i

FINAL REPORT


MID-ATLANTIC UNIVERSITIES TRANSPORTATION
CENTER





Evaluation of Load Transfer Efficiency Measurement







Samir N. Shoukry, Ph.D.
Departments of Mechanical and Aerospace/
Civil and Environmental Engineering
Tel: (304) 293-3111 Ext 2367
Fax: (304) 293-6689
Email: shoukry@cemr.wvu.edu

Gergis W. William, Ph.D., P.E.
Mourad Y. Riad, MSCE
Department of Civil and Environmental Engineering


West Virginia University
College of Engineering and Mineral Resources






The contents of this report reflect the views of the authors, who are responsible for the facts and
accuracy of the information presented herein. This document is designated under the sponsorship of
West Virginia Department of Transportation, Division of Highways in the interest of information
interchange. The U.S. Government assumes no liability for the contents or use thereof. This report
does not constitute a standard, specification, or regulation. The contents do not necessarily reflect
the official views or policies of the State or Federal Highway Administration.




ii
Technical Report Documentation Page
1. Report No. WVU-2002-04
2. Government Accession No. 3. Recipients Catalog No.
5. Report Date July 2005
4. Title and Subtitle
Evaluation of Load Transfer Efficiency Measurement
6. Performing Organization Code
7. Author(s)
Samir N. Shoukry , Gergis W. William, Mourad Y. Riad
8. Performing Organization Report No.
10. Work Unit No. (TRAIS)
9. Performing Organization Name and Address
West Virginia University,
Department of Civil and Environmental Engineering
Morgantown, WV 26505-6103.
11. Contract or Grant No.
13. Type of Report and Period Covered

12. Sponsoring Agency Name and Address

Mid-Atlantic Universities Transportation Center
MAUTC-WVU
14. Sponsoring Agency Code


15. Supplementary Notes
Sponsored by Mid-Atlantic Universities Transportation Center

16. Abstract
This report documents an evaluation of load transfer efficiency (LTE) of dowel jointed concrete
pavements. Measurement of load transfer efficiency of transverse joints in concrete pavements is universally
conducted using FWD device. LTE is an important parameter affecting pavement performance. Due to the
importance of the results for maintenance decisions, the accuracy of the measurement technique is investigated in
this report.
The availability of instrumented dowel fitted concrete slabs in West Virginia Smart Road (Corridor H,
Route 33 as well as Goshen Road, West Virginia), offers a unique opportunity to examine the accuracy of
determining the load transfer efficiency of transverse joints using FWD. For this purpose, FWD tests were
conducted on both pavement sites at different times during years 2003, 2004, and 2005. Thirty transverse joints
were tested along the slab edges as well as along the wheel-path. Trend analysis was performed to evaluate the
effect of design features and site conditions on LTE. Key findings from this study:
1. Load transfer efficiency was found to be a complex parameter that depends on many factors that include load
position, testing time, slab temperature, and load transfer device.
2. Testing time and season was found to have a significant effect on the measured load transfer efficiency.
3. The slabs fitted with 32 mm (1.25 in) diameter dowels displayed higher variability of the measured load
transfer efficiency than those fitted with 38 mm (1.5 in) diameter dowels.
4. Joint opening changes daily and seasonally as the ambient temperature changes. As the amount of joint
opening increases due to slab contraction during winter, the measured load transfer efficiency generally
decreases.
5. Poor correlation was found between the deflection-based load transfer efficiency and the percentage of the
load transferred through the load transferring devices mounted across the transverse joint.


17. Key Words
Dowel bars, Transverse joints, Load transfer Efficiency, Joint Testing,

18. Distribution Statement

19. Security Classif. (of this report)
Unclassified

20. Security Classif. (of this page)
Unclassified

21. No. of Pages


22. Price

Form DOT F 17007.7 (8-72) Reproduction of completed page authorized
iii
TABLE OF CONTENTS


ABSTRCT
TABLE OF CONTENT
LIST OF FIGURES
LIST OF TABLES

CHAPTER ONE INTRODUTION
1.1 Background
1.2 Mechanistic Modeling of Load Transfer Efficiency
1.3 Measurement of Load Transfer Efficiency
1.4 Objective
1.5 Research Methodology

CHAPTER TWO PAVEMENT SECTION AND DATA COLLECTION
2.1 Pavement Test Section
2.2 FWD Loading Tests
2.3 Effect of Joint Opening
2.4 Effect of Slab Temperature
2.5 Effect of Temperature Gradient

CHAPTER THREE MEASUREMENT OF DOWEL SHEAR FORCES
3.1 Introduction
3.2 Goshen Road Pavement Site
3.3 FWD Testing
3.4 Dowel Bar Shear
3.5 Distribution of Shear Forces among Dowel Bars

CHAPTER FOUR SUMMARY AND CONCLUSIONS

REFERENCES


ii
iii
iv
v

1
1
3
5
6
6

8
8
13
20
20
21

22
22
22
23
24
25

31

32















iv
LIST OF FIGURES


Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5

Layout of Elkins Instrumented Pavement Site
Typical Slab Instrumentation System
Recorded Temperature Differential and Dowel Bending Moment
Transverse Joint Opening along Instrumented Pavement Section
Recorded Mean Slab Temperature and Joint Opening
Load Transfer Efficiency Versus Joint Opening
Load Transfer Efficiency Versus Slab Temperature
Temperature Differential Versus Load Transfer Efficiency
Construction of Instrumented Slabs
FWD Testing of Transverse Joints
Typical Shear Force History in Dowel Bars
Shear Forces Distribution in Dowel Bars
Shear Forces Distribution in Shokbars
9
10
10
11
12
20
21
21
23
24
25
27
28
v
LIST OF TABLES


TABLE 2.1
TABLE 2.2
TABLE 2.3
TABLE 2.4
TABLE 2.5
TABLE 2.6
TABLE 3.1
TABLE 3.2
TABLE 3.3


Load Transfer Efficiency Data Measured on September 2, 2003
Load Transfer Efficiency Data Measured on March 22, 2004
Load Transfer Efficiency Data Measured on April 28, 2004
Load Transfer Efficiency Data Measured on June 15, 2004
Load Transfer Efficiency Data Measured on August 30, 2004
Load Transfer Efficiency Data Measured on June 15, 2005
Shear Forces in Instrumented Dowels and Shokbars
Shear Forces in Dowel bars and Shokbars (N)
Comparison of Deflection Load Transfer Efficiency and Actual Load Transfer
Efficiency
14
15
16
17
18
19
26
29

29




1

CHAPTER ONE

INTRODUCTION



1.1 Background

As a wheel load is applied near a transverse doweled joint in a PCC pavement, both
loaded and unloaded slab deflect since a portion of the load applied to the loaded slab is
transferred to the unloaded one through the dowel bars. As a result of the presence of load
transfer devices, deflections and stresses in the loaded slabs may be significantly less than those
induced in slab with free edge. The magnitude of reduction in stress and deflections by a joint
depends on its load transfer efficiency.

The term load transfer efficiency is used to express the ability of a joint to transmit part of
the applied load on the loaded slab to the adjacent unloaded one (Ioannides and Krovesis, 1992).
Several formulae for calculating load transfer efficiency have been adopted by various
researchers to provide quantitative measures of pavement-system response:


1. Deflection load transfer efficiency, LTE

:

This is the most common measure for load transfer efficiency.
l
u
d
d
LTE =

.. (1.1)
where d
u
and d
l
are the vertical displacements of the unloaded and loaded slabs respectively
measured at the joint on the top of the slabs. Another definition for the deflection load transfer
efficiency is the one first proposed by Teller (Teller and Sutherland, 1936; Teller and Cashell,
1959) and is still in use by researchers (American Concrete Paving Association, 1991, McGhee,
1995, Colley and Humphrey, 1967; Raja and Snyder, 1991; Khazanovich and Gotlif, 2003) as:

l u
u
d d
d
LTE
+
=

.. (1.2)
If a joint exhibits a poor ability to transmit load, the deflection of the unloaded slab will
be much less than that of the loaded one. In this case, both LTE indexes will have values close
to zero. On the other hand, if the joint has a good ability to transfer load, the deflection of both
sides will be close to each other and both indexes have values close to 1.0. The two indexes are
related to each other by the equation:

+
=

LTE
LTE
1
1
1 2 . (1.3)

Therefore, these two indexes are equivalent, and if one is known, the other can be easily
calculated. In this study, we will adopt the first index defined by Equation 1 since it is much
2
more widely used and accepted by the AASHTO Pavement Design Guide (1993) and its
supplement for Concrete Pavement Design (AASHTO, 1998).


