U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
|Publication Number: FHWA-HRT-12-035 Date: November 2012|
Publication Number: FHWA-HRT-12-035
Date: November 2012
To validate the findings and conclusions presented in chapter 3, the following additional analyses were performed:
Evaluation of maintenance and repair impacts, intended to demonstrate that maintenance and repair activities on the selected test sections had little, if any, impact on the study findings and conclusions.
Review of IRI time-history data, intended to confirm that the change between the first and last IRI survey dates used in the analyses was reasonable and not due to outliers in the first or last dates.
Review of deflection time-history data, intended to confirm that the change between the first and last FWD (SN or Deff) test dates used in the analyses was reasonable and not due to outliers on the first or last dates.
Assessment of PCC warping and curling, intended to address the effects of warping and curling on the study findings and conclusions for the PCC pavement test sections.
Each of these analyses is explained in this chapter. The first three are based on data specific to the study, while the fourth is based on the findings and conclusions from other literature and another FHWA study carried out at approximately the same time as this one.
Information on maintenance and repair activities for each of the test sections during the time periods in the study was obtained from the LTPP database to assess the potential impact of these activities on ride quality and structural adequacy. The following discussion addresses each of the maintenance activities reported along with the extent of the section affected (length or percent area of the test section). The main types of maintenance activities identified included crack sealing, seal coats, joint sealing, patching (partial-depth, full-depth, skin, and strip patching), and grinding.
Crack sealing is intended to help reduce moisture infiltration through surface cracks into the layers below, which could accelerate pavement deterioration. This maintenance activity is not expected to improve the ride quality or structural adequacy of the pavement. Table 27 summarizes test sections where crack sealing was reported. As shown in the table, for sections where the quantity of crack sealing was reported, the quantity of treatment applied was small compared to the total length of the section.
|040502, Arizona||5/1/2002||Information not available in the LTPP database|
|050217, Arkansas||12/3/2002||Information not available in the LTPP database|
|270509, Minnesota||6/15/1991||10 ft|
|320101, Nevada||5/1/2002||98 ft|
|040123, Arizona||5/1/2001||248 ft|
|040123, Arizona||4/2/2002||Information not available in the LTPP database|
The application of a fog seal coat was reported for LTPP section 040502 in Arizona, where it was applied three times: May 1998, August 2001, and April 2003. However, no information was provided about the area over which the treatment was applied. Similar to crack sealing, this maintenance activity is not expected to impact the ride quality or structural adequacy of the pavement. As noted in the next section, IRI over time data for the section do not indicate a reduction in IRI as a result of the fog seal coats.
Section 050217 in Arkansas is reported to have received joint sealing in February 1997. This maintenance activity helps reduce pumping, faulting, joint spalling, and blowups. Joint sealing will not improve the structural adequacy of a section. However, deflection values are consistent for this section through October 2003, which could be attributed, at least partially, to the application of the joint sealing in 1997.
Another maintenance activity reported for the sections in question is patching, which includes partial-depth patching (for PCC pavements), full-depth patching, skin patching, and strip patching. Patching helps temporarily or permanently address localized pavement distresses, which could improve structural adequacy and affect ride quality, depending on the type of patching. Table 28 summarizes the sections that received patching. As shown in the table, for those sections where the quantity of patching has been reported, the area of patching was relatively small compared to the total section. It is anticipated that only those sections that received full-depth patches may have affected ride quality and structural adequacy; however, no information on the patching extent is available to confirm this.
|LTPP Section||Maintenance Activity Date||Maintenance Activity Description||Quantity|
|050217||12/3/2002||Partial-depth patching of PCC pavements at joints||2 ft2|
|240505||4/1/2009||Machine premix patch (placing premix with paver roller)||Information not available in LTPP database|
|270509||8/1/2001||Skin patching*||76 ft2|
|190101||6/1/2001||Strip patching**||2,800 ft2|
|190103||7/1/2001||Strip patching||2,480 ft2|
|010102||4/17/2003||Strip patching||Information not available in LTPP database|
|010102||4/17/2003||Full-depth patching*** of AC pavement||Information not available in LTPP IMS database|
|050119||5/15/1999||Full-depth patching of AC pavement||560 ft2|
|190108||7/1/2001||Strip patching||3,120 ft2|
*Skin patching is using hand tools or hot pot to apply liquid asphalt and aggregate.
