Available online at www.sciencedirect.

com

ScienceDirect
International Journal of Pavement Research and Technology 9 (2016) 414–423
www.elsevier.com/locate/IJPRT

Optimal design on dowel length for cement concrete pavement

Chichun Hu, Jiexian Ma, Yuan Yu ⇑, Yi Luo
College of Civil and Transportation Engineering, South China University of Technology, Wushan Road, Guangzhou 510641, China

Received 24 March 2016; received in revised form 18 October 2016; accepted 23 October 2016
Available online 4 November 2016

Abstract

In order to compute the optimal dowel length in cement concrete pavement, semi-infinite beam on elastic foundation was deduced and
modified for the analysis of dowel bars. The dowel deflection, bending moment and shearing were analyzed for dowel bars under the
traffic loading, dowel length based on the second inflexion distance was computed and a relationship between dowel length and dowel
diameter was established. The theoretical analysis found that the dowel length in Chinese specification is conservative. A finite element
model was also established to simulate the dowel load system. Based on the result of maximum value and variation tendency of mises
stress for different dowel diameters and dowel lengths, it is feasible to shorten dowel length specified in JTG D40-2002 by 50%. However,
considering the construction tolerances in the making and sawing of joints in new pavement construction, which might add 50–150 mm
to the required overall dowel length, it’s more appropriate to reduce the dowel length by 20% in practice.
Ó 2016 Chinese Society of Pavement Engineering. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Dowel bar; Mechanical analysis; Length design; Cement concrete pavement; Finite element model

1. Introduction Some studies have been conducted on dowel bar design.

The performance of cement concrete pavement depends
largely upon how well the joint was constructed. The joint
function is greatly related to the dowel bars [1]. Three main
damage types of cement concrete pavement (slabcracking,
faulting and joint deficiencies) are all relevant with dowel
bars. It is significant to conduct the dowel length design
for pavement joint. Too short a dowel bar will restrict
the loading transfer capability across the joint, which will
potentially cause the uneven settlement between slabs and
even rehabilitation, while too long a dowel bar will be a
waste of resource. Since early 2000, the price of steel dowel
bar is always going higher and higher [2].
⇑ Corresponding author.
E-mail addresses: cthu@scut.edu.cn (C. Hu), ma.jiexian@mail.scut.
edu.cn (J. Ma), 546433568@qq.com (Y. Yu), debulo@scut.edu.cn
(Y. Luo).
Peer review under responsibility of Chinese Society of Pavement
Engineering.
Jiang [3] analyzed the strain distribution along the dowel
length based on finite element method. Zhang [4] calculated
the deflection of dowel bar based on energy variation prin-
ciple and pointed out both bending moment and shearing
force should be taken into account for the dowel analysis.
Bradury [5] assumed the stress along the dowel bar bean
was linear while the assumption was not demonstrated.
Friberg [6] established a model for dowel bar analysis
including five parameters: loading, dowel bar diameter,
elastic modulus of dowel bar, bending rigidity of dowel
bar and joint width. The model proposed by Friberg was
intended for expansion joints initially, while the contrac-
tion joint is much shorter than the expansion joint. There-
fore, the Friberg model is not suitable for the pavement
contraction joint analysis [7–9].
In this paper, the Timoshenko model of infinite beam on
elastic foundation was modified to analyze dowel bar.
Based on the theoretical solution for deformation, bending
moment and shearing force of dowel bar, the second

http://dx.doi.org/10.1016/j.ijprt.2016.10.003
1996-6814/Ó 2016 Chinese Society of Pavement Engineering. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423 415

inflexion distance was proposed to be used as dowel
embedment length. Moreover, in order to verify the theo-
retical result, a finite element model was established to ana-
lyze the effect of dowel length on dowel-concrete bearing
stress. Optimal dowel length was obtained based on the
above calculation.
There is no significant difference for spacing require-
ments by foreign and domestic specifications, while dowel
length requirement varies among different specifications.
Therefore, it is necessary to conduct the analysis on dowel
bars to get an optimized dowel length, which will guide the
joint and dowel design within cement concrete pavement.