2. Stress Load Transfer Efficiency, LTE

Load transfer efficiency can be computed based on stress using formulae similar to those
developed for deflection-based load transfer efficiency. The most commonly used equation for
stress-based load transfer efficiency is given as:

l
u
LTE

= ... (1.4)
where
u
and
l
are bending stresses of the unloaded and loaded slabs respectively. Also, the
following expression was proposed by Sutherland and Cashell (1945):
i f
j f
LTE

= ... (1.5)
Where
f
= stress for a given load applied at a free edge,

j
= stress for a given load applied at the crack or joint edge,

i
= stress for a given load applied at the slab interior.

The stress-based load transfer efficiency indicates the reduction in bending stress at the
joint caused by the presence of load transfer devices. Previous studies indicated that there is no
one-to-one relationship between stress-based and deflection-based load transfer efficiency
indexes. Korbus and Barenberg (1979) developed a relationship that can be used to correlate
these two efficiency parameters. Such a relationship was subsequently adopted by the AASHTO
Pavement Design Guide (AASHTO, 1993). Because of the difficulty inherited in stress
measurement in concrete slabs, the fact that stress in concrete slab is influenced by the slab
geometry and applied load configuration, and the ease of rapidness of deflection measurement,
the deflection-based load transfer efficiency is commonly used to measure the load transfer in
concrete pavements.


3. Transferred Load Efficiency, TLE

TLE, a dimensionless variable, was introduced by Ioannides et al. (1990) to express the
ratio of the load transferred across the entire length of joint P
T
to the total applied load P, and
given by:
P
P
TLE
T
= .. (1.6)
Additionally, Ioannides et al. (1990) presented the load distribution factor, f
d
, that indicates the
load share of any given dowel bar from the total transferred load through a joint as expressed by
the equation:
T
i
d
P
P
f = .. (1.7)
3
where P
i
is the load transferred by a particular dowel bar.


1.2 Mechanistic Modeling of Load Transfer Efficiency

Westergaard (1929) presented the first rational method of designing joints in concrete
pavements by providing a theoretical means for computing the shearing force in each active
dowel bar as well as the critical stresses that occur in the slab under the edge load. However, his
analysis was limited to the hypothetical case in which the deflections of the two sides of the joint
are equal (LTE=100%). Westergaard concluded that only the first two dowel bars from each
side are active in transferring the load to the adjacent slab. Friberg (1938) assumed that dowels
at a distance greater than 1.8 times the radius of the relative stiffness, l , from the center of the
load are inactive in transferring any load through the joint. Westergaard (1926) defined the
radius of the relative stiffness as:

4
2
3
1 12 k ) (
Eh
l

= (1.8)

Where:
E= Elastic modulus of concrete.
h= Thickness of the concrete slab.
= Poisson's ratio of concrete.
k= Modulus of subgrade reaction.

Friberg (1938) also assumed that the shear force in an active dowel decreases as a linear
function of distance. Stresses induced in concrete were calculated based on Timoshinko's
solution for the dowel bar as an infinite beam encased in an elastic medium. The concrete
surrounding the dowel bar was idealized with a single parameter, the modulus of dowel support,
K. Values of K ranging between 300,000 and 3,000,000 pci were suggested. A value of
1,500,000 pci is commonly used as a typical value (Yoder and Witczak 1975 and Haung 1993).
The relative stiffness, , of a bar embedded in concrete is given by:

4
4EI
Kb
= ... (1.9)

Where:
b= diameter of the dowel.
E= modulus of elasticity of the dowel.
I= moment of inertia of the dowel.

If the joint width is designated w, Friberg calculated the deflection of the dowel at the
joint face, y
0
, and the bearing pressure (
b
) on the concrete as:
( ) w
EI
P
y
t
o

+ = 2
4
3
.. (1.10)
4
( ) w
EI
P K
y K
t
o o

+ = = 2
4
3
. (1.11)

It is significant to point out that the above stress is a compressive stress that is well below the
concrete compressive strength.

The lack of sound experimental or theoretical procedures for determining both the
modulus of dowel bar support K and the amount of load transferred in any dowel made it
impossible to accurately calculate the stresses or the deflection using the Firbergs approach.
Kushing and Fremont (1940) suggested that the range of distribution of load transfer hardly
extended over more than l each side of the load, while Sutherland (1940) reported experimental
results that support Westergaard's conclusion that only the dowels near the load are effective in
load transfer.

In an attempt to better understand the role of dowels in load transfer, Two Dimensional
Finite Element (2D-FE) was introduced. Several finite element programs have been developed
including: ILLI-SLAB (Tabatbaie and Barenberg, 1978; Krovosis, 1990), JSLAB (Tayabji and
Colley, 1986), KENSLABS (Huang and Wang, 1974), WESLIQUID (Chou, 1981),
WESLAYER (Chou, 1981), KENLAYER (Huang, 1993), FEACONS IV (Tia et al., 1987; Wu
and Larsen, 1993), and ISLAB2000 (Khazanovich et al., 2000, Beckemeyer et al., 2002). All
these programs simulated the traffic loads as a static load. Earlier 2D FE programs utilized the
single parameter K to simulate the dowel-concrete interaction. Using ILLI-SLAB, Tabaabaie
and Barenberg (1980) supported linear decrease of shearing forces in dowels with distance from
center of load application; however the distribution length was reduced to 1.0l . The use of
reduced distribution length was also supported by the results of Heinrichs et al. (1989) and
Huanng (1993). Results from J-SLAB (Tayabji et al. 1986) indicated that only the dowel
diameter, modulus of subgrade reaction, and modulus of dowel support significantly effected
pavement stresses and deflections.

In 2D FE programs, dowel bars were simulated using beams elements or linear and
torsion spring elements. In such a representation, it is assumed that the displacements at the ends
of the beam or spring element are equal to those slab elements. Nishizawa et al. (1989) rejected
this assumption and suggested modeling dowels using a combination of linear and torsional
springs that simulated both the dowel bar and the aggregate interlock. Guo et al. (1995)
integrated the modulus of dowel support in a two-dimensional finite element formulation in
order to simplify dowel bar simulation in 2DFE models of rigid pavements and to account for
dowel looseness. The 2DFE-calculated shear force in the dowel bar is substituted into Equation
11 to calculate the bearing stress in the concrete surrounding that dowel. However, the triaxial
state of stress that takes place around the dowel could not be viewed. Further, the horizontal
friction force between the dowel bars and the surrounding concrete cannot be modeled.

Channakeshava et al. (1993) developed a nonlinear static 3D-FE model to study the
combined effect of a linear thermal gradient and a static wheel load. Dowel bars were simulated
as beam elements mounted on elastic springs of constants equal to the modulus of dowel support
K. Effect of different extents of loss of subgrade support and different dowel-concrete interface
characteristics were considered. The study concluded that nighttime curling is a critical loading
5
case since truck traffic is heaviest during the night; in addition, it causes loss of support under the
joints. They also reported the loss of support under transverse joint due to nighttime curling
would result in an increase of the load transfer efficiency of the joint. The authors reported a
large concentration of tensile stresses in the elements below the dowel bars. A finding that does
not explain the small horizontal cracks observed in Fribergs (1938) experiments.

Davids et al. (1997, 1998) developed a 3D-FE program specifically for rigid pavements.
Concrete slabs were modeled using 8-node solid brick elements. An embedded beam element
formulation was developed to precisely locate the dowel bar within the finite element mesh
irrespective of the slab mesh lines. This model also permits rigorous treatment of dowel
looseness, which affects the load transfer efficiency. Later, the program was modified so that a
Winkler foundation can be specified between the dowel and concrete instead of explicitly
modeling dowel looseness (Davids, 2000). Examining the load transfer efficiency at doweled
transverse joints for both curled and flat slabs revealed that slab curling has a large influence on
dowel-concrete bearing stress (Davids, 2001). However, both simplified dowel modeling
techniques employed in these models are still incapable of predicting the triaxial state of stress
that develops in the concrete surrounding the dowel bars.

