Construction and Building Materials 55 (2014) 212–219

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Premature distresses at transverse construction joints (TCJs) in

continuously reinforced concrete pavements
Wujun Zhou a, Pangil Choi a,⇑, Sureel Saraf a, Sung Woo Ryu b, Moon C. Won a
a
Department of Civil and Environmental Engineering, Texas Tech University, Lubbock, TX 79409, United States
b
Korea Expressway & Transpiration Research Institute, Hwa-Sung, Republic of Korea

h i g h l i g h t s

 Poor construction practices and the use of additional tie bars at TCJs are the major causes of premature distress.
 Behavior of longitudinal steel and additional tie bars at TCJs is different.
 Premature distress at TCJs can decrease as use of relief transverse saw cut joints near TCJs.

a r t i c l e i n f o a b s t r a c t

Article history: The long-term performance of continuously reinforced concrete pavement (CRCP) has been quite satis-
Received 24 September 2013 factory in Texas, providing an important network of highways with heavy truck traffic and minimal main-
Received in revised form 16 January 2014 tenance. However, certain forms of distresses were observed within a few years after construction,
Accepted 16 January 2014
requiring repair of the distresses and causing traffic delays. Distresses at transverse construction joints
Available online 8 February 2014
(TCJs) were one of the most frequently observed distresses, and efforts were made to identify the mech-
anisms of those distresses and develop or improve design standards or specifications. Field evaluations of
Keywords:
concrete and steel behavior near TCJs were made using various gages. The analysis of the data indicates
Continuously reinforced concrete pavement
CRCP
that the current practice of placing additional tie bars at TCJs, along with poor construction practices, are
Punchout the major causes of the distress. There is a large difference in the behavior of reinforcement between
Premature distress additional tie bars and longitudinal steel that is continuous through transverse construction joints. The
Transverse construction joint premise that additional tie bars at TCJs will behave the same way as longitudinal steel, thus reducing steel
stresses of longitudinal steel at TCJs, keeping the joint widths tight, and improving TCJ performance, is not
necessarily correct. Also, the use of additional tie bars makes the consolidation of concrete at TCJs difficult
due to the reduced spacing between longitudinal reinforcing steel. Not placing additional tie bars at TCJs,
and the use of relief transverse saw cut joints near TCJs, along with improved construction practices at
TCJs, will minimize the distresses.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction 50 years. Based on the excellent performance of CRCP in Texas,

As of 2012, Texas Department of Transportation (TxDOT) has
over 13,000 lane miles (20,930 km) of continuously reinforced con-
crete pavement (CRCP), which represents about 6.8% of the high-
way lane mileage in the state, primarily in interstate highways
and US highways where traffic volume is high. The extensive use
of CRCP is based on its quite satisfactory performance, in terms
of performance life and the low frequency of repairs. The statewide
punchout distress development rate of CRCP is, on an average, less
than 0.2 distresses per year per lane mile. If 10 punchouts per mile
(1.61 km) is considered as the terminal condition of CRCP, the
statewide average performance life of CRCP would be about
⇑ Corresponding author. Tel.: +1 806 742 3523.
E-mail address: pangil.choi@ttu.edu (P. Choi).
TxDOT made it a policy to utilize CRCP when a rigid pavement is
selected for a project. However, distresses do occur in CRCP and
Fig. 1 illustrates the classifications of the distresses in CRCP in
Texas [1,2].
In Fig. 1, distresses are those classified as punchouts in TxDOT
Pavement Management Information System (PMIS) [3]. It shows
that about half of the total distresses are due to large spalling,
about 20% at repair joints, 20% at transverse construction joints
(TCJs), and only 15% are real punchouts that are due to structural
deficiency of the CRCP. Except for real punchouts, the rest of the
distresses usually occur within a few years of new or repaired
CRCP, which makes them premature distresses. Premature dis-
tresses are not due to structural deficiency of the CRCP; rather,
they are related to deficiencies in design details and the quality
of materials and construction. In Texas, large spalling occurs when