**Strip patching is using a spreader and distributor to apply hot liquid asphalt and aggregate.
***Full-depth patching includes removing damaged material and repairing the supporting layer.
Another maintenance activity reported for the sections in question was grinding performed at two AC sections. Grinding, which can improve skid resistance and ride quality, is reported to have been applied over the entire area for sections 310117 and 310118 in Nebraska on July 12, 2000. The grinding had a reported average depth of 0.5 inches, which could have slightly reduced structural adequacy. Both of these sections are in group 2 and were not considered in the analysis because of grinding that was carried out at the sections (see sections 3.10.3 and 3.10.4).
In summary, limited maintenance and repair activities were applied to some of the test sections during the time periods considered in the analysis. Those activities included crack sealing, PCC joint sealing, patching (skin, strip, and full-depth patching at AC sections and partial-depth patching at a section), and grinding. Of these, only full-depth patching and grinding may have affected ride quality, structural adequacy, or both. However, the impact depends on the extent of the patching. Further information on whether patching affected ride quality is provided in section 4.3. The effect of patching on the observed deflections is discussed in section 4.4.
In chapter 3, the change in IRI over time that occurred at a test section was evaluated by considering IRI at the first profile date and at the last profile date (i.e., the last date available in the database). The time-sequence IRI values at the test sections that were used in this study were evaluated to investigate if any sudden drops in IRI could be noticed, which would indicate that maintenance or a rehabilitation activity had occurred at that site.
The changes in IRI that occurred at the test sections were evaluated by preparing a table and then using the data from this table to prepare plots, with time = 0 being assigned to the construction date of the test section. The construction date was assumed to be the date the section was opened to traffic except for a few sections. For LTPP sections 050119 (Arkansas), 190108 (Iowa), 190101 (Iowa), 190103 (Iowa), 050114 (Arkansas), and 050116 (Arkansas), the first FWD date was before the traffic open date. The first FWD date for these sections was assigned as the construction date. For LTPP section 390205 (Ohio), the first profile date was before the traffic open date, and the first profile date was assigned as the construction date for this section. Table 29 shows the construction dates assigned to the sections as well as the first profile and first FWD dates.
|Group||LTPP Section||State||Construction Date||First IRI Date||First FWD Date|
Table 30 shows the time-sequence right wheel path IRI values of the test sections. The IRI values shown are the average IRI obtained from the five profiler runs performed at the test section on each test date. The time shown in the table is the time from the construction date, with time being zero at the construction date. Appendix C includes plots of the time-sequence IRI values.
|Group||Section||Profile Date||Time (years)||Right Wheel Path IRI (inches/mi)|
Lateral variations in the profiled paths can affect the IRI value.(26) Hence, some variability in the time-sequence IRI values is expected. Maintenance activities, such as patching and spall repairs (PCC pavements) that repair pavement distress, can cause a reduction in IRI. Placement of an AC overlay will cause a significant reduction in IRI. After maintenance or rehabilitation, the IRI values will show a gradual increase with time. An IRI value that is greater than the previous and subsequent values is likely the result of lateral variability in the profiled path. Such a phenomenon can occur in PCC pavements due to curling and warping effects.
The time-sequence IRI values were evaluated to determine if evidence of maintenance or rehabilitation could be detected. A reduction in IRI greater than 10 inches/mi between two adjacent profile dates was noted in the time-sequence IRI values at the following sections:
Group 1, section 010102 (Alabama): IRI for March 14, 2001, March 10, 2002, and January 29, 2003, were 91, 56, and 140 inches/mi, respectively. The cause of the sudden drop in IRI for March 10, 2002, is not clear. The tables do not indicate maintenance was carried out on this section between March 14, 2001, and March 10, 2002.