The tendency to simplify the dowel bar representation in 3DFE models using beam
elements supported on springs whose modulus is the modulus of dowel support became the
common feature of the majority of 3DFE models developed during the past decade (Kuo et al.,
1996; Kennedy and Everhart, 1996, 1997, 1998; Sargand and Breegle, 1998, Davids, 2001).
This simplification is justified by the fact that such models were used to investigate the overall
slab stresses and deflections induced under mechanical and/or thermal loads. However, when
the modeling objective is to examine dowel-concrete contact stresses, detailed modeling of
dowel bars and their interfaces with concrete becomes essential.

A 3DFE model of dowel-jointed concrete pavement that featured detailed three
dimensional modeling of dowel bars was developed by Shoukry and William (1998) to examine
the effect of dowel looseness on load transfer efficiency and slab stresses. The general 3D-FE
code LS-DYNA was used as the equation solver. All pavement layers as well as dowel bars
were modeled using 8-node solid brick elements. Dowel-concrete interfaces as well as slab-base
interface were simulated using sliding interfaces with friction that allows for slab-base
separation. The results indicated that the localized high stresses induced in the concrete
surrounding the dowels causes an elastic deformation. As a result, for a joint with intact dowel
bars, the maximum recorded load transfer efficiency was 91 percent. The model results also
indicated that such high stresses around dowels initiate cracks around the dowel, which is the
main reason for dowel looseness (Shoukry et al., 2002).


1.3 Measurement of Load Transfer Efficiency

Nondestructive deflection devices are currently used to evaluate the in situ load transfer
efficiency of rigid pavement joints as well as cracks slabs (AASHTO, 1993). Such techniques
allow the engineer to evaluate the actual joint performance in the field relative to their expected
performance in the design phase.
6

All nondestructive deflection devices are suitable for evaluating the load transfer
efficiency at any joint provided that deflection sensors can be mounted close enough to each
other across the joint to measure vertical displacement of both loaded and unloaded slabs. Once
the unloaded and loaded slab deflections are measured, deflection-based load transfer efficiency
can be directly determined using Equation 1.

Falling Weight Deflectometer (FWD) is the most commonly used nondestructive testing
device for load transfer efficiency measurements. Because of its ease of use, rapidness and
accuracy of measurements, the AASHTO Pavement Design Guide (AASHTO, 1993)
recommends the use of such a device for deflection measurements required for designing the
required thickness of the asphalt overlay or bonded concrete overlay. The AASHTO Design
Guide (1993) recommends measuring the joint deflection along the outer wheel-path. The
loading plate should be placed on one side of the joint with its edge touching the joint.
Deflections are measured at the loading plate center and at 12 inches from the center (6 inches
from the joint edge on the adjacent slab). Deflection load transfer Efficiency can be computed
from the following Equation:
B 100

=
l
u
d
d
LTE

.. (1.12)

Where LTE

= deflection load transfer efficiency, percent,

d
u
= unloaded side deflection, inches,
d
l
= loaded side deflection, inches,
B = slab bending Correction factor.

The slab bending correction factor, B, is required since the deflections d
0
and d
12
are
measured 12 inches apart, which would not be equal even if measured in the interior of a slab.
An appropriate value for the correction factor may be determined from the ratio d
0
to d
12
for
typical center slab deflection basin measurements. Typical values for B are within the range
from 1.05 to 1.15 (AASHTO, 1993).


1.4 Objective

The main objective of this study is to examine the accuracy of measuring load transfer
efficiency of transverse joints in dowel jointed concrete pavements using FWD and examine how
such a deflection-based load transfer efficiency correlates with the actual shear force transmitted
through mechanical load transfer devices (traditional dowel bars or shokbars) across transverse
joints.


1.5 Research Methodology

The study was performed on West Virginia instrumented pavement sections located on
Route 33 near Elkins and in Goshen Road, West Virginia. Each instrumented pavement section
is fitted with thermistor tree that enables accurate measurement of the temperature gradient
7
profile at the time of load transfer efficiency measurement. The slabs are also fitted with
instrumented dowels that allow measurement of the actual shear load transmitted by the dowels
during the application of impact loading. FWD tests were conducted on slabs fitted with 1.5 inch
diameter epoxy coated dowel bars and shokbars. The deflection-based load transfer efficiency
was measured at the corner dowels, wheel-path dowels and center dowels. The results obtained
from FWD sensors were compared with the direct measured results using bonded strain gauges
on the dowel bar. Measurements were conducted during different times of the year to investigate
the seasonal effect on the measured data.

8

CHAPTER TWO

PAVEMENT SECTION AND DATA COLLECTION



2.1 Pavement Test Section

A heavily instrumented concrete pavement section was constructed in September 2001 along the
east bound lanes of Robert C. Byrds Highway (Route 33) near Elkins, West Virginia, USA. The
135 m (450 ft) long section consists of 30 consecutive dowel jointed concrete slabs. Each slab is
0.28 m (11 in.) thick, 4.5 m (15 ft) long, and 3.6 m (12 ft) wide. The slabs are placed on top of a
0.10 m (4 in.) asphalt-stabilized free-drainage base constructed over 0.15 m (6 in.) compacted
gravel. Six slabs out of the thirty were fitted with systems of sensors designed for continuous
monitoring of slab response to temperature variations as shown in Figure 2.1. The
instrumentation plan was set in order to collect data for several key performance parameters that
evaluate the behavior of pavement slabs due to diurnal, seasonal, and structural inputs. For this
purpose, strains in various locations along slab centerlines, joint openings, temperature profiles,
dowel bar moments and axial forces were recorded at specific intervals starting at the time of
placing concrete. The instrumentation plan was aimed to focus sensors in selected slabs so as to
capture the full behavior of slabs in a mechanistic fashion rather than collecting redundancy in
data. The selection of the sensors was based on their functional characteristics, size, sensitivity,
accuracy, and past experience with successful performance. Figure 2.2 illustrates the
arrangement of the sensors installed in a typical instrumented slab prior to concrete placement. A
detailed instrumentation plan as well as a description of the sensors and data acquisition systems
could be found in (Shoukry et al., 2004 a, b).

The state of curling of the slab is recorded through dowel bending measurements of
instrumented epoxy coated dowels 38.1 mm in diameter and 45.7 cm long located at the slab
transverse joints. For collecting bending strains, two uniaxial strain gages were mounted on the
top and bottom of the dowel within the bar center. Each instrumented dowel went through a
careful procedure to mount the strain gages in their proper locations, with the goal of producing
the instrumented dowel with almost no change in its characteristics or its surface texture. Each
instrumented dowel was calibrated in the laboratory to check the accuracy of the collected
readings and the theoretically calculated straining actions. This was achieved by applying pure
bending moments of known magnitudes, and verifying the response of the bars against
theoretical calculations. To calculate bending moments of the dowels the following equation was
used:

y
EI
M
b t

=
2
) (
(2.1)
Where: M = Moment of dowel
E = Youngs Modulus of steel
I = Moment of inertia of dowel

t
= Collected strain on top of dowel
9

b
= Collected strain on bottom of dowel
y = Distance to strain gage from neutral axis
The calibration procedure showed a maximum variation of 10 % between the actual applied
moment and that calculated from the gages for a range of bending moments up to 226 N.m (2000
lb.in). Temperature gradients across the slab depth are measured by a thermistor tree including
11 thermistors located at an interval of 2.54 cm (1 in) from the slab bottom. Figure 2.3 illustrates
the time histories of the temperature gradient and the corresponding bending moment measured
at the corner dowel.
T
r
a
f
f
i
c

D
i
r
e
c
t
i
o
n

(
T
o
w
a
r
d

E
a
s
t
)

Figure 2.1 Layout of Elkins Instrumented Pavement Site.
Shok bar
steel 1.25
Shok bar polymer
conc. 1.25
Shok bar polymer
conc. 1.5
Steel
dowel 1.5
Shok bar
steel 1.5
Steel
dowel
Instrumentation box
housing long term
monitoring data
acquisition systems
Instrumentation box
housing dynamic strain
gage connectors
Instrumented slab
Total of 30 slabs
10

Figure 2.2 Typical Slab Instrumentation System.
1400
1400
Figure 2.3. Recoded Temperature Differential and Dowel Bending Moment.
0
100
200
400
T
e
m
p
e
r
a
t
u
r
e

D
i
f
f
e
r
e
n
t
i
a
l
.

