You are on page 1of 626

POTENTIAL BENEFITS OF GEOSYNTHETICS

FLEXIBLE PAVEMENTS
PRELIMINARY DRAFT
FINAL REPORT
Prepared for

National Cooperative Highway Research Program
Transportation Research Board
Nationai Research Council

TRANSPORTAT!ON RESEARCH BOARD
NAS-N RC
EELY1LEQa1=LMEL11
This report, not released for publication, is furnished only for
review to members of or participants in the work of the
National Cooperative Highway Research Program. !t is to be
regarded as fully priviledged, and dissemination of the information included herein must be approved by the NCHRP.

Richard D. Barksdale
Georgia institute of Technology
Atlanta, Georgia
Stephen F. Brown
University of Nottingham
Nottingham, England

GTRI Project E20-672

June 1988

SCEGIT-88-102

THIS IS A DRAFT REPORT FOR REVIEW ONLY

Acknowledgment
This work was sponsored by the American Association of State
Highway and Transportation Officials, in cooperation with
the Federal Highway Administration, and was conducted in the
National Cooperative Highway Research Program which is
administered by the Transportation Research Board of the
National Research Council.

Disclaimer
This copy is an uncorrected draft as submitted by the research
agency. A decision concerning acceptance by the Transportation
Research Board and publication in the regular NCHRP series will
not be made until a complete technical review has been made and
discussed with the researchers. The opinions and conclusions
expressed or implied in the report are those of the research
agency. They are not necessarily those of the Transportation
Research Board, the National Research Council, or the Federal
Highway Administration, American Association of State Highway
and Transportation Officials, or of the individual states
participating in the National Cooperative Highway Research
Program.

TABLE OF CONTENTS
Page
LIST OF FIGURES
LIST OF TABLES
ACKNOWLEDGEMENTS
ABSTRACT
1

SUMMARY
INTRODUCTION AND RESEARCH APPROACH
Objectives of Research
Research Approach

7
8
9

CHAPTER II FINDINGS
Literature Review - Reinforcement of Roadways
Analytical Study
Large-Scale Laboratory Experiments
Summary and Conclusions

13
15
28
84
142

CHAPTER III SYNTHESIS OF RESULTS, INTERPRETATION, APPRAISAL
AND APPLICATION
Introduction
Geosynthetic Reinforcement
Separation and Infiltration
Filter Selection
Durability

144
144
146
202
235
241

CHAPTER I

250
250

CHAPTER IV CONCLUSIONS AND SUGGESTED RESEARCH
Introduction
Overall Evaluation of Aggregate Base Reinforcement
Techniques
Separation and Filtration
Durability
Suggested Research

250
265
266
267

APPENDIX A REFERENCES

A-1

APPENDIX B Properties of Materials Used in Large-Scale
Pavement Test Facility
Laboratory Testing of Materials .
References

B-1
B-2
B-25

APPENDIX C PRELIMINARY EXPERIMENTAL PLAN FOR FULL-SCALE
FIELD TEST SECTIONS

C-1

LIST OF FIGURES
Page

Figure
1

General Approach Used Evaluating Geosynthetic
Reinforcement of Aggregate Bases for Flexible
Pavements

15

Effect of Reinforcement on Behavior of a Subgrade
Subjected to Vertical Stress (After Bender &
Barenberg, Ref. 3)

17

Maximum Surface Deformation as a Function of Traffic
(After Barker, Ref. 38)

17

Comparison of Strain at Bottom of Asphalt Surfacing
With and Without Mesh Reinforcement (After Van Grup
and Van Hulst, Ref. 41)

23

Deflection and Lateral Strain Measured in Nottingham
Test Facility (After Brown, et al., Ref. 37) .

25

Resilient Modulus Relationships Typically Used for a
Cohesive Subgrade and Aggregate Base ..

31

Idealization of Layered Pavement Structure for
Calculating Rut Depth (After Barksdale, Ref. 50) .

47

Comparison of Measured and Computed Permanent Deformation Response for a High Quality Crushed Stone Base:
1000 Load Repetitions

47

Comparison of Measured and Computed Permanent Deformation Response for a Low Quality Soil-Aggregate Base:
100,000 Load Repetitions

50

Comparison of Measured and Computed Permanent Deformation Response for a Silty Sand Subgrade: 100,000 Load
Repetitions

50

Pavement Geometries, Resilient Moduli and Thicknesses
Used in Primary Sensitivity Studies .

53

12

Typical Variations of Resilient Moduli with CBR .

55

13

Variation of Radial Stress at Top of Subgrade with
Radial Distance from Centerline (Tension is Positive) .

67

Equivalent Base Thickness for Equal Strain: 2.5 in. AC,
Es = 3.5 ksi

68

2

3
4

5
6
7
8

9

10

11

14

ii

LIST OF FIGURES (continued)
Page

Figure

Equivalent Base Thickness for Equal Strain: 6.5 in.
AC, E s = 3.5 ksi

68

Equivalent Base Thickness for Equal Strain: 2.5 in.
AC, E s = 12.5 ksi

69

Variation in Radial Strain in Bottom of Aggregate Base
(Tension is Positive)

69

18

Equivalent Base Thicknesses for Equal Strain: S g 1/3 Up

74

19

Equivalent Base Thicknesses for Equal Strain: S g 2/3 Up

74

20

Geosynthetic Slack Force - Strain Relations Used in
Nonlinear Model

75

Variation of Radial Stress a r With Poisson's Ratio
(Tension is Positive)

75

Theoretical Influence of Prestress on Equivalent Base
Thickness: c r and cv Strain Criteria . .

83

23

Gradation Curve for Aggregates Used in Asphaltic Mixes

87

24

Gradation Curves for Granular Base Materials

90

25

Typical Layout of Instrumentation Used in Test Track
Study

92

Profilameter Used to Measure Transverse Profiles on
Pavement

95

15
16
17

21
22

26

.

Triple Legged Pneumatic Tamper Used on Subgrade

28

Single Legged Pneumatic Compactor Used on Subgrade

97

29

Vibrating Plate Compactor

97

30

Vibrating Roller

97

31

Woven Geotextile with 1 in. Diameter Induction Strain
Coils

100

32

Geogrid with 1 in. Diameter Induction Strain Coils

10D

33

Method Employed to Stretch Geogrid Used to Prestress
the Aggregate Base - Test Series 4 . .

103

iii

.

97

27

LIST OF FIGURES (continued)
Figure

211t
Static Cone Penetrometer Test on Subgrade

35

Dynamic Cone Penetrometer Test on Subgrade

106

36

Nuclear Density Meter

106

37

Clegg Hammer

106

38

Pavement Test Facility

111

39

Distribution of the Number of Passes of Wheel Load in
Multiple Track Tests

113

Variation of Rut Depth Measured by Profilometer with the
Number of Passes of 1.5 kips Wheel Load - All Test
Series

121

Pavement Surface Profiles Measured by Profilometer at
End of Tests - All Test Series

122

Variation of Vertical Permanent Deformation in the
Aggregate Base with Number of Passes of 1.5 kip Wheel
Load - All Four Test Series . ..

123

Variation of Vertical Permanent Deformation in the
Subgrade with Number of Passes of 1.5 kip Wheel Load All Four Test Series

125

Variation of Permanent Surface Deformation with Number
of Passes of Wheel Load in Single Track Tests - All
Four Test Series

128

Variation of Vertical Permanent Strain with Depth of
Pavement for All Four Test Series .

129

Variation of Vertical Resilient Strain with Depth of
Pavement for All Test Series ...

131

Variation of Longitudinal Resilient Strain at Top and
Bottom of Granular Base with Number of Passes of
1.5 kip Wheel Load

134

Variation of Transient Vertical Stress at the Top of
Subgrade with Number of 1.5 kips Wheel Load - All
Test Series

136

40

41
42

43

44

45
46
47

48

iv

.

106

34

LIST OF FIGURES (continued)
Page

Figure

Variation of Transient Longitudinal Stress at Top and
Bottom of Granular Base with Number of Passes of 1.5
kips Wheel Loads - All Test Series

137

Variation of Permanent Surface Deformation with Number
of Passes of Wheel Load in Supplementary Single Track
Tests

139

Pavement Surface Condition at the End of the MultiTrack Tests - All Test Sections

141

52

Basic Idealized Definitions of Geosynthetic Stiffness

149

53

Selected Geosynthetic Stress-Strain Relationships

149

54

Variation of Subgrade Resilient Modulus with Depth
Estimated from Test Results

154

Reduction in Response Variation as a Function of Base
Thickness

154

Variation of Radial Stress in Base and Subgrade with
Base Thickness

159

Superposition of Initial Stress and Stress Change
Due to Loading .

159

Reduction in Permanent Deformation Due to Geosynthetic
for Soil Near Failure

161

59

Reduction in Subgrade Permanent Deformation

168

60

Reduction in Base Permanent Deformation

168

61

Improvement in Performance with Geosynthetic Stiffness

175

62

Improvement in Performance with Geosynthetic Stiffness

175

63

Influence of Base Thickness on Permanent Deformation:
Sg = 4000 lbs/in

177

Influence of Subgrade Modulus on Permanent Deformation:
S = 4000 lbs/in

177

49

50

51

55
56
57
58

64

V

LIST OF FIGURES (continued)
Page

Figure
Theoretical Effect of Slack on Force in Geosynthetic:
2.5 in. AC/9.72 in. Base

180

Free and Fixed Direct Shear Apparatus for Evaluating
Interface Friction

184

Influence of Geosynthetic Pore Opening Size on Friction
Efficiency

184

68

Reduction in Rutting Due to Prerut with Geogrid .

192

69

Reduction in Rutting Due to Prerut - No Reinforcement .

192

70

Variation of Shear Stress Along Geosynthetic Due to
Initial Prestress Force on Edge

192

Influence of Added Fines on Resilient Modulus of Base
(After Jorenby, Ref. 104)

203

Influence of Subgrade Water Content and Geosynthetic
on Stone Penetration (After Glynn & Cochrane, Ref. 84)

203

Variation of Vertical Stress on Subgrade with Initial
Compaction Lift Thickness and Roller Force

208

Bearing Capacity Failure Safety Factor of Subgrade
During Construction of First Lift

208

Mechanisms of Slurry Formation and Strain in
Geosynthetic

217

Electron Microscope Pictures of Selected Geotextiles:
Plan and Edge Views (94x)

219

Variation of Geosynthetic Contamination with Number of
Load Repetitions (After Saxena and Hsu, Ref. 98) . .

222

Variation of Geosynthetic Contamination with
Geosynthetic Apparent Opening Size, 0 95 (After Bell,
et al., Ref. 79)

222

Variation of Geosynthetic Contamination Approximately
8 in. Below Railroad Ties with Geosynthetic Opening
Size (After Raymond, Ref.80)

224

65

66

67

71

72

73

74

75

76

77

78

79

vi

LIST OF FIGURES (continued)
Page

Figure

Variation of Geosynthetic Contamination with Stress
Level and Subgrade Moisture (After Glynn & Cochrane,
Ref. 84)

224

Observed Variation of Geosynthetic Contamination with
.
Depth Below Railway Ties (After Raymond, Ref. 80)

226

Variation of Vertical Stress with Depth Beneath
Railroad Track and Highway Pavement

226

Cyclic Load Triaxial Apparatus for Performing
Filtration Tests (Adapted from Janssen, Ref. 101)

229

Economic Comparison of Sand and Geosynthetic Filters
for Varying Sand Filter Thickness .

229

85

Observed Strength Loss of Geosynthetics with Time

245

86

Approximate Reduction in Granular Base Thickness as a
Function of Geosynthetic Stiffness for Constant
Radial Strain in AC: 2.5 in. AC, Subgrade CBR = 3

258

Approximate Reduction in Granular Base Thickness as a
Function of Geosynthetic Stiffness for Constant
Vertical Subgrade Strain: 2.5 in. AC, Subgrade CBR = 3 .

258

Approximate Reduction in Granular Base Thickness as a
Function of Geosynthetic Stiffness for Constant
Radial Strain in AC: 2.5 in. AC, Subgrade CBR = 3

259

Approximate Reduction in Granular Base Thickness as a
Function of Geosynthetic Stiffness for Constant
Vertical Subgrade Strain: 6.5 in. AC, Subgrade CBR = 3 .

259

Approximate Reduction in Granular Base Thickness as a
Function of Geosynthetic Stiffness for Constant
Radial Strain in AC: 2.5 in. AC, Subgrade CBR = 3

260

Break-Even Cost of Geosynthetic for Given Savings in
Stone Base Thickness and Stone Cost .

260

Placement of Wide Fill to Take Slack Out of
Geosynthetic

263

80

81
82
83
84

87

88

89

90

91
92

vii

LIST OF FIGURES (continued)
Page

Figure

The Relationship Between Stiffness and CBR for
Compacted Samples of Keuper Marl for a Range of Stress
Pulse Amplitudes (After Loach)

B-3

Results From Suction-Moisture Content Tests on
Keuper Marl (After Loach)
...

B-6

Permanent Axial and Radial Strain Response of Keuper
Marl for a Range of Stress Pulse Amplitudes (After
Bell)

B-8

Stress Paths Used in Cyclic Load Triaxial Tests for
Granular Materials

B-10

Permanent Axial and Radial Strains Response of Sand
and Gravel During Repeated Load Triaxial Test . .

B-11

Permanent Axial and Radial Strains Response of
Dolomitic Limestone During Repeated Load Triaxial
Test at Various Moisture Contents (w) and Degree
of Saturation (Sr)

B-12

Results of Standard Compaction Tests for the Granular
Materials

B-15

Relationship Between Normal and Maximum Shear Stress
in Large Shear Box Tests

B-16

Variation of Axial Strain with Load in Wide-Width
Tensile Tests

B-19

Results of Creep Tests at Various Sustained Loads
for the Geosynthetics During the First 10 Hours .

B-20

B-11

Summary of Hot-Mix Design Data by the Marshall Method .

B-21

B-12

Gradation Curves for Aggregates Used in Marshall Tests .

B-22

C-1

Tentative Layout of Proposed Experimental Plan

C-3

C-2

Preliminary Instrument Plan for Each Test Section

B-1

B-2
B-3

B-4
B-5
B-6

B-7
B-8
B-9
B-10

viii

.

C-7

LIST OF TABLES
Page

Table
1
2

3
4
5
6
7
8
9
10
11
12
13

14
15

Summary of Permanent Deformation in Full-Scale
Pavement Sections on a Compacted Sand Subgrade .

23

Comparison of Measured and Calculated Response for
a Strong Pavement Section: 3.5 in. Asphalt
Surfacing; 8 in. Crushed Stone Base

38

Anisotropic Material Properties Used for Final
Georgia Tech Test Study

39

Comparison of Measured and Calculated Response
for Nottingham Series 3 Test Sections

39

Aggregate Base Properties Used in Cross- Anisotropic
Model for Sensitivity Study

43

Nonlinear Material Properties Used in Sensitivity
Study

43

General Physical Characteristics of Good and Poor
Bases and Subgrade Soil Used in the Rutting Study .

51

AASHTO Design for Pavement Sections Used in
Sensitivity Study

55

Effect of Geosynthetic Reinforcement on Pavement
Response: 2.5 in. AC, E s = 3500 psi

58

Effect of Geosynthetic Reinforcement on Pavement
Response: 6.5 in. AC, E s = 3500 psi

60

Effect of Geosynthetic Reinforcement on Pavement
Response: 2.5 in. AC, E s = 6000 psi

62

Effect of Geosynthetic Reinforcement on Pavement
Response: 2.5 in. AC, E s = 12,500 psi

64

Effect of Geosynthetic Reinforcement Position on
Pavement Response: 2.5 in. AC, E s = 3500 psi

71

Effect of Initial Slack on Geosynthetic
Performance

77

Effect of Base Quality on Geosynthetic Reinforcement Performance

77

ix

LIST OF TABLES (continued)
Page

Table
Effect of Prestressing on Pavement Response:
2.5 in. AC, E s = 3500 psi

81

17

Summary of Test Sections

85

18

Specification of Hot Rolled Asphalt and Asphaltic
Concrete

88

19

Properties of Geosynthetics Used

93

20

Layer Thickness of Pavement Sections and Depth of
Geosynthetics From Pavement Surface

105

Summary of Construction Quality Control Test Results
for All Test Series

108

Summary of Results from Falling Weight Deflectometer
Tests Performed on Laboratory Test Sections

109

Transverse Loading Sequence Used in Multiple Track
Test Series 2 through 4

114

Description of Test Sections Used in Laboratory
Experiment and Purpose of the Supplimentary Single
Track Tests

117

Summary of Measured Pavement Response Data Near the
Beginning and End of the Tests for All Test Series

119

Summary of Measured Pavement Response for All Test
Series

127

Summary of Lateral Resilient Strain in Geosynthetics
and Longitudinal Resilient Strain at Bottom of AsphaltAll Test Series

133

Tentative Stiffness Classification of Geosynthetic
for Base Reinforcement of Surfaced Pavements

150

Influence of Geosynthetic Position on Potential
Fatigue and Rutting Performance . ...

166

Influence of Asphalt Thickness and Subgrade Stiffness
on Geosynthetic Effectiveness

167

Influence of Aggregate Base Quality on Effectiveness
of Geosynthetic Reinforcement
. ..

173

16

21

22

23

24

25

26

27

28

29

30

31

21) 230 Aggregate Gradations Used by Pennsylvania DOT for Open-Graded Drainage Layer (OGS) and Filter Layer (2A).S. Polypropylene and Polyester Polymers 243 45 Effect of Environment on the Life of a Polypropylene 246 B-1 Results of Classification Tests for Keuper Marl . B-4 B-2 Summary of Resilient Parameters for Granular Materials Obtained from Cyclic Load Triaxial Tests B-13 Summary of Large Shear Box Tests B-17 32 33 34 37 38 39 40 41 44 B-3 xi . Ref. 232 Separation Number and Severity Classification Based on Separation/Survivability 232 Guide for the Selection of Geotextiles for Separation and Filtration Applications Beneath Pavements 238 Pavement Structural Strength Categories Based on Vertical Stress at Top of Subgrade 240 42 Partial Filtration Severity Indexes 240 43 General Environmental Characteristics of Selected Polymers 243 Summary of Mechanisms of Deterioration. Advantages and Disadvantages of Polyethylene. . 106) 206 35 Preliminary Subgrade Strength Estimation 214 36 Vertical Stress on Top of Subgrade for Selected Pavement Sections 214 U.LIST OF TABLES (continued) Page Table Typical Friction and Adhesion Values Found for Geosynthetics Placed Between Aggregate Base and Clay Subgrade 188 Beneficial Effect on Performance of Prestressing the Aggregate Base 197 Design Criteria for Geosynthetic and Aggregate Filters (Adapted from Christopher & Holtz. Army Corps of Engineers Geosynthetic Filter Criteria (Ref.

LIST OF TABLES (continued) Table B-4 Page Comparison of Marshall Test Data for Two Asphaltic Mixes xii B-24 .

_ .

William S. the University of Nottingham. the University of Nottingham. universities and the geosynthetics industry who made valuable contributions to this project. Richard D. Geosynthetics were supplied by Netlon Ltd. The following Research Assistants at Georgia Tech participated in the study: Jorge Mottoa. Francis Chan performed the experimental studies at the University of Nottingham. Brown. Lan Yisheng and Mike Greenly gave much valuable assistance in analyzing data. was Principal Investigator. Research Assistant. the University of Nottingham. Professor of Civil Engineering. Professor Brown and Francis Chan. Stephen F. Orr. the Georgia Institute of Technology. The Georgia Institute of Technology was the contractor for this study. and the Nicolon Corporation. University of Nottingham was Co-Principal Investigator. and the Department of Civil Engineering. sincere appreciation is extended to the many engineers with state DOT's. and Yan Dai performed the numerical calculations. Brodrick. Barksdale. Department of Civil Engineering. Barry V. The authors of the report are Professor Barksdale.. .ACKNOWLKDGMENTS This research was performed under NCHRP Project 10-33 by the School of Civil Engineering. Finally. Georgia Tech. gave valuable assistance in setting up the experiments. Professor of Civil Engineering. The work performed at the University of Nottingham was under a subcontract with the Georgia Institute of Technology.

.

geogrid stiffness should be at least 1500 lbs/in. A linearly elastic finite element model having a crossanisotropic aggregate base gave a slightly better prediction of response than a nonlinear finite element model having an isotropic base. (2) prestressing the aggregate base by pretensioning a geosynthetic. xiv . The greatest benefit of reinforcement appears to be due to small changes in radial stress and strain in the base and upper 12 in.ABSTRACT This study was primarily concerned with the geosynthetic reinforcement of an aggregate base of a surfaced. Both largescale laboratory pavement tests and an analytical sensitivity study were conducted. Greatest improvement occurs when the material is near failure. Reinforcement appears to be effective for reducing rutting in light sections (SN < 2. flexible pavement. A geogrid performed considerably better than a much stiffer woven geotextile. Prerutting without reinforcement gave performance equal to that of prestressing.5 to 3) placed on weak subgrades (CBR < 3). and significantly better than just reinforcement. filtration and durability were also considered. and (3) prerutting the aggregate base with and without reinforcement. Specific methods of reinforcement evaluated included (1) reinforcement placed within the base. Both prerutting and prestressing the aggregate base were found experimentally to significantly reduce permanent deformations. Separation. of the subgrade. Prerutting is inexpensive to perform and deserves further evaluation.

- .

The term geosynthetic as used in this study means either geotextiles or geogrids manufactured from polymers. (2) prestressing the aggregate base by means of pretensioning a geosynthetic. The accurate prediction of tensile strain in the bottom of the base was found to be very important. (150200 mm) thick aggregate base. Specific methods of improvement evaluated included (1) geotextile and geogrid reinforcement placed within the base. The analytical sensitivity study considered a wide range of pavement structures.5. (25-38 mm) thick asphalt surfacing placed over a 6 or 8 in. flexible pavement. The subgrade was a silty clay subgrade having a CBR of about 2. A 1500 lb. (6.0 to 1.5 in. A linearly elastic finite element model having a cross-anisotropic aggregate base was found to give a slightly better 1 . Extensive measurements of pavement response from this study and also a previous one were employed to select the most appropriate analytical model for use in the sensitivity study. The large-scale pavement tests consisted of a 1.7 kN) moving wheel load was employed in the laboratory experiments. Analytical Modeling. Larger strains cause greater forces in the geosynthetic and more effective reinforcement performance. REINFORCEMENT Both large-scale laboratory pavement tests and an analytical sensitivity study were conducted. subgrade strengths and geosynthetic stiffnesses.SUMMARY This study was primarily concerned with the geosynthetic reinforcement of an aggregate base of a surfaced. and (3) prerutting the aggregate base either with or without geosynthetic reinforcement.

Reductions in rutting due to reinforcement occur in only about the upper 12 in.5 to 3) on a weak subgrade (CBR < 3) potentially can significantly reduce the permanent deformations in the subgrade and/or the aggregate base. As the strength of the pavement section increases and/or the materials become stronger. strain and deflections are all relatively small for pavements designed to carry more than about 200. Reinforcement of a thin pavement (SN = 2. The low resilient modulus existing at the top of the subgrade causes a relatively large tensile strain in the bottom of the aggregate base. A modest improvement in fatigue life can be gained from geosynthetic reinforcement. 2 . (300 mm) of the subgrade. As a result. As a result. the state of stress in the aggregate base and the subgrade moves away from failure. Forces developed in the geosynthetic are relatively small. and hence much larger forces in the geosynthetic than for a subgrade whose resilient modulus is constant with depth. Mechanisms of Reinforcement.000 equivalent 18 kip (80 kN) single axle loads. typically being _less than about 30 lbs/in. (0. The resilient modulus of the subgrade was found to very rapidly increase with depth. The effect of geosynthetic reinforcement on stress. The greatest beneficial effect of reinforcement appears to be due to small changes in radial stress and strain together with slight reductions of vertical stress in the aggregate base and on top of the subgrade.prediction of tensile strain and other response variables than a nonlinear finite element model having an isotropic base.37 N/m). geosynthetic reinforcement of an aggregate base will have relatively little effect on overall pavement stiffness. the improvement caused by reinforcement rapidly becomes small.

‘ 3500 psi. Light to moderate strength sections placed on weak subgrades having a CBR < 3 (E s . the minimum stiffness to be used for aggregate base reinforcement applications should be about 1500 lbs/in. From the experimental and analytical findings.5 to 3 if reduction in subgrade rutting is to be achieved by geosynthetic reinforcement. sections with asphalt surface thicknesses much greater than about 2. (4. the improvement in performance due to reinforcement should rapidly become greater. (25-50 mm).5 to 3. (1.5 in.9-7. Also. 24 MN/m 2 ) are most likely to be improved by geosynthetic reinforcement. For weak subgrades and/or 3 .3 kN/m) can give reductions in base thickness on the order of 10 to 20 percent based on equal strain criteria in the subgrade and bottom of the asphalt surfacing.5 times as great as the geogrid.Type and Stiffness of Geosynthetic.8 kN/m) for geogrids and 4000 lbs/in. Strong pavement sections placed over good subgrades would not in general be expected to show any significant level of improvement due to geosynthetic reinforcement of the type studied. Light sections on weak subgrades reinforced with geosynthetics having equivalent stiffnesses of about 4000 to 6000 lbs/in. The experimental results indicate that a geogrid having an open mesh has the reinforcing capability of a woven geotextile having a stiffness approximately 2. (4. For light sections this corresponds to actual reductions in base thickness of about 1 to 2 in.9 kN/m) for woven geotextiles. Reinforcement Improvement. The structural section in general should have AASHTO structural numbers no greater than about 2. As the structural number and subgrade strength decreases below these values.3-4. Improvement Levels. (64-90 mm) would in general be expected to exhibit relatively little improvement even if placed on relatively weak subgrades.

and significantly better performance compared to the use of stiff to very stiff. Considerably more reduction in rutting occurs for the thinner sections on weak subgrades than for heavier sections on strong subgrades. The total expense associated with 4 .low quality bases. (64-90 mm) in thickness for the reinforcement to be most effective. reduce rutting. The cost of prerutting an aggregate ba'e at one level would be on the order of 25 percent of the inplace cost of a stiff geogrid (S g = 1700 lbs/in. Geosynthetic Position. nonprestressed reinforcement. For light pavement sections constructed with low quality aggregate bases. Stress relaxation over a long period of time. Geosynthetic reinforcement of a low quality aggregate base can.. 2. particularly if a good subgrade is present.5 to 3.1 kN/m). and the base is of high quality and well compacted. The asphalt surface should in general be less than about 2. This would be particularly true if the subgrade is known to have rutting problems. however. For pavements constructed on soft subgrades. under the proper conditions. total rutting in the base and subgrade of light sections might under ideal conditions be reduced on the order of 20 to 40 percent. the reinforcement should be in the middle of the base. might significantly reduce the effectiveness of prestressing the aggregate base. The laboratory experiments indicate prerutting without reinforcement should give performance equal to that of prestressing. the reinforcement should probably be placed at or near the bottom of the base. PRERUTTING AND PRESTRESSING Both prerutting and prestressing the aggregate base were found experimentally to significantly reduce permanent deformations within the base and subgrade.5 in. Low Quality Base.

When an open-graded drainage layer is placed above the subgrade. angular open-graded aggregates placed directly upon a soft or very soft subgrade are most critical with respect to separation. The severity of erosion is greatly dependent upon the structural thickness of the pavements. 5 . which determines the stress applied to the subgrade. Sand filters generally perform better than geoextile filters. although satisfactorily performing geotextiles can usually be selected. Thick nonwoven geotextiles perform better than thin nonwovens or wovens. partly because of their three-dimensional effect. A very severe environment with respect to subgrade erosion exists beneath a pavement which includes reversible. Full-scale field experiments should be conducted to more fully validate the concept of prerutting. possibly turbulent flow conditions. DURABILITY Under favorable conditions the loss of strength of typical geosynthetics should be on the order of 30 percent in the first 10 years. Large.prestressing an aggregate base would be on the order of 5 and more likely 10 times that of prerutting the base at one level when a geosynthetic reinforcement is not used. SEPARATION AND FILTRATION Separation problems involve the mixing of an aggregate base/subbase with the underlying subgrade. and develop appropriate prerutting techniques. Either a sand or a geotextile filter can usually be used to maintain a reasonably clean interface. Both woven and nonwoven geotextiles have been found to adequately perform the separation function. the amount of contamination due to fines moving into this layer must be minimized by use of a filter. They usually occur during construction of the first lift of the granular layer.

geogrids might exhibit a lower strength loss. Whether better reinforcement performance will result has not been demonstrated. ` 6 . geosynthetics. if selected to fit the environmental conditions. filtration and pavement reinforcement applications. should generally have a 20 year life. For reinforcement applications geosynthetic stiffness is the most important structural consideration. For separation.because of their greater thickness. Some geosynthetics become more brittle with time and actually increase in stiffness.

1. Both nonwoven and woven geotextile fabrics are made from polypropylene. and have an open mesh with typical rib spacings of about 1.7 x 10 9 m 2 ) of geotextiles annually. With time many new uses for geosynthetics have developed including the reinforcement of earth structures such as retaining walls.5. geotextiles have a very wide application in civil engineering in general and transportation engineering in specific. erosion control and haul road or railroad construction [3. These fabrics have widely varying material properties including stiffness. Geogrids are manufactured by a special process.5 inches (38-114 mm). Because of their great variation in type. nylon and polyethylene. 7 .6]. strength. The introduction of geogrids which are stiffer than the commonly used geotextiles has lead to the use of the general term "geosynthetic" which includes both geotextiles and geogrids. More recently polyethylene and polypropylene geogrids have been introduced in Canada and then in the United States [2]. and resulting material properties. composition. and creep characteristics [1] (1) . Early civil engineering applications of geosynthetics were primarily for drainage. The numbers given in brackets refer to the references presented in Appendix A.4].5 to 4. polyester.CHAPTER I INTRODUCTION AND RESEARCH APPROACH The geotextile industry in the United States presently distributes about 2000 million square yards (1. slopes and embankments [2. Growth rates in geotextile sales during the 1980's have averaged about 20 percent each year.

et al.13]. relatively little factual information has been developed concerning the utilization of geosynthetics as reinforcement in the aggregate base. The specific objectives of the project are as follows: S . OBJECTIVES OF RESEARCH One potential application of geosynthetics is the improvement in performance of flexible pavements by the placement of a geosynthetic either within or at the bottom of an unstabilized aggregate base. An important need presently exists for establishing the potential benefits that might be derived from the reinforcement of the aggregate base. flexible pavement structure. The overall objective of this research project is to evaluate from both a theoretical and practical viewpoint the potential structural and economic advantages of geosynthetic reinforcement within a granular base of a surfaced.The application of geosynthetics for reinforcement of many types of earth structures has gained reasonably good acceptance in recent years. Several investigations have also been conducted to determine the effect of placing a geogrid within the asphalt layer to prolong fatigue life [12. particularly with respect to the use of stiff geogrids as reinforcement in the asphalt surfacing. however. At the present time. The results of these studies appear to be encouraging. and the conditions necessary for geosynthetic reinforcement to be effective. Mitchell. [6] have recently presented an excellent state-of-the-art summary of the reinforcement of soil structures including the use of geosynthetics. A number of studies have also been performed to evaluate the use of geosynthetics for overlays [7-11]. Considerable interest presently exists among both highway engineers and manufacturers for using geosynthetics as reinforcement for flexible pavements.

3. consideration had to be given to the large number of factors potentially affecting the overall longterm behavior of a geosynthetic reinforced. Perform an analytical sensitivity study of the influence due to reinforcement of pertinent design variables on pavement performance. Develop a preliminary experimental plan including layout and instrumentation for conducting a full-scale field experiment to verify and extend to practice the most promising findings of this study. 4. Verify using laboratory tests the most promising combination of variables. flexible pavement structure. The potential effect on performance of geosynthetic slack which might develop during construction and also slip between the geosynthetic and surrounding materials were also included in the study. and the overall strength of the pavement structure. installation and longterm durability aspects. 9 . stiffness and strength.1. Of these factors the more important ones appeared to be geosynthetic type. Develop practical guidelines for the design of flexible pavements having granular bases reinforced with geosynthetics including economics. RESEARCH APPROACH To approach this problem in a systematic manner. geosynthetic location within the aggregate base. and (2) prerutting the geosynthetic. 2. Longterm durability of the geosynthetic was also felt to be an important factor deserving consideration. Techniques to potentially improve geosynthetic performance within a pavement deserving consideration in the study included (1) prestressing the aggregate layer using a geosynthetic.

The large-scale laboratory tests made possible not only verifying the general concept and mechanisms of reinforcement. Further. This analytical model permitted the inclusion of a geosynthetic reinforcing membrane at any desired location within the aggregate layer. First the most important variables affecting geosynthetic performance were identified. Emphasis in the investigation was placed on identifying the mechanisms associated with reinforcement and their effect upon the levels of improvement. but also permitted investigating in an actual pavement factors such as prerutting and prestressing of the geosynthetic which are hard to reliably model theoretically. and hence require verification. including both design and construction related factors. feedback 10 . As a result. the influence of the geosynthetic reinforcement is relatively small in terms of its effect on stresses and strains within the pavement. caution must be exercised in a study of this type in distinguishing between conditions which will and will not result in improved performance due to reinforcement. The general research approach taken is summarized in Figure 1.The potential importance of all of the above factors on pavement performance clearly indicates geosynthetic reinforcement of a pavement is a quite complicated problem. As the analytical study progressed. Then an analytical sensitivity study was conducted followed by large-scale laboratory tests. The analytical sensitivity studies permitted carefully investigating the influence on performance and design of all the important variables identified. The analytical studies were essential for extending the findings to include practical pavement design considerations. A nonlinear. isotropic finite element pavement idealization was selected for use in the sensitivity study.

from the test track study and another previous laboratory investigation showed that adjustments in the analytical model were required to yield better agreement with observed response. the role of geosynthetic stiffness. 11 . Lateral tensile strain developed in the bottom of the aggregate base and the tensile strain in the geosynthetic were considered to be two of the more important variables used to verify the cross-anisotropic model. This synthesis includes all important aspects of reinforcement such as the actual mechanisms leading to improvement. As a result. A detailed synthesis of the results was then assembled drawing upon the findings of both this study and previous investigations. This important feedback loop thus improved the accuracy and reliability of the analytical results. Permanent deformation in both the aggregate base and the subgrade was also evaluated. cross-anisotropic model was employed for most of the sensitivity study which agreed reasonably well with the observed experimental test section response. The equivalent designs were based on maintaining the same strain in the bottom of the asphalt surfacing and the top of the subgrade. The analytical model was employed to develop equivalent pavement structural designs for a range of conditions comparing geosynthetic reinforced sections with similar non-reinforced ones. equivalent structural designs and practical considerations such as economics and construction aspects. a linear elastic. The analytical results were then carefully integrated together with the large-scale laboratory test studies.

General Approach Used Evaluating Geosynthetic Reinforcement of Aggregate Bases for Flexible Pavements.Identify Reinforcement Problems (1) Design Variables (2) Practical Aspects Construction Durability Analytical Study Predict Response Develop Equivalent Sections Failure Mechanisms Fatigue Rutting V Verify Analytical Model Large-Scale Laboratory Tests Type Ceosynthetic Pavement Structure Prerutting Prestress Define Performance Mechanisms Synthesis of Results General Benefits Equivalent Designs Construction Aspects Durability Economics F. Overall Evaluation Figure 1. 12 .

The potential beneficial effects of aggregate base reinforcement are investigated in this study using both an analytical finite element model. The only position of the reinforcement considered is within an unstabilized aggregate base. Pavement thickness. 3. and by a large scale laboratory test track study. The analytical investigation permits considering a very broad range of variables including developing structural designs for reinforced pavement sections. The important general pavement variables considered in this phase of the investigation were as follows: 1. perhaps at least partially due to the emphasis presently being placed in this area by the geosynthetics industry. Presently the important area of reinforcement of pavements is rapidly expanding. Type and stiffness of the geosynthetic reinforcement. Location of the reinforcement within the aggregate base.CHAPTER II FINDINGS The potential beneficial effects are investigated in this Chapter of employing a geosynthetic as a reinforcement within a flexible pavement. The laboratory investigation was conducted to verify the general analytical approach. 2. Unfortunately. relatively little factual information is now available with which the designer can reliably access the proper utilization of geosynthetics for pavement reinforcement applications. Quality of subgrade and base materials as defined by their resilient moduli and permanent deformation characteristics. and to also study important selected reinforcement aspects in detail using simulated field conditions including a moving wheel loading. 13 . 4.

In the analytical studies equivalent pavement designs were developed for geosynthetic reinforced structural sections compared to similar sections without reinforcement. 14 . and also the individual rutting in the base and subgrade. Slip at the interface between the geosynthetic and surround materials. Potential improvement in performance is evidenced by an overall reduction in permanent deformation and/or improvement in fatigue life of the asphalt surfacing.5. 6. an analytical procedure was also employed to evaluate the effects of geosynthetic reinforcement on rutting permanent deformations. 8. Prestressing the aggregate base using a geosynthetic as the pretensioning element. 7. A detailed synthesis and interpretation of the many results presented in this chapter is given in Chapter III. pavement performance was accessed primarily by permanent deformation including the total amount of surface rutting. Equivalent sections were established by requiring equal tensile strain in the bottom of the asphalt layer for both sections. Prerutting the geosynthetic as a simple means of removing slack and providing a prestretching effect. Finally. For the test track study. Influence of slack left in the geosynthetic during field placement. constant vertical subgrade strain criteria were also used to control subgrade rutting.

REINFORCEMENT OF ROADWAYS UNSURFACED ROADS Geosynthetics are frequently used as a reinforcing element in unsurfaced haul roads. Beneficial effects are greatest when construction is on soft cohesive soils.LITERATURE REVIEW . typically characterized by a CBR less than 2. prevention of the contamination of the fill by fine particles. Mechanisms of Behavior Bender and Barenberg [3] studied both analytically and in the laboratory the behavior of soil-aggregate and soil-fabric-systems. the reinforcement of the fill layer was attributed primarily to the high tensile modulus of the geotextile element. The following four principle mechanisms of improvement were identified by by Bender and Barenberg when a geosynthetic is placed between a haul road fill and a soft subgrade: 1. Also. The economics of justifying the use of a geosynthetic must. be considered for each application [26]. in large scale test pits [16-20]. and 4.14]. Tests involving the reinforcement of unsurfaced roads have almost always shown an improvement in performance. 15 . and full-scale field trials [21-26]. confinement of the subgrade 3. however.13. This finding would of course apply for either geotextile or geogrid reinforcement. it is usually not as great when stronger and thicker subbases are involved [24]. Although improved performance may still occur. separation of the subgrade and fill layer. These tests have been conducted at the model scale in test boxes [3. confinement and reinforcement of the fill layer 2.

et al.Bender and Barenberg [3] concluded for relatively large movements. a reinforcing element confines the subgrade by restraining the upheaval generally associated with a shear failure. use of another geotextile increased failure to 10. [21] found plastic strains in the subgrade to be reduced by the presence of a geotextile.000 repetitions. The work of Bender and Barenberg [3] indicated that over ground of low bearing capacity having a California Bearing Ratio (CBR) less than about 2. Nevertheless. the conservative 16 . The importance of developing large rut depths (and hence large fabric strain) was later confirmed by the work of Barenberg [27] and Sowers. Structural Performance . frequently referred to as the tensioned membrane effect. et al. increases the bearing capacity of the soil as illustrated in Figure 2. Later work by Barenberg [27] and Lai and Robnett [29] emphasized the importance of the stiffness of the geotextile.Full-Scale Experimental Results Relatively few full-scale field tests have been conducted to verify the specific mechanisms which account for the observed improvement in performance of geosynthetic reinforced haul type roads. Confinement. with greater savings being achieved with the use of a stiffer reinforcement. Webster and Watkins [25] observed for a firm clay subgrade that one geotextile reinforcement increased the required repetitions to failure from 70 to 250 equivalent 18-kip (80 kN) axle loads.. uniform bases on very soft subgrades. Ramalho-Ortigao and Palmeira [26] found for a geotextile reinforced haul road constructed on a very soft subgrade that approximately 10 to 24 percent less cohesive fill was required when reinforcement was used. Ruddock. Another 2 to 3 inch (50-70mm) reduction in base thickness was also possible since aggregate loss did not occur during construction of coarse. [28]. the use of a geotextile could enable a 30 percent reduction in aggregate depth.

c- Se -‘. 3). 38).. r.„.. . • . — Seberado „ I lc 1/ .• i • • ist.":". Note: 1 in.. 17 — .." '-hoveranl fore• on matured* due to telattorclna element (1) foal teed eacc/oa watts relegate.%`tee' L 5 . 10 100 Traffic Coverages. // pa / lebstaaa . Effect of Reinforcement on Behavior of a Subgrade Subjected to Vertical Stress (After Bender & Barenberg. Maximum Surface Deformation as a Function of Traffic (After Barker..Zotolall • • /. 0 a se .o'..T. --// .• pui .°1‘.1 -1. N Figure 3. Ref..stella aaaaa eaa la ••• / • tf //f ///.:c 0.25 nun 3 0 ° N. • cp-„..4m Figure 2. Ref.4 • ‘' • .. Upheaval dua CO poach/go !allure • 11u/ / 17. (a) Meal earl *melee wi thout relalorterse .• • • •—• .

the resilient surface deflections and also resilient stresses and strains within the pavement structure are only slightly reduced by the inclusion of a reinforcing element. [35] and a theoretical study by Raad [36]. but even indicated that greater permanent deformations could occur as a result of geotextile reinforcement. To be effective as a reinforcing element. [39] whose findings indicate no measurable increase in pavement stiffness due to reinforcement. Further. These results are supported by the work of Barker [38] and also Forsyth et al. [37] not only agreed with this finding. the important reinforcing effects observed in unsurfaced haul roads are considerably less apparent. the geosynthetic must undergo tensile strain due either to lateral stretching or else large permanent deformations. Vokas and Stoll [33] and Barksdale and Brown [34] indicate that for low deformation pavements. Both a laboratory study by Barvashov et al. several studies have indicated that under the proper conditions geosynthetic reinforcement can 18 . A full scale field study by Ruddock et al. however. In apparent conflict with these findings. Theoretical studies by Thompson and Raad [32]. SURFACE PAVEMENTS For surfaced pavements which undergo a small level of permanent deformation. have shown that prestressing the aggregate layer using a membrane greatly alters the stress state and potentially could result in improved pavement performance. a large scale laboratory study by Brown et al. [21.recommendation was made that no reduction in aggregate thickness should be allowed.30] on a reasonably heavy pavement section with a moderately thick bituminous surfacing has shown reinforcement to have little measurable effect on resilient pavement response.

and hence the potential for improvement in fatigue performance.result in improved performance.Thick Bituminous Surfacing Full-scale experiments conducted by Ruddock.30] included two sections having a 6. and clearly demonstrate that additional study is required to define the mechanisms and level of improvement associated with geosynthetic reinforced flexible pavements. The above findings appear to be somewhat conflicting. The geotextile used was stiff (S g @ 5 percent 1 3400 lbs/in..(10 mm) thick asphaltic seal.3 in. A more detailed summary of some of the experimental findings involving geosynthetic reinforcement is given in the following subsections. The road experienced considerably reduced permanent surface deformations. The primary effect on pavement response was an important reduction in tensile strain in the bottom of the asphalt. but dynamic response was unchanged by the presence of the geogrid.8 percent. (160 mm) thick bituminous surfacing and a 12 in./in. Field Tests . if desired. A field investigation by Barker [37] and a laboratory study by Penner et al. be skipped without loss of continuity. [40] have also shown that geogrid reinforcement can result in reduced permanent deformations. the woven geotextile had a strength of about 474 lb. (300 mm) thick crushed granite base. (83 kN/m) in each direction.. Pappin [23] has reported a pavement reinforcing experiment carried out in New South Wales. One of these sections had a woven multi-filament polyester geotextile reinforcement in the bottom of the granular base. Potter and McAvoy [21.4 in. A recent study by van Grup and van Hulst [41] involved placing a steel mesh at the interface between the asphalt and the aggregate base. 600 kN/m) and 19 . A stiff geogrid was placed at the bottom of an aggregate base of a pavement surfaced with a 0. This discussion could. and an elongation at failure of 14.

transient stress and strain in the subgrade. the pavement should have been loaded sufficiently to cause rutting to develop. unstabilized aggregate base reinforced by a stiff to very stiff geogrid. that resilient vertical subgrade stresses and strains were not significantly changed by fabric inclusions. Field Tests . Barker [38] has studied the performance of a pavement having an open-f.5 percent at a depth of 11.5 kN) was applied for 4600 repetitions.3 in (160 mm) bituminous surfacing. Ruddock et al.9 kips (97. and transient tensile strain in the bottom of the bituminous layer. raded. A rear axle load of 21. Because of the use of a thick bituminous surfacing.8 in. permanent strain in the geotextile. (300 mm). with the axle loading being increased to 30 kips (133 kN) for an additional 7700 passes. no difference in structural performance was observed between the geotexile reinforced sections and the control section.7 percent at the top to 3.had an elastic modulus of about 72. found in the trials at Sandleheath. 2 (500 kN/m 2 ). The geosynthetic stiffness S g is defined as the force applied per unit length of geosynthetic divided by the resulting strain. The sections were constructed on a London clay subgrade having a CBR increasing with depth from about 0. although transverse resilient strains were somewhat reduced. Measurements made included surface deformations. The test sections consisted of a 3 in. For the conditions of the test which included a 6. (75 mm) asphalt surfacing overlying a 20 .Geogrid and Heavy Loading Recently. Loading was applied by a two-axle truck having dual rear wheels.000 lbs/in. it is doubtful that the conclusions reached would have been significantly changed. The geogrid was placed at the center of the aggregate base. To demonstrate if some improvement in permanent deformation could be achieved due to reinforcement. however.

The granular base.000 load repetitions. The non-conventional pavement section studied at WES had (1) a very open-graded granular base. Falling weight deflectometer (FWD) tests showed the stiff to very stiff reinforcement did not affect the measured FWD deflection basins throughout the experiment.000 repetitions of a heavy moving aircraft load. (700 kN/m).5 m) width. and about 21 percent less at a rut depth of 2 in. A 6 in. The 27-kip (120 kN) load applied to the pavement consisted of a single tire inflated to 265 psi (1./in. (50 mm) as shown in Figure 3. was loose and unstable to most traffic. Maximum observed rutting of the reinforced section was about 8 percent less than the unreinforced section at a rut depth of 1 in. (150 mm) thick. even after compaction. (25 mm). very open-graded base consisting of No. These factors greatly complicate translating the test results to conventional pavements. 21 . 57 crushed limestone. (150 mm) cement stabilized clay-gravel subbase was constructed to provide a strong working platform for the open-graded base.000 lbs. Also. This finding indicates similar stiffnesses and effective layer moduli of the reinforced and unreinforced sections. The pavement was trafficked over a 60 in. the reinforcement was placed in the middle of the granular base. (2) a cement stabilized supporting layer and (3) was subjected to a very high wheel load and tire pressure. (1.8 MN/m 2 ). An unstable base of this type would appear to be a good candidate for reinforcing with the stiff geogrid.6 in. The general condition of the two pavements appeared similar after 1. The pavement was subjected to 1. The subgrade was a sandy silt having a CBR of 27 percent. This geogrid used had a secant stiffness at 5 percent strain of about 4. Subsequent trench studies indicated that most of the permanent deformation occurred in the subgrade and not the base.

and the rutting which developed as a function of load repetitions is given in Figure 4. Steel Mesh Reinforcement A hexagonal wire netting of steel was placed at the interface between a crushed rubble aggregate base and the asphalt surfacing in a large scale test track experiment described by van Grup and van Hulst [41]. Seven different reinforced sections were 22 . coarse sand.For this well constructed pavement. Reinforcement of a weak section which did not have an aggregate base resulted in a 40 percent reduction in rutting at about 0. would have a very significant beneficial effect on fatigue performance. This large level of reduction in strain. important reductions in permanent deformations occurred due to reinforcement only after the development of relatively large permanent deformations. Large-Scale Laboratory Tests . however. Rutting. however. A summary of the test conditions is given in Table 1. observed in the bottom of the asphalt surfacing. Nonwoven Geotextiles Brown.4 in.5 in (12 mm) rut depth. et al. Better performance might have been obtained had the reinforcement been placed at the bottom of the base. About an 18 percent reduction in tensile strain was. The asphalt surfacing was 2.Low Stiffness. primarily occurred in the subgrade. if permanent. [37] investigated the effect of the placement of a nonwoven geotextile within and at the bottom of the aggregate base of bituminous surfaced pavements. (200-400 mm). The reinforcement was placed at the center of the aggregate base to improve its performance. (60 mm) thick. and the aggregate base varied in thickness from 8 to 16 in. Reinforcement made little difference in rutting performance for the stronger sections having rubble aggregate bases. The subgrade consisted of a compacted.

Note: 1.3 0. The steel mesh reinforcement was placed at the aggregate base/asphalt surfacing interface. 41). 200 400 600 800 1000 Asphalt Strain Section #1 (pm/m) Figure 4. 23 ./ @ Top of Base NO NO NO NO YES YES Crushed Rubble 0 7. GA 4.8 0 47.2 35.9 11.5 47.3 35. 1000 E Loading Time : 0.4 2.49 0.4 2.4 2.) @ 140.000 Reps.Table 1 Summary of Permanent Deformation in Full-Scale Pavement Sections on a Compacted Sand Subgrade LAYER THICKNESSES AND PERMANENT DEFORMATION OF SECTIONS (in.98 Sand Clayey Sand Permanent Surface Deformation (in.4 31.1 200 r-I C 0.4 Steel Mesh Reinf.) 600 C 400 C.7 11. Ref.4 2.55 0.55 0.4 2.2 Force : 6 kN Up 800 C 0 4.2 39.4 - - - - - - 1.) LAYER 1 2 3 4 5 6 Dense Asphaltic Concrete 2.8 15.44 0. Comparison of Strain at Bottom of Asphalt Surfacing With and Without Mesh Reinforcement (After Van Grup and Van Hulst.

9 in. Both vertical and lateral resilient and permanent strains were also found to be greater in the base and subgrade of all of the reinforced sections (Figure 5b). (107-175 mm). for each condition a similar control section was also tested without reinforcement. and using geotextiles strengthened by stitching. Two layers of reinforcement were also employed in some tests. (78 kN/m).4 kip (15 kN). nonwoven. The inclusion of the nonwoven geotextiles in the aggregate base in most tests appeared to cause a small increase in rutting (Figure 5a). Maximum surface rutting was less than about 1 in. (25 mm). In light of more recent findings. The pavements rested on a silty clay subgrade having a CBR that was varied from 2 to 8 percent. (220 kN/m) and 445 lbs/in. the relatively low geosynthetic stiffness probably also helps to explain the results.1 in. The crushed limestone base was varied in thickness from 4. several factors suggest compaction of the aggregate above the geosynthetic may not have been as effective when the geotextile was present. melt bonded geotextiles were used in the study. Two very low to low stiffness. which resulted in relatively small strains in the geosynthetic. A moving wheel load was used having a magnitude of up to 3. The experiments included placing the geotextiles within the granular layer./in. and no increase in effective elastic stiffness of the granular layer. The bituminous surfacing of the seven test sections varied in thickness from 1.studied. (37-53 mm). Finally.5 to 2.2 to 6. The poor performance of the reinforced sections was attributed to a lack of adequate aggregate interlock between the base and the geotextiles. 24 . These geotextiles had a secant stiffness at one percent strain of about 1270 lbs.

Deflection and Lateral Strain Measured in Nottingham Test Facility (After Brown. Ref. et al. 7. 25 mm Figure 5. 37). 200 Subgrade Subgrade 300 300 (b) Lateral Strains Note: 1 in. E3 Reinforced • V 4 100 Granular Base V 1 100 V • . Granular ej Base 9 1=1 200 O . • Unreinforced —5 Bit. 25 . 0 LEGEND Bit.• Radial Offset Distance (mm) 450 600 300 S0 150 0 150 300 450 600 I o c O 10 LJ Reinforced Unreinforced (a) Surface Deflections Lateral Resilient Strain (millistrain) 5 0 '• Permanent Lateral Strain..

After testing. (300 mm) diameter plate. the CBR of Loop 3 was found to have increased by a factor of about 2 or even more. The subgrade was a fine beach sand having a CBR of typically 4 to 8 before the tests.Large-Scale Laboratory Tests Using Stiff Geogrids Penner. (0. (75 or 100 mm). Placement of the geogrid within the granular base was found to result in a significant reduction in pavement deformation when placed in the middle or near the bottom of the base. sections having geogrid reinforcement at the bottom and midheight exhibited only about 32 percent of the 0. A stationary. (312 kN/m). An increase in CBR might also have occurred in other sections although the researchers assumed for analyzing test results an increase did not occur. 26 .2 in. (100-300 mm). The asphalt surface thickness was either 3 or 4 in.9 m) deep. Important improvements in performance were found in this test for deformations of the reinforced section as small as 0. (200 mm) granular base and 3 in. (15 mm) deformation observed in the unreinforced section. In contrast with the above findings. The aggregate base was well-graded and was varied in thickness from 4 to 12 in. (75 mm) asphalt surfacing. peat was mixed with the fine sand at a high water content to give a very weak subgrade having an initial CBR of only 0. In one series of tests. S g at 5 percent strain of the geogrid used in the experiment was about 1780 lbs/in. The secant stiffness.6 in. laboratory CBR testing indicated a CBR value of 100 percent or more. For one section having an 8 in.8 to 1. (5 mm). 9 kip (40 kN) cyclic load was applied through a 12 in.2 percent. et al. Little improvement was observed when the reinforcement was located at the top of the base. The base had a reported insitu CBR value of 18 percent. [40] studied in the laboratory the behavior of geogrid reinforced granular bases using a shallow plywood box 3 ft.

27 .use of geogrid reinforcement in under-designed sections on weak subgrades showed no apparent improvement until permanent deformations became greater than roughly 1 in. (25 mm).

43]. cross-anisotropic model were used to idealize selected pavement sections reinforced with a geosynthetic. The stiffness of a geosynthetic used for pavement reinforcement applications is an important but often underrated or overlooked aspect that has a considerable effect upon the ability of reinforcement to improve performance. Therefore the capabilities of this comprehensive program are only 28 . As a result the cross- anisotropic formulation was selected after considerable study as the primary model for this study. S gcan be determined by stretching it. MODELING PAVEMENTS WITH GEOSYNTHETIC REINFORCEMENT The GAPPS7 finite element model has been described in detail elsewhere [42.ANALYTICAL STUDY The analytical study was performed using a comprehensive finite element program called GAPPS7. The units of geosynthetic stiffness S g are. for example. The stiffness should be determined at strains no larger than 2 to 5 percent for pavement reinforcement applications. and dividing the force per unit length applied by the corresponding induced strain. Most geosynthetics suitable for pavement reinforcement for practical purposes can be assumed to perform in a linear manner for the small levels of geosynthetic strain that should develop within pavements designed for small levels of permanent deformation. The GAPPS7 finite element program was developed previously to predict the response of surfaced or unsurfaced pavements reinforced with a geosynthetic [16. The cross-anisotropic model was found to in general give better agreement with observed pavement response than the isotropic. The stiffness of the geosynthetic.43]. pounds per inch. Both a nonlinear elastic-plastic model and a linear. nonlinear model.

The GAPPS7 program models a general layered continuum reinforced with a geosynthetic and subjected to single or multiple load applications.e. and 6. The main program handles the input. Provision for solving either plane strain or axisymmetric problems. GAPPS7 consists of a main program and twelve subroutines. The ability to consider either small or large displacements which might for example occur under multiple wheel loadings in a haul road. be changed for each loading cycle to allow considering time and/or load dependent changes in properties. and repetitive loadings. Material properties can. plastic and failure response.briefly summarized in this section. An automatic finite element mesh generation program MESHG4 is used to make the GAPPS7 program practical for routine use. and calls the appropriate subroutines. the load does not move across the continuum). however. 3. Only axisymmetric. performs the needed initializations. A no-tension analysis that can be used for granular materials. 4. 5. A two dimensional flexible fabric membrane element which can not take either bending or compression loading. Important features of the GAPPS7 program include: 1. 2. The ability to model materials exhibiting stress dependent behavior including elastic. are applied at a stationary position (i. In addition to 29 . The twelve subroutines perform the actual computations. The GAPPS7 program does not consider either inertia forces or creep. small displacement analyses were performed for this study using a single loading. Modeling of the fabric interfaces including provisions to detect slip or separation. when used.

MESHG4 completely generates the finite element mesh from a minimum of input data. Using this model the resilient modulus is usually considered to very rapidly decrease linearly as the deviator stress increases a small amount above zero. the resilient modulus stops decreasing and may even very slightly increase in a linear manner. A plotting program called PTMESH can be used to check the generated mesh. The most commonly used nonlinear model for the resilient modulus of cohesionless granular base materials is often referred to as the k-6 model (Figure 6b) which is represented as E r = K ao N (1) where E r = resilient modulus of elasticity. Resilient Properties Three different models can be utilized in the GAPPS7 program to represent the stress dependent elastic properties of the layers. often referred to as contour models.47] to more 30 . After a small threshold stress is exceeded. These supplementary programs greatly facilitate performing finite element analyses and checking for errors in the data.handling material properties. have been developed by Brown and his co-workers [46. and assist in interpreting the large quantity of data resulting from the application of the program. The stress dependent resilient modulus E r of the subgrade is frequently given for cohesive soils as a bi-linear function of the deviator stress al-a3 as shown in Figure 6. sometimes called M r detrminfolabtryesing k and e = material constants determined from laboratory testing 8 = sum of principle stresses. a l + a2 + a3 In recent years several improved models.

31 G e . (log scale) G -0 1 3 (a) Subgrade (b) Base Figure 6. •• = —0 vi 3 •o m a) o (E3. tle U1 0 cu E =K r G Sum of Principal Stresses.Res ilien t Mo du lus. G 3) r d 0 E r d 2 ) Deviator Stress. Resilient Modulus Relationships Typically Used for a Cohesive Subgrade and Aggregate Base.

[49] have shown that the material constants n and m are approximately related to G1 as follows: n = 0.31 m = 0.n. and their use will be discussed subsequently.03) 2 (03 _ 01) 2 Laboratory tests by Jouve et al.m = material properties evaluated in the laboratory from special cyclic loading stress path tests The model described by Equations (2) and (3) will be referred to throughout this study as the simplified contour model.028 G (5) 0.31 I (6) The bulk modulus K as given by equation (2) is always greater than zero which neglects the dilation phenomenon which can cause computational difficulties.03 G0.707 where J 2 Jr (a l - (4) 02 )2 4.G1.accurately characterize granular base materials. 32 . The contour model as simplified for routine use by Mayhew [48] and Jouve et al. [49] was used in this study. All three of the above nonlinear models for representing resilient moduli were employed in the present study. (02 . (a l + 02 + 03)/3 q = shear stress K1. For a general state of stress the shear stress q can be defined as q = 0. Following their approach the bulk modulus K and shear modulus G of the base can be calculated from the simplified relations K = K G = G1 (2) p (1-n) {1 + y ( 2 ) 2 } 1 p (1-m) (3) where: K = bulk modulus G = shear modulus p = average principal stress.

Partway through this study it became apparent that the usual assumption of material isotropy. Two independent comparison studies were performed to both verify the analytical model selected for use. The importance of the role which tensile strain developed in the reinforcing layer plays became very apparent as the analytical study progressed. strain and deflection response variables. a supplementary investigation was undertaken to develop modified models that could more accurately predict the tensile strain and hence the response of geosynthetic reinforced pavements. To be able to reliably predict the tensile strain in an unstabilized granular base is quite important in a study involving granular base reinforcement.PREDICTED PAVEMENT RESPONSE Little work has been carried out to verify the ability of theoretical models to accurately predict at the same time a large number of measured stress. An accurate prediction of tensile strain is required since the level of tensile strain developed in the base determines to a large extent the force developed in the geosynthetic and hence its effectiveness. Therefore. and the usually used subgrade and base properties including the k-6 type model were in general not indicating the level of improvement due to reinforcement observed in the weak section used in the first laboratory test series. well instrumented test section without geosynthetic reinforcement tested 33 . The first study involved theoretically predicting the response including tensile strain in the aggregate base of a high quality. and to assist in developing appropriate material parameters. The presence of a tensile reinforcement and relatively thick granular layers which have different properties in tension compared to compression greatly complicate the problem of accurately predicting strain in the aggregate layer.MODEL VERIFICATION .

The simplified. A homogeneous material has the same properties at every point in the layer.9 percent above optimum.5 kip (29 kN) uniform. (200 mm) thick granular base were cyclically loaded to failure [44. In the verification study a number of models were tried including the nonlinear finite element k-B and contour models. (90 mm) asphalt surfacing and an 8 in.5 in. A comparison of the observed and measured pavement response variables for each model is given in Table 2. circular loading were applied at a primary and six secondary positions. cross anisotropic model were selected as having the most premise. several pavement sections having a 3. The second study used the extensive measured response data collected from Test Series 3 of the large scale laboratory pavement tests conducted as a part of the present study. A cross-anisotropic representation has different elastic resilient material properties in the horizontal and vertical directions.previously by Barksdale and Todres [44. High Quality Aggregate Base Pavement As a part of an earlier comprehensive investigation to evaluate aggregate bases. nonlinear contour model and a linear elastic.4 million applications of a 6. An isotropic model has the same material properties such as stiffness in all directions.45].45]. A total of about 2. manual trial and error procedure was used to select material properties that gave the best overall fit of all of the measured response quantities. These results indicate that a cross 34 . A tedious. These sections were placed on a micaceous silty sand subgrade compacted to 98 percent of AASHTO T-99 density at a water content 1. High quality materials were used including the asphalt and the crushed stone base which was compacted to 100 percent of AASHTO T180 density. Unreinforced.

0 0 E 1. N 04 el 01 un E O t+. cn a) ao N 124 a E (x • ( -OI 9 a) a 0 subgra de is Es . 1..1.. IN •I ... 00a > O 0 0 0 Res ilien t mo du lu s o fbase is Eb .0 CO 0 • 4-1 0 .. 0 0 0 0 . • E IL 0 •••• 4•I 41 0 > -0 I3. IN v. -1 u-1u 0 N N N • .4 STRA IN ." -0 L13 .) 0. 0 .. .• CA .1 Z. 1--4 0 •r-I O CO .1"1 NI 0 .--/ 0 •••• 0 cn in (1) TI (1) E •• PP '0 '0 11.• 41 -2' 4. x CI I-I 10 c1 01 el h. . .I STR BOTTO L6-1 Measu re d an d Ca lcu 3. t ha t ... c•A RAINN OF BA _ 0 •.. .•1 mmmi 1) 2 o Measu re d N .0 IN 0 00 tri 1.. Aspha lt • c0 E-1 o co N RTI CAL S UBGR STRES S/ STRAI a.1 '4 I ■.•■ 11.0 0 (1. IN Z 1..0 en en N. 5 in...-1 O C..1 .1 ••-.. Cross ...--4 o 0 • • CO CO In rn I -Y 0 00 .An isotrop ic E s Con s tan t E Var iable s Fin ite Elemen t 0 NI 35 Mo de l ( 2 ) r un c0 0 o o CONDITION cd 14 co co . NI 0 0 . N r•I -1 01 r1 4-4 o 4-4 o co 0 01 70 N 01 O.. N 0 0 • • CO SI I re s ilien t p rop er t ies g iven in Ta ble 6.• ./1 N . 0 ..•"I ^ N N N • • V V . in t he lower t hir d o f t he base t he modu lus was ta ken as 40% o f b V 0 0 .1 N ... <1 01 '..1 UJ▪ 0) c0 CD AS E -6 10 ) O cn }.4..

the stiffness of the silty sand subgrade underwent a significant increase with depth. Later after the sensitivity study was underway it was discovered that the tensile strain in the base greatly increased if the subgrade modulus increases with depth. The isotropic. the average stiffness of the subgrade was quite high. probably much larger than generally believed at the present time. the one exception was the tensile strain in the bottom of the base which was about 30 percent too low. nonlinear finite element method could not predict at the same time large tensile strain in the bottom of the aggregate base. The cross-anisotropic material properties employed in the sensitivity study are summarized in Table 3. This important difference in measured strain is readily explained if the 36 . and the small observed vertical strains in the bottom and upper part of that layer. The significant decrease in strain and increase in confinement with depth probably account for most of this observed increase in stiffness with depth [62]. They are similar to those used for the homogeneous subgrade comparison in Table 2.anisotropic model is at least equal to. The better agreement with measured pavement response when using a subgrade resilient modulus that rapidly increases with depth is shown in Table 2. isotropic subgrade resilient modulus was used. Thus the important finding was made that the resilient modulus of the subgrade near the surface had to be quite low as indicated by the very large measured vertical strains on the subgrade. Since the total measured surface deflections were relatively small. At the time this comparison was made a homogeneous. homogeneous subgrade was able to predict measured variables to within about ± 20 percent. and perhaps better than the simplified contour model for predicting general pavement response. The cross-anisotropic model using an isotropic. Therefore.

Table 3 Anisotropic Material Properties Used for Final Georgia Tech Test Study Location in Pavement Resilient Modulus Vertical Poisson's Ratio Horizontal Vertical Horizontal Aggregate Base (Anisotropic) Top 1.s and Eh (avg) =35.818E Bottom b b 1.000 psi.10 0.15 0.0227E b b 0.4 0. Eb = resilient modulus of base as shown in table 2.4 0.0852E 0.4 0.75 where E =8000 psi h s .875E s = average resilient modulus of elasticity of subgrade.375E s 0.375E Middle 0. the numerical average of the three vertical resilient moduli of base = 38.75E s s 1.875E Bottom Note: 1. E s s 0.420E Middle lE b 0.75E s 1.4 0.4 0.200 psi.43 0.15 0.136E b 0.4 Subgrade (Isotropic) Top 0.45 0. Modular ratio E (avg)/E = 4.43 0. 37 .

Response of Geosynthetic Reinforced Sections The measured pavement response obtained from the three sections included in Test Series 3 of the laboratory tests provided an excellent opportunity to verify whether a cross-anisotropic model can be successfully used to predict the response of the two geosynthetic reinforced sections and the non-reinforced control section included in the study. Once again.7 kN) at a tire pressure of 80 psi (0. A soft clay subgrade (CL) was used having an average inplace CBR before trafficking of about 2. and a crushed stone base thickness of about 8.9. These comprehensive experiments. The wheel loading was 1.5 kips (6. (30 mm).2 in (208 mm). but under 38 .7 to 2. The comparison between the anisotropic model using the best fit material properties and the measured response is shown in Table 4 for each section. Also the cross-anisotropic model gave a much better estimate of the vertical stress on the subgrade and the vertical surface deflection than did the nonlinear model. The strain in the geosynthetic was over predicted by about 33 percent when the geosynthetic was located in the bottom of base.actual stiffness of the aggregate base is considerably greater in the vertical than the horizontal directions. These test sections had an average asphalt surface thickness of about 1. based on the measured strains. are described in detail in the last section of this chapter. the theory did a fair job of predicting observed response. the conclusion was reached that the resilient modulus of subgrade was quite low near the surface but rapidly increased with depth. which included the measurement of tensile strain in the aggregate base and also in the geosynthetic.6 MN/m 2 ).2 in. These sections were constructed over a subgrade having a very low back estimated average resilient modulus of about 2000 psi. Overall.

0 •1 7 •-. 1 0 ol uA soa o . aa ■-■ 0 < .. Stress / Str a in S 41:7_ 6' 7 E* 6 3 39 SUS 00 75 - 1 590Z 6 091.C1 71 1 E .I Measu re d Mo de l 1 Mo de l2 1371EC96 C00(1- Measur e d Mo de l1 116 11 OC LI 8617 097E00 71 - 011 1T au p 1 I 0081 SSC 7 L5 C 700Z9- at 70. . ..C OI Z 1. . CEOSYNTH ETIC IN MIDDL E O F BAS E CEOSYNT RETI C IN BOTTOM OF BASE Ra dia l ) E r (10 -6 r CONTROL S ECT ION . .1 b.0 • • • ...I1 cla I.0 . Me asu re d 087109 910051. .0 1) 1 3 I uI Pe R Stra in Bo ttom 2080 4 61Si 0087LLC 700 9 S- " id aA (9-01) A l 00C S 0099 79 91 7777..NO CEOSYNTHET I C 81 81 £ 86 7 0 Mo de l 1 0091 0065 66 S Z 7CC 7 (0 0079 . 1 +." W ..0 .. C. ... . . .•-• 1..0 79 0' 0 99 0 . .0 n n 1 4 1 Su bg..-.0 0 .' 0C6I 0051- fEO Z 0007 - A aft (9-00 Stress Stra in o z (p s i ) t v ( 1 0 -6 ) 0972 09 0 7 000S OCS Z OZ 9Z 00 09 Stra in Top o f (9.) 0 • .■ UtlU t (z) (' uT )A y I 090' 0 R0' 0 . JD n n I 1 1 C E •r -. 01 (u r is cil)" ) u Ter d s I C90' 0 09 0 ... ..0 9/0' 0 I ssa 1 1 5 (u t is q00 I 0081 0807 CC 4 7 > -Y VI . . . -.... • C ci 0 VD . 000 rep e t it ion s... 1 1 1 Ve rt.—.

This finding suggests relative comparisons of should generally be reasonably good.000 load repetitions at permanent deflections of 1.predicted by about 14 percent when located in the middle of the layer. perhaps due to a build up of pore pressures.4 million heavier load repetitions for the stronger sections on a better subgrade used in the first comparison. only strain was measured in the geosynthetic. were quite weak and exhibited very large resilient deflections.9. Nevertheless. Larger radial strains were measured in the bottom of the aggregate base than calculated by about 50 percent. these pavement sections. (38-50 mm) as compared to about 2. as originally planned. In summary.5 to 2 in. the subgrade used in Test Series 3 probably performed like one having a CBR less than the measured value of 2. These sections only withstood about 70. As a result the computed vertical strain at the top of the subgrade was too small by about the same amount. the calculated relative changes in observed response between the three sections did indicate correct trends. Undoubtedly the analytical studies are susceptible to greater errors as the strength of the pavement sections decrease toward the level of those used in the laboratory studies involving the very weak subgrade. The vertical stress on the top of the subgrade was about 50 percent too small. The cross anisotropic model did not do nearly as good in predicting the pavement response of the weak Test Series 3 sections compared to the stronger sections previously described. The postulation is presented that under repetitive loading. A reasonably strong section would in general be more commonly used in the field. Of considerable interest is the fact that the largest calculated geosynthetic stress was about 10 lbs/in (12 N/m). strains and stresses. and indicate correct relative trends of performance.7 to 2. 40 .

The subgrade used was isotropic and homogeneous. recall that strains in the aggregate base and geosynthetic were actually overpredicted for the tests involving a very soft subgrade. this important finding was not made until the sensitivity study was almost complete. Unfortunately. Also. simplified contour model was also used as the secondary method for general comparison purpose and to extend the analytical results to include slack in the geosynthetic and slip between the geosynthetic and base and subgrade. 41 .5 times the value reported. the stiffness of the geosynthetics actually used in the analytical sensitivity studies was 1.MODEL PROPERTIES USED IN SENSITIVITY STUDY The cross-anisotropic model was selected as the primary approach used in the sensitivity studies to investigate potential beneficial effects of geosynthetic reinforcement. A supplementary analytical study did show using a higher geosynthetic stiffness with a homogeneous subgrade gives comparable results to a model having a subgrade modulus increasing with depth. Tensile strains in the aggregate base and geosynthetic can be calculated directly by assuming a subgrade stiffness that increases with depth. The measured strain in the bottom of the aggregate base in the test section study that withstood 2.6 times the value calculated using the cross-anisotropic base model.4 million load repetitions (Table 2) was about 1. To approximately account for this difference in strain. In an actual pavement the development of large tensile strains in the granular base than predicted by theory would result in the reinforcing element developing a greater force and hence being more effective than indicated by the theory. the nonlinear. Also.

Therefore. Cross-Anisotropic Model Material Properties. 130 kN/m) were used in the theoretical analyses. Because of the small stresses and strains developed within the geosynthetics. The relative values of crossanisotropic elastic modulu and Poisson's ratios of the aggregate base used in the study are summarized in Table 5. The resilient modulus of the asphalt surfacing used in the sensitivity study was 250.35.5 scaling factor be 1000. 88. In turn the level of tensile strain in the aggregate base determines at least to a great extent the force developed in the geosynthetic.5 was about 42 .000 psi (1700 MN/m2 ). (22. For the primary sensitivity study the modular ratio used was 2. The ratio of the resilient modulus of the base to that of the subgrade has a significant influence on the tensile strain developed in the base for a given value of subgrade resilient modulus. For this study the cross-anisotropic modular ratio was defined as the vertical resilient modulus of the center of the base divided by the uniform (or average) resilient modulus of the subgrade. Since the force in the geosynthetic significantly influences the improvement in behavior of the reinforced pavement system. 6000 and 12.5. The resilient moduli of the subgrade included in the sensitivity analyses were 2000. The modular ratio of 2.500 psi (14. 41. 58. 24.Using the above engineering approximation. 88 kN/m). the corresponding stiffnesses reported as those of the sections would using the 1. actual geosynthetic stiffnesses. 3500. 6000 and 9000 lbs/in. they remain well within their linear range. S g = 1500. Hence nonlinear geosynthetic material properties are not in general required for the present study. using a modular ratio comparable to that actually developed in the field is very important. 4000 and 6000 lbs/in (15. 86 MN/m 2 ). The corresponding Poisson's ratio was 0.

43 0.10 Table 6 Nonlinear Material Properties Used in Sensitivity Study 1.0458E 0.640 3.43 0.45 0.300 4.925E 0.300 0 1.375E 0.15 Bottom 0.000 psi (isotropic) 2.180 0.138E 0. Granular Base: Position in Base Kl GI Y Very Good Crushed Stone Base Upper 2/3 14. Average Subgrade C s = 6.950 0.300 4.000 psi. .Table 5 Aggregate Base Properties Used in Cross-Anisotropic Model for Sensitivity Study Poisson's Ratio Resilient Modulus Location in Base Vertical Horizontal Vertical Horizontal Top 1.100 7. E r = 250.825E 0.5 30. Subgrade: Typical Subgrade E s (psi) given below (see Figure 6) 1P.14 Lower 1/3 5.620 0.0E 0.050 0.0 .000 3 800 4.000 2 750 4.) Point 03 (psi) Resilient Moduli Top Middle Bottom I 1300 16.000 16.12 Lower 1/3 1.320 1.15 Middle 1.4 43 . 1. ti = 0.35 2. ■ 0. Asphalt Surfacing: Isotropic.14 Poor Quality Gravel/Stone Base Upper 2/3 3.000 4. u .12 3.

if the resilient modulus of the entire subgrade was reduced calculated surface deflections were too small. The nonlinear options in the GAPPS7 program. Both studies comparing predicted and measured pavement response indicate the base performs as a cross anisotropic material. The resilient properties of the asphalt surfacing were the same as used in the cross-anisotropic model. Supplementary sensitivity studies were also carried out using modular ratios of 1. However. however.5 and 4.the value back calculated from the measured response of the test pavement on the very soft subgrade having an average modulus of about 2000 psi (14 MN/m2 ) as shown in Table 4. The modular ratio of 4. For example. only permit the use of isotropic properties. Nonlinear Properties The material properties used in the nonlinear finite element analyses were developed by modifying typical nonlinear properties evaluated in the past from laboratory studies using the measured response of the two test pavement studies previously described.5. 44 . The radial tensile strain in the bottom of the granular base could be increased by 1. Therefore some compromises were made in selecting the resilient simplified contour model properties of the aggregate base. Decreasing the resilient modulus of the top of the subgrade.5 was about that observed for the better subgrade which had an average resilient modulus of about 8000 psi (55 MN/m 2 ) as shown in Table 2. the small vertical strain and large lateral tensile strain in the aggregate base could only be obtained using the cross anisotropic model.

[49] multiplied by 1. Developing as good of comparisons with measured responses as shown in Table 2 and 4 for both the cross-anisotropic and nonlinear models required a considerable amount of effort. Reducing this resilient modulus caused the calculated vertical strain in the layer to be much greater than observed. For such relative comparisons the material properties developed are considered to be sufficiently accurate. only the relative response is required of pavements with and without geosynthetic reinforcement. 45 . A better match of calculated and measured response could probably be developed by further refinement of the process. For this sensitivity study.5 gave better values of vertical strain in the base. The compromise selected gave weight to increasing the radial tensile strain in the granular base as much as believed to be practical. The nonlinear material properties used in the upper two-thirds of the aggregate base are essentially the best and worst of the material properties given by Jouve et al. as well as the aggregate base properties. were developed from the tedious trial and error procedure used to match the measured response variables with those calculated. Increasing the stiffness by 1. Decreasing the resilient modulus of the lower part of the base. The subgrade material properties.4. The resilient properties used in the lower third of the base were obtained by multiplying the properties used in the upper portion of the base by 0. The nonlinear subgrade material properties used in the study are also summarized in Table 6.2. The nonlinear material properties used in the simplified contour model are given in Table 6.5.

Therefore the layer strain method proposed by Barksdale [501 was selected as an appropriate alternate technique for estimating the relative effect on rutting of using different stiffnesses and locations of reinforcement within the aggregate layer. the layer strain method consists of dividing the base and upper part of the subgrade into reasonably thin sublayers as illustrated in Figure 7. Total permanent deformation (rutting) is calculated for each sublayer by multiplying the permanent 46 . it was found that the GAPPS7 program in its present form is not suitable for predicting the effects on rutting due to the relatively small changes in lateral stress. The vertical permanent strain. c. Shear stresses do not act on these planes which greatly simplifies the analysis. is then calculated in each element knowing an accurate relationship between the permanent strain E p and the existing stress state acting on the element. During the numerous preliminary nonlinear computer runs that were performed early in this study. In summary.Estimation of Permanent Deformation The presence of the geosynthetic in the granular base was found to cause small changes in vertical and somewhat larger changes in lateral stresses (at least percentage-wise) within the granular layer and the upper portion of the subgrade. The representative element is located beneath the center of the loading where the stresses are greatest. respectively. The complete stress state on the representative element within each sublayer beneath the center of loading is then calculated using either the cross-anisotropic or the nonlinear pavement model. the principal stresses ci and 03 are orientated vertically and horizontally. For this location. Residual compaction stresses must be included in estimating the total stress state on the element.

.P 413 Figure 7...61° 3 I I 0. I Calculated 1 40 a o..10 psi -c!) .. •'"'"' .0 f 03= 3 psi ct) = 58 ° 0 = 5. 50). Idealization of Layered Pavement Structure for Calculating Rut Depth (After Barksdale. Comparison of Measured and Computed Permanent Deformation Response for a High Quality Crushed Stone Base: 100. . 45 P 5 a n6 A • P e5 fi 4 41E.70 K = 1880 c = 0 R = 1.8 E (Percent) Figure 8..WHEEL LOADING ri-rri-r1 SUBLAYER 1 I • 42 P .. 0 3 = 10 psi a =5 psi 3 ret 0 • 0 =3 psi 3 • 0 0 04 Permanent Strain.000 Load Repetitions... . Ref....' 1 n = 0. 47 .2 €3 3 AGGR EGA TE BAS E h2 • na °3 A ss • • 0.

The hyperbolic expression for the permanent axial strain for a given number of load repetitions is 48 . when the element is highly stressed can greatly reduce permanent deformations under certain conditions.03 = vertical deviator stress Although the changes in confining stress are relatively small. The sum of the permanent deformations in each sublayer gives an estimation of the level of rutting within the layers analyzed.03 and confining stress 03: E p = f(0 1 - (7 3'3 ) ) where: s p = vertical permanent strain which the element would undergo when subjected to the stress state 03 and 01 . The hyperbolic permanent strain model proposed by Barksdale [50] for permanent deformation estimation gives the required sensitivity to changes in both confining pressure and deviator stress. To predict accurately the effects of these small changes in stress on rutting the permanent strain E must be expressed as a continuous function of the deviator stress 01 .03 a l = major principal stress acting vertically on the specimen below the center of the load 03 = lateral confining pressure acting on the specimen below the center of the load a l .strain within each representative element by the corresponding sublayer thickness. Placement of even a stiff geosynthetic within the aggregate base causes small changes in confining pressure on the soil and also small vertical stress changes. these changes.

K.(a l . Two different quality crushed stone bases were modeled for use in the the sensitivity studies [50]: (1) an excellent crushed granite gneiss base having 3 percent fines and compacted to 100 percent of T-180 density and (2) a low quality soil-aggregate base consisting of 40 percent of a nonplastic.sin(p where: 0 and c = quasi angle of internal friction 0 and cohesion c determined from cyclic loading testing Rf. friable soil and 60 percent crushed stone compacted to 100 percent of T-180 density. A comparison of the stress-permanent strain response predicted by the hyperbclic relationship given by equation 8 and the actual measured response for the two bases and the subgrade are shown in Figures 8 through 10 for 49 • . n and Rf) used in the expression must be determined from at least three stress-permanent strain relationships obtained from at least nine cyclic load triaxial tests. 0.osp+sin0) 1 . The silty sand subgrade used in the comparative study was compacted to 90 percent of T-99 density. k and n = material constants determined from cyclic load testing All of the material constants (c. The subgrade had a liquid limit of 22 percent and a plasticity index of 6 percent.Ep = (0 1 - 03 )/K o-3 n 1 .o-3 ) . The resulting stresspermanent strain curves are then treated similarly to static stress-strain curves. R f 3 2(c. Three different confining pressures would be used in these tests. . The soil-aggregate blend was about three times more susceptible to rutting than the high quality crushed stone base.

Comparison of Measured and Computed Permanent Deformation Response for a Low Quality Soil-Aggregate Base: 100. 5 psi Calculated a 0 co a 6 6 2 6 20 GY 0 0 8 1 6 2.77 n = 0.000 Load Repetitions.8 Permanent Strain. 0 r-o Dev ia to r Stress.8° 40 R = 0.77 a3= 3.265 0 10 psi 3 .5° — 4) = 41. E p (Percent) Figure 10.4 Permanent Strain.5 psi ' 0 3= 10 psi 1 1 90% of T-99 density Wet of Optimum o. c (Percent) Figure 9.. K = 680 4) = 46."111111111111111111 111 0 0 1 6 0. 50 .0c f 03= 50. 0 =10psi 20 5psi =3psi .000 Load Repetitions.4 ° 4 o = 40 -53° 3 0.n = K = 1750 R = 1. Comparison of Measured and Computed Permanent Deformation Response for a Silty Sand Subgrade: 100.

Table 7 summarizes the general material properties of the base and subgrade.0 - - / 1. the soil was a gray. 2. The granite gneiss crushed stone had OZ passing the No. 3. 4.000 cyclic load applications. and A-4(1). plasticity index 6. 10 sieve. 200 sieve. 51 . Classification SM-141. The actual material parameters used in the hyperbolic model are given in the figures. A-2-4(0)]. Table 7 General Physical Characteristics of Good and Poor Bases and Subgrade Soil Used in the Rutting Study(l) GRADATION BASE DESCRIPTION 1 11 3/4 10 60 200 COMPACTION T-180 co )( Kt 'max (Pcf) (%) LA WEAR (%) s (3) 2 40-60 Soil/Crushed Granite Gneiss Blend( 2 ) 99 85 42 25 13 138 5. Data from Barksdale (50]. 22%.2 50 47 1 Slightly Clayey Silty Sand (4) 100 100 100 63 40 115. Degree saturation in percent as tested.100. 40 and 20% < No.4 13. The theoretical curve given by the hyperbolic model is seen to agree quite nicely with the actual material response. nonplastic with 73% < No. liquid limit.5 73 45 6 Crushed Granite Gneiss 100 60 25 9 3 137 4. silty fine sand (SM.

Pavement Geometries. and vertical strain on top of the subgrade. cross anisotropic and nonlinear finite element sensitivity analyses were performed during the study. and subgrade stiffnesses. pavement geometries. 52 . Aggregate tease quality was also investigated. The basic pavement condition investigated (Figure 11a) consisted of light to moderate strength pavements resting on a subgrade have stiffnesses varying from 2000 to 12. Both linear. and the potential beneficial effect of prestressing the aggregate base and subgrade using a geosynthetic (Figure 11c). Pavement geometries and subgrade stiffnesses used in the primary sensitivity investigations are summarized in Figure 11. The potential beneficial effects of geosynthetic reinforcement are also more clearly quantified in terms of the reduction in aggregate base thickness and the relative tendency to undergo rutting in both the base and the upper portion of the subgrade. The general effect upon response of placing a geosynthetic within the aggregate layer is demonstrated including its influence on vertical and lateral stresses.500 psi (14-86 MN/m 2 ). The effect of prestressing the aggregate base is also considered for geosynthetic pretensioning load positions at the middle and bottom of the aggregate layer. Sensitivity studies were also conducted to determine the effect of geosynthetic position (Figure 11b). tensile strain in the bottom of the asphalt layer. the geosynthetic was located in the bottom of the base.ANALYTICAL SENSITIVITY STUDY RESULTS Sensitivity Study Parameters The results of the analytical sensitivity study are summarized in this section including predicted response for a range of geosynthetic stiffnesses. Other supplementary sensitivity studies were performed to evaluate various effects including slip at the geosynthetic interfaces.

. 3.5 s CBR 2r3 10 Note: 1. (c) Prestress (b) S gPosition (a) S at Bottom 2. hY.9 = 0.. . 53 .0 12. E /E b s • 2..9 9. 3.7 7. p =120 psi 16.4 12.5 9.0 15..5 1 E • 250...E1W2We /ffiin44!ii14/ E Es 3' 5 CBR 3 CBR s = 3... ..5 PRESTRESS @ BOTTOM .4 Figure 11. %) .• .. '• 12 15 .7 7.5 E • 3.P =8.I.5 sL 201b/in.7 9...000 lbs.5 6. Initial [Prestress 45 in.000 A 2.. Units of inches and kips unless shown / /Al .5 16. Resilient Moduli and Thicknesses Used in Primary Sensitivity Studies.3 12. Pavement Geometries. 3 .5 12 PRESTRESS @ CENTER ..0 9.6 7. . .L..5 6 3 5 12.

their response was taken to be linear. Equivalent AASHTO Design Sections. A CBR value of 10 was considered to be a realistic upper bound on the strength of subgrade that might possibly be suitable for geosynthetic reinforcement. (7. 5 and 10 were selected for use. Geosynthetic Stiffness. 41 MN/m 2 ). except in a limited sensitivity study to investigate its effect upon reinforcement behavior. 86 kN/m 2 ) were selected from Figure 12 for use in the cross-anisotropic sensitivity 54 . 18 kips (80 kN) single axle loadings (ESAL's). 4000 and 6000 lbs/in. Therefore. 41.000 equivalent. 6 and 12. Three levels of geosynthetic reinforcement stiffness Sg were used in the sensitivity study.000 and 2.35. 500. Subgrade support values and other constants used in the 1972 AASHTO design method are given in Table 8. Sg = 1000. Preliminary analyses indicated that the geosynthetic reinforcement of heavy sections (or lighter sections on very good subgrades) would probably have relatively small beneficial effects.5 ksi (24. structural pavement sections were selected for use in the study having light to moderate load carrying capacity.5. and the value of Poisson's ratio of the geosynthetic. Poisson's ratio was assumed to be 0.slack in the geosynthetic. Subgrades having CBR values of 3.000. 28. To reduce the number of computer runs to a manageable level. Since small values of stress and strain were found to develop in the geosynthetic. Selected pavement thickness designs are shown in Table 8 for 200.000. Average subgrade resilient moduli of 3. all three levels of geosynthetic stiffness were only used in selected studies. The equivalent axle loads which these sections can withstand serve as a convenient reference for acccessing the strength of the sections used in the sensitivity study.

9 5. AC B RESILIENTSUGRADEMODULUS.35 TAC T AC > 3.50E4 1.5 in.85 2.2 3.44 a 2 . T (in.00 9.00 6.5 6.Table 8 AASHTO Design for Pavement Sections Used in Sensitivity Study ■ SECTION TRAFFIC LOADING(2) (x10 3 ) (Z) (ksi) STRUCT.5 2. Es (PSI) 2.0 12.Ts (in. 1 SUBGRADE CBR Es 3. S 3.5 2.5 6.00 SUBGRADE CBR (PERCENT) Figure 12.2 4.00 0. 55 12 00 . T + < 12 in.9 9. for T in excess of 3. single axle loadings.9 5.6 6. BASE SOIL SUPPORT.7 7.5 2.) AGG.5 15.45 2.0. elu-.5 in.5 3.00 6000.00 3000.3 120 8 9.5 Regional Factor .20E4 9000. NO.55 THICKNESS. Typical Variations of Resilient Moduli with CBR. AC TB T + T > 12 in.5 2. Equivalent 18 kip.5 4 5 6 500 500 500 3 5 10 3. 1.0 12.18 a 4 = 0.5 11. Design Assumptions: Present Serviceability Index = 2.00 0.00 3.4 2 1 200 200 200 .0 7 2000 3 3.14 0.5 Asphalt Surfacing: al Aggregate Base: a3 = 0.2.2 3.5 3.) a 1.5 2.1.5 12. (SN) SURFACE THICKNESS.0 2.

Therefore. Almost all presently used mechanistic design procedures are based upon (1) limiting the tensile strain in the bottom of the asphalt concrete surfacing as a means of controlling fatigue and (2) limiting the vertical compressure strain at the top of the subgrade to control subgrade rutting [51. the effect of geosynthetic reinforcement on permanent deformation in the aggregate base and upper part of the subgrade was independently considered using the previously discussed layer strain approach and hyperbolic permanent strain model. Limiting the vertical compressure strain on the subgrade is an indirect method for controlling permanent deformation of only the subgrade. Cross-Anisotropic Sensitivity Study Results Geosynthetic at Bottom of Aggregate Layer. Structural pavement sections for the primary sensitivity study were analyzed using the previously discussed 56 . These results are presented in Chapter III. The beneficial effect was accounted for by establishing the reduction in base thickness due to reinforcement. the procedure followed was to determine for a reinforced section the required aggregate base thickness that gives the same critical tensile and compressive strains as calculated for similar sections without reinforcement.studies to characterize subgrades having CBR values of 3. An important objective of the sensitivity study was to establish pavement sections reinforced with a geosynthetic that structurally have the same strength as similar non-reinforced sections. Equivalent pavement sections with and without reinforcement are hence identical except for the thickness of the aggregate base. 5 and 10. In keeping with these accepted design concepts.52]. respectively. Separate reductions in base thickness are presented based on equal resistance to fatigue and rutting as defined by this method.

These variables include tensile strain in the bottom of the asphalt.7 in. (1.3 in (200-400 mm). (18 mm) horizontally outward from the center of the load. 57 . Tables 9 through 12 give a detailed summary of the effect of reinforcement on the stress. The force developed in the geosynthetic reinforcement is also shown.5 ksi (24-86 MN/m 2 ). to a maximum of about 18 lbs/in. All response variables given in the table are those calculated by the finite element model 0. each table is separated into two parts. These sections had an asphalt surface thickness of 2. strain and deflection response of each pavement layer. varying from less than 1 lb/in. Downward deflections are negative. The force mobilized in the geosynthetic is also small. given on successive pages. Because of the large quantity of information given for the sections.5 to 12.2-22 N/m) depending upon the structural section and subgrade strength. The pavement response under the exact center of the loading can not easily be determined using a finite element representation. The percent difference is also given between the particular response variable for a reinforced section compared to the corresponding non-reinforced section.cross-anisotropic finite element model.5 in to 15. In these tables a positive stress or strain indicates tension. An examination of the results given in Tables 9 through 12 show that the effect of the geosynthetic reinforcement is in general relatively small in terms of the percent change it causes in the response variables usually considered to be of most importance. subgrade resilient moduli were varied from 3. and a negative value compression. and vertical deflections. Also refer to the notes given at the bottom of the table for other appropriate comments concerning this data. vertical subgrade stress and strain.5 in (64 mm) and aggregate base thicknesses varying from 7.

0 U) 0 a.4 0 4-1 4.0 0 = en C N N it N 0. O O Cl 11. RTI CAL ST 0 In II 0 O 0.1 U O 0 T 0 CJ C.? 0 a 0 C C . en O 1 4-1 0 1-1 C • "0 O 0 ^0 0 0 0 e . cr) a) En 0 O O 0 a.4 O C) C ■•■••■ ro 0 CJ 0 IZ4 ro A 0 CD g 0' 0. cr.1 0 O rl • O S.4 U 0 0 0 1. •••1.1 •-1 4-1 4-1 CD U r 4-1 0 0 • -1 4-i 9-1 "0 0 • -1 Cr) C CD (1. ria • 0 0 O 41 • 9"1 (=1 O • C.1 04 $4 CU VI 4-4 O 0. 0. O VER T. SUR FACE DEFLEC TION 4-4 O O O O O O J1 O O O O 0 0 O a • 58 0 a. C Lei cO O 0 C (1) O 0.4 0 $.1 '711 0-) 01 S-1 S-I 0 a-t 4-1 sC O 0 en t•CI .• Z 7.1 Cl 0 0 0 4.1 OD 0 O H $.0 0 ✓ y O 9-1 C/1 C•1 S.11 Q Z E fa.1 O 0 rl Z 0 IN 10 9-1 •.1 0.1 Cl •• VI 11. C. 0 cu U • Q) a) .1 ▪ • ten en N 1.1 C) 4-1 0 .I CL1 0 U- a. O •-■ II Qi 0.