2. Domestic and foreign dowel bar sizes 3. Theoretical model

Before 2003, there was no stringent requirement for The deformation of beam (Fig. 1) due to bending
dowel bar within the contraction joint, which led to many moment can be expressed as:
serious problems in China. Therefore, the pavement design
specification (JTG D40-2002) required all contraction joint 1=q ¼ M=EI ð1Þ
to be installed with dowel bars for heavy volume highways
where q is curvature radius of neutral surface, M is bending
[10] and the dowel diameter was increased. The spacing and
moment, E is elastic modulus of beam, I is second axial
size of dowel bar specified in the new and old specification
moment.
[10–11] are listed in Table 1.
Considering the beam deformation is very small com-
The dowel bars size suggested by Portland Cement
pared to the beam height, the curvature can be calculated
Concrete Pavement Association [12] is listed in Table 2.
with Eq. (2)
The dowel bar size required by several states [12] is listed
in Table 3. 1 d2 y
¼ 2 ð2Þ
q dx
Table 1
Dowel size and spacing. where y is beam deflection, x is distance from joint face.
Slab Dowel Minimum Maximum Generally, the upper concave curvature was negative
thickness diameter length spacing
(cm) (cm) (cm) (cm) d2 y
EI ¼ M ð3Þ
JTG D40–2002 22 2.8 40 30 dx2
24 3 40 30
26 3.2 45 30 In the case of Fig. 2
28 3.5 45 30 dM
30 3.8 50 30 ¼Q ð4Þ
dx
JTJ 012–94 <20 2 40 30
21–25 2.5 45 30 where Q is shear force.
26–30 3 50 30 Assume that a beam AB, loaded by a distributed load q,
is supported along its entire length by a continuous elastic
foundation (See Fig. 3). When the beam is deflected, the
Table 2
Dowel size recommended by Portland Cement Association. reaction of elastic foundation at every section is propor-
tional to the deflection at that section. Under such condi-
Slab thickness Dowel Dowel embedment Dowel length
(cm) diameter (cm) length (cm) (cm) tion, the reaction per unit length of the beam can be
represented by the expression ky in which k, called the
12.5 1.6 12.5 30
15 1.9 15 35 modulus of the foundation denoting the reaction per unit
20 2.2 15 35 length.
20 2.5 15 35 In order to study the beam deformation, second
22.5 2.8 17.5 40 derivation of Eq. (3) was taken:
25 3.1 18.8 45
27.5 3.5 20 45 d4 y d2 M dQ
30 3.8 22.5 50 EI 4
¼  2
¼ ð5Þ
dx dx dx

Table 3
Dowel size in different states in the U.S. ρ

State Dowel Dowel Dowel

diameter (cm) length (cm) spacing (cm)
Alabama 3.8 45.7 30.5
Florida 1.9–3.2 38.1 30.5
Minnesota 3.2–3.8 38.1 30.5
Indiana 3.2–4.4 30.5 30.5
New Jersey 3.2 43.2 20.3–30.5
Fig. 1. Pure bending of the beam.
416 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423

where b is relative stiffness of the dowel bar encased in con-

crete, P is shear force carried by the critical dowel, M0 is
bending moment carried by the critical dowel at joint face
which is equal to the product of P and half joint width.
The slope angle of dowel bar deformation can be calcu-
lated based on the first derivation of y:
dy ebx
¼ 2 ½ðP  M 0 bÞðcosbx  sinbxÞ þ ðM 0 bÞðsinbx þ cosbxÞ
dx 2b EI
ð9Þ
The moment at any section can be calculated as follows:
Fig. 2. Bending and shearing of small specimen inside the beam.
d2 y ebx
M ¼ EI ¼ ½P sinbx  bM 0 ðcosbx þ sinbxÞ ð10Þ
dx 2 b
The shearing force can be calculated as follows:
dM
Q¼ ¼ ebx ½ð2bM 0  P Þsinbx þ P cosbx ð11Þ
dx