Group 1, section 040123 (Arizona): IRI ranged between 56 and 59 inches/mi from January 23, 1997, (3d profile date) to March 2, 2003, (10th profile date), except for November 6, 2001, (8th profile date), when the IRI value was 70 inches/mi. The reason for a high IRI value on November 6, 2001, is not clear. The IRI on March 2, 2003, (10th profile date) was 56 inches/mi. The IRI values for the profile dates subsequent to March 2, 2003 (March 10, 2004, March 15, 2005, and March 27, 2006) were 93, 69, and 113 inches/mi, respectively. These values show a drop in IRI of 24 inches/mi between March 10, 2004, and March 15, 2006, and an increase in IRI of 44 inches/mi between March 15, 2005, and March 27, 2006. IRI values for the five runs on each of these dates showed significant variations, and the operator comments indicated there was cracking and raveling along the wheel-paths. The maintenance tables do not indicate repairs were carried out on this section during this time frame. Accordingly, the variations in IRI noted at this section seem to be caused by lateral variations in the profiled path.
Group 1, section 190108 (Iowa): The last four IRI values at this section (May 29, 2001, August 6, 2001, November 22, 2002, and September 22, 2004) were 134, 119, 123, and 139 inches/mi, respectively. These values show a drop in IRI of 15 inches/mi between May 29, 2001, and August 6, 2001. The maintenance tables indicate that strip patching was performed on this section on July 1, 2001. The reduction in IRI seen between May 29, 2001, and August 6, 2001, may be due to this patching.
Group 2, section 310117 (Nebraska): IRI ranged between 63 and 68 inches/mi for the seven profile dates between November 1, 1995, and March 20, 2000. However, IRI at the second profile date on April 17, 1996, was 81 inches/mi. The reason for the high IRI value is not clear. IRI on May 16, 2001 (eighth profile date) and April 24, 2002 (ninth and last profile date) were 75 and 54 inches/mi, with a reduction in IRI of 21 inches/mi occurring between the two test dates. The maintenance tables indicate grinding was performed on this section around this time period, and the reduction in IRI is attributed to this grinding. IRI at the first profile date for this section was 68 inches/mi, and IRI at the last profile date was 54 inches/mi, which was less than the IRI at the first profile date.
Group 2, section 310118 (Nebraska): IRI for March 20, 2000 (seventh profile date), May 16, 2001 (eighth profile date), and April 24, 2002 (ninth and last profile date) were 82, 63, and 54 inches/mi, respectively, with a reduction in IRI of 19 inches/mi occurring between March 20, 2000, and May 16, 2001. IRI reduced further by 9 inches/mi between May 16, 2001, and April 24, 2002. The maintenance tables indicate grinding was performed on this section around this time period, and the reduction in IRI probably occurred because of the grinding. The IRI at the first profile date for this section was 74 inches/mi, and IRI at the last profile date was 54 inches/mi, which was less than the IRI at the first profile date.
Group 4, section 240505 (Maryland): IRI from June 11, 1992 (1st profile date) to June 30, 2004 (11th profile date) gradually increased from 66 to 116 inches/mi. However on October 20, 1999 (seventh profile date), IRI was 118 inches/mi, with the IRI previous to this date being 96 inches/mi and the IRI subsequent to this date being 95 inches/mi. The maintenance tables do not indicate any activity performed between October 20, 1999, and December 5, 2000, on this section. Hence, the cause for the high IRI on October 29, 1999, is not clear. The IRI for the last three profile dates (April 26, 2005, September 22, 2005, and June 15, 2006) were 134, 115, and 207 inches/mi, respectively. The maintenance tables do not indicate any maintenance activities occurred within this section between April 26, 2005, and September 22, 2005. The profile operator comments indicated that cracking was present in the section, and a reduction in IRI was noted between April 26, 2005, and September 22, 2005. The increase in IRI from September 22, 2005, to June 15, 2006, may have occurred because of lateral variability in the profiled path.