AC / 9. 01- 19E' 0I NC O '71 0 rs c-.J iG0}r14lk. 5IN. 0 1 1 C VE'IL TERI - 916I-1 8 4 LI6 - ZOI' 0 E RS' 0 9' 118 /' SI T 1 .777711 L8 9' 91. CA ta . VERTI CAL STRAI N 1 1-71 Z• ST S' £ h' SZ6 L 'SlSC OT I C. 0- 9' 1 01 1I '7 1078' 9. CE- 6S1Z - BUZ - ZISZ .1 01 r-.I 8/9 ..—. IL- 7' 91 91 1/9• 0 Z 9'7'7' 8 - 1zcu e- 1 /91 .I 9 .1IT 9T - 3 AGGR EGATE BASE S UB( 9"76- LOO' S - 1 L6LI- T8LT- 1-7771 L6LIE9LI0.. RADI AL STRAI N SUBGRADE RAD IAL S TRES S 7S 9' 0 '3JTU % ( Con t inue d) 6' Si 9. 011 OSZ' O 8. I 1 I I . 0 I —.7 7 AU( % 0006 1 I 0009 1 °°gT 59 0 1 0 I 0 .Z 9' 0 81 .I 1891 I 0' 0E0' 9- I C BI- ZT 8T IZIZ - '710Z - 9807 I Lt.II I 187 I 9ZI' 0- T19 .0 6.. I I 6 .—r . 0AGGREGAT EBA SES UBGRADE E s = 3 500 PS I I I' 61- I I I I I mMM 109 . 6 ..7C• 0 9791 8 8 .c z• z L.1 6. 1 I FT "v s- LOZ' S9 05 .1 -1 I 8' TTI .S90' 0 177 £E 6' V- 80 r7. 11 I 1os oz. 6- 6. 0 (Tsd ) l o I EtS . I VERTICAL STRAINI VERT ICAL STRES S I 2.0 .. '71 - T9 . I 6 91' 0 gm■■ IZ S *0 1 Z' ZZ 0 . 9Z - 1 I 0 . 0 %Di ff. 00 ch . 9T - 999i- IMM1 I I 1 I VERT ICAL STR ES SI[ RAD IAL STRE SSI V EZ I E . 0- IEZ 9 . AC/I 2.I 0' 97 . 0 I I• Z 69 81 9S6T S' S 1ZS 91 91 009 . 8 - (9- BO TTOM 1 /3 OFAGGREGATE BA S E 07'71 1 C 'ZZ - 777777 E 80 . 77. 0 2. 0 I LII 0 I. ° 6I ZZ RADIALSTRA IN 9 . 9 - 17777771 I IhTZ - 0I) 1 3 1..1 SD 6.771 I OLEI I6ZT Z' ST- 91ST 6SET - ELET- Er'? = I h' CZ- 8. ° 0006 I - 0051 I 00 06I I [_OOS T 10009 1 I 8 75 ' 6 - VI ZI- 0009I 1 97 7T- 1119 665 .1 1 7771 - ■ (Tsd ) Z 0' 76' 71 - '78 .1—(71 009Z - 79707177771 I r--77777 777TF77777g1 77. '1 11 1I z• sIE' LZ I EBOZ Z0r 0 8891 - 1791 - 69. 7 2 AGGREGATE BA S ESUBC % D i ff. 9 Z19 . 5IN.

C C) C 1 C) O . E ffe c t o fGeo syn the t ic Re info rcemen I. ti C C.t on Pave me n t Resp on s e : 6. 5in.) C O P.

0 I I 6.. 0- 1 hI L' O - I I cl 6. 0 S 60 . 0363 60 .0 -1 £50 -1 - I (Ts 0 10 I 8' 1 81 CET 9 . 09. 961 I 9. /81 9' 661 -i 0' 001 SUBCRADE RADIALSTRESS S' 66 IC0' 0CSO' Co - 'JITU % Ef fec t o f Geo syn the t ic Re in for c emen t on Pavemen t Resp ons e : AD IALSTRAIN I FIT rg CI S Z' f1 [ 1 • 18 [FIT Z 9' SS6 0' 16 [77 S . AGGRECAT 9• 6I I 8' L 91 I I L.1 00S T 8T Z' I - I 0006I 0009 I 00g.9 8• ST I 9 . 0 IN.13L R I 190 190 -1- .. 7767 VE RTICAL STRAIN I 1-1716' 16 1 - 1 3SV9311/931199V30I /IW011011 I I 00 06 I 0009I 0051 . 91 L' Lll 1 . 0- 198 ..0 0 H00 ul .8 I9 ..1 £ 6. VI 1..-1- LOS • I T I (I S' I- IZS• I- 11S -1- (Tsd ) Z o VERTICALSTRES SI 6 .. AGGREGAT E RADIAL STRAINI VERTI CAL STRAIN c r (1 0-6 ) 118 . 91 I 9' 51 U0' 0 1 610 . 5 IN. AC /8.. AC/ 1 6. S9I 0. 991CTI L 9' 901 9• ZIT (900' 3 ( Con t inued) I 9' 0T •7.1 0' 0- 100 . 0 8E8' 0 - 6' 511 - 9 . 61 1) IC 2 Z • >•• 1.) V/ 1 1 . 01 6.1 I I I '7 • L 0' 91 1. 811 . I C11' 1 .I 9• 9 . ACI 1 2. 5 IN. 01Z - I S• II O' hE l' IJTO 1 1771 76717777177177-7 777- L' h 6 FEU - 19 .. 8hT - wx O' Z 669' 0- 8' 0 9• ZI 9 10' 0 I' 69I - 1 -11 h• T 1 . ° RADIALSTRESS 6' 81 S' OL 0' 006 O' SIT I 1. . 4 2 I N. 0 I 1 1' 9 '7 0 S' I 9I I TTL' O .0 141 F0 101 . 91I 6• 1 11 90/ .• C.I I- v(z . 0 I 6. 5 IN. 1 ..3J -Ta% 690 . 0£18' 0- 1I '7 . 03 92 0' 0 509 LS0 . ° '0 FT L' 0 I FO I I " CCU- 0. 0- VC O' i- 77T I I I 0-1 SIO' 0 1Th0 .. 901 - URI I 6-£ Z . 11 9910' 0 96 10' 0 I 9 . 8ST I L9 . 91 Z- I 9' 99 I 1 1 600' 0 - VERTI CAL STR ES S 1 I9TF O - '7 .. 0TN. 1I1 .1 0006 I ISO' I-4 0009I 1 090 -1.295 0.I I' IST- Z• ET ri3ST L• SEI Z' E LI- 1' 9 T' ZLT - 1. 0- 8• C ro COS I - 1• 0 .1 S .E' 01 1 C10• 0 Z' T 8 . 0810.::::::J 1 I IC 0 . 01L 0 . EI Z 0 . AGGR EGATE BASESU BGRADE E s 3 500PS I 47-E '70 Z• Z FI /I - I 0 . /81 I 5 .

O O O I a. 73 9000 . 04 3. 51+1. 5 I N. 3 . 8 6000 .1 . 9 22.4 0 4-+ 0 a) In O a) 14-1 0 14-t CU C34 6 0.-■ N . 1 .t •. 83 - 0 . AC /1 2.1. 0 I O O 62 O O N -.-■ ..-I N . 0 10 2.0. 11 202 - O O 0 N. AGGRE GATE BAS E Ca ( ps i) I % Di ff.0.) a) a) Ci z 0 6 z (in ) I%Di ff.1 00 00 to ‘D. 3. 1919 6 . 0 1 0 on 0 en o V1 CO 0.0. AGGRE GATEBA S E -6 00 0P SI 03 O 0 O O O 00 -5 in r. 59 31 . -.0. 46 .6.6. D n-I 0 CI. . 461 . 5I N. rl . 1.. O O N O 03 N CI. 5 .2..0. 9 .7 O en Cf.0 -2 I . tl ro In 1:4 •. ) a) 1/3OFAGG REGATEBASE 0 SUBGRAD E VERTICAL S URFACE DEF LECTI ON V VERTICAL ERTICALS TRE SS ON V ON EFL ECTI VEE RT. .2 438 . 09 4 .1.1 6. AGGR EGATEBASE s O 0' ro 1 I a.-.4 0 4-I - en 9 36 G - 11.7 0 N 0.1 O -1 0. 86 . 1 2 1 6.1.24 40 . 0 on a.0.1 6.0. 0 .1 I a. 5 cn 00 .1.0.4 0. 0.4 7.-1 .46.0 . AC / 7.243 91 3 . 5 5. 03 52 .4 oo ..1 5.1 f.1 5. 0. 8 44 11 7 7 . 28 8 +1. 5 . 04 .4 7. O 0 re.-I . 05 2 7 . z %Di ff.2439 en . CO CO O N 1 -1 0 0 0 0 0 on 0 di f fe ren ce SU B GRAD EE 11.5. 8 5 IN.7 0 . CO -4.47. O • 0 0 I O 0 .1 en N 1 NI 1 . 45 3.2 0.0. 03 63- VERT I CALST RA I N o . 66 9 ./3 en rn 0 rn C. N I 0 N -1 0 to 1. -4.0. 8 . 511 63 ..0. 453. 7 N 0 on N 0. AC /9. Res i lien t O SUBGRADE E s -600 0 P SI N IN.1 03 N ti O 0 O CO a.1a1.3.2.0.7 I 1 0 0. 04 +0. . 0 ON N O en C. C.3. 9 . 0 .0. 0 RAD IALSTR E S S O 6( in. 71 vM . O 0 I 0 N 0..1.11 1 O en en 0 0 O O O 0 0 1 0 2. 5 IN.4.2 8. 75 I N. rn -7 N I 15 00 . 05 29- -6000 P S I RAD IALS TRA IN I 0 a) s en SUBGRADEE cn 2.. 0521 . 2 .2. 9 09- 1 213. O Os Vi tT 0 N 0 OD 03 c0 O 0 O 0 O et. 6. 03 55 .0.ce -7 i -2 0 0 en 0 In ro CI.00 --••••■■■•••110.-1 0 es . uy -1 N a. 4. 0 51 7 . 795 . 03 61 0.

"- uL I 06 100 LL 00gI I 00091 I I Z• TL 7 [ 9 rmm777-77 OgLI- IZLI9' 0- -I I9LI BO TTOM 1/ 3 AGGREGATE BA SE 8' S T- 8 .1 -1(1% . AC / 9.E tg• O I OLZI I611T I I 7 15 1 14 1 - I RADIAL STRESS IN.I I TS . 0 VERTI CAL STRESS Z' O L - 1 I 76 ET - S OLI - 1.0 7I- LEW7I B' ZI791 . 7I 1 .I078' 0 I 6. 11 9 I EC OT - E- I I09gI 819 .I187' 0 I 0' EI8 . I 1 s =w 6000 P SI I % D UI.IZISI 9. • 2 --• 0 f -.mmmmmm 06' 91 - 60 .0 ZI 99' 0 g 01• 0 915 . I Z 'Z 1 I 7' 0- I I LZ. 6gg EE6 .• ("1 41 (fl U I I I I I I 1 1 I I I . 01- 0006 1 000 9 1 OOS I I 0006 1 I 00091 I 00g TI I OS' ZI .mmm. 2. 1I- I Tg 9I- VERTICAL STRA IN 1 T• I - R • g19 L7Z• 0 6• Z - 1• 6E- 7 70Z E• Oge 7-077177771 I I 6'LI I I I— 0* Z- 86 I• L- l• Z . 5 IN. 0- 6' 6- /* SU- 6E61 - [5 0 . VERT ICAL STRA IN AGGREGATE BASE RAD IAL STRAIN 76r° Ill"?- EL R' 9- vei- 1691 891 .I'VS67 IO' IS . ( Con t inue d) 1 O' I T - Z • Z- I SL' 7- 1E 1' 0 I7' 709 I1' 96 . 75 IN. AGGREGATE BASE SURFACE E 2.IIOg' 0 EIZZT 1 I' S . 0 88 5' 01 0I EI L .I I I SI . E1- 1 97' 0 709' 0 I 9g .777 . ° I 7 . 189 131 7T- 881• 0 7g7 LS' IT - 90 7' 0 067' 6 - ZgE . 0 7 .IIECC O 6' 6 9 - 9 . I 6' 7 - 7' 0 - g' l- 96' ZI - OT' ET- 1 9 . %I 0 0 I 1 HIM 2 I I 1 I LROZ - t it: I 1 • •. "- g6E• 6- L 7S I BT' ZI - 777] Ef fec t o fGeosy n t he t ic Re in fo rc emen t o n Paveme n t L 7119811 - REIZT- I6S ZT - 6 E0' 6 - I Table 11.1I- 1- 91 .7 - I0. 5 AC/ 7. 0- I L' 60L - RB9I - L' 98T - 1• 711 9L' II- 9• EI- E 891- rE - 6' 5 91 57I . 01- g• / ZIT I O' LT. 0I. 6 - 61 71- I E./ - O' I- IE' 0 1- 11• I 5 . 9 - I Z" I 7g I I S' TI . 5 g' IE- 1 BE 7T 7 . 01 - I 9' I - IO' IZ- m.E Z' 7-7 9L OT E' II- CSTE• 1- gROI- TOIT LOTI 9101 Grow / IN.LI- . VERTICAL STRESS 1 RADIAL STRAIN SUBGRADE RADIAL S TR ESS S URFA C E E s =600 0 PSI 1 % D1ff. 1386' 0 E -E - Immm77777 67ET 1.1 7' 0- E. EI- 61' 91- mm.1 99' 9 - F. 7T.

0 U a) • ••• W "LI (1) ..71 CO ■. ❑ Na F v N i I. Na ▪ -a 1-) Ls F 4. Na LL. O O O 0 0 0 0 0 771 C1..7 . Na 0 Ca 0 VERTI CAL.. STRA I N 0 r-7.4 0 .1 OE 02 41 10 .: . 10 CRECATE ni Res i lien t RADIALSTR ESS fr 0 4-i 0 0 C.0 7* 0. Na Na 0 •-■ U U z 8 01 24 ccC •-.) 0 W 2 OD 11 1-1 • I • N. E CC 14 > '-' 1■1 0 U IA O.1 cc o N H .4 7.. Z 6. 0. C.4 ro •-I • 0 • ro Na -c) o -c) L0 Na O )4-i 4-1 Cll E. U 64 0 s.0 rn 4.24 Z U II 24 1/3 c.

.0 1 1 L -.I6I' E - : 57 L' 1 91 6 .I 0' 9 1' 01 1- - I Z • 1VE Z1 - [ I I' I - VERT ICAL ST RAI N I . 19' 991 p]ITO % 65 6' 61E . 9Z1 - '1 . 1 1 1 I 'MU Z __ . 890' 0- 777 00E' S - Z' OT1- 9 7771 977 1 1 Z . 9 - LE' S 60 . 9" IE 6.' /IWO. 0 '0 011 RADIAL STRESS I O' EI691' 0 rh 7 i 09E' V 0 11' 1 - 911' 5 - I T' 6 01. AGGREGATEBASE SUBGRAD E E 9 .77FT71r70717. 1- V 15E- I 9' 10 1- 1 I 1 . ° Z/E' 0 9 E• 9 I T5' 9 - 000 9 0006 1 1 ES *9 I I I.°'0. L6E 77677 1 i 6 . 0 .1TO % 30V9090S . .I I Z' 0- I L I 1 0 I VI- E' RI - 9 1E • 0 I 5ZE . ___.. I1I 8 5I Er 1c I h' 191' 0- 181' S - 9' 1 'I iI 0Z Table 1 2..—. AC/ 6. 5 IN. 1- [-TT 5' 1 1E- EC E01 - I RADIAL STRAIN BOTTOM1/3OF AGGREGATEB ASE 1 s = 1 2. 5 00 PSI 2. C0 i 1 [. . s. 0 81 0' 0 L' 6 1111 . 91 1 - 1 9 . 5T- I 005 1 -1 9I' OT- o (`) 61V E - I C. 0 I 0 . 1- 5• 1 91 . 500 PSI 0. r I 1 I I fr IEE - I V . E9 1 - I-- 0 . 5 IN. 10E I I I 5 .11 5' SI - z 'E 691' 0 VOE' 0 TEE .F9' Z - 0' 15 1 !VERTICAL STRESS V ERTICAL STRAIN 'VERTICAL STRES S [RADIAL STRESS E' ZIT I I0I' 0. AC/ 7.n O F. T .0 (con t inue d) 9 I I•9I L' T I E . 99E LAT E BASE 2. I1I VLOT 6' L - 1 _0' 1 91 - 6 . 1I811 . 0 61E• 0 9' 0- 6' 0- 'OZ - 1L£ .0.I__ 6'0 I I I SZZ* 0 - IE' - . I' E T- E 9E .8- L I I 6. 0 IN. 5IN. . 8- I CTI- 71 7I 767.7 .7 . E f fec t o f Geosy n t he t ic Re in forcemen t on Pavemen t r". 0 E 0. 56Z . 1 00091 00061 0 OOST 0009 1 0006 95' 9 . E • T_ S ._ V6EI E .77-EI7Z7* L' 99Z - V 9EE CZEE - VI I 9' 01E - E' 6 - Z• 0- 9I'OI - c ( Ps 590 .. 0 EZE • E - S ZZ .. 10I 1 RAD IAL STRAI N E' EL Z- OZIZ 09/I - s oo' o+ 9' 91 lEs z' o-I CE E' 0 7.—• y . 5' 0i- 9 .3717 5' 95Z - 1' 1IT E 1•E I 08E - I I I T. 0 'fi- B'f- 96E . AGG R EGATE BASE SUBGRADE E = 1 2.1/1I TZ.. 0I - 0.I0 051 1 - 1S . REE- 5 . ". ZZE 5' 6 ZE S' OSE frE 9E . 1 .

tended to be in the wrong direction. 86 kN/m 2 ). even though quite small. The important variation in radial stress which can occur within the upper part of the subgrade is illustrated in Figure 13. Figures 14 through 16 summarizes the effect of geosynthetic reinforcement on the tensile strain in the bottom of the asphalt and the vertical compressive strain on top of the subgrade. The presence of the geosynthetic can have a small but potentially important beneficial effect upon the radial and tangential stresses and strains developed in the aggregate base and upper portion of the subgrade due to the externally applied loading. General Response. Equivalent structural sections can be readily estimated as shown in Figure 66 .5 in. and as the subgrade becomes softer.5 in. (165 mm) AC surfacing are very small initially. The radial stresses caused by loading in the heavier section having a 6. This is especially true considering the magnitude of the initial stress that would exist in the layer due to overburden and compaction effects. and permanent deflections increased. Further these changes in stress. The change in both radial stress and radial strain expressed as a percentage of that developed in a section without reinforcement is appreciable for all three sections shown including one with a 6. relatively small radial stresses occur regardless of the presence of geosynthetic reinforcement. Thus. Even when lighter sections are placed upon a good subgrade having a CBR of about 10 (Es = 12. the change in stress resulting from the geosynthetic has a negligible effect on performance. they tend to become less compressive due to reinforcement which means confinement perhaps would be reduced.The force developed in the geosynthetic increases as the thickness of the structural section decreases. That is.500 psi. (165 mm) thick asphalt surfacing.

30 0. Variation of Radial Stress at Top of Subgrade with Radial Distance from Centerline (Tension is Positive).0 2.30 00 2.00 —0.0 60 DISTANCE FROM CENTERLINE. R (INCHES) (a) Subgrade E s = 3500 psi RADIALSTR ESS.0 4.RADIAL STR ESS.500 psi Figure 13.0 6.15 00 4. ar ( PSI) 0.05 —0.90 0.60 0.0 80 DISTANCE FROM CENTERLINE. ar ( P S I) 0.05 —0.15 0. R (INCHES) (b) Subgrade E s =12. 67 .

.-4 CO Z cr LT4 o .4 0 I) Z 9'4 IC C.. .■ M $.I ..r) -a = 0 la! tJ 2 co CZ/ I- 0 0 VI .■ 0 0 1 ) A3 Radia lEr in AC (9 . N SA W 8 1--1 • Ci 4:4 cd 0 O O N N O N tp (g —0 I) 13 "31/ i101108 NMJJS 7V10118 O in ( 9 _01) /3 "ON N401108 68 O In CI) 1-1 •ri 4..• ▪ (b) Ve r t ica lEv on Subgra de 0 a W o ■-4 OD — N c.0 0 •..a •..-I RI IA- ---. '. 44 (AI N (9_01—) "3 Iardoens JO dal WAILS -mum loYsoans ao d01 WOOS -mum $.1— 4 .. CO Z ..) X U-I • .4 0 1.) o e4 tr.-t ca ). C.) 0 u -4 --4 VI .1 . 4..0-••. fe) ii a ttl u.4...

0 I—) A3 1avt19ens 40 d01 MILLS IVOU113A O 0 (MCI) j13 P') NIVILLS 7VICIV8 (MCI) 'ID i SS3NIS 69 0 O ismais 8 8 8 cq 0 O d E r on Su bg ra de Subg ra de E s =12. 5ks i Ra dia l Stra in in Bo ttom o f Agg reg a te Base ( Tens ion is 8 Var ia t ion 0 DISTANCE FROMCENTERLINE..THICKN ESS OF BASE. T (INCHES) ..... R (INCHES) 0 . 0 0 ri 8 (9 .... 0 0 N (9-00 13 o 0 "o-' muse I .

Consider a section with a 2. The reductions in required base thickness are even smaller based on constant tensile in the bottom of the asphalt surfacing. To develop a set of design curves for the three levels of geosynthetic stiffnesses requires a total of twelve finite element computer analyses. The theoretical effect of reinforcement position on the major response variables is summarized in Table 13 for the three levels of geosynthetic 70 . An exception is the section having the stiff subgrade where the difference was much less.(4. (64 mm) thick asphalt surfacing. The reduction in base thickness for constant vertical compressive subgrade strain decreases from about 12 to 5 percent as the subgrade stiffness increases from 3500 to 12. but the stresses were very small.75 in. Figure 17 shows for the same sections as compared in Figure 14 the reduction in radial stress caused in the bottom of the aggregate base due to reinforcement. Geosynthetic Position. The pavement response was also determined for geosynthetic reinforcement locations at the lower 1/3 and upper 2/3 positions withili the aggregate base in addition to the bottom of the base.9 kN/m). The actual magnitude of the change in radial stress in the bottom of the aggregate base is about 10 to 20 percent of that occurring in the subgrade.500 psi (24-86 MN/m 2 ).5 in. (250 mm) aggregate base that is reinforced with a geosynthetic having a stiffness of 4000 lbs/in. The results summarized in Tables 11 and 12 indicate that the beneficial effects of geosynthetic reinforcement decrease relatively rapidly as the stiffness of the subgrade increases from 3500 to 12. and a 9.500 psi (24-86 MN/m 2 ).14 by selecting a reduced aggregate base thickness for a reinforced section that has the same level of strain as in the corresponding unreinforced section.

. I 0 C 0 C -37. 89' 7- I P7'7 . 7 1 1 P.r1 ro 89' 6 091 1 67' 7 I t'?1 I I WIT 011 1 I 1 C aJ aJ 41 4-1 (13 (1) 107' 8- 1 '0 I 10 I 0' 0 I 6.O' h OR'71 I SE' l67'71 091 E- I •E- 00 7 101- I I 1 09 11 -I 89'71 2771 1 0' 0 - h' 7- I9I E - I I 90' 0- 191 E- 71 a) 1' 0- I i ■ C U 0 • et I I 90' 0- ■ (1) 0 4a r ■ I I 6[ 7' 11 9/ h' E I I I 1 O C CIO C 0 U w I I I I I I I I I I 0 P•I C 7 U • <13 -: I I I I I I I — Sa C I I — $4 • 00 XI 0 • •••i I I _r• Cr.° I 6' 1 6. 1 Ior T .1 R .9' 9 1 - I I 1 'I 6' C O* 0 4-1 . 7 I c. SEIE- I 0 971 c si - 8 • 61 8' 6 7 Z•6 -7 "0 I 711' 1 CNJ ^0 )1 I 1h1C- Rhlf6cI E - U a) I I 1 A I • et (10 -6 R• 0 Sf• 017- 576. (0.07 6' 117/ .1.777''' - Conven t io n: Tens ion I 1SEI Z hh I 06'71 89 E1 71 hi Q1 _1 r-i 4•1 fP 7 10 'n - 1 c81 1 0' n. AC/1 2. 017'70' 0-1 001'70' 0 ] 6 07'. 07 0' 1 6F1 I' I . AGGREGATE BAS E SUBG RADEsE= 35 I I 77' 9 78' h - 1 1CS' II % Dill.. Xi I I TENSILE STRAIN BUTTON OF AC •• . 7 6 '0 1 861 . N Y C/ 177: 9' 0- I 9' 1I 1 Sh' 9I8'7' 9E- 17* (169' 9E- NI 0111 0911 0511 071 1 01 1 1 0 91 1 Ohl!.) •1 6. 7 9' 7- I U 0S1 1 # 11' 71/:1 Res i lien t Mo du lus o f hOFI 1' 1 7- I E.r) (1.r. -1 8' 1 /F1 /1C 70 • I f ' 1 c/1 !n' 1. 0 ' - . 0 IN.0 aJ 0 GEOSYN THEFIC 1/ 3 UP 2. 5 IN.71I I 6 c8" 0 1 • 4-3 a) I 9' 1 - 10' 1Eh6' 9E- . /6- 099- I 8C0' 0 S1h• O 1 IhE' T 11 1 .

1. ..11■1.1 AG GR EGATEBA S • o ••O o O en S CO O O O (al fg r-. a) 0 0.4 CJ NI 0. o 0.1 0 r en 0 CO (. 11. O O O en O ti •il r O Lr. • TIC Z • co ow ■—■ F 00 o) 0 O 0 L•-• 0 0 0 0 0 0 cY. CJ (-7 ti 4.4 (6. 0 (.■ E".

The geosynthetics used in the analysis having the 0. (20. slack levels of 0. Slack levels of 0. 1.4 percent correspond to about 0. three different levels of slack in the geosynthetic were analyzed using the nonlinear finite element model.75 and 1. This type geosynthetic load-strain behavior was modeled using a smoothly varying interpolation function as shown in Figure 20 for the 0. The influence of reinforcement position on horizontal tensile strain in the bottom of the asphalt and vertical compressive strain on top of the subgrade is given in Figures 18 and 19 for the 1/3 up from the bottom of the aggregate base position and the 2/3 position.75 percent levels of slack would have a constant stiffness Sg = 6000 lbs/in (88 kN/m) after all slack is removed. The actual displacement that would exist in the geosynthetic as a result of slack is equal to the width of the geosynthetic times the level of slack expressed in decimal form. 0. 2. 28.75 percent slack level.7 m). A geosynthetic stiffness of 9000 lbs/in (130 kN/m) was used with the 0. with the rate at which it is picked up increasing with the applied strain level. To determine the effect on performance. 0. The higher stiffness geosynthetic was employed because it would be more likely to pick up load 73 .25 and 0.2 and 4 in. As wheel load is applied in the field. for a width of 24 ft. The effect of position was only studied for sections having a subgrade stiffness E s = 3500 psi (24 MN/m 2 ).25.8. Slack.4.75 percent slack level. (3.75 and 1. The force would be expected to go on the geosynthetic slowly at first.1 and 2 in. (7 m) the corresponding amounts of slack are or 0.25. the geosynthetic would gradually start to deform and begin picking up some of this load.stiffness used in the study. Hence. 50 mm) for a geosynthetic width of 12 ft.4 percent strain were chosen for the analysis.

_0 L) 13 '•o v 1101108 NIYLLLS "MINN 74 _0L) 3 mime NPALLS 1VIGV)1 . nimus •cum $.4 0 (4-1 0 § N (9_01 —) AO ' 3CM1J8/7 40 dOL NMILS 1V311Al3A P=1 a) cr) r-I Z0 f cti 0 0 0 74 I0 ••• •••• (9 (9.oo (b 0 8 O 0 0 csi N 0I0 N (2_01 —) Al 0 0 0 0 o° Iaveraens AO dOl.

00 30.00 12.00 3.00 13.00 10.00 Figure 20.00 1600 THICKNESS OF BASE. e (10 -3) 6. 75 . crr (PSI) 0. RADIALSTRESS. Variation of Radial Stress o With Poisson's Ratio (Tension r is Positive).GEOSYNTHETIC FORCE.00 0.00 GEOSYNTHETIC STRAIN.00 0.150 7.300 0.150 0. T (INCHES) Figure 21.00 9.00 10.00 20.000 —0. Geosynthetic Slack Force-Strain Relations Used in Nonlinear Model. F (LBS/IN) 40.

the reductions in tensile strain in the asphalt surfacing and vertical subgrade strain were less than 0. (64 mm) of asphalt surfacing.0 lbs/in.25 percent slack in the geosynthetic results in at most about 20 percent of the stress that would be developed in an initially tight geosynthetic. The resulting radial stress in the top of the 76 . The light pavement sections used consisted of a 2. for v = 0. (64 mm) thick asphalt surfacing. and the subgrade was represented by the bilinear model (Figure 6a).2 and 1 percent. A limited sensitivity study was therefore conducted for Poisson's ratios of v = 0.4 ksi (85 MN/m 2 ) and 3. The relative effects of slack were found to be similar for both subgrade stiffnesses.2 to 12. and a homogeneous subgrade with E s = 3500 psi (24 MN/m 2 ). For a Poisson's ratio variation from v = 0. as would be expected. The base was characterized using the good nonlinear simplified contour model material properties (Table 6).3 and 0. respectively.5 in. a cross-anisotropic base of variable thickness. or its effect on the response of a reinforced pavement.4. (88 kN/m).5 in.9 lbs/in. (250 mm) aggregate base and a subgrade having an average resilient modulus of E s = 12. 0.4. The geosynthetic force varied from 10.for this high level of slack than a geosynthetic having S g = 4000 lbs/in.4. The slack sensitivity study was performed for a light pavement section consisting of a 2. The literature was found to contain little information on the value of Poisson's ratio of geosynthetics. has a very significant effect on geosynthetic performance.2 to 0. (12 N/m) for v = 0.(7 kN/m). The results of the sensitivity study for the stronger subgrade aresummarized in Table 14. a 9. A geosynthetic was used having an actual stiffness of 6000 lbs/in.5 ksi (24 MN/m 2 ).75 in. Thus slack. an increase of 29 percent. Poisson's Ratio. A 0.2.

3 None 0. The numbers 2. rather than zero.25 0.) Total Rutting( 2 ) AC Rad'al e r Vert.72. (2) 3500 PSI SUBGRADE 15. (2) 2.5 -11 -22 -2. (7.) 12.8 -4. 8 3. Table 15 Effect of Base Quality on Geosynthetic Reinforcement Performance (1) REDUCTION IN RUTTING REDUCTION IN RASE THICKNESS BASE THICK.4 12.Table 14 Effect of Initial Slack on Geosynthetic Performance Design (3) E subg. AC SURFACING Poor Base Diff.6 - - 0 6000 6. Subg. (1) Good Base Diff.75 -11 -14 -15 -12 -19.0 Slack (1) SR 0(2) - 0.5/12. asphalt surfacing and a 9. 2.5 - (lbs71n.4 9000 8.5/15. Cross-anisotropic analysis.72 12. (2) Poor Base Diff.5 in. 2.75 1.5 IN. AC surfacing.0 -11 -12 -10 . The initial stiffness of each geosynthetic was assumed to be S R =300 lbs/in.5 ksi subgrade. 3. aggregate base. Modular ratio E b /E .8 -39 3-7 -10 Note: 1.5 in. indicate a 2.72 in. il .4 6000 10.45.4 1.3 -11 -12 -8 -6.4 Notes: 1.3 2.9 0.9 0 (2) 9000 13. this strain level corresponds to the slack indicated.3 0.1 -30 -2. Stiffness (avg) (ksi) 2.3 1.5/9. 77 1. (2) Poor Base Diff.4 2. Reduction in permanent deformation of the aggregate base and suberade.) Poor Base Diff.6 -6 9. T (in. The stiffnesses shown are ta limiting stiffnesses at the strain level where all the slack has been taken out.34 - 0 (2) 9000 10. (1) Good Base Base Rutting Diff. c v Good Base Diff. 2. Zero stress is inferred from the results obtained from the results for S 9000 lbs/in. for example.3 - - 0 6000 8.0 -4 12.5/9. (2) Good Base Diff.

5 in. Base Quality. which used a modular ratio of 1. A sensitivity study was then performed to determine the effect of aggregate base quality on reinforcement performance. a geosynthetic reinforced higher quality base would require about 10 to 25 percent less base thickness than a low quality base having a lower modular ratio. 78 . a low quality base reinforced with a geosynthetic would permit. Based on vertical strain on the subgrade. indicated that for the structural sections analyzed. A nonlinear finite element analysis indicated that a low quality base has a modular ratio between the aggregate base (Eb) and the subgrade (E s ) of about Eb/E s = 1 to 1. (250-400 mm) and a geosynthetic stiffness of 4000 lbs/in.subgrade as a function of Poisson's ratio of the geosynthetic is shown in Figure 21. compared to higher quality reinforced bases. Once again the light reference section was used having a 2. (5 kN/m) was used. 0.45 (Table 15). The base thickness was varied between 9. and would potentially have very little identifiable effect upon permanent deformation. A supplementary sensitivity study was conducted to determine the effect of base quality on the performance of geosynthetic reinforced pavements. The reduction of rutting percentage-wise is less for the low quality b.3 in.075 psi.8 as compared to the average Eb/E s = 2.-se compared to the high quality base. the use of a thinner reinforced aggregate base by about 20 to 25 percent with respect to fatigue. For this study the subgrade used had a resilient modulus E s = 3500 psi (24 MN/m 2 ).75 and 15. The changes in radial stress are relatively small (about 0.5 MN/m 2 ). (64 mm) thick asphalt surfacing and a subgrade with E s = 3500 psi (24 MN/m 2 ). The results of this study. however.5 used as the standard modular ratio in the cross-anisotropic analysis.

In the analytical model the effect of the prestretched geosynthetic was simulated entirely by applying appropriately concentrated forces at node points.Prestressed Aggregate Base An interesting possibility consists of prestressing the aggregate base using a geosynthetic to apply the prestressing force [35. A net prestress force of either 10. This effect was neglected in the prestress analysis. 50 N/m) of geosynthetic was applied in the model at a distance of 45 in (1140 mm) from the center of loading.36]. The shear stresses transferred from the geosynthetic to the pavement can be simulated by applying statically equivalent concentrated horizontal forces at the node points located along the horizontal plane where the geosynthetic is located. Once again. shear stresses are developed along the length of the geosynthetic as soon as it tries to return to its unstretched position. and a homogeneous subgrade having a resilient modulus E s = 3500 psi (24 MN/m 2 ). Theory shows that the force in a stretched axisymmetric membrane should vary linearly from zero at the center to a maximum value along the edges. provided slip of the geosynthetic does not occur. axisymmetric finite element formulation was once again used for the prestress analysis. These shear stresses vary approximately linearly from a maximum at the edge to zero at the center. 24. the same light reference pavement section was used consisting of a 2. Upon releasing the pretensioning force on the geosynthetic. (64 mm) asphalt surfacing. The cross anisotropic.5 in. a variable thickness aggregate base. (12. An external wheel load would cause a tensile strain in the geosynthetic and hence affect performance of the prestressed system. 20 or 40 lbs/in. The prestressing effect was simulated in the finite element model at both the bottom and the middle of the aggregate base. The geosynthetic membrane 79 .

The benefits derived from prestressing should actually fall somewhere between a fixed and free exterior boundary condition. This was accomplished by placing rollers along the exterior vertical boundary of the finite element grid. Comparisons of tensile strain in the asphalt layer and vertical compressive strain in the top of the subgrade are given in Figure 22 for a geosynthetic stretching force of 20 lbs/in.effect that was neglected would reduce the prestress force. In the prestress model the outer edge of the finite element mesh used to represent the pavement was assumed to be restrained in the horizontal directions. and deflections within each layer of the pavement is summarized in Table 16. On the other hand. 80 . prestressing the middle of the layer is more effective than prestressing the bottom. Edge restraint gives conservative modeling with respect to the level of improvement caused by the geosynthetic. but improve performance due to the reinforcing effect of the membrane. The important effect of prestressing either the middle or the bottom of the aggregate base on selected stresses. prestressing the bottom of the layer is most effective. To reduce tensile strain in the asphalt surfacing or reduce rutting of the base. strains. if subgrade deformation is of concern. (24 N/m).

I 7' 01 - I I I L.4 v . 017 t• L- - 76• SE - I18 ...• 1 - 6011 - OCZ• C- I HSI - 71-7 7EEI . SIN... . 7 I I i I I I I or I I •— --1 7+ 1 'WO 2 .. . I-• . A C/ 12..76' 1 17' 9E - 18 '0E - 0 167 S0* .'' 7 I -. I 1 I I I I I I I ■ I I I .. - 666 T0 — 7 . AC/1 5..01 1 DI I . I 1 i I ...m... 1 .7* E- R6E1 1 - 714 . AGGREGATE BA SE SUBGRADEE.‘1 CL 1.!f0* - A P[7 70 H I la I I' S (' 96 9 .I *MU _. ".. AC /7. 7771 I i I I t • Effec t o f Pres tress ing on Pavemen t 9' 7- I ±22±I SR7 I - SOLI 98 LT- 1 Table 1 6 I II ?I 61 I 9LIf 27 61 Z I 7 I "1 . I I 1 1 _. 3 IN. = 3500PSI 77 7. I I I I.'2 I 1 I 1 imman s _: .mmmmmm VERTI CAL IN oz sc- 71 . ...1 N VEFJ.-.qR' 9 16' 1C - I 2.I 8071 S' 6 7 [S 7.05093 I SHOI 1 . 5 IN. .7 6/SE - ■■ mm. D EFLECTION 'VERTI CALSTRES S I I I S' E 7..PR ESTRESS 0BOTTOM: 7' EN I UMW - I I co st a .1111 1: 81 .. 'VERTI CALSTRESS RADIAL STRESS RADIAL STRA IN TOP1/3OFAGGREGATEBAS E 1 2. -- 7 1 -I' I I .. I 1 --. 1-. 0 VI .....1 . 1 i I I I I I I C.1-7 I 6' IR- 1I I VERT. PRESTRESS eBOTTOM: 6801 2.. I I I 1 I ■ I / I I I I 1 I 1 I I I I 1 I I .. O 111 10 . 5IN..I fl' Z - TENSI LE STRA IN BOTTOM OF AC WM- "ff- IS 7S . 7 • S- SE. AGGR EGATE BASE SUB CRADE E s 3500PS I 6111 6 '6J - CH6- I . 0 I N. 5 I N.. e Z 14 . 7 . AGG REGA TEBASE SUBGRA DE E." 1 1 I .. _I 7 1 0 I L.!-I. SU RFACE I 6' 1 6 E •9 - 96' 6Z- I woz.. I. 7 .. .. I I . 5 4_ '3JI G 1 1 1 RSS 90' I 70C- 11910' - 77f1 0' _ 17 07 690' - 16790' - Lt 69 0' - .1 1 C8C ELIV E - 8' 7 - PRES TRESS 0 BOTTOM: 7ECE - I 7.

.▪ n1 2- 2• 2. .° • rn ^I ^I 0 82 .1..•-- . z 2- 7. E f fec t o fPr e s tr e s s ing o n Pa vemen t 30 .• nl 0 0 CO n . 4.< I z ".

5 IN A.0 10. v on Subgrade Figure 22.5 KSI aoo 70 13. Es=3.5 IN AC. Es=3.0 70 THICKNESS OF BASE.6) 1260 NO REINFORCEMENT OR PRESTRESS 1180 PRESTRESS 20 LBS/IN MIDDLE OF LAYER 1100 PRESTRESS 2.5 KSI 20 LBSAN BOTTOM OF LAYER 1020 16.C. er ( 1 0.C. r v 83 .RADIAL STRAINBOTTOMA.0 THICKNESS OF BASE..0 13.0 10. T (INCHES) (b) Vertical Strain E. Theoretical Influence of Prestress on Equivalent Base Thickness: E and c Strain Criteria. T (INCHES) (a) Radial Strain r in AC ur 3200 0 7 PRESTRESS 20 LBS/IN MIDDLE OF LAYER 0 2400 NO REINFORCEMENT OR PRESTRESS PRESTRESS 1-Z 1600 20 LBS/IN BOTTOM OF LAYER 2.

Four series of experiments were carried out.5 kip (7 kN) wheel loading moving at a speed of 3 mph (4. Important variables included in the investigation were (1) geosynthetic type. and to supplement and assist in verifying the analytical results previously presented. The test sections used in this study and their designations are summarized in Table 17. is represented by either M (middle of base) or B (bottom of base). an aggregate base (with or without geosynthetic reinforcement) and a soft silty clay subgrade. (2) location of geosynthetic within the aggregate base. A section name is generally preceded by the letters PR (prerutted) or PS (prestressed) if prerutting or prestressing is involved. This designation is then followed by the letters GX (geotextile) or GD (geogrid) which indicates the type of geosynthetic used. A knowledge of the notation used to designate the sections will be helpful later when the observed results are presented.LARGE-SCALE LABORATORY EXPERIMENTS Large-scale laboratory experiments were conducted to explore specific aspects of aggregate base reinforcement behavior. These large scale tests were performed in a test facility 16 ft. each consisting of three pavement sections. The pavement sections included a thin asphalt surfacing. Following 84 .9 by 2. (4) prestressing the aggregate base using a geosynthetic and (5) pavement material quality.4 m) in plan using a 1. A large number of potentially important variables exist which could influence the performance of an asphalt pavement having a geosynthetic reinforced aggregate base. Using up to 70.000 repetitions of wheel loading were applied to the sections in a constant temperature environment.8 km/hr). (3) prerutting the reinforced and unreinforced sections. The location of the geosynthetic which follows. (4. Therefore several compromises were made in selecting the variables included in the 12 sections tested. by 8 ft.

5 in. Crushed Limestone Section Designation Details of Geosynthetic and Section Specification PR. Sand & Gravel Base 1.75 in. Subgrade prerutted by 0. A. Prerutting carried out at single track test location GX-M Geotextile placed at middle of Base GX-M Same as GX-M (Series 3). CONTROL Control Section. 8 in. no geosynthetics and no prerutting GX-B Same as PR-GX-B. Subgrade prerutted by 0. Prerutting carried out at single track test location GD-M Same as GX-M but use geogrid PS-GD-M Prestressed Geogrid placed at middle of base Notes for section designation: PR = Prerutted PS = Prestress GX = Geotextile GD = Geogrid B = Bottom of Base M = Middle of Base 85 . A.C.C.GX-B Geotextile placed at bottom of Base. CONTROL Control Section GD-B Same as PR-GD-B. 6 in.4 in.no prerutting GX-B Geotextile placed at bottom of Base CONTROL Control Section.Table 17 Summary of Test Sections I Test Series 1 2 3 4 Proposed Geometry 1 in. no prerutting PR-GD-B Geogrid placed at bottom of Base.

The same soft silty clay subgrade was employed throughout the entire project. Figure C-1. the section PR-GD-B would indicate it is a prerutted section having a geogrid located at the bottom of the aggregate base. The properties of the pavement materials used in construction of the test pavements were thoroughly evaluated in an extensive laboratory testing program. For quality control during construction. prepared in accordance with the British Standard 594 [55]. placed and tested to insure as uniform of construction as possible. water content and cone penetration resistance were frequently measured and evaluated during and after the construction of the test sections. A brief description of the materials used in the experiments is given in the following subsections. Two different asphalt surfacings. INSTRUMENTATION AND CONSTRUCTION Materials All materials were carefully prepared. An asphaltic concrete mix was employed for the remaining three series of tests. These quality control tests are fully described subsequently. aggregate bases and geosynthetic reinforcement materials were used in the tests. MATERIALS. a gap-graded. The granite aggregate gradation curves used in each bituminous mix is shown in Figure 23. The asphaltic concrete mix was prepared in accordance with the Marshall design results given in Appendix C. described in detail in Appendix C. some of the readily measurable material properties such as density. and the specifications of both mixes are summarized in Table 18. 86 .this notation. Asphalt Surfacing. During the first series of tests. Hot Rolled Asphalt (HRA) mix was used.

4 0 (j) O O ■". 1-1 •-■ U F.? ONISSVd 30VIN30d3ci 87 ° 0 Fig ure 23. .67 ■ 0 II II • =III % Elam 3 PARTI CLES IZ E ( MM) ro u. 0 8g 09 (?) 1.7) 4 :. w ca to ti no no C 1.[ alMID1\103 DLLTAlciS51 ■■ .

88 . 100g. The viscosity of the asphalt cement was 4600 poises at 140 ° F 2.75 0.75 Delivery Temperature 110 ° C 160 ° C Rolling Temperature 80° C 120 °C 1.Table 18 Specification of Hot Rolled Asphalt and Asphaltic Concrete Property Hot Rolled Asphalt Asphaltic Concrete Binder Penetration 100 50 Binder Content (% by weight) 8 6. 5 sec.5 Maximum Aggregate Size (in. Test performed at 77 ° F (25 ° C).) 0.

The dolomitic limestone had a maximum particle size of 1. The limestone aggregate was slightly angular and non-flaky. low plasticity. as shown in Figure 24. Subgrade. The grading of the granular material. The upper 18 in. a weak granular base was used during the first series of tests. conforms with the British Standard Type 2 subbase specification [56]. silty clay known locally as Keuper Marl.5 ft. (20 mm).1 m) thick layer of drier and hence stiffer silty clay subgrade obtained previously from the same quarry. with a maximum particle size of about 3/4 in. This latter type of granular material is widely used in British highway construction. and about 3 percent passing the 75 micron sieve. as shown in Figure 24. The gravel base sections used in Test Series 1 exhibited extremely poor performance as evidenced by a very early failure at 1690 repetitions of wheel load.Aggregate Base. Both granular materials were compacted in the test facility at optimum moisture content to generally between 96 and 100 percent of the maximum dry density as determined by laboratory compaction tests. To enhance the benefit of a geosynthetic inclusion in the pavement structure. The subgrade used in this project was an inorganic.6. The clay subgrade was transported to the test facility in the form of unfired wet bricks from a local quarry.5 in. As a result. (1. (450 mm) of this soft clay was placed over an existing 3. lay within the British Standard Type 1 subbase specification. (38 mm) and about 7 percent fines passing the 75 micron sieve. The grading. This base consisted of rounded sand and gravel. A layer 18 in. and a 89 . the gravel was replaced for the remaining three test series by a crushed dolomitic limestone. (450 mm) of the soft subgrade had an in-place CBR value of about 2.

118 i 100 75 k i i/ i 90 PERCEN TAGEPASSIN G (MM) 10 20 37.1 1.112111 .40.S... iit % 01 44_ 60 50 40 Al. . 90 COITUS .. Gradation Curves for Granular Base Materials.„ 110 . . SIEVES (pm) 75 -5 300 600 1.5 ■1 80 70 Ilt NIi III DOLOMITIC LIMESTONE-) 'rill .0 10 100 PARTICLE SIZE (MM) F I'M I SILT C FIMI SAND C F IM IC GRAVEL Figure 24.B. .. ■ -■ 30 A.. widerIPL_SAND & 20 GRAVEL d SI Feelli Pi 10 tillilli 0 (101 0.dtrA .„CIF .

Both geosynthetics. All the instruments were placed directly beneath the center line of each test section in the direction of wheel travel. The magnitude and distribution with depth of the transient and permanent vertical strains in both the granular base and the subgrade.2 kN/m) and a weight of 28. Instrumentation was installed to measure the following parameters: 1. Transient and permanent longitudinal strain at the bottom of the asphaltic layer.moisture content of 18 percent. were manufactured from polypropylene. and copper-constantan thermocouples. The CBR of the underlying stiffer subgrade was found to be about 8 to 10. additional pressure cells and strain coils were installed in both the top and bottom of the aggregate base. (5. This additional instrumentation assisted in validating the analytical results. Beginning with the third series of tests. The other geosynthetic was a medium to high stiffness biaxial geogrid having a stiffness S g = 1600 lbs/in. Bison type inductance strain coils [58]. Geosynthetic Reinforcement. 2.9 kN/m) and a weight of 6 oz/yd 2 (203 gm/m 2 ). Details of instrument calibration have been described in the literature [59]. beginning with Test 91 . Instrumentation All the sections were instrumented using diaphragm pressure cells [57]. One geosynthetic was a very stiff. Two types of geosynthetics were used in the study (Table 19). (1. The arrangement of instrumentation installed in each pavement section was similar.5 oz/yd 2 (970 gm/m 2 ). however. woven geotextile having at 5 percent strain a stiffness S g = 4300 lbs/in. Details of one particular test section is shown in Figure 25.

c 0 wheel travel 40 32 pressure cell r. Figure 25. Typical Layout of Instrumentation Used in Test Track Study. strain coils 41 C direction of . 92 . during the 2nd to 4th test series when the base thickness was increased to 8 in. 2) The pressure cells and the pair of 1" strain coils at the top and bottom of the granular layer were available only during the 3rd and 4th test series. cols attached to geotextile 6 X coils in granular base 1 2I0 18 16 4 ' 12 12 10 thermocouple 6 4 6 4 2 0 2 distance from centre (ins) 8 10 12 14 16 18 20 Notes: 1) The vertical distance between the 2" strain coils in the granular layer was increased to 4 in. 6 asphalt colts 4 2 2 r dia. strain coils 2" dia.SIDE VIEW direc lion of wheel travel linch strain coils mobile strain coil 2— 0 thermocouples 10 — granular base pressure ccl s Keuper Marl subgrade a 14— 16— I s- 11111l11 20 18 16 14 12 10 8 4 2 0 2 4 6 distance from centre (ins) 6 6 1 10 12 14 16 18 20 48 PLAN pressure cell in subgrode pressure cell in granular base.

8 Grid Size (in.175 kN/m 1 oz/yd2 = 33.) n/a Conversions: 1 lb/in = 0.5 5.9 g/m2 n/a 1. Sg at 5% Strain (lb/in) 4300 1600 % Open Area 2 . Property Polymer Composition Geotextile Polypropylene Geogrid Polypropylene Weight/ area (oz/yd2) 28.Table 19 Properties of Geosynthetics Used.99 Tensile Strength (lb/in) 886 119 Stiffness.22 X 1.56 . X in.

However. Transient and permanent lateral strain in the geosynthetic. 5. 18 in. In addition to the instrumentation installed within the pavement materials. and at the complimentary location in the control section. 4. Transient stress near the top of the subgrade. was used to measure the surface profile. and Beginning with the Third Test Series the transient longitudinal stress was measured at both the top and bottom of the granular layer. Temperature in each pavement layer. During the construction of the first series of pavement sections. The silty clay subgrade (Keuper Marl) was installed as 7 layers of wet bricks. Each layer was compacted by using a triple legged pneumatic tamper (Figure 27) which had sufficient energy to destroy the joints in the bricks. (450 mm) of fresh silty clay was placed after the same thickness of existing stiff subgrade material had been removed. The surface elevation of the subgrade was established by measuring the distance from a reference beam to various locations on the subgrade surface. Pavement Construction Subgrade.Series 3 longitudinal strain was also determined at both the top and bottom of the granular base layer. The fresh silty clay subgrade used in the first series of tests was reused for all subsequent tests. a profilometer (Figure 26) consisting of a linear potentiometer mounted on a roller carriage. since the design thickness for both the aggregate base and asphalt surfacing was increased after the first 94 . 3. The final subgrade surface was then leveled with a single legged pneumatic compactor (Figure 28) before the aggregate material was placed over it.

.95 Pro fi lame te r Use d to Meas u re Trans ve rse Pro files on Pavem en t.

an additional 2. Aggregate Base Material. (100 mm) layers. respectively. The moisture content and dry density remained relatively constant throughout. All holes and horizontal layer surfaces were scarified as installation proceeded to give good bonding of materials. The 6 in. compaction was performed on the two 4 in. The first two layers each received 5 passes of the compactor. (50 mm) layers at a moisture content of 7 percent by means of a vibrating plate compactor (Figure 29). This was partly due to the fact that it was covered most of the time with a moist aggregate base. preventing drying out and stiffening of the subgrade. the condition of the subgrade remained constant throughout the study. The last layer was continuously compacted until no further densification was apparent. The aggregates used in the base were brought up to their optimum moisture content prior to placing and compacting. (150 mm) thick layer of sand and gravel base employed in the first test series was compacted in three 2-in.5 in. This value increased slightly to 3. The finished subgrade had an average CBR of 2.3 after it was first placed. Pressure cells and strain coils in the subgrade were placed in holes which were cut with special tools designed to ensure minimum disturbance around the instruments [61]. Compaction was performed at a moisture content of 7 percent by 96 .test series. For the 8 in. (200 mm) layer of crushed limestone base used after the first test series. The subgrade density of 111 pcf (1778 kg/m 2 ) corresponds to about 95 percent of the maximum dry density of this subgrade material as obtained in the British Standard compaction test [60]. (64 mm) of the newly installed silty clay was removed before construction of the pavement sections used in the second test series. In general. at about 18 percent and 111 pcf (1778 kg/m 3 ).2 at the end of the last series of tests.

Vibrating Plate Compactor Figure 30.Figure 27. Single Legged Pneumatic Compactor Used on Subgrade. Figure 28. Vibrating Roller 97 . Triple Legged Pneumatic Tamper Used on Subgrade. Figure 29.

To do this.S. The plastic bolt. Compaction of the two layers was continued until no rut was detected in the wheel path of the roller. The dolomitic limestone employed in the second series of tests was reused in the third series after the bottom 2 in. the geosynthetic was placed after all pressure cells and induction strain coils had been installed in the subgrade. a set of plastic nuts and bolts were used.using an 840 lb. In the last series. with the fine sand backfill held in position over the diaphragm with a thin plastic film. The induction strain coils were attached to the underside of the woven geotextile. The vertically oriented pressure cells were placed in the excavated hole in a prepacked condition. was tightened against a small nut 98 . The geotextile was stretched tight by hand-pulling at the edges while the granular base material was being placed. The geogrid was held in place by small U-shaped steel anchors after it was stretched tight by hand. (50 mm) of material contaminated by the subgrade was replaced. a fine sand passing the B. No. or within the aggregate base below the level of geosynthetic. A similar installation procedure was used during pressure cell calibration in a large triaxial specimen. To prevent large aggregate particles from damaging or influencing the output of the cells. For each pavement section. (200 mm) of base was replaced with fresh limestone aggregate. which passed through the central hole of the strain coil and between the filaments of the geotextile. holes were excavated after compaction of the layer was completed. however. 7 sieve (212 micron) was placed and carefully tamped around the instruments. Typical compacting time per layer was about 30 minutes. To install pressure cells and strain coils in the aggregate base. all 8 in. Geosynthetic Reinforcement. (380 kg) hand operated vibrating roller (Figure 30).

The wheel load was discontinued when a specified amount of rut was established at the surface of the subgrade. a very small hole was drilled through the thick junction of the grid before the coil was attached using the plastic nut and bolt (Figure 32). thereby potentially increasing its effectiveness as a reinforcing element. 99 . but prior to the construction of the asphalt surfacing. To prevent the strain coils from interlocking with the surrounding soil or granular material. The purpose of prerutting was to induce a tensile force in the geosynthetic. Sometimes prerutting was performed down the center of the pavement as a primary test variable. To carry out prerutting. they were covered on the underside by a small piece of geotextile. The geosynthetic used in each test section was carefully examined and stored after each series of tests were completed.located on the upper side of the geotextile (Figure 31). For the geogrid.1 kips (5 kN) for the sand and gravel base (Test Series 1) to 2 kips (9 kN) for the crushed dolomitic limestone used in the remaining three test series. a moving wheel load from the Pavement Test Facility was applied directly onto the surface of the granular base layer of the pavement section. The applied wheel loadings varied from 1. In other instances prerutting was carried out along the edge of the pavement as a supplementary study. Prerutting. When prerutting was conducted along the center line of the section under which strain coils were installed. no geosynthetic materials were reused. This simulated the traffic condition during construction when heavily loaded trucks pass over the aggregate base. Prerutting was carried out in every series of tests after the aggregate base was placed. vertical permanent deformations were monitored during prerutting of both the aggregate surface and the subgrade.

d3 tj).. _ t • "s• • ._ ... ..• :"44.. . • .. • I...-4.. s'..S■ '"44t 1 1. . .-.1.i..*• ..i1 . .. '-. . ... .f: ..• t • . .I r .. . ... -I .. •_ ..••• 'it .1.._ 4.. P.... .... ‘:i's''-'11.. -•.4. ....'... . Diameter Induction Strain Coils.r. 7:::... c 1... .1. . '06.) i.._0.w- t . e _...■ ....: ...:::!: 4 -`..-. ' ..........--• : . . 0 .. .r .. 4..ii.. 4. .: r•• .1...' -. ..:—... • . . 1::: : 7„.: --- 1.i.. •i. . ..... .. .' -- ' -. .. .-. . ..... e ---•. i t ' . ' .I. ..• ..i:::.. _ •... ... ."c"'''- - ...-.. c.._ _ _ _. .. Figure 32.-- - _. .. .' PR . 1 00 .. . . ......:4-0 I. 4•..• .-:-... •••4 . •L * - . ..._ . ... .... .„.„.... .. kr. . ..' 71.. t. ■ '" ' . . IT.....-.. lia .1 71 N" . Diameter Induction Strain Coils.. ..*.. „ ...i. i.: ... s...? -'. . Geogrid with 1 in... • ._. .._. . 4 .44i. -....- •... i tE. • as i i 411“ir. to. ' . - . 4 ii. r .. • '. . i - _ A' ' '... : .. . A i •1i ' I I: -o - v 0. 'A . .. ' :- ... ........ . .. : • ... cl 4 ..• " l' .. . .._...‘.' 1 :1 . 1 •.. ' ^ ._-4... 1. Woven Geotextile with 1 in.-... r ' 1 I 1. .I.. -.1 .• " -- . 1 '.. . .Figure 31..• ' ..'• .7...._.. ...

the load from the hydraulic jack was then released applying a prestress to the base and 101 . Criteria to discontinue the wheel load was then based on an accumulation of about 2 in. One section included in the Fourth Test Series had a prestressed aggregate base. By jacking against a steel column which was firmly bolted to the concrete floor. the entire rut in the base was refilled and carefully compacted with aggregate preconditioned to the proper moisture content. The geogrid then went through a set of rollers and was connected. In performing the clamping operation. Very often during prerutting. some additional tensile force may have been created in the geogrid. With the exception of the sand-gravel base. (7 kN/m) was achieved. prerutting generally resulted in local densification of the aggregate base.When prerutting was carried out in areas away from the centerline of the pavement section. a second clamping bar was used to lock the geogrid in position thus maintaining its tensioned state. a tension force was generated and transferred to the geogrid. A schematic diagram showing the prestressing arrangement used in the laboratory tests is given in Figure 33. the rut created in the aggregate layer needed to be partially refilled because the ram used to force the tire against the pavement had a limited amount of travel. to a hydraulic jack. by way of a load transfer steel bar and steel cable. After the pretensioning force was "locked in" the geogrid. After the first layer of granular material was placed and compacted. the second layer of aggregate base was immediately placed and compacted. Prestressing was accomplished using the stiff geogrid. (50 mm) of rut at the surface of the aggregate layer. Upon completion of prerutting. the geogrid was clamped to the side wall of the pavement using the clamping system detailed in Figure 33. Prestressing Aggregate Base. only the surface rut could be monitored. As soon as the target force of 40 lb/in.

Compaction was carried out in both the longitudinal and transverse directions of the pavement area. The excess material helped to prevent rapid loss of heat during transportation. they were covered with a fine asphalt mix before placement of the main bulk of material. which used 50 Pen binder. To protect the strain coils placed on top of the aggregate layer. was about 320°F (160°C) at the time of placement. Compaction of the single layer was performed using the same vibrating roller that was employed for the aggregate base. The temperature of the hot rolled asphalt (HRA) mix used in the first series. when the hydraulic jack was in action) was about one hour. which used 100 Pen binder. the material was transferred to the test sections by using preheated wheelbarrows. Three tons (2730 kg) of material were delivered for each test series. was about 230°F (110°C) when it was being placed.e. This quantity of asphalt is about three times the amount required for the single lift construction. The total period of stretching the geosynthetic (i. Rolling was continued until no further movement or indentation was observed on the surface. Asphalt Surfacing.. Both asphalt surface mixes used in this project were transported by truck from the same quarry located about 22 miles (35 km) from the test facility.subgrade. Upon arrival at the test facility. 102 . The temperature for the AC mix. The whole sequence of construction of the asphalt layer took about 35 minutes. All exposed cables were also protected when the mix arrived by covering them with carefully selected material from which relatively large aggregate particles had been removed. The first pass was made without using vibration to avoid creating large distortions.

4 . a) Cn 0 1-) a) ao 103 ai .4 U1 V $. a -C 0 - u 0 CU ro CU $.•U CU 0a U Cd 0 >.0 bD r. 0u (11 4-e 0 0 c 0 a) E 5.4 E' Uc 0o _a 0 0 a) a) a.4 a) O 4-1 _ a. "0 CU • 0 la. -c) g atn c.

(2) strain coil readings. 104 . After placement of the subgrade. The finished profiles for all 12 sections are summarized in Table 20. four insitu CBR tests (ASTM D4429) and two dynamic cone penetrometer tests (Figure 35) were performed at the surface. Construction Quality Control Construction of the subgrade during each series of tests was closely monitored. the thickness of the layers of the completed pavement were not exactly as specified. For the first test series. However. The thickness of individual layers was obtained by several technique including: (1) core samples of the asphalt surfacing. static cone penetrometer tests (Figure 34) were performed to determine the corresponding CBR values after compaction of each layer of fresh silty clay. The nuclear density tests were complimented by regular laboratory moisture and density tests using four 2. the nuclear density meter (Figure 36) was used to determine the density and moisture content at various locations. variations between sections within one series of tests were within acceptable tolerances.Pavement Surface Profile Despite great care during construction. This was probably due to difficulties in judging the quantity of material required for a specified compacted thickness.5 in. The moisture content was also measured at a number of locations for each layer of subgrade. (64 mm) diameter tube samples. In addition. and (4) cross sections taken during trench excavation at the end of each test series. generally less than 10 percent. all of the tests described above were repeated on the subgrade surface both before and after the wheel loading tests. After the first test series. with the exception of the insitu CBR tests. (3) measurements from a reference beam to points on the top and bottom of each layer.

1 1...c:t.. ) a) From Pave men t a) aSeg 4...1 ..1 Ln• -. • Ln tn Ln (N (N CO • • • in '4' -11 Ill Cr) toO• t.1 Sect ion* CU Pavement Tab le 20 Depth of Geosynthet ic ro a.13 R K80 X cr 080 c6 T 0 .. ) So ft Subgd r•-• (c1 (-N3 en N CTS cm aNzul m ul CO N ir) 0 • • • I) h Ln cs% kr) • -.: al 0 @ DI .) (4 ) (N) N0-1 (N) CNI r•1 1. Th ickness of Layer ( in.0 • r•••4 1-4 ..1 • .. ..•4 In to -1-4 I-1 r•••1 Cr) N • • • m • c14 to CO • .1 ■.-4 1/40 tn ■ 0 CO O COOC O OCO• CO N WI (N1 (v. • 0`&44d0 V? T 0 5-1 R .-1 v-1 Ln Ln tn CY) CO ..0 CO C O O CO tn Lr1 C • • %.(f) CO 'Cr ro F-4 (N 105 . rt:1 ..0c6 ci. • • • • In r-1 I-1 "..1 P .•1 r-1 1 .„ • fa —4 • --._ • • • • N h h "-I . • 0 ..-4 In C'") .. • .from Sur face ( in.•1 cn sa Tias 01 qsaI C...

Static Cone Penetrometer Test on Subgrade. Clegg Hammer Figure 36. Figure 34.Figure 35. Dynamic Cone Penetrometer Test on Subgrade. Nuclear Density Meter 106 . Figure 37.

at least ten core samples were taken to determine the compaction. In addition. Density of the asphalt surfacing immediately after compaction was measured by the nuclear density meter. void ratio and density. however. in most cases. In general. and at the end of the wheel loading test. At the end of each test series. convergence of the analysis was not possible. On delivery of the asphalt surfacing. As a result. nine Clegg Hammer tests (Figure 37). A summary of the results obtained from the various quality control tests just described is given in Table 21. the shape of the recorded deflection bowl was different from those encountered outside the PTF.Gradation tests were performed on the aggregate base when they were delivered. Falling Weight Deflectometer (FWD) tests (Figure 37) were carried out on the test sections. The results of these tests. after compaction. and nine nuclear moisture-density tests were performed on the aggregate base before and after each test series. The test results were further complicated by the fact that the test facility was constructed on and surrounded by thick concrete which reflected abnormal signals to the geophones of the FWD. no significant change in grading was noticed as a result of the various operations. The high deflections created difficulties in reliably back-calculating the stiffness of individual layers. as shown in Table 22. At least two dynamic cone penetrometer tests. Tests were performed directly on the aggregate base as well as on the asphalt surfacing. 107 . appeared to be unsatisfactory due to the fact that very high deflections were obtained from the impact load of the FWD. six samples were taken to determine the aggregate gradation and binder content.

O CO 0 0 ts) .-1 ■-1 .31 .• 11 CO .. Av Afte r (NO o fBlows/ 10cm ) Materia l type lsayToi woo ATFIr010Jo add.--i •--1 s in •d• s cr CO C.Z1 112 1 11 17.2 8 . Air Void 2(%)Be fore 1:1 4-1 X CU al >1 0 s-i •.--i 0 4-1 C. Binder Conten t( %) Max.. 9 U) 1„1 al 4-1 .) 108 a.414 0 z 8 ()ra 8 0.--1 •—• CO .:I• rn 'CI' .ga tn rn cr) • 7 as r- ✓a kr) rn 0 .3' in CO • ••1• cn 0 011 N • i't) 0 )4 CV C.3.▪ • ■ V' V 6171 T VT S' 9 cr "g 8 u) to • 7 al s 03 1.) n o 1. 3 1 8. fore Be ing Read ic Cone Dynam e.) • . Q) • 7:14 a) . —i 04 ■.) .-1 4-4 V 0 g . g rUl 4-1 oT l i vie sy '8 g 8 2 (1.--i 03 2:1 > CY . • . re) en • • cl tri 111 0 .4 . 710 •••••■ w . Aggrega te S ize (mm ) Ave.. Agg rega te S ize (mm ) %Fine r than .. Dry Dens ity (ps f)Be fore After 3 Be fore tent Con (% ) ture Mois Ave Afte r Be fore Ave...--1 03 CO . Compacted Dens ity (pc f) Be fore After ] Ave.•1 V* 9 Si I T VT VT S .1-) .1• .0 r) b •0. 9 Qs •-1 a) rCS C r0 U) S•4 ••■ O >4 0 5 u-i0 w >-■ N4. 075mm Ave.e.) C4 111 14 tn IS U) 2 •4' . —I 0 .) t)E u) Max. 6171 Till VT s• 9 Ser ies ra 0 04 0 ••-1 IA 0 4J C.--1 . . 7 2 86 204 2. C legg Hammer Read ing After .1. .9 Test a) 1 8 4 4-) u-■ u) co 0 • s .0 0 V) en rn • • 0 0 0 tn )4 E rn s .-1 O.11.. .--1 C4 .1* rn S rel CO the read ing by 110.) ••-I .2 5 4-1 -1-) >I 0 1 11 Ta' C.-1 . •zr s •cr .1 )u1 0 • • . rn O' L.

0S6T ZLT S -XD In CE) c/4 0 CD CO 0 CO m 0 u1 r-.1 r-1 r.--1 tfl 1". 9179Z LLE 8-V0.1 r-f C'. l'N CO r-i CNI 6662 LT T 60TE 9ZT 01fcr M l'N CV Contact Stress (kPa ) De flect ion**(micron ) .co .0 0 .--i . I ■ Sect ion Des ignat ion* Contact Stress ( kPa ) Deflect ion**(micron ) r1 Test Series Tab le 22 M in f".1 r-1 C.ON .-4 CNI 109 .--1 ..--i WON .1 01 OLVZ LIE OE LZ 8TT .I CV 1'- 0 N ko N tn 0 kr) C'.-Hd ttE rICEJNOD 68SZ 8--(19-Hcl E9T rIOELINCO OSOZ rICRLIOD •zP CI' 0 %.

The maximum wheel load that can be achieved by the PTF is about 3. It consisted of a 250pass cycle. Table 23 summarizes the loading sequence adopted for the last three series of tests. [61]. This facility has been described in detail by Brown. across the pavement to nine specified positions (four on each side of the center line). et al.PAVEMENT TEST PROCEDURES Load Application The pavement tests were conducted at the University of Nottingham in the Pavement Test Facility (PTF) as shown in Figure 38. A realistic simulation can be obtained of actual loading where traffic wander exists. The beams in turn are mounted on end bogies which allow the whole assembly to traverse across the pavement. A load feedback servo-mechanism is incorporated in the system to maintain a constant wheel loading. Loading was applied to the surface of the pavement by a 22 in. 6 in. starting with 55 passes along the center of the section (Position 5). The whole assembly is housed in an insulated room having temperature control. The carriage moves on bearings between two support beams which span the long side of the rectangular test pit. (150 mm) wide loading wheel fitted to a support carriage. followed by 15 passes at position 8. Two ultra low friction rams controlled by a servo-hydraulic system are used to apply load to the wheel and lift and lower it.4 kips (15 kN). then 7 passes at 9 (refer to Table 23) until it finished back at the center line where the cycle was • 110 . in a random sequence. At each position a predetermined number of wheel passes is applied. The spacing between wheel positions was set at a constant step of 3 in. with a speed range of 0 to 10 mph (0 to 16 km/hr). Multiple Track Tests The moving wheel in the PTF can be programmed to traverse. (75 mm). (560 mm) diameter.

111 'avemen t Te s t Fac i i t .

Nonetheless. In the first series of tests. single track tests were carried out along one or both sides of the main test area where the pavement had not been previously loaded. 100.690. During the scheduled recording of output from the instrumentation. Only surface rut depth was measured.repeated. because of the rapid deterioration and very early failure of the pavement sections. The distribution of these passes across each loading position is shown in Figure 39. the numbers shown in Table 23 and Figure 39 apply only to the center of each track position. Hence. respectively.690. These special tests normally involved the use of a much higher wheel load. during the test.070 and 106. these tests helped greatly to confirm trends observed in the development of permanent deformation during the multi-track tests.300. so that the deterioration of the pavement structure would be greatly accelerated. since instruments were not located beneath the loading path. the wheel constantly overlapped two tracks at any one time. The total number of passes in the multiple track tests for the second to fourth series of tests were 69. Therefore. The single track tests also made possible extra comparisons of the performance 112 . the loading program described above could not be executed. Stress and strain data were not obtained for these single track tests. Note that the width of the tire is larger than the distance between each track position. and their distribution is shown in Figure 39. This procedure ensured that consistent and compatible outputs were recorded from the instruments installed below the center line of the pavement. The total number of wheel load passes for this test series was 1. Single Track Tests On completion of the main multi-track tests. the center line track was given an additional 100 passes of wheel load before actual recording began.

▪ • .+ 4 CO b 0 0 ci (0001 0 Csi 0 - •—• 0 x) awn 'min do sassva 113 30 usrakm In Dis tr ibu t ion o f t he Numbe r o f Pass es o f Whee l Load in Mult ip le Trac k Tes ts. C+1 4.■•••••••■ .14 z rl 0 UI 0 0 in N 0 z in 0 H E-I H 0 a.

Ce) CO Cr tr) 0 z i a) 0 114 . tf) Cr CO 0 Cs.0 CO N in ■Co Ch Ln Ce) N ■S) mm) dis tan ce.

The load exerted by the rolling wheel on the pavement during Test Series 2 through 4 of the multi-track tests was 1.5 kips (6. Three additional single track tests were performed during the second to fourth test series. only minor variations of load occurred. This load variation was probably also due to the unevenness in the longitudinal profile of the pavement. due to the rapid deterioration of the pavement and hence large surface deformations.of pavement sections tested in the prerutted and non-prerutted condition. tire wall strength. Wheel Loads Bidirectional wheel loading was used in all tests.7 to 2.6 kN).8 kips (8 kN) was used for the First Test Series. a wheel load of 1.5 kips (3 to 11 kN). Details of these tests and their purposes are shown in Table 24. generally less than 10 percent of the average value. In the first series of tests. difficulties were encountered at an early stage of the test in maintaining a uniform load across the three pavement sections which underwent different amounts of deflection. much stronger pavement sections were constructed. Therefore.5 kips (6. the actual load varied from 0. For all other test series a 2 kip (9 kN) load was applied. Bidirectional loading means that load was applied on the wheel while it moved in each direction. As a result. while the average load was 1. all of these supplementary tests employed bidirectional loading.6 kN). Based on a previous investigation of the effect of wheel tread. In the single track tests. In subsequent test series. The designations of the test sections follow those for the multitrack tests previously described. With the exception of the single track test carried out during the first series. 115 . however. and refinements were made in the servo-system which controlled the load. The tire pressure was maintained at 80 psi (550 kN/m 2 ).

respectively. the temperature of the asphalt in the wheel track could increase by as much as 9°F (5 ° C) due to the repeated loading by the wheel. During the actual loading. In addition. (68 and 76 mm).6 °F (20 ± 2°C) throughout the testing.2 to 4. However.tire pressure and load. assuming them to be circular.8 km/hr) with slight variations between forward and reverse direction. the contact pressures acting on the pavement from a 1. respectively.6 and 9 kN) wheel load were estimated to be 67 and 73 psi (460 and 500 kN/m2 ). Therefore. a slower speed was sometimes necessary to maintain constant loading.5 and 2 kip (6. position and speed. of 2. All pressure cells could be recorded continuously. Near the end of the test when the pavement surface became uneven. The wheel moved at a speed of about 2 to 3 miles per hour (3. but it was only possible to record one strain coil pair at a time. elevations of all the reference points at the surface of the sections along the center line were measured and checked. The temperature inside the PTF was kept at 68 ± 3. it normally required about 100 to 200 passes of wheel load at the center line were normally required to obtain a complete set of 116 . Temperatures at the asphalt surface and within the aggregate base and the subgrade were found to be about 2 to 4 °F (1 to 2oC) lower than that of the air. These gave radii of contact areas. it was previously observed that during long continuous runs of the PTF. resilient strains and transient stresses were recorded on an Ultra Violet Oscillograph which also recorded wheel load. Data Recording Procedure The transverse profile and permanent strain readings from the aggregate base and silty clay subgrade were taken at appropriate intervals during testing of all pavement sections to establish their deformation characteristics under loading.7 and 3 in.

)c •c c.base preruttedby 2in.13 117 PrestressedGeogrid at middle of non-preru tted base Control sect ion..)g • Ul ftC 14 a. . use geogrid c4 a g T — a a • ri8 C.Control Sect ion 8in.) 4.-i CU $4 to C .) 5 .LNICO E.0 •rt • U) a. Geotext ile at bottom o f base Same as the 1s t GD-B .4-) (I) -.. Same as PR-GX-M. Crushed Limes tone Ul -4-1 0 Geotext ile at m iddle of Base Same as GX-M.. B-GO S-Go Sec t ions Used Table 24 Todeterm ine per formance of reinforcedand-prerutted and prestress edbut non-prerutted sect ions To compare performance of non-prerutreinforced and prerutunreinforced sect ions To compare per formance o f reinforced and un-rein forced unbound pavement sec t ions qSalJOe SOO-Ind u) cot Cf) cl. base preru tted by 2in. Geogr id at bottom of Base Sect ion limil vnthet ic Details of sGeo Sect ion rICRI.

based on the relative behavior of the similar control section in each test series. and also near the end of each test series is given in Table 25.. were monitored regularly by means of a readout device. permanent vertical deformation. were excluded from the averaging process. Air temperature of the PTF was obtained from a thermometer placed inside the facility. or with depth in the pavement structure at a particular time. an average value has been reported in the tables and figures.induction strain coil readings. permanent vertical strain. usually only relatively small changes of this variable occurred with time. In addition.). Direct comparisons can be made between each test section within a given series. The outputs from the thermocouples. A "peak hold" data acquisition system was later used to record the peak values of the stress and strain pulses. 118 Er7.e. Unless indicated. . Whenever there is more than one value of data available (i. TEST RESULTS A summary of important measured pavement response variables recorded at both an early stage of loading. all the results were obtained from multi-track tests. after relatively large permanent deformations had developed.atic data. number of load cycles). comparisons can be made between test series if appropriate adjustments are made in observed responses. The permanent strain results were obtained near the end of the test. which measured temperature at selected depths in the pavement structure. Most of the results presented show either variation of test data with time (i. Vertical resilient strains are given at early stages of the test when the pavement structure was still undamaged. subgrade stress. etc. however.e..

) f§1 4:6 .44 5 ILI) a+ C.1 4) tri CQ a■ 5 ro 5 0 a) g 4.4 U) C) U) ••••I 0 •ri CU Lit gj int •nl Cl■ .-.) (13 II IV 4J 44 4J 4t) rti B .2 23 T) .Z .48 ai.§O N 0 LM •.7 • E 0 a) . '8 8 4 c 61 01 0) 07 IM M 4J 0 CC C) Cl/ L-1(.) •cl• 1.)-1 01 cu 8.8" am 4 a) W0 4J 4.4 0 E> 0 M 4.7 it) CI) .-4 cn z ._) < a) )-4 E ct• •-• 119 • o # # .U) C) 1-0 CI) tr) N iJ 0 1-1 W V) .8 2 4. V) (1) C)E 4j 1S) a9 1-4 4-1 II 0.C ro E a) ID 01 N0 W 1-1 (11 0 0.9 V) 4.4 a) .2 4.z rj 4j ccl 4j a:1 4. 3-4 L ) .1 r' L4 4) 23 °V1 .C el 4J to 44 0 4-4 0N 1-1 1-1 (. • f .

However. prerutting does not appear to improve the overall rutting performance of the weak pavement section compared to the geotextile reinforced section which was not prerutted. and also in Table 25.Permanent Vertical Deformation In this study the permanent vertical surface deformation of the pavement is taken as the primary indicator of performance. The permanent deformation occurring in the base and subgrade are shown in Figures 42 and 43. respectively. as shown in Figure 40. as will be discussed subsequently for the single test track results. since similar 120 . These results indicate that the inclusion of a stiff to very stiff geotextile at the bottom of the very weak sand-gravel base reduces the amount of rut by about 44 percent for a rut depth of 0. prerutting of the reinforced section was quite effective. Thus. Furthermore. This section was reinforced with a geogrid at the bottom of the base and resulted in a 66 percent reduction in total rutting of the base and subgrade. This finding by itself is misleading. the prerutted section in the second series performed best. Figure 40 clearly shows that the pavement sections used in the first test series are very weak. (11 mm) in the control section.43 in. Because of the use of a higher quality aggregate base and thicker base and surfacing. the life for the pavement sections of the other three series of tests was considerably longer. in contrast to the results of the first test series. with large deformations developing in less than 2000 passes of wheel load. The accumulation of surface rutting measured by the profilometer is shown in Figure 40. Permanent vertical deformation in both layers was calculated from the changes in distance between the pairs of induction strain coils. Profiles showing the permanent deflection basin at the end of the tests are given in Figure 41.

1-1 )-1 0+ a) cn ›. 0 . o a) . 11 al . 0 :1? (f) 0.. a) ./1 ul 0 .0 aJ r.) it II X CU >LD' CIt-D CI) 0 . 11 r-4 P-..-.71 f -1 4-1 no 0 0 a) a) cD c.- '.I 4-I in • ce . i53 ai 0 g `1] rsi UI 0 UI 0 2 F— 0 HIdaa 121 o 0 in 0 ( 'uT) m 0 on ..tC a) . 14-1 4-1 4 cn 7:1 a) a) E.= 7:1 4-I al R.1 IZ .--1 0 .-i )-) U) 1-1 0 0 in N ri lD In 0 O O 0 0 O N 0 0 0 aJ E 0 a) )--I 0 ._.-4 1-1 r-I 0 f:4 in 2 o —1 au a) fil co 4-I 1 "0 -1 1(1) 1-) P.--1 4-I a) a) up P.--1 )--1 $.7 o —1 4-1 0 Z 4-) Cr) a) )-) P. 1 Z t.--1 as R 0 ..

5 1.5 1.0 0.0 0. Pavement Surface Profiles Measurt.1st SERIES CONTROL GX-B PR-GX-B 2nd SERIES CONTROL GD-B PR-GD-B 3rd SERIES caumr.0 GX-B GX-M 0.All Test Series.d 1)7 Profilometer at End of Tests .5 4th SERIES Note: PR= Prerutted GX-M GD-M _pS-GD-M GX= Geotextile M= Middle of Base PS= Prestressed GD= Geogrid B= Botban of Base Figure 41.0 _ 1. 122 . 0.

4 -1C3 4-4 O Z 0 f-.) II II O L.7 C..-i Cr) a P. 0 4-4 3.--i o . 0 r-I 11 CI G RI CV 0 = .0 ..-.4 v... •-■ 0 O 0 1 4..• 4o 0 4-4 0 k a.1 14 ..O. 00 0..J Iv k •o-f 0 0 CU S-r P-4 4-1 Er} (.4 CU 3.-4 •. Z -C.I) •-■ N 3.--I 4-1 4.n cn al cu na 11:1 rn 0 .) E-+ '"'1 .. '4-4 a) O o N (LI E 0 75 44 tI.-o (s) r:=1 I.1 . = I I II 0 S-4 (1.-. oi) co a0 id d .-.-1 F.4 a b.4 CJ 0 d .3 1-1 4-1 O cn g .-r oo-I 4.. ._1 1:3-o (1) 2 3 R] ..1 1:1' 11 T m0 .--f 4-1 0 0 P3 CU oC=1 CO 0 •-i (1. 14. or. tri C) ( 'UT) NOLIMAIOZICI Inammemd rivaardA 123 • > o..■ co 1-4 • "Ct 4-.1.-1 ct E o—I $.4 4:C N0 ON 3.: Cil N. O ai ILI a) ca r..0 in r. Iti 0 ..() PE4 r.3 4-1 0 .4 O .:r2 C... T 1 ( 4-1 OD 1.4 [15 c1. E = z ri 0 .-1 Fig ure 42. 0 r-4 0 Z 7:3 CV IZJ (i) 11 i C.

An important reduction in subgrade deformation was evident when a geosynthetic was placed directly on top of the subgrade. it follows that the pattern of permanent deformation as a function of load repetitions observed in the base was very similar to that observed at the pavement surface as can be seen by comparing Figure 40 with Figure 42.very good performance was also observed for prerutted sections which were not reinforced. use of geogrid reinforcement.. particularly for the prerutted sections. This section was not prerutted. resulted in better performance than a higher stiffness. despite its lower stiffness. Therefore. 5.. most of this improvement occurred within the aggregate layer (Table 25. the amount of rutting was reduced by a total of 28 percent. woven geotextile (Sg = 1600 lbs/in. 5. Permanent vertical deformation in the subgrade was relatively small compared to that occurring in the base. Figure 40c). A large portion of the total permanent deformation occurred within the aggregate base. Figure 40c). 124 . Only an 8 percent reduction in rutting was observed for the geogrid reinforced section used in Test Series 2 which was not prerutted (Figure 40b). Results from the last series of tests indicate that prestressing the aggregate base appears to improve performance compared with a nonprestressed section having the same geogrid reinforcement (Table 25. Figure 40c). A similar relatively low level of improvement with respect to rutting (13 percent reduction) was observed for the section in Test Series 3 reinforced with a stiff to very stiff geosynthetic (S g = 4300 lbs/in. When the location of the geotextile was raised to the middle of the aggregate base in Test Series 3.2 kN/m) located at the bottom of the layer (Table 25.2 kN/m) when both were placed at the middle of the granular layer (Figure 40d). Further.

.) C. 4-) • 5r:14 1 ... 7-1 1 cf) 04 1-1 CU . r-4 1) r-1 cv til .--i O0 O 0 • 4) Cf) Ca a) co 1:13 I O 0 ▪0 —4 I 0 0 N O 0 0 0 0 0 C.) = r-I II 11 X i=i 4-I a) 0 "0 Ca Cti 0 /-1 ■ 4 "0 cll '10 Cil cu cn ).. ry 0 ( 'UT) noiiramarca araialsA avaNvmaci 125 O 8 0 0 > .1 a) a-) i-• 4-■ 1-r ci7 a) 0.-i N m O .1 rt.-1 0 0 LI-I (1) CU CU r-I C. LE) L4-i O +.) O 0 4J g (j] .) I) CU -r-i a..-1 0 0 .) C.-I cn G N-1 0 O rt N O 0 8 0 O N 0 O .4 O Ln o ...1 C/) 0 45 .1 7:.-1 S. .:.0 c.) S-4 C. O (J) 4. 8 G.0 Number o f •• 0 0 N z ri 0 .L....4 4-i ty..) 14 0 O C.1-1 E-4 P.4 1 0 'RI X **-1 0.

At the same time (with the exception of the first series results). a substantial decrease in permanent vertical strain occurs 126 . a redistribution of vertical permanent strain is seen to occur due to the presence of the reinforcement. Figure 45 shows that as a result of placing the geotextile at the middle of the aggregate base. In general. Other interesting results that can be obtained from these figures reveal the following differences between pavement sections: 1.as shown in Table 25 and Figure 43. a decrease of strain is generally observed near the top of the subgrade. decreasing rapidly with depth towards the subgrade. Permanent Vertical Strain The variation of permanent vertical strain with depth for all the sections at the end of testing is shown in Figure 45. 2. For sections with the geosynthetic reinforcement placed at the bottom of the granular base. the pattern of results is very similar for all test series. with large permanent strain at the top of the granular base. The average values of strain are plotted at the mid-point between the two strain coils which measured the corresponding vertical movement. The trend in the development of total permanent deformation in all 12 sections of the four test series in the multi-track loading tests was generally confirmed by the single track studies (Figure 44). Reductions in subgrade rutting of 25 to 57 percent were observed for this condition. When comparing results from the geosynthetic reinforced and control sections. an increase in permanent strain occurred in the top half of the granular base.

5kips whee l1 S' 9 6' S E' 9 SI' 0 LEO E0' 0 Or0 SZ' O 170' 0 0' 9 9' 8 0' 8 9 —(T Z' OT 5' 9 77 88LE IV6Z 9L9Z 626E 03n )4 3 4Iettdsii 409 0 ** • 127 4./ d0 . Sect ion I S' 0 Z9' 0 L L' O 9E' T SS' T 95' 0 55' 0 V6' 0 £ 9' 0 PA S peoTTaawi scITX S' I Josassed 0SI4e 24 e Isa gassay Sect ion SI* 0 U* 0 ET' O OI* T LO' T SE' 0 69' 0 6 5' 0 °SPE( Data at 1 262 pas ses of1..11 ti — .-I 0 o O 6 c) .-4 ✓ 0) . 5kips whee lload 3 V a(ps i) t (u E ) c (ps i) 3 V 8 ci.-.-1 0.c4. 1/8.1 .1 8 2 d . 2 12ES11V.1'1 2 '.1.1 OZ* 0 6Z' O BZ* 0 N Ch 1/4.) •-4 W 4...1..): 71-1°4.--1 . 3 1.1 '0 -•■-■ LI] Z* 9 0' 9 6' 9 a Permanen t De formation ( in ) 1 Subg rade Aspha lt 0 •H '0 E▪ $_./ ). 1 2• 5 731 -.'1 o t 8 2 2 to' 124 t-4 < > ■-4 E E V ° u 41 1. 6 . 5k ips wheelload 29 17 ** Da ta a t 70000 pa s s es o f1.) 0 (1) L0 • g i41 4) 'CI • ... 5 1.9 M I-I 44 0 44 0 (./ • E E cn 7 0 0 ..I C I H-XD ricuusx0 8-0D ' rICU.8 '1 Q'3 il12 5. 6/8..4 t .0 •. ... .Q " .4 2 .7-14 E 8 :LI II Q.0 VE* 0 4-1 11) 0 4.Fij 2 .W . 2/8..1000 sa Taasqsay -XD 1CUIOOD 8 -XD-Hd Tes t Ser ies VZ' O EV* 0 ° ET 9L' 0 £8* 0 8Z' O [Da ta a t 10000 pas s es o f1. 2/8.0 .. 3 1.. a.—.411)5-24117-IL131 0...(11 5' S S' L 8' L SLibgrade •—I *86TZ 4E 86Z *SSEZ ££.I CU . 01 as ...4 En i .4E.'.ri 111 90' 0 L0' 0 E0' 0 ‘8 ..1 'CI "4 8 •••••■ ("4 • •••■.4.. r* 2 B ° 01 tri 0 (0 ul 41 --..4 N -.17t I 9LE BE LE 0' 8 S' L E' 9 aS PU Pe rmanen t De format ion ( in ) 9 Z' O ZP* 0 89' 0 OL* 0 06' 0 86' 0 . 4. 4. 4/ 8.. 6 L' L/£.1.....7$.--I 00 q 74 TZ* 0 £0' 0 60' O ZO' 0 s igna t ion ' 'De Des .4 '8 2 0. 5/8.. lii ti .c. 5 1.-I CI 14 ). U 4/ (0 . / LVOE .I..tit 0)) L‘) o ck• U q(U A -g.-1 "B '2.4 09' 0 L5' 0 IVO ST' 0 TE* 0 8Z' O Geometry 2 • ....T E' 8/Z' T T' 8/Z' T I' 9/E' T B' S/V' T E' 9/Z' T 1.4 0 F4 tft 0.' •./t •Zr sr 41 . 8 •-•I 2 ..4. 4) 01 0'1' 0 81' 0 9Z' O 8Z .1: 1.

.-+ -.0 O •-■ 4-J -r-i 0 0 --4 • 0 0 v ..-.) ri PR PS e-I 0 0 w0 .) cl 14-4 3-■ 0 . .0 Z 1 (11 0 m I•.c .4 Li tani .... c..--\ . \ e z (1 z f \A Ei t 'b .) 4-1 64) O 0 C) = 41 Cr) cll cl) (-7 C.-. -1 4-) 7:1 X •.) trl E +-I i-i CA O . 2 R i.) II II 0 0 0 c-I .--. .-1 z II zas 0A•0 OA 'JD cr.. (3 6 O st 0 —a-- e-XD —44— rICALLICO—B- MUS PSE(D sa—}p-÷•- EI-XD-Lld (2 `1CUJNAD -4"r- 511113Sls T al O Ill . CJ •C .j IC.w =1 I u) Cl.= 3 0 --4 Li-t 0 cn m • 0 0 CIJ cn ro • NIN 0 0 ( N l 0 °0 .21 O 'uT) NOLLWRICIIZICI DVd is moNwiaci 128 0 O clC)...1II O •in en 0 m 0 G.... 0 m m.:11 o v .. C11 0 . .7 '4-$.J ix 1 0 mi a.c1 C..-I 0 0 N rl UI O O O ------_.• ti Ls-..) z (s) cli CU E E-I 1-1 CU P. c...:-4 0 rl 0... el c...7.-.) 4-... .0 C.-4 ••I 4-1 0 •ri . .-I w . O m cs O ..-■ C..r..

_ .I' . ...gr I / 0 0 1...-..‹.. L A.....z. d) 1. 10 1 ■■■■ /2\ 5 ..C• / / f 10 A •A.. 4th SERIES —a— Gx-m GD-M --.1 ... .... .... " „.. „. 15 0 ..... EASE \\.. • 5 o 0 SUBGBADE / 3rd SERIES 15 _e_ casrpor..-.....L.. z.. \.. 129 ..i..z... / c) • — / SUBGPADE 15 .x .. t . .....C......t-a3-13 -c -ca T1 =TM cis 1= — 15 — (3)( 13 - 20 --...C. . / ..--5 i''' BASE BASE i • _ 10 ...PR-GX-B -1. ..4„.../ 5 20 15 _ SUBGPADE 3 SUBGPADE I b) 2nd SERIES I 10 a) 1st SERIES 15 ° dui —19--PR-CD-B —E3-........ ...-- 5 10 5 20 15 10 ....VERTICAL PERMANMT STRAIN ( % ) 25 20 15 5 0 10 5 0 ..GX-B GX-M PS-¢)-m 20 2 Note: PR= Prerutted GX= Gectextile M= Middle of base PS= Prestressed GCS Geogrid B= Bottom of base Figure 45.......... Variation of Vertical Permanent Strain with Depth of Pavement for All Four Test Series.

uniform conditions across all the three sections could not be maintained while the resilient response of all the sections was being measured. within one series of tests. As a result. Prestressing of the geogrid appears to reduce the development of permanent vertical strain in both the granular base and the subgrade layer. In general. Vertical Resilient Strain The variations of vertical resilient strain with depth for all the pavement sections are shown in Figure 46. the 100 to 200 passes of wheel load required to complete the recording procedure did not have a significant influence on the consistency of the results. Nevertheless. Figure 45 shows that the resilient strain profile for all the sections have a similar shape and. however. The results for the first series of tests are considered unreliable because the pavement structure deteriorated rapidly at quite an early stage of the experiment. The vertical permanent strains for the two prerutted sections are in general smaller than those in the nonprerutted sections with or without reinforcement. as shown in Figures 45a and 45b. a similar magnitude of strain. For other series of tests. it is believed that the recorded strain values shown in Figure 46a at least in the correct trends. 4. The only exception is the permanent strain developed within the prerutted sandgravel base which shows a greater value than its nonprerutted counterparts. while permanent strain at the top of the subgrade increased. 3.immediately below the geotextile. large strains were obtained at the top of both the 130 .

a.A.GX-M .M SUBGRADE • SUBGRADE 15 15 20 20 or M • Note: PR= Prerutted GX= Geotextile 11= Middle of base B= Bottom of base PS= Prestressed GD= Geogrid Figure 46.VERTICAL RESILIENT STRAIN (millistrain) • 10 Ig i 0 0 a) 15 2 0 1st SERIES _.-----'■ e 9 SUBGRADE ." GX-M F7— 41 ..PR-GD-B 4 CCUTRCL 8 2nd SERIES ' b) /// l' BASE 5 --i-.Gx-B __x__GD-B ---. A.I.7%-..C.C. 4 . Variation of Vertical Resilient Strain with Depth of Pavement for All Test Series.- 3 4 6 A. /7 . 10 10 d) 4th SERIES BASE 5 -._Gx-B . / SUBGRADE 15 c-------77' gr / / i - 15 0 h 2 0 " A.. __.R..Car2P01.T.. PR-GX-B -9--8. 6 • 1 7:- —e— CCN.. 6 b 10 _ --■ ---.C. 131 8 . ' r ____ ( PS-GD. 5 4 ) 1 20 2 6 3rd SERIES BASE .

Both prestressing and prerutting appear to reduce significantly the resilient strain at the top of the subgrade.aggregate base and subgrade. Lateral Resilient Strain Lateral resilient strains were only recorded from the strain coils installed on the geosynthetics and in the complimentary location of the control sections. Unlike the vertical resilient strain. However. respectively. Longitudinal Resilient Strain The results of the resilient longitudinal strain for the asphalt surfacing and the aggregate base are shown in Table 27 and Figure 47. In general. Beginning with the third test series they were also measured in two of the three sections at both the top and bottom of the aggregate layer. The non-reinforced control sections (with the exception of the first series of tests) normally exhibited slightly higher resilient strain than the reinforced sections. the longitudinal resilient strain varied greatly throughout the test. The lateral resilient strains recorded during the 4 test series are shown in Table 27. No consistent trend emerged regarding the effect of geosynthetic stiffness and location of the reinforcement on the measured resilient lateral strain. Longitudinal resilient strains at the bottom of the asphalt surfacing were measured for all the sections. but that in the non-reinforced control section tends to be considerably higher. for a given test series the magnitude of the resilient lateral strain in the geosynthetic reinforcement of both sections is quite similar. regardless of its location within the pavement structure. overall resilient response of the pavement sections does not seem to be significantly influenced by the geosynthetic reinforcement. Generally longitudinal resilient strain increased in the top and bottom of 132 .

