National Academies Press: OpenBook

Enhancement of the Practice for Certification of Inertial Profiling Systems (2023)

Chapter: Chapter 2 - Review of Existing Practices

« Previous: Chapter 1 - Introduction
Page 10
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 - Review of Existing Practices." National Academies of Sciences, Engineering, and Medicine. 2023. Enhancement of the Practice for Certification of Inertial Profiling Systems. Washington, DC: The National Academies Press. doi: 10.17226/27182.
×
Page 23

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

10 Review of Existing Practices This chapter discusses the various factors relevant to inertial profiler certification, including the factors associated with the AASHTO R 56-14 and other factors not considered in that stan- dard practice. The first section reviews the certification process, the second section reviews factors associated with the measurement system (equipment and operator), and the third section covers the certification environment, including test sections, reference devices, and analysis approaches used in the certification process. Certification Process Before considering specific aspects of certification of inertial profilers, it may be helpful to review the philosophy behind profiler certification. Typically, certification is to follow a traceable system of review (UNIDO 2019). In other words, the approach for certification of equipment is to be conducted by reviewing the output against a known value that has itself been identified as “known” (UNIDO 2019). Within the United States, the Office of Weights and Measures (OWM) was developed to provide confidence in commerce (Judson 1976). The OWM identifies that “exactness in units must have its basis in standards that are as permanent and exact as possible” (Judson 1976). This certification approach means that the IRI is accurate and repeatable and assures reproducibility between devices and operators. These parameters of certification are depicted in Figure 5. On this basis, consider the scope of the current AASHTO R 56-14: This practice describes a certification procedure for test equipment used to measure a longitudinal surface elevation profile of highways based on an inertial reference system that is mounted on a host vehicle. The minimum requirements stipulated herein are intended to focus on the need for accurate and repeatable profile measurements for construction quality control/quality assurance, acceptance and network-level data collection. Another way to interpret this scope is that AASHTO R 56-14 is intended to evaluate equip- ment used for routine data collection at either the project level (e.g., construction acceptance) or network level. The scope does not identify that it is for use in identifying that a device may meet the standards for a reference device. The scope also does not identify that it is appropriate for the research and development of new equipment. Based on NCHRP Synthesis 526: Inertial Profiler Certification for Evaluation of International Roughness Index¸ 38 state DOTs indicated that they use an IRI-based smoothness specification for construction acceptance, and 27 of these state DOTs indicated that a certified device must be used for the collection of the data used for construction acceptance (Perera 2018). The proce- dures used by these state DOTs are summarized in Table 1. In the table, “Out-of-State” indicates C H A P T E R   2

Review of Existing Practices 11   that the state DOT uses the certification facility and procedure in use by another state DOT. The state DOT-developed procedures were further classified by those incorporating cross-correlation analysis to assess whether the device being certified collects data at the required levels of precision and accuracy. For network-level data collection, 15 state DOTs require certification of the profiling device used. Table 2 identifies the various procedures these state DOTs use, including a distinction between which state DOT-developed procedures incorporate cross-correlation. Based on the synthesis results, it is clear that AASHTO R 56-14 is not in common use across the nation at the time of the survey conducted for the synthesis. Location Procedure Adopted Number of Responding State DOTs In-State AASHTO R 56-14 6 ASTM E950 (2008) 1 State DOT-developed procedure using cross-correlation 14 State DOT-developed procedure not using cross-correlation 4 Out-of-State State DOT-developed procedure 2 Procedure Adopted Number of Responding State DOTs AASHTO R 56-14 5 ASTM E950 1 State DOT-developed procedure using cross-correlation 5 State DOT-developed procedure not using cross-correlation 4 Figure 5. Parameters for equipment certification (Lundberg and Sjögren 2004). Table 1. Certification procedures used by state DOTs at project-level data collection (Perera 2018). Table 2. Certification procedures used by state DOTs at network-level data collection (Perera 2018).