0950-0618/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2014.01.042
 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219
Fig. 1. CRCP punchout classification in Texas.
coarse aggregates with a high coefficient of thermal expansion
(CoTE) are used. TxDOT now requires that for concrete to be used
in CRCP, the CoTE value of concrete must be less than 5.5  106/
°F (9.9  106/°C). This requirement is expected to reduce large
spalling distresses in Texas [4].
As for the distresses at transverse repair joints (TRJs), there
could be several causes. One is the failure to restore base support
during the repair, and another is the inability to restore the conti-
nuity of longitudinal steel at TRJs when tie bars are epoxy grouted.
Lately, TxDOT improved specifications for CRCP repairs, and devel-
oped a training manual as well [5]. The implementation of the new
specifications and training of personnel conducting CRCP repairs is
expected to minimize the occurrence of distresses at repair joints.
Fig. 2 illustrates a typical distress at TCJs. Most of the distresses
at TCJ are usually observed at a distance of two feet (0.61 m) from
the TCJ. Currently, 50% additional longitudinal steel is placed at the
TCJs in the form of 50 in. (127 cm) long tie-bars, which is accom-
plished by placing a tie bar at every other spacing as shown in
Fig. 3. In Texas, CRCP design standards dating back as early as
1961 provide details of the use of tie-bars at TCJs. Although no doc-
umented evidence exists on the rationale behind the introduction
of additional steel at the TCJs, the reasoning could have been that
if the longitudinal steel amount is deficient, stresses at longitudinal
steel at TCJs can be elevated above an allowable level, resulting in
larger joint widths and distresses. Lower load transfer efficiency
(LTE) and/or larger deflections at TCJs could be considered as po-
tential causes for the distresses at TCJs. However, extensive deflec-
tion testing conducted in a research study, where a total 27
sections were selected and deflection testing using Falling Weight
Deflectometer (FWD) was conducted twice a year, summer and
winter, for 8 years, indicates no substantial difference in LTE and
deflections between at cracks and TCJs [2].
Fig. 2. Typical distress near transverse construction joint.
213
Fig. 3. Typical layout of a test section [10 = 0.305 m; 100 = 25.4 mm].
In order to minimize the distresses at TCJs by identifying their
mechanisms and improving design standards and specifications,
TxDOT initiated a research study, which consisted of investigating
mechanistic behavior of CRCP near the TCJs with various gages, and
evaluating construction and material related issues near TCJs.
2. Field experimental program
Volume changes in concrete in CRCP due to temperature and moisture varia-
tions are almost fully restrained, except for end areas, by longitudinal steel and base
friction, which causes transverse cracks to develop and relieve stresses [6]. The
behavior of concrete and steel in CRCP in almost fully restrained areas has been
investigated [7]. However, a CRCP segment placed near the end of the concrete
placement of the day is not restrained until concrete placed on the other side of a
TCJ gains adequate strength. Also, there is a difference in stiffness of concrete at
both sides of TCJs at early ages. Accordingly, the behavior of CRCP near TCJs might
be different from those of CRCP in other areas. To investigate the behavior of con-
crete and steel in CRCP near TCJs, field experiments were conducted in two projects
as shown in Table 1.
In each project, concrete and steel strains, along with concrete slab displace-
ments in the longitudinal direction, were evaluated. Fig. 3 illustrates a typical test-
ing plan. Concrete strains in the longitudinal direction were measured at 1 ft
(0.305 m), 2 ft (0.61 m), and 4 ft (1.22 m) from TCJs with vibrating wire strain gages
(VWSGs). Steel strains in the longitudinal steel and additional tie bars in the longi-
tudinal direction were obtained at different locations at the joint with electrical
steel strain gages (SSGs). Concrete slab displacements in the longitudinal direction
were obtained at both sides of the slab, as well as across the TCJs, with crackmeters.
The size of longitudinal steel and tie bars used in this project was 19 mm.
2.1. Lubbock test section
This section is located on US 82 (eastbound) in the Lubbock District between IH
27 and US 84. The CRCP structure consists of 13 in. (330 mm) of concrete slab on
4 in. (100 mm) of asphalt stabilized base. Two TCJs were evaluated – one between
CRCPs built on November 15, 2010 and on December 15, 2011 (LBB-I), and the other
between CRCPs built on December 15, 2011 and December 21, 2011 (LBB-II).
The age difference between the concretes at the joint LBB-I is one year and one
month and this joint presents a unique opportunity to investigate the interactions
at a joint between old and new concretes. Since the old section was built before the
research study was initiated, concrete gages were installed only in the new concrete
section. When there is a sufficient time gap between the old and new CRCP place-
ments at TCJs, it is normal practice in TxDOT to place female-piece tie bars at the old
concrete side of TCJs. Prior to the placement of concrete on the other side of a TCJ,
male-piece tie bars are installed at the TCJ and the continuity of longitudinal rein-
forcement is provided by placing and tying longitudinal steel to male-piece tie bars.
2.2. Brownwood test section
The test section denoted as BWD-I was constructed in the Brownwood District,
10 miles (16.1 km) east of Ranger, on IH 20 (west bound). The CRCP structure con-
sists of 13 in. (330 mm) of concrete slab on 4 in. (100 mm) of asphalt stabilized
214 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219

Table 1
Field test section details [1.0 in. = 25.4 mm].