1 <0N 071 1.1.N N N 0 0 N cr) U1 I".4 r. •• CI la E-4 Cl U) 41111. •--I i t0 r-i Cr) 4.1' CO N ••1 m r 0 •cr c0 .0 CV m m In 8 ul 0 tO a) •-1 ed.11"0!ammammr--- 133 .to 0 N CI1 03 111 U) 01 %.0 (Nt tO N N o-I Ln ° 8in IA 8 N 0 0 Ul 0 CD If) tO CO 0 r-I CV 6 O ttl cn tn N to O U) O O O O (4) (1) 0 0 O O 8 O O O O O to •.).

. •••-• ..1... TOP PS-GD-M. ..- ....--_-__ '1'.CNTFOL .BOTTCM= Strain. 'Li-- 0 4 10 103 IX-M.-•----=.:1P • .<. TOP 6 0 U . 134 .0 PS-GD-M..!' ■■ __-. 4:— .s.. bottom of base Figure 47... \ 2 .-1 ----31-..-.5 kip Wheel Load.. Variation of Longitudinal Resilient Strain at Top and Bottom of Granular Base with Number of Passes of 1.. BOITOM Eit . bottom of base 2) For location of strain ReasurerentTOP..8 a) 3rd SERIES r. . ..'- \ .._.0 10 ''.. - 4 .•••• N 3.rITROL . ■ . '11-... measured at top. 2.:. 5 >: IX-B.GX-M. .--- ..---__.. r- GX-M....0 10 2 10 10 10 NUMBER OF PASSES CF WHEEL LOAD Note: 1) For section designation- PS= Prestressed GX= Geotextile GD= Geogrid B= Geosynthetics placed at middle.... BC/1M1 10 0 . BOTTI:M .----.C. TOP b) 4th SERIES t. BOITCNI A-.

This could be at least partly due to the slight differences in the finished thickness of the surfacing and base and small differences in material properties. with the magnitude of vertical stress typically varying from about 6 to 9 psi (42 to 63 kN/m 2 ). Only resilient strains at the beginning of the test are shown in Table 27. For the first series of tests. Figure 49a also suggests that the inclusion of geosynthetic reinforcement at the middle of the aggregate base may result in a slower rate of increase in horizontal stress at the top of the layer. Transient Stresses The variation of transient vertical stress at the top of the subgrade during each test for all the pavement sections is shown in Figure 48. however. For resilient longitudinal strains measured within the aggregate base. The results. there did not appear to be a consistent development trend. Longitudinal. Longitudinal strain at the bottom of the asphalt surfacing also varied from one series of tests to another. as shown in Figure 49. The horizontal stress at the bottom of the aggregate base. on the other 135 . horizontal transient stress (in the direction of wheel traffic) at both the top and bottom of the aggregate base was measured in the third and fourth test series. A consistent influence of geosynthetic reinforcement on vertical subgrade stress was not observed in any of the test series. indicate that the horizontal stress at the top of the granular layer increased throughout each test.the aggregate base as the pavement started to deteriorate. The subgrade stress for the last three test series remained reasonably constant throughout the test. the subgrade stress rapidly increased as the pavement developed large permanent deformations early in the experiment.

(Tsd) rivonkran ssauas ameans 136 .

sorram 0 10 2 3 10 10 12 10 4 10 5 C b) 4th SERIES 0___----____ G„. 5 IX-M. POTTCM 1 0 10 2 10 3 1o4 10 10 6 NUMBER OF PASSES OF WHEEL LOAD Note: 1) For section designationPS= Prestressed GX= Geotextile GO= Geogrid M. bottom of base Figure 49. Variation of Transient Longitudinal Stress at Top and Bottom of Granular Base with Number of Passes of 1.---- 10 R. -M. Cr) BOTTOM -t-----4 ----- CONPROL.30 a) 3rd SERIES 25 / 20 / / / / c ..7 -r . TOP PS-GD-M.All Test Series. B= Geosynthetics placed at middle. '10P 15 3. TOP 0 8 6 .C( CLNWROL. 2 GX-M.... BOTTOM= Stressmeasured at top.---.TOP ( IX-M.-4 II-. bottom of base 2) For location of stressrreasurenent TOP. BOrICM --. 137 .------ PS-CD-M.5 kips Wheel Loads .

Single Track Supplementary Tests After performing the multiple track tests in Test Series 2 through 4. which are valid for the conditions existing in these tests. The following observations. permanent deflections in the reinforced sections were actually slightly greater throughout most of the test. These tests were conducted where wheel loads had not been previously applied during the multiple track tests. single track tests were then performed along the side of the test pavements. The single track tests performed are described in Table 24. For these tests the permanent vertical deformation in two reinforced sections and the unreinforced control section were all very similar. The single track tests consisted of passing the moving wheel load back and forth in a single wheel path. did not appear to be influenced by the progress of the test. 2. nor by the presence of a geosynthetic at the center of the layer.hand. This test series was conducted before the sections were surfaced. Placement of a geogrid at the bottom of the aggregate base did not have any beneficial influence on the performance of the unsurfaced pavement in Test Series 2 (Figure 50a). and the results of these tests are presented in Figure 50. A surfaced pavement section which has been prerutted during construction but is not reinforced can perform better than a similar section reinforced with a very stiff geotextile placed at the middle of the aggregate 138 . can be drawn from these experimental findings: 1. These special supplementary tests contributed important additional pavement response information for very little additional effort.

•)-1 0 0 0 ri o4-. U) US CU ((I CO cU II O •■•-1 11 CI.1-1 4. P.-1 Q.1 W I-1 cn O E-..--1 0 .-..) CO C..i 4-1 0 al a) 4.• 0 0 0 a) 4-1 0 0 01. RI C.1 • 0 (i) 4-4 4-I CD il II U CL) Ca (1) >4 A E-■ a) C.= •1•1 C/1 . 0.. cif 4..) S.H 4-I 4.H o z go n ii go z R U) cu 0 m o • N 0 U) /4 ri) o W IX .) o .1 cl) CQ CO cli ca P0 00 4-1 4.-I 0 0 0 0 0 0 . 0. "a 4-) PI • -1 a 4-1 CU E 0 ...-•I O 4-J O CZ I-1 CIS CU O 0. 0 in 0 0 0 ( 'U) 0 NOILWRI032E1 aDV12. 3.-I 7:1 11 Vt) 4-1 (1) CU 4. 0 0 N 0 . Ct ..) -0 r-z. 4.) . 0 . o ..) 'a 4-1 c11 0..•I CZ Z > •H . al 3.)0 0 0 CU ED E /. >.—i . II 0 z 0 N a= z 0 0 0 a) .1f1S JZZINWIdad 139 0 .11 04 Cil P.

Heaving outside of the rut was generally not observed for the sections with crushed limestone base. 140 .base. When a geogrid was placed on the subgrade. The improvement in performance is greater due to a combination of prerutting and geosynthetic reinforcement at the middle of the aggregate base than is prestressing the same geogrid at the same location within the aggregate base (Figure 50d). Placement of the very stiff geotextile at the middle of the layer did result. however. During the single track tests. upward soil migration appeared to be the dominant mechanism of contamination. no Class 1 cracks developed within the wheel track during the multi-track tests. Surface Condition and Soil Contamination Surface Condition at End of Test. With the exception of the first test series. Contamination occurred as a result of both stone penetration into the subgrade and the subgrade soil migrating upward into the base. in important reductions in rutting compared to placing the same reinforcement at the bottom of the layer. heaving along the edge was evident for the three sections of Test Series 1 using the sand-gravel base. surface cracks were observed along the shoulder of the deeper ruts. However. Contamination of the aggregate base by the silty clay subgrade was evident in most sections except those where a geotextile was placed directly on top of the subgrade. for the conditions of the test. The surface condition of the pavement sections at the end of the tests is shown in Figure 51. Soil Contamination. 3. but has not been prerutted (Figure 50b).

10 C CJ a) 711 P.cn V) cn CJ Ei 1 4-1 E.) 3-1 C C. CZ 3-1 E-1 *1-1 C JJ O CJ 4_1 4-) 0 O C2 C. 141 .

the accurate prediction of tensile strain in the bottom of the base was found to be very important. Both the laboratory and analytical studies. Hence. the elastic cross-anisotropic model was used as the primary analysis method in the sensitivity study. The resilient modulus of the subgrade was found to very rapidly increase with depth. field testing methods that measure stiffness such as the falling weight deflectometer tend not to be effective for evaluating the potential improvement due to reinforcement. A finite element model having a crossanisotropic aggregate base was found to give a slightly better prediction of tensile strain and other response variables than a nonlinear finite element model having an isotropic base. (25 to 38 mm).5 in. Extensive measurements of pavement response from this study and also a previous one were used to select the most appropriate analytical model for use in the sensitivity study. show that placing a geosynthetic reinforcement within the base of a surfaced pavement has a very small effect on the measured resilient response of the pavement.Depth of soil contamination of the base was found to be in the range of 1 to 1. In modeling a reinforced aggregate base. The low resilient modulus existing at the top of the subgrade causes a relatively large tensile strain in the bottom of the aggregate base. SUMMARY AND CONCLUSIONS Both large-scale laboratory tests and an analytical sensitivity study were performed to evaluate the performance of surfaced pavements having geosynthetic reinforcement within the unstabilized aggregate base. Hence. as well as full-scale field measurements. Larger strains cause greater forces in the geosynthetic and more effective reinforcement performance. 142 .

Reinforcement can. cause changes in radial and vertical stress in the base and upper part of the subgrade that can reduce permanent deformations and to a lessor degree fatigue in the asphalt surfacing. 143 . under the proper conditions.

_ - A - .

and resist clogging of the filter over the design life of the pavement. the geosynthetic must minimize the possibility of erosion of soil into the drainage layer.CHAPTER III SYNTHESIS OF RESULTS INTERPRETATION. At the same time.to maintain a clean interface between an aggregate layer and the underlying subgrade. 2. APPRAISAL AND APPLICATION INTRODUCTION The use of geosynthetics in pavements has dramatically increased over the last 10 years. Reinforcement . polyethylenes and polyesters. nonwoven and open grids type products manufactured from polymers such as polypropylenes. geogrids and other similar synthetic materials used in civil engineering applications. Geosynthetics are considered to include woven and nonwoven geotextiles. Separation . 3. The present study is concerned with the utilization of a geosynthetic within the unstabilized aggregate base of a surfaced. and allow the rapid dissipation of excess subgrade pore pressures caused by traffic loading. flexible pavement. Filtration .to aid in improving subsurface drainage. Geosynthetics may be included within the aggregate base of a flexible pavement structure to perform the following important functions: 1. 144 . Geosynthetics can be defined as woven.to structurally strengthen the pavement section by changing the response of the pavement to loading.

filters for longitudinal drains. These mechanisms are of considerable importance because of the many new innovations in reinforcement that will have to be evaluated in the future. For example. The emphasis of this study was placed on the reinforcement aspects of surfaced pavements.Potential geosynthetic applications in pavements not considered in this study include their use in overlays to retard reflection cracking. the use of steel reinforcement has been introduced as an alternative to geosynthetic reinforcement as the present project was being carried out. and how these changes influence overall structural fatigue and rutting performance. strain and deflection within the pavement. Variables associated with geosynthetic reinforcement of importance are shown to be geosynthetic stiffness. reinforcement of an asphalt surfacing mixture. The large-scale laboratory test track study and comprehensive theoretical sensitivity analyses both performed as a part of this investigation clearly show that the potential beneficial effects due to reinforcement decrease rapidly as the strength of the subgrade and overall structural strength increases. and in the repair of potholes and for other maintenance operations. Probably the greatest effect upon performance due to reinforcement is the change in lateral stress. 145 . Some emphasis is placed on developing an understanding of the fundamental mechanisms of improvement of geosynthetic reinforcement. overall strength of the pavement section and strength of the subgrade. particularly in the subgrade. Relatively little is presently known about the influence of geosynthetic reinforcement on pavement response. This influence can be expressed as changes in stress.

The separation function is shown to be relatively easily achieved using a wide range of geosynthetics. 146 . the numerous findings of all portions of the study are interpreted and appraised considering other available experimental results. However. Where a higher degree of modeling accuracy is required. Finally.Both the separation and filtration mechanisms of geosynthetics are analyzed relying mainly upon the existing literature as a part of the general synthesis of the use of geosynthetics within aggregate base layers. separation and filtration functions of geosynthetics used in aggregate bases. For reinforcement to be effective. and design recommendations are presented. GEOSYNTHETIC REINFORCEMENT The response of a surfaced pavement having an aggregate base reinforced with a geosynthetic is a complicated engineering mechanics problem. The important advantage of using a simplified linear elastic model of this type is the relative ease with which an analysis can be performed of a pavement structure. The filtration function is shown to be quite complicated. As will be demonstrated subsequently. a linear elastic. the present state-of-the-art of durability aspects are considered. cross-anisotropic finite element model can be successfully used to model geosynthetic reinforcement of a pavement structure. Therefore. because of its great importance. it must be sufficiently durable to serve its intended function for the design life of the facility. and put in perspective from the standpoint of reinforcement. with performance depending upon a number of important variables. analyses can be performed on pavement structures of this type using theoretical approaches similar to those employed for non-reinforced pavements but adapted to the problem of reinforcement.

For example. Geosynthetic stiffness should be used since the modulus of elasticity of a thin geosynthetic has relatively little meaning unless its thickness is taken into consideration. Under certain conditions it is an important consideration in insuring the success of an installation. as will be discussed later. and also the effect that prestressing the geosynthetic has on the behavior of the pavement. Geosynthetic stiffness S g is equivalent to the modulus of elasticity of the geosynthetic times its average thickness. the geosynthetic strength and ductility are important factors when used as a filter layer between open-graded drainage layer consisting of large. accumulation of permanent deformation. Use of a finite element analysis gives reasonable accuracy in modeling a number of important aspects of the problem including slack in the geosynthetic. angular aggregate and a soft subgrade. the first step should be to establish the stiffness of the geosynthetic to be used. Use of the grab type tension test to 147 . In evaluating potential benefits of reinforcing an aggregate base. The ultimate strength of a geosynthetic plays at most a very minor role in determining reinforcement effectiveness of a geosynthetic. The stiffness of a relatively thin geotextile can be determined in the laboratory by a uniaxial extension test. slip between the geosynthetic and the surrounding material. This does not imply that the strength of the geosynthetic is not of concern. The wide width tension test as specified by ASTM 61-201 (tentative) is the most suitable test at the present time to evaluate stiffness. GEOSYNTHETIC STIFFNESS The stiffness of the geosynthetic is the most important variable associated with base reinforcement that can be readily controlled.a more sophisticated but time consuming nonlinear finite element analysis can be employed.

S. the stiffness of the geosynthetic must be presented for a specific value of strain. be defined as the uniformly applied axial stretching force F (per unit width of the geosynthetic) divided by the resulting axial strain in the geosynthetic. a very low stiffness and also a low stiffness geosynthetic does not have the ability to cause any significant change in stress within the pavement and hence is not suitable for use as a reinforcement. the geosynthetic should be stiff to very stiff.30 to $0. but not all. (1 kN/m) and costs about $0. with in general S g > 1500 lbs/in. This value of strain has been employed for example by the U. geosynthetics the stiffness decreases as the strain level increases. Several selected geosynthetic stress-strain curves are shown in Figure 53 for comparison.evaluate geotextile stiffness is not recommended.59/yd 2 ). For low deformation pavement structural reinforcement applications. Since many geosynthetics give a quite nonlinear load-deformation response. For most. Classification System. This table includes typical ranges of other properties and also approximate cost. Let the secant geosynthetic stiffness S g . for at least low deformation conditions. A strain level of 5 percent has gained some degree of acceptance. as shown in Figure 52. As discussed later. A very low stiffness geosynthetic has a secant modulus at 5 percent strain of less than 800 lb/in.36-0.50/yd 2 (0.8 kN/m). A geosynthetic classification based on stiffness for reinforcement of aggregate bases is shown in Table 28. 148 . Army Corps of Engineers in reinforcement specifications. (1. Use of a 5 percent strain level is generally conservative for flexible pavement reinforcement applications that involve low permanent deformations usually associated with surfaced pavements.

) 400 0. = 4.)/Strain 2 Strain = LL/L Axial Strain in Percent./in.04 0.24 Ceosynthetic Strain.5 N Axial Force F(lbs.C / .) Secant S = F 5% 2Z g F 5Z Strain L o . t_ Figure 53./in.A sz.16 0.0 Yield . / wu 4. o Figure 52. Selected Geosynthetic Stress-Strain Relationships. 149 .08 0.12 0. &. //// — Geosynthetic Stiffness: S = F(lbs. 500 Lo a d( l bs / in.20 0.Note: 1 in. Basic Idealized Definitions of Geosynthetic Stiffness. = 25 mm 1 lb..

30-0. The properties given in addition to stiffness are typical ranges of manufacturers properties and do not indicate a material specification./in. s (2) g (lbs./in.50-3. Alternately a 57.00-$7. 150 A .50 1500-4000 20-400 85-1000 10-35 0.50 Low 800-1500 15-50 60-200 10-60 0.) Failure Elongation (% Initial Length) Typical Cost Range ($/yd 2 ) Very Low < 800 10-30 50-150 10-100 0.40-0.Table 28 Tentative Stiffness Classification of Geosynthetic for Base Reinforcement of Surfaced Pavements (1) Stiffness Description Secant Stiffness @ 2% Strain.) Elastic Limit (lbs. secant modulus could be used./in.00 4000-6500 > 300 Stiff /ery Stiff 350-500 (or more) - 5-15 $3.) Tensile Strength (lbs.00 NOTES: 1. 2.

the actual changes in stress. however.0) resting on a soft CBR = 3 subgrade. develops relatively low tensile strain in the aggregate base and hence low geoysnthetic forces. As tensile strain in the aggregate base increases. The changes in response of the pavement are for the most part determined by the tensile strain developed in the geosynthetic. the theory used to model the pavement must be able to predict reasonably well the compressive and tensile strains in the aggregate base in addition to the conventionally used response parameters including tensile strain in the bottom of the asphalt concrete.51. This increase in geosynthetic force continues to become greater until either the full loading is reached. Consider the behavior of a surfaced pavement as a wheel loading is applied. the force developed in the geosynthetic also increases and the beneficial effects of reinforcement become greater. strain and deflection in pavements of usual strength are relatively small due to the development of small tensile strain. 151 .REINFORCEMENT MODELING Problems associated with modeling the behavior of aggregate bases which can take only limited tension are well-known [16.5 to 3. For an aggregate base reinforced with a geosynthetic the problem is even more complicated due to the presence of a stiff to very stiff geosynthetic inclusion which acts as an abrupt discontinuity. Hence. A surfaced flexible pavement of low to moderate structural strength (AASHTO structural number SN 1 2.62].46-49. As a result of geosynthetic reinforcement. or else slip occurs between the geosynthetic and the materials in which it is sandwiched. Modeling. vertical compressive strain on the subgrade and overall surface pavement deflection.

and also the commonly used response parameters such as tensile strain in the bottom of the asphalt. As a result. 152 .45] were used involving fully instrumented. These tests were previously described in Chapter II. First the results of an earlier study by Barksdale and Todres [44. The model can predict reasonably well the measured strain state in the aggregate base. This study was accomplished by directly comparing the analytically calculated pavement response with that observed in well instrumented pavement sections having aggregate bases. After the experimental findings of the present study became available. Cross-Anisotropic Model. full-scale pavements constructed in the laboratory. it became apparent that problems existed in the conventionally used analytical models as presently applied in predicting the response of geosynthetic reinforced pavements. a special study was undertaken to verify the analytical techniques including material properties used to model geosynthetic reinforced pavements. Measured vertical and horizontal strains from both of the well-instrumented laboratory studies clearly indicate the aggregate base performs much stiffer in the vertical direction than in the horizontal direction. strain and deflection measurement data were used in developing suitable analytical models.As the project progressed. These large scale laboratory tests involved applying a moving wheel loading to both geosynthetic reinforced and non-reinforced pavement sections. Therefore. These results can only be explained if the aggregate base behaves as a cross-anisotropic solid. and vertical compressive strain on top of the subgrade. the extensive stress. cross-anisotropic finite element model appears to give the best overall predictions of pavement response (Tables 2 and 4). a linear elastic.

respectively. In the upper one-third of the base the horizontal resilient modulus was taken to be about 80 percent of the vertical modulus. of the average resilient subgrade modulus as shown in Figure 54.The best agreement with observed response was found for a crossanisotropic model whose vertical resilient moduli of the base became about 40 percent smaller in going from the upper one-third to the lower one-third of the aggregate base. the resilient subgrade modulus near the surface appeared to be about 10 and 20 percent. isotropic base underpredicted the base tensile strain by about 30 to 40 percent. the horizontal modulus decreased to about 3 percent of the vertical resilient modulus at that location (refer to Tables 3 and 5). Linear elastic models having a homogeneous. This was true for either the cross-anisotropic model or the nonlinear finite element models. and hence shows much better agreement with observed pavement response (Table 2). These increases in subgrade resilient 153 . In the lower one-third of the aggregate base. Use of a subgrade whose resilient modulus increases significantly with depth greatly increases calculated tensile strains in the aggregate base. A linear elastic cross-anisotropic model having a homogeneous. For the micaceous silty sand and silty clay subgrades used in the validation studies. isotropic subgrade tended to underpredict tensile strain in the bottom of the aggregate base by a factor of about three. best agreement was found when the model became progressively more cross-anisotropic with depth. Nonlinear models employing typical resilient subgrade properties of the type shown in Figure 6 also under-predicted tensile strain. the resilient modulus of the soft silty clay subgrade apparently did not increase as much as that of the micaceous silty sand subgrade. As expected. Also.

Z ( IN) -10 T Fla (•ax) ///3.-7/1/3:-7/ TOP SUBCRADE (50%) (39. 3E4 2E4 3E4 RESILIENT SUBGRADE MODULUS. 1E4 5000 • I da 2E4 .C. Es =3. Reduction in Response Variable As A Fuction of Base Thickness.5 KSI 20 cr BOTTOM BASE z LL1 0 ec z 0 ev SUBGRADE 10 0 a ■ • dv SURFACE Er A. T (INCHES) Figure 55./// fll (27%) SU BGRAD EDEPTH. 30 2. —50 0 . /. Es (KSI) Figure 54. 154 16 . .5 IN A. Variation of Subgrade Resilient Modulus With Depth Estimated From Test Results.310 —20 MICACEOUS SILTY SAND Nog —30 (160.C. 0 5 10 12 14 BASE THICKNESS.■ SILTY CLAY RIGID RIGID ./ .4%) -.

The cross-anisotropic model appears to give slightly better results and was more economical to 155 . (1. however. be made to predict reasonably well the tensile strain in the aggregate base.modulus are quite large. and the small measured vertical resilient strain observed throughout the aggregate layer. Nonlinear Isotropic Model. aggregate base quality and permanent deformation.2 m) in thickness. The isotropic nonlinear analysis cannot.49] appeared to give better results than the often used K-O type model for the aggregate base. predict at the same time both the large tensile strain measured in the bottom of the aggregate base. upon proper selection of material parameters. particularly considering that the subgrades were only about 4 ft. Summary. and also the commonly used response parameters. However. Use of a simplified contour model [48. Reasonably good response was obtained using both the linear cross-anisotropic model and the nonlinear model. it is believed not to be a dominant factor effecting the increase in modulus with depth. A discussion of the increase in resilient modulus with depth has been given by Brown and Dawson [89]. A nonlinear isotropic model was used in the sensitivity study primarily to investigate the effect of special variables such as geosynthetic slip. The nonlinear. The rigid layer which was located below the subgrade in the instrumented studies may have had some influence on performance. The nonlinear resilient modulus was therefore adjusted to approximately agree with the variation of modulus with depth shown in Figure 54. the nonlinear analysis underpredicted vertical strain in the subgrade. isotropic finite element model used can. When the nonlinear properties originally selected for the subgrade were employed.

30] also confirm this finding. The analytical results shown in Figure 55 (and also in Tables 9 through 11 of Chapter II) indicate radial strain in the asphalt surfacing and surface deflection are changed usually less than 5 percent. some usually modest improvement can be derived from reinforcement following the commonly employed design approaches of limiting vertical subgrade strain and radial tensile strain in the asphalt. be sufficiently accurate to give acceptable results that can show the potential relative effect of aggregate base reinforcement. Field measurements by Ruddock. Pavement response is defined in terms of the resilient stresses. Considerable progress was made in this study in developing appropriate techniques to model both reinforced and nonreinforced aggregate bases. however.use.000 equivalent 18 kip (80 kN) single axle loads results in relatively small changes in the resilient response of the pavement. Even though the changes in response are relatively small. et al. Therefore it was the primary method of analyses employed in the sensitivity study. and vertical subgrade strain less than 10 percent. This level of change holds true even for relatively light structural sections placed on a soft subgrade and reinforced with a very stiff geosynthetic having S g = 4000 lbs/in. (7 kN/m). The analytical models and material properties used in the sensitivity study should. strains and displacements caused by the applied loadings. IMPROVEMENT MECHANISMS The analytical and experimental results show that placement of a stiff to very stiff geosynthetic in the aggregate base of a surfaced pavement designed for more than about 200. 156 . Better models could probably now be developed using the results of this study. [21. Specific benefits resulting from reinforcement using these criteria are discussed later.

The experimental measurements show the strain in the geosynthetic to be on the order of one-half the corresponding 157 .Pavement Stiffness The structural strength of a pavement section is frequently evaluated using the falling weight deflectometer (FWD) or Dynaflect devices.38. The laboratory test results also indicate no observable improvement in pavement stiffness. even when a very stiff geosynthetic is used as reinforcement. Both the laboratory and analytical results indicate the change in radial stress and strain as a result of base reinforcement to probably be the most important single factor contributing to improved pavement performance. Variations in pavement thickness and/or material quality including subgrade stiffness could account for the difference in overall pavement stiffness observed for the one series of tests in Texas. The analytical results of this study indicate the overall increase in stiffness of the pavement will be increased less than about three percent.39]. The improvement in stiffness resulting from geosynthetic reinforcement is therefore too small to reliably measure in either a full-scale or laboratory pavement. Radial Stress and Strain. These findings therefore indicate stiffness is a poor indicator of the potential benefit of geosynthetic reinforcement on performance. while another indicated no observable difference. Dynaflect measurements in Texas described by Scullion and Chou [63] showed one section to be stiffened when a geosynthetic is added.30. These devices measure the deflection basin and the overall stiffness of the pavement [62]. The overall stiffness of a structural section can be defined as the force applied from a loading device such as a falling weight deflectometer (FWD) divided by the resulting deflection. The results of several field studies also tend to substantiate this finding [21.

the actual value of change in radial stress is relatively small. As the pavement section becomes moderately strong (structural number SN Z. 4. Considering just the large percent change in radial stress. As a result. First.7 kN/m 2 ) as shown in Table 10. Recall that tension is positive so the decrease in stress shown in Figure 56 actually means an increase in confinement. 158 .1 psi (0. or even more. must be superimposed upon the initial stresses resulting from body weight and compaction effects as illustrated in Figure 57.strain in a non-reinforced aggregate base (Table 27).0 psi (3-7 kN/m 2 ) for relatively light sections. The analytical studies performed on stronger sections indicate changes in radial strain in the bottom of the base to be about 4 to 20 percent for sections having low to moderate structural numbers. The initial stress in the base is likely to be at least twice as large. Changes in radial stress determined from the analytical study typically vary from about 10 percent to more than 100 percent of the corresponding radial stress developed in an unreinforced section (Figure 56). typically being less than about 0. does not give the full picture of the potential beneficial effect of reinforcement. the radial stresses. the changes in radial stress typically become less than about 0. the strain in the geosynthetic becomes greater. the beneficial effects of changes in radial stress caused by reinforcement are reduced but not eliminated. Secondly. As a result improvement also becomes greater. As the resilient modulus of the subgrade and the ratio between the base modulus and subgrade modulus decreases. than the radial stress caused by the external loading. however. however. including the relatively small changes resulting from reinforcement.5 to 1.5).

T (INCHES) Figure 56.0 —1 0 cn n < Lu (X x a 0 0 —0.0 0. Wheel Load Wheel Load Da a3 &Jr 0 vo y ///1_--. 159 .v zh° (a) Weight and Compaction Stress (b) Wheel Load Stresses (c) Combined Stress Figure 57. Variation of Radial Stress in Base and Subgrade With Base Thickness. // Ko.1. Superposition of Initial Stress and Stress Change Due to Loading.5 1.5 16 14 12 10 BASE THICKNESS.

[16] on unsurfaced pavements tend to verify that the depth 160 . The small beneficial changes in radial stress due to reinforcement can have under the proper conditions important effects on rutting permanent deformation. was relatively light. the depth of reduction in permanent strain in the subgrade was about 12 in. The addition of reinforcement causes a small but important change in radial stress. were only changed to a depth of about 6 to 7 in. as the stress state becomes less severe. The tire loading in this case. the beneficial effect of reinforcement becomes disproportionately less.Permanent Deformation. Because of the highly nonlinear stress-permanent strain response of the subgrade or base (Figure 58). et al. (150-180 mm) below the surface of the subgrade. If the initial stress state is near failure. As a result both the confining pressure and deviator stress on an element of subgrade soil are decreased slightly. Conversely. For the heavy load used in the analytical study. the top of the subgrade. (300 mm). Findings by Barksdale. Depth of Subgrade Improvement. a small increase in confining pressure and decrease in deviator stress can lead to a significant reduction in permanent deformation when near failure. The largest beneficial effects by far are realized when the stress state is close to failure on an element of material in. when reduced. The laboratory studies indicate both resilient and permanent strains in the subgrade. The reduction in permanent deformation becomes disproportionately larger as the stress state in the top of the subgrade (or bottom of the base) moves closer to failure. very important reductions in permanent deformation can occur as illustrated in Figure 58. and also a slight reduction in vertical stress. for example. When examining Figure 58 remember that permanent deformation is proportional to the permanent strain developed in a thin sublayer of material. however.

E Figure 58.Note: AG = change in radial stress r due to reinforcement Permanent Strain. 161 . Reduction in Permanent Deformation Due to Geosynthetic for Soil Near Failure.

The greatest beneficial effect of reinforcement appears to be due to changes in radial stress and strain together with small reductions of vertical stress in the aggregate base and on top of the subgrade. As a result. strain and deflections are all relatively small for pavements designed to carry more than about 200. however. Summary The effect of geosynthetic reinforcement on stress.5 to 3) on a weak subgrade (CBR < 3) potentially can significantly reduce the permanent deformations in the subgrade and/or the aggregate base. In this comparison a geotextile reinforcement was located in the middle of the base.of improvement in the subgrade is relatively shallow. the difference in tensile strain due to reinforcement appeared to disappear and eventually the tensile strain in the nonreinforced sections was less than in the reinforced one. Reinforcement of a thin pavement (SN = 2. A modest improvement in fatigue life can be gained from reinforcement as discussed subsequently. geosynthetic reinforcement of an aggregate base will have relatively little effect on overall pavement stiffness. The changes in radial stresses appear to be due to the reduction in tensile strain caused in the lower part of the aggregate base. Tensile Strain Variation with Load Repetitions. With increasing numbers of load repetitions. the 162 . Strain measurements made in the third test series of the experimental study show at low load repetitions a very large reduction in tensile strain in the bottom of the aggregate base due to reinforcement. the state of stress in the aggregate base and the subgrade moves away from failure. As a result. As the strength of the pavement section increases and/or the materials become stronger.000 equivalent 18 kip (80 kN) single axle loads.

The influence of reinforcement on the required pavement thickness is studied considering both fatigue and permanent deformation (rutting) mechanisms. 1.3 kN/m). This range of subgrade stiffness approximately corresponds to a variation of CBR from 3 to 10.2-7.500 psi (86 MN/m 2 ). REINFORCEMENT EFFECTS The primary factors associated with aggregate base reinforcement are discussed including their interaction with each other and the overall pavement. Geosynthetic reinforcement levels included in the analytical sensitivity study vary from low to high stiffness (S g = 1000 to 6000 lbs/in. The layer strain method and the permanent strain materials properties employed in the analysis are described in Chapter II. Alternate thicknesses are given from the analytical sensitivity study for subgrade strengths varying from a resilient modulus of 3500 psi (24 kN/m 2 ) to 12.improvement caused by reinforcement would be expected to rapidly become small. Effects of reinforcement on permanent deformations that might occur in the base are also considered. In the analytical sensitivity study the reduction in aggregate base thickness as a result of geosynthetic reinforcement was determined using an equal strain approach for controlling fatigue and rutting. 163 . A similar procedure was employed to determine the reduction in base thickness for equal vertical strain near the top of the subgrade. An estimate of reduction in rutting in the aggregate base and subgrade was also made using the layer strain method.. and a number of practical aspects are examined such as slack and slip of the geosynthetic. A reduction in base thickness due to reinforcement was established by requiring the reinforced section to have the same tensile strain in the bottom of the asphalt surfacing as the non-reinforced section.

A geogrid. Permanent Deformation .070 load repetitions were applied by a 1. When placed in the bottom of the aggregate base. (30 mm) thick. The results of Test Series 2 also tends to verify these findings. This test series was terminated when the total permanent deformation reached about 1 in.5 kip (6. 5 kN/m).7 kN) wheel load. A 52 percent reduction in permanent subgrade deformation did. In contrast. A stiff to very stiff woven geotextile was used (S g = 4300 lb/in. were reduced by only 14 percent. when the same geotextile was placed in the middle of the aggregate base. when placed in the bottom of the base. Subgrade permanent deformations..Experimental Findings. Test Series 3 was constructed using a stiff asphalt surfacing mix 1. (200 mm) crushed limestone base. however. and at the center of the base in another section. 164 . and an 8 in. permanent deformation in the base was reduced by 31 percent. however. did not decrease the permanent deformation in the base (measurements suggested an increase of 5 percent). (25 mm). but only a 3 percent reduction of permanent deformation in the aggregate base (Table 25). A total of 100. The geotextile was placed at the bottom of the base in one section. The experimental findings of Test Series 3 nicely demonstrates the effect of position on performance with respect to permanent deformation.2 in.Optimum Geosynthetic Position The laboratory pavement tests together with the results of the analytical sensitivity study can be used to establish the optimum positions for placement of geosynthetic reinforcement within an aggregate base. apparently occur. A control section without reinforcement was also present. the stiff to very stiff geotextile caused a 57 percent reduction in permanent deformation in the subgrade.

The agreement between the observed and calculated reductions in rutting are reasonably good. 30 and also other tables and figures in this chapter frequently express improvement due to reinforcement in terms of a reduction in base thickness. Geotextile reinforcement positions were at the bottom and center of the layer. (30 mm) compared to 2.Analytical Results. however. Tables 29. The results of this analytical study for the standard reference section having a 2. poorer than for standard reference sections. Conversely.Permanent Deformation . (64 mm) for the analytically developed relations shown in the figures. the reduction of permanent deformation in the base becomes much greater as the reinforcement is moved upward in the base (Figure 60). Also. The reduction in subgrade deformation gradually goes from about 45 percent to 10 percent as the geosynthetic location goes from the bottom of the base to a location 2/3 up from the bottom.5 in. the asphalt thickness of the experimental sections were only 1.5 in. expressed of course as a decimal. 165 . 24 MN/m 2 ) are summarized in Figures 59 and 60. The actual reduction in base thickness would be equal to the base thickness without reinforcement indicated in the table or figure multiplied by the percent reduction. In Figures 59 and 60 the bold solid symbols indicate observed reductions in rutting from the previously described Test Series 3 experiment. Material properties of the test sections were. An analytical study was also performed to establish the effect of geosynthetic position on the reduction in rutting in the base and subgrade (Tables 29 and 30). The maximum measured reductions in rutting are greater than calculated values for similar pavement base thicknesses. (64 mm) thick asphalt surfacing and a relatively soft subgrade (E s = 3500 psi.2 in.

.. 8I ZZZI - I SL* 6 J 6' I I SL . • • I7. 7 ------ .-.1 I .I 5' 9- I GEOSYNTHETIC STIFFNESS.I —.I 97I ZZ - 12 GEOSYN.8' 1 - II f SURFACINi ... ZI . CHANGE IN BASE THICKNESS ( 2) CONSTANT TENSIL E E L 8Z- I CZ- 11 I0009 ...-4 = = L.1 . .. ) ---------------------- Z I. • C. 5 -7... c ■ CZSi- 000 7 0 I. POSITION I- I "I • . 3 0 ... 5 Z- CO- FTI7 - - - r..: a.7 . LA V7 0009 < . 6 eBOTTOM LI - Z1- Z6 .' GEOSYN II RI - z i- 7... II t •S I . 8 .....2. 'A DS NI VI/ 3SV8NOOd NI39 NVHD Z .7.4. 5 IN..3. II .) . '. - 8J - f t- I I • 166 2 ' • . r ' 61- IN • a 4.■•I -.) SI S . AC SURFACING 35 00psi SUBGRADE RI - ZI- 9I-7 pmmommy 6£ - ZI- i 01... 97—TZT-2. • LIZ Y I 30V11941 05 O NY 3SV8 I • -- ( • • • !- Ch a . Sg (l bs /in.

30•^4 1:0= •••'' CC Q.) . 0 r`I v.' CC C crl C O O O Co en 7-1 CC 111 1.0 1-1 C 0 0 O Co C. O O O O C CU +Z..▪ ▪• ▪ O O O IY 7 O O O rn 7 INRUT T I NGO F BAS EAND 1..■•••- In 0 Co C z rn PS ISU BCRA I • O O O Cn O 0.f) . Cr.. N CC tri O C C. C CC CC Cl CC N. N. O CO Co • V1 1.1 f. N 4 lrl CO ni 1.• CI 2 C r.-1 •■•• • 0 7 0 O a.1. 0 @BOTTO M 6. O INBA SETHI CKN ESS z rl O C 0 C C z O O O E O C O O C • O •-■ co ti C E z NO z Co 0 0 0 C C O C 1 0 M r- C i• VI C. • O O O CO CO U... C C U C C.) C C. csa 167 C C .

5 IN A.2 IN A.2 IN A.0 12.0 ./8 IN BASE S9 -4000 LBS/IN • LAB S9 AT BOTTOM 1.0 16./8 IN BASE 59 =4000 LBS/IN z 50 w IA C.5 KSI S9 =4000 LBS/IN POOR BASE • GEOTEXTILE AT MIDDLE ■GEOTEXTILE AT BOTTOM 40 • 30 - S 9 POSMON O F_ 0 < 20 rx Er 0 IA 10 CI 0 60 8.2 IN A.0 14.0 12.C.C.- 60 • w 2. 60 TEST SECTION 1.) cn < Co 0Z ic 0 I= 0 Es = 3.0 BASE THICKNESS. 3. 168 16. Reduction in Subgrade Permanent Deformation.5 KSI S9 =4000 LBS/IN POOR BASE S 9 POSMON Es = BOTTOM Lc/ CO Lu 40 1/3 UP Z Z 0 ° 20 cr w ix LILA 0 2/3 UP 0 ■ LAB S9 AT MIDDLE 1.C.0 BASE THICKNESS.0 10. T (INCHES) Figure 60. T (INCHES) Figure 59./8 IN BASE S g =4000 LBS/IN —20 60 8.C.0 10. Reduction in Base Permanent Deformation.0 14.

the measurements from Test Series 2 showed the strain in the bottom of the layer to be higher due to the placement of a stiff geogrid at the bottom of the layer. These small changes in tensile strain. et al. specific material properties and loading conditions. however. If only fatigue is of concern. A greater beneficial effect will also be realized for this higher location of reinforcement with respect to fatigue of the asphalt surfacing. The maximum calculated changes in tensile strain in the asphalt were less than about 3 percent. In contrast. The analytical results (Table 29) show from a fatigue standpoint that placing the reinforcement 1/3 to 2/3 up in the base is better than at the bottom. The optimum position of the geosynthetic with respect to minimizing permanent deformation depends upon the strength of the section. or perhaps as high as 2/3 up as indicated by the analytical study. 169 . the optimum reinforcement position is near the middle of the base. Summary. [41] did indicate an extremely stiff steel mesh reinforcement placed at the top of the aggregate base can under certain conditions reduce tensile strains by about 18 percent. Consideration should be given to placing the reinforcement at this location where low quality aggregate bases are used known to have rutting problems. cause reductions in required base thickness of up to about 20 percent (Table 29) for light pavements on a subgrade having a low resilient modulus E s = 3500 psi (24 MN/m 2 ). the reinforcement should be placed at the top of the base.Fatigue. To minimize rutting in the aggregate base. Full-scale measurements made by van Grup. It is hard to tell if the analytically calculated reductions in strain in the bottom of the asphalt surfacing are valid. Strain measurements from Test Series 3 indicate that placement of a stiff to very stiff geotextile in the middle of the aggregate base reduced the tensile strain by about 26 percent.

The purpose of this reinforcement would be to reduce rutting within a soft subgrade typically having a CBB<3. Base Quality Use of a low quality base can cause a significant reduction in the level of pavement performance due to increased permanent deformation and surface fatigue as a result of a lower resilient modulus. A low quality base might be caused by achieving a compaction level less than 100 percent of AASHTO T-180 density. but it is also frost susceptible [74]. probably reinforcement. Nevertheless. Use of a high fines content base cannot only result in rutting and fatigue. The reduction in tensile strain in general should be considerably less for full size sections than the 26 percent reduction observed for Test Series 3.The analytical results indicate that when high quality base materials and good construction practices are employed. The analytical results indicate to minimize fatigue cracking of the asphalt surfacing. Low quality aggregate bases would include those having a fines content greater than about 8 percent and also gravels. even small reductions in tensile strain in the bottom of the asphalt can cause disproportionately large reductions in required aggregate base thickness. the reinforcement should be placed somewhere between the middle and the top of the layer. or by using low quality materials. Both the laboratory tests and the analytical study indicates placement of the reinforcement at the bottom of the layer should be most effective where a soft subgrade is encountered. particularly if it is known to have problems with rutting. Also. reductions in tensile strain indicated by the analytical theory due to reinforcement might not be as great as actually occur in the pavement. should be placed in the bottom of the base. 170 . sand-gravels and soil-aggregate mixtures. if used.

placement of the same reinforcement at the center of the base resulted in a 31 percent reduction of permanent deformation within the base. In contrast. confirming it was a higher quality pavement than used in the first series.000 load repetitions. 5 kN/m) at the bottom of the base had a corresponding permanent deformation of only 0.35 in. The pavement of Test Series 3 withstood about 100.4 in. permanent deformation within the base was reduced by only 3 percent compared to 50 percent for the lower quality pavement of Test Series 1. When the very stiff geosynthetic reinforcement was placed at the bottom of the base. geosynthetic reinforcement can reduce base rutting up to about 50 percent as observed in Test Series 1.Observed Test Section Improvements. (9 mm). At this time the base of the control section without reinforcement had a permanent deformation of 0. (150 mm) thick sand-gravel base. The same very stiff geotextile was used in Test Series 3 as for Test Series 1. As previously discussed. the reduction in base rutting was also about 50 percent. Analytical Results. (18 mm). (200 mm) crushed limestone base.. the sections included in Test Series 3 were considerably stronger than the first series. Thus for under-designed sections having low quality bases.69 in. the ratio of the average resilient modulus of the base to that of the subgrade (Eb/E s ) is on the average about 171 . The companion section having a very stiff geotextile (S g = 4300 lbs/in. Results of a nonlinear finite element analysis indicate that for low quality bases. and had an asphalt surfacing rather than the rolled asphalt used in the first series. Test Series 3 sections had a thicker 8 in. (36 mm) bituminous surfacing and 6 in. Of interest is the finding that at about one-half of the termination rut depth. The pavement used in Test Series 1 had a 1. The pavement failed after about 1262 wheel repetitions (Table 25).

The experimental findings. and develop a better understanding of the. The analytical results indicate little change in permanent deformation developed in the base with position of the geosynthetic. Use of a geosynthetic reinforced low-quality aggregate base results in about 3 times greater reduction in actual permanent displacement (expressed in inches) in the base than for a high quality base. Both a high quality base (indicated in the tables as a "good" base). this was done simply to extend the results. however. Therefore. reductions in rutting in the light reference pavement previously described were developed for both of the above values of modular ratios (Table 31). Strictly speaking.5 for high quality materials for the sections studied. 172 .45 suitable for lower quality base pavements was only given in Table 31. influence of reinforcement on permanent deformation. show reinforcement at the middle of the base to be most effective and is preferred to reduce base rutting.5. A complete description of the layer strain approach and the permanent strain material properties were previously given in Chapter II. the lower quality base properties should probably not have been used with the stress states obtained from analyses for Eb/E s = 2. The stress state within the pavement was first calculated using the cross-anisotropic analysis and these modular ratios. Calculated permanent deformations are given in Tables 29 and 30 for both the poor and good bases for a modular ratio Eb/E s = 2. and a low quality base (indicated as a "poor" base) were included in the layer-strain analyses (Table 31).1. The results for a lower modular ratio Eb/E s = 1. The layer strain approach was then employed together with appropriate permanent strain properties to calculate permanent deformations.45 compared to about 2.5.

090. 06 0. 03 .36 -6 7 1 1 2. 3. 50. D EF. 03 . 20. in. 0. S = 6000 E /E S= 4000lbs / in. 3 Poo r 2.1 7.2 8 2 -2 -9 4 -8. 3 9. / lbs BASE DEF.2. 1 1 .1 1. REDUCT IONINRUTT ING (PERCENT ) TOTALBASESUBG.1 6 . 50.38. 50.109. 3. 1 8 1 2 -41. 5 1 0. 1 1 Po or . 03 -8. 230. 23 36 29 6 1 0. 7 5 Poo r 2. 8. 03 2.. (in.2. 50.5. lbs / in. Poo r 2. 1 40.5 52. 5 2 1 1 2 0. 06 -6 38 -4 -15 Good1 5. 03 5 1 5.200. 4 50. 1 40. 3.4 2-8.362. 7 5 2 30. 15. Goo d1 5.2 Poo r .9. 1 1 . 0 3 -8. ) ( in.2 1.1. 3. 1 1 . 50. 50.2 8 -8 5 0. 03 . 1 5. 030. 1 40. 03 1 5. 1 40. 0. ) BAS ETH IC KS =1000 rade Subg se Ba Q UA L ITYT ade Subgr Ba s e Subg r ade ( in.47. ) Base REGAT E BASE GEO SYN THETI CAT BOTTOMOFAGG 0. 3 9. 50. M M o-I M r••1 •■•1 0 0 • • • 0 0 0 173 .6. 060.50 -6 6 0. 7 5 Poor 2. 3 1 47 1 0 . ) b s ( in. 1 1 . 7 M GEOSY NTHET IC2 / 3UP FROM BOTTO 0. 3.3. 75 Po o r 2. 4 5 Goo d 0. 1 1 . 50.4 Po o r 1-4 L. 0 7 1. 1 20.10.0 so O -I 0 0 0 0 0 0 0 0 .9 13 6 1.1 Poo r -482.3. 7 5.5 5 5 1 0 2 Good0.5.1 62. DEF. 0 6 1 1 2.1 0 1. 5 Poor 2. 0 1 2.2 6. 450. 50.4 8 7 3. 1 20. 1 3 34 7 9. 4 M GEOS YNTHE TIC1 / 3UPFROMBOTTO 0.22. 5 . 1 20. 105 .

(1. The pavement sections given in Figures 61 and 62 have an asphalt surface thickness of 2.2-7. The equal vertical subgrade strain analytical approach (Figure 61) indicates that allowable reductions in base thickness for this design would increase from about 3 to 16 percent as the geosynthetic stiffness increases from 1000 to 6000 lbs/in. (64 mm) and a subgrade with a resilient modulus of 3500 psi (24 MN/m 2 ) corresponding to a CBR of about 3. This would be true even for the relatively light structural sections shown in Figures 61 and 62. (4.3 kN/m).9 kN/m). (250-390 mm). For these conditions an AASHTO design for 200. The experimental results suggest the levels of improvement in rutting shown on Figure 62 might be too high for the pavement section used in the comparison.5 in. Structural Strength The beneficial effect of reinforcement in terms of reduction in base thickness and rutting was found to decrease as the overall base thickness 174 .75 to 15. the rate of change in improvement with increasing stiffness appears to decrease.Geosynthetic Stiffness The analytical results indicate that geosynthetic stiffness has an important effect upon the level of improvement as shown in Figures 61-62 (refer also to Tables 29 and 30). For stiffnesses greater than about 4000 lbs/in.. Base thicknesses varied from 9.000 equivalent 18 kip (80 kN) single axle loads (ESAL t s) would have a base thickness of about 12 in. 1 kN/m) would be expected to have from a practical viewpoint no noticeable effect on pavement performance. (300 mm). Permanent deformations as determined by layer strain theory would be reduced from 12 to 36 percent for a similar variation in geosynthetic stiffness (Figure 62a).3 in. These results indicate that very low stiffness geosynthetics (Sg < 800 lb/in.

C1N30213d) ly SS3N>1011-11 3SV8 NI NOLL3f1032i (1N33213d) QV •9N111112J NI NO110(1038 175 Imp rovemen t in Per formanc e w it h Geo syn t he t ic St i ffnes s. Imp rov emen t on Pe r forman ce w it hGeo sy n the t ic St i ffnes s.) a. 0 II rn rn U) 0 O 0. (a ) Re duc t ion in Base Thickness O (1N30213c1) QV `NOIIVF160330 3cmioens NI N0113r1C3e1 O To ta l De format ion 6 AA z 9 O O O O O r.z 0u_ 5 0 O 0 O (1N301:13d) QV NI NO110110321 z z 0 0 9 0 0 0 U7 a N 9 z in a z U ta O tn 5 0 0 0 II U7 . O z (b) Po s it ion co. . ( b) Subg ra de De forma t ion O . (.

the total reduction in permanent deformation increases from about 10 to 55 percent as the thickness of the pavement decreases from 15 to 6 in.000 ESAL's. (250 mm) to 15. very weak and placed on a poor subgrade (E s = 2000 psi.5 to 6. 64 mm. The subgrade resilient modulus will remain the same with E s = 3500 psi (24 MN/m 2 ).5 in. For a section having this structural strength. Let the asphalt surfacing increase from 2.3 in. Increasing the base thickness from 9. both the laboratory and analytical results indicate if the system is weak enough so that stresses are close to failure.becomes greater when all other variables were held constant. E s = 3500 psi.9 kN/m). for the first series of laboratory pavement tests.5 in. Consider the light reference pavement described in the previous section (2. However. As shown in Figure 63. the observed reduction in rutting due to reinforcement was about 44 percent. (381-150 mm). Thus. Reductions in rutting of the base and subgrade computed by layer strain theory were from 39 to 22 percent. (400 mm) results. important reductions in permanent deformations can be achieved by base reinforcement. 24 MN/m 2 ).8 MN/m 2 ). AC.000 ESAL's of the previous example to perhaps a more typical value of 2. 13. with reinforcement in the bottom having an S g = 4000 lbs/in.75 in. with an aggregate base thickness of about 12. (315 mm). in a very small reduction in base thickness decreasing from 14 to 12 percent (Figure 61a). relatively small changes in stress result from the applied loading 176 . These sections tested were thin. Now consider the effect of significantly increasing the load carrying capacity of the pavement from the 200. The results of Test Series 2 and 3 suggest actual levels of improvement in permanent deformation for the sections shown in Figures 61 and 62 might not be as great as indicated by layer strain theory.4 in.000. based on subgrade strain criteria. (54-165 mm). (4.

R EDUCTIO NINRUTTING, M (PERC ENT)

60

40

20

0
50

10.0

15.0

BASE THICKNESS, T (IN)
Figure 63. Influence of Base Thickness on Permanent Deformation:
S = 4000 lbs/in.

g

16

cif
Lai

12

C.)
ta

co w

z0-

z

0
C.)
CD
CG

0

9

6

3

12

SUBGRADE MODULUS, E s (KSI)

Figure 64. Influence of Subgrade Modulus on Permanent Deformation:
S = 4000 lbs/in.

g

177

either with or without reinforcement (Table 10). For example, the total
change in radial stress due to loading near the top of the subgrade is less
than 0.1 psi (0.7 kN/m 2 ). As shown in Table 30, at best very little
reduction in rutting occurs as a result of reinforcement. This conclusion
is in agreement with previous observations of Brown, et al. [37] for largescale laboratory pavements and by Ruddock, et al. [21,30] for a full-scale
pavement having a comparable bituminous thickness to the section above.
Subgrade Strength. A decrease in the strength of the subgrade as defined
by the subgrade stiffness E s has a very dramatic beneficial effect on the
level of improvement due to reinforcement that can be expected based on the
fatigue and rutting equal strain comparisons. Consider a pavement having an
asphalt surface thickness of 2.5 in. (64 mm), and a base thickness of 9.7
in. (250 mm). Figure 64 summarizes the beneficial effect of reducing the
subgrade stiffness for this pavement from E s = 12,500 psi (86 MN/m 2 ) to 3500
psi (24 MN/m 2 ). This reduction in stiffness caused the percent decrease in
base thickness due to reinforcement to increase from about 5 to 14 percent
for a stiff geosynthetic having S g = 4000 lbs/in. (4.9 kN/m). For a similar
section having a reinforcement stiffness S g = 6000 lbs/in. (7.3 kN/m), the
corresponding decrease in base thickness went from 6 to 16 percent as the
stiffness of the subgrade decreased. These comparisons are both for equal
vertical subgrade strain criteria. This criteria gives the grestest
reductions in base thickness.
For a given structural section, the layer strain theory would also show
a significant increase in beneficial effect with regard to rutting as the
strength of the subgrade decreases. For all computations of permanent
deformation using the layer strain approach, however, the same subgrade
permanent strain properties were used, regardless of the resilient modulus
178

employed in the analysis. Suitable permanent deformation properties for
other subgrades were not available.
The laboratory test track and sensitivity studies both indicates that
an important improvement in performance with respect to rutting can be
obtained when a weak section is constructed on a soft subgrade with a CBR<3,
provided a suitable stiff to very stiff reinforcement is placed at the
correct location. Even when a subgrade is present having a CBR<3, the
economics associated with geosynthetic reinforcement compared to other
alternatives must be carefully evaluated as discussed later.

Slack
During installation of a geosynthetic slack in the form of wrinkles and
irregularities may develop in the reinforcement. As a result, its ability
to provide reinforcement may be significantly reduced as indicated by a
supplementary nonlinear finite element sensitivity study. Figure 65 shows
that even a small amount of slack in a geosynthetic theoretically can result
in a very significant reduction in the force developed in the reinforcement.
The rate of reduction in geosynthetic force becomes less as the amount of
slack increases.
As used in this study, slack is defined in terms of a strain in the
geosynthetic. Hence, slack expressed as a displacement equals a
geosynthetic length, such as its width, times the slack expressed as a
decimal. A slack of 0.1 percent corresponds to 0.14 in. (3.6 mm) in a
distance of 12 ft. (3.6 m). Slack in a geosynthetic as small as about 0.1
percent of its width can reduce the geosynthetic force by about 60 percent,
and a slack of 0.4 percent can cause a 90 percent reduction in force (Figure
65).

179

GEOSYNTHETICFORCEDEVEL OPED ( LBS/I N)

10

5

0
0.0

0.2

0.4

0.6

0.8

SLACK IN GEOSYNTHETIC-PERCENT OF LENGTH
Figure 65. Theoretical Effect of Slack on Force in Geosynthetic:
2.5 in. AC/9.72 in. Base.

180

1.0

In an actual installation the effect of slack may not be quite as great
as indicated by theory. This would be due to the geosynthetic generally
being in full contact with the surrounding materials after construction has
been completed. In laboratory tests, such as those performed for this
study, slack can be easily removed by hand stretching the small pieces of
geosynthetic required in these tests. In full-scale field installations,
slack is an important practical consideration, and it must be minimized
through proper construction practices as discussed later.
Poisson's Ratio.

The value of Poisson's ratio of the geosynthetic was

found to have a moderate effect on the force developed in the geosynthetic.
As the value of Poisson's ratio increases, the force developed in the
geosynthetic also becomes larger, and hence the effectiveness of the
reinforcement increases. For light pavement sections on a weak subgrade,
increasing Poisson's ratio v from 0.2 to 0.4 resulted in a 29 percent
increase in the force developed in the geosynthetic; corresponding
reductions in tensile strain in the asphalt surfacing and vertical
compressive strain on the subgrade were less than 0.2 and 1 percent,
respectively. Further, the compressive increase in radial stress was about
0.075 psi (0.5 MN/m2 ). A Poisson's ratio of 0.3 was used in all other
sensitivity analyses.
In summary, if all other factors are equal, the geosynthetic having the
greatest value of Poisson's ratio should perform best. The improvement in
performance for moderate increases in Poisson's ratio should be reasonably
small. Such improvements would be very hard to detect experimentally
because of variability in the results. Practically no information is
presently available concerning the value of Poisson's ratio for
geosynthetics.

181

Geosynthetic Slip
A slip failure can occur along the interfaces between the geosynthetic
and the materials above and below. The occurrence of interface slip reduces
the effectiveness of the geosynthetic reinforcement to improve pavement
performance. As the rutting beneath the geosynthetic increases, the
tendency to slip would also increase. Whether slip occurs depends upon (1)
the shear strength T that can be developed between the geosynthetic and the
materials in contact with it, and (2) the level of shear stress developed
along the interface due to the external load applied to a particular
pavement structure. The level of applied shear stress is related to both
the resilient and permanent deformations in the pavement, including the
shape of the deflection basin.
Slip may occur directly at the interface between the geosynthetic and
the adjacent soil, or by sliding of soil on soil immediately adjacent to the
interface. The resulting ultimate interface shear stress, T for sliding at
the interface can be predicted by the expression:

T =

where:

T =

C a + on tans

(9

)

ultimate shearing resistance along the
interface

an = stress acting normal to the geosynthetic
C a = adhesion
= friction angle
The contact efficiency e between the geosynthetic and the surrounding
material is defined as e = 6/0 and is expressed as either a percent of or
in decimal form [651. Angular, well-graded sands and silty sands have been

182

found to exhibit high efficiencies when in contact with most geotextiles.
Angular soil grains exhibit better friction performance than rounded grains.

Testing Methods. The interface friction characteristics of a geosynthetic
to be used for aggregate base reinforcement can be best evaluated using a
direct shear test [64-68] as compared to a pullout type test [65,69,70].
Either a free or a fixed type direct shear test can be used. The free type
direct shear test appears, however, to be preferable to the fixed test. In
the free type direct shear test, one end of the geosynthetic is left free as
shown in Figure 66. The same materials to be used in the field should be
placed below and above the geosynthetic, and carefully compacted to the
densities expected in the field. When large size base course aggregates are
used, the apparatus should be at least 8 and preferably 12 in. (200-300 mm)
on a side. Frequently the materials are saturated before performing the
test.
In the fixed shear test preventing strain in the geosynthetic,
particularly if it has a relatively low in-plane stiffness, can have an
important effect on the interface friction developed [70]. Also, bonding
the geosynthetic to a rigid block hampers natural soil grain penetration and
interaction with the underlying material. Nevertheless, Ingold [70] found
relatively small differences between fixed and free type tests.
Interface Behavior.

A slip type failure tends to develop under low

confining stress and for smooth, stiff geosynthetics which resist
penetration of soil grains into the surface [64]. For conditions where soil
grains penetrate into the surface, failure develops a small distance from
the geosynthetic within the soil. Failure occurs in this case by adhesion
and rolling, sliding, dilation, and interlock of soil grains [64]. Cohesive

183

Normal
Stress

PAWAVAPAPWAN

Shear Stress

Clamp

Geosynthetic

Clamp

Geosynthetic
(b) Free

(a) Fixed

Figure 66. Free and Fixed Direct Shear Apparatus for Evaluating
Interface Friction.

FRICTION EFFIC IENCY. (5/43 ( PERCENT)

120

DATA FROM COWOS ET. AL. [64]


100


80

DIRECT SHEAR TEST
CRUSHED GRAVEL ON
BOTH SIDES
g0=53°
60
01

10.0

1.0

100.0

GEOSYNTHET1C OPENING SIZE, (mm)
Figure 67. Influence of Geosynthetic Pore Opening Size on Friction
Efficiency.

184

soils require less surface roughness than cohesionless materials to result
in a soil on soil failure immediately adjacent to the geotextile.
The contact efficiency for loose sands in contact with a wide range of
geotextiles is close to the angle of internal friction, with the range in
contact efficiency typically varying from about 90 to 100 percent of 0 [71].
For dense sands the contact efficiency is lower, typically varying from
about 75 to 90 percent, but it can be as great as 100 percent [66,71].
When the effective grain size of the soil on the side which has
relative movement is smaller than the pore openings of the geosynthetic,
contact efficiency is high. Factors that otherwise would be important have
in general only minor influence on the friction behavior. As pore openings
of the geosynthetic increase (or the grain size of the soil decreases),
better penetration occurs of the grains into the pores of the geosynthetic,
and hence the friction angle 6 becomes greater as illustrated in Figure 67
for a crushed gravel. When the material particle size is less than the
openings of the reinforcement, the contact efficiency may be greater than
100 percent (i.e., 6/0 > 1). A high contact efficiency would therefore be
achieved for most materials placed against very open reinforcement such as
geogrids. Clays also have a high contact efficiency [65].
A geotextile that is compressible in the direction perpendicular to the
plane of the fabric allows better penetration of particles; this has been
observed for nonwoven, needle-punched geotextiles by Martin, et al. [66].
The inplane stiffness of the geotextile also affects interface friction
behavior. Consider two geotextiles having the same size pore openings. The
geotextile having the higher inplane stiffness reaches the peak interface
shear stress at a much lower deformation than the lower modulus

185

geosynthetic. The lower stiffness geosynthetic, however, eventually reaches
a higher peak shear stress [65].

Aggregate Bases.

Collios, et al. [65] found for tests involving stone on

stone the contact efficiencies e of three different large stones to be 86
percent for crushed gravel and 66 percent for rounded gravel compared to 84
percent for sand. These friction test results would be applicable when a
geotextile is placed within a granular layer, since stone was located both
above and below the geosynthetic.
Usually the geosynthetic has been placed at the interface between the
granular base or subbase and the subgrade. To simulate field conditions,
the subgrade soil should be compacted in the bottom of the shear box, and
the coarse base or subbase aggregate in the top [68,72].
The relative displacement required to develop full shear strength at a
ballast-geosynthetic interface was found by Saxena and Budiman [68] to be
about 1.6 in. (41 mm). This large displacement was about three times that
required at the soil-geosynthetic interface on the other side. Upon cycling
the shear stress back and forth, up to 40 percent loss of interface shear
strength was observed. The loss of shear strength appeared to be due to the
ballast pulling the fibers, and causing severe deterioration of the
geotextile.
The deflection required to reach peak shear stress is a function of the
particle size and the normal stress. Typically displacements of 0.1 to 0.4
in. (3-10 mm) are required [64]. However, for large base course aggregate
or very rough geosynthetics, as much as 1 to 2 in. (25-50 mm) of
displacement may be necessary to mobilize full interface strength [68].
Hence for the pavement problem where deformations are small, full interface
strength would probably not be mobilized.

186

Robnett and Lai [72] have determined typical values of adhesion and
friction angle for geotextiles exhibiting both good and poor friction
characteristics (Table 32). The occurrence of relatively large adhesion for
slippage at both the soil and the stone-geotextile interface is in agreement
with the findings of Saxena and Budiman [68].
Grid Reinforcement.

Both metallic and polymer type grid reinforcements

have large openings. As a result well-graded base coarse aggregates
protrude through the openings and hence exhibit a high contact efficiency.
The high contact efficiency has in the past been attributed for granular
materials to aggregate interlock. Jewell, et al. [73] have presented an
excellent discussion of the interaction between geogrids and soil and give
contact efficiencies for seven aggregates. In addition to the mechanisms
previously discussed, a bearing capacity type failure may occur in front of
the transverse members of the grid.
Ingold [70] has found the contact efficiency of a geogrid for the free,
direct shear test to be about 106 percent, compared to 88 percent for the
fixed shear test. A medium to coarse sand with some gravel was used in the
comparison.

Slip in Reinforced Pavements. The shear stresses developed at the
geosynthetic interface become larger, and hence a greater tendency to slip
occurs as the total deflection of the geosynthetic increases. Also, the
laboratory shear test results show a relative movement of up to 2 in. (50
mm) between a geosynthetic and a soft cohesive soil is required to mobilize
full friction. Nonlinear finite element analyses indicate that slip is not
likely to occur for sections of moderate strength or subgrades with a CBR ?
3.

187

Table 32
Typical Friction and Adhesion Values Found for Geosynthetics
Placed Between Aggregate Base and Clay Subgrade

RANGE OF VALUES
GEOSYNTHETIC
CLASSIFICATION

INTERFACE
ADHESION

FRICTION
ANGLE, d
(DEGREES)

TYPICAL VALUES
ADHESION

FRICTION
ANGLE, d
(DEGREES)

High Friction

Soil Geosyn.
Stone-Geosyn.

(0.6-0.8)c
(0.4-0.7)c

0-12
19-23

0.8c
0.5c

6
20

Low Friction

Soil-Geosyn.
Stone-Geosyn.

(0.2-0.3)c
(-0.3-+0.3)c

6-13
11-30

0.2c
0.2c

9
20

/

188

For lighter sections and/or lower strength subgrades. (5 kN/m) for the very stiff geotextile. (2. Results of the two supplementary single track test studies (Figures 44c and 50c) appears to suggest that perhaps the stiff. From the previous discussion of friction.25 in.1 kN/m) compared to about 4300 lbs/in. Problems with slip and also separation can occur at deformations less than 0. but probably not as good as a geogrid. despite its lower stiffness. The better performance of the geogrid under the relatively light wheel loading could be caused by better interface friction characteristics due to interlocking between the geosynthetic and the aggregate base. 189 . The stiffness of the geogrid was about 1700 lbs/in. This seems to indicate that the strengthening observed in the tests was not due to membrane effects. A geogrid and a woven geotextile were placed at the center of the base in two different sections in Test Series 4. The geogrid. a nonwoven needle-punched geosynthetic should have better frictional characteristics than a woven. gave better performance than the much stiffer woven geotextile (refer for example to Table 25 and Figures 40d and 41). These results indicate that only geosynthetics with good friction characteristics should be used for reinforcement. Type Geosynthetic Reinforcement Reinforcement. but rather due to local reinforcement probably caused by small increases in lateral confining pressure. (6 mm) if the full friction in the geosynthetic is not mobilized. woven geotextile used in this project required a much higher deformation to mobilize an equal level of reinforcing potential. The experimental results showing that a stiff geogrid performed better than a very stiff woven geotextile supports this finding. slip does appear to become a problem.

The amount of subgrade soil contamination of the base in sections having the geotextile was negligible. The woven geotextile performed better than the very open mesh geogrid in performing as a separator between subgrade and base. slack in the geosynthetic can very significantly reduce its effectiveness as a reinforcement.Separation. Geogrids were of course not developed to perform the function of separation. prerutting was continued until a surface rut of about 2 in. A geotextile and a geogrid were placed at both the bottom and middle of the aggregate base of different sections. The performance of a number of prerutted sections both reinforced and non-reinforced were evaluated during the laboratory phase of this investigation. PRERUTTING As previously discussed. Prerutting was performed by applying repeated repetitions of a wheel load to the top of the aggregate base before the asphalt surfacing was applied.5 in. (50 mm) was achieved in the 8 in. The separation effect is not considered to be significant for this study in regard to improvement in pavement performance. prerutting was continued until a rut depth of about 0. The loading was carried out along a single wheel path until the desired level of rutting was developed. (200 mm) thick aggregate base.4 to 0. One very efficient method of removing slack and even applying some pretensioning to the geosynthetic is by means of prerutting as demonstrated by Barenberg [75]. 190 . (10-19 mm) was developed at the top of the subgrade.75 in. (38 mm). When loading was conducted above instrumentation. If instrumentation was not present. Prerutting was carried out in both a sand-gravel and a crushed dolomitic limestone base. while in geogrid sections it was as great as 1.

Prerutting alone has more benefit than placing a geosynthetic at an effective location 191 . The presence of a geosynthetic reinforcement appears not to affect the efficiency of prerutting. however. and as a result a stiffer zone. at the top of the aggregate layer. Reinforced sections which have been prerutted can reduce surface rutting on the order of 30 percent or more compared to non-prerutted sections. Supplementary tests show. Prerutting therefore appears to be most effective in reducing the permanent deformation in the soft subgrade. prerutting alone is the mechanism which explains the observed improvement in performance. Prerutting normally results in the formation of a denser. but can also significantly reduce rutting in an aggregate base. that prerutting a non-reinforced section is just as effective as prerutting one which is reinforced (Figure 69). Also. Prerutting is beneficial because of the additional compactive effect applied to the aggregate base. The results from Test Series 2 (Table 25) indicate an 85 percent reduction in subgrade rutting. Improved resistance to permanent deformation and less rutting are thus achieved. Therefore. Prerutting appears to reduce vertical resilient and permanent strains in the base and subgrade (Figures 45(a) and (b) and Figure 46(a) and (b). the vertical stress on the subgrade appears to remain relatively constant with number of load repetitions until the pavement has been severely damaged (Figure 48a). The vertical subgrade stress developed in non-prerutted sections tended to increase at a gradually increasing rate throughout the test. and a 60 percent reduction in base rutting apparently due to prerutting.The experimental results of Test Series 2 (Figure 68) indicate that prerutting an aggregate base reinforced with a geosynthetic results in an important overall reduction in surface rutting. similar to that from a rubber tire roller.

. Reduction in Rutting Due to Prerut .) Geosynthetic Figure 70..0 1...5 2nd SERIES 3rd SERIES —e--15R-M-B 0 GX-M / 1.. Shear Stress T F 10 Fo 0 Initial Prestress Force (lbs/in. Variation of Shear Stress Along Geosynthetic Due to Initial Prestress Force on Edge. 0 10 Figure 68..... 4 5 10 0..2..5 0. 100 10 1 10 2 10 3 4 Figure 69... Reduction in Rutting Due to Prerut with Geogrid. 192 10 5 ..5 -1 .001NMMIL / // 74/W/ 0.-mHOID-B S _.0 0 10 10 1 10 2 10 3 .No Reinforcement.

If sufficient friction cannot be developed to hold the geosynthetic in place. The prestress force can be applied using the geosynthetic as the prestressing element by the following procedure: (1) first stretch the geosynthetic to a desired load level.36]. The friction developed between the geosynthetic and the surrounding soils restrains the geosynthetic from moving. Care must be taken. This mechanism was indicated by a high permanent deformation in the weak aggregate layer of the prerutted section in the first test series (Figure 42a). 193 . PRESTRESSED AGGREGATE BASE/SDBGRADE Basic Prestressing Concepts One potential approach for improving pavement performance is to prestress the aggregate base and the subgrade of the pavement. Although the prestressing of concrete slabs and beams has been performed for many years. it is a relatively new idea for flexible pavements with unstabilized aggregate bases [35. in prerutting a weak granular base which tends to shear rather than densify under a concentrated wheel load. however. the geosynthetic prestressing element tries to return to its original. The mechanism of load transfer to the aggregate base and subgrade is through the shear stress developed along the sides of the geosynthetic. and then (3) release the prestress force. Upon release. (2) hold the geosynthetic in the stretched position until sufficient material is above it to prevent slip. unstretched condition. The formation of shear planes or a weakened zone within the aggregate layer as a result of prerutting can have a detrimental effect on pavement performance.(Figure 69). As a result. the force from the geosynthetic is transferred to the surrounding soil as a compressive lateral stress.

2. and a very stiff woven geotextile (Sg = 4300 lbs/in.. After construction the clamps were removed. Stress relaxation is a loss of force in the geosynthetic occurring when it is prevented from undergoing any deformation. and then the sides were rigidly clamped against the walls of the test facility during construction of the aggregate base and asphalt surfacing. (24 N/m). The improvement of pavement performance due to prestressing the aggregate was clearly indicated by the results of the fourth test series as shown in Figures 40 and 41 (refer also to Table 27). 5 kN/m) reinforced section. stress relaxation can be visualized as the inverse of creep. Experimental Findings The same stiff polypropylene geogrid was employed as the prestressing element that was used in the other experiments.000 load repetitions the prestressed geogrid pavement had about 30 percent less permanent deformation than the 194 . The loss of prestressing effect through stress relaxation is unavoidable. The geogrid was initially stretched to a force of 40 lbs/in..1 kN/m). Prestress loss due to stress relaxation probably reduced the effective applied prestress force to perhaps 20 lbs/in. Important losses of prestress force are also developed through stress relaxation. Stress relaxation in geosynthetics can be quite large.part of the beneficial effect of prestressing is lost through slippage along the interface of the geosynthetic. The shear stress distribution developed along the geosynthetic is approximately as shown in Figure 70. The prestressed pavement performed better than both a non-prestressed section reinforced with a stiff geogrid (S g = 1700 lb/in. which was the prestress level used in the analytical study. At 10. and is highly dependent upon the material type with less stress relaxation occurring in polyester geosynthetics. (50 N/m).

however. (24 N/m). (50 N/m). or (2) loss of prestress with increase in lapsed time from construction.000 load repetitions was about 30 percent less than in a geotextile reinforced section not prestressed (Table 25). which performed next to best. The measured strain in the bottom of the asphalt surfacing of the prestressed section at 10. It then follows from the other results of the experimental study that prerutting a section without a geosynthetic should be just as effective in terms of reducing permanent deformation as prestressing (Figures 50c and 50d). the difference in measured strain was only about 5 percent. Analytical Results In the analytical study of prestress effects. Of considerable practical importance is the finding that the prerutted section having a very stiff geotextile in the middle performed equally well compared to the prestressed section. If the beneficial effect of prestressing on tensile strain was a result of general pavement deterioration. an effective prestress force was applied of 20 lb/in. This conclusion is valid for the conditions of the study including using a polypropylene geogrid with S g = 1700 lbs/in. prestressing would probably not be effective in a field installation for a pavement having a life of 10 to 20 years or more. An important unknown is whether the apparent loss of the beneficial effect of prestressing on strain was due to (1) general deterioration of the pavement as a result of reaching the end of its life.corresponding non-prestressed geogrid reinforced section. then prestressing should be quite effective in increasing fatigue life. (2. On the other hand. By 70. if the loss of prestress was due to stress relaxation with time. This represents the net force 195 .000 repetitions.1 kN/m) initially stressed to 40 lbs/in.

The experimental results. (300 mm). For general comparison. the analytical results tend to support the experimental finding that prestressing the middle of the aggregate base should greatly improve rutting of the base and fatigue performance.existing after all losses including stress relaxation. The analytical results indicate prestressing the center of the layer would have little effect on the vertical subgrade strain.5 in. With this exception. The standard reference section was used consisting of a 2. particularly in the subgrade. For the reference section reductions in permanent deformation were obtained varying from 30 to 47 percent. demonstrate that prestressing the center of the layer can also lead to important reductions in permanent deformation of both the base and subgrade. (64 mm) asphalt surfacing. reduction in rutting of the subgrade would also be small. The analytical study indicates prestressing the bottom of the layer is quite effective in reducing permanent deformation.9 in. and reductions in base thickness based on vertical subgrade strain of about 35 percent (Table 33). however. a variable thickness base. and it might even increase a small amount. Prestressing the center of the aggregate base based on tensile strain in the asphalt surfacing resulted in large reductions in base thickness varying from about 25 to 44 percent (Table 33). expected reductions in total permanent deformation would be on the order of 20 to 45 percent. geotextile reinforced section with reinforcement at the center. the observed reductions in total rutting of the lighter prestressed test section was about 60 percent compared to the nonprestressed. and a subgrade with E s = 3500 psi (24 MN/m2 ). For a base thickness of 11. The analytical results indicate prestressing the bottom of the base is not 196 .

.

Allowance must also be made for all other prestress losses resulting between the time pretensioning is carried out and the prestress force is transferred to the aggregate base. the geosynthetic should be prestressed in the direction transverse to that of the vehicle movement. which is the force used in the laboratory tests. Probably the most efficient method would be to apply the pretensioning force to the other side of the geosynthetic using an electrically powered wench attached to a loaded truck. Pretensioning: Practical Field Considerations To achieve the demonstrated potential for an important improvement in performance. however. as prestressing the middle with respect to tensile strain in the asphalt surfacing. would be a reasonable starting point for additional field studies. Proper allowance should be made for prestress loss due to stress relaxation. One approach that could be employed for applying the pretensioning force would be to place sufficient stakes through loops into the ground along one side of the geosynthetic to firmly anchor it. similar to the way the other side would be anchored during prestressing as described next. These losses would be related to the method used to apply and maintain the prestress force. The truck would supply the dead weight reaction necessary to develop the pretensioning force. A rigid longitudinal rod or bar would be attached along the side of the geosynthetic to 198 . (24 N/m). Probably an initial pretensioning force on the order of 40 lbs/in.as effective. and the skill and care of the crew performing the work. which would depend upon the type and composition of the geosynthetic. and the initial applied stress level. An alternate approach would be to use a dead weight anchor such as a loaded vehicle.

If base construction was not progressing rapidly. To minimize bending in the rod or bar attached to the geosynthetic. and used to pretension other segments of the geosynthetic. consider a soft subgrade having an undrained shear strength of about 500 psf (24 kN/m 2 ). the cable leading to the wench would be attached to the bar at two locations to form a "V" shape. Prestressing the base would most likely be carried out where the subgrade has a CSR less than 3 to 4.9 m) would be required to hole a light initial pretensioning load of only about 20 lbs/in. For example. For conditions where a soft subgrade exists. The pretensioning force could be applied by one wench to about a 10 to 15 ft.9 m) in length having a spacing of about 2.6 m) length of geosynthetic. (24 N/m). (3-4. Thus the practicality is questionable of applying by means of temporary anchors even a light pretensioning force to pavements constructed on soft subgrades having undrained shear strengths less than about 500 psf (24 199 . The cost to just apply this light level of pretensioning to a geogrid by an experienced contractor would probably be about 1 to 1. temporary anchorage of the geosynthetic becomes a serious problem. (50 by 50 mm) by 3 ft. It might be desirable to pretension two or more lengths of geosynthetic at a time. (0.0 to 3 ft. it would be necessary to anchor the side of the geosynthetic being pretensioned using stakes. The pretensioning force could then be maintained on the geosynthetic until sufficient aggregate base is placed and compacted over the geosynthetic to provide the necessary friction force to prevent slippage. Wood stakes 2 in.5 times the geogrid cost. by 2 in. or where a low quality aggregate base is used. The wench and cable system could then be removed. (0. as would likely be the case.distribute the pretensioning force uniformly to the geosynthetic.5-0.

Prerutting would also be considerably less expensive than prestressing. Prestressing the center of the layer would also have the most potential for improving performance with respect to fatigue of the asphalt surfacing. Even moving equipment over very soft soils to provide dead weight. and should be effective over an extended period of time. be achieved through prestressing the aggregate base. particularly in the subgrade (Table 33). Reduction in base thickness due to fatigue in general would be less than if the prestressing is placed at the center of the layer. The analytical results indicate placing the prestress in the bottom of the base will result in the most reduction in permanent deformations in the subgrade. The analytical results further show that placing the prestressing element at the bottom of the base has the potential for greatly reducing permanent deformations.kN/m2 ). The experimental results indicate prerutting the base without the use of a geosynthetic should give equally good performance as prestressing at least with respect to reducing permanent deformations. The experimental results on the prestressed sections were obtained for short-term tests performed under idealized conditions. temporary anchorage would probably not be practical. Prestressing the center of the base should be most effective in reducing rutting within the base. Limited strain 200 . under at least idealized conditions. Loss of prestress effect in the field and prestress loss due to long-term stress relaxation effects are certainly important practical considerations that can only be fully evaluated through full-scale field studies. Summary The experimental and analytical results indicate that important reductions in rutting can. The experimental findings show permanent subgrade deformations should also be reduced by prestressing the middle.

measurements made in the bottom of the asphalt surfacing of the prestressed section indicates an important loss of benefit occurs with either time or else deterioration of the pavement. 201 .

subbases and drainage layers in pavements. the effective thickness and strength of the aggregate layer is reduced. an open graded layer. when the level of contamination becomes sufficiently great. Separation . As shown in Figure 71. Other work. however. and in may cases even a more densely graded layer. silt and clay may with time contaminate the lower portion of the drainage layer. The intrusion of fines into an aggregate base or subbase results in (1) a loss of stiffness. If.A poor physical separation of the base/subbase and subgrade can result in mechanical mixing at the boundary when subjected to load. an increase in fines of up to 6 percent has been found to have a minor effect upon the resilient modulus [104]. In either case.SEPARATION AND FILTRATION INTRODUCTION In recent years considerable interest has been shown in using opengraded aggregate layers as bases. indicates contamination of a portion of an aggregate layer with 2 to 6 percent clay can cause reductions in shear strength on the order of 20 to 40 percent [76]. and (4) a reduction in permeability. A well-designed drainage system has the potential for increasing the life of a flexible pavement by a factor of forty or more [85]. Filtration . clay and fine sand size particles) may form at the top of the subgrade when water is present and under 202 .A slurry of water and fines (primarily silt. is placed directly on the subgrade. Contamination due to the intrusion of fines into the base or subbase can be caused by the following two mechanisms: 1. (3) increased susceptibility to frost action and rutting. however. (2) loss of shear strength. 2.

100 80 60 Bulk Stress. 203 . Ref. Influence of Added Fines on Resilient Modulus of Base (After Jorenby. 20 LEGEND • 11 Nonreinforced Woven Q Thin Nonwoven Thick Nonwoven ❑ Composite 0 10 0 •• S 0 24 20 28 Water Content. 84). Influence of Subgrade Water Content and Geosynthetic on Stone Penetration (After Glynn & Cochrane. 40 rl 20 10 1 5 20 15 10 Added Fines (Percent) Figure 71. 104). Ref. 6 (psi) 0 . w (Percent) Figure 72.

have not yet been developed for aggregate • 204 . (2) limit the amount of fines passing through the separator so as not to significantly change the physical properties of the overlying layer. the geotextile must be sufficiently strong. and (3) the separator must not become sufficiently clogged with fines so as to result in a permeability less than that of the underlying subgrade. Unfortunately. Comprehensive state-of-the-art summaries of the separation and filtration problem have been given by Dawson and Brown [78]. ductile and abrasion resistant to survive construction and in service loading. however. because of the relatively harsh environment which can exist beneath a pavement. If the filtration capacity of the layer above the subgrade is not sufficiently great. a filter cake probably does not develop in the soil adjacent to the filter [90-92]. the slurry will move upward under pressure into the aggregate layer and result in contamination. the classical Terzaghi filter criteria used for steady state filter design are not applicable for at least severe levels of pulsating loading. Formal filter criteria. FILTER CRITERIA FOR PAVEMENTS To perform properly for an extended period of time the filtration/ separation aggregate filter or geotextile must (1) maintain a distinct separation boundary between the subgrade and overlying base or subbase. Finally.pressure due to repeated traffic loading. Jorenby [104] and more recently by Dawson [105]. In harsh environments some clogging and loss of fines through the geosynthetic will Occur. such as occurs beneath pavements where the flow may be turbulent and also reversing. For these severe conditions.

The classical Terzaghi filter criteria were developed for uniform. cohesionless soils in contact with an aggregate filter. with the subgrade squeezing upward into the pores of an open-graded stone as it penetrates downward. A separation type failure can occur either during construction or later after the pavement has been placed in service. Under unfavorable conditions such as a heavy loading and a very weak 205 . The total thickness of the contaminated zone is typically up to about 2 times the diameter of the aggregate which overlies the subgrade [21. This table was taken from the excellent work of Christopher and Holtz [106] who give a comprehensive general discussion of the engineering utilization of geotextiles.80]. including filter criteria and infiltration. The geotextile selection criteria given by Christopher and Holtz is also summarized in Table 34 for both steady state and cyclic flow conditions.77]. are summarized in Table 34. SEPARATION Maintaining a clean separation between the subgrade and overlying aggregate layer is the first level of protection that can be provided to the base. Separation Failure Mechanisms Contamination of the base occurs as a result of the aggregate being mechanically pushed into the subgrade.or geotextile filters placed at the interface between the base and subgrade of a pavement.79. Most serious separation problems have developed when relatively opengraded aggregates have been placed on very soft to soft subgrades [76. The Terzaghi criteria. which assumes steady state flow conditions to exist.76.

Porosity requirement based on graded granular filter porosity.14 (soil) I. old 1511.18 0tif lnletr3211. 2. GEOSYNTHETIC FILTERS I.Otomulnimmm PimmeliMM Relleiramnti OMmmise Remairemeres MM scremenketW pip anodes 0 s IllnerKS Oss 11•111 0 is (1111m1314 On Wei Cho 011ier)%25 Oss (. 2. Des U. Suggested performance test method: Gradient Ratio L.050. 50 sieve ✓ynomic.S. For example. III. Select fabric an the basis of largest opening value required (smallest AOS) II. II. Qualifiers In potential clogging condition situations (e. and perform soil/fabric filtration tests before specification. sea conicles ere. Oss MOW 3 OA Is 1. gap•groded soils and silty type soils) where filtration is of concern. 200 sieve. Additional Qualifier (Optional): 0 95 3(20 5 301 4. prsquolifying the fabric.. D.73 ■ Ml.2 le 1.5 D85 85 Nonwoven: 0 95 I. fabric with maximum opening size possible (lowest AOS No. and Cyclic Flaw 0 95 f5 (If soil can move beneath fabric) 0 50 4 04 a 0 50 85 0. Ref. Additional Qualifier (Opt ional): 015 '130 15 3015 Notet 1. 3. AGGREGATE FILTERS . Permeability should be based on the actual fabric open area available for flow. 206 . ar *fief selection before bid closing. No. II. Filtration tests ore pertormance tests and cannot be performed by the manufacturer as they depend on specific soil and design conditions. Tests to be performed by specifying agency at his representative. dienweet Mures 015. CLOGGING CRITERIA A. Whenever passible. Effective Open Area Qualifiers 2 : Woven fabrics: Percent Open Area: Nonwoven fabrics: Porosity 3 3. SOIL RETENTION (PIPING RESISTANCE CRITERIA) 1 Sails WINN Stalls Flow csoi. Nes Hogan of to. No. PERMEABILITY CRITERIA I A. Passim? U. Less Critical/Less Severe and (with Clean Medium to Coarse Sands and Grovels) k (febtfe). time item lees groin Me. the effective flow area is reduced by 50%. eel O ss • the Minnow et wl SintleSes. use only the gradation of soil passing the U. Critical/Severe Applications i Select fabries meeting I. Alternative: use approved list specification for filtration applications.Table 34 Design Criteria for Geosynthetic and Aggregate Filters (Adapted Christopher and Holtz. if 50% of fabric area is to be covered by flat concrete Weeks.SNo20siev 1. et whet. 2. Pulsating.8 85 AOS No.) from retention criteria should be specified. 1811.. 200 sieve AOS Des 095 Cu42er48 9 e I 450% Passing 24C444 El 0.g.S.4) • IMO wide. b dm weight.Severe Applications I.S.3 B. Note: experience required to obtain reproducible results in gradient ratio test. No. I. (fabric) IND. Less Critical/Non. When the protected soil contains particles from I inch size to those passing the U. 4 sieve in selecting the fabric. 106). 1118. Critical/Severe Applications It (fabric) = 110 k (soil) B.5 Cu Ii‘CuLe B Woven: 095 .

the stone penetration is about equal to the radius of the aggregate. Squeezing of the subgrade was observed to also be equal to about the radius. et al. As the moisture content of the subgrade increases above the optimum value. the depth of contamination could be even more. Because of the presence of the soft subgrade.subgrade.5 in. giving a total contamination depth of approximately one diameter. smooth wheel vibratory roller. • 207 . Linear elastic layered theory was used in developing these relationships. Construction Stresses The critical time for mixing of the subgrade with the aggregate layer is when the vertical stress applied to the subgrade is the greatest.5 ton. (460 mm). are both important factors affecting stone penetration. and as a result the subgrade moisture content. The largest vertical subgrade stresses probably occur during construction of the first lift of aggregate base. [76] found for a very large 4. It might also occur later as construction traffic passes over the base before the surfacing has been placed. The common practice is to compact an aggregate layer with a moderate to heavy. Smooth drum vibratory rollers develop dynamic vertical forces varying from 4 tons (or less) for a small. The subgrade strength. (110 mm) diameter aggregate. Even a reasonably light roller applies relatively large stresses to the top of the subgrade when an initial construction lift is used of even moderate thickness. smooth drum vibratory roller for initial lift thicknesses up to 18 in. 8 and 17. Bell. light roller to as much as 15 to 20 tons for very large rollers. the tendency for aggregate to penetrate into it greatly increases as illustrated in Figure 72. Figure 73 summarizes the vertical stress caused at the subgrade interface by a typical 4. the modulus of elasticity of the first 6 in.

0 4.5 sc 2. fn 1.0 C = undrained shear strength of subgrade 2.) Figure 73. Variation of Vertical Stress on Subgrade with Initial Compaction Lift Thickness and Roller Force.5 0 6 12 18 24 Compacted Lift Thickness Figure 74.o 3 0 1. -5 40 0 oc .60 N t... 208 . 1?"' 1111161. vi 0. Bearing Capacity Failure Safety Factor of Subgrade During Construction of First Lift. T (in. 1111111111111 la 20 7 . k 0 6 12 Compaction Lift Thickness. 3.o e„-1.5 . /..

The ultimate bearing capacity of a cohesive subgrade can be expressed as [17]: quit where: = 5.2c (10) quit = ultimate bearing capacity of the subgrade c = undrained shear strength of a cohesive subgrade The above equation is for plane strain conditions such as would exist beneath a long vibratory roller.(150 mm) thickness of the initial life was assumed to be 1. Large pore openings are. Each successive 6 in. the actual average vertical stress developed on large stone particles at the subgrade interface is greater than the average stress predicted by conventional stress distribution theories. a local bearing failure occurs below the tips of the aggregate. The vertical stress at the subgrade interface predicted by conventional layered theory requires continuous contact on a horizontal plane between the two layers. The actual average vertical stress o z * for an open-graded base would be approximately equal to: o * = a /(1-n) z z (11) 209 .(150 mm) thickness within the lift was assigned an elastic modulus equal to 1. however.5 times that of the material underlying it. the externally applied stress level must be near the ultimate bearing capacity of the subgrade. As a result. and the soil would squeeze upward between the aggregate into the open pores.5 times the modulus of elasticity of the subgrade. the ultimate bearing capacity is about 20 percent greater than given by equation (10). present in coarse. Actually. When the load is applied over a circular area. opengraded granular materials. Bearing Capacity Analysis For a separation problem to develop. which is approximately the case for a wheel loading.

where:

a * = actual average stress developed on the stone particles
z
a
= theoretically calculated vertical stress
n
= porosity of the granular layer
n

The problem is further complicated by the fact that the aggregate
particles are both three-dimensional and irregular in shape. Therefore,
until penetration of the aggregate particles into the subgrade occurs,
contact stresses between the aggregate and subgrade will be even higher than
the average stress given by Equation (11).
For conditions of a wet, weak subgrade, the irregular-shaped aggregates
will be readily pushed into the subgrade, usually during the construction
phase. When stone penetration equals about the effective radius of the
stone, the average contract stress between the stone and soil becomes close
to that given by equation (11). The bearing capacity is probably somewhat
greater than (11) which does not consider the resistance to flow of soil
through the pores of the stone above which is required for a bearing failure
to occur.
Several additional factors further complicate the aggregate penetration
problem. Under a dynamic loading the strength of a cohesive subgrade is
greater than under a slow loading. However, several passes of the roller
may result in reduction in strength due to the build-up of pore pressures in
the subgrade. Also, the possibility exists that the pores in the lower,
tensile portion of the aggregate layer open slightly as the external load
moves over [105]. Because of the overall complexity of the problem, a
rigorous theoretical prediction of soil intrusion is quite difficult.
Therefore, until more research is performed in this area, a simplified
approach can be taken using equation (10) for performing a general
assessment of the severity of the aggregate penetration problem.

210

Construction Lift Thickness
For an initial lift thickness of 6 in (150 mm), the average vertical
stress at the top of the subgrade varies from about 16 to 32 psi (110-220
kN/m2 ) as the dynamic vibratory roller force increases from 4 to 17.5 tons
(Figure 73). These stress levels are sufficient, based on equation 10, to
cause a general bearing capacity failure of a very soft to soft subgrade
having an undrained shear strength less than about 400 to 800 psf (19-38
kN/m2 ), respectively. Aggregate penetration would occur at even lower
stress levels.
Where very soft subgrades are present, frequently the first lift to be
constructed is placed at a greater thickness than used for succeeding lifts
because of subgrade instability problems caused by the construction
equipment. A lift thickness of 12 in. (300 mm) is probably reasonably
typical. For this lift thickness, the average vertical subgrade stress
varies from about 8 to 16 psi (55-110 kN/m 2 ) as the dynamic roller force
increases from 4 to 17.5 tons. For these conditions, a general bearing
capacity failure as predicted by equation (11) could occur for undrained
shear strengths less than about 200 to 400 psf (10-20 kN/m 2 ).
Separation Case Histories
Mixing of the subgrade with an aggregate base has been reported at
several sites where geosynthetics have not been used. At one site wellgraded aggregate with about a 1.25 to 1.5 in.(30-38 mm) top size and 5
percent fines was observed during construction to intrude up to a depth of
about 1 to 2 in. (25-50 mm) into a soft subgrade [21,81]. For the
conditions existing at the site, the calculated safety factor for a general
bearing capacity type failure varied from about 0.8 to 1.4.

211

At two sites where intrusion occurred, the ratio D15/d85 varied from 17
to 20. For comparison, the frequently used Terzaghi filter criteria for
steady seepage requires D15/d85 5. 5. Hence, conventional static filter
criteria was significantly exceeded at these two sites. Under severe
conditions of loading, intrusion may also occur even if conventional
Terzaghi filter criteria are satisfied [82,83].
Separation Design Recommendations

The following tentative design criteria are proposed to minimize
problems with separation between an aggregate layer and the underlying
subgrade. Most problems involving separation will occur where soft to very
soft cohesive subgrades are encountered typically having undrained shear
strengths less than about 500 psf (24 kN/m 2 ).
1.

