Relating Ride Quality and Structural Adequacy for Pavement Rehabilitation/Design Decisions

CHAPTER 3. DATA AVAILABILITY AND DATA ASSESSMENT

3.1 INTRODUCTION

Roughness indices computed from the profile data collected at LTPP sections are stored in the LTPP database. The roughness indices available in the LTPP database are IRI, root-mean-square vertical acceleration (RMSVA) for base lengths of 1, 2, 4, 8, 16, 32, 64, and 128 ft, slope variance, and Mays output. The IRI of the left and right wheel paths and the mean IRI (average of the left and right wheel paths) are available in the LTPP database. Although RMSVA, slope variance, and Mays output have been used in the past as measures of roughness, these indices are not currently used.

An assessment of the availability of IRI values and FWD data for SPS-1, SPS-2, and SPS-5 was performed to obtain a general idea about the length of time over which data have been collected at these projects.

Thereafter, a few sections were selected from each experiment to investigate the change in ride quality and structural strength that had occurred over time and to investigate if there was a relationship between the change in ride quality and change in structural strength. The primary emphasis was placed on analyzing data from SPS-1 projects. Hence, most of the test sections analyzed were SPS-1 sections.

3.2 DATA AVAILABILITY

Table 6 shows the following information for the SPS-1 projects: State where the project is located, construction date of the project, last available profile date in the database, and number of times profile data were collected at the project. Similar information for SPS-2 and SPS-5 is shown in table 7 and table 8, respectively. The information shown in table 6 through table 8 was obtained from work performed for a previous research project, and more recent roughness data are expected to be available in the most recent LTPP data release. There is a possibility that monitoring of one or more sections in a SPS project may have stopped at an earlier date than that listed as the last available profile date in these tables because of rehabilitation being performed on the test sections. It appears that most SPS-1 test sections that had a weak pavement structure (e.g., sections 2 and 13) failed within a very short time period after being opened to traffic.

Based on the information shown in table 6 through table 8, adequate ride quality data are available to evaluate changes in roughness at SPS-1, SPS-2, and SPS-5 sections. A review of the FWD data for these projects also indicated that adequate test data are available in the LTPP database to evaluate changes in structural strength over time.

Table 6. Roughness data availability for SPS-1 projects.

State	Construction Date	Last Available Profile Date	Number of Times Profiled
Alabama	3/1/1993	5/4/2005	9
Arizona	8/1/1993	3/27/2006	13
Arkansas	3/15/1994	5/22/2007	8
Delaware	5/1/1996	6/13/2006	14
Florida	2/1/1995	7/26/2006	8
Iowa	5/19/1992	9/22/2004	11
Kansas	11/1/1993	3/15/2004	9
Louisiana	7/1/1997	8/7/2006	3
Michigan	11/1/1995	6/2/2006	10
Nebraska	1/1/1995	4/24/2002	9
Nevada	9/1/1995	8/27/2006	12
New Mexico	11/1/1995	4/24/2006	7
Ohio	9/1/1995	8/9/2006	12
Oklahoma	6/1/1997	4/11/2007	8
Texas	6/1/1997	3/19/2007	11
Virginia	11/28/1995	12/2/2006	18

Table 7. Roughness data availability for SPS-2 projects.

State	Construction Date	Last Available Profile Date	Number of Times Profiled
Arizona	10/1/1993	08/11/2006	12
Arkansas	12/1/1995	4/2/2005	6
California	2000	11/7/2004	7
Colorado	11/1/1993	06/02/2006	10
Delaware	5/1/1996	06/11/2006	14
Iowa	12/1/1994	10/29/2004	9
Kansas	8/1/1992	06/05/2006	14
Michigan	11/1/1993	08/16/2006	17
Nevada	9/1/1995	12/3/2003	10
North Carolina	1/1/1994	06/14/2006	18
North Dakota	11/1/1994	5/18/2004	6
Ohio	8/14/1996	08/08/2006	12
Washington	11/1/1995	06/07/2006	11
Wisconsin	1/1/1997	9/12/2005	9

Table 8. Roughness data availability for SPS-5 projects.

State/Province	Construction Date	Last Available Profile Date	Number of Times Profiled
Alabama	12/19/1991	8/3/2006	10
Alberta	10/3/1990	6/12/2006	16
Arizona	4/20/1990	3/24/2006	13
California	4/25/1992	3/20/2007	12
Colorado	10/3/1991	4/24/2000	9
Florida	4/5/1995	7/27/2006	8
Georgia	6/7/1993	5/2/2005	8
Maine	6/20/1995	7/29/2004	8
Manitoba	9/1/1989	6/8/2006	15
Maryland	3/31/1992	6/15/2006	15
Minnesota	9/15/1990	6/6/2005	13
Mississippi	9/24/1990	4/13/1999	5
Montana	9/11/1991	6/7/2005	14
New Jersey	8/18/1992	6/10/2006	18
New Mexico	9/11/1996	4/24/2006	7
Oklahoma	7/8/1997	4/11/2007	8
Texas	9/1/1991	4/10/2007	11

3.3 RIDE QUALITY PARAMETER SELECTED FOR STUDY

In selecting an appropriate ride quality parameter to be used in this study, parameters including PSD plots, smoothness indices developed for selecting locations for WIM scales, and IRI were investigated.

A road profile encompasses a spectrum of sinusoidal wavelengths. A PSD function is a statistical representation of the importance of the various wavelengths contained in the profile.⁽²⁰⁾ A Fourier transform is used to generate a PSD plot from profile data. A PSD plot can assist in data interpretation by detecting cases where a significant portion of roughness is concentrated in a specific waveband and by detecting the type of content (i.e., short or long wavelengths) that dominates the profile.⁽²¹⁾ Profile data collected on the same road section at two different times can be used to investigate the changes that have occurred for a specific wavelength. However, a PSD plot will not give any information about the location within a pavement section where changes have occurred. Hence, an investigation to determine whether there is a correlation between locations within a test section where changes in structural strength have taken place and changes in the ride quality of that area cannot be performed using a PSD plot. If changes in a pavement profile have occurred only at a few localized locations, there is a possibility for a PSD plot to give misleading results, as the PSD plot is generated by fitting sine waves to the pavement profile.

Smoothness indices for finding a suitable location to place a WIM scale were developed under a research project sponsored by FHWA.⁽²²⁾ These indices were developed to see if the smoothness of the pavement before the WIM site is suitable such that the axle hop and body bounce motions of a truck will not have an influence on the weights recorded. Two indices were developed, one based on short wavelengths that cause axle hop and the other based on long wavelengths that cause body bounce motion. A location was deemed suitable for establishing a WIM scale if the pavement before the WIM location met smoothness criteria based on the short and long wavelength smoothness indices. These WIM indices and IRI had a high correlation.⁽²²⁾ Hence, using the WIM indices as the smoothness index in this study would not provide additional information that is not already provided by IRI.

IRI is widely used throughout the world as a ride quality indicator. It is a measurement of the roughness in the road that impacts the vehicle response and, thereby, the riding comfort of passengers in the vehicle.⁽²⁰⁾ IRI correlates well with the vertical passenger acceleration, which is related to the ride quality experienced by a passenger in a vehicle.⁽²⁰⁾ State highway agencies in the United States use IRI to track the ride quality of their pavement network. The profile data on LTPP sections are collected along the left and right wheel path using an inertial profiler, and these data are used to compute the IRI of the wheel paths. The computed IRI values and the profile data are stored in the LTPP database. Locations of high IRI values within a test section can be determined by using a continuous IRI plot (described later in this report). Because of the wide acceptance of IRI by State highway agencies and the ability of a continuous IRI plot to pinpoint locations within a section where high changes in IRI have occurred, IRI was selected as the ride quality parameter to be used in this study.

3.4 STRUCTURAL STRENGTH PARAMETER SELECTED FOR STUDY

3.4.1 Flexible Pavements

The parameters considered to evaluate the structural strength of the pavements in this study include the SN_eff of the pavement and the Structural Adequacy Index (SAI). The SN concept was developed during the AASHO Road Test and is widely used in the United States to design and evaluate the structural strength of flexible pavements. SN_eff can be calculated using deflection data collected by an FWD.

Several agencies use SAI as a means to assess the structural ability of the pavement to support the anticipated traffic loads. Typically, the maximum allowable deflection is determined as a function of the anticipated traffic and compared against the measured deflection to determine SAI. To use this method, the anticipated traffic must be known. Although the LTPP database contains observed traffic information, future traffic estimates are not available. Hence, SN_eff was selected as the parameter to represent the structural strength of the pavement in this study.