Highway District TCJ ID Slab thickness (in.) Construction date Construction time Construction season
US 82 Lubbock LBB-I 13 11/15/2010 Existing Winter
13 12/15/2011 Morning
LBB-II 13 12/15/2011 Evening Winter
13 12/21/2011 Morning
IH 20 Brownwood BWD-I 13 08/14/2012 Evening Summer
13 08/16/2012 Morning

base. The two CRCP slabs at the TCJ were constructed on August 14, 2012 and Au-
gust 16, 2012. At 6 ft from the inside LCJ and 1 ft and 4 ft from the TCJ, VWSGs were
installed at the mid-depth of the slab in both the morning and evening construction
sides. Six SSGs were also installed in the morning side at BWD-I. Three of these SSGs
were installed on the longitudinal steel and the other 3 on top of additional tie-bars
at a distance of 3 ft, 6 ft and 10 ft from the inside LCJ.
3. Results and discussion
3.1. Behavior of additional and longitudinal steel at transverse
construction joint
Fig. 4(a) and (b) represents the steel strain behavior at LBB-I in
the early ages after construction and 3 months after construction,
respectively. The legends in Figs. 4–6 consist of two parts: the first
indicates the transverse distance from free edge and the second is
the type of bar (L for longitudinal steel and T for additional tie bar).
Concrete temperature at the mid-depth of slab was used for the
temperature analysis. At early ages, the concrete temperature drop
from December 20, 2011 to December 24, 2011 was 22 °F (12.3 °C)
overall, decreasing from 62 °F (16.7 °C) to 40 °F (4.4 °C). This drop
in concrete temperature is rather large. However, the variations
in steel strains were minimal, which indicates that the steel strains
at TCJ were relieved while the concrete temperature was decreas-
ing, potentially due to creep of concrete and the development of
transverse cracks. On the other hand, daily variations in steel
strains follow temperature variations quite well – in the morning
when the concrete temperature is getting lower, steel strains move
to the tension side, while in the afternoon when the concrete tem-
perature increases, steel strains move in the compression direction.
This implies that daily concrete volume changes due to tempera-
ture variations affected steel stresses at TCJ, while the overall trend
of concrete temperature variations during the several days did not

