspex sheet with a length of 2.25m fully covers the sand layer,

which simulates both the dike body and the surface clay layer. which simulates both the dike body and the surface clay layer.

The joints between the perspex sheet and the flume is sealed up by glThe joints between the perspex sheet and the flume is sealed up by glue. The outflow can only run out from the circular hole.ue. The outflow can only run out from the circular hole.

3.1 Set Up of Models3.1 Set Up of Models

The first is to keep the hole open and to The first is to keep the hole open and to

increase the hydraulic head step by step from a increase the hydraulic head step by step from a

low original water level. low original water level.

3.2 Two methods are used in the tests respectively3.2 Two methods are used in the tests respectively

The second is to plug the hole by a rubber The second is to plug the hole by a rubber

stopper before the test and then raise the water stopper before the test and then raise the water

level to a relatively high point. Then the plug is level to a relatively high point. Then the plug is

pulled out to begin the test. pulled out to begin the test.

The propagation process, scope and form of piping The propagation process, scope and form of piping channel in the two-stratum foundation are similar to that in channel in the two-stratum foundation are similar to that in the single-stratum foundation.the single-stratum foundation.

3.3 Test Process3.3 Test Process

Piping in the two-stratum foundation only occurs on the top Piping in the two-stratum foundation only occurs on the top

of the sand layer after the surface soil layer is penetrated by of the sand layer after the surface soil layer is penetrated by

the upward seepage flow, and piping propagates horizontally the upward seepage flow, and piping propagates horizontally

under the seepage force. under the seepage force.

3.4 From the tests, it can be inferred that:

The observed piping channel development process in the two-The observed piping channel development process in the two-

stratum foundation is nearly the same as single-stratum stratum foundation is nearly the same as single-stratum

foundation.foundation.

For the two-stratum dike foundation, the critical average For the two-stratum dike foundation, the critical average

horizontal seepage gradient is about 0.214. This value is smaller horizontal seepage gradient is about 0.214. This value is smaller

than that of the single-stratum foundation, which is caused by than that of the single-stratum foundation, which is caused by

more intense 3D effects in the two-stratum foundation.more intense 3D effects in the two-stratum foundation.

4 Tests on Piping in the Three-stratum Dike Foundation

Tests also adopted the same method as the second one Tests also adopted the same method as the second one for the two-stratum foundation.for the two-stratum foundation.

The conditions of The conditions of

different thickness of different thickness of

sand layer and gravel sand layer and gravel

layer are considered.layer are considered.

4.1 Set Up of Models4.1 Set Up of Models

Conditions and results of different test schemes in the three-stratum foundationConditions and results of different test schemes in the three-stratum foundation

NoNo..

Thickness distribution (cm)Thickness distribution (cm)Critical Critical average average

horizontal horizontal gradientgradient

Pattern of piping Pattern of piping developmentdevelopment

Sand layerSand layer Gravel Gravel layerlayer

Percentage Percentage of sand of sand layer layer thickness thickness

11 44 5656 6.7%6.7% 0.0780.078 deep-seated deep-seated pipingpiping

22 66 5454 10%10% 0.1160.116 deep-seated deep-seated pipingpiping

33 1818 4242 30%30% 0.1650.165 shallow-seated shallow-seated pipingpiping

44 6060 00 100%100% 0.2140.214 shallow-seated shallow-seated pipingpiping

4.2 Test Schemes4.2 Test Schemes

Several stages (1)

4.3 Scheme No. 1 and No.24.3 Scheme No. 1 and No.2

As soon as the stopper is pulled out, some grains will be brought out from the hole along with the seepage flow, and a large sand ring will be formed around the hole. In the hole, sand boils appear. Beyond the sand ring, no deformation is observed.

When the hydraulic head is high, both sand boils and shallow When the hydraulic head is high, both sand boils and shallow channels appear. They propagate toward the river side together. If channels appear. They propagate toward the river side together. If the hydraulic head is lower than the critical value, this process the hydraulic head is lower than the critical value, this process can stop. can stop.