The procedure to determine the SN of a flexible pavement from FWD measurements is described in the AASHTO Guide for the Design of Pavement Structures.⁽²³⁾ In this procedure, the deflection data are first used to determine the subgrade modulus. Then, the effective pavement modulus (E_p), which is the effective modulus of all pavement layers above the subgrade, is determined. E_p is calculated as a function of the deflection measured directly beneath the load, the applied pressure, the load plate radius of the FWD, the total thickness of the pavement layers above the subgrade, and the subgrade modulus, as shown in figure 17.

Figure 17. Equation. d₀.

Where:

SN is estimated from E_p using the equation in figure 18.

Figure 18. Equation. SN_eff.

Where:

SN can be determined for each FWD test location. Because the deflection measured at the center of the load plate depends on the temperature of the asphalt layer, this deflection is adjusted to a standard temperature of 68 °F before computing SN.

3.4.2 Rigid Pavements

Effective slab thickness was the parameter selected in this study to define the structural strength of rigid pavements. The FWD data were used to compute an effective concrete slab thickness using the computer program ILLI-BACK. This program computes an effective slab thickness using FWD data based on a user input elastic modulus for concrete. A concrete modulus value of 4.5 million psi was used for this analysis. The program can use either an elastic solid or a dense-liquid foundation below the PCC slab. A dense-liquid foundation was used in this study. The program estimates a modulus of subgrade reaction (k-value) and the effective concrete slab thickness for each FWD sensor at a test location and outputs an average k-value and effective slab thickness.

3.5 CONTINUOUS ROUGHNESS PLOT

IRI is reported in units of inches per mile. The IRI of a pavement segment can be computed over any desired length (e.g., 500 ft, 0.1 mi, 0.5 mi, 1 mi, etc.). The test sections used in the LTPP program for both GPS and SPS studies are 500 ft long except for two sections in the SPS-6 experiment that are 1,000 ft long. For each LTPP section, the IRI of the left and right wheel path computed from the profile data collected at the section are stored in the LTPP database. This stored IRI value represents the average IRI of the 500-ft section. The average IRI does not provide any information about the variability of IRI within a test section.

Consider a pavement section that is 0.1 mi long with a right wheel path IRI of 96 inches/mi. As previously described, this IRI value represents the average IRI over the 0.1-mi length. This section can be divided into 10 52.8-ft segments. The IRI values of these 10 segments are shown in figure 19. The IRI of the 0.1-mile long segment is the average IRI of these 10 52.8-ft segments. Figure 19 shows that the IRI of segment 6 is significantly higher than the rest of the segments and that segment 9 has the lowest IRI of all segments. As shown in the figure, the IRI is not uniform within this 0.1 mi segment, with a significant localized roughness event occurring in segment 6.

Figure 19. Graph. IRI of 52.8-ft segments in 0.1-mi pavement section.

Instead of using a single value to show the IRI of a road over a fixed distance, a continuous roughness plot can be used to show how IRI varies with distance along a roadway. Figure 20 shows a continuous IRI plot based on a 25-ft base length for the same data that were used to compute the IRI of 52.8-ft segments shown in figure 19. The IRI shown at any location in figure 20 is the average IRI over a 25-ft length (i.e., base length of the continuous roughness plot) that is centered at that location. For example, the IRI shown in this plot at a distance of 100 ft is the average IRI from 87.5 to 112.5 ft. Any base length can be used for generating a continuous IRI plot. Typically, a base length of 25 ft is used for locating areas of localized roughness. A detailed description of continuous roughness plots is presented by Sayers.⁽²⁴⁾

Figure 20. Graph. Continuous roughness plot based on 25-ft base length.

3.6 EVALUATING CHANGES IN RIDE QUALITY AND STRUCTURAL CAPACITY OVER TIME

3.6.1 Changes in IRI

Figure 21 shows the right wheel path continuous IRI plots (based on a 25-ft base length) for SPS-1 section 050119 in Arkansas for profile runs performed on July 6, 1995, and March 18, 2004. In 1995, data on this test section were collected immediately after construction. The IRI values of the right wheel path for the 500-ft LTPP section for July 6, 1995, and March 19, 2004, were 47 and 98 inches/mi, respectively. The continuous IRI plot shows most of the increase in IRI occurred between about 150 and 260 ft. This information would not be known if only the average IRI of the section was evaluated. Hence, as shown in this example, continuous IRI plots provide a method to determine the locations within the section where significant changes in IRI occur over time.

Figure 21. Graph. Continuous IRI plots for two test dates.

3.6.2 Changes in Structural Strength—Flexible Pavements

As described earlier, SN was used to characterize the structural strength of flexible pavements. FWD testing is performed on flexible pavements in the LTPP program at 50-ft intervals. An SN value can be computed at each FWD test location. Figure 22 shows the SN values computed from FWD data collected along the right wheel path of SPS-1 section 050119 in Arkansas for March 15, 1994, and May 24, 2004. Such a plot can be used to evaluate the change in SN that occurred at each FWD test location over time.

Figure 22. Graph. SN from FWD tests performed at two test dates.

3.6.3 Change in Structural Strength—Rigid Pavements

As previously described, the effective concrete slab thickness was used to characterize the structural strength of rigid pavements. The data from FWD tests performed at the center of the slab were used for this analysis, and the effective slab thickness was computed at each center slab FWD test location. Similar to the SN analysis, the effective slab thickness determined from tests performed at different time periods was used to evaluate the change in this parameter. Distress within a PCC slab is expected to result in an increase in IRI values, while the effective slab thickness is expected to decrease if distress is located within a slab. An important point to remember is that distress frequently occurs at joints in PCC pavements (e.g., faulting and spalling), which causes an increase in IRI. However, such distress will not impact the deflection obtained at the center of the slab, which is used to compute the effective slab thickness.

3.7 METHODOLOGY FOR COMPARING RIDE QUALITY-STRUCTURAL CAPACITY RELATIONSHIP

For each LTPP test section, profile data are available for five repeat runs for each monitoring date. One representative profile run was selected for generating the IRI values for analysis in this study. The structural capacity data used for analysis are described throughout this section according to pavement type.

3.7.1 Flexible Pavements

For flexible pavements, four load levels (6,000, 9,000, 12,000, and 16,000 lb) were used at each FWD test location, with four drops for each load level. This test sequence resulted in 16 deflection basins for each test location. SN values were computed for all 16 drops and then averaged. The average SN at each test location was used for analysis.

FWD testing on LTPP flexible pavements was performed along the center of the lane and the right wheel path at 50-ft intervals. Profile data at LTPP test sections were collected along the left and right wheel path. Hence, for flexible pavements, the SN values computed from FWD data collected along the right wheel path were compared with a continuous IRI plot for the right wheel path to examine if there was a relationship between the change in IRI and change in SN. Profile testing and FWD testing were not performed on the same day at LTPP test sections. Therefore, IRI and FWD data collected on dates close to each other were selected from the LTPP database for analysis.

Figure 21 shows that a large increase in IRI occurred along the right wheel path of the test section between 150 and 260 ft. If this increase in IRI occurred because of pavement distress along the right wheel path, a reduction in the SN within these limits is expected when SN data for the two test dates are compared. Locations with a low subgrade modulus can result in subgrade settlement or movement, causing a dip in the pavement and resulting in an increase in IRI. The subgrade modulus at each FWD test location was computed using the procedure described in the AASHTO Guide for Design of Pavement Structures and plotted versus distance to evaluate if locations where high changes in IRI occur can be related to locations with a low subgrade modulus.⁽²³⁾ The deflection below the load of the FWD represents the response due to the pavement structure and the subgrade. The deflections below the load of the FWD were also plotted versus the distance and compared with the continuous IRI plot to see if there was any relationship between high changes in roughness and the deflection below the load.

For all flexible pavements considered in this study, the continuous IRI and SN data were plotted on the same graph to evaluate changes and compare the relationship between the two data parameters. Figure 23 shows an example of such a plot, with SN and IRI data for two test dates.

Figure 23. Graph. SN and IRI data on same plot.

Thereafter, normalized plots were created to visualize the percent changes in IRI and SN. Figure 24 shows an example of a normalized plot. The normalized IRI values were computed by dividing the continuous IRI value at each location for the initial and final test dates by the average initial IRI of the test section (i.e., average IRI for the entire section) and then expressing the computed value as a percentage. The normalized SN values were computed by dividing SN at each test location for the initial and final test dates by the average initial SN (i.e., average SN of the section for the initial test date) and then expressing the computed value as a percentage.

Figure 24. Graph. Normalized SN and IRI data on same plot.

3.7.2 Rigid Pavements

For rigid pavements, three load levels (9,000, 12,000, and 16,000 lb) were used at each FWD test location, with four drops for each load level. This test sequence resulted in 12 deflection basins for each test location. The effective concrete slab thickness (D_eff) values were computed for all 12 drops and then averaged. The average D_eff at each test location was used for analysis.