(a) Early Age Steel Strain at BWD-I Test Section

(a) Early Age Steel Strain at LBB-I Test Section

(b) Later Age Steel Strain at BWD-I Test Section

(b) Later Age Steel Strain at LBB-I Test Section
Fig. 5. Steel strain behavior with time-short term, BWD-I [°C = (°F  32)/1.8;
Fig. 4. Steel strain behavior with Time, LBB-I [°C = (°F  32)/1.8; 1 ft = 0.305 m]. 1 ft = 0.305 m].
 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219
(a) Steel Strain at BWD-I right after construction
(b) Steel Strain at BWD-I 3 months after construction
Fig. 6. Steel strain variation with TIME-Long Term, BWD-I [°C = (°F  32)/1.8; 1ft = 0.305 m].
have appreciable effects on steel stresses. It is shown that steel
strains in 2 bars exceeded yield points. As shown in Fig. 4(b), as
the concrete temperature increased from 50 °F (10.0 °C) to 70 °F
(21.1 °C) from March 21, 2012 to March 25, 2012, which is a rather
large temperature increase, steel strain variations were relatively
small. Also, it is noted that, compared with steel strains at the early
ages shown in Fig. 4(a), those about 3 months after concrete place-
ment decreased substantially. Also, comparisons of steel strains at
the same concrete temperature, for example at 60 °F (15.5 °C)
(December 21, 2011) at the early ages and 3 months later (March
21, 2012), indicate significant decrease in steel strains for those
3 months. This finding has a technical significance, because in CRCP
research, it is assumed that steel stresses at transverse cracks or at
concrete discontinuities such as TCJs, are maintained quite high,
and crack widths could be large enough to decrease load transfer
efficiency if steel stresses are excessive. This assumption was made
with the premise that concrete behavior is elastic. It is well known
that concrete exhibits visco-elastic behavior when subjected to
slow loading such as gradual temperature variations. The informa-
tion in Fig. 4(a) and (b) indicates that visco-elastic nature of con-
crete has an effect on concrete slab behavior in CRCP and should
be considered in a mechanistic analysis of CRCP behavior due to
temperature and moisture variations [8].
Fig. 5(a) and (b) illustrates the early and later age behavior of
steel strain at BWD-I. SSGs were installed at 3 ft (0.92 m), 6 ft
(1.83 m) and 10 ft (3.05 m) from the pavement edge. The gage in-
stalled in the longitudinal steel at 10 ft (3.05 m) from free edge did
not provide data after construction and hence was not included in
215
(c) Steel Strain at BWD-I 5 months after construction
(d) Steel Strain at BWD-I 7 months after construction
the data analysis. As can be seen in Fig. 5(a), concrete temperature
dropped from 70 °F (21.1 °C) to 55 °F (12.8 °C) from October 1,
2012 to November 30, 2012. However, the steel strains in the lon-
gitudinal steel, as well as the additional tie-bars, remained rela-
tively steady, while experiencing daily variations due to daily
temperature variations, as observed in the LBB-I section. As shown
in Fig. 5(b), the steel strains in the longitudinal steel at 3 ft (0.92 m)
and 6 ft (1.83 m) from the free edge remained steady at around
1000 le and 500 le in tension, respectively. However, after the
temperature increase from March 7, 2012, steel strains in the tie-
bar at 6 ft (1.83 m) moved into compression, as did those at the
tie-bars at 3 ft (0.92 m) and 10 ft (3.05 m). On the other hand,
strains in the longitudinal steel remained in tension, which indi-
cates that the behavior of longitudinal steel and additional tie bars
is quite different. The primary reason for this discrepancy in the
behavior appears to be the ‘‘continuity’’ condition of the bars.
When the concrete temperature increases, the volumes of both
concrete and steel bars try to increase as well. In general, there is
a good bond between longitudinal steel and surrounding concrete,
and the concrete and longitudinal steel behave as a near composite
material. On the other hand, additional tie bars are not continuous,
and there is a discontinuity at the end of tie bars. When the con-
crete and tie bar volume expands due to temperature increase,
concrete could push tie bars longitudinally towards TCJs, causing
tie bars in compression and, if excessive, tensile stress in concrete
in the vertical direction near the tie bars, potentially resulting in
horizontal cracks or delaminations at the depth of tie bars. This
mechanism of horizontal cracking in concrete slabs has been
216 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219

observed at transverse contraction joints in jointed concrete slabs,

where dowels cause vertical tensile stresses in concrete when tem-
perature increase is excessive. Based on the available CRCP design

º
standards, TxDOT has used additional tie bars as early as 1960, and
it is difficult to form a logical or technical justification for the use of
additional tie bars at TCJs. It could be that a premise was made at
that time that additional tie bars would behave the same way as
longitudinal steel and thus ‘‘complement’’ longitudinal steel at
TCJs, reducing stress level in longitudinal steel at TCJs and improv-
ing overall performance. The data from the field experimentation
indicates that the premise may not be valid.
Fig. 6(a)–(d) provides a more exhaustive display of the steel
strains at BWD-I test section from the early age until 7 months
after construction. They show that the variations in steel strains
in longitudinal steel at 3 ft (0.92 m) and 6 ft (1.83 m) from the
pavement edge remained relatively small and in tension, whereas
the steel strains in tie bars gradually moved towards compression. (a) Rate of concrete strain change at LBB-I TCJ
This disparity in steel strains between longitudinal steel and addi-
tional tie bars provides valuable technical information which
should be considered for improved design of TCJs in CRCP.