Group 4, section 270509 (Minnesota): IRI on September 20, 2000 (eighth profile date) and August 19, 2001 (ninth profile date) were 124 and 100 inches/mi, respectively. Thereafter, from December 2, 2002 (10th profile date) to June 6, 2005 (13th and last profile date), IRI showed a gradual increase from 100 to 115 inches/mi. These IRI trends indicate some type of maintenance activity was performed on this section between September 20, 2000, and August 19, 2001. The maintenance tables indicate skin patching was performed on August 1, 2001, which probably caused the reduction in roughness.
Group 5, section 040213 (Arizona): IRI increased from 94 to 114 inches/mi from January 25, 1994, to December 12, 2004, with this section being profiled 11 times during this period. The time-sequence IRI values were variable, with four reductions in IRI greater than 10 inches/mi between subsequent profile dates during this period. The cause for this reduction in IRI as well as the variability in the IRI values is not clear and may be due to curling and warping effects of the PCC pavement.
Group 5, section 050217 (Arkansas): IRI was 89 inches/mi on April 9, 2003, 127 inches/mi on April 3, 2002, and 135 inches/mi on March 12, 2004. The cause for the low IRI value on April 9, 2003, is not clear and may be due to curling or warping effects of the PCC slab.
Group 5, section 390205 (Ohio): IRI was 100 inches/mi on November 4, 2001, 92 inches/mi on August 16, 2000, and 90 inches/mi on December 6, 2002. The cause for the high IRI value on August 16, 2000, is not clear and may be due to curling or warping effects of the concrete slab.
As described, there were several cases where an IRI value for a particular date was higher than the preceding and subsequent IRI values. Also, variable IRI values over time were noted at one PCC section. An evaluation of the profile data to investigate the cause for the high IRI or variable IRI was not performed because it was outside the scope of this study.
For sections 310117 and 310118 in Nebraska, the IRI on the last profile date was less than the IRI at the first profile date because of grinding. These two sections were not used in the analysis to determine the relationship between ride quality and structural strength.
For the remaining sections, a reduction in ride quality between subsequent profile dates could be attributed to maintenance activities for only two test sections: sections 190108 (Iowa) and 270509 (Minnesota). In the case of the former, strip patching is believed to have caused IRI to decrease by 15 inches/mi (from 134 to 119 inches/mi), but IRI increased to 139 inches/mi in 3 years. For the latter, skin patching is believed to have caused an IRI reduction of 24 inches/mi (from 124 to 100 inches/mi). Subsequently, IRI increased to 115 inches/mi in 4 years. Neither maintenance activity is expected to have a significant effect on the structural adequacy of the pavement.
Based on the review of the time-sequence IRI data, using the IRI at the first and last profile dates to determine the change in IRI that occurred at the test section appears reasonable. The selected data values do not appear to be influenced by errors in data collection that would have influenced the IRI value.
In chapter 3, the change in SN that occurred over time at a test section was evaluated by considering SN determined from the first and last FWD dates (i.e., last FWD date available in the database). The time-sequence average deflection values obtained below the center of the load plate for a 9,000-lb load at the test sections were studied to evaluate the changes in deflection over time. Such a plot shows if the deflections were affected by patching that was performed at the test sections and also verifies that that the first and last SN values used in the analysis were not outliers.
The modulus of the AC layer changes with temperature, and deflections obtained on AC-surfaced pavements depend on the temperature. The deflection below the load was adjusted to a standard temperature of 68 °F before evaluating the deflections. The deflection adjustment was performed using the procedure outlined in the 1993 AASHTO Guide for Design of Pavement Structures.(23) The AC modulus versus temperature relationship depends on the properties of the AC mix. However, only one set of adjustment factors are provided in the guide and applied to all AC surfaces irrespective of mix properties. The guide presents adjustment factors for AC temperatures between 30 and 120 °F. The validity of these adjustment factors when correcting the deflection from a temperature close to the limits (i.e., 30 and 120 °F) could be questionable. Seasonal effects that cause changes in moisture conditions in the base/subbase and subgrade also can have a significant effect on the deflection measured below the load.