12 Enhancement of the Practice for Certification of Inertial Profiling Systems System Factors Equipment Checks AASHTO R 56-14 requires that the profiler successfully pass a block check and a bounce test as outlined in AASHTO R 57, Standard Practice for Operating Inertial Profiling Systems. The block check is used to verify that the height sensors accurately identify relative changes in the height of the pavement surface using a series of blocks. The bounce test checks that the accelerometer and the laser height sensor work together to cancel the vehicle’s motion from the relative surface elevation measurement (Simpson et al. 2002). These tests are specific to an inertial profiler and may not be relevant for future technologies. However, their purpose is to ensure that these individual components are in good working con- dition before attempting the full-scale certification process. These tests are an essential evaluation of inertial profiling equipment. In comparison, these tests may need to be varied based on the device’s exact nature. Some versions of these two tests will likely be emulated for any other device. These tests also provide a means for evaluating operator proficiency. For example, the North Carolina DOT (NCDOT) identifies that the operator must demonstrate the ability to perform the equipment checks, including both the bounce test and block check, to be certified by the DOT (2018). Operator Certification AASHTO R 56-14 identifies a need for the operators to be certified. This certification includes both written and practical examinations. It outlines elements for inclusion in these examina- tions and operator certification but provides no specifics for these elements, including equip- ment operation, data collection, and data processing. The procedure outlines the recommended knowledge and frequency of this certification. While AASHTO R 56-14 identifies critical elements, it lacks detail on the requirements for operator certification. Equipment Requirements Equipment is required to meet the specifications of AASHTO M 328. AASHTO M 328-14 specifies the components of an inertial profiler along with their resolution. Making AASHTO M 328 a requirement for certification means that only inertial profilers may be certified per AASHTO R 56-14 and does not allow for consideration of non-inertial equipment that may meet the precision and bias requirements identified within the procedure. AASHTO R 56-14 requires that data collection be automatically triggered at the start of the section. At the end of each test section, a marker may be used to check the accuracy and repeat- ability of the distance measurement completed by the profiler. AASHTO R 56-14 recommends using a line laser or large footprint sensor for coarse-textured pavement surfaces (e.g., open-graded or porous friction course, chip seal). A note indicates the single spot lasers have been shown to provide inaccurate data on coarse pavement surface textures. Ideally, the larger footprint provides a bridging algorithm across the coarse texture, lessening the impact of the texture on the collected profile. Test Speed AASHTO R 56-14 requires that the equipment to be certified should collect a minimum of five runs at the operational speed of the profiler. For devices that may be used to collect data at

Review of Existing Practices 13   varying speeds, AASHTO R 56-14 recommends collecting five runs at each of two speeds, with these speeds representing the lower and upper ends of the speed range for that device. Speeds are assumed to be as constant as possible over the section tested. Therefore, there is no requirement to evaluate the equipment over changing speed. In other words, the certification process does not address what happens to the collection as the equipment slows down, speeds up, or comes to a stop, which is often necessary when the equipment is used in an urban environment. NCHRP Research Report 914: Measuring, Characterizing, and Reporting Pavement Roughness of Low-Speed and Urban Roads summarizes research conducted to address concerns with profiling and monitoring roughness on urban and low-speed roadways (Karamihas et al. 2019). Urban and low-speed profiling challenges include the potential for frequent braking and acceleration/ deceleration, stop-and-go data collection, and built-in features that affect pavement profiles, such as utility and drainage covers/structures, pavement crowns, sudden grade changes, and matching curb and gutter. Specifically, NCHRP Research Report 914 findings resulted in recommendations for mea- surement practices to ensure valid data are collected in urban and low-speed environments. Several revisions to AASHTO R 56-14 (as well as to AASHTO M 328-14 (2022)) are presented to improve profiler certification testing to “identify the lowest valid operating speed of a profiler and discern the range of distance within a measured profile that should be marked as invalid in the vicinity of braking or stop-and-go operation for computation of IRI” (Karamihas et al. 2019). While manufacturers should ideally design profilers to collect and process data under these con- ditions, these additional certification practices will help verify this capability for a given profiler, regardless of design. The proposed revisions to AASHTO R 56-14 (and AASHTO M 328-14) are not a means to extract valid data from a longitudinal profile collected under these conditions, but rather a means to verify the limits under which a profiler can collect data under urban and low-speed conditions and the limits of valid data within a given profile trace when collected under urban and low-speed conditions (Karamihas et al. 2019). NCHRP Research Report 914 noted that these recommendations were intended to place requirements on profiler system performance rather than specific profiler components (e.g., accelerometer and height sensor) to help improve profiler designs themselves (Karamihas et al. 2019). The primary revisions recommended for AASHTO R 56-14 in NCHRP Research Report 914 were the addition of four types of “dynamic certification testing,” specifically for inertial profilers to be used in urban and low-speed environments to identify: • Minimum valid operating speed: This procedure requires passes at a minimum speed for correct operation proposed by the operator for certification. The profiler qualifies as valid for operation at the proposed speed if the measured profiles meet the required thresholds for repeatability and accuracy. • Maximum valid operating deceleration: This procedure requires braking at a deceleration level proposed by the operator for certification. The profiler qualifies as valid for deceleration up to the testing level if the measured profiles with braking reproduce profiles measured at a constant speed with the same level of repeatability established for regular operation. • Invalid range near deceleration: This procedure requires passes with braking at an average deceleration of 0.26 g. The range to mark as invalid is based on the bias in the short-interval roughness profile compared to a pass at a constant speed. • Invalid range near stops: This procedure requires passes with a stop. The range to mark as invalid is based on the bias in the short-interval roughness profile compared to a pass at a constant speed (Karamihas et al. 2019). Specific proposed revisions to the various sections of AASHTO R 56-14 are provided in an appendix to NCHRP Research Report 914. In addition to these dynamic certification tests,