If the safety factor with respect to a general bearing
capacity failure is greater than 2.0, no special
precaution is needed with respect to separation. For
very open-graded granular bases or subbases, a limited
amount of punching of the aggregate into the subgrade
will occur for a safety factor of 2. The depth of
punching should be equal to or less than approximately
the radius of the maximum aggregate size.

2.

For a bearing capacity safety factor between about 1.3
and 2.0, either conventional Terzaghi filter criteria
should be satisfied, or else a geotextile should be used
as a separator. Specific recommendations concerning the
selection of a geotextile are given in the next section.

3.

If the safety factor is less than 1.3, use of a
geotextile is recommended regardless of whether filter

212

criteria are satisfied. Consideration in this case
should also be given to satisfying filter criteria,
particularly if a very open-graded stone is to be used
for drainage applications. If the granular filter
material satisfies filter criteria, the geotextile will
serve primarily as construction aid.
The above recommendations are given to avoid contamination of the
granular layer due to intrusion and subsequent mixing. Drainage
applications where filtration is important are discussed in the next
section.
Figure 74 gives the bearing capacity safety factor as a function of
construction lift thickness for selected vibratory rollers and undrained
subgrade shear strengths. This figure shows for a moderate vibratory roller
weight of 8 tons and lift thicknesses of 12 in. (300 mm), separation could
become a problem for subgrades having undrained shear strengths less than
about 500 psf (24 kN/m2 ). This subgrade strength corresponds to a standard
penetration resistance (SPT-value) of approximately 4 blows/ft.(13 b/m).
A very substantial increase in shear strength of a soft to very soft
subgrade will in most cases occur reasonably rapidly after placement of the
pavement structure [42]. This increase in strength should be considered in
estimating the bearing capacity safety factor for long-term traffic loading
conditions. The initial undrained shear strength of the subgrade can be
estimated from vane shear tests, undrained triaxial shear tests, or from the
results of cone penetrometer tests. For preliminary design purposes, Table
35 can be used when reliable estimates of the shear strength based on
testing are not available.

213

Table 35
Preliminary Subgrade Strength Estimation

Subgrade
Description

Field Condition

Standard
Penetration
Resistance,N
(blows/ft.)

Approximate
Undrained
Shear Strength,C
(psf)

Very Soft

Squeezes between
fingers

0-1

0-250

Soft

Easily molded
by fingers

2-4

250-500

Firm

Molded by strong
pressure of
fingers

5-8

500-1000

Stiff

Dented by strong
pressure of
fingers

9-15

1000-1500

Very Stiff

Dented slightly
by finger
pressure

15-30

1500-2000

Hard

Dented slightly
by pencil point

>30

>2000

Table 36
Vertical Stress on Top of Subgrade
for Selected Pavement Sections

Section

A.C.
Surface
(in.)

Very Light

1.5

Light

Granular
Base
(in.)

Vertical
Subgrade
Stress (psi)

6

21

3.5

a

10

Medium

6

8

6

Heavy

8

14

3

i
Notes: 1. Dual wheel loading of 4.5 kips/wheel at 100 psi tire
pressure.
2. Moduli/Poisson's Ratio: AC- 200,000 psi/v=0.2;
Granular Base - 10,000 pji/
v=0.35;
Subgrade - 4000 psi/v=0.4.
3. Analysis - Linear elastic; linear elastic vertical
subgrade stress increased by 12 percent
to give good agreement with measured test
section subgrade stress.

214

Selection of an actual geosynthetic or aggregate filter to use as a
separator is considered later in the section on Filter Selection.

FILTRATION
Some general requirements for intrusion of a slurry of subgrade fines
into an open-graded aggregate layer can be summarized from the early work of
Chamberlin and Yoder [86]:
1.

A saturated subgrade having a source of water.

2.

A base more permeable than the subgrade with large
enough pores to allow movement of fines.

3.

An erodable subgrade material. Early laboratory work by
Havers and Yoder [94] indicate a moderate plasticity
clay to be more susceptible to erosion than a high
plasticity clay. Silts, fine sands and high plasticity
clays that undergo deflocculation are also very
susceptible to erosion.

4.

The applied stress level must be large enough to cause a
pore pressure build-up resulting in the upward movement
of the soil slurry.

Although the work of Chamberlin and Yoder [86] was primarily for concrete
pavements, a similar mechanism similar to movement of slurry also occurs for
flexible pavements.
Filtration Mechanisms
Repeated wheel load applications cause relatively large stresses to be
developed at the points of contact between the aggregate and the subgrade.
As loading continues, the moisture content in the vicinity of the projecting
aggregate points, for at least some soils, increases from about the plastic

215

limit to the liquid limit [97]. The moisture content does not, however,
significantly increase in the open space between aggregates (Figure 75). As
a result the shear strength of the subgrade in the vicinity of the point
contacts becomes quite small. Hoare [97] postulates the increase in
moisture content may be due to local shearing and the development of soil
suction. When a geotextile is used, soil suction appears to be caused under
low stress levels by small gaps which open up upon loading [98]. The gaps
apparently develop because the geotextile rebounds from the load more
rapidly than the underlying soil. Remolding may also play a role in the
loss of subgrade strength.
Due to the application of wheel loadings, relatively large pore
pressures may build up in the vicinity of the base-subgrade interface
[87,99,1001. As a result, in the unloaded state the effective stress
between particles of subgrade soil become negligible because of the high
residual pore water pressures. These pore pressures in the subgrade results
in a flow of water upward into the more permeable aggregate layer. The
subgrade, in its weakened condition, is eroded by the scouring action of the
water which forms a slurry of silt, clay and even very fine sand particles.
The slurry of fines probably initiates in the vicinity of the point contacts
of the aggregate against the soil [761. This location of slurry initiation
is indicated by staining of geotextiles used as separators where the
aggregates contact the fabric.
The upward distance which fines are carried depends upon (1) the
magnitude of induced pore pressure which acts as the driving force, (2) the
viscosity of the slurry, and (3) the resistance encountered to flow due to
both the size and arrangement of pores. Fine particles settle out in the
filter or the aggregate layer as the velocity of flow decreases either

216

Stress Concentration

Aggregate RotationCauses Large Local
...../Ille Strain in Fabric

Slurry Movement
//1-.

.

4•

:

I .*:

fie

/W.5,/
/it .r. •so...
•e•
11%0

Increase in
Water Content
Under Stress Concentration
(Fabric stained in this area)
SLURRY INITIATION HERE

Figure 75. Mechanisms of Slurry Formation and Strain in Geosynthetic.

217

locally because of obstructions, or as the average flow velocity becomes
less as the length of flow increases. Some additional movement of material
within, or even out of, the base may occur as the moisture and loading
conditions change with time [86].

Geotextile Filters
Geotextile filters have different inherent structural characteristics
compared to aggregate filters. Also, a considerable difference can exist
between geotextiles falling within the same broad classification of woven or
nonwoven materials due to different fiber characteristics. Nonwoven
geotextiles have a relatively open structure with the diameter of the pore
channels generally being much larger than the diameter of the fibers. In
contrast, aggregate filters have grain diameters which are greater than the
diameter of the pores [92]. Also, the porosity of a nonwoven geotextile is
larger than for an aggregate filter.
The following review of factors influencing geotextile filtration
performance are primarily taken from work involving cyclic type loading.
Electron microscope pictures showing the internal structure of several
non-woven geosynthetics are given in Figure 76. None of these geosynthetics
were considered to fail due to clogging during 10 years of use in edge
drains [107]. Their approximate order of ranking with respect to clogging
from best to worst is from (a) to (d).
Thickness. The challenging part of modifying granular filter criteria for
use with fabrics is relating soil retention characteristics on a geotextile
with those of a true three-dimensional granular filter. Heerten and
Whittmann [92] recommend classifying geotextiles as follows:

218

Top

&a),

•■■

Av.‘4

1..

ilib

IISIMININIL
_ \lir"

EdeWilialk . ‘74, / fasts...L..
-41111CoNs A
:641111.111111■••

is.■ ,_...6
".....mn V.I'..".

(a) Nonwoven, Needle 4.5 oz/yd , (b) Nonwoven, Needled 5.3 oz/yd 2
Heat Bonded, 60 mil.
75mil.
/
V
_ Top

00I I111117 A IIPPP•b.■rr

Top
V
' ■v

amflie.A.0:71,00
t
7

' ,= ■111•1111W

(c) Nonwoven 4.5 oz/yd 2 , 30 mil. (d) Spun-Bonded, 15 mil.
Figure 76. Electron Microscope Pictures of Selected Geotextiles:
Plan and Edge Views (94x).

219

1.

Thin:

thickness t<2 mm and geotextile weights up to 9 oz./yd 2
2 ).

2.

Thick:

(30g/m

single layer, needle punched: thickness t>2 mm and
geotextile weights up to 18 oz./yd 2 (600 g/m2 ).

3.

Thick multi-layer, needle punched geotextiles.

Earlier work by Schober and Teindl [90] found wovens and non-wovens
less than 1 mm in thickness to perform different than non-wovens greater
than 2 mm, which gives support to the above classifcation scheme.
As the thickness of a nonwoven, needle punched geotextile increases,
the effective opening size decreases up to a limiting thickness similarly to
an aggregate filter [92]. Thick needle punched geotextiles have been found
to provide a three-dimensional structure that can approach that of an
aggregate filter; thin geotextiles do not. Also, soil grains which enter
the geotextile pores reduce the amount of compression which occurs in a
nonwoven, needle punched geotextile subjected to loading.
As the thickness of the geotextile increases, the effective opening
size decreases and fines in suspension have a harder time passing through
the geotextile because of the three-dimensional structure [91,98,102]. The
fines which do pass through the geotextile may be deposited above the fabric
in a thin layer that can significantly reduce the effective permeability of
the layer. A layer of fines forming a cake below the geotextile has also
been observed. When open-graded granular materials are located above the
geotextile, the fines passing through would probably to be pumped into the
voids of the stone resulting in stone contamination. The load on the
aggregates in contact with the geotextile can result in a significant amount
of stretching of the fabric and a temporary increase in pore diameter, which
allows more fines to pass through. If, however, the geotextile has pores

220

which are too small in diameter or the porosity is too small, clogging can
occur, and the geotextile is not self-cleaning.
Self-Cleaning Action. Laboratory tests have shown a change in the direction
of flow through a geotextile can cause an increase in its permeability
[98,101]. Hence, partial flushing of fines from a geotextile is apparently
possible under conditions of reversing flow. The permeability, however,
does not go back to its original value upon flow reversal. Flushing was
found by Saxena and Hsu [98] to be more effective for heavier, nonwoven
geotextiles. Whether self-cleansing can actually occur in the field has not
been demonstrated.
Load Repetitions. The quantity of fines migrating upward through a
geotextile filter is directly related to the log of the number of load
applications [91,98] as illustrated in Figure 77. The Soil Contamination
Value (SCV) quantifies soil loss through a geotextile. SCV is the weight of
soil per unit area passing through the geotextile [91].
Apparent Opening Size. The Apparent Opening Size (AOS) quantifies at least
approximately the effective pore opening size of a geosynthetic. The
apparent opening size (AOS) of a geotextile is defined as the minimum
uniform, spherical particle size of a uniform shape that allows 5 percent or
less of the particles to pass through the geotextile [106]. For a given
weight, geotextiles having a small fiber size, and as a result a smaller
effective opening, allow less material to be washed through [92]. Some
general findings by Carroll [93] involving AOS as related to geotextile
filtration are as follows:
1.

The apparent opening size (AOS) of the geotextile cannot
be used alone to directly compare the retention ability
221

4000 N 3000 E ta 0 O t't. logN Figure 77. 0 95 (pm) Figure 78.5 4. V. et al. Ref.. 1 Hoare (1982) Saxena & Hsu (nonwoven) 8 oz/yd 2 t=1 mm AOS = 70 16 oz/yd 2 t =2mm AOS= 100 3000 2000 1 00 0 0 5. Variation of Geosynthetic Contamination with Number of Load Repetitions (After Saxena and Hsu. 0 95 (After Bell. C. 98).5 3.5 Number of Load Repetitions.1 (gm/m 2) So i lCon tam ina t ion Va lue. S. Ref. Variation of Geosynthetic Contamination with Geosynthetic Apparent Opening Size. 2000 • O • 1000 0 A • • • •• Zis 0 100 200 300 400 500 Apparent Opening Size.5 2. 222 . 79).

AOS values can be related to the retention ability of geotextiles provided proper consideration is given to the other significant factors. In the laboratory tests performed by Bell.of a nonwoven and woven geotextile. soil contamination is the percent of soil trapped within the 223 . et al. This extensive field study also indicates increasing soil contamination of the geotextile occurs with increasing apparent opening size (AOS) as shown in Figure 79. 2. Woven and thin nonwovens should have different filter criteria than thick wovens. the AOS of woven monofilaments and nonwoven geotextiles should not in general be compared since they will not have the same filtration efficiency [93]. Fabric pore size. [79]. The AOS measures the maximum "straight through" openings in a woven geotextile. The uniformity coefficient of the soil being protected has an important influence on the filter criteria. as estimated using the method of Schober and Teindl [90]. The sand layer also had the smallest apparent opening size. As defined in this figure. pore structure and filtration capacity are not accurately defined by AOS. Also. 3. The quantity of fines trapped by the filter layer when subject to cyclic loading generally increases with increasing apparent opening size (AOS) of the filtering media (Figure 78). Soil contamination of geotextiles removed from beneath railroad tracks has been reported by Raymond [80]. the least amount of contamination was observed when a thin sand layer was employed compared to the geotextiles tested. 4.

3000 2500 50 2000 150 0 5 00 30 Soil Contamination Value Contours. N 70 3500 Ave rag e Stres s Leve l.S. SCV (gm/m 2 ) 16 20 24 28 32 Subgrade Moisture Content. Ref. 84).800 So i l Con ten t ( Per c en t o f Or ig ina lWe ight ) Needle Punched Geosynthetics i t 0 600 400 200 Most Suitable Limits for High k 400 270 200 140 100 70 50 40 30 U. 80). Ref. w (percent) Figure 80. Below Railroad Ties with Geosynthetic Opening Size (After Raymond. Sieve Standard Sieve Number Figure 79. 224 . Variation of Geosynthetic Contamination Approximately 8 in. Variation of Geosynthetic Contamination with Stress Level and Subgrade Moisture (After Glynn & Cochrane.

As pointed out by Raymond [80].5 kip (20 kN) dual wheel loads are applied to the surface of the pavement.geotextile compared to the uncontaminated dry geotextile weight. Since the applied vertical stress also decreases with increasing depth. The curve relating variation of soil content with depth (Figure 81) is similar in general shape to a typical vertical stress distribution curve. this alternate definition is more closely related to classical filter criteria that limits the amount of soil which can enter the filter. and the tires are inflated to 100 psi (0. Figure 79 shows results for an alternate definition of AOS based on 95 percent of the uniform particles being retained on the surface of the geotextile [103]. so does the quantity of fines migrating through the geotextile (Figure 80) and the amount of contamination. Data obtained from field studies (Figure 81) shows that the level of contamination rapidly decreases below a railroad track structure with increasing depth [80]. a comparison was made of the vertical stress developed beneath a heavily loaded railroad track with the stress developed at the top of the subgrade for typical pavement sections. As the applied stress level on the geosynthetic increases. contamination of a geotextile in the field is indeed dependent upon stress level. Stress Level. Undoubtedly the scatter in data in Figure 79 is at least partly because soil contamination is not only related to AOS but also to a number of other factors as previously discussed. 225 . Assume 4.7 MN/m2 ). Let the critical railroad loading be simulated by a fully loaded cement hopper car. To approximately translate the extensive findings of Raymond [80] for geotextiles placed below railroad track installations to pavements.

Vertical Stress on Top of Highway Subgrade or Beneath Railroad Track (psi) Heavy Pvmt.5 pci/rail Figure 82..51c. 226 . ■ma 400 600 800 Soil Content (percent original weight) Figure 81... 80). O • 200 • tq 0 a 30 0 A - ./.m.0 0 • C ..•A • /..."---. Ref.... Variation of Vertical Stress with Depth Beneath Railroad Track and Highway Pavement. Observed Variation of Geosynthetic Contamination with Depth Below Railway Ties (After Raymond.. US 297 Sieve 100 a • .. Resin Treat 13-15 oz/yd 2 AOS... Needle Punched. US 40 Sieve AOS. x 1 in. 15 10 a a o a a 40 a Vertical Stress o z BenathCros-Ti 3 a a a A a a cr..5 ir J 9 :n.... Light Pvmt. . Medium Pvmt. x 8. a a Doubled Over Geotextile • 40 I i ■■••■ 0 200 wl.. . Tie: k = I9. 80 Corrected O z Railroad Stress Loaded Cement Hopper Car 71.

light. Laboratory Testing Methods Laboratory studies to observe the migration of fines through both granular filter layers and geotextile filters have most commonly employed a constant gradient test which simulates steady state. medium and heavy highway pavement section.Figure 82 shows the approximate equivalent depths below the railroad cross-ties that corresponds to the vertical stress at the top of the subgrade for a typical light. vertical stresses from pavement type loadings spread out relatively quickly. respectively. typical very light. a heavy structural pavement section would be subjected to a much less severe stress condition. (200-300 mm) beneath the tie which corresponds to a vertical stress level on the order of 14 psi (96 kN/m 2 ). and (2) a highway type pavement should exhibit a wide variation in performance with respect to filtration depending. unidirectional seepage conditions [91. In contrast. among other things. The results obtained from constant gradient tests. geotextiles are generally placed at a depth of about 8 to 12 in. Very thin pavement sections would probably be subjected to an even more severe vertical stress and hence more severe infiltration condition than for a typical railroad ballast installation. The practical implications of these findings are that (1) the railroad type loading is considerably more severe compared to most structural sections used for pavements. 227 . In contrast. 41. upon the thickness and strength of the structural section. 6 and 3 psi (138. 21 kN/m 2 ). For railroad track rehabilitation.93]. A heavy train loading causes relatively large vertical stresses which spread out slowly with depth. 69. For comparison. 10. medium and heavy pavement sections (Table 36) would have maximum vertical stresses at the base-subgrade interface on the order of 21.

Selected Practices Corps of Engineers Filter Criteria. the Corps of Engineers recommends the criteria show in Table 37. For unidirectional. In contrast to other tests. with the filter layer and base material above. An improved test [101] has been developed by Dempsey and Janssen for evaluating the relative effectiveness of different geotextiles (Figure 83). bound for establishing design criteria for pavement infiltration applications. Most frequently dynamic testing to simulate pavement conditions has been carried out in cylindrically shaped." 228 . non-turbulent conditions of flow. possibly unsafe. A cyclic loading is then applied to the top of the specimen through a rigid loading platen. The Corps instructions cautions about using filter materials in inaccessible areas indicating that their use "must be considered carefully. the subgrade soil is placed on top of the geotextile filter. as well as the permeability of the geotextile as a function of load repetitions. The test is performed in a triaxial cell at a realistic confining pressure. one million load repetitions are applied. rigid cells which may consist of either a steel mold [76. Dawson [105] has pointed out the important need for performing tests at realistic vertical stress levels comparable to those existing in pavements.which do not use a cyclic load. To evaluate long-term performance.96] or a plexiglass cylinder [98].84. serve as an upper. Water is continuously passed downward through the specimen at a constant hydraulic gradient as a repeated loading is applied. He also shows that three dimensional pavement tests are more appropriate than the conventional one-dimensional test. The quantity of fines washed through the geotextile is measured. The subgrade soil is generally placed in the bottom of the mold.

Dual Water Reservoirs Air Pressure Cyclic Load Mercury Manometer to Measure Pressure Difference Water Flow Air Flushing Valve Confining Pressure Soil Specimen Confining Pressure Geotextile Marbles Lucite Plates (indented to trap sediment) Water Flow Filter Paper Overflow Collection Reservoir Figure 83. 4 INPLACE COSTS T=SAND FILTER THICKNESS 3 T=6 IN T=4 IN 2 T=3 IN TYPICAL RANGE IN FABRIC COST 1 • EXAMPLE (T=4 IN) 0 0 4 6 12 16 FILTER SAND COST (DOLLARS/TON) Figure 84. Ref. 101). Cyclic Load Triaxial Apparatus for Performing Filtration Tests (Adapted from Janssen. 229 . Economic Comparison of Sand and Ceosynthetic Filters for Varying Sand Filter Thickness.

200 Sieve) Piping (1) (2) Permeability Woven Non-Woven (4) (3) n85 ( mm ) PO4 > 10% 5% to 50% E0S(mm) < D 85 ( mm ) PO4 >4% 50% to 85% (a) EOS(mm) < D 85 (mm) POA > Less than 5% 1. E0S(mm) < kG G — 5k S (2) 4% k > 5k G — S k G > 5k S (b) Upper Limit on EOS is EOS (mm) < . (4) k G is the permeability of the non-woven geotextile and k s is the permeability of the protected soil. S. Standard Sieve) Footnotes For Table No.S. G (3) 1. 120 U. (2) These protected soils may have a large permeability and thus the POA or k may he a critical design factor.125 mm (No. 4 sieve in selecting the EOS of the geotextile. Army Corps of Engineers Geosynthetic Filter Criteria (Ref. 70 U. Standard Sieve) >85% (a) E0S(mm) < D 85 (mm) k G — > 5k S (b) Lower Limit on EOS is EOS (mm) > .212 mm (No. 121) Protected Soil (Percent Passing No.Table 37 U. 4 sieve use only the soil passing the No. T. 230 .)85 is the grain size in millimeters for which 85 percent of the sample by weight has smaller grains. 1 (1) When the protected soil contains appreciable quantities of material retained on the No.

For fine grained soils having 50 or more percent passing the number 200 sieve. Pennsylvania DOT Filtration/Separation Practices. the two layers are each 3 in. If a 6 in (150 mm) thick subbase is used. If a geotextile is used. (75 mm) in thickness. trapezoidal tear strength > 75 lbs (0. the open graded aggregate drainage layer is placed directly on the geotextile. 231 . puncture > 110 lbs (0. The open graded base must be well graded. 70 U. this criteria requires that the AOS generally be between the No. (300 mm) subbase is used the two layers are each 6 in. (150 mm) thick. 120 U. grab tensile strength a.5 kN). The Pennsylvania DOT uses as a standard design an open graded subbase (OGS) to act as a blanket drain (Table 38). and is equal in thickness to the full depth of the subbase. Standard Sieve. To maintain separation a more densely graded Class 2A stone separation layer is placed beneath the open graded drainage course. 70 and No. if a 12 in.3 kN). 270 lbs (0. The geotextile separator used typically has a weight of about 16 oz/yd 2 (380 gm/m2 ). Both woven and non-woven geotextiles are allowed.S. To exhibit some stability during construction. the open graded base is required to have a minimum of 75 percent crushed particles with at least two faces resulting from fracture. It also has the additional mechanical properties: AOS smaller than the No. To permit adequate drainage and to resist clogging.3 kN).3 kN). wovens must have a percent open area greater than 4 percent for soils having 5 to 85 percent passing the number 200 sieve.S. and greater than 10 percent for soils having less than 5 percent fines. Sieve. and an abrasion resistance 40 lbs (0. non-woven geotextiles must have a permability greater than 5 times that of the soil. An approved geotextile may be substituted for the separation layer. For similar reasons. grab elongation ? 15 percent.

Top Size 1/2-4 in.-. Top Size >51 Fines N=4 1. Rounded gravels can be given a separation number one less than indicated.4 < SF < 10 SF < 1.. N — 1-2 in./ SEPARATION NUMBER( 1 ). Top Size Aggr. 232 . if desired. Uniform (no fines N 1) GEOTEXTILE SEVERITY CLASSIFICATION Low Moderate Severe Very Severe 3.0 . Well-graded Uniform N=3 (No Fines) N=2 1/2-2 in. Tests indicate the proposed gradation should have a permeability of about 200 to 400 ft/day.4 2 1 - 4 3 2 1 - 3. Table 39 Separation Number and Severity Classification Based on Separation/Survivability SEARING CAPACITY SAFETY FACTOR 1.4 < SF < 2 1. Angular. 2-4 in. 1-5% Aggr.4 - . Angular. Top Size Angular.Table 38 Aggregate Gradations Used by Pennsylvania DOT For Open-Graded Drainage Layer (OGS) and Filter Layer (2A) AASHTO DRAINAGE LAYER (OCS) SEPARATION LAYER AYER (2A) New Proposal (1) 100 2 Old 100 100 3/4 52-100 52-100 52-100 3/8 36-70 36-65 36-65 #4 24-50 20-40 8-40 #8 16-38 - #16 30-70 3-10 0-12 #30 - 0-5 0-8 #50 - 0-2 - <10 0-2 <5 #200 - / \ Note: 1. Fines..

In addition. They require for blanket drains a nonwoven geotextile having a minimum weight of 4 oz. et al.120] have performed a comprehensive study of open graded aggregate and bituminous stabilized drainage layers. These studies involved wetting the pavement sections and observing their performance in a circular test track. New Jersey/University of Illinois. These geotextile material requirements are in general much less stringent than those used by the Pennsylvania DOT.5 to 0.and have a uniformity coefficient C u = D60/D10 ? 4. [75. In one instance. the grab tensile strength must be ? 100 lbs. The open graded base is placed using a spreader to minimize segregation.83. dense-graded aggregate subbase or lime-flyash treated base. The subgrade used was a low plasticity silty clay./yd 2 (95 gm/m2 ). These studies indicated good performance can be achieved by placing an open-graded aggregate base over a sand filter. and the toughness (percent grab elongation times the grab tensile strength) ? 4000 lbs (18 kN). The California DOT allows the use of geotextiles below open graded blanket drains for pavements and also for edge drains. grab tensile test elongation ? 30 percent.4 kN). (12-19 mm) of intrusion of sand occurred into the open-graded base. An open-graded bituminous stabilized layer was found to be an effective drainage layer. although the opengraded drainage layer/sand filter used met conventional static filter criteria. A significant amount of intrusion of subgrade soil also occurred into an open-graded control section which was placed directly on the subgrade. 233 . but rutted more than the nonstabilized drainage material. (0. about 0. Barenberg.75 in. California DOT.

needle bonded geotextiles with loose filament crossings have a sufficiently high elongation to withstand heavy railroad loadings without puncturing. As a result of this study. Wehr [82] concluded that only non-woven. The extensive work of Raymond [80] was for geotextiles placed at a shallow depth (typical about 8 to 12 in. these nonwovens did better than spun bonded geotextiles having little needling.. The findings of Raymond appears to translate to the most severe conditions possible for the problem of filtration below a pavement including a thin pavement section. having large tex fibers on the inside and low tex fibers on the outside. nonwoven geotextiles were found by Raymond to perform better than thin heat bonded geotextiles which behaved similarly to non-wovens. Stone penetration into the lime modified subgrade was approximately equal to the diameter of the drainage layer stone. heavy. Raymond also found the best performing geotextile to be multi- layered.Lime modifications of the subgrade was also found to give relatively good performance. For the reversible. non-steady flow conditions existing beneath a railway track. open-graded drainage layer placed over a dense graded aggregate filter [109]. 200-300 mm) below a railroad track structure. This condition constitutes a very harsh environment including high cyclic stresses and the use of large. resin treated. the New Jersey DOT now uses as standard practice a non-stabilized. particularly with an open-graded base having a finer gradation. The drainage layer/filter interface is designed to meet conventional Terzaghi type static filter criteria. Well needle punched. non-woven geotextiles having a low AOS less than 55 pm 234 . uniformly graded angular aggregate above the geotextile. Abrasion of thick spun bonded geotextiles caused them not to perform properly either as a separator or as a filter. Also. Harsh Railroad Track Environment.

105. The strength of the pavement section placed over the filter/separator determines the applied stresses and resulting pore pressures generated in the subgrade.110]: 1. existing moisture conditions and undrained shear strength are all important.109. A large. Aggregate Base/Subbase. Sieve No. Approximately extrapolating Raymond's work based on vertical stress indicates for structural numbers greater than about 4 to 4. of about 100 to 140 should result in roughly the same level of contamination and clogging when a large uniformly graded aggregate is placed directly above. which provides important lateral drainage. (300 mm) the needle punched geotextile should have a weight of at least 20 oz. The top size. 270 sieve size) were found to provide the best resistance to fouling and clogging. a geosynthetic having a U. (300 mm) in a track structure corresponds approximately to a geosynthetic placed at the subgrade of a pavement having an AASHTO structural number of about 2. Pavement Section Strength. angularity and uniformity of the aggregate placed directly over the filter. 3. and low plasticity clays should be most susceptible to erosion and filtration problems.93.(U. Use of a low AOS was also found to insure a large inplane permeability. Low cohesion silts. for continuous welded rail. FILTER SELECTION INTRODUCTION Factors of particular significance in the use of geotextiles for filtration/separation purposes below a pavement can be summarized as follows [79. dispersive clays. The type subgrade.5. Full scale field tests by Wehr [82] indicate for low plasticity clays and highly compressible silts. A depth of 12 in. 2. and preferably more.S./yd2 (480 gm/m 2 ).S. Raymond [80] recommends that at a depth below a railway tie of 12 in. that primarily sand and silt erodes into the geotextile.75 based on vertical stress considerations (Figure 82). No. Subgrade. 235 .80.

Separation. The apparent opening size (AOS) is at least approximately related to the level of base contamination and clogging of the geotextile. Thin (t < 1 mm) non-woven geotextiles do not perform as well as thicker. Geosynthetic Thickness. for example.83. The steps for selection of a geosynthetic for separation and 236 .105. 4. needle punched non-wovens (t 2 2 mm). Clogging. uniformly graded aggregates are placed directly above. 9. 5. large strains are locally induced in a geosynthetic when big. fiber structure and also internal pore size are all important. 8.80.angular uniform drainage layer. Granular filters are thicker than geosynthetics and hence have more three dimensional structural effect. In providing filtration protection particularly for silts and clays some contamination and filter clogging is likely to occur. GEOTEXTILE SELECTION Where possible cyclic laboratory filtration tests should be performed as previously described to evaluate the filtering/clogging potential of geosynthetic or aggregate filters to be used in specific applications.108]. A preliminary classification method is presented for selecting a geosynthetic based on the separation/survivability and filtration functions for use as drainage blankets beneath pavements. Most studies conclude that needle punched.84].92. Sand aggregate filters are superior to geotextiles. Strain. Fiber size. Wehr [82] found strains up to 53 percent were locally developed due to the spreading action of the aggregate when subjected to railroad loads. 7. For conditions of a very soft to soft subgrade. particularly under severe conditions of erosion below the pavement [76. non-woven geotextiles perform better than wovens. Aggregate Filters.107. Apparent Opening Size (AOS). 6. and greater reductions occur [80. constitutes a particularly severe condition. Non-Wovens. Reductions in permeability of 1/2 to 1/5 are common. The filter criteria given in Table 34 can serve as a preliminary guide in selecting suitable filters for further evaluation.

Enter Table 40 with the appropriate geotextile SEVERITY CLASSIFICATION and read off the required minimum geotextile properties. Both sand filter layers and geotextiles can effectively maintain a clean separation between an open-graded aggregate layer and the subgrade. Also for a Separation Number of 4.81-83]. however. Sieve sizes greater than the No. a filter layer is probably not required if the bearing capacity safety factor is greater than 1. compressible geotextiles than for thinner ones. 70. The choice therefore becomes primarily a matter of economics.survivability are as follows: 1.78. 2.S. A wide range of both nonwoven and woven geotextiles have been found to work well as just separators [76. Where filtration is not of great concern. been found by Glynn and Cochran [84] to be considerably greater for thicker.3. Select from the upper part of Table 39 the appropriate column which the Separation Number N falls in based on the bearing capacity of the subgrade. and for a SEPARATION NUMBER of 3 or more it is not required if the safety factor is greater than about 1. Most geosynthetics when used as a separator will reduce stone penetration and plastic flow [84]. gradation and angularity of the aggregate to be placed above the filter.0. More care is perhaps required for the design of an adequate aggregate filter to maintain separation than is necessary for the successful use of a 237 . The reduction in penetration has. Figure 74 provides a simple method for estimating subgrade bearing capacity.7. the requirements on apparent opening size (AOS) can be relaxed to permit the use of geotextiles with U. Estimate from the bottom of Table 39 the SEPARATION NUMBER N based on the size. 3. Read the SEVERITY CLASSIFICATION from the top of the appropriate column. A separation layer is not required if the bearing capacity safety factor is greater than 2.

. cl . . ••C -I -It 7 1O ..1 1 0.•1 • • 11) D.1 .13 0 00 C ■1...-I t.i 0 .. ■0 ..2 A 'A m 7....... 7: ... HI 238 tl) N m • 1-4 0 0 VI E o j IJ > 11 C . .t .. 0 ..... E M . CO CI CL 1:. VI • . •:0 •.. 0 . 0 Z a.i ..1 . 'V .11 ID 0.. . •-■ 0 /14 x at .0 t..0 4 Cll. .. .. ..I... C . 1' ....-.-4 M1 1 '7 0 0 'I-. H 11 I X V ...-1 4.-4 1 01 IM M .. 0I F. ....0 0 •-.4.1 I r...n .-: C a) rb DO •rn c r. .5 '4 ...c . 2._. 0 .. Q > ml O 0. r..-. ..4 0„.• Z.4 .-• .: .. •1 •11 DJ .1 Z .. .1-' . •D3 m C •. .1 0 . ..1 .: cn a) •H yC O 11 t)-1 O O fi lt e r c r ite r ia s ma l le r op e n ing crl 0 0 CO 'D m DO '''•-• .0 O )./ 06 o . . .0 .. ... . 1 . I 0- .. ... ._4 4J D ..0 ... Y m --• O ..-• l 0 N c 1. 7. c 1...0 7: Mk-. )-. .... u.) •=4 C O 0 0 ▪ CO O O )-1 O O r-- T O r-1 V ti 0 L. al C 0 ..a t. M Z C 0 ... .5 ..... c M al .- . cV „ .I I..034 .) C) a) cu cu 4. .1 C v.1 ... -• •.7. CZ r"... C. t ..• - 4. 0 ..c c 3 n ct : "3 ..01 4.-.-. ...0 ... ... •11 n....../ 1 3 'C o .4 .e. C 0 'Z. 1-.1 .9 .-..+ IA . ... 4-. ... ■••• . 11 r1 .. .'....--.. Ve ry Se v e r e O 4-) .) 0 c o ndi t io n s..2 ..' --4 -1 .. r-I L.....0 -• ml •••••• CO . . ..-4 . 'n . ••-• 7.. CO 5 C. m •-• 0 5 . c. cn 4. ..i. . N 0 .•••■ ...-I > ....- '4 ..) O ..7• 1-' t 'f . 5 • ■••• 0 O ... . O W. 1.. .. O 0 E 0. NC I 4i... -'..s .0 > -.. 1. 11 IA . 1.• 0. ry r..1 P.. lj ....... •••0 0. 1 ' V...0 . • ) 0 11 H • •: 7.1 i-.u•1 W-3 ... MI a E-• HI 11. 1... 4.• . tj. r•-• V) re) . N cl) U1 $-+ 0 W O O 0 'A . II 0 --• 3. T. .. Deg r a da t io n a) large r s ie ve O ro r 0 0 .1 0 cu a) cd as cc) a) cu cn ro C..71) •-• 0.. . tc .I H •-• :11 4.. C al MI 1.• 7.n ..n . .. ••.-. CO 5 ."1 0 ▪ tJ 0 0 ro )-4 a. 7 1...

Filtration. Economics. (114 mm) diameter aggregates can punch through a thin.30 Moderate 15 . The geotextile selected based on filtration considerations should also satisfy the previously given requirements for separation/ survivability. Figure 84 can be used to quickly determine whether a geosynthetic is cheaper to use as a filter or separator than a sand filter layer. Estimate the pavement structural strength category from Table 41 based on its AASHTO structural number. The steps for selection of a geosynthetic for filtration considerations are as follows: 1. Enter Table 40 with the appropriate filtration Severity Level. Estimate the filtration SEVERITY CLASSIFICATION as follows: FILTRATION SEVERITY CLASSIFICATION FILTRATION INDEX Very Severe > 30 Severe 25 . et al. Add up the appropriate partial filtration severity numbers given in parentheses in each column of Table 42 to obtain the FILTRATION INDEX.25 Low < 15 4. 2. (50 mm) sand layer into a soft cohesive subgrade. and determine the required characteristics of the geotextile. 239 .geotextile. uncompacted 2 in. 3.5 in. [76] found that large 4. A granular filter layer having a minimum thickness of 3 to 4 in. (75-100 mm) is recommended. Bell.

of year (5) Medium (5) Firm (3) Silty sands. >60% sand.5-5 <5 >4.5 Approximate Vertical Subgrade Stress (psi) >14 Light 2. < 151 sand (10) Wet through year (9) Light (12) Soft (10) Low cohesion silts.25 14-9. Wet more than 3 me.25-4.5 9. Stiffer Stronger (0) e (-26) N 240 . Very fine sands (6) Periodically wet Medium to coarse sands and gravels (0) Rarely wet (21) 1 Heavy (3) .5 Heavy Table 42 Partial Filtration Severity Indexes CLASSIFICATION PAVEMENT STRUCTURE (Table ) SUBGRADE SHEAR STRENGTH (Table ) Very Light (20) Very Soft (20) SUBGRADE TYPE SUBGRADE MOISTURE CONDITION Dispersive Clays. clays. sandy silts (8) Frequently wet.Table 41 Pavement Structural Strength Categories Based on Vertical Stress at Top of Subgrade Category Approximate Structural Number (SN) Very Light <2.5-3. Law Cohesion Silts.5 Medium 3.

which are known primarily as polypropylenes and polyethylenes. When a geosynthetic is used as reinforcement for a permanent pavement. For severe separation applications. maintaining a high strength over a period of time for reinforcement would appear not to be as important as retaining the stiffness of the geosynthetic. the overall life may be as great as 40 years or more. the durability properties of the individual fibers may be 241 . The rate at which the loss occurs. can vary greatly between the various polymer groups or even within a group depending upon the specific polymer characteristics such as molecular weight. is generally significantly greater than required. Most mechanical properties of geosynthetics such as grab strength. maintaining strength and ductility would be more important than for most reinforcement applications. Also. chainbranching.DURABILITY PAVEMENT APPLICATIONS The commonly used geosynthetics can be divided into two general groups: (1) the polyolefins. which should be used for reinforcement. The geosynthetic. Therefore. and (2) the polyesters. Considering possible future pavement rehabilitation. Their observed long-term durabiilty performance when buried in the field is summarized in this section. (50-70 kN/m). a high level of stiffness must be maintained over a large number of environmental cycles and load repetitions. and specific manufacturing process employed. burst strength and tenacity will gradually decrease with time when buried beneath a pavement. additives. except when used for moderate and severe separation applications. Most flexible pavements are designed for a life of about 20 to 25 years. The strength of a stiff to very stiff geosynthetic. however. is subjected to forces that should not in general exceed about 40 to 60 lb/in.

Aging by ultraviolet light before installation. Changes in mechanical properties with time occur through very complex interactions between the soil. 2. Stiffness in some instances has been observed to become greater by Hoffman and Turgeon [107] and Christopher [108] as the geosynthetic becomes more brittle with age. Whether some geosynthetics actually become a more effective reinforcement with time has not been shown. as long as a safe working stress of the geosynthetic is not exceeded as the strength decreases. Some general characteristics of polymers are summarized in Table 43 and some specific advantages and disadvantages are given in Table 44. or from chemicals having an external origin such as chemical pollutants or fertilizers from agricultural applications. the ability of the geosynthetic to act as a reinforcement might improve with time for some polymer groups. 4. geosynthetic and its environment and are caused by a number of factors including: 1. Sustained stress acting on the geosynthetic which through the mechanism of environmental stress cracking can significantly accelerate degradation due to chemical micro-organisms and light mechanisms. Chemical reactions resulting from chemicals in the soil in which it is buried. 242 . As a result. Micro-organisms. 3.significantly different than the durability of the geosynthetic manufactured from the fibers.

etc.iron. Adsorption of Liquid.11111 Hsall Table 44 Summary of Mechanisms of Deterioration. manganese. 243 . zinc. zinc. Advantages and Disadvantages of Polyethylene. copper. lime and fertilizers Polypropylene Environmental stress cracking catalized by (2) an oxidizing environment. Environmental stress cracking is adversely affected by the presence of stress risers and residual stress. Environmental stress cracking Degradation due tc oxidation catalized by heavy metals iron. Physical properties in general should be evaluated of the geosynthetic which can have different properties than the fibers. environmental stress Degradation due to oxidation catalized by heavy metals . Oxidation Adsorption of Liquid Anti-oxidants usually added Good resistance to low pH environments Good resistance to fuels Susceptible to creep and stress relaxation.takes In water Good creep and stress relaxation properties Attacked by strong alkaline environment N Notes: 1. Anti-oxidants usually added Good resistance to low and high pH environments Susceptible to creep and stress relaxation. Oxidation. copper.G. Polypropylene and Polyester Polymers( 1 ) ) POLYMER TYPE MECHANISMS OF DETERIORATION GENERAL ADVANTAGES IMPORTANT DISADVANTAGES Polyethylene Environmental stress cracking catalized by an oxidizing environment.Table 43 General Environmental Characteristics of Selected Polymers RIMMOWDOWWW00000wo: Low u mk. manganese Degradation in strong alkaline environment such as concrete. 2.4. wroni. May be attacked by hydrocarbons such as fuels with time Polyester Hydrolysis .

and from one geosynthetic to another is almost impossible due to the very complex interaction of polymer structure and environment. wet-dry cycles. Hoffman and Turgeon [107] have reported the change in grab strength with time over 6 years. the basic mechanisms affecting degradation for each material under consideration must be understood. but gradually decrease thereafter. The reduction in fiber tensile strength in one series of burial tests has been found by Scotten [112] to be less than ten percent. After six years the nonwoven polyester geotextile 244 . Different environments including pH. Hence. In evaluating a geosynthetic for use in a particular environment. and chemical pollutants will have significantly different effects on various geosynthetics. heavy metals present.SOIL BURIAL Full validation of the ability of a geosynthetic used as a reinforcement to withstand the detrimental effects of a soil environment can only be obtained by placing a geosynthetic in the ground for at least three to five years and preferably ten years or more. One study has indicated that the strength of some geosynthetics might increase after about the first year of burial [107]. Long-term burial tests should be performed on the actual geosynthetic rather than the individual fibers from which it is made. geosynthetic structure and bonding can have an important effect on overall geosynthetic durability which has also been observed in other studies [113]. Translation of durability performance data from one environment to another. Relatively little of this type data presently exists. The geosynthetic should be stressed to a level comparable to that which would exist in the actual installation. The overall strength loss of the geotextile was up to 30 percent.

Temperature was held constant at 20°C. In apparent contradiction to this study. All geotextiles (except one nonwoven polypropylene) underwent a decrease in average elongation at failure varying up to 32 percent. A statistically significant decrease in burst strength was not observed over the seven year period for any of the 245 . Alternating cycles of wetting and drying were found to be particularly severe compared to other conditions.studied exhibited no loss in strength in the machine direction (a 26 percent strength loss was observed in the cross-direction). needle-punched nonwoven geotextiles were conducted by Colin.6. et al. Slit tape polypropylenes placed in aerated. The four polypropylenes exhibited losses of strength varying from 2 to 45 percent (machine direction). and 30 percent for polypropylene. [116]. moving seawater were found to undergo a leaching out of anti-oxidants if the tape is less than about eight microns thick [115]. The geotextiles were buried in a highly organic.6 geotextiles lost about 30 percent of their strength [114]. 15 percent for nylon 6. The test specimens consisted of monofilaments of polypropylene. moist soil having a pH of 6. they were not subjected to any significant level of stress during the study. Table 45 shows for these conditions the important effects that anti-oxidants.7. After one year of burial in peat. polyethylene and a mixture of polypropylene and polyamide-coated polypropylene filaments. geosynthetics exposed for at least seven years showed average tenacity losses of 5 percent for polyethylene. metals and condition of submergence can have on the life of a polypropylene. Since the geosynthetics were used as edge drains. Burial tests for up to seven years on spunbonded. hence these geotextiles became stiffer with time. but polyester and nylon 6. no loss in strength was observed for a polypropylene.

CM CM 5-1 4.- S-1 a) c./ • O 1-n o r-1 O cr) 4.-1 tn aJ r-f 0 (C1 246 CD U 4./ 4-1 w O 4-1 14.1 Z 4_i 'H a) O O .1.4 •.4 O O O cr) O O O c\I a) O O 0 ro O Cr) CM CM 1-1 a) 5-1 O O O 1-4 N C11 ..) aJ a) 4-1 r-I CI C.) G • • L1-1 cu f:4 O • C W N to o f--I ro .

This data was obtained from numerous sources including [107. polyester nonwovens were found to be quite susceptible to degradation. Polyester and polyproylene. are given on the figures for the 85 and 95 percent levels. Summary of Test Results. however. The loss of tenacity of a number of geotextiles buried under varying conditions for up to ten years in France and Austria has been summarized by Schneider [107.118]. became embrittled during this time.112. when buried for up to 32 months. NaOH. One polypropylene geotextile did indicate a nine percent average loss of burst strength.108. Both low and high density polyethylene. Confidence limits. underwent from five to as much as seventy percent loss in mechanical properties. 247 .116. showing strength losses of 43 to 67 percent for the polyesters compared to 12 percent for polypropylene [117]. did not undergo any significant loss of mechanical properties [118]. which admittedly are rather crude for this data. and about 30 percent after ten years of burial. Scatter diagrams showing observed long-term loss of strength as a function time are given in Figure 85 primarily for polyproylene and polyester geotextiles. Typically the better performing geotextiles lost about 15 percent of their strength after five years.117.108. The level of significance of the data was generally very low except for the nonwoven polypropylene geotextiles where it was 73 percent.116. in any of these materials. sunlight and burial. however.119]. When exposed to a combination of HCL. when subjected to stress in the field. Schneider [117] indicates geotextiles buried in one study for between four months and seven years.samples. Stabilizers were not used.117.

11 iMit s .11 SignE. Observed Strength Loss of Geosynthetics with Time.- 957 . • C.11 Signf. 9/ R.= 0. Age (Years) (a) 807 • fon . r . = 757 • as • • rJc • 0 0 • • t a 4.onfid en A .n 2. A• ce Limi ts Age (Years) (d) All Geosvnthetics Age (Years) Polyester Figure 85. 248 A 807 95Z 4 .11 Signf.2 r = 0. L ts 807 Age (Yclrs) Nonwoven Polvpropvleno (b) Woven and Nonwoven Polypropytene r = 0. • • On 807 . = 757 • • A • 952 A 807 • • • • • • • 11. • A as • • - • • • • • • • • *. = 737 95 .

With two exceptions. grab strength and tenacity. loss of strength was measured by a number of different tests including burst strength.In these comparisons. As a result. the polyester geosynthetics showed long-term performance behavior comparable to the polypropylenes. test methods and environments included in this data probably account for at least some of the large scatter and poor statistical correlations observed. only general trends should be observed from the data. The wide range of geosynthetics. the 85 percent confidence limit indicates a strength loss of about 30 percent. The results indicate after 10 years the typical reduction in strength of a polypropylene or polyester geotextile should be about 20 percent. 249 .

To evaluate the use of geosynthetics as reinforcement. Existing literature was relied upon for the separation. Geosynthetics are manufactured from polymers and include woven and nonwoven geotextiles and also geogrids which generally have an open mesh. Finally. (2) prestressing the aggregate base (and also as a result the 250 . the important question is briefly addressed concerning the durability of geosynthetics when buried for a long period of time.CHAPTER IV CONCLUSIONS AND SUGGESTED RESEARCH INTRODUCTION This study was primarily concerned with the geosynthetic reinforcement of an aggregate base of a flexible pavement. OVERALL EVALUATION OF AGGREGATE BASE REINFORCEMENT TECHNIQUES In studying new methods for improving pavement performance. Specific methods of improvement evaluated included (1) geotextile and geogrid reinforcement placed within the base. filtration and durability portions of the study. an analytical sensitivity study and large-scale laboratory experiments were performed on selected pavement sections. Separation and filtration is considered timely to include because of the present interest in employing open-graded drainage layers which frequently require a filter layer. Therefore. A geotextile reinforcement may at the same time serve the functions of separation and/or filtration. all important factors must be carefully integrated together to develop a realistic overall evaluation. In this study methods were investigated involving the reinforcement of an unstabilized aggregate base to be used beneath a surfaced flexible pavement. these aspects were also included in the study.

Type and Stiffness of Geosynthetic. Geosynthetics having stiffnesses much less than the above values would not have the ability to 251 .8 kN/m) for geogrids and 4000 lb/in. Comparative tests were not performed on nonwoven geotextiles which might have better reinforcing characteristics than wovens due to improved friction characteristics. A general assessment of the above improvement techniques is made including their overall benefit. important levels of improvement will only be derived for relatively light sections placed on weak subgrades or having low quality aggregate bases. it appears at this time that the minimum stiffness to be used for aggregate base reinforcement applications should be about 1500 lbs/in. GEOSYNTHETIC REINFORCEMENT BENEFITS The laboratory and analytical results indicate that geosynthetic reinforcement of an aggregate base can.9 kN/m) for woven geotextiles. Some specific findings from the study are as follows: 1. their relative potential. improve pavement performance with respect to both permanent deformation and fatigue. (4.subgrade) by means of pretensioning a geosynthetic. under the proper conditions. (1. The term geosynthetic as used in this study means either geotextiles or geogrids. The experimental results suggest that a geogrid having an open mesh has the reinforcing capability of a woven geotextile having a stiffness approximately 2. From the experimental and analytical findings.3-4. and an economic evaluation.5 times as great as the geogrid. and (3) prerutting the aggregate base either with or without geosynthetic reinforcement. In general.

For pavements constructed on soft subgrades. Placement of the reinforcement at the middle of the base will also result in better fatigue performance than at the bottom of the layer.effectively perform. 2. The experimental study showed important improvements of subgrade rutting when a very stiff geotextile was placed at the bottom of an extremely weak section. This would be particularly true if the subgrade is known to have rutting problems. For light pavement sections constructed with low quality aggregate bases. the reinforcement should probably be placed at or near the bottom of the base. the preferred position for the reinforcement should be in the middle of the base. Geosynthetic Position. Placing geosynthetics having the above stiffnesses within pavements would not be expected to increase the overall stiffness of the system as indicated for example by the falling weight deflectometer (FWD) or Dynaflect methods. Almost no 252 . and the base is of high quality and well compacted. The analytical approach indicated placing the reinforcement at the bottom of the base would be most effective in minimizing permanent deformations in the subgrade. even on weak pavements. The experimental results show that placing the reinforcement in the middle of a thin aggregate base can reduce total permanent deformations. as a reinforcement. particularly if a good subgrade is present.

3. 253 . Also. As the structural number and subgrade strength of the pavement decreases below the above values.5 to 3. The structural section in general should have AASHTO structural numbers no greater than about 2.5 to 3 if reduction in subgrade rutting is to be achieved by geosynthetic reinforcement. sections with asphalt surface thicknesses greater than about 2. Pavement Strength. 4. Strong pavement sections placed over good subgrades would not in general be expected to show any significant level of improvement due to geosynthetic reinforcement of the type studied. Light to moderate strength sections placed on weak subgrades having a CBR < 3 (E s = 3500 psi. (64-90 mm) would be expected to exhibit little improvement even if placed on weak subgrades. for a stronger section having a stiff geogrid at the bottom. the improvement in performance due to reinforcement should rapidly become greater.improvement was observed. Subgrade Rutting. compared to a soft subgrade being located on the bottom. and hence reduction of rutting in the subgrade would be harder to validate. 24 MN/m 2 ) are most susceptible to improvement by geosynthetic reinforcement. The possibility does exist that the geogrid may be more effective when aggregate is located on both sides. however. In these tests most of the rutting occurred in the base.5 in.

For light sections this corresponds to actual reductions in base thickness of about 1 to 2 in. Light sections on weak subgrades reinforced with geosynthetics having equivalent stiffnesses of about 4000 to 6000 lbs/in. Considerably more reduction in rutting occurs. 6. (4. Geosynthetic reinforcement of a low quality aggregate base can. however. The experimental results are inconclusive as to whether fatigue is actually affected less by 254 . (25-50 mm) for light sections. This is true for reinforcement locations at the center and bottom of the base. The analytical results indicate that improvements in permanent base and subgrade deformations may be greater than the improvement in fatigue life.5 to 3.9-7.5. The asphalt surface should in general be less than about 2. under the proper conditions. 7. when these improvements are expressed as a percent reduction of required base thickness. Low Quality Base.3 kN/m) can give reductions in base thickness on the order of 10 to 20 percent based on equal strain criteria in the subgrade and bottom of the asphalt surfacing. For weak subgrades and/or low quality bases. reduce rutting. for the thinner sections on weak subgrades than for heavier sections on strong subgrades. Improvement Levels. Fatigue. total rutting in the base and subgrade might under ideal conditions be reduced on the order of 20 to 40 percent.5 in. (64-90 mm) in thickness for the reinforcement to be most effective.

The fines content of aggregate bases should be kept as low as practical.reinforcement than rutting. and cost-effective method for significantly improving performance of light pavements constructed on weak subgrades. Prerutting may also be 255 . The optimum position of geosynthetic reinforcement from the standpoint of fatigue appears to be at the top of the base. geosynthetic reinforcement should not be used as a substitute for good construction and quality control practices. The experimental findings of this study indicate that prerutting is equally effective with or without the presence of geosynthetic reinforcement. permanent. Prerutting without a geosynthetic provides the potential for a quick. Improvement in fatigue performance perhaps might be greater than indicated by the analytical analyses. Finally. and compacting aggregate bases to a minimum of 100 percent of AASHTO T180 density. preferably less than 8 percent. The analytical results also show prestressing to be quite effective. however. fatigue life should also be significantly improved if the center of the layer is prestressed. Stress relaxation of a long period of time. Good construction practices would include proper subgrade preparation including proofrolling and undercutting when necessary. PRERUTTING AND PRESTRESSING Both prerutting and prestressing the aggregate base was found experimentally to significantly reduce permanent deformations within the base and subgrade. could significantly reduce the effectiveness of prestressing the aggregate base.

prerutting does not have the present uncertainties associated with prestressing an aggregate base. ECONOMIC CONSIDERATIONS Prerutting and Prestressing.found effective where low quality aggregate bases are used. Prerutting without reinforcement is cheap and appears to be quite effective. woven geotextile.. 256 . prestressing the aggregate base using the same geogrid would result in a total cost equal to at least 2 times (and more likely 2. or where reasonably strong pavement sections are placed on weak subgrades. including whether prestressing will prove effective over a long period of time.1 kN/m). at least with regard to reducing permanent deformations. Further. the total expense associated with prestressing an aggregate base would be on the order of 10 times that of prerutting the base at one level when a geosynthetic reinforcement is not used. Therefore. Full-scale field experiments should therefore be conducted to more fully validate the concept of prerutting. 2. Prerutting without reinforcement should give performance equal to that of prestressing. Further. The cost of prerutting an aggregate base at one level would be on the order of 25 percent of the inplace cost of a stiff geogrid (S g = 1700 lbs/in. and significantly better performance compared to the use of stiff to very stiff non-prestressed reinforcement. Recall that a stiff geogrid apparently has the equivalent reinforcing ability equal to or even greater than a very stiff. and develop appropriate prerutting techniques. The most promising potential method of improvement studied appears at this time to be prerutting a non-reinforced aggregate base.5 times) the actual cost of the geogrid.

Certainly full-scale field studies are needed to validate the findings of this study. Geosynthetic reinforcement may also be economically feasible for other combinations of structural designs and material properties where rutting is a known problem. Base = 10 in. particularly for subgrade rutting problems. the results of the entire study including the uncertainties associated with it should be integrated together considering the specific unique conditions and features associated with each design. E s = 3500 257 . the local inplace cost of aggregate base.5 in. or where low quality bases are used beneath relatively thin asphalt surfacings. General guidance concerning the level of improvement that can be achieved using geosynthetic reinforcement of the aggregate base is given in Figures 86 to 90 (refer also to Tables 29.. CBR = 3. The use of geosynthetic reinforcement is in general considered to be potentially economically feasible only when employed in light pavements constructed on soft subgrades. Figure 91 gives the relationship between the inplace geosynthetic cost (or the cost of some other type improvement).Geosynthetic Reinforcement. The results presented in this study were developed for specific conditions including material properties and methodology. This figure serves as an aid in evaluating the economics of using aggregate base reinforcement. 30 and 33). Consider as a hypothetical example.. and the corresponding reduction in aggregate base thickness that would be required for the reinforcement to be comparable in cost to a non-reinforced aggregate base. In estimating potential levels of improvement for a specific pavement. the economics of reinforcing a pavement having a light to moderate structural section constructed on a relatively weak subgrade (AC = 2.

---4 o 0 7 C o O C C 4.4 W 00 M G cr) e4 >.1. Ax le Lo a ds • 7: C. pi° Su r fa c e CBR = 3.4 CO Cv O ri (•11T) 17 'ssampTu asug uT uoT17npaH al •..anpall '0 CI3 ‘24 W M ti •7 • • .) C <4 0 L) m 0 ct w un CC M MCN E >CC 0 S. uiv.4 O — cn L cC cn C.1 co CC CU 0 FC1 w ct cwIl CC T 4-1 r-4 C to e4 pa o G 01 C.) •0 C. 0 e4 Bas e Thic k. U w Ctl • •?-1 3. • cc 0 CO rfa ce R -=3. E=3500 p s i W C.-.0 3 5 .C.1 "0 • . E=35 00r • un .7 Ctl 0 -.) CC • o CC Cl) E bD 258 ro O ) (•LIT) I.$ 0 CO 0 ▪ U W 4 O *..7 •-1 a) 1-1-4 O O 1 CU CC 1-1 CU 1-D 4.. 'ssaulDNI asvg Lir u0T.) 1.

* ■-•-■ . T( in.' 0 0 0 .. AC Su Su bg ra de CBR= Con t -.z. . 0 0 „c •uT) 17 . Re du c t ion in Granu la r a Fun c t ion o f Base Thic kness.■ 413 0 (•LIT) Iv 6ssompTql asug uT uoT1Dnp Stif fne ss for Cons tan t AC.r — . ..1 8kip Loa ds ig n : ..oPu =12 . 5 in.c IA A . 40 21 c■ CILI .7o ls s ire dBas e : k.0 0 CI 0 0 0 ■•■1 0 259 0 C 0 . ) rl '. Su bgra de 11 6.). .. 'ssampTyi uT uoTlanpau lit co oo to O . SDV ' u T • .

-4 20 Bre ak-Even Co s t o f 4-4 U 'A :I R. 14 '1* C 0 0 47 . T(in..4 .-4 col I i L U.I 0 . (O'CisluTI 0 a) CN 1E00 -1T7atiluXsoao aDeidu/ aS 0. CO 0 0 (•UT) iv 'ssampiu aseq u! uoti3npaH 260 Bas e Th ic knes s. ) bI Re duc t ion in Gra nu la r Base ic e Ston e Co s t ($ /ton ) I .

psi. with the amount of strain required depending upon the desired level of improvement and the stiffness of the geosynthetic. 64 mm.3 in. (33 mm). (25-33 mm). (4. If the geosynthetic is placed in the field so as to have slack or wrinkles.2 percent of the width of the geotextile can render it essentially ineffective. (21 kN/m). the required inplace cost of stone base to make the geosynthetic economically comparable to an aggregate base would be about $15 per ton. The geogrid should perform equal to or somewhat better than a very stiff woven geotextile based on the experimental results of Test Series 4. Considering fatigue might be improved more than the analytical approach indicates. From Figures 86 and 87. CONSTRUCTION ASPECTS Stretching Geosynthetic in the Field.19/m2 )) and performs about the same as a geotextile having a stiffness of 4000 lbs/in. The use of a grid reinforcement could help to decrease rutting.3 in. Further. then considerable deformation is required in the form of rutting before the strain is developed to mobilize sufficient tensile force in the geosynthetic necessary to make it effective. this aspect should not be overlooked making the final decision concerning reinforcement. From Figure 91. To remove wrinkles and irregularities. the geosynthetic must undergo strain. The results of this study show that to be effective as a reinforcement.9 kN/m). particularly if poorer materials were involved. a geogrid is to be used having a stiffness of about 1700 lbs/in. the geosynthetic should be stretched as tight as practical by hand during placement [42]. Then a 261 . Theory indicates that even a small amount of slack on the order of 0. 250 mm. Assume the geogrid costs inplace $1/yd 2 ($1. assume the allowable reduction in base thickness is 1. the reduction in base thickness should be about 1. 24 MN/m 2 ).0 to 1.

(12-30 m) ahead of the main area of fill placement between the fingers.special fork. (18 m) or more in width and requiring several feet of fill (Figure 92). Usually the fingers are extended out about 40 to 100 ft. The fingers of fill pushed out are typically 20 to 30 ft. (6-9 m) in width. (60-90 m) in length. A technique was previously suggested in Chapter III involving applying the pretensioning force to the geosynthetic by means of wenches and 262 . The geosynthetic is first spread out over an area of about 200 to 300 ft. should be used to at least lightly stretch the geosynthetic. and any necessary seams made. effective and efficient methods must be devised for pretensioning the geosynthetic in the field. This technique is particularly effective where soft subgrade soils are encountered in eliminating most of the slack in the geosynthetic. Pretensioning. and serve to anchor the two ends of the geosynthetic. A simple but relatively effective method can be readily used in practice for stretching a geosynthetic when used across a roadway or embankment about 60 ft. or other device. Use of a top plate on the stake is recommended to prevent a geogrid from lifting up off the stake.42]. and even may place a little initial stretch in the material. The geosynthetic should then be fastened down with wood or metal stakes to give the best performance and most uniform strain distribution within the geosynthetic [26. When fill is placed in the center area. particularly when a soft cohesive subgrade is present. If the aggregate base is to be prestressed. the resulting settlement stretches the geosynthetic. Fingers of fill are then pushed out along the edges of the geosynthetic covered area in the direction perpendicular to the roll direction. Wide Geosynthetic Widths. The material is rolled out in the short direction.

Geosynthetic Fill Placement Area Being Stretched by Fill Settlement on Weak Subgrade Figure 92. 263 . Placement of Wide Fill to Take Slack Out of Geosynthetic.

cables. (300 mm). The actual rut spacing used would be dependent upon the wheel configuration selected to perform the prerutting. On the other hand. Thus. perhaps spaced apart about 12 in. it is generally not practical to place geotextiles on windy days. Wind Effects. (200 mm) base thickness would be a good starting point. Prerutting is just an extension of proofrolling. 264 . additional wrinkling and slack may result in the material. Prerutting. If geotextiles are placed during even moderate winds. Prerutting of course could be performed at more than one level within the aggregate base. A moderate wind will readily lift a geotextile up into the air. (50 m) was found to be effective in reducing rutting in both the 8 in. Caution should be exercised to avoid excessive prerutting. (200 mm). Development of a total rut depth of about 2 in. however. Probably prerutting in the field an 8 in. geogrids are not lifted up by the wind due to their open mesh structure. Prerutting in the laboratory was carried out in a single rut path for a base thickness of 8 in. and hence can be readily placed on windy days [42]. (200 mm) aggregate base and also the subgrade. can only be developed and refined through development studies including field trials. For actual pavements it may very likely be found desirable to prerut along two or three wheel paths. Effective methods of pretensioning. Wind can further complicate the proper placement of a geotextile. Appropriate techniques for prerutting the aggregate base in the field need to be established.

Sand filters generally perform better than geotextile filters. Thick nonwoven geotextiles perform better than thin nonwovens or wovens partly because of their three-dimensional effect. 265 . low cohesion silts and clays. applied stress level and the size. angular open-graded aggregates placed directly upon a soft or very soft subgrade result in a particularly harsh environment with respect to separation.SEPARATION AND FILTRATION The level of severity of separation and filtration problems varies significantly depending upon many factors including the type subgrade. Separation problems involve the mixing of an aggregate base/subbase with the underlying subgrade. Separation problems are most likely to occur during construction of the first lift of the aggregate base/subbase or perhaps during construction before the asphalt surfacing has been placed. When an open-graded drainage layer is placed above the subgrade. Both woven and nonwoven geotextiles have been found to adequately perform the separation function. angularity and grading of the aggregate to be placed above the subgrade. moisture conditions. the amount of contamination due to fines moving into this layer must be minimized by use of a filter to insure adequate flow capacity. possibly turbulent flow conditions. The severity of erosion is greatly dependent upon the structural thickness of the pavements. Large. which determines the stress applied to the subgrade. Also. although satisfactorily performing geotextiles can usually be selected. When separation is a potential problem. dispersive clays and silty fine sands are quite susceptible to erosion. either a sand or a geotextile filter can be used to maintain a reasonably clean interface. A very severe environment with respect to subgrade erosion exists beneath a pavement which includes reversible.

Under favorable conditions the loss of strength of typical geosynthetics should be on the order of 30 percent in the first 10 years. because of their greater thickness. (50 N/m). Polypropylenes and polyethylenes are susceptible to degradation in oxidizing environments catalized by the presence of heavy minerals such as iron. zinc and manganese. Consideration should be given to the environment in which it will be used. geogrids may exhibit a lower strength loss. Limited observations indicate that some geosynthetics will become more brittle with time and actually increase in stiffness. geosynthetics should have at least a 20 year life. Guidance is also given in selecting suitable geotextiles for use beneath pavements.Semi-rational procedures are presented in Chapter III for determining when filters are needed for the separation and filtration functions. Polyesters are attacked by strong alkaline and to a lessor extent strong acid environments. Whether a sand filter or a geotextile filter is used would for most applications be a matter of economics. These procedures and specifications should be considered tentative until further work is conducted in these areas. For reinforcement applications geosynthetic stiffness is the most important structural consideration. Whether better reinforcement performance will result has not been demonstrated. For separation and filtration applications. Most 266 . The typical force developed in a geosynthetic used for aggregate base reinforcement of surfaced pavements should be less than about 40 lbs/in. they are also susceptible to hydrolysis. DURABILITY Relatively little information is available concerning the durability of geosynthetics when buried in the ground for long periods of time. copper.

Prerutting in the large-scale experiments was found to be both effective and also inexpensive. Geosynthetic reinforcement of light pavement sections constructed on weak subgrades having a CSR less than 3. filtration. or reinforcement can last for 40 or 50 years has not been demonstrated. The geosynthetic reinforcement of an unstabilized. particularly in the subgrade. Prerutting a non-reinforced aggregate base appears to have the best overall potential of the methods studied for improving pavement performance. shows some promise for reducing permanent deformations. and preferably less than 2.geosynthetics would initially be strong enough to undergo significant strength loss for at least 20 years before a tensile failure of the geosynthetic might become a problem for pavement reinforcement applications. Prerutting. 3. 2. Whether geosynthetics used for separation. low quality aggregate base appears to offer promise as one method for reducing permanent pavement deformation of pavements having thin asphalt surfacings. 267 . SUGGESTED RESEARCH Reinforcement The laboratory investigation and the sensitivity analyses indicate the following specific areas of base reinforcement which deserve further research: 1. Low Quality Aggregate Base. Weak Subgrade.

The recommendation is therefore made that additional an experimental investigation be conducted to further evaluate these three techniques for potentially improving pavement performance. 268 . Such a study would be applicable to mechanically stabilized earth reinforcement applications in general. ductility and chemical composition. This investigation should consist of carefully instrumented. A very important need presently exists for conducting long-term durability tests on selected geosynthetics known to have good reinforcing properties. Properties to be evaluated as a function of time should include changes in geosynthetic strength. Geosynthetic Durability. and buried in several different carefully selected soil environments. Admittedly. Soil environments to include in the experiment should be selected considering the degradation susceptibility of the polymers used in the study to specific environments. stiffness. A description of a proposed experimental plan for this study is presented in Appendix C. The geosynthetics used should be subjected to varying levels of stress. Separation/Filtration Important areas involving separation and filtration deserving further study are: 1. each geosynthetic product has a different susceptibility to environmental degradation. full-scale field test sections. Tests should run for at least 5 years and preferably 10 years.

and also at lateral edge drains.Nevertheless. The tests should probably be performed in a triaxial cell by applying cyclic loads as water is passed through the sample. 269 . At least 1x10 6 load repetitions should be applied during the test to simulate long-term conditions. 2. Filtration. A formal study should be undertaken to evaluate the filtration characteristics of a range of geotextiles when subjected to dynamic load and flowing water conditions likely to be encountered both beneath a pavement. a considerable amount of valuable information could be obtained from a long-term durability study of this type.

APPENDIX A REFERENCES .

Bell. W. Transportation Research Board. D... Kamel. J. Vol. 1987. "Fabric-Reinforced Aggregate Roads Overview". R.. . A. Knight. "Laboratory Testing of Fabric Interlayers for Asphalt Concrete Paving: Interim Report". January. Transportation Research Record 916.K... 671. Bonaparte.. R.L. D.. D. June.W. Report FHWA-RD-80-021. of Transportation.. Transportation Research Board. paper submitted to Transportation Research Board Annual Meeting. 1985. 64-75. 1983. Federal Highway Administration.J. "Heavy Duty Membranes for the Reduction of Reflective Cracking in Bituminous Concrete Overlays".C. and Epps. R. 4. Research Record No. 9..H. of Transportation. 1983. 6. Dept.. Penna. E.D.. Research and Development Bureau. Dept. 7. and Villet.. Bell. 1980. and Lai. J. "Reinforcement of Earth Slopes and Embankments". 10. Bureau of Bridge and Roadway Technology.. J. "Treatment for Reduction of Reflective Cracking of Asphalt Overlays of Jointed-Concrete Pavements in Georgia". 1982. J.. Barret. Robnett. "Geotextile Earth Reinforced Retaining Wall Tests: Glenwood Canyon. 1978. Report 113. April. Bender.. Colorado". 1983. N.S. "Use of Geogrids in Soil Reinforcement".S. N. C. D. Washington. 11.A. R. Research Project 79-6. pp. Q. "Test Methods and Use Criteria for Filter Fabrics". 2. Transportation Research Board.A. Transportation Research Board. U.A. Texas Transportation Research.. Dixon. "Field Evaluation of Fabric Interlayers". January. Transportation Research Board.R. Smith. Transportation Research Record 875. "Design and Behavior of SoilFabric-Aggregate Systems". August.. p. J. New York State Department of Transportation Engineering .C. and Barenberg. 1984. p. 4-5.. pp. J. 1983. 19. 37 p.I. Mitchell. Ruckman. et al. 3.K..E. Button.. 8.C.. Washington. 1984. 2. D. J. NCHRP Report 290. Gulden.. Transportation Research Record 916. 5. Frederick. Transportation Research Board. "Stress Relieving Interlayers for Bituminous Resurfacing".B. for presentation at the 62nd Annual Meeting. 1-6. 6-18. No. April.APPENDIX A REFERENCES 1. and Brown.

375-380. IBID. J. Conf. 13. 71-76.W... H. and Love. J.. II. pp. IBID. 15. Second International Conference on Geotextiles. Chaddock. Project Record 245. 19. Division of Applied Geomechanics.C. pp. 55-65. 93-98. "Deformation of a Haul Road Reinforced with a Geomesh". 21. p. Robnett. Gourc. R. Washington. J.A. II. 1981. 1983. . Vol. 1975. 1982.. pp. Q.W. "Development of Stresses in Reinforcement and Subgrade of a Reinforced Soil Slab". A. University of Nottingham. NSW". New York. Barksdale. 151-154. and Brodrick.E. Bell.V. D. and Phang. 1983.. "Geotextiles in Unpaved Roads: Design Considerations". Hughes.12. Transportation Research Board.C. Milligan.D.. Brown. Proceedings.P. Rigo.F. "Chargement Cyclic d'un Bicouche Renforce par Geotextile".. Proceedings. I.P. W. and Pefetti.. London. Second International Conference on Geotextiles. 1977.. "The Effect of a Fabric Membrane on the Structural Behavior of a Granular Road Pavement". 17. B. 24. J. Proceedings. "Report on the Construction and Performance of a Full-Scale Experimental Road at Sandleheath.W. 399-404. Greenway..C.R. G.. Haas. Hants". E. November..S.H.. CSIRO. "Construction and Analysis of a Fabric Reinforced Low Embankment on Muskeg". "Pavement Evaluation Project. Griffith. E. Part 1. A. D. J. Ruddock. Vol.L. A.H.. Transport and Road Research Laboratory. and McAvoy. Lai. J. Vol.A. 1982.R.. Melbourne. Riondy.. Potter. Las Vegas.R. pp. J. First Int. W. paper 4.. on Use of Fabrics in Geotechnics.J. 16. First Int. and Currer. Report LR 996.M. Potter.F. pp. Report submitted to Netlon Ltd and SERC. 22. Conference on Use of Fabrics in Geotechnics. 1982. Proceedings. P. CIRCIA. 417-422. 1.F. 45 p.2.. "Model Testing of Geogrids Under an Aggregate Layer on Soft Ground". 1982. S. 14..M.. 1979 (4th Edition).. MacMillan. 20. "INTRODUCTORY SOIL MECHANICS AND FOUNDATIONS. Sowers. J.. Research Record 949.. "Experimental and Theoretical Behavior of Geotextile Reinforced Aggregate Soil Systems". pp. Raumann.F.. G. 1977. and Vischerm. R. J. B. 23. 1985. "Grid Reinforcement of Asphalt Pavements and Verification of Elastic Theory". "Grid Reinforcement for Asphalt Pavements".O. C.. Petrix. G.. Perrier.B. 1984. Proceedings.. Project Report 2.. & Zeevaert-Wolf.. Pappin. 18. Vol. Halim.. Second Symposium on Unbound Aggregates in Roads. G.

... Thompson. U.. E.R. 341-346. "Deformation of Soil Foundations Reinforced with Prestressed Synthetic Fabric". S. "Investigation of Construction Techniques for Tactical Bridge Approach Roads Across Soft Ground". A. R.R. "Geosynthetic Reinforcement of Aggregate Bases of Surfaced Pavements". UIL-ENG80-2019. Halliday. Second International Conference on Geotextiles. Budanov. pp.. 1981.25. D. "A Full-Scale Experience on Granular and Bituminous Road Pavements Laid on Fabrics".. "The Performance of a Flexible Pavement Constructed on a Strong Fabric".F..F. "Reinforcement of Transportation Support Systems through Fabric Prestressing"...L. Report LR1123. October. Potter. II. and Palmeira. Fomin. Technical Report S-77-1. January. 1. 1977. 36.. paper presented at the 66th Annual TRB Meeting. Vol. Transportation Research Record 810. M. "Design and Use of Geotextiles in Road Construction". Vicksburg. Collins. "Mechanisms of Geotextile-Aggregate Support in Low Cost Roads". Taiwan. S. V. and Robnett. 29..I. and Stoll. II. Ruddock. A. January.G.. 1977. Barksdale. Transportation Research Board.D. and McAvoy. Barvashov.. 31. Vol.. Transportation Research Record 755. Sowers..A. Vol. August.R. S. First International Conference on Use of Fabrics in Geotechnics.F.J. 35.. Proceedings. 57-60. 1987. 27. pp.. Las Vegas.M. Ramalho-Ortigao.Aggregate Systems with Mirafi 500X Fabric". 26. Lai. Q. 49-51..A. Barenberg. Third Conference on Road Engineering Association of Asia and Australia..A.A. A-4 ..S. and Potter. and Watkins. 1981. J. C. Second International Conference on Geotextiles. Perkov. August. 6770.G. 1982. and Miller.S.R. 1982. V. "Geotextile Performance at an Access Road on Soft Ground Near Rio de Janiero".N.F. 34. 365-370. 32. Army Engineering Waterways Experiment Station. 1980. Vol. University of Illinois.. and Raad. L.. E. p.L. 1984. G.. Raad. Nevada.E. 30. A. II. 1987. February. J.. 33. Proceedings. 1982. J. R. E. Las Vegas. p. Vokas. and Pushkin. V. Proceedings. "Reinforced Elastic Layered Systems". paper presented at the 66th Annual TRB Meeting. Second International Conference on Geotextiles. J. "Fabric Used in Low-Deformation Transportation Support Systems".D. Mississippi.C. J. J. Proceedings. L. and Brown. Transport and Road Research Laboratory. pp. 28.. "Design Procedures for Soil Fabric . Proceedings. Webster.. 1980.

. Atlanta.F. Barker. 169 p. 39. 50-64. 38. Transportation Research Record 954. S.V... Ga.D. Atlanta. Constitution of Civil Engineers. 267 p... J.B.E. R.E. Barksdale. Transport and Road Research Lab.W. "Resilient Properties of Unbound Roadbase Under Repeated Loading". Brown. School of Civil Engineering. pp. "Crushed Stone Base Performance". Hannon. 1983. 43. 47. A.A. 1984. 541-563 . and Pappin. 1980. W. 1982.P... Paute. 48. Waterways Experiment Station... 1982. E.. J. 45-51.. School of Civil Engineering.. "Geogrid Reinforcement of Granular Bases". Mayhew. Sixth International Conference on the Structural Design of Asphalt Pavements. Penner.D. September.D. Christ.R. J. 1988. A-5 . P. Barksdale. "Open-Graded Bases for Airfield Pavements". 1988. B. "The Modeling of Granular Materials in Pavements". Transportation Research Record 1011. Transportation Research Board. 1985. and Pappin.P. PhD Thesis. "Finite Element Formulations for the Analysis of Interfaces....L.A. "Reinforcement at Asphalt-Granular Base Interface". and Prendergast. Nonlinear and Large Displacement Problems in Geotechnical Engineering". July.37. February. Jouve.D. R. "Analysis of Pavements with Granular Bases".F. and Todres. Transportation Research Board. R. Nokes. 42. "A Field Study of the Performance of a Tensar Reinforced Haul Road". W. Jones. and Ragneau.W. and van Hulst. pp. 40.. 46. R.. 49.. H. J. 1981. Barksdale.. Forsyth. H. pp. paper presented at the 67th Annual Meeting. Proceedings. 73.. Brown. 1987. "A Study of Factors Affecting Crushed Stone Base Performance".. Presented to Roads and Transportation Association of Canada Annual Conference. Georgia Institute of Technology. Misc. Georgia Institute of Technology. January. Ann Arbor. Walls. pp. pp.. August. Brown. Report LR 1088. Final Report.M. van Grup. 173 p. J... S. "Use of Nonwoven Fabrics in Permanent Road Pavements". Georgia Institute of Technology. Martinez. Transportation Research Board. S.A. 1985.. Vol. Part 2. 41... Proceedings.. Transportation Research Board.F. R. R. 1986. 17-22. Zeevaert. Transportation Research Record 810. R. Paper GL-86. Sept.S. paper submitted to Journal of Geotextiles and Geomembranes. 44. C. Haas.C. J. 45. and Brodrick. "Incremental Design of Flexible Pavements". 7887. "Rational Model for the Flexible Pavements Deformations".M. Vancouver. 1985. School of Civil Engineering.. A. J.

Brown. B. 61 Brown. 1977. "Theme Lecture: Pavement Design and Materials". 57. "Design Practice for Bituminous Pavements in the United Kingdom". 1987.. Final Technical Report to ERO U.D.W. 161-174.D. 1975. 1973. "Methods of Testing Soils for Civil Engineering Purposes". R. submitted to Transportation Research Board. "Performance of a Thin-Surfaced Crushed Stone Base Pavement". 51 Brown. Bison Instrument. Research Report to•Scientific Research Council. Brodrick.. II (in publication).F. A.M.. January.W.. 1972.. 62.. Vol. S. N.S. and Machemehl. Lofti. pp. Brown. Schwartz. "Specification for Rolled Asphalt (hot process) for Roads and Other Paved Areas". "Developments to the Nottingham Analytical Design Method for Asphalt Pavements". pp. Greene.A. B. pp.. Sixth International Conference on the Structural Design of Asphalt Pavements. S. "Compaction Specification for the Control of Pavement Subgrade Rutting". 59.50 Barksdale. England. C. C. 58. Barksdale. Proceedings. 1976. 220-231.M. J. and Brunton. Ann Arbor. "Laboratory Evaluation of Rutting in Base Course Materials". Sixth International Conference on the Structural Design of Asphalt Pavements..V. 1987.F. R. and Pappin. B. BRITISH STANDARDS INSTITUTION..V. Bush... 55. BS594..D. University of Nottingham.. 54. J. DEPARTMENT OF TRANSPORT. Army. and Witczak. "Specification for Road and Bridge Works".. 60 BRITISH STANDARDS INSTITUTION. Newcastle. Lister. "Permanent Deformation of Flexible Pavements".F. BS1377. Proceedings. Proceedings.. 1987. 52 Brown. "The Performance of Stress and Strain Transducers for Use in Pavement Research". 366-377. and Powell. R. August. H. ASTM Symposium on the Implications of Aggregate. Ann Arbor.W. M.F. and Brodrick. Conference on Measurements in Civil Engineering. and Brodrick. and Barksdale.. Sixth International Conference on the Structural Design of Asphalt Pavements.. New Orleans (submitted for publication).. Soil Strain Gage Model 410A". 1973.W. 3rd International Conference on Structural Design of Asphalt Pavements. 1980. August. University of Nottingham. 1987. BISON INSTRUMENT INC. 56. HMSO.F.D. R. United Kingdom.D.. London.. "Instructions Manual. S.V. S. W. "Stress and Strain Measurements in Flexible Pavements". 53... . S.

University of Illinois. 1982. N. 72.P. 3. pp. "Thickness Design for Effective Crushed Stone Use". 6. 1773-1775.P. J. April. June 1. "A Study of Typar Non-Woven and Other Fabrics in Ground Stabilization Applications". Barenberg. 65. 1985.. Ground Engineering. August.W. Scullion. R. Collois. of Civil Engineering. "Field Evaluation of Geotextiles Under Base Courses . 1987.K.. "The Shear Strength Behavior of Certain Materials on the Surface of Geotextiles". Geosynthetic '87 Conference. Eleventh International Conference on Soil Mechanics and Foundation Engineering".E. 74. pp.. "Experimental Friction Evaluation of Slippage Between Geomembranes.. Geotextiles and Soil".. Arlington. National Crushed Stone Assoc. 1986.. Sarsby. Portland. Oregon. Proceedings.L. E. Vol. 71. Jewell.. and Giroud.. on Crushed Stone. and Lai. and Dubois.. No. 21-112. Martin. T. "Evaluation of Interface Friction Properties Between Geosynthetics and Soil". Barksdale.M. School of Civil Engineering. Proceedings. New Orleans. 191-196.S.. October. T. 66.. P. April 17. Georgia Institute of Technology. pp. and Brown.A. Research Report. 1801-1804. J.D..63. Oregon. Geotechnical Testing Journal.. Research Report 414-IF (Suppliment). pp. pp..P. Milligan.W. "Interaction Between Soil and Geogrids". J. 1984. 1984. R. Koerner. 1985. Bell. 3. Ingold.. Formazin. Proceedings. Q.S.D. 1984. 18-29. April 17. 100-111. and Budiman.. ASCE National Convention.. 73. November. "Laboratory Pull-Out Testing of Grid Reinforcement in Sand". August. and Whitty. A. 616-627.. "Interface Response of Geotextiles".. and Houlihan. 3. Ingold. . J. VII-1 through VI-32. Thomas Telford. J. pp. T..E. Proceedings. San Francisco. Williams. M. 69.F..S. "Soil Fabric Friction Testing". Robnett.. ASCE National Convention. 67. J. Portland. 70. Polymer Grid Reinforcement.. E. Sept. D.Suppliment". 1980. 68. 80-177.. 1980. 53-73. J. Conf. S. Denver. 1983. "A Laboratory Investigation of Soil-Geotextile Friction". C. "The Use of Geotextiles for Soil Improvement". 1981. Delmas. Dept. G.A. Eleventh International Conference on Soil Mechanics and Foundation Engineering.. D. GTJODJ.. Texas Transportation Institute. International Conference on Geomembranes. Vol. pp. "Modeling of Effects of Moisture and Drainage of NJDOT Flexible Pavement Systems".. J. J. Vol. R. 1984. Goore. and Batereau. Saxena. San Francisco. pp. and Chou. pp.. 64. 75.S.. R.

D.. 435-440. Highway Research Board. McCullough.P.. Bell. pp. 44..J.76. Third International Conference on Geotextiles. S. pp. Transportation Research Record 849. "Laboratory Investigation and Field Studies of Channeling and Pumping".W. Wehr. 1-12. Proceedings. University of Nottingham. Vol. p.. Geosynthetics. and Yoder. Potter. Proceedings. A. No. "Separation Function of Non-Woven Geotextiles in Railway Construction". Transport and Road Research Laboratory. J. 75 p. Civil Engineering. "Long-Term Behavior of Geotextiles Installed in Road Constructions in Finland Since 1973". H. 967-971.. E.M. 77. Dawson. and Currer. Session Conference on Engineering Materials. 84.. Proceedings..M. "The Effect of a Fabric Membrane on the Structural Behavior of a Granular Road".. and Tayabji.... E. McCullough. Series 10. September. B. . 1978. 88. 1984. "Geotextiles in Road Foundations". 7882. pp. 259-276. "The Effect of a Fabric Membrane on the Structural Behavior of a Granular Road". Vienna. 83. Dempsey.F. "Evaluation of Typical Pavement Drainage Systems Using Open-Graded Bituminous Aggregate Mixture Drainage Layers". Engr. G. and Cochrane. Las Vegas. Glynn. "Research on Geotextiles for Heavy Haul Railroads". Vag-och Vattenbyggaren 7-8.J.S. 77 p. S. and Currer. 355-365. J. 1980 (in English).R.. A. L.D.R. 1958. Transport and Road Research Laboratory. Barenberg..F. "The Behavior of Geotextiles as Separating Membrane on Glacial Till Subgrades"..J. University of Illinois. Second International Conference on Geotextiles.... "An Experimental Investigation of Sub-base Protection Using Geotextiles".H. K. H. Transportation Research Board. 86. UILU-ENG-74-2009..A. Potter.T.R.. NSW. Pavement. 1981. Snaith. 80. Sept..I. E. Transp. Australia. 9.L.F. La. 87.H. W. "Clay Contamination in Crushed Rock Highway Sub-Bases". 1981. 1986. 1987. 1974.W. New Orleans. 78. Research report to ICI Fibres Geotextiles Group. M. Bell. "Water: Key Cause of Pavement Failure". 85. Raymond. Cedergren. H. and Brown. 1974. A.P. Volume 21. E. Austria. L.. and Godfrey. 79. pp.. p.. Proceedings. J. Canadian Geotechnical Journal. 1981. Proceedings. February. "Effect of Base Course Gradations on Results of Laboratory Pumping Tests". S. 82. Rathmayer. 81. TRRL Report 996.. 1984. Chamberlin.. and Gregory. 1982. Pavement. TRRL Report 996..

Saxena. 468-492. Geot. 36. 134-135. A. Vol.. 466-468. and Bell. 31-51. pp. L.J. 1983. Proceedings. D. 9th European Conf. Discussion of "An Experimental Comparison of the Filtration Characteristics of Construction Fabrics Under Dynamic Loading". pp. "The Filtration Behavior of 'Construction Fabrics Under Conditions of Dynamic Loading".. 1986..H.. and Dawson. 1985.. 2. "A Theoretical and Laboratory Investigation of Alternating Flow Filtration Criteria for Woven Structures". A.... Schober. and Stiffens. 1. 97. 4763. Geotextiles and Geomembranes. 1979. Highway Research Board. 93. Highway Research Board. "A Study of Interactions of Selected Combinations of Subgrade and Base Course Subjected to Repeated Loading". 1984. and Hsu.. and Yoder. Proceedings.. Brighton. pp. R. 1983. pp.J. "Geotextile Filter Criteria". England.S.L. and Yoder. pp. Hoare. Carroll. Research Record 39. Snaith. T. Vol. Disc. Ingold. 28.. M. T. E. 1987. D. Transportation Research Board. Brown.S. S. No. and Wittmann. J. D.J. Proceedings. 90. 32-37. 1958. "The Effects of Groundwater on Pavement Foundations"... No. 31-45. Heerten. Havers. Vol. 98. Geotextiles and Geomembranes.. Vol. 92.S.89.G.. Janssen. and Teindl. 1983.. 1985. . 100. 94. G. "Filter Criteria for Geotextiles". pp.K. Geotechnique. 96. Geotechnique. 4.. International Conference on Design Parameters in Geotechnical Engineering. 95. "Effects of Repeated Loading on Gravel and Crushed Stone Base Course Materials Used in AASHO Road Test". Geotextiles and Geomembranes. "Filtration Properties of Geotextile and Mineral Fillers Related to River and Canal Bank Protection". pp. Transportation Research Board. 657-660. 37. Highway Research Board.. Transportation Research Record 916.T. 1963.. on Soil Mechanics and Foundation Engineering.. 1957. E. Washington.J. 693721. pp. H. "Pore Pressures in Base Courses".S. Transportation Research Record 916. S. Barber.A. Vol. "Permeability of Geotextile-Included Railroad Bed Under Repeated Load". 4. 443-478. 2.F. G. "Dynamic Test to Predict Field Behavior of Filter Fabrics Used in Pavement Subdrains". 2.R. J. 34.C. 99.. Vol. 101. Hoare. p. 91. pp. W. Haynes. Vol. Vol. E.

"The Durability of Tensar Geogrids". "Geotextile Use as a Separation Mechanism".. 1987. 1986. and Raymond.C.S. Correspondence and Unpublished Report. on Soil Mechanics and Foundation Engineering. New Jersey Department of Transportation Report No. Congress Papers. A. Nonwovens for Technical Applications (EDANA). "Instrumented Case Histories of Fabric Reinforced Embankments over Peat Deposits". Brussels. L. GT1. R. 79-88. 1976.F. 150 p. Proceedings. B. M. Christopher.R. GT4.S.. ASCE... Federal Highway Administration. "The Effects of Groundwater on Pavement Foundations". B. Dunnigan. 175 pp. 1986. Transportation Research Board. B.. June 1983. Vol. S. 53-63.P.N. N.L. 104. pp.. Transportation Research Record 916.L.F.. 103.. August. J.102.. Sotton. and Decker. "Long-Term In Situ Properties of Geotextiles". 102. pp. Vienna. III. R.D.. Vol. January. Civil Engineering. Gerry. Netlon Limited. No. 107.. pp. III..D. Vol.R. Transportation Research Board.P. 1983. . Sherard.W. Austria. J.19. 111. May. 102. Barsvary. April. G. R. Dawson.L. G. 1984. G. A. 105. and Holtz. "Evaluation of Two Geotextile Installations in Excess of a Decade Old". 89-93. "Improved Drainage and Frost Action Criteria for New Jersey Pavement Design". and Turgeon.. and Steele. 9th European Conf.E. Sherard. GTJODJ.. Phillips Fibers Corp. Hoffman. pp. "Equivalent Opening Size of Geotextiles". Vol. Oregon State University.. Proceedings. G. 112. 287-301. 1982. Decker.. Geotechnical Testing Journal.. Vol. 1976.. Index 81. Christopher.K. 1985. and Brown. "Pinhole Test for Identifying Dispersive Soils". pp. E. 2. "The Role of Geotextiles in Controlling Subbase Contamination". 109.S. March. Dawson. 69-85.. 1986. Seneca. B. Wrigley. Strobeck. England. 84-015-7740.. 110. "Identification and Nature of Dispersive Soils". Second International Conference on Geotextiles. "Geotextile Engineering Manual".. 2. pp. Las Vegas. 647-652.P. R. M. Transportation Research Record 916. 1984. Proceedings. Jorenby.. TRR84-4. April.. Third International Conference on Geotextiles. 114. 657-660.R.. A. 106.. 115. 593-598. Dunnigan. S. "Long-Term Durability". 6. 1981.. 108. L. Kozlov. 16. ASCE. Draft Report. and McLean. 1983.S. pp. Vol. 113. pp.

553-558. 1986. Colin. 119. Journal of Applied Polymer Science. Mitton..J. "The Effect of Soil Burial Exposure on Some Geotechnical Fabrics".. Carlsson. Office of the Chief..J. 1986.116.L. M.. Las Vegas. 117. Colin.J. and Fayoux. Cooney. Barenberg. May.. D. D. 26. Proceedings.T. 1982. . and Wiles. D. Carlsson... LeClerc.M. 120. Proceedings. 1981. July. pp. Conference on Geotextiles.. August. "Some Answers Components on Durability Problem of Geotextiles". CW-02215. Vol. H. J. "Effects of Moisture and Drainage on Behavior and Performance of NJDOT Rigid Pavements". "Civil Works Construction Guide Specifications for Geotextiles Used as Filters". 77-84. Research Report. III. 1985. Second International Conference on Geotextiles. and Wiles. D.M. 118. Dept. Paute..D. pp. Singapore. 1982. Schneider. pp. Department of the Army. 3. p. D. Geotextiles and Geomembranes. "Durability of Geotextiles". Civil Works Construction Guide Specification. M. of Civil Engineering... 509. March.. Vol. 121. Vol.. Sotton. G. J. University of Illinois. B. 60-75. G. E..

.

APPENDIX B PROPERTIES OF MATERIALS USED IN LARGE-SCALE PAVEMENT TEST FACILITY .

The work carried out by Loach (B-4) on compacted samples of Keuper Marl was of most relevance to the current project.B-2) and also as the subgrade in the PTF (B-3). has been used extensively at Nottingham in earlier research projects on repeated load triaxial testing (B-1. The basic material properties of Keuper Marl used in the current project is given in Table B-1. This indicates the relationship between resilient modulus and CBR for compacted samples of Keuper Marl and clearly shows the influence of shear stress on the relationship (i.APPENDIX B LABORATORY TESTING OF MATERIALS GENERAL An extensive laboratory testing program was carried out to characterize all the pavement material used in this project. a few index tests and four repeated load triaxial tests were carried out on samples of material used during the project in order to characterize the particular index and mechanical properties. the nonlinear stiffness characteristic of the soil). The tests were carried out in accordance with either (1) existing ASTM and British Standards. One result obtained from Loach's tests is shown in Fig. or (3) established and published testing procedures adopted by individual laboratories (for the cyclic load triaxial test). Despite the large amount of data accumulated from previous tests on Keuper Marl. . Tests on Silty Clay Subgrade The silty clay. known as Keuper Marl. B-1. (2) tentative standards and procedure in their proposal stage (for the geosynthetics).e..

11 W u1 4 a)..) U. 4. (4_1 4. 4_1 0 . 1113111S38 B-3 1.1-1 co Cr CI_ `-"r UW X (r) X O cr O O C4.1 I I (r) C- II O 0 0 CO O rC5 cOn co 0 '7' tr) tr) CC CrJ 0 • 10 (edw) sninook.7t cc • ' 6cr .