Several stages (2)

Finally, if the hydraulic head exceeds the critical value, the Finally, if the hydraulic head exceeds the critical value, the channel will develop in an accelerated way to the river side channel will develop in an accelerated way to the river side until it connects to the water tank. until it connects to the water tank.

In this case, the development of piping is mainly driven by In this case, the development of piping is mainly driven by the vertical hydraulic gradient, and the depth of piping the vertical hydraulic gradient, and the depth of piping channels is larger than that of the former two types of dike channels is larger than that of the former two types of dike foundations. This piping phenomenon is suggested to be foundations. This piping phenomenon is suggested to be called “deep-seated piping”. In contrast, piping developing called “deep-seated piping”. In contrast, piping developing in surface areas in horizontal direction as in the former two in surface areas in horizontal direction as in the former two types of dike foundation is suggested to be called “shallow-types of dike foundation is suggested to be called “shallow-seated piping”.seated piping”.

Several stages (3)

4.4 Scheme No. 3

If the sand layer in the three-stratum dike If the sand layer in the three-stratum dike foundation is thick enough as in the scheme foundation is thick enough as in the scheme No. 3, the mechanism of piping will be similar No. 3, the mechanism of piping will be similar to that in the single-stratum and the two-to that in the single-stratum and the two-stratum foundations and can be classified as stratum foundations and can be classified as “shallow-seated piping”.“shallow-seated piping”.

0

0.05

0.1

0.15

0.2

0.25

0.3

0102030405060708090100percent of the thickness of sand layer（%）

crit

ical

hor

izon

tal

grad

ient

It is indicated that the average critical gradient of piping in the three-stratum dike foundation will decrease when there exists a gravel layer. What’s more, the piping development mechanism is changed in this case. These effects will be enhanced when the sand layer is getting thinner.

4.54.5

5 Tests on seepage control effect of suspended 5 Tests on seepage control effect of suspended cutoff in the two-stratum dike foundationcutoff in the two-stratum dike foundation

5.1 Setup of Models5.1 Setup of Models

No.No. LocationLocation X/LX/L PenetrationPenetrationt/Tt/T

Critical Critical average average

horizontal horizontal gradientgradient

Percent ofPercent of incrementincrement

Vertical Vertical efficiencyefficiency

11 no cutoffno cutoff ---- 0/600/60 0.2140.214 0%0% ----

1(a)1(a)riversideriverside

35/14035/140 6/606/60 0.2540.254 19%19% 4.44.4

1(b)1(b) 35/14035/140 12/6012/60 0.4070.407 90%90% 10.510.5

2(a)2(a)

landsidelandside

105/140105/140 6/606/60 0.3100.310 45%45% 10.510.5

2(b)2(b) 105/140105/140 12/6012/60 0.4310.431 101%101% 11.811.8

2(c)2(c) 105/140105/140 18/6018/60 0.5510.551 157%157% 12.212.2

2(d)2(d) 125/140125/140 6/606/60 0.340.34 59%59% 13.713.7

5.2 Conditions and results of tests with suspended cutoffs 5.2 Conditions and results of tests with suspended cutoffs

Effect of the Depth of CutoffsEffect of the Depth of Cutoffs

Effect of the Location of CutoffsEffect of the Location of Cutoffs

The critical average horizontal The critical average horizontal gradient of piping failure gradient of piping failure increases along with the increases along with the increase of the cutoff depth increase of the cutoff depth penetrated into the sand layer at penetrated into the sand layer at a definite location. a definite location.

The critical average The critical average horizontal gradient of piping horizontal gradient of piping failure is higher when the cutoff failure is higher when the cutoff is placed at landside than that is placed at landside than that placed at riverside with a definite placed at riverside with a definite penetration.penetration.