14 Enhancement of the Practice for Certification of Inertial Profiling Systems revisions to requirements for the certification test sections are also provided. Specifically, a mini- mum of 1,000 ft test section length will be necessary for performing these additional checks (Karamihas et al. 2019). Some additional notes related to the proposed revisions to AASHTO R 56-14 documented in NCHRP Research Report 914 (Karamihas et al. 2019) include: • The additional dynamic testing requirements for AASHTO R 56-14 treat stop-and-go runs as a distinct type of adverse condition rather than a combination of deceleration and operation below the valid profiling speed because the magnitude and extent of the length of erroneous IRI values were different for the two conditions. • The proposed dynamic testing in AASHTO R 56-14 requires profiler operators to select a minimum test speed and maximum deceleration level for testing. It is recommended that pro- filer operators attempt the recommended testing before the official certification. It is anticipated that profiler manufacturers will know the limitations of their equipment, particularly after experiencing the recommended testing. • Until the specific limitations of a given inertial profiler are established, default values for minimum test speed, maximum deceleration, and the range of invalid profile near stops and severe braking may have to be set using conservative estimates. Conservative observations from the testing performed during the research suggest a minimum valid test speed of 25 mph (40 km/hr) and maximum deceleration of 0.16 g. Based on testing from the research, the fol- lowing default settings for removing areas from the calculation of the IRI are recommended: (1) within 155 ft (47.3 m) upstream of the location of a stop, (2) within 255 ft (75.6 m) down- stream of the location of a stop, (3) within the area where deceleration of 0.16 g and above is detected, and (4) up to 51 ft (15.5 m) downstream of the location where deceleration passed below 0.16 g. Less restrictive settings may be justified for a particular profiler design using the recommended dynamic testing in AASHTO R 56-14. Profiler performance under these dynamic certification tests is highly dependent on the pro- filer host vehicle (Karamihas et al. 2019). Mounting a profiler to a host vehicle that resists changes in pitch and roll orientation is encouraged. Although host vehicle properties were not explicitly examined in this research, there are vehicle characteristics that can maintain a consistent orientation, including a long wheelbase, a wide track, a low center of gravity, stiff suspensions, suspension anti-dive, and a high-suspension roll center (Karamihas et al. 2019). Environment Factors AASHTO R 56-14 does not identify any specific weather requirements for data collection. However, it is widely understood that rain and other precipitation can impact collected data (Simpson et  al. 2002). Research has indicated that ambient temperature changes have little impact on a sensor’s ability to collect profile data (Karamihas et al. 1999). However, the tempera- ture may be expected to impact the pavement profile, particularly for jointed concrete pavements (Karamihas and Senn 2012). AASHTO R 56-14 attempts to account for temperature impacts by requiring that reference profile data be collected immediately before measurements are made with the device being certified when testing jointed concrete pavements. Test Sections The CPAR report (Karamihas 2005) recommends six surface types to be included in the certification process: • Dense-graded asphalt with small aggregate (9.5-mm or less); • Fresh chip seal with positive texture;