FWD tests on rigid pavements were performed at the center of the slab. The center slab deflections were used to compute D_eff. The continuous IRI plots for the right wheel path were then compared to D_eff to evaluate if a decrease in effective slab thickness was noted at locations that had a high increase in IRI. In addition, the plot showing deflection below the load was compared with the continuous IRI plot to see if any relationship could be detected between changes in IRI and changes in deflection.

Thereafter, normalized plots were created to visualize the percent changes in IRI and D_eff. The normalized IRI values were computed by dividing the continuous IRI value at each location for the initial and final test dates by the average initial IRI of the test section (i.e., average IRI for the entire section) and then expressing the computed value as a percentage. The normalized D_eff values were computed by dividing the D_eff at each test location for the initial and final test dates by the average initial D_eff (i.e., average D_eff of the section for the initial test date) and then expressing the computed value as a percentage.

3.8 TEST SECTIONS SELECTED FOR ANALYSIS

The pavement sections selected for the study were divided into the following five groups:

• Group 1: The sections in this group were selected by evaluating the change in IRI that occurred at SPS-1 sections. The change in IRI since construction was computed for all SPS-1 sections. Thereafter, seven sections were selected for analysis such that the change in IRI for the sections covered a wide range.

• Group 2: A previous study evaluated the change in effective pavement thickness at SPS-1 sections over time.⁽²⁵⁾ The effective pavement thickness in that study was computed from FWD data. Four sections that showed a significant decrease in the effective pavement thickness were selected for analysis.

• Group 3: In a previous study, SPS-1 sections with different base types appeared to behave differently from a structural point of view.⁽²⁵⁾ Based on this observation, a pair of sections with aggregate base and ATB was selected from two SPS-1 projects for analysis.

• Group 4: This group consists of flexible pavements that have been subjected to an AC overlay. Three test sections from the SPS-5 experiment were selected for analysis. The selected sections have had a significant increase in IRI since the overlay.

• Group 5: The sections in this group were selected by evaluating the change in IRI that occurred at SPS-2 sections. The change in IRI since construction was computed for all SPS-2 sections. Thereafter, three sections that showed a significant change in IRI since construction were selected for analysis.

3.9 ANALYSIS OF GROUP 1 SECTIONS

Table 9 shows the following parameters for the test sections selected for analysis: LTPP section number, date corresponding to first IRI (first IRI date), date corresponding to last IRI (last IRI date), FWD test date close to the first IRI date (first FWD date), FWD test date close to the last IRI date (last FWD date), and the time difference between the first and last dates for IRI and FWD data collection.

Table 9. Group 1 sections selected for analysis.

LTPP Section	IRI Date		FWD Date		Time Difference Between First and Last Date (years)
	First Date	Last Date	First Date	Last Date	IRI Data	FWD Data
050119, Arkansas	7/6/1995	3/19/2004	3/15/1994	5/24/2004	8.7	10.2
480114, Texas	9/8/1997	3/19/2007	11/17/1997	6/24/2004	9.5	6.6
310113, Nebraska	11/1/1995	3/20/2000	8/3/1995	10/14/1999	4.4	4.2
010102, Alabama	8/25/1994	5/4/2005	6/21/1995	4/29/2005	10.7	9.9
390112, Ohio	8/14/1996	5/5/2005	11/6/1996	9/1/2004	8.7	7.8
040123, Arizona	1/27/1994	3/27/2006	2/16/1994	4/7/2005	12.2	11.2
190108, Iowa	10/15/1993	9/22/2004	5/19/1993	6/27/2005	10.9	13.9

Table 10 shows the IRI values for the 500-ft-long sections for the first and last profile dates for the selected profile runs as well as the average SN values for the section for the first and last FWD date.

Table 10. IRI and SN for group 1 sections.

LTPP Section	IRI (inches/mi)		SN		Change Between First and Last*
	First Date	Last Date	First Date	Last Date	IRI (in/mi)	SN
050119, Arkansas	48	98	4.51	5.40	50	-0.89
480114, Texas	46	65	10.90	10.39	19	0.51
310113, Nebraska	93	126	5.73	5.82	33	-0.09
010102, Alabama	62	204	2.86	2.19	143	0.67
390112, Ohio	57	101	7.83	7.39	44	0.44
040123, Arizona	52	138	7.94	7.26	86	0.68
190108, Iowa	49	141	8.30	9.94	93	-1.64

*A negative change in SN indicates that the SN increased from the first to the last date.

Table 11 shows the average pavement layer thickness of the selected test section computed from the layer thickness data in the LTPP database.

Table 11. Pavement layer thickness for group 1 sections.

LTPP Section	Layer Type	Layer Thickness (inches)	Material Type
050119, Arkansas	AC	6.8	Hot-mixed, dense-graded
	PATB	3.4	Open-graded, hot laid, central plant mix
	Granular base (GB)	4.1	Crushed stone
480114, Texas	AC	6.8	Hot-mixed, dense-graded
	GB	12.1	Crushed stone
	Treated subgrade (TS)	23	Lime-treated soil
310113, Nebraska	AC	5.1	Hot-mixed, dense-graded
	GB	8	Crushed stone
	Granular subbase (GS)	24	Fine-grained soils, lean inorganic clay
010102, Alabama	AC	4.2	Hot-mixed, dense-graded
	GB	12	Crushed stone
390112, Ohio	AC	4	Hot-mixed, dense-graded
	ATB	11.8	HMAC
	PATB	4	Open-graded, hot laid, central plant mix
040123, Arizona	AC	6.8	Hot-mixed, dense-graded
	ATB	7.9	HMAC
	PATB	3.8	Open-graded, hot laid, central plant mix
190108, Iowa	AC	6	Hot-mixed, dense-graded
	PATB	4.5	Open-graded, hot laid, central plant mix
	GB	8	Crushed stone
	GS	24	Fine-grained soils, lean clay with sand

3.9.1 LTPP Section 050119 (Arkansas)

The first and last profile dates for section 050119 (Arkansas) were July 6, 1995, and March 19, 2004, and the IRI of the right wheel path increased by 50 inches/mi (from 48 to 98 inches/mi) during this 9-year period. The first and last FWD dates for this section were March 15, 1994, and May 24, 2004, and the average SN for this section increased by 0.89 (from 4.51 to 5.40) during this 10-year period. Figure 25 shows the continuous IRI plots for this section for the first and last profile dates, and figure 26 shows the SN values for the first and last FWD dates. Figure 27 shows the deflection measured below the load for a 9,000-lb load for the first and last FWD dates. Figure 28 shows the subgrade modulus estimated from the FWD data for the first and last FWD dates. Figure 29 shows the first and last IRI and SN values on a single plot. The straight, solid line in this plot corresponds to the average IRI at the last profile date. This line can be used as a reference to identify areas that have a high IRI value at the last profile date. Figure 30 shows the normalized IRI and SN data for the first and last dates. For all other sections, only the SN and IRI single plot and the normalized plot are shown. Plots for deflection below the load for a normalized load of 9,000 lb and the subgrade modulus for all sections are included in appendix A.

Figure 25. Graph. IRI for two test dates, section 050119 (Arkansas).

Figure 26. Graph. SN for two test dates, section 050119 (Arkansas).

Figure 27. Graph. Deflection below load for 9,000 lb for two test dates, section 050119 (Arkansas).

Figure 28. Graph. Subgrade modulus for two test dates, section 050119 (Arkansas).

Figure 29. Graph. SN and IRI for two test dates, section 050119 (Arkansas).

Figure 30. Graph. Normalized SN and IRI for two test dates, section 050119 (Arkansas).

For this test section, SN corresponding to the last test date was higher than SN corresponding to the first test date at all test locations. Figure 29 and figure 30 show that a large increase in IRI occurred in this section between about 160 and 240 ft. Only one FWD test point is located within these limits (at 200 ft), and SN corresponding to that test point does not show any significant difference when compared to SN obtained at the other test locations. The average increase in SN for this section between the two FWD test dates was 0.89, with the SN at 200 ft increasing by 0.85. The deflection below the load as well as the subgrade modulus at 200 ft does not show any trends that suggest a weaker subgrade was present at this location. Figure 27 shows the deflection below the load was lower at the last FWD date compared to the first FWD date at all test locations. Figure 28 shows the subgrade modulus for the last FWD date was lower than that for the first FWD date at all locations.

3.9.2 LTPP Section 480114 (Texas)

The first and last profile dates for section 480114 (Texas) were September 8, 1997, and March 19, 2007, and the IRI of the right wheel path increased by 19 inches/mi (from 46 to 65 inches/mi) during this 10-year period. The first and last FWD dates for this section were November 17, 1997, and June 24, 2004, and the average SN of this section decreased by 0.51 (from 10.90 to 10.39), during this 7-year period.