Table B-1. .5 * According to British Standard 1377 (B-8).(%) 18 Liquid Limit (%) 37 Plasticity Index 19 Maximum Dry Density* (pcf) 117 Optimum Moisture Content* (%) 15. Unified Soil Classification CL Specific Gravity 2.69 % Clay 33 Plastic Limit . Results of classification tests for Keuper Marl.

Loach also carried out repeated load triaxial tests on compacted 3 in.3 (Croney) Er = Elastic Stiffness in kPa qr = Repeated deviator stress in kPa A = 2740 B = 2. The results of these tests generally conformed with those obtained by Loach who suggested the following equation to model the elastic stiffness of compacted Keuper Marl: Er B qr u+ctri ) A.35 psi (0 to 30 kPa) and 0 to 10. Using a similar procedure to that adopted by Loach and with the aid of a computer-controlled servo-hydraulic testing system. respectively.1 .B-6. The ranges of cell pressure and repeated deviator stress he used during these tests were 9 to 4.Cyclic Load Triaxial Test. It has been found (B-5. Therefore. four additional tests were performed on recompacted samples obtained from the pavement test sections. (76 mm) diameter cylindrical samples of Keuper Marl. Loach (B-4) carried out some soil suction tests on samples of compacted Keuper Marl at their original moisture contents using the Rapid Suction Apparatus developed at the Transport and Road Research Laboratory (B-9). in order to determine the general resilient properties of Keuper Marl.( q - r where: u = suction in kPa p = cell pressure in kPa a = 0. B-2.B-7) that relationships exist between soil suction and elastic stiffness for saturated and near saturated clay.15 psi (0 to 70 kPa). The results of his tests are shown in Fig. a series of soil suction and cyclic load triaxial tests are required.

X x O O U CY) (N. (Ad) uo!401-1S .

2 to 10. hence allowing permanent strain to accumulate. therefore. The permanent strain tests were carried out at a frequency of 4 Hz and with a 2 second rest period. B-3. The increase of permanent axial and radial strains with number of cycles for the tests are summarized in Fig. Comparison of the index properties between Bell's soil and the one used in the current project showed them to be similar.2 psi (15 to 70 kPa) were used. 2) A permanent strain test where the stress path was considerably closer to the failure condition. sieve analyses and other ■. For the permanent strain behavior of Keuper Marl.8 kPa) and repeated deviator stresses in the range of 2. Each cyclic load triaxial test was subdivided into: 1) A resilient strain test where the stress paths were far away from failure with the resulting strain essentially recovered during unloading and. Cyclic Load Triaxial Test.■ index tests. A cell pressure of 0.Both A and B are constants derived from experiments. The results of earlier testing showed that resilient behavior of a granular material under repeated loading was very stress dependent and. Tests on Granular Base Material Laboratory tests performed on the granular materials consisted mainly of cyclic load triaxial tests. A total of six tests were carried out on recompacted 6 in. (150 mm) diameter samples of the two types of material at various moisture contents. Details of procedure and equipment for carrying out cyclic load triaxial tests on granular material were described by Pappin (B-10) and Thom (B-11). compaction tests.26 psi (1. the results obtained by Bell (B-3) was found to be the most applicable. .

B-8 ._ -- 30 04 q = 30 .....Z1 CycTES ''.moisture content 17-4% 5 o Scatter Ver t ica l Perm a nen t S train q q = 70 in 141\l/m2 0 3 50 2 1 Per man e ntStr a in (% ) q= 100 101 •41111111 --- q =15 2 - lk -.. Permanent axial'and radial strain response of Keuper Marl for a range of stress pulse amplitudes (after Bell). \ q = 50 q = 70 Figure B-3..

Fig. For permanent strain tests. the results of the resilient tests are summarized in Table B-2.nonlinear.13(q/p') 2 } where P' = (aa + 2o c ) q = 1/2(a a . K. each of the six tests used 20 stress paths. as a function of both p'. a cell pressure of 7. Up to 2000 stress cycles at a frequency of about 1 hz were applied to the test samples. The results of the resilient strain tests were interpreted by means of Boyce's model (B-12) which expressed the bulk modulus. The dry densities of the test samples are shown in Table B-2. Hence. to characterize resilient strain. Figure B-5 indicates that the sand and gravel has a rather low resistance to permanent deformation. and q.n ) K = K1p' (1-n) /(1 . Based on the above equations. B-4.3 psi (50 kPa) and a repeated deviator stress of 0 to 20 psi (0 to 200 kPa) were used.G1.n and PI are constants to be determined by experiments. B-5 and B-6. The ranges of repeated cell pressure and repeated deviator stress used in the tests were 0 to 36 psi (0 to 250 kPa) and 0 to 29 psi (0 to 200 kPa). the deviator stress. The results are presented in the form of change of permanent axial and radial strains with the number of stress cycles. The results for the permanent strain tests for the two types of granular material are shown in Figs. The equations which Boyce used in the interpretation of results are as follows: G = G i p' (1. For the dolomitic limestone. the mean normal effective stress. as shown in Fig.ad and K1. and the shear modulus. B-6 indicates that the rate of development of permanent deformation varies with B-9 . respectively. G.

- . Stress paths used in cyclic load triaxial tests for granular materials.Stress Paths for Elastic Stiffness Testing le Stress Path for Plastic Strain Testing Deviator Stress (kPa) 200- 100- 200 100 Mean Normal Stress (kPa) Figure B 4.

- Permanent axial and radial strains response of sand & gravel during repeated load triaxial test.c3 = 50 kPa r = 200 kPa - /// 6 / // .. B-11 .E) \ \ c' L\ 0 10 1 100 6 b) RADIAL..9 a) AXIAL STRAIN a // .STRAM1 1 /0 5 a 3 2 .-4D-----12 1 a0I 1 100 10 NUMBER OF CYCLES Figure B 5.--.

--..--R-----. .4 2..0 w 4 0 27.1 --s-8.0 51. 10 ---' --.7.4 27. _ : 4 o 40.3. I eC P-4 0.....0 ------.5 10 100 1000 w(%) Sr (%) _.--- e.e.--av- 7 - ----0-- .:4P- 0. 0 • ..7 1 a c = 50 kPa qr = 200 kPa 94.5 = .-v 1.7 63.-- 0.0 ____I -4-6.3.1 :_ -4 rH 2...3 40..2 -4-6.4 94.8 10000 b) RADIAL S1RAIN 1. B-12 ..3 _.5 _Ei.-6. 1 10 100 1000 10000 NUMBER OF CYCLES Figure B-6...5 a) AXIAL STRAIN 1 ep / II / / .0 w(%) Sr(%) 2.. Permanent axial and radial strains response of dolomitic limestone during repeated load triaxial test at various moisture contents (w) and degree of saturation (Sr)._.2 6 0 51.2 63.0 1 2..0 il 0.2 -.3.5 1.5 .8 1.

. • c•-• cr) 0 0 0 0to0 C*1 cc') 4J 4-7 un 0 N P1 co cs) v!) te CO N M CO 0\ CV r.) }4 4.7 $2) 0 .LT)■ (1) SA • i-I (.1. a) a) 0 / 0 '4' 4.) rc ca u) al )4 ..4 .) 0) a) C.) N N -I • CY) • M • Cr) • 0 Lin cn •H 1.1 g u) wo O ''1 ): rip.11 N •Cl" r4 k0 CY) H • CO cs% cn cn M c) • un CO 0 I--i O CO.H 0 0 a) a) 4.. CY) CY) CV 84 s) ro c) LS 3-4 cl) ›rc5 ( (.) C9 '18 19). 9$.-I2 `91-1 • H rFJ g u) O)4 g in in U 14 r-I CNJ Cn Ul .• • Cr) M M m Cr) Cr) un (11 p 4-) 4.

(170 mm) high. while the PI of the fines for the dolomitic limestone was found to be 3 percent. very rapid increase in the rate of deformation will occur. the index varied from 3. The result of the test indicated an index of 9 percent overall while for individual size fractions. Compaction Tests. Tests on Geosynthetics Large Direct Shear Box Tests. The shear box used for these tests measured 11. Two plasticity index tests were carried out for the fines (less than 425 micron) of each of the two granular materials.7 in. The results of the tests for the two materials are shown in Fig. One flakiness index test BS812 (B-14) was performed on the crushed dolomitic limestone used in the third series of tests.8 in. Index Tests. B-7. In general. the British Standard Vibrating Hammer method (B-8) was adopted.moisture content and as the material approaches saturation. the same material was used in both the upper and lower half of the shear box. Twenty-four large direct shear box tests were performed on the two geosynthetic materials in conjunction with the soil and granular materials. Details of the tests and the results are shown in Table B-3 and Fig. In each test. the test was carried out according to the ASTM D-1557 test method (B-13) while for the dolomitic limestone. A series of compaction tests were carried out in order to determine the optimum moisture content and maximum dry density of the compacted material. For most of the B-14 . For the sand and gravel. The fines for the sand and gravel were found to be non-plastic. Compaction was carried out by using a hand-held vibrating hammer.1 percent. (300 mm) square by 6.8 to 16. respectively. the moisture content and dry density of the material at the time of the large scale pavement test were simulated. B-8.

/"...7 0 G SAND & GRAVEL 2. uses the British Standard vibrating hammer test method (B-14). .•145 ZERO AIR VOID 140 ■ G =2 S IX)LaLITIC LIMESTONE ._..6 S= L.._.) 10 12 Note: Sand & Gravel are compacted according to ASTM D-1557 test method (B-13) while dolomitic limestone.-..11/:7--: 125 '''lit 120 0 a 3 MOISTURE CONTENT (7. Results of standard compaction tests for the granular materials. Figure B-7.

.\\■ 0 N 0 Lrt s \\ i c) \ \\ \\ \ \ I 8 8 0 f—i 0 r-1 0 0 0 kr) 0 0 0 0 8 0 r- 0 0 In S AND & GRAVEL 8 0 8 .----- ■11111 1 0 t 0 N zulAPO SSMIS IBMS fl EDAM 8 0 . Ai 1 Ains [ \ :OTEXTILE 4. )11.

30 22 23 24 Keuper Marl 105 107 108 16.14 .30 .2 0.07 2.4 0.06 .06 .27 2.95 1.10 0.7 5.15 2.55 1.30 19 20 21 Tensar/Keuper Marl 106 109 111 16.75 1.54 1.5 16.3 0.38 0.18 0.10 2.2 16.6 16.21 0.47 0.75 .2 3.62 1.3 16.48 0.30 .4 3.0 4.0 0.30 .65 1. Summary of large shear box tests.06 .06 2.30 Nicolon/Limestone 138 137 138 5.06 13 14 15 Crushed Limestone 138 140 138 5.55 1.30 U1 MIN 140 138 138 Cs1 MN Nicolon/Sand&Gravel 1 r--GOcrt .39 .3.35 0.7 4.06 .57 1.06 .46 .36 0.30 .22 2.55 1.29 2.06 .10 2.30 .30 .18 0.06 .54 1.55 1. / Test No Type of Geosynthetic/Soil Dry Density (pcf) Moisture Content (%) Normal Stress (tsf) Shear Stress (tsf) Shear Rate (mm/min) 3.6 5.30 .Table B.8 3.18 0.75 1.00 .30 .30 16 17 18 Nicolon/Keuper Marl 107 109 110 16.70 1.2 3.20 0.9 16.30 .4 0.10 2.06 10 11 12 Tensar SS1/Limestone 139 139 141 5.75 1.9 5.99 1.18 0.54 1.0 4.10 2.4 0.12 2.8 16.48 .9 0.46 0.6 0.30 .30 .06 Sand & Gravel 138 136 136 3.

a recipe grading B-18 . For the Hot Rolled Asphalt. (10 mm). because of the requirement of a much higher failure load. (12 mm) with grading as shown in Fig. were performed. The aggregate used in the design mix had a maximum particle size of 0. B-10. the use of a faster rate of 7. One series of Marshall tests (ASTM D1559) was carried out for the design of the asphaltic concrete mix. B-12.tests involving granular material. maximum shear stress was obtained at a horizontal displacement of less than 0. Background and details of the test was reported by Murray and McGown (B-16). each with a different sustained load. These tests were carried out at the University of Strathclyde where specialist apparatus was available (B-15). All tests were conducted at a standard test temperature of 68 °F (20°C) and were continued until rupture occurred. (30 mm) was required to achieve maximum shear stress. lasted for 1000 hours. B-11. Tests on Asphaltic Materials Marshall Tests.2 in. All tests were carried out at 68 ° F (20 °C) and. B-9.5 in. in most cases. The results of the tests for both materials are shown in Fig. For each geosynthetic material. The results of the two sets of tests during the first 10 hours are shown in Fig. All creep tests were carried out in isolation with no confining media. A grade 50 Pen binder was used. Creep Test.4 in. A standard shearing rate of 2 percent per minute was used for the geogrid but for the stiff geotextile. up to five separate tests.5 percent per minute was necessary. for tests with Keuper Marl. The result of the test is summarized in Fig. For the geogrid. However. Wide Width Tensile Test. a horizontal displacement of up to 1. the maximum sustained load corresponded to 60 percent of the tensile strength of the material.

. 20 .....5%/min rar''s 4k---. 20 .----- a------ Er.200 . a) GEOTEXTILE 150 /a .--.. .0 // . . E< IY 10 5 0 20 15 25 25 b) GEOGRID .. O 5 a■ 0 5 10 AXIAL STRAIN (%) 15 Figure B-9../ Li 50 Strain rate= 7.. Variation of axial strain with load in wide-width tensile tests.. Strain rate= 2%/Min.--0.. .

)7 .... Results of creep tests at various sustained loads for the geosynthetics during the first 10 hours..3 kN/rn n 12 .-.-"Fc'l( --.5 kNATI ----./rn .- t..- --v--.' :1.=. a) GEOTEXTILE --E9. / / E.5 kN.4.1 kNim -. v .8._ -T-- - ._.- 7 .A..--E-.i+ AXIA L S TRA IN --.2 k Nim 15 --.:' .7." .!D. c-.i-4/ - 0 "El- 6 4 8 TIME (HOUR) Figure B-10...4 :--- 4 Lxi-4 -E-rB----4) 0 8 4 0 I /•BREAKAGE b) GEOGRID . B-20 10 ..15 kH7rn 1---------.12.

. OF MIX 9 4 7 8 5 6 (%) AC BY WGT. OF MIX 9 Figure 3-1LSurrmary of Hot-mix design data by the Marshall method. OF MIX 9 2200 21 7-1 1900 Ed 20 -J -J 4 5 6 7 8 (%) AC BY WGT. OF MIX 9 19 1600 (r) 1300 4 5 6 7 8 (%) AC BY WGT. OF MIX 9 16 35 — 30 7.12 154 %A IR VO ID UNITWE IGH T 152 150 148 4 4 0 8 5 6 7 (%) AC BY WGT. 25 20 LL 15 103 8 4 5 6 7 (%) AC BY WGT.

S I H ASP HALTICCONCREi 8 /1.LIE 300C6 JA B.1 —J 2 V 75 (p m ) CC a_ 1 O 2 LL 0i 0 8 v 0 0 L'?-) O (-9) O ° ONISSVci ]0\-11N30a3d O En . S.■ —J LLJ ›- < O 0 0 ZOT .

For comparison purposes. six Marshall samples. The viscosity at 140°F (60°C) was found to be about 4600 poises. B-12 with 8 percent of 100 Pen binder was used.as shown in Fig.5 percent. Also shown in the table are the test results obtained from an asphaltic concrete sample with a binder content of 6. a specification which was used for the last three series of tests. made out of the HRA used in the first series were tested. . Two viscosity tests were carried out by the Georgia Department of Transportation on the 50 Pen binder used for the asphaltic concrete mix. The average test results of the six samples are shown in Table B-4. Viscosity Test.

5 144 152 6 2. - Binder Content (% by weight) Mix Density (pcf) Air Void (%) Hot Roller Asphalt Asphaltic Concrete 8 6.5 19 VMA (%) 23. Comparison of Marshall test data for two asphaltic mixes.6 • Corrected Stability (lb) 2028 2150 Flow (1/100 in.5 18 Conversion : 1 pcf = 16.02 kg/m3 1 lbf = .Table B 4.00445 kN .) 16.

. and Monismith. B-5 Croney. B-11 Thom. B-3 Bell. University of Nottingham.. B-8 British Standards Institution.F.1. University of Nottingham. 1972. B-7 Brown. The Journal of Geotechniques.J. PhD Thesis. Vol. PhD Thesis. 1975. 1982. B-6 Finn.. 1979. HMSO. PhD Thesis. B-4 Loach.L.W. C. B-12 Boyce..APPENDIX B REFERENCES B-1 Hyde. "Design of Road Foundations". S.F. "The Prediction of Permanent Deformation in Flexible Pavements". "Repeated Loading of Fine Grained Soils for Pavement Design". B-10 Pappin. B-2 Overy. K. University of Nottingham. A. "Application of Theory in the Design of Asphalt Pavements".H. "Repeated Load Triaxial Testing of a Silty Clay". The Journal of Soil Science.. PhD Thesis..F. "Repeated Load Testing of Soils".A. S. G.. No. Vol. PhD Thesis. 19. B-13 ASTM Standard. 1977. 25. J..08. and Brown. London.. 1987.F. Vol. 1982. 1. C. Lashine.L... University of Nottingham. Nair. "Methods of Testing Soils for Civil Engineering Purposes". B-25 . "Soil Suction by the Rapid Method on Apparatus with Extended Range". Vol. M.C.F. "Methods for Determining the Flakiness Index of Coarse Aggregate". 1987. 04. University of Nottingham. 1.F. B-9 Dumbleton. Building Stones. S. N. R. 1985. J. BS1377. of 3rd Int. BS 812. Geotextiles". Proc.. PhD Thesis. on the Structural Design of Asphalt Pavements. Conf. Sections 105. 1972. 1976. "Characteristics of a Granular Material for Pavements Analysis". and West. 1987. 1985.. London. A. B-14 British Standards Institution. and Hyde. "The Behavior of Anisotropically Consolidated Silty Clay Under Cyclic Loading". University of Nottingham.. A. "The Design and Performance of Road Pavements". Standard D-1557.R.K.L.N. 1975.. D. "Soil and Rock. University of Nottingham. F. Interim Report to SCRC.. "The Behavior of a Granular Material Under Repeated Loading"..

"Geotextile Test Procedures Background and Sustained Load Testing". and McGown. PhD Thesis. B-16 Murray. K. 1985.. "The Behavior of Polymeric Grid Used for Soil Reinforcement".T. 1987. University of Strathclyde. R.B-15 Yeo. .L.. TRRL Application Guide 5. A..

APPENDIX C PRELIMINARY EXPERIMENTAL PLAN FOR FULL-SCALE FIELD TEST SECTIONS .

however. 2. Because of the high cost of prestressing. a prestressed section could be readily added to the test program as pointed out in the discussion. Prestressing was also found to give similar reductions in permanent deformations of the base and subgrade as prerutting.APPENDIX C PRELIMINARY EXPERIMENTAL PLAN FOR FULL-SCALE FIELD TEST SECTIONS INTRODUCTION An experimental plan is presented for evaluating in the field the improvement in pavement performance that can be achieved from the more promising techniques identified during the NCHRP 10-33 project. If desired. a prestressed test section was not directly included in the proposed experiment. the inclusion of a non-woven geosynthetic reinforced section would be a possibility if sufficient funds and space are available to compare its performance with the geogrid reinforcement proposed. TEST SECTIONS The layout of the ten test sections proposed for the experiment are shown in Figure C-1. Prerutting the unstabilized aggregate base without reinforcement. The experiment is divided into two parts involving (1) five test sections constructed using a high quality aggregate base. Also. A control section is included as one of the test sections for each base type. and (2) five test sections constructed using a low quality aggregate base susceptible to rutting. Geogrid Reinforcement of the unstabilized aggregate base. C-2 . The methods of improvement selected are as follows: 1.

4 O a.1. 4.. ro cia O ai 1-1 ro on 0) (cs) •r-I 4-+ C4) ro Q.rl O E C E Y.4 0 O 1. CD CO No ta t ion: Se c t ions @ 1 00 ft. T ro O C 00 Fig u re C. af ro E s. Eac h (no t inc luding trans it ions ) r LI . eu) C 0 O3 as Co Ql CI) J.

An even stronger structural section might be included in the experiment if sufficient space and funds are available.5 in. The reinforced sections should have similar stiffnesses as the control sections. The test sections should be placed over a soft subgrade having a CBR of about 2. should be used to evaluate the as-constructed stiffness of each section. Extensive vane shear. are to be constructed using a 2. (250 mm) unstabilized aggregate base. (32 m) in length with a short 25 ft.All test sections.0.5 to 3. except Section 10. The test sections should be a minimum of 100 ft. High Quality Base Sections. Two prerutted sections and two reinforced sections are included in the high quality base experiment. The high quality base section study is designed to investigate the best pattern to use for *C-4 . A falling weight deflectometer. (114 mm) thick asphalt surfacing and an 8 in. (200 mm) low quality aggregate base. high quality construction is achieved for each test section. cone penetrometer or standard penetration resistance tests should be conducted within the subgrade at close intervals in each wheel track of the test sections. A careful quality control program should be conducted to insure uniform. Test Section 10 is to have a 4. (8 m) transition between each section. The high quality base experiment could be placed on one side of the pavement and the low quality base experiment on the other to conserve space.5 in. Measurements should also be made to establish as-constructed thicknesses of each layer of the test sections. (64 mm) asphalt concrete surfacing and a 10 in. or similar device. The purpose of these tests are to establish the variability of the subgrade between each section. The falling weight tests will serve as an important indictor of any variation in strength between test sections.

8 kN/m) would be placed at the center of the base. which is also prerutted. additional aggregate would be added to bring the base to final grade. Low Quality Base Section.5 in. After prerutting.. while Section 7 should be prerutted to a depth of about 3 to 3. Section 6 should be prerutted to a depth of about 2 in. (50 mm). A good subgrade could be used rather than a weak one for this experiment. In Section 9 a geogrid reinforcement (S g > 1500 lbs/in. Sections 2 and 3 have geogrid reinforcement at the center and bottom of the base. prerutting should be continued until a rut depth of approximately 2 in. Prerutting would be carried out for an aggregate base thickness of about 7 in. In each test section. In Section 5. a single rut should be formed in each side of the lane. C-5 . it could also be included in this study. If desired. (75-90 mm). A heavy vehicle having single tires on each axle should be used. The ruts would be about 12 in. (1. and also the optimum position for geosynthetic reinforcement. Section 2 could be prestressed.prerutting. respectively. 1. (50 mm) is developed. Optimum depth of prerutting is studied in the low quality base experiment. Two prerutted sections are included in the study to allow determining the influence of prerut depth on performance. and then densified again by a vibratory roller. Section 10 is included in the experiment to verify if improved performance due to prerutting is not obtained for heavier pavement sections. (200-300 mm) apart. Prerutting would be accomplished in Test Section 1 by forming two wheel ruts in each side of the single lane test section. (180 mm). This experiment is included in the study to establish in the field the improvement in performance that can be obtained by either prerutting or reinforcing a low quality base.8 kN/m). The minimum stiffness of the geogrid should be S g = 1500 lbs/in.

The following instrumentation should be used for each test section. this was demonstrated during the experiments conducted as a part of this study. Tensile strain in the bottom of the asphalt concrete should be measured using embedment type wire resistance strain gages. wire resistance strain gages should also be used to directly measure strain in the geogrid reinforcement. Two pressure cells should be used to measure vertical stress on top of the subgrade. The instrumentation layout for one test sections should be similar to that shown in Figure C-2. Although quite desirable. and preferably two. . of strain coils should be placed in the bottom of the aggregate base to measure lateral tensile strain. In addition to using strain coils. is recommended in addition to the manual measurement of rut depth. similar to the one described in Chapter 2. Much valuable information can be gained through a carefully designed instrumentation program. In general.MEASUREMENTS The primary indicators of pavement test section performance are surface rutting and fatigue cracking. The embedment gages should be oriented perpendicular to the direction of the traffic. a duplicate set of instruments is provided to allow for instrumentation loss during installation and instrument malfunction. Bison type strain coils should be employed to measure both permanent and resilient deformations in each layer (Figure C-2). the two vertical oriented pressure cells in the base shown in Figure C-2 could be omitted for reasons of economy. An instrumentation program similar to the one used in this study is therefore recommended. Use of a surface profilometer. Both of these variables should be carefully measured periodically throughout the study. At least one pair.

Preliminary Instrument Plan for Each Test Section.Embedment Strain Gage SIDE VIEW enabt:e alma 2'1 or in. Strain Coil Stack direelian 0! vowel travel Coils in Aggregate Base to Measure Tension <al pre Sure cell in granular base Coils and/or Strain Cages Ar:ached to Geogrid 20 11 II 14 12 10 pressure cen therfflOCOUpte 0 2 ri 111 12 16 16 II 20 6 4 2 distance tram centre (Ins) Figure C-2. Dia. Strain Coils . M 4 3 e0. . Diameter Strain Coil \ 2 in..1 2 IS 1412106 6 4 2 0 2 4 6 distance from centre (ns) a 16 12 1% IS II 20 61 PLAN Embedment Strain Gage pressure cell in subgrade Embedment Strain Gage 2 in.

Stress-strain and strength of the geogrid reinforcement as determined by a wide width tension test. MATERIAL PROPERTIES The following laboratory material properties should be evaluated as a part of the materials evaluation program: 1. 3. Shear strength and water content of the subgrade beneath each test sections. Placement of moisture gages in the subgrade would also be desirable. and measurements made each time readings are taken. Friction characteristics of the geogrid reinforcement as determined by a direct shear test. Resilient and permanent deformation characteristics of the low and high quality aggregate base and subgrade. . Mix design characteristics of the asphalt concrete surfacing.Thermocouples for measuring temperature should be placed in each section. 5. 2. 4.

Nation/al Cooperative Highway Researrh Progyam Type Agreement Contract No HR 10-33 Award Period: From 1/6/86 To 7/5/88 Sponsor Amount: This Change • (Reports) Total to Date Estimated: $ $ 100. 476 . DC 20418 (202) 334-2254 Defense Priority Rating: Military Security Classification: N/A (or) Company/Industrial Proprietary: N/A N/A RESTRICTIONS See Attached Government Supplemental Information Sheet for Additional Requirements. Procurement/EES Supply Services curit Services AI Research cvmmuniconti 12) GTRC Library Project File Other JaneRnsmal .00 Cost Sharing Amount: $ Title: (Performance) 1/5/88 25. Office of Contracts and Grants Washington. NO.W.000. 5 / 1 / 86 DATE GT RCAICT Sch ooll1Mi C i vil Enginee r i p g Project Director: 1)r Ri r hard D. Crawford Jenks Ann Fisher NationalAcademy of Sciences Contract Specialist 2101'Constitution Avenue. Equipment: Title vests with Sponsor COMMENTS. COPIES TO: Project Director Research Administrative Network Research Property Management Accounting SPONSOR'S I. Barksdal e tA ( Sponsor: National Academy of Sciences. 00 $ Cost Sharing No: E-20311 Potential Benefits of Geosynthetics in Flexible Pavements Systems ADMINISTRATIVE DATA 1) Sponsor Technical Contact: OCA Contact E. D. Faith Gleason X4820 2) Sponsor Admin/Contractual Matters: Mr. Domestic travel requires sponsor approval where total will exceed greater of $500 or 125% of approved proposal budget category.OFFICE OF CONTRACT ADMINISTRATION GEORGIA INSTITUTE OF TEAlik1OLZGe fROJECT ADMINISTRATION DATA SHEET F X -1 ORIGINAL Project No (R6116-0A0) E-20-672 Ei REVISION NO.000. S.00 Funded: $ 100. Travel Foreign travel must have prior approval — Contact OCA in each case. DC 20418 National Academy of Sciences 2101 Constitution Avenue • Washington.

Includes Subproject No(s) . . GTRC HR 10-33 GIT Prime Contract No. Barksdale 318/89 R6116-0A0 School/Lab CE National'Arademy of Sripnces Contract/Grant No.of Inventions and/or. Distribution: Continued by Project No. Potential Benefits of Geosynthetics 'in Flexible Pavements Systems Effective Completion Date 12/15/88 (Performance) 12/15/88 (Reports) Closeout Actions Required: None Final -Invoice or. . Government Property Inventory & Related Certificate Classified Material Certificate Release and Assignment ter.Questionnaire . Copy of Last' Invoice FinalReport .„--. Project Director x Administrative Network Accounting x Procurement/GTRI Supply Services x Research Property Management Research Security Services Reports Coordinat•r (OCA) Project File Contract Support Division (OCA)(2) Other . F-20-672 R. Pro ect. Continues Project No. Director Sponsor Center No.Subcontracis-Patent. Subproject Under Main Project No. . Title .747711111 Prw!IMIRIper —T73'7.D.I. ---7 FPNIMPror■Pty 1 GEORGIA INSTITUTE OF TECHNOLOGY OFFICE OF CONTRACT ADMINISTRATION NOTICE OF PROJECT CLOSEOUT Date Project No. to P.

upp ement to Report 315 etear Richard D. Barksdalf3 eorgia Institute of :TechnOlo Atlanta. Georgia .

Supplement to NCHRP Report 315 POTENTIAL BENEFITS OF GEOSYNTHETICS IN FLEXIBLE PAVEMENTS APPENDICES B-H .

. Resilient Properties Model Verification . .. Unreinforced. Nonlinear Properties Estimation of Permanent Deformation .TABLE OF CONTENTS Page LIST OF FIGURES iii LIST OF TABLES vii APPENDIX B .TEST SECTION MATERIALS.Geogrid and Heavy Loading Steel Mesh Reinforcement Large-Scale Laboratory Tests . Model Properties Used in Sensitivity Study .. High Quality Aggregate Base Pavement Response of Geosynthetic Reinforced Sections . .LABORATORY TESTING OF MATERIALS .Low Stiffnesses...Predicted Pavement Response .DEVELOPMENT OF ANALYTICAL MODELS USED TO PREDICT • • REINFORCED PAVEMENT RESPONSE ..EXPERIMENTAL STUDIES OF SURFACED PAVEMENTS REINFORCED WITH A GEOSYNTHETIC B-2 Field Tests . • • • • • • APPENDIX D . . References B-7 B-9 B-10 C-2 C-3 C-6 C-7 C-11 C-14 C-17 C-19 C-25 D-2 Materials Instrumentation Pavement Construction Pavement Surface Profile Construction Quality Control References APPENDIX D . References . INSTRUMENTATION AND CONSTRUCTION Tests on Silty Clay Subgrade Tests on Granular Base Material Tests on Geosynthetics Tests on Asphaltic Materials . References APPENDIX C . Nonwoven Geotextiles • • Large-Scale Laboratory Tests Using Stiff Geogrids . .Thick Bituminous Surfacing Field Tests . B-2 B-3 B-4 D-2 D-7 D-10 D-19 D-20 D-26 • • • • E-2 E-2 E-7 E-14 E-18 E-23 .

.. • APPENDIX G . . F-2 .... ..DURABILITY • F-2 F-4 F-5 F-5 F-7 F-9 F-11 F-11 F-12 F-13 F-15 F-17 F-19 F-29 F-31 F-38 F-39 F-45 G-2 Pavement Applications . APPENDIX H . Introduction Test Sections Measurements Material Properties G-2 G-5 G-11 H-2 H-2 H-2 H-6 H-8 ii ...SEPARATION AND FILTRATION Introduction Filter Criteria for Pavements Separation Separation Failure Mechanisms Construction Stresses Bearing Capacity Analysis Construction Lift Thickness Permanent Deformation Separation Case Histories Separation Design Recommendations Filtration Filtration Mechanisms Geotextile Filters Laboratory Testing Methods Selected Practices Filter Selection Geotextile References ... Soil Burial References . . ..PRELIMINARY EXPERIMENTAL PLAN FOR FULL-SCALE FIELD TEST SECTIONS . ..TABLE OF CONTENTS (continued) Page APPENDIX F .

000 Load Repetitions C-23 Gradation Curve for Aggregates Used in Asphaltic Mixes D-3 D-6 D-2 Gradation Curves for Granular Base Materials D-3 Typical Layout of Instrumentation Used in Text Track Study D-9 Profilometer Used to Measure Transverse Profiles on Pavement D-11 D-4 D-5 Triple Legged Pneumatic Tamper Used on Subgrade D-6 Single Legged Pneumatic Compactor Used on Subgrade .. Ref. Ref. • • D-12 D-12 • • D-12 D-12 iii . D-7 Vibrating Plate Compactor D-8 Vibrating Roller .LIST OF FIGURES Figure B-1 B-2 B-3 C-1 C-2 C-3 C-4 C-5 D-1 Page Maximum Surface Deformation as a Function of Traffic (After Barker. et al.000 Load Repetitions C-20 Comparison of Measured and Computed Permanent Deformation Response for a Low Quality Soil-Aggregate Base: 100. C-9) . Ref. B-5) B-8 Resilient Modulus Relationships Typically Used for a Cohesive Subgrade and Aggregate Base C-4 Idealization of Layered Pavement Structure for Calculating Rut Depth (After Barksdale. C-20 Comparison of Measured and Computed Permanent Deformation Response of a High Quality Crushed Stone Base: 100. B-3) B-5 Comparison of Strain at Bottom of Asphalt Surfacing With and Without Mesh Reinforcement (After Van Grup and Van Hulst. . B-4) B-6 Surface Deformation and Lateral Strain Measured in Nottingham Test Facility (After Brown.000 Load Repetitions C-23 Comparison of Measured and Computed Permanent Deformation Response for a Silty Sand Subgrade: 100. Ref.

Test Series 4 D-18 D-12 Static Cone Penetrometer Test on Subgrade D-22 D-13 Dynamic Cone Penetrometer Test on Subgrade D-22 D-14 Nuclear Density Meter D-22 D-15 Clegg Hammer used on Aggregate Base D-22 E-1 The Relationship Between Stiffness and CBR for Compacted Samples of Keuper Marl for a Range of Stress Pulse Amplitudes (After Loach) E-3 Results from Suction-Moisture Content Tests on Keuper Marl (After Loach) E-6 Permanent Axial and Radial Strain Response of Keuper Marl for a Range of Stress Pulse Amplitudes (After Bell) E-8 Stress Paths Used in Cyclic Load Triaxial Tests for Granular Materials E-11 Permanent Axial and Radial Strain Response of Sand and Gravel During Repeated Load Triaxial Test . • • D-15 D-10 Geogrid with 1 in. D-15 D-11 Method Employed to Stretch Geogrid Used to Prestress the Aggregate Base . • .LIST OF FIGURES (continued) Figure D-9 Page Woven Geotextile with 1 in. Diameter Induction Strain Coils . E-12 Permanent Axial and Radial Strain Response of Dolomitic Limestone During Repeated Load Triaxial Test at Various Moisture Contents (w) and Degree of Saturation (Sr) E-13 Results of Standard Compaction Tests for the Granular Materials E 15 Relationship Between Normal and Maximum Shear Stress in Large Shear Box Tests E-17 Variation of Axial Strain with Load in Wide-Width Tensile Tests E-19 Results of Creep Tests at Various Sustained Loads for the Geosynthetics During the First 10 Hours E 20 E-2 E-3 E-4 E-5 E-6 E-7 E-8 E-9 E-10 iv - - . . Diameter Induction Strain Coils . .

. (After Jorenby. Ref. F-31) F-3 Variation of Vertical Stress on Subgrade with Initial Compaction Lift Thickness and Roller Force . F-25) . Ref. Ref. . F-30 Economic Comparison of Sand and Geosynthetic Filters for Varying Sand Filter Thickness F-30 F-2 F-3 F-4 F-5 F-6 F-7 F-8 F-9 F-10 F-11 F-12 F-13 F-14 . E-21 E-12 Gradation Curves for Aggregates Used in Marshall Tests . F-28 Variation of Vertical Stress with Depth Beneath Railroad Track and Highway Pavement F-28 Cyclic Load Triaxial Apparatus for Performing Filtration Tests (Adapted from Janssen. F-8 Bearing Capacity Failure Safety Factor of Subgrade During Construction of First Lift . Ref. 095 (After Bell. F-28) . Ref. F-10) F-23 Variation of Geosynthetic Contamination Approximately 8 in. et al. Below Railroad Ties with Geosynthetic Opening Size (After Raymond. F-8 Mechanisms of Slurry Formation and Strain in Geosynthetic F-18 Electron Microscope Pictures of Selected Geotextiles: Plan and Edge Views (84x) F-21 Variation of Geosynthetic Contamination with Number of Load Repetitions (After Saxena and Hsu. Ref. F-23 Variation of Geosynthetic Contamination with Geosynthetic Apparent Opening Size. Ref.LIST OF FIGURES (continued) Figure Page E-11 Summary of Hot-Mix Design Data by the Marshall Method . E-22 F-1 Influence of Added Fines on Resilient Modulus of Base . F-11) F-26 Variation of Geosynthetic Contamination with Stress Level and Subgrade Moisture (After Glynn & Cochrane. F-31) F-26 Observed Variation of Geosynthetic Contamination with Depth Below Railway Ties (After Raymond.. Ref. F-2) F-3 Influence of Subgrade Water Content and Geosynthetic on Stone Penetration (After Glynn & Cochrane. F-11) .

Use of Longer Sections and More Variables are Encouraged H-3 H-2 Preliminary Instrument Plan for Each Test Section vi . G-9 H-1 Tentative Layout of Proposed Experimental Plan .• LIST OF FIGURES (continued) Page Figure G-1 Observed Strength Loss of Geosynthetic with Time . H-7 .

5 in. C-4 E-1 Results of Classification Tests for Keuper Marl E-2 Summary of Resilient Parameters for Granular Materials Obtained from Cyclic Load Triaxial Tests E-10 E-3 Summary of Large Shear Box Tests E-16 E-4 Comparison of Marshall Test Data for Two Asphaltic Mixes. (Adapted Christopher and Holtz. Ref. B-6 Comparison of Measured and Calculated Response for a Strong Pavement Section: 3. Crushed Stone Base C-9 Anisotropic Material Properties Used for Final Georgia Tech Test Study C-10 Comparison of Measured and Calculated Response for Nottingham Series 3 Test Sections C-13 Aggregate Base Properties Used in Cross-Anisotropic Model for Sensitivity Study C-16 C-5 Nonlinear Material Properties Used in Sensitivity Study C-16 C-6 General Physical Characteristics of Good and Poor Bases and Subgrade Soil Used in the Rutting Study C-24 C-4 Specification of Hot Rolled Asphalt and Asphaltic Concrete D-4 D-2 Properties of Geosynthetics Used D-8 D-3 Layer Thickness of Pavement Sections and Depth of Geosynthetics From Pavement Surface D-21 Summary of Construction Quality Control Test Results for All Test Series D-24 Summary of Results from Falling Weight Deflectometer Tests Performed on Laboratory Test Sections D-25 D-1 D-4 D-5 . Asphalt Surfacing. E-26 F-1 Design Criteria for Geosynthetic and Aggregate Filters . 8 in. F-9) F-6 vi i .LIST OF TABLES Page Table B-1 C-1 C-2 C-3 Summary of Permanent Deformation in Full-Scale Pavement Sections on a Compacted Sand Subgrade .

G-6) G-7 G-2 G-3 viii . . F-41 Pavement Structural Strength Categories Based on Vertical Stress at Top of Subgrade . Polypropylene and Polyester Polymers G-4 Effect of Environment on the Life of a Polypropylene (After Wrigley. Army Corps of Engineers Geosynthetic Filter Criteria (Ref. F-16 F-2 Preliminary Subgrade Strength Estimation . Ref. Ref. F-34) F-34 Aggregate Gradations Used by Pennsylvania DOT for Open-Graded Drainage Layer (OGS) and Filter Layer (2A) . S. F . F-35 Separation Number and Severity Classification Based on Separation/Survivability F-35 Guide for the Selection of Geotextiles for Separation • • and Filtration Applications Beneath Pavements .9) F-33 U. F-43 F-10 Partial Filtration Severity Indexes G-1 General Environmental Characteristics of Selected Polymers G-4 Summary of Mechanisms of Deterioration. F-43 F-4 F-5 F-6 F-7 F-8 F-9 • • • . F-3 Vertical Stress on Top of Subgrade for Selected Pavement Sections F-16 Recommended Minimum Engineering Fabric Selection Criteria in Drainage and Filtration Applications AASHTO-AGG-ARTBA Task Force 25 (After Christopher and Holtz. Advantages and Disadvantages of Polyethylene.LIST OF TABLES (continued) Table Page .

APPENDIX B EXPERIMENTAL STUDIES OF SURFACED PAVEMENTS REINFORCED WITH A GEOSYNTHETIC .

/in.3 in. Potter and McAvoy [B-1.000 lbs/in. 600 kN/m) and had an elastic modulus of about 72. (83 kN/m) in each direction. and transient tensile strain in the bottom of the bituminous layer. Loading was applied by a two-axle truck having dual rear wheels. transient stress and strain in the subgrade. The sections were constructed on a London clay subgrade having a CBR increasing with depth from about 0. The woven geotextile had a strength of about 474 lb. (300 mm). Ruddock et al.5 kN) was applied for 4600 repetitions..5 percent at a depth of 11. For the conditions of the test which included a 6.B2] included two sections having a 6. with the axle loading being increased to 30 kips (133 kN) for an additional 7700 passes. no difference in structural performance was observed between the geotexile reinforced sections and the control section.3 in.APPENDIX B EXPERIMENTAL STUDIES OF SURFACED PAVEMENTS REINFORCED WITH A GEOSYNTHETIC Field Tests . permanent strain in the geotextile. The geosynthetic stiffness S gis defined as the force applied per unit width of geosynthetic divided by the resulting strain.(160 mm) bituminous surfacing. and an elongation at failure of 14.8 percent.Thick Bituminous Surfacing Full-scale experiments conducted by Ruddock.8 in. 2 (500 kN/m2 ).9 kips (97. A rear axle load of 21. Measurements made included surface deformations. The geotextile used was stiff (S g @ 5 percent = 3400 lbs/in. (300 mm) thick crushed granite base.7 percent at the top to 3. One of these sections had a woven multi-filament polyester geotextile reinforcement in the bottom of the granular base. found that resilient vertical subgrade stresses and strains were not significantly B-2 . (160 mm) thick bituminous surfacing and a 12 in.

Falling Weight Deflectometer (FWD) tests showed the stiff to very stiff reinforcement did not affect the measured deflection basins throughout the experiment. The granular base./in. (700 kN/m). although transverse resilient strains were somewhat reduced.000 lbs. the pavement should have been loaded sufficiently to cause rutting to develop. Field Tests . (1. (150 mm) thick. however. The geogrid was placed at the center of the aggregate base. The geogrid used had a secant stiffness at 5 percent strain of about 4. A 6 in. 57 crushed limestone. The 27-kip (120 kN) load applied to the pavement consisted of a single tire inflated to 265 psi (1. even after compaction. An unstable base of this type would appear to be a good candidate for reinforcing with a stiff geogrid.5 m) width. The subgrade was a sandy silt having a CBR of 27 percent.Geogrid and Heavy Loading Recently. The test sections consisted of a 3 in.000 repetitions of a heavy moving aircraft load. The pavement was trafficked over a 60 in. was loose and unstable to most traffic [B-3]. (75 mm) asphalt surfacing overlying a 6 in. very open-graded base consisting of No. Because of the use of a thick bituminous surfacing. unstabilized aggregate base reinforced by a stiff to very stiff geogrid.changed by fabric inclusions.8 MN/m 2 ). The pavement was subjected to 1. (150 mm) cement stabilized clay-gravel subbase was constructed to provide a strong working platform for the open-graded base. To demonstrate if some improvement in permanent deformation could be achieved due to reinforcement. This finding indicates similar stiffnesses and effective layer moduli for the reinforced and unreinforced sections. Barker [B-3] has studied the performance of a pavement having an open-graded. it is doubtful that the conclusions reached would have been significantly changed. The general condition of the two pavements B-3 .

appeared similar after 1.4 in. The subgrade consisted of a compacted. These factors greatly complicate translating the test results to conventional pavements. The reinforcement was placed in the middle of the granular base. The asphalt surfacing was 2.000 load repetitions. however. The non-conventional pavement section studied at WES had a very opengraded granular base. Rutting. and the rutting which developed as a function of load repetitions is given in Figure B-2. . Steel Mesh Reinforcement A hexagonal wire netting of steel was placed at the interface between a crushed rubble aggregate base and the asphalt surfacing in a large scale test track experiment described by van Grup and van Hulst [B-4]. A summary of the test conditions is given in Table 8-1. (25 mm). Subsequent trench studies indicated that most of the permanent deformation occurred in the subgrade and not the base. Maximum observed rutting of the reinforced section was about 8 percent less than the unreinforced section at a rut depth of 1 in. The reinforcement was placed at the center of the aggregate base to improve its performance. coarse sand. Better performance might have been obtained had the reinforcement been placed at the bottom of the base. and about 21 percent less at a rut depth of 2 in. (60 mm) thick. primarily occurred in the subgrade. For this well constructed pavement. and the aggregate base varied in thickness from 8 to 16 in. important reductions in permanent deformation occurred due to reinforcement only after the development of relatively large deformations. (200-400 mm). (50 mm) as shown in Figure B-1. a cement stabilized supporting layer and was subjected to a very high wheel load and tire pressure.

Maximum Surface Deformation as a Function of Traffic (After Barker. N 1 0 00 Figure B-1. 100 TRAFFIC COVERAGES.av ( I N) SURFACEDEFORMATION. . B-3). Ref.

8 0 47.4 2. 1.4 31.98 A Note: 1.3 0.9 11.3 35.49 0.2 39. 0.4 2.4 Steel Mesh Reinf.55 0.4 2. Ref. B-4).7 11. B-6 . The steel mesh reinforcement was placed at the aggregate base/asphalt surfacing interface.4 2. 1000 E Up lb 800 C 0 44 4) U 600 0 Cl) C •4 (0 400 CA ri (0 200 200 400 600 BOO 1000 Asphalt Strain Section fl 400 Figure B-2.) @ 140.5 47./ @ Top of Base NO NO NO NO YES YES Crushed Rubble 0 7.4 2.) LAYER 1 2 3 4 5 6 Dense Asphaltic Concrete 2.44 0.000 Reps.8 15.2 35. Comparison of Strain at Bottom of Asphalt Surfacing With and Without Mesh Reinforcement (After Van Grup and Van Hulst.55 .4 - - - - Sand ----------- Clayey Sand Permanent Surface Deformation (in.Table B-1 Summary of Permanent Deformation in Full-Scale Pavement Sections on a Compacted Sand Subgrade LAYER THICKNESSES AND PERMANENT DEFORMATION OF SECTIONS (in.

A reduction in tensile strain of about 18 percent was. which did not have an aggregate base.Low Stiffness. (220 kN/m) and 445 lbs/in./in. The inclusion of the nonwoven geotextiles in the aggregate base in most tests appeared to cause a small increase in rutting (Figure B-3a). [B-5] investigated the effect of the placement of a nonwoven geotextile within and at the bottom of the aggregate base of bituminous surfaced pavements. These geotextiles had a secant stiffness at one percent strain of about 1270 lbs. observed in the bottom of the asphalt surfacing. Two very low to low stiffness.4 kip (15 kN). Seven different reinforced sections were studied. however. (37-53 mm).5 to 2. The pavements rested on a silty clay subgrade having a CBR that was varied from 2 to 8 percent. This large level of reduction in strain. Both vertical and lateral resilient and permanent strains were also found to be greater in the base and subgrade of all of the reinforced sections (Figure B-3b). Large-Scale Laboratory Tests . nonwoven. et al. for each condition a similar control section was also tested without reinforcement. The crushed limestone base was varied in thickness from 4. (78 kN/m). A moving wheel load was used having a magnitude of up to 3.5 in (12 mm) rut depth.1 in. resulted in a 40 percent reduction in rutting at about 0. The bituminous surfacing of the seven test sections varied in thickness from 1. Reinforcement made little difference in rutting performance for the stronger sections having rubble aggregate bases. and no increase in effective elastic stiffness of the granular layer.9 in.Reinforcement of a weak section. (107-175 mm). would have a very significant beneficial effect on fatigue performance. melt bonded geotextiles were used in the study. if maintained.2 to 6. Nonwoven Geotextiles Brown. The experiments included placing the geotextiles within the granular B-7 .

• Unreintorced B . Ref. % 0 5 ■ Reinforced •1• • . it • 0 -5 LEGEND je • • Granular • Base ► 200— • • ► • 1 /I = /// • Subgrade 300 Subgrade 300 — Note: 1 in. Surface Deformation and Lateral Strain Measured in Nottingham Test Facility (After Brown.e. .DEFLECTION (mm ) • ▪ RADIAL OFFSET DISTANCE (mm) 600 450 300 150 0 150 300 450 600 0 5 REINFORCED UNREINFORCED 10 (A) SURFACE DEFLECTION Lateral Resilient Strain Permanent Lateral Strain.. ./ • • • • 4 Granular Base • 200 • • • . B-5). • • • . • 25 mm (b) Lateral Strains Figure B-3. 100 Depth (mm) • •• e slip 100 Depth (mm) 46. Bit.0/ a . et al.

In one series of tests. several factors suggest compaction of the aggregate above the geosynthetic may not have been as effective when the geotextile was present. After testing. The base had a reported insitu CBR value of 18 percent but laboratory CBR testing indicated a value of 100 percent or more. Two layers of reinforcement were also employed in some tests. peat was mixed with the fine sand at a high water content to give a very weak subgrade having an initial CBR of only 0. B-9 . The poor performance of the reinforced sections was attributed to a lack of adequate aggregate interlock between the base and the geotextiles. although the researchers assumed for analyzing test results an increase did not occur. (0. S g of the geogrid at 5 percent strain was about 1780 lb/in. et al. In the light of more recent findings. (100-300 mm). (312 kN/m). 9 kip (40 kN) cyclic load was applied through a 12 in. The aggregate base was well-graded and was varied in thickness from 4 to 12 in. which resulted in relatively small strains in the geosynthetic. An increase in CBR might also have occurred in other sections. The subgrade was a fine beach sand having a CBR of typically 4 to 8 percent before the tests.2 percent. Maximum surface rutting was less than about 1 in. A stationary. Finally. Large-Scale Laboratory Tests Using Stiff Geogrids Penner. [B-6] studied the behavior of geogrid reinforced granular bases in the laboratory using a shallow plywood box 3 ft. (300 mm) diameter plate. the CBR of Loop 3 was found to have increased by a factor of at least 2.layer and using geotextiles strengthened by stitching. (75 or 100 mm). the relatively low geosynthetic stiffness probably also helps to explain the results. The secant stiffness.8 to 1. The asphalt surface thickness was either 3 or 4 in. (25 mm).9 m) deep.

. Little improvement was observed when the reinforcement was located at the top of the base.. 73. pp.. Christ. S.. B-4 van Grup.R. Part 2. Proceedings. II.2 in. "Reinforcement at Asphalt-Granular Base Interface". "Use of Nonwoven Fabrics in Permanent Road Pavements". For one section having an 8 in.D.F. CIRCIA.. Hants". Potter. "A Full-Scale Experience on Granular and Bituminous Road Pavements Laid on Fabrics". A. paper submitted to Journal of Geotextiles and Geomembranes. A.F.P. "Report on the Construction and Performance of a Full-Scale Experimental Road at Sandleheath.C. R. and Brodrick. Misc. W.C. 1982..P.. and McAvoy.. J. APPENDDCB RFYKRENCES B-1 Ruddock.L. Jones. 541-563.. V. pp. (5 mm). Las Vegas. Project Record 245.R. (15 mm) deformation observed in the unreinforced section. B-5 Brown.. Proceedings. 1982. "Open-Graded Bases for Airfield Pavements". Important improvements in performance were found in this test for deformations of the reinforced section as small as 0. Sept. February.M. C.. A. B-2 Ruddock. E.M. (25 mm). 1982. E. B. 365-370. Vol. 1986. B-3 Barker. Potter. . July. (75 mm) asphalt surfacing. 1988. Second International Conference on Geotextiles. and van Hulst. Vol. In contrast with the above findings.Placement of the geogrid within the granular base was found to result in a significant reduction in pavement deformation when placed in the middle or near the bottom of the base. use of geogrid reinforcement in under-designed sections on weak subgrades showed no apparent improvement until permanent deformations became greater than about 1 in. Paper GL-86.. (200 mm) granular base and 3 in.R.6 in. J. Waterways Experiment Station..F. Institution of Civil Engineers. London. and McAvoy. sections having geogrid reinforcement at the bottom and mid-height exhibited only about 32 percent of the 0.

"Geogrid Reinforcement of Granular Bases". Haas.. . presented to Roads and Transportation Association of Canada Annual Conference. 1985. J.. Walls. September. R.. Vancouver.B-6 Penner. R.

APPENDIX C DEVELOPMENT OF ANALYTICAL MODELS USED TO PREDICT REINFORCED PAVEMENT RESPONSE .

The ability to consider either small or large displacements which might.e. are applied at a stationary position (i. Only . and 6. 2. occur under multiple wheel loadings in a haul road. The GAPPS7 program does not consider either inertia forces or creep. for example. Important features of the GAPPS7 program include: 1. Modeling of the fabric interfaces including provisions to detect slip or separation. be changed for each loading cycle to allow considering time and/or load dependent changes in properties to be considered. A two dimensional flexible fabric membrane element which can not take either bending or compression loading. and repetitive loadings. 5. the load does not move across the continuum). Therefore. plastic and failure response. 3. 4. A no-tension analysis that can be used for granular materials. when used. The GAPPS7 program models a general layered continuum reinforced with a geosynthetic and subjected to single or multiple load applications. The ability to model materials exhibiting stress dependent behavior including elastic. Material properties can. Provision for solving either plane strain or axisymmetric problems. however. the capabilities of this comprehensive program are only briefly summarized in this section.APPENDIX C DEVELOPMENT OF ANALYTICAL MODELS USED TO PREDICT REINFORCED PAVEMENT RESPONSE The GAPPS7 finite element model has been described in detail elsewhere [C-1].

axisymmetric. These supplementary programs greatly facilitate performing finite element analyses and checking for errors in the data. small displacement analyses were performed for this study using a single loading. and calls the appropriate subroutines. An automatic finite element mesh generation program MESHG4 is used to make the GAPPS7 program practical for routine use. the resilient modulus stops decreasing and may even very slightly increase in a linear manner. The most commonly used nonlinear model for the resilient modulus of cohesionless granular base materials is often referred to as the k-O model (Figure C-lb) which is represented as C-3 . The stress dependent resilient modulus E r of the subgrade is frequently given for cohesive soils as a bi-linear function of the deviator stress a1-a3 as shown in Figure C-1. The main program handles the input. GAPPS7 consists of a main program and twelve subroutines. For this model the resilient modulus is usually considered to very rapidly decrease linearly as the deviator stress increases a small amount above zero. After a small threshold stress is exceeded. When a nonlinear model was used the subgrade was characterized following this approach. A plotting program called PTMESH can be used to check the generated mesh and assist in interpreting the large quantity of data resulting from the application of the program. MESHG4 completely generates the finite element mesh from a minimum of input data. Resilient Properties Three different models can be utilized in the GAPPS7 program to represent the stress dependent elastic properties of the layers. In addition to handling material properties. The twelve subroutines perform the actual computations. performs the needed initializations.

Resilient Modulus. Resilient Modulus Relationships Typically Used for a Cohesive Subgrade and Aggregate Base. E r E . E r (logscae) Resilient Modulus. cre .• DEVIATOR STRESS.173 (a) Subgrade KO Sum of Principal Stresses. (Ti . (log scale) (b) Base Figure C-1.

the deviator stress q can be defined as q = 0.G1. sometimes called M r . Following their approach the bulk modulus (K) and shear modulus (G) of the base can be calculated from the simplified relations K = v1 L1/4 -(1-n) {1 + P G = G1 p T (C-2) 2 p( ) 21.m = material properties evaluated in the laboratory from special cyclic loading stress path tests The model described by Equations (C-2) and (C-3) is referred to throughout this study as the simplified contour model.C-4] to more accurately characterize granular base materials. et al.Er = K cre N (C-1) where E r = resilient modulus of elasticity. often referred to as contour models.(03 . (01 + 02 + 03)/3 q = shear stress K1. 01 + 0 2 + 03 In recent years several improved models. (1-m) (C-3) where: K = bulk modulus G = shear modulus p = average principal stress.707 where (C-4) 2 J2 = (01 .n.02 )2 4.(02 . [C-6] was employed in this study. For a general state of stress.03 )2 -I. The contour model as simplified for routine use by Mayhew [C-5] and Jouve.u l ) C-5 . determined from laboratory testing k and 6 = material constants determined from laboratory testing u 8 = sum of principle stresses. have been developed by Brown and his co-workers [C-3.

The importance of the role which tensile strain developed in the reinforcing layer plays became very apparent as the analytical study progressed. including the k-131 type model.31 n = 0.PREDICTED PAVEMENT RESPONSE Little work has been carried out to verify the ability of theoretical models to accurately predict at the same time a large number of measured stress. The presence of a tensile reinforcement and relatively thick granular layers which have different properties in tension compared to compression greatly complicate the problem of accurately predicting strain in the aggregate layer.31 1 The bulk modulus (K) as given by equation (C-2) is always greater than zero which neglects the dilation phenomenon which can cause computational difficulties.6) m = 0.Laboratory tests by Jouve et al. An accurate prediction of tensile strain is required since the level of tensile strain developed in the base determines to a large extent the force developed in the geosynthetic and hence its effectiveness. All three of the above nonlinear models for representing resilient moduli were employed in the present study and their use will be discussed subsequently. [C-6] have shown that the material constants n and m are approximately related to G 1 as follows: 0. MODEL VERIFICATION . Partway through this study it became apparent that the usual assumption of material isotropy. were in general not C-6 . To be able to reliably predict the tensile strain in an unstabilized granular base is quite important in a study involving granular base reinforcement.028 G0. strain and deflection response variables. and the usually used subgrade and base properties.03 G 1 (C-5) (c.

including tensile strain in the aggregate base.5 kip (29 kN) uniform. of a high quality. well instrumented test section without geosynthetic reinforcement tested previously by Barksdale and Todres [C-7. The simplified. The second study used the extensive measured response data collected from Test Series 3 of the large scale laboratory pavement tests conducted as a part of the present study.4 million applications of a 6.indicating the level of improvement due to reinforcement observed in the weak section used in the first laboratory test series. Unreinforced.5 in.C-8]. In the verification study a number of models were tried including the nonlinear finite element k-13 and contour models. nonlinear C-7 . High quality materials were used including the asphalt and the crushed stone base which was compacted to 100 percent of AASHTO T-180 density. These sections were placed on a micaceous silty sand subgrade compacted to 98 percent of AASHTO T-99 density at a water content 1. Therefore. A total of about 2. several pavement sections having a 3. circular loading were applied at a primary and six secondary positions. Two independent comparison studies were performed to both verify the analytical model selected for use and to assist in developing appropriate material parameters.C-8]. High Quality Aggregate Base Pavement As a part of an earlier comprehensive investigation to evaluate aggregate bases. a supplementary investigation was undertaken to develop modified models that could more accurately predict the tensile strain and hence the response of geosynthetic reinforced pavements. (90 mm) asphalt surfacing and an 8 in.9 percent above optimum. (200 mm) thick granular base were cyclically loaded to failure [C-7. The first study involved theoretically predicting the response.

A cross-anisotropic representation has different elastic material properties in the horizontal and vertical directions. it was discovered that the tensile strain in the base greatly increased if the subgrade modulus increases with depth. A manual trial and error procedure was used to select material properties that gave the best overall fit to all of the measured response quantities. isotropic subgrade resilient modulus was used. cross anisotropic model were selected as having the most promise. the average stiffness of the subgrade was C-8 . The usually used isotropic model has the same material properties such as stiffness in all directions. Thus the important finding was made that the resilient modulus of the subgrade near the surface had to be quite low as indicated by the very large measured vertical strains on the subgrade. after the sensitivity study was under way. A comparison of the observed and measured pavement response variables for each model is given in Table C-1. The cross-anisotropic model using an isotropic. A homogeneous material has the same properties at every point in the layer. They are similar to those used for the homogeneous subgrade comparison in Table C-1. These results indicate that a cross anisotropic model is at least equal to. At the time this comparison was made a homogeneous. homogeneous subgrade was able to predict measured variables to within about ± 20 percent. Since the total measured surface deflections were relatively small. Later. The cross-anisotropic material properties employed in the sensitivity study are summarized in Table C-2. and perhaps better than the simplified contour model for predicting general pavement response. the one exception was the tensile strain in the bottom of the base which was about 30 percent too low.contour model and a linear elastic.

..• C .. S UBGR /STRA I 0 N 0 •••1 o .0 .4 4. rN .I 0 N 0 I.1 -4. .I A. *7 .11 0 0 1.1 O. 0 5 1• . ••0 MI CLI IX vi ..... .1 14 Y fa 0 > 42 (x 0 rn •••1 un co N .. .n trt 0./ o■ I 00 0.7 I..4:1 ui Le% 0 N 0 00 00 %. Cons tan t E s Va r iable STR . ■ c•.1 01 Q 0 1:1 al a 9 0 0 mmmmmm .. 1.°. . 70.4 -. 00 N • • 0• res i lien t modu lus E r BOTTO O.. rs • rs ..4 • .1 4.. N 0 0 • • 0 0 N 0 • 0 0 U E CM 0 O d 01 . 4141 0 ... I.■ 0 > • •••I La O etj > 0 Z MI 0 0o 72. 4 til 10Q M-I.. Cru she d 4..0 P AS E -6 10 ) STRAIN • 0 30 wi 40 N wi . Sec t ion : Table C... .I 00 N mmmmm 4BOT BASE • rl N . ..4 •4 -. > Y.. . Cros s ..') 0) cal 9 0 14 00 .01 Y ...4 N 1...4 x ' 'a 1..o•S O. r. so . .-s . 0 .} .. 7 (1 9A) I 1. r•$ 04 m ..? MI Fin ite E lemen t Mo de l ( 2 ) Measur e d NOI LI GNOD C-9 0 .-• NO 40 •••■ ..0 al 41 .11 '..7.J hi Z 01 .. at) c0 1-1 ri . r. r. 0 0 • . N -7 N .I .- TI in as eq (• 111m) 3 mm (T s 1) 2 qns lItte ) 3 I .r...-1 . .4 00 .-1 '0 E 0 0 +I 4.An iso trop ic E.1 C) ( -0 1 9 in.7 O... 00 . rs. .• 0 Q I..0 r-f 0 r-I 1. .4 0 0 .44 4..

45 0. 2.4 0.420Eb Middle lE b 0.375E s 0.4 0. E s s 0.15 0.000 psi. Eb = resilient modulus of base as shown in Table C-1.10 0.75E s s 1.15 0. .Table C-2 Anisotropic Material Properties Used for Final Georgia Tech Test Study Location in Pavement Poisson's Ratio Resilient Modulus Vertical I Horizontal Vertical Horizontal Aggregate Base (Anisotropic) Top 1.75 where E =8000 psi and Els (avg) = 35.4 0. Modular ratio Eb (avg)/E = 4.0852E b 0.136E b 0.4 0.200 psis the numerical average of the three vertical resilient moduli of base= 38.4 Subgrade (Isotropic) Top 0.0227Eb 0.818E Bottom b 1.43 0.4 0.375E Middle 0.75E s 1.43 0.875E s = average resilient modulus of elasticity of subgrade.875E Bottom Note: 1.

The isotropic.quite high. The measured pavement response obtained from the three sections included in Test Series 3 of these laboratory tests provide an excellent opportunity to verify the theory.2 in (208 mm). nonlinear finite element method could not predict at the same time large tensile strain in the bottom of the aggregate base and the small observed vertical strains in the bottom and upper part of that layer.5 kips (6. This important difference in measured strain is readily explained if the actual stiffness of the aggregate base is considerably greater in the vertical than the horizontal directions. The cross-anisotropic model gave a much better estimate of the vertical stress on the subgrade and the vertical surface deflection than did the nonlinear model. probably much larger than generally believed at the present time. Response of Geosynthetic Reinforced Sections A total of 12 well-instrumented laboratory test sections were tested as a part of this study.2 in. The significant decrease in strain and increase in confinement with depth probably account for most of this observed increase in stiffness with depth IC-10]. the stiffness of the silty sand subgrade underwent a significant increase with depth. A cross-anisotropic model was used to predict the response of the two geotextile reinforced sections and the non-reinforced control section included in the study. (30 mm). These test sections had an average asphalt surface thickness of about 1. Therefore. are described in detail in the last section of this chapter. which included the measurement of tensile strain in the aggregate base and also in the geosynthetic.7 kN) C-11 . The wheel loading was 1. and a crushed stone base thickness of about 8. These comprehensive experiments. The better agreement with measured pavement response when using a subgrade resilient modulus that rapidly increases with depth is shown in Table C-1.

A soft clay subgrade (CL) was used having an average inplace CBR before trafficking of about 2.9 percent. The postulation is presented that. the theory predicted observed response reasonably well.8 percent. only strain was measured in the geosynthetic. Once again. These sections only withstood about C-12 . The cross anisotropic model was less satisfactory in predicting the pavement response of the weak Test Series 3 sections compared to the stronger sections previously described. Of considerable interest is the fact that the largest calculated geosynthetic stress was about 10 lbs/in (17 N/m). The strain in the geosynthetic was over predicted by about 33 percent when the geosynthetic was located in the bottom of base. Overall. the subgrade used in Test Series 3 probably performed like one having a CBR less than the measured value of 2. were quite weak and exhibited very large resilient deflections.7 to 2. based on the measured strains. strains and stresses. It was under predicted by about 14 percent when located in the middle of the layer. under repetitive loading.6 MN/m 2 ). In summary. perhaps due to a build up of pore pressures. As a result. the conclusion was reached that the resilient modulus of subgrade was quite low near the surface but rapidly increased with depth. The comparison between the anisotropic model using the best fit material properties and the measured response is shown in Table C-3 for each section. The vertical stress on the top of the subgrade was about 50 percent too small. these pavement sections. These sections were constructed over a subgrade having a very low average resilient modulus that was back-calculated to be about 2000 psi (15 MN/m2 ).at a tire pressure of 80 psi (0. Larger radial strains were measured in the bottom of the aggregate base than calculated by about 50 percent. the computed vertical strain at the top of the subgrade was too small by about the same amount. as originally planned.

1 .. . n rot I I 0 .77-7 1 (( 010001- r4 n .11.1' Z UtfU t '3 443 (' u l) A9 C C vt• •••• 0 •1 3.. 590' 0 090' 0 59 .. . . . [7.. . .) 0 L.0 .. -.17.. Subg.0 .-• •-∎ 0 •C ■•• 44 WI 1.. Stress /Str a in . 5 902 OZ9 Z 000 9 •ll aft ■ I 0081 0807 1IWOI C' O I 00ES 0099 "iaaA nn I 11 se se ..1 ••••• I (9* Z 7.0 ..-6 ... I 81 81 (1) 00 79 ( 13 6 1 777771:77:77 E' 6 777 Geosy n t h et ic ES77t.. 'Ve r t.. S 90' 0 090' 0 1 1 990' 0 91 0' 0 1 0081 0110Z 1 1 1 1 E9' Z 71' 7 I • . w.0 .. C ia 0 . .0 00 10 0 > 4 01 7 •-•• 4..014 la PoN I Mode l 1 Mo de l2 Me asu re d Mode lI Mo de l 2 0 9ZE- I I Stra in CEO SYNT HET1C IN MIDDLE OF BAS E 1730 I 1 790 1 91a(96(- . Stress 0091 191 ( CEO SYNTHETICI NBOTTOMOFBAS E C S IZ I ORBI 66SZ 0091 I 08110991- I I /500019 - Condit io n 09 ZZ OROZ (1610(61- II I 1 CONTROL SECTION Ra dia l 1 NO GEOSYNTHETIC 717471:717- 008 7[1E 7- 01 9 2 Str a in Top o f 1862 1 579 1 M S91ZS- VCE . ■■ • . .0 ._ Z 1 . 00 • -.0 42 .? I .. 0..

The nonlinear. Nevertheless. and indicate correct relative trends of performance.6 times the value calculated using the cross-anisotropic base model. simplified contour model was also employed as the secondary method for general comparison purposes and to extend the analytical results to include slack in the geosynthetic and slip between the geosynthetic and the base and subgrade. Undoubtedly the analytical studies are susceptible to greater errors as the strength of the pavement sections decrease toward the level of those used in the laboratory studies involving the very weak subgrade. the stiffness of the . (38-50 mm) as compared to about 2. The subgrade used was isotropic and homogeneous. The measured strain in the bottom of the aggregate base in the test section study that withstood 2.5 to 2 in.70. A reasonably strong section would in general be more commonly used in the field. MODEL PROPERTIES USED IN SENSITIVITY STUDY The cross-anisotropic model was selected as the primary approach used in the sensitivity studies to investigate potential beneficial effects of geosynthetic reinforcement.4 million heavier load repetitions for the stronger sections on a better subgrade used in the first comparison. the calculated relative changes in observed response between the three sections did appear to indicate correct trends. To approximately account for this difference in strain.4 million load repetitions (Table C-1) was about 1. In an actual pavement the development of larger tensile strains in the granular base than predicted by theory would result in the reinforcing element developing a greater force and hence being more effective than indicated by the theory. This finding suggests relative comparisons should be reasonably good.000 load repetitions with permanent deflections of 1.

Hence nonlinear geosynthetic material properties are not required for the present study. actual geosynthetic stiffnesses. The resilient moduli of the subgrade included in the sensitivity analyses were 2000. Using the above engineering approximation. 1000. The resilient modulus of the asphalt surfacing used in the sensitivity study was 250. the level of tensile C-15 . 6000 and 12.500 psi (14. 6000 and 9000 lbs/in. The ratio of the resilient modulus of the base to that of the subgrade has a significant influence on the tensile strain developed in the base for a given value of subgrade resilient modulus.5 times the value reported. this important finding was not made until the sensitivity study was almost complete. 1000 kN/m). A supplementary analytical study using a higher geosynthetic stiffness with a homogeneous subgrade gave comparable results to a model having a subgrade stiffness increasing with depth. they remain well within their linear range.5 scaling factor. the corresponding stiffnesses reported as those of the sections would. Because of the small stresses and strains developed within the geosynthetics. In turn.geosynthetics actually used in the analytical sensitivity studies was 1. 4000 and 6000 lbs/in (170. 41. The corresponding Poisson's ratio was 0. (260. The relative values of crossanisotropic elastic moduli and Poisson's ratios of the aggregate base used in the study are summarized in Table C-4. 24. 1600 kN/m) were used in the theoretical analyses. Unfortunately. Tensile strains in the aggregate base and geosynthetic can be calculated directly by assuming a subgrade stiffness that increases with depth.35. Cross-Anisotropic Model Material Properties. 86 MN/m 2 ). 700. 3500. S g = 1500. be 1000. Therefore. using the 1.000 psi (1700 MN/m2 ).

12 3. Horizontal Vertical Vertical Horizontal 1.0458E 0.35 2.000 4.300 4.5 3 800 4.138E 0.15 Middle 1.925E 0. .825E 0. Granular Base: Position in Base Kl G1 'I' Very Coed Crushed Stone Base Upper 2/3 14.000 2 750 4.100 7.45 0.0. C-1)( 1) . .Table C-4 Aggregate Base Properties Used in Cross-Anisotropic Model for Sensitivity Study Location in Base Resilient Modulus Poisson's Ratio .10 Top Table C-5 Nonlinear Material Properties Used in Sensitivity Study 1.0 1 0 .320 1.14 1 0.14 Poor Quality Gravel/Stone Base Upper 2/3 3.0.640 3. Asphalt Surfacing: Isotropic. Er * 250.000 16.050 0. 1.620 0.000 psi (isotropic) 2.43 0. Average Subgrade E s • 6.950 Lower 1/3 5.300 4. v.000 1.375E 0.180 0.0E 0.43 0.12 Lower 1/3 1. Subgrade: Typical Subgrade E s (psi) given below (see Fig.4 C-16 .000 psi. Point 03 (psi) Resilient Moduli Top Middle Bottom 1300 16.300 30.15 Bottom 0.

using a modular ratio comparable to that actually developed in the field is very important. however.5 and 4.5.5 was about that observed for the full-scale test sections having the better subgrade.5. This was approximately the value back calculated from the measured response of the test pavement on the very soft subgrade having an average resilient modulus of about 2000 psi (14 MN/m2 ) as shown in Table C-3.strain in the aggregate base determines to a great extent the force developed in the geosynthetic. For example. The nonlinear options in the GAPPS7 program. The resilient properties of the asphalt surfacing were the same as used in the cross-anisotropic model. Both studies comparing predicted and measured pavement response indicate the base performs as a cross anisotropic material. The modular ratio of 4. Supplementary sensitivity studies were also carried out using modular ratios of 1. the small vertical strain and large lateral tensile strain in the aggregate base could only be obtained using the cross anisotropic model. For the primary sensitivity study the modular ratio used was 2. Nonlinear Properties The material properties used in the nonlinear finite element analyses were developed by modifying typical nonlinear properties evaluated in the past from laboratory studies using the measured response of the two test pavement studies previously described. only permit the use of C-17 . For this study the cross-anisotropic modular ratio was defined as the vertical resilient modulus of the center of the base divided by the uniform (or average) resilient modulus of the subgrade. Since the force in the geosynthetic significantly influences the improvement in behavior of the reinforced pavement system. the average resilient modulus of the subgrade was about 8000 psi (55 MN/m 2 ) as shown in Table C-1.

However. Reducing this resilient modulus caused the calculated vertical strain in the layer to be much greater than observed. Decreasing the resilient modulus of the lower part of the base. The compromise selected gave weight to increasing the radial tensile strain in the granular base as much as practical.4. A considerable amount of effort was required to develop the reasonably good comparisons with measured responses shown in Table C-1 and C-3 for both the cross-anisotropic and nonlinear models. The nonlinear subgrade material properties used in the study are also summarized in Table C-5. as well as the aggregate base properties. The nonlinear material properties used in the upper two-thirds of the aggregate base are essentially the best and worst of the material properties given by Jouve et al.5 gave better values of vertical strain in the base. 2. The nonlinear material properties employed in the simplified contour model are given in Table C-5.isotropic properties. A better match of calculated and measured response could probably be developed by further refinement of .5. The resilient properties used in the lower third of the base were obtained by multiplying the properties used in the upper portion by 0. Increasing the stiffness by 1. were developed from the trial and error procedure used to match the measured response variables with those calculated. [C-6] multiplied by 1. The subgrade properties. if the resilient modulus of the entire subgrade was reduced calculated surface deflections were too large. Therefore. The radial tensile strain in the bottom of the granular base could be increased by 1. Decreasing the resilient modulus of the top of the subgrade. some compromises were made in selecting the simplified contour model resilient properties of the aggregate base.

Estimation of Permanent Deformation The presence of the geosynthetic in the granular base was found to cause small changes in vertical stresses and somewhat larger changes in lateral stresses (at least percentage-wise) within the granular layer and the upper portion of the subgrade. Shear stresses do not act on these planes which greatly simplifies the analysis. The representative element is located beneath the center of the loading where the stresses are greatest.the process. it was found that the GAPPS7 program in its present form is not suitable for predicting the effects on rutting due to the relatively small changes in lateral stress. For this sensitivity study. the layer strain method consists of dividing the base and upper part of the subgrade into reasonably thin sublayers as illustrated in Figure C-2. For such relative comparisons the material properties developed are considered to be sufficiently accurate. For this location. only the relative response is required of pavements with and without geosynthetic reinforcement. During the numerous preliminary nonlinear computer runs that were performed early in this study. Therefore the layer strain method proposed by Barksdale C9] was selected as an appropriate alternate technique for estimating the relative effect on rutting of using different stiffnesses and locations of reinforcement within the aggregate layer. the principal stresses of and 03 are orientated vertically and horizontally. The complete stress state on the representative element within each sublayer beneath the center of loading is then calculated using either the cross-anisotropic or the nonlinear pavement model. Residual compaction stresses must be included in estimating the total stress state on the element. In summary. C-19 . respectively.

WHEEL LOADING h1 h2 •• • • •SUBLAYER 1 • p • • • 2 • • : t • • ASPHALT SURFACE • • • • • •• • E4)2 ••••• • iMmIlommie h3 h4 h5 03 3 D 0 1 4 5 A A V A AGGREGE BASE (33 C 4 Cl \. Comparison of Measured and Computed Permanent Deformation Response of a High Quality Crushed Stone Base: 100. En a_ CALCULATED -----7 In 40 b 173=10 PSI a3=5 PSI b cr) cr3=3 PSI Ce 20 ce 0 0 K=1880 Rf=1.70 c=0 a =3 PSI 3 0. C-20 .8 12 PERMANENT STRAIN.000 Load Repetitions.4 n=0. E p (PERCENT) Figure C-3.1 • 0 3 eP5 ••=1■0 01 h6 6 03 *a ' e6 03 SUBGRADE Figure C-2. Ref.1 0 PSI 40=61 4 0 00 0.0 10=58 ° c3 =5. Idealization of Layered Pavement Structure for Calculating Rut Depth (After Barksdale. C-9).

To predict accurately the effects of these small changes in stress on rutting. 01 . can greatly reduce permanent deformations under certain conditions. the permanent strain must be expressed as a continuous function of the deviator stress (q) and confining stress 03: c = f(q . The sum of the permanent deformations in each sublayer gives an estimation of the level of rutting within the layers analyzed. Placement of even a stiff geosynthetic within the aggregate base causes only small changes in confining pressure on the soil and also small vertical stress changes.The vertical permanent strain. The hyperbolic permanent strain model proposed by Barksdale [C-9] for permanent deformation estimation gives the required sensitivity to changes in both confining pressure and deviator stress.03 al = major principal stress acting vertically on the specimen below the center of the load a3 = lateral confining pressure acting on the specimen below the center of the load q = deviator stress. e. is then calculated in each element knowing an accurate relationship between the permanent strain E p and the existing stress state acting on the element. Total permanent deformation (rutting) is calculated for each sublayer by multiplying the permanent strain within each representative element by the corresponding sublayer thickness. these changes. when the element is highly stressed.03 Although the changes in confining stress are relatively small.03) (C-7) where: E = vertical permanent strain which the element would undergo when subjected to the stress state 03 and 0 1 . The hyperbolic expression for the permanent axial strain for a given number of load repetitions is C-21 .

000 cyclic load applications. n and Rf) used in the expression must be determined from at least three stress-permanent strain relationships obtained from at least nine cyclic load triaxial tests. The resulting stress-permanent strain curves are then treated similarly to static stress-strain curves. The soil-aggregate blend was about three times more susceptible to rutting than the high quality crushed stone base. The theoretical curve given by C-22 . The silty sand subgrade used in the comparative study was compacted to 90 percent of T-99 density.Ep = ( al 1 .8) 1 .sin0 where: 0 and c = quasi angle of internal friction 0 and cohesion c determined from cyclic loading testing Rf. Three different confining pressures are used in these tests. A comparison of the stress-permanent strain response predicted by the hyperbolic relationship given by equation C-8 and the actual measured response for the two bases and the subgrade are shown in Figures C-3 through C-5 for 100. 0. k and n = material constants determined from cyclic load testing All of the material constants (c.03 )/K c 3n 0 (a l 3 Rf 2(c. friable soil and 60 percent crushed stone compacted to 100 percent of T-180 density. K.cos0 + cr3 sin0) (C. Two different quality crushed stone bases were modeled for use in the sensitivity studies [C-9]: (1) an excellent crushed granite gneiss base having 3 percent fines and compacted to 100 percent of T-180 density and (2) a low quality soil-aggregate base consisting of 40 percent of a nonplastic. The subgrade had a liquid limit of 22 percent and a plasticity index of 6 percent.

e p (PERCENT) 24 DEVIATO RSTRESS. al —a3 ( PSI) 40 n=0. 40 20 0 00 0.265 c=0 a =10 PSI 3 473=3. 5 PSI 0 3 =10 PSI 20 0• 00 0. Comparison of Measured and Computed Permanent Deformation Response for a Silty Sand Subgrade: 100. Comparison of Measured and Computed Permanent Deformation Response for a Low Quality Soil-Aggregate Base: 100.8 1.000 Load Repetitions. al —a3 (P SI) Figure C-4.8 1.DEVIATORSTRESS.000 Load Repetitions.6 PERMANENT STRAIN. e p (PERCENT) Figure C-5.6 PERMANENT STRAIN. .

the hyperbolic model agrees quite nicely with the actual material response. The granite gneiss crushed stone had 0% passing the No. 200 sieve. liquid limit 22%. . plasticity index b.0 - 1. A-2-4(0)I. Table C-6 General Physical Characteristics of Good and Poor Bases and Subgrade Soil Used in the Rutting Study (1) GRADATION BASE DESCRIPTION 1 11 3/4 10 60 200 COMPACTION T-180 Y max (pcf) 4W (3) S (1) LA WEAR (%) 2 40-60 Soil/Crushed Cranite Gneiss Blend( 2) 99 85 42 25 13 138 5. 2. 10 sieve. 4. Table C-6 summarizes the general material properties of the base and subgrade.2 50 47 1 Slightly Clayey Silty Sand( 4) 100 100 100 63 40 115. silty fine sand ISM. the soil was a gray.5 73 45 6 Crushed Granite Gneiss 100 60 25 9 3 137 4. The material parameters used in the hyperbolic model are given in Figures C3 to C-5. Classification SM-ML and A-4(1). 40 and 20% < No. nonplastic with 73% < No. Data from Barksdale [C-9].4 13. Degree saturation in percent as tested. 3.

.. S. New Orleans (submitted for publication)... School of Civil Engineering. 1980.. pp. Ann Arbor. E. pp. C-2 Barksdale. Las Vegas.. C-3 Brown. 17-22. 1982.S. 1984. H. R. Robnett.. Lai. Nonlinear and Large Displacement Problems in Geotechnical Engineering". PhD Thesis. and Ragneau. R. Paute. C-5 Mayhew. 3rd International Conference on Structural Design of Asphalt Pavements. "Performance of a Thin-Surfaced Crushed Stone Base Pavement". J. August. pp. Transportation Research Record 954. Greene. Transport and Road Research Lab. "Crushed Stone Base Performance". R. pp. "A Study of Factors Affecting Crushed Stone Base Performance". 169 p. R. J.F..APPENDIX C REFERENCES C-1 Zeevaert.. J..C...D. 1987. and Pappin.. Report LR 1088. 161-174. "Laboratory Evaluation of Rutting in Base Course Materials". C-8 Barksdale. 50-64. Proceedings.. Atlanta. C.F. and Todres. R..D. 1981.D. R. Q.W. "Rational Model for the Flexible Pavements Deformations".D.D. Vol. C-10 Barksdale.. C-6 Jouve.. Georgia Institute of Technology. Transportation Research Record 810.L. "Finite Element Formulations for the Analysis of Interfaces. 375-380. A. 1987. S. II. and Machemehl. and Zeevaert-Wolf. pp. Georgia Institute of Technology. School of Civil Engineering. Transportation Research Record 810. Sixth International Conference on the Structural Design of Asphalt Pavements. .. Second International Conference on Geotextiles. J. and Pappin.M.D. Proceedings... Atlanta. Bush. "The Modeling of Granular Materials in Pavements".A. A. Transportation Research Board. 1983. 1972.E. pp.. Transportation Research Board. J. Transportation Research Board. Martinez.W.. C-9 Barksdale. 1982. 7887. Ga. H. "Experimental and Theoretical Behavior of Geotextile Reinforced Aggregate Soils Systems". P. 1981. "Resilient Properties of Unbound Roadbase Under Repeated Loading". 267 p. 17-22. C-4 Brown. A.S. "Analysis of Pavements with Granular Bases". ASTM Symposium on the Implications of Aggregate. Proceedings.. C-7 Barksdale..

INSTRUMENTATION AND CONSTRUCTION .APPENDIX D TEST SECTION MATERIALS.

some of the readily measurable material properties such as density. The granite aggregate gradation used in each bituminous mix is shown in Figure D-1. described in detail in Appendix E. placed and tested to insure as uniform construction as possible. Figure Ell.APPENDIX D TEST SECTION MATERIALS. water content and cone penetration resistance were frequently determined during and after the construction of the test sections. The properties of the pavement materials used in construction of the test pavements were thoroughly evaluated in an extensive laboratory testing program. a weak granular base was used during the first series of D-2 . aggregate bases and geosynthetic reinforcement materials were used in the tests. a gap-graded. Asphalt Surfacing. For quality control during construction. prepared in accordance with the British Standard 594 [D-1]. and the specifications of both mixes are summarized in Table D1. Hot Rolled Asphalt (HRA) mix was used. These quality control tests are fully described subsequently. The asphaltic concrete mix was prepared in accordance with the Marshall design results given in Appendix E. During the first series of tests. To enhance the benefit of a geosynthetic inclusion in the pavement structure. An asphaltic concrete mix was employed for the remaining three series of tests. INSTRUMENTATION AND CONSTRUCTION Materials All materials were carefully prepared. Two different asphalt surfacings. A brief description of the materials used in the experiments is given in the following subsections. Aggregate Base. The same soft silty clay subgrade was employed throughout the entire project.

co cus quality control tion.moisture-density tests . les were taken to t. As a 0 0 0 0 0 0 0 0 Co/ tfl N 0 NISSVd 30V11 ■133ad different from those .) Dn the asphalt O LU w' 1:5 L) tared to be 1.1 oo :ions were obtained from o he high deflections Or < stiffness of CL analysis was not 0 ther complicated by the O 7rounded by thick 'ones of the FWD. Density of the neasured by the nuclear Least ten core samples O O Lnd density. Falling Weight O st sections. Tests were o.er each test series.

f DOLOM AO . 60 CD 50 Of TIC 01 .0 0.5 10 2037.° Ir.. Gradation Curves for Granular Base Materials..B. SIEVES (P.ii 10 0 • 001 1.° LIMESTONE 40 1 6 30 20 0.1 PARTICLE SIZE (MM) 10 100 Figure D-2.S.M) 75 30300118 (MM) ...5 75 100 90 C5 80 z 05 ul 70 SAND & GRAVEL a. .

Table D-2 Properties of Geosynthetics Used. Geotextile Geogrid Polypropylene Polypropylene Weight/ area (oz/yd2) 28. X in.8 n/a Grid Size (in.99 Tensile Strength (lb/in) 886 119 Stiffness at 5% Strain (lb/in) 4300 1600 % Open Area 2 .5 5.22 X 1.56 .) n/a Polymer Composition 1.

was used to measure the surface profile. However. Transient stress near the top of the subgrade. Pavement Construction Subgrade. (450 mm) of fresh silty clay was placed after the same thickness of existing stiff subgrade material was removed. Beginning with the Third Test Series the transient longitudinal stress was measured at both the top and bottom of the granular layer. Each layer was compacted by using a triple legged pneumatic tamper (Figure D-5) which had sufficient energy to destroy the joints in the bricks. In addition to the instrumentation installed within the pavement. 4. During the construction of the first series of pavement sections. a profilometer (Figure D-4) consisting of a linear potentiometer mounted on a roller carriage. Temperature in each pavement layer. and at the complimentary location in the control section. 18 in. The final subgrade surface was then leveled with a single legged pneumatic compactor (Figure D-6) before the aggregate material was placed over it. The silty clay subgrade (Keuper Marl) was installed as 7 layers of wet bricks. The surface elevation of the subgrade was established by measuring the distance from a reference beam to various locations on the subgrade surface. Transient and permanent lateral strain in the geosynthetic. D-10 . The fresh silty clay subgrade employed in the first series of tests was reused for all subsequent tests. an additional 2. 5.3. since the design thickness for both the aggregate base and asphalt surfacing was increased after the first test series. (64 mm) of the newly installed silty clay was removed before construction of the Second Series pavement sections.5 in.

Single Legged Pneumatic Compactor Used on Subgrade. Vibrating Plate Compactor. D-12 . Figure D-7.Figure D-5. Tamper Used on Subgrade. Vibrating Roller. Triple Legged Pneumatic Figure D-6. Figure D-8.

APPENDIX E PROPERTIES OF MATERIALS USED IN LARGE-SCALE PAVEMENT TEST FACILITY .

.4 .tt tal 0 X 0 L Z 0 4. 0 ..1 0 <4 al `.. Fig ure E ....I • ■-I Ln 1..- t-N (edW) sninoow / 0 0 . VI 1..14 o LA A4 0 0 in . 0 0 X 0 tip Z co .1.1••• 1N3111S38 0 tr.. Ul LH Li 14.

E-6.15 psi (0 to 70 kPa). Loach also carried out repeated load triaxial tests on compacted 3 in. It has been found (E-5. Using a similar procedure to that adopted by Loach and with the aid of a computer-controlled servo-hydraulic testing system.Cyclic Load Triaxial Test.35 psi (0 to 30 kPa) and 0 to 10. Apparatus developed at the Transport and Road Research Laboratory (E-9). The results of his tests are shown in Fig. Therefore. The ranges of cell pressure and repeated deviator stress he used during these tests were 0 to 4. respectively.3 (suggested by Croney) Er = Elastic Stiffness in kPa q r = Repeated deviator stress in kPa A = 2740 B = 2. (76 mm) diameter cylindrical samples of Keuper Marl. Loach (E-4) carried out some soil suction tests on samples of compacted Keuper Marl at their original moisture contents using the Rapid Suction.E-7) that relationships exist between soil suction and elastic stiffness for saturated and near saturated clay. four additional tests were performed on recompacted samples obtained from the pavement test sections. E-2. The results of these tests generally conformed with those obtained by Loach who suggested the following equation to model the elastic stiffness of compacted Keuper Marl: Er q r ( u+ap ) A q B r where: u = suction in kPa p = cell pressure in kPa a = 0. a series of soil suction and cyclic load triaxial tests are required.1 . in order to determine the general resilient properties of Keuper Marl.

compaction tests. (150 mm) diameter samples of the two types of material at various moisture contents. Details of procedure and equipment for carrying out cyclic load triaxial tests on granular material were described by Pappin (E-10) and Thom (E-11).2 to 10. The results of earlier testing showed that resilient behavior of a granular material under repeated loading was very stress dependent and. A total of six tests were carried out on recompacted 6 in. hence allowing permanent strain to accumulate. sieve analyses and other index tests. For the permanent strain behavior of Keuper Marl. A cell pressure of 0.2 psi (15 to 70 kPa) were used. E-3. The permanent strain tests were carried out at a frequency of 4 Hz and with a 2 second rest period.8 kPa) and repeated deviator stresses in the range of 2. Cyclic Load Triaxial Test. Comparison of the index properties between Bell's soil and the one used in the current project showed them to be similar. The increase of permanent axial and radial strains with number of cycles for the tests are summarized in Fig. the results obtained by Bell (E-3) was found to be the most applicable. . Tests on Granular Base Material Laboratory tests performed on the granular materials consisted mainly of cyclic load triaxial tests. 2) A permanent strain test where the stress path was considerably closer to the failure condition.26 psi (1.Both A and B are constants derived from experiments. Each cyclic load triaxial test was subdivided into: 1) A resilient strain test where the stress paths were far away from failure with the resulting strain essentially recovered during unloading and. therefore.

.„.ST ... 3 ....1 moisture content 17. .... 0 0 -2 T \....... Permanent Axial and Radial Strain Response of Keuper Marl for a Range of Stress Pulse Amplitudes (After Bell). ..t: n ... 20 E q = 50 a. q = 30 U) .10377:1-204 ---..s q = 50 2 tt 1 q = 30 10 0 1 61 q =15 102____4...r4 = 70 cc Figure E-3..494 5 — I Scatter q t in 101/m2 q' 70 Ver t ic a l Permanen t S tr ain 4 xi .Z— C Ere' s 2 ——— .

n ) K = K i p. The equations which Boyce used in the interpretation of results are as follows: G = G1p' (1. E-4. respectively.gq/p1)2} where p' = 1/3 (aa + 20c ) q = 1/2(aa . and the shear modulus. Figure E-5 indicates that the sand and gravel has a rather low resistance to permanent deformation. to characterize resilient strain. The results for the permanent strain tests for the two types of granular material are shown in Figs. For permanent strain tests. Based on the above equations. For the dolomitic limestone. each of the six tests used 20 stress paths. K. The dry densities of the test samples are shown in Table E-2. E-5 and E-6. The ranges of repeated cell pressure and repeated deviator stress used in the tests were 0 to 36 psi (0 to 250 kPa) and 0 to 29 psi (0 to 200 kPa).ac) and K1.nonlinear. Up to 2000 stress cycles at a frequency of about 1 hz were applied to the test samples. as shown in Fig. the deviator stress.n)/{1 . The results are presented in the form of change of permanent axial and radial strains with the number of stress cycles. the results of the resilient tests are summarized in Table E-2. Fig. The results of the resilient strain tests were interpreted by means of Boyce's model (E-12) which expressed the bulk modulus. a cell pressure of 7. as a function of both p'.G1. E-6 indicates that the rate of development of permanent deformation varies with E -9 .n and 3 are constants to be determined by experiments.3 psi (50 kPa) and a repeated deviator stress of 0 to 29 psi (0 to 200 kPa) were used. Hence. the mean normal effective stress. G.( 1. and q.

.1 C') N H m rICI 0 w 4-1 WW V0 0. op C' cc) CMc) cc) . —{ .—( 0 cn 8 H m m N s cn Cn .) W IS) . Cr) 0 4-1 d' 01 N 0 0 CO cc) cc) 0 s •rt cc. It N 0 • •l• m N co.-1 cc) .-4 cn In . Ln ki) . cn m en cc) M ..) - 0 ca) a) 4-) g 0 •• 0 . co 0 .--4 v > ED ••i 0 .1 ZE ).—( . cc) d am' 0 Cr) N OO CO N.Volumetric Strain Coefficients Mois ture Content (%) Type of Material I ( 3m') AqTsuaa Aaa I 111 Q TD Shear Strain Coefficients ■ z s2 `. 00 ZE 1-4 -.-4 h N %:Ir . is') N . 0% W 0 o cc) •:). m m .I W C') ■—( .: M cc) V m cc) 0 0 M 0 O Ln N N CI\ cc) 0 .I r. OD s N 00 .-1 00 Z5 1 1 -.--1 (JO (NI rn cl■ in iD Qs . cf) M .) V 0 g coa) (nal Cr) H w g u) a) 0 a) 4-) A u) AN $.-1 00 ZE L.0.—I W to Is H N H C**) H a) a) ro a) 44 g u) ra a) 4.I O)-4 ZE $4 ••.

