Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home   |   Pavements Home
Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-06-106
Date: September 2009

Design and Evaluation of Jointed Plain Concrete Pavement With Fiber Reinforced Polymer Dowels

Previous | Table of Contents | Next

CHAPTER 7. CONCLUSIONS AND RECOMMENDATIONS

INTRODUCTION

In this research, FRP dowel bars with 3.81- and 2.54-cm (1.5- and 1.0-inch) diameters spaced at different intervals as load transferring devices in JPCP were evaluated under static and fatigue loads corresponding to HS25 trucks. Their responses were compared with JPCP consisting of steel dowels under laboratory and field conditions. Performance of JPCP rehabilitated with FRP and steel dowels was also evaluated.

Analysis and discussions corresponding to experimental results and theoretical calculations are summarized in this chapter.

CONCLUSIONS FOR LABORATORY TESTS

Laboratory evaluations were carried out on contraction joints (figure 105) with FRP and steel dowels similar to field implementation. During laboratory tests, crack formations were noted under the joint (referred to as proper crack formation, figure 18) or away from the joint near dowel edges (figure 19 and figure 20). Proper crack formation at joint location was noted in slabnumber 1 (15.24-cm (6-inch)-c/c spacing for 2.54-cm (1.0-inch)-diameter FRP dowels), slab number 4 (30.48-cm (12-inch)-c/c spacing for 3.81-cm (1.5-inch)-diameter FRP dowels), and slab number 5 (30.48-cm (12-inch)-c/c spacing for 3.81-cm (1.5-inch)-diameter steel dowels).

Crack formation at locations away from the joint and close to the edge of the dowel end was observed in slab number 2 (15.24-cm (6-inch)-c/c spacing for 2.54-cm (1.0-inch) diameter for steel dowels) and slab number 3 (15.24-cm (6-inch)-c/c spacing for 2.54-cm (1.5-inch) diameter for steel dowels).

Results mainly focused on slabs with proper crack formation at the mid-joint. Conclusions on laboratory tests are provided with respect to the following:

RD.
LTE.
Pavement pumping.
Strains on dowels.

Joint RD

For Specimens (Numbers 1, 4, and 5) Having Proper Crack Formation at Joint Locations (Table 6)

For static testing, RD decreased by decreasing the dowel bar diameter and spacing (0.0577 cm (0.00227 inch), i.e., 2.54-cm (1.0-inch) diameter at 15.24 cm (6 inches) c/c versus 0.00738 cm (0.029 inch) for 3.81-cm (1.5-inch)-diameter FRP dowel at 30.48-cm (12-inch)-c/c spacing). RD for slab number 4 reduced (0.00738 to 0.0635 cm (0.029 to 0.025 inch)) with the progression of the fatigue load from 0 to 2 million cycles. RD for slab number 5 remained the same (0.0278 cm (0.011 inch)). This is attributed partly to compaction/settlement of aggregate base underneath the pavement slabs with increasing fatigue cycles, resulting in smaller RD. It can also be attributed to similarities in stiffness between FRP (≈ 3.79 × 104 MPa (5.5 × 106 psi)) and concrete (≈ 2.64 × 104 MPa (3.8 × 106 psi)). Benefits of these reductions may be more evident with freeze-thaw variations. In static tests (table 14), RD in slab number 1 with two 2.54-cm (1.0-inch)-diameter FRP dowels had low RD (0.0577 cm (0.00227 inch)) corresponding to HS25 loading. Analytical values for RD were found to be larger than the experimental value (0.0212 cm (0.0084 inch) for the contraction joint and 0.2903 cm (0.01143 inch) for expansion joint). In static tests (table 16), slab number 4 with 3.81-cm (1.5-inch)-diameter FRP dowel had a larger RD (0.0738 cm (0.029 inch)) than slab number 5 with 3.81-cm (1.5-inch)-diameter steel dowel (0.029 cm (0.011 inch)) where the ratio of RDs was 2.64 (0.029/0.011) corresponding to HS25 loading.