Figure 31 shows the first and last IRI and SN values, and figure 32 shows the normalized IRI and SN plots. The first 120 ft of the test section had a higher increase in IRI compared to the rest of the section. However, the decrease in SN that occurred at the test locations within these limits does not appear to be significantly different from the decrease in SN that occurred at the other test points. The average decrease in SN for this section was 0.51, and the average decrease in SN at the first three FWD locations was 0.46. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 31. Graph. SN and IRI for two test dates, section 480114 (Texas).

Figure 32. Graph. Normalized SN and IRI for two test dates, section 480114 (Texas).

3.9.3 LTPP Section 310113 (Nebraska)

The first and last profile dates for section 310113 (Nebraska) were November 1, 1995, and March 20, 2000, and the average IRI of the right wheel path increased by 33 inches/mi (from 93 to 126 inches/mi) during this 4-year period. The first and last FWD dates for this section were August 3, 1995, and October 14, 1999, and the average SN of this section increased by 0.09 (from 5.73 to 5.82) during this 4-year period.

Figure 33 shows the first and last IRI and SN values for this section, and figure 34 shows the normalized IRI and SN plots. No major changes in IRI occurred at any localized location within the test section. A slight increase in SN occurred over time at about half of the test locations, and the other half showed a slight decrease in SN. No test location showed a change in SN that was vastly different than the rest of the test locations. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 33. Graph. SN and IRI for two test dates, section 310113 (Nebraska).

Figure 34. Graph. Normalized SN and IRI for two test dates, section 310113 (Nebraska).

3.9.4 LTPP Section 010102 (Alabama)

The first and last profile dates for section 010102 (Alabama) were August 25, 1994, and May 4, 2005, and the average IRI of the right wheel path increased by 143 inches/mi (from 62 to 204 inches/mi) during this 11-year period. The first and last FWD dates for this section were June 21, 1995, and April 29, 2005, and the average SN of this section decreased by 0.67 (from 2.86 to 2.19) during this 10-year period.

Figure 35 shows the first and last IRI and SN values, and figure 36 shows the normalized IRI and SN plots. The IRI of the last 300 ft of the test section showed a high increase in IRI when compared to the first 200 ft. IRI of the first 200 ft of the section increased from 75 to 114 inches/mi, and IRI of the section from 200 to 500 ft increased from 53 to 271 inches/mi. The highest increase in IRI occurred between about 260 and 380 ft. The average decrease in SN for the first 200 ft was 0.59, and average decrease in SN between 250 and 500 ft was 0.74. Although the last 300 ft of the section had a higher increase in IRI and showed a larger decrease in SN when compared to the first 200 ft, the decrease in SN of the last 300 ft was only slightly higher than that for the first 200 ft. In addition, a small SN decrease of 0.24 was noted at 300 ft, which was within the limits that showed the highest increase in IRI for this section. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 35. Graph. SN and IRI for two test dates, section 010102 (Alabama).

Figure 36. Graph. Normalized SN and IRI for two test dates, section 010102 (Alabama).

3.9.5 LTPP Section 390112 (Ohio)

The first and last profile dates for section 390112 (Ohio) were August 14, 1996, and May 5, 2005, and the IRI of the right wheel path increased by 44 inches/mi (from 57 to 101 inches/mi) during this 9 year period. The first and last FWD dates for this section were November 6, 1996, and September 1, 2004, and the average SN of this section decreased by 0.44 (from 7.83 to 7.39) during this 8-year period.

Figure 37 shows the first and last IRI and SN values, and figure 38 shows the normalized IRI and SN plots. SN corresponding to the last FWD date was slightly higher than SN corresponding to the first FWD date at the first test location. At all other test locations, SN decreased, with the decrease ranging from 0.37 to 0.76 and an average value of 0.51. High IRI values at the last profile date were noted between approximately 180 to 210 ft, 230 to 310 ft, and 390 to 430 ft. The lowest initial and final SN at this section occurred at 200 ft (within 180 to 210 ft), which was the area that had a high increase in IRI. Hence, for this particular area, a high increase in IRI was associated with a low initial SN. However, the decrease in SN at 200 ft was 0.51, which was equal to the average decrease in SN for this section. Two FWD test points are located between 230 and 310 ft, and the average decrease in SN at these two points was 0.49, while the decrease in SN at 400 ft (within 390 to 430 ft) was 0.45. The decrease in SN at these locations was less than the average decrease in SN for this section. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 37. Graph. SN and IRI for two test dates, section 390112 (Ohio).

Figure 38. Graph. Normalized SN and IRI for two test dates, section 390112 (Ohio).

3.9.6 LTPP Section 040123 (Arizona)

The first and last profile dates for section 040123 (Arizona) were January 27, 1994, and March 27, 2006, and the average IRI of the right wheel path increased by 86 inches/mi from (52 to 138 inches/mi) during this 12-year period. The first and last FWD dates for this section were February 16, 1994, and April 7, 2005, and the average SN of this section decreased by 0.68 (from 7.94 to 7.26) during this 11-year period.

Figure 39 shows the first and last IRI and SN values, and figure 40 shows the normalized IRI and SN plots. A higher decrease in SN was observed between 0 and 200 ft compared to 250 to 500 ft. All test locations between 0 and 200 ft showed a decrease in SN, with an average decrease of 1.48. Four of the six test locations between 250 and 500 ft showed an increase in SN. For the two points that showed a decrease in SN, the average decrease was 0.40. A significant increase in IRI occurred within the section between the following approximate limits: about 80 to 140 ft, 210 to 260 ft, and 300 to 350 ft. An FWD test point that was close to a peak IRI location occurred only at 100 ft. At this location, a decrease in SN of 1.14 was noted. The average decrease in SN at test locations between 0 and 200 ft was 1.57 and was more than the decrease in IRI noted at 100 ft. However, the change in IRI seen at these test locations was much less when compared to the change in IRI seen at 100 ft. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 39. Graph. SN and IRI for two test dates, section 040123 (Arizona).

Figure 40. Graph. Normalized SN and IRI for two test dates, section 040123 (Arizona).

3.9.7 LTPP Section 190108 (Iowa)

The first and last profile dates for section 190108 (Iowa) were October 15, 1993, and September 22, 2004, and the IRI of the right wheel path increased by 93 inches/mi (from 49 to 141 inches/mi) during this 11-year period. The first and last FWD dates for this section were May 19, 1993, and April 24, 2007, and the average SN of this section increased by 1.64 (from 8.30 to 9.94) during this 14-year period.

Figure 41 shows the first and last IRI and SN values, and figure 42 shows the normalized IRI and SN plots. SN at the last FWD date was greater than SN at the first FWD date at all test locations. The peaks in the IRI plot for the last IRI date are locations where major increases in IRI had occurred. Except at one location (450 ft), the FWD data points are located outside the peak IRI areas. The average increase in SN between the FWD test dates at this section was 1.65, with SN at 450 ft showing an increase of 2.13 between the two test dates. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus.

Figure 41. Graph. SN and IRI for two test dates, section 190108 (Iowa).

Figure 42. Graph. Normalized SN and IRI for two test dates, section 190108 (Iowa).

3.10 ANALYSIS OF GROUP 2 SECTIONS

Table 12 shows the following parameters for the test sections selected for analysis: LTPP section number, date corresponding to first IRI (first IRI date), date corresponding to last IRI (last IRI date), FWD test date close to the first IRI date (first FWD date), FWD test date close to the last IRI date (last FWD date), and the time difference between first and last dates for IRI and FWD data collection.

Table 12. Group 2 sections selected for analysis.

LTPP Section	IRI Date		FWD Date		Time Difference Between First and Last Date (years)
	First Date	Last Date	First Date	Last Date	IRI Data	FWD Data
320101, Nevada	12/3/1996	8/7/2006	3/27/1996	8/28/2006	9.7	10.4
390106, Ohio	8/14/1996	8/9/2006	11/5/1996	7/15/2008	10.0	11.7
310117, Nebraska	11/1/1995	4/24/2002	8/2/1995	7/9/2002	6.5	6.9
310118, Nebraska	11/1/1995	4/24/2002	8/3/1995	7/10/2002	6.5	6.9

Table 13 shows the IRI values for the 500-ft-long sections for the first and last profile dates for the selected profile runs as well as the average SN for the first and last FWD date. Table 14 shows the average pavement layer thickness of the group 2 sections computed from the layer thickness data available in the LTPP database.

Table 13. IRI and SN for group 2 sections.