5.3 Effects of Depth and Location of Cutoffs5.3 Effects of Depth and Location of Cutoffs

piping holepiping hole

The seepage vector at the front and back sides of the cutoff is The seepage vector at the front and back sides of the cutoff is changed from horizontal to vertical directions. As a result, the anti-changed from horizontal to vertical directions. As a result, the anti-permeability strength of the dike foundation can be significantly permeability strength of the dike foundation can be significantly improved. improved.

5.4 The Action Mechanism of Cutoffs5.4 The Action Mechanism of Cutoffs

The cutoff can increase the The cutoff can increase the seepage path after piping seepage path after piping occurrence.occurrence.

The cutoff can prevent the The cutoff can prevent the piping channel from piping channel from propagation toward riverside.propagation toward riverside.

When the piping channel propagates to landside of the cutoff, the When the piping channel propagates to landside of the cutoff, the effective seepage path will be composed of two parts, namely the effective seepage path will be composed of two parts, namely the horizontal distance from the river water to the cutoff, as well as horizontal distance from the river water to the cutoff, as well as the product of the vertical efficiency and the cutoff depth. It can the product of the vertical efficiency and the cutoff depth. It can be lengthened if the cutoff is placed at landside in contrast to be lengthened if the cutoff is placed at landside in contrast to riverside. Therefore, it can be concluded that cutoffs at landside riverside. Therefore, it can be concluded that cutoffs at landside will be more effective than at riverside.will be more effective than at riverside.

6 Conclusions

If the hydraulic head is larger than the critical value, piping will If the hydraulic head is larger than the critical value, piping will propagate persistently toward riverside and it will eventually propagate persistently toward riverside and it will eventually cause dike failure for all the three types of dike foundations. cause dike failure for all the three types of dike foundations. However, piping can only extend to a limited scope and it will However, piping can only extend to a limited scope and it will stop. In this case, piping will not cause dike failure. Therefore it stop. In this case, piping will not cause dike failure. Therefore it is suggested that it is not necessary to fight against piping is suggested that it is not necessary to fight against piping wherever it appears.wherever it appears.

For the single-stratum and two-stratum foundations, as well as For the single-stratum and two-stratum foundations, as well as the three stratum foundation when the sand layer is thick the three stratum foundation when the sand layer is thick enough, piping propagates horizontally along the surface of the enough, piping propagates horizontally along the surface of the sand layer. However, for the three-stratum foundation with a sand layer. However, for the three-stratum foundation with a very thin sand layer, “deep-seated piping” will occurs, and the very thin sand layer, “deep-seated piping” will occurs, and the damaged zone will be large in both width and depth.damaged zone will be large in both width and depth.

Critical average horizontal seepage gradients are obtained for differCritical average horizontal seepage gradients are obtained for different types of dike foundations in experimental conditions. These can ent types of dike foundations in experimental conditions. These can be referenced as a basis for the establishment of the horizontal conbe referenced as a basis for the establishment of the horizontal control criteria of dike-foundation piping. It can be used in the dike projtrol criteria of dike-foundation piping. It can be used in the dike project design and emergency management such as the determination ect design and emergency management such as the determination of landside berm width as well as the reasonable scope in which wof landside berm width as well as the reasonable scope in which whether it is necessary or not to fight against piping.hether it is necessary or not to fight against piping.

Suspended cutoffs can effectively increase foundation resistance to Suspended cutoffs can effectively increase foundation resistance to piping, and they are more efficient when placed at landside than at ripiping, and they are more efficient when placed at landside than at riverside. Therefore it is suggested that suspended cutoffs can be useverside. Therefore it is suggested that suspended cutoffs can be used for seepage control in future design. In addition, it is recommended d for seepage control in future design. In addition, it is recommended that suspended cutoffs placed near landside dike toe or the end of lathat suspended cutoffs placed near landside dike toe or the end of landside berm will be taken as an alternative in the selection of possiblndside berm will be taken as an alternative in the selection of possible seepage control measures.e seepage control measures.

6 Conclusions

Thank you!