Review of Existing Practices 15   • Stone matrix asphalt; • Transversely tined jointed concrete; • Longitudinally ground concrete; and • Longitudinally tined concrete. The selected sections need to cover multiple levels of roughness, including: • 30 to 65 in./mile; • 65 to 95 in./mile; • 95 to 160 in./mile; and • 160 to 250 in./mile. Although not found in practice, this matrix results in a combination of up to 24 potential test sections to be used in certification (Karamihas 2005) to provide a complete matrix of testing equipment. The 24 sections represent a full factorial of the combination of surface types and levels of roughness. A limited approach would be to identify sections containing the four levels of roughness within the six surface types resulting in a minimum of six test sections. Additionally, if all six surface types identified are not used by the agency performing certification, the number of test sections may be limited to four, representing the four levels of roughness. Sections should be 528 ft in length except for the dense-graded asphalt, which should be at least 1,056 ft. The longer section allows for evaluating the longer wavelength content of the profile (Karamihas 2005). Other work has shown that the profile is influenced by various factors related to the pavement surface and geometry (Karamihas et al. 1999). These factors include a dependence on the trans- verse location within the lane, as minor variations in the profile across the lane may impact the results measured (Karamihas et al. 1999). Profiles may also be impacted by daily temperature variations, primarily on jointed concrete pavements, where the curling of the slabs may vary significantly throughout the day (Karamihas and Senn 2012). Other elements of the pavement that impact the collected profile include (Karamihas et al. 1999): • Distress; • Curves, which result in changes in lateral acceleration that may contaminate accelerometer readings; • Hills and grades that may also impact accelerometer readings, particularly with changes in grade; • Surface moisture resulting in standing water on the pavement surface; • Surface contaminants such as litter, construction debris, and leaves; and • Pavement markings with high retroreflectivity that may impact the measurements by the laser height sensors. This list of factors could result in many additional considerations for the certification procedure if all of them are included. However, not all of these factors may be safely evaluated, as perform- ing certification on a non-tangent section, for example, may result in risks to the traveling public, requiring traffic control when public travel ways are used for certification. Construction and maintenance of sites meeting all these factors where a private facility may be used would likely be costly. The length of the section is another geometric factor that must be considered in selecting a test section location. AASHTO R 56-14 requires a minimum length of 528 ft for the test sections. In addition to the 528 ft used for the analysis, sufficient lead-in distance is required by AASHTO R 56-14 before the section for equipment stabilization and sufficient lead-out distance for the safe stopping of the equipment. The lead-in distance should be roughly equivalent to the high-pass

16 Enhancement of the Practice for Certification of Inertial Profiling Systems filter wavelength used in the data collection (i.e., approximately 300 ft) (El-Korchi 2000). Based on the range of speed for data collection of up to 60 mph, the stopping distance could be as much as 200 ft (El-Korchi 2000). These distances suggest a minimum total length of a test section of 1,028 ft. AASHTO R 56-14 specifies that test sections must be relatively flat with no significant grade or change in grade. They must also be relatively straight with no significant horizontal curvature or superelevation. Section roughness levels are specified as smooth, with an IRI ranging from 30 to 75 in./mile, and medium-smooth, with an IRI ranging from 95 to 135 in./mile. If the equipment being certified will be used on pavements with more distress, then a third, medium- rough section is recommended with an IRI up to 200 in./mile. The geometric requirements can make it challenging to identify test sections with the necessary length within the specific smoothness ranges. On the other hand, a pavement with a 200 in./mile IRI may have sufficient elements within the surface that make it very difficult to meet the requirements for accuracy and repeatability set by AASHTO R 56-14. The minimum roughness identified within AASHTO R 56-14 is 30  in./mile. This level of roughness on an open-graded surface may be difficult to meet (Karamihas and Perera 2014). The cross-correlation representing an approximate 5 percent error in IRI suggests that the maximum difference observed in the IRI from the multiple runs is 1.5 in./mile. For comparison, the bounce test allows a difference in IRI of 5 in./mile, suggesting that 30 in./mile may be too smooth for equipment certification, or the cross-correlation may need to be increased based on some mini- mum level of IRI. Similarly, an IRI of 200 in./mile allows for a difference of 10 in./mile based on the cross-correlation level. For practical purposes of evaluating a project, an IRI of 190 in./mile may essentially mean the same thing as a 200 in./mile value. Still, in evaluating the accuracy and repeatability of the equipment, these differences may be too large. AASHTO R 56-14 requires that the sections incorporate surface types (and macrotexture) typically used by a state DOT. This requirement increases the number of sections that a state DOT must identify with concrete and asphalt pavements in their network. Some state DOTs, such as the Minnesota Department of Transportation (MnDOT), have elected not to include jointed concrete pavement (JCP) sections due to the diurnal slab curling and warping impact (Perera 2018). Some state DOTs also have requirements for open-graded or coarse-grained asphalt surfaces on higher-volume roadways but fine-grained mixes on low-volume roadways. Furthermore, it is common for state DOTs to change approaches over time such that older pave- ments may be constructed of different surfaces than those used on newer sections. These factors result in the multiplication of the number of sections required for equipment certification. Per NCHRP Synthesis 526, state DOTs are currently using varying numbers of sections in the certification process (Perera 2018). Table 3 summarizes the number of sections by surface Surface Type Number of Sections Number of Responding State DOTs Asphalt 0 1 1 13 2 7 3 or more 4 PCC 0 18 1 4 2 1 3 or more 2 Table 3. Number of sections used in certification (Perera 2018).

