Stage 4 Distribution, March 31, 2009 November 2008 Technical Memorandum: UCPRC-TM-2008-13 A A A n n n A A A s s s s s s e e e s s s s s s m m m e e e n n n t t t o o o f f f A A A u u u t t t o o o m m m a a a t t t e e e d d d P P P a a a v v v e e e m m m e e e n n n t t t D D D i i i s s s t t t r r r e e e s s s s s s I I I d d d e e e n n n t t t i i i f f f i i i c c c a a a t t t i i i o o o n n n T T T e e e c c c h h h n n n o o o l l l o o o g g g i i i e e e s s s i i i n n n C C C a a a l l l i i i f f f o o o r r r n n n i i i a a a Authors: Bruce Steven, John T. Harvey, and Bor-Wen Tsai Work Conducted as part of Partnered Pavement Research Center Strategic Plan Element 3.3: Support for Implementation of Caltrans PMS PREPARED FOR: California Department of Transportation (Caltrans) Division of Research and Innovation Division of Pavements PREPARED BY: University of California Pavement Research Center Davis and Berkeley
70
Embed
An Assessment of Automated Pavement Distress ... · An Assessment of Automated Pavement Distress Identification Technologies in California Authors: Bruce Steven, John T. Harvey, and
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Title: An Assessment of Automated Pavement Distress Identification Technologies in California
Author: Bruce Steven, John T. Harvey, and Bor-Wen Tsai
Prepared for: California Department of Transportation, Division of Research and Innovation, Division of Pavements
FHWA No.
Work Submitted December 3, 2008
Date:March 31, 2009
Strategic Plan Element No: 3.3
Status: Draft
Version No:Stage 4 review draft
Abstract: This technical memorandum describes work undertaken to evaluate the state of the practice for automated image capture and subsequent rating of pavement distresses. The Caltrans Pavement Condition Survey (PCS) crew, UCPRC personnel, and six automated condition survey vendors evaluated ten 500-foot long sections in an area near Davis, California that encompassed a wide range of pavement types and distresses. The distress evaluations were performed using a newly developed distress matrix containing elements that are based on the underlying mechanisms of pavement distress, elements of which have been incorporated into the 2008 Caltrans PCS. Use of this new matrix will facilitate development of network-level pavement performance models, and will also provide “feedback” to Mechanistic-Empirical pavement design models. It was found that the vendors are able to collect images of sufficient resolution and clarity at highway speeds to enable distress evaluations to be performed. It was also found that the different distress-type definitions need to be refined by Caltrans/UCPRC to be simpler and more precise for the new automated data collection PCS manual current under development.
Proposals for implementation: Develop new distress-type definitions that are based on identification of crack geometries and locations. Include IRI in vendor RFQ process.
Related documents:
Signatures:
B. Steven 1st Author
J. Harvey Technical Review
D. Spinner Editor
J. Harvey Principal Investigator
J. Holland Caltrans Contract Manager
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 iii
DISCLAIMER
The contents of this report reflect the views of the authors who are responsible for the facts and accuracy
of the data presented herein. The contents do not necessarily reflect the official views or policies of the
State of California or the Federal Highway Administration. This report does not constitute a standard,
specification, or regulation.
ACKNOWLEDGMENTS
This work was funded and managed by the California Department of Transportation, Division of
Research and Innovation, under the direction of Nick Burmas, Joe Holland, Michael Samadian, and
Alfredo Rodriguez. The technical lead for Pavement Standards Team was Peter Vacura of the Division of
Pavement Management. The authors would also like to thank members of the Caltrans Pavement
Condition Survey crew for their help in refining distress definitions and in performing the manual survey
of the test sections, and the members of the Caltrans PMS Expert Task Group for their many suggestions
and comments. The authors would also like to thank Erwin Kohler for help in setting up the test sections.