C

D
o
w
e
l

B
e
n
d
i
n
g

M
o
m
e
n
t
.

N
.
m

0 200 400 600 800 1000 1200
-200
Time, days
-100
0 200 400 600 800 1000 1200
Time, days
-8
-4
0
4
8
12
(a) Temperature Differential
(b) Dowel Bending Moment
-300
300
11

Figure 2.4 Transverse Joint Opening Along Instrumented Pavement Section.
0 1 2 3 4
5
Joint Opening, mm
0
5
10
15
20
25
30
35
J
o
i
n
t

N
u
m
b
e
r







After 1 day
After 2 days
After 3 days
After 7 days
12

Transverse joints were formed by saw-cutting a shallow groove approximately 50 mm in
depth five hours after concrete placement. As the concrete slab starts shrinking, cracks initiate
from the tip of such opened groove and propagate downward. Visual inspection of the opened
cracks at transverse joints during the seven days following concrete placement indicated that the
crack width at transverse joints is not constant along the pavement section as illustrated in Figure
2.4. This indicates that the amount of edge constraint differs from one transverse joint to another
along the road. Therefore, the concrete slab does not contract or expand symmetrically around
its center. Figure 2.4 also indicates that the amount of joint opening at each transverse joint
increases with time during the first week of pavement life.

To have a better insight into the change of the amount of joint opening with time, the
time-history of the joint opening recorded by the crack-meter installed at transverse joint No. 13
is shown in Figure 2.5 (b) together with the corresponding mean slab temperature shown in
Figure 2.5 (a). It can be noticed that after an initial increase in the amount of joint opening
0 200 400 600 800 1000 1200 1400
-10
0
10
20
30
40
Time, days
T
e
m
p
e
r
a
t
u
r
e
,

C

0 200 400 600 800 1000 1200 1400
Time, days
0
0.4
0.8
1.2
1.6
J
o
i
n
t

O
p
e
n
i
n
g
,

m
m

(a) Mean Slab Temperature
(a) Joint Opening
Figure 2.5 Recorded Mean Slab Temperature and Joint Opening.
13
during the first week to about 0.20 mm, it starts increasing as the temperature decreases during
winter until it reaches a maximum value of about 0.6 mm after 180 days, it then decreases during
summer to fluctuate around 0.2 mm (the initial opening) after 300 days. Such a change in the
amount of joint opening with time indicates a change in the amount of constrains along the slab
edges due to the presence of dowel bars. The constraining effect of the dowel bars at the
transverse joint opening is evident from observing the change in recorded dowel bending
moment due to the change in the temperature gradient. Figure 2.3 (a) illustrates the time history
of the measured difference in slab top and bottom temperatures. Over the monitoring period of
400 days (from September 2001 to June 2005), the difference between slab top and bottom
temperatures varied between 8 C to +10 C. The time history of dowel bending moments
developed in the corner dowel nearest to the shoulder during the same period of pavement life is
illustrated in Figure 2.3 (b).

The magnitude of the joint opening as well as the amount of its curling due to the
temperature gradient through its thickness will influence the joint load transfer efficiency. For
example, if the joint is tested while it is fully opened, the measured load transfer efficiency is
expected to be less than that measured if the joint is closed. Also, the load transfer efficiency
measured while the joint suffers upward curling will differ from that measured if the joint suffers
downward curling and is in contact with the base layer.


2.2 FWD Loading Tests

The thirty transverse joints were tested using FWD to measure load transfer efficiency.
Deflection tests were conducted along the pavement edge as well as along the wheel-path. The
tests were conducted on several days during the years 2003, 2004, and 2005. At each joint test,
three load drops are applied and the deflection basin is recorded for each drop. The load transfer
efficiency is calculated for each drop and an average value was taken as the measured load
transfer efficiency.

The results of the load transfer efficiency obtained from different tests are summarized in
Tables 2.1 to 2.6. The following observations can be made from the results presented in Tables
2.1 to 2.6:

The slabs fitted with 32 mm (1.25 in) diameter dowels displayed higher variability of the
measured load transfer efficiency than those fitted with 38 mm (1.5 in) diameter dowels.
Load transfer efficiencies measured along the wheel-path are generally higher than those
measured along the slab edges.
Shokpolymer out performs the shoksteel and traditional dowel bars. The lower values of
load transfer efficiencies measured for shoksteel are expected because of the clearance
between the sleeves and dowel.
The value of the measured load transfer efficiency along the wheel-path is higher than
that measured along the pavement edge.
Both Shoksteel and Shokpolymer designs preserved the load transfer efficiencies.
Although their measured load transfer efficiencies are close to those of the steel dowels,
14
they enabled a more uniform joint opening through their sections than regular steel
dowels.

TABLE 2.1 Load Transfer Efficiency Data Measured on September 2, 2003.

Slab mean temperature = 22.84 C.
Temperature Differential = -1.069 C.
Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 88.07% 92.56%
2 51.22% 91.37%
3 88.07% 88.60%
4 77.90% 89.85%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 78.72%
76.79% 15.11%
91.81%
90.84% 1.59% 0.0564
6 53.07% 85.11%
7 87.57% 88.28%
8 83.91% 90.61%
9 86.10% 91.52%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 -
77.66% 16.46%
87.83%
88.67% 2.52% 0.1050
11 83.81% 83.81%
12 86.81% 89.09%
13 92.90% 81.27%
14 85.36% 80.20%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 91.13%
88.00% 3.86%
80.12%
82.90% 3.77% 0.0718
16 90.45% 75.42%
17 90.45% 85.44%
18 88.83% 86.31%
19 84.44% 85.92%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 69.47%
84.73% 8.87%
65.44%
79.71% 9.18% 0.0701
21 83.15% 80.41%
22 88.26% 80.12%
23 83.55% 67.00%
24 91.96% 83.16%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 84.22%
86.23% 3.79%
55.70%
73.28% 11.66% 0.0222
26 91.29% 76.43%
27 83.35% 66.11%
28 89.12% 81.56%
29 94.60% 72.86%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 87.16%
89.10% 4.24%
65.49%
72.49% 6.85% 0.0016
15
TABLE 2.2 Load Transfer Efficiency Data Measured on March 22, 2004.

Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 77.4% 91.96%
2 87.2% 90.59%
3 88.5% 90.75%
4 68.0% 79.79%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 86.8%
81.6% 8.80%
89.10%
88.4% 4.94% 0.0297
6 88.1% 86.61%
7 86.5% 86.90%
8 86.2% 88.43%
9 84.2% 81.26%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 85.2%
86.0% 1.47%
88.55%
86.4% 2.98% 0.3998
11 91.1% 88.42%
12 82.6% 86.83%
13 86.2% 92.38%
14 93.1% 93.35%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 95.5%
89.7% 5.23%
92.97%
90.8% 2.96% 0.2876
16 87.7% 90.87%
17 95.4% 90.66%
18 72.9% 80.98%
19 83.4% 90.17%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 72.9%
82.5% 9.72%
86.79%
87.9% 4.20% 0.0752
21 75.7% 89.84%
22 75.7% 85.59%
23 63.9% 82.10%
24 86.3% 86.69%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 88.3%
78.0% 9.78%
85.10%
85.9% 2.80% 0.0617
26 86.6% 84.63%
27 80.7% 88.47%
28 73.4% 84.63%
29 63.5% 71.36%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 82.3%
77.3% 9.06%
87.45%
83.3% 6.89% 0.0267

Slab mean temperature = 4.30 C.
Temperature Differential = 1.05 C.