Review of Existing Practices 17   type used by various state DOTs in completing the certification process. Most state DOTs use at least one asphalt section, and most do not use a concrete section for certification. From NCHRP Synthesis 526, 11 state DOTs currently use just one test section for certification. From a review of various state procedures, Wisconsin and Iowa each use one concrete-surfaced and one asphalt-surfaced section (Perera 2018). Other state DOTs provide limited information about the sections to be used for certification, predominantly indicating that the state DOT will identify the section(s) for certification. Alabama DOT (ALDOT) Procedure 448 identifies four specific test sections for use in the certification process (ALDOT 2015). Three of these sections represent three levels of roughness on dense-graded asphalt concrete. The fourth section is a smooth section with an open-graded asphalt surface. The profiler is required to make 10 runs, but the standard does not identify changes in speed. In addition to the analysis using cross-correlation to compare the longitudinal profiles, the ALDOT procedure specifies that the IRI is within 5 percent of the IRI produced by the reference device (ALDOT 2015). The California Department of Transportation (Caltrans) was noted in NCHRP Synthesis 526 as following a procedure other than AASHTO R 56-14 but including the cross-correlation analysis approach. Caltrans Procedure 387 governs the certification of inertial profilers used on Caltrans projects. This method indicates that two sections are used, one asphalt section and one jointed concrete section. The equipment is required to collect 10 repeat runs of the sections. The method also notes that the jointed concrete section may not be used based on the availability of that section. The plan also requires the analysis to include the cross-correlation approach and a comparison of the IRI values to those from the reference device with a maximum error of 5 percent (Caltrans 2016), while the comparison of the IRI is not included in the analysis in AASHTO R 56-14. Colorado DOT (CODOT) uses an approach based on the process documented in AASHTO R 56-14 (2022). The Colorado method allows for collection of 12 runs and the substitution of the first two runs with the last two runs based on the analysis results. If the 12 runs do not provide a successful certification, then the equipment must re-collect 12 runs, and the analysis must be repeated (CODOT 2020). This number of runs is more than the required runs in AASHTO R 56-14, and AASHTO R 56-14 does not identify an approach for the event of the equipment failing the dynamic certification. Michigan DOT (MIDOT) requires that the certification section be measured 10 times by the equipment being certified. The data analysis approach is similar to that identified in AASHTO R 56-14. These data are used to identify 15 bumps or dips within the profile with a deviation greater than 0.2 in. The mean and standard deviation of each deviation’s height or depth is cal- culated, and the mean of the standard deviation is required to be less than 0.1 in. The bias is reviewed by comparing the heights and depths for the 15 deviations to rod and level survey data. Analyses are also conducted by reviewing the standard deviation of the IRI calculated from the 10 profiles and the difference in these IRI values to the IRI determined from the rod and level survey (MIDOT 2020). The Ohio DOT (OHDOT) procedure does not identify specifics about the sections used for the certification of devices. The analysis requires a minimum cross-correlation of 94 percent for both repeatability and accuracy (OHDOT 2020). Similarly, Oregon DOT (ORDOT) uses a different set of requirements of 90 percent for repeatability and 88 percent for accuracy (ORDOT 2019). These limits for both Ohio and Oregon are different from those recommended in AASHTO R 56-14. The Texas DOT (TxDOT) procedure uses 10 runs from multiple test sections of varying sur- face texture and roughness. In place of the cross-correlation analysis specified by AASHTO R 56,