º
3.2. Concrete strain behavior near transverse construction joint

Concrete strain variations derived from VWSGs installed in

LBB-I test section are shown in Fig. 7(a) in the form of rate of
concrete strain variations with temperature until about 500 days
from construction. The rate of concrete strain variations is the
largest at 1 ft (0.305 m) from the TCJ, and no practical difference
between at 2 ft (0.61 m) and 4 ft (1.22 m) is observed. Right
after construction, during winter, the rate of concrete strain
change was between 3.5 le/°F (6.3 le/°C) and 4.0 le/°F
(7.2 le/°C). As the concrete temperature increased during sum-
mer, the rate of strain change dropped to between 1.5 le/°F
(2.7 le/°C) and 2.0 le/°F (3.6 le/°C). In the following winter, rate
(b) Rate of concrete strain change at LBB-II TCJ
of concrete strain change increased again to between 4.0 le/°F
(7.2 le/°C) and 5.0 le/°F (9.0 le/°C). With the increase in
temperature during the summer, more than 400 days after
construction, the rate of concrete strain change again dropped,
º

even though the reduction is not as large as in the first cycle

of winter to summer. This pattern of increase and reduction in
the rate of concrete strain changes during winter and summer,
respectively, can be explained by the space (joint width) avail-
able for concrete slab to expand and contract at the TCJ. In
the summer, concrete expands and the space available for con-
crete slab to move at TCJ is restricted, resulting in a smaller rate
of concrete strain changes. On the other hand, in the winter,
concrete contracts and more space becomes available for con-
crete slab to move at TCJ, with a larger rate of concrete strain
changes. However, with time, continued concrete drying shrink-
age appears to provide more space (larger joint width) at TCJ
and the rate of concrete strain changes increases compared to
the previous summer.
(c) Rate of concrete strain change at BWD-I TCJ
Fig. 7(b) and (c) displays the variations in the rate of concrete
strain over time at TCJs constructed in the winter (LBB-II) and sum- Fig. 7. Concrete strain variation at TCJs for each test section [°C = (°F  32)/1.8;
mer (BWD-I), respectively. In Fig. 7(b), little difference in the rate 1ft = 0.305 m].
of concrete strain change is observed at all locations, except in
the second summer. The variations over time are quite similar to
the pattern shown in LBB-I (Fig. 7(a)). winter, around 100 days after construction, the rate of strain
In the summer construction section at BWD-I, Fig. 7(c) shows a change increases to between 2.0 le/°F (3.6 le/°C) and 3.0 le/°F
similar pattern of low concrete strain change rate initially in the (5.4 le/°C).
summer, but the rate increases as winter approaches. These con- It is postulated that when the concrete is placed at the TCJ dur-
crete strain rate changes are at a distance of 1 ft (0.305 m) and ing winter construction, the concrete setting temperature is low,
4 ft (1.22 m) from the TCJ in the evening and morning sections at and as the concrete temperature increases in the summer, concrete
BWD-I. The rate of concrete strain change right after construction expands and the width of the TCJ becomes quite small, thus
of the BWD-I test section in summer 2012 is between 0.3 le/°F minimizing concrete strain variations due to temperature varia-
(0.5 le/°C) and 1.5 le/°F (2.7 le/°C). As the temperature drops in tions. In the following winter seasons, as the concrete temperature
 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219 217

decreases, the width of TCJ becomes large, allowing more concrete

strains due to temperature variations.

3.3. Longitudinal slab movement at transverse construction joint

Slab movement in the longitudinal direction at the TCJ was

evaluated using crackmeters at LBB-I TCJ. Crackmeters were
installed individually to each slab to determine longitudinal dis-
placements of each slab. Also, a crackmeter was installed across
the TCJ to determine the joint opening and closing as shown in
Fig. 8(a).
Fig. 8(b) shows the slab displacements obtained from the crack-
meters. As the concrete temperature dropped after concrete place-
ment on December 15, 2011, the newly placed slab, as well as the
one placed one year and one month earlier, moved towards the
right in Fig. 8(a) (minus displacement in the graph) by almost (a) Unit weight of cores taken on US 287
the same amount for about five days. This indicates composite
behavior between old and new slabs provided by the continuity
of longitudinal reinforcement. It also shows that the newly placed
slab was pulling the existing pavement, possibly due to the drying
shrinkage of the newly placed concrete. It is to be noted that, even
though the modulus of elasticity of concrete in the newly placed
slab should be lower than that in the existing slab at early ages,
the difference in stiffness becomes quite small in a few days, and
the drying shrinkage of the new slab has a dominant effect. In
Fig. 8(b), ‘Across-Relative Displacement’ illustrates the relative dis-
placement measured from the crackmeter installed across the TCJ,
shown in Fig. 8(a), where movement in the positive direction

(b) Dynamic Young’s Modulus of cores taken on US 287

Fig. 9. Concrete density and modulus on US 287 in Wichita Falls District [1 lb/
ft3 = 16.02 kg/m3; 1 psi = 6.894 kPa].

