Evaluation of Dowel Alignment Constructability in Portland Cement Concrete Pavements

Shreenath Rao, Kyle Hoegh, Thomas Yu, and Lev Khazanovich

Although dowel bars are an essential design feature of jointed concrete pavements to prevent pumping and faulting, few studies dealing with the constructability of dowel alignment have been conducted. With the emergence of the Magnetic Imaging Tools (MIT) Scan-2 as a nondestructive, robust method of locating dowels, it is now possible to efficiently evaluate the construction quality of dowel bar placement without damaging the newly constructed pavement. A brief review is presented of the results of recent eld studies aimed at examining the constructability of doweled joints in concrete pavements. The dowel bar alignment data collected in this study encompass a wide range of design and environmental conditions and show that alignment within reasonable levelsconsidering equipment, mix design, as well as workmanship and consistencyis achievable in the eld. Distress and performance analyses show that within these levels of misalignment, the pavement performance is not signicantly affected. However, misalignment at a greater level may not be detrimental to pavement performance, and therefore performancebased guidelines may be less strict than the constructability levels collected in this study. Nevertheless, the results shed light on the level of alignment that can be reasonably achieved in the eld.

longitudinally on the transverse joint and placed at middepth to provide maximum shear load transfer capacity in the concrete slab. The following major categories of dowel misalignment were identied by Tayabji (3) (Figure 2): Horizontal translation, Longitudinal translation, Horizontal skew, Vertical translation, and Vertical tilt.

EQUIPMENT Dowel Bar Placement Dowel bars are installed in concrete pavement by using a mechanical dowel bar inserter (DBI) mounted on the paving machine or prefabricated baskets placed on grade before concrete placement. Different factors can affect the constructability in both methods, and both methods of construction require care and attention to numerous details to achieve good results.

Joints are introduced in portland cement concrete (PCC) pavements to allow for thermal expansion and contraction as well as shrinkage after construction. Many transportation agencies include dowel bars in these joints to improve load transfer capacity across the joints, thereby reducing slab deections and stresses (1). The load transfer concept is shown in Figure 1 (2). When a wheel load is applied to an undoweled joint, greater corner deections, and thus transverse joint stresses, are experienced. Dowel bars reduce the critical deections and stresses by transferring the load between the slabs.

Measuring Dowel Bar Location In the past, dowel alignment was rarely measured on newly constructed PCC pavements because there was no simple way to perform this operation. A project sponsored by the California Department of Transportation (Caltrans) in 2003 identified some devices that have recently become available for estimating or measuring dowel position (4 ): Magnetic Imaging Tools (MIT) GmbH Scan-2; Impact echo; Profometer; Micro covermeter; Covermaster; Rebar locator; Fisher Model M-101 rebar locator; Refor 3, Ferroscan; Ground-penetrating radar (GPR); and Kansas State University dowel bar locator.

DOWEL MISALIGNMENT Dowel bars should be placed parallel to both the pavement surface and the longitudinal axis of the pavement in order to minimize longitudinal restraint of the transverse joints. The dowel bar should be centered
S. Rao, Applied Research Associates, Inc., 100 Trade Centre Drive, Suite 200, Champaign, IL 61820. K. Hoegh and L. Khazanovich, Department of Civil Engineering, University of Minnesota, Minneapolis, MN 55455. T. Yu, Federal Highway Administration, 1200 New Jersey Avenue SE, Washington, DC 20590. Corresponding author: S. Rao, srao@ara.com. Transportation Research Record: Journal of the Transportation Research Board, No. 2098, Transportation Research Board of the National Academies, Washington, D.C., 2009, pp. 8693. DOI: 10.3141/2098-09

Although all the devices mentioned can be used for locating dowels or reinforcements in concrete, the only devices that are frequently used for identifying dowel misalignment are the MIT Scan-2, the
86