. Stress Paths Used in Cyclic Load Triaxial Tests for Granular Materials. Figure E-4.Stress Paths for Elastic Stiffness Testing Stress Path for Plastic Strain Testing Deviator Stress (kPa) 200- 10 0- 0 0 100 200 Mean Normal Stress ( kPa) .

10 NUMBER OF CYCLES 100 Figure E-5. E -12 ..ce z Z< / 0o . 6 c b) RADIAL STRAIN' 'em. Permanent Axial and Radial Strains Response of Sand and Gravel During Repeated Load Triaxial Test..- I Z X ffi CL 0 1 . ff! 3 9 a) AXIAL STRAIN ( cc = 50 kPa q r = 200 kPa •E z6 (1) 5k<3 ara z z 0- cp 0 13 1 100 10 .. 70 5 = E a / a z tx Z 3 ' _..--ir..IF .

2 6. ' 2.0 1 b 100 1000 o 1 0000 2.0 * 8. 2 + 6.0 a 0.1 o 8.5 • I-- .5 1.ir. .0 c w(%) Sr(%) o TY 40.0 •0 6.5 w(%) Sr(6) 0 I3 45747 a x 4.8 b) RADIAL STRAIN F-5 1.8 cr la)AXIAL STRAIN oc = 50 kPa qr = 200kPa 1. 6-7 63.1.0 1 10 100 1000 10000 NUMBER OF CYCLES Figure E-6.4 94.5 x 4.2 4. Permanent Axial and Radial Strain Response of Dolomitic Limestone During Repeated Load Triaxial Test at Various Moisture Contents (w) and Degree of Saturation (Sr).7 611 2.5 cr 0.50.5 0.0 51.2 rg 2._3.7=r- 1-0 ag w <0.0 51.0 27.4 94.0 27.

In each test.moisture content and as the material approaches saturation. E-7. while the PI of the fines for the dolomitic limestone was found to be 3 percent. The shear box used for these tests measured 11. Tests on Geosynthetics Large Direct Shear Box Tests. In general. (170 mm) high. Details of the tests and the results are shown in Table E-3 and Fig.1 percent. For the sand and gravel. A series of compaction tests were carried out in order to determine the optimum moisture content and maximum dry density of the compacted material. (300 mm) square by 6. very rapid increase in the rate of deformation will occur. Index Tests. the index varied from 3. the British Standard Vibrating Hammer method (E-8) was adopted. The fines for the sand and gravel were found to be non-plastic.7 in. The result of the test indicated an index of 9 percent overall while for individual size fractions. The results of the tests for the two materials are shown in Fig. respectively. Twenty-four large direct shear box tests were performed on the two geosynthetic materials in conjunction with the soil and granular materials. Two plasticity index tests were carried out for the fines (less than 425 micron) of each of the two granular materials. the moisture content and dry density of the material at the time of the large scale pavement test were simulated. For most of the E -14 . Compaction Tests. One flakiness index test BS812 (E-14) was performed on the crushed dolomitic limestone used in the third series of tests. the same material was used in both the upper and lower half of the shear box. the test was carried out according to the ASTM D-1557 test method (E-13) while for the dolomitic limestone. Compaction was carried out by using a hand-held vibrating hammer.8 to 16. E-8.8 in.

Figure E-7. Results of Standard Compaction Tests for the Granular Materials.2 4 6 MOISTURE CONTENT (%) 10 12 Note: Sand and gravel are compacted according to ASTM D-1557 test method [E-13] while dolomitic limestone uses the British Standard vibrating hammer test method [E-8]. .

55 1.55 1.54 1.30 16 17 18 Nicolon/Keuper Marl 107 109 110 16.54 1.35 0.9 16.75 .06 10 11 12 Tensar SS1/Limestone 139 139 141 5.54 1.21 0.4 0.30 LA irm Moisture Content (%) N mo Dry Density (pcf) Test No Normal Stress (tsf) 11111= rs.18 0.4 0.75 1.2 3.12 2.27 2.0 4.06 .30 .39 .99 1.18 0.57 1.6 0.46 0.5 16.9 0.62 1.2 0.36 0.0 0.30 Nicolon/Limestone 138 137 138 5.47 0.06 2.75 1.8 16.2 3.75 1. Summary of Large Shear Box Tests.06 .2 16.06 Sand & Gravel 138 136 136 3.15 2.30 .55 1.06 .20 0.30 .30 .10 0.06 13 14 15 Crushed Limestone 138 140 138 5.30 .95 1. 1.OD ON ■ .3 16.06 .30 .4 3.06 .10 2.6 5.10 2.18 0.22 2.7 4.38 0.10 2.30 .55 1.65 1.30 .3 0. Type of Geosynthetic/Soil Shear Stress (tsf) Shear Rate (mm/min) Nicolon/Sand&Gravel 140 138 138 3.7 5.14 .06 .70 1.30 .48 .46 .00 .30 22 23 24 Keuper Marl 105 107 108 16.0 4.18 0.30 .6 16.07 2.8 3.10 2.Table E-3.30 19 20 21 Tensar/Keuper Marl 106 109 111 16.48 0.06 .30 .4 0.9 5.29 2.

..GEOTEXTI LE •S .‘• 0 o 53 • 0 col ' 0 / NM) SS32:L1S 2:1V31—IS Figure E.GEOTEXT I LE O 0 0 (NI vs zUJ 0 to 0 q■ 0 if) LP c ai cu 3 1.. 0 CNI N ' I Csi .J ••••••■• -J CC 0•5 I L ..Ct cu cu bD ro C 8 as LA L.8.-- I o o o o 0 in o 0 0 in —1 w J.3 CO = 4. 5 O CD .\ 4.+.

E-9. All tests were conducted at a standard test temperature of 68°F (20°C) and were continued until rupture occurred. each with a different sustained load. maximum shear stress was obtained at a horizontal displacement of less than 0. All tests were carried out at 68°F (20 °C) and.5 in.2 in. E-10. These tests were carried out at the University of Strathclyde where specialist apparatus was available (E-15).4 in. For the Hot Rolled Asphalt. The aggregate used in the design mix had a maximum particle size of 0.5 percent per minute was necessary. The results of the two sets of tests during the first 10 hours are shown in Fig. A grade 50 Pen binder was used. The result of the test is summarized in Fig. in most cases. E-12. (12 mm) with grading as shown in Fig. the maximum sustained load corresponded to 60 percent of the tensile strength of the material. the use of a faster rate of 7. A standard shearing rate of 2 percent per minute was used for the geogrid but for the stiff geotextile. lasted for 1000 hours. E-11. up to five separate tests. One series of Marshall tests (ASTM D1559) was carried out for the design of the asphaltic concrete mix. Background and details of the test was reported by Murray and McGown (E-16). Creep Test. because of the requirement of a much higher failure load. All creep tests were carried out in isolation with no confining media. Tests on Asphaltic Materials Marshall Tests. The results of the tests for both materials are shown in Fig. However. For the geogrid.tests involving granular material. Wide Width Tensile Test. for tests with Keuper Marl. a recipe grading E -18 . For each geosynthetic material. (10 mm). were performed. a horizontal displacement of up to 1. (30 mm) was required to achieve maximum shear stress.

.'' 00 10 15 20 25 25 b) GEOGRID 20 - I 2 15Strain rate = 2%/min a 10 910- 5 5 10 AXIAL STRAIN ( %) 15 20 Figure E-9..5%/min }. . Variation of Axial Strain with Load in Wide-Width Tensile Tests.200 a) GEOTEXTILE 150 - Z0 7 Strain rate = 7." 0 ° .

.BREAKAGE -J <X R 6 a 0 a 1 4 6 TIME (HOUR) 10 Figure E-10.1 kN/m x 8.8 6 a1GEOTEXTILE ▪ 5 kN/m • 7. . Results of Creep Tests at Various Sustained Loads for the Geosynthetics During the First 10 Hours.5kN/m • 15 kN/m 0 00 0 18 4 2 6 b) GEOGRID a 4.3 kN/m 8 10 x...2 kN/m 12.

OF MIX 9 Figure E-11. OF MIX g: 30 25 o 2° 15 10 3 4 5 6 7 8 (%) AC BY WGT. OF MIX 35 184 5 6 7 8 9 WO AC BY WGT.UNITWEIGHT(lb / ft3 ) • 154 12 152 0 8 ct 150 148 4 5 6 7 8 9 eYo AC BY WGT. OF MIX o4 5 6 7 8 9 (%)AC BY WGT. . OF MIX 212200 21 Q1 900 o • 20 isoo 19 F- x13004 (%) 5 6 7 8 9 BY WGT. Summary of Hot-mix Design Data by the Marshall Method.

..... I a o..1 2... . X Figure E.._ ER 53 ° (53 5i (Z? 2 R 0 0 a ASE "''''' .:. C.. 0 ow ..■•■.) X z Q a in U. X U..) . . ?Oill SIE C. Gra da t ion Curves for Agg rega tes Use d in Ma rsha ll Tes ts. . N_ 0 I • a l a ONISSIld 30VIN3083d I Li.

C. . S. Conf. A.L.F. E-3 Bell. University of Nottingham.5 percent. E-8 British Standards Institution. Nair.. "Repeated Loading of Fine Grained Soils for Pavement Design". and Hyde. E-7 Brown. "The Behavior of Anisotropically Consolidated Silty Clay Under Cyclic Loading". 25.. K.. 1972.. 1972. S. PhD Thesis. University of Nottingham.F.F. E-2 Overy. A. The average test results of the six samples are shown in Table E-4. C.N. E-6 Finn. "Application of Theory in the Design of Asphalt Pavements". F. Two viscosity tests were carried out by the Georgia Department of Transportation on the 50 Pen binder used for the asphaltic concrete mix. 1982. PhD Thesis. For comparison purposes. a specification which was used for the last three series of tests. 1982. "Methods of Testing Soils for Civil Engineering Purposes". HMSO. Lashine. C. made out of the HRA used in the first series were tested. "The Prediction of Permanent Deformation in Flexible Pavements".. University of Nottingham. PhD Thesis. 1987.A.F.K.. PhD Thesis. BS1377. Proc. D.. A. The Journal of Geotechniques..F. R. Also shown in the table are the test results obtained from an asphaltic concrete sample with a binder content of 6. Vol.L. University of Nottingham. APPENDIX 13 REFERENCES E-1 Hyde. E-5 Croney. "The Design and Performance of Road Pavements". London. Viscosity Test. and Monismith. Vol. 1977. 1. "Repeated Load Testing of Soils".. E-12 with 8 percent of 100 Pen binder was used. "Repeated Load Triaxial Testing of a Silty Clay". 1975.L.as shown in Fig. The viscosity at 140°F (60°C) was found to be about 4600 poises.. London. on the Structural Design of Asphalt Pavements.. 1987. six Marshall samples. E-4 Loach. of 3rd Int.

) 16. Hot Roller Asphalt Asphaltic Concrete 8 6.Table E -4.6 19 Corrected Stability (BD) 2028 2150 Flow (1/100 in.5 144 152 6 2.5 18 ■ Binder Content (% by weight) Mix Density (pcf) Air Void (%) .5 VMA (%) 23. Caparison of Marshall test data for two asphaltic mixes.

1985...08. "Geotextile Test Procedures Background and Sustained Load Testing". J. BS 812. and West..J. "Soil and Rock. Building Stones. University of Nottingham. 1975. Standard D-1557. "Soil Suction by the Rapid Method on Apparatus with Extended Range". University of Nottingham. 19.W.H. "Methods for Determining the Flakiness Index of Coarse Aggregate". "Characteristics of a Granular Material for Pavements Analysis". 1979. E-16 Murray. 1985. Vol. E-14 British Standards Institution. "The Behavior of a Granular Material Under Repeated Loading". A. and Brown. 1987. Sections 105. J.L.R.. Geotextiles".F. E-12 Boyce. E-13 ASTM Standard. University of Nottingham. .. PhD Thesis. 1987. TRRL Application Guide 5..1. "The Behavior of Polymeric Grid Used for Soil Reinforcement".. "Design of Road Foundations". and McGown. Vol. M. PhD Thesis. University of Strathclyde. 1976. The Journal of Soil Science. N. 1985. No.. 1.T.. PhD Thesis. E-15 Yeo.E-9 Dumbleton. K. Interim Report to SCRC. E-10 Pappin. E-11 Thom. R. 04. S. G.

APPENDIX F SEPARATION AND FILTRATION .

silt and clay may with time contaminate the lower portion of the drainage layer. subbases and drainage layers in pavements. If. Figure F-1 shows that an increase in fines of up to 6 percent can have a minor effect upon the resilient modulus [F-2].A poor physical separation of the base/subbase and subgrade can result in mechanical mixing at the boundary when subjected to load.A slurry of water and fines (primarily silt. in many cases even a more densely graded layer) is placed directly on the subgrade. clay and fine sand size particles) may form at the . The intrusion of fines into an aggregate base or subbase results in (1) Loss of stiffness. and (4) Reduction in permeability. when the level of contamination becomes sufficiently great. 2.APPENDIX F SEPARATION AND FILTRATION INTRODUCTION In recent years. the effective thickness and strength of the aggregate layer is reduced. however. Separation . an open graded layer (and. In either case. considerable interest has been shown in using opengraded aggregate layers as bases. A well-designed drainage system has the potential for increasing the life of a flexible pavement by a factor of forty or more [F-1]. (2) Loss of shear strength. Other work. indicates contamination of a portion of an aggregate layer with 2 to 6 percent clay can cause reductions in shear strength on the order of 20 to 40 percent [F3]. Filtration . however. (3) Increased susceptibility to frost action and rutting. Contamination due to the intrusion of fines into the base or subbase can be caused by the following two mechanisms: 1.

Influence of Subgrade Water Content and Geosynthetic on Stone Penetration (After Glynn & Cochrane. F-2).-0 WOVEN 0.A THICK NONWOVEN CI••• CO1IPOSITE /C) . . / 10 .100 0 0 10 5 15 20 ADDED FINES (PERCENT) Figure F-1. Ref. F-31).- Q.. fA w /. 0 0.• -0 THIN NONWOVEN A . Influence of Added Fines on Resilient Modulus of Base (After Jorenby..• NONREINFORCED • -. 3 - 76 •Ir ' • •4ro 0 15 20 24 28 WATER CONTENT. 20 E LEGEND o 15 • • • .. w (PERCENT) Figure F-2. Ref..

the filtration/ separation aggregate filter or geotextile must: (1) Maintain a distinct separation boundary between the subgrade and overlying base or subbase. the slurry will move upward under pressure into the aggregate layer and result in contamination. ductile and abrasion resistant to survive construction and in service loading.top of the subgrade when water is present and under pressure due to repeated traffic loading. and (4) Because of the relatively harsh environment which can exist beneath a pavement. Formal filter . Unfortunately. If the filtration capacity of the layer above the subgrade is not sufficiently great. (3) Must not become sufficiently clogged with fines so as to result in a permeability less than that of the underlying subgrade. the geotextile must be sufficiently strong. the classical Terzaghi filter criteria used for steady state filter design are not applicable for severe levels of pulsating loading. In harsh environments some clogging and loss of fines through the geosynthetic will occur. (2) Limit the amount of fines passing through the separator so as not to significantly change the physical properties of the overlying layer. Comprehensive state-of-the-art summaries of the separation and filtration problem have been given by Dawson and Brown [F-4]. a filter cake probably does not develop in the soil adjacent to the filter [F-6 through F-8]. such as occur beneath pavements where the flow may be turbulent and also reversing. Jorenby [F-2] and more recently by Dawson [F-5]. FILTER CRITERIA FOR PAVEMENTS To perform properly for an extended period. For these conditions.

have not yet been developed for aggregate or geotextile filters placed at the interface between the base and subgrade of a pavement. with the subgrade squeezing upward into the pores of an open-graded stone as it penetrates downward. including filter criteria and infiltration. SEPARATION Maintaining a clean separation between the subgrade and overlying aggregate layer is the first level of protection that can be provided to the base. A separation type failure can occur either during construction or later after the pavement has been placed in service. are summarized in Part III of Table F-1.F10. Christopher and Holtz give a good general discussion of the engineering utilization of geotextiles. The classical Terzaghi criteria were developed for uniform. however.criteria. This type problem is described in the report as a separation failure. Contamination due to washing of fines into the base from seepage is referred to as filtration. The geotextile selection criteria given by Christopher and Holtz is also summarized in Table F-I for both steady state and cyclic flow conditions. Separation Failure Mechanisms Contamination of the base occurs as a result of the aggregate being mechanically pushed into the subgrade. cohesionless soils in contact with an aggregate filter. These criteria.F-I1]. The total thickness of this contaminated zone as a result of separation problems (as opposed to filtraton) is typically up to about 2 times the . Most serious separation problems have developed when relatively opengraded aggregates have been placed on very soft to soft subgrades [F-3. which was taken from Christopher and Holtz [F-9]. which assumes steady state flow conditions.

Suggested performance test method: Gradient Ratio < 3 Loss Critical/Non-Severe Applicatons B. AGGREGATE where: and 0. Additional Qualifier (Optional): 005> 3D13 4. Less Critical/Leas Severe and (with Clean Medium to Coarse Sands and Gravels): k (fabric) > k (soil) 1.4) x slot width 055 (filter) > (1.5 Cu 4 < Cu < 3 1•11/Cu >50% Passing Woven: 093 < U13 059 < 0. use only the gradation of soil passing the U. 1111. 50 sieve 1. No. Pulsating. Ref. the effective flow area is reduced by 50%. or after selection before bid closing. Tests to be performed by specifying agency or his representative. Filtration tests are performance tests snd cannot be performed by the manufacturer as they depend on specific soil and design conditions. CEOSYNTHETIC FILTERS I. Critical/Severe Applications: k (fabric) > 10 k (soil) B.0 to 1. respectively. II. and S5%. by dry weight.S. Note: experience required to obtain reproducible results in gradient ratio test. 200 sieve AOS -. For example.6 Dim AOS No. gap-graded soils and silty type soils) where filtration is of concern. prequalifying the fabric. and perform soil/fabric filtration tests before specification. 50%. When the protected soil contains particles from 1 inch size to those passing the U. 4 sieve in selecting the fabric.S. 2. (fabric) > No. 015. D of which 15%. if 50% of fabric area to be covered by flat concrete blocks. No.5 • the diameter of soil particles.g. Qualifiers in potential clogging condition situations (e. Select fabric on the basis of largest opening value required (smallest AOS) II. Critical/Severe Applications1 Select fabric meeting I. PERMEABILITY CRITERIA" A. SOIL RETENTION (PIPING RESISTANCE CRITERLA) 1 Soils Steady State Plow <50% Passing= U. Permeability should be based on the actual fabric open area available for flow. F-9) I.5 Ds3 U.TERZAGHI CRITERIA TOE STEADY TUN Piping Requirement: D13 (filter) < 5 D o (soil) Permeability Requirement: 013 (filter) > 5 D 15 (soil) Uniformity Requirement: DSO (filter) < 25 Dso (soil) Well screens/slotted pipe criteria: 095 (filter) > (1. and Cyclic Flow 095 < 013 (If soil can move beneath fabric) or 050 < 0. 3.S.2) x hole diameter II. 200 Sieve Nonwoven: 095 < 1.095 < 3 0s5 Dynamic. 200 sieve. Additional Qualifier (Optional): 0 15 > 3D53 Note: 1. finer than that grain size. . III.S. Whenever possible. 2. Alternative: use approved list specification for filtration applications. No. fabric with maximum opening size possible (lowest AOS No.Table F-1 Design Criteria for Geosynthetic and Aggregate Filters (Adapted Christopher and Holtz. Porosity requirement based on graded granular filter porosity FILTERS . 1. Effective Open Area Qualifiers = : Woven fabrics: Percent Open Area: > 4% Nonwoven fabrics: Porosity > 30% 3. of the soil particles are. CLOGGING CRITERIA A. 2.) from retention criteria should be specified. 050.2 to 1. No.5 D os Cu < 2 or > 6 3 ■ 1 2 < Cu < 4 1■0.

light roller to as much as 15 to 20 tons for very large rollers. 4. The subgrade strength. are both important factors affecting stone penetration. et al. Bell. and as a result the subgrade moisture content. Under unfavorable conditions such as a heavy loading and a very weak subgrade. the depth of contamination could be even more. It might also occur later as construction traffic passes over the base before the surfacing has been placed. [F-3] found for a very large. (460 mm). smooth wheel vibratory roller.diameter of the aggregate which overlies the subgrade [F-3. Figure F-3 summarizes the vertical stress caused at the subgrade interface by a typical 4. (110 mm) diameter aggregate. smooth drum vibratory roller for initial lift thicknesses up to 18 in. giving a total contamination depth of approximately one aggregate particle diameter. Even a reasonably light roller applies relatively large stresses to the top of the subgrade when an initial construction lift is used of even moderate thickness. The largest vertical subgrade stresses usually occurs during construction of the first lift of aggregate base.5 ton. Construction Stresses The critical time for mixing of the subgrade with the aggregate layer is when the vertical stress applied to the subgrade is greatest. As the moisture content of the subgrade increases above the optimum value. Linear elastic F-7 .F-13]. the stone penetration to be about equal to the radius of the aggregate. the tendency for aggregate to penetrate into it greatly increases as illustrated in Figure F2.5 in.F-12. Smooth drum vibratory rollers develop dynamic vertical forces varying from 4 tons (or less) for a small. A similar amount of squeezing of the subgrade was also observed. 8 and 17. The common practice is to compact an aggregate layer with a moderate to heavy.

F-8 . 6 12 18 COMPACTED UFT THICKNESS 24 Figure F-4. T (IN) 18 Figure F-3. Variation of Vertical Stress on Subgrade with Initial Compaction Lift Thickness and Roller Force. -I ix I 20 ce 0 0 6 12 COMPACTION LIFT THICKNESS. Bearing Capacity Failure Safety Factor of Subgrade During Construction of First Lift.50 0N b LI 40 03 cn cn L.

Hence. a local bearing failure occurs below the tips of the aggregate.1) gulf = ultimate bearing capacity of the subgrade c = undrained shear strength of a cohesive subgrade The above equation is for plane strain conditions such as would exist beneath a long vibratory roller. the ultimate bearing capacity is about 20 percent greater than given by equation (F-1). the modulus of elasticity of the first 6 in. Because of the presence of the soft subgrade. As a result. present in coarse.layered theory was used in developing these relationships. (150 mm) thickness of the initial lift was assumed to be 1. Bearing Capacity Analysis For a separation problem to develop.5 times that of the material underlying it.2c where: (F. The ultimate bearing capacity of a cohesive subgrade can be expressed as [F-14]: quit = 5. opengraded granular materials. the actual average vertical stress developed on large stone particles at the subgrade interface is greater than the average stress predicted by conventional stress distribution theories. The vertical stress at the subgrade interface predicted by conventional layered theory requires continuous contact on a horizontal plane between the two layers. however. Large pore openings are. and the soil squeezes upward between the aggregate into the open pores. When the load is applied over a circular area. the externally applied stress level must be near the ultimate bearing capacity of the subgrade.5 times the modulus of elasticity of the subgrade.(150 mm) thickness within the lift was assigned an elastic modulus equal to 1. . Each successive 6 in. which is approximately the case for a wheel loading.

the irregular-shaped aggregates will be readily pushed into the subgrade. The possibility exists that the pores in the lower. contact stresses between the aggregate and subgrade will be even higher than the average stress given by Equation (F-2). Therefore. The bearing capacity is probably somewhat greater than obtained from applying equation (F-1) which does not consider the resistance to flow of soil through the pores of the stone. When stone penetration equals about the effective radius of the stone. weak soil. tensile portion of the aggregate layer open slightly as the external load moves over [F-5]. usually during the construction phase. until more research is performed in this area. a simplified approach can be F-10 .The actual average vertical stress a* for an open-graded base is approximately equal to: (F 2) a * = a /(1-n) z z where: - a * = actual average stress developed on the stone particles z a = theoretically calculated vertical stress n n = porosity of the granular layer The problem is further complicated by the fact that the aggregate particles are both three-dimensional and irregular in shape. until penetration of the aggregate particles into the subgrade occurs. the strength of a cohesive subgrade is greater than under slow loading. Several additional factors further complicate the aggregate penetration problem. Therefore. For conditions of a wet. However. Because of the overall complexity of the problem. the average contract stress between the stone and soil becomes close to that given by equation (F-2). a rigorous theoretical prediction of soil intrusion is quite difficult. Under dynamic loading. several passes of the roller may result in reduction in strength due to the build-up of pore pressures in the subgrade.

(300 mm) is probably reasonably typical. the permanent deformations in the subgrade increases with each load repetition. A lift thickness of 12 in. For these conditions a general bearing capacity failure. could occur for undrained shear strengths less than about 200 to 400 psf (10-20 kN/m 2 ). respectively. frequently the first lift to be constructed is placed at a greater thickness than used for succeeding lifts because of subgrade instability problems caused by the construction equipment.(150 mm). Where very soft subgrades are present. Aggregate penetration and excessive permanent deformations during construction can occur at even lower stress levels. These permanent deformations are due to accumulation of permanent strains at stress levels below failure but above the permanent strain yield stress of the material. Construction Lift Thickness For an initial lift thickness of 6 in. the average vertical subgrade stress varies from about 8 to 16 psi (55-110 kN/m 2 ) as the dynamic roller force increases from 4 to 17.5 tons.5 tons (Figure F-3). These stress levels are sufficient. to cause a general bearing capacity failure of a very soft to soft subgrade having undrained shear strength less than about 400 to 800 psf (19-38 kN/m2 ). Permanent Deformation Under repeated loading at a stress level below failure. For this lift thickness. based on equation (F-1). . as predicted by equation (F-1). as predicted by equation (F-1). the average vertical stress at the top of the subgrade varies from about 16 to 32 psi (110-220 kN/m2 ) as the dynamic vibratory roller force increases from 4 to 17.taken using equation (F-1) for performing a general assessment of the severity of the aggregate penetration problem.

be much smaller and conventional pavement design theory can be used to avoid problems with permanent deformations. in general.3. With reinforcement the safety can decrease somewhat. . the over-consolidation ratio). large permanent strains rapidly develop under the application of repeated loadings. At one site wellgraded aggregate with about a 1. These results indicate that a suitable safety factor must be used with equation (F-1) to avoid accumulation of excessive permanent deformations. For the conditions existing at the site.(30-38 mm) top size and 5 percent fines was observed during construction to intrude up to a depth of about 1 to 2 in. the threshold stress (a z /c) was found by Bender and Barenberg to increase above this level.5 to 2 for a relatively few number of loadings and an unreinforced aggregate layer. that the soil beneath the load first starts to fail locally at an applied loading of 3. the calculated safety factor for a general bearing capacity type failure varied from about 0..8 to 1. After construction the stress on the subgrade would.Equation (F-1) predicts the required load to cause a general bearing failure under the application of a single load. Yielding of the soil occurs at even lower stresses and is greatly influenced by the initial stress state in the soil (i. By using a light fabric.25 to 1. The safety factor during construction should be a minimum of 1. Bender and Barenberg [F-33] found for non-reinforced aggregate bases if a z /c 1 3.14c. however. (25-50 mm) into a soft subgrade [F-12. Jurgenson [F-21] has shown.F-13].4.5 in. Separation Case Histories Mixing of the subgrade with an aggregate base has been reported at several sites where geosynthetics have not been used.e.

0. conventional static filter criteria was significantly exceeded at these two sites. a limited amount of punching of the aggregate into the subgrade will occur for a safety factor of 2.4 and 2. either conventional Terzaghi filter criteria should be satisfied or a geotextile should be used as a F-13 . For comparison.F-17]. 2. the ratio D15/d85 varied from 17 to 20. The depth of punching should approach the radius of the maximum aggregate size. Separation Design Recommendations The following tentative design criteria are proposed to minimize problems with separation between an aggregate layer and the underlying subgrade and to avoid excessive permanent deformation during construction. If the safety factor with respect to a general bearing capacity failure is greater than 2. intrusion may also occur even if conventional Terzaghi filter criteria are satisfied [F-16. Most problems involving separation will occur where soft to very soft cohesive subgrades are encountered typically having undrained shear strengths less than about 500 psf (24 kN/m 2 ). Problems such as excessive permanent subgrade deformations during construction or aggregate penetration would also occur on firm subgrades under more severe loading conditions. no special precaution is needed with respect to separation or excessive permanent deformations during construction. For a bearing capacity safety factor between about 1. For very open-graded granular bases or subbases. Hence. the Terzaghi filter criteria for steady seepage requires D15/d85 S 5. 1.0. Under severe conditions of loading.At two sites where intrusion occurred.

Consideration should also be given to satisfying filter criteria. Figure F-4 gives the bearing capacity safety factor as a function of construction lift thickness for selected vibratory rollers and undrained subgrade shear strengths. the geotextile will serve primarily as a construction aid. separation could become a problem for subgrades having undrained shear strengths less than . (300 mm). use of a geotextile is recommended regardless of whether filter criteria are satisfied.separator.0 to 2. If the granular filter material satisfies filter criteria. If the safety factor is less than about 1.5 should be provided to avoid excessive deformations. particularly if a very open-graded stone is to be used for drainage applications. The above recommendations are given to avoid contamination of the granular layer due to intrusion and subsequent mixing and also prevent excessive permanent deformations from construction traffic on the unsurfaced aggregate layer. Construction traffic should not be permitted for this condition. Drainage applications where filtration is important are discussed in the next section. This criteria should also avoid permanent deformation problems from compacting the first lift.4. Specific recommendations concerning the selection of a geotextile are given in a later section. If a small to modest amount of construction traffic is to use the initial construction lift. This figure shows for a moderate vibratory roller weight of 8 tons and lift thicknesses of 12 in. 3. then a safety factor of at least 2.

Early laboratory work by Havers and Yoder [F-20] indicate a moderate plasticity clay to be more susceptible to erosion than a high plasticity clay. for the existing soil conditions. in most cases.about 500 psf (24 kN/m 2 ). Table F-2 can be used when reliable estimates of the shear strength based on testing are not available. A saturated subgrade having a source of water. This increase in strength should be considered in estimating the bearing capacity safety factor for long-term traffic loading conditions. A very substantial increase in shear strength of a soft to very soft subgrade will. This subgrade strength corresponds to a standard penetration resistance (SPT-value) of approximately 4 blows/ft. The initial undrained shear strength of the subgrade can be estimated from vane shear tests. occur reasonably rapidly after placement of the pavement structure [F-181. FILTRATION Some general requirements for intrusion of a slurry of subgrade fines into an open-graded aggregate layer can be summarized from the early work of Chamberlin and Yoder [F-19]: 1. Silts.(13 b/m). be even more critical and in general unacceptable. An erodable subgrade material. A base more permeable than the subgrade with large enough pores to allow movement of fines. 3. For preliminary design purposes. or from the results of cone penetrometer tests. 2. undrained triaxial shear tests. Heavy construction traffic on this thickness would. Selection of an actual geosynthetic or aggregate filter to use as a separator is considered later in the section on Filter Selection. fine sands and high plasticity .

) Very Light 1.C.-I 10 Medium 6 Heavy 8 Granular Base (in. linear elastic vertical subgrade stress increased by 12 percent to give good agreement with measured test section subgrade stress. Surface (in. Moduli/Poisson's Ratio: AC.N (blows/ft.) Approximate Undrained Shear Strength. 3.5 aoco-7 .35.2. 2. Granular Base-10.C (psf) A Very Soft Squeezes between fingers 0-1 0-250 Soft Easily molded by fingers 2-4 250-500 Firm Molded by strong pressure of fingers 5-8 500-1000 Stiff Dented by strong pressure of fingers 9-15 1000-1500 Very Stiff Dented slightly by finger pressure 15-30 1500-2000 Hard Dented slightly by pencil point >30 >2000 Table F-3 Vertical Stress on Top of Subgrade for Selected Pavement Sections A.000 psi/v=0.000 psi/ v-0.5 kips/wheel at 100 psi tire pressure. Dual wheel loading of 4.4. Subgrade .200.) Vertical Subgrade Stress (psi) Section 6 3 _ / Notes: 1. F-16 .4000 psi/v 0.5 43 21 Light 3. Analysis - Linear elastic.Table F-2 Preliminary Subgrade Strength Estimation Subgrade Description Field Condition Standard Penetration Resistance.

for at least some soils. Although the work of Chamberlin and Yoder [F-19] was primarily for concrete pavements. Remolding may also play a role in the loss of subgrade strength. similar mechanisms associated with the formation and movement of slurry also occurs for flexible pavements. Due to the application of wheel loadings.F-23. however. soil suction appears to be caused under low stress levels by small gaps which open up upon loading [F-25]. The gaps apparently develop because the geotextile rebounds from the load more rapidly than the underlying soil. The moisture content does not. 4. increases from about the plastic limit to the liquid limit [F-7]. As a result the shear strength of the subgrade in the vicinity of the point contacts becomes quite small. When a geotextile is used. Filtration Mechanisms Repeated wheel load applications cause relatively large stresses to be developed at the points of contact between the aggregate and the subgrade. significantly increase in the open space between aggregates (Figure F-5). in the unloaded state the effective stress between particles of subgrade soil become negligible because of the high F-17 . As a result. relatively large pore pressures may build up in the vicinity of the base-subgrade interface [F22. Hoare [F-7] postulates the increase in moisture content may be due to local shearing and the development of soil suction.clays that undergo deflocculation are also very susceptible to erosion. the moisture content in the vicinity of the projecting aggregate points.F-24]. As loading continues. The applied stress level must be large enough to cause a pore pressure build-up resulting in the upward movement of the soil slurry.

Stress Concentration Slurry Movement 4/3 Aggregate RotationCauses Large Local . Mechanisms of Slurry Formation and Strain in Geosynthetic.: Increase in Water Content Under Stress Concentration (Fabric stained in this area) SLURRY INITIATION HERE Figure F-5. .wor'Strain in Fabric ni= A/ 1..

is eroded by the scouring action of the water which forms a slurry of silt. Some additional movement of material within. the porosity of a nonwoven geotextile is larger than for an aggregate filter. or as the average flow velocity becomes less as the length of flow increases. The subgrade. clay and even very fine sand particles.residual pore water pressures. a considerable difference can exist between geotextiles falling within the same broad classification of woven or nonwoven materials due to different fiber characteristics. or even out of. Nonwoven geotextiles have a relatively open structure with the diameter of the pore channels generally being much larger than the diameter of the fibers. (2) the viscosity of the slurry. The slurry of fines probably initiates in the vicinity where the aggregate tips press against the soil [F-3]. Also. The upward distance which fines are carried depends upon (1) the magnitude of induced pore pressure which acts as the driving force. and (3) the resistance encountered to flow due to both the size and arrangement of pores. Fine particles settle out in the filter or the aggregate layer as the velocity of flow decreases either locally because of obstructions. the base may occur as the moisture and loading conditions change with time [F-19]. In contrast. Also. in its weakened condition. F-19 . These pore pressures in the subgrade result in the flow of water upward into the more permeable aggregate layer. This location of slurry initiation is indicated by staining of geotextiles in the immediate vicinity of where the aggregates contact the fabric. Geotextile Filters Geotextile filters have different inherent structural characteristics compared to aggregate filters. aggregate filters have grain diameters which are greater than the diameter of the pores [F-8].

The challenging part of modifying granular filter criteria for use with fabrics is relating soil retention characteristics on a geotextile with those of a true three-dimensional granular filter. 3./yd 2 g /m 2 ). needle punched geotextile increases. the effective opening size decreases up to a limiting thickness which is also true for an aggregate filter [F-8].Electron microscope pictures showing the internal structure of several non-woven geosynthetics are given in Figure F-6. None of these geosynthetics were considered to fail due to clogging during 10 years of use in edge drains [F-26]. 2. Thin: thickness t<2 mm and geotextile weights up to 9 oz. Thick multi-layer./yd 2 (600 g/m2 ). soil grains which enter the geotextile pores reduce the amount of compression which occurs in a nonwoven. (30 . The approximate order of ranking with respect to clogging from best to worst is from (a) to (d) for similar geotextiles. thin geotextiles do not. which gives support to the above classifcation scheme. needle punched geotextile subjected to loading. needle punched geotextiles. Thick needle punched geotextiles have been found to provide a three-dimensional structure that can approach that of an aggregate filter. Heerten and Whittmann [F-8] recommend classifying geotextiles as follows: 1. needle punched: thickness t>2 mm and geotextile weights up to 18 oz. Thick: single layer. Also. The following review of factors influencing geotextile filtration performance are primarily taken from work involving cyclic type loading. As the thickness of a nonwoven. Thickness. Earlier work by Schober and Teindl [F-6] found wovens and non-wovens less than 1 mm in thickness to perform different than non-wovens greater than 2 mm.

(b) Nonwoven.5 oz/yd 2 . 60 mil. Figure F-6. 15 mil. (d) Spun-Bonded. Needle 5.3 oz/yd 2 . 30 mil. (c) Nonwoven 4. F-21 .(a) Nonwoven. Electron Microscope Pictures of Selected Geotextiles: Plan and Edge Views (84x). 75 mil. Heat Bonded. Needle 4.5 oz/yd 2 .

The Soil . does not go back to its original value upon flow reversal. Flushing was found by Saxena and Hsu [F-25] to be more effective for heavier. partial flushing of fines from a geotextile is apparently possible under conditions of reversing flow. Load Repetitions.As the thickness of the geotextile increases. however. however. nonwoven geotextiles. The quantity of fines migrating upward through a geotextile filter is directly related to the log of the number of load applications [F-7. Self-Cleaning Action. Hence. A layer of fines forming a cake on the downstream side of the geotextile has also been observed. The load on the aggregates in contact with the geotextile can result in a significant amount of stretching of the fabric and a temporary increase in pore diameter. and the geotextile is not self-cleaning.F-25] as illustrated in Figure F-7. The permeability. Laboratory tests have shown a change in the direction of flow through a geotextile can cause an increase in its permeability [F25. Whether self-cleansing can actually occur in the field has not been demonstrated. The fines which do pass through the geotextile may be deposited on the upstream side of the fabric in a thin layer that can significantly reduce effective permeability. If.F-28]. the effective opening size decreases and fines in suspension have a harder time passing through because of the three-dimensional structure [F-7. which allows more fines to pass through.F-25. clogging can occur. When open-graded granular materials are located above the geotextile. the fines passing through would probably be pumped into the voids of the stone resulting in stone contamination.F-27]. the geotextile has pores which are too small in diameter or the porosity is too small.

Variation of Geosynthetic Contamination with Number of Load Repetitions (After Saxena and Hsu. 0 95 (After Bell.0. F-23 .5 NUMBER OF LOAD REPETITIONS.1 mm AOS ■170 CJ C6 0 16 0Z/YD2 mm AOS∎ 100 m. 2000 z 0 Z. et al..1000 0 C. F-25). Variation of Geosynthetic Contamination with Geosynthetic Apparent Opening Size. 4000 (Ci E woo / 0 / • • 2000 • ° 1000 • •• • • 0 0 100 200 300 400 500 APPARENT OPENING SIZE. Ref.) 0 25 O 4. logN Figure F-7.5 3. 095 (pan) Figure F-8.% E ti HOARE (1982) 1:71 3000 - - SAXENA & HSU (NONWOVEN) 0 8 OZPD2 t.5 5. Ref. F-10).

3. SCV is the weight of soil per unit area passing through the geotextile [F-7]. The apparent opening size (AOS) of a geotextile is defined as the minimum uniform. The uniformity coefficient of the soil being protected has an important influence on the filter criteria. 4. Apparent Opening Size. Fabric pore size. Also. The Apparent Opening Size (AOS) quantifies at least approximately the effective pore opening size of a geosynthetic. pore structure and filtration capacity are not accurately defined by AOS. spherical particle size of a uniform shape that allows 5 percent or less of the particles to pass through the geotextile [F-9]. 2.Contamination Value (SCV) quantifies soil loss through a geotextile. the AOS of woven monofilaments and nonwoven geotextiles should not in general be compared since they will not have the same filtration efficiency [F-29]. Some general findings by Carroll [F-29] involving AOS as related to geotextile filtration are as follows: 1. For a given weight. The AOS measures the maximum "straight through" openings in a woven geotextile. geotextiles having a small fiber size. allow less material to be washed through [F-8]. The apparent opening size (AOS) of the geotextile cannot be used alone to directly compare the retention ability of a nonwoven and woven geotextile. and as a result a smaller effective opening. AOS values can be related to the retention ability of geotextiles provided proper consideration is given to the other significant factors. .

As pointed out by Raymond [F-11]. this alternate definition is more closely related to classical filter criteria that limits the amount of soil which can enter the filter. As the applied stress level on the geosynthetic increases. In the laboratory tests performed by Bell. Soil contamination of geotextiles removed from beneath railroad tracks has been reported by Raymond [F-11]. as estimated using the method of Schober and Teindl [F-6]. The sand layer also had the smallest apparent opening size. soil contamination is the percent of soil trapped within the geotextile compared to the uncontaminated dry geotextile weight. Since the applied F-25 . Stress Level. et al. Undoubtedly the scatter in data in Figure F-9 is at least partly because soil contamination is not only related to AOS but also to a number of other factors as previously discussed.The quantity of fines trapped by the filter layer when subject to cyclic loading generally increases with increasing apparent opening size (AOS) of the filtering media (expressed in units of length and not sieve size) (Figure F-8). [F10]. so does the quantity of fines migrating through the geotextile (Figure F-10) and the amount of contamination. This extensive field study also indicates increasing soil contamination of the geotextile occurs with increasing apparent opening size (AOS) as shown in Figure F-9. As defined in this figure. Data obtained from field studies (Figure F-11) show that the level of contamination rapidly decreases below a railroad track structure with increasing depth [F-10]. Figure F-9 shows results for an alternate definition of AOS based on 95 percent of the uniform particles being retained on the surface of the geotextile [F-30]. the least amount of contamination was observed when a thin sand layer was employed compared to the geotextiles tested.

SCV (gm/m2) 20 16 20 24 28 32 SUBGRADE MOISTURE CONTENT. F-31). AVERAGESTRESS LEVEL a(KN/m2) 80 3500 ---3000 2500 1500 SOIL CONTAMINATION VALUE CONTOURS.S. F-26 . SIEVE STANDARD SIEVE NUMBER Soil Content Based on 5% Passing (AOS Definition) Based on 95% Retained Figure F-9.800 600 SOIL CONTENT PERCENT OF ORIGINAL WEIGHT 400 200 0 400 270 200 140 100 70 60 40 30 U. Below Railroad Ties with Geosynthetic Opening Size (After Raymond. CPERCENT) Figure F-10. Ref. Variation of Geosynthetic Contamination Approximately 8 in. Ref. Variation of Geosynthetic Contamination with Stress Level and Subgrade Moisture (After Glynn & Cochrane. F-11).

vertical stress also decreases with increasing depth. 69. 41. The curve relating variation of soil content with depth (Figure F-11) is similar in general shape to a typical vertical stress distribution curve. geotextiles are generally placed at a depth of about 8 to 12 in. typical very light. To approximately translate the extensive findings of Raymond [F-I0] for geotextiles placed below railroad track installations to pavements. (200-300 mm) beneath the tie which corresponds to a vertical stress level on the order of 14 psi (96 kN/m2 ). 21 kN/m2 ). a comparison was made of the vertical stress developed beneath a heavily loaded railroad track with the stress developed at the top of the subgrade for typical pavement sections. contamination of a geotextile in the field is indeed dependent upon stress level. Loss of integrity of the geotextile due to abrasion and also breakdown of the aggregate may also play an important role in aggregate contamination. and the tires are inflated to 100 psi (0. 6 and 3 psi (138. vertical stresses from pavement type loadings spread out relatively quickly because of the small diameter of the loaded area. light. medium and heavy pavement sections (Table F-3) have maximum vertical stresses at the base-subgrade interface on the order of 21. 10. A heavy train loading causes large vertical stresses which spread out slowly with depth. For comparison. Let the critical railroad loading be simulated by a fully loaded cement hopper car. respectively. In contrast. For railroad track rehabilitation. medium and heavy highway pavement section. . Assume 4. Figure F-I2 shows the approximate equivalent depths below the railroad cross-ties that corresponds to the vertical stress at the top of the subgrade for a typical light.5 kip (20 kN) dual wheel loads are applied to the surface of the pavement.7 MN/m2 ).

F-29]. bound for establishing design criteria for pavement infiltration applications. The results obtained from constant gradient tests. A cyclic loading is then applied to the top of the specimen through a rigid loading platen.F-32] or a plexiglass cylinder [F-33]. In contrast to other tests.The practical implications of these findings are that (1) the railroad type loading is considerably more severe compared to most structural sections used for pavements. among other things. In contrast. The test is performed in a triaxial cell at a realistic confining pressure. serve as an upper. Most frequently dynamic testing to simulate pavement conditions has been carried out in cylindrically shaped. and (2) a highway type pavement should exhibit a wide variation in performance with respect to filtration depending.F-31. with the filter layer and base material above. Laboratory Testing Methods Laboratory studies to observe the migration of fines through both granular filter layers and geotextile filters have most commonly employed a constant gradient test which simulates steady state. than for a typical railroad ballast installation. unidirectional seepage conditions [F-7. An improved test [F-28] has been developed by Dempsey and Janssen for evaluating the relative effectiveness of different geotextiles (Figure F13). The subgrade soil is generally placed in the bottom of the mold. which do not use a cyclic load. upon the thickness and strength of the structural section. rigid cells which may consist of either a steel mold [F-3. the subgrade soil is placed on top of the geotextile filter. Very thin pavement sections are probably subjected to an even more severe vertical stress. Water is continuously passed downward through the F-29 . a heavy structural pavement section would be subjected to a much less severe stress condition. and hence more severe infiltration condition. possibly unsafe.

F-28). T=8 IN INPLACE COSTS T=SAND FILTER THICKNESS T=6 IN T=4 IN Tas3 IN IYPICAL RANGE IN FABRIC COST EXAMPLE (T=4 IN) 4 a 12 16 FILTER SAND COST (DOLLARS/TON) Figure F-14.Cyclic Load Triaxial Apparatus for Performing Filtration Tests (Adapted from Janssen. Economic Comparison of Sand and Geosynthetic Filters for Varying Sand Filter Thickness. Ref.Dual Water Reservoirs Air Pressure Cyclic Load Air Flushing Valve Mercury Manometer to Measure Pressure Difference Water Flow Confining Pressure Soil Specimen Confining Pressure Ceotextile Marbles Lucite Plates (indented to trap sediment) Cair A - Water Flow Filter Paper Overflow Collection Reservoir Figure F-13. .

specimen at a constant hydraulic gradient as a repeated loading is applied. and relief wells. Intended applications for the Task Force 25 criteria are as follows: edge of pavement drains. interceptor drains. They shall be formed into a network such that the filaments or yarns retain dimensional stability relative to each other. Selected Practices Task Force 25 Criteria. Over about the last five years Task Force 25 has developed comprehensive specification guidelines for drainage geotextiles. The current version of the Task Force 25 criteria requires that: "Fibers used in the manufacture of geotextile. and the threads used in joining geotextiles by sewing. F-31 . AGC. one million load repetitions are applied. universities and the geotextile industry. To evaluate long-term performance. recharge basins. wall drains. including selvedges". The quantity of fines washed through the geotextile is measured. shall consist of long chain synthetic polymers composed of at least 85% by weight polyolefins. As a result this task force has a wide range of experience and backgrounds. He also shows that three dimensional pavement tests are more appropriate than the conventional one-dimensional test. Dawson [F-51 has pointed out the important need for performing tests at realistic vertical stress levels comparable to those existing in pavements. polyesters. ARTBA. or polyamides. as well as the permeability of the geotextile as a function of load repetitions. Task Force 25 has representatives from a number of organizations including AASHTO.

this criteria requires that the AOS generally be between the No. and greater than 10 percent for soils having less than 5 percent fines. Both woven and non-woven geotextiles are allowed. To maintain separation a more densely graded Class 2A stone separation layer is placed beneath the open graded drainage course. The Pennsylvania DOT uses as a standard design an open graded subbase (OGS) to act as a blanket drain (Table F-6). (300 mm) subbase is used the two layers are each 6 in. (150 mm) thick. 70 and No. wovens must have a percent open area greater than 4 percent for soils having 5 to 85 percent passing the number 200 sieve. if a 12 in. For similar reasons. (75 mm) in thickness.S. To permit adequate drainage and to resist clogging. the open graded aggregate drainage layer is placed directly on the geotextile. For unidirectional. An approved geotextile may be substituted for the separation layer. and is equal in thickness to the full depth of the subbase. The geotextile separator used typically has a weight of about . 120 U. non-turbulent conditions of flow. If a 6 in (150 mm) thick subbase is used. Standard Sieve." For fine grained soils having 50 or more percent passing the number 200 sieve.Task Force 25 geotextile criteria are summarized in Table F-4. the Corps of Engineers recommends the criteria show in Table F-5. If a geotextile is used. Corps of Engineers Filter Criteria. Pennsylvania DOT Filtration/Separation Practices. The Corps [F-34] cautions about using filter materials in inaccessible areas indicating that their use "must be considered carefully. non-woven geotextiles must have a permability greater than 5 times that of the soil. the two layers are each 3 in.

S. All other conditions are said to be unprotected. Examples of each condition are: Protected: highway edge drains. 4 Sieve in selecting the fabric. PERMEABILITY Critical/Severe Applications* Normal Applications k(fabric) > k (soil) k(fabric) > 10k (soil) *'Woven monofilament fabrics only. rock. . PIPING RESISTANCE (soil retention . When the protected soil contains particles from 1 inch size to those passing the U. interceptor drains on cut slopes.S. No. should be specified. or polyamides. blanket drains. CHEMICAL COMPOSITION REQUIREMENTS/CONSIDERATIONS A. Puncture Strength (ASTM-D-751-68) 2 3 Burst Strength (ASTM D-751-68) Trapezoid Test (ASTM D-1117) (Any direction) 290 psi 50 lbs. smooth stable trenches < 10 feet in depth.all applicationa) A. etc.S. IV. use only the gradation of soil passing the U.Table F-4 (1) Recommended Minimum Engineering Fabric Selection Criteria in Drainage and Filtration Applications -AASHTO-AGG-ARTBA Task Force 25 (After Christopher and Holtz. Ref. 1 All numerical values represent minimum average roll values (i. Unprotected: stabilization trenches.e. These fabrics shall resist deterioration from ultraviolet exposure. 80 lbs. percent open area > 4. In trenches. polyesters. 130 psi 25 lbs. 3 Diaphram Test Method 4 Fabric is said to be protected when used in drainage trenches or beneath/ behind concrete (Portland or asphalt cement) slabs. II. Note: these values are normally 20% less than manufacturers typically reported values. Soils with 50% or less particles by weight passing U. PHYSICAL PROPERTY REQUIREMENTS (all fabrics) Fabric Unprotected Fabric Protected` Grab Strength (ASTM D-1682) (Minimum in either principal direction) 180 lbs. F-9) I. (fabric) 2 30 sieve Note: 1. 2. No.. No. 25 lbs. any roll in a lot should meet or exceed the minimum values in the table). B. 200 Sieve. Fibers used in the manufacture of civil engineering fabrics shall consist of long chain synthetic polymers. Soils with more than 50% particles by weight passing U. concrete. Steel ball replaced with a 5/16 inch diameter solid steel cylinder with hemispherical tip centered within the ring clamp. III. 2 Tension Testing Machines with Ring Clamp. (fabric) > 50 sieve EOS No. 200 Sieve: B. No. 80 lbs. 200 Sieve: EOS No. in which the aggregate is extra sharp additional puncture resistance may be necessary. rocky or caving trenches or smooth stable trenches > 10 feet in depth.S. Whenever possible. < 100 sieve. composed of at least 85% by weight of polyolephins. fabric with the lowest possible EOS No. The engineering fabric shall be exposed to ultraviolet radiation (sunlight) for no more than 30 days total in the period of time following manufacture until the fabric is covered with soil.0 and EOS No.

(4) k c. 4 sieve in selecting the EOS of the geotextile. 120 U.212 mm (No. 70 U.2. (2) These protected soils may have a large permeability and thus the POA or kG may be a critical design factor. S.S. S. Standard Sieve) (1) When the protected soil contains appreciable quantities of material retained on the No.125 mm (No. 200 Sieve) (2) Piping (1) Permeabilit y Woven ' Non-Woven (4) (3) EOS(mm) < 085 (mm) POA > 10% kG > Sk s 5% to 50% EOS(mm) < D85 (mm) POA > 4% kG > Sk s 50% to 85% (a)EOS(mm) < 0 85 (mm) POA > 4% k G > Sk s Less than St (2) (b)Upper Limit on EOS is EOS (mm) < . 4 sieve use only the soil passing the No. Army Corps of Engineers Geosynthetic Filter Criteria (Ref. (3) D 85 is the grain size in millimeters for which 85 percent of the sample by weight has smaller grains.F-34) Protected Soil (Percent Passing N. Standard Sieve) >85% (a)EOS(mm) < 085 (mm) kG >— SK S (b)Lower Limit on EOS is EOS (mm) > . is the permeability of the non-woven geotextile and k s is the permeability of the protected soil.Table F-5 U. .

Angular.Table F-6 Aggregate Gradations Used by Pennsylvania DOT For Open-Graded Drainage Layer (OGS) and Filter Layer (2A) \ DRAINAGE LAYER (OGS) SEPARATION LAYER AYER (2A) AASHTO 2 New Proposal(1) 100 100 100 Old 3/4 52-100 52-100 52-100 3/8 36-70 36-65 36-65 #4 24-50 20-40 8-40 #8 16-38 - - #16 30-70 3-10 0-12 #30 - 0-5 0-8 #50 - 0-2 - <10 0-2 <5 #200 Note: 1.3 1/2-2 in. Rounded gravels can be given a separation number one less than indicated. if desired. Well-graded N.-1 SEPARATION NUMBER( 1 ). Top Size >5% Fines N.. N 2-4 in. Top Size Aggr.4 < SF < 2 1. Top Size Aggr. Uniform (No Fines) 1/2-4 in.0 SF < 1.4 < SF < 1.0 Low Moderate Severe Very Severe 3.4 - .4 2 1 - 4 3 2 1 - 3. . . 1-5T Fines. 1. Uniform (no fines N.1) 1-2 in. Tests indicate the proposed gradation should have a permeability of about 200 to 400 ft/day.. Top Size Angular. Table F-7 Separation Number and Severity Classification Based on Separation/Survivability BEARING CAPACITY SAFETY FACTOR GEOTEXTILE SEVERITY CLASSIFICATION .4 N. Angular.2 1..

grab tensile test elongation 1 30 percent. and have a uniformity coefficient C u = D60/D10 1 4. puncture > 110 lbs (0. California DOT. 70 U. The California DOT allows the use of geotextiles below open graded blanket drains for pavements and also for edge drains. although the open- .4 kN).S./yd 2 (95 gm/m2 ).3 kN). The open graded base must be well graded.3 kN). In addition. and an abrasion resistance 1 40 lbs (0. These geotextile material requirements are in general much less stringent than those used by the Pennsylvania DOT. These studies indicated good performance can be achieved by placing an open-graded aggregate base over a sand filter.5 kN). (0. et al. Barenberg. In one instance. The open graded base is placed using a spreader to minimize segregation. and the toughness (percent grab elongation times the grab tensile strength) 1 4000 lbs (18 kN). These studies involved wetting the pavement sections and observing their performance in a circular test track. The subgrade used was a low plasticity silty clay. the open graded base is required to have a minimum of 75 percent crushed particles with at least two faces resulting from fracture. New Jersey/University of Illinois. grab tensile strength 1 270 lbs (0.F-361 have performed a comprehensive study of open graded aggregate and bituminous stabilized drainage layers. Sieve. [F-35.F-17. grab elongation 1 15 percent. It also has the additional mechanical properties: AOS smaller than the No. dense-graded aggregate subbase or lime-flyash treated base. They require for blanket drains a nonwoven geotextile having a minimum weight of 4 oz.16 oz/yd 2 (380 gm/m 2 ). To exhibit some stability during construction. the grab tensile strength must be 1 100 lbs. trapezoidal tear strength > 75 lbs (0.3 kN).

The extensive work of Raymond [F-11] was for geotextiles placed at a shallow depth (typical about 8 to 12 in. Harsh Railroad Track Environment. nonwoven geotextiles were found by Raymond to perform better than thin heat bonded geotextiles which behaved similarly to non-wovens. Also.. Lime modifications of the subgrade was also found to give relatively good performance. open-graded drainage layer placed over a dense graded aggregate filter [F-37]. Abrasion of thick spun bonded geotextiles caused them not to perform properly either as a separator or as F-37 . Well needle punched. 200-300 mm) below a railroad track structure. resin treated.graded drainage layer/sand filter used met conventional static filter criteria. uniformly graded angular aggregate above the geotextile.5 to 0. particularly with an open-graded base having a finer gradation. these nonwovens did better than spun bonded geotextiles having little needling. This condition constitutes a very harsh environment including high cyclic stresses and the use of large. about 0. A significant amount of intrusion of subgrade soil also occurred into an open-graded control section which was placed directly on the subgrade. (12-19 mm) of intrusion of sand occurred into the open-graded base. the New Jersey DOT now uses as standard practice a non-stabilized. but rutted more than the nonstabilized drainage material.75 in. The findings of Raymond translates to a very severe condition for the problem of filtration below a pavement including a thin pavement section. The drainage layer/filter interface is designed to meet conventional Terzaghi type static filter criteria. As a result of this study. An open-graded bituminous stabilized layer was found to be an effective drainage layer. Stone penetration into the lime modified subgrade was approximately equal to the diameter of the drainage layer stone.

and preferably more. non-woven geotextiles having a low AOS less than 55 pun (U. Sieve No.S. which provides important lateral drainage. A depth of 12 in.F-38]: . Raymond also found the best performing geotextile to be multi- layered.F10. needle bonded geotextiles with loose filament crossings have a sufficiently high elongation to withstand heavy railroad loadings without puncturing.5. Raymond [F-11] recommends that at a depth below a railway tie of 12 in.a filter. non-steady flow conditions existing beneath a railway track.F-37.F-11. 270 sieve size) were found to provide the best resistance to fouling and clogging. Approximately extrapolating Raymond's work based on vertical stress indicates for structural numbers greater than about 4 to 4. having large tex fibers on the inside and low tex fibers on the outside. No.75 based on vertical stress considerations (Figure 1-12). For the reversible. for continuous welded rail. Wehr [F-16] concluded that only non-woven. heavy.F-29. (300 mm) in a track structure corresponds approximately to a geosynthetic placed at the subgrade of a pavement having an AASHTO structural number of about 2. (300 mm) a needle punched geotextile should have a weight of at least 20 oz. of about 100 to 140 should result in roughly the same level of contamination and clogging when a large uniformly graded aggregate is placed directly above.S./yd 2 (480 gm/m 2 ). FILTER SELECTION INTRODUCTION Factors of particular significance in the use of geotextiles for filtration purposes below a pavement can be summarized as follows [F-6. a geosynthetic having a U. Use of a low AOS was also found to insure a large inplane permeability.

8. Properly designed sand aggregate filters are superior to geotextiles. and low plasticity clays should be most susceptible to erosion and filtration problems. non-woven geotextiles perform better than wovens.F-31]. and greater reductions occur [F-5. 7. for example. Low cohesion silts. uniformly graded aggregates are placed directly above. 3.F-17. Non-Wovens. Most studies conclude that needle punched. needle punched non-wovens (t 2 2 mm). Granular filters are thicker than geosynthetics and hence have more three dimensional structural effect. 4. Aggregate Base/Subbase. large strains are locally induced in a geosynthetic when big. Aggregate Filters. 6. constitutes a particularly severe condition when placed over a subgrade. Full scale field tests by Wehr [F-16] indicate for low plasticity clays and highly compressible silts. existing moisture conditions and undrained shear strength are all important. Geosynthetic Thickness.F-8. 5. angular uniform drainage layer. The top size. A large. The type subgrade. dispersive clays. 9. angularity and uniformity of the aggregate placed directly over the filter all affect performance. particularly under severe conditions of erosion below the pavement [F-3. Thin (t < 1 mm) non-woven geotextiles do not perform as well as thicker. for example. The strength of the pavement section placed over the filter/separator determines the applied stresses and resulting pore pressures generated in the subgrade. 2. fiber structure and also internal pore size are all important. In providing filtration protection particularly for silts and clays some contamination and filter clogging is likely to occur. Pavement Section Strength. Wehr [F-16]. The apparent opening size (AOS) is at least approximately related to the level of base contamination and clogging of the geotextile. that primarily sand and silt erodes into the geotextile. Reductions in permeability of 1/2 to 1/5 are common. Fiber size.F-11.1.F-39].F-26. found strains up to 53 percent were locally developed due to the spreading action of the aggregate when subjected to railroad loads. For conditions of a very soft to soft subgrade. GEOTEXTILE Where possible cyclic laboratory filtration tests should be performed as previously described to evaluate the filtering/clogging potential of F-39 . Clogging. Strain. Subgrade.F-11. Apparent Opening Size (AOS).

Also for a Separation Number of 4.7. Enter Table F-8 with the appropriate geotextile SEVERITY CLASSIFICATION and read off the required minimum geotextile properties. Select from the upper part of Table F-7 the appropriate column which the Separation Number N falls in based on the bearing capacity of the subgrade. A layer to maintain a clean interface (separation layer) is not required if the bearing capacity safety factor is greater than 2.0. The steps for selection of a geosynthetic for separation and survivability are as follows: 1. larger opening size). the requirements on apparent opening size (AOS) can be relaxed to permit the use of geotextiles with U. Where filtration is not of great concern.e. Figure F-5 provides a simple method for estimating subgrade bearing capacity. an intermediate layer is probably not required if the bearing capacity safety factor is greater than 1. Read the SEVERITY CLASSIFICATION from the top of the appropriate column. and for a SEPARATION NUMBER of 3 or more it is probably not required if the safety factor is greater than about 1.4. 70 (i. gradation and angularity of the aggregate to be placed above the filter. Separation. Survivability is defined as the ability of the geotextile to maintain its integrity by resisting abrasion and other similar mechanical forces during and after construction.. Sieve sizes smaller than the No. The filter criteria given in Table F-1 can serve as a preliminary guide in selecting suitable filters for further evaluation. 3. 2. Estimate from the bottom of Table F-7 the SEPARATION NUMBER N based on the size.geosynthetic or aggregate filters to be used in specific applications.S. A preliminary classification method is presented for selecting a geosynthetic based on the separation/survivability and filtration functions for use as drainage blankets beneath pavements. .

70% s treng t h re ta ine dfor a ll c lasses (h) p( > OV Oh OS OVI App a re n t Op e n ing Size (AOS) U. Li Cl CI o CO . — STM D.4 35 5 A Pe rmea bi lity (c m /sec ) As T10. u s e t he filter cr iter ia g iven in Ta ble F. 0 . 200 s ieve 1 Oh1. S.1...1 4.0 U 4.o do 0 o r •■•• ra.1 C 4.g Ultra v io le t Deg ra da t ion at 150hrs.1 682 I OSI 8. re X 4l C 0 0 coarse g r a in e d so ils. 4+ . ■-■ • .4491.I 682 1Seve re 00 Z. S.65 1 kg > 10 k .2 0 0 0 14 4. I co to 4' . ••1' 4.1 • 0 4.9 Gra bTens i le E long a t ion ( 1) A STM D.1 1-44•1 0 0 O. Less t ha n U.1 C > m CO 0.1 c at 0 01 0 144 1C4'.c 3 c 0 4) > 4) 0 C 0 .7 m 11) 0 .) M tC •. 70 s ie ve means a s im ilar op en ing s ize a n dhe nce large r s ie ve nu mber.0 11 for severe and severe ap p l ica t ions.9I 007.0 w > Qf 61 V C • c 0 0 0 0. Sie ve S ize (3 ) So ils w it h mo re t ha n 50Z pa ss ing Nu.001 > ( V) St 09 SE 0/Z 09 101 : Abras ion Re s is ta nce ( lb) ASTMD.1 X 0.04 0 0 ■-•1 0 CO .1for revers ib le flow condit ions.0 .001 > ( V) OS SL 06 OS E OS I Modera te (2) -Z1 S LZ (t)91 ILow (2) I 1 GEOT EXTILE SEVERITY CLASSIFICATION SS 08 OSI 00S 09 00 1 ZE.11 75 an d D.1 C 0 0 0 011 0.0Z1 > ( V) (/) azan a gki wi M si •••■ CO .

[F-3] found that large 4. however. The suggested steps for selection of a geosynthetic for filtration considerations are as follows: 1. The choice therefore becomes primarily a matter of economics. Bell. An intermediate granular layer between the subgrade and base or subbase having a minimum thickness of 3 to 4 in.e. (114 mm) diameter aggregates can punch through a thin.F-17]. Most geosynthetics when used as a separator will reduce stone penetration and plastic flow [F31].Both sand filter layers and geotextiles can effectively maintain a clean separation between an open-graded aggregate layer and the subgrade. excessive permanent subgrade deformations may occur during construction of the aggregate base as a result of loads applied by construction traffic. Estimate the pavement structural strength category from Table F-9 based on its AASHTO structural number. washing of fines from the subgrade into the base or subbase) should also satisfy the previously given requirements for separation/ survivability.. compressible geotextiles than for thinner ones.F-4. (75-100 mm) is recommended. uncompacted 2 in. More care is perhaps required for the design of an intermediate aggregate layer to maintain separation than is necessary for the successful use of a geotextile. been found by Glynn and Cochran [F-31] to be considerably greater for thicker. (50 mm) sand layer into a soft cohesive subgrade. . This potentially important aspect must be considered separately as discussed in the separation section. Finally. The geotextile selected based on filtration considerations (i. A wide range of both nonwoven and woven geotextiles have been found to work well as just separators [F-3. Filtration. et al.5 in. The reduction in penetration has.F-13.F-16.

Micaceous Silty Sands and Sandy Silts 20 Well-graded cohesionless gravel-sand-silt mixtures (PI<6)/ Medium plasticity. very uniform fine cohesionless sands (Pl<6). Clay tinder may be present. T-31 for indications of susceptibility to erosion.25-4.25 18 13 7 4 Medium 3.5 Medium 3.5 Susceptibility to Erosion Rarely Wet Description (1) / Partial Index (7) Dispersive clays. F-43 00 .25 14-9. Low PI clays 12 Nondisporsive clays of high plasticity (pi>25).25-4.5 Approximate Vertical Subgrade Stress (psi) >14 Light 2.5-5 Heavy <5 >4. F-20.5-3. Y-15.5-3.Table F-9 Pavement Structural Strength Categories Based on Vertical Stress at Top of Subgrade Category Approximate Structural Number (SN) Very Light <2.5 9.5 (3) Frequently Wet. of Year Periodically Wet 0) (6) 25 17 9 5 Light 2. Wet More Than 3 mo.5 Table F-10 Partial Filtration Severity Indexes Subgrade Moisture Condition: Partial Index Pavement Structure Description (1) Very Light Wet Entire Year SW (2) <2. Gravels 3 Mote: 1. See for example References F-2. Coarse sands.3 13 9 6 3 10 7 4 2 Heavy >4.

The proposed procedures for considering separation. gradation and angularity of the aggregate. Stress level in turn is determined by the strength of the structural section placed above the subgrade. Add the appropriate Partial Filtration Severity Indexes given in Table F-10 given for the appropriate subgrade moisture condition and pavement structural strength (Add one number from one of columns (3) through (6) to the partial index (one number) given in column (8) corresponding to the subgrade soil present). and (2) subgrade strength and applied stress level at the subgrade. and subgrade moisture content. In making a final geotextile selection good judgment and experience should always be taken into consideration. The addition of these two numbers gives the FILTRATION SEVERITY INDEX. filtration and permanent subgrade deformations during construction are intended to illustrate some of the fundamental parameters of great importance in selecting geotextiles for separation/filtration applications. and determine the required filtration characteristics of the geotextile. FILTRATION INDEX Very Severe > 36 Severe 28-35 Moderate 18-27 Low S 17 Enter Table F-8 (third row from bottom) with the appropriate FILTRATION SEVERITY CLASSIFICATION . In separation problems important variables include (1) size. 3. it has been shown earlier that filtration and contamination levels are significantly influenced by the magnitude of the subgrade stress. F-44 . For example.2. Estimate the filtration SEVERITY CLASSIFICATION as follows: FILTRATION SEVERITY CLASSIFICATION 4. It would seem illogical not to consider these important parameters in selecting a geotextile for use beneath a pavement. number of load repetitions.

N. 355-365.The primary purpose of presenting the proposed procedure for geotextile selection was. April. "Water: Key Cause of Pavement Failure".R... "Geotextile Use as a Separation Mechanism". The interaction between some variables such as stress level and number of load repetitions was through necessity estimated. 1974. the previously presented effects of stress level. when good judgement and experience is applied. 9. Proceedings. 77 p. offers a reasonable approach to semi-rationally select a suitable geotextile. 1984. F-4 Dawson. 1981. and Gregory..I. A.F.. University of Nottingham. and Godfrey. No.R. number of load repetitions (both of which are related to structural number) and moisture content were used in developing the semi-rational procedures presented here. For example. 7882. pp.. and Brown. Figure F-14 can be employed to quickly determine whether a geosynthetic is cheaper to use as a filter or separator than a sand filter layer. "Geotextiles in Road Foundations". Nevertheless. 175 pp. Vol. S. K. to encourage engineers to begin thinking in terms of the variables that are known to be significant. Civil Engineering. "Clay Contamination in Crushed Rock Highway Subbases".A. pp. 44. J. hopefully. Sept. Oregon State University. Civil Engineering. McCullough. L. TRR84-4.... NSW. . Australia. A. The procedures presented were developed during this study using presently available data. September. Research Report to ICI Fibres Geotextiles Group. H. 1984. it is felt that the proposed procedure. Session Conference on Engineering Materials. F-2 Jorenby. APPENDIX F REFERENCES F-1 Cedergren.M. F-3 Bell. Economics. B.

Third International Conference on Geotextiles. MacMillan. p. CIRCIA. Second International Conference on Geotextiles. Austria. and Tayabji.D. Volume 21.F.. "Geotextile Engineering Manual"... Engr.. G. p. University of Illinois. Gizienski. No. R.J. 134-135. Austria. Vienna. 1963. 1985.. and Holtz. E.. "Earth and Earth-Rock Dams". F-13 Potter. E. McCullough. F-15 Sherard. Las Vegas. 967-971. pp. London. Potter. "Research on Geotextiles for Heavy Haul Railroads". 1979 (4th Edition). J. Transport and Road Research Laboratory. W. and McAvoy. 1984. Vol.. Proceedings.J. F-14 Sowers. Proceedings.P.. F-10 Bell. 435-440. "The Role of Geotextiles in Controlling Subbase Contamination". Geotechnique. 1986. pp.S. A. H. J. "Filtration Properties of Geotextile and Mineral Fillers Related to River and Canal Bank Protection". INTRODUCTORY SOIL MECHANICS AND FOUNDATIONS. Federal Highway Administration. "Separation Function of Non-Woven Geotextiles in Railway Construction". 75 p. Project Record 245. Pavement. Hants. 1. F-8 Heerten.R. G.R. England. L. 1978. A. International Conference on Design Parameters in Geotechnical Engineering.F-5 Dawson. .W. "Report on the Construction and Performance of a Full-Scale Experimental Road at Sandleheath.. G.. Brighton... F-9 Christopher. L. 1986. Geotextiles and Geomembranes. New York.. pp. Canadian Geotechnical Journal. 1979... Clevenger. 1984.F. Vol. R. W. Third International Conference on Geotextiles. "Filter Criteria for Geotextiles". Discussion of "An Experimental Comparison of the Filtration Characteristics of Construction Fabrics Under Dynamic Loading".. 1974. F-17 Barenberg. F-6 Schober.. "An Experimental Investigation of Subbase Protection Using Geotextiles". New York. Transp. H..M. 1982. and Currer. Snaith. A.. 1981.. 259-276. TRRL Report 996.. 1985.. 4763. F-12 Ruddock. "Evaluation of Typical Pavement Drainage Systems Using Open-Graded Bituminous Aggregate Mixture Drainage Layers". D. Woodward. Vienna.H. UILU-ENG-74-2009. B. and Wittmann.C...D.L. and Teindl. 2. S. F-16 Wehr. M. F-11 Raymond. pp.J. S. Series 10. 34.F.F.A.. John Wiley. "The Effect of a Fabric Membrane on the Structural Behavior of a Granular Road". F-7 Hoare. J. Proceedings. E. 593-598.L.

F-27 Dawson. pp. "Effects of Repeated Loading on Gravel and Crushed Stone Base Course Materials Used in AASHO Road Test". D. E.J. Boston. 1983. 4.. pp. 1985.J. pp. "Permeability of Geotextile-Included Railroad Bed Under Repeated Load". Transportation Research Board. Highway Research Board. Proceedings. 657-660. 32-37. 1-12. and Yoder.J. pp. "A Study of Interactions of Selected Combinations of Subgrade and Base Course Subjected to Repeated Loading". 37. and Stiffens. 148-183. on Soil Mechanics and Foundation Engineering. R. and Yoder. "A Field Study of the Performance of a Tensar Reinforced Haul Road". 1982. 2. Vol.J. F-20 Havers. Mass. 9th European Conf. 1983. F-19 Chamberlain.. and Hsu. F-25 Saxena.. Research Record 39. F-24 Haynes. E. D.. and Prendergast... F-23 Barber. E.. Transportation Research Record 916. 1986. . 693721. J. Transportation Research Board. Vol. 31-51. Geotextiles and Geomembranes. pp.. and Turgeon. W. Transportation Research Board. Transportation Research Record 849. A.. E. Contributions to Soil Mechanics 1925-194. G.F. pp. "Dynamic Test to Predict Field Behavior of Filter Fabrics Used in Pavement Subdrains". and Brown. School of Civil Engineering. 1987. Transportation Research Board. G. F-28 Janssen.. 36. Highway Research Board.A. 89-93. Final Report.. Washington. Transportation Research Record 916.T.. "Pore Pressures in Base Courses". pp. J. "Laboratory Investigation and Field Studies of Channeling and Pumping".G.P.. S. Boston Society of Civil Engineers. 173 p... Vol.E. "Long-Term In Situ Properties of Geotextiles".. B. "Effect of Base Course Gradations on Results of Laboratory Pumping Tests". 468-492. Highway Research Board.. 1983.S. 1963.L. R. "Geotextile Filter Criteria". R. Vol. F-21 Jurgenson.F-18 Barksdale. and Yoder. 1957.. Highway Research Board. F-29 Carroll. 1958. F-26 Hoffman. Proceedings. "The Application of Theories of Elasticity and Plasticity to Foundation Engineering".J.. Proceedings.. 1958..D.R.C. 443-478. J. T.S.. pp. Transportation Research Record 916. S. L.H. F-22 Dempsey. Georgia Institute of Technology. "The Effects of Groundwater on Pavement Foundations". p.K.

Vol..S.. Proceedings. F-36 Barenberg. G. "Effects of Moisture and Drainage on Behavior and Performance of NJDOT Rigid Pavements". and Barenberg. "Design and Behavior of SoilFabric-Aggregate Systems". of Civil Engineering.J. Research Record No. 287-301. and Bell. 1981. Vol. D. 6. No. Proceedings. and Decker.A.. Transportation Research Board. F-37 Kozlov.R. F-39 Christopher.P. June.. G. Dunnigan.F-30 Gerry. New Jersey Dept. "Identification and Nature of Dispersive Soils".. 64-75. 2.L. F-33 Bender.L. J.. pp.. 466-468. 28. 84-015-7740.S. "Modeling of Effects of Moisture and Drainage of NJDOT Flexible Pavement Systems". B..S. F-31 Glynn. GT4. Department of the Army. B. E. E. and Raymond. F-38 Sherard. Research Report. April. 1983. 1982. 671. D. "Improved Drainage and Frost Action Criteria for New Jersey Pavement Design". No. Geotechnical Testing Journal. 1984. 1978. R. of Civil Engineering. pp.. Vol. E. pp. April. March. "Evaluation of Two Geotextile Installations in Excess of a Decade Old". 102. . CW-02215. University of Illinois. 53-63.L.. and Brown. 4. ASCE. "The Behavior of Geotextiles as Separating Membrane on Glacial Till Subgrades".S.J.. March. 79-88. Dept. III. 1976.J. F-35 Barenberg. L.. Research Report.. "Equivalent Opening Size of Geotextiles". F-34 Office of the Chief. GTJODJ. and Cochrane.. Transportation Research Board. Geotechnique. University of Illinois. Dept. 1983.P. A. S.Febury F-32 Snaith.R. M... July. pp. 150 p.1987NewOrlans. Civil Works Construction Guide Specification. Vol.T. D. 1986. pp. "The Filtration Behavior of Construction Fabrics Under Conditions of Dynamic Loading". of Transportation Report No. "Civil Works Construction Guide Specifications for Geotextiles Used as Filters". Geosynthic. Transportation Research Record 916.

APPENDIX G DURABILITY .

maintaining a high strength over a period of time for reinforcement would appear not to be as important as retaining the stiffness of the geosynthetic. Considering possible future pavement rehabilitation. For severe separation applications. a high level of stiffness must be maintained over a large number of environmental cycles and load repetitions. burst strength and tenacity will gradually decrease with time when buried beneath a pavement. Therefore. The strength of a stiff to very stiff geosynthetic. The rate at which the loss occurs. however. except when used for moderate and severe separation applications. and (2) the polyesters. maintaining strength and ductility would be more important than for most pavement reinforcement applications. which are known primarily as polypropylenes and polyethylenes.APPENDIX G DURABILITY PAVEMENT APPLICATIONS The commonly used geosynthetics can be divided into two general groups: (1) the polyolefins. Most mechanical properties of geosynthetics such as grab strength. which should be used for pavement reinforcement applications. can vary greatly between the various polymer groups or even within a group depending . usually these forces will be less. the overall life may be as great as 40 years or more. (7-10 kN/m). is generally significantly greater than required. The geosynthetic. Most flexible pavements are designed for a life of about 20 to 25 years. When a geosynthetic is used as reinforcement for a permanent pavement. is subjected to forces that should not in general exceed about 40 to 60 lb/in. Their observed long-term durabiilty performance when buried in the field is summarized in this section.

Some general characteristics of polymers are summarized in Table G-1 and some specific advantages and disadvantages are given in Table G-2. 4. .upon the specific polymer characteristics such as molecular weight. Stiffness in some instances has been observed by Hoffman and Turgeon [G-1] and Christopher [G-2] to become greater as the geosynthetic becomes more brittle with age. Micro-organisms. Also. Chemical reactions resulting from chemicals in the soil in which it is buried. geosynthetic and its environment and are caused by a number of factors including: 1. Changes in mechanical properties with time occur through very complex interactions between the soil. or from chemicals having an external origin such as diesel fuel. Whether some geosynthetics actually become a more effective reinforcement with time has not been shown. 3. the durability properties of the individual fibers may be significantly different than the durability of the geosynthetic manufactured from the fibers. Aging by ultraviolet light before installation. additives. chemical pollutants or fertilizers from agricultural applications. chainbranching. 2. Sustained stress acting on the geosynthetic which through the mechanism of environmental stress cracking can significantly accelerate degradation due to chemical micro-organisms and light mechanisms. as long as a safe working stress of the geosynthetic is not exceeded as the strength decreases. the ability of the geosynthetic to act as a reinforcement might improve with time for some polymer groups. and the specific manufacturing process employed. As a result.

copper. Adsorption of Liquid. May be attacked by hydrocarbons such as fuels with time Hydrolysis . environmental stress Degradation due to oxidation catalized by heavy metals . manganese Degradation in strong alkaline environment such as concrete.1111111 POIVellter Fungt: viiew.! fr'IV# 7 'Acosta. copper.•41 &maims tologorepacess: I. Environmental stress cracking is adversely affected by the presence of stress risers and residual stress. Physical properties in general should be evaluated of the geosynthetic which can have different properties than the fibers. Ea Heati NNW Table G-2 Summary of Mechanisms of Deterioration. zinc. Polv. Polypropylene and Polyester Polymers(l) POLYMER TYPE MECHANISMS OF DETERIORATION GENERAL ADVANTAGES IMPORTANT DISADVANTAGES Polyethylene Environmental stress cracking catalized by an oxidizing environment. anne C:k •:4t• • xv.ertlen• Peovonotione oinergfiew Un• WU" V •Kte.iron. Q Nabitted Stabiitw :•. Anti-oxidants usually added Good resistance to low and high pH environments Susceptible to creep and stress relaxation. moniuno ben POPOIlde Avalon fuel Imams . 2.Table G-1 General Environmental Characteristics of Selected Polymers Enwonntentel lifter lory misong etimi) sawn absorrp Aca. . Oxidation.takes on water Good creep and stress relaxation properties Polyester - Attacked by strong alkaline environment Notes: 1. Advantages and Disadvantages of Polyethylene. Environmental stress cracking Degradation due to oxidation catalized by heavy metals iron. zinc. etc. lime and fertilizers Polypropylene Environmental stress cracking catalized by (2) an oxidizing environment. Oxidation Adsorption of Liquid Anti-oxidants usually added Good resistance to low pH environments Good resistance to fuels Susceptible to creep and stress relaxation. manganese.

heavy metals present. The overall strength loss of the geotextile was up to 30 percent. One study has indicated that the strength of some geosynthetics might increase after about the first year of burial [G-1]. the basic mechanisms affecting degradation for each material under consideration must be understood. Relatively little of this type data presently exists. and from one geosynthetic to another is almost impossible due to the very complex interaction of polymer structure and environment. After six years the nonwoven polyester geotextile G-5 . The reduction in fiber tensile strength in one series of burial tests was found by Sotten [G-3] to be less than ten percent. Long-term burial tests should be performed on the actual geosynthetic rather than the individual fibers from which it is made. Hence. Different environments including pH. Translation of durability performance data from one environment to another. but gradually decrease thereafter.SOIL BURIAL Full validation of the ability of a geosynthetic used as a reinforcement to withstand the detrimental effects of a soil environment can only be obtained by placing a geosynthetic in the ground for at least three to five years and preferably ten years or more. and chemical pollutants will have significantly different effects on various geosynthetics. In evaluating a geosynthetic for use in a particular environment. geosynthetic structure and bonding can have an important effect on overall geosynthetic durability which has also been observed in other studies [G-4J. Hoffman and Turgeon [G-1] have reported the change in grab strength with time over 6 years. wet-dry cycles. The geosynthetic should be stressed to a level comparable to that which would exist in the actual installation.

Slit tape polypropylenes placed in aerated. [G-7]. Since the geosynthetics were used as edge drains. The four polypropylenes exhibited losses of strength varying from 2 to 45 percent (machine direction). The geotextiles were buried in a highly organic. Table G-3 shows for these conditions the important effects that anti-oxidants. After one year of burial in peat. geosynthetics exposed for at least seven years showed average tenacity losses of 5 percent for polyethylene.6 geotextiles lost about 30 percent of their strength [G-5]. et al. polyethylene and a mixture of polypropylene and polyamide-coated polypropylene filaments. hence these geotextiles became stiffer with time. In apparent contradiction to this study.studied exhibited no loss in strength in the machine direction (a 26 percent strength loss was observed in the cross-direction). Temperature was held constant at 20°C. moving seawater were found to undergo a leaching out of anti-oxidants if the tape is less than about eight microns thick [G-6]. no loss in strength was observed for a polypropylene. Alternating cycles of wetting and drying were found to be particularly severe compared to other conditions. A statistically significant decrease in burst strength was not observed over the seven year period for any of the . 15 percent for nylon 6. The test specimens consisted of monofilaments of polypropylene. Burial tests for up to seven years on spunbonded.6.7. All geotextiles (except one nonwoven polypropylene) underwent a decrease in average elongation at failure varying up to 32 percent. but polyester and nylon 6. and 30 percent for polypropylene. they were not subjected to any significant level of stress during the study. metals and condition of submergence can have on the life of a polypropylene. moist soil having a pH of 6. needle-punched nonwoven geotextiles were conducted by Colin.

.5 0 yrs.3.-_. Norma l An t iOxi dant Min imum Exp ec ted L ife t ime in Mar it ime App lica t ion s. ■_ • • Po lyp ropy lene Fabr ic a t an Averag e Temp era ture o f 1 0° C 1200 y rs. 200. 'Low Leach ' An t i-Ox idan t 400-600 y rs.To ta l Under Wa ter Wit hou t Me ta l In fluenc e Withou t Me ta l In fluence With Me ta l In fluence 1 .„ . Inc lu ding Some Steep in Ly e L ife o f a Po lyp ropy lene 's -14 OOT Table G.r. 3 0. 41 60-1 00 y rs. 200 y rs. 's -14 00 9 1 4-4 .300 y rs.

did not undergo any significant loss of mechanical properties [G-9]. however. when buried for up to 32 months. showing strength losses of 43 to 67 percent for the polyesters compared to 12 percent for polypropylene [G-8]. and about 30 percent after ten years of burial. NaOH. Stabilizers were not used.G-2. Summary of Test Results. when subjected to stress in the field. This data was obtained from numerous sources including [G-1. Schneider [G-8] indicates geotextiles buried in one study for between four months and seven years. When exposed to a combination of HCL. Typically the better performing geotextiles lost about 15 percent of their strength after five years. Polyester and polyproylene.G-8. The loss of tenacity of a number of geotextiles buried under varying conditions for up to ten years in France and Austria has been summarized by Schneider [G-8]. underwent from five to as much as seventy percent loss in mechanical properties. are given on the figures for the 80 and 95 percent levels. Scatter diagrams showing observed long-term loss of strength as a function time are given in Figure G-1 primarily for polyproylene and polyester geotextiles. in any of these materials. .G-10]. Confidence limits. polyester nonwovens were found to be quite susceptible to degradation.G-7. became embrittled during this time. Both low and high density polyethylene. The level of significance of the data was generally very low except for the nonwoven polypropylene geotextiles where it was 73 percent. sunlight and burial. One polypropylene geotextile did indicate a nine percent average loss of burst strength. however. which admittedly are rather crude for this data.samples.

H. 4 • qo3 • 44 %Of • • i = La U 8 0 c z 8 • 1 q oco q q q q q o o o 2 $ N N I (x) .1-11.441 4 • 401 • q * 4 q N q 2 ig 03 • 4•• CONFIDENCEUNITS • . it w) co WI g CONFIDENCE UNITS 4 .1.5 40 SSO1 q ? q 0 CONFIDENCE UNITS .44 .4 • • fli 4% • g it to 03 co q q0 Y o d N i 4 4 41 • 4:1 • q o o q° 0 N / I I N o C q q 2 O 8 R (94 I HION.0N38. • 41 4! 0 q 0 co q q 9 2 2 N q 0 0 CI N (x) .L3N3NIS JO SSO1 R / i q 2 1 0 Obs erve d Streng th Loss o f Geosyn the t ics with Time..laS JO (%) 1-11DN321. V • 1• • • 4 in co co 0 411 • li1 N • 41 • • • q 2 q 2 o d N o 0 si • ..LS JO SSO1 SSO1 0 d it ig in 0 m CO 1 ye q ea • • • • 1.

In these comparisons. The wide range of geosynthetics. . test methods and environments included in this data undoubtedly account for at least some of the large scatter and poor statistical correlations found. the polyester geosynthetics showed long-term performance behavior comparable to the polypropylenes. As a result. grab strength and tenacity. loss of strength was measured by a number of different tests including burst strength. The results indicate after 10 years the typical reduction in strength of a polypropylene or polyester geotextile should be about 20 percent. only general trends should be observed from the data. the 80 percent confidence limit indicates a strength loss of about 30 percent. With two exceptions.

B. LeClerc.R. G-9 Colin. pp. Proceedings. Las Vegas. 1986. 647-652.. D. Second International Conference on Geotextiles. G-5 Barsvary. "Some Answers Components on Durability Problem of Geotextiles".. Las Vegas.. pp. Congress Papers. 1981.L. Transportation Research Record 916. Transportation Research Record 916. B. G. Paute. August. M. pp. May. Proceedings.E.W.. pp. Carlsson. 1983. May.. Mitton. Draft Report. August. Netlon Limited. 1982. G. 3. Conference on Geotextiles. M. A.L. 77-84. III. G-4 Strobeck. D. Geotextiles and Geomembranes. 509.. J. and Wiles.M.. Correspondence and Unpublished Report. Second International Conference on Geotextiles...M. Brussels. pp. D. G. "The Effect of Soil Burial Exposure on Some Geotechnical Fabrics". Nonwovens for Technical Applications (EDANA)...J.APPENDIX G REFERENCES G-1 Hoffman. G-6 Wrigley. Journal of Applied Polymer Science. 26. Proceedings.. Vol. "The Durability of Tensar Geogrids".. and Wiles. 60-75. Vol.. and Fayoux. J. 1981. Vol. England. G-2 Christopher. G. M..C. G-7 Colin.K. . 89-93.T. 16. 1982.D. 79-88.. p. 1986.. G-3 Sotton. 553-558. "Evaluation of Two Geotextile Installations in Excess of a Decade Old"... Seneca. pp. and McLean. N.19. G-10 Sotton. Transportation Research Board. G-8 Schneider. Phillips Fibers Corp. III. R. "Instrumented Case Histories of Fabric Reinforced Embankments over Peat Deposits". and Turgeon. Index 81. M. 1983.J. "Durability of Geotextiles". "Long-Term Durability". Singapore.. S. D. Vol. H. Cooney. Transportation Research Board. 1986... "Long-Term In Situ Properties of Geotextiles".. 1985. D.D. Carlsson.

APPENDIX H PRELIMINARY EXPERIMENTAL PLAN FOR FULL-SCALE FIELD TEST SECTIONS .

The stiffness of the geotextile should be at least 1500 lbs/in. (260 kN/m). (500700 kN/m). however. Prestressing was also found to give similar reductions in permanent deformations of the base and subgrade as prerutting. The experiment is divided into two parts involving (1) five test sections constructed using a high quality aggregate base.APPENDIX H PRKL1KINA1 Y EXPERIMENTAL PLAN FOR FULL-SCALE FIELD TEST SECTIONS INTRODUCTION An experimental plan is presented for evaluating in the field the improvement in pavement performance that can be achieved from the more promising techniques identified during the NCHRP 10-33 project. and (2) . If desired. Prerutting the unstabilized aggregate base without reinforcement. 2. Because of the high cost of prestressing. a prestressed test section was not directly included in the proposed experiment. Geogrid Reinforcement of the unstabilized aggregate base. These methods of improvement are as follows: 1. The minimum stiffness of the geogrid should be Sg = 1500 lbs/in. it could be readily added to the test program as pointed out in the discussion. (260 kN/m) and preferably 3000 to 4000 lbs/in. The inclusion of a non-woven geosynthetic reinforced section would be a possibility if sufficient funds and space are available to compare its performance with the geogrid reinforcement proposed. TEST SECTIONS The layout of the ten test sections proposed for the experiment are shown in Figure H-1.

4.■ 00 c--) c0 (1) cn o •. Each(no t inc luding trans it ion s ) oG E c0 C..4 11 E ►.) LI 0 cl:1 1-1 0 0 • "0 CO o a) IJ 4-4 ccl 0 0 0 C a1Q 11 U1 G II C.tl 1 0 Sec t ions @ 1 00ft.) $. 0 Po Ft L. 171. 0 .4 a.) "0 co I.-1 u. a12 0 C..4 X CU X CU $-I 0 "0 Z cf) 0 $4 vii cZ X C14 4-4 0 1-I 4-■ G 0 0 "01 u. 4-1 —4 . 44 Figu re H .) E 0 c3 X ft. 0 cu C..6.1.1 1:10 0 I. u ao H No ta t io n : CL 00 t.

(8 m) in length between each section. H-4 . Longer test sections are encouraged. Although not shown. except Section 10.5 in. is to have a 4.five test sections constructed using a low quality aggregate base susceptible to rutting. A Falling Weight Deflectometer (FWD). which is to be prerutted. The test sections should be a minimum of 100 ft. separation and filtration capability. Extensive vane shear. Also. A control section is included as one of the test sections for each base type. An even stronger structural section might be included in the experiment if sufficient space and funds are available.5 to 3. All test sections. The high quality base experiment could be placed on one side of the pavement and the low quality base experiment on the other to conserve space. The purpose of these tests is to establish the variability of the subgrade between each section. (114 mm) thick asphalt surfacing and an 8 in. (200 mm) low quality aggregate base.5 in. it would be quite desirable to include a companion control section. (32 m) in length with a transition at least 25 ft. use of a geogrid and nonwoven fabric together could be studied to provide reinforcement.0 percent. (64 mm) asphalt concrete surfacing and a 10 in. Measurements should also be made to establish as-constructed thicknesses of each layer of the test sections. Test Sections 1 to 5 should be placed over a soft subgrade having a CBR of about 2. are to be constructed using a 2. cone penetrometer or standard penetration resistance tests should be conducted within the subgrade at close intervals in each wheel track of the test sections. high quality construction is achieved for each test section. A careful quality control program should be conducted to insure uniform. (250 mm) unstabilized aggregate base. Test Section 10.

A heavy vehicle having single tires on each axle should be used. High Quality Base Sections. a single rut should be formed in each side of the lane. If desired. Prerutting would be accomplished in Test Section 1 by forming two wheel ruts in each side of the single lane test section. the improvement in performance that can be obtained by either prerutting or reinforcing a low quality base.device. Optimum depth of prerutting is studied in the low quality base experiment. Sections 2 and 3 have geogrid reinforcement at the center and bottom of the base. Prerutting would be carried out for an aggregate base thickness of about 7 in. additional aggregate would be added to bring the base to final grade. After prerutting. The FWD tests will serve as an important indicator of any variation in pavement strength between test sections. should be used to evaluate the as-constructed stiffness of each section. respectively. (200-300 mm) apart. which is also prerutted. (260 kN/m). Section 2 could be prestressed. in the field. reinforced sections should have similar stiffnesses to the control sections. and then densified again by a vibratory roller. The. number of passes required. Two prerutted sections and two reinforced sections are included in the high quality base experiment. (50 mm) is developed. A good subgrade . prerutting should be continued until a rut depth of approximately 2 in. In each test section. In Section 5. (180 mm). it could also be included in this study. Low Quality Base Section. and the optimum position for geosynthetic reinforcement. The minimum stiffness of the geogrid should be S g = 1500 lbs/in. The high quality base section study is designed to investigate the best pattern to use for prerutting. This experiment is included in the study to establish. The ruts would be about 12 in.