LTPP Section	IRI (inches/mi)		SN		Change Between First and Last*
	First Date	Last Date	First Date	Last Date	IRI (inches/mi)	SN
320101, Nevada	57	57	11.91	11.23	0	0.68
390106, Ohio	71	114	6.94	5.04	43	1.90
310117, Nebraska	68	53	8.32	7.30	-15	1.02
310118, Nebraska	74	54	9.21	7.62	-20	1.59

*A negative IRI indicates IRI at the last date was less than the IRI at the first date.

Table 14. Pavement layer thickness, group 2 sections.

LTPP Section	Layer Type	Layer Thickness (inches)	Material Type
320101, Nevada	AC	7.2	Hot-mixed, dense-graded
	GB	8.5	Crushed gravel
	GS	22.8	Soil-aggregate mixture (predominantly coarse-grained)
	TS	12	Lime-treated soil
390106, Ohio	AC	6.7	Hot-mixed, dense-graded
	ATB	7.9	HMAC
	GB	3.8	Crushed stone
310107, Nebraska	AC	7.1	Hot-mixed, dense-graded
	PATB	3.8	Open-graded
	GB	4	Crushed stone
	GS	24	Fine-grained soils, lean inorganic clay
310118, Nebraska	AC	4.3	Hot-mixed, dense-graded
	ATB	8.2	HMAC
	GB	4	Crushed stone
	GS	24	Fine-grained soils, lean inorganic clay

3.10.1 LTPP Section 320101 (Nevada)

The first and last profile dates for section 320101 (Nevada) were December 3, 1996, and August 7, 2006, and the IRI of the right wheel path remained the same at 57 inches/mi over this 10-year period. The first and last FWD dates for this section were March 27, 1996, and August 28, 2006, and the average SN of this section decreased by 0.68 (from 11.91 to 11.23) over this 10-year period.

Figure 43 shows the first and last IRI and SN values, and figure 44 shows the normalized IRI and SN plots. SN decreased from the first to the last FWD date at all test locations, with the decrease in SN ranging from 0.13 to 1.41. However, as seen in figure 43 and figure 44, there is virtually no noticeable change in IRI throughout the section.

Figure 43. Graph. SN and IRI for two test dates, section 320101 (Nevada).

Figure 44. Graph. Normalized SN and IRI for two test dates, section 320101 (Nevada).

3.10.2 LTPP Section 390106 (Ohio)

The first and last profile dates for section 390106 (Ohio) were August 14, 1996, and August 9, 2006, and the IRI of the right wheel path increased by 43 inches/mi (from 71 to 114 inches/mi) during this 10-year period. The first and last FWD dates for this section were November 5, 1996, and July 15, 2008, and the average SN of the section decreased by 1.90 from (6.94 to 5.04) during this 12-year period.

Figure 45 shows the first and last IRI and SN values, and figure 46 shows the normalized IRI and SN plots. An increase in IRI was observed at most locations within the section, with the highest change in IRI occurring between about 140 and 190 ft. The highest increase in IRI occurred close to 150 ft, where an FWD test was performed. SN decreased from the first to the last FWD dates at all test locations. The decrease in SN at 150 ft was 2.06, slightly higher than the average decrease in SN of 1.90 for this section. A decrease in SN greater than the 2.06 decrease that was observed at 150 ft was noted at 50, 100, 200, and 450 ft. The magnitude of the IRI change at these locations was much less than that observed at 150 ft. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus (see appendix A).

Figure 45. Graph. SN and IRI for two test dates, section 390106 (Ohio).

Figure 46. Graph. Normalized SN and IRI for two test dates, section 390106 (Ohio).

3.10.3 LTPP Section 310117 (Nebraska)

The first and last profile dates for section 310117 (Nebraska) were November 1, 1995, and April 24, 2002, and the IRI of the right wheel path decreased by 15 inches/mi (from 68 to 53 inches/mi) during this 7-year period. The LTPP database indicated this section was diamond ground on July 12, 2000. The reduction in IRI is attributed to the diamond grinding. The first and last FWD dates for this section were August 2, 1995, and July 9, 2002, and the average SN of this section decreased by 1.02 (from 8.32 to 7.30) during this 7-year period. IRI and SN plots were not developed for this section because it was subjected to diamond grinding.

3.10.4 LTPP Section 310118 (Nebraska)

The first and last profile dates for section 310118 (Nebraska) were November 1, 1995, and April 24, 2002, and the IRI of the right wheel path decreased by 20 inches/mi (from 74 to 54 inches/mi) during this 7-year period. This section is in the same SPS-1 project as section 310117 (Nebraska). The LTPP database indicated the section was diamond ground on July 12, 2000. The reduction in IRI is attributed to the diamond grinding. The first and last FWD dates for this section were August 3, 1995, and July 10, 2002, and the average SN of this section decreased by 1.59 (from 9.21 to 7.62) during this 7-year period. IRI and SN plots were not developed for this section because it was subjected to diamond grinding.

3.11 ANALYSIS OF GROUP 3 SECTIONS

Table 15 shows the following parameters for the test sections selected for analysis: LTPP section number, date corresponding to first IRI (first IRI date), date corresponding to last IRI (last IRI date), FWD test date close to the first IRI date (first FWD date), FWD test date close to the last IRI date (last FWD date), and the time difference between first and last dates for IRI and FWD data collection.

Table 15. Group 3 sections selected for analysis.

LTPP Section	IRI Date		FWD Date		Time Difference Between First and Last Date (years)
	First Date	Last Date	First Date	Last Date	IRI Data	FWD Data
190101, Iowa	10/15/1993	9/22/2004	5/19/1993	6/28/2005	10.9	12.1
190103, Iowa	10/15/1993	9/22/2004	5/19/1993	6/29/2005	10.9	12.1
050114, Arkansas	7/7/1995	4/6/2005	3/16/1994	5/12/2005	9.8	11.2
050116, Arkansas	7/7/1995	4/6/2005	3/16/1994	5/11/2005	9.8	11.2

Table 16 shows the IRI values for the 500-ft-long section for the first and last profile dates for the selected profile runs as well as the average SN for the section for the first and last FWD date. Table 17 shows the pavement layer thickness for group 3 sections.

Table 16. IRI and SN for group 3 sections.

LTPP Section	IRI (inches/mi)		SN		Change Between First and Last*
	First	Last	First	Last	IRI (inches/mi)	SN
190101, Iowa	83	159	7.48	7.62	76	-0.14
190103, Iowa	46	92	9.39	8.68	46	0.71
050114, Arkansas	49	72	4.26	5.13	23	-0.87
050116, Arkansas	68	63	7.81	10.30	-5	-2.49

*A negative change in SN indicates SN at the last date was higher than SN at the first date. A negative change in IRI indicates IRI at the last date was lower than IRI at the first date.

Table 17. Pavement layer thickness, group 3 sections.

LTPP Section	Layer Type	Layer Thickness (inches)	Material Type
190101, Iowa	AC	7.7	Hot-mixed, dense-graded
	GB	8	Crushed stone
	GS	25	Fine-grained soils, clay with gravel
190103, Iowa	AC	3.8	Hot-mixed, dense-graded
	ATB	8.4	HMAC
	GS	24	Fine-grained soils, clay with gravel
050114, Arkansas	AC	6.9	Hot-mixed, dense-graded
	GB	11	Gravel (uncrushed)
050116, Arkansas	AC	4.1	Hot-mixed, dense-graded
	ATB	11.8	HMAC

Sections 190101 and 190103 are both test sections in the Iowa SPS-1 project. Section 190101 has DGAB, and section 190103 has ATB. Section 050114 and 050116 are test sections in the Arkansas SPS-1 project. Section 050114 has DGAB, and section 050116 has ATB. In each project, the test section with the DGAB base showed a higher increase in IRI between the first and the last dates when compared to the test section with ATB.

3.11.1 LTPP Section 190101 (Iowa)

The first and last profile dates for section 190101 (Iowa) were October 15, 1993, and September 22, 2004, and the IRI of the right wheel path increased by 76 inches/mi (from 83 to 159 inches/mi) during this 11-year period. The first and last FWD dates for this section were May 19, 1993, and June 28, 2005, and the average SN of this section increased by 0.14 (from 7.48 to 7.62) during this 12-year period.

Figure 47 shows the first and last IRI and SN values, and figure 48 shows the normalized IRI and SN plots. SN at the last FWD date was greater than SN at the first FWD date at all test locations, except for the last three (400, 450, and 500 ft). There are five peaks in the IRI plot for the last IRI date, and, except for two, the peak IRI locations fall outside locations where FWD testing was conducted. SN values are available between 380 to 410 ft and 440 to 460 ft, where two of the peaks occur. The reductions in SN at 400 and 450 ft were 0.31 and 1.39, respectively. Therefore, a reduction in SN was seen at the locations that had a high increase in IRI. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus (see appendix A).

