Application of Laboratory Evaluation to Develop Stiffness Values and Layer Coefficients for Design

1
Application of Laboratory Evaluation to Develop
Stiffness Values and Layer Coefficients for Design
Topic 1

• Dante Fratta
• UW-Madison

North Central Pavement Research Partnership
2
Outline

• Introduction
• Seismic modulus/resilient modulus comparison
• large box tests
• simple tests
• Large Scale Testing
• gravel equivalency
• Geogrid/soil interaction
• rotations
• modulus change

3
Introduction

• The strain-stress dependency of elastic modulus
  can be described with a backbone curve (Seed and
  Idriss 1970 Hardin and Drnevich 1972).
• The backbone curve describes the ratio of a
  modulus at a given strain to the low-strain
  modulus as a function of strain.
• Kokusho (1980) examined some of the properties
  affecting modulus including confining stress.
  Higher moduli are expected with a higher
  confining pressure while the strain level remains
  constant.

4
Previous Studies

• Kim and Stokoe (1992) and Tanyu et al. (2003)
  demonstrated the effect of strain level in
  evaluating the resilient modulus of subgrade and
  various working platforms.
• Sawangsuriya et al. (2005) used the strain
  dependency of elastic modulus to predict the
  strain level of modulus in the soil stiffness
  gauge (SSG) for medium sand and crushed rock.
• Schuettpelz (2008) described the stress and
  strain dependency of modulus by finding a
  correlation between low-strain elastic modulus
  (from seismic tests) and laboratory resilient
  modulus.

5
Introduction

• Different test equipment and methods result in
  various stress and strain levels, so both strain
  dependency and stress levels are taken into
  account in comparing moduli provided from
  different test methods and materials.
• The effect of confining pressure on the modulus
  along with low strain modulus was measured at
  different bulk stresses by means of seismic
  tests. The predicted modulus of the material from
  numerical models was normalized with the low
  strain modulus for a specific bulk stress.

6
Seismic modulus/resilient modulus comparison
7
Wave Propagation in Geomaterials

• Stiffness - Hertz Theory
• Youngs modulus
• Bulk modulus

8
Wave Propagation in Geomaterials

• Stiffness and Effective Stresses - Hertz Theory

where ß 1/6 for perfectly shaped elastic
spheres ß 1/4 for cone tip on a plane
surface ß 1/4 for plastic yielding at
spherical contact
(Duffy and Mindlin 1957)
9
Wave Propagation in Geomaterials
(Santamarina et al. 2001)
10
P-wave Velocity vs. Resilient Modulus
?s
so
so
so
so
?s
11
Materials
Portage sand
Grade 2 gravel
Class 5 gravel
RPM
Pit run sand and gravel
Breaker run
11
12
Test Setup Elastic wave velocity measurements
Load
MEMS accelerometers

• 90 cm x 60 cm x 60 cm box
• Velocity measured by exciting P-waves with a
  hammer

L
Soil
Large wood test cell
13
Research Hypothesis

• Modulus can be estimated by measuring geophysical
  parameters

• Modulus ? P-wave velocity
• Modulus depends on
• ? effective stress
• ? void ratio
• ? strain level
• ? water content, saturation, particle
  shape, etc.

Low-strain
High-strain
Emax
s
E
e
14
Relationship Between Velocity and Modulus
Modulus based on Velocity
Modulus based on Traditional Mr Test
D is the constraint modulus
mechanistic comparison
Background - Seismic Modulus/Resilient Modulus
Comparison
15
Resilient Modulus (Mr)
Resilient Modulus, Mr Test, Mr (MPa)
Data Camargo (2008)
Bulk Stress (kPa)
16
Modulus Based on P-wave Velocity
Shaded area is range of Mr curves
Modulus based on P-wave Velocity (MPa)
Bulk Stress (kPa)
17
Mechanistic Modulus Comparison
not performed in this research
18
I Stress Correction
All Data
Mr 0.174(Ds) 33.2 R2 0.69
Resilient Modulus, Mr Test (MPa)
Constraint modulus based on P-wave velocities
corrected for stress (MPa)
18
19
II Void Ratio Correction
All Data
Mr 0.173(Ds, e) 37.7 R2 0.79
Resilient Modulus, Mr Test (MPa)
Constraint modulus based on P-wave velocities
corrected for stress, void ratio (MPa)
19
20
III - Strain Correction Backbone Curve
Modulus based on P-waves Dmax
Mr Test
E/Dseismic (MPa/MPa)
General Correction
10-6
10-5
10-4
10-3
10-2
10-1
Shear Strain, g (mm/mm)
21
III - Strain Correction
11
All Data
Mr 0.59(Ds, e, e) 38.8 R2 0.88
Resilient Modulus, Mr Test (MPa)
Constraint modulus based on P-wave velocities
corrected for stress, void ratio, strain (MPa)
22
IV Conversion Constraint Modulus to Youngs
Modulus
11
All Data
Resilient Modulus, Mr Test (MPa)
Mr 0.95(Es, e, e) 38.0 R2 0.88
Youngs modulus based on P-wave velocities
corrected for stress, void ratio, strain (MPa)
23
Summary Modulus Comparison
Traditional resilient modulus test
Modulus based on mechanistic analysis and
individual correction factors Modulus based
on mechanistic analysis and global correction
factors
Modulus (MPa)
Portage sand
Grade 2
Class 5
RPM
Pit Run
Breaker Run
Soil Type
24
Simple Test
500 g mass
Load Plate
Direction of wave propagation along soil surface
MEMS accelerometer
Soil
5-gallon bucket
Seismic Modulus/Resilient Modulus Comparison
25
Simple Test Results Comparison
Traditional resilient modulus test
Modulus based on mechanistic analysis and global
correction factors large tests Modulus
based on mechanistic analysis and global
correction factors simple tests
Modulus (MPa)
Grade 2
Pit Run
Soil Type
26
Lessons Learned -Seismic Modulus/Resilient
Modulus Comparison-