Rao, Hoegh, Yu, and Khazanovich

87

ered from the 2004 reports by the Midwest Concrete Consortium state representatives (10), e-mail from state representatives, and other sources (11, 12). However, recent improvements in nondestructive dowel location technology have led to development of guidelines based on alignment and limited performance data (8, 13). In the Ontario Ministry of Transportation (MTO) study, MIT Scan-2 was used to measure the misalignment of dowels in a section with suspected misalignment (due to a DBI malfunction) on Hwy 417 east of Ottawa. Typical pavement sections were evaluated for dowel location on Hwy 401 east of Windsor as well as on Hwy 417 on sections paved after the malfunctioning DBI was replaced with a properly working device. As a result of these evaluations, the following tolerance and rejection misalignment levels were formulated (11):
FIGURE 1 Effectiveness of load transfer ( 2 ). ( L loaded slab deflection; U unloaded slab deflection).

profometer, and GPR. Numerous evaluations of the accuracy of the MIT Scan-2 have shown the device to be a reliable tool for locating dowels with high accuracy when dowels inserted with cut dowel baskets or DBIs are considered (4, 5). The accuracy of the device was conrmed in studies by Hossain and Elno in Virginia (6 ) as well as by Leong in Ontario (7 ). The efficiency and ease of use of the device were also found to be robust for application to investigations that require large amounts of measurements in a wide range of environments (8). The robustness of the device caused FHWAs Concrete Pavement Technology Program to initiate a loan program encouraging implementation of this technology, and the device has been identied by FHWA as a practical, implementation-ready product that has the potential to improve the quality of concrete pavements (9).

CURRENT DOWEL ALIGNMENT TOLERANCES Although dowel alignment tolerances were not based on substantial performance data, most agencies have strict limits on them, as shown in Table 1. The dowel alignment tolerance data were gath-

Plan

Plan

Plan

Tolerance criteria: Rotational misalignment (horizontal and vertical): 0.59 in. (15 mm) per 18-in. (457-mm) dowel length; Longitudinal translation: 1.57 in. (40 mm) for an 18-in. dowel; and Vertical translation as follows: 1. Concrete 7.87 in. (200 mm) thick: 0.24 in. (6 mm) toward the base, 0.24 in. (6 mm) toward the surface; 2. Concrete 8.86 in. (225 mm) thick: 0.57 in. (14.5 mm) toward the base, 0.26 in. (6.5 mm) toward the surface; 3. Concrete 9.84 in. (250 mm) thick: 1.10 in. (28 mm) toward the base, 0.47 in. (12 mm) toward the surface; and 4. Concrete 10.24 in. (260 mm) thick: 1.30 in. (33 mm) toward the base, 0.59 in. (15 mm) toward the surface. According to the Ontario specications, the entire transverse joint should be removed and replaced if any individual bar is outside of the following rejectable range: Rotational misalignment: >0.98 in. (25 mm) per 18-in. dowel length; Longitudinal translation: >1.97 in. (50 mm) for a dowel with 18-in. length; and Vertical translation as follows: 1. Concrete 7.87 in. (200 mm) thick: 0.35 in. (9 mm) toward the base, 0.35 in. (9 mm) toward the surface; 2. Concrete 8.86 in. (225 mm) thick: 0.81 in. (20.5 mm) toward the base, 0.49 in. (12.5 mm) toward the surface; 3. Concrete 9.84 in. (250 mm) thick: 1.34 in. (34 mm) toward the base, 0.59 in. (15 mm) toward the surface; and 4. Concrete 10.24 in. (260 mm) thick: 1.54 in. (39 mm) toward the base, 0.71 in. (18 mm) toward the surface. The MTO dowel alignment tolerances were developed with data from several pavement sections in Ontario. This paper presents an analysis of dowel alignment data collected for a large number of sections in the United States.

Horizontal translation

Longitudinal translation

Horizontal skew

Section

Section

FIELD EVALUATION
Vertical translation Vertical tilt

Data Overview In this study dowel alignment data were evaluated for more than 28,000 dowel bars in more than 2,300 joints at 60 pavement sections representing projects in 17 states (Arizona, California, Colorado,

FIGURE 2 Types of dowel misalignment [adapted from work by Tayabji ( 3 )] (1 in. 25 mm).