Figure 47. Graph. SN and IRI for two test dates, section 190101 (Iowa).

Figure 48. Graph. Normalized SN and IRI for two test dates, section 190101 (Iowa).

3.11.2 LTPP Section 190103 (Iowa)

The first and last profile dates for section 190103 (Iowa) were October 15, 1993, and September 22, 2004, and the IRI of the right wheel path increased by 46 inches/mi (from 46 to 92 inches/mi) during this 11-year period. The first and last FWD dates for this section were May 19, 1993, and June 29, 2005, and the average SN of this section decreased by 0.71 (from 9.39 to 8.68) during this 12-year period.

Figure 49 shows the first and last IRI and SN values, and figure 50 shows the normalized IRI and SN plots. SN at the last FWD date was lower than SN at the first FWD date at all test locations. An increase in IRI is noted throughout the section, with a major increase in IRI close to 250 ft. An FWD test was performed at 250 ft, and the decrease in SN observed at this location was 0.32, which was less than the average decrease in SN at this section. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus (see appendix A).

Figure 49. Graph. SN and IRI for two test dates, section 190103 (Iowa).

Figure 50. Graph. Normalized SN and IRI for two test dates, section 190103 (Iowa).

3.11.3 LTPP Section 050114 (Arkansas)

The first and last profile dates for section 050114 (Arkansas) were July 7, 1995, and April 6, 2005, and the IRI of the right wheel path increased by 23 inches/mi (from 49 to 72 inches/mi) during this 10-year period. The first and last FWD dates for this section were March 16, 1994, and May 12, 2005, and the average SN of this section increased by 0.87 (from 4.26 to 5.13) during this 11-year period.

Figure 51 shows the first and last IRI and SN values, and figure 52 shows the normalized IRI and SN plots. SN at the last FWD date was greater than SN at the first FWD date at all test locations. The increase in SN at the FWD test locations ranged from 0.55 to 1.25 and averaged 0.87. An increase in IRI between the test dates is noted for the entire section, but there were no major localized roughness increases in this section. The greatest increase in IRI within the section occurred between 350 to 400 ft, which borders the FWD tests that were performed at 350 and 400 ft. However, a FWD test point is not located within these limits. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus (see appendix A).

Figure 51. Graph. SN and IRI for two test dates, section 050114 (Arkansas).

Figure 52. Graph. Normalized SN and IRI for two test dates, section 050114 (Arkansas).

3.11.4 LTPP Section 050116 (Arkansas)

The first and last profile dates for section 050116 (Arkansas) were July 7, 1995, and April 6, 2005, and the IRI of the right wheel path decreased by 5 inches/mi (from 68 to 63 inches/mi) during this 10-year period. The first and last FWD dates for this section were March 16, 1994, and May 11, 2005, and the average SN of this section increased by 2.49 (from 7.81 to 10.30) during this 11-year period.

Figure 53 shows the first and last IRI and SN values, and figure 54 shows the normalized IRI and SN plots. SN at the last FWD date was greater than SN at the first FWD date at all test locations. The increase in SN ranged from 0.92 to 3.13 and averaged 2.49. The overall IRI of this section decreased from 68 to 63 inches/mi. The continuous IRI plots show little change in IRI, with the IRI for the last date showing a slight increase at some locations and a slight decrease at other locations. The highest IRI for the first and last IRI dates occurred around 270 ft. No relationship was noted between the increase in IRI and the deflection below the load or the subgrade modulus (see appendix A).

Figure 53. Graph. SN and IRI for two test dates, section 050116 (Arkansas).

Figure 54. Graph. Normalized SN and IRI for two test dates, section 050116 (Arkansas).

3.12 ANALYSIS OF GROUP 4 SECTIONS

The test sections in group 4 are SPS-5 sections. FWD testing and roughness data collection were performed at these test sections before and immediately after rehabilitation. The date corresponding to the data collection performed immediately after rehabilitation is referred to as the first IRI date or first FWD date. Table 18 shows the following information for the test sections selected for analysis: LTPP section number, date when IRI data were collected before rehabilitation, date corresponding to first IRI (first IRI date), date corresponding to last IRI (last IRI date), date when FWD data were collected before rehabilitation, FWD test date close to the first IRI date (first FWD date), FWD test date close to the last IRI date (last FWD date), and the time difference between first and last dates for IRI and FWD data collection.

Table 18. Group 4 sections selected for analysis.

LTPP Section	IRI Date			FWD Date			Time Difference Between First and Last Date (years)
	Before Overlay	First Date	Last Date	Before Overlay	First Date	Last Date	IRI Data	FWD Data
040502, Arizona	2/5/1990	9/21/1990	3/24/2006	1/18/1990	10/3/1991	9/15/2008	16.1	18.7
240505, Maryland	1/24/1992	6/11/1992	6/15/2006	2/20/1992	8/25/1992	4/7/2009	14.4	17.1
270509, Minnesota	5/24/1990	7/13/1993	6/6/2005	7/22/1990	11/6/1990	6/7/2005	15.0	14.9

Table 19 shows the IRI values for the 500-ft-long sections for the selected profile runs and the average SN for the section for the three test dates shown in table 18. Table 20 shows the average pavement layer thickness of the three test sections before rehabilitation obtained from the data in the LTPP database. Table 21 shows the milling depths and overlay thickness for the three sections.

Table 19. IRI and SN for group 4 sections.

LTPP Section	IRI (inches/mi)			SN			Change Between First and Last Date
	Before Overlay	First Date	Last Date	Before Overlay	First Date	Last Date	IRI (inches/mi)	SN
040502, Arizona	139	60	244	3.39	5.91	4.78	184	1.13
240505, Maryland	148	69	228	6.15	8.13	7.40	159	0.73
270509, Minnesota	197	54	117	4.52	5.94	5.78	63	0.16

Table 20. Pavement layer thickness before rehabilitation, group 4 sections.

LTPP Section	Layer Type	Layer Thickness (inches)	Material Type
040502, Arizona	AC	0.9	Hot-mixed, open-graded
	AC	4.2	Hot-mixed, dense-graded
	GB	14.7	Soil-aggregate mixture (predominantly coarse-grained)
240505, Maryland	AC	1.1	Hot-mixed, open-graded
	AC	3.5	Hot-mixed, dense-graded
	TB	3.7	Cement aggregate mixture
	GS	5.9	Crushed stone
	TS	8.9	Cement-treated soil
270509, Minnesota	AC	6.9	Hot-mixed, dense-graded
	GB	4.7	Crushed gravel
	GS	12.6	Gravel (uncrushed)

Table 21. Mill and overlay thickness.

Section	Milled Thickness (inches)	Overlay Thickness (inches)
040502, Arizona	1.4	2.7
240505, Maryland	—	2.1
270509, Minnesota	1.9	3.6

— Indicates milling was not performed.

3.12.1 LTPP Section 040502 (Arizona)

IRI before rehabilitation of section 040503 (Arizona) was 139 inches/mi, and IRI immediately after the overlay was 60 inches/mi. Rehabilitation resulted in IRI decreasing by 79 inches/mi. The first and last profile dates for this section were September 21, 1990, and March 24, 2006, and IRI of the right wheel path increased by 184 inches/mi (from 60 to 244 inches/mi) during this 16-year period.

SN before rehabilitation of this section was 3.39, and it increased to 5.91 after rehabilitation. The first and last FWD dates for this section were October 3, 1991, and September 15, 2008, and the average SN decreased by 1.13 (from 5.91 to 4.78) during this 17 year period.

Figure 55 shows the continuous IRI plots for before the overlay (February 5, 1990), immediately after the overlay (September 21, 1990), and at the last IRI date (March 24, 2006). The continuous IRI plot had higher values at the last date compared to the before overlay plot except at a few locations. Figure 56 shows the SN plots for before the overlay (January 8, 1990), immediately after the overlay (October 3, 1991), and at the last SN date (September 15, 2008). The rehabilitation resulted in an increase in SN at all test locations. SN at all test locations for the last date was lower than SN immediately after rehabilitation (first date) at all test locations.

Figure 55. Graph. IRI for before overlay, after overlay, and last test date, section 040502 (Arizona).

Figure 56. Graph. SN for before overlay, after overlay, and last test date, section 040502 (Arizona).

Figure 57 shows the IRI and SN values after the overlay and at the last test date, and figure 58 shows the normalized IRI and SN plots. IRI increased throughout the section, with the highest increases between 80 and 110 ft and 420 and 460 ft. SN at 100 ft had a decrease of 1.06, which was lower than the average decrease in SN at the section. SN at 450 ft had the lowest decrease for this section, which was 0.31.

Figure 57. Graph. SN and IRI for two test dates after overlay, section 040502 (Arizona).