For Specimens Having Crack Formation away from Joint Location (Table 7)

Slightly larger RD were noted under static tests (table 15) for slab number 2 (2.54-cm (1.0-inch) steel dowels) and slab number 3 (1.5-inch) FRP dowels) due to crack formation at dowel edges. In addition to higher RD, the load transferred by dowels was less due to that crack formation.

For Base Material Properties (Table 6)

Experimental results on RD were sensitive to supporting base stiffness (k). During tests, the base property (modulus of subgrade reaction, k) value changed from 11.072 to 22.144 kg/cm3 (400 to 800 pci) after 5 million fatigue cycles were applied on slab number 4. RD was expected to increase due to the increase in value of subgrade reaction k.

Joint LTE

AASHTO characterizes LTE value greater than 70 percent as "very good."(1) APCA suggests that 75 percent of joint effectiveness (E) is sufficient for heavy traffic load, which corresponds to 60 percent of LTE as defined by AASHTO.(11)

For Specimens (Numbers 1, 4, and 5) Having Proper Crack Formation at Joint Location

In static tests, all slabs provided good LTE, which was greater than AASHTO and APCA requirements. Slab number 1 (2.54-cm (1.0-inch)-diameter FRP dowels with 15.24-cm (6-inch)-c/c spacing) had a greater than 90 percent LTE. Both slab number 4 (3.81-cm (1.5-inch)-diameter FRP dowel with 30.48-cm (12-inch)-c/c spacing) and slab number 5 (3.81-cm (1.5-inch)-diameter steel dowels with 30.48-cm (12-inch)-c/c spacing) had an LTE of more than 88 percent.

The LTE was found to be 93.79 percent when 1.5 times design load (HS25 loading) was applied for slab number 1 (2.54-cm (1.0-inch)-diameter FRP dowel at 15.24-cm (6-inch)-c/c spacing) after finishing 1 million fatigue cycles. Later, an increased joint width (increased from 0.635 to 1.02 cm (0.25 to 0.4 inch)) after 2 million cycles was considered; the observed LTE in slab number 1 reduced from 93.79 percent to 71.57 percent, but it was still higher than 60 percent of LTE (corresponding to 75 percent of joint effectiveness, E) (ACPA).

In fatigue tests, slab number 4 provided good LTE, greater than 80.5 percent after 5 million fatigue load cycles when the base surface and base material under the slab remained in good condition. When the base aggregates were crushed and some of the aggregates were pushed outside of the slab-base contact area (poor base condition), LTEs were found to decrease to about 55 percent, but they were still around 92.1 percent of the 60 percent LTE, which corresponds to the ACPA recommended value on joint effectiveness, E, 75 percent. Detailed data are shown in figure 43.

At 2 million cycles, slab number 4 containing FRP dowels provided slightly lower LTE (85.1 percent versus 90.21 percent) than slab number 5 containing steel dowels (table 21). With poor base condition (as defined in the previous finding), LTE of the slab with FRP dowels over the slab with steel dowels was 50.71 percent versus 90.21 percent. The most plausible reason causing the low LTE was poor base condition. It should be noted that the modulus of subgrade reaction k changed from 11.072 to 22.144 kg/cm3 (400 to 800 pci) after 5 million cycles.

Compaction of the base was noted especially under the loaded side of slabs during the tests, which could have created slight concave surfaces under the slabs, possibly leading to reduction in LTE. It is suggested to check the base property (such as k, modulus of subgrade reaction) before and after each fatigue test.

For Specimens Having Crack Formation away from Joint Location (Table 7)

Due to the crack formation, LTEs from slab number 2 (2.54-cm (1-inch)-diameter steel dowels with 15.24-cm (6-inch) spacing) and slab number 3 (3.81-cm (1.5-inch)-diameter FRP dowels with 15.24-cm (6-inch) spacing) were slightly lower than that from slab number 1 in static tests (table 18).

Investigation of Pavement Pumping Problem

LTEs were observed in cases investigated for simulated pavement pumping problems with supporting base removal up to a certain length near the joint (figure 47 and figure 50). LTEs were not less than the LTEs obtained from intact base condition (table 20). However, under fatigue load cycles, LTE was expected to reduce significantly for specimens without support near the joint.