88

Transportation Research Record 2098

TABLE 1

Dowel Bar Tolerances for Various Agencies Vertical Tilt in. per 18 in. 0.25 0.25 0.25 Horizontal Skew in. per 18 in. 0.25 Longitudinal Translation in. per 18 in. N/A Vertical Translation (in.) N/A

Agency Arkansas Connecticut Federal Aviation Administration Hawaii Idaho Kentucky Minnesota Texas Utah Wisconsin Nebraska Iowa Michigan Montana North Dakota Tennessee Ministry of Transportation, Ontario Nevada Missouri Kansas Indiana North Carolina Illinois Delaware South Carolina Georgia Germany Alabama Great Britain New York Ohio Pennsylvania

0.50 0.25 0.24 0.24 0.59

0.50 0.25 0.59

0.50 0.50 0.38

0.50 0.50 0.38

N/A 0.50 N/A

N/A 1.00 10 pavement depth N/A N/A 0.75 N/A N/A N/A N/A 0.26 0.50 1.00
1

0.19 0.5625 0.75 0.25 0.39 N/A N/A 0.23

0.19 0.5625 0.75 0.69 0.39 0.16 N/A 0.23

N/A 3.00 N/A 2.00 N/A N/A 0.25 0.50 1.00

Georgia, Indiana, Illinois, Kansas, Michigan, Minnesota, Missouri, Nevada, North Carolina, Ohio, South Dakota, Virginia, Washington, and Wisconsin), allowing for a broad range of design, construction, climate, and traffic variables: 1. Climatic region: eight sections in dryfreeze, 24 sections in dry nonfreeze, 22 sections in wetfreeze, six sections in wetnonfreeze; 2. Pavement thickness: ve sections with thickness less than 8 in. (203 mm), ve sections with thickness between 8 and 9 in. (203 and 229 mm), 10 sections with thickness between 9 and 10 in. (229 and 254 mm), 20 sections with thickness between 10 and 11 in. (254 and 279 mm), and 20 sections with thickness greater than 11 in. (279 mm); 3. Dowel size: 16 sections with 1.25-in. (32-mm) diameter, 42 sections with 1.5-in. (38-mm) diameter, and two sections with other values of diameter [1 and 1.125 in. (25 and 29 mm)];

4. Dowel installation procedure: 35 sections that used basket assemblies, 23 sections that used a DBI, and 2 sections using retrot; 5. Construction year: four sections before 1990, 22 sections between 1991 and 1995, 10 sections between 1995 and 2000, 20 sections between 2000 and 2006, and four sections in 2007; and 6. Average daily traffic (ADT): 16 sections had ADT 15,000, 12 sections had ADT between 15,000 and 30,000, 19 sections had ADT between 30,000 and 60,000, and 12 sections had ADT 60,000. Typical Alignment

Vertical Translation (Concrete Cover)

Average vertical translation for individual projects ranged from 1.1 in. (28 mm) to +0.9 in. (23 mm); the distribution of these averages is shown in Figure 3, where it can be seen that although most

Rao, Hoegh, Yu, and Khazanovich

89

Percent of Sections 35 30 25 20 15 10 5 0 < -1.0 -1.0 to -0.5 -0.5 to 0.0 0.0 to 0.5 0.5 to 1.0 Vertical Depth Deviation, in. > 1.0

FIGURE 3 Average vertical translation distribution of sections using actual values (negative values are closer to surface; positive values are closer to base).

projects are within vertical translation limits of 0.5 in. (13 mm), it is possible for entire projects to have an average vertical translation level of more than 0.5 in. This nding can stem from using the wrong dowel baskets (wrong height) when baskets are used. When DBIs are used, this result can be because of improper settings or concrete mix- or slump-related issues. This finding can also be a result of contractors making the pavements thicker than specied. More than 95% of the projects have an average vertical translation level within 1.0 in. (25 mm) of the slab middepth. With absolute values when all bars from all projects are considered together, the average vertical depth deviation is 0.46 in. (11.6 mm) with a standard deviation of 0.6 in. (15 mm).

