# USE OF FWD

Document Sample

```					     In-Situ Validation of
Mechanistic Pavement
Finite Element Modeling

Mostafa A. Elseifi
Pyeong J. Yoo
Acknowledgement

VTRC/VDOT
Kevin McGhee and Tom Freeman

Michelin Research and Development
Ibrahim Janajreh

VTTI/VT
Introduction
• The multi-layer elastic theory is the
most popular method to predict
• Various software utilizes the multi-layer
elastic solution (e.g., BISAR3, VESYS,
etc.)
• Many assumptions are questionable, but
are usually accepted for simplicity
Assumptions
• Homogeneous, isotropic, and elastic
layers: E, n
• Weightless materials
• Infinite layers horizontally
• Finite layer thicknesses; except for
• Static loading over a circular area
• Compatibility of strains and stresses
satisfied at interfaces
The Finite Element Method
• The FE method “approximates the
behavior of a continuum by an assembly
of finite elements”
• More powerful and flexible approach to
simulate pavement response to
• Has been used successfully for
pavement applications
Theoretical Methodologies
Method         Layered
Finite Element
Theory
Dimension       Axis           2D, 3D
Pressure       Uniform    Unif. or Nonunif.
Bonding          Full       Full/Friction
Dynamic          No             Yes
Material       Elastic     Elastic/Visco-
elastic …
Objectives
• Determine the accuracy and
convergence of a typical FE model
• Compare the results against the
elastic layered theory and field-
measured pavement responses
• Modify the FE models to better
simulate pavement response to

• Adaptable Facility for Transportation
Research and Evaluation
• 2.5km Long: 12 Flexible Pavement Sections
and One CRCP
• Instrumentation during Construction
• Rain and Snow Towers
•   3 Load Levels (L1, L2, L3)
•   3 Tire Inflation Pressures (80,
95, 105 psi) (550, 655, 725 kPa)
•   4 Speeds (5, 15, 25, 45 mph) (8,
25, 40, 72 km/h)
Instrument Responses
400
Transverse Strain

300
Time Retardation
(mm/m)

200

100

0
0.0    0.5      1.0       1.5                      2.0
Time (sec)
Longitudinal Strain   600

450
(mm/m)

300

150

0
0.00   0.10   0.20   0.30   0.40   0.50
-150
Time (sec)
Instrument Responses
600                                                             Strain
Transverse Strain

400                                                          Accumulation
(mm/m)

200

0
0.0   0.5      1.0                 1.5        2.0
Time (sec)
Vertical Stress

100
80
(kPa)

60
40
20
0
0.0            0.5        1.0
Time (sec)
Pavement Responses
• Instrument responses indicate the
viscoelastic nature of HMA: time
retardation, relaxation with time, and
asymmetry of the response
• Pavement response is different in the
transverse direction than in the
longitudinal direction
• Longitudinal strain was greater than
transverse strain for the dual-tire
configuration
Theoretical Formulation
• Phase I: Sensitivity analysis to ensure
convergence of the developed FE model
– Compare the results to the layered theory
– Compare the results to field measurements

• Phase II: Introduce modifications to the FE
model to improve the prediction of
pavement responses
Section B

Surface Mix (SM-9.5D – 38mm)
• Typical interstate
highway pavement                Base Mix (BM-25.0 – 150mm)

design
Asphalt-Treated Drainage Layer
• Heavy instrumentation:                (OGDL – 75mm)
strain gages, pressure
21A Cement Stabilized Base layer
cells, thermocouple, etc.        (21B – 150mm)

21B Aggregate Subbase Layer
(21B – 175mm)
FE Model (ABAQUS)
250     622     250

500

Area

500
Y
1122              X
Sensitivity Analysis
• Two criteria for convergence:
 The finite element model converges to the
continuum solution
 Jumps in stresses across inter-element
boundaries are acceptable
Case ID Thickness    dof      Elements Computation
Time (sec)
A       38.10     172,476     42,916     3145
B       19.05     212,226     55,656     4871
C        9.53     291,726     81,136     6599
D        4.76     450,726    132,096    18565