16
TABLE 2.3 Load Transfer Efficiency Data Measured on April 28, 2004.

Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 74.2% 87.92%
2 77.6% 89.76%
3 44.9% 61.80%
4 65.1% 84.23%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 78.9%
68.1% 14.07%
83.98%
81.5% 11.31% 0.0026
6 61.8% 79.09%
7 46.1% 51.81%
8 77.4% 85.03%
9 55.6% 81.68%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 61.7%
60.5% 11.41%
86.97%
76.9% 14.36% 0.0093
11 61.1% 83.68%
12 55.8% 78.32%
13 64.5% 87.19%
14 70.0% 87.70%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 70.3%
64.3% 6.16%
78.74%
83.1% 4.47% 0.0012
16 75.3% 89.43%
17 72.7% 86.14%
18 80.8% 93.66%
19 76.3% 94.51%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 78.6%
76.7% 3.07%
87.76%
90.3% 3.66% 0.0004
21 54.9% 72.38%
22 69.4% 85.35%
23 53.0% 77.24%
24 68.7% 82.95%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 61.1%
61.4% 7.57%
84.01%
80.4% 5.44% 0.0003
26 40.2% 72.27%
27 65.5% 72.27%
28 62.5% 86.68%
29 65.0% 84.09%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 64.5%
59.5% 10.89%
85.51%
80.2% 7.26% 0.0037

Slab mean temperature = 12.19 C.
Temperature Differential = 2.42 C.
17
TABLE 2.4 Load Transfer Efficiency Data Measured on June 15, 2004.

Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 57.9% 90.26%
2 86.5% 94.33%
3 62.1% 95.23%
4 80.2% 91.06%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 57.6%
68.9% 13.54%
78.49%
89.9% 6.70% 0.0081
6 86.7% 91.44%
7 80.2% 88.49%
8 84.4% 91.29%
9 81.1% 89.92%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 87.0%
83.9% 3.14%
89.09%
90.0% 1.31% 0.0038
11 88.5% 91.80%
12 93.6% 90.19%
13 87.6% 90.13%
14 93.7% 95.25%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 90.6%
90.8% 2.83%
92.43%
92.0% 2.09% 0.1880
16 95.3% 97.72%
17 92.0% 91.27%
18 74.4% 92.32%
19 91.0% 93.28%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 88.5%
88.2% 8.13%
91.87%
93.3% 2.58% 0.0998
21 86.7% 90.73%
22 91.1% 92.42%
23 83.0% 91.90%
24 88.7% 88.94%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 84.0%
86.7% 3.31%
91.21%
91.0% 1.34% 0.0292
26 88.9% 86.35%
27 87.0% 91.87%
28 79.6% 91.71%
29 80.1% 91.13%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 85.2%
84.2% 4.15%
89.60%
90.1% 2.30% 0.0436

Slab mean temperature = 27.02 C.
Temperature Differential = 3.33 C.

18
TABLE 2.5 Load Transfer Efficiency Data Measured on August 30, 2004.

Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 90.1% 91.81%
2 59.3% 67.69%
3 88.5% 89.55%
4 72.9% 87.42%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 72.3%
76.6% 12.82%
87.29%
84.8% 9.72% 0.0269
6 54.3% 61.68%
7 70.8% 88.26%
8 70.3% 84.42%
9 74.1% 89.78%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 81.1%
70.1% 9.86%
91.03%
83.0% 12.19% 0.0011
11 82.0% 82.47%
12 88.7% 89.78%
13 93.0% 90.75%
14 90.9% 84.61%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 89.5%
88.8% 4.15%
91.87%
87.9% 4.11% 0.2920
16 83.9% 92.06%
17 94.9% 93.69%
18 92.8% 91.74%
19 88.8% 91.01%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 69.7%
86.0% 10.04%
87.50%
91.2% 2.29% 0.1110
21 84.4% 88.84%
22 75.1% 90.05%
23 84.8% 87.59%
24 88.8% 92.41%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 76.7%
82.0% 5.81%
82.40%
88.3% 3.73% 0.0231
26 86.2% 89.58%
27 84.7% 84.77%
28 91.2% 91.36%
29 87.7% 90.99%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 73.9%
84.7% 6.52%
87.09%
88.8% 2.79% 0.0851

Slab mean temperature = 28.15 C.
Temperature Differential = 2.85 C.
19
TABLE 2.6 Load Transfer Efficiency Data Measured on June 15, 2005.

Edge Wheel-path
Type
Joint
No.
LTE Average
Standard
Deviation
LTE Average
Standard
Deviation
T Test
1 79.1% 87.35%
2 52.6% 71.75%
3 69.5% 81.81%
4 64.0% 80.06%
S
h
o
k
s
t
e
e
l

1
.
2
5

i
n
.

5 76.1%
68.3% 10.52%
82.35%
80.7% 5.67% 0.0033
6 59.3% 53.71%
7 74.6% 83.45%
8 78.4% 82.54%
9 80.1% 88.14%
S
h
o
k
p
o
l
y
m
e
r

1
.
2
5

i
n
.

10 84.2%
75.3% 9.59%
85.81%
78.7% 14.15% 0.1290
11 83.5% 78.70%
12 79.4% 89.48%
13 87.9% 83.77%
14 89.9% 83.92%
S
h
o
k
p
o
l
y
m
e
r

1
.
5

i
n
.

15 87.2%
85.6% 4.16%
89.10%
85.0% 4.45% 0.4286
16 90.5% 91.47%
17 90.2% 92.60%
18 84.7% 85.87%
19 81.6% 85.94%
S
t
e
e
l

D
o
w
e
l


1
.
5

i
n
.

20 71.9%
83.8% 7.65%
80.72%
87.3% 4.81% 0.0361
21 82.6% 82.30%
22 82.5% 88.03%
23 87.0% 84.69%
24 86.4% 86.41%
S
h
o
k
s
t
e
e
l

1
.
5

i
n
.

25 81.6%
84.0% 2.47%
85.54%
85.4% 2.13% 0.1983
26 86.4% 90.46%
27 84.3% 86.62%
28 80.9% 87.39%
29 85.6% 82.19%
S
t
e
e
l

D
o
w
e
l
s

1
.
2
5

i
n
.

30 81.8%
83.8% 2.40%
83.40%
86.0% 3.30% 0.1257

20
Loading position has a significant effect on resulting LTEs of individual joints especially
if the joints were fitted with 32 mm diameter dowels (T-Test values are less than 0.1).
The effect of the loading position on the load transfer efficiency is less in the case of joint
fitted with 38 mm diameter bars.

2.3 Effect of Joint Opening
The effect of change in joint opening on LTE was examined. Figure 2.6 presents the joint
opening recorded at the time of testing versus the deflection-based load transfer efficiency for the
joint fitted with regular 38 mm diameter dowels. As the joint opening increases, the load transfer
efficiency decreases. The limited amount of data available from this study indicates a linear
relationship with a moderate correlation between LTE and joint opening.


2.4 Effect of Slab Temperature
Figure 2.7 presents the mean slab temperature recorded at the time of testing versus the
deflection-based load transfer efficiency for the joint fitted with regular 38 mm diameter dowels.
As the slab temperature increases, the load transfer efficiency increases. The limited amount of
data available from this study indicates a linear relationship with a moderate correlation between
LTE and joint opening.

0
10
20
30
40
50
60
70
80
90
100
0 0.2 0.4 0.6 0.8 1.0 1.2
Joint Opening, mm
L
T
E
,

%
LTE = -7.7352 + 94.565
R
2
= 0.1738
LTE = -18.143 + 94.994
R
2
= 0.3502
Figure 2.6 Load Transfer Efficiency Versus Joint Opening.
Wheel-Path
Edge
21
2.5 Effect of Temperature Gradient
Figure 2.8 presents the temperature differential through slab thickness recorded at the
time of testing versus the deflection-based load transfer efficiency for the joint fitted with regular
38 mm diameter dowels. As the temperature differential increases, the load transfer efficiency
increases if the load is positioned at the corner, while it decreases if the load is positioned along
the wheel-path. The limited amount of data available from this study indicates a linear
relationship with a moderate correlation between LTE and joint opening.