Longitudinal Translation (Embedment Length)

The average longitudinal translation for all projects was 0.86 in. (21.9 mm). The standard deviations of the longitudinal translation

within individual projects ranged from 0.4 in. (9 mm) to 1.9 in. (49 mm). This finding suggests that some projects were constructed with excellent longitudinal position control, whereas a few projects had significant variability in the longitudinal position of the dowel bars. A key factor influencing longitudinal translation is the accuracy of the joint marking and saw-cut operation during construction. The average standard deviation for all projects was 0.9 in. (23 mm). The individual project maximum with respect to average longitudinal translation reached 1.9 in. (49 mm), resulting in the lowest average embedment length of 7.1 in (180 mm). The standard deviation of the longitudinal translation for all the individual dowels tested was 1.2 in. (30 mm). Figure 4 shows the longitudinal translation distribution for all bars from all projects. The bar graph shows that more than 91% of all bars are within 2 in. (50 mm) from the transverse joint. About 98% of all bars are within 3 in. (75 mm) from the transverse joint.

Percent of Sections 25

20

15

10

0
0.0 to 0.25 0.5 to 0.75 1.0 to 1.25 1.5 to 1.75 2.0 to 2.25 2.5 to 2.75 3.0 to 0.25 to 0.5 0.75 to 1.0 1.25 to 1.5 1.75 to 2.0 2.25 to 2.5 2.75 to 3.0 6.5

Longitudinal Translation, in.

FIGURE 4 Distribution of longitudinal translation of individual dowel bars (all sections).

90

Transportation Research Record 2098

Percent of Sections 70 60 50 40 30 20 10 0
< 0.25 0.25 to 0.5 0.5 to 0.75 0.75 to 1.0 1.0 to 1.25 1.25 to 1.5 1.5 to 3.5

Horizontal Skew, in. FIGURE 5 Distribution of horizontal skew measurements for individual dowel bars (all sections).

Horizontal Skew
Average absolute horizontal skew for individual projects ranged from 0.13 in. (3.3 mm) to 0.41 in. (10.4 mm). The average absolute horizontal skew for all projects was 0.23 in. (5.9 mm). The horizontal skew standard deviations for individual projects ranged from 0.1 in. (2.6 mm) to 0.34 in. (8.7 mm). The average standard deviation for all projects was 0.19 in. (4.7 mm). When all bars from all projects are considered together, the average horizontal skew is 0.24 in. (6.1 mm) with a standard deviation of 0.21 in. (5.3 mm). Fewer than 80% of all bars fell within the current typical 38-in. (9-mm) specications. Figure 5 shows the horizontal skew distribution for all bars from all projects. Almost 90% of the dowel bars have a horizontal skew less than 0.50 in. (12 mm), almost 98% of dowel bars have a horizontal skew less than 0.75 in. (19 mm), and almost 99.5% of dowel bars have a horizontal skew less than 1.0 in. (25 mm).
Percent of Sections 70 60 50 40 30 20 10 0
< 0.25

Vertical Tilt
Average absolute vertical tilt for individual projects ranged from 0.11 in. (2.9 mm) to 0.51 in. (13.1 mm). The average vertical tilt for all projects was 0.24 in. (6.1 mm). The standard deviation of vertical tilt measurements for individual projects ranged from 0.1 in. (2.7 mm) to 0.53 in. (13.5 mm). The average standard deviation for all projects was 0.19 in. (4.9 mm). When all bars from all projects are considered together, the average vertical tilt is 0.23 in. (6 mm) with a standard deviation of 0.21 in. (5.4 mm). Overall these values are almost identical to the horizontal skew values, suggesting no systematic differences between achievability of horizontal skew and vertical tilt. Approximately 80% of all bars fell within the current typical 3 8-in. (9-mm) specications. Figure 6 shows the vertical skew distribution for all bars from all projects. About 91% of dowel bars have vertical tilt less than 0.50 in. (12 mm), 98% of dowel bars have