PROJECT OBJECTIVES
The purpose of the work included in this technical memorandum is to evaluate the state of the practice
with respect to automated methods for recording and measuring pavement distress, and to make
recommendations for changes to the Caltrans Pavement Condition Survey manual for use with automated
data collection.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 iv
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 v
TABLE OF CONTENTS Acknowledgments ...................................................................................................................................... iii Project Objectives ...................................................................................................................................... iii List of Tables .............................................................................................................................................vii List of Figures............................................................................................................................................. ix 1 Introduction.................................................................................................................................... 1
1.1 Pavement Condition Survey Techniques Used by Caltrans Prior to 2008 .................................. 1 1.2 Pavement Condition Survey Techniques Implemented in 2008 PCS.......................................... 3 1.3 Future Pavement Condition Survey Techniques Used by Caltrans............................................. 3 1.4 Need to Assess State-of-the-Technology for Automated Distress Collection............................. 4 1.5 Objective of this Technical Memorandum .................................................................................. 4
3.1 Rodeo Details .............................................................................................................................. 9 3.2 Manual Survey and Desktop Survey ......................................................................................... 11 3.3 North Carolina DOT Data Collection Rodeo and Workshop .................................................... 12
4 Survey Results .............................................................................................................................. 17 4.1 Section 1: Old Davis Road On-Ramp to NB SR113 ................................................................. 17
4.1.1 Alligator A and B Cracks ...................................................................................................... 18 4.1.2 Cracks Between and Outside the Wheelpaths....................................................................... 19 4.1.3 Roughness and Rutting ......................................................................................................... 21
4.2 Section 2: NB SR113................................................................................................................. 22 4.2.1 Alligator A and B Cracks ...................................................................................................... 22 4.2.2 Cracks Between and Outside the Wheelpaths....................................................................... 24 4.2.3 Roughness and Rutting ......................................................................................................... 25
4.3 Section 3: NB SR113................................................................................................................. 25 4.3.1 Rigid Transverse Cracks ....................................................................................................... 26 4.3.2 Rigid Joints and Cracks Filled with Sealant.......................................................................... 27 4.3.3 Rigid Faulting, Roughness, and Rutting ............................................................................... 27
4.4 Section 4: NB SR113................................................................................................................. 27 4.4.1 Rigid Corner Cracks.............................................................................................................. 28 4.4.2 Rigid Joints and Cracks Filled with Sealant.......................................................................... 29 4.4.3 Rigid Faulting, Roughness, and Rutting ............................................................................... 29
4.6 Section 6: WB SR16.................................................................................................................. 29 4.6.1 Alligator A and B Cracks ...................................................................................................... 30 4.6.2 Cracks Between and Outside the Wheelpaths....................................................................... 31 4.6.3 Roughness and Rutting ......................................................................................................... 33
4.9 Section 9: EB I80....................................................................................................................... 36 4.9.1 Rigid Longitudinal Cracks .................................................................................................... 37 4.9.2 Rigid Transverse Cracks ....................................................................................................... 37 4.9.3 Rigid Joints and Cracks Filled with Sealant and Joint/Crack Spalling ................................. 37 4.9.4 Rigid Faulting, Roughness, and Rutting ............................................................................... 37
4.10 Section 10: EB I80..................................................................................................................... 38 4.10.1 Alligator A and B Cracks.................................................................................................. 38 4.10.2 Cracks Between and Outside the Wheelpaths................................................................... 39 4.10.3 Digouts and Patches.......................................................................................................... 41 4.10.4 Roughness and Rutting ..................................................................................................... 41
4.11 Section 11: EB I80..................................................................................................................... 41 4.11.1 Alligator A and B Cracks.................................................................................................. 42 4.11.2 Cracks Between and Outside the Wheelpaths................................................................... 44 4.11.3 Roughness and Rutting ..................................................................................................... 45
4.12 Roughness Results..................................................................................................................... 45 5 5 Findings and Conclusions......................................................................................................... 53
