MTE Experiences with Performance Testing

Andrew Hanz, MTE Services Inc.FHWA Mix ETGFall River, MAMay 9, 2018

Discussion Points• Overview• Summary of Performance Test Efforts

–DCT Test–SCB Test: I-FIT and Jc

• Challenges–Standardization–Implementation & Setting Limits

• Next Steps

Performance Testing Experiences - MTE• WI & MN: High recycle projects on state highways or

county roads.–Internally developed specification that includes Hamburg,

SCB, DCT.• Iowa: Surface mixes and Interlayer

–State specifications for Hamburg and Beam Fatigue• BMD Approach

–WI & MN: Tier 3. Volumetric requirements remain, mix expected to meet or exceed the performance of a conventional mix.

– Iowa: State sets volumetric and performance test limits.

Disc Shaped Compact Tension (DCT) TestThermal Cracking Resistance

0.0

1.0

2.0

3.0

4.0

0 0.5 1 1.5 2 2.5 3 3.5 4

Mea

sure

d L

oad

(kN

)

CMOD(fit) (mm)

Binder 2Binder 1

Fracture Energy = Area Under Curve to

Test Implementation - DCTProcedure and Specification

Test Procedure

• Temperature: LT PG + 10°C • Aging: AASHTO R30 Short Term

– Long term aging done for research.

• Air voids: Design + 3%• Detailed procedures in place for

conditioning time and duration samples can be held at low temps.

Specification

• Previous iterations included a min. fracture energy of 690 J/m2

• MTE draft specification compares to a conventional mix.

Test Implementation – DCTEvaluation #1 – Effect of Mix Design Factors

Statistic Peak Load (kN)Time at Peak Load (secs)

Fracture Energy (J/m2)

Average 3.27 7.58 542Range 0.23 0.38 43Std Dev 0.08 0.15 15.67COV 2.5% 2.0% 2.9%

PBRPG 58-34 (LT

Continuous Grade -34.9)

15 -32.730 -30.650 -27.7

Max Deviation from Plan Grade (°C) 6.3

Factors Studied• Binder Replacement (RAP): 15%, 30%, 50%• Aging: Short Term (4 hrs @ 135C), Long Term

(12 hrs @ 135C)• Polymer Modification: PG 58S-34, PG 58V-34

Recovered Binder Data

Results

Bahia et. al, WHRP 15-04 Study (3)

Test Implementation – DCTEvaluation #1 – Effect of Mix Design Factors

FactorGeneral Trend

Peak Load Time to Peak Load Fracture Energy

Increase PBRPG 58-28: No trend.PG 58-34: No trend.

PG 58-28: Decrease (0.70s)PG 58-34: Decrease (0.33s)

PG 58-28: Increase (31 J/m2)PG 58-34: Decrease (42 J/m2)

Increase Aging

PG 58-28: Increase (0.06 kN)PG 58-34: Decrease (0.05 kN)

PG 58-28: Decrease (0.15s)PG 58-34: Decrease (0.37s)

PG 58-28: Decrease (48 J/m2)PG 58-34: Decrease (14 J/m2)

Use of Modification

PG 58-28: Increase (0.26 kN)PG 58-34: Increase (0.08 kN)

PG 58-28: Decrease (0.60s)PG 58-34: Decrease (0.11s)

PG 58-28: Decrease (24 J/m2)PG 58-34: Decrease (28 J/m2)

Highlight = Inconsistent Trend Between Binder Grades Bahia et. al, WHRP 15-04 Study (3)

Test Implementation – DCTEvaluation #2 – Aging and Aggregate Type

Granite• LAR @ 500 = 18.3• Fracture Energy (12 hr)

= 551 J/m2

Limestone• LAR @ 500 = 32.0• Fracture Energy (12 hr) =

360 J/m2

• Factor driving fracture energy depends on aggregate type.

• Hard aggregate = Mastic Failure.• Soft Aggregate = Coarse aggregate

fracture.

Refer to TRB Paper by Braham (2001)

Test Implementation – DCTEvaluation #2 – Aging and Aggregate Type

• Limestone Aggregate: LAR @ 500 = 32%

• Gravel Aggregate: LAR @ 500 = 15%

• All aging was loose mix at 135°C

0

100

200

300

400

500

600

700

800

Limestone 12.5mmE1

Limestone 19.0 mmE0.3

Gravel 12.5 mm E3

Frac

ture

Ene

rgy

(J/m

2)

Reheat 12 hour Aging 24 Hour Aging

Test Implementation – DCTEvaluation #3 – MnDOT Report

Test Implementation – DCTObservations and Discussion

• Recent discussion has suggested reducing fracture energy requirements from initial targets:

– Benefits: Accommodates “soft” aggregates such as limestone.– Risks: If hard aggregate is used there is potential that an inferior mix

(i.e. high binder replacement or low binder content) would still pass specification limit.