0.25 to 0.5 0.5 to 0.75 0.75 to 1.0 1.0 to 1.25 1.25 to 1.5

1.5 to 4.0

Vertical Tilt, in. FIGURE 6 Distribution of vertical tilt measurements for individual dowel bars (all sections).

Rao, Hoegh, Yu, and Khazanovich

91

TABLE 2 Students t -Test Results for Significance of Differences in Measures of Dowel Alignment Between Joints Adjacent to Cracked Panels and Joints Adjacent to Intact Slabs in Section 1-AZ3 Measure Vertical translation Longitudinal translation Vertical tilt Horizontal skew Groupa A (n = 16) B (n = 17) A (n = 16) B (n = 17) A (n = 16) B (n = 17) A (n = 16) B (n = 17) Mean, mm 13.67 13.51 17.82 17.23 4.03 3.97 4.98 7.55 SD, mm 4.02 3.17 10.32 9.84 1.17 1.80 1.97 2.72 t-Stat. 0.1252 0.1666 0.1061 3.0921 t-Critical (95% 2-tail) 2.0395 2.0395 2.0395 2.0395 P-Value 0.901 0.869 0.916 0.004

Group A: adjacent to slabs with transverse cracking; Group B: no transverse cracking on adjacent slabs.

vertical tilt less than 0.75 in. (19 mm) and 99% of dowel bars have vertical tilt less than 1.0 in. (25 mm).

Pavement Performance

Transverse Cracking and Joint Spalling

Considering pavement sections with signicant levels of cracking and spalling, the effect of misalignment at the variable alignment points within a section can be evaluated. Project-level analysis was performed for four projects. The project-level analysis of one of these projects (1-AZ3) is presented here. In Project 1-AZ3, 30% of the slabs exhibited transverse cracking and none of the joints had any major spalling. The project-level analysis included a statistical analysis that compared the dowel alignment of joints adjacent to slabs that exhibited transverse cracks with that of joints adjacent to slabs that did not exhibit any transverse cracking. Sixteen of the 33 test section joints were adjacent to slabs with transverse cracking (Group A) and 17 of the 33 joints were adjacent to slabs without any transverse cracking (Group B). Students t-tests were conducted to determine whether there were any statistically signicant differences between these two sets of joints with regard to the average absolute values of vertical and longitudinal translation, vertical skew, and horizontal tilt at the individual joints. The results are summarized in Table 2. Table 3 shows that there is no statistical difference in average vertical translation, average longitudinal translation, and average
TABLE 3 Paired t -Test Results for Faulting and LTE Analysis Performance Indicator Faulting (n = 10) Load transfer efficiency (LTE) (n = 5) Faulting (n = 14) Load transfer efficiency (LTE) (n = 4) Faulting (n = 14) Load transfer efficiency (LTE) (n = 6) Faulting (n = 14) Load transfer efficiency (LTE) (n = 4)

vertical tilt between joints that are adjacent to slabs exhibiting transverse cracking and joints adjacent to intact slabs. There is a statistically signicant difference between the two groups with respect to horizontal skew: the joints adjacent to the intact slabs had higher levels of average horizontal skew than did the joints adjacent to cracked slabs, which is counterintuitive (i.e., higher levels of horizontal misalignment should be associated with higher levels of cracking rather than lower levels if there is any correlation at all). This nding is believed to be a statistical anomaly for this particular section. The actual levels of misalignment for both groups are less than 0.3 in. (8 mm), which suggests that within these levels of misalignment for the eld conditions in the evaluated pavement sections, the performance does not appear to be affected. Similar analyses for test sections 1-AZ9, 1-IL2, and 1-WI2 also show that there is no statistically signicant difference between the average vertical translation, average longitudinal translation, and average vertical tilt in joints adjacent to cracked slabs and those measures for dowels adjacent to uncracked slabs for these sections.