5.1 Image Capture ........................................................................................................................... 53 5.2 Image Analysis .......................................................................................................................... 54 5.3 Recommended Changes to the Distress Definitions.................................................................. 55 5.4 Recommended Changes Regarding Quality Assurance for IRI Measurements ........................ 56 5.5 Recommendations for the Vendor Evaluation Part of the RFQ Process ................................... 57 5.6 Overall Recommendations ........................................................................................................ 57
Appendix: Summary of IRI (m/km) Values for All Vendors and All Sections ................................... 58
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 vii
LIST OF TABLES Table 2.1: Distress Descriptions in PCS Manual for 2008 Used for Rodeo ................................................ 8 Table 3.1: Site Locations ........................................................................................................................... 10 Table 3.2: Alphabetical List of Participating Vendors............................................................................... 11 Table 3.3: Detailed Descriptions of Vendor Equipment............................................................................ 13 Table 4.1: Section 1, Alligator A Cracking ............................................................................................... 19 Table 4.2: Section 1, Alligator B Cracking................................................................................................ 19 Table 4.3: Section 1, Combined Total for Alligator A and B Cracking .................................................... 19 Table 4.4: Section 1, Alligator C Cracking................................................................................................ 20 Table 4.5: Section 1, Short Transverse Cracks Outside the Wheelpath..................................................... 21 Table 4.6: Section 1, Transverse Cracks.................................................................................................... 21 Table 4.7: Section 1, Longitudinal Cracks................................................................................................. 21 Table 4.8: Section 2, Alligator A Cracking ............................................................................................... 23 Table 4.9: Section 2, Alligator B Cracking................................................................................................ 23 Table 4.10: Section 2, Combined Total for Alligator A and B Cracking .................................................. 23 Table 4.11: Section 2, Alligator C Cracking.............................................................................................. 24 Table 4.12: Section 2, Short Transverse Cracks Outside the Wheelpath................................................... 24 Table 4.13: Section 2, Transverse Cracks.................................................................................................. 24 Table 4.14: Section 2, Longitudinal Cracks............................................................................................... 25 Table 4.15: Section 3, Rigid Transverse Cracks ........................................................................................ 27 Table 4.16: Section 4, Rigid Corner Cracks .............................................................................................. 28 Table 4.17: Section 6, Alligator A Cracking ............................................................................................. 31 Table 4.18: Section 6, Alligator B Cracking.............................................................................................. 31 Table 4.19: Section 6, Combined Total for Alligator A and B Cracking .................................................. 31 Table 4.20: Section 6, Alligator C Cracking.............................................................................................. 32 Table 4.21: Section 6, Short Transverse Cracks Outside the Wheelpath................................................... 32 Table 4.22: Section 6, Transverse Cracks.................................................................................................. 32 Table 4.23: Section 6, Longitudinal Cracks............................................................................................... 32 Table 4.24: Section 8, Rigid Transverse Cracks ........................................................................................ 35 Table 4.25: Section 9, Rigid Transverse Cracks ........................................................................................ 37 Table 4.26: Section 10, Alligator A Cracking ........................................................................................... 39 Table 4.27: Section 10, Alligator B Cracking............................................................................................ 39 Table 4.28: Section 10, Combined Total for Alligator A and B Cracking ................................................ 39 Table 4.29: Section 10, Alligator C Cracking............................................................................................ 40
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 viii
Table 4.30: Section 10, Short Transverse Cracks Outside the Wheelpath................................................. 40 Table 4.31: Section 10, Transverse Cracks................................................................................................ 40 Table 4.32: Section 10, Longitudinal Cracks............................................................................................. 40 Table 4.33: Section 10, Digouts................................................................................................................. 41 Table 4.34: Section 11, Alligator A Cracking ........................................................................................... 43 Table 4.35: Section 11, Alligator B Cracking............................................................................................ 43 Table 4.36: Section 11, Combined Total for Alligator A and B Cracking ................................................ 43 Table 4.37: Section 11, Alligator C Cracking............................................................................................ 44 Table 4.38: Section 11, Short Transverse Cracks Outside the Wheelpath................................................. 44 Table 4.39: Section 11, Transverse Cracks................................................................................................ 44 Table 4.40: Section 11, Longitudinal Cracks............................................................................................. 45 Table 4.41: Summary of the Means and Standard Deviations of Ln(IRI). ................................................. 49
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 ix