• Mechanistic approach increases complexity, but
  improves correlation between P-wave velocity
  results and resilient modulus test results
• Resilient modulus is about 29 of constraint
  modulus based on P-wave velocity measurements
  when corrected for bulk stress and void ratio
• Large grain soils have a higher modulus than
  small grain soils
• Simple velocity tests are effective and moduli
  compare well to those of the more developed,
  large scale test and resilient modulus tests (14
  lower)

26
27
Large Scale Model Experiments
27
28
Introduction

• The base course elastic modulus measured in large
  scale model experiments is sensitive to the
  thickness of the layer being evaluated (i.e.,
  thicker layers have a higher elastic modulus at a
  given bulk stress).
• The sensitivity of modulus to layer thickness
  reflects the varying levels of strain in the
  layers having different thicknesses, which is
  known to affect the elastic modulus of granular
  materials (Seed and Idriss, 1970, Hardin and
  Drnevich 1972, Edil and Luh 1978).

29
Recycling of Pavement Materials

• Full Depth Reclamation involves pulverizing a
  deteriorated surface layer and mixing with
  existing base layer and then paving over this new
  recycled base material.
• When upgrading a gravel road to a paved road, use
  existing road surface gravel as base course.
• Improve engineering properties of recycled base
  course materials with the addition of Class C fly
  ash (a by-product of coal combustion)
• Time and money saved, while less impact to the
  environment.

30
Objective of Research Study

• Determine the Gravel Equivalency (GE) of recycled
  pavement material (RPM) and road surface gravel
  (RSG) with and without fly ash.
• GE factors of these recycled materials can then
  be used with MnDOT design methods
• GE factors are determined from the resilient
  modulus obtained from a Large Scale Model
  Experiment (LSME)
• RPM is a 50/50 mix of pulverized asphalt concrete
  and underlying granular base course

31
Recycled Materials Tested

• Class 5 Gravel
• RPM (recycled pavement material)
• RSG (road surface gravel)

32
Material Properties
33
Gravel Equivalency

• GE factors provide a means of equating the
  structural performance of all bituminous and
  aggregate courses constituting a pavement
  structure with respect to the structural
  performance of a select, high-quality, aggregate
  base (MnDOT Class 5 gravel).
• (HMA) (base) (subbase)
• GE a1D1 a2D2 a3D3 where a is GE factor and
  D is thickness of layer

34
Gravel Equivalency

• The GE factor can be determined from resilient
  modulus using the relationship for determining
  the layer coefficient of granular base materials
  from the AASHTO structural number (SN) method
• a2 0.249 log Mr 0.977
• where Mr is the summary resilient modulus (psi)
  at a bulk stress of 208 kPa.
• Then equating base course terms from the GE
  equation of Class 5 gravel (a conventional base
  material with subscript c) with GE factor equal
  to 1.00 to an alternative recycled material
  (with subscript a) and solving for this unknown
  GE factor

35
LSME Tests used to Determine Mr
36
LSME Setup and Operation

• Load Applied to Base Kenlayer analysis
  performed to determine the load at the base layer
  surface as a result of a 700 kPa tire load on a
  12.7 cm asphalt surface
• Applied for 10,000 cycles with a Haversine pulse
  (0.1 sec pulse / 0.9 sec rest)

• Deflection Data
• LVDTs measured deflections up to 0.005 mm at base
  surface and subgrade surface
• A back calculation was performed in MICH-PAVE, a
  finite-element program, to determine Mr

36
37
LSME test results
Resilient modulus of different recycled materials
w w/o fly ash Compared with Class 5 gravel
37
38
Resilient modulus vs. layer thickness
38
Note Resilient modulus did not change with
thickness for the material mixed with fly ash
39
Granular Equivalent Factor
39
40
Lessons Learned

• The main objective was to develop the Gravel
  Equivalency (GE) of recycled materials with and
  without fly ash stabilization.
• The GE for RPM was determined to be equal to 1.07
  and did not vary with base layer thickness. This
  response is similar to that of Class 5 gravel
  having a GE factor of 1.00.
• The GE factor of RSG varied with thickness and
  was less than 1.00 indicating that RSG has less
  desirable structural properties than Class 5
  gravel.
• The GE factor of the fly ash stabilized materials
  decreases with increasing base layer thickness
  with the constant modulus assumed for these
  materials, becoming approximately equal to 1.00
  at a thickness of 0.55 m.