E        2.38     760,776    231,468    30551
Sensitivity Analysis
100
Jump in Vertical Stress

3D FE    Accuracy Limit
80
Jump in Vertical
60                                                                    Stress at the
(kPa)

40                                                                     Interfaces
20

0
38.10    19.05   9.53    4.76    2.38
730
Element Thickness (mm)
720
Vertical Stress (kPa)
710                             FE       Bisar 3.0
700
Convergence of the
690
Vertical Stress with                                  680
Mesh Refinement                                      670
660
650
38.10    19.05    9.53      4.76        2.38
Element Thickness (mm)
Sensitivity Analysis
Vertical Stress (kPa)
0           200           400           600   800
0
20       Kenlayer
40
Depth (mm)

3D FE
60
80       Bisar
100
120
140
160
180
200
Measured vs. Calculated

Layered Theory = -0.18
3D FE = -0.995
200
Measured
160
Strain (mstrain)

BISAR 3.0
120                                       3D FE

80

40

0
Under BM-25.0 -   Under BM-25.0 -   Under OGDL - TS
TS                LS
Depth
Results of Phase I
• The sensitivity analysis indicate that an
element thickness of 9.53mm provides an
acceptable level of accuracy
• The layered theory and its analogous FE
model poorly predict pavement responses
• In order to improve pavement prediction,
several modifications are necessary
Suggested Modifications
• Quasi-static analysis (Speed = 8km/h)
• Friction contact between the layers
(Mohr-Coulomb Theory)
• Linear springs to simulate subgrade
support
• Viscoelastic time hardening model to
simulate HMA constitutive behavior
T1
692kPa

T2
904kPa

T3
859kPa

T4
895kPa
T5
692kPa
T5

692kPa
T4

904kPa
Tire Contact Area

T3

859kPa
T2

895kPa
T1

692kPa

1
0.9
Normalized Vertical Stress

0.8                      Dual Tire
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.08 -0.06 -0.05 -0.03 -0.02       0   0.016 0.032 0.048 0.064 0.08
Time (sec)
Model Prediction
Material Characterization
• Viscoelastic constitutive behavior based
on variable creep load test:
– Apply 10 different load levels while holding
each load for 60sec
– Impose a recovery rest period of 350sec


  A t
   c

 c = Creep strain rate
m   n 1

   = Stress
A, m, n = fitting parameters
Measured vs. Calculated
Vertical Stress (kPa)

800
Measured
600                  BISAR 3.0
Modified 3D FE
400

200

0
Under   Under   Under                   Under
WS     OGDL      21A                    21B
Depth                                                  Measured
Strain (mstrain)   200
160                         BISAR 3.0
Modified 3D FE
120
80
40
0
Under BM-    Under BM-    Under OGDL
25.0 - TS    25.0 - LS      - TS
Depth
Model Results
80              T1
70
T2
Transverse Strain

60
T3
(mstrain)

50
40              T4
30              T5
20
10
0
0   0.05    0.1     0.15   0.2                    0.25    0.3   0.35
Time (sec)

100          T1
Longitudinal Strain

T2
T3
(mstrain)

50
T4
T5
0
0    0.05   0.1   0.15   0.2   0.25   0.3   0.35
-50
Time (sec)
Surface Transverse Strain (T=25°C)
Section B
80.0
Transverse Surface Strain

60.0

40.0

20.0

0.0
0.0   200.0       400.0       600.0   800.0
-20.0

-40.0

-60.0

-80.0
Distance (mm)
HMA Rutting Damage
0.00

Surface Deflection (mm)
-0.02

-0.08

-0.10

-0.12
0   48 96 144 192 240 288 336 384 432 480 528
Distance from the left edge of the tire (mm)

Conclusions
• The multi-layer elastic theory
poorly predicted pavement
• Several modifications are
necessary to improve FE
simulation
Conclusions
• Anisotropic characteristics of HMA are
critical in simulating pavement
responses in the lateral directions

• A time hardening model was
successfully used to simulate time
retardation of HMA in the transverse
direction, and fast relaxation in the
longitudinal direction

```
DOCUMENT INFO
Shared By:
Categories:
Stats:
 views: 19 posted: 3/28/2011 language: English pages: 32