Two prerutted sections are included in the study to allow determination of the influence of prerut depth on performance.. Section 10 is included in the experiment to determine whether or not improved performance due to prerutting is obtained for heavier pavement sections. Section 6 should be prerutted to a depth of about 1. MEASUREMENTS The primary indicators of pavement test section performance are surface rutting and fatigue cracking. In general.could be used rather than a weak one for this experiment. Two pressure cells should H-6 . while Section 7 should be prerutted to a depth of about 3 in. Inductance Bison strain coils should be employed to measure both permanent and resilient deformations in each layer (Figure H-2). In Section 9 a geogrid reinforcement (S g > 1500 lbs/in. similar to the one described in Appendix D. At least one pair of strain coils (preferably two) should be placed in the bottom of the aggregate base to measure lateral tensile strain. 260 kN/m) would be placed at the center of the base. Use of a surface profilometer. The following instrumentation should be used for each test section. (37-50 mm). The instrumentation layout for one test section should be similar to that shown in Figure H-2. (90 mm). Both of these variables should be carefully measured periodically throughout the study. a duplicate set of instruments is provided to allow for instrumentation loss during installation and instrument malfunction. is recommended in addition to the manual measurement of rut depth. Such a program is therefore recommended.5-2 in. Much valuable information can be gained through a carefully designed instrumentation program demonstrated during the experiments conducted as a part of this study.

/// 6 in.) 12 24 PLAN VIEW Figure H-2.5 in. Subgrade 2. Strain Coils ELEVATION VIEW Embedment Strain Cage Pressure Pressure Cell Cell in in Subgrade Aggregate B se Embedment Strain Coil 2 in. Pressure Cells 6 in.5 or 3 in. Preliminary Instrument Plan for Each Test Section. Dia. Strain Coil Stack Direction of soligrommnfong Wheel Travel 2 in.Mobile Strain Coil Embedment Strain Gage Asphaltic Concrete 2. H-7 . Diameter Strain Coil Strain Coils in Aggregate Base to Measure Tens Pressure Cell Thermocouple Strain Coils and/or Strain Gages Attached to Geogrid S g 24 12 0 Distance from Center (in. 10 in.

Resilient and permanent deformation characteristics of the low and high quality aggregate base and also of the subgrade. Shear strength and water content of the subgrade beneath each test sections. Mix design characteristics of the asphalt concrete surfacing. Thermocouples for measuring temperature should be placed in each section. Stress-strain and strength of the geogrid reinforcement as determined by a wide width tension test. Tensile strain in the bottom of the asphalt concrete should be measured using embedment type wire resistance strain gages. In addition to using strain coils. and measurements made each time readings are taken. 2. wire resistance strain gages should also be employed to directly measure strain in the geogrid reinforcement. The embedment gages should be oriented perpendicular to the direction of the traffic. Although quite desirable. 5. 3.be used to measure vertical stress on top of the subgrade. 4. . Placement of moisture gages in the subgrade would also be desirable. MATERIAL PROPERTIES The following laboratory material properties should as a minimum be evaluated as a part of the materials evaluation program: 1. Friction characteristics of the geogrid reinforcement as determined by a direct shear test. the two vertically oriented pressure cells in the base shown in Figure H-2 could be omitted for reasons of economy.

England January 1989 . Georgia Stephen F. Brown University of Nottingham NottiViarn.POTENTIAL BENEFITS OF GEOSYNTRETI FLEXIBLE PAVEMENTS FINAL REP ationai Richard D. Barksdale _Georgia institute of Technology tianta.

A decision concerning acceptance by the Transportation Research Board and publication in the regular NCHRP series will not be made until a complete technical review has been made and discussed with the researchers. They are not necessarily those of the Transportation Research Board.Acknowledgment This work was sponsored by the American Association of State Highway and Transportation Officials. Disclaimer This copy is an uncorrected draft as submitted by the research agency. or of the individual states participating in the National Cooperative Highway Research Program. the National Research Council. . American Association of State Highway and Transportation Officials. The opinions and conclusions expressed or implied in the report are those of the research agency. in cooperation with the Federal Highway Administration. or the Federal Highway Administration. and was conducted in the National Cooperative Highway Research Program which is administered by the Transportation Research Board of the National Research Council.

Reinforcement of Roadways Analytical Study Large-Scale Laboratory Experiments Summary and Conclusions 14 16 20 51 83 CHAPTER III SYNTHESIS OF RESULTS. INTERPRETATION.TABLE OF CONTENTS Page LIST OF FIGURES LIST OF TABLES ACKNOWLEDGMENTS ABSTRACT 1 SUMMARY INTRODUCTION AND RESEARCH APPROACH Objectives of Research Research Approach 8 9 10 CHAPTER II FINDINGS Literature Review . APPRAISAL AND APPLICATION Introduction Geosynthetic Reinforcement Summary 85 85 86 138 CHAPTER I CHAPTER IV CONCLUSIONS AND SUGGESTED RESEARCH Introduction Overall Evaluation of Aggregate Base Reinforcement Techniques Separation and Filtration Durability Suggested Research 140 140 APPENDIX A REFERENCES 161 140 155 156 158 .

AC/Es = 3.. 39 Equivalent Base Thicknesses for Equal Strain: S g = 1/3 Up 44 Equivalent Base Thicknesses for Equal Strain: Sg = 2/3 Up 44 Geosynthetic Slack Force .5 in..5 ksi 39 Variation in Radial Strain in Bottom of Aggregate Base (Tension is Positive) .Strain Relations Used in Nonlinear Model 45 Variation of Radial Stress o r With Poisson's Ratio (Tension is Positive) 45 2 3 4 7 8 9 10 11 12 13 14 ii . . 23 5 Typical Variations of Resilient Moduli with CBR .5 in.5 ksi 38 Equivalent Base Thickness for Equal Strain: 2. AC/E s = 12. AC/Es = 3. 18 Pavement Geometries.5 ksi 38 Equivalent Base Thickness for Equal Strain: 6.LIST OF FIGURES Figure 1 Page General Approach Used Evaluating Geosynthetic Reinforcement of Aggregate Bases for Flexible Pavements 12 Effect of Reinforcement on Behavior of a Subgrade Haul Road Section Without Reinforcement 18 Effect of Reinforcement on Behavior of a Subgrade Haul Road Section With Reinforcement . 36 Equivalent Base Thickness for Equal Strain: 2.5 in. Resilient Moduli and Thicknesses Used in Primary Sensitivity Studies . 25 6 Variation of Radial Stress at Top of Subgrade with Radial Distance from Centerline (Tension is Positive) .

All Test Series .LIST OF FIGURES (continued) Page Figure 15 Theoretical Influence of Prestress on Equivalent Base Thickness: E r and E v Strain Criteria 50 16 Pavement Test Facility 54 17 Distribution of the Number of Passes of Wheel Load in Multiple Track Tests 54 Variation of Rut Depth Measured by Profilometer with the Number of Passes of 1.5 kip Wheel Load . .Third and Fourth Series . 78 Variation of Permanent Surface Deformation with Number of Passes of Wheel Load in Supplimentary Single Track Tests .5 kip Wheel Load . 71 Variation of Vertical Resilient Strain with Depth of Pavement for All Test Series 72 Variation of Longitudinal Resilient Strain at Top and Bottom of Granular Base with Number of Passes of 1.5 kip Wheel Load . . . 75 Variation of Transient Vertical Stress at the Top of Subgrade with Number of 1.Second to Fourth Test Series 80 18 19 20 21 22 23 24 25 26 27 28 iii .5 kip Wheel Loads .All Test Series 63 Pavement Surface Profiles Measured by Profilometer at End of Tests . 64 Variation of Vertical Permanent Deformation in the Aggregate Base with Number of Passes of 1.5 kip Wheel Load All Four Test Series 66 Variation of Permanent Surface Deformation with Number of Passes of Wheel Load in Single Track Tests All Four Test Series 69 Variation of Vertical Permanent Strain with Depth of Pavement for All Four Test Series . 65 Variation of Vertical Permanent Deformation in the Subgrade with Number of Passes of 1.Third and Fourth Series .All Test Series 77 Variation of Transient Longitudinal Stress at Top and Bottom of Granular Base with Number of Passes of 1.All Four Test Series .5 kips Wheel Load .

89 31 Selected Geosynthetic Stress-Strain Relationships . 55). .5 in. Base: S g = 6000 lbs/in.. 132 42 43 44 45 46 iv . 97 Superposition of Initial Stress and Stress Change Due to Loading 97 Reduction in Permanent Deformation Due to Geosynthetic for Soil Near Failure 99 33 34 35 36 37 Reduction in Subgrade Permanent Deformation . Ref.72 in. 106 . 118 Free and Fixed Direct Shear Apparatus for Evaluating Interface Friction . ... AC/9. . 93 Reduction in Response Variable as a Function of Base Thickness 93 Variation of Radial Stress in Base and Subgrade With Base Thickness. et al.LIST OF FIGURES (continued) Page Figure 29 Pavement Surface Condition at the End of the MultiTrack Tests . 115 Theoretical Effect of Slack on Force in Geosynthetic: 2. 89 32 Variation of Subgrade Resilient Modulus With Depth Estimated From Test Results . 82 30 Basic Idealized Definitions of Geosynthetic Stiffness.... 122 Variation of Shear Stress Along Geosynthetic Due to Initial Prestress Force on Edge . 115 Influence of Subgrade Modulus on Permanent Deformation: S g = 4000 lbs/in. 122 Influence of Geosynthetic Pore Opening Size on Friction Efficiency (Data from Collios. ..All Test Sections .. 106 . 38 Reduction in Base Permanent Deformation 39 Improvement in Performance with Geosynthetic Stiffness 113 40 Improvement in Performance with Geosynthetic Stiffness 113 41 Influence of Base Thickness on Permanent Deformation: S g= 4000 lbs/in.

150 Placement of Wide Fill to Take Slack Out of Geosynthetic . 149 Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Vertical Subgrade Strain: 6. Subgrade CBR = 10 150 Break-Even Cost of Geosynthetic for Given Savings in Stone Base Thickness and Stone Cost .5 in. Subgrade CBR = 3 . 149 Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Radial Strain in AC: 2. AC. AC. Subgrade CBR = 3 . Subgrade CBR = 3 . 154 v . Strain in AC: 2.5 in. 148 Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Vertical Subgrade Strain: 2.5 in. Subgrade CBR = 3 .5 in.LIST OF FIGURES (continued) Pag e Figure 47 48 49 50 51 52 53 Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Radial Strain in AC: 2. 148 Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Radial .5 in. AC. AC. . AC.

40 7 Effect of Initial Slack on Geosynthetic Performance 43 8 Effect of Base Quality on Geosynthetic Reinforcement Performance 43 Effect of Prestressing on Pavement Response: 2.LIST OF TABLES Table 1 2 3 4 5 6 Page AASHTO Design for Pavement Sections Used in Sensitivity Study 25 Effect of Geosynthetic Reinforcement on Pavement Response: 2. AC. Es = 3500 psi 48 10 Summary of Test Sections 52 11 Transverse Loading Sequence Used in Multiple Track Test Series 2 through 4 56 Description of Test Sections Used in Laboratory Experiment and Purpose of the Supplimentary Single Track Tests 58 Summary of Measured Pavement Response Data Near the Beginning and End of the Tests for All Test Series 61 Summary of Lateral Resilient Strain in Geosynthetics and Longitudinal Resilient Strain at Bottom of Asphalt Test Series 1 and 2 74 Summary of Lateral Resilient Strain in Geosynthetics and Longitudinal Resilient Strain at Bottom of Asphalt Test Series 3 and 4 74 9 12 13 14 15 vi . AC. 33 Effect of Geosynthetic Reinforcement Position on Pavement Response: 2. AC.5 in.5 in. .5 in.5 in.5 in. AC. E s = 3500 psi 29 Effect of Geosynthetic Reinforcement on Pavement Response: 2. E s = 3500 psi . E s = 12.5 in. E s = 6000 psi 31 Effect of Geosynthetic Reinforcement on Pavement Response: 2. E s = 3500 psi 27 Effect of Geosynthetic Reinforcement on Pavement Response: 6. AC.500 psi . AC.

LIST OF TABLES (continued) Table 16 17 18 19 20 21 Page Tentative Stiffness Classification of Geosynthetic for Base Reinforcement of Surfaced Pavements . 90 Influence of Geosynthetic Position on Potential Fatigue and Rutting Performance 104 Influence of Asphalt Thickness and Subgrade Stiffness on Geosynthetic Effectiveness 105 Influence of Aggregate Base Quality on Effectiveness of Geosynthetic Reinforcement 111 Typical Friction and Adhesion Values Found for Geosynthetics Placed Between Aggregate Base and Clay Subgrade 126 Beneficial Effect on Performance of Prestressing the Aggregate Base 135 vii . .

The following Research Assistants at Georgia Tech participated in the study: Jorge Mottoa. gave valuable assistance in setting up the experiments. Research Assistant. The Georgia Institute of Technology was the contractor for this study. Dawson. Richard D. Lan Yisheng and Mike Greenly gave much valuable assistance in analyzing data. the University of Nottingham. William S. Lecturer in Civil Engineering at Nottingham gave advice on the experimental work and reviewed sections of the report. and the Nicolon Corporation. the Georgia Institute of Technology. Barksdale. the University of Nottingham.ACKNOWLEDGMENTS This research was performed under NCHRP Project 10-33 by the School of Civil Engineering. Barry V. Stephen F. the University of Nottingham. universities and the geosynthetics industry who all made valuable contributions to this project. Professor of Civil Engineering.. Andrew R. was Principal Investigator. Georgia Tech. Orr. Professor of Civil Engineering. sincere appreciation is extended to the many engineers with state DOT's. Brodrick. Geosynthetics were supplied by Netlon Ltd. Department of Civil Engineering. and Yan Dai performed the numerical calculations. Francis Chan performed the experimental studies at the University of Nottingham. and the Department of Civil Engineering. Finally. Professor Brown and Francis Chan. University of Nottingham was Co-Principal Investigator. viii . The authors of the report are Professor Barksdale. Brown. The work performed at the University of Nottingham was under a subcontract with the Georgia Institute of Technology.

ABSTRACT This study was primarily concerned with the geosynthetic reinforcement of an aggregate base of a surfaced. whether reinforcement is effective in reducing rutting where low quality bases and/or weak subgrades are present needs to be established by field trials. geogrid stiffness should be at least 1500 lbs/in. A linear elastic finite element model having a crossanisotropic aggregate base gave a slightly better prediction of response than a nonlinear finite element model having an isotropic base. Separation. As the strength of the section increases. Both prerutting and prestressing the aggregate base were found. Both large-scale laboratory pavement tests and an analytical sensitivity study were conducted. flexible pavement. of the subgrade. A geogrid performed differently and considerably better than a much stiffer woven geotextile. The greatest benefit of reinforcement appears to be due to small changes in radial stress and strain in the base and upper 12 in. for a woven geotextile. and (3) prerutting the aggregate base with and without reinforcement. compared to about 4000 lbs/in. to significantly reduce permanent deformations. Greatest improvement occurs when the material is near failure. (2) pretensioning a geosynthetic placed within the base. Prerutting without reinforcement gave performance equal to ix .5 to 3 placed on weak subgrades (CBR < 3 percent). Specific methods of reinforcement evaluated included (1) reinforcement placed within the base. experimentally. For somewhat stronger sections. the potential benefits of reinforcement decrease. filtration and durability were also considered. Reinforcement is effective for reducing rutting in light sections having Structural Numbers less than 2.

that of prestressing and significantly better than just reinforcement. . Prerutting is relatively inexpensive to perform and deserves further evaluation.

SUMMARY This study was primarily concerned with the geosynthetic reinforcement of an aggregate base of a surfaced.0 to 1. (2) pretensioning a geosynthetic placed within the base. and (3) prerutting the aggregate base either with or without geosynthetic reinforcement. The term geosynthetic as used in this study refers to either geotextiles or geogrids manufactured from polymers.5 percent. The silty clay subgrade used had a CBR of about 2. A linear elastic finite element model having a cross-anisotropic aggregate base was found to give a slightly better prediction of tensile strain and other response variables than a nonlinear 1 . subgrade strengths and geosynthetic stiffnesses. Larger strains cause greater forces in the geosynthetic and more effective reinforcement performance. (6. Analytical Modeling. (25-38 mm) thick asphalt surfacing placed over a 6 or 8 in.7 kN) moving wheel load was employed in the laboratory experiments. Specific methods of improvement evaluated included (1) geotextile and geogrid reinforcement placed within the base.5 in. The accurate prediction of tensile strain in the bottom of the base was found to be very important. A 1500 lb. REINFORCEMENT Both large-scale laboratory pavement tests and an analytical sensitivity study were conducted. The analytical sensitivity study considered a wide range of pavement structures. flexible pavement. The large-scale pavement tests consisted of a 1. (150200 mm) thick aggregate base. Extensive measurements of pavement response from this study and also a previous one were employed to select the most appropriate analytical model for use in the sensitivity study.

5 to 3) on a weak subgrade (CBR 5. Reinforcement of a thin pavement (SN = 2. the improvement caused by 2 . More analytical and experimental research is required to define the mechanisms of improvement associated with geogrids and develop suitable models. strain and deflection are all relatively small for pavements designed to carry more than about 200. The resilient modulus of the subgrade was found to very rapidly increase with depth. This models a membrane such as a woven geotextile.000 equivalent 18 kip (80 kN) single axle loads. Geogrids. A modest improvement in fatigue life can be gained from geosynthetic reinforcement. As a result. As the strength of the pavement section increases and/or the materials become stronger. The model assumed a membrane reinforcement with appropriate friction factors on the top and bottom. the state of stress in the aggregate base and the subgrade moves away from failure. The effects of geosynthetic reinforcement on stress. 3 percent) can potentially reduce the permanent deformations in the subgrade and/or the aggregate base by significant amounts. Mechanisms of Reinforcement.finite element model having an isotropic base. The greatest beneficial effect of reinforcement appears to be due to small changes in radial stress and strain together with slight reductions of vertical stress in the aggregate base and on top of the subgrade. As a result. geosynthetic reinforcement of an aggregate base will in general have relatively little effect on overall pavement stiffness. however. were found to perform differently than a woven geotextile. The low resilient modulus existing at the top of the subgrade causes a relatively large tensile strain in the bottom of the aggregate base and hence much larger forces in the geosynthetic than for a subgrade whose resilient modulus is constant with depth.

Reductions in rutting due to reinforcement occur in only about the upper 12 in. (300 mm) of the subgrade. Also. Therefore. As the structural number and subgrade strength decreases below these values. (5 kN/m). (700 kN/m) for woven geotextiles. be expected to show any significant level of improvement due to geosynthetic reinforcement of the type studied. Reinforcement Improvement. The experimental results indicate that a geogrid having an open mesh has the reinforcing capability of a woven geotextile having a stiffness approximately 2. in general. typically being less than about 30 lbs/in. the improvement in performance due to reinforcement should rapidly become greater. Geosynthetic stiffness S g is defined as the force in the geosynthetic per unit length at 5 percent strain divided by the corresponding strain.5 to 3 if reduction in subgrade rutting is to be achieved by geosynthetic reinforcement. the minimum stiffness to be used for aggregate base reinforcement applications should be about 1500 lbs/in. 24 MN/m 2 ) are most likely to be improved by geosynthetic reinforcement. Forces developed in the geosynthetic are relatively small.5 times as great as the geogrid. From the experimental and analytical findings. Strong pavement sections placed over good subgrades would not. Light to moderate strength sections placed on weak subgrades having a CBR 3 percent (E s = 3500 psi. 3 . The structural section in general should have AASHTO Structural Numbers no greater than about 2. (260 kN/m) for geogrids and 4000 lbs/in. Hence geogrids perform differently than woven geotextiles.5 times the actual one should be used in the figures and tables.reinforcement rapidly becomes small. Type and Stiffness of Geosynthetic. a reinforcement stiffness 2. in determining the beneficial effects of geogrids.

Low Quality Base. For weak subgrades and/or low quality bases. particularly if a good subgrade is present. Light sections on weak subgrades reinforced with geosynthetics having woven geotextile stiffnesses of about 4000 to 6000 lbs/in. Considerably more reduction in rutting occurs for the thinner sections on weak subgrades than for heavier sections on strong subgrades. Geosynthetic Position. Geosynthetic reinforcement of a low quality aggregate base can. be reduced on the order of 20 to 40 percent. the reinforcement should be in the middle of the base to minimize rutting. For pavements constructed on soft subgrades.sections with asphalt surface thicknesses much greater than about 2. under the proper conditions.5 in. For light pavement sections constructed with low quality aggregate bases. The asphalt surface should in general be less than about 2. the reinforcement should be placed 4 . this corresponds to actual reductions in base thickness of about 1 to 2 in. (700-1000 kN/m) can give reductions in base thickness on the order of 10 to 20 percent based on equal strain criteria in the subgrade and bottom of the asphalt surfacing. but this needs to be established by field trials. (64-90 mm) in thickness for the reinforcement to be most effective.5 in. Some stronger sections having low quality bases and/or weak subgrades may be improved by reinforcement. (64-90 mm) would in general be expected to exhibit relatively little improvement even if placed on relatively weak subgrades. under ideal conditions. Improvement Levels.5 to 3. For light sections. reduce rutting. Field trials are required to establish the benefits of reinforcing heavier sections having low quality bases. total rutting in the base and subgrade of light sections might.5 to 3. (25-50 mm).

The total expense associated with prestressing an aggregate base would be on the order of 5 or more times that of prerutting the base at one level when a geosynthetic reinforcement is not used. SEPARATION AND FILTRATION Separation problems involve the mixing of an aggregate base/subbase with an underlying weak subgrade. 300 kN/m).at or near the bottom of the base. and significantly better performance compared to the use of stiff to very stiff. This would be particularly true if the subgrade is known to have rutting problems. however.. angular open-graded aggregates placed directly upon a soft or very soft subgrade are most critical with respect to separation. non-prestressed reinforcement. They usually occur during construction of the first lift of the granular layer. experimentally. Full-scale field experiments should be conducted to more fully validate the concept of prerutting and develop appropriate prerutting techniques. Stress relaxation over a long period of time. might significantly reduce the effectiveness of prestressing the geosynthetic. PRERUTTING AND PRESTRESSING Both prerutting and prestressing the geosynthetic were found. Either a properly designed sand or geotextile filter can be used to maintain a reasonably clean interface. Large. The laboratory experiments indicate prerutting without reinforcement gives performance equal to that of prestressing. and the base is of high quality and well compacted. to significantly reduce permanent deformations within the base and subgrade. Both woven and 5 . The cost of prerutting an aggregate base at one level would be on the order of 50 to 100 percent of the inplace cost of a stiff geogrid (S g = 1700 lbs/in.

should generally have at least a 20 year life. DURABILITY Strength loss with time is highly variable and depends upon many factors including material type. although satisfactorily performing geotextiles can usually be selected. flow conditions. 6 . Sand filters used for filtration. Under favorable conditions the loss of strength of geosynthetics on the average is about 30 percent in the first 10 years. geosynthetics. For reinforcement applications. geosynthetic stiffness is the most important structural consideration. if selected to fit the environmental conditions. geogrids might exhibit a lower strength loss. The severity of erosion is dependent upon the structural thickness of the pavement. may perform better than geoextile filters. A very severe environment with respect to subgrade erosion exists beneath a pavement which includes reversible. filtration and pavement reinforcement applications. which determines the stress applied to the subgrade and also the number of load applications. stress level. possibly turbulent. when properly designed. For separation. Thick nonwoven geotextiles perform better than thin nonwovens or wovens. because of their greater thickness. the amount of contamination due to fines being washed into this layer must be minimized by use of a filter. and the local environment in which it is placed.nonwoven geotextiles have been found to adequately perform the separation function. Whether better reinforcement performance will result has not been demonstrated. partly because of their three-dimensional effect. Some geosynthetics become more brittle with time and actually increase in stiffness. manufacturing details. When an open-graded drainage layer is placed above the subgrade.

Mechanistically. geogrids perform differently than the analytical model used in this study to develop most of the results. A proposed preliminary guide for conducting field tests is given in Appendix H. Additional research is also needed to better define the durability of geosynthetics under varying stress and environmental conditions. Therefore.ADDITIONAL RESEARCH Geogrid reinforcement and prerutting the base of non-reinforced sections appears to be the most promising methods studied for the reinforcement of aggregate bases. 7 . the recommendation is made that full-scale field tests be conducted to further explore the benefits of these techniques.

5 to 4. which are stiffer than the commonly used geotextiles. nylon and polyethylene. has lead to the use of the general term "geosynthetic" which can include both geotextiles.5. Growth rates in geotextile sales during the 1980's have averaged about 20 percent each year. Because of their great variation in type.6]. and have an open mesh with typical rib spacings of about 1. 1. The numbers given in brackets refer to the references presented in Appendix A. Geogrids are manufactured by a special process. geosynthetics refer to geotextiles and geogrids. strength. polyester. geonets and geomembranes.5 inches (38-114 mm). however.85 x 10 9 m2 ) of geotextiles annually. composition. geogrids. geocomposites. slopes and embankments [2. More recently polyethylene and polypropylene geogrids have been introduced in Canada and then in the United States [2]. As used in this report. 8 . Early civil engineering applications of geosynthetics were primarily for drainage. These fabrics have widely varying material properties including stiffness. geotextiles have a very wide application in civil engineering in general and transportation engineering in particular. and creep characteristics [1] (1) .CHAFFER INTRODUCTION AND RESEARCH APPROACH The geotextile industry in the United States presently distributes over 1000 million square yards (0. and resulting material properties. Both nonwoven and woven geotextile fabrics are made from polypropylene. The introduction of geogrids. With time many new uses for geosynthetics have developed including the reinforcement of earth structures such as retaining walls. erosion control and haul road or railroad construction [3.4].

The specific objectives of the project are as follows: 9 . flexible pavement structure. Mitchell. [6] have recently presented an excellent state-of-the-art summary of the reinforcement of soil structures including the use of geosynthetics. the potential structural and economic advantages of geosynthetic reinforcement within a granular base of a surfaced. OBJECTIVES OF RESEARCH One potential application of geosynthetics is the improvement in performance of flexible pavements by the placement of a geosynthetic either within or at the bottom of an unstabilized aggregate base. Considerable interest presently exists among both highway engineers and manufacturers for using geosynthetics as reinforcement for flexible pavements. The results of these studies appear to be encouraging.13]. At the present time. An important need presently exists for establishing the potential benefits that might be derived from the reinforcement of the aggregate base and the conditions necessary for geosynthetic reinforcement to be effective. A number of studies have also been performed to evaluate the use of geosynthetics for overlays [7-12]. Several investigations have also been conducted to determine the effect of placing a geogrid within the asphalt layer to prolong fatigue life and reduce rutting [12. particularly with respect to the use of stiff geogrids as reinforcement in the asphalt surfacing.The application of geosynthetics for reinforcement of many types of earth structures has gained reasonably good acceptance in recent years. from both a theoretical and practical viewpoint. et al. however. relatively little factual information has been developed concerning the utilization of geosynthetics as reinforcement in the aggregate base. The overall objective of this research project is to evaluate.

4. 3. Of these factors. stiffness and strength. Longterm durability of the geosynthetic was also felt to be an important factor deserving consideration. Develop a preliminary experimental plan including layout and instrumentation for conducting a full-scale field experiment to verify and extend to practice the most promising findings of this study. 2. Perform an analytical sensitivity study of the influence due to reinforcement of pertinent design variables on pavement performance. installation and longterm durability aspects.1. Verify using laboratory tests the most promising combination of variables. consideration had to be given to the large number of factors potentially affecting the overall longterm behavior of a geosynthetic reinforced. Develop practical guidelines for the design of flexible pavements having granular bases reinforced with geosynthetics including economics. 10 . and the overall strength of the pavement structure. geosynthetic location within the aggregate base. The potential effect on performance of geosynthetic slack which might develop during construction and also slip between the geosynthetic and surrounding materials were also included in the study. Techniques to potentially improve geosynthetic performance within a pavement deserving consideration in the study included (1) prestressing the geosynthetic. RESEARCH APPROACH To approach this problem in a systematic manner. the more important ones appeared to be geosynthetic type. and (2) prerutting the geosynthetic. flexible pavement structure.

They also permitted investigation. This analytical model permitted the inclusion of a geosynthetic reinforcing membrane at any desired location within the aggregate layer. A nonlinear. The large-scale laboratory tests made possible verification of the general concept and mechanisms of reinforcement. Further. The analytical sensitivity studies permitted carefully investigating the influence on performance and design of all the important variables identified. including both design and construction related factors. isotropic finite element pavement idealization was selected for use in the sensitivity study. feedback 11 .The potential importance of all of the above factors on pavement performance clearly indicates geosynthetic reinforcement of a pavement is a quite complicated problem. As a result. caution must be exercised in a study of this type in distinguishing between conditions which will and will not result in improved performance due to reinforcement. As the analytical study progressed. of factors such as prerutting and prestressing of the geosynthetic which are difficult to model theoretically and hence require verification. the influence of the geosynthetic reinforcement is relatively small in terms of its effect on stresses and strains within the pavement. The general research approach taken is summarized in Figure 1. followed by large-scale laboratory tests. The most important variables affecting geosynthetic performance were first identified. in an actual pavement. Emphasis in the investigation was placed on identifying the mechanisms associated with reinforcement and their effect upon the levels of improvement. An analytical sensitivity study was then conducted. The analytical studies were essential for extending the findings to include practical pavement design considerations.

12 .Identify Reinforcement Problems (I) Design Variables (2) Practical Aspects Construction Durability Analytical Study Predict Response Develop Equivalent Sections Failure Mechanisms Fatigue Rutting Verify Analytical Model Large-Scale Laboratory Tests Type Ceosynthetic Pavement Structure Prerutting Prestress Define Performance Mechanisms Synthesis of Results General Benefits Equivalent Designs Construction Aspects Durability Economics i Overall Evaluation Figure 1. General Approach Used Evaluating Geosynthetic Reinforcement of Aggregate Bases for Flexible Pavements.

cross-anisotropic model was employed for most of the sensitivity study which agreed reasonably well with the observed experimental test section response. The equivalent designs were based on maintaining the same strain in the bottom of the asphalt surfacing and at the top of the subgrade. The analytical results were then carefully integrated together with the large-scale laboratory test studies. This synthesis includes all important aspects of reinforcement such as the actual mechanisms leading to improvement. Permanent deformation in both the aggregate base and the subgrade was also evaluated. A detailed synthesis of the results was then assembled drawing upon the findings of both this study and previous investigations. Lateral tensile strain developed in the bottom of the aggregate base and the tensile strain in the geosynthetic were considered to be two of the more important variables used to verify the cross-anisotropic model.from the laboratory test track study and previous investigations showed that adjustments in the analytical model were required to yield better agreement with observed response. This important feedback loop thus improved the accuracy and reliability of the analysis. the role of geosynthetic stiffness. The analytical model was employed to develop equivalent pavement structural designs for a range of conditions comparing geosynthetic reinforced sections with similar non-reinforced ones. As a result. 13 . a linear elastic. equivalent structural designs and practical considerations such as economics and construction aspects.

Type and stiffness of the geosynthetic reinforcement. relatively little factual information is available to assist the designer with the proper utilization of geosynthetics for pavement reinforcement applications. and by a large scale laboratory test track study. perhaps at least partially due to the emphasis presently being placed in this area by the geosynthetics industry. The laboratory investigation was conducted to verify the general analytical approach and to also study important selected reinforcement aspects in detail using simulated field conditions including a moving wheel loading. 14 . Location of the reinforcement within the aggregate base. The potentiaL beneficial effects of aggregate base reinforcement are investigated in this study using both an analytical finite element model. The important general pavement variables considered in this phase of the investigation were as follows: 1. Pavement thickness. 3. 4. The analytical investigation permits a broad range of variables to be considered including development of structural designs for reinforced pavement sections. Quality of subgrade and base materials as defined by their resilient moduli and permanent deformation characteristics. Presently the important area of reinforcement of pavements is rapidly expanding.CHAPTER II FINDINGS The potential beneficial effects of employing a geosynthetic as a reinforcement within a flexible pavement are investigated in this Chapter. 2. Unfortunately. The only position of the reinforcement considered is within an unstabilized aggregate base.

Potential improvement in performance is evidenced by an overall reduction in permanent deformation and/or improvement in fatigue life of the asphalt surfacing. Slip at the interface between the geosynthetic and surrounding materials. In the analytical studies. constant vertical subgrade strain criteria were also used to control subgrade rutting. 8. Prestressing the geosynthetic. Influence of slack left in the geosynthetic during field placement. 15 . an analytical procedure was also employed to evaluate the effects of geosynthetic reinforcement on permanent deformations. 6. equivalent pavement designs were developed for geosynthetic reinforced structural sections compared to similar sections without reinforcement. 7. and also the individual rutting in the base and subgrade. For the laboratory test track study. Equivalent sections were established by requiring equal tensile strain in the bottom of the asphalt layer for both sections. Finally. pavement performance was accessed primarily by permanent deformation including the total amount of surface rutting. Prerutting the geosynthetic as a simple means of removing slack and providing a prestretching effect.5. A detailed synthesis and interpretation of the many results presented in this chapter is given in Chapter III.

Tests involving the reinforcement of unsurfaced roads have almost always shown an improvement in performance. This finding would of course apply for either geotextile or geogrid reinforcement. for relatively large movements. however.18-20]. Bender and Barenberg [3] concluded. a reinforcing element confines the subgrade by restraining the upheaval 16 .14]. Confinement and reinforcement of the fill layer 2.13.REINFORCEMENT OF ROADWAYS UNSURFACED ROADS Geosynthetics are frequently used as a reinforcing element in unsurfaced haul roads. typically characterized by a CBR less than 2 percent.42]. The economics of justifying the use of a geosynthetic must. in large scale test pits [16. These tests have been conducted at the model scale in test boxes [3. Mechanisms of Behavior Bender and Barenberg [3] studied the behavior of soil-aggregate and soil-fabric-systems both analytically and in the laboratory. be considered for each application [26]. Prevention of contamination of the fill by fine particles. They identified the following four principal mechanisms of improvement when a geosynthetic is placed between a haul road fill and a soft subgrade: 1. The reinforcement of the fill layer was attributed primarily to the high tensile modulus of the geotextile element. Confinement of the subgrade 3. Separation of the subgrade and fill layer 4. and full-scale field trials [21-26.LITERATURE REVIEW . Beneficial effects are greatest when construction is on soft cohesive soils. it is usually not as great when stronger and thicker subbases are involved [24]. Although improved performance may still occur.

17 . et al. for a geotextile reinforced haul road constructed on a very soft subgrade. et al. Structural Performance . The work of Bender and Barenberg [3] indicated that over ground of low bearing capacity having a CBR less than about 2 percent. the use of a geotextile could enable a 30 percent reduction in aggregate depth. increases the bearing capacity of the soil as illustrated in Figures 2 and 3.Full-Scale Experimental Results Relatively few full-scale field tests have been conducted to verify the specific mechanisms which account for the observed improvement in performance of geosynthetic reinforced haul roads. use of another geotextile increased failure to 10. Nevertheless. uniform bases on very soft subgrades. Ruddock. Confinement. with greater savings being achieved with the use of a stiffer reinforcement. Ramalho-Ortigao and Palmeira [26] found. The importance of developing large rut depths (and hence large fabric strain) was later confirmed by the work of Barenberg [27] and Sowers. Webster and Watkins [25] observed for a firm clay subgrade that one geotextile reinforcement increased the required repetitions to failure from 70 to 250 equivalent 18-kip (80 kN) axle loads. [28]. the conservative recommendation was made that no reduction in aggregate thickness should be allowed. frequently referred to as the tension membrane effect. that approximately 10 to 24 percent less cohesive fill was required when reinforcement was used.000 repetitions.. Later work by Barenberg [27] and Lai and Robnett [29] emphasized the importance of the stiffness of the geotextile. [21] found plastic strains in the subgrade to be reduced by the presence of a geotextile. Another 2 to 3 inch (50-70mm) reduction in base thickness was also possible since aggregate loss did not occur during construction of coarse.generally associated with a shear failure.

0. 191E89 .71 7' 6 [ou t 2.71 111'7' 9E=1 00 0 9 CU 00 S 1 27 [6 9' 1 O z • 191 E- cr ..6' 0 0911 61 . 042301 Res i lien t •■•1 76' 9E- 691' 1 [' I- 9 67 C 9' 9E- I L I CE S' I 79 0' 6 - 9' 0 k 91 10' 0 - (8710' 0 - 1 1. AC/ 1 2. A GGR EGAT E BASE SUBGRADE E s . 5 I N.35 00 PSI I .0. 073 231 I• Z .. 7 97' i 519' 9- 177' 9 - 6' 0 6 -7 I I I 1I r . 0 41 44 I 0' 0 1 9 710' 01 I [0. 90' 0- 90.. 0 IN. 0- E ffec t o f Geosy n t he t ic Re in fo rcemen t • ri l 9L h' E • Iu • • • .-• C.7• 7- r." ..7 .

• :. . .. 9- 606' 9 - E V II - tOZ' S- 911 . 0Z- 01E1 TE1' 6- 1 Z' ST- '0 E97 Z' Z ZI ZI Ioszt- 9 7 11- 19.I RADIAL STRAIN • • RADIAL STRAINI VERTI CAL STRAI NI SUBGRADE I 07'71 Z ...I 8 991 L Z9 I . 611- 0.I 9 ... 7777 6SEI- SIO• 0 72AGGREGATE BASE S WIG ■3500PSI [C IE- L 9 . "- / I. 717 1 11' 8- L_ I 6 .I 61 1Z 9802 9• 8- CALI- I00' S- IOS UO I 1 T -E 6. 0 I CBS .7 :- 77:771 107' O I I00 06 1 REST 13 17 . La rA pp i Le.I V8' 11- 76• 1 00S 1 005 1 0009 170.I 6971- I' II Z.71 • 7 7Z ZI. 0I vzil l C ' SE- 195' 9 - 77777 BOTTOM 1/ 3ORA( .777 0 AGGREGATE BASE SUBGRADEE s • ■3500 PS I CC6' I- TEBASESUBGRADEE. I i 1 I 17 . 0- I jI [' SIT- I Tt- I• C O I I "9E 1 .o CZE . 0 11 .0 T• 1 1' 9 9' 7 977' 8 - 6 / 81 51 91 L6LI- 18 11 -- O' Ll I 905' 9 - 119 . o o9 z- (9_ 01)' 1' zj I • i . II9 F77 f70713. 0L S VERT ICAL STRAIN 'VERT ICAL STRE SS 8•C B . 0C- [ 6* Z IZ SI 91S' O 0 . 00- C/1 . >•46 VI . 0- 1 "I Ct I 13. 0I % I S•C Shl 1 RIE Z. 0 IIE' R I- I9561 IC l3r 0 Z . I CV7 .• .. 0 OIC' 0 I 6' i 9 Z 66S' 6- 1115 . I_ '7 'S- I' l l9' 91-I I II 6I . 6- 001 09 0006 ■ . 'S! 9• 17 .8 . 17• 116 I' S Z I5 vsi- I Z' S - ■mmimmimmill 6S H `l C ROZ -II IC R1 ..I ZI S 7 - I 4 '77 1.) —• . r 16 11 h • CZ - 1 1 . 11B 6 . 9 OZ8 . 0 II1 Z . ° Izoe .0% SSRHIS 191 091 I 28 I I • I .1.I07( .. . 91- 8/T' S- I11 9 . 0 £ 891 I0 . ZZ OSOZILOZ 1 •0 LI S* 0 D9 -1 CL5 . 0 I I 1---. 11 I I I pm IZZ Z . S CO T_1 IS9 . 0 1I 7- 19191 ELC I • 4 1I II - I' SE.I . SR' 01 Ni t 690 1 ' M U %I SS 3R1S 1 V101V1 3 .1 h• I 777779 ="17771 L STR ESS 6' S I I 9' 01 1 MMMMMMMMMM 8' 1[ 1 1' 1 99 C80 .. 6- 1ZI Z . 01 + .I 000 9 I Si ll- I8L• 6- 711 • 0 67( .7Z- 1181 7Z5 . 0 I ZOI .

. 9Z - I' S SE 0' Z- O' S RE - 7' 6 . 4.0 093 2 -. 009 83 ISE cT..ET9T--1-97071 I C7.8996' Z 9 9E S 9E SU BGRADE E s DIf f.I 809"7 - SS 7 00. I 1 1 i I 1 0 z I. 00449 1 I 1 I 0009 .I 1 VERTI CAL.ISLE ( Tsd ) z o VERT. . DEFLECTION 777' 9IE 1 I - F IS' 9Z - Z• 1+ 19' LZ-1I ILE 1 I 0' 7E 9 - CO Z 0. I I I I i I I k 1 i I a I I i i I . ■ 0 CK A I41 0 . 7 I90' 09' 11- rut- 1 0' 0- I90' 0- 17I' E 11 8 .I 9E 6' 7 O U- I 70 1 .0 ULa. 00 542 1-77 06 1' 9.9S . E89 .0 0 2 3 9L E 9' 6- ZE L• E E ' 0+ I C '9- 7 16' E 6' i - Z9' L Z - 1166' E 1 F071.. SUBGRADE Es IX VE RT ICALSTRESS f 1 2.--. 01I EE ZOI O* 0- OOSI 000 9 0006 I - .8Z900' 0.—. 104- 81Z' Z 0613 . Res ilie n t Mo du lus o f re in force d an d a AGGREGATE 2. - 87500' OS S00" - -.I 8' LK S' I . SZ• S Z I 110L' Z E . AC/ 6. 006245 TC7 + 1 I Z 9E • I TENSILE STRAIN BOTTOM OF AC I 5 IN. 009 79 0• 1- I I EZ Z* 9 - CI - -.1 1O' Z - I 1 -. 1 10' 0- 70' 0 - 6' E L 9. 0 X 1 (9_0 0 1 I H MO NIVVIS 1 V10041I 3amens 33 I 0..-- 0' E 89 . AC /7. AGGREGATE BA SE BLE '/ J1 0 X 1 . Z E• I- Z . 0098 8 6Z' SZ ISE ZSL ESE 06 L' S- 'p- '3 3T OX I 7' 07- I I 6 18' S- 9E 0' 9960' 9- Lit' L- OLE • II 999' L - h' E- I16' 0- 9L E' L- 7771 -0.3 •—•• CA I—.1 0 -1 -. . 0092 9 -. 88E I 7• 0- 9' 1. 00453 -. I• 1- 1E 600" - 6 E6 0 0' - 0006 1 OOS II I000 9 1 1 . STRAIN =12.- Z 'O I z• I I"- lc -0 I -.I 69E 010' 0] 6L E 900' 01 SE 900' 01 I z0 I [16 600 .1 0006 I I I OOSI -..I 96 0" E Z' 0- I I E• 06E ' Ina X 1 1 11 .Z89- 1 VERTICALSURFACE 10' 0- 0' 7E9. 500 p s i TOP1 /3 OF AGGREGATE BASE RADIAL S TRE SS 'IITO X ] (9 . Z C P.: !.T. 5 I N. 171 - L' 28 9- 91 7 ... 0 0 53 9 I 6' 0- E• I- 199E 01 0' 0] SIZ010 . L.I 785' 0 1I l7CE I 70 .I I I 89I' E 229• E 6' 0+ 69' L Z .00446 di fferen ce F -. 5IN. •••••• .7 •—•1 pp al 0 I—I <A A Lii V5 I I • • . 500PS I ■12 9L' LZ- 8' 0. 0 10 . 0- 959' 1 I 6LL' O I 1 I 9 . AGGREGATE BA SE r I 9. P. .I Z' ISE I I 9" 8I E frZSE 8' 14 9- 0 .) 0 0 1 I i 0 ••••• ••••••.--- ISU Z *J ITO X F77:71 I 1773:- 7V 0- E 9E U] VERTICALSTRESS Z' S I C 86I' Z 757 57-7 I SL' 92 - 8I' S Z - 7:CT. 0 IN..0.

30 S9 -4000 PSI 0.500 psi Figure 6.0 4.05 —0. ar ( PSI) 0.0 60 DISTANCE FROM CENTERLINE.90 0 PSI v) o_ 2.30 00 6.15 00 4.. Variation of Radial Stress at Top of Subgrade with Radial Distance from Centerline (Tension is Positive). R (INCHES) (b) Subgrade E s = 12. R (INCHES) 2.6000 PSI —0.00 S .0.5 IN A./15.0 60 (a) Subgrade E s = 3500 psi RADIALSTR ESS.5 KSI 0.60 S9 -1000 PSI b" Cii cn w cc iv) __I < tO < cc 0.05 —0.0 2.3 IN Es =3. 36 .15 0.0 DISTANCE FROM CENTERLINE.C.

with the stresses being very small. General Response. small radial stresses occur regardless of the presence of geosynthetic reinforcement (Figure 6).500 psi. 37 . An exception is the section having the stiff subgrade where the difference was much less. 86 kN/m 2 ). The pavement response was also determined for geosynthetic reinforcement locations at the lower 1/3 and upper 2/3 positions within the aggregate base in addition to the bottom of the base. To develop a set of design curves for the three levels of geosynthetic stiffnesses requires a total of twelve finite element computer analyses. Equivalent structural sections can be readily estimated as shown in Figures 7 and 8 by selecting a reduced aggregate base thickness for a reinforced section that has the same level of strain as in the corresponding unreinforced section. Geosynthetic Position.subgrade having a CBR of about 10 percent (E s = 12. Figures 7 through 9 summarize the effect of geosynthetic reinforcement on the tensile strain in the bottom of the asphalt and the vertical compressive strain on top of the subgrade. The actual magnitude of the change in radial stress in the bottom of the aggregate base is about 10 to 20 percent of that occurring in the subgrade. Figure 10 shows for the same sections as compared in Figure 7 the reduction in radial tensile stress caused in the bottom of the aggregate base due to reinforcement. The theoretical effect of reinforcement position on the major response variables is summarized in Table 6 for the three levels of geosynthetic stiffness used in the study. The effect of position was only studied for sections having a subgrade stiffness Es = 3500 psi (24 MN/m 2 ).

1 (9 _1:IL—) A3 '30YeiCERS 10 101 NW IS 1■0011113A (moo I) A 3avaosns 40 141 NNW TratIMA )' co Equ iva len t Base Thicknes s U . 5KS I (b) Ve r t ic a l Ey on Subg ra de 41.-1 r4 — VD •• cci co 1.) -4 • •.rt 0.. pp U H tn o V7 Z u1 ▪ O yj > U tr) Il c.1 cn O PQ 8 z h' uz rW co 0 8z 2 (a ) Radia l Cr in AC 0 0 0 a (g—oi) 33 "ON NIO11013 NMI'S 1VICYLI 0 0 (9— 0 t ) 43 "O'Y 1101108 NMI'S IVIOVII 38 o .• "c1 a:7 • AC /E s =3.4 CO1 a ed 2 el 4.

.0 (8-0 • r.4 (9_oi) 13 . T (INCH ES) ..uoe 8 .1 o .303Y80eflS 40 d01 PO WS TOLIMA NIVUIS Es Ve r t ic a l C r on Subg ra de ( b) Su bg rade E s = 1 2..31/ mr: wo. 5ks i Equ iva le n t Base Thic kn ess for EqualStr a in : THICKNESSOFBASE. n o 8 0 fiD wava (ISd) 39 1D ISS3HIS IVICMI 0 0 (ISd) j a sSlas WM'S o 8 0 ..

.7. ■ . 5in.ZL- 0'711 1 r051 1 WO . 7 g 1 1 1 I I ■mo . I 't' 1 . I 1 I p .0 6' 0 11 61 70' - . 7- RFICI 717 1 F 711 70971 OR71 I cf 1C - STRAINI 1 R7If6 51f 09 1 1- I 1 • (p4 11 ) 7. 0 IN.II L7• 91- I '3(IV11E)R05 R .0 .I 6. — . R• fl 1 09I f- 19 IC191fR L7T 1 . I I90' 0 - 7. 6' 11 ?UR- 6 771 13 . 1 [' SI7 1 L' 1 I I Of - 1 9..R7' 91 - 1.I 077 0* R07 70 V • [c cI t o' ... 6C 1C- . 5 IN. 169' 9C . I J71' 1 6 1. AC/ I2..fIl 6L0 /0 : 1 ns11 0 -I 6' 1 - 9 .1• • 10' 0 - 0.71 Z' 6 R' 67 7" 7. 9S - I7f• 1 0 70' 1 6R1' I M* 6- IC I10' - I F I 0. ) .. 67 77' I 1 9' °- 1 01171 .1 -10 :171V. .1 . -. = ..1 11ARINA N0 T) I di f fe r ence be tween a I Of II 167' 91.= E ffec t o f Geosy n the t ic Re in fo rc emen t Pos it ion on Pav emen t Resp ons e : 2. ... .1 NO I ..1 II 1 .c co) -7.0 WO- TICO' 0 517' 0 I Re s ilien t R' L6 f .T1 LO' 0 - o7 tw.1 I fII7W .0 c141 [ 0' 0.1I 1 1' 0 - 17 . 5 I N. . - 1 I [1 110 % I ( 9 _01) ' ' 1 . 11 1 0 VIT. 11- I 0 •0 WO 6' 0 7' 0 6. /fLO' LC - 6' 0.0 '' 0 -. -7.1 — -.1IS • .6 LI 05 11 091 1 0111 I 67' 7 87' 1 C19 .P0911 I L. AC/I2.R97 I I L ' 0- R7' 9 C - 1 1 Z771 si t' I 7 115' 0 l' IZ- 0' 99 6S14' 0 ..[ OLII iIff V11 AIV :EMEnV ' NI0' 71 0V ' NI17 L . ( (0° l J 00 06 I 0006I .1 607 70' - [001 70' 0 1 - 71 .L):11 . vufins 40 170' 7 6/ 7' 1 1 9/7 .7 . 0• 0 -] 01 770 • 1 9 1.I '"fo --1 6' 1 1 97 [0' 0 .l. 'TITO I 5. ACCRECATE BASE S OBCRADE E s =3500P SI I !I I1/ 7 '. I I 00cf. 1 I Lf S' I 0111 7.-. 6f f • C7. 6 7LR' R - 177' 8 - C 07 ..- 0911 MI 1 071 1 I I 1 C. 1 1 1W ..1 [ MU "! I T ' 0 !) • I 7011 15f. IWO- I I -C. - 9 C9' 9C - 76' 9 C - I L0' 11. 1 .. 0 - DEF1 F el'ION VERTI CAL STRESS I TEN SIL ESTRAIN BOTTOM OF AC 1 RADIAL CEOS YNT HETI C0 BOTTOM2. 0 IN.. .1 TTf LW0 . ACCR ECATE RA SE S ORCRADE E s = 3 500 PSI 0' 0 7' 0 - CEOSYNTHET IC 1 / 1 UP 2.1.- .

0 101 . 11/ 9' 1/9 7 .7' 0 RADIA LSTRESS RAD IAL STRA IN %DI U] VERT ICAL. . 11- I I 0051 1 00091 0006] th ... I I i I .. 11 f R OZ - % k _01)' ' 1 I I $ 41 I 0 0 I I I 991 1 • 1 $ . 9. 9 - 019' 9 .9- 1611 - 1 9' 9 - 7191 6 '7 9 1 C .1 I £7 11 - II9tI 1 611 - • 131 0 %1( 9-01) . 711.1 I 7' 9 yet- GEOSYNTHET IC1 /3UP2.1 1211..[ 907 .... 0'7E' 0 1' 1 r 1 ( 81' 0 d (Isd ) I05 1I - I 1 016 .. zI- I I 00 06 6' Z 7'111- 6* ( 7 - h Z91 - 1E 9 . '7 EI 33FazI(Tsd ) z o_ BOTTOM1 /3OFAGGREGATEBASE RAD IA L STRAIN I PSI • 1 1 .■ E69' 0 I 96 11 . 0 1 9 5' 0 109 • 0 1 1-77777-1 777771 =7 E . 0 IN. 5IN.. 11 "5 9561 '1110 % I ( 9_ 01) 00 1 I 66( .. AC/ I 2.Dirt 51' 11- os . 1h - Fit '7 • 0 E . AGGREGAT EBA SESUBGRADE E s 3 500 PS I 1. _0 tj t. 0 .i 5 E1 1 '0 Z ' 11 .7 — /9 113 . AGGR EGATE BASE SUBGRADE..1 1 .0 51 05' 0 5 Z6.) U1 .. 5 in. Z • P. VERTICA L STR ESS E ' E9 .. AC/ I 2.1 156' 9 - 905' 9 Z19 .I . r. 11 - hh' 11- 0009 1 0006 1 01' 11- 006 1 1 •■■•=1. 0 ("7 . C. E s 3 500 1I I 15' 9 •. I I .6 CH - USE ± I o. STRAI NI VERT I CALSTRES S I RAD IAL STRESS O(p s 1)1 GEOSYNTH ETI CATBOTTOM2. AC /1 2. 516 01) I hf 0 1- 1I. 1... . I 971Z- I 1 . 0IN. 0 5991 - 1 ' SC 6' 91 - 9' O zs' o -ma 91. LA Is.. I. 0 AGGR EGATE BASE SUBCRAD EE s 3 500 PS I 1 %I(Isd ) 1 0 0 0009 Z . 01 1 •1- I '7' 1 0' 1 Z' O 115' 0 GEOSYNTHETIC2/3 UP 2.) I $ 1'7 11 - I Effec t o f Geo sy n t he t ic Re in fo rcemen t Po s it ion on Pavemen t Re sp on se : 2.1■ 13E . I 91 02 9' 1 Z6 11' 5 I E' ZI- E' 1 7 % 9' 91- T' ZZ - 1 1 6_ 1'76 hI61 .. . '7 61 9' 011' 69 9I 1 I 1 1910 . £1 81 Ir zz a 7111: ZE Z' 1 E.71' 0 tt 9 wo 1 56 .. ZI6 1 .ro556' 9 - 656' 9 - 956' 9 - 77 t• o- I I E° CI 6' 1 161 1-- hi tt - '79/1 - 9' 1 - £ . 9 - 91 9' 9- I959' 9 . 4- L '7' I 1. 116I1 f' /E6 Z• Z9 • 0- 191 1 .I i 0 .79 I 06 91' 0 I NI . 5 I N.—.116 0 6 06' 9- . I I RAIZ - 9• / - 9 . 19 - E' LZ- 3OVHDROS f. 5IN.. 1 55 1 I - 86 61 h• Z I 9611 - r ....I I 777 £' 0 6' 7. 9- 771-1 (611 1 591 6Z61 '7961 I I I S' EI -1 95'7' 0I (1 5 .7 l I Z . 0 10 1 E . DO ...[ 11 61- 95 11 - 1 Jib I I I 51 [-JIZT ICAL ST RA IN I 1 I 7 .

(1 MN/m). The results of the slack sensitivity study for the stronger subgrade is summarized in Table 7.2. 0. A limited sensitivity study was therefore conducted for Poisson's ratios of v = 0. The relative effects of slack on force in the geosynthetic were found to be similar for the stiff subgrade shown in Table 7 and also a weaker subgrade having E s = 3. 0.75 percent slack level. The resulting radial stress in the top of the subgrade as a function of Poisson's ratio of the geosynthetic is shown in Figure 14. or its effect on the response of a reinforced pavement.The influence of reinforcement position on horizontal tensile strain in the bottom of the asphalt and vertical compressive strain on top of the subgrade is given in Figures 11 and 12 for the 1/3 up from the bottom of the aggregate base position and the 2/3 position. The literature was found to contain little information on the value of Poisson's ratio of geosynthetics. three different levels of slack in the geosynthetic were analyzed using the nonlinear finite element model. A geosynthetic was used having an actual stiffness of 6000 lbs/in.5 ksi (24 MN/m 2 ). with the rate at which it is picked up becoming greater with the applied strain level. Poisson's Ratio. To determine the effect on performance.4.75 and 1. Slack. Slack levels of 0. As wheel load is applied in the field. A supplementary sensitivity study was conducted to determine the effect of base quality on the performance of geosynthetic reinforced 42 . This type of geosynthetic load-strain behavior was modeled using a smoothly varying interpolation function as shown in Figure 13 for the 0.3 and 0. Base Quality. the geosynthetic would gradually start to deform and begin picking up some of this load. The force on the geosynthetic should increase slowly at first.25.25 and 0.4 percent strain were chosen for the analysis.

5 ksi subgrade. Base characterized using high quality properties (Table C-5.6 -6 -39 3-7 -10 \ 7 Note: 1.75 -11 -11 -11 -12 -12 -14 Good Base Diff. (2) Good Base Diff.5 -22 -11 -. 43 ■ 1.45.4 - 0 9000 8.3 0. .3 6000 10.8 -19.8 -12 -4 -2. (B) Poor Sage Diff.4 0. Zero stress is inferred from the results obtained from the results for S 8 • 9000 lbs/in. Table 8 Effect of Base Quality on Geosynthetic Reinforcement Performance REDUCTION IN RUTTING REDUCTION IN BASE TRICKINESS BASE THICK. The atiffnesses shown are to limiting stiffnesses at the strain level where all the slack has been taken out.Table 7 Effect of Initial Slack on Geosynthetic Performance E Design 2. .72. Nodular ratio E b /Es of the aggregate base and subgrade.3 12.3 1. e v Poor lase Diff. for example. 2.72 in. 3. Cross-anisotrooio analysis.-4.5/9. Reduction in permanent deformation 3.5 - - 0.5/12.0 Stiffness SI (lbs7in.5/9. 2.5 in.4 2. asphalt surfacing and a 9.0 9.) 12. 0 0 (2) 0 (2) Notes: 1. indicate a 2.3 Slack (Percent) (1) subg.9 - 0 (2) 13.5 in.4 None . T (in..72 2.1 . Appendix C).9 - 0. (avg) (ksi) 12. 4. (2) Diff.) Vert. AC surfacing. 2. 2.25 0.4 9000 12.34 - 9000 10. The initial stiffness of each geosynthetic was assumed to be S s • 300 lbs/in.75 1.3 6000 6. 5. (2) AC SURFACING -8 -10 -13 Base Rutting Total Rutting (2 ) AC Radial t r Good Saes (1) Poor Base Diff.4 1. rather than zero. (2) Poor Bass (%) Diff. Appendix C). (I) 3500 PSI SUBGRADE -6.6 - - 6000 6. Subg.0 / -30 -2.5 IN. this strain level corresponds to the slack indicated. The numbers 2. (B) Good Rase Diff.5/15. 15. Subgrade characterized by bilinear properties (Table C-5. aggregate base.

§

r

(9-01 —) 43 •30`0119611S 30 d01 NIV21.1.5 'NOWA ,

cr

(IN

8
.4

THICKNESS OF BASE. T (INCHES)

(9_01—) N '3001L.181.1 30 d01 NIMIS 1V3LI1GA

w

a

S

0
N
.,....
....

(9_01.) /I * *31f Y101108 NIVHIS TAM!

(g—OL) 43 "ON 11011013 NIVaLS 'MUM

44

Eq u iva len t Base Thic knesses for Eq ua l Stra in :

i

8

F igure 1 2.

8

THICKNESS OFBASE. T (INCHES)

8

GEOSYNTH ETIC FORCE. F (LBS/IN)

9.00

6.00

3.00

CEOSYNTHET1C STRAIN. a (10 -3)

RADIALSTRESS, crr (Psi)

Figure 13. Geosynthetic Slack Force - Strain Relations Used in Nonlinear
Model.

10.00

13.00

16.00

THICKNESS OF BASE. T (INCHES)

Figure 14. Variation of Radial Stress a With Poisson's Ratio (Tension
r
is Positive).
45

pavements. For this study the subgrade used had a resilient modulus E s =
3500 psi (24 MN/m 2 ). A nonlinear finite element analysis indicated that a
low quality base has a modular ratio between the aggregate base (E b ) and the
subgrade (E s ) of about Eb/E s = 1 to 1.8 as compared to the average Eb/E s =
2.5 used as the standard modular ratio in the cross-anisotropic analyses.
The results of this study, which employed a modular ratio of 1.45, are
summarized in Table 8.

Prestressed Geosynthetic
An interesting possibility consists of prestressing the aggregate base
using a geosynthetic to apply the prestressing force [35,36]. The
prestressing effect was simulated in the finite element model at both the
bottom and the middle of the aggregate base. Once again, the same light

- reference pavement section was used consisting of a 2.5 in. (64 mm) asphalt
surfacing, a variable thickness aggregate base, and a homogeneous subgrade
having a resilient modulus E s = 3500 psi (24 MN/m-). The cross-anisotropic,
axisymmetric finite element formulation was once again used for the
prestress analysis. A net prestress force on the geosynthetic of either 10,
20 or 40 lbs/in. (2,4,7 kN/m) was applied in the model at a distance of 45
in (1140 mm) from the center of loading.
Theory shows that the force in a stretched axisymmetric membrane should
vary linearly from zero at the center to a maximum value along the edges.
Upon releasing the pretensioning force on the geosynthetic, shear stresses
are developed along the length of the geosynthetic as soon as it tries to
return to its unstretched position. These shear stresses vary approximately
linearly from a maximum at the edge to zero at the center, provided slip of
the geosynthetic does not occur.

The shear stresses transferred from the

geosynthetic to the pavement can be simulated by applying statically
46

equivalent concentrated horizontal forces at the node points located along
the horizontal plane where the geosynthetic is located.
In the analytical model the effect of the prestretched geosynthetic was
simulated entirely by applying appropriately concentrated forces at node
points. The external wheel load which was applied would cause a tensile
strain in the geosynthetic and hence affect performance of the prestressed
system. The tensile strain in the geosynthetic caused by the load was
neglected in the prestress analysis; other effects due to the wheel loading
were not neglected. The geosynthetic membrane effect due to external
loading that was neglected would reduce the prestress force, but improve
performance due to the reinforcing effect of the membrane.
In the prestress model the outer edge of the finite element mesh used
to represent the pavement was assumed to be restrained in the horizontal
directions. This was accomplished by placing rollers along the exterior
vertical boundary of the finite element grid. Edge restraint gives
conservative modeling with respect to the level of improvement caused by the
geosynthetic. The benefits derived from prestressing should actually fall
somewhere between a fixed and free exterior boundary condition.
The important effect of prestressing either the middle or the bottom of
the aggregate base on selected stresses, strains, and deflections within
each layer of the pavement is summarized in Table 9. Comparisons of tensile
strain in the asphalt layer and vertical compressive strain in the top of
the subgrade are given in Figure 15 for a geosynthetic stretching force of
20 lbs/in. (3.5 kN/m). To reduce tensile strain in the asphalt surface or
reduce rutting of the base, prestressing the middle of the layer is more
effective than prestressing the bottom. On the other hand, if subgrade

47

9[E'ill
rt16' St i

11 . 1C-1

I

VS' 6C- I

611 1

,- ,_ I611 0 1
1

7107

ZZZ I

/* CI -

1 If 0( 0' -I

I
17014771

1771
6' 6 Z-

VS L CO' -1

. 3 5001.5 1

BOTTOM: 2. 5 IN. AC/I2. 0IN. AGGREGATE BASE SUSCRADE E. 3500 PS

ISi ll

1' 9- 1-

V 't-

11' 1.

Li

19' 0E-

I

I • I-

I

11/ 1 90' -

(6 0S0' -1

ISS 90' -I

16190' -

[7:17.71

I"4'1r

Z1 910'

17771
I 0,1

1

I
1771

r-

'S I OZ690 . -11-7-1
Ti..7146 19Z00' - OI
Z' ZI-

BOTTOM: 2. 5 IN. AC/I 5. 1 I N. AGGREGATE BASESUOGRADC E.

I

4' 11

B* 5-

PRESTRESS (1

,• ,_ 1

" 1.

1 "1
t e r. ill

11611-

PRESTRESS D

I

I

°MOW-

4

60E1- 1

1' 11

-1

11 9' S

6' 1 11-

I

S• 4 -

I

I
[Z`:

I

MI-

911
1 •116..

4

19101- II

I,Su l
1-21-11

1

. AC/ 7. 5 I

co

v zI
z i- s- ]

-I

I

a

T' S-

S' 61]

I

4

I
1

0

C

toll-

SOLI

[

6' 1 6 1
S69* C-

I

a

I

4' 611 I

1

I

a

1c m1

a1

7' 1. 1IWt. .1

116 1

Z' ZI

48

a

I
1

I
I
a

61St-

[

...

III a

I

vs

I
a
I
I
a
a

N
0

N

O

N
N

0

0

N

O
0

0

N
_

co
e...

O

N.
I

VI
•=1

...

I

el

O

..7
II

In

VI
GO

In

N

••

GO
0

a.
L

11

fT

101

8

O

49

1250

tV
2

1180

0
mi
1100

1020
70

15.0

13.0

10.0

THICKNESS OF BASE, T (INCHES)
(a) Radial Strain C

7.0

r

in AC

13.0

10.0

THICKNESS OF BASE, T (INCHES)
(b) Vertical Strain E v on Subgrade
Figure 15. Theoretical Influence of Prestress on Equivalent Base
Thickness: E and E v Strain Criteria.
r

50

deformation is of concern, prestressing the bottom of the layer is most
effective.

LARGE-SCALE LABORATORY EXPERIMENTS
Large-scale laboratory experiments were conducted to explore specific
aspects of aggregate base reinforcement behavior, and to supplement and
assist in verifying the analytical results previously presented. These
large scale tests were performed in a test facility 16 ft. by 8 ft. (4.9 by
2.4 m) in plan using a 1.5 kip (7 kN) wheel loading moving at a speed of 3
mph (4.8 km/hr). Up to 70,000 repetitions of wheel loading were applied to
the sections in a constant temperature environment.
Four series of experiments were carried out, each consisting of three
pavement sections. The pavement sections included a thin asphalt surfacing,
an aggregate base (with or without geosynthetic reinforcement) and a soft
silty clay subgrade. A large number of potentially important variables
exist which could influence the performance of an asphalt pavement having a
geosynthetic reinforced aggregate base. Therefore several compromises were
made in selecting the variables included in the 12 sections tested.
Important variables included in the investigation were (1) geosynthetic
type, (2) location of geosynthetic within the aggregate base, (3) prerutting
the reinforced and unreinforced sections, (4) prestressing the aggregate
base using a geosynthetic and (5) pavement material quality. The test
sections included in this study and their designations are summarized in
Table 10. A knowledge of the notation used to designate the sections is
helpful later when the observed results are presented. A section name is
generally preceded by the letters PR (prerutted) or PS (prestressed) if
prerutting or prestressing is involved. This designation is then followed
by the letters GX (geotextile) or GD (geogrid) which indicates the type of
51

Table 10
Summary of Test Sections

Test
Series
1.

2

3

4

Proposed
Geometry
1 in. A.C.
6 in. Sand &
Gravel Base

1.5 in. A.C.
8 in. Crushed
Limestone

Section
Designation

Details of Geosynthetic
and Section Specification

PR--0C--B

Geotextile placed at bottom
of Base; Subgrade prerutted
by 0.75 in.

CONTROL

Control Section; no geosynthetics and no prerutting

GX-B

Same as PRIX B; no prerutting

PR-GD-B

Geogrid placed at bottom of
Base; Subgrade prerutted by
0.4 in.

CONTROL

Control Section

GD-B

Same as PR7GD-B;no prerutting

GX-B

Geotextile placed at bottom
of Base

CONTROL

Control Section; Prerutting
carried out at single track
test location

GX-M

Geotextile placed at middle
of Base

GX-M

Same as GX-M (Series 3); Prerutting carried out at single
track test location

GD-M

Same as GX-M but use geogrid

PS-GD-M

Prestressed Geogrid placed at
middle of base

,
Notes for section designation: PR = Prerutted PS m Prestress

GX = Geotextile GD = Geogrid
B = Bottom of Base
M = Middle of Base

52

geosynthetic used. The location of the geosynthetic which follows, is
represented by either M (middle of base) or B (bottom of base). Following
this notation, the section PR-GD-B indicates it is a prerutted section
having a geogrid located at the bottom of the aggregate base.
Materials, instrumentation and construction procedures used in the
laboratory tests are described in Appendix D. A summary of the material
properties are presented in Appendix E.

PAVEMENT TEST PROCEDURES
Load Application
The pavement tests were conducted at the University of Nottingham in
the Pavement Test Facility (PTF) as shown in Figure 16. This facility has
been described in detail by Brown, et al. [66]. Loading was applied to the
surface of the pavement by a 22 in. (560 mm) diameter, 6 in. (150 mm) wide
loading wheel fitted to a support carriage. The carriage moves on bearings
between two support beams which span the long side of the rectangular test
pit. The beams in turn are mounted on end bogies which allow the whole
assembly to traverse across the pavement. Two ultra low friction rams
controlled by a servo-hydraulic system are used to apply load to the wheel
and lift and lower it. A load feedback servo-mechanism is incorporated in
the system to maintain a constant wheel loading. The maximum wheel load
that can be achieved by the PTF is about 3.4 kips (15 kN), with a speed
range of 0 to 10 mph (0 to 16 km/hr). The whole assembly is housed in an
insulated room having temperature control.

Multiple Track Tests
The moving wheel in the PTF can be programmed to traverse, in a random
sequence, across the pavement to nine specified positions (four on each side

53

Figure 16. Pavement Test Facility.

0.8

8

20

a/1st SERIES

bNWSEMES

0.6

15 •

0.4

10

9
0-2
0.0

—C—
1 2 3 4

7

n

5 6 7 8 9

1 2 3

ul 25 cnrd SERIES
TO .

20

u.
O

15-

15

icri

10 •

10

5

5

0

4 5 6 7 8 9

25,dl4th SERIES I

0 E
12 3 4 5 6 7 89
12
WHEEL POSITION NUMBER

b.
3 4 5 6 7 8 9

Figure 17. Distribution of the Number of Passes of Wheel Load in Multiple
Track Tests.

54

of the center line). At each position a predetermined number of wheel
passes is applied. The spacing between wheel positions was set at a
constant step of 3 in. (75 mm). A realistic simulation can be obtained of
actual loading where traffic wander exists. Table 11 summarizes the loading
sequence adopted for the last three series of tests. It consisted of a 250pass cycle, starting with 55 passes along the center of the section
(Position 5), followed by 15 passes at position 8, then 7 passes at 9 (refer
to Table 11) until it finished back at the center line where the cycle was
repeated. During the scheduled recording of output from the
instrumentation, the center line track was given an additional 100 passes of
wheel load before actual recording began. This procedure ensured that
consistent and compatible outputs were recorded from the instruments
installed below the center line of the pavement. The total number of passes
in the multiple track tests for the second to fourth series were 69,690,
100,070 and 106,300, respectively. The distribution of these passes across
each loading position is shown in Figure 17. Note that the width of the
tire is larger than the distance between each track position. Therefore,
during the test, the wheel constantly overlapped two tracks at any one time.
Hence, the numbers shown in Table 11 and Figure 17 apply only to the center
of each track position.
In the first series of tests, because of the rapid deterioration and
very early failure of the pavement sections, the loading program described
above could not be executed. The total number of wheel load passes for this
test series was 1,690, and their distribution is shown in Figure 17.

Single Track Tests
On completion of the main multi-track tests, single track tests were
carried out along one or both sides of the main test area where the pavement
55

0 Cr) LA . 4) 4) (3) LA 0 LA Ln 0 0 4) CO 0 U) 0 CL la A CU 0 0 $-3 4) CO U) 61 8 I CO LLI Ctl •• 4J 0 56 .0 CU 4J 1-4 N Z 4) 1.• 0 C*1 ■••■ LA 0 CO mm) dis tance. cl a) .1 CNI 4) tn 4) U) 4) • %.4 4) CO U) cd 0 0.54 CJ CO N ▪ 1-1 %ID O 4) 1-1 0. 1—r 4.

Only surface rut depth was measured. only minor variations of load occurred. Nonetheless. difficulties were encountered at an early stage of the test in maintaining a uniform load across the three pavement sections which underwent different amounts of deflection. Therefore. Bidirectional loading means that load was applied on the wheel while it moved in each direction. Stress and strain data were not obtained for these single track tests.5 kips (3 to 11 kN). In the first series of tests. and refinements were made in the servo-system which controlled the load. These special tests normally involved the use of a much higher wheel load.5 kips (6. As a result.6 kN). The load exerted by the rolling wheel on the pavement during Test Series 2 through 4 of the multi-track tests was 1. Three additional single track tests were performed during the second to fourth test series. the actual load varied from 0. Wheel Loads Bidirectional wheel loading was used in all tests.7 to 2. In subsequent test series. The designations of the test sections follow those for the multitrack tests previously described. so that the deterioration of the pavement structure would be greatly accelerated. since instruments were not located beneath the loading path.6 kN). due to the rapid deterioration of the pavement and hence large surface deformations. these tests helped greatly to confirm trends observed in the development of permanent deformation during the multi-track tests. The single track tests also made possible extra comparisons of the performance of pavement sections tested in the prerutted and non-prerutted condition.5 kips (6. Details of these tests and their purposes are shown in Table 12. while the average load was 1. generally less than 10 percent of the 57 . however.had not been previously loaded. much stronger pavement sections were constructed.

5.) 4.) W o 4..) 'a a . a.) C) a) r.4 C (1) g cn s O /-4 C.4 4.• a) cn O a./ 0 a) U) U) c 0 0 4.o en 8 0 ••i 4.1 0..) co 4) . az ).) E.-I 4-5 u 0 13 as a) • E-1 o ca no CNI r-I e•1 C >-. U cn A 38 U cn o 4. G 4) a) E 4-4 >4 co a.

7 and 3 in. For all other test series a 2 kip (9 kN) load was applied. all of these supplementary tests employed bidirectional loading. respectively. Near the end of the test when the pavement surface became uneven. of 2. (68 and 76 mm). tire pressure and load. With the exception of the single track test carried out during the first series. The wheel moved at a speed of about 2 to 3 miles per hour (3.2 to 4. Temperatures at the asphalt surface and within the aggregate base and the subgrade were found to be about 2 to 4°F (1 to 2°C) lower than that of the air. the contact pressures acting on the pavement from a 1. This load variation was probably also due to the unevenness in the longitudinal profile of the pavement.5 and 2 kip (6. The tire pressure was maintained at 80 psi (550 kN/m 2 ). respectively. a slower speed was sometimes necessary to maintain constant loading. Data Recording Procedure The transverse profile and permanent strain readings from the aggregate base and silty clay subgrade were taken at appropriate intervals during testing of all pavement sections to establish their deformation 59 .8 km/hr) with slight variations between forward and reverse direction.6 and 9 kN) wheel load were estimated to be 67 and 73 psi (460 and 500 kN/m2 ).6°F (20 t 2 °C) throughout the testing. Based on a previous investigation of the effect of wheel tread. the temperature of the asphalt in the wheel track could increase by as much as 9°F (5°C) due to the repeated loading by the wheel. These gave radii of contact areas. In the single track tests. it was previously observed that during long continuous runs of the PTF. assuming them to be circular.8 kips (8 kN) was used for the First Test Series.average value. tire wall strength. However. The temperature inside the PTF was maintained at 68 t 3. a wheel load of 1.

The outputs from the thermocouples. position and speed. Air temperature of the PTF was obtained from a thermometer placed inside the facility. number of load cycles). and also near the end of each test series is given in Table 13. Direct comparisons can be made between each test section within a given series. All pressure cells could be recorded continuously. it normally required about 100 to 200 wheel load passes at the center line to obtain a complete set of strain coil readings. The permanent strain results were obtained near the end of the test. Unless indicated. elevations of all the reference points at the surface of the sections along the center line were measured and checked. after relatively large permanent deformations had developed. which measured temperature at selected depths in the pavement structure. In addition. resilient strains and transient stresses were recorded on an Ultra Violet Oscillograph which also recorded wheel load. comparisons can be made between test series if appropriate adjustments are made in observed responses.e.. TEST RESULTS A summary of important measured pavement response variables recorded at both an early stage of loading. or with depth in the pavement structure at a particular time. were monitored regularly by means of a readout device. In addition. Therefore. based on the 60 . During the actual loading. Most of the results presented show either variation of test data with time (i. Vertical resilient strains are given at early stages of the test when the pavement structure was still undamaged. A "peak hold" data acquisition system was later used to record the peak values of the stress and strain pulses. usually only relatively small changes of this variable occurred with time.characteristics under loading. all the results were obtained from multi-track tests. but it was only possible to record one strain coil pair at a time.

o 41:1 • -1 ✓ V •• -....—..11 .. .4 4. 4i 4c .) cra 12 ult II CI:1 13 .. C 0 i ..i...40cE8E2n:5 Tes t Series 1 ta• W •-•4 N N a) L o O I I 1 61 1 . .. ..-t C 0 no ••••t 4... " 1..11 a RI N .4 U) .4 191 JJ 4...1 ti •••■ 0 .1 111 to 4..o 8j . 0 .S:.l ..a) al •--i al V in ..1 tn 0..1 .. .1 4>:I si 0 4 0 0 to 0 . 4.c C ••4 c Nu 04 D 0 in .... ..0 0 0 Is o • —.z.1/1 j '4-1 ( 0 : 5T iii 4..-t .. tv) 'I' 4t •—• 4..I 14 Ca /13) Vol L T..t 14 0 0 4-} 10 1■Ll) 5° 'Era 4:'0.3) Ti. 1.4 8 5 1-1 fa r-I 44 0-1 22 .-t J-4 • •r9 ..-1 t.2 8 ...4 5 C a) Cr) rj 1...1 44 3 CI) 4-1 1...) •-4 X co 0 a) 4-4 u) 4-. g >. T'B T . g a) 'W 4. ..'..... au4 >.1...4 4-1 [) 1.0 0 14 2..! a) In ar w ". .J 0 CL) 411 • 4 . 44 al to 0 0 .-1 to 4.) E 0 ..t. U) g.c 0 0 4) Z6 1al I'4. I 0 V ►4 111 4-I 44 i 151 14 0 0 • 4 >1 •.—. 0 A 0 t•I 0 . • 4..4 .... .: : CI.

permanent vertical deformation.. etc. Permanent Vertical Deformation In this study the permanent vertical surface deformation of the pavement is taken as the primary indicator of performance. however. as shown in Figure 18. the prerutted section in the second series performed best. prerutting does not appear to improve the overall rutting performance of the weak pavement section compared to the geotextile reinforced section which was not prerutted.). This section was reinforced with a 62 .e. Because of the use of a higher quality aggregate base and thicker base and surfacing. The accumulation of surface rutting measured by the profilometer is shown in Figure 18. Permanent vertical deformation in both layers was calculated from the changes in distance between the pairs of strain coils. subgrade stress. were excluded from the averaging process. However.43 in. (11 mm) in the control section. Figure 18 clearly shows that the pavement sections used in the first test series are very weak. Furthermore. The permanent deformation occurring in the base and subgrade are shown in Figures 20 and 21. Profiles showing the permanent deflection basin at the end of the tests are given in Figure 19. Whenever there is more than one value of data available (i.relative behavior of the similar control section in each test series. the life for the pavement sections of the other three series of tests was considerably longer. respectively. with large deformations developing in less than 2000 passes of wheel load. and also in Table 13. These results indicate that the inclusion of a stiff to very stiff geotextile at the bottom of the very weak sand-gravel base reduces the amount of rut by about 44 percent for a rut depth of 0. permanent vertical strain. in contrast to the results of the first test series. Erratic data. an average value was reported in the tables and figures.

0 to 0 •4 0 0 N to rb 0 0 0 '4 63 13 0 by Pro filome ter with the Number o f Passes o f 1. 5 kip s 113 I O mo O Load .All o o o o'o N — (ww) H1430 In CO. 00 0 .

.0 CONTROL GX.1st SERIES =1■■••■■■ 41 0.B PR-GX-B . 04. GD-M PS-GD-M Note : PR = Prerutted GX= Geotextile PS = Prestressed GD= Geogrid M= Middle of Bose M = Middle of Base Figure 19.. Pavement Surface Profiles Measured by Profilometer at End of Tests — All Test Series.8.. 64 .0 2nd SERIES 11..1100 4th SERIES CONTROL GD.5 0i 1.B PR.GD-B GX-M .

J 0 w LU cb Li? 0 i r-ki? o o i o1 6 o 6 c5 u.-.1 ci.) . ti4 .../) re . cr I? b..-- 0 0 it I I X CI CD CD cy aj Ct et 0.) «-1 0. . taup of 0) 4.I .0. .J 0..3 $. in 4..I •t4 0 II Cr c.i.4 0.3 75 a ai E 2_ 0 7:3 .1 PM ".01 4.3 0 z z = 4.. . 1 a) .._..4 1.4 01 03 CO > a.-.-a 0O u_.i t —C a.c •r4 4.. UJ • a) 113 1-. 0 N 1.1..•• 0 34 0.:4 O E 0 •r4 °Lcs .1 ao ao a1 0J .-1 CD I-I Cti $.01 14 •r4 OJ .4 01 3.4 E O 14 o O T. a).J as ac al $. O a1 . 0 tij ID 03 u li X CO 0.0 .1 A ..1 • 0 . Q_ b Z O in 6 sip o 6 6 6 ('u!) NOLLV14180d30 IN3NVIA4 3d 1V01183A . . z •r1 65 .1 Ill X u)0 in . V§ 4..o I .

.0 CU El A *0 Clo W 44 0 a...1 vi:1 11 II 1.. 0 14 .)00 oc.. Q. g ••-4 • co c as o — 0 -8 c: 0 131 E L.. a in . w-Li0 x l'In .1 la • • cb 0 /-- CO 44 • 0 ... E a 1" 0 0 X co C3 it it -.the Subgra de w it h Number o f Passes • Ul O InO Clt 00 in IA g CO° ID b r rb 8 a.-- Z ct.• ••I = go •■•4 • 4.. O ( (11-111 ) in - at) 0 a 0 x t> .t Cr) N I-NOILVIA42:10d30 1V011:13A 1N3NVIA183d 66 al c0 u 0 9-1 I-1 4-I 0 rb "-CD )...1 X CO "-CD W W 0 O -.1 )..4 r••l CU 0 > as (2) CO NND . a0 ..• <4 cu D ai iP .2 N O In O o 0 1 al .IJ cu RI Cr) E )4 . — -0 X.6) 11 11 o a) 0 t44 cU V") X %0 < C7D C. 0 rb 0 et (1) CL CL w... .1-Acox 28 )) mc.13 Irb or -o in 0 112 7.1 1.

. Only an 8 percent reduction in rutting was observed for the geogrid reinforced section used in Test Series 2 which was not prerutted (Figure 18b). since similar very good performance was also observed for prerutted sections which were not reinforced. most of this improvement occurred within the aggregate layer (Table 13. Therefore. woven geotextile when both were placed at the middle of the granular layer (Figure 18d). Further. use of geogrid reinforcement. When the location of the geotextile was raised to the middle of the aggregate base in Test Series 3. Results from the last series of tests indicate that prestressing the geosynthetic appears to improve performance compared with a non-prestressed section having the same geogrid reinforcement (Table 13. This section was not prerutted. Thus. 280 kN/m) resulted in better performance than a higher stiffness. the amount of rutting was reduced by a total of 28 percent.geogrid at the bottom of the base and resulted in a 66 percent reduction in total rutting of the base and subgrade. Figure 18d).. This finding by itself is misleading. Figure 18c). 750 kN/m) located at the bottom of the layer (Table 13. despite its lower stiffness (S g = 1600 lbs/in. as will be discussed subsequently for the single test track results. A large portion of the total permanent deformation occurred within the aggregate base. prerutting of the reinforced section was quite effective. Figure 18c). it follows that the pattern of permanent deformation as a function of load repetitions observed in the base was very similar to that observed at the pavement surface as can be seen by comparing 67 . A similar relatively low level of improvement with respect to rutting (13 percent reduction) was observed for the section in Test Series 3 reinforced with a stiff to very stiff geosynthetic (S g = 4300 lbs/in.

68 . a decrease of strain is generally observed near the top of the subgrade. the pattern of results is very similar for all test series. Reductions in subgrade rutting of 25 to 57 percent were observed for this condition.Figure 18 with Figure 20. The trend in the development of total permanent deformation in all 12 sections of the four test series in the multi-track loading tests was generally confirmed by the single track studies (Figure 22). a redistribution of vertical permanent strain is seen to occur due to the presence of the reinforcement. Permanent Vertical Strain The variation of permanent vertical strain with depth for all the sections at the end of testing is shown in Figure 23. decreasing rapidly with depth towards the subgrade. Other interesting results that can be obtained from these figures reveal the following differences between pavement sections: 1. At the same time (with the exception of the first series results). In general. an increase in permanent strain occurred in the top half of the granular base. as shown in Table 13 and Figure 21. The average values of strain are plotted at the mid-point between the two strain coils which measure the corresponding vertical movement. An important reduction in subgrade deformation was evident when a geosynthetic was placed directly on top of the subgrade. For sections with the geosynthetic reinforcement placed at the bottom of the granular base. with large permanent strain at the top of the granular base. Permanent vertical deformation in the subgrade was relatively small compared to that occurring in the base. When comparing results from the geosynthetic reinforced and control sections. particularly for the prerutted sections.

ci.—.J k Ce W ig Z 4. _ ("I .. 4:00 7.. Ct 0 G) I 0 11 IS w -.-I i Q. E-4 •.5d O U co O E-I .1 1.— CD •-"Eff 0 la. 4.1 sb . CO X ..4 . a) fb $ CI9 tE 6 15 `7:.-4 a) 2 0 tr).1 ELI "ID 0 X Et• cb 0 'Crl NI A 0 0 0 O r1 0 ciD T c? i Cg z d V P 0 EL 0 w c LU .) a) 0 .) a..7.1 0.L. 0 .) cu Figure 22. 8 al I-I $4 E g 0 4.41:2 w az 0 t. .0 .1 0 X :3..0 4-1 4f.3 si .. m2— (. I i 1? 00 x x 01 .-1 1.0 4. IP Cu!) NO11VV480330 30VdTifIS IN3NVIAl2:33 cd 69 22 $.-1 co = $4 .4 a.. al E C:) 2 g 144 0 O a) 0 to 03 114 4-4 cl ./1 cc u) a. Ei get W 6 L. . LL cu E-4 ID 0 C.± C2 0 . 0 Lk 0 k 44 0 Cil v-4 . (1) `1--1 (..--f •. MUOL.1 rn 0 ti..-- co O .1 • vi-1 0 • o co •...

I w 100 100 BASE tcj 200 200 300 300 400 500 c)3 rd SERIES c3 CONTROL x Gx-B v GX-M 400 SUE3GRADE d)4th SERIES o GX-M x GD-M ✓PS-GO-M 500 Note : PR = Prerutted GX= Geotextile M= Middle of Base B = Bottom of Base PS = Prestressed GD= Geogrid Figure 23. Variation of Vertical Permanent Strain with Depth of Pavement for All Four Test Series. BASE 200 cr u. C.- z 11/ b) 2nd SERIES o PR-GD-B 400 ■ GD-B 7 CONTROL 500 w 0 10 15 20 0 5 15 10 A.C.300 20 100 100 w 15 B 1/77 SUBGRADE 300 0)1st SERIES 0 PR-GX-B x CONTROL 7 GX-B 1. 70 .0 O VERTICAL PERMANENT STRAIN 1%) 5 10 15 20 25 0 10 5 0 FIkA 200 a A.

however. The vertical permanent strains for the two prerutted sections are in general smaller than those in the nonprerutted sections with or without reinforcement. 4. 71 . Prestressing of the geogrid appears to reduce the development of permanent vertical strain in both the granular base and the subgrade layer. while permanent strain at the top of the subgrade increased. Figure 23 shows that as a result of placing the geotextile at the middle of the aggregate base. As a result.2. as shown in Figures 23a and 23b. The only exception is the permanent strain developed within the prerutted sandgravel base which shows a greater value than its nonprerutted counterparts. the 100 to 200 passes of wheel load required to complete the recording procedure did not have a significant influence on the consistency of the results. The results for the first series of tests are considered unreliable because the pavement structure deteriorated rapidly at quite an early stage of the experiment. uniform conditions across all the three sections could not be maintained while the resilient response of all the sections was being measured. a substantial decrease in permanent vertical strain occurs immediately below the geotextile. Vertical Resilient Strain The variations of vertical resilient strain with depth for all the pavement sections are shown in Figure 24. it is believed that the recorded strains shown in Figure 24a at least show the correct trends. 3. For other series of tests. Nevertheless.

Variation of Vertical Resilient Strain with Depth of Pavement for All Test Series.VERTICAL RESILIENT STRAIN (milli strain) o° HRA 5 i 100 BASE 10 15 oust SERIES PR-GX-B • CONTROL GX-B 0 2 4 6 10 200 E 300 LY. 0 01 100 L 2 AC 4 6 8 [E13rd SERIES CCNTROL BASE 2 4 6 8 100 cc OO 200 t 300 300 400 400 GX-B 500 Note : PR = Prerutted GX= Geotextile M = Middle of Base PS = Prestressed GD= Geogrid B = Bottom of ease Figure 24. 72 .

No consistent trend emerged regarding the effect of geosynthetic stiffness and location of the reinforcement on the measured resilient lateral strain. The lateral resilient strains recorded during the 4 test series are shown in Tables 14 and 15. In general. Beginning with the third test series they were also measured in two of the three sections at both the top 73 . regardless of its location within the pavement structure. but that in the nonreinforced control section tends to be considerably higher. for a given test series the magnitude of the resilient lateral strain in the geosynthetic reinforcement of both sections is quite similar. Longitudinal resilient strains at the bottom of the asphalt surfacing were measured for all the sections. The non-reinforced control sections (with the exception of the first series of tests) normally exhibited slightly higher resilient strains than the reinforced sections. respectively. In general. Both prestressing and prerutting appear to reduce significantly the resilient strain at the top of the subgrade. overall resilient response of the pavement sections does not seem to be significantly influenced by the geosynthetic reinforcement. large strains were obtained at the top of both the aggregate base and subgrade. Lateral Resilient Strain Lateral resilient strains were only recorded from the strain coils installed on the geosynthetics and in the complimentary location of the control sections. a similar magnitude of strain. Longitudinal Resilient Strain The results of the resilient longitudinal strain for the asphalt surfacing and the aggregate base are shown in Tables 14-15 and Figure 25. within one series of tests. However.Figure 24 shows that the resilient strain profile for all the sections has a similar shape and.

GX-B 2 250 40000 PR-GD-B CONTROL GD-B 1585 3130 2616 PR-GD-B CONTROL 00-8 1730 3410 2852 / 2047 / 3725 3860 4121 GC= Geotextile M2 Middle of Base Note: • PR= Prerutted B= Bottom of base VS= Prestressed GC= Geogrid *• In the control sections. Middle of Base Note: • PR= Prerutted . the measured strain is that of the soil. 74 .Test Series 3 and 4.Test Series 1 and 2. / 3 4 400 GX41 01/11436 GXHM 1413 6871 2103 70000 CC-B CONTROL GX44 1609 4765 2242 400 G7NM GD-M PS-GD-M 2550 1500 1500 46000 GX-M GD-M PS-GD-M 1650 1800 2050 2355 2983 2198 2800 / 1800 GX= Geotextile M. Lmngitudinal Resilent Strain in Geosynthetic** (mm) at bottom of asphalt (pc) r 1675 ccurRor.Table 14 Summary of Lateral Resilient Strain in Geosynthetics and Longitudinal Resilient Strain at Bottom of Asphalt . Table 15 Summary of Lateral Resilient Strain in Geosynthetics and Longitudinal Resilient Strain at Bottom of Asphalt . of Passes Section Designations 1 50 PR-CDC-B COMM G)C-8 1480 4740 1200 PR-GX-B 2317 11340 2561 Lateral Resilient Strain . Test Series No.PS= Prestressed GO ■ Geogrid Be Bottom of base In the control sections. the measured strain is that of the soil.

Third and Fourth Series.**.4tB.. For section designationPS = Prestressed GX=Geotextile GD= Geogrid M. TOP CONTROL.. TOP .a) 3rd SERIES 6 cr-CONTROL TOP GX.0' 102 "'".. 75 . BOTTOM = strain measured at top. BOTTOM 103 104 NUMBER OF PASSES OF WHEEL LOAD 105 Note: 1.B = Geosynthetics placed at middle.. bottom of base 2 For location of strain measurementTOP..M.....0 - 0.. BOTTON x GX..e CFr-GD-M. BOTTOM 0 103 H I— 104 b)4th SERISL/GitM.1.. Variation of Longitudinal Resilient Strain at Top and Bottom of Granular Base with Number of Passes of 1. 105 ..77----v/- E o z 31. 0..3.0 tri w cc z 2. BOTTOM Ps-co-m.M.5 kip Wheel Load .bottom of base Figure 25.

the subgrade stress rapidly increased as the pavement developed large permanent deformations early in the experiment. This could be at least partly due to the slight differences in the finished thickness of the surfacing and base and small differences in material properties. shown in Figure 27. The results. indicate that the horizontal stress at the top of the granular layer increased throughout each test. horizontal transient stress (in the direction of wheel traffic) at both the top and bottom of the aggregate base was measured in the third and fourth test series. Transient stress is that change in stress caused by the moving wheel load. For resilient longitudinal strains measured within the aggregate base. Only resilient strains at the beginning of the test are shown in Tables 14 and 15. Longitudinal strain at the bottom of the asphalt surfacing also varied from one series of tests to another. with the magnitude of vertical stress typically varying from about 6 to 9 psi (42 to 63 kN/m2 ). the longitudinal resilient strain varied greatly throughout the test. A consistent influence of geosynthetic reinforcement on vertical subgrade stress was not observed in any of the test series. Transient Stresses The variation of transient vertical stress at the top of the subgrade during each test for all the pavement sections is shown in Figure 26. Generally longitudinal resilient strain increased in the top and bottom of the aggregate base as the pavement started to deteriorate. Figure 27a also suggests that the inclusion of 76 . however.and bottom of the aggregate layer. Longitudinal. The subgrade stress for the last three test series remained reasonably constant throughout the test. Unlike the vertical resilient strain. there did not appear to be a consistent development trend. For the first series of tests.

) E-' o .-4 ..I 4..0 CO .1 .44 it 0 ..._ . . 4.1 ce ei TO Z CI.S 30.2 O o E-1 CI 0 II 11 a... 0 E V) CD W ti) 11 0 on 0 N < U. z 0 0 (NI a) 4. > 4. . 0:1 •ri > .▪ 44 a) az 0 0 Z O r3 21 •ri 3 0 0 -0 aE 75 o cdI.. O. 0 • 8 CO ca .0tt 0 X CO 4.4 ( zwitsin SS38. tV 77 .. 0 I1V31183A 0 to ID .(8091S O 0 0 cr.M rs.-I *I 34 O.-1 al 44 u a) 4.. . 0 a) s. ... .4 0 < O —J It W W = 4- • • a O O 0 C•4 O 0 ft 03 12 O 0 eti 0 1" C U.1 II '10 W ta 0 1 IA "C/ E-1 0 0 44-1 34 0 z0 W(11 Ti § Ca •.4 0 o 4.1 • co o et V) O 0 .. a o.I.

comamg GX-M. bottom of base Figure 27.TOP //. 78 . BOTTOM 20 0 102 lo' 105 106 NUMBER OF PASSES OF WHEEL LOAD Note :1. bottom of base 2.5 kips Wheel Loads . Variation of Transient Longitudinal Stress at Top and Bottom of Granular Base with Number of Passes of 1. For section designationPS= Prestressed GX= Geotextile GD=Geogrid M.G D-M.TOP z 100 _x 44) cr 50 CONTROL.E30170M +---+ G M.a) 3rd SERIES 150 CONTROL.Third and Fourth Series.TOP 5 BO ps.BOTTOM z w (7) 0 102 g 104 b) 4th SERIES /GX-M.GX-M. BOTTOM —+ PS.BOTTOM =Stress measured at top. For location of stress measurement TOP.B= Geosynthetics placed at middle.