18 Enhancement of the Practice for Certification of Inertial Profiling Systems profiles are compared using an elevation-based approach. The elevation measurements are compared at each reporting interval, and the square root of the mean variance of these measure- ments is calculated to assess the profiler repeatability and accuracy. The reference device used to evaluate accuracy includes a combination of rod and level survey and other devices, such as an inclinometer-based device or other walking profiler. The rod and level data are used to correct these data at specific intervals over the length of the profile collected. The accuracy is based on a point-to-point comparison of elevation measurements (TxDOT 2018). Reference Device The first step in a certification process to meet the approach for a traceable reference must be to establish a reference that is maintained as a “known” value or that may be compared against a “known” value. This approach may be achieved by either building a pavement with a known profile or measuring an existing pavement with a device meeting known standards. Building a pavement with a known profile is expected to be impractical. One difficulty in achieving a pave- ment of known profile is that the pavement profile may be impacted by seasonal or even daily variations in temperature (Karamihas and Senn 2012). Therefore, using a reference device with the ability to produce a known value for comparison is the better option. The CPAR report (Karamihas 2005) identifies the minimum requirements for a reference profiler as follows: • Error < 1 percent for wavelengths from 0.5 to 1.16 ft; • Error < 0.25 percent for wavelengths from 1.16 to 118 ft; • Error < 1 percent for wavelengths from 118 to 220 ft; • No phase distortion over the range of wavelengths from 0.5 to 220 ft; and • Distance measurement error within 0.1 percent. It is necessary to consider how those requirements establish repeatability and accuracy requirements on a reference device. Suppose the precision is placed on the profile elevation; it results in an uneven response in the IRI because the IRI filter does not have a uniform gain on the profile elevation for all wavelengths (Karamihas and Gillespie 2002). In other words, the stipulating repeatability on the profile elevation does not directly relate to a standard error in IRI. It is possible to have an error with a wavelength of 7.61 ft that adds as much as 96 in./mile to the IRI but still falls within the ASTM E950-98 precision limit (Karamihas 2011). The actual amount of error added will depend highly on the actual profile. The CPAR report notes that using accuracy criteria based on elevation will place undue influence on the long-wavelength content of the data and may “ignore critical levels of error in the measurement of short-wavelength features” (Karamihas 2005). Gain limits provide a means to evaluate profiler repeatability and accuracy that result in the error limits described previously. However, this approach requires a benchmark device that is very accurate for evaluating the reference profilers. Ultimately, the report recommends a cross-correlation of at least 98 percent for a reference class device (Karamihas 2005). Various devices have been considered reference devices and were evaluated by FHWA (Karamihas 2011). This study evaluated five different reference devices on six different surface types, including a dense-graded asphalt, a chip seal, an open-graded asphalt, a transversely tined concrete, a longitudinally tined concrete, and a diamond-ground concrete (Karamihas 2011). In the evaluation conducted in 2010, most of the devices were able to pass most of the concrete sections, except in the short-waveband evaluation. The diamond-ground section appears to have been the most difficult of the concrete sections for these devices to pass (Karamihas 2011). The chip seal and open-graded asphalt sections appear to have been the most problematic for these devices to pass overall, including the asphalt and concrete sections (Karamihas 2011).

Review of Existing Practices 19   As one example of a reference device requirement outside of the United States, VTI recom- mends the longitudinal profile be sampled at a 100-mm interval and does not identify this reference as a “true” road profile. However, this device is still expected to identify unevenness within wavelengths of 0.5 to 100 m. Ride indexes computed from these data include IRI and root mean square (RMS), with the RMS computed on four ranges of wavelength: 0.5 to 1 m, 1 to 3 m, 3 to 10 m, and 10 to 30 m (Lundberg and Sjogren 2004). VTI uses the Primal device to collect reference data, as depicted in Figure 6. The Primal device measures a 10-m profile, and the shorter profiles are connected using a geodetic elevation mea- surement at the end of each 10-m location. The Primal device consists of a trolley that moves parallel to the direction of travel. The trolley contains a measurement wheel that follows the road profile as the carriage travels along the defined path. The system measures the distance between the measurement wheel and the laser beam that defines the trolley path. This device is classified as a stationary reference plan method (Lundberg and Sjogren 2004). Lundberg and Sjogren (2004) do not indicate how VTI identified that this device met the standard of a reference class device. Beyond just the device, the reference device operator may significantly impact the collected data (de Léon Izeppi and Toom 2014). Training the operators in the proper preparation and operation of the devices is vital in achieving repeatable data from the reference device (de Léon Izeppi and Toom 2014). The reference profile requirements are more than may be identified within a single item of AASHTO R 56-14. Therefore, a separate standard for defining the reference device may be required, because a single device is not currently available that meets the identified level of repeatability and accuracy (Karamihas 2011). AASHTO R 56-14 is not clear on what constitutes a reference device. It specifies that reference profiles be collected using a device that can meet the repeatability and accuracy criteria identi- fied in the Benchmark Test Evaluation Report (Karamihas 2011). This report identifies multiple requirements for a reference device that may be beyond incorporation into a standard on equip- ment certification to be used for routine data collection. AASHTO R 56-14 recommends three runs using the reference device, with data collection occurring immediately prior to data collection with the inertial profiler on JCPs due to poten- tial shape changes caused by temperature effects. These three runs are expected to meet an average coefficient of cross-correlation of 98 percent if the index of interest is IRI. Any of these three runs may then be used to assess the accuracy of the device being certified. This reference selection may be an issue when three coefficients of cross-correlation are, for example, 99, 98, and 97 percent. The reference profile associated with the pairs of cross-correlation values of 97 and 98 percent may not be a good selection. The profile associated with the 99 and 98 percent Figure 6. VTI’s Primal device for reference profile measurements (Sjogren et al. 2008).