Table 31 shows the time-sequence average deflection below the load for a 9,000-lb load corrected to a temperature of 68 °F for each section. The temperature at the mid-depth of the surface layer (i.e., AC for AC-surfaced pavements and PCC for PCC-surfaced pavements) at the time of testing is also shown in this table. For AC-surfaced pavements, the deflections shown are those obtained along the right wheel path, and the value shown is the average of all deflections obtained within the section. For PCC sections, the average deflections obtained at the center of the slab are shown. The time-sequence deflection plots are included in appendix D. For each section, a plot showing the time-sequence temperature values at the mid-depth of the surface layer is shown below the deflection plot.
|Group||Section||FWD Test Date||Time (years)||Avg. Deflection Below Load for
9,000 lb Corrected for Temperature (mil)
|Temperature of AC (°F)||Deflections|
|Std Dev (mil)||COV
COV = Coefficient of variation.
Evaluation of the time-sequence data showed that major changes in deflections had not occurred at most sites. At 15 of the 21 sections evaluated in this study, the coefficient of variation (COV) of the time-sequence deflection values was 20 percent or less. COV values greater than 20 percent were obtained at the following sites: 010102 in Alabama (23 percent), 390106 in Ohio (27 percent), 310118 in Nebraska (33 percent), 050114 in Arkansas (22 percent), 040502 in Arizona (36 percent), and 040213 in Arizona (38 percent). Section 310118 in Nebraska was not used in the analysis.
The following observations were made at the other sections that had COV values greater than 20 percent:
010102 (Alabama): Overall, the time-sequence deflection data showed a decrease with time with some scatter, which caused the high COV value.
390106 (Ohio): The time-sequence deflection data showed an increase over time, which caused the high COV value.
050114 (Arkansas): The average deflection for the first FWD date was much higher than the average deflection at the other test dates and is attributed as the cause for the high COV. The cause for the high deflections at the first FWD date is not clear.
040502 (Arizona): The time-sequence deflections showed large variability. The temperature at time of testing at this section for the different test dates ranged from 54 to 126 °F. The variability in the deflections may be related to the effects of the temperature adjustment factor.
040213 (Arizona): This is a PCC section, and the deflection in March 1996 was more than double the deflections obtained at the other test section, which resulted in a high COV value. The high deflection may be due to excessive downward curing at the time of testing.
Table 28 shows the patching activities performed on the test sections. There are two cases where full-depth patching was performed. If this patching was along the right wheel path where FWD deflections were performed, the SN value of the pavement may show an increase over the SN obtained for the previous FWD date. However, the increase in SN depends on the quantity of patching that was performed. Section 050119 had 560 ft2 of full-depth patching, and the average deflection after the patching was only 2 percent less than that before patching. Hence, no increase in SN was expected at this section due to patching. The amount of full-depth patching performed on section 190108 is unknown. Skin patching and strip patching are expected to have little, if any, any impact on the structural strength of the pavement. The reason strip patching was performed is not known; filling of ruts is a possible reason.
Based on the review of the time-sequence deflection values, the use of the first and last FWD deflections at a site to characterize the change in SN appears to be reasonable.
The shape of PCC slabs in a pavement can vary depending on the temperature and moisture gradient within the slab. A concrete slab can take the following shapes: (1) curled up, when the joints are at a higher elevation than the center of the slab, (2) curled down, when the center of the slab is at a higher elevation than the joints, or (3) flat. Changes in the slab shape caused by temperature effects are referred to as curling, while changes in the slab shape due to moisture effects are referred to as warping.