• Implications of changing/eliminating aggregate sources.• Recommended Action: Universal limit is not feasible. Compare to mixes

of known performance.

SCBIntermediate Temperature Cracking Test• Test Methods Evaluated

– ASTM D8044: LSU Procedure, 3 notch depths, 0.5 mm/min loading rate.– AASTHO TP 124: I-FIT, one notch depth, 50 mm/min loading rate.

• Factors Evaluated: Binder Replacement, Polymer Modification, Aging

0.00.20.40.60.81.01.2

20.0 25.0 30.0 35.0 40.0

Stra

in E

nerg

y (k

J)

Notch Depth (mm)

SCB - LSUAdjusting for a Northern Climate• Test temperature was adjusted to PG Inter. Temp to account for use of

softer grades in WI. • Three independent studies:

– WisDOT High RAM Pilot Program (AAPT 2015)– WHRP Performance Testing Feasibility Project (2016)– WHRP Durability Project – Bonaquist (2016 & 2016 AAPT Paper)

• Studies found the test was insensitive to the variables studied.• Example of the localized development of these tests and potential

complications in implementation.

SCB – D8044ILS 1424 – Phase 1

• Three samples:

• 12 laboratories

Steel Sample Plastic Sample Plastic Sample w/notchLoading Rate 0.5mm/min Loading Rate 0.5mm/min Loading Rate 0.5mm/min

Sampling Rate 10/sec Sampling Rate 10/sec Sampling Rate 10/secLoad limit 500N, 1000N, 2500N Load limit 1500N Load limit 1000N

Gage Length 127mm (5") Gage Length 127mm (5") Gage Length 127mm (5")Temperature Room Temperature Temperature Room Temperature Temperature Room Temperature

Pre-load 45± 10N Pre-load 45± 10N Pre-load 45± 10N

Procedure

Steel SamplePlastic SamplePlastic Sample w/notch

Loading Rate0.5mm/minLoading Rate0.5mm/minLoading Rate0.5mm/min

Sampling Rate10/secSampling Rate10/secSampling Rate10/sec

Load limit500N, 1000N, 2500NLoad limit1500NLoad limit1000N

Gage Length127mm (5")Gage Length127mm (5")Gage Length127mm (5")

TemperatureRoom TemperatureTemperatureRoom TemperatureTemperatureRoom Temperature

Pre-load45± 10NPre-load45± 10NPre-load45± 10N

ProcedureZero load, and load sample into fixture (with or without teflon)

Apply pre-load. Once achieved zero load again and zero displacement.

Start test and run until set load limit is reached.

Repeat at least 3 tests at each load limit

ReportGraph Load vs Displacement and add a linear trendline to each test to get slope.

Graph Displacement vs time to show displacement rate. Average entire displacement rate column to get average displacement rate.

Report slope for each test and average slope between all tests.

Template

Name:e-mail:

Lab Name:Tel. No.:

Date:

Testing Frame (manufacturer, model):

Temperature Chamber (if available: manufacturer, model):

Testing Temperature:Room or Temperature Chamber Controlled?

ASTM ILS 1424

Time ExtensionLoadDisplacement RateTime ExtensionLoadDisplacement RateTime ExtensionLoadDisplacement Rate

ERROR:#DIV/0!

ERROR:#DIV/0!

ERROR:#DIV/0!ERROR:#DIV/0!ERROR:#DIV/0!

ERROR:#DIV/0!ERROR:#DIV/0!

ERROR:#DIV/0!ERROR:#DIV/0!