20 Enhancement of the Practice for Certification of Inertial Profiling Systems cross-correlations is likely to be the best representation of the pavement surface given that it is most like the other two profiles. The number of repeats and repeatability of the reference profile identified within AASHTO R 56-14 is recommended under Note 2 in Appendix A, but specific requirements for the reference device or collecting data with that reference device are not included. Additionally, no provision is made in AASHTO R 56-14 for data collected on consecutive days. For example, if data are collected using a reference device and three runs are collected meeting the repeat requirements identified within AASHTO R 56-14, additional equipment is to be reviewed the following day. Based on the current version of the standard practice, the reference device would need to collect three repeat runs of the profile again on a subsequent day. However, if the first run closely matches the previous day’s data, this may be sufficient evidence that these data are appropriate for use in certifying equipment on a subsequent day. Analysis Approach Ride Index AASHTO R 56-14 focuses on the IRI with minimal reference to other ride indices. It identifies where a change may be necessary if the profile data will be used to calculate a different ride index. Section 8.3.2 of AASHTO R 56-14 identifies that the IRI should be compared to verify that the equipment software is capable of accurately calculating the IRI. NCHRP Research Report 914 provides recommendations for urban and low-speed profiling based on the use of IRI for the certification process, even though the research identified a temporal intensity of roughness index, the “Golden Car” average rectified velocity (GCARV), that would permit better characterization of roughness at lower simulated speeds (Karamihas et al. 2019). Profile Analysis AASHTO R 56-14 does not identify a timeline for providing data to the state DOT or other agency providing the certification. The lack of a timeline or an extended timeline may allow for extraneous post-processing of the data, but a short timeline may not allow for sufficient review by those performing data collection to identify concerns they have with their own data. The recommended analysis approach in AASHTO R 56-14 is cross-correlation. This analysis approach provides a means to evaluate the overall roughness and spatial distribution. The analysis provides a coefficient of cross-correlation ranging from 0 to 100 percent, with 100 per- cent indicating perfect agreement between the two profiles compared (Karamihas 2004). Implementing the cross-correlation method in AASHTO R 56-14 is facilitated with FHWA ProVAL software. The ProVAL Profiler Certification Module (PCM) uses the following process to implement the cross-correlation analysis algorithm detailed in the appendix of AASHTO R 56-14 (2022) (FHWA ProVAL Workshops 2018): 1. Filter both profiles with the IRI quarter car filter, focusing on the waveband of interest. 2. Convert to slope to avoid the masking effects as in the elevation-based method. 3. Interpolate one profile to the interval of another when the sample intervals of the two profiles are different. 4. Remove the mean, casting the profiles around the mean value. 5. Apply the adjustment factor, scaling the comparison score between 0 and 100 percent. The cross-correlation algorithm in the appendix of AASHTO R 56-14 allows the sampling interval to be adjusted within ±0.1 percent to maximize the result. In the current version of AASHTO R 56-14, the cross-correlation threshold is 92 percent for the repeatability score and 90 percent for the accuracy score.