Figure 58. Graph. Normalized SN and IRI for two test dates after overlay, section 040502 (Arizona).

3.12.2 LTPP Section 240505 (Maryland)

IRI before rehabilitation of this section was 148 inches/mi, and IRI immediately after the overlay was 69 inches/mi. The rehabilitation resulted in IRI decreasing by 79 inches/mi. The first (after rehabilitation) and last profile dates were June 11, 1992, and June 15, 2006, and IRI of the right wheel path increased by 159 inches/mi (from 69 to 228 inches/mi) during this 14-year period.

SN before rehabilitation of this section was 6.15, and it increased to 8.13 after rehabilitation. The first (after rehabilitation) and last FWD dates for this section were August 25, 1992, and April 7, 2009, and the average SN decreased by 0.73 (from 8.13 to 7.40) during this 17-year period.

Figure 59 shows the IRI plots for before the overlay (January 24, 1992), immediately after the overlay (June 1, 1992), and at the last IRI date (June 15, 2006). Six distinct peaks are shown in the IRI plot for last IRI date, and, except for one, these peaks generally correspond to the peaks in the before-overlay IRI plot. Figure 60 shows the SN plots for before the overlay (February 20, 1992), immediately after the overlay (August 25, 1992), and at the last FWD date (April 27, 2009). SN at all test locations increased after rehabilitation. SN at the last FWD date was lower than SN immediately after rehabilitation (first FWD date) for all test locations.

Figure 59. Graph. IRI for before overlay, after overlay, and last test date, section 240505 (Maryland).

Figure 60. Graph. SN for before overlay, after overlay, and last test date, section 240505 (Maryland).

Figure 61 shows the IRI and SN values for the first (immediately after rehabilitation) and last test dates, and figure 62 shows the normalized IRI and SN plots. High increases in IRI were noted between 120 and 150 ft, 180 and 210 ft, 230 and 250 ft, 270 and 310 ft, and 330 and 360 ft. FWD test locations were present at 200, 300, and 350 ft. The decreases in SN at 200, 300, and 350 ft were 0.19, 0.52, and 0.91 respectively, with the decrease at 200 ft greater than the average decrease in SN for the section (0.73).

Figure 61. Graph. SN and IRI for two test dates after overlay, section 240505 (Maryland).

Figure 62. Graph. Normalized SN and IRI for two test dates after overlay, section 240505 (Maryland).

3.12.3 LTPP Section 270509 (Minnesota)

IRI before rehabilitation of section 270509 (Minnesota) was 197 inches/mi, and IRI immediately after the overlay was 54 inches/mi. Rehabilitation resulted in IRI decreasing by 143 inches/mi. The first and last profile dates for this section were July 13, 1993, and June 6, 2005, and IRI of the right wheel path increased by 63 inches/mi (from 54 to 117 inches/mi) during this 12-year period.

SN before rehabilitation of this section was 4.52, and it increased to 5.94 after rehabilitation. The first (after rehabilitation) and last FWD dates for this section were November 6, 1990, and June 27, 2005, and the average SN decreased by 0.16 (from 5.94 to 5.78) during this 15-year period. Figure 63 shows the IRI plots for before the overlay (May 24, 1990), immediately after the overlay (July 13, 1993), and at the last IRI date (June 6, 2005). A significant reduction in IRI occurred due to the overlay. Figure 64 shows the SN plots for before the overlay (July 22, 1990), immediately after the overlay (November 6, 1990), and at the last SN date (June 7, 2005). The overlay resulted in an increase in SN at all test locations.

Figure 63. Graph. IRI for before overlay, after overlay, and last test date, section 270509 (Minnesota).

Figure 64. Graph. SN for before overlay, after overlay, and last test date, section 270509 (Minnesota).

Figure 65 shows the IRI and SN at the first test date (after overlay) and the last test date, and figure 66 shows the normalized IRI and SN plots. An increase in IRI is noted throughout the section, but no localized areas showed a very high increase in IRI. Some test locations showed an increase in SN, while others showed a decrease, ranging from 0.06 to 0.57.

Figure 65. Graph. SN and IRI for two test dates after overlay, section 270509 (Minnesota).

Figure 66. Graph. Normalized SN and IRI for two test dates after overlay, section 270509 (Minnesota).

3.13 ANALYSIS OF GROUP 5 SECTIONS

The test sections in group 5 are SPS-2 test sections. Table 22 shows the following information for the test sections selected for analysis: LTPP section number, date corresponding to first IRI (first IRI date), date corresponding to last IRI (last IRI date), FWD test date close to the first IRI date (first FWD date), FWD test date close to the last IRI date (last FWD date), and the time difference between first and last dates for IRI and FWD data collection.

Table 22. Group 5 sections selected for analysis.

LTPP Section	IRI Date		FWD Date		Time Difference Between First and Last Date (years)
	First Date	Last Date	First Date	Last Date	IRI Data	FWD Data
040213, Arizona	1/25/1994	8/11/2006	2/8/1994	12/15/2004	12.6	10.9
050217, Arkansas	2/6/1997	4/2/2005	11/14/1996	9/20/2004	8.2	7.9
390205, Ohio	8/14/1996	8/8/2006	12/30/1996	9/9/2004	10.0	7.7

Table 23 shows the IRI values for the 500-ft-long sections and the average effective slab thicknesses for the test dates shown in table 22. Table 24 shows the average pavement layer thickness for the test sections determined from the data in the LTPP database.

Table 23. IRI and effective slab thickness for group 5 sections.

LTPP Section	IRI (inches/mi)		Effective Slab Thickness (inches)		Change Between First and Last Date*
	First Date	Last Date	First Date	Last Date	IRI (inches/mi)	Effective Slab Thickness (inches)
040213, Arizona	95	165	9.45	9.00	70	0.45
050217, Arkansas	86	160	13.80	16.80	74	-3.00
390205, Ohio	89	161	9.81	10.20	72	-0.39

*A negative effective slab thickness indicates the effective slab thickness at the last date was higher than that for the first date.

Table 24. Pavement layer thickness for group 5 sections.

LTPP Section	Layer Type	Layer Thickness (inches)
040213, Arizona	PCC	7.9
	Aggregate base	5.8
050217, Arkansas	PCC	8.3
	Lean concrete base (LCB)	6.3
390205, Ohio	PCC	8.0
	LCB	6.2

Plots of deflection measured below the load and at a distance of 60 inches from the center of the load plate for the first and last FWD date for all sections are included in appendix B. For the September 20, 2004, FWD test for test section 050217, non-decreasing deflections were encountered at several test locations. The ILLI-BACK program computed high and unreasonable effective slab thickness at these locations. Some of the extremely high effective slab thicknesses were not considered in computing the average, yet the average was higher than the initial effective slab thickness. Sections 050217 and 390205 both had a LCB.

The ILLI-BACK program appears to have combined the PCC and LCB thicknesses when determining the effective slab thickness for these two sections. The computed effective slab thickness for section 040213 was also higher than the actual slab thickness.

3.13.1 LTPP Section 040213 (Arizona)

The first and last profile dates for section 040213 (Arizona) were January 25, 1994, and August 11, 2006, and the IRI of the right wheel path increased by 70 inches/mi (from 95 to 165 inches/mi) during this 13-year period. The first and last FWD test dates for this section were February 8, 1994, and December 15, 2004, and the average effective slab thickness of this section showed a decrease of 0.45 inches (from 9.45 to 9.00 inches) during this period.

Figure 67 shows the first and last IRI and effective slab thickness, and figure 68 shows the normalized IRI and D_eff plots. An increase in IRI is seen across the entire section. The effective slab thickness at the last FWD date shows an increase for the first three data points and a decrease for the rest of the points. There are three locations within the section where large changes in IRI occurred, but these locations fall between FWD test locations. No relationship between change in IRI and deflections (below the load and 60 inches from the load) was observed.

Figure 67. Graph. D_eff and IRI for two test dates, section 040213 (Arizona).

Figure 68. Graph. Normalized D_eff and IRI for two test dates, section 040213 (Arizona).

3.13.2 LTPP Section 050217 (Arkansas)

The first and last profile dates for section 050217 (Arkansas) were February 16, 1997, and April 2, 2005, and the IRI of the right wheel path increased by 74 inches/mi (from 86 to 160 inches/mi) during this 8-year period. The first and last FWD dates for this section were November 14, 1996, and September 20, 2004, and the average effective slab thickness of this section increased by 3 inches (from 13.80 to 16.80 inches). As previously described, non-decreasing deflections were noted at many test locations for the September 20, 2004, test, and these deflections caused high effective slab thickness values to be computed at such locations. This resulted in a much higher effective slab thickness for the last test date than for the first test date.