The LTE obtained in test case two (30.48 cm (12 inches) base material removal under both sides of slabs) was greater than 90 percent at 13.345 kN (3 kips) loading, and, after loading exceeded 13.345 kN (3 kips), two joint faces would bear against each other. Thus, case two (30.48 cm (12 inches) base removal under both slabs) with unsupported slab areas on both sides of the slab was more detrimental than case one (60.96 cm (24 inches) base removal under loaded slab).

Strains on Dowels

Strain values at the unloaded side of dowels during static tests from slab number 1 and slab number 5 were 513.04 and 376.43 microstrains, respectively. Both values were less than those from analytical evaluation (ranges from 1,000 to 1,200 microstrains) (appendix C). However, strain values are typically not used for LTE calculation.

CONCLUSIONS FOR FIELD APPLICATIONS AND TEST RESULTS

Conclusions for FRP Dowels Used for New Highway Pavement Construction

Effect of Dowel Spacing

For dowel groups in pavement joints 2 and 3 that had the same dowel diameter (3.81 cm (1.5 inches)), joint 2, with smaller dowel spacing (22.86-cm (9-inches)), had a higher LTE (94 percent) than that provided by joint 3, with 30.48-cm (12-inch) dowel spacing (81.58 percent)

Joint 2, with 3.81-cm (1.5-inch) diameter and 22.86-cm (9-inch) spacing, had smaller RD (6.35 × 10-3 cm (0.25 × 10-3 inch)) than joint 3, with same diameter bar and 30.48-cm (12-inch) spacing (17.78 × 10-3 cm (0.70 × 10-3 inch)).

Joint 2, with 3.81-cm (1.5-inch)-diameter dowels and 22.86-cm (9-inch) spacing, provided a 15.4-percent increase in LTE in addition to a 64.3-percent reduction in RD over joint 3, with 3.81-cm (1.5-inch)-diameter dowels and 30.48-cm (12-inch) spacing; refer to table 28.

For pavement joints 5 and 6 with 2.54-cm (1.0-inch) dowel diameters, 20.32- and 15.24-cm (8.0- and 6.0-inch) dowel spacing, respectively, the LTEs were very close (95 percent and 94.44 percent). Relative joint deflections were also identical (2.54 × 10-3 cm (1 × 10-3 inch)).

For JPCP with 2.54- or 3.81-cm (1.0- or 1.5-inch) FRP dowels, larger dowel spacings of 30.48 cm (12 inches) (for 2.54-cm (1.0-inch)-diameter dowels) or 20.32 cm (8 inches) (for 2.54-cm (1.0-inch)-diameter dowels) resulted in higher dowel strains compared to those with 22.86- or 15.24-cm (9- or 6-inch) spacing (for 3.81-cm (1.5-inch) diameter dowels and 2.54-cm (1.0-inch)-diameter dowels, respectively) under AASHTO Type 3 truck loading.

  • For example, for 2.54-cm (1.5-inch)-diameter FRP dowels (A1 and A2), dowels A2 with larger spacing (30.48 cm (12 inches)) had greater strain change (31 microstrain) than dowels A1 with 22.86-cm (9-inch) spacing (strain change of 9 microstrain). Similarly, for FRP dowels C5 and C6 with the same 2.54-cm (1.0-inch) diameter, the dowel C5 with smaller spacing (15.24 cm (6 inches)) showed a small strain change (3 versus 60 microstrain) compared to dowels C6 with 20.32-cm (8-inch) spacing.
  • Decreasing the spacing by 25 percent (30.48 to 22.86 cm (12 to 9 inches) and 20.32 to 15.24 cm (8 to 6 inches)) resulted in more dowels sharing the load within the radius of relative stiffness (lr, see equation 15), leading to 30 percent or higher strain reductions in dowels.