41
Geogrid/Base Course Interaction
42
Engineering Problem

• Evaluate use of geogrid in the flexible pavement
  system to
• Reduce surface rutting
• Prevent pavement cracking

National Road Maintenance Condition Survey, 2003
5 cm
42
43
Engineering Hypotheses
Interlocking

• Geogrids
• increase bearing capacity
• enhance lateral resistance
• improve modulus

Base
Base or subgrade
Bender and Barenberg, 1978
43
44
Engineering Objectives

• Determine mechanistic relationship between
  low-strain modulus and resilient modulus
• Quantify the interaction zone between geogrids
  and base course
• Quantitatively assess benefits of geogrids for
  use as reinforcement in paved and unpaved roads

44
45
Research Hypotheses

• The interaction zone between geogrids and base
  course can be estimated from geophysical
  parameters

Modulus ? P-wave velocity Lateral confinement ?
Rotation
46
Modeled Rotation Tensor (PLAXIS)
Measured Rotation Angle (Lab Test)
Vin 5 V (split between accelerometer axes)
PLAXIS calibrated based on surface deflections
47
Measuring Rotation at the Plate Edge

• Test Setup

Load Plate
150 mm
MEMS accelerometers
20 - 25 mm
Geogrid
TENSION APPLIED
25 cm
Geogrid/soil Interaction
48
Shear Strain Results (g)
Load plate
0
geogrid
Shear Strain, g (mm/mm)
10
40
20
Depth (cm)
30
gmax 36 mm/mm
gmax 35 mm/mm
40
0
geogrid
10
geogrid
20
Depth (cm)
0
30
gmax 27.8 mm/mm
gmax 38.6 mm/mm
40
0
10
20
30
40
50
0
10
20
30
40
50
Distance from Load Plate (cm)
Distance from Load Plate (cm)
49
Rotation Results Zone of Influence 7 mm
surface displacement - grade 2 gravel
No Geogrid 7.5 cm
10 cm 15 cm
results
model
PLAXIS
50
Rotation Results Portage Sand
Zone of Influence?
51
Velocity Results Modulus Change

550 kPa surface load
550 kPa surface load
DVunreinforced
DVreinforced
DVreinforced/DVunreinforced 7.5 cm 1.2x 10 cm
1.6x 15 cm 1.2x
DEreinforced/DEunreinforced 7.5 cm 1.4x 10 cm
2.6x 15 cm 1.4x
52
Lessons Learned -Geogrid/Soil Interaction-

• Rotation results most effectively constrain the
  zone of influence
• The zone of influence is lt5 cm on either side
  of geogrid reinforcement
• The zone of influence depends on the vertical
  position of the geogrid and shifts up with
  increasing depth of reinforcement
• Soil rotation (shear) depends on the interlock
  between the geogrid and soil (i.e. particle size)
  
• Velocity results do not appear to effectively
  constrain a zone of influence around geogrid
  reinforcement, although a change in modulus is
  visible (up to 2.6x more than stress changes)

53
Acknowledgements

• Prof. Tuncer Edil
• Prof. Craig H. Benson
• Ali Ebrahimi
• Craig Schuettpelz
• Brian Koostra
• Xiadong Wang
• WisDOT WRHP (H. Bahia and A. Henz)
• MRUTC

54
(No Transcript)
55
Recommendations

• Does the presence accelerometers have an
  influence on results?
• Perform field seismic tests to determine modulus
  in the field and compare to lab tests
• Perform tests with many different geogrids and
  plate sizes to determine dimensionless influence
  of geogrid

56
NON-LINEAR Grain Contact Behavior-(Hertz Theory)-
P 0
P gtgt 0
d 2a2/R
R
s 0
a
s f (P, a, r)
Area ? ?(PR)
Area 0
Background
56
57
Surface Deflections

Geogrid/soil Interaction
57
58
Horizontal Displacement Results (ux)
Load plate
0
geogrid
10
Horizontal Displacement, Ux (mm)
2
20
Depth (cm)
30
umax 1.5 mm
Umax 1.5 mm
umax 1.6 mm
40
0
geogrid
10
geogrid
0
20
Depth (cm)
30
umax 1.8 mm
umax 2.0 mm
40
0
10
20
30
40
50
0
10
20
30
40
50
Distance from Load Plate (cm)
Distance from Load Plate (cm)
59
Picking Arrival Times
Akaike Information Criteria
Cross Correlation
Manual Picking
Materials/methods
59
60
Comparison of Picking Schemes
Velocity (m/s)
Materials/methods
60