Figure 69 shows the first and last IRI and effective slab thickness, and figure 70 shows the normalized IRI and D_eff plots. Some of the high D_eff values computed for the second test date are not shown in these plots, and no FWD data were available after 382 ft for the first FWD date. Most of the increase in IRI at this section occurred between 130 and 330 ft. The effective slab thickness at the last FWD date was higher than that for the first date for all test locations. No relationship between the increase in IRI and the effective slab thickness can be seen in the plot. No relationship between change in IRI and deflections (below the load and at 60 inches from the load) could be observed.

Figure 69. Graph. D_eff and IRI for two test dates, section 050217 (Arkansas).

Figure 70. Graph. Normalized D_eff and IRI for two test dates, section 050217 (Arkansas).

3.13.3 LTPP Section 390205 (Ohio)

The first and last profile dates for section 390205 (Ohio) were August 14, 1996, and August 8, 2006, and the IRI of the right wheel path increased by 72 inches/mi (from 89 to 161 inches/mi) during this 10-year period. The first and last FWD dates for this section were December 30, 1996, and September 9, 2004, and the average effective slab thickness of this section showed an increase of 0.39 inches (from 9.81 to 10.20 inches) during this 8-year period.

Figure 71 shows the first and last IRI and effective slab thickness, and figure 72 shows the normalized IRI and D_eff plots. Most of the increase in IRI at this section occurred between the start of the section and about 200 ft. Within the first 200 ft, the effective slab thickness at the last FWD date was lower than that for the first date at three locations and the same at one location. Among the rest of the FWD test locations within this section, half showed a decrease in effective slab thickness, and the other half showed an increase between the first and last FWD dates. No relationship between change in IRI and deflections (below the load and at 60 inches from the load) could be observed.

Figure 71. Graph. D_eff and IRI for two test dates, section 390205 (Ohio).

Figure 72. Graph. Normalized D_eff and IRI for two test dates, section 390205 (Ohio).

3.14 SUMMARY

Table 25 shows all flexible pavement sections used in the analysis. As previously described, the IRI and FWD data collections were not performed at the same time at LTPP sections. The first IRI and FWD dates correspond to the first time IRI and FWD data were collected at a test section, and the last dates were selected such that the last IRI and FWD dates were close to each other. Overall, the time period used to evaluate IRI and SN changes at a section were reasonably close to each other.

Table 25. Flexible pavement sections used in study.

Group	LTPP Section	IRI Date		FWD Date		Time Difference Between First and Last Date (years)
		First Date	Last Date	First Date	Last Date	IRI Data	FWD Data
1	050119	7/6/1995	3/19/2004	3/15/1994	5/24/2004	8.7	10.2
	480114	9/8/1997	3/19/2007	11/17/1997	6/24/2004	9.5	6.6
	310113	11/1/1995	3/20/2000	8/3/1995	10/14/1999	4.4	4.2
	010102	8/25/1994	5/4/2005	6/21/1995	4/29/2005	10.7	9.9
	390112	8/14/1996	5/5/2005	11/6/1996	9/1/2004	8.7	7.8
	040123	1/27/1994	3/27/2006	2/16/1994	4/7/2005	12.2	11.1
	190108	10/15/1993	9/22/2004	5/18/1993	6/27/2005	10.9	12.1
2	320101	12/3/1996	8/7/2006	3/27/1996	8/28/2006	9.7	10.4
	390106	8/14/1996	8/9/2006	11/5/1996	7/15/2008	10.0	11.7
	310117	11/1/1995	4/24/2002	8/2/1995	7/9/2002	6.5	6.9
	310118	11/1/1995	4/24/2002	8/3/1995	7/10/2002	6.5	6.9
3	190101	10/15/1993	9/22/2004	5/19/1993	6/28/2005	10.9	12.1
	190103	10/15/1993	9/22/2004	5/19/1993	6/29/2005	10.9	12.1
	050114	7/7/1995	4/6/2005	3/16/1994	5/12/2005	9.8	11.2
	050114	7/7/1995	4/6/2005	3/16/1994	5/11/2005	9.8	11.2
4	040502	9/21/1990	3/24/2006	10/3/1991	9/15/2008	15.5	17.0
	240505	6/11/1992	6/15/2006	8/25/1992	4/7/2009	14.0	16.6
	270509	7/13/1993	6/6/2005	11/6/1990	6/7/2005	11.9	14.6

Table 26 shows the changes in IRI and SN that had occurred at the sections between the first and last dates expressed as magnitude of the change as well as the percent change. Only three sections showed a decrease in IRI between the first and last dates (310117 (Nebraska), 310118 (Nebraska), and 050116 (Arkansas)). The decrease in IRI at sections 310117 and 310118 in Nebraska occurred because of a treatment that was applied on the pavement. Of the 18 sections evaluated in this study, 12 sections showed a decrease in SN between the first and last dates, and 6 sections showed an increase in SN.

Table 26. Change in IRI and SN at evaluated sections.

Group	LTPP Section	IRI (inches/mi)		SN		Change Between First and Last Date*		Increase in IRI (percent)**	Decrease in SN (percent)**
		First Date	Last Date	First Date	Last Date	IRI (inches/mi)	SN
1	050119	48	98	4.51	5.40	50	-0.89	105	-20
	480114	46	65	10.90	10.39	19	0.51	42	5
	310113	93	126	5.73	5.82	33	-0.09	35	-2
	010102	62	204	2.86	2.19	143	0.67	232	23
	390112	57	101	7.83	7.39	44	0.44	77	6
	040123	52	138	7.94	7.26	86	0.68	163	9
	190108	49	141	8.30	9.94	93	-1.64	191	-20
2	320101	57	57	11.91	11.23	0	0.68	0	6
	390106	71	114	6.94	5.04	43	1.90	61	27
	310117	68	53	8.32	7.30	-15	1.02	-22	12
	310118	74	54	9.21	7.62	-20	1.59	-27	17
3	190101	83	159	7.48	7.62	76	-0.14	92	-2
	190103	46	92	9.39	8.68	46	0.71	100	8
	050114	49	72	4.26	5.13	23	-0.87	47	-20
	050116	68	63	7.81	10.30	-5	-2.49	-7	-32
4	040502	60	244	5.91	4.78	184	1.13	307	19
	240505	69	228	8.13	7.40	159	0.73	230	9
	270509	54	117	5.94	5.78	63	0.16	117	3

*A positive change in IRI indicates that IRI of the section increased. A positive change in SN indicates that SN of the section decreased.
**Percent changes between first and last dates.

Some sections showed an increase in SN between the first and last dates. Some of the factors that may have contributed to the increase include hardening of the AC layer with age resulting in an increase in the structural layer coefficient for the asphalt layer, changes in moisture conditions in the GB/SB layers that resulted in an increase in the structural layer coefficient for these layers, or the effect of the temperature adjustment factor. The deflection below the load was adjusted to a standard temperature of 68 °F before computing the SN values using the temperature adjustment factors shown in the 1993 AASHTO Guide for Design of Pavement Structures.⁽²³⁾ However, each AC mix is expected to have unique adjustment factors, and the difference in the adjustment factors presented in the AASHTO Guide for Design of Pavement Structures with the actual behavior of the mix may result in errors when the deflections are adjusted to 68 °F.⁽²³⁾

Figure 73 shows the relationship between changes in IRI and SN observed at the sections, and figure 74 shows this relationship for percent changes. Data for sections 310117 and 310118 in Nebraska, which had a treatment that caused a decrease in SN, are not shown in these plots.

Figure 73. Graph. Relationship between changes in IRI and SN.

Figure 74. Graph. Relationship between percent changes in IRI and SN.

As shown in these figures, there is no relationship between change or percent change in IRI and the change or percent change in SN. Because only three rigid pavements were evaluated, sufficient data were not available to conduct a similar comparison as presented for flexible pavements.

At all of the analyzed flexible pavement sections, continuous IRI plots for first and last dates were compared with SN values estimated from FWD measurements at the first and last dates. If there was a relationship between IRI and SN, it was hypothesized that areas where large changes in IRI had occurred should show large decreases in SN. However, such a relationship was not seen in the analyzed sections. Generally, no major changes in SN associated with localized areas were seen at the evaluated sections. However, in most sections, major IRI changes were noticed at localized areas within the section. At these areas, the changes in SN were not significantly different from changes in SN that had occurred at areas where major changes in IRI had not occurred. Comparison of continuous IRI plots with deflection below the FWD load and estimated subgrade modulus plots also did not show any relationship between changes in IRI and deflection or subgrade modulus or between changes in IRI and changes in deflection or subgrade modulus.

For the rigid pavements, the effective slab thickness instead of SN was used as the structural strength parameter in this study. Similar to flexible pavements, no relationship was seen between changes in IRI and changes in effective slab thickness.