4. Dowel length design

(a) 4.1. Modification of the semi-infinite beam on elastic

foundation

The forementioned analysis was toward the deformation

of semi-infinie beam. Although dowels are of measurable
finite length, Albertson and others [14,15] have shown that
the theory can be applied to dowel bars given that Leb is
greater than or equal to 2 (where Le is the embedment
length of the dowel within the slab) [16]. Therefore, dowel
bar in concrete pavement could be regarded as the modifi-
cation of semi-infinite beam.
To show variation along dowel bar, deflection, shearing
(b) force and bending moment for critical dowel at joint edge
were computed using the above analytic solutions. 3.2 cm
Fig. 3. Beam on elastic foundation (a) beam, (b) an infinitely small unit.
dowel diameter, 0.5 cm joint width and 30 cm dowel spac-
ing were taken. The traffic loading and the slab thickness
were 100 KN and 26 cm, respectively.
As for an infinitely small unit, dQ/dx can be obtained In this paper, inflexion point is defined as the zero point
through derivation of shearing force on x. The equilibrium of dowel deformation. The second inflexion distance refers
equation of the unit due to shearing force can be written as to the distance between joint face and the second inflexion
dQ point. As shown in Figs. 4–6, the shearing force and
¼ ky  q ð6Þ
dx
Substitute Eq. (6) into Eq. (5)
d4 y
EI ¼ q  ky ð7Þ
dx4
Eq. (7) is the general equation of beam deformation on
elastic foundation [13].
Considering a semi-infinite dowel bar encased in con-
crete is equivalent to a semi-infinite beam supported by
elastic foundation, Eq. (7) can be applied to solve dowel
deformation. The solution for semi-infinite dowel bar
encased in concrete pavement can be obtained:
ebx
y¼ ½ðP  M 0 bÞcosðbxÞ þ ðM 0 bÞsinðbxÞ ð8Þ
2b3 EI Fig. 4. Dowel deflection.
 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423 417

Table 4
Influence of dowel diameter on the second deflection inflexion.
Dowel The second inflexion The second inflexion
diameter (cm) distance (cm) distance/dowel diameter
0.5 4.8 9.6
1 8.2 8.2
2 13.92 7.0
2.8 17.98 6.4
3.2 19.89 6.2
3.6 21.72 6.0
5 27.89 5.6

Fig. 5. Shearing force distribution of dowel.

Fig. 7. Relationship between the second deflection inflexion and dowel

diameter.

where Lemb is dowel length embedded in the concrete slab, z

Fig. 6. Bending distribution of dowel. is joint width.
The joint width is too small compared with the second
bending moment at the second inflexion point were 142.3 N
inflexion distance, thus it can be ignored. Dowel length
and 6.1 N m, around 1.10% and 2.84% of the maximum
based on the second inflexion distance can be obtained:
value along dowel bar, respectively.
Concerning that dowel bar is mainly used to transfer Ldowel ¼ 2Lemb þ z  2L2 ¼ 10:1568D þ 6:6732 ð13Þ
shear load, the contribution of dowel embendment length
where Ldowel is dowel bar length, D is dowel diameter.
larger than second inflexion distance to load transfer can
be neglected. Therefore, it is feasible to cut the semi-
infinite beam at the second inflexion and take the second 4.3. Analysis of the dowel length in the current Chinese
inflexion distance as dowel embendment length. specification

Based on the modified dowel bar model, the calculated

4.2. Influence of dowel diameter on the second inflexion
dowel length is listed in Table 5 together with the required
distance
length in Chinese specification. As can be seen in Table 5,
the dowel length based on the second inflexion distance was
In order to study the influence of the dowel diameter on
88%-95% of the minimum dowel length specified in JTG
the second inflexion distance, the second inflexion distances
D40-2002, indicating that dowel length in the specification
corresponding with various dowel diameters were calcu-
is consevative. It is recommended that dowel length can be
lated and the result is listed in Table 4.
reduced by 10%.
As can be seen in Table 4, the second inflexion distance
increased with the dowel diameter, while the ratio of the
second inflexion distance over the dowel diameter 5. Finite element method
decreased. Fitting between the second inflexion distance
and the dowel diameter was conducted (see Fig. 7). 5.1. Modelling
Meanwhile,
In order to verify the theoretical results, a finite element
L2 ¼ Lemb þ z=2 ð12Þ model based on AbaqusÒ was established for stress
418 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423