indicates the joint opening. As the drying shrinkage progressed,

the joint opened little in the early age, although there was signifi-
cant temperature drop. The joint opening did not exceed 0.45 mm
(18 mils), which is less than the maximum crack width value rec-
ommended by the AASHTO [9]. It was not feasible to obtain long
term joint width variations because the gages had to be removed
for the placement of tied concrete shoulder. However, it appears
that the continuity of the reinforcing steel at the TCJ and proper
(a) Crackmeter Installation at LBB-I Test Section bonding between concrete and longitudinal reinforcing steel kept
the joint width quite small, at least at early ages.
Not all the information from the field experiment is presented
in this paper. For more complete information, refer to TxDOT re-
search report 0-6687-1: ‘‘Minimizing Premature Distresses in Con-
tinuously Reinforced Concrete Pavement’’ [10].
º

4. Evaluation of material quality and construction practices at

TCJs

In order to evaluate the difference between the concrete near

(b) Slab Movement at LBB-I Test Section
Fig. 8. Slab Movement Evaluation [1.0 mil = 0.0254 mm; °C = (°F  32)/1.8].
and away from the TCJ, concrete cores were taken at two test sec-
tions, one on IH 20 in the Brownwood District and the other on US
287 in the Wichita Falls District. At the Brownwood section, two
cores were taken on both sides of the TCJ, and one core was taken
away from the TCJ. In Wichita Falls, two cores were taken at loca-
tions exhibiting surface distresses (denoted by B1 and B2 in
Fig. 9(a)) near the TCJ, and one core was taken at the location with
no surface distress (denoted by G1 in Fig. 9(a)). Unit weight and
218 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219

is substantial. The dynamic Young’s modulus of the core B2 is

43.8% lower than that of the core G1, which could have an effect
on the behavior of concrete near TCJs.
Fig. 10 shows the unit weight and dynamic modulus test results
of concrete cores taken on US 287 in Wichita Falls District. As can
be seen in Fig. 10(a), the unit weight of cores near the construction
joint in both the evening and morning construction sections are
lower by more than 1% compared to the core taken away from
the TCJ. Here, the evening side represents the concrete placed at
the end of the day and the morning side, the concrete placed in
the following or later day. Fig. 10(b) shows the dynamic Young’s
modulus of the three cores. The dynamic Young’s modulus of the
cores in the morning construction section is almost 9% less, and
that of the evening construction section is around 5% less, than that
of the core taken away from the TCJ.
These results could be due to the inadequate concrete consoli-
dation near TCJs and/or construction practices near TCJs that are
(a) Unit Weight of cores taken on different from other areas. Normally, concrete placement and
IH 20 in Brownwood District vibration near TCJs is done manually as shown in Fig. 11(a), since
the slip-form paver cannot start or end paving right at the TCJ. If
proper manual vibrations near the TCJ are not performed as per
specifications, improper consolidation would lead to different con-
crete material properties, as shown in Figs. 9 and 10. Also, a high
concentration of longitudinal steel in the form of regular steel
and additional tie bars act as an obstacle for proper concrete con-
solidation at TCJs. Insufficient consolidation in concrete could lead
to the presence of air voids in concrete and potentially surrounding
reinforcing bars, resulting in non-uniform steel stresses at TCJs, re-
duced soundness of concrete, and potential distresses near TCJs.
For example, in a CRCP construction project in Texas, in the con-
struction of southbound lanes, tie bars were installed at every
spacing of longitudinal bars at TCJs, i.e., 100% additional tie bars
as opposed to 50% required by TxDOT design standards. In the con-
struction of northbound lanes, 50% additional tie bars were placed
in accordance with TxDOT design standards. Distresses occurred in
almost all TCJs in the southbound lanes, but no distresses in the
northbound lanes. Fig. 11(b) illustrates the repairs done for those
(b) Dynamic Young’s Modulus of cores taken on distresses at TCJs in the southbound lanes. It appears that the
IH 20 in Brownwood District 100% additional tie bars resulted in poor consolidation of concrete
at TCJs and distresses.
Fig. 10. Concrete density and modulus on IH 20 in Brownwood District [1 lb/
ft3 = 16.02 kg/m3; 1 psi = 6.894 kPa].