Load vs Displacement

Run 1Run 2Run 3

Displacement (mm)

Load (N)

Displacement Rate

Run 1Run 2Run 3

Time (secs)

Displacement (mm)

SCB – D8044ILS 1424 – Phase 1

Results to date from 6 laboratories• Testing Devices

1- AMPT1- MTS2 – Brovold1 – Instron1- IPC-Global1 - Instrotek

• Testing Fixtures7 – fixed rollers3 – rollers with springs1 – 36mm roller

Material x¯ r-COV% R-COVValidator 300 24.3 57.0

Plastic with notch 58 26.6 102.1Plastic with notch 111 45.5 122.7

average 32.1 93.9

Sheet1

E691 Precision and Bias Analysis - SCB Strain Energy

Calibrator

Sample 1Sample 2Sample 3Sample 4Sample 5Average sdhk

LMLab 1308.2300.9300.9303.34.23.00.050.16

LALab 2325.8320.2328.5324.84.224.40.430.16

BLab 3323.6318.3323.7321.93.121.50.370.12

MILab 4311.8306.6307.8308.82.78.40.150.10

MI-sprLab 5445.9374.4366.6395.643.795.21.661.67

MI-spr3Lab 6338.3336.7336.2337.11.136.70.640.04

PGLab 7194.7211.8200.2202.28.8-98.1-1.710.34

IBLab 8234.6234.0231.5233.41.7-67.0-1.170.06

AILab 9278.5275.6275.1276.41.8-24.0-0.420.07

Average of Cell Averages, x¯300

Standard Deviation of Cell Averages, sx¯57COV%95% Conf

Repeatability Standard Deviation, sr268.724.3

Reproducibility Standard Deviation, sR6120.357.0

95% Repeatability Limit, r73

95% Reproducibility Limit, R171

Plastic- Notch

Sample 6Sample 7Sample 8Sample 9Sample 10Average sdhk

LMLab 157.658.353.756.52.5-1.0-0.00.5

LALab 265.860.963.063.22.55.70.30.5

BLab 340.139.239.439.50.5-18.0-0.90.1

MILab 466.264.463.564.71.47.20.40.3

MI-sprLab 562.759.158.260.02.42.50.10.4

MI-spr3Lab 669.466.665.567.22.09.60.50.4

PGLab 7100.688.1103.197.38.139.71.91.5

IBLab 823.724.224.024.00.3-33.5-1.60.0

AILab 945.544.945.445.30.3-12.2-0.60.1

Average of Cell Averages, x¯57.5

Standard Deviation of Cell Averages, sx¯20.5COV%95% Conf

Repeatability Standard Deviation, sr5.59.526.6

Reproducibility Standard Deviation, sR21.036.5102.1

95% Repeatability Limit, r15.3

95% Reproducibility Limit, R58.7

Plastic

Sample 11Sample 12Sample 13Sample 14Sample 15Average sdhk

LMLab 196.4150693.0825492.9418694.12.0-17.3-0.40.1

LALab 2132.338121.1045111.2045121.510.610.10.20.6

BLab 370.02864.016563.674565.93.6-45.5-1.00.2

MILab 4144.7496139.3661138.8242141.03.329.60.60.2

MI-sprLab 5141.6595134.0093136.3845137.43.925.90.60.2

MI-spr3Lab 6130.6145125.6587125.1613127.13.015.70.30.2

PGLab 7222.9531166.9393185.2165191.728.680.31.71.6

IBLab 833.93932.385531.325532.61.3-78.9-1.70.1

AILab 992.9132590.8160190.5595191.41.3-20.0-0.40.1

Average of Cell Averages, x¯111.4

Standard Deviation of Cell Averages, sx¯46.5COV%95% Conf

Repeatability Standard Deviation, sr18.116.245.5

Reproducibility Standard Deviation, sR48.843.8122.7

95% Repeatability Limit, r50.7

95% Reproducibility Limit, R136.7

h-value Tablek-value Table

Material AMaterial BMaterial CMaterial AMaterial BMaterial C

Lab 10.05-0.05-0.37Lab 10.160.450.11

Lab 20.430.280.22Lab 20.160.450.58

Lab 30.37-0.88-0.98Lab 30.120.090.20

Lab 40.150.350.64Lab 40.100.250.18

Lab 51.660.120.56Lab 51.670.440.22

Lab 60.640.470.34Lab 60.040.370.17

Lab 7-1.711.941.72Lab 70.341.481.58

Lab 8-1.17-1.64-1.69Lab 80.060.050.07

Lab 9-0.42-0.60-0.43Lab 90.070.060.07

Precision Statistics

2ds2ds2ds2ds

Materialx¯sx¯srsRrRr (Within)R (Between)r-COV%r-COV

A30057266173171206.8484.124.357.0

B5820521155943.2166.126.6102.1

C11147184951137143.4386.845.5122.7

Average131.1345.632.193.9

2ds2ds

Materialx¯r-COV%R-COV

Validator30024.357.0

Plastic with notch5826.6102.1

Plastic with notch11145.5122.7

average32.193.9

h-Table

Lab 1

Material AMaterial BMaterial C5.1552044853934574E-2-4.8448383874926754E-2-0.37106824316525122Lab 2