For FRP dowels with 2.54-cm (1.0-inch) diameters (C5 and C6), the increase of spacing from 15.24 to 20.32 cm (6 to 8 inches) had a higher influence on strain value change (3 versus 60 microstrain) than the increase of spacing from 22.86 to 30.48 cm (9 to 12 inches) for dowels (A1 and A2) with 3.81-cm (1.5-inch) diameters (9 versus 31 microstrain).

Dowels with different diameters and spacing could be compared with one another based only on strain value. Because FRP dowels act as a group, spacing and diameter are both important factors for the group action. It should be noted that FRP dowels with smaller diameters typically had better mechanical properties per unit area than larger diameter dowels due to shear lag effects (refer to chapter 6).

Effect of Dowel Diameter

Both 3.81-cm (1.5-inch)-diameter and 2.54-cm (1.0-inch)-diameter FRP dowel groups with spacing varying from 30.48 to 15.24 cm (12 to 6 inches) provided very good LTE (81 percent and higher)—they had LTE greater than 60 percent, which corresponds to ACPA’s 75 percent joint effectiveness (E) value.

Relative Deflection

Currently, there is no requirement or limitation for the RD from AASHTO’s Guide for Pavement Structures.(1) From field tests the maximum RD was 17.78 × 10-3 cm (0.70 × 10-3 inch), corresponding to AASHTO Type 3 truck loading, but, for the laboratory test, the maximum value was 0.109 cm (43 × 10-3 inch) (table 15) corresponding to HS25 loading. It should be noted that joint width due to different joint models (contraction versus expansion joint models) and thermal variables also affected field LTE values.

Conclusions for FRP Dowels Used for Highway Pavement Rehabilitation

Pavement rehabilitation was successfully carried out using FRP dowels near the junction of Routes 119 and 857, University Avenue, Morgantown, WV. After nearly 7 years of rehabilitation, this pavement is performing well without any pavement distress. Strains were monitored on this pavement, which is one of the busiest traffic routes in Morgantown, WV.

  • Strains at loaded and unloaded status from FRP dowels (A and B) (28.17 and 36.24 microstrain) were greater than those from steel dowels (C) (11.49 microstrain), which conforms to analytical findings of shorter FRP dowel length required than the length required for steel dowels and higher deflection in FRP dowels (refer to figure 108).
  • The strain value ratio from the same gauge at unloaded to loaded status does not represent a proper measure of LTE. It is suggested that LTE should be calculated from the pavement deflection measurement.

CONCLUSIONS FOR ANALYTICAL EVALUATION

Calculations have been carried out for dowel diameters (3.81 and 2.54 cm) (1.5 and 1.0 inches) for both FRP and steel dowels. The base modulus of subgrade reaction k = 11.072 kg/cm3 (400 pci), fc′ = 31.026 MPa (4,500 psi), and other parameters considered for calculation are listed in examples 1 through 4.

Conclusions for 3.81-cm (1.5-inch)-Diameter Dowels with 30.48-cm (12-inch)-c/c Spacing

Effect of Dowel Material

Based on current equations, for dowels with the same spacing, steel dowels provided lower values of maximum dowel deflection ( y0 ), dowel shear deflection (s), RD (s), and bearing stress (s) as compared to FRP dowels.

For the same-design spacing (30.48 cm (12 inches)), the maximum bending deflection ( y0 ) of FRP dowels was 56 percent more than those from steel dowels (0.00948 versus 0.00606 cm (3.731 versus 2.386 milli-inches)). Due to a larger shear deflection, the total RD of FRP dowels was 1.95 times the value of RD from steel dowels (0.0237 versus 0.0121 cm (9.33 versus 4.78 milli-inches)).

Effect of Dowel Spacing

Current equations to evaluate JPCP response to wheel load do not include dowel material properties. The radius of relative stiffness, number of effective dowels, and critical dowel load remained identical for FRP and steel dowels for a given spacing.

  • For example, for pavement with 30.48-cm (12-inch)-c/c spacing, FRP and steel dowels with fc′ = 31.026 MPa (4,500 psi) and a joint width of 0.635 cm (0.25 inch) had identical lr (82.342 cm (32.4180 inches)), number of effective dowels (1.89), and critical dowel load (2.161 metric tons (4,763.15 lb)).