0
10
20
30
40
50
60
70
80
90
100
0 5 10 15 20 25 30
Slab Temperature,
L
T
E
,

%
LTE = 0.3T + 83.332
R
2
= 0.3423
LTE = 0.4844T + 72.791
R
2
= 0.3268
Figure 2.7 Load Transfer Efficiency versus Slab Temperature.
Wheel-Path
Edge
LTE = 1.9534T + 85.649
R
2
= 0.4297
LTE = -1.0994T + 83.833
R
2
= 0.0498
0
10
20
30
40
50
60
70
80
90
100
-2 -1 0 1 2 3 4
Temperature Differential, C
L
T
E
,

%
Wheel-Path
Edge
Figure 2.8 Temperature Differential Versus Load Transfer Efficiency.
22
CHAPTER THREE

MEASUREMENT OF DOWEL SHEAR FORCES


3.1 Introduction

Both 38-mm diameter steel dowels as well as Shokbars are fitted with two miniature
strain gages embedded on top and bottom to measure dowel bending as presented in Figure 2.3
(b) and strain rosette on the side to measure the shear force. Attempting to measure the dynamic
shear force in the dowels in Elkins, West Virginia during the FWD tests failed since the signals
were contaminated with a high noise level that prevented extracting any useful information out
of them. The reasons for such a nose can be attributed to the long cable lengths as well as the
aging of strain gages. For this reason, the dynamic tests were conducted on another newly
constructed concrete pavement site in Goshen Road, West Virginia. This Chapter describes this
pavement site as well as the data collected from the dynamic testing of dowel bars and shock
bars.


3.2 Goshen Road Pavement Site

A set of seven test concrete slabs were cast for experimental study in a designated open
area at the parking lot of the WVDOT Maintenance Shop at Goshen road, West Virginia. The
slabs were instrumented with a variety of sensors for long-term monitoring of the slab response
to various loading conditions including seasonal and daily temperature changes. The sensory
system is designed to provide continuous data from key-performance parameters that formulate
the behavior of the slabs such as the distribution of strains along and across the slab centerlines,
joint openings, temperature profiles through the slab thickness, dowel bending moments,
dynamic shear and normal forces, strains at the concrete-dowel interface along with a continuous
record of weather conditions.

In total seven full scale slabs (4.57 m 3.65 m 0.25 m) were constructed. Three of the
seven slabs were not instrumented and act as support or joint slabs. They were laid in September
2002. The remaining four slabs are instrumented and were poured in October 2003. Data
collection began in October 2003 at 12:20 pm when the first of the four instrumented slabs was
poured. The site has a total of seven slabs; five jointed slabs with their joints fitted with either
regular dowels or Shokbars and two free slabs.

Special care was taken while constructing the base layer of the instrumented slabs in
order to provide the required base/slab friction characteristics. The base consisted of a 10 cm
thick layer of crushed stones with an average size of 1.25 cm. Three bleeders about 50 cm in
width were constructed along the 36 meters long base for the purpose of rain water drainage. The
base layer was extended 30 cm from each side of the slab edges to provide uniform distribution
to the slab support. The top surface of the base layer was carefully treated in order to provide
minimum amount of slab/base friction under 6 slabs. A 2.5 cm thick layer of cement paste was
spread on the top surface of the base layer, finished to a smooth surface and set to cure for 15
23
days. This layer was then drilled at staggered spacing of about 30 cm throughout the surface with
openings 1 cm in diameter to allow better water drainage. A thick fabric sheet was then placed
on top of the finished surface and concrete slabs were cast on top of the fabric sheets within
wooden forms. In contrast, one free slab was placed directly on the crushed stones base layer for
the purpose of studying the effect of slab/base friction. The construction of slabs was carried out
through 2 stages. The first stage consisted of placing 3 slabs that served to provide joints to the
instrumented ones. Those 3 slabs were anchored to the ground to simulate the extended
continuity of pavements, i.e. maximum resistant to movement. In each of the fixed slabs, 4 steel
guard rail posts 1.8 meters long were driven along the transverse center line of the slab. The
extended portion of the posts above the finished surface of the base layer measured 25.4 cm
which is exactly the thickness of the slabs. The second stage of construction consisted of placing
the instrumentation system and casting the 4 instrumented slabs in between the wooden forms as
shown in Figure 3.1.


3.3 FWD Testing

The West Virginia Department of Transportation provided a Dynatest Falling Weight
Deflectometer (FWD) to apply dynamic loads at transverse joints as shown in Figure 3.2. The
tests were conducted along three lines: slabs edge, wheel-path, and slab centerline. The
transverse joints in this instrumented pavement section were constructed to be fully opened using
a foam separator. This construction technique ensures that dowel bars or shokbars are the only
means of load transfer device between adjacent slabs. This means that the shear force measured
Figure 3.1 Construction of Instrumented Slabs.
Free slab on
crushed stones
Free slab on
Frictionless base
Doweled slabs on
frictionless base
Weather station
24
in this study represents the highest value of the shear force that can be transferred by the dowels
as the FWD is applied at the transverse joint edge.
The FWD impacts are applied at the locations of the instrumented dowels as the strain
rosette output was recorded. The data acquisition system was programmed to collected data
from all sensors at a rate of 1000 sample per second during the FWD testing. The system was
activated as the FWD was about to drop and stopped after impact. Three FWD impacts were
applied and recorded at each testing location. The deflection-based load transfer efficiencies
were calculated from the surface deflections measured using the FWD device. The tests were
performed in June 2005.


3.4 Dowel Bar Shear

Under the effect of the impact load of the FWD, dowel bars are subjected to bending
moment and shear force. In this analysis, we will focus on the shear force since it represents the
actual force transferred by the dowels to the unloaded slabs. Equation 3.1 was used to calculate
vertical shear forces from the strain measured by the strain gage rosette as the FWD loads were
applied at the joint edge:
( )
2 1
4
3
side side
A G P = (3.1)
Where P = Vertical shear force
G = Shear modulus of the dowel
A = Cross-sectional area of the dowel

side 1
= strain measured in the rosette leg directed 45 upward.

side 2
= strain measured in the rosette leg directed 45 downward.
Figure 3.2 FWD Testing of Transverse Joints.
25

Figure 3.3 illustrates a typical measured time history of the shear forces recorded in the
instrumented dowel bar located at the slab corner due to applying the FWD loading pulse at the
same location. The maximum magnitudes of the shear forces for all tests at different positions
are listed in Table 3.1. It can be noticed that the maximum shear forces are recorded in the
corner dowels when the FWD loading plate is positioned just over these dowels. It can be
noticed that the shear force recorded in the shokbar, located at the corner, is approximately five
times that recorded in the corner dowel. This clearly demonstrates the ability of the shokbar in
transferring shear forces between adjacent slabs.



3.5 Distribution of Shear Forces among Dowel Bars

Figure 3.4 illustrates the distribution of the shear forces among the dowel bars at transverse joint
due to the application of the FWD load at three locations: corner, wheel-path and joint center. It
can be noticed that the magnitude of the dowel shear force decreases linearly as the distance
from the loading position decreases on both sides, which agrees with the assumption of Friberg
(1938) and the finding of Tabatabie and Bernberg (1980). The same observation can also be
made on the shokbars presented in Figure 3.5. However, the magnitudes of the shear force
transmitted through the shokbars are higher than those transmitted through traditional dowels and
the number of the active bars is less. The magnitude of the shear forces induced in the
0 0.5 1.0 1.5 2.0 2.5
-100
-50
0
50
100
150
200
Time (seconds)
D
o
w
e
l

s
h
e
a
r
i
n
g

f
o
r
c
e

(
l
b
s
)
Shear
Force
Figure 3.3 Typical Shear Force History in Dowel Bars.
26
uninstrumented dowels can be calculated through the linear interpolation of the forces shown in
Figures 3.4 and 3.5.

TABLE 3.1 Shear Forces in Instrumented Dowels and Shokbars.