The horizontal stress at the bottom of the aggregate base. single track tests were then performed along the side of the test pavements. Single Track Supplementary Tests After performing the multiple track tests in Test Series 2 through 4. did not appear to be influenced by the progress of the test. 79 .geosynthetic reinforcement at the middle of the aggregate base may result in a slower rate of increase in horizontal stress at the top of the layer. This test series was conducted during the excavation of test series 2 pavement after the surfacing was removed. For these tests the permanent vertical deformation in the two reinforced sections and the unreinforced control section were all very similar. on the other hand. These special supplementary tests contributed important additional pavement response information for very little additional effort. which are valid for the conditions existing in these tests. and the results of these tests are presented in Figure 28. The single track tests performed are described in Table 12. permanent deflections in the reinforced sections were actually slightly greater throughout most of the test. Placement of a geogrid at the bottom of the aggregate base did not have any beneficial influence on the performance of the unsurfaced pavement in Test Series 2 (Figure 28a). These tests were conducted where wheel loads had not been previously applied during the multiple track tests. The single track tests consisted of passing the moving wheel load back and forth in a single wheel path. nor by the presence of a geosynthetic at the center of the layer. can be drawn from these experimental findings: 1. The following observations.

Depth of soil contamination of the base was found to be in the range of 1 to 1. In modeling a reinforced aggregate base. heaving along the edge was evident for the three sections of Test Series 1 using the sand-gravel base. upward soil migration appeared to be the dominant mechanism of contamination.However. When a geogrid was placed on the subgrade. The resilient modulus of the subgrade was found to very rapidly increase with 83 . A finite element model having a crossanisotropic aggregate base was found to give a slightly better prediction of tensile strain and other response variables than a nonlinear finite element model having an isotropic base. SUMMARY AND CONCLUSIONS Both large-scale laboratory tests and an analytical sensitivity study were performed to evaluate the performance of surfaced pavements having geosynthetic reinforcement within the unstabilized aggregate base.5 in. the accurate prediction of tensile strain in the bottom of the base was found to be very important. Larger strains cause greater forces in the geosynthetic and more effective reinforcement performance. Hence. the elastic cross-anisotropic model was used as the primary analysis method in the sensitivity study. Extensive measurements of pavement response from this study and also a previous one were used to select the most appropriate analytical model for use in the sensitivity study. Contamination of the aggregate base by the silty clay subgrade was evident in most sections except those where a geotextile was placed directly on top of the subgrade. Contamination occurred as a result of both stone penetration into the subgrade and the subgrade soil migrating upward into the base. (25 to 38 mm). Soil Contamination.

Hence. 84 . The low resilient modulus existing at the top of the subgrade causes a relatively large tensile strain in the bottom of the aggregate base. show that placing a geosynthetic reinforcement within the base of a surfaced pavement has a very small effect on the measured resilient response of the pavement.depth. under the proper conditions. a geogrid will provide considerably better reinforcement than a woven geotextile. Reinforcement can. Both the laboratory and analytical studies. as well as full-scale field measurements. cause changes in radial and vertical stress in the base and upper part of the subgrade that can reduce permanent deformations and to a lessor degree fatigue in the asphalt surfacing. field testing methods that measure stiffness such as the Falling Weight Deflectometer tend not to be effective for evaluating the potential improvement due to reinforcement. The experimental results show that for a given stiffness.

Therefore. it must be sufficiently durable to serve its intended function for the design life of the facility. These aspects are considered in Appendices F and G. because of its great importance. analyses can be performed on pavement structures of this type using theoretical approaches similar to those employed for non-reinforced pavements but adapted to the problem of reinforcement. For reinforcement to be effective. a more sophisticated but time consuming nonlinear finite element analysis 86 . The important advantage of using a simplified linear elastic model of this type is the relative ease with which an analysis can be performed of a pavement structure. cross-anisotropic finite element formulation can be successfully used to model geosynthetic reinforcement of a pavement structure. Both the separation and filtration mechanisms of geosynthetics are considered as a part of the general synthesis of the use of geosynthetics within aggregate base layers. a linear elastic. GEOSYN7EITIC REINFORCEMENT The response of a surfaced pavement having an aggregate base reinforced with a geosynthetic is a complicated engineering mechanics problem. As will be demonstrated subsequently. the use of steel reinforcement in the base has been introduced as an alternative to geosynthetics as the present project was being carried out. Where a higher degree of modeling accuracy is required. existing literature was heavily relied upon for this portion of the study. the present state-of-the-art of durability aspects are considered and put in perspective. However. These mechanisms are of considerable value because of the many new innovations in reinforcement that will have to be evaluated in the future. For example.understanding of the fundamental mechanisms of geosynthetic reinforcement.

angular aggregate. a very minor role in determining reinforcement effectiveness of a geosynthetic. The wide width tension test as specified by ASTM Test Method D-4595 is the most suitable test at the present time to evaluate stiffness. The ultimate strength of a geosynthetic plays. The stiffness of a relatively thin geotextile can be determined in the laboratory by a uniaxial extension test. Note that ASTM Test Method D-4595 uses the term "modulus" rather than stiffness S gwhich is used throughout this 87 . Use of a finite element analysis gives reasonable accuracy in modeling a number of important aspects of the problem including slack in the geosynthetic. evaluating potential benefits of reinforcing an aggregate In base the first step should be to establish the stiffness of the geosynthetic to be used. the geosynthetic strength and ductility are important factors when it is used as a filter layer between a soft subgrade and an open-graded drainage layer consisting of large. Under certain conditions it is an important consideration in insuring the success of an installation. This does not imply that the strength of the geosynthetic is not of concern. accumulation of permanent deformation and the effect of prestressing the geosynthetic. at most. Geosynthetic stiffness should be used since the modulus of elasticity of a thin geosynthetic has relatively little meaning unless its thickness is taken into consideration.was employed in the study. GEOSYNTHETIC STIFFNESS The stiffness of the geosynthetic is the most important variable associated with base reinforcement that can be readily controlled. as will be discussed later. Geosynthetic stiffness S g as defined here is equivalent to the modulus of elasticity of the geosynthetic times its average thickness. For example. slip between the geosynthetic and the surrounding material.

For low deformation pavement reinforcement applications. Use of the grab type tension test to evaluate geotextile stiffness is not recommended. the geosynthetic should in general have a stiffness exceeding 1500 lbs/in. both the ASTM "modulus" and the stiffness as used here have the same physical meaning. and hence is not suitable for use as a reinforcement. Classification System.30 to $0. A geosynthetic classification based on stiffness for reinforcement of aggregate bases is shown in Table 16. As discussed later. Army Corps of Engineers in reinforcement specifications.59/m2 ). A strain level of 5 percent has gained some degree of acceptance. This table includes typical ranges of other properties and also approximate 1988 cost. 88 . a low stiffness geosynthetic does not have the ability to cause any significant change in stress or strain within the pavement. A very low stiffness geosynthetic has a secant modulus at 5 percent strain of less than 800 lb/in. The secant geosynthetic stiffness S g is defined in Figure 30 as the uniformly applied axial stretching force F (per unit width of the geosynthetic) divided by the resulting axial strain in the geosynthetic. the stiffness of the geosynthetic must be presented for a specific value of strain. Since many geosynthetics give a quite nonlinear load-deformation response. for low deformation conditions. (260 kN/m).S.study. This value of strain has been employed for example by the U. Use of a 5 percent strain level is generally conservative for flexible pavement reinforcement applications that involve low permanent deformations.50/yd 2 (0. Several selected geosynthetic stress-strain curves are shown in Figure 31 for comparison.360. For most but not all geosynthetics the stiffness decreases as the strain level increases. (140 kN/m) and costs about $0.

0 GEOSYNTHETIC STIFFNESS: S • FiLBS/INVSTRAIN 9 STRAIN • LIAO Figure 30. SOO 400 z S 300 g 200 100 0 0.08 0. Basic Idealized Definitions of Geosynthetic Stiffness. Figure 31.AXIAL FORCE F (LBS/IN) I I1 I I 1 I I 1 I I I I AXIAL FORCE FIL13/1N) Lo 5 0 AXIAL STRAIN IN PERCENT. 89 .04 0.12 0.24 GEOSYNTHETIC STRAIN.00 0.18 0.20 0. Selected Geosynthetic Stress-Strain Relationships.

90 . The properties given in addition to stiffness are typical ranges of manufacturers properties and do not indicate a material specification.50 Low 800-1500 15-50 60-200 10-60 0.) Failure Elongation (2 Initial Length) Typical Cost Range ($/yd 2 ) Very Low < 800 10-30 50-150 10-100 0. s8 ../in.40-0.00 NOTES: 1.00 4000-6500 > 300 Stiff /erg Stiff _ 350-500 (or more) 5-15 $3.30-0./in.50-3.Table 16 Tentative Stiffness Classification of Geosynthetic for Base Reinforcement of Surfaced Pavements( 1 ) Stiffness Description Secant Stiffness 8 52 Strain.00-$7. Tensile Strength (lbs.50 1500-4000 20-400 85-1000 10-35 0.) Elastic Limit (lbs. (lbsain.

5 to 3. Appendix C). however. A reinforced aggregate base presents an even more challenging problem. A surfaced flexible pavement of low to moderate structural strength (AASHTO structural number SN =:= 2. Use of a subgrade where the resilient modulus increases significantly with depth greatly increases calculated tensile strains in the aggregate base and shows much better agreement with observed pavement response (Table C-1). Measured vertical and horizontal strains from two well-instrumented laboratory studies described in Chapter II and Appendix C clearly indicate the aggregate base exhibits much higher stiffness in the vertical direction than in the horizontal direction. The many problems associated with modeling the behavior of a non-reinforced aggregate base which can take only tension are well known [16. and the model became progressively more crossanisotropic with depth (refer to Tables C-2 and C-4. The best agreement with observed response was found for a crossanisotropic model where vertical stiffness of the base became about 40 percent smaller in going from the upper one-third to the lower one-third of the aggregate base. This was true for either the cross-anisotropic model or the nonlinear 91 . Appendix C). These results can only be explained if the aggregate base behaves as a cross-anisotropic solid. Cross-Anisotropic Model.48.44.0) resting on a soft subgrade (CBR = 3 percent). As a result. a linear elastic. develops relatively low tensile strain in the aggregate base and hence low geosynthetic forces.49]. Changes in response of the pavement are for the most part determined by the tensile strain developed in the geosynthetic.REINFORCEMENT MODELING Modeling. cross-anisotropic finite element model appears to give the best overall predictions of pavement response (Tables C-1 and C3.

As expected. Use of a simplified contour model for aggregate bases [51. The rigid layer. the resilient subgrade modulus near the surface appeared to be about 10 and 20 percent. which was located below the subgrade in the instrumented pavement studies. may have had some influence on performance. the nonlinear analysis underpredicted vertical strain in the subgrade. the resilient modulus of the soft silty clay subgrade apparently did not increase as much as that of the micaceous silty sand subgrade. aggregate base quality and permanent deformation. of the average resilient subgrade modulus as shown in Figure 32. and also the other commonly used response parameters. isotropic finite element model which was used can. The nonlinear resilient modulus was therefore adjusted to approximately agree with the variation of modulus with depth shown in Figure 32. respectively. For the micaceous silty sand and silty clay subgrades used in the two validation studies. A nonlinear isotropic model was used in the sensitivity study primarily to investigate the effect of special variables such as geosynthetic slip. however. The isotropic nonlinear analysis cannot.52] appeared to give better results than the often used K-9 type of model. predict at the same time both the large tensile strain measured in the bottom of the aggregate base and the small measured vertical resilient strain observed throughout the aggregate layer. 92 . upon proper selection of material parameters. A discussion of the increase in resilient modulus with depth has been given by Brown and Dawson [50]. When the nonlinear properties originally selected for the subgrade were employed. Nonlinear Isotropic Model.finite element models. The nonlinear. but should not have been a dominant factor. predict reasonably well the tensile strain in the aggregate base.

0 10 12 14 BASE THICKNESS.C. 30 2.C.SUBGRADEDEPTH. T (INCHES) Figure 33. E3 (KSI) 3E4 Figure 32.5 IN A. 93 16 . 2 ( I N) 0 —10 —20 —30 -40 —50 0 3E4 2E4 5000 1E4 2E4 RESIDENT SUBGRADE MODULUS. Variation of Subgrade Resilient Modulus With Depth Estimated From Test Results. Es=3.5 KSI 20 Er BOTTOM BASE ev SUBGRADE dv SURFACE Er A. Reduction in Response Variable As A Fuction of Base Thickness.

[21. et al. The analytical results shown in Figure 33 (and also in Tables 2 through 4 of Chapter II) indicate radial strain in the asphalt surfacing and surface deflection are generally changed by less than 5 percent. results in relatively small changes in the resilient response of the pavement. (700 kN/m). Therefore.000 equivalent 18 kip (80 kN) single axle loads.30] confirm this finding. Considerable progress was made in this study in developing appropriate techniques to model both reinforced and nonreinforced aggregate bases. Field measurements by Ruddock. resilient strains and displacements caused by the applied loadings. This level of change applies even for relatively light structural sections placed on a soft subgrade and reinforced with a very stiff geosynthetic having Sg = 4000 lbs/in. and vertical subgrade strain by less. The cross-anisotropic model appears to give slightly better results and was more economical to use. Specific benefits resulting from reinforcement using these criteria are discussed later. 94 . Even though the changes in response are relatively small. than 10 percent when the geosynthetic is present. IMPROVEMENT MECHANISMS The analytical and experimental results show that placement of high stiffness geosynthetic in the aggregate base of a surfaced pavement designed for more than about 200. simplified contour model. Reasonably good response was obtained using both the linear cross-anisotropic model and the nonlinear. it was the primary method of analyses employed in the sensitivity study.Summary. some modest improvement can usually be derived from reinforcement following the commonly employed design approaches of limiting vertical subgrade strain and radial tensile strain in the asphalt. Pavement response is defined in terms of the transient stresses.

Variations in pavement thickness and/or material quality including subgrade stiffness could account for the difference in overall pavement stiffness observed for the one series of tests in Texas. even when a very stiff geosynthetic is used as reinforcement. The experimental measurements show the strain in the geosynthetic to be about 50 percent of the corresponding 95 .Pavement Stiffness The structural strength of a pavement section is frequently evaluated using the Falling Weight Deflectometer (FWD) or Dynaflect devices.38. The overall stiffness of a structural section can be defined as the force applied from a loading device. The laboratory test results also indicate no observable improvement in pavement stiffness. Both the laboratory and analytical results indicate the change in radial stress and strain as a result of base reinforcement to probably be the most important single factor contributing to improved pavement performance. Dynaflect measurements in Texas described by Scullion and Chou [53] showed one section to be stiffened when a geosynthetic was added. such as the FWD. while another indicated no observable difference.30. divided by the resulting deflection. These devices measure the deflection basin from which the overall stiffness of the pavement and of its constituent layers can be determined [49]. too small to be reliably measured in either a full-scale or laboratory pavement. The analytical results of this study indicate the overall increase in stiffness of the pavement will be less than about three percent.39]. These findings therefore indicate stiffness is a poor indicator of the potential benefit of geosynthetic reinforcement on performance. The results of several field studies also tend to substantiate this finding [21. therefore. The improvement in stiffness resulting from geosynthetic reinforcement is. Radial Stress and Strain.

The analytical studies performed on stronger sections indicate changes in radial strain in the bottom of the base to be about 4 to 20 percent for sections having low to moderate structural numbers. Secondly. the radial stresses. does not give the full picture of the potential beneficial effect of reinforcement. the beneficial effects of changes in radial stress caused by reinforcement are reduced but not eliminated. the strain in the geosynthetic becomes greater.5 to 1. The initial stress in the base due to body weight and compaction is likely to be at least twice as large as the radial stress caused by the external loading. the changes in radial stress usually become less than about 0. however.1 psi (0.7 kN/m 2 ) as shown in Table 3. Consequently. As the resilient modulus of the subgrade and the ratio between the base modulus and subgrade modulus decreases.strain in a non-reinforced aggregate base (Table 15). must be superimposed upon the initial stresses resulting from body weight and compaction effects as illustrated in Figure 35. First. Recall that tension is positive so the decrease in stress shown in Figure 34 actually means an increase in confinement. typically being less than about 0. including the relatively small changes resulting from reinforcement. however. 96 . As the pavement section becomes moderately strong (structural number SN = 4. the actual value of change in radial stress is relatively small.5). Changes in radial stress determined from the analytical study typically vary from about 10 percent to more than 100 percent of the corresponding radial stress developed in an unreinforced section (Figure 34).0 psi (3-7 kN/m 2 ) for relatively light sections. Considering just the large percent change in radial stress. As a result improvement also becomes more pronounced.

Variation of Radial Stress in Base and Subgrade With Base Thickness. 0 —1.5 g a. T (INCHES) Figure 34. U1 rt 0. Superposition of Initial Stress and Stress Change Due to Loading. 97 . Wheel Load BODY STRESS I 71 I C/ h h Oh Wheel Load 2 • 0 = ? = kOv z Of (a) Weight and Compaction Stress (b) Wheel Load Stresses (c) Combined Stress Figure 35.5 Z o b Z Cr.5 0.0 C) 5g Li —0.0 8 10 16 14 12 BASE THICKNESS.1.

the top of the subgrade. As a result. By far the largest beneficial effects are realized when the stress state is close to failure on an element of material in. for example. (150-180 mm) below the surface of the subgrade. were only changed to a depth of about 6 to 7 in. 98 . very important reductions in permanent deformation may occur as illustrated in Figure 36. The reduction in permanent deformation becomes disproportionately larger as the stress state in the top of the subgrade (or bottom of the base) moves closer to failure. If the section is weak and hence the initial stress state is near failure. as the stress state becomes less severe. the depth of reduction in permanent strain in the subgrade was about 12 in. The addition of reinforcement under the proper conditions causes a small but potentially important increase in compressive radial stress and a slight reduction in vertical stress. For the heavy load used in the analytical study. The large scale laboratory tests indicate both resilient and permanent strains in the subgrade. (300 mm). Because of the highly nonlinear stress-permanent strain response of the subgrade or base (Figure 36). Depth of Subgrade Improvement. The tire loading in this case. the beneficial effect of reinforcement becomes significantly less. was relatively light. The small beneficial changes in radial stress due to reinforcement can have important effects on permanent deformation under the proper conditions. a small increase in compressive confining pressure and decrease in deviator stress can lead to a significant reduction in permanent deformation when the element of material is near failure. when reduced. the deviator stress on an element of subgrade soil is decreased slightly.Permanent Deformation. When examining Figure 36 remember that permanent deformation is proportional to the permanent strain developed in a thin sublayer of material. however. Conversely.

DEVIATORSTR ESS. Reduction in Permanent Deformation Due to Geosynthetic for Soil Near Failure. ep Note: Acr =change in radial stress due to reinforcement r Figure 36. 99 . . PERMANENT STRAIN. a1 -a3 a'3 +Ar NO REINFORCEMENT MP m •• •mallp• 0'3 REINFORCED a t a a I I a a T1 .

In this comparison a geotextile reinforcement was located in the middle of the base. With increasing numbers of repetitions. eventually. Summary The effects of geosynthetic reinforcement on stress. et al. The changes in radial stresses due to reinforcement appear to be caused by the reduction in tensile strain in the lower part of the aggregate base. As a result. Tensile Strain Variation with Load Repetitions.Findings by Barksdale.000 equivalent 18 kip (80 kN) single axle loads. the difference in tensile strain resulting from reinforcement appeared to disappear and. strain and deflection are all relatively small for pavements designed to carry more than about 200.5 to 3) on a weak subgrade (CBR < 3 100 . geosynthetic reinforcement of an aggregate base will have relatively little effect on overall pavement stiffness. [16] on unsurfaced pavements tend to verify that the depth of improvement in the subgrade due to reinforcement is relatively shallow. The greatest beneficial effect of reinforcement appears to be due to changes in radial stress and strain together with small reductions of vertical stress in the aggregate base and on top of the subgrade. Strain measurements made in the third test series of the experimental study show a very large reduction in tensile strain in the bottom of the aggregate base due to reinforcement at low load repetitions. The increase in confining pressure caused by the geosynthetic would make the upper portion of the subgrade more resistant to liquefaction. A modest improvement in fatigue life can be gained from reinforcement as discussed subsequently. the tensile strain in the nonreinforced sections was less than in the reinforced one. Reinforcement of a thin pavement (SN = 2. however.

Geosynthetic reinforcement levels included in the analytical sensitivity study varied from low to high stiffness (S g = 1000 to 6000 lbs/in. and a number of practical aspects are examined such as slack and slip of the geosynthetic. A reduction in base thickness due to reinforcement was established by requiring the reinforced section to have the same tensile strain in the bottom of the asphalt surfacing as the non-reinforced section. Alternate thicknesses are given from the analytical sensitivity study for subgrade strengths varying from a resilient modulus of 3500 psi (24 kN/m2 ) to 12. As the strength of the pavement section increases and/or the materials become stronger. 170-1000 kN/m). the states of stress in the aggregate base and the subgrade move away from failure. the improvement caused by reinforcement would be expected to rapidly become small.percent) can potentially reduce the permanent deformations in the subgrade and/or the aggregate base by significant amounts. As a result. the reduction in aggregate base thickness as a result of geosynthetic reinforcement was determined using an equal strain approach for controlling fatigue and rutting. In the analytical sensitivity study. REINFORCEMENT EFFECTS In this section the primary factors associated with aggregate base reinforcement are discussed including their interaction with each other and the overall pavement response.. Effects of reinforcement on permanent deformations that might occur in the base are also considered. A similar procedure was employed to determine the reduction in base thickness for equal vertical 101 . This range of subgrade stiffness approximately corresponds to a variation of CBR from 3 to 10 percent. The influence of reinforcement on the required pavement thickness was studied considering both fatigue and permanent deformation (rutting) mechanisms.500 psi (86 MN/m 2 ).

This test series was terminated when the total permanent deformation reached about 1 in. the stiff to very stiff geotextile caused a 57 percent reduction in permanent deformation in the subgrade but only a 3 percent reduction of permanent deformation in the aggregate base (Table 13). When placed in the bottom of the aggregate base. The layer strain method and the permanent strain materials properties employed in the analysis are described in Appendix C. A total of 100. Optimum Geosynthetic Position The laboratory pavement tests together with the results of the analytical sensitivity study can be used to establish the optimum positions for placement of geosynthetic reinforcement within an aggregate base.Experimental Findings.5 kip (6.7 kN) wheel. and an 8 in. (25 mm).2 in. when the same geotextile was placed in the middle of the aggregate base. permanent deformation in the base was reduced by 31 percent.strain near the top of the subgrade. (200 mm) crushed limestone base. Subgrade permanent deformations. were reduced by only 14 percent. however. A control section without reinforcement was also present.. The geotextile was placed at the bottom of the base in one section and at the center of the base in another section. 102 . (30 mm) thick. A stiff to very stiff woven geotextile was used (S g = 4300 lb/in. Test Series 3 was constructed using a stiff asphalt surfacing mix 1. An estimate of reduction in rutting in the aggregate base and subgrade was also made using the layer strain method. In contrast.070 load repetitions were applied by a 1. The experimental findings of Test Series 3 demonstrate the effect of geosynthetic position on performance with respect to permanent deformation. Permanent Deformation . 750 kN/m).

when placed in the bottom of the base. A 52 percent reduction in permanent subgrade deformation was observed in this test series. 24 MN/m 2 ) are summarized in Figures 37 and 38. Geotextile reinforcement positions were at the bottom and center of the layer.5 in. A geogrid. expressed as a decimal. An analytical study was also performed to establish the effect of geosynthetic position on the reduction in rutting in the base and subgrade (Tables 17 and 18). Conversely. The actual reduction in base thickness is equal to the base thickness without reinforcement indicated in the table or figure multiplied by the percent reduction. Permanent Deformation . the reduction of permanent deformation in the base becomes much greater as the reinforcement is moved upward in the base (Figure 38).The results of Test Series 2 also tend to verify these findings. In Figures 37 and 38 the solid symbols indicate observed reductions in rutting from the previously described Test Series 3 experiment. did not decrease the permanent deformation in the base (measurements suggested an increase of 5 percent). The results of this analytical study for the standard reference section having a 2. Agreement between the observed and calculated reductions in rutting is reasonably good.Analytical Results. Material 103 . The reduction in subgrade deformation gradually goes from about 45 percent to 10 percent as the geosynthetic location moves from the bottom of the base to a location 2/3 up from the bottom. (64 mm) thick asphalt surfacing and a relatively soft subgrade (E s = 3500 psi. The maximum measured reductions are greater than calculated values for similar pavement base thicknesses. Improvements due to reinforcement in terms of a reduction in base thickness are apparent from the data in Tables 17 and 18 and other tables and figures in this chapter.

• C/1 0 •. C O ea.RUTTI NG OF BASE AND SUBGRA DE z ao C '0 Un 0 C 0 0 C 0 O C 0 CC rl 0 O C O O Ci C co tC fn a. AC SURFA CI CHANCE IN BASE THICKNESS GEOSYNTHETIC '0 a. 01 Se 0 Z • C" U -••• C < `•• 3 0 •••• GC1 = ••••• a. O CI 0 8 IN. O O C O 0 z U1 0. C O a. F • 2 2 0 0. ti L 0• C tal U 0 .1 O O O C O • .

nod Base /P 1.. 0 ry 0. O 0 0 tri C O Ot H C BASE THICKNESS %we Ca Ca c.1 0 O O O z z z •-■ Lc) :41 O O U U O C C O I.POOR BA SE/FA C O O O O O O O O .D O O z C sr cn 1.1 Y C c/1 cf) < o NEM • •• z 0 C Z C t..1 LI 0 z z c. cc O O C C O C O O ce ea C C Cf) o-1 Cf) (01 Cd A.. Lo te = 2 .. • ) •••■ .-• 0 3 ay er Stra in Me t K BA SE/FA IR SUBC.1.r) O CD 105 >•■ z C O .

C.0 14. T (INCHES) Figure 38. Reduction in Subgrade Permanent Deformation.C.0 8./8 IN BASE S9 -4000 LBS/IN 0 • LAB S9 AT BOTTOM 1.So w 2. T (INCHES) Figure 37.0 12.0 14./8 IN BASE S9 =4000 LBS/IN -20 6.2 IN A./8 IN BASE 59 =4000 LBS/IN 2. 60 • 50 TEST SECTION 1. Es= 3.0 BASE THICKNESS.5 KSI S9 =4000 LBS/IN POOR BASE • GEOTEXTILE AT MIDDLE ■GEOTEXT1LE AT BOTTOM 2E5 0.0 BASE 'THICKNESS. 16.2 IN A.2 IN A.0 10. Es= 3. 40 Z do • o z I= 0 S9POSMON 2/3 UP < 20 W MO U. uj 0 10 BOTTOM L 0 60 8.C.G.0 16.5 IN A. Reduction in Base Permanent Deformation.0 .C.5 KSI S9 =4000 LBS/IN • POOR BASE s POSITION • BOTTOM L6j LLI CO CL M (13 1g Z 40 1/3 UP 2 0 °0 2/3 UP cc w 0 ■ LAB S9 AT MIDDLE 1.0 10.0 12.5 IN A.

Strain measurements from Test Series 3 indicate that placement of a stiff to very stiff geotextile in the middle of the aggregate base reduced the tensile strain by about 26 percent.5 in. In contrast. (30 mm) compared to 2. Fatigue. the reinforcement should be placed at the top of the base. Also. Full-scale measurements made by van Grup. specific material properties and loading conditions.properties of the test sections were. The maximum calculated changes in tensile strain in the asphalt were less than about 3 percent. These small changes in tensile strain. The analytical results (Table 17) show for increasing fatigue life placing the reinforcement 1/3 to 2/3 up in the base is better than at the bottom. however. the measurements from Test Series 2 showed the strain in the bottom of the asphalt layer to be higher due to the placement of a stiff geogrid at the bottom of the layer. however. cause reductions in required base thickness of up to about 20 percent (Table 17) for light pavements on a subgrade having a low resilient modulus E s = 3500 psi (24 MN/m 2 ). poorer than for standard reference sections. The optimum position of the geosynthetic with respect to minimizing permanent deformation depends upon the strength of the section. Summary.2 in. [41] did indicate an extremely stiff steel mesh reinforcement placed at the top of the aggregate base can reduce tensile strains by about 18 percent under certain conditions. the asphalt thickness of the experimental sections were only 1. The analytically calculated reductions in strain in the bottom of the asphalt surfacing were not validated by the experimental results which were inconsistent. The optimum depth 107 . If only fatigue is of concern. (64 mm) for the analytically developed relations shown in the figures. et al.

Nevertheless. the optimum reinforcement position is near the middle of the base. should be placed in the bottom of the base. Consideration should be given to placing the reinforcement near the middle of the base when low quality aggregate bases are used which are known to be susceptible to rutting. the reinforcement should be placed somewhere between the middle and the top of the layer. A greater beneficial effect will also be realized for this higher location of reinforcement with respect to fatigue of the asphalt surfacing. reinforcement. particularly if it is known to be susceptible to rutting. The analytical results indicate to minimize fatigue cracking of the asphalt surfacing. 108 . Both the laboratory tests and the analytical study indicates placement of the reinforcement at the bottom of the layer should be most effective where a soft subgrade is encountered. even small reductions in tensile strain in the bottom of the asphalt can give for equal fatigue performance large reductions in required aggregate base thickness.might also be dependent upon the width of wheel load although this variable was not investigated. The experimental results of van Grup and van Hulst [41] which used steel mesh reinforcement are quite promising for the reduction of fatigue cracking. when used. or perhaps as high as 2/3 up as indicated by the analytical study. Reductions in tensile strain indicated by the analytical theory due to reinforcement might not be as great as actually occur in the pavement. The purpose of this reinforcement would be to reduce rutting within a soft subgrade typically having a CBR<3 percent. The reduction in tensile strain in general should be considerably less for full size sections than the 26 percent reduction observed for Test Series 3. To minimize rutting in the aggregate base. The analytical results indicate that when high quality base materials and good construction practices are employed.

Base Quality
Use of a low quality base can result in a significant reduction in the
level of pavement performance due to increased permanent deformation and
asphalt fatigue as a result of a lower resilient modulus. A low quality
base might be caused by achieving a compaction level less than 100 percent
of AASHTO T-180 density, or by using low quality materials. Low quality
aggregate bases would include those having a fines content greater than
about 8 percent and also gravels, sand-gravels and soil-aggregate mixtures.
A high fines content base may also be frost susceptible [54].

Observed Test Section Improvements.

The pavement used in Test Series 1 had

a 1.4 in. (36 mm) bituminous surfacing and 6 in. (150 mm) thick sand-gravel
base. The pavement failed after about 1262 wheel repetitions (Table 13).
At this time the base of the control section without reinforcement had a
permanent deformation of 0.69 in. (18 mm). The companion section having a
very stiff geotextile (S g = 4300 lbs/in.; 750 kN/m) at the bottom of the
base had a corresponding permanent deformation of only 0.35 in. (9 mm).
Thus, for under-designed sections having low quality bases, geosynthetic
reinforcement can reduce base rutting up to about 50 percent as observed in
Test Series 1. Of interest is the finding that at about one-half of the
termination rut depth, the reduction in base rutting was also about 50
percent.
The same very stiff geotextile was used in Test Series 3 as for Test
Series 1. As previously discussed, the sections included in Test Series 3
were considerably stronger than the first series. Test Series 3 sections
had a thicker 8 in. (200 mm) crushed limestone base and an asphalt surfacing
rather than the rolled asphalt used in the first series. The pavement of

109

Test Series 3 withstood about 100,000 load repetitions, confirming it was a
higher quality pavement than used in the first series.
When the very stiff geosynthetic reinforcement was placed at the bottom
of the base, permanent deformation within the base was reduced by only 3
percent compared to 50 percent for the lower quality pavement of Test Series
1. In contrast, placement of the same reinforcement at the center of the
base resulted in a 31 percent reduction of permanent deformation within the
base.

Analytical Results.

Results of a nonlinear finite element analysis

indicate that for low quality bases, the ratio of the average resilient
modulus of the base to that of the subgrade (Eb/E s ) averages about 1.45
compared to about 2.5 for high quality materials for the sections studied.
Therefore, reductions in rutting in the light reference pavement previously
described were developed for both of the above values of modular ratios
(Table 19). The stress state within the pavement was first calculated using
the cross-anisotropic analysis and these modular ratios. The layer strain
approach was then employed together with appropriate permanent strain
properties to calculate permanent deformations.
Both a high quality base (indicated in the tables as a "good" base),
and a low quality base (indicated as a "poor" base) were included in the
layer-strain analyses (Table 19). A complete description of the layer
strain approach and the permanent strain material properties are given in
Appendix C.
Calculated permanent deformations are given in Tables 17 and 18 for
both the poor and good bases for a modular ratio Eb/E s = 2.5. This was done
to extend the results and develop a better understanding of the influence of
reinforcement on permanent deformation. Strictly speaking, the lower

110

tf1

U

• •
XI W
7 W ••e
N 0

O

L.1 • •
WC
.14 lal ••4

4, .0
C C.

0

0

0 0

0 0

UI N In
•••1
••1

r4

0

0 0

0

ir
0* 0 0
. • .

0 0 0

1,1
0 101
••1

I

0

M

. CD.
I c*
CI 0

IN .1
.4 .4

0

0 0

0

.1 .0

P1 J

9.1

N

•ri

a)
el .4

U
rl

N

•M • •
I-• fa.
0 47C
•••4
0-. 0 •••

U


0

N
0

OG


0

12,
0

0


0 0

I

0

0 0

U
1"
1• ■
.0 id1 I" ••,/ 101

V
'0

••■
C.
CO
.0
■••

0
0
0

U

•-•1

It

0
4./
tfS

(t)

az
U

.0

REDUCT IONIN RUTT I NG( PERCENT)

U
47
4.4

00
.0

01
w
M
00

(11


C
•••
•••••
2
.0
a

1.
00
.0
2

0.

In

V
La

44


C

41
•-•

al

BASE
GEO SYNTHE T ICATBOTTOMOF AGGREGATE

O

.11

IPS
att UI
I
I

N A rrt



‘`,1 e.4
r4

" .5 " "

4

eV

ell

00 00

0
.0
01 el
I
I

•et
I

I

r.
I

I

03 00

en en

I •C
I

I

I

•-• O "
I
1

4, 0
•••

•••

I

UI

•/, CO CO
C•I
I
/

I

••••
I

•11

N

I

.0

1

I

I

I

ri
I

I

I

I

N

ee
ee

I

/

I

I

.4 co O.
I

1.1
N 4

I

(a
01

0 0
N

00
.0
7

I

Vat 4/1

c••

el I v. .4 I 4 ■1
I
I
a

O
- 413

I

I

I

I


N
I

a

$.1

00

e1

44

tn"

0
41
U

AN N ,1 N IN •4

1.4

40-1

*el

GEOSYNTHETIC2 / 3

GEOS YNTHET IC1 / 3

••••
.0
101

O

2
••■

ta

•••4 1.• C
1.•

r•

•••

I

•ft

0
I

I

I

" O 0 0

I

UI



r.. N. r.




• N N N wr,
Ch 0. O ....

d

00

I

I

0.1 r•


•••■
UI

/.1

1,1

N.


10..

r1 e1


*0.1
N

•••

4.4

../.•••••■■■•.

/•4

0.
1. •3
I.
0 o o 0 o o 0
0 CZ

V

CZ CZ

111

1.•
0 Q

3 0. 4.)

N 1. '0
0 0 0

c?.

3

0.
5.•
0 0
0
0.
0.

quality base properties should probably not have been used with the stress
states obtained from analyses for Eb/E s = 2.5. The results for a lower
modular ratio E b/Es = 1.45, which are more suitable for lower quality base
pavements, are given in Table 19.
Use of a geosynthetic reinforced low-quality aggregate base causes
about 3 times greater reduction in actual permanent displacement in the base
than for a high quality base. The analytical results indicate little change
occurring in permanent deformation in the base as the position of the
geosynthetic was varied. The experimental findings, however, show
reinforcement at the middle of the base to be most effective and is
preferred to reduce base rutting.
Geosynthetic Stiffness
The analytical results indicate that geosynthetic stiffness has an
important effect upon the level of improvement as shown in Figures 39 and 40
(refer also to Tables 17 and 18). For stiffnesses greater than about 4000
lbs/in. (700 kN/m), the rate of change in improvement with increasing
stiffness appears to decrease.
The pavement sections given in Figures 39 and 40 have an asphalt
surface thickness of 2.5 in. (64 mm) and a subgrade with a resilient modulus
of 3500 psi (24 MN/m2 ) corresponding to a CBR of about 3 percent. Base
thicknesses varied from 9.75 to 15.3 in. (250-390 mm).
For these conditions, an AASHTO design for 200,000 equivalent 18 kip
(80 kN) single axle loads (ESAL's) has a base thickness of about 12 in. (300
mm). The equal vertical subgrade strain analytical approach (Figure 39)
indicates that allowable reductions in base thickness for this design
increase from about 3 to 16 percent as the geosynthetic stiffness increases
from 1000 to 6000 lbs/in. (170-1000 kN/m). Permanent deformations as
112

(LN3083d) 117
SS3NN311-113SV9 NI NO11011038

UN30h3d) PV 'ONLLIflel NI NO1l.3flCMI

UN30213d) cIV 'ONIIStel NI NOLIOnG321

113

(b) Po s it io n

( b) Subg r ade De fo rma t ion
Imp rovemen t in Pe r fo rmanc e ' wit h Geosyn t he t ic St i ffness.

(1N33e13d) QV I NOINPIelOACI
30Ve108fIS NI NOLLTION

Fig ure 40.

O

Imp rovemen t in Pe r formanc e w it h Geosy n t he t ic St i ffnes s.

0

To ta lDe forma t ion

0

Fig ure 3 9.

(a ) Re duc t ion in Base Thickne ss

O

determined by layer strain theory are reduced from 12 to 36 percent for a
similar variation in geosynthetic stiffness (Figure 40a). The experimental
results suggest the levels of improvement in rutting shown in Figure 40
might be too high for the pavement section used in the comparison.
From a practical viewpoint, these results indicate that very low stiffness
geosynthetics (S g < 800 lb/in.; 140 kN/m) would be expected to have no
noticeable effect on pavement performance. This would be true even for the
relatively light structural sections shown in Figures 39 and 40.

Structural Strength
The beneficial effect of reinforcement in terms of reduction in base
thickness and rutting decreases as the overall base thickness becomes
greater when all other variables are held constant. Consider the light
reference pavement described in the previous section (2.5 in. AC, E s = 3500
psi; 64 mm, 24 MN/m 2 ), with reinforcement in the bottom having an S g = 4000
lbs/in. (700 kN/m). Increasing the base thickness from 9.75 in. (250 mm) to
15.3 in. (400 mm) results in a very small reduction in base thickness
decreasing from 14 to 12 percent based on the subgrade strain criteria
(Figure 39a). Reductions in rutting of the base and subgrade computed by
layer strain theory were from 39 to 22 percent. The total reduction in
permanent deformation increases from about 10 to 55 percent as the thickness
of the pavement decreases from 15 to 6 in. (381-150 mm) as shown in Figure
41.
The results of Test Series 2 and 3 suggest actual levels of improvement
in permanent deformation for the sections shown in Figures 39 and 40 might
not be as great as indicated by layer strain theory. However, for the first
series of laboratory pavement tests, the observed reduction in rutting due
to reinforcement was about 44 percent. These sections were thin, very weak
114

60

ce
a
.40

(.6

20
0

D

0
w

0

0

10.0

15.0

BASE 'THICKNESS, T (IN)
Figure 41.

Influence of Base Thickness on Permanent Deformation:
Sg = 4000 lbs/in.

18

0
3

9

6

12

SUBGRADE MODULUS, Es (KSI)

Figure 42. Influence of Subgrade Modulus on Permanent Deformation:
S = 4000 lbs/in.
g
115

and placed on a poor subgrade (E s

2000 psi; 13.8 MN/m 2 ). Thus, both the

laboratory and analytical results indicate if the system is weak enough so
that stresses are close to failure, important reductions in
permanent deformations can be achieved by base reinforcement.
Now consider the effect of significantly increasing the load carrying
capacity of the pavement from the 200,000 ESAL's of the previous example to
perhaps a more typical value of 2,000,000 ESAL's. The subgrade resilient
modulus will remain the same with E s = 3500 psi (24 MN/m 2 ). Let the asphalt
surfacing increase from 2.5 to 6.5 in. (54-165 mm), with an aggregate base
thickness of about 12.4 in. (315 mm). For a section having this structural
strength, relatively small changes in stress result from the applied loading
either with or without reinforcement (Table 3). For example, the total
change in radial stress due to loading near the top of the subgrade is less
than 0.1 psi (0.7 kN/m 2 ). Hence, as shown in Table 18, very little
reduction in rutting occurs as a result of reinforcement. This conclusion
is in agreement with previous observations of Brown, et al. [37] for largescale laboratory pavements and by Ruddock, et al. [21,30] for a full-scale
pavement having a comparable bituminous thickness to the section above.
Subgrade Strength. A decrease in the strength of the subgrade as defined

by the subgrade stiffness E s has a very dramatic beneficial effect on the
level of improvement due to reinforcement that can be expected based on the
fatigue and rutting equal strain comparisons. Consider a pavement having an
asphalt surface thickness of 2.5 in. (64 mm), and a base thickness of 9.7
in. (250 mm). Figure 42 shows that a reduction in subgrade stiffness from
E s = 12,500 psi (86 MN/m 2 ) to 3500 psi (24 MN/m2 ) causes the decrease in
base thickness due to reinforcement to increase from about 5 to 14 percent
for a stiff geosynthetic having S g = 4000 lbs/in. (700 kN/m). For a similar
116

section having a reinforcement stiffness S g = 6000 lbs/in. (1000 kN/m), the
corresponding decrease in base thickness is from 6 to 16 percent as the
stiffness of the subgrade decreases. These comparisons are both for equal
vertical subgrade strain; this criterion gives the greatest reductions in
base thickness.
For a given structural section, the layer strain theory would also show
a significant increase in beneficial effect with regard to rutting as the
strength of the subgrade decreases. For all computations of permanent
deformation using the layer strain approach, however, the same subgrade
permanent strain properties were used, regardless of the resilient modulus
employed in the analysis. Suitable permanent deformation properties for
other subgrades were not available.
Slack
During installation of a geosynthetic, slack in the form of wrinkles
and irregularities may develop in the reinforcement. As a result, its
effectiveness as a reinforcement may be significantly reduced as indicated
by a supplementary nonlinear finite element sensitivity study. Figure 43
shows that even a small amount of slack in a geosynthetic theoretically can
result in a very significant reduction in the force developed in the
reinforcement. The rate of reduction in geosynthetic force becomes less as
the amount of slack increases.
For the purposes of this study, slack was defined in terms of strain in
the geosynthetic. Hence, slack expressed as a displacement equals a
geosynthetic length, such as its width, times the slack expressed as a
decimal. A slack of 0.1 percent corresponds to 0.14 in. (3.6 mm) in a
distance of 12 ft. (3.6 m). Slack in a geosynthetic as small as about 0.1

117

percent of its width reduces the geosynthetic force by about 60 percent, and
a slack of 0.4 percent causes a 90 percent reduction in force (Figure 43).
In an actual installation, the effect of slack may not be quite as
great as indicated by theory. This would be due to the geosynthetic
generally being in full contact with the surrounding materials after
construction has been completed. In laboratory tests, such as those
performed for this study, slack can be easily removed by hand stretching the
small pieces of geosynthetic required in these tests. In full-scale field
installations, slack is an important practical consideration which must be
minimized through proper construction practices as discussed later.
Poisson's Ratio.

The value of Poisson's ratio of the geosynthetic was

found to have a moderate effect on the force developed in the geosynthetic.
As the value of Poisson's ratio increases, the force developed in the
geosynthetic also becomes larger, and hence the effectiveness of the
reinforcement increases. For light pavement sections on a weak subgrade,
increasing Poisson's ratio v from 0.2 to 0.4 results in a 29 percent
increase in the force developed in the geosynthetic; corresponding
reductions in tensile strain in the asphalt surfacing and vertical
compressive strain on the subgrade are less than 0.2 and 1 percent,
respectively. Further, the compressive increase in radial stress is only
about 0.075 psi (0.5 MN/m 2 ) as shown in Figure 14. A Poisson's ratio of 0.3
was used in all other sensitivity analyses.
In summary, if all other factors are equal, the geosynthetic having the
greatest value of Poisson's ratio should perform best. The improvement in
performance for moderate increases in Poisson's ratio should be reasonably
small. Such improvements would be very hard to detect experimentally
because of variability in the results. Practically no information is
119

presently available concerning the value of Poisson's ratio for
geosynthetics.

Geosynthetic Slip
A slip failure can occur along the interfaces between the geosynthetic
and the materials above and below. The occurrence of interface slip reduces
the effectiveness of the geosynthetic reinforcement. As the rutting beneath
the geosynthetic increases, the tendency to slip also increases. Whether
slip occurs depends upon (1) the shear strength (T) that can be developed
between the geosynthetic and the materials in contact with it, and (2) the
level of shear stress developed along the interface due to the external load
applied to a particular pavement structure. The level of applied shear
stress is related to both the resilient and permanent deformations in the
pavement, including the shape of the deflection basin.
Slip may occur directly at the interface between the geosynthetic and
the adjacent soil, or by sliding of soil on soil immediately adjacent to the
interface. The resulting ultimate interface shear stress, (T) for sliding
at the interface can be predicted by the expression:

= Ca

where:

T =

an tans

(1

)

ultimate shearing resistance along the
interface

an

=

stress acting normal to the geosynthetic

ca

=

adhesion

d

=

friction angle

The contact efficiency e between the geosynthetic and the surrounding
material is defined as e = d/0 and is expressed as either a percent of 0 or
in decimal form [55]. Angular, well-graded sands and silty sands have been
120

found to exhibit high efficiencies when in contact with most geotextiles.
Angular soil grains exhibit better friction performance than rounded grains.

Testing Methods. The interface friction characteristics of a geosynthetic
to be used for aggregate base reinforcement can be best evaluated using a
direct shear test [55-59] as compared to a pullout type test [55,60,61].
Either a free or a fixed type direct shear test can be used. The free type
direct shear test appears, however, to be preferable to the fixed test. In
the free type direct shear test, one end of the geosynthetic is left free as
shown in Figure 44. The same materials to be used in the field should be
placed below and above the geosynthetic, and carefully compacted to the
densities expected in the field. When large size base course aggregates are
used, the apparatus should be at least 8 and preferably 12 in. (200-300 mm)
on a side. Frequently the materials are saturated before performing the
test.
In the fixed shear test development of strain in the geosynthetic is
prevented and this can have an important effect on the interface friction
developed [61] particularly if it has a relatively low in-plane stiffness.
Bonding the geosynthetic to a rigid block is another technique which has
been used but this hampers natural soil grain penetration and •nteraction
with the underlying material. Nevertheless, Ingold [61] found relatively
small differences in results between fixed and free type tests.

Interface Behavior.

A slip type failure tends to develop under low

confining stress and for smooth, stiff geosynthetics which resist
penetration of soil grains into the surface [56]. For conditions where soil
grains penetrate into the surface, failure develops a small distance from
the geosynthetic within the soil. Failure occurs in this case by adhesion

121

122 . FRICTIONEFFICIENCY. Influence of Geosynthetic Pore Opening Size on Friction Efficiency (Data from Collios.1 1. o/+(PERCENT) 120 • • • 100 • • 80 • • DIRECT SHEAR TEST CRUSHED GRAVEL ON BOTH SIDES *u353° 60 0. Ref.0 100.0 GEOSYNTHETIC OPENING SIZE.NORMAL STRESS C I' SHEAR STRESS STONE GEOSYNTHETIC (b) Free (a) Fixed Figure 44. 55). Free and Fixed Direct Shear Apparatus for Evaluating Interface Friction.. (mm) Figure 45.0 10. et al.

The contact efficiency for loose sands in contact with a wide range of geotextiles is close to the angle of internal friction. As pore openings of the geosynthetic increase (or the grain size of the soil decreases). The geotextile having the higher inplane stiffness reaches the peak interface 123 . When the effective grain size of the soil on the side which has relative movement is smaller than the pore openings of the geosynthetic.e. and hence the friction angle (6) becomes greater as illustrated in Figure 45 for a crushed gravel. This has been observed for nonwoven. and interlock of soil grains [56]. Cohesive soils require less surface roughness than cohesionless materials for development of a "soil on soil" failure immediately adjacent to the geotextile. better penetration of the grains into the pores of the geosynthetic occurs. achieved for most materials placed against very open reinforcement such as geogrids. Factors that otherwise would be important have in general only minor influence on the friction behavior. contact efficiency is high. A high contact efficiency is. needle-punched geotextiles by Martin.and rolling.. typically varying from about 75 to 90 percent but it can be as great as 100 percent [57. the contact efficiency may be greater than 100 percent (i. dilation. When the material particle size is less than the openings of the reinforcement. Clays also have a high contact efficiency [55]. with the range in contact efficiency typically varying from about 90 to 100 percent of 0 [62].62]. Consider two geotextiles having the same size pore openings. therefore. A geotextile that is compressible in the direction perpendicular to the plane of the fabric allows better penetration of particles. The inplane stiffness of the geotextile also affects interface friction behavior. sliding. 6/0 > 1). et al. [57]. For dense sands the contact efficiency is lower.

however. [55] found for tests involving stone on stone the contact efficiencies of three different large stones to be 86 percent for crushed gravel and 66 percent for rounded gravel compared to 84 percent for sand. eventually reaches a higher peak shear stress [55].6 in.4 in. up to 40 percent loss of interface shear strength was observed. the subgrade soil should be compacted in the bottom of the shear box. for large base course aggregate or very rough geosynthetics. and the coarse base or subbase aggregate in the top [59. The relative displacement required to develop full shear strength at a ballast-geosynthetic interface was found by Saxena and Budiman [59] to be about 1. Upon cycling the shear stress.63]. The loss of shear strength appeared to be due to the ballast pulling the fibers and causing severe deterioration of the geotextile. Aggregate Bases. et al. since stone was located both above and below the geosynthetic. Typically displacements of 0. Collios.shear stress at a much lower deformation than the lower modulus geosynthetic. To simulate field conditions. However. full interface strength would probably not be mobilized. as much as 1 to 2 in. The deflection required to reach peak shear stress is a function of the particle size and the normal stress.1 to 0. Usually the geosynthetic has been placed at the interface between the granular base or subbase and the subgrade. Hence for the pavement problem where deformations are small. (3-10 mm) are required [56]. These friction test results are applicable when a geotextile is placed within a granular layer. This large displacement was about three times that required at the soil-geosynthetic interface on the other side. 124 . The lower stiffness geosynthetic. (25-50 mm) of displacement may be necessary to mobilize full interface strength [59]. (41 mm).

a greater tendency to slip occurs as the total deflection of the geosynthetic increases. These results changed into a slightly different form are given in Table 20. As a result well-graded base coarse aggregates protrude through the openings and hence exhibit a high contact efficiency. a bearing capacity type failure may occur in front of the transverse members of a grid. et al. compared to 88 percent for the fixed shear test. Grid Reinforcement. (50 mm) between a geosynthetic and a soft cohesive soil is required to mobilize full friction. Nonlinear finite element analyses indicate that 125 . Both metallic and polymer type grid reinforcements have large openings. In addition to the mechanisms previously discussed. The shear stresses developed at the geosynthetic interface become larger and. Slip in Reinforced Pavements. [64] have presented an excellent discussion of the interaction between a geogrid and soil and give contact efficiencies for seven aggregates. Jewell. Ingold [61] has found the contact efficiency of a geogrid for the free. A medium to coarse sand with some gravel was used in the comparison. direct shear test to be about 106 percent. The occurrence of relatively large adhesion for slippage at both the soil and the stone-geotextile interface is in agreement with the findings of Saxena and Budiman [59]. The high contact efficiency has in the past been attributed for granular materials to aggregate interlock. The laboratory shear test results show that a relative movement of up to 2 in. hence.Robnett and Lai [63] have determined typical values of adhesion and friction angle for geotextiles exhibiting both good and poor friction characteristics.

2c 9 20 126 .2-0. 6 (DEGREES) Soil Geosyn.3)c (-0. (0. Stone-Geosyn. (0. Stone-Geosyn.3-4.4-0.0.5c 20 Low Friction Soil-Geosyn.2c 0. 6 (DEGREES) TYPICAL VALUES ADHESION FRICTION ANGLE.8)c (0.7)c 0-12 0.Table 20 Typical Friction and Adhesion Values Found for Geosynthetics Placed Between Aggregate Base and Clay Subgrade RANGE OF VALUES GEOSYNTHETIC CLASSIFICATION INTERFACE ADHESION FRICTION ANGLE.3)c 6-13 11-30 0.6-0.8c 6 High Friction 19-23 0.

The stiffness of the geogrid was about 1700 lbs/in. Results of the two supplementary single track test studies (Figures 22c and 28c) appears to suggest that the stiff.slip is not likely to occur for sections of moderate strength or subgrades with a CBR 3 percent. but probably not as high a friction as a geogrid. These results indicate that only geosynthetics with good friction characteristics should be used for reinforcement. gave better performance than the much stiffer woven geotextile (refer for example to Table 13 and Figures 18d and 19). despite its lower stiffness. slip does appear to become a problem. (300 kN/m) compared to ' about 4300 lbs/in. woven geotextile used in this project required a much higher deformation to mobilize an equal level of reinforcing potential. From the previous discussion of friction. For lighter sections and/or lower strength subgrades. A geogrid and a woven geotextile were placed at the center of the base in two different sections in Test Series 4. The experimental results showing that a stiff geogrid performed better than a very stiff woven geotextile supports this finding. The geogrid.25 in. This seems to indicate that the strengthening observed in the tests was not due to membrane effects but rather to local 127 . The better performance of the geogrid under the relatively light wheel loading might be caused by better interface friction characteristics due to interlocking between the geosynthetic and the aggregate base. a nonwoven needle-punched geosynthetic should have better frictional characteristics than a woven geotextile. (6 mm) if the full friction in the geosynthetic is not mobilized. (750 kN/m) for the very stiff geotextile. Type of Geosynthetic Reinforcement Reinforcement. Problems with slip and also separation can occur at deformations less than 0.

Prerutting was carried out in both a sand-gravel and a crushed dolomitic limestone base. Separation. A geotextile and a geogrid were placed at both the bottom and middle of the aggregate base of different sections. The amount of subgrade soil contamination of the base in sections having the geotextile was negligible. slack in the geosynthetic can very significantly reduce its effectiveness as a reinforcement. When loading was conducted above instrumentation.5 in. PRERUTTING As previously discussed. The loading was carried out along a single wheel path until the desired level of rutting was developed. This conclusion is supported by the work of Panner. probably caused by small increases in lateral confining pressure. These results show that special consideration must be given in an analytical study of sections having geogrid compared to geotextile reinforcement. The separation effect is not considered to be significant for this study in regard to improvement in pavement performance. (38 mm).reinforcement. One very efficient method of removing slack and even applying some pretensioning to the geosynthetic is by means of prerutting as demonstrated by Barenberg [65]. Haas and Walls [40]. 128 . The performance of a number of prerutted sections both reinforced and non-reinforced were evaluated during the laboratory phase of this investigation. The woven geotextile performed better than the very open mesh geogrid in performing as a separator between subgrade and base. Prerutting was performed by applying applications of a wheel load to the top of the aggregate base before the asphalt surfacing was applied. while in geogrid sections it was as great as 1. Geogrids were of course not developed to perform the function of separation.

prerutting was continued until a surface rut of about 2 in. The results from Test Series 2 (Table 13) indicate an 85 percent reduction in subgrade rutting. The vertical subgrade stress developed in non-prerutted sections tended to increase at a gradually increasing rate throughout the test. This level of rutting was approximately equivalent to a 0. For the subsequent stronger test series where instrumentation was present a subgrade rut depth of 0. Supplementary tests showed. The vertical stress on the subgrade appears to remain relatively constant with number of load repetitions until the pavement has been severely damaged (Figure 26a). however. (10 mm) was developed.000 and 10. The experimental results of Test Series 2 (Figure 22b) indicate that prerutting an aggregate base reinforced with a geosynthetic results in an important overall reduction in surface rutting of the completed pavement.prerutting was continued until a rut depth was developed on the subgrade of about 0. (200 mm) thick aggregate base. (19mm) for the first test series which involved very weak sections. Therefore. prerutting alone is the mechanism which explains the observed improvement in performance. and a 60 percent reduction in base rutting apparently due 129 .4 in. The presence of a geosynthetic reinforcement appears not to affect the efficiency of prerutting. Prerutting appears to reduce vertical resilient and permanent strains in the base and subgrade (Figures 23(a) and (b) and Figure 24(a) and (b). If instrumentation was not present. (50 mm) was achieved in the sections having an 8 in.4 in. The number of load repetitions required to accomplish prerutting was between 5. Reinforced sections which have been prerutted can reduce surface rutting by 30 percent or more compared to non-prerutted sections.75 in.000. that prerutting a non-reinforced section is just as effective as prerutting one which is reinforced (Figure 28b). (10 mm) subgrade rut.

Improved resistance to permanent deformation and less rutting are thus achieved. and then (3) Release the prestress force. PRESTRESSED GEOSYNTHETIC Basic Prestressing Concepts One potential approach for improving pavement performance is to prestress the geosynthetic [35. the geosynthetic prestressing element tries to return to its original. similar to that from a pneumatic-tired roller. Prerutting therefore appears to be most effective in reducing the permanent deformation in the soft subgrade but can also significantly reduce rutting in an aggregate base. stiffer zone at the top of the aggregate layer. This can be achieved by the following procedure: (1) Stretch the geosynthetic to a desired load level. Prerutting is beneficial because of the additional compactive effect applied to the aggregate base. This mechanism was indicated by a high permanent deformation in the weak aggregate layer of the prerutted section in the first test series (Figure 20a). Upon release. in prerutting a weak granular base which tends to shear rather than densify under a concentrated wheel load. Prerutting alone has more benefit than placing a geosynthetic at an effective location (Figure 28b). The friction developed between the geosynthetic and the surrounding soils restrains the geosynthetic from 130 . however.36]. The formation of shear planes or a weakened zone within the aggregate layer as a result of prerutting can have a detrimental effect on pavement performance. Care must be taken. Prerutting normally results in the formation of a denser. (2) Hold the geosynthetic in the stretched position until sufficient material is above it to prevent slip.to prerutting. unstretched condition.

Stress relaxation is a loss of force in the geosynthetic occurring when it is prevented from undergoing any deformation. Prestress loss due to stress relaxation probably reduced the effective applied prestress force to perhaps 20 lbs/in. the force from the geosynthetic is transferred to the surrounding soil as a compressive lateral stress. which was the prestress level used in the analytical study. The shear stress distribution developed along the geosynthetic is approximately as shown in Figure 46.5 kN/m). Stress relaxation in geosynthetics can be quite large and is highly dependent upon the material type with less stress relaxation occurring in polyester geosynthetics. part of the beneficial effect of prestressing is lost through slippage along the interface of the geosynthetic. stress relaxation can be visualized as the inverse of creep. the clamps were removed. Experimental Findings The stiff polypropylene geogrid was used for the prestressing experiments. The improvement of pavement performance was clearly indicated by the results of the fourth test series as shown in Figures 18 and 19 (refer also to Table 15). The geogrid was initially stretched to a force of 40 lbs/in. The mechanism of load transfer to the aggregate base and subgrade is through the shear stress developed along the sides of the geosynthetic. After construction. (3. Important losses of prestress force are also developed through stress relaxation. (7 kN/m) and then. If sufficient friction cannot be developed to hold the geosynthetic in place. As a result.moving. The pavement with prestressed geogrid performed better than both a non-prestressed section reinforced with a stiff 131 . The loss of prestressing effect through stress relaxation is unavoidable. the sides were rigidly clamped against the walls of the test facility during construction of the aggregate base and asphalt surfacing.

. ► ■■ 4— 4. .SHEAR STRESS T / i • • • am■ F 04. +O.. F 0 initial Prestress Force fibs/in. 4+ 4+ 4• 4+ 4+ 4+ 4- 4• 4+ 4+ .■••• — • m■ • • • • • rn■ •• • • • • • I ■ +W gm.) i GEOSYNTHETIC Figure 46.....+0.. 132 .. em +10 . Variation of Shear Stress Along Geosynthetic Due to Initial Prestress Force on Edge.

however. 133 . On the other hand. prestressing would probably not be effective in a field installation for a pavement having a life of 10 to 20 years or more. 300 kN/m). By 70. This conclusion is valid for the conditions of the study including using a polypropylene geogrid with S g = 1700 lbs/in. the difference in measured strain was only about 5 percent.A10.loadreptins the prestressed geogrid pavement had about 30 percent less permanent deformation than the corresponding non-prestressed geogrid section.geogrid (S g = 1700 lb/in. The measured strain in the bottom of the asphalt surfacing of the prestressed section at 10.000 repetitions. If the beneficial effect of prestressing on tensile strain was a result of general pavement deterioration.000 load repetitions was about 30 percent less than in a geotextile reinforced section not prestressed (Table 13). (7 kN/m). which was the next most satisfactory one.. then prestressing should be quite effective in increasing fatigue life. (300 kN/m) initially stressed to 40 lbs/in.. if the loss of prestress was due to stress relaxation with time. or loss of prestress with increase in lapsed time from construction. and a very stiff woven geotextile (S g =430lbs/in. Of considerable practical importance is the finding that the prerutted section having a very stiff geotextile in the middle performed equally well compared to the prestressed section.75kNm)reinfocdst. It then follows from the other results of the experimental study that prerutting a section without a geosynthetic should be just as effective in terms of reducing permanent deformation as prestressing (Figures 28c and 28d). An important unknown is whether the apparent loss of the beneficial effect of prestressing on strain was due to general deterioration of the pavement as a result of reaching the end of its life.

and reductions in base 134 . With this exception. reduction in rutting of the subgrade would also be small. an effective prestress force of 20 lb/in. The analytical results indicate prestressing the center of the layer would have little effect on the vertical subgrade strain and may even increase it by a small amount. expected reductions in total permanent deformation are on the order of 20 to 45 percent. The analytical study indicates prestressing the bottom of the layer is quite effective in reducing permanent deformation. For a base thickness of 11. particularly in the subgrade. geotextile reinforced section with reinforcement at the center. For general comparison. the analytical results tend to support the experimental finding that prestressing the middle of the aggregate base should greatly improve rutting of the base and fatigue performance. a variable thickness base and a subgrade with E s = 3500 psi (24 MN/m2 ). The experimental results. The standard reference section was used consisting of a 2. the observed reductions in total rutting of the lighter prestressed experimental section was about 60 percent compared to the non-prestressed.5 kN/m) was applied.9 in. however. For the reference section reductions in permanent deformation were obtained varying from 30 to 47 percent.5 in. (300 mm). demonstrate that prestressing the center of the layer can for very light sections also lead to important reductions in permanent deformation of both the base and subgrade. This represents the net force existing after all losses including stress relaxation. Prestressing the center of the aggregate base based on tensile strain in the asphalt surfacing resulted in large reductions in base thickness varying from about 25 to 44 percent (Table 21). (64 mm) asphalt surfacing.Analytical Results In the analytical study of prestress effects. (3.

SURFACING .18 -12 -9 .7 -4 .30 -39 -9 -1 2 -18 .1 . CHA NCE IN RUTTI NG OF BASE AND S US.33 . 5 IN. -12 8 C •-• PRESTRESS 8 DOT. •-• V •0 La • La Let 1-1 . 9 2 0 0 0 0 .2.. ”. vs •-• . 9 zz 0 -16 6. 5 IN.0 . 75 . MOD BAS E /FAIR S URG.4.28 -7 .tr -14 U.15 3 500PSI SUBGRADE (REINFO RCED) 1 0 0 CEOSYN @ BOTTOM 15.. V. ■ .18 r.g X •-• 21 2 U F. AC SURFACING 2.46 .27 3500 PSI SUBGRADE (PRESTRESSED SECTION) -12 - -6.11 . 5 0 3500 PSI SUBGRA DE ( PRESTRESSED S ECTI ON) . 9 0 POOR BASE/FA IRSW. AC SURFACING .22 .2 O .36 . 3 C 2) C O 2... 0. 8 1:0 .16 •■••• . 6 .mm.23 . 8 . z IE d O C z 0 U z O C 0S Ca/ \ 0 O --.. 5 I N.19 .1.Y c C 0 0 •-• 135 2. 3 tif 11.3. z 0 .12 .3.7 .. 0.-C 0v 9.

which would depend upon the type and composition of the geosynthetic and the initial applied stress level. Allowance must also be made for all other prestress losses resulting between the time pretensioning is carried out and the prestress force is transferred to the aggregate base. The analytical results indicate prestressing the bottom of the base is not as effective. The truck would supply the dead weight reaction necessary to develop the pretensioning force. A rigid longitudinal rod or bar would be attached along the side of the geosynthetic to distribute the pretensioning force uniformly.weight anchor such as a loaded vehicle. Proper allowance should be made for prestress loss due to stress relaxation. (7 kN/m). One approach that could be employed for applying the pretensioning force would be to place sufficient stakes through loops into the ground along one side of the geosynthetic to firmly anchor it. The pretensioning force could 136 . Pretensioning: Practical Field Considerations To achieve the demonstrated potential for an important improvement in performance. would be a reasonable starting point for additional field studies. the geosynthetic should be prestressed in the direction transverse to that of the vehicle movement. which is the force used in the laboratory tests. These losses would be related to the method used to apply and maintain the prestress force and the skill and care of the crew performing the work. Probably an initial pretensioning force on the order of 40 lbs/in. Probably the most efficient method would be to apply the pretensioning force to the other side of the geosynthetic using an electrically powered winch attached to a loaded truck. An alternate approach would be to use a dead . however. as prestressing the middle with respect to reducing tensile strain in the asphalt surfacing.thickness based on vertical subgrade strain of about 35 percent (Table 21).

(3-4. For example.9 m) would be required to hold a light initial pretensioning load of only about 20 lbs/in. Thus. Even moving equipment 137 . The pretensioning force could then be maintained on the geosynthetic until sufficient aggregate base is placed and compacted over it to provide the necessary friction force to prevent slippage. Wood stakes 2 in. It might be desirable to pretension two or more lengths of geosynthetic at a time.5 times the geogrid cost. If base construction was not progressing rapidly. (3.6 m) length of geosynthetic. by 2 in.9 m) in length having a spacing of about 2 to 3 ft. (50 by 50 mm) by 3 ft. consider a soft subgrade having an undrained shear strength of about 500 psf (24 kN/m 2 ).be applied by one winch to about a 10 to 15 ft. The winch and cable system could then be removed and used to pretension other segments of the geosynthetic.5 kN/m). Prestressing the base would most likely be carried out where the subgrade has a CBR less than 3 to 4 percent. The cost to just apply this light level of pretensioning to a geogrid by an experienced contractor would probably be about 1 to 1.5-0. To minimize bending in the rod or bar attached to the geosynthetic. the cable leading to the winch would be attached to the bar at two (or more) locations to form a "V" shape. temporary anchorage of the geosynthetic becomes a serious problem. For conditions where a soft subgrade exists. or where a low quality aggregate base is used. (0. (0. it would be necessary to anchor the side of the geosynthetic being pretensioned probably using stakes. the practicality of applying even a light pretensioning force to pavements constructed on soft subgrades having undrained shear strengths less than about 500 psf (24 kN/m2 ) is questionable. as would likely be the case.

138 . (150 to 300 mm) of the subgrade. anchorage would probably not be practical. be achieved through prestressing the aggregate base. (260 kN/m) for base reinforcement to start to become effective. For light pavement sections (SN . this aspect needs to be established using full-scale field tests. reinnforcement can have an important effect on reducing rutting in the base and upper part of the subgrade. SUMMARY The presence of geosynthetic reinforcement causes a small but potentially important increase in the confining stress and reduction in vertical stress in the base and upper 6 to 12 in. The experimental and analytical results indicate that important reductions in rutting can.over very soft soils to provide temporary dead weight. The stiffness of the geosynthetic is an important factor. at least under idealized conditions.5 to 3) where stresses are high. (260 kN/m) performs about the same as a woven geotextile having a stiffness of about 4000 lb/in. A geogrid performs differently than a woven geotextile reinforcement. reinforcement may be beneficial where low quality bases or weak subgrades are present. 2. however. and should be greater than 1500 lb/in. The experimental results indicate that prerutting the base without the use of a geosynthetic is equally effective at least with respect to reducing permanent deformations. For heavier sections the potential beneficial effect of reinforcement tends to decrease rapidly. The laboratory tests indicate that a geogrid having a stiffness of about 1500 lbs/in. Prerutting would very likely be less expensive than prestressing and should be effective over an extended period of time. (700 kN/m). In heavier sections.

The experimental results on the prestressed sections were obtained for short-term tests performed under idealized conditions. 139 . Loss of prestress effect in the field and prestress loss due to long-term stress relaxation effects are certainly important practical considerations that can only be fully evaluated through full-scale field studies. Limited strain measurements made in the bottom of the asphalt surfacing of the prestressed section indicates an important loss of benefit occurs with either time or deterioration of the pavement.

in general. prerutting is a more positive treatment than prestressing. Prerutting without reinforcement should give performance equal to that of prestressing and significantly better performance compared to the use of stiff to very stiff non-prestressed reinforcement. therefore. The most promising potential method of improvement appears to be prerutting a non-reinforced aggregate base. Geosynthetic Reinforcement. Therefore. or where low quality bases are used beneath relatively thin asphalt surfacings. 18 and 21). General guidance concerning the level of improvement that can be achieved using geosynthetic reinforcement of the aggregate base is given in Figures 47 to 51 (refer also to Tables 17. at least with regard to reducing permanent deformations. Geosynthetic reinforcement may also be economically feasible for other combinations of structural designs and material properties where rutting is a known problem. Further. Further. Prerutting without reinforcement is relatively cheap and appears to be quite effective. be conducted to more fully validate the concept of prerutting and develop appropriate prerutting techniques. The cost of prerutting an aggregate base at one level might be as small as 50 percent of the inplace cost of a stiff geogrid (S g = 1700 lbs/in.ECONOMIC CONSIDERATIONS Prerutting and Prestressing. prestressing the same geogrid would result in a total cost equal to about 2 times the actual cost of the geogrid.. considered to be economically feasible only when employed in light pavements constructed on soft subgrades. the total expense associated with prestressing might be as great as 5 times that of prerutting the base at one level when a geosynthetic reinforcement is not used. The use of geosynthetic reinforcement is. 300 kN/m). Full-scale field experiments should. The results 147 .

0 REQUIRED BASE THICK. SURFACE SUBGRADE CBR a 3.0 0 1000 0 ct 0.5 IN A C.0 vi S9(LBSAN. Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Vertical Subgrade Strain: 6.1 Y 6000 2.5 in.0 6 8 10 12 14 18 18 BASE THICKNESS. T (IN) Figure 49. 3.<1 DESIGN: 6. Subgrade CBR = 3.18 KIP AXLE LOADS Y 2.0 4000 1. E Er CONTROLS 3. C. 149 . z DESIGN: 6. .0 6 8 lo 12 14 18 BASE 'THICKNESS. T (IN) Figure 50. AC. E = 3500 PSI p- <. Approximate Reduction in Granular Base Thickness as a Function of Geosynthetic Stiffness for Constant Radial Strain in AC: 2. Subgrade CBR= 3.5 IN A.5 in. AC.0 la 3500 PSI 5 X 105 2 X 105 S ( LBSAN.) 6000 4000 1..) LL. SURFACE SUBGRADE CBR = 3.0 0 LLI CC 1000 0.

Appropriate techniques for prerutting the aggregate base in the field need to be established. and serve to anchor the two ends of the geosynthetic. If the geosynthetic is to be pretensioned. Prerutting in the laboratory was carried out in a single rut path for a base thickness of 8 in. the resulting settlement stretches the geosynthetic. Pretensioning. Suggestions were made in Chapter III involving application of the pretensioning force by means of winches and cables. (200 mm) aggregate base and also the subgrade. The material is rolled out in the short direction and any necessary seams made. (5-8 m) in width. Prerutting. When fill is placed in the center area. (60-90 m) in length. (18 m) in width and requiring several feet of fill (Figure 53). The geosynthetic is first spread out over an area of about 200 to 300 ft. This technique is particularly effective in eliminating most of the slack in the geosynthetic where soft subgrade soils are encountered. can only be developed and refined through studies including field trials.roadway or embankment about 60 ft. perhaps 153 . Effective methods of pretensioning. Prerutting is just an extension of proofrolling and should probably be carried out with a reasonably heavy loading. Development of a total rut depth of about 2 in. a suitable technique must be developed. and may even place a little initial stretch in the material. Fingers of fill are then pushed out along the edges of the geosynthetic covered area in the direction perpendicular to the roll. (200 mm). The fingers of fill pushed out are typically 15 to 20 ft. (50 m) was found to be effective in reducing rutting in both the 8 in. For full-scale pavements. (12-30 m) ahead of the main area of fill placement between the fingers. it may be found desirable to prerut along two or three wheel paths. however. Usually the fingers are extended out about 40 to 100 ft.

Placement of Wide Fill to Take Slack Out of Geosynthetic.FILL PLACEMENT GEOSYNTHETIC AREA BEING STRETCHED BY FILL SETTLEMENT ON WEAK SUBGRADE Figure 53. 154 .

Caution should be exercised to avoid excessive prerutting. angularity and grading of the aggregate to be placed above the subgrade. either a sand or a geotextile filter can be used to maintain a reasonably clean interface. A moderate wind will readily lift or "kite" a geotextile. (300 mm) apart. Wind Effects. Prerutting an 8 in. applied stress level and the size. geogrids are not lifted up by the wind due to their open mesh structure and hence can be readily placed on windy days 142].spaced about 12 in. On the other hand. additional wrinkling and slack may occur in the material. (200 mm) base lift thickness in the field would be a good starting point. Wind can complicate the proper placement of a geotextile. moisture conditions. Large. Both woven and nonwoven 155 . Separation problems involve the mixing of an aggregate base or subbase with the underlying subgrade. It is therefore generally not practical to place geotextiles on windy days. If geotextiles are placed during even moderate winds. When separation is a potential problem. SEPARATION AND FILTRATION The level of severity of separation and filtration problems varies significantly depending upon many factors. including the type of subgrade. as discussed in Appendix F. angular open-graded aggregates placed directly upon a soft or very soft subgrade result in a particularly harsh environment with respect to separation. Separation problems are most likely to occur during construction of the first aggregate lift or perhaps during construction before the asphalt surfacing has been placed. Prerutting could be performed at more than one level within the aggregate base. The actual rut spacing used would be dependent upon the wheel configuration selected to perform the prerutting.

should perform better than geotextile filters with regard to filtration. the amount of contamination due to fines moving into this layer must be minimized by use of a filter to ensure adequate flow capacity and also strength. Semi-rational procedures are presented in Appendix E for determining when filters are needed for the separation and filtration functions. Thick nonwoven geotextiles perform better than thin nonwovens or wovens partly because of their three-dimensional structure. These procedures and specifications should be considered tentative until further work is conducted in these areas. Whether a sand filter or a geotextile filter is used would be a matter of economics for most applications. Several studies are currently underway which should contribute to an understanding of durability. when properly designed. Durability is discussed in Appendix G. When an open-graded drainage layer is placed above the subgrade. DURABILITY Relatively little information is available concerning the durability of geosynthetics when buried in the ground for long periods of time. The severity of erosion is greatly dependent upon the thickness of the pavements which determines the stress applied to the subgrade. dispersive clays and silty fine sands are quite susceptible to erosion. A very severe environment with respect to subgrade erosion exists beneath a pavement which includes reversible. possibly turbulent flow conditions. although satisfactorily performing geotextiles can usually be selected. Sand filters.geotextiles have been found to adequately perform the separation function. 156 . Low cohesion silts and clays. Guidance is also given in selecting suitable geotextiles for use beneath pavements.

For separation and filtration applications. copper. geosynthetics should have at least a 20 year life. 157 . (7 kN/m). zinc and manganese. Whether geosynthetics used for separation. Most geosynthetics would initially be strong enough to undergo significant strength loss for at least 20 years before a tensile failure of the geosynthetic might become a problem for pavement reinforcement applications. geosynthetic stiffness is the most important structural consideration. Limited observations indicate that some geosynthetics will become more brittle with time and actually increase in stiffness. Whether better reinforcement performance will result has not been demonstrated. or reinforcement can last for 40 or 50 years has not been clearly demonstrated. they are also susceptible to hydrolysis. Under favorable conditions the loss of strength of typical geosynthetics should be on the average about 30 percent in the first 10 years. Because of their greater thickness. Polyesters are attacked by strong alkaline and to a lessor extent. The typical force developed in a geosynthetic used for aggregate base reinforcement of surfaced pavements should be less than about 40 lbs/in. Polypropylenes and polyethylenes are susceptible to degradation in oxidizing environments catalized by the presence of heavy minerals such as iron. filtration. strong acid environments. geogrids may exhibit a lower strength loss although this has not been verified. For reinforcement applications.Consideration should be given to the environment in which they will be used.

2. full-scale field test sections. Prerutting a non-reinforced aggregate base appears to have the best overall potential of the methods studied for improving pavement performance. The geosynthetic reinforcement of an unstabilized. whether reinforcement of heavier sections will reduce permanent deformations needs to be further studied in the field. Prerutting. Geosynthetic reinforcement of light pavement sections constructed on weak subgrades shows promise for reducing permanent deformations particularly in the subgrade. Weak Subgrade. Prerutting in the large-scale experiments was found to be both effective and is also inexpensive. Therefore geogrid reinforcement is recommended as the primary 158 . The recommendation is therefore made that an additional experimental investigation be conducted to further evaluate these three techniques for potentially improving pavement performance. Geogrid reinforcement was found to perform better than a much stiffer woven geotextile. 3. low quality aggregate base appears to offer promise as one method for reducing permanent pavement deformation of pavements having thin asphalt surfacings.SUGGESTED RESEARCH Reinforcement The laboratory investigation and the sensitivity analyses indicate the following specific areas of base reinforcement which deserve further research: 1. This investigation should consist of carefully instrumented. Low Quality Aggregate Base.

Separation/Filtration Important areas involving separation and filtration deserving further study are: 1. stiffness. A very important need presently exists for conducting long-term durability tests on selected geosynthetics known to have good reinforcing properties. A description of a proposed experimental plan for this study is presented in Appendix H. A formal study should be undertaken to evaluate the filtration characteristics of a range of 159 . ductility and chemical composition. Soil environments to include in the experiment should be selected considering the degradation susceptibility of the polymers used in the study to specific environments. Tests should run for at least 5 years and preferably 10 years. Filtration. Geosynthetic Durability. The geosynthetics used should be subjected to varying levels of stress and buried in several different carefully selected soil environments. Such a study would be applicable to mechanically stabilized earth reinforcement applications in general.reinforcement for use in this study. 2. Each geosynthetic product has a different susceptibility to environmental degradation. Properties to be evaluated as a function of time should include changes in geosynthetic strength. and a considerable amount of valuable information could be obtained from a long-term durability study of this type.

At least 10 6 load repetitions should be applied during the test to simulate long-term conditions. The tests should probably be performed in a triaxial cell by applying cyclic loads as water is passed through the sample.geotextiles when subjected to dynamic load and flowing water conditions likely to be encountered both beneath a pavement. and also at lateral edge drains. 160 .

APPENDIX A RKFERENCES 161 .

January. J. Q..K. "Fabric-Reinforced Aggregate Roads Overview". Dept. Federal Highway Administration.. U. D. 37 p. J. C. 9.E. Transportation Research Board. August. 8. 6. 1983. 1984. "Treatment for Reduction of Reflective Cracking of Asphalt Overlays of Jointed-Concrete Pavements in Georgia". R. Frederick.A. 1985. D. January. J.A. 19.R. 4-5. 4. and Brown.. R.APPENDIX A REFERENCES 1. A. J.. Research Record No. and Lai.C.. Bonaparte. April. and Barenberg. "Use of Geogrids in Soil Reinforcement".. D. April. Button.. 1983. Transportation Research Board..H. Transportation Research Record 916. Dept. p.. and Epps. 1984. Report 113. Penna. June. Transportation Research Record 875.. Gulden. Dixon.. 1982. 3.A. J. p. J. of Transportation. Vol. D. Report FHWA-RD-80-021. Transportation Research Board.D. Kamel.. 6-18... 1987. R. 64-75.. 671.. 10.. Research and Development Bureau. Colorado". NCHRP Report 290.C.J. "Test Methods and Use Criteria for Filter Fabrics".K. and Villet. Bender. Mitchell. "Field Evaluation of Fabric Interlayers". N. Texas Transportation Research. Ruckman. Transportation Research Board. No. E. "Heavy Duty Membranes for the Reduction of Reflective Cracking in Bituminous Concrete Overlays". 1978. Washington.B. 1980.S. of Transportation. "Reinforcement of Earth Slopes and Embankments".. Smith. "Laboratory Testing of Fabric Interlayers for Asphalt Concrete Paving: Interim Report"..C. Bell. pp. Washington. 1-6. R. pp. W. for presentation at the 62nd Annual Meeting. "Design and Behavior of SoilFabric-Aggregate Systems". 2..I. 2. Barret. "Stress Relieving Interlayers for Bituminous Resurfacing". Bureau of Bridge and Roadway Technology. Bell. J. 1983. Knight. Robnett. Transportation Research Board. D.S. Research Project 79-6... 162 . 1983. 7. New York State Department of Transportation Engineering . 11. paper submitted to Transportation Research Board Annual Meeting. "Geotextile Earth Reinforced Retaining Wall Tests: Glenwood Canyon.W. N. Transportation Research Record 916.L. Transportation Research Board. et al. 5.

and Phang..W. "Pavement Evaluation Project. Research Record 949.F.M. Second International Conference on Geotextiles.R. 22. Project Record 245. D. November. 18.. and Currer. 1975. Vol. A.R. Perrier. on Use of Fabrics in Geotechnics. Melbourne. 21.. "Development of Stresses in Reinforcement and Subgrade of a Reinforced Soil Slab". and Vischerm. 375-380. CIRCIA. Part 1. E. 1982. Vol.. "Report on the Construction and Performance of a Full-Scale Experimental Road at Sandleheath. 1983.J. Proceedings. S. G. 45 p. Potter.. R. B. 399-404. Proceedings. Barksdale.. Second International Conference on Geotextiles. 163 . Proceedings..D. H. Conf. II. Proceedings.. A. "Model Testing of Geogrids Under an Aggregate Layer on Soft Ground". D. IBID. J. 1.. NSW". "Grid Reinforcement of Asphalt Pavements and Verification of Elastic Theory".C. MacMillan.V.W. Division of Applied Geomechanics. 1984. 71-76. and Brodrick.. 23. and Pefetti. "INTRODUCTORY SOIL MECHANICS AND FOUNDATIONS. pp.E. CSIRO.A.. G. 15. Lai.. 20.C.H. Sowers. 417-422. 1985. 55-65. Rigo. J. Conference on Use of Fabrics in Geotechnics.. Vol. 1981.. G. 1977. Transport and Road Research Laboratory. "Construction and Analysis of a Fabric Reinforced Low Embankment on Muskeg". Riondy. 93-98. II. "The Effect of a Fabric Membrane on the Structural Behavior of a Granular Road Pavement". and Love. pp.. Halim. Griffith. Project Report 2. Report submitted to Netlon Ltd and SERC.F.L.. 1983. IBID. Second Symposium on Unbound Aggregates in Roads. 1977.M. paper 4. Q. Transportation Research Board. 17. W. 14..12. Las Vegas. Gourc.P. 1982. Chaddock. London. University of Nottingham. Robnett. Bell.. 13.F.. J. "Deformation of a Haul Road Reinforced with a Geomesh". 24.O. Vol. J. pp.H. A. P.P.. Ruddock. B. Hants". First Int. First Int.. E. pp. D... Potter. and McAvoy. Raumann.. 1979 (4th Edition). New York. I. 151-154. R.F. Washington.A. J. Haas.S.. W. Milligan. "Grid Reinforcement for Asphalt Pavements". 16. 1982. Brown. Hughes. "Experimental and Theoretical Behavior of Geotextile Reinforced Aggregate Soil Systems"... J. & Zeevaert-Wolf.. J. Petrix..C. J..W. p. "Geotextiles in Unpaved Roads: Design Considerations". J.R.2. Greenway. G.B. pp. 1982. Proceedings. pp. Pappin. "Chargement Cyclic d'un Bicouche Renforce par Geotextile". 19. Report LR 996.

G.R. Vicksburg. C. Transportation Research Record 755.. Report LR1123. Ramalho-Ortigao.. "Investigation of Construction Techniques for Tactical Bridge Approach Roads Across Soft Ground". J. Army Engineering Waterways Experiment Station. 1. Potter. 34. Ruddock..G. Lai.. Sowers. Proceedings. February.A.F. 1982. V..E. paper presented at the 66th Annual TRB Meeting.. "Reinforced Elastic Layered Systems".Aggregate Systems with Mirafi 500X Fabric". 1977. and McAvoy... II. II.A. J... Mississippi. Taiwan. Barenberg.. 35. "Mechanisms of Geotextile-Aggregate Support in Low Cost Roads". University of Illinois. and Watkins.J. Halliday. S. D. pp.L.F. and Raad. and Potter. Proceedings. "Fabric Used in Low-Deformation Transportation Support Systems". A. 1987. Barksdale. First International Conference on Use of Fabrics in Geotechnics. 1982. S. pp. Las Vegas. E. "The Performance of a Flexible Pavement Constructed on a Strong Fabric". V.. 57-60.N. 164 . 29. "Reinforcement of Transportation Support Systems through Fabric Prestressing". Perkov. 1982. and Robnett. II.D. Vokas.. Las Vegas. Second International Conference on Geotextiles. "Design Procedures for Soil Fabric . p. and Palmeira. 365-370. L. 49-51. 30.. Nevada.. paper presented at the 66th Annual TRB Meeting. R.R. "Geotextile Performance at an Access Road on Soft Ground Near Rio de Janiero". 1980. Second International Conference on Geotextiles.. 26. J. J... S..A. and Stoll. Vol. G.D. Transportation Research Record 810. Proceedings. V. 33. Collins. "Geosynthetic Reinforcement of Aggregate Bases of Surfaced Pavements".. M. 31. 1984. Barvashov. Raad. Transport and Road Research Laboratory. Technical Report S-77-1. E.. Vol.I. Thompson. p.S. 1981. October. Vol. J.. Q. Webster.A. Third Conference on Road Engineering Association of Asia and Australia. UIL-ENG80-2019.. and Brown. J. E.M. 27. August.F. January. 1981. Fomin. R. A. January. L. August. 36. 1980. and Pushkin.. Budanov.S. "A Full-Scale Experience on Granular and Bituminous Road Pavements Laid on Fabrics". Proceedings.. and Miller. U. A.25. "Design and Use of Geotextiles in Road Construction". 1977. Vol. 32.R.F. Proceedings.L. Transportation Research Board.R. "Deformation of Soil Foundations Reinforced with Prestressed Synthetic Fabric". Second International Conference on Geotextiles. pp. 6770.C.. 1987. 28. 341-346.

D. 42. 1986.W. W. 47. Proceedings. C. Transportation Research Record 1011. Vol. Waterways Experiment Station. II (in publication).. Vancouver. Georgia Institute of Technology.. Ann Arbor. August.F. pp.A. Lister. "Theme Lecture: Pavement Design and Materials". Proceedings. Penner.W. R. 267 p. "Incremental Design of Flexible Pavements".E.. R.. "Use of Nonwoven Fabrics in Permanent Road Pavements". S. H.37. School of Civil Engineering.D. pp. 45-51.. 73. and Brunton. "Finite Element Formulations for the Analysis of Interfaces. Christ. J. Sixth International Conference on the Structural Design of Asphalt Pavements.. Brown.R. 43..F. 1987.L.M. W.. Brown. Atlanta. Ann Arbor. Nokes. B. and Barksdale. "Developments to the Nottingham Analytical Design Method for Asphalt Pavements"..E. Sixth International Conference on the Structural Design of Asphalt Pavements.. 1987. Final Report. Haas. Brown. Nonlinear and Large Displacement Problems in Geotechnical Engineering". "Compaction Specification for the Control of Pavement Subgrade Rutting".B.P. September.W. Lofti. S.M.F. Schwartz.. 1988. Transportation Research Board.. S. N. Walls.. and Powell. 48. Paper GL-86. and Witczak.. pp.. and van Hulst.. Zeevaert. Barker. R. Hannon. 1985. 40... January. August. 220-231. Vol. Forsyth. January. C.. 173 p. 1987.D. Brown.. School of Civil Engineering..P. 366-377. submitted to Transportation Research Board. R.A. 45. Misc. J.F. 39. pp.. 41.A. Part 2. "Design Practice for Bituminous Pavements in the United Kingdom". "The Modeling of Granular Materials in Pavements". 1985. van Grup. J. paper submitted to Journal of Geotextiles and Geomembranes.. Barksdale. PhD Thesis.. "Geogrid Reinforcement of Granular Bases". "Reinforcement at Asphalt-Granular Base Interface". 44. Georgia Institute of Technology.D. A. July. 1985. 1988.. M. Presented to Roads and Transportation Association of Canada Annual Conference. and Brodrick.W. paper presented at the 67th Annual Meeting. Sept. Jones. R. 1980. Transportation Research Board. 165 . 541-563. February. "Open-Graded Bases for Airfield Pavements". S. Sixth International Conference on the Structural Design of Asphalt Pavements. A. W. 1982. 46.V. and Pappin. "A Field Study of the Performance of a Tensar Reinforced Haul Road". and Prendergast. J. 38. J. Constitution of Civil Engineers. R.M...

50-64.. and Budiman. 57. "Resilient Properties of Unbound Roadbase Under Repeated Loading". Geotextiles and Soil"..C. 1984. ASCE National Convention.. August. Geosynthetic '87 Conference. Proceedings.. 1987. August. 53. and Dawson. 59. R. and Ragneau.M. "Performance of a Thin-Surfaced Crushed Stone Base Pavement". 166 . Conf. Sixth International Conference on the Structural Design of Asphalt Pavements. 3. 2. Jouve. Research Report 414-IF. A. ASTM Symposium on the Implication of Aggregate.. 52. pp. VII-1 through VI-32.D. T. Mayhew. pp.D. and Whitty. J. 1983. Arlington. Martinez. and Giroud.K. A. June 1. Vol.. N. and Machemehl. Koerner. 1987. New Orleans (submitted for publication).. pp. 3.S. 60. "Interface Response of Geotextiles".. S. Barksdale. Barksdale. P. and Chou. 100-111. J... S.D.S.. pp... "Evaluation of Interface Friction Properties Between Geosynthetics and Soil". Saxena. GTJODJ.49. Sept.. 1984. Greene.P. J.. pp.. 191-196. "The Shear Strength Behavior of Certain Materials on the Surface of Geotextiles". 1987.. Portland. 1986. 657-660. 3. 1801-1804. San Francisco.. Report LR 1088.. 58. E.F. R.Francisco. Formazin. R. Oregon... 1980. A. 1773-1775. "The Use of Geotextiles for Soil Improvement". Williams. Transport and Road Research Lab. National Crushed Stone Assoc. Proceedings. pp.E. P. "The Effects of Groundwater on Pavement Foundations". 6. M. J. Goore. 54. on Crushed Stone. No.P. Ann Arbor. Proceedings. T. pp. Scullion. New Orleans.. pp. Vol. 1987.. Bush.. 1985. E. 616-627. pp. Brown. Ingold. R.R. Proceedings..M.P. and Batereau. 9th European Conf. 1985. San. "Laboratory Pull-Out Testing of Grid Reinforcement in Sand". J. 1983. International Conference on Geomembranes. April 17. C.. J. August.D. Delmas. 51. 55. Geotechnical Testing Journal. "Field Evaluation of Geotextiles Under Base Courses . "Thickness Design for Effective Crushed Stone Use". 80-177. Martin. Proceedings.. "Rational Model for the Flexible Pavements Deformations". J. Eleventh International Conference on Soil Mechanics and Foundation Engineering". 56. Texas Transportation Institute. on Soil Mechanics and Foundation Engineering. Eleventh International Conference on Soil Mechanics and Foundation Engineering. J.. Denver.S. Collois.Suppliment". C. Paute.. 50.F. Vol. and Houlihan. 53-73. (Suppliment).. "Experimental Friction Evaluation of Slippage Between Geomembranes. Vol. H.

Milligan. "Permanent Deformation of Flexible Pavements".S.. of Civil Engineering. "A Study of Typar Non-Woven and Other Fabrics in Ground Stabilization Applications". Final Technical Report to ERO U.F.L. "A Laboratory Investigation of Soil-Geotextile Friction". 1984. 1981.J. D.E. J. and Lai..A.61.. School of Civil Engineering.. 1984.S. October. November. G.. 1980. and Pappin. "Soil Fabric Friction Testing". "Interaction Between Soil and Geogrids". J. 66.W. S. Army. Q. 18-29. Ground Engineering. Barenberg.. R. April. ASCE National Convention. 65. Sarsby.A. April 17. University of Illinois.W.W. Research Report. pp. E.. Dept. Brown.S. 1980. 1982. R. pp. Thomas Telford. B. Polymer Grid Reinforcement.. Robnett. Oregon. T. and Dubois. 62.. Brodrick. Bell. Jewell. Georgia Institute of Technology. D. "Modeling of Effects of Moisture and Drainage of NJDOT Flexible Pavement Systems". 64. 63. and Brown.. 167 . 21-112.V.. Ingold.. University of Nottingham. J.. Portland.