Material AMaterial BMaterial C0.426149695404568760.278249272821019580.21766506677663616Lab 3

Material AMaterial BMaterial C0.37486920986762523-0.87708043083420351-0.97779795233303879Lab 4

Material AMaterial BMaterial C0.146250791788091220.350206684163343880.63513251517749381Lab 5

Material AMaterial BMaterial C1.66243628809120560.120571159101740650.55716759794467885Lab 6

Material AMaterial BMaterial C0.640595044788427210.47054913084704780.33788956735954417Lab 7

Material AMaterial BMaterial C-1.71342504133332191.93908656010768991.7248982354574955Lab 8

Material AMaterial BMaterial C-1.169972583361035-1.6356339485416866-1.6944469798136654Lab 9

Material AMaterial BMaterial C-0.41845545009948981-0.5975000437900253-0.4294398074038871

k-Table

Lab 1Material AMaterial BMaterial C0.162117545330949940.450358818630673340.10860773900107845Lab 2Material AMaterial BMaterial C0.162401464209355970.454086834400570180.58415564191477265Lab 3Material AMaterial BMaterial C0.11792718696768168.7822342053975821E-20.19742421888573772Lab 4Material AMaterial BMaterial C0.104489111035639720.250120251171783270.18097540493214206Lab 5Material AMaterial BMaterial C1.6731138005923440.438147248168169250.21632182855251569Lab 6Material AMaterial BMaterial C4.1986667050792591E-20.37167559420356350.16657123319614361Lab 7Material AMaterial BMaterial C0.335479444028734311.47814434518806341.5780793109372613Lab 8Material AMaterial BMaterial C6.4229476606337252E-24.7066672069199025E-27.2620174096601778E-2Lab 9Material AMaterial BMaterial C6.9096434016894387E-25.8884991810675177E-27.1337333314237458E-2

SCB – D8044ILS 1424 – Phase 1

Results to date from 6 laboratories• Testing Devices

1- AMPT1- MTS2 – Brovold1 – Instron1- IPC-Global1 - Instrotek

• Testing Fixtures7 – fixed rollers3 – rollers with springs1 – 36mm roller

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

Material A Material B Material C

k-Table

SCB-IFITInitial Evaluations • Benefits

– Identifies mixes that are too stiff. – Verifies design vs. production– Provides a relatively easy way to evaluate mix composition.

• Concerns– Repeatability– Polymer modification (Discussed in Fall 2017 meeting)– Aging– Refining limits.

SCB - IFITRAP/RAS Content & Volumetrics

Mix Design

AB (%)

%AV at Ndes VMA VFA

RCY AB (%) ABR

RAP RAS Total RAP RAS Total

N50 5.8 3.6 15.1 73.5 1.2 0.8 2.00 20.3 14.0 34.3

N70 5.9 3.5 15.3 73.9 0.6 0.0 0.6 9.6 0 9.6

• Aggregate structure • Recycled products and ABR values for mix designs:

‒ N50 has 34% PBR, 40% of the binder replacement is from RAS.

Differences

SCB I-FITEffect of PBR - Flexibility Index

241 81 260

0.1 6

0.1 4

0.1 2

0.1 0

0.08

0.06

0.04

0.02

0.00

5.730 2.564 321 3.56 4.895 32

Mean StDev N

F

ytisneD

xednI ytilibixel

NepyT xiM

07N05

H lamroN

xednI ytilibixelF fo margotsi

SCB I-FitPotential Benefits to Monitor Production

Factors Levels

Mix Design Variables

Aggregate Source: Granite, Gravel, Limestone

Mix Traffic Level Medium Traffic and High Traffic

Production Variables

Asphalt Binder Content Design – 0.3%, Design, Design + 0.3%

P200 Content Design -2%, Design, Design + 2%

SCB I-FitPotential Benefits to Monitor Production

Factors Flexibility Index Post-Peak Slope Fracture Energy

Mix Type/Modification P-value Sig? P-Value Sig? P-value Sig?

Aggregate Source

SCB I-FIT ConcernsRepeatability – Single Lab

Base BinderPolymer

PG 64-22PG 58-28SBSNone51 7051 60SBSNone51 7051 60

1 2

1 0

8

6

4

2

Flex

ibili

ty In

dex

Boxplot of Flexibility Index

Base BinderPolymer

PG 64-22PG 58-28SBSNone51 7051 60SBSNone51 7051 60

30

25

20

1 5

1 0

Flex

ibili

ty In

dex

Boxplot of Flexibility Index

N50 N70

Four replicates tested for all mixes.