When the temperature of the pavement surface is lower than the temperature at the bottom of the slab in the night and early morning hours, the edges of the slab tend to curl upward. When the surface of the slab heats up during the day and is at a higher temperature than the bottom of the slab, the edges of the slab tends to curl downward. These changes in slab shape due to temperature are superimposed on the original shape of the slab. Usually, a pavement can either be curled upward or downward, and the temperature gradient influences the severity of the curvature.(26) A PCC slab can change shape from a curled up position during the early morning hours to a curled down position in the afternoon as the surface of the pavements heats up. The curvature that is built into the slab depends on factors such as mix properties, base support, slab length, slab thickness, and temperature and moisture conditions during curing.(26) Changes in the moisture within a slab can also cause a slab to change shape, and these changes occur over time. For example, if the moisture content in the top portion of the slab reduces over time and is less than at the bottom of the slab, the edges of the slab will curl upward.
Curvature in a slab (either upward or downward) contributes to roughness. Byrum et al. developed an equation to predict the IRI due to slab curvature.(27) From an analysis of GPS-3 data in the LTPP database, Byrum et al. noted undoweled slabs have a higher level of curvature compared to doweled slabs.
As discussed in the previous section, the shape of a PCC slab can change during the day due to temperature variations, which causes changes in roughness. Karamihas et al. presented IRI changes that occurred because of temperature-related slab curling at 11 test sections in the Michigan SPS-2 project.(28) Profile measurements were obtained on March 28, 1997, at this SPS 2 project starting at 5:07 a.m. and ending at 3:42 p.m., with measurements obtained over approximately 10 h. The air temperature during this period increased from 51 to 74 °F. Three of the pavement sections in this project showed no change in IRI over the period. The mean IRI (average of left and right wheel paths) of six sections decreased continuously during testing, with the reduction in IRI ranging from 10 to 25 inches/mi. All of these sections were curled up, and the degree of the upward curling reduced as the air temperature increased, resulting in a decrease in IRI. Two sections showed increases in IRI ranging from 6 to 12 inches/mi. Both of these sections were curled downwards, and the degree of downward curling increased with the increase in air temperature, causing IRI to increase.
Karamihas and Senn studied the progression of IRI at 21 test sections (12 LTPP sections and 9 State supplemental sections) at the LTPP SPS-2 project in Arizona over 16 years after construction.(29) Two of the supplemental sections were surfaced with AC, and the other seven sections were surfaced with PCC. Four of the PCC sections were non-doweled, and the other three were doweled and had designs that were of interest to the Arizona Department of Transportation. The results obtained by Karamihas and Senn at the 12 LTPP SPS-2 sections are presented in this section.(29)
Investigation of the profiles obtained at the SPS-2 sections indicated roughness in some test sections was caused by longitudinal and transverse cracking and some built-in localized defects. However, slab curvature had a significant contribution to roughness in some sections and, in some cases, was the dominant contributor to roughness. Evaluation of the time-sequence roughness values at the test sections showed significant variability in roughness from year to year at many sections because of changes in slab curvature due to diurnal and seasonal effects. In this study, objective profile analysis methods were used to quantify the level of slab curvature in each section. A procedure for estimating the gross strain gradient needed to deform the slab to the shape noted in the profile was developed to produce a pseudo strain gradient (PSG) value. The level of slab curvature in the section was then summarized by the average PSG value.