LIST OF FIGURES Figure 3.1: Map of rodeo circuit. ............................................................................................................... 11 Figure 3.2: Downward pavement images of a flexible pavement.............................................................. 15 Figure 3.3: Downward pavement images of a rigid pavement. ................................................................. 16 Figure 4.1: Photo of Section 1, Old Davis Road On-Ramp to NB SR 113................................................. 18 Figure 4.2: Photo of Section 2, NB SR113. ............................................................................................... 22 Figure 4.3: Photo of Section 3, NB SR113. ............................................................................................... 26 Figure 4.4: Photo of Section 4, NB SR113. ............................................................................................... 28 Figure 4.5: Photo of Section 6, WB SR16. ................................................................................................ 30 Figure 4.6: Photo of Section 7, SB I505. ................................................................................................... 33 Figure 4.7: Photo of Section 8, EB I80. ..................................................................................................... 34 Figure 4.8: Photo of Section 9, EB I80. ..................................................................................................... 36 Figure 4.9: Photo of Section 10, EB I80. ................................................................................................... 38 Figure 4.10: Photo of Section 11, EB I80. ................................................................................................. 42 Figure 4.11: Summary of IRI (m/km) values for all vendors, all sections, and all subsections.................. 46 Figure 4.12: Summary of IRI (m/km) values for all vendors and all subsections (with
mean-trend lines). .......................................................................................................................... 47 Figure 4.13: Mean IRI (m/km) versus mean standard deviation across vendors for each subsection. ...... 49 Figure 4.14: Mean IRI (m/km) versus mean standard deviation across vendors for each
asphalt-surfaced subsection. .......................................................................................................... 50 Figure 4.15: Mean IRI (m/km) versus mean standard deviation across vendors for each
PCC-surfaced subsection. .............................................................................................................. 50 Figure 4.16: IRI distributions of all the vendors across all the sections: (a) Histogram and
7 Joint spalling length/depth/width across slab %joints 8 Cracks filled with sealant 9 Joints sealed 10 CRC Crack Spacing Average spacing (m) 11 CRC Punchouts # punchouts/segment # 12 CRC Crack Sealed 13 Surface spalling found % area Flexible Pavements 14 Alligator A cracking in wheelpaths Crack width (mm) % wp area 15 Alligator B cracking in wheelpaths Crack width (mm) %wp area 16 Alligator C cracking in lane Crack width (mm) %area outside wp 31 Short transverse cracks outside wp Crack width (mm) Area (m2) 17 Transverse cracking Crack width (mm) # 18 Longitudinal cracks outside wp Crack width (mm) Total crack length
(m) 19 Asphalt pavement binder bleeding Presence % wp area 20 Asphalt pavement has been dug out and
replaced Area (m²) Area
21 Patched (potholed) pavement in this location Area (m²) Area 22 Shoulder joint sealed No joint /open/ sealed %segment length Composite Pavements 23 Corner Reflective Cracking (AC/PCC or
AC/CTB) Crack width (mm) #
24 Longitudinal Reflective Cracking (AC/PCC or AC/CTB)
Crack width (mm) %segment length
25 Transverse Reflective Cracking (AC/PCC or AC/CTB)
Crack width (mm) #
Flexible and Composite Pavements 26 Block cracking Vendors were asked to rate
Notes: CRC is continuously reinforced concrete pavements; wp is wheelpath; # is number; MPD is mean profile depth measured using a high-speed laser in the wheelpath; IRI is International Roughness Index.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 9
3 PAVEMENT CONDITION RODEO
It must be noted that the rodeo was planned and completed in a time span of just a few months. This left
the participating vendors little time to make any changes for the data collection, which was intentional so
that UCPRC/Caltrans might see what current technology could deliver. The vendors also had little time to
analyze the data according to the distress elements as defined for the 2008 PCS. They therefore had
difficulty changing automated or semiautomated distress analysis software and processes to match the
Caltrans distress definitions, and they had no time to go through a normal iterative process to be sure that
they understood all the distress definitions as UCPRC/Caltrans meant them. They were also not
compensated to make major changes in any software used for image analysis, or to spend much time
training their staff in the manual interpretation of distresses from the images.
It is certain that the schedule caused differing interpretations of the PCS distress definitions among the
different vendors, and this resulted in differences in the reported distresses for the test sections.
3.1 Rodeo Details
The route for the rodeo was chosen so that vendors would be required to survey a range of surface types
and conditions. The route included eleven 500-ft long sites chosen by UCPRC personnel. There were four
rigid and seven flexible or composite sites, described in Table 3.1 and shown in Figure 3.1.
A list of eighteen pavement condition survey service providers was compiled and an invitation was issued
to participate in the rodeo. Vendors were compensated by Caltrans for coming with a fixed fee. Six
vendors elected to participate (Table 3.2). The vendors completed the rodeo at their convenience between
mid-April and early June. The vendors were supplied with a detailed description of the route and the
locations of the sections. The vendors also had the opportunity to make a presentation and to demonstrate
their equipment to interested personnel from Caltrans and the UCPRC. The vendors completed the rodeo
circuit with only their own personnel in the survey vehicle. The equipment that each vendor used in the
rodeo is listed in Table 3.3.
The vendors were required to submit their results in a standard format, which five of the six vendors did.
The sixth vendor was unable to meet this requirement in the six months following the rodeo. This vendor
did submit detailed reports of its analysis but the cracking measurements were not in a form that could be
parsed into the standard format. The roughness and rutting measurements from this vendor were in a
format that enabled them to be incorporated into the results.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 10
Brief descriptions of the distress definitions were given to the vendors; as the data analysis progressed,
UCPRC received specific questions from the vendors and their answers were distributed to the other
vendors. However, it is likely that the time frame was insufficient for Caltrans/UCPRC and the different
vendors to arrive at complete agreement in understanding the definitions.
Images from the same pavement from four vendors are shown in Figure 3.2 and Figure 3.3 for flexible
and rigid pavements, respectively. It can be seen that distresses are readily visible in the images. One
vendor reported unhappiness with the aperture settings on its equipment and that the images were not
typical of its equipment.