Review of Existing Practices 21   This analysis approach is superior to the direct comparison of profile elevation values because it compares agreement with the waveband of interest of the profiles (Karamihas 2004). Further, it is improved over a direct comparison of IRI values because it compares the spatial distribution of roughness between two profiles to avoid the compensating effects of IRI errors over the length of the test section (Karamihas 2004). The ProVAL PCM produces a summary report and a detailed report (FHWA 2016). The summary report includes six tables of cross-correlation results of repeatability and accuracy for the left and right wheel track profiles, as shown in Figure 7 (FHWA 2016). The detailed report consists of the comparison results for each profile pair, including the resulting cross-correlation, shape coefficient, roughness coefficient, offset, basis, comparison IRI, and IRI difference in percent, as shown in Figure 8 (FHWA 2016). If the two profiles are from the same device, then the cross-correlation provides a measure of repeatability (Karamihas 2004). If the two profiles are from two different devices, the cross- correlation provides reproducibility. If one of the profiles was collected by a reference device or otherwise identified as the “true profile,” the result provides a measure of the accuracy of the device as measured against the reference device. The recommended levels of cross-correlation are 98 percent for a “research” class device, 94 percent for a “project” class device, and 88 percent for a “network” class device (Karamihas 2004). This analysis approach may also be used to perform diagnostics using various filtering tech- niques (Karamihas 2004). Where two profiles do not agree, filtering the profiles to various levels (e.g., filtering out all wavelengths except those in the range of 1 to 10 ft) provides an under- standing of where the disagreement occurs between the two profiles and ultimately may help in identifying the root cause of disagreement between the two profiles (Karamihas 2004). The criteria for a reference device are a cross-correlation of at least 98 percent with an IRI filter, 98 percent for the long waveband, 98 percent for the medium waveband, and 94 percent for the short waveband, compared to the benchmark profile (Karamihas 2005). The same levels of cross-correlation are expected for repeat measurements as well. The CPAR report recommends a cross-correlation level of 94 percent as a minimum for certification of a device for routine collection of longitudinal profiles (Karamihas 2005). The FHWA conducted studies in 2009 and 2010 to identify devices that may serve as reference devices for evaluating devices for collecting longitudinal profile data. While the studies identi- fied a device achieving the minimum repeatability and accuracy requirements over the broadest range of conditions, the device did not achieve passing scores for all conditions (Karamihas and Perera 2014). Further, no device was demonstrated as meeting all the reference device requirements (Karamihas and Perera 2014). Data analysis is based on first filtering the data to focus on the profile features of interest. In other words, the data are to be filtered based on the index to be used in analyzing the ride quality. Both the reference and device profiles are to be filtered using this same filter. In most cases, the IRI is the index of interest (thus using the “Golden Quarter Car” model as a prefilter). Still, the state DOT may be interested in using a different index in some cases. The processed profiles are compared using a cross-correlation analysis. An agreement score of 92 percent is suggested within the profile for repeatability. This score is the mean of the cross- correlation values calculated between each pair of traces collected by the device being certified. The score of 92 percent is based on the 2004 Profiler Rodeo study (Karamihas 2005), identifying that this level of cross-correlation results in IRI agreement within 5 percent for a 95 percent confidence level where the IRI is less than 150 in./mile. For sections with a higher IRI, the requirement for cross-correlation may be lowered. During the cross-correlation computation, a penalty function is applied to evaluate where specific features coincide to avoid errors masked by compensation effects.

Figure 7. The summary report of the ProVAL PCM (FHWA ProVAL workshops 2018).

Review of Existing Practices 23   Per AASHTO R 56-14, each of the filtered profiles is compared to the reference profile to assess accuracy. The mean value of ten cross-correlation values must meet a value of 90 percent or greater to achieve an IRI value within 5 percent of the reference level at a 95 percent confidence level for each wheel path. A state DOT may have differing requirements for network-level data than for project-level data. AASHTO R 56-14 provides no allowance for considering the risk associated with the error in the data compared to the potential extra expense of using equipment that meets higher requirements for accuracy and repeatability. Summary Significant effort has been expended to identify appropriate means to certify profiling devices. This research has identified a set of parameters that result in a need for up to 24 potential test sections to accurately understand the equipment’s level of accuracy and precision on the range of surface textures and levels of roughness on each surface type that may be encountered. However, additional effort is needed to optimize these factors to assist state DOTs with developing a realistic approach to the certification process. Figure 8. The detailed report of the ProVAL PCM (FHWA ProVAL workshops 2018).

Next: Chapter 3 - Analysis of Certification Data »
Enhancement of the Practice for Certification of Inertial Profiling Systems Get This Book
×
 Enhancement of the Practice for Certification of Inertial Profiling Systems
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Inertial profilers are used by state departments of transportation and others to produce an accurate and repeatable measure of the longitudinal pavement profile, which can be analyzed to produce various smoothness statistics such as the International Roughness Index.

NCHRP Research Report 1057: Enhancement of the Practice for Certification of Inertial Profiling Systems, from TRB's National Cooperative Highway Research Program, proposes revisions to AASHTO R 56-14 to enhance the practice for certification of inertial profiling systems.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!