Variations in the average PSG over time were able to explain the changes in roughness over time at the sections that had a PCC design flexural strength of 550 psi. A good correlation existed at these sections between the change in IRI and the change in PSG. Table 32 shows the change in IRI that occurred at the test sections that were designed for a flexural strength of 550 psi over a 16-year period, with a positive change indicating an increase in IRI. This table also shows the IRI when the IRI contribution caused by slab curvature was removed. All of these sections were curled up, and the degree of upward curling increased with time. The increase in upward curling of the slabs was the major contributor to the increase in IRI at these test sections. Although not stated by Karamihas and Senn, the increase in upward curling with time was likely related to slab warping.
|PCC flexural strength (psi)||550||550||550||550||550||550|
|PCC thickness (inches)||8||8||8||11||11||11|
|Direction of curl||Up||Up||Up||Up||Up||Up|
|IRI change (inches/mi), left||20||-12||9||34||20||28|
|IRI change (inches/mi), right||101||-7||11||38||24||22|
|IRI change curl removed (inches/mi), left||-3||12||2||-6||6||-6|
|IRI change curl removed (inches/mi), right||72||9||2||-5||12||-5|
Table 33 shows the changes in IRI that occurred at the test sections that were designed for a PCC flexural strength of 900 psi. Little change in IRI was noticed at most of the sections, and a good relationship between the change in PSG and change in IRI could not be established for these sections. Hence, the IRI values with the curling effects removed are not shown in this table.
|PCC flexural strength (psi)||900||900||900||900||900||900|
|PCC thickness (inches)||8||8||8||11||11||11|
|Direction of curl, initial (0.32 years)||Up||Up||Up||Up||Up||Up|
|Direction of curl, initial (16.32 years)||Down||Up||Up||Down||Up||Up|
|IRI change (inches/mi), left||15||-14||4||3||3||20|
|IRI change (inches/mi), right||19||-15||-1||18||3||16|
Upward curling of the pavement at the edges of the slab does not have an effect on FWD tests conducted at the center of the slab, as the center of the slab will be in contact with the base. However, downward curling of the slab where the mid-slab is at a higher elevation when compared to edges will have a significant impact on mid-slab FWD testing. This occurs because the center of the slab will not be in contact with the base when the slab is curled downward, and the degree of downward curling is expected to have a significant impact on the deflections obtained at the center of the slab.
The information presented in the previous section shows that diurnal variations in the shape of a PCC slab can have a significant effect on the IRI value. Additionally, changes in slab curvature can occur over time that can cause an increase in roughness. These changes in roughness have no relationship to the structural capacity of the pavement. However, it should be noted that an increase in downward curling over time can result in premature mid-panel cracking initiating at the bottom of the slab, while an increase in upward curling over time can result in top-down cracking adjacent to the joints.
Information on the maintenance and repair activities applied to each of the test sections during the time periods considered in the study was obtained from the LTPP database to assess the potential impact of these activities on ride quality and structural adequacy. Only a limited amount of maintenance was performed. The maintenance activities carried out at the test sections were crack sealing, fog sealing, joint sealing, partial-depth patching (at a PCC section), grinding, and patching at AC sections (full-depth, skin, and strip). Only grinding and full-depth patching may have an impact on ride quality, structural adequacy, or both.
Based on the review of the time-sequence IRI data, using IRI at the first and last profile dates to determine the change in IRI that occurred at a test section appears reasonable. The selected data values do not appear to be influenced by errors in data collection. Little change in deflections occurred over time at many sections. The COV of the deflections values was less than 20 percent at 15 of the 21 sections evaluated in this study. At the sections that had a COV greater than 20 percent, the high COV values were attributed to: (1) deflections showing an increase with time, (2) deflections showing a decrease with time, (3) possible influence of the temperature adjustment factor on the adjusted deflections, and (4) one deflection in the time-sequence deflections being much higher than the other values.
The two AC sections where grinding was performed were not considered in this study. The limited maintenance activities that were performed on other sections appeared to have little or no influence on either the time-sequence IRI or deflection values. Based on the review of the time-sequence IRI and time-sequence deflection values, the use of the first and last profile and FWD dates to characterize changes in IRI and SN appears to be reasonable.
A past study has shown that significant diurnal variations in IRI can occur. A recently concluded FHWA study showed that the degree of curvature in a slab can increase over time, and curvature of the slab can have a significant effect on the IRI. Changes in IRI that occur due to changes in slab shapes have no relationship with the structural adequacy of the pavement.