Table 3.1: Site Locations
Site # Route Direction Lane Pavement Surface PM 1 CA 113 NB Old Davis Rd to Woodland
onramp NB 2 HMA or possibly
semi-rigid (HMA on CTB)
-
2 113 NB 3 (of 4) HMA (Composite) 0.0 3 113 NB 2 PCC 4.30 4 113 NB 2 PCC 7.50 51 16 WB 1 HMA 37.62 6 16 WB 1 HMA 36.4 7 505 SB 2 HMA (Composite) 9.5 8 80 EB 3 PCC 32.48 9 80 EB 3 PCC 33.2 10 80 EB 3 HMA 40.2 11 80 EB 4 HMA 42.4 1 Site was eliminated from rodeo due to rehabilitation between the first and second vendors completing their surveys.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 11
Figure 3.1: Map of rodeo circuit.
Table 3.2: Alphabetical List of Participating Vendors
Vendor Applied Research Associates Inc.
EBA Engineering Consultants Ltd. Dynatest Consulting Inc.
Mandli Communications Inc. Pathway Services Inc. Roadware Group Inc.
3.2 Manual Survey and Desktop Survey
In order to provide an indication of the relationship of the reduced manual distress matrix and the full
automated matrix, the Caltrans PCS crew manually surveyed the sections in the rodeo. In addition, the
UCPRC performed a “desktop” survey to measure the cracking and other distresses that were visible in
the images provided by the vendors. The UCPRC survey was used as the baseline to provide a point of
reference for assessing the automated technologies.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 12
3.3 North Carolina DOT Data Collection Rodeo and Workshop
The North Carolina Department of Transport (NCDOT) conducted an exercise similar to the one
described in this technical memorandum in the summer of 2008. The NCDOT was assessing the available
technologies with a view to automating their current manual evaluation methods. The NCDOT manages
the complete road network in North Carolina, approximately 80,000 centerline miles, hence its desire to
move toward an automated system for both safety and logistical reasons. The NCDOT required the
automated outputs to follow the same format and style as its current manual rating system (0 to
100 overall score) as the department wished to maintain continuity with historical data.
The main findings from the NCDOT exercise were:
• Only two vendors participated in the rodeo.
• The four experienced manual raters who provided the “ground truth” had considerable
variation in their individual ratings. “Ground truth” is defined as the base measurement
against which all other measurements are compared. It was found that as the extent/severity of
distress increased, the variability in the ratings increased. This was due to a rater having to
make more complex interpretations and decisions about distress type and severity.
• There were difficulties in translating the output from the automated systems to match that
required by the historical overall rating score.
• There was considerable variation between the vendor and ground truth results. This could be
attributed partly to the fact that neither training nor feedback was given to the vendors, apart
from the NCDOT rating manual.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 13
Table 3.3: Detailed Descriptions of Vendor Equipment
ARA* Dynatest* EBA Mandli Communication*
Pathway* Fugro-Roadware
Profiling Systems
Rutting Laser, 5-sensor Dynatest Mk III – 7 sensor
11 point laser INO LRMS1 INO LRMS1 INO LRMS1
Roughness Laser Dynatest Mk III Yes (Selcom) Dynatest Mk IV Yes Yes Texture 32kHz Lasers No texture sensor in
this profilometer Yes (Selcom) Optional Optional Optional
Imaging Systems
System ICC INO LRIS2 integrated by Dynatest
In-house INO LRIS2 In-house In-house
Image type Line scan Line scan Line scan Line scan Line scan Area scan No. of cameras 1 2 1 2 1 2 Illumination Visible (white)
4000 pixels per 4000mm of pavement (width) = 1.0 mm/pixel
4000 pixels per 4000 mm of pavement (width) = 1.0 mm/pixel
4000 pixels per 4000 mm of pavement (width) = 1.0 mm/pixel
4000 or 6000 pixels per 4000 mm of pavement (width) = 1.0 or 0.67 mm/pixel
2784 pixels per 3600mm of pavement (width) = 1.3 mm/pixel
Longitudinal resolution
1.0 mm cracks visible
1.0 mm/pixel 1.5 mm Continuous, 100% coverage
2 mm, 1 mm optional
Location system DMI, DGPS & Applanix INS
Trimble AG-132 receiver + Applanix POS LV 220
Applanix POS LV
Applanix POS LV DMI/GPS Applanix POS LV420
Digital imaging ROW
3 Cameras (ICC) currently 1300 x 1028 pixels each. Upgrading soon.
Panasonic HVX-200 high-resolution, digital, color camera
Single camera 1600 x 1200 pixels, 2 cameras
Up to 3 cameras Single high definition camera, 1920 x 1080 pixel resolution
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 14
Table 3.4: Detailed Descriptions of Vendor Equipment (con’t.)