Location
Force of
FWD Drop
(kN)
Shear in
Corner Dowel
(N)
Shear in
wheel-path
Dowel (N)
Shear in
Center Dowel
(N)
70.03 1700 1250 670
70.27 1100 1400 700
Joint 1 Corner
(Regular Dowels)
70.34 860 1350 600
71.73 800 1250 550
71.70 800 1320 550
Joint 1 Wheel-path
(Regular Dowels)
71.82 750 1370 560
69.91 50 850 2200
69.95 50 820 2100
Joint 1 Center
(Regular Dowels)
69.75 50 900 2100
68.00 940 125
67.49 940 120
Joint 2 Corner
(Regular Dowels)
67.37 900 118
70.85 750 2400 350
70.78 700 2300 350
Joint 2 Wheel-path
(Regular Dowels)
70.66 680 2100 340
71.45
71.91
Joint 2 Center
(Regular Dowels)
71.93
69.03 10000 600 0
69.36 9200 750 0
Joint 3 Corner
(Shokbar)
69.10 9300 770 0
69.95 2300 3000 500
70.63 1220 1400 500
Joint 3 Wheel-path
(Shokbar)
70.51 1360 1360 500
68.97 400 800 1500
69.14 350 800 1200
Joint 3 Center
(Shokbar)
68.80 260 750 1500


27
y = 903.49x + 670
R
2
= 0.982
y = -905.6x + 2050
R
2
= 0.9935
0
200
400
600
800
1000
1200
1400
1600
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
y = -672.57x + 1787.5
R
2
= 0.9982
0
500
1000
1500
2000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
y = 1415.9x - 189.47
R
2
= 0.999
0
500
1000
1500
2000
2500
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
a. Corner Load
y = 903.49x + 670
R
2
= 0.982
y = -905.6x + 2050
R
2
= 0.9935
0
200
400
600
800
1000
1200
1400
1600
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
b. Wheel-path Load
c. Joint Center Load
Figure 3.4 Shear Forces Distribution in Dowel Bars.
28
y = -1E-12x + 1360
R
2
= #N/A
y = -984.25x + 2150
R
2
= 1
0
500
1000
1500
2000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
y = -13993x + 11433
R
2
= 1
0
2000
4000
6000
8000
10000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
0
500
1000
1500
2000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Distance along Joint, m
S
h
e
a
r

F
o
r
c
e
,

N
c. Joint Center Load
Figure 3.5 Shear Forces Distribution in Shokbars.
a. Corner Load
b. Wheel-path Load
29

The magnitudes of the shear forces induced in the dowel bars at transverse joints are
listed in Table 3.2. The results in Table 3.2 show that the shear force transmitted through the
shokbars due to corner loading is as twice as that transmitted by regular dowels. This indicates
that the use of shokbars at transverse joints would reduce the corner breaks. The maximum shear
force transmitted through the dowels or shokbars is obtained when the load is applied near the
joint center. The results indicate the superiority of the shokbars in transmitting the corner and
wheel-path loads.

TABLE 3.2 Shear Forces in Dowels and Shokbars (N).

Dowel Bar Location measured from Corner (m)
Load
Position
0.15 0.46 0.76 1.07 1.37 1.68 1.98 2.29 2.59 2.90 3.20 3.51
Sum
(N)
Dowel Bars
Corner 1685 1480 1275 1070 865 660 455 250 45 7785
Wheel-Path 808 1083 1358 1084 808 532 256 5929
Joint Center 26 458 889 1321 1753 2184 2184 1753 1321 889 458 26 13262
Shokbars
Corner 9300 5035 770 15106
Wheel-Path 1400 1380 1360 1100 800 500 200 6740
Joint Center 258 506 754 1002 1250 1498 1498 1250 1002 754 506 258 10536


TABLE 3.3 Comparison of Deflection Load Transfer Efficiency and Actual Load Transfer
Efficiency.

Load Position
Total Force
Transmitted by
Dowels
(KN)
Force of
Drop
(KN)
Load
Transferred
(%)
Deflection-Based
Load Transfer
Efficiency
(%)
Dowel Bars
Corner 7.785 70.21 11.1% 77.03%
Wheel-Path 5.929 71.75 8.3% 89.84%
Joint Center 13.262 69.87 19.0% 78.51%
Shokbars
Corner 15.106 69.17 21.8% 81.16%
Wheel-Path 6.74 70.29 9.6% 87.31%
Joint Center 10.536 68.97 15.3% 77.37%


The percentages of the load transferred through the dowels and/or shokbars are compared with
the deflection load transfer efficiency calculated based from the deflection basins as illustrated in
Figure 3.3. Despite the high values of deflection-based load transfer efficiencies, the percentages
of the actual shear forces transmitted through dowels/Shokbars are very low. This finding agrees
with measurements reported by Sargand (2000), who reported low values for the shear forces
measured in dowel bars irrespective to the high values of deflection-based load transfer
30
efficiency. The results in Table 3.3 indicate that there is no one-to-one relationship between load
transfer efficiency and the deflection-based load transfer efficiency.

The large discrepancies between the deflection-based load transfer efficiencies and load
transfer efficiencies illustrated in Table 3.3 can be explained by the following reasons:
Slab deflections increase in the presence of dowel bar looseness while dowel shear forces
decrease.
The FWD tests were conducted between 10:00 AM and 12:00 PM. During this time, the
slabs are subjected to positive temperature gradient, which makes the joints in contact
with the base layer. This reduces the shear forces transmitted by the dowel bars or the
shokbars (Channakeshava et al., 1993).
31

CHAPTER FOUR

SUMMARY AND CONCLUSIONS


The analyses of the trends reported in this study were intended to calculate deflection-
based load transfer efficiency (LTE) and the percentage of load transferred through the load
transferring devices mounted across the transverse joint. The effects of the seasonal temperature
variations and the position of load application were examined. The following are highlights of
the results obtained from this study:

1. Load transfer efficiency was found to be a complex parameter that depends on many factors
that include load position, testing time, slab temperature, and load transfer device.

2. Testing time and season were found to have a significant effect on the measured load transfer
efficiency. For the same joint, the load transfer efficiency measured in winter was found to
be less than that measured in summer.

3. The slabs fitted with 32 mm (1.25 in) diameter dowels displayed higher variability of the
measured load transfer efficiency than those fitted with 38 mm (1.5 in) diameter dowels.

4. Joint opening changes daily and seasonally as the ambient temperature changes. As the
amount of joint opening increases due to slab contraction during winter, the measured load
transfer efficiency generally decreases.

5. As the slab temperature increases, the load transfer efficiency increases.

6. Poor correlation was found between the deflection-based load transfer efficiency and the
percentage of the load transferred through the load transferring devices mounted across the
transverse joint.

32

REFERENCES


1. American Association of State Highway and Transportation Officials (1993).
AASHTO Guide for Design of Pavement Structures.

2. American Association of State Highway and Transportation Officials (1998).
Supplement: AASHTO Guide for Design of Pavement Structures-Part II Rigid
Pavement design and Rigid Pavement Joint Design.

3. American Concrete Pavement Association (1991). Design and Construction of Joints
for Concrete Highways. Technical Bulletin-010.0, American Concrete Pavement
Association, Portland Cement Association.

4. Beckemeyer, C.A., L. Khazanovitch, and H.T. Yu (2002). Determining the Amount of
Built-in Curling in JPCP. Presented at the 81
st
Annual Meeting of Transportation
Research Board, National Research Council, Washington, D.C.

5. Channakeshava, C., and F. Barzegar, and G.Z. Voyiajis (1993). Nonlinear FE Analysis
of Plain Concrete Pavements with Doweled Joints. Journal of Transportation
Engineering, ASCE, Vol. 119, No. 5, pp.763-781.

6. Chou, Y.T. (1981). Structural Analysis Computer Programs for Rigid Multicomponent
Pavement Structures with Discontinuities: WESLIQUID and WESLAYER. Technical
Report GL-81-6. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station.

7. Colley, B.E., and H.A. Humphrey. Snyder (1967). Aggregate Interlock at Joints in
Concrete Pavements. Highway Research Record No. 189, Transportation Research
Board, National Research Council, Washington, D.C., pp. 1-18.

8. Davids, W.G., and G.M. Turkiyyah (1997). Development of Embedded Bending
Member to Model Dowel Action. Journal of Structural Engineering, ASCE, Vol.
123, No. 10, pp.1312-1320.

9. Davids, W., G.M. Turkiyyah, and J. Mahoney (1998). EverFE-Rigid Pavement 3D
Finite Element Analysis Tool. Transportation Research Record No. 1629,
Transportation Research Board, National Research Council, Washington, D.C., pp. 41-
49.

10. Davids, W.G. (2000). Effect of Dowel Looseness on Response of Jointed Concrete
Pavements. Journal of Transportationl Engineering, ASCE, Vol. 126, No. 1, pp. 50-
57.

33
11. Davids, W.G. (2001). 3D Finite Element Study on Load Transfer at Doweled Joints in
Flat and Curled Rigid Pavements. The International Journal of Geomechanics, Vol. 1,
No. 2, pp. 309-323.