Current equations for evaluating JPCP response to wheel load have smaller dowel spacing design results in the same values for the radius of relative stiffness, larger values for the number of effective dowels, and lower values for critical dowel load as compared to a larger dowel spacing.

  • For example, for pavement with 15.24-cm (6-inch)-c/c spacing, FRP and steel dowels with fc′ = 31.026 MPa (4,500 psi) and joint width of 0.635 cm (0.25 inch) had an identical lr (82.342 cm (32.4180 inches)), a larger number of effective dowels (3.2238 versus 1.8895), and a lower critical dowel load (1.266 versus 2.161 metric tons (2,791.76 versus 4,763.15 lb)) as compared to dowels with 30.48-cm (12-inch)-c/c spacing.

Spacings between FRP dowels (3.81-cm (1.5-inch) diameter) less than 17.78 cm (7 inches) provided a maximum dowel bending deflection ( y0 ) (0.006 302 cm (2.481 × 10-3 inch)) that was close to the value (0.00606 cm (2.386 × 10-3 inch)) provided by steel dowels with 3.81-cm (1.5-inch) diameter and 30.48-cm (12-inch) spacing (refer to table 33).

Effect of Joint Width

Shear deflection depended on joint width and was significant for FRP dowels.

For example, when joint width z = 0.635 cm (0.25 inches) with 30.48-cm (12-inch) dowel spacing, shear deflection of the FRP dowels ( s ) was 1.872/9.33 = 20.06 percent of the total RD. Shear deflection of the steel dowels with 0.635-cm (0.25-inch) joint width was only 0.15 percent (0.007/4.78) of its total RD. When joint width was reduced to 0.07938 cm (0.03125 inch) (FRP* case in table 33), the shear deflection of FRP dowels was only 3.30 percent (0.234/7.10) as compared to 20.06 percent with 0.635-cm (0.25-inch) joint width (refer to table 33).

For same-diameter (3.81-cm (1.5-inch)) FRP dowels, use of the contraction joint model greatly reduced the shear effect of dowels.

  • For example, RD for joints with FRP dowel bars were greatly reduced (0.03896 versus 0.019812 cm (15.34 versus 7.80 milli-inches) with joint width of 0.635 cm (0.25 inch) reduced to 0.0794 cm (0.03125 inch).

Effect of Dowel Length

The required FRP length for 3.81-cm (1.5-inch)-diameter dowels was only 64.7 percent (11/17) of that of steel dowels with the same diameter. Based on inflexion points (figure 108), the minimum total length needed for steel dowels was 43.18 cm (17 inches) (2 by 21.59 cm (2 by 8.5 inches)), whereas FRP dowel bars needed 27.94 cm (11 inches) (2 by 13.97 cm (2 by 5.5 inches)).

Effect on Bearing Stress

For a given set of pavement properties in terms of fc′, thickness, joint width, dowel diameter, and spacing, pavement with FRP and steel dowels showed significant differences in deflection and bearing stress value.

For current analytical models, bearing stress around the dowel-concrete (3.81-cm (1.5-inch) diameter) interface is only associated with maximum bending deflection. In order to meet the bearing stress limit (25.856 MPa (3,750 psi) in this case), spacings for steel dowels (3.81-cm (1.5-inch) diameter) cannot exceed 30.48 cm (12 inches), whereas spacing for FRP dowels (3.81-cm (1.5-inch) diameter) should not be more than 17.78 cm (7 inches) (expansion joint) or 19.05 cm (7.5 inches) (contraction joint) (refer to table 33).

Peak bearing stress at the joint location did not take into account the stiffness match between FRP dowel and concrete, which allowed better distribution of bearing stress, leading to reduced bearing stress concentration. Theoretical calculations indicated that allowable stress was exceeded. However, for 3.81-cm (1.5-inch)-diameter FRP dowel bars with 30.48-cm (12-inch)-c/c spacing, the average bearing stress was only 35.4 percent (distance from the joint face to the first inflexion point) and 60.98 percent (within 2.54-cm (1-inch) distance from the joint face) of the peak bearing stress (table 34 and table 35).