ARA* Dynatest* EBA Mandli Communication*
Pathway* Fugro-Roadware
Data Management In-vehicle data management/control software
ICC WinPro Dynatest RSP (Road Surface Profiler)
NS Mandli’s DVX data collection software, developed in-house
PathRunnerXP including Database/GPS Navigation/VoiceFeedback
ARAN 3
Image/post-processing software
ICC WinRP, ICC Work Station Software & Others depending upon client criteria
Waylink Crack-scope Semi-automatic 20% auto 80% manual
PathView II WiseCrax, D/V-Rate, Step3 (Data compiling software)
Data turnaround time
Depends upon criteria and data being produced
16 days NS 2 weeks 2 weeks 6 weeks
*Information confirmed with vendor NS Information not supplied ICC International Cybernetics Corporation INO Canadian optics/photonics development and manufacturing institute INS Inertial Navigation System LRMS Laser Rut Measurement System – 2-laser illumination with line scan camera, 1280 points/lane, up to 4000 mm wide LRIS Laser Road Imaging System ROW Right of Way Line Scan camera Takes a picture with a resolution of one pixel by n pixels. Area Scan camera Takes a picture with a resolution of n pixels by m pixels. The same a regular camera
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 15
Note: The upper right image is in reverse order with respect to the other images due to the way the vendor’s software arranges the line scans for viewing (For three of the vendors, the images show increasing distance as going from the bottom to the top of the image. The fourth vendor shows
increasing distance from the top to the bottom of the image). Figure 3.2: Downward pavement images of a flexible pavement.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 16
Note: The upper left image is in reverse order with respect to the other images due to the way the vendor’s software orders the line scans for viewing.
Figure 3.3: Downward pavement images of a rigid pavement.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 17
4 SURVEY RESULTS
The results of the survey are presented separately for each of the eleven sections. The vendors are only
identified by a number, which remains the same throughout all the results. It was decided to use blind
reporting of the results as the vendors were not given feedback on their submitted data, and as such, there
may have been discrepancies between the UCPRC and vendor interpretations of distresses and the
reporting format. In a formal contracting environment, the vendor would be asked to submit a test set of
data at the beginning of the contract and this would be reviewed by the client and vendor to ensure that
both parties were “seeing” the same distresses in the pavement.
The vendors were asked to rate each section as a series of 100-ft subsections and to present the results
separately for each subsection. Only data for the distresses that were present in each section are given in
this document. During the analysis of the rigid sections it was found that the vendors had assigned
differing numbers of slabs to each subsection, thus affecting the percentages reported for each subsection
and making vendor-to-vendor comparisons difficult. By analyzing the reported percentages, the number
of cracks per section could be estimated using the assumption that the distress counts were integer
quantities. For example if a vendor reported 14 percent for a distress, it could be assumed that there were
7 distress counts in the subsection (100 / 14 = 7.14 ≈ 7). Once the number of cracks for each subsection
was known, the totals for each subsection were summed to give a total for the complete section and the
totals were then reported.
4.1 Section 1: Old Davis Road On-Ramp to NB SR113
This section was a flexible pavement with cracking. The majority of the cracks were sealed, with some
locations being completely covered by sealant. The types of distress present in this section could create
difficulty for a rating process because it could be argued that the cracks are a series of longitudinal and
transverse cracks rather than Alligator A, B, and C cracks. Figure 4.1 shows a photo of Section 1.
The feedback from the results from Section 1 is that it is important to have more detailed definitions of
the various crack types, e.g., the minimum separation distance between isolated cracks, and what
constitutes a wheelpath-related crack. The definitions in the 2008 PCS manual provided to the vendors do
not have that level of detail.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 18
Figure 4.1: Photo of Section 1, Old Davis Road On-Ramp to NB SR 113.
4.1.1 Alligator A and B Cracks
The vendor and manual results are shown in Table 4.1 for Alligator A cracks and in Table 4.2 for
Alligator B cracks. Because the cracks tended to meander in this section, a crack could drift in and out of
the defined wheelpath, thus moving a crack between Alligator A and C or Alligator A and Short
Transverse. For the Alligator A cracks, two vendors over-reported cracking when compared to the manual
measurements and one vendor was closer to the manual results. For the Alligator B cracks, the vendor
results were lower than the manual measurements, with only one vendor close to the manual results. The
PCS manual shows that Alligator A and B cracking are mutually exclusive, i.e., the combined percentage
should not exceed 100 percent. The sum of Alligator A and B for Vendor 1 is greater than 100 percent,
indicating a misinterpretation of the definitions. The combined totals for Alligator A and B cracks
appearing in Table 4.3 show that Vendors 3 and 4 are closest to the manual totals, but still with up to a
20 percent difference. Vendor 2 appears to have used a different interpretation of distress definitions for
asphalt surfaced sections than the other vendors, including Section 1.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 19
Considering this section’s extensive and complex cracking in and between the wheelpaths (possibly
including reflective cracking from a shrinkage-cracked cement-treated base), it is clear that the responsive
vendors (Vendors 1, 3, 4, and 5) all identified extensive Alligator cracking in the section (as shown in
Table 4.3), and that the definitions of Alligator A and B need to be more precise in the PCS manual.
Table 4.1: Section 1, Alligator A Cracking
Vendor 0-100' 100-200' 200-300' 300-400' 400-500' % of wheelpath
UCPRC/CT PCS 100 99 78 71 80 ** Misinterpretating PCS definition, Vendor 1 counted both Alligator A and B in the same locations in the wheelpath if both were present, whereas in the intended definition there can only be one type of Alligator cracking in the same location.
4.1.2 Cracks Between and Outside the Wheelpaths
Because of potential misinterpretation of the crack definitions, the results for percent of area with the
different types of cracking outside the wheelpaths defined in the manual—Alligator C, Short Cracks
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 20
Outside the Wheelpath, Transverse, and Longitudinal cracks—are shown back-to-back in this technical
memorandum. The results appear in Table 4.4 through to Table 4.7. As there was no clear trend regarding
whether the vendors over- or underreported when compared to the manual results, this reinforces the
feedback from the rodeo that the definitions provided in the 2008 PCS manual need more detail. Because
vendors overreported transverse cracks, this would explain the underreporting for the other categories.
These results also suggest that an option is to simplify many of the outside-the-wheelpath cracking
categories, since the mechanisms for some are not clear. For example, Alligator C and Short Cracks
Outside the Wheelpath do not have clear connections to pavement damage mechanisms, and are most
likely the result of advanced aging and cracking, and/or reflection of shrinkage cracks from underlying
cemented layers. In addition, the maintenance strategies for this cracking are essentially the same.
Longitudinal cracks may be due to asphalt construction joints or to reflection from underlying PCC joints.
Transverse cracks, on the other hand, are clearly tied to either low-temperature cracking or to reflective
cracking of transverse joints in underlying PCC. Therefore, both of these types of cracking can be
assigned to a particular mechanism from information regarding the pavement structure.
In general it can also be concluded that the analysis systems that the vendors use are able to detect and
rate cracks that have been sealed.
Table 4.4: Section 1, Alligator C Cracking
Vendor 0-100' 100-200' 200-300' 300-400' 400-500' % of area between wheelpaths
All vendors reported rutting in the section, however there was variability as to the rut depth measurement
category. Four vendors reported a maximum rut depth of 6.4 mm with a random distribution between the
rut depths categories of less than 1.2 mm, and 1.2 to 6.4 mm; the other vendor reported rut depth
distributions in these categories, plus the next more severe category (12.5-to-19 mm).
The roughness values are summarized in Section 4.12.
4.12 Roughness Results
A table showing the roughness measures (IRI) for all vendors and all subsections is included in the
Appendix. The table shows the mean, standard deviation, and total range (maximum minus minimum) of
IRI for each section and subsection across the different vendors, the mean of IRI for each vendor across
all of the subsections in each section, and the mean and standard deviation of all the vendor’s means for
each section is included in the Appendix.
Trellis graphs show the data in Figure 4.11 and Figure 4.12 for each section and each vendor. Each
column in Figure 4.11 shows measured IRI for all six vendors for the same section with each subsection
shown as an individual circle from left to right. Figure 4.12 shows the values for each vendor, with all
values for each of the subsections (circles) stacked on top of each other in each of the 11 sections and a
mean trend line between them. It can be seen that the trends for IRI values follow same patterns and have
similar values across the vendors. The results also show that even within the short 500-ft sections, there is
considerable variation between the 100-ft subsections.
Because IRI is an average value, shorter reporting lengths tend to have a higher variability due to the
influence of individual profile perturbations. The effect of isolated perturbations is attenuated over longer
section lengths as a result of the averaging process.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 46
1
2
3
4
5
1 2 3 4 5
1vendor1
2vendor1
1 2 3 4 5
3vendor1
4vendor1
1 2 3 4 5
6vendor1
7vendor1
1 2 3 4 5
8vendor1
9vendor1
1 2 3 4 5
10vendor1
11vendor1
1vendor2
2vendor2
3vendor2
4vendor2
6vendor2
7vendor2
8vendor2
9vendor2
10vendor2
1
2
3
4
5
11vendor2
1
2
3
4
5
1vendor3
2vendor3
3vendor3
4vendor3
6vendor3
7vendor3
8vendor3
9vendor3
10vendor3
11vendor3
1vendor4
2vendor4
3vendor4
4vendor4
6vendor4
7vendor4
8vendor4
9vendor4
10vendor4
1
2
3
4
5
11vendor4
1
2
3
4
5
1vendor5
2vendor5
3vendor5
4vendor5
6vendor5
7vendor5
8vendor5
9vendor5
10vendor5
11vendor5
1vendor6
1 2 3 4 5
2vendor6
3vendor6
1 2 3 4 5
4vendor6
6vendor6
1 2 3 4 5
7vendor6
8vendor6
1 2 3 4 5
9vendor6
10vendor6
1
2
3
4
5
1 2 3 4 5
11vendor6
IRI (
m/k
m)
Figure 4.11: Summary of IRI (m/km) values for all vendors, all sections, and all subsections.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 47
Section
IRI (
m/k
m)
1
2
3
4
5
2 4 6 8 10
vendor4 vendor3
2 4 6 8 10
vendor5
vendor2
2 4 6 8 10
vendor6
1
2
3
4
5
vendor1
Section
IRI (
m/k
m)
1
2
3
4
5
2 4 6 8 10
vendor4 vendor3
2 4 6 8 10
vendor5
vendor2
2 4 6 8 10
vendor6
1
2
3
4
5
vendor1
Figure 4.12: Summary of IRI (m/km) values for all vendors and all subsections (with mean-trend lines).
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 48
The standard deviation between vendors’ means for each very short 100-ft subsection is somewhat
correlated with the mean roughness for the subsection, as can be seen in Figure 4.13, with rougher
subsections having greater variation between vendors. Comparison of Figure 4.14 (asphalt surfaces) with
Figure 4.15 (PCC surfaces) indicates that the variation between vendors is somewhat greater on the
subsections with asphalt surfaces than on the subsections with PCC surfaces.
The results in the appendix also show that the vendor outliers are well distributed across all test sections.
These results indicate that there is randomness across sections, but half the vendors are responsible for
most of the outliers. The results in the table also indicate that there is no pattern of one vendor being high
or low relative to the mean.
Each subsection was short and had its own specific roughness, and each test car may have followed a
slightly different line along the wheelpath. It must be remembered that IRI is being used as a network-
level measure, not to find precise measures of roughness for sections that are 100 ft long. This is borne
out by comparison of the overall IRI summed across the ten sections (IRI is linearly cumulative1), each
500 ft long, as if it were one section with a total length of 5,000 ft (nearly a mile) and a wide range of
roughness within each of the section’s 100-ft subsections, shown below. These results indicate that when
summed across longer sections, the differences in measured IRI on are lower than those indicated by
individual subsection values; the vendors generally show similar variability. A large part of this may be
the averaging of the drivers’ lines through the wheelpaths over longer sections.
Vendor 1 2 3 4 5 6
Mean IRI (m/km) 2.14 2.10 1.92 2.12 2.02 2.07
Mean IRI (inches/mile) 136 133 122 135 128 132
Std Dev IRI (m/km) 1.03 1.03 0.83 0.99 0.89 1.08
Maximum (m/km) 5.48 5.12 4.47 4.27 4.89 5.22
Minimum (m/km) 0.69 0.64 0.65 0.69 0.63 0.69
A normal plot (QQ-plot) of the Studentized residuals is used to check for outliers. If there are vertical
jumps near either end of such a plot, or even if the plot turns sharply upward or downward near the ends,
we might have points that should be flagged for further investigation.”2 Figure 4.16 plots the IRI
1 Sayers, M., and S. Karamihas. 1998. The Little Book of Profiling. University of Michigan. 2 Sen, A., and Srivastava, M. 1990. Regression Analysis: Theory, Methods, and Applications. Springer-Verlag, New
York.
Stage 4 Distribution, March 31, 2009
UCPRC-TM-2008-13 49
distributions of all the vendors across all the sections in terms of histogram and QQ-plot. The IRI value is
expressed in natural logarithm. The plots indicate that the distribution of IRI of each vendor seems to be
lognormal distribution. All of the vendors have similar QQ plots. Table 4.41 summarizes the means and
standard deviations of all the vendors across all the sections.
Table 4.41: Summary of the Means and Standard Deviations of Ln(IRI).