Table 5
Comparison of computed dowel length and the length in Chinese specification.
Slab thickness Dowel diameter The second inflexion distance Dowel length (2L2) Minimum dowel length Lmin in JTG 2L2/Lmin
(cm) (cm) L2 (cm) (cm) D40-2002 (cm) (%)
22 2.8 17.98 35.96 40 89.9
24 3 18.95 37.9 40 94.8
26 3.2 19.89 39.78 45 88.4
28 3.5 21.29 42.58 45 94.6
30 3.8 22.66 45.32 50 90.6

analysis of dowel bar. The model was meshed by structured
technique, and C3D8R unit (eight-node brick element with
reduced integration) was used. Moreover, to save the com-
putational time and ensure the convergence of multiple
step method for solving plastic damage model of concrete,
the final number of mesh was controlled within 30,000 by
establishing symmetric 1/2 model and setting symmetry
constraints in the symmetry plane. Simulations on more
successively finer grids were also performed to examine
the mesh convergence of the model.
For the purpose of obtaining a smooth stress curve
around dowel-concrete contact surface, an integrated
model was established first, and then the dowel-concrete
interface was cut out. Therefore, uniform mesh could be
applied to achieve meshing alignment, effectively avoiding
local separation of contact surface and stress singularity.
The numerical model for the parametric study was sub-
jected to static load only. The load and the joint width were
taken as 12 kN and 5 mm, respectively.
5.2. Design of model control group

5.2.1. Geometric parameters of the model

The overall model is shown in Fig. 8. According to the
Chinese cement concrete pavement design standard, the
pavement thickness corresponding to dowel diameters of
28 cm, 30 cm, 32 cm, 35 cm, 38 cm for different traffic levels
were selected, as shown in Table 6.
In order to evaluate the influence of dowel length on
stress distribution along the dowel, the models for different
dowel lengths were also established (see Table 7).

5.2.2. Mesh convergence analysis

In the region where stresses changed sharply, the accu-
racy of stresses was often affected by the size of the mesh.
Therefore, the stresses for different mesh densities were cal-
culated to indicate whether convergence was achieved or
not. For the purpose of the comparison, 30 mm dowel
diameter and 60% dowel length were taken. As shown in
Fig. 9, a mesh refinement approach was conducted, in
which the number of grids on every 90 degrees circumfer-
Fig. 8. The overall model. ence of the dowel was increased from 10 to 15.

Table 6
Geometric parameters and material properties of the model.
Component Geometric parameters for different traffic level (mm) Material properties
Level 1 Level 2 Level 3 Level 4 Level 5 Elasticity Poisson’s
modulus (MPa) ratio
Dowel bar Diameter 28 30 32 35 38 210,000 0.3
Length 400 400 450 450 500
Cement concrete Width 300 300 300 300 300 28,000 0.2
Thickness 220 240 260 280 300
Length 197.5 197.5 222.5 222.5 247.5
Cement treated base Width 300 300 300 300 300 14,000 0.3
Thickness 200 200 200 200 200
Length 197.5 197.5 222.5 222.5 247.5
 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423 419

Table 7
Dowel length for different diameter.
Diameter (mm) Dowel length (mm)
100% (Minimum dowel length 60% 50% 40% 30%
in JTG D40-2002)
28 400 240 200 160 120
30 400 240 200 160 120
32 450 270 225 180 135
35 450 270 225 180 135
38 500 300 250 200 150

Fig. 9. Mesh convergence analysis (a) before mesh refinement, (b) after mesh refinement.

5.2.3. Model constraint condition analysis 5.3. Results and discussion

To establish control group of the constraint condition,
the displacement normal to the back of each model (see
Fig. 10) was constrained and unconstrained, respectively.
In reality, the constraint state of the backside is between
these two situations, so it can be inferred that the actual
stress distribution of the dowel is also between these two
situations. The diameter of the dowel was taken as 28 mm.
In addition to dowel corrosion, the main reason for the
failure of dowel load transfer system is the crush of the sur-
rounding concrete caused by excessive dowel-concrete
bearing stress. Therefore, curves of mises stress of concrete
below and on top of dowel bar were plotted against the dis-
tance from joint to evaluate bearing stress.

5.3.1. Comparison of stress before and after mesh refinement

After mesh refinement, as can be seen in Fig. 11, the
maximum stress below the dowel was increased from

Fig. 10. Backside of the model. Fig. 11. Comparison of mises stress before and after mesh refinement.
420 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423
Fig. 12. Comparison of mises stress below the dowel with and without
constraint
Fig. 13. Comparison of mises stress on top of the dowel with and without
constraint
28.5 MPa to 29.5 MPa, and then two curves rapidly over-
lapped. On top of the bar, the peak of the stress curve
shifted slightly to the right. While the mesh number at
Fig. 14. Mises stress nephogram (a) dowel bar, (b) concrete slab.
the interface between concrete and dowel was increased
by 50%, the stress was increased by 3.6%, which indicated
that the results did not depend on the mesh size and
rougher mesh was fine enough.
5.3.2. Comparison of stress with and without constraint
For 28-mm diameter dowel bar, the result of mises stress
of concrete below the dowel when the backside of the
model being constrained and unconstrained is shown in
Fig. 12. In the case of unconstraint, as the dowel length
increased from 30% to 60%, the maximum stress under-
neath the dowel bar kept decreasing (35.0 MPa,
34.4 MPa, 32.6 MPa and 32.5 MPa), almost unchanged
after more than 50%.
The result presented in Fig. 13 clearly showed that the
dowel length has distinct effect on the stress distribution
on top of dowel bar. When dowel length was 30%, the
stress along the dowel changed sharply, and the maximum
mises stress without constraint increased significantly com-
pared to the stress with constraint. As the dowel length
reached 40%, the impact of constraint condition was
reflected in the change of stress distribution at the end of
the dowel bar and the slightly increased maximum stress.
It can also be observed that when the length is greater than
or equal to 50%, the variation tendency of bearing stress
along dowel bar and the maximum stress under two differ-
ent boundary conditions were pretty close. With the
increase of dowel length, the differential stress between
two boundary conditions decreased.
The above numerical results of stress with and without
constraint indicate that using dowel length of greater than
50% of the minimum dowel length in JTG D40-2002 for
28-mm diameter dowel will not affect the zone of maximum
bearing stress.
5.3.3. Stress distribution analysis along the dowel for
different diameters
The mises stress nephogram of the dowel load transfer
system is shown in Fig. 14.
 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423 421

The curves of mises stress below and on top of dowel for
different diameters, when the displacement normal to the
backside of the model being constrained, are presented in
Figs. 15 and 16.
As shown in Fig. 15, for the same diameter, when dowel
length less than 50%, the maximum mises stress of concrete
below dowel bar increased slightly as dowel length increas-
ing. When dowel length is greater than 50%, the mises
stress curves almost overlaped. Comparing maximum
stress for different dowel diameters, while the diameters
were 28 mm, 30 mm, 32 mm, 35 mm, 38 mm, the maximum
mises stress of concrete below dowel bar was 32.2 MPa,
28.5 MPa, 25.4 MPa, 21.2 MPa, 17.9 MPa, respectively. It
is suggested that dowel bar with large diameter can effec-
tively reduce the dowel-concrete bearing stress.
The result of mises stress for concrete on top of dowel
bar is shown in Fig. 16. What deserves special mention is
that a distinct difference exists between the variation
tendency of mises stress versus distance from joint face
for 28-mm, 30-mm diameter dowel and that for 32-mm,
35-mm, 38-mm diameter dowel, especially when dowel
length smaller than 50%. As dowel length is greater than
50%, with the increase of diameter, the maximum stress
value at the bottom were 3.72 MPa, 3.66 MPa, 3.55 MPa,
3.50 MPa, 3.44 MPa, respectively.
On the whole, it is acceptable to take dowel length
greater than or equal to 50% of the minimum dowel length
in JTG D40-2002 as limit value, based on the result of finite
element analysis on maximum value and variation ten-
dency of mises stress.

Fig. 15. Mises stress below the dowel for different diameters (a) 28 mm, (b) 30 mm, (c) 32 mm, (d) 35 mm, (e) 38 mm.
422 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423

Fig. 16. Mises stress on top of the dowel for different diameters (a) 28 mm, (b) 30 mm, (c) 32 mm, (d) 35 mm, (e) 38 mm.

6. Conclusions dowel diameter was established. Calculation result

(1) The general equation of beam deformation on elastic
foundation was deduced in this paper. Considering
dowel bar encased in concrete pavement could be
regarded as the modification of semi-infinite beam,
the solution for deformation, bending moment and
shearing force of dowel bar was obtained.
(2) As dowel bar is mainly used to transfer shear load,
the contribution of dowel embendment length lar-
ger than second inflexion distance to load transfer
can be neglected. Therefore, it is feasible to cut
the semi-infinite beam at the second inflexion and
take the second inflexion distance as dowel
embendment length. A relationship between dowel
length based on the second inflexion distance and
indicated that dowel length in the specification is
conservative and dowel length is suggested to be
10% shorter.
(3) In order to verify the theoretical results, a finite ele-
ment model was established for dowel-concrete bear-
ing stress analysis. Based on the result of maximum
value and variation tendency of mises stress for differ-
ent dowel diameters and dowel length, it is feasible to
shorten dowel length specified in JTG D40-2002 by
50%. However, considering the construction
tolerances in the making and sawing of joints in
new pavement construction, which might add
50–150 mm to the required overall dowel length.,
it’s more appropriate to reduce the dowel length
specified in JTG D40-2002 by 20% in practice.
 C. Hu et al. / International Journal of Pavement Research and Technology 9 (2016) 414–423
(4) Future experimental research is needed to be carried
out to verify the theoretical and numerical results
presented in the paper.
Acknowledgements
Supports provided by National Natural Science Foun-
dation (51308228, 51578248), Chongqing Jiaotong Univer-
sity (LHSYS-2013-002), Guangdong DOT (2004-02-005),
and Fundamental Research Funds for Central Universities
(2015ZZ006) are greatly appreciated.
References
[1] Federal Highway Administration (FHWA), Concrete pavement
joints, Technical Advisory T 5040.30. http://www.fhwa.dot.gov/
legsregs/directives/techadvs/t504030.htm, 1990 (accessed 16.08.01).
[2] D. Wang, C. Hu, Y. Wang, B. Tan, Try of fiber reinforced polymer as
dowel bar alternatives, Highway 01 (2007) 32–36.
[3] Y. Jiang, Study of Concrete Pavement Materials and Structure under
Heavy-Load Transportation (Doctoral dissertation), Chang’an
University, Xi’an, China, 2005.
[4] W. Zhang, W. Yi, Analytic solution of the deflection of dowel steel in
rigid pavement joint, J. Hunan Univ. (Natural Sci.) 11 (2007) 25–28.
[5] R.D. Bradbury, R.B. Gage, Design of joints in concrete pavements,
in: Proceedings of the 12th Annual Meeting of the Highway Research
Board, Washington D.C., 1932.
423
[6] B.F. Friberg, Design of dowels in transverse joints of concrete
pavements, Trans. Am. Soc. Civil Eng. 195 (1940) 1076–1095.
[7] D. Wang, C. Hu, R. Robert, Assessment of grouted glass
fiber polymer (GFRP) tubes as dowel bar alternatives, in: The
25th Annual Southern African Transport Conference, South Africa,
2006.
[8] O. Selezneva, J. Jiang, S.D. Tayabji, Preliminary evaluation and
analysis of LTPP faulting data, Report No. FHWA-RD-00-076,
ERES Consultants Inc., Columbia, MD, 2000.
[9] American Concrete Pavement Association (ACPA), Joints and
jointing practices, Technical Reports, 2005.
[10] D.L. Bischoff, Random skewed joints with and without dowels,
Report No. WI/FEP-07-96, Wisconsin Department of Transporta-
tion, Madison, WI, 1996.
[11] S.S. Mannava, Behavior of Dowel Bars Embedded in Concrete
Pavements (Master thesis), University of Oklahoma, Norman,
Oklahoma, 1997.
[12] JTG D40–2002, Design specifications of cement concrete pavement,
Ministry of Transport of the PR China, Beijing, 2011.
[13] D. Wang, C. Hu, Dowel bar alternatives in rigid pavements, in: The
International Symposium on Innovation & Sustainability of Struc-
tures in Civil Engineering, Shanghai, China, 2007.
[14] M.D. Albertson, Fiber composite and steel pavement dowels (Master
thesis), Iowa State University, Ames, Iowa, 1992.
[15] M.L. Porter, B. Barnes, B. Hughes, K. Viswanath, Non-corrosive tie
reinforcing and dowel bars for highway pavement slabs, Report No.
HR-343, Iowa State University, Ames, IA, 1993.
[16] M.L. Porter, J.K. Cable, F.S. Fanous, J.F. Harrington, N.J. Pierson,
Laboratory study of structural behavior of alternative dowel bars,
Report No. DTFH61-01-X00042, Center for Transportation
Research and Education, Iowa State University, Ames, IA, 2006.