Bearing stress (s) could be reduced by increasing dowel diameter (d), pavement slab thickness (h), and concrete strength (fc), or it could be reduced by decreasing dowel spacing (b), joint width (z), modulus of subgrade reaction (k), and modulus of dowel support (K0 ).

Conclusions for 2.54-cm (1.0-inch)-Diameter Dowels with 15.24-cm (6-inch)-c/c Spacing

Effect of Dowel Material

For same-design spacing (15.24 cm (6 inches)), the maximum bending deflection ( y0 ) of FRP dowels was 54.03 percent more than those from steel dowels (0.01138 versus 0.007381 cm (4.479 versus 2.906 milli-inches)). Due to larger shear deflection, the total RD of FRP dowels was 1.96 times the value from steel dowels (0.02903 versus 0.01478 cm (11.43 versus 5.82 milli-inches)).

Effect of Dowel Spacing

Trends of the effect of dowel spacing were found to be similar to those of 3.81-cm (1.5-inch)-diameter FRP dowel bars.

Spacings between FRP dowels (2.54-cm (1.0-inch) diameter) less than 8.89 cm (3.5 inches) provided a dowel maximum bending deflection ( y0 ) (7.132 × 10-3 cm (2.808 × 10-3 inch)) that was close to the value (7.381 × 10-3 cm (2.906 × 10-3 inch) of steel dowels (3.81-cm (1.5-inch) diameter) with 15.24-cm (6-inch) spacing (refer to table 36).

Effect of Joint Width

Shear deflection depended on joint width and was significant for FRP dowels.

  • For example, when joint width z = 0.635 cm (0.25 inch) with 15.24-cm (6-inch) dowel spacing, shear deflection of FRP dowels ( s ) was 2.468/11.43 = 21.59 percent of the total RD. Shear deflection of steel dowels with 0.635-cm (0.25-inch) joint width was only 0.15 percent (0.009/5.82) of its total RD. When joint width was reduced to 0.07938 cm (0.03125 inches) (FRP* case in table 32), the shear deflection of FRP dowels was only 3.70 percent (0.309/8.35) as compared to 21.59 percent (refer to table 33).

For same-diameter (2.54-cm (1.0-inch)) FRP dowels, use of the contraction joint model greatly reduced the shear effect of dowels.

  • For example, RD for joints with FRP dowel bars were greatly reduced (0.0290 versus 0.0212 cm (11.43 versus 8.35 milli-inches)) with joint width of 0.635 cm (0.25 inch) reduced to 0.0794 cm (0.03125 inch).

Effect of Dowel Length

The required length of 2.54-cm (1.0-inch)-diameter FRP dowels was only 69.23 percent (9/13) of the required length of steel dowels with the same diameter. Based on inflexion points (figure 109), the minimum total length for steel dowel was 33.02 cm (13 inches) (2 × 16.51 cm (2 × 6.5 inches)), whereas, for FRP dowels, the minimum length was 22.86 cm (9 inches) (2 × 11.43 cm (2 × 4.5 inches)).

Effect on Bearing Stress

In order to meet the bearing stress limit (31.026 MPa(4,500 psi) in this case), spacings for steel dowels (2.54-cm (1.0-inch) diameter) should be limited to 30.48 cm (12 inches), whereas spacing for FRP dowels (2.54-cm (1.0-inch) diameter) should not be more than 8.89 cm (3.5 inches) (expansion joint) or 10.16 cm (4.0 inches) (contraction joint) (table 36).

For 2.54-cm (1.0-inch)-diameter FRP dowel bars with 15.24-cm (6-inch)-c/c spacing, the average bearing stress was only 34.94 percent (distance from joint face to the first reflection point) and 50.44 percent (within 2.54 cm (1 inch) distance from joint face) of the peak bearing stress (table 37 and table 38). Hence, based on stress redistribution, peak bearing stress did not appear to have damaged the concrete surface.

Previous | Table of Contents | Next

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration