National Academies Press: OpenBook

Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)” (2014)

Chapter: Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)"

Page 1
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 1
Page 2
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 2
Page 3
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 3
Page 4
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 4
Page 5
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 5
Page 6
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 6
Page 7
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 7
Page 8
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 8
Page 9
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 9
Page 10
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 10
Page 11
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 11
Page 12
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 12
Page 13
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 13
Page 14
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 14
Page 15
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 15
Page 16
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 16
Page 17
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 17
Page 18
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 18
Page 19
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 19
Page 20
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 20
Page 21
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 21
Page 22
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 22
Page 23
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 23
Page 24
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 24
Page 25
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 25
Page 26
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 26
Page 27
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 27
Page 28
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 28
Page 29
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 29
Page 30
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 30
Page 31
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 31
Page 32
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 32
Page 33
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 33
Page 34
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 34
Page 35
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 35
Page 36
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 36
Page 37
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 37
Page 38
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 38
Page 39
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 39
Page 40
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 40
Page 41
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 41
Page 42
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 42
Page 43
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 43
Page 44
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 44
Page 45
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 45
Page 46
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 46
Page 47
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 47
Page 48
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 48
Page 49
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 49
Page 50
Suggested Citation:"Precision Estimates of AASHTO T 324, "Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)" ." National Academies of Sciences, Engineering, and Medicine. 2014. Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)”. Washington, DC: The National Academies Press. doi: 10.17226/22242.
×
Page 50

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.

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM Responsible Senior Program Officer: Edward T. Harrigan October 2014 Research Results Digest 390 CHAPTER 1—INTRODUCTION AND RESEARCH APPROACH 1.1 Background The Hamburg wheel tracking test (HWTT) has been extensively used by state DOTs and industry to identify mix- tures prone to rutting or moisture damage. AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA),” describes the procedure for testing asphalt mixture samples using the HWTT device. The method specifies the testing of submerged, compacted asphalt mixture in a reciprocating rolling-wheel device (1). The test results provide infor- mation about the rate of permanent defor- mation from a moving concentrated load. The test accommodates both linearly kneaded slab and gyratory-compacted specimens. Alternatively, field cores of 150-mm, 250-mm, or 300-mm in diam- eter or saw-cut slab specimens may be tested. 1.2 Problem Statement The accurate and precise measurement of asphalt mixture properties is an important aspect of designing and selecting appropri- ate mixtures for various pavement projects. AASHTO T 324 has been extensively used in recent years for detecting rutting, mois- ture susceptibility, or both, of asphalt mix- tures. However, there is no information on the precision of the test method, includ- ing the allowable differences between two replicate measurements in one laboratory or measurements in two laboratories. In addition, important aspects of the test are not sufficiently specified in the test method; these include position of the wheel with respect to specimen, verification of the location of the measurements, specimen preparation and assembly, and analysis and reporting of test data. Because these fac- tors could significantly affect HWTT mea- surements and performance verification of asphalt mixtures, it is important to identify PRECISION ESTIMATES OF AASHTO T 324, “HAMBURG WHEEL-TRACK TESTING OF COMPACTED HOT MIX ASPHALT (HMA)” NCHRP Project 10-87, Task Order #2B This digest presents results of Task Order #2B of NCHRP Project 10-87, “Precision Statements for AASHTO Standard Methods of Test.” This work was conducted to update precision estimates of AASHTO T 324. Using the computed precision estimates, new precision statements for the test method have been prepared and are presented in this digest. The research was conducted by the AASHTO Materials Reference Laboratory. Dr. Haleh Azari was the Principal Investigator. C O N T E N T S Chapter 1—Introduction and Research Approach, 1 1.1 Background, 1 1.2 Problem Statement, 1 1.3 Research Objectives, 2 1.4 Scope of Study, 2 Chapter 2—Design and Conduct of the Study, 2 2.1 Materials Selection, 2 2.2 Test Samples, 3 2.3 Test Machine, 3 2.4 Specimen Preparation, 3 2.5 Selection of Laboratories for ILS, 3 2.6 Specimen Shipment, 4 2.7 Instructions for Interlabora- tory Study, 4 Chapter 3—Interlaboratory Study Test Results and Analysis, 5 3.1 Test Properties, 5 3.2 Number of Data Sets, 5 3.3 Results of the ILS, 5 3.4 Bulk Specific Gravity Results, 5 3.5 Deformation Versus Number of Passes, 8 3.6 Deformation Versus Measure- ment Location, 8 3.7 Difference in Deformation from Right and Left, 10 3.8 Difference in Laboratory Results, 10 3.9 Percent Error in Measurement Location Data, 12 3.10 Comparison of Properties of Various Mixture/Specimen Types, 13 Chapter 4—Precision Estimates, 22 4.1 Method of Analysis of ILS Test Results, 22 4.2 Statistical Comparisons, 22 4.3 Results of Analysis, 23 Chapter 5—Findings and Proposed Changes to AASHTO T 324 and the HWTT Equipment, 46 5.1 Findings, 46 5.2 Proposed Changes to AASHTO T 324 and the HWTT Equipment, 47 References, 47 Appendixes A Through H, 48 Appendix I: Recommended Precision Estimates for AASHTO T 324, 48 Precision Statement for AASHTO T 324, Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA), 48 Acknowledgments, 49

2a. Analyze data received from laboratories to determine variability of the HWTT measurements. b. Statistically compare variability of gyra- tory and slab specimens. c. Statistically compare variability of mea- surements from all measurement locations with those measured using (1) all except three middle measurement locations and (2) all except the two measurement loca- tions at each end. d. Determine which variances are not statisti- cally different and therefore can be pooled together. e. Prepare a precision statement for AASHTO T 324. 4. Conduct a research study to identify the causes of variability of the AASHTO T 324 test results. 5. Identify measures for improving accuracy and precision of the test results. 6. Prepare findings and proposed changes to AASHTO T 324 and the HWTT device based on the research results. CHAPTER 2—DESIGN AND CONDUCT OF THE STUDY The availability of precision estimates for AASHTO T 324 test method is essential for reliable laboratory determination of the rutting and mois- ture susceptibility of asphalt mixtures. In addition, aspects of the test method not yet standardized could be sources of variability. These sources need to be identified and further specified in the test method. An interlaboratory study (ILS) was designed and conducted in which variability of the test for two different types of mixtures and two methods of compaction were examined. The following sections present the details of the ILS. 2.1 Materials Selection Given that determining the level of rutting and moisture susceptibility of HMA is a main aspect of AASHTO T 324, two mixtures with different levels of rutting and moisture susceptibility were selected for the study. The rutting- and moisture-sensitive (WY) mixture, which was mixed and compacted in laboratory, consisted of 9.5 mm nominal maximum aggregate size (NMAS) gravel stones from Wyo- ming and PG 64-22 asphalt binder. The rutting- and the factors causing variability of measurements and further specify their limits in the test method. 1.3 Research Objectives The objective of this study was to determine pre- cision estimates for AASHTO T 324. To accomplish this objective, the research • Determined the variability of (1) the deforma- tion measurements after specified number of load passes and (2) the creep slope for well- performing mixtures. • Determined the variability of (1) the number of passes to threshold deformation, (2) creep slope, (3) stripping slope, and (4) number of passes to the stripping inflection point for poorly performing mixtures. • Compared the mean and variance of the measured properties of gyratory and slab specimens. • Compared the mean and variance of proper- ties measured using all measurement loca- tions with those measured using (1) all except the three middle measurement loca- tions and (2) all except two measurement locations at each end. • Identified causes of variability of the test results. • Proposed modifications to the test method for (1) optimum use of the deformation mea- surements, (2) improvement to the specimen preparation and assembly, and (3) necessary adjustments to the machine components. 1.4 Scope of Study The project encompassed the following major steps: 1. Select materials and mixture design for the interlaboratory study (ILS). 2. Design and conduct the ILS: a. Prepare instructions for preparing and test- ing the ILS specimens. b. Identify the laboratories participating in the ILS. c. Prepare gyratory and slab asphalt mixture samples. d. Provide the compacted samples and in- structions to the participating laboratories. 3. Develop precision estimates of AASHTO T 324:

3of gyratory specimens, Location 6 should be at the joint where the two adjoining samples abut. The wheel makes 52 ± 2 passes across the specimen per minute. The maximum speed of the wheel (0.305 m/s) is reached at the midpoint of the specimen. 2.4 Specimen Preparation Preliminary work was conducted to determine the appropriate weight of the mixtures for compact- ing gyratory and slab specimens with 7.0% ± 1.0% air voids based on the original job mix formulas. The gyratory samples were prepared using an IPC gyratory compactor (Servopac) following AAS- HTO T 312 (2). The slabs were compacted using a PMW linear kneading slab compactor. WY samples were mixed at 165°C and subsequently conditioned at 135°C for 4 hours according to AASHTO R 30 (3) before compaction. Field samples were reheated to 135°C before compaction. All samples were com- pacted to the height of 60 mm. A total of 280 gyratory and 60 slab specimens were compacted for shipment to the participating laboratories. Given that the percent water absorp- tion of aggregates of both Field and WY mixtures was less than 1.5%, the maximum specific gravities (Gmm) of both mixtures were determined accord- ing to the weighing-in-water method (Method A) described in Section 9 of AASHTO T 209 (4). The Gmm of the Field and WY mixtures are provided in Table 2-1. The bulk specific gravity of the samples was measured according to AASHTO T 166 (SSD) (5) and AASHTO T 331 (Corelok) (6) before send- ing the specimens to the participating laboratories. The average absorption of Field samples was 1.49% and of WY samples was 1.89%. Given that water absorption of the compacted samples was less than 2%, the target air voids of 7.0% ± 1.0% was achieved based on the AASHTO T 166 procedure. The sam- ples were dried using CoreDry® to a constant weight before they were packaged for shipment. 2.5 Selection of Laboratories for ILS State DOT and industry laboratories operating the HWTT device on a regular basis were contacted to participate in the study. All participating laboratories were AASHTO accredited for test methods related to AASHTO T 324. Thirty-five laboratories agreed to participate in the ILS. Twenty-eight laboratories moisture-resistant (Field) mixture was produced at the Aggregate Industries plant in Maryland and consisted of 19.0 mm NMAS limestone aggregates and PG 64-22 asphalt binder. Table 2-1 provides the aggregate gradation and asphalt content of the two mixtures. 2.2 Test Samples Given that AASHTO T 324 allows testing of both slab and gyratory-compacted specimens, the effect of specimen type on the test results was also inves- tigated. For this purpose, both 150-mm × 60-mm Superpave gyratory specimens and 265.5- × 331- × 60-mm slab specimens were prepared for the study. 2.3 Test Machine The wheel track testing machines included in the ILS were either one-wheel or two-wheel Hamburg Wheel Track Testers manufactured by Precision Metal Works (PMW). Linear variable displace- ment transducers (LVDTs) measure deformation at 11 locations referred to as measurement locations along the specimen. Location 1 is the furthest from the wheel gear and Location 11 is the closest to the wheel gear as shown in Figure 2-1. Location 6 is at the midpoint of the test specimen by design. In case Table 2-1 Volumetric properties of Wyoming laboratory and Maryland field mixtures. Sieve Opening (mm) US Sieve Size % Passing Maryland (Field) % Passing Wyoming (WY) 25 1" 100 100 19 ¾" 98 100 12.5 ½" 87 97 9.5 ³⁄8" 74 87 4.75 #4 37 51 2.36 #8 27 35 1.18 #16 20 25 0.60 #30 15 17 0.30 #50 10 13 0.15 #100 7 9 0.075 #200 5.1 6.2 Aggregate Water Absorption 0.8 0.6 Pb, % 4.5 4.4 Gmm 2.510 2.459

4the data. Given that preparation of gyratory and slab specimens is different, different sets of instructions were prepared for the two types of specimens. The preparation of gyratory specimens by the laboratories included cutting across the height of the specimens so that when the two cut specimens were adjoined, there would be a gap of no more than 7.5 mm between the two polyethylene molds holding the specimens in place (Figure 2-1). The laboratories were also asked to measure the air voids of the gyratory specimens before preparing them for the wheel track test. The slab specimens were surrounded by plaster of Paris to form their holder. Air voids measurements were not requested for the slab specimens. To reduce the size of data files collected during testing, the laboratories were asked to follow these data sampling intervals: every 20th cycle for the first 1000 cycles, every 50th cycle for the second 4000 cycles, and every 100th for the remainder of the test (up to 20,000 cycles). In addition to the output data file, the laborato- ries were asked to report back (1) the rut depths at returned results on at least one specimen type (gyra- tory and slab). 2.6 Specimen Shipment Each laboratory received four gyratory and two slab specimens from each of the WY and Field mix- tures. Slab specimens were only sent to the 15 labo- ratories capable of testing slabs. The shipment of the two different mixture types was done at a 2-month interval to allow receipt of the results from the first set of materials before the second set of specimens were sent. The reason for sending the compacted samples, rather than raw materials, was to separate the variability in sample preparation from the vari- ability associated with the test configuration and test equipment. 2.7 Instructions for Interlaboratory Study Participants were provided with instructions and data sheets for performing the tests and collecting Wheel Stopping Position (Measurement Location 1) Wheel Starting Position (Measurement Location 11) Figure 2-1 Starting and stopping positions of HWTT wheel and the first and last measurement locations shown on the schematic of the HWTT mounting system from AASHTO T 324. (1)

5• Twenty-two laboratories sent complete sets of data on the properties of the gyratory- compacted WY mixture. • Eleven laboratories sent complete sets of data on the properties of the slab-compacted WY mixture. Table 3-1 and Figure 3-1 show the number of laboratories that provided results for each combina- tion of material and specimen type. Table 3-1 also shows the number of wheels (two or one) on the Hamburg wheel track tester in each participating laboratory. 3.3 Results of the ILS The results received from the participating labo- ratories include the measurements of the bulk specific gravity of gyratory specimens and HWTT properties of the gyratory and slab samples. These results are discussed in the following sections. 3.4 Bulk Specific Gravity Results The statistics of the air voids measured prior to shipment of samples and the air voids measured by participating laboratories for both WY and Field mixtures using SSD and Corelok are shown in Table 3-2. The average water absorption of the WY and Field mixtures were 1.89% and 1.49%, respectively, which were under 2%. Therefore, for both mixtures, the 7% ± 1% air voids speci- fied in AASHTO T 324 for the HWTT samples was achieved based on the SSD air voids. The measure- ment of air voids by AMRL was made 24 hrs after compaction; for the WY samples, they averaged 6.86% and ranged between 6.51% and 7.49%; for the Field samples, they averaged 6.94% and ranged between 6.48% and 7.52%. The air voids measured by participating laboratories averaged 6.44% for WY samples and ranged from 5.72% to 7.00%. For Field samples, the average air voids was 6.86%, ranging between 6.25% and 7.45%. Despite the dif- ference between average SSD values of AMRL and the participating laboratories for the WY samples, the Corelok values were similar (averaged 7.73% and 7.54%, respectively), which indicates that the difference in the SSD values may be due to the subjectivity in SSD determination. The distribu- tion of SSD air voids for both mixtures, measured by AMRL, is shown in Figure 3-2. pass counts of 5, 10, 15, and 20 thousands; (2) the creep slope; (3) the stripping slope; (4) the number of cycles to threshold deformation; and (5) the num- ber of passes to stripping inflection point. A copy of each set of instructions for preparing and test- ing gyratory and slab specimens and the data sheets for entering measurement results are provided in Appendix A, which is not published herein but is available on the TRB website where it can be found by searching for NCHRP Research Results Digest 390. CHAPTER 3—INTERLABORATORY STUDY TEST RESULTS AND ANALYSIS Before determining the precision estimates of the measurements from the results of the ILS, graphical comparisons of the averages and standard deviations of the AASHTO T 324 test properties for different mixture types, specimen types, wheel side, pass num- ber, deformation threshold level, and measurement locations were performed. The test properties, num- ber of data sets, and observed results are explained in the following sections. 3.1 Test Properties The following test properties were computed from the data received from the participating labo- ratories and compared for the two mixtures and the two specimen types: • Deformation (rut depth) at 5000, 10000, 15000, and 20000 wheel passes; • Number of wheel passes to 6 mm and 12 mm rut depth; • Creep slope; • Stripping slope; and • Pass number and deformation at the Stripping Inflection Point. 3.2 Number of Data Sets The following number of laboratories provided completed data sets for the four specimen types (two mixtures x two specimen types): • Nineteen laboratories sent complete sets of data on the properties of the gyratory-compacted Field mixture. • Seven laboratories sent complete sets of data on the properties of the slab-compacted Field mixture.

6Evaluation of the difference between the SSD and Corelok values in Table 3-2 might indicate the method that is more reliable for measuring the air voids of HWTT samples. For the WY samples, at 7% SSD air voids, Corelok air voids were 0.8% higher (7.8%) as measured at AMRL. The differ- ence was similar (0.9%) when measured by par- ticipating laboratories. The difference between Corelok and SSD air voids for Field samples con- ducted by participating laboratories was 1.1%. The Corelok air voids of the Field mixture were not measured at AMRL due to the press for time to send the samples within 48 hrs after the compac- tion. Figure 3-3 shows the Corelok and SSD air Table 3-1 Mixture/specimen type associated with the results sent and the corresponding number of HWTT wheels for each participating laboratories. Laboratories No. of Wheels Field- Gyratory Field- Slab WY- Gyratory WY- Slab Alliance Geotechnical Group 1 ü ü ü ü AMEC Earth & Environmental 1 ü ü APAC TX, Inc. 1 ü ü California DOT, Sacramento, CA 2 ü ü ü ü Colorado DOT, Denver, CO 2 ü ü Florida DOT, Gainesville, FL 2 ü ü ü Iowa DOT, Ames, IA 2 ü Jones Bros. Dirt & Paving Contractors, Inc. 1 ü ü Kansas State University— Manhattan 2 ü ü Louisiana State University 2 ü ü Mathy Technology & Engineering Services 2 ü ü Nactech 2 ü ü ü Oklahoma DOT—Oklahoma City 2 ü ü Pave Tex 2 ü ü Road Science, LLC 2 ü ü ü ü Texas A&M University 2 ü Texas DOT—Childress District 2 ü Texas DOT—Paris 1 ü Texas DOT—San Marcos 2 ü Texas DOT—Uvalde Field Lab 1 ü U. of Massachusetts—Dartmouth 1 ü ü University of Texas—Austin 2 ü University of Texas—El Paso 2 ü Utah DOT—Salt Lake City 2 ü ü Utah DOT—Ogden Lab 2 ü Vulcan Materials Co. 1 ü ü ü Washington State DOT, Pullman 2 ü ü Wyoming DOT—Cheyenne 2 ü 0 5 10 15 20 25 Gyratory Slab Gyratory Slab Field 19 mmWY Figure 3-1 Number of laboratories that provided results.

7Table 3-2 Air voids of Field and WY samples measured by AMRL and participating laboratories. Mixture Lab Test Average STD Min Max N WY AMRL SSD 6.86 0.24 6.51 7.49 51 Corelok 7.73 0.17 7.42 8.15 51 Participating Labs SSD 6.44 0.32 5.72 7.00 62 Corelok 7.54 0.36 6.95 8.32 22 Field AMRL SSD 6.94 0.19 6.48 7.52 95 Corelok - - - - - Participating Labs SSD 6.86 0.27 6.25 7.45 63 Corelok 8.05 0.37 7.60 8.70 19 0 2 4 6 8 10 12 6.4 6.5 6.6 6.7 6.8 6.9 7 7.1 7.2 7.3 7.4 to to to to to to to to to to to 6.49 6.59 6.69 6.79 6.89 6.99 7.09 7.19 7.29 7.39 7.49 N o. o f S am pl es SSD Air Voids, % 0 5 10 15 20 25 30 6.4 6.5 6.6 6.7 6.8 6.9 7 7.1 7.2 7.3 7.4 to to to to to to to to to to to 6.49 6.59 6.69 6.79 6.89 6.99 7.09 7.19 7.29 7.39 7.49 N o. o f S am pl es SSD Air Voids, % Figure 3-2 Distribution of SSD air voids of WY samples (top) and Field samples (bottom) measured at AMRL.

8website). Some general observations can be made from the graphs: 1. The Field mixture has a small deformation versus number of passes (low creep slope). 2. Other than two outlier results, the Field mixture does not exhibit an inflection point. Loosening of the bolts holding specimens in test trays was reported by the laboratories as the reason for the outlier data. 3. The WY mixture clearly shows a stripping inflection point. 4. The inflection point of the WY mixture occurs after a greater number of passes in the slab specimens than in the gyratory specimens. 5. In each mixture, slab and gyratory specimens show similar trends, but the deformation curves of slabs seem less noisy than those of gyratory specimens. 6. For the WY mixture, the stripping slopes (2nd slope) are generally larger in gyratory specimens than in slab specimens. 3.6 Deformation Versus Measurement Location Figure 3-6 shows the deformation profile from the last wheel pass at Location 11 of the HWTT for the four mixture/specimen combinations. The x-axis shows the measurement locations and the y-axis shows the deformation measurements in mm. The top and bottom graphs for each combination show voids from measurements made at AMRL (only WY mixture) and Figure 3-4 shows the Corelok and SSD air voids from measurements made by participating laboratories (both WY and Field mix- tures). As indicated from the figures, the SSD and Corelok air voids are distinctly different for both mixtures. Considering the level of absorption of 1.89% and 1.49% of WY and Field mixtures, it is suggested that bulk specific gravity of samples with absorption level of above 1.0% measured using the Corelok method. The data shown in Figure 3-3 include air voids of samples prepared for the study but either not sent to the participating laboratories or sent but not tested by any laboratories. Examples of these samples are those with SSD air void val- ues between 6.2% and 6.5% in Figure 3-3. On the other hand, Figure 3-4 includes only air voids of samples measured by both the SSD and Core- lok methods. Not all laboratories measured bulk specific gravity of the samples according to both SSD and Corelok; therefore, fewer number of data points than the number of sent samples are included in Figure 3-4. 3.5 Deformation Versus Number of Passes Graphs of average deformation versus number of passes for the four material/specimen combina- tions from all laboratories are provided in Figure 3-5. Graphs of the individual tests are provided in Appen- dix B (not published herein but available on the TRB 6 6.5 7 7.5 8 8.5 6 6.2 6.4 6.6 6.8 7 7.2 7.4 Ai r V oi ds , % SSD Air Voids, % SSD Corelok Figure 3-3 Air voids of WY samples using SSD and Corelok measured at AMRL.

9the measurements from the right and left wheels in a two-wheel machine or replicate measurements in a one-wheel machine. Several observations can be made from the profiles: 1. For the well-performing Field mixture, the deformation profiles of gyratory and slab specimens appear similar. 2. For the poorly performing WY mixture, as indicated from the deformation profiles, the deformations from different measurement locations are more consistent for the slab specimens than for the gyratory specimens. 3. The maximum deformations for WY gyra- tory specimens mostly occur at Locations 7 and 8, rather than Location 6, which is the midpoint. 4. For the WY gyratory specimens, a maximum deformation typically occurs at or around the midpoint of the specimen (Locations 6, 7, or 8). However, for the slabs only a few pro- files show a maximum deformation around the center. This might indicate that the mid- point of gyratory specimens, where the two samples join, is the weakest part of the test specimen. 6 6.5 7 7.5 8 8.5 6 6.2 6.4 6.6 6.8 7 7.2 7.4 A ir V oi ds , % SSD Air Voids, % SSD Corelok 6 6.5 7 7.5 8 8.5 6 6.2 6.4 6.6 6.8 7 7.2 7.4 A ir V oi ds , % SSD Air Voids, % SSD Corelok Figure 3-4 Air voids of WY samples (top) and Field samples (bottom) using SSD and Corelok measured at participating laboratories.

10 the mixtures from individual laboratories in Appen- dix B (which is available on the TRB website). 3.8 Difference in Laboratory Results Close examination of the deformation history (deformation versus number of passes) and defor- mation profiles (deformation versus measurement location) presented in the previous sections found that the results could be grouped into two catego- ries: (1) a group of laboratories with very similar deformation profiles to each other and (2) a group of laboratories with different deformation profiles from each other and from those in the first group. Figure 3-8 shows the deformation measurements of gyratory specimens of the Field mixture. The left graph shows the deformation measurements from the laboratories with similar results and the right graph shows the deformation measurements from the labo- 3.7 Difference in Deformation from Right and Left The top and bottom of Figure 3-7 show the mea- surement locations versus deformation (deformation profile) and number of passes versus deformation (deformation history) for the WY gyratory mixture reported by one of the laboratories. As indicated from the deformation profile (top), the magnitudes of the maximum deformations of the left and right wheels are the same; however, the maximum deformation occurred at Location 6 for the right wheel and Loca- tion 9 for the left wheel. This shows that either rep- licate samples do not always wear similarly or the measurement locations are not the same on the two sides of the machine. The deformation history from Location 7, shown at the bottom of the figure, indi- cates that the deformations from right and left wheels are very different. Similar problems can be observed from deformation profiles and deformation history of Figure 3-5 Deformation (mm) versus number of passes for (a) Field gyratory, (b) Field slabs, (c) WY gyratory, and (d) WY slabs received from laboratories.

a b c d Figure 3-6 Deformation profiles of (a) Field gyratory, (b) Field slabs, (c) WY gyratory, and (d) WY slabs received from laboratories. Figure 3-7 Deformation versus measurement locations and versus number of passes for the gyratory specimens of WY reported by Laboratory R.

12 ratories with different results from each other and from those in the first group. The large spread in the defor- mation measurements of the laboratories in the second group suggests problems with either the calibration or alignment of the HWTT device or the specimen-mold assembly in those laboratories. This finding empha- sizes the need for regular calibration checks of the machines and standardization of the specimen-mold assembly to reduce variability of the data. 3.9 Percent Error in Measurement Location Data Figure 3-9 shows the % error in deformation signals caused by electrical and mechanical inter- ferences (noise) in HWTT, determined from labo- ratories’ data. The percent error is the same as coefficient of variation, which is standard devia- tion of signal amplitude divided by the mean sig- nal amplitude, times 100. The percent error is the reciprocal of Signal to Noise Ratio (SNR), which describes how much noise is in the output of a device, in relation to the signal level. To evaluate the quality of the HWTT data, a threshold % error needed to be established. From the analysis of the data, it was experienced that when percent error is less than 5%, the least amount of filtering and averaging was required for deter- mining the properties of the test. In addition, several literatures show that a typical SNR threshold for an acceptable signal quality is 20 (7, 8, 9), which is equivalent to 5% signal error (inverse of 20). There- fore, a threshold value of 5% was selected for evalu- ating the quality of the signal data. The graphs in Figure 3-9 represent the average percent error from readings of Locations 4 through 8 of Passes 5,000 through 10,000 of the four mixture/ specimen types. As indicated in the figure, the per- cent error is as small as 1% in one laboratory and as large as 25% in another laboratory. Considering the acceptable percent error of 5%, this threshold has been exceeded in more than 30% of the laboratories, especially for the WY mixture. The percent error in deformation signals could be a major source of measurement variability. When the noise level is low, the parameter of the test could be Figure 3-8 Deformation profile and deformation history of the gyratory specimens of Field mixture; the left graph shows the laboratories with similar results and the right graph shows laboratories with different results from each other and from laboratories in the left graph.

13 easily determined without major manipulation of the signal data. However, if the noise in the data is high, significant smoothing and averaging are required to determine the value of the parameters. This would result in estimated value of the property that is dif- ferent from the actual value, therefore, causing high variability of the measured properties especially when measured in different laboratories. Reducing the % error in the signal data is another step in reducing variability of the measurements. Figure 3-10 shows the data from Laboratory F. Although the deforma- tion profiles and history of the right and left wheels are very similar, the percent error of the deformation signals from the two wheels is very high. 3.10 Comparison of Properties of Various Mixture/Specimen Types The deformation curves in Figure 3-6 dem- onstrate that the preferred HWTT measurement parameters for the well performing and the poorly performing mixtures are likely to be different. For well performing mixtures, where the test could be continued for the specified number of passes, the deformation at those passes is a meaningful test parameter as is the slope of the deformation curve before the end of the test, also known as creep slope. For the poorly performing mixtures, where defor- mation is large and the duration of the test is ulti- mately limited by the degree of deformation, the number of cycles to a specified threshold deforma- tion is a meaningful test parameter. Additionally, given that poorly performing mixtures have a clear inflection point, the slope of the deformation curve before and after the inflection point (the creep and striping slopes) and the number of cycles to the inflection point are also useful test parameters. The choice of test parameters for a given mix- ture is not made a priori, but is based on the observed performance of the mixture in the HWTT. 3.10.1 Comparison of Properties of Gyratory and Slab Specimens of Field Mixture The properties of the well performing mixture include creep slope, deformation at specified num- ber of passes, and deformation at the end of the test. 0 2 4 6 8 10 12 14 16 A B C D E F G H I J K L M N O P Q R % Er ro r Laboratory ID Field-Gyratory Left Right 0 2 4 6 8 10 12 B C D E F G H I L M N O P Q R S T U V % Er ro r Laboratory ID WY-Gyratory Left Right 0 1 2 3 4 5 6 7 8 9 A B C D E F G H % Er ro r Laboratory ID WY-Slab Left Right 0 5 10 15 20 25 30 A B C D E G H % Er ro r Laboratory ID Field-Slab Left Right Figure 3-9 % error in sensor data corresponding to the deformation measurements of the four material/specimen types.

14 The comparison of the properties of the gyratory and slab specimens is explained as follows. 3.10.1.1 Creep Slopes of Gyratory and Slab. Fig- ure 3-11 shows the average and standard devia- tion of the creep slope for the gyratory and slab specimens of the well performing Field mixture. For this mixture, the creep slope represents the rate of deformation before the end of the test. As indicated from the figure, the average and standard deviation of creep slope of gyratory specimens is only slightly smaller than those of slab specimens. This suggests that for well performing mixtures, gyratory specimens may provide a better estimate of rutting performance of the mixture than slab specimens. 3.10.1.2 Deformation of Gyratory and Slab Speci- mens at End. Figure 3-12 shows the average and standard deviation of deformation of the Field mixture at the end of the test. The criteria for the test termination are either 20,000 passes or 25 mm of deformation, whichever comes first. For the well performing mixture, which experienced a small deformation, tests were ended after 20,000 passes. As indicated from the figure, the deformation of the gyratory specimens is an average 0.4 mm less than the deformation of slab specimens at the end of the test. This also indicates that gyratory speci- mens may provide a better estimate of rutting per- formance of well performing mixtures than slab specimens. 3.10.1.3 Deformation of Gyratory and Slab Speci- mens after Specified Number of Passes. Figures 3-13 and 3-14 show the average and standard deviation of deformation for the gyratory and slab specimens of the well-performing Field mixture after 1000, 2000, 5000, 10,000, and 20,000 passes. The graph shows that after each set of passes, slab specimens have experienced slightly more deformation than the gyratory specimens. The standard deviations of the deformation of the slab specimens are shown to be larger than those of gyratory specimens after 5,000 passes. This indicates that for the well per- forming mixtures, gyratory specimens are slightly more resistant to rutting and moisture and provide slightly less variable results than slab specimens. Figure 3-10 Deformation profiles and deformation history for gyratory specimens of WY, Laboratory F.

15 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 Gyratory Slab Cr ee p S lo pe , m m /p as s Specimen Type Average of Slope1 StdDev of Slope1_2 Figure 3-11 Comparison of average and standard deviation of creep slopes of gyratory and slab specimens of well-performing mixture. 0 0.5 1 1.5 2 2.5 3 3.5 4 Gyratory Slab D ef or m ati on , m m Specimen Type Average of Def_End StdDev of Def_End2 Figure 3-12 Average deformation of gyratory and slab specimens of the well-performing mixture at the end of the test.

16 0 0.5 1 1.5 2 2.5 3 3.5 4 Gyratory Slab Field 19 mm D ef or m ati on , m m Specimen Shape/Material Average of Def_1000 Average of Def_2000 Average of Def_5000 Average of Def_10000 Average of Def_End Figure 3-13 Average deformation of the Field mixture after various number of passes. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Gyratory Slab Field 19 mm D ef or m ati on , m m Specimen Shape/Material StdDev of Def_1000 StdDev of Def_2000 StdDev of Def_5000 StdDev of Def_10000 StdDev of Def_End Figure 3-14 Standard deviation of deformation of Field mixture after various number of passes.

17 3.10.2 Comparison of Properties of Gyratory and Slab Specimens of Wyoming Mixture Test properties for the poorly performing WY mixture include number of passes to threshold rut depth, creep and stripping slopes, and inflection point. Different state DOTs specify different rut depth thresholds to define test failure. The more commonly used failure criteria are 6-mm and 12-mm rut depths. Herein, the number of passes to these two failure criteria was compared for the gyratory and slab specimens of the WY mixture. 3.10.2.1 Creep Slopes of Gyratory and Slab of Wyo- ming Mixture. Figure 3-15 shows the average and standard deviation of the creep slope for the gyra- tory and slab specimens of the WY mixture. The creep slopes represent the rate of deformation before the inflection point. As indicated from the figure, for the WY mixture, the average and standard deviation of the creep slope of gyratory specimens is larger than those of slab specimens. The fact that gyratory specimens are less resistant to rutting and moisture damage might indicate that the rate of deformation of the poorly performing mixture is underestimated using gyratory specimens. 3.10.2.2 Number of Passes to 6-mm Deformation. Figure 3-16 shows the average and standard devia- tion of the number of passes to 6-mm rut depth for gyratory and slab specimens of WY mixture. A greater number of passes was needed to achieve the same amount of deformation in the slab than in gyratory specimens (12,000 versus 7,000 passes). Although the standard deviation of the number of passes is larger for the slab specimens, considering the larger number of passes, the coefficient of varia- tion for the slab specimens would be smaller. This shows that a poorly performing mixture is more vul- nerable to rutting and moisture damage when tested in the form of gyratory specimens than slab speci- mens. The weaker performance of gyratory speci- mens of the poorly performing mixture is speculated to be caused by the cut cross-sections of the jointed gyratory specimens. 3.10.2.3 Number of Passes to 12-mm Deformation. Figure 3-17 shows the average and standard devia- tion of the number of passes to 12-mm rut depth for the WY specimens. Similar to the observation above, a greater number of passes was needed to achieve 12-mm rut depth in slabs than in gyratory specimens (17,000 versus 10,000), indicating more vulnerability of gyratory specimens to rutting and moisture damage. The standard deviation and con- sequently the coefficient of variation of number of passes to 12-mm deformation is smaller for slab specimens than for the gyratory specimens. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Gyratory Slab Cr ee p sl op e, m m /p as s Specimen Type WY - Average of Slope1 WY - StdDev of Slope1_2 Figure 3-15 Comparison of average and standard deviation of creep slopes of gyratory and slab specimens of the poorly performing mixture.

18 3.10.2.4 Number of Passes to Inflection Point. Fig- ure 3-18 shows the average and standard deviation of the number of passes to the inflection point for the gyratory and slab specimens of the WY mixture. The graph indicates that the gyratory specimens exhibit an inflection point around 4000 passes while the slab specimens exhibit an inflection point around 7000 passes. The variability of this parameter for gyratory and slab specimens is comparable consid- ering that the higher number of passes were required to develop the inflection point in the slab specimens. These results also indicate that for poorly perform- ing mixtures, gyratory specimens are more vulner- able to rutting and moisture damage than the slab specimens, probably due to the cut cross-sections of the jointed samples. 3.10.2.5 Deformation at Inflection Point. Fig- ure 3-19 provides the average and standard deviation of deformation at the inflection points of the WY spec- imens. As indicated from the figure, although the inflec- tion point occurs after a different number of passes for gyratory and slab specimens, as was shown in the previous section, the average deformations at the inflection point are not very different (around 2.5 mm) for the two specimen types. This might indicate that slope of the deformation curve before the inflec- tion point (creep slope) is a better test parameter than 0 2000 4000 6000 8000 10000 12000 14000 Gyratory Slab # of P as se s Specimen Type WY - Average of Pass_Rut_6 WY - StdDev of Pass_Rut_6_2 Figure 3-16 Comparison of the number of passes to 6-mm deformation. 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 Gyratory Slab # of P as se s Specimen Type Average of Pass_Rut_12 StdDev of Pass_Rut_12 Figure 3-17 Comparison of the number of passes to 12-mm deformation in WY mixture.

19 deformation and number of passes because creep slope explains how fast mixtures reach the same level of deformation. 3.10.2.6 Stripping Slopes of Gyratory and Slab of Wyoming Mixture. Figure 3-20 shows the average and standard deviation of the stripping slopes for the gyratory and slab specimens of WY mixture. The stripping slopes represent the rate of deformation after the inflection point. As shown in the figure, the aver- age and standard deviation of the stripping slope of gyratory specimens is larger than that of slab speci- mens, indicating a faster degradation of the gyratory specimens of the poorly performing mixture after the inflection point. 3.10.3 Measurement Locations of Maximum Deformation Figure 3-21 shows the distribution of the maxi- mum deformation at the measurement locations from all laboratories. As indicated from the figure, for the gyratory specimens, maximum deformation occurs most frequently at Locations 7 and 8; while for the slab specimens, frequency of maximum deformation is relatively equal at all measurement locations. This clearly shows that despite the maxi- mum speed of the wheel at the midpoint, maximum deformation for gyratory specimens occur most fre- quently at or around the midpoint due to the weak- ness at the joint. 0 1000 2000 3000 4000 5000 6000 7000 8000 WY WY Gyratory Slab # of P as se s Mixture/Specimen Type Average of Pass_Inf StdDev of Pass_Inf Figure 3-18 Number of passes to the inflection point for the WY mixture. 0 0.5 1 1.5 2 2.5 3 WY WY Gyratory Slab D ef or m ati on a t In fl ec ti on P oi nt Mixture/Specimen Type Average of Def_Inf StdDev of Def_Inf Figure 3-19 Average deformation at the inflection point.

20 Another observation from Figure 3-21 is that the most frequent readings of maximum deformation occur at Locations 7 and 8 and not at Location 6, which is the midpoint. This indicates that there is a possibility that the positions of the measurement locations (and therefore the spacing between mea- surement locations) are not consistent among dif- ferent machines. An in-house investigation into this matter was conducted and the results are discussed in Appendix C (not published herein but available on the TRB website). 3.10.4 Effect of Left and Right Wheels on Replicates’ Variability Figures 3-22 and 3-23 show average and stan- dard deviation of rut depth from one-wheel and 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Gyratory Slab St rip pi ng S lo pe , m m /p as s Specimen Type WY - Average of Slope2 WY - StdDev of Slope2 Figure 3-20 Comparison of average and standard deviation of stripping slopes of the gyratory and slab of poorly performing mixture. 0 5 10 15 20 25 30 1 2 3 4 5 6 7 8 9 10 11 Fr eq ue nc y Sensor Number Gyratory Slab Figure 3-21 Number of maximum deformation at each measurement location.

21 two-wheel HWTT machines for the WY gyratory specimens. Figure 3-22 shows that the two-wheel HWTT causes about a 10% greater average rut depth than one-wheel HWTT. Figure 3-23 indicates lower variability for two-wheel HWTT below 10,000 cycles and similar variability at 10,000 cycles; however, the two-wheel HWTT’s variability at the end of the test is twice as much as that from the one- wheel machine. This may be due to the dynamics of the wheels and the dynamic effect of one wheel on the other as the specimens’ rut depth significantly increases. This was usually after 10,000 passes for the WY mixture. Figure 3-24 shows the standard deviation of the rut depths for Field mixture specimens. Lower stan- dard deviations for two-wheel than for one-wheel machines are seen throughout the test. The dynamic effect is less evident from the Field mixture given Figure 3-22 Average impression of one-wheeler and two-wheeler HWTT for WY gyratory specimens. Figure 3-23 Standard deviation of one-wheel and two-wheel HWTT for WY gyratory specimens.

22 that this material does not rut significantly, even after 10,000 cycles. The two-wheel system may produce more pre- cise replicate measurements for well performing mixtures with low rut depths; however, the variabil- ity between replicates increases significantly with the increased rut depth of the specimens, probably due to the dynamic effect of one wheel on another. If this hypothesis is true, then having separate mechan- ical systems for each wheel may be warranted. CHAPTER 4—PRECISION ESTIMATES 4.1 Method of Analysis of ILS Test Results The ILS test results were analyzed for precision in accordance with ASTM E691, “Standard Practice for Conducting an Interlaboratory Study to Deter- mine the Precision of a Test Method” (10). Prior to the analysis, partial sets of data were eliminated by following the procedures described in E691 for determining repeatability (Sr) and reproducibility (SR) estimates of precision. Data exceeding the critical h and k statistics, representing the threshold values for the within- and between-laboratory vari- ability, were eliminated from the analysis. The h and k statistics are provided in Appendixes D through H (not published herein but available on the TRB website). The measured data and the computed sta- tistics for each mixture and specimen type are also provided in the tables and displayed in the figures of Appendixes D through H. The shaded cells in the tables indicate data eliminated from the analysis because they exceeded the critical h and k statis- tics. The graphical display of the data received from laboratories and their associated error bars are pro- vided in the appendixes. For each replicate data set, the bottom bar represents the minimum value, the top bar represents the maximum value, and middle point represents the median. The spacing between the median and the top and bottom values indicate the degree of dispersion. This is a useful technique for summarizing the data and determining how vari- able the data are in each laboratory and among vari- ous laboratories. 4.2 Statistical Comparisons The measurements according to AASHTO T 324 were collected at 11 measurement locations on two different specimen types, gyratory and slab, of well performing and poorly performing asphalt mix- tures. The analysis of the measured data was con- ducted with respect to different sets of measurement locations and specimen types. To prepare precision estimates of the properties, variability correspond- ing to the various measurement locations, specimen types, number of passes to various threshold rut depth criteria, and the rut depths after various num- bers of wheel passes were compared statistically. Figure 3-24 Standard deviation of one-wheel and two-wheel HWTT for Field gyratory specimen.

23 Those variability values that were not statistically significantly different were pooled to prepare the precision estimates. Statistical t- and F-tests were used to examine the significance of the following differences: 1. Difference between statistics of gyratory and slab specimens 2. Difference between statistics calculated from all measurement locations, all except the mid- dle three measurement locations, and all except two measurement locations at each end 3. Difference between variability of rut depth after 10,000, 15,000, and 20,000 passes (for well performing mixture) 4. Difference between variability of number of passes to 6-mm and 12-mm rut depth and to the inflection point (for poorly performing mixture) The rejection probability of the computed t- and F-statistics would indicate if the differences from the above comparisons are significantly different. For a 5% level of significance, a rejection probabil- ity (p) of less than 0.05 is an indication of significant difference. In the preparation of the precision esti- mates, those standard deviations that are not signifi- cantly different (P > 0.05) would be pooled together. Given that the parameters of the wheel track test are different for the well and poorly performing mix- tures, separate analyses were conducted for the well performing Field mixture and the poorly performing WY mixture. For the well performing mixture, the parameters of the test are deformation after 10,000, 15,000, and 20,000 passes and the creep slope. For the poorly performing mixture, the parameters of the test are number of passes to either 6-mm or 12-mm deformation, the creep and stripping slopes, and the number of passes to the inflection point. 4.3 Results of Analysis 4.3.1 Well Performing Field Mixture Table 4-1 provides the statistics of the rut depth after 10,000, 15,000, and 20,000 passes and the sta- tistics of the creep slope for the gyratory and slab specimens of the well performing Field mixture. The statistics are calculated using data from all measurement locations, all except the three middle measurement locations (Locations 5, 6, and 7), and all except the two measurement locations at each end (Locations 1 and 2, and 10 and 11). The statisti- cal tests were conducted to compare the averages and variability of the properties measured: (1) from different sets of measurement locations and (2) mea- sured on gyratory and slab specimens. 4.3.1.1 Comparison of Statistics from Various Mea- surement Locations. A review of the statistics in Table 4-1 indicates relationships between the aver- ages and standard deviations. Therefore, comparison of variability is based on the coefficient of variation (COV). Figures 4-1 and 4-2 show the averages and COV of the measurements from various measure- ment locations. Table 4-2 through Table 4-4 provide the results of statistical comparison of the averages and the repeatability/reproducibility COVs of the properties measured using different sets of measure- ment locations. In the figures and tables, the com- parisons corresponding to the gyratory specimens come first followed by the comparisons correspond- ing to the slab specimens. The observations are as follows: 1. For the gyratory specimens, excluding the readings from the three middle measurement locations resulted in slight, but not statistically significant, decreases in average rut depth and creep slope. This is because the deformations at the locations of the middle measurement locations are larger than those at other loca- tions. There is no trend of change in repeat- ability COV; however, there is increase in reproducibility COV of the properties from excluding the readings of the middle three measurement locations. No differences are statistically significant. 2. For the gyratory specimens, excluding the readings of the end measurement locations resulted in slight, but not statistically signifi- cant, increases in average rut depths and creep slope. This is because the deformations at the location of the end measurements are smaller than the deformations at other measurement locations. There is an increase in repeatability and a decrease in reproducibility COV of the properties from excluding the readings of the end measurement locations; however, none of the differences are statistically significant. 3. For the slab specimens, excluding the read- ings from the three middle measurement loca- tions resulted in slight, but not statistically significant, increases in average creep slope and average rut depth after 10,000, 15,000, or

24 Table 4-1 Summary of statistics of rut depth (mm) and creep slope (mm/pass) of gyratory and slab specimens of Field material from average of all measurement locations, average of all except middle three measurement locations, and average of all except two measurement locations at each end. # of Labs Repeatability Reproducibility Condition Property Average STD CV% STD CV% Sx Field gyratory (all measurement locations) Rut after 10,000 cycles 18 2.26 0.275 12.2 0.594 26.3 0.561 Rut after 15,000 cycles 18 2.53 0.334 13.2 0.665 26.3 0.621 Rut after 20,000 cycles 18 2.71 0.386 14.2 0.729 26.9 0.676 Creep Slope 18 0.089 0.014 15.8 0.023 25.7 0.021 Field gyratory (except middle measurement locations) Rut after 10,000 cycles 19 2.22 0.309 13.9 0.616 27.7 0.575 Rut after 15,000 cycles 18 2.46 0.318 12.9 0.677 27.6 0.639 Rut after 20,000 cycles 18 2.63 0.360 13.7 0.739 28.1 0.694 Creep Slope 18 0.086 0.013 15.7 0.023 27.3 0.021 Field gyratory (except end measurement locations) Rut after 10,000 cycles 18 2.36 0.328 13.9 0.601 25.5 0.554 Rut after 15,000 cycles 18 2.65 0.392 14.8 0.669 25.3 0.609 Rut after 20,000 cycles 18 2.85 0.459 16.1 0.744 26.1 0.669 Creep Slope 18 0.095 0.017 18.0 0.024 25.5 0.021 Field slab (all measurement locations) Rut after 10,000 cycles 6 2.60 0.333 12.8 0.606 23.3 0.558 Rut after 15,000 cycles 6 2.99 0.443 14.8 0.762 25.5 0.694 Rut after 20,000 cycles 6 3.27 0.532 16.3 0.889 27.2 0.805 Creep Slope 6 0.112 0.029 26.4 0.039 34.8 0.033 Field slab (except middle measurement locations) Rut after 10,000 cycles 6 2.62 0.338 12.9 0.587 22.4 0.536 Rut after 15,000 cycles 6 3.00 0.443 14.8 0.735 24.5 0.665 Rut after 20,000 cycles 6 3.28 0.528 16.1 0.849 25.8 0.762 Creep Slope 6 0.113 0.029 25.6 0.037 32.6 0.031 Field slab (except end measurement locations) Rut after 10,000 cycles 6 2.56 0.312 12.2 0.613 24.0 0.573 Rut after 15,000 cycles 6 2.94 0.414 14.1 0.780 26.6 0.723 Rut after 20,000 cycles 6 3.23 0.517 16.0 0.924 28.6 0.848 Creep Slope 6 0.109 0.029 26.9 0.041 37.6 0.035

25 Figure 4-1 Graphical comparison of average properties of Field mixture measured using data from all measurement locations, all except middle three measurement locations, and all except two measurement locations at each end. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles A ve ra ge , m m Average Deformation, Gyratory Field gyratory (all sensors) Field gyratory (except middle sensors) Field gyratory (except end sensors) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles A ve ra ge , m m Average Deformation, Slab Field Slab (all sensors) Field slab (except middle sensors) Field slab (except end sensors) 0.080 0.082 0.084 0.086 0.088 0.090 0.092 0.094 0.096 Creep Slope A ve ra ge , m m /p as s Average Creep Slope, Gyratory Field gyratory (all sensors) Field gyratory (except middle sensors) Field gyratory (except end sensors) 0.107 0.108 0.109 0.110 0.111 0.112 0.113 0.114 Creep Slope A ve ra ge , m m /p as s Average Creep Slope, Slab Field Slab (all sensors) Field slab (except middle sensors) Field slab (except end sensors) Figure 4-2 Graphical comparison of coefficients of variation (COV) of properties of Field mixture measured using data from all measurement locations, all except middle three measurement locations, and all except two measurement locations at each end. 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Repeatability COV, Gyratory Field gyratory (all sensors) Field gyratory (except middle sensors) Field gyratory (except end sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Reproducibility COV, Gyratory Field gyratory (all sensors) Field gyratory (except middle sensors) Field gyratory (except end sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Repeatability COV, Slab Field Slab (all sensors) Field slab (except middle sensors) Field slab (except end sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Reproducibility COV, Slab Field Slab (all sensors) Field slab (except middle sensors) Field slab (except end sensors)

26 20,000 passes. This could be because defor- mation at and around the midpoint of the slab, where the speed of the wheel is the highest, is the smallest. There is no trend of change in the repeatability and a slight, but not signifi- cant, decrease in the reproducibility COV of the properties from excluding the data from the middle three measurement locations of the slab specimens (Figure 4-2). 4. For the slab specimens of the well perform- ing mixture, excluding the readings of the end measurement locations resulted in slight, but not statistically significant, decreases in average rut depths and average creep slope. This indicates that in the slabs, contrary to gyratory specimens, the deformations at the ends are slightly larger than the deformation at other locations. There is no trend of change in the repeatability; however, there is a slight increase in the reproducibility coefficients of variation. None of the differences are statisti- cally significant. From the above it can be concluded that all measurement locations are equally important for measurement of properties of either gyratory and slab specimens of well performing mixtures. Therefore, it is proposed that for well-performing mixtures, the readings from all measurement locations be aver- aged when analyzing the data from the HWTT. 4.3.1.2 Comparison of Statistics from Gyratory and Slab Specimens. Figures 4-3 and 4-4 compare the Table 4-2 Statistical t-test on the average rut depth (mm) after 10,000, 15,000, and 20,000 cycles and creep slope (mm/pass) of Field mixture for the comparison of measurements from various sets of measurement locations. Comparison Property Averages S T df Critical t P Decision Field gyratory (all measurement locations) vs. Field gyratory (except middle measurement locations) Rut after 10,000 cycles 2.26 vs. 2.22 0.568 0.17 35 1.69 0.435 Accept Rut after 15,000 cycles 2.53 vs. 2.46 0.630 0.34 34 1.69 0.367 Accept Rut after 20,000 cycles 2.71 vs. 2.63 0.685 0.35 34 1.69 0.363 Accept Creep Slope 0.089 vs. 0.086 0.021 0.50 34 1.69 0.311 Accept Field gyratory (all measurement locations) vs. Field gyratory (except end measurement locations) Rut after 10,000 cycles 2.26 vs. 2.36 0.558 -0.54 34 1.69 0.297 Accept Rut after 15,000 cycles 2.53 vs. 2.65 0.615 -0.58 34 1.69 0.284 Accept Rut after 20,000 cycles 2.71 vs. 2.85 0.672 -0.62 34 1.69 0.270 Accept Creep Slope 0.089 vs. 0.095 0.021 -0.81 34 1.69 0.211 Accept Field slab (all measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 2.6 vs. 2.62 0.547 -0.07 10 1.81 0.473 Accept Rut after 15,000 cycles 2.99 vs. 3 0.680 -0.03 10 1.81 0.490 Accept Rut after 20,000 cycles 3.27 vs. 3.28 0.784 -0.04 10 1.81 0.485 Accept Creep Slope 0.112 vs. 0.113 0.032 -0.08 10 1.81 0.468 Accept Field slab (all measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 2.6 vs. 2.56 0.565 0.13 10 1.81 0.451 Accept Rut after 15,000 cycles 2.99 vs. 2.94 0.709 0.13 10 1.81 0.451 Accept Rut after 20,000 cycles 3.27 vs. 3.23 0.827 0.07 10 1.81 0.473 Accept Creep Slope 0.112 vs. 0.109 0.034 0.13 10 1.81 0.451 Accept

27 averages and the COVs of the measurements from slab and gyratory specimens. Table 4-5 through 4-7 provide the results of statistical comparison of the averages and repeatability/reproducibility COVs of deformation and creep slope from gyratory and slab specimens. In the figures and tables, the first com- parison corresponds to all measurement locations, the second comparison corresponds to all except the middle three measurement locations, and the third comparison corresponds to all except the two mea- surement locations at each end. The following are observed from the tables: 1. Regardless of the sets of measurement loca- tions used, the average deformation and creep slope of the slab specimens of the well per- forming mixture are always larger than those of the gyratory specimens. This indicates that gyratory specimens of well performing mix- tures are more resistant to rut and moisture damage than slab specimens. 2. When all measurement locations are used, the average creep slope of slab specimens is significantly larger than that of gyratory specimens (Table 4-5). 3. When the middle three measurement loca- tions are excluded, the average rut depths after 15,000 and 20,000 passes and the aver- age creep slope of slab specimens are statisti- cally larger than those of gyratory specimens. The significant differences are shown as the shaded cells in Table 4-5. 4. When the four end measurement locations are excluded, the differences between rut Table 4-3 Statistical F-test on repeatability coefficients of variation (COV) of rut depth (mm) after 10,000, 15,000, and 20,000 cycles and of creep slope (mm/pass) of Field mixture for the comparison of measurements from various sets of measurement locations. Comparison Property COV, % F Critical F df1 df2 P Decision Field gyratory (all measurement locations) vs. Field gyratory (except middle measurement locations) Rut after 10,000 cycles 12.2 vs. 13.9 1.29 2.26 18 17 0.300 Accept Rut after 15,000 cycles 13.2 vs. 12.9 1.04 2.27 17 17 0.467 Accept Rut after 20,000 cycles 14.2 vs. 13.7 1.08 2.27 17 17 0.440 Accept Creep slope 15.8 vs. 15.7 1.02 2.27 17 17 0.485 Accept Field gyratory (all measurement locations) vs. Field gyratory (except end measurement locations) Rut after 10,000 cycles 12.2 vs. 13.9 1.30 2.27 17 17 0.295 Accept Rut after 15,000 cycles 13.2 vs. 14.8 1.26 2.27 17 17 0.319 Accept Rut after 20,000 cycles 14.2 vs. 16.1 1.28 2.27 17 17 0.306 Accept Creep Slope 15.8 vs. 18 1.29 2.27 17 17 0.303 Accept Field slab (all measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 12.8 vs. 12.9 1.01 5.05 5 5 0.494 Accept Rut after 15,000 cycles 14.8 vs. 14.8 1.01 5.05 5 5 0.497 Accept Rut after 20,000 cycles 16.3 vs. 16.1 1.03 5.05 5 5 0.489 Accept Creep Slope 26.4 vs. 25.6 1.06 5.05 5 5 0.474 Accept Field slab (all measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 12.8 vs. 12.2 1.11 5.05 5 5 0.456 Accept Rut after 15,000 cycles 14.8 vs. 14.1 1.10 5.05 5 5 0.458 Accept Rut after 20,000 cycles 16.3 vs. 16.0 1.04 5.05 5 5 0.485 Accept Creep Slope 26.4 vs. 26.9 1.04 5.05 5 5 0.481 Accept

28 depth and creep slope of gyratory and slab specimens become smaller. This is because by excluding the end measurement locations, the average deformation of gyratory speci- mens slightly increases and average defor- mation of slab specimens slightly decreases, resulting in smaller differences between prop- erties of the two specimen types. However, as indicated from Table 4-5, none of the differ- ences are statistically significant. 5. Regardless of the sets of measurement loca- tions used, both the repeatability and repro- ducibility COV of the creep slope from the slab specimens is larger than that of the gyra- tory specimens. However, the differences are not statistically significant. 6. There appears to be a relationship among the differences between the COV of rut depths from gyratory and slab specimens, number of passes, and the measurement locations. As indicated from Tables 4-6 and 4-7, prior to 10,000 passes, slab specimens provide either the same or lower repeatability/reproducibility COVs than gyratory specimens. However, variability of rut depth corresponding to the slab specimens increases as the number of passes increases. On the other hand, the dif- ference between the variability of measure- ments corresponding to gyratory and slab specimens decreases when the data from the end measurement locations are excluded from the analysis. However, none of the dif- Table 4-4 Statistical F-test on reproducibility coefficients of variation (COV) of rut depth (mm) after 10,000, 15,000, and 20,000 cycles and of creep slope (mm/pass) of Field mixture for the comparison of measurements from various sets of measurement locations. Comparison Property COV, % F Critical F df1 df2 P Decision Field gyratory (all measurement locations) vs. Field gyratory (except middle measurement locations) Rut after 10,000 cycles 26.3 vs. 27.7 1.10 2.26 18 17 0.420 Accept Rut after 15,000 cycles 26.3 vs. 27.6 1.10 2.27 17 17 0.423 Accept Rut after 20,000 cycles 26.9 vs. 28.1 1.09 2.27 17 17 0.428 Accept Creep Slope 25.7 vs. 27.3 1.12 2.27 17 17 0.406 Accept Field gyratory (all measurement locations) vs. Field gyratory (except end measurement locations) Rut after 10,000 cycles 26.3 vs. 25.5 1.06 2.27 17 17 0.450 Accept Rut after 15,000 cycles 26.3 vs. 25.3 1.08 2.27 17 17 0.437 Accept Rut after 20,000 cycles 26.9 vs. 26.1 1.06 2.27 17 17 0.452 Accept Creep Slope 25.7 vs. 25.5 1.02 2.27 17 17 0.484 Accept Field slab (all measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 23.3 vs. 22.4 1.08 5.05 5 5 0.466 Accept Rut after 15,000 cycles 25.5 vs. 24.5 1.08 5.05 5 5 0.467 Accept Rut after 20,000 cycles 27.2 vs. 25.8 1.11 5.05 5 5 0.456 Accept Creep Slope 34.8 vs. 32.6 1.14 5.05 5 5 0.444 Accept Field slab (all measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 23.3 vs. 24 1.06 5.05 5 5 0.476 Accept Rut after 15,000 cycles 25.5 vs. 26.6 1.09 5.05 5 5 0.464 Accept Rut after 20,000 cycles 27.2 vs. 28.6 1.10 5.05 5 5 0.459 Accept Creep Slope 34.8 vs. 37.6 1.16 5.05 5 5 0.436 Accept

29 ferences between variability of gyratory and slab specimens are statistically significant. From the above observations it can be concluded that the type of specimens used for the HWTT should be recorded along with the test results, given that the average of one or more properties could be signifi- cantly different depending on which measurement location data are used in the analysis. However, if the end measurement locations are excluded from the analysis, the estimate of mixture performance from the gyratory and slab specimens would not be different. Given that the differences in variability of mea- surements using gyratory and slab specimens are not statistically significant, the precision estimates for the properties of well-performing mixtures were prepared by pooling together the COV of the prop- erties of gyratory and slab specimens. 4.3.2 Poorly Performing Wyoming Mixture Table 4-8 provides statistics on the properties of gyratory and slab specimens of the poorly per- forming Wyoming mixture. The properties include number of passes to 6-mm and 12-mm threshold rut depths, creep slope, stripping slope, and the number of cycles to the inflection point. The comparison of statistics from various measurement locations and from gyratory and slab specimens are discussed in the following sections. A review of the data in Table 4-8 indicates that there is a strong relationship between 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles De fo rm ati on , m m Average Rut Depth (Except Middle Sensors) Field gyratory (except middle sensors) Field slab (except middle sensors) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles De fo rm ati on , m m Average Rut Depth (Except End Sensors) Field gyratory (except end sensors) Field slab (except end sensors) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles De fo rm ati on , m m Average Rut Depth (All Sensors) Field gyratory (all sensors) Field Slab (all sensors) 0.000 0.020 0.040 0.060 0.080 0.100 0.120 Creep Slope Cr ee p Sl op e, m m /p as s Average Creep Slope (Except Middle Sensors) Field gyratory (except middle sensors) Field slab (except middle sensors) 0.000 0.020 0.040 0.060 0.080 0.100 0.120 Creep Slope Cr ee p Sl op e, m m /p as s Average Creep Slope (Except End Sensors) Field gyratory (except end sensors) Field slab (except end sensors) 0.000 0.020 0.040 0.060 0.080 0.100 0.120 Creep Slope Cr ee p Sl op e, m m /p as s Average Creep Slope (All Sensors) Field gyratory (all sensors) Field Slab (all sensors) Figure 4-3 Graphical comparison of average of the properties of gyratory and slab specimens of the Field mixture measured using data from all measurement locations, all except middle three measurement locations, and all except four measurement locations at each end.

30 using different measurement locations: all measure- ment locations, all except three middle measure- ment locations, and all except two measurement locations at each end. In each table, the first two comparisons correspond to gyratory specimens and the third and fourth comparisons correspond to the slab specimens. The following are observed from Figure 4-5 and Table 4-9. 1. For the gyratory specimens, excluding the data from the three middle locations resulted in an increase in the average number of cycles to both 6-mm and 12-mm rut depth, decreases in the creep and stripping slopes, and an increase in the number of cycles to the inflec- tion point. This is because the deformations averages and the standard deviations. Therefore, the statistical comparison has been performed on the averages and repeatability/reproducibility COV. 4.3.2.1 Comparison of Statistics from Different Measurement Locations. Figures 4-5 and 4-6 show the averages and COV of the properties from vari- ous measurement locations. The results of statisti- cal comparisons are provided in Tables 4-9 through 4-13. Discussion of the results follows. StatiStical compariSon of average valueS. Figure 4-5 compares the average values using dif- ferent measurement locations. Table 4-9 provides the results of statistical comparison of the averages of various properties of gyratory and slab specimens 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Repeatability COV (All Sensors) Field gyratory (all sensors) Field Slab (all sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Repeatability COV (Except Middle Sensors) Field gyratory (except middle sensors) Field slab (except middle sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Repeatability COV (Except End Sensors) Field gyratory (except end sensors) Field slab (except end sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Reproducibility COV (Except Middle Sensors) Field gyratory (except middle sensors) Field slab (except middle sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Reproducibility COV (Except End Sensors) Field gyratory (except end sensors) Field slab (except end sensors) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 Rut after 10,000 cycles Rut after 15,000 cycles Rut after 20,000 cycles Creep Slope CO V, % Reproducibility COV (All Sensors) Field gyratory (all sensors) Field Slab (all sensors) Figure 4-4 Graphical comparison of coefficients of variation (COV) of the properties of gyratory and slab specimens of the Field mixture measured using data from all measurement locations, all except middle three measurement locations, and all except four measurement locations at each end.

31 Table 4-5 Statistical t-test on averages of rut depth (mm) after 10,000, 15,000, and 20,000 passes and of creep slope (mm/pass) corresponding to gyratory and slab specimens of Field mixture. Comparison Property Averages S T df Critical t P Decision Field gyratory (all measurement locations) vs. Field slabs (all measurement locations) Rut after 10,000 cycles 2.26 vs. 2.6 0.56 -1.30 22 1.72 0.103 Accept Rut after 15,000 cycles 2.53 vs. 2.99 0.64 -1.53 22 1.72 0.070 Accept Rut after 20,000 cycles 2.71 vs. 3.27 0.71 -1.67 22 1.72 0.055 Accept Creep Slope 0.089 vs. 0.112 0.02 -2.02 22 1.72 0.028 Reject Field gyratory (except middle measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 2.22 vs. 2.62 0.57 -1.49 23 1.71 0.074 Accept Rut after 15,000 cycles 2.46 vs. 3 0.64 -1.79 22 1.72 0.043 Reject Rut after 20,000 cycles 2.63 vs. 3.28 0.71 -1.95 22 1.72 0.032 Reject Creep Slope 0.086 vs. 0.113 0.02 -2.49 22 1.72 0.011 Reject Field gyratory (except end measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 2.36 vs. 2.56 0.56 -0.77 22 1.72 0.225 Accept Rut after 15,000 cycles 2.65 vs. 2.94 0.64 -0.97 22 1.72 0.171 Accept Rut after 20,000 cycles 2.85 vs. 3.23 0.71 -1.14 22 1.72 0.133 Accept Creep Slope 0.095 vs. 0.109 0.02 -1.25 22 1.72 0.111 Accept Table 4-6 Statistical F-test for comparison of the repeatability COV of rut depth (mm) after 10,000, 15,000, and 20,000 passes and of creep slope (mm/pass) corresponding to gyratory and slab specimens of the Field mixture. Comparison # of Passes COV, % F Critical F df1 df2 P Decision Field gyratory (all measurement locations) vs. Field slab (all measurement locations) Rut after 10,000 cycles 12.2 vs. 12.8 1.10 2.81 5 17 0.395 Accept Rut after 15,000 cycles 13.2 vs. 14.8 1.26 2.81 5 17 0.326 Accept Rut after 20,000 cycles 14.2 vs. 16.3 1.31 2.81 5 17 0.305 Accept Creep Slope 15.8 vs. 26.4 2.77 2.81 5 17 0.052 Accept Field gyratory (except middle measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 13.9 vs. 12.9 1.16 4.58 18 5 0.477 Accept Rut after 15,000 cycles 12.9 vs. 14.8 1.30 2.81 5 17 0.308 Accept Rut after 20,000 cycles 13.7 vs. 16.1 1.38 2.81 5 17 0.282 Accept Creep Slope 15.7 vs. 25.6 2.65 2.81 5 17 0.060 Accept Field gyratory (except end measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 13.9 vs. 12.2 1.31 4.59 17 5 0.411 Accept Rut after 15,000 cycles 14.8 vs. 14.1 1.10 4.59 17 5 0.500 Accept Rut after 20,000 cycles 16.1 vs. 16 1.02 4.59 17 5 0.545 Accept Creep Slope 18 vs. 26.9 2.24 2.81 5 17 0.097 Accept

32 not show any consistent trend of decrease or increase in the average properties. This might be because the deformation of slabs is more uniform among various measurement loca- tions than those of gyratory specimens. The stripping slope is shown to be significantly decreased by excluding the three middle measurement locations. However, the physi- cal significance of this difference is not clear, given that an increase in stripping slope is expected when the smaller deformation at the location of the three middle measurement locations are excluded from the analysis. StatiStical compariSon of variability. Tables 4-10 through 4-13 provide the results of sta- tistical comparison of the repeatability and reproduc- ibility COV of the number of passes to 6-mm and 12-mm rut depth, creep slope, stripping slope, and number of cycles to inflection point using different measurement locations: all, all except middle three, and all except two at each end. The COV values are shown in Figure 4-6. As indicated by Tables 4-10 through 4-13, there are no specific trends of decrease or increase in variability by excluding data from any at the location of three middle measurement locations are larger than those at other mea- surement locations and, therefore, excluding them would result in an estimate of greater resistance of the mixture to deformation. The effect of excluding the readings from the three middle locations is statistically signifi- cant for the stripping slope (Table 4-9). 2. For the gyratory specimens, excluding the data from the end measurement locations resulted in decreases in the average number of cycles to 6-mm and 12-mm rut depth and the inflec- tion point and an increase in the creep and stripping slopes. This is because the defor- mations at the ends are smaller than those at other locations and excluding them yields an estimate of less resistance of the mixture to deformation. Among the comparisons, the dif- ferences between number of passes to 12-mm rut depth and between the stripping slopes are statistically significant. 3. For the slab specimens, excluding the data from the three middle measurement locations or the four end measurement locations does Table 4-7 Statistical F-test on reproducibility COV of rut depth (mm) after 10,000, 15,000, and 20,000 cycles and of creep slope (mm/pass) of gyratory and slab specimens of the Field mixture. Comparison # of Passes COV, % F Critical F df1 df2 P Decision Field gyratory (all measurement locations) vs. Field slab (all measurement locations) Rut after 10,000 cycles 26.3 vs. 23.3 1.28 4.59 17 5 0.425 Accept Rut after 15,000 cycles 26.3 vs. 25.5 1.06 4.59 17 5 0.519 Accept Rut after 20,000 cycles 26.9 vs. 27.2 1.02 2.81 5 17 0.434 Accept Creep Slope 25.7 vs. 34.8 1.83 2.81 5 17 0.160 Accept Field gyratory (except middle measurement locations) vs. Field slab (except middle measurement locations) Rut after 10,000 cycles 27.7 vs. 22.4 1.53 4.58 18 5 0.338 Accept Rut after 15,000 cycles 27.6 vs. 24.5 1.27 4.59 17 5 0.428 Accept Rut after 20,000 cycles 28.1 vs. 25.8 1.18 4.59 17 5 0.463 Accept Creep Slope 27.3 vs. 32.6 1.43 2.81 5 17 0.265 Accept Field gyratory (except end measurement locations) vs. Field slab (except end measurement locations) Rut after 10,000 cycles 25.5 vs. 24 1.13 4.59 17 5 0.486 Accept Rut after 15,000 cycles 25.3 vs. 26.6 1.10 2.81 5 17 0.394 Accept Rut after 20,000 cycles 26.1 vs. 28.6 1.20 2.81 5 17 0.351 Accept Creep Slope 25.5 vs. 37.6 2.17 2.81 5 17 0.105 Accept

33 Table 4-8 Summary of statistics of HWTT properties for gyratory and slab specimens of WY mixture computed from all measurement locations, all except the middle three measurement locations, and all except the end measurement locations. Specimens Type/ Measurement Locations Set Property # of Labs Average Repeatability Reproducibility STD COV, % STD COV, % Sx WY gyratory (all measurement locations) Cycles to 6 mm 25 7619 1180 15.5 1928 25.3 1738 Cycles to 12 mm 25 11879 2030 17.1 2686 22.6 2270 Creep Slope 24 0.36 0.057 16.0 0.116 32.4 0.106 Stripping Slope 24 1.09 0.186 17.1 0.229 21.0 0.172 Cycles to Inflection Point 24 4605 1091 23.7 1510 32.8 1219 WY gyratory (except middle measurement locations) Cycles to 6 mm 25 8193 1262 15.4 2022 24.7 1815 Cycles to 12 mm 19 12919 2225 17.2 2902 22.5 2438 Creep Slope 24 0.32 0.063 19.6 0.100 30.9 0.089 Stripping Slope 24 0.91 0.151 16.5 0.177 19.4 0.141 Cycles to Inflection Point 25 4756 1093 23.0 1469 30.9 1250 WY gyratory (except end measurement locations) Cycles to 6 mm 25 7041 1138 16.2 1843 26.2 1659 Cycles to 12 mm 25 10517 1883 17.9 2492 23.7 2106 Creep Slope 24 0.38 0.054 14.1 0.106 27.8 0.099 Stripping Slope 24 1.36 0.250 18.4 0.274 20.2 0.210 Cycles to Inflection Point 24 4290 1161 27.1 1525 35.5 1285 WY slab (all measurement locations) Cycles to 6 mm 10 11870 1620 13.6 2385 20.1 2092 Cycles to 12 mm 5 16540 858 5.2 1478 8.9 1347 Creep Slope 10 0.21 0.031 14.7 0.048 22.4 0.040 Stripping Slope 10 0.69 0.120 17.4 0.163 23.7 0.131 Cycles to Inflection Point 10 7540 1555 20.6 2214 29.4 1814 WY slab (except middle measurement locations) Cycles to 6 mm 9 11544 1414 12.3 1713 14.8 1391 Cycles to 12 mm 5 17460 728 4.2 1793 10.3 1717 Creep Slope 10 0.22 0.026 12.3 0.047 21.9 0.043 Stripping Slope 9 0.59 0.085 14.5 0.096 16.3 0.075 Cycles to Inflection Point 10 7495 1478 19.7 2181 29.1 1914 WY slab (except end measurement locations) Cycles to 6 mm 10 11480 1795 15.6 2292 20.0 1908 Cycles to 12 mm 6 16017 1244 7.8 1794 11.2 1563 Creep Slope 9 0.20 0.046 23.6 0.043 21.8 0.027 Stripping Slope 10 0.79 0.175 22.0 0.190 23.9 0.144 Cycles to Inflection Point 10 7160 1809 25.3 2506 35.0 2156

34 Figure 4-5 Comparison of the average properties measured using all measurement locations, all except middle three measurement locations, and all except the end measurement locations. 0 2000 4000 6000 8000 10000 12000 14000 Cycles to 6 mm Cycles to 12 mm Cycles to Inflection Point # of P as se s Average, # of Passes WY gyratory (all sensors) WY gyratory (except middle sensors) WY gyratory (except end sensors) 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 Creep Slope Strip Slope De fo rm ati on p er p as s, m m /p as s Average, Creep and Strip Slopes WY gyratory (all sensors) WY gyratory (except middle sensors) WY gyratory (except end sensors) 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 Cycles to 6 mm Cycles to 12 mm Cycles to Inflection Point # of P as se s Average, # of Passes WY slab (all sensors) WY slab (except middle sensors) WY slab (except end sensors) 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 Creep Slope Strip Slope D ef or m ati on p er p as s, m m /p as s Average, Creep and Strip Slopes WY slab (all sensors) WY slab (except middle sensors) WY slab (except end sensors) Figure 4-6 Comparison of the repeatability and reproducibility COV of properties of the poorly performing mixture using all measurement locations, all except middle three measurement locations, and all except the end measurement locations. 0 5 10 15 20 25 30 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point CO V, % Repeatability COV, % WY slab (all sensors) WY slab (except middle sensors) WY slab (except end sensors) 0 5 10 15 20 25 30 35 40 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point Co v, % Reproducibility COV, % WY slab (all sensors) WY slab (except middle sensors) WY slab (except end sensors) 0 5 10 15 20 25 30 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point CO V, % Repeatability COV, % WY gyratory (all sensors) WY gyratory (except middle sensors) WY gyratory (except end sensors) 0 5 10 15 20 25 30 35 40 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point Co v, % Reproducibility COV, % WY gyratory (all sensors) WY gyratory (except middle sensors) WY gyratory (except end sensors)

35 Table 4-9 Statistical t-test for comparison of the average # of cycles to 6-mm and 12-mm rut depths, creep and stripping slopes, and # of cycles to inflection point of WY gyratory and slab specimens from various measurement location sets. Comparison Property Averages S T df Critical t P Decision WY gyratory (all measure- ment locations) vs. WY gyratory (except middle measurement locations) Cycles to 6 mm 7619 vs. 8193 1777 -1.14 48 1.68 0.130 Accept Cycles to 12 mm 11879 vs. 12919 2344 -1.46 42 1.68 0.076 Accept Creep Slope 0.36 vs. 0.32 0.098 1.24 46 1.68 0.111 Accept Stripping Slope 1.09 vs. 0.91 0.157 3.86 46 1.68 0.000 Reject Cycles to Inflection Point 4605 vs. 4756 1235 -0.43 47 1.68 0.335 Accept WY gyratory (all measurement locations) vs. WY gyratory (except end measurement locations) Cycles to 6 mm 7619 vs. 7041 1699 1.20 48 1.68 0.117 Accept Cycles to 12 mm 11879 vs. 10517 2190 2.20 48 1.68 0.016 Reject Creep Slope 0.36 vs. 0.38 0.102 -0.72 46 1.68 0.237 Accept Stripping Slope 1.09 vs. 1.36 0.192 -4.84 46 1.68 0.000 Reject Cycles to Inflection Point 4605 vs. 4290 1252 0.87 46 1.68 0.194 Accept WY slab (all measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 11870 vs. 11544 1796 0.39 17 1.74 0.349 Accept Cycles to 12 mm 16540 vs. 17460 1543 -0.94 8 1.86 0.187 Accept Creep Slope 0.21 vs. 0.22 0.042 -0.15 18 1.73 0.442 Accept Stripping Slope 0.69 vs. 0.59 0.108 2.03 17 1.74 0.029 Reject Cycles to Inflection Point 7540 vs. 7495 1865 0.05 18 1.73 0.479 Accept WY slab (all measurement locations) vs. WY slab (except end measurement locations) Cycles to 6 mm 11870 vs. 11480 2002 0.44 18 1.73 0.334 Accept Cycles to 12 mm 16540 vs. 16017 1471 0.59 9 1.83 0.286 Accept Creep Slope 0.21 vs. 0.2 0.035 1.04 17 1.74 0.157 Accept Stripping Slope 0.69 vs. 0.79 0.138 -1.70 18 1.73 0.053 Accept Cycles to Inflection Point 7540 vs. 7160 1992 0.43 18 1.73 0.337 Accept Table 4-10 Statistical F-test on repeatability COV of number of cycles to 6-mm and 12-mm rut depth, and number of cycles to inflection point of gyratory and slab specimens of Wyoming mixture measured using different measurement locations sets. Comparison Property COV,% F Critical F df1 df2 P Decision WY gyratory (all measurement locations) Vs. WY gyratory (except middle measurement locations) Cycles to 6 mm 15.5 vs. 15.4 1.01 1.98 24 24 0.490 Accept Cycles to 12 mm 17.1 vs. 17.2 1.02 2.05 18 24 0.478 Accept Cycles to Inflection Point 23.7 vs. 23 1.06 1.99 23 24 0.441 Accept WY gyratory (all measurement locations) Vs. WY gyratory (except end measurement locations) Cycles to 6 mm 15.5 vs. 16.2 1.09 1.98 24 24 0.418 Accept Cycles to 12 mm 17.1 vs. 17.9 1.10 1.98 24 24 0.411 Accept Cycles to Inflection Point 23.7 vs. 27.1 1.31 2.01 23 23 0.263 Accept (continued on next page)

36 Table 4-11 Statistical F-test on repeatability COV of creep slope and stripping slope of gyratory and slab specimens of Wyoming mixture measured using different measurement locations sets. Comparison Property COV, % F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY gyratory (except middle measurement locations) Creep Slope 16 vs. 19.6 1.50 2.01 23 23 0.170 Accept Stripping Slope 17.1 vs. 16.5 1.07 2.01 23 23 0.437 Accept WY gyratory (all measurement locations) vs. WY gyratory (except end measurement locations) Creep Slope 16 vs. 14.1 1.28 2.01 23 23 0.278 Accept Stripping Slope 17.1 vs. 18.4 1.17 2.01 23 23 0.357 Accept WY slab (all measurement locations) vs. WY slab (except middle measurement locations) Creep Slope 14.7 vs. 12.3 1.43 3.18 9 9 0.303 Accept Stripping Slope 17.4 vs. 14.5 1.45 3.39 9 8 0.306 Accept WY slab (all measurement locations) vs. WY slab (except end measurement locations) Creep Slope 14.7 vs. 23.6 2.58 3.23 8 9 0.090 Accept Stripping Slope 17.4 vs. 22 1.60 3.18 9 9 0.247 Accept Table 4-10 (Continued) Comparison Property COV,% F Critical F df1 df2 P Decision WY slab (all measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 13.6 vs. 12.3 1.24 3.39 9 8 0.386 Accept Cycles to 12 mm 5.2 vs. 4.2 1.55 6.39 4 4 0.341 Accept Cycles to Inflection Point 20.6 vs. 19.7 1.09 3.18 9 9 0.449 Accept WY slab (all measurement locations) vs. WY slab (except end measurement locations) Cycles to 6 mm 13.6 vs. 15.6 1.31 3.18 9 9 0.346 Accept Cycles to 12 mm 5.2 vs. 7.8 2.24 6.26 5 4 0.227 Accept Cycles to Inflection Point 20.6 vs. 25.3 1.50 3.18 9 9 0.278 Accept

37 Table 4-12 Statistical F-test on reproducibility COV of number of cycles to 6-mm and 12-mm rut depth and number of cycles to inflection point of gyratory and slab specimens of Wyoming mixture measured using different measurement locations sets. Comparison Property COV, # of Cycles F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY gyratory (except middle measurement locations) Cycles to 6 mm 25.3 vs. 24.7 1.05 1.98 24 24 0.452 Accept Cycles to 12 mm 22.6 vs. 22.5 1.01 2.15 24 18 0.496 Accept Cycles to Inflec- tion Point 32.8 vs. 30.9 1.13 1.99 23 24 0.387 Accept WY gyratory (all measurement locations) vs. WY gyratory (except end measurement locations) Cycles to 6 mm 24.7 vs. 26.2 1.12 1.98 24 24 0.388 Accept Cycles to 12 mm 22.6 vs. 23.7 1.10 1.98 24 24 0.410 Accept Cycles to Inflec- tion Point 32.8 vs. 35.5 1.17 2.01 23 23 0.351 Accept WY slab (all measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 20.1 vs. 14.8 1.83 3.39 9 8 0.203 Accept Cycles to 12 mm 8.9 vs. 10.3 1.32 6.39 4 4 0.397 Accept Cycles to Inflec- tion Point 29.4 vs. 29.1 1.02 3.18 9 9 0.490 Accept WY slab (all measurement locations) vs. WY slab (except end measurement locations) Cycles to 6 mm 20.1 vs. 20 1.01 3.18 9 9 0.493 Accept Cycles to 12 mm 8.9 vs. 11.2 1.57 6.26 5 4 0.341 Accept Cycles to Inflec- tion Point 29.4 vs. 35 1.42 3.18 9 9 0.304 Accept Table 4-13 Statistical F-test on reproducibility COV of number of creep slope and stripping slope of gyratory and slab specimens of Wyoming mixture measured using different measurement locations sets. Comparison Property COV, % F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY gyratory (except middle measurement locations) Creep Slope 32.4 vs. 30.9 1.10 2.01 23 23 0.410 Accept Stripping Slope 21 vs. 19.4 1.18 2.01 23 23 0.347 Accept WY gyratory (all measurement locations) vs. WY gyratory (except end measurement locations) Creep Slope 32.4 vs. 27.8 1.36 2.01 23 23 0.235 Accept Stripping Slope 21 vs. 20.2 1.08 2.01 23 23 0.425 Accept (continued on next page)

38 the averages and repeatability/reproducibility sta- tistics of the properties of the gyratory and slab specimens. Tables 4-14 through 4-18 provide the results of statistical comparison of the averages and variability of the properties of gyratory and slab specimens. The COV values are the basis of repeatability/reproducibility precision estimates given that there are strong relationships between the averages and standard deviations. In each table, the first comparison corresponds to all mea- surement locations, the second comparison corre- sponds to all except the three middle measurement locations, and the third comparison corresponds to all except two measurement locations at each end. The following are observed from the graphs and tables: 1. The comparison of the average properties of gyratory and slab specimens in Table 4-14 and Figure 4-7 indicates that regardless of the measurement locations used, the slab specimens of the poorly performing mixture are more resistant to rutting and moisture damage than the gyratory specimens. The difference between average properties of slab and gyratory specimens become statically significant when the three middle measure- ment locations or the four end measurement locations are excluded from the analysis. This suggests that for the poorly performing mixtures, unlike well-performing mixture, gyratory specimens are less resistant to rut and moisture damage. This is because for the well-performing mixture, the mold for measurement location sets. Moreover, none of the dif- ferences between the COVs corresponding to differ- ent measurement locations are statistically significant. In summary, for the gyratory specimens of the poorly performing mixture, excluding the data from the four end measurement locations provides signif- icantly smaller average number of passes to 12-mm rut depth and larger average stripping slope, which are a more conservative estimate of mixture perfor- mance. On the other hand, excluding the data from the three middle measurement locations provided a significantly smaller stripping slope, which is a less conservative estimate of the mixture’s performance. In terms of variability, excluding the measurements from the end or the middle measurement loca- tions did not significantly improve the variability of the properties. The variation of the deformation along with various measurement locations can be improved by reducing the confinement at the ends and increasing the confinement around the midpoint of gyratory specimens, as discussed in Appendix C. Thus, it can be concluded that the precision esti- mates of AASHTO T 324 should be prepared by pooling the statistics from all sets of measurement locations. Considering that at various measurement locations the deformations are interdependent, excluding the deformation from any measurement location is not recommended. An average deforma- tion from all measurement locations would provide a more comprehensive representation of the entire deformation basin. 4.3.2.2 Comparison of Statistics from Gyratory and Slab Specimens. Figures 4-7 and 4-8 present Table 4-13 (Continued) Comparison Property COV, % F Critical F df1 df2 P Decision WY slab (all measurement locations) vs. WY slab (except middle measurement locations) Creep Slope 22.4 vs. 21.9 1.04 3.18 9 9 0.476 Accept Stripping Slope 23.7 vs. 16.3 2.12 3.39 9 8 0.152 Accept WY slab (all measurement locations) vs. WY slab (except end measurement locations) Creep Slope 22.4 vs. 21.8 1.06 3.39 9 8 0.475 Accept Stripping Slope 23.7 vs. 23.9 1.02 3.18 9 9 0.489 Accept

39 Tables 4-15 through 4-18 indicate that the COVs of the majority of the properties of slab specimens are significantly smaller than those of gyratory specimens. However, this could be attributed to the significantly smaller degrees of freedom (the number of values in the final calculation of F statistics) of slab specimens than those of gyratory specimens. In summary, given that depending on the mea- surement locations used, the average of the prop- erties measured using gyratory and slab specimens could be significantly different, the type of speci- mens used should be recorded along with the wheel track test results of poorly performing mixtures. gyratory specimens provides confinement higher than the confinement for slabs; so gyratory specimens perform better. However, for the poorly performing mixture, the high confinement of gyratory specimens causes increased differential deformation between the midpoint and the ends. This is because the material is not allowed to move laterally at the ends but free to move at the center. When deformation increases beyond a certain level, the wheels’ dynamic for gyratory specimens intensifies resulting in more deformation and poorer performance of gyratory than slab specimens. 2. The comparison of variability of prop- erties of gyratory and slab specimens in Figure 4-7 Comparison of average properties of gyratory and slab specimens of WY mixture measured using all measurement locations, all except middle three measurement locations, and all except the end measurement locations. 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 Cycles to 6 mm Cycles to 12 mm Cycles to Inflection Point # of P as se s Average, # of Passes, All Sensors WY gyratory (all sensors) WY slab (all sensors) 0.00 0.20 0.40 0.60 0.80 1.00 1.20 Creep Slope Strip Slope De fo rm ati on p er p as s, m m /p as s Average, Creep and Strip Slopes, All Sensors WY gyratory (all sensors) WY slab (all sensors) 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 Cycles to 6 mm Cycles to 12 mm Cycles to Inflection Point # of P as se s Average, # of Passes, Except Middle Sensors WY gyratory (except middle sensors) WY slab (except middle sensors) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Creep Slope Strip Slope De fo rm ati on p er p as s, m m /p as s Average, Creep and Strip Slopes, Except Middle Sensors WY gyratory (except middle sensors) WY slab (except middle sensors) 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 Cycles to 6 mm Cycles to 12 mm Cycles to Inflection Point # of P as se s Average, # of Passes, Except End Sensors WY gyratory (except end sensors) WY slab (except end sensors) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Creep Slope Strip SlopeDe fo rm ati on p er p as s, m m /p as s Average, Creep and Strip Slopes, Except End Sensors WY gyratory (except end sensors) WY slab (except end sensors)

40 4.3.3 Pooled Statistics Precision estimates were prepared for the proper- ties of the two types of mixtures. For the well perform- ing mixture, the precision estimates were prepared for deformation after specific numbers of passes and for creep slope. For the poorly performing mixture, the precision estimates were prepared for the num- ber of passes to the threshold rut depth, creep slope, stripping slope, and number of passes to the inflec- tion point. Given that creep slope is a common prop- erty for both well and poorly performing mixtures, statistical analysis will be conducted to determine if the statistics of creep slope from the two mixture types are the same and can be pooled together. Preci- sion estimates of all other properties will be prepared The differences between properties of gyratory and slab specimens can be reduced by decreas- ing the confinement at the ends and increasing the confinement around the midpoint of gyratory specimens. The significantly smaller COV of the number of passes to 12-mm rut depth for the slab specimens than for the gyratory specimens is most probably due to the significantly smaller number of slab spec- imens compared to gyratory specimens. Therefore, in preparing the precision estimates of the number of passes to 12-mm rut depth, the COV corresponding to gyratory specimens were used. For other proper- ties, where the COVs associated with the gyratory and slab specimens are not significantly different, they were pooled together. Figure 4-8 Comparison of coefficients of variation (COV) of properties of gyratory and slab specimens measured using all measurement locations, all except middle three measurement locations, and all except the end measurement locations. 0 5 10 15 20 25 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point CO V, % Repeatability % COV, All Sensors WY gyratory (all sensors) WY slab (all sensors) 0 5 10 15 20 25 30 35 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point Co v, % Reproducibility % COV, All Sensors WY gyratory (all sensors) WY slab (all sensors) 0 5 10 15 20 25 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point CO V, % Repeatability % COV, Except Middle Sensors WY gyratory (except middle sensors) WY slab (except middle sensors) 0 5 10 15 20 25 30 35 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point Co v, % Reproducibility % COV, Except Middle Sensors WY gyratory (except middle sensors) WY slab (except middle sensors) 0 5 10 15 20 25 30 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point CO V, % Repeatability % COV, Except End Sensors WY gyratory (except end sensors) WY slab (except end sensors) 0 5 10 15 20 25 30 35 40 Cycles to 6 mm Cycles to 12 mm Creep Slope Strip Slope Cycles to Inflection Point Co v, % Reproducibility % COV, Except End Sensors WY gyratory (except end sensors) WY slab (except end sensors)

Table 4-14 Statistical t-test for comparison of average properties of gyratory and slab specimens of WY mixture using various measurement location sets. Comparison Property Averages S T df Critical t P Decision WY gyratory (all measurement locations) vs. WY slab (all measurement locations) Cycles to 6 mm 7619 vs. 11870 1841 -1.10 33 1.69 0.139 Accept Cycles to 12 mm 11879 vs. 16540 2163 -1.58 28 1.70 0.063 Accept Creep Slope 0.358 vs. 0.213 0.093 1.31 32 1.69 0.100 Accept Stripping Slope 1.09 vs. 0.69 0.161 3.77 32 1.69 0.000 Reject Cycles to Inflection Point 4605 vs. 7540 1412 -0.37 32 1.69 0.355 Accept WY gyratory (except middle measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 8193 vs. 11544 1719 -5.02 32 1.69 0.000 Reject Cycles to 12 mm 12919 vs. 17460 2324 -3.89 22 1.72 0.000 Reject Creep Slope 0.323 vs. 0.215 0.079 3.62 32 1.69 0.000 Reject Stripping Slope 0.914 vs. 0.589 0.127 6.53 31 1.70 0.000 Reject Cycles to Inflection Point 4756 vs. 7495 1461 -5.01 33 1.69 0.000 Reject WY gyratory (except two ends) vs. WY slab (except two ends) Cycles to 6 mm 7041 vs. 11480 1730 -6.86 33 1.69 0.000 Reject Cycles to 12 mm 10517 vs. 16017 2023 -5.98 29 1.70 0.000 Reject Creep Slope 0.38 vs. 0.196 0.086 5.46 31 1.70 0.000 Reject Stripping Slope 1.357 vs. 0.794 0.193 7.73 32 1.69 0.000 Reject Cycles to Inflection Point 4290 vs. 7160 1579 -4.83 32 1.69 0.000 Reject Table 4-15 Statistical F-test for comparison of repeatability coefficients of variation (COV) of number of cycles to 6-mm and 12-mm rut depth and to the inflection point for gyratory and slab specimens of WY mixture using various measurement location sets. Comparison Property COV of # of Cycles F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY slab (all measurement locations) Cycles to 6 mm 15.5 vs. 13.6 1.29 2.90 24 9 0.361 Accept Cycles to 12 mm 17.1 vs. 5.2 10.85 5.77 24 4 0.016 Reject Cycles to Inflection Point 23.7 vs. 20.6 1.32 2.91 23 9 0.345 Accept WY gyratory (except middle measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 15.4 vs. 12.3 1.58 3.12 24 8 0.257 Accept Cycles to 12 mm 17.2 vs. 4.2 17.06 5.82 18 4 0.007 Reject Cycles to Inflection Point 23 vs. 19.7 1.36 2.90 24 9 0.328 Accept WY gyratory (except end measurement locations) vs. WY slab (except end measurement locations) Cycles to 6 mm 16.2 vs. 15.6 1.07 2.90 24 9 0.487 Accept Cycles to 12 mm 17.9 vs. 7.8 5.31 4.53 24 5 0.036 Reject Cycles to Inflection Point 27.1 vs. 25.3 1.15 2.91 23 9 0.436 Accept

42 Table 4-16 Statistical F-test for comparison of repeatability COV of creep and stripping slope of gyratory and slab specimens of WY mixture using various measurement location sets. Comparison Property COV (%) of Slope F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY slab (all measurement locations Creep Slope 16 vs. 14.7 1.19 2.91 23 9 0.414 Accept Stripping Slope 17.1 vs. 17.4 1.04 2.32 9 23 0.439 Accept WY gyratory (except middle measurement locations) vs. WY slab (except middle measurement locations) Creep Slope 19.6 vs. 12.3 2.54 2.91 23 9 0.075 Accept Stripping Slope 16.5 vs. 14.5 1.30 3.12 23 8 0.367 Accept WY gyratory (except end measurement locations) vs. WY slab (except end measurement locations) Creep Slope 14.1 vs. 23.6 2.78 2.37 8 23 0.026 Reject Stripping Slope 18.4 vs. 22 1.43 2.32 9 23 0.234 Accept Table 4-17 Statistical F-test for comparison of reproducibility coefficients of variation (COV) of number of cycles to 6-mm and 12-mm rut depth and to the inflection point for gyratory and slab specimens of WY mixture using various measurement location sets. Comparison Property COV of # of Cycles F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY slab (all measurement locations) Cycles to 6 mm 25.3 vs. 20.1 1.59 2.90 24 9 0.240 Accept Cycles to 12 mm 22.6 vs. 8.9 6.41 5.77 24 4 0.042 Reject Cycles to Inflec- tion Point 32.8 vs. 29.4 1.25 2.91 23 9 0.381 Accept WY gyratory (except middle measurement locations) vs. WY slab (except middle measurement locations) Cycles to 6 mm 24.7 vs. 14.8 2.77 3.12 24 8 0.069 Accept Cycles to 12 mm 22.5 vs. 10.3 4.78 5.82 18 4 0.070 Accept Cycles to Inflec- tion Point 30.9 vs. 29.1 1.13 2.90 24 9 0.450 Accept WY gyratory (except end measurement locations) vs. WY slab (except end measurement locations) Cycles to 6 mm 26.2 vs. 20 1.72 2.90 24 9 0.201 Accept Cycles to 12 mm 23.7 vs. 11.2 4.48 4.53 24 5 0.051 Accept Cycles to Inflec- tion Point 35.5 vs. 35 1.03 2.91 23 9 0.511 Accept independent of each other. The following sections explain which statistics were pooled in determining the precision estimates of the properties. 4.3.3.1 Well-Performing Mixture. For the rutting- and moisture-resistant mixture, the statistical com- parisons in Tables 4-3 and 4-4 indicated that the COV of the properties measured from any sets of measurement locations are not significantly differ- ent. Therefore, they are pooled together. Addition- ally, the COV of the properties of the gyratory and slab specimens, as shown in Tables 4-6 and 4-7 are not significantly different. For the rut depth after each set of pass numbers, the COVs are pooled from different specimen types as presented in Table 4-19. However, for the creep slope, although the difference between COVs corresponding to gyratory and slab specimens are not statistically significant, the COVs are not pooled. This is because the rejection prob- ability for the comparison of the repeatability COV of creep slope of gyratory and slab specimens is only slightly larger than 0.05% (0.052% in Table 4-6) and

43 considering the magnitude of the difference between the variability of creep slope of gyratory and slab specimens, this difference is considered significant from a practical stand point. Given that the number of gyratory specimens is larger than the number of slabs, the COVs measured from gyratory specimens are considered more accurate and, therefore, the pre- cision estimates of creep slope are determined using the COVs corresponding to gyratory specimens as presented in Table 4-19. A statistical comparison of the repeatability and reproducibility of COVs of the rut depth after various numbers of passes was conducted to determine if they are the same and can be pooled together. The results are shown in Table 4-20. As shown in the table, the COVs of the rut depths after various numbers of passes are not significantly different. Therefore, the COVs of rut depth after 10,000, 15,000, and 20,000 are pooled together, resulting in the 1s repeatability COV of 14.2% and 1s reproducibility COV of 26.0%. Table 4-18 Statistical F-test for comparison of reproducibility coefficient of variations (COV) of creep and stripping slope of gyratory and slab specimens of WY mixture using various measurement location sets. Comparison Property COV (%) of Slope F Critical F df1 df2 P Decision WY gyratory (all measurement locations) vs. WY slab (all measurement locations) Creep Slope 32.4 vs. 22.4 2.09 2.91 23 9 0.125 Accept Stripping Slope 21 vs. 23.7 1.27 2.32 9 23 0.304 Accept WY gyratory (except middle measurement locations) vs. WY slab (except middle measurement locations) Creep Slope 30.9 vs. 21.9 1.98 2.91 23 9 0.144 Accept Stripping Slope 19.4 vs. 16.3 1.41 3.12 23 8 0.319 Accept WY gyratory (except end measurement locations) vs. WY slab (except end measurement locations) Creep Slope 27.8 vs. 21.8 1.63 3.12 23 8 0.243 Accept Stripping Slope 20.2 vs. 23.9 1.40 2.32 9 23 0.244 Accept Table 4-19 Pooled COV of deformation after 10, 15, and 20 thousand number of passes and of creep slope for well performing mixture. Property Repeatability COV, % Reproducibility COV, % Rut after 10,000 passes (mm) 13.0 24.9 Rut after 15,000 passes (mm) 14.1 25.9 Rut after 20,000 passes (mm) 15.4 27.1 Creep slope (mm/pass) 16.5 26.2 Table 4-20 Statistical comparison of the pooled COV of deformation after 10, 15, and 20 thousands number of passes for well-performing mixture Comparison, Passes Statistics F Critical F df1 df2 P Decision 10,000 vs. 15,000 Repeatability 1.18 1.50 66 67 0.25 Accept Reproducibility 1.09 1.50 66 67 0.37 Accept 15,000 vs. 20,000 Repeatability 1.19 1.50 66 66 0.24 Accept Reproducibility 1.09 1.50 66 66 0.36 Accept 10,000 vs. 20,000 Repeatability 1.41 1.50 66 67 0.08 Accept Reproducibility 1.19 1.50 66 67 0.24 Accept

44 A statistical comparison of the repeatability and reproducibility COV of the number of passes to 6-mm and 12-mm rut depth and to the inflection point was conducted to determine if the COVs are the same and can be pooled together. The results are provided in Table 4-26. As indicated from the table, the COV of the number of passes to 6-mm and 12-mm rut depth are the same and can be pooled together. However, the COV of the number of passes to the inflection point is significantly differ- ent from those of the number of passes to 6-mm and 12-mm rut depth. Therefore, in preparing the preci- sion statement, a separate set of precision estimates is provided for the number of passes to inflection point. The resulting 1s repeatability/reproducibility COVs of the number of passes to threshold rut depth are 16.6% and 24.2%. 4.3.4 Comparison of COV of Creep Slopes of the Two Mixture Types The COVs of the creep slope corresponding to well-performing and poor-performing mixtures were statistically compared to investigate if they are statisti- cally the same and can be pooled together. The results 4.3.3.2 Poorly Performing Mixture. For the poorly performing mixture, the statistical comparisons in Tables 4-10 through 4-13 show that COV of the properties measured from various measure- ment location sets are not significantly different. Therefore, they are pooled together as presented in Tables 4-21 and 4-22. A statistical comparison of the variability of properties of gyratory and slab specimens in Tables 4-21 and 4-22 was conducted to determine if the COVs are the same and can be pooled. Tables 4-23 and 4-24 provide the results. As indicated from the tables, the repeatability/reproducibility COVs for the number of passes to 12-mm rut depth are sig- nificantly different and the reproducibility COVs of passes to 6 mm and of creep slope are significantly different. Considering the smaller number of slab specimens compared to gyratory specimens, the COV of the number of passes to 6-mm and 12-mm rut depth and of the creep slope corresponding to gyratory specimens are considered more accurate and therefore used for preparing the precision esti- mates. For other properties, COVs are not signifi- cantly different and are pooled together. The pooled COVs are presented in Table 4-25. Table 4-21 Pooled coefficients of variation (COV) of number of cycles to 6-mm and 12-mm rut depth and to inflection point for gyratory and slab specimens of the poorly performing mixture. Specimen Type Property Repeatability STD, # of Cycles Reproducibility STD, # of Cycles Gyratory Passes to 6-mm 15.7 25.4 Passes to 12-mm 17.4 22.9 Passes to Inflection Point 24.6 33.1 Slab Passes to 6-mm 13.8 18.3 Passes to 12-mm 5.7 10.1 Passes to Inflection Point 21.9 31.2 Table 4-22 Pooled coefficients of variation (COV) of creep and stripping slopes of gyratory and slab specimens of the poorly performing mixture. Specimen Type Property Repeatability COV, % Reproducibility COV, % Gyratory Creep Slope, mm/pass 16.6 30.4 Stripping Slope, mm/pass 17.3 20.2 Slab Creep Slope, mm/pass 16.9 22.0 Stripping Slope, mm/pass 18.0 21.3

45 Table 4-23 Results of statistical comparison of repeatability COVs of the properties of gyratory and slab specimens. Property COV F Critical F df1 df2 P Decision Cycles to 6 mm 16 vs. 14 1.28 1.79 72 26 0.242 Accept Cycles to 12 mm 17 vs. 6 9.30 2.29 66 13 0.000 Reject Cycles to Inflection Point 25 vs. 22 1.26 1.77 70 27 0.253 Accept Creep Slope 17 vs. 17 1.03 1.66 26 69 0.440 Accept Stripping Slope 17 vs. 18 1.07 1.76 18 69 0.395 Accept Table 4-24 Results of statistical comparison of reproducibility COVs of the properties of gyratory and slab specimens. Property COV F Critical F df1 df2 P Decision Cycles to 6 mm 25 vs. 18 1.93 1.79 72 26 0.032 Reject Cycles to 12 mm 23 vs. 10 5.12 2.29 66 13 0.001 Reject Cycles to Inflection Point 33 vs. 31 1.13 1.77 70 27 0.375 Accept Creep Slope 30 vs. 22 1.90 1.79 69 26 0.035 Reject Strip Slope 20 vs. 21 1.11 1.76 18 69 0.359 Accept Table 4-25 Pooled coefficients of variation (COV) of properties of poorly performing mixture. Property Repeatability COV, % Reproducibility COV, % # of Cycles to 6-mm 15.7 25.4 # of Cycles to 12-mm 17.4 22.9 # of Cycles to Inflection Point 23.2 32.1 Creep Slope, mm/pass 16.7 30.4 Stripping Slope, mm/pass 17.7 20.8 Table 4-26 Statistical comparison of the pooled COV of number of passes to 6-mm and 12-mm deformation and to the inflection point for poorly performing mixture. Comparison Statistics COV, % F Critical F df1 df2 P Decision 6-mm vs. 12-mm Repeatability 15.7 vs. 17.4 1.23 1.44 66 98 0.173 Accept Reproducibility 25.4 vs. 22.9 1.23 1.46 98 66 0.188 Accept 6-mm vs. Inflection Point Repeatability 15.7 vs. 23.2 2.19 1.40 97 98 0.000 Reject Reproducibility 25.4 vs. 32.1 1.60 1.40 97 98 0.011 Reject 12-mm vs. Inflection Point Repeatability 17.4 vs. 23.2 1.78 1.46 97 66 0.007 Reject Reproducibility 22.9 vs. 32.1 1.96 1.46 97 66 0.002 Reject of the analysis are provided in Table 4-27. As shown in the table, the differences between repeatability/ reproducibility COVs of creep slope corresponding to the two mixtures are not significantly different and can be pooled together. The resulting 1s repeatability COV of creep slope is 16.6% and 1s reproducibility COV is 28.3%. 4.3.5 Precision Estimates of AASHTO T 324 Table 4-28 provides the precision estimates for AASHTO T 324 developed in this research. The table includes repeatability and reproducibility COVs for various properties of HWTT. A single set of pre- cision estimates was prepared for the properties

46 CHAPTER 5—FINDINGS AND PROPOSED CHANGES TO AASHTO T 324 AND THE HWTT EQUIPMENT 5.1 Findings This report presents the results of an inter- laboratory study (ILS) to determine precision esti- mates for AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA).” The ILS included preparing and sending four replicates of Superpave gyratory and two replicates of linearly kneaded compacted slab specimens of each of a rut- ting and moisture resistant (well-performing) and a rutting and moisture susceptible (poorly performing) mixture to laboratories participating in the ILS to be tested according to AASHTO T 324. Using the results reported by the laboratories, the precision estimates for properties of the two mixtures were prepared. The precision estimates include the within- and between- laboratory precisions for deformation, the number of passes to threshold rut depths, the creep and strip slopes, and the number of passes to the inflection point. The effect of measurement locations used in the analysis and the effect of specimen type on the mean and variance of the HWTT properties were also exam- ined. The properties of the mixtures measured using of both gyratory and slab specimens, either by com- bining the COVs corresponding to the gyratory and slab specimens or by using COVs corresponding to the gyratory specimens given that there were a larger number of gyratory specimens than slab specimens. The proposed precision statement that includes the developed precision estimates is provided in Appendix I. The variability computed in this research only reflects the variability from the HWTT and the test specimen assembly because test specimens were fabricated at AMRL. The variability of mea- surements is attributed to the factors such as the dynamic effect of the wheels, position of the wheel with respect to specimen, the actual measurement locations compared to the design locations, lack of confinement at the joint between gyratory samples, and the effect of the dynamics of the right and left wheels on each other, as discussed in Appendix C (available on the TRB website). To minimize the variability of the test measurements, factors such as position of the wheel with respect to specimen and position of measurement locations should be regu- larly verified. Improving the specimen assembly and mold geometry would also help reduce the variability of the test. Table 4-27 Statistical comparison of the COVs of creep slope of well performing and poorly performing mixtures. Statistics COV, % F Critical F df1 df2 P Decision Repeatability 16.6 vs. 16.5 1.47 1.55 51 69 0.069 Accept Reproducibility 30.4 vs. 26.2 1.35 1.55 51 69 0.122 Accept Table 4-28 Precision estimates for AASHTO T 324. Properties Single-Operator Multilaboratory COV, (%) Acceptable Range of Two Test Results (Percent of Mean)a COV (%) Acceptable Range of Two Test Results (Percent of Mean)a Deformation (mm) 14.2 40.2 26.0 73.6 Number of Passes to Threshold Rut Depth 16.6 47.0 24.2 68.5 Number of Passes to Inflection Point 23.9 67.6 32.1 90.9 Creep Slope (mm/cycle) 16.6 47.0 28.3 80.1 Strip Slope, mm/pass 17.7 50.0 20.8 58.8 aThese values represent the 1s and d2s limits described in ASTM Practice C670.

47 5. The possibility of increasing the specimen length should be explored. This will result in a greater distance between the wheels and the ends of specimens, reduction in the confine- ment, and more even wear of the sample. 6. A means of confining around the joint of the two adjoined gyratory specimens needs to be investigated. A new mold can be designed for this purpose. The use of plaster of Paris is a possible solution for confining the gyratory specimen around the joint using the existing mold configuration. This will also prevent the movement of the molds that might be a cause of loosening of bolts during the test. 7. The expansion of the polyethylene mold due to increase in temperature was discussed as another possible cause of the tray bolts loos- ening. Retightening of the tray bolts at the end of 30-min temperature conditioning is recommended. 8. Exploring a material for the mold with smaller coefficient of thermal expansion than poly- ethylene is suggested. 9. Due to the possible deficiencies in the equip- ment and test setup that could affect the accuracy and precision of the test results, the results from HWTTs should be occasionally verified against the test results of reference specimens with known properties. Testing reference specimens can identify problems with the machine or test setup and remove any anomalies. It is expected that this ref- erence testing can significantly reduce the variability of the test results between partici- pating laboratories. 10. Considering that the deformations across vari- ous measurement locations are inter dependent, excluding the deformation from any mea- surement location is not recommended. An average deformation from all measurement locations would provide a more comprehen- sive representation of the entire deformation basin. REFERENCES 1. AASHTO, Designation T 324, “Standard Method of Test for Hamburg Wheel-Track Testing of Com- pacted Hot Mix Asphalt (HMA),” Standard Speci- fications for Transportation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. all measurement locations were statistically compared with those measured using all except three middle measurement locations and those measured using all except the two measurement locations at each end. Moreover, the statistics of the properties of gyra- tory and slab specimens were statistically compared. These results along with the precision estimates are presented in Chapter 4. The precision statement that includes the developed precision estimates is pro- vided in Appendix I. In addition to developing precision estimates, the data from the ILS and from in-house research were used to gain insight into the causes of variability of the test results. The effects of various components of the wheel track tester and the effect of specimen assembly on the test measurements were investi- gated. These results are presented in Chapter 3. The results of the cause and effect study are presented in Appendix C (available on the TRB website). 5.2 Proposed Changes to AASHTO T 324 and the HWTT Equipment The results of the ILS suggest that the repeat- ability and reproducibility of measurements from the HWTT may be improved by these proposed changes: 1. The current AASHTO T 324 does not address key factors affecting performance such as starting location of the wheel, alignment of the wheel with respect to specimen, and the mea- surement locations used in the analysis. These factors, which significantly affect variability of measurements, need to be standardized. 2. The operation of the equipment should be periodically verified by the manufacturer to identify any machine-related deficiencies. 3. Reducing the confinement at the ends of the two gyratory specimens and increasing the confinement at midpoint around the joints may achieve a more consistent deformation profile. Currently, there is high confinement at the ends and little or no confinement at the midpoint causing differential wear in the wheel path, which would result in bias and high variability in measurements. 4. The variability in cutting the gyratory speci- mens may affect the measured performance of mixtures (especially that of poorly per- forming mixtures). The possibility of elimi- nating the cut should be investigated.

48 9. http://en.wikipedia.org/wiki/Signal-to-noise_ratio_ (imaging) 10. ASTM E691, “Standard Practice for Conducting an Interlaboratory Study to Determine the Precision of a Test Method,” Vol. 14.02, Quality Control Stan- dards, West Conshohocken, PA, 2009. APPENDIXES A THROUGH H These are not published herein but are available on the TRB website where they can be found by searching for NCHRP Research Results Digest 390. APPENDIX I: RECOMMENDED PRECISION ESTIMATES FOR AASHTO T 324 Precision Statement for AASHTO T 324, Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA) 1 Precision and Bias 1.1 Precision—Criteria for judging the accept- ability of deformation after certain number of passes, number of passes to threshold rut depth, number of passes to inflection point, creep slope, and strip slope obtained by this test method are given as follows: 1.1.1 Single-Operator Precis ion (Repeatability)—The single-operator coefficients of variation (1s limit) is shown in Table 1, Column 2. The results of two properly conducted tests obtained in the same laboratory, by the 2. AASHTO, Designation T 312, “Standard Method of Test for Preparing and Determining the Density of Hot-Mix Asphalt (HMA) Specimens by Means of the Superpave Gyratory Compactor,” Standard Specifi- cations for Transportation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. 3. AASHTO, Designation R30, “Standard Practice for Mixture Conditioning of Hot Mix Asphalt (HMA),” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. 4. AASHTO, Designation T 209, “Standard Method of Test for Theoretical Maximum Specific Gravity and Density of Hot-Mix Asphalt Paving Mixtures,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. 5. AASHTO, Designation T 166, “Standard Method of Test for Bulk Specific Gravity (Gmb) of Compacted Hot-Mix Asphalt (HMA) Using Saturated Surface- Dry Specimens,” Standard Specifications for Trans- portation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. 6. AASHTO, Designation T 331, “Bulk Specific Gravity (Gmb) and Density of Compacted Hot Mix Asphalt (HMA) Using Automatic Vacuum Sealing Method,” Standard Specifications for Transportation Materials and Methods of Sampling and Testing, 33rd Edition, AASHTO, Washington, DC, 2013. 7. http://www.svi.nl/SignalToNoiseRatio 8. http://www.wireless-nets.com/resources/tutorials/ define_SNR_values.html Table 1 Precision estimates for AASHTO T 324. Properties Single-Operator Multilaboratory COV (%) Acceptable Range of Two Test Results (Percent of Mean)a COV (%) Acceptable Range of Two Test Results (Percent of Mean)a Deformation (mm) 14.2 40.2 26.0 73.6 Number of Passes to Threshold Rut Depth 16.6 47.0 24.2 68.5 Number of Passes to Inflection Point 23.9 67.6 32.1 90.9 Creep Slope (mm/cycle) 16.6 47.0 28.3 80.1 Strip Slope, mm/pass 17.7 50.0 20.8 58.8 aThese values represent the 1s and d2s limits described in ASTM Practice C670 Note—The precision estimates are based on the analysis of test results from an AMRL interlaboratory study (ILS), which involved testing of gyratory and slab specimens prepared with one lab-mixed, lab-compacted mixture with poor performance and one plant-mixed, lab-compacted mixture with good performance tested at 50°C using PMW wheel track testers. The details of this analysis are presented in the main text of NCHRP Research Results Digest 390.

49 tory study. Their willingness to volunteer their time and conduct the testing under tight time constraints at no cost to the study is most appreciated. Alliance Geotechnical Group AMEC Earth & Environmental APAC TX, Inc. California DOT, Sacramento, CA Colorado DOT, Denver, CO Florida DOT, Gainesville, FL Iowa DOT, Ames, IA Jones Bros. Dirt & Paving Contractors, Inc. Kansas State University—Manhattan Louisiana State University Mathy Technology & Engineering Services Nactech Oklahoma DOT—Oklahoma City Pave Tex Road Science LLC Texas DOT—Austin Texas DOT—Chico Texas DOT—Childress District Texas DOT—Paris Texas DOT—San Marcos Texas DOT—Uvalde Field Lab University of Texas—El Paso Utah DOT—Salt Lake City Utah DOT—Ogden Lab Vulcan Materials Co. Wyoming DOT, Cheyenne Washington State DOT, Pullman University of Texas—Austin University of Massachusetts—Dartmouth Texas A&M University same operator using the same equip- ment, in the shortest practical period of time, should not be considered suspect, unless the difference in the two results, expressed as a percent of their mean, exceeds the single-operator precision limits given in Table 1, Column 3. 1.1.2 Mult i laboratory Precis ion (Reproducibility)—The multi- laboratory coefficients of variation (1s limit) is shown in Table 1, Column 4. Two results submitted by two different operators testing the same material in different laboratories shall not be considered suspect unless the differ- ence in the two results, expressed as a percent of their mean, exceeds the multilaboratory precision limits given in Table 1, Column 5. Bias—No information can be presented on the bias of the procedure because no comparison with the material having an accepted reference value was conducted. ACKNOWLEDGMENTS The research reported herein was performed under Task Order #2B of NCHRP Project 10-87 by the AASHTO Materials Reference Laboratory (AMRL). Dr. Haleh Azari was the principal investi- gator on the study. The author wishes to acknowledge the follow- ing laboratories that participated in this interlabora-

Transportation Research Board 500 Fifth Street, NW Washington, DC 20001 These digests are issued in order to increase awareness of research results emanating from projects in the Cooperative Research Programs (CRP). Persons wanting to pursue the project subject matter in greater depth should contact the CRP Staff, Transportation Research Board of the National Academies, 500 Fifth Street, NW, Washington, DC 20001. COPYRIGHT INFORMATION Authors herein are responsible for the authenticity of their materials and for obtaining written permissions from publishers or persons who own the copyright to any previously published or copyrighted material used herein. Cooperative Research Programs (CRP) grants permission to reproduce material in this publication for classroom and not-for-profit purposes. Permission is given with the understanding that none of the material will be used to imply TRB, AASHTO, FAA, FHWA, FRA, FTA, or Transit Development Corporation endorsement of a particular product, method, or practice. It is expected that those reproducing the material in this document for educational and not-for-profit uses will give appropriate acknowledgment of the source of any reprinted or reproduced material. For other uses of the material, request permission from CRP. ISBN 978-0-309-30812-0 9 780309 308120 9 0 0 0 0 Subscriber Categories: Materials

Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)” Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Research Results Digest 390: Precision Estimates of AASHTO T 324, “Hamburg Wheel-Track Testing of Compacted Hot Mix Asphalt (HMA)” presents new precision statements for the Hamburg wheel tracking test, which is a test used to identify mixtures prone to rutting or moisture damage.

  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!