Joint Faulting
For the faulting analyses that follow, only older pavements that exhibited some signicant amount of faulting were considered.
Vertical Translation Paired t-tests were used to compare faulting and load transfer efficiency (LTE) at joints with dowels that were centered within 0.25 in. (6.35 mm) (on average) of middepth with

Comparison of Joints Average vertical translation < 0.25 in. and average vertical translation > 1.0 in. closer to the slab surface Average longitudinal translation < 0.5 in. and joints with average longitudinal translation > 2.0 in. (but < 3 in.) Average vertical tilt < 0.25 in. and joints with average vertical tilt > 0.75 in. Average horizontal skew < 0.25 in. and joints with average horizontal skew > 0.75 in.

Pearsons Correlation 0.636 0.654 0.5366 0.5845 0.8846 0.9630 0.7969 0.9617

t-Stat. 0.1605 0.9848 0.2723 1.6775 2.5415 0.7732 0.7797 2.2284

t-Critical (95%, 2-tail) 2.2622 2.7764 2.1603 3.1824 2.1604 2.5706 2.1604 3.1824

P-Value 0.876 0.380 0.790 0.192 0.024 0.474 0.4496 0.1122

92

Transportation Research Record 2098

those that had dowels centered more than 1.0 in. (25.4 mm) closer (on average) to the pavement surface. The average vertical translation at each joint at each project was computed. For a given project, the average faulting of all joints with the smaller level of vertical translation (Group 1) was paired with the average faulting of all joints with the higher level of vertical translation (Group 2). Faulting measurements were taken in the wheelpath and at the slab edge, and the maximum of the two values was used as the representative value for a given joint. Ten projects with sufficient data points were available to perform this paired t-test, which allows for evaluating the effects of just the dowel alignments; climate, traffic, design, materials, layers, and support conditions were the same for a pair but differed between different pairs. The same procedure was followed for joint LTE: the average LTE of all joints in Group 1 was paired with the average LTE of all joints in Group 2 for each project. Five projects with sufficient data points were available for performing this paired t-test. The summary of the t-test results is shown in Table 3. The relatively high P-values suggest that there are no statistically signicant differences in faulting or LTE between the two groups [i.e., joints with average vertical translation < 0.25 in. and joints with average vertical translation >1.0 in. (25.4 mm) closer to the slab surface]. It should be noted that the faulting levels considered in this study were extremely low, and the vertical translations greater than 1.0 in. closer to the surface were on thick slabs with sufficient cover [i.e., the dowels were still 4 to 5 in. (102 to 127 mm) from the surface of the slab].
Longitudinal Translation Paired t-tests were used to compare faulting and LTE at joints with dowels that were centered within 0.5 in. (12.7 mm) (on average) of the transverse joints with those that had dowels that were centered more than 2.0 in. (5 mm) (on average) from the transverse joints. It should be noted that longitudinal translations more than 3 in. (76.2 mm) were observed on some newly constructed pavements and were not included in the performance analysis because they were too young. Fourteen projects with sufficient data points were available to perform this paired t-test. The same procedure was followed for joint LTE, where the average LTE of all joints in Group 1 was paired with the average LTE of all joints in Group 2 for each project. Only four projects with sufficient data points were available to perform this paired t-test. A summary of the results of the t-test analysis is shown in Table 3. The high P-values suggest that there is no statistically signicant difference in faulting or LTE between the two groups [i.e., joints with average longitudinal translation <0.5 in. of the transverse joint and average longitudinal translation >2.0 in. (but less than 3.0 in.) of the transverse joint]. Therefore, within sections with an average minimum embedment length of more than 6 in. (152.4 mm), the longitudinal translation did not have an effect on faulting. It should be noted that the faulting levels measured on this study were extremely low. Vertical Tilt

points were available to perform this paired t-test. A summary of the results of the t-test analysis is shown in Table 3. The P-value of 0.024 for faulting suggests that there is a statistically signicant difference in faulting between the two groups (joints with average vertical tilt <0.25 in. and joints with average vertical tilt >0.75 in.) with respect to faulting. The joints with higher average vertical tilts had higher levels of average faulting. It should be noted that the faulting levels are low, and only a small number of joints at each section had an average tilt >0.75 in. There was no statistically signicant difference in LTE between the two groups, as indicated by the relatively high P-value of 0.474.
Horizontal Skew Paired t-tests were used to compare faulting and LTE at joints with dowels that had horizontal skews of less than 0.25 in. per 18 in. (457 mm) (on average) with those that had dowels with horizontal skews greater than 0.75 in. per 18 in. (on average). The average horizontal skew at each joint at each project was computed. Fourteen projects with sufficient data points were available to perform this paired t-test. The same procedure was followed for joint LTE, where the average LTE of all joints in Group 1 was paired with the average LTE of all joints in Group 2 for each project. Four projects with sufficient data points were available to perform this paired t-test. A summary of the results of the t-test analysis is shown in Table 3. The P-value of 0.45 suggests that there is no statistically signicant difference in faulting between the two groups (joints with average horizontal skew <0.25 in. and joints with average horizontal skew >0.75 in.). The P-value of 0.11, however, suggests that there is moderate statistical signicance in the differences in LTE between groups of joints with these different levels of horizontal skew. The joints with higher average horizontal skews had slightly lower joint LTE. It should be noted that (a) the faulting levels are extremely low, (b) only a small number of joints at each section had an average skew >0.75 in., and (c) the number of sections available for LTE comparisons is small. Any conclusions drawn from the foregoing analysis should consider these factors.

Constructability Analyses of the eld data from 60 projects in this study indicated that the joints had dowel misalignment constructability (90% of all dowel bars) within the following limits: Vertical translation: 0.5 in. for pavements less than 12 in. (305 mm) thick, Horizontal skew: 0.5 in. per 18 in. (457 mm), Vertical tilt: 0.5 in. per 18 in., and Longitudinal translation: 2 in. for 18-in. dowels. These constructability levels have been shown to be realistically achievable with regard to consistent placement of the dowels within these levels of misalignment, taking into account limitations that include constructability factors such as the ability of construction workers and variability in the equipment used and in the mix design and factoring in any environmental variation from site to site. Dowel alignment within those constructability levels should not affect pavement performance either. In this study pavement sections with somewhat higher levels of misalignment also did not exhibit higher levels of distresses. However, within individual sections, several joints with dowels exhibiting higher levels of vertical tilt

Paired t-tests were used to compare faulting and LTE at joints with dowels that had vertical tilts less than 0.25 in. (6.35 mm) (on average) with those that had dowels with vertical tilts greater than 0.75 in. (19.05 mm) (on average). The average vertical tilt at each joint at each project was computed. Fourteen projects with sufficient data points were available to perform this paired t-test. The same procedure was followed for joint LTE, where the average LTE of all joints in Group 1 was paired with the average LTE of all joints in Group 2 for each project. Four projects with sufficient data

Rao, Hoegh, Yu, and Khazanovich

93

(i.e., greater than 0.75 in.) developed higher levels of faulting, and some joints with dowels exhibiting significant levels of horizontal skew (i.e., greater than 0.75 in.) developed somewhat greater losses of LTE. Since it is feasible to consistently construct dowels within the constructability levels described, any misalignment outside of these levels could be considered for evaluation. However, misalignment levels outside of these constructability levels may or may not affect pavement performance, and thus performance-based dowel misalignment limitations could be less strict. Furthermore, the primary concern with dowel misalignment is the risk of failure of the pavement before the expected design life. Therefore, dowel misalignment outside of these constructability levels should be evaluated on the basis of the effect on long-term pavement performance.

ACKNOWLEDGMENTS The research described in this paper was conducted under NCHRP Project 10-69. Amir Hanna served as NCHRP Program Officer.

REFERENCES
1. McGhee, K. H. NCHRP Synthesis of Highway Practice 211: Design, Construction, and Maintenance of PCC Pavement Joints. TRB, National Research Council, Washington, D.C., 1995. 2. Concrete Paving: 100 Years of Progress Through Innovation. Concrete in Highway Transportation, No. 10. Portland Cement Association, Skokie, Ill., Nov. 1991. 3. Tayabji, S. D. Dowel Placement Tolerances for Concrete Pavements. In Transportation Research Record 1062, TRB, National Research Council, Washington, D.C., 1986, pp. 4754. 4. Khazanovich, L., H. T. Yu, and R. Stubstad. Nondestructive Dowel Bar Detection in Existing Rigid Concrete Pavement Slabs. California Department of Transportation, Sacramento, 2003. 5. Yu, H. T., and L. Khazanovich. Dowel Bar Alignments of Typical In-Service Pavements. R&D Serial No. 2894. Portland Cement Association, Skokie, Ill., 2005. 6. Hossain, S., and M. K. Elno. Field Demonstration of Magnetic Tomography Technology for Determination of Dowel Bar Position in Concrete Pavement. VTRC 06-R40. Virginia Transportation Research Council, Charlottesville, June 2006. 7. Leong, P. An Assessment of MIT-Scan Accuracy and Repeatability. Thesis. University of Waterloo, Ontario, Canada, 2006. 8. Best Practices for Dowel Placement Tolerances. TechBrief. FHWA, U.S. Department of Transportation, 2007. 9. Use of Magnetic Tomography to Evaluate Dowel Bar Placement. TechBrief. FHWA, U.S. Department of Transportation, 2005. 10. State Representative Reports. Midwest Concrete Consortium (MCC), Madison, Wisc., April 2004. http://www.ctre.iastate.edu/mcc. 11. Lane, B., and T. Kazmierowski. Use of MIT Scan Data for Improved Dowel Bar Tolerances. Proc., International Conference on Long-Life Concrete Pavements, Chicago, Ill., FHWA, U.S. Department of Transportation, 2006. 12. Lechner, B. Joint Design and Joint Performance of Plain Concrete Pavements (JPCP): Investigations and Experiences in Germany. Proc., 8th International Conference on Concrete Pavements, Colorado Springs, Colo., 2005. 13. Construction Specication for Concrete Pavement and Concrete Base. OPSS 350. Ontario Provincial Standards for Roads and Public Works, Ontario, Canada, 2007.
The Portland Cement Concrete Pavement Construction Committee sponsored publication of this paper.

CONCLUSIONS Although the emergence of MIT Scan-2 has resulted in some recent studies on dowel misalignment in the eld, most of these studies have been performed to evaluate sections with suspected misalignment or to confirm the accuracy of the device, with very few studies being focused on the constructability of dowel alignment. One of these recent studies was conducted for the conditions in Ontario (13), and another study was conducted with limited performance data (8). In the study described here, the constructability of concrete pavements was evaluated by using data from 60 pavement sections with a wide range of design, construction, climate, and traffic variables, and a performance evaluation was made of the dowel alignments within the achievable limits. This evaluation resulted in constructability levels for each type of misalignment that are achievable within worker capabilities, equipment limitations, and design variability and are also shown to have a negligible effect on performance. These constructability levels differ substantially from the current strict tolerances for most highway agencies (Table 1). Since the constructability levels identied in this paper are regularly achievable in the eld and have a negligible effect on performance, it is feasible to evaluate the effect of misalignment on pavement performance outside of these levels. A comprehensive evaluation of the effects of dowel bar misalignment on pavement performance that included eld and laboratory investigations as well as analytical modeling was conducted under NCHRP Project 10-69 to provide guidelines on dowel placement tolerance.