5. Conclusions and recommendations

dynamic Young’s modulus of elasticity were evaluated for each of
the coring samples [10].
As shown in Fig. 9(a), the unit weight of core B1 is 2.8% lower
and that of B2 is 3.6% lower than that of the core from the location
G1. As shown in Fig. 9(b), the difference in dynamic Young’s mod-
ulus of elasticity between the locations with and without distress
In order to minimize distresses at TCJs in CRCP, field evaluations
of concrete and steel behavior near TCJs were made using various
gages with the objective of improving current CRCP design stan-
dards or specifications. The following conclusions can be derived
on the basis of field investigations and evidence collected:

(a) Concrete Consolidation at TCJ (b) Repair of Distress at TCJ due to

Excessive Additional Steel

Fig. 11. Concrete consolidation and distress at TCJs.

 W. Zhou et al. / Construction and Building Materials 55 (2014) 212–219
1. Longitudinal tie bars at TCJs behave quite differently from lon-
gitudinal steel, primarily because longitudinal steel is continu-
ous while tie bars are terminating at about 25 in. (0.635 m)
from TCJs. When the concrete temperature increases substan-
tially from the setting temperature of concrete, tie bars actually
could be in compression, potentially causing horizontal crack-
ing of concrete at the depth of the steel placement and
distresses.
2. Steel stresses at TCJ decrease over time, potentially due to the
creep of concrete and the development of transverse cracks,
which might invalidate the benefit of additional tie bars at TCJs.
3. The placement of additional tie bars at TCJs could interfere with
concrete consolidation operations near TCJs, creating poor qual-
ity concrete and potentially causing distresses in TCJs.
4. The rate of concrete strain variation with temperature near a
TCJ is higher during the winter than during the summer season,
due to the space available or joint widths at TCJs at different
seasons. In the summer, joint widths get smaller, restricting
the slab displacements when concrete temperature increases.
On the other hand, joint widths get larger in the winter, allow-
ing slab displacements with less restriction. However, this sea-
sonal effect becomes less with pavement age.
5. When there is a significant time difference between the place-
ments of concrete on either side of the construction joint, the
newly placed concrete seems to pull the slab placed earlier on
the other side of the TCJ due to drying shrinkage of newly
placed concrete. Also, the old and new concrete slabs behave
as a near composite structure with quite a small joint width,
at least at early ages.
6. Due to the presence of additional steel at TCJs, the consolidation
of concrete near TCJs is hindered, resulting in formation of air
voids and surface distresses near TCJs.
219
Based on the conclusions derived from the field testing and
investigations conducted, it is recommended that no tie bars be in-
stalled at TCJs. Also, in order to reduce stresses in longitudinal steel
at TCJs, it is further recommended that transverse saw cuts be
made at 5 ft (1.52 m) and 10 ft (3.05 m) from the TCJs in the morn-
ing placement side.
References
[1] Won MC. Just in time technology transfer for Portland cement concrete
pavement construction and rehabilitation. Lubbock (TX): Texas Tech
University; 2012.
[2] Ryu SW, Saraf S, Won MC. Project level performance database for rigid
pavements in Texas, Part II. Austin (TX): Texas Department of Transportation;
2012.
[3] Texas Department of Transportation. 2010 Pavement Management
Information System. Texas Department of Transportation; 2010.
[4] Ryu SW, Choi PG, Zhou W, Saraf S, Senadheera S, Hu J, et al. Optimizing
concrete pavement type selection based on aggregate. Austin (TX): Texas
Department of Transportation; 2012.
[5] Ryu SW, Choi PG, Zhou W, Won MC. Improvements of partial and full-depth
repair Practices for CRCP distresses. Austin (TX): Texas Department of
Transportation; 2012.
[6] ARA Inc., ERES Consultants Division. Guide for mechanistic-empirical design of
new and rehabilitated pavement structures. Part3-Design analysis.Champaign
(IL); 2004.
[7] Nam JH. Early-age behavior of CRCP and its implications for long-term
performance. Austin (TX): The University of Texas at Austin; 2005.
[8] Yeon JH, Choi S, Ha S, Won MC. Effects of creep and built-in curling on stress
development of Portland cement concrete pavement under environmental
loadings. J Transport Eng 2012;139(2):147–55.
[9] AASHTO. AASHTO guide for design of pavement structures. Washington
(DC): American Association of State Highway and Transportation Officials;
1993.
[10] Saraf S, Choi PG, Won MC. Minimize premature distresses in continuously
reinforced concrete pavement. Austin (TX): Texas Department of
Transportation; 2012.