12. Friberg, B. F. (1938). Load and Deflection Characteristics of Dowels in Transverse
Joints of Concrete Pavements. Proceedings of Highways Research Board No. 18,
National Research Council, Washington, D.C., pp. 140-154.

13. Guo, H., J. Sherwood, and M. Snyder (1995). Component Dowel Bar Model for Load
Transfer System. Journal of Transportation Engineering, ASCE, Vol. 121, No. 3,
pp.289-298.

14. Heinriches, K.W., Liu, M.J., Darter, M.I., Carpenter, S.H., and Ioannides, A.M. (1989).
Rigid Pavement Analysis and Design. Federal Highway Administration Report No.
FHWA-RD-068.

15. Huang, Y.H. and S.T. Wang (1974). Finite Element Analysis of Rigid Pavements with
Partial Subgrade Contact. Transportation Research Record No. 485, Transportation
Research Board, National Research Council, Washington, D.C., pp. 39-54.

16. Huang, Y.H. (1993). Pavement Analysis and design, Prentice Hall, Englewood, N.J.

17. Ioannides, A.M., Y.H. Lee and L. M.I. Darter (1990). Control of Faulting Through
Shear Load Transfer Mechanism. Transportation Research Record No. 1286,
Transportation Research Board, National Research Council, Washington, D.C., pp. 49-
56.

18. Ioannides, A.M. and G.T. Krovesis (1992). Analysis and Design of Doweled Slab-on-
Grade Pavement System. Journal of Transportation Engineering, ASCE, Vol. 118,
No. 6, pp.745-768.

19. Kennedy, J.C., and R.D. Everhart (1996). Comparison of Predicted Pavement
Structural Response with Field Measurement Data. Final Report to Turner Fairbanks
Highway Research Center, Contract No. DIFH 61-93-C-00055, Technical Work
Request No. 3.

20. Kennedy, J.C., and R.D. Everhart (1997). Thermal Effects on Pavement Response.
Final Report to Turner Fairbanks Highway Research Center, Contract No. DIFH 61-
93-C-00055, Technical Work Request No. 5.

21. Kennedy, J.C., and R.D. Everhart (1998). Warping and Curling of Rigid Pavements, A
Mechanistic Approach Implemented in UMPAD. Final Report to Turner Fairbanks
Highway Research Center, Contract No. DIFH 61-93-C-00055, Technical Work
Request No. 6.

34
22. Khazanovitch, L., H.T. Yu, and C. Beckemeyer (2000). Application of ISLAB2000
for Forensic Studies. Proceeding of the 2
nd
International Symposium of 3D Finite
Element for Pavement Analysis, Design, and Research, Charleston, West Virginia, pp.
433-450.

23. Khazanovich, L. and A. Gotlif (2003). Evaluation of Joint and Load Transfer. Final
Report No. FHWA-RD-02-088, Federal Highway Administration, McLean, Virginia.

24. Korbus, L. and E.J. Barenberg (1979). Longitudinal Joint Systems in Slip-Formed
Rigid Pavements-Volume IV. Final Report No. DOT/FAA/RD-79/4, Federal Aviation
Administration, U.S. Department of Transportation.

25. Korovesis, G.T. (1990). Analysis of Slab-On-Grade Pavement Systems Subjected to
Wheel and Temperature Loadings. Ph.D. Dissertation, University of Illinois at
Urbana, Illinois.

26. Kuo, C.M., K.T. Hall, and M.I. Darter (1996). Three Dimensional Finite Element
Model for Analysis of Concrete Pavement Support. Transportation Research Record
No. 1505, Transportation Research Board, National Research Council, Washington,
D.C., pp. 119-127.

27. Kushing, J.W., and W.O. Fremont (1940). Design of Load Transfer Joints in Concrete
Pavements. Proceedings of Highways Research Board No. 20, National Research
Council, Washington, D.C., pp. 481-493.

28. McGhee, K.H. (1995). Design, Construction, and Maintenance of PCC Pavement
Joints, NCHRP Synthesis No 211, National Cooperative Highway Research Program,
Transportation Research Board, National Research Council, Washington, D.C.

29. Nishizawa, T., T. Fukuda, and S. Matsuno (1989). A Refined Model of Doweled
Joints for Concrete Pavement Using FEM Analysis. Proceedings of the 4
th

International Conference on Concrete Pavement Design and Rehabilitation, Perdue
University, West Lafayette, Indiana.

30. Raja, Z.I. and M.B. Snyder (1991). Factors Affecting Deterioration of Transverse
Cracks in Jointed Reinforced Concrete Pavements. Transportation Research Record
No. 1307, Transportation Research Board, National Research Council, Washington,
D.C., pp. 162-168.

31. Sargand, S.M., and D.J. Breegle (1998). Three-Dimensional Finite Element Software
Development and Verification Case Study. Paper Presented at the First National
Symposium of 3D Finite Element for Pavement Analysis and Design, Charleston, West
Virginia.

32. Sargand, S.M. (2000). Performance of Dowel Bars and Rigid Pavement. Final Report
for Federal Highway Administration, Ohio University, Athens. Ohio.
35

33. Shoukry, S.N., and G.W. William (1998). 3D FEM Analysis of Load Transfer
Efficiency. Proceedings of the First National Symposium of 3D Finite Element for
Pavement Analysis and Design, Charleston, West Virginia, pp. 40-46.

34. Shoukry, S.N., M.Y. Riad, G.W. William (2004-a). Field Performance of Shok Bar.
Draft Final Report No. RP # 170 submitted to West Virginia Division of Highways.

35. Shoukry, S.N., G.W. William, M. Riad (2004-b). Validation of 3DFE Model of
Jointed Concrete Pavement Response to Temperature Variations. The International
Journal of Pavement Engineering, Vol. 5, No. 3, pp. 123 - 136.

36. Sutherland, E.C. (1940). Discussion of "Design of Load Transfer Joints in Concrete
Pavements, by J.W. Kushing and W.O. Fremont. Proceedings of Highways Research
Board No. 20, National Research Council, Washington, D.C., pp. 494-497.

37. Tabatbaie, A.M., and E.J. Barenberg (1978). Finite-Element Analysis of Jointed or
Cracked Concrete Pavements. Transportation Research Record No. 671,
Transportation Research Board, National Research Council, Washington, D.C., pp. 11-
19.

38. Tabatbaie, A.M. and E.J. Barenberg (1980). Structural Analysis of Concrete Pavement
Systems. Journal of Transportation Engineering, ASCE, Vol. 106, No. 5, pp.493-506.

39. Tayabji, S.D. and Colley, B.T. (1986). Analysis of Jointed Concrete Pavements.
Technical Report FHWA-RD-86-041, Federal Highway Administration, McLean,
Virginia.

40. Teller, L.W. and H.D. Cashell (1959). Performance of Dowelled Joints Under
Repetitive Loading. Highway Research Bulletin No. 217, Highway Research Board,
National Research Council, Washington, D.C.

41. Teller, L.W. and E.J. Sutherland (1936). A Study of Structural Action of Several
Types of Transverse and Longitudinal Joint Design. Public Roads, Vol. 17, No. 7.

42. Tia, M., J.M. Armaghani, C. Wu, S. Lei, and K.L. Toye (1987). FEACONS III
Computer Program for Analysis of Jointed Concrete Pavements. Transportation
Research Record No. 1136, Transportation Research Board, Washington, D.C.

43. Westergaard, H.M. (1926). Stresses in Concrete Pavements Computed by Theoretical
Analysis. Public Roads, Vol. 7, No. 2, pp. 25-35.

44. Westergaard (1929). Spacing of Dowels. Proceedings of Highway Research Board
No. 8, National Research Council, Washington, D.C., pp. 154-158.

36
45. Wu, C., T.J. Larsen (1993). Analysis of Structural Response of Concrete Pavements
under Critical Thermal-Loading Conditions. Proceedings of the Fifth International
Conference on Concrete Pavement Design and Rehabilitation, Purdue University,
West Lafayette, Indiana.

46. Yoder, E.J. and M.W. Witczak (1975). Principles of Pavement Design, Second Edition.
John Wiley & Sons, Inc., New York, N.Y.