SCB I-FIT ConcernsAging

• Is Flexibility Index a discriminating property on long term aged samples?• Data represents ~12 designs. 4% AV and 3.0% Regressed AV contents.

SCB IFIT ConcernsEffects of Loading Rate and Aging

• After aging Flexibility Index values collapse due to stiffness effect.

ChallengesStandardization• HMA acceptance based on volumetrics causes plenty of disputes due

to multi-lab variability and differing practices.• Adoption of even simple performance tests introduces more

complexity.• Have been successful in generating test procedures.

– Need to understand precision and bias, ruggedness, etc.

• Assign testing responsibility and at what point in the process it will occur.

ChallengesImplementation Approach and Value Added

1. Maintaining current specifications and adding performance test requirements.a. Pro: Good for initial data gathering.b. Cons: Limited flexibility for mix adjustments.

Performance Test Limits

Volumetric Requirements

Long Term OutcomesChange in specifications based on initial performance test results.

OrBMD Approach #2: Relax volumetric criteria and add performance test requirements.

ChallengesOne Size Does Not Fit All – Setting Limits

• Dense Graded Mixes– Surface Layers vs. Lower Layers– Mix Traffic Level– Should effects of load/moisture be combined?

• Specialty Mixes– Interlayer, Thinlay, SMA

• Example– Flexibility Index = 7.0 (Dense Graded Mix) or 20.0 (SMA).

• Different Tests may be better suited for different applications.

Next StepsIdeal Process

Rutting (Stiffness)Cracking (Compliance)

Moisture DamageWhich test?

Aging Protocol?Aging Evaluation – Mixture or

Binder?

1. Selection of Binder Grade and Binder Content

2. Quality Assessment

Next StepsQuality Assessment - Gaps

• Moisture Damage– Combined with Rutting by using wet Hamburg with very high # of

passes.– Is specification promoting dry/stiff mixes?– Should effects of load/moisture be combined?

• Aging/Durability– Significant debate on which aging method to use and aging binder or

mix.– Many index cracking tests have not been developed at the levels of

aging currently under consideration.– Interim solution? Binder properties (i.e. ΔTc, G-R) have shown good

correlation to field performance.

Remarks/Discussion Points• There are benefits to single loading rate/single temperature

tests, but they cannot solve all problems. – Evaluates the mix as a system & Provides a control for mix

stiffness.• A solution for aging resistance is still a major research need.

– Accelerated load correlations indicate load associated cracking.• States are looking for guidance on how to incorporate these

tests into practice.

Thank You!Andrew Hanz, Ph.D.Technical DirectorMTE Services Inc.

andrew.hanz@mteservices.com

mailto:andrew.hanz@mteservices.com

References

1. Braham, A., Buttlar, W. G. & Marasteanu, M., 2007. Effect of Binder Type, Aggregate, and Mixture Composition on Fracture Energy of Hot-Mix Asphalt in Cold Climates. Transportation Reserach Record, Journal of the Transportation Research Board, Volume 2001, pp. 102-109.

2. Johanneck, L. et al., 2015. DCT Low Temperature Fracture Testing Pilot Project, St. Paul Minnesota: Minnesota Department of Transportation.

3. Bahia, et. al, “Analysis and Feasibility of Asphalt Pavement Performance Based Specifications for WisDOT”. Wisconsin Highway Research Program Proj. 0092-15-04. December 2016.

4. Mandal, T., Hanz, A. Bahia, H., “Challenges in using the DCT test to Determine the Role of Asphalt Mix Design Variables in Cracking Resistance at Low Temperatures.” International Journal of Engineering, Taylor & Francis 2017. ISSN: 1029-8436.

References

5. Hanz, A., Dukatz, E., Reinke, G., “Use of Performance-Based Testing for High RAP Mix Design and Production. Journal of Road Materials and Pavement Design, Volume 18: Papers from 91st AAPT Meeting, 2017.

6. Bonaquist, R., Paye, B., Johnson, C. “Application of InterdmediateTemperature SCB Test Results to Design Mixtures with Improved Load Associated Cracking Resistance.: AAPT Meeting 2017.

7. Bonaquist, Critical Factors Affect.ing Asphalt Concrete Durability. Wisconsin Highway Research Program Report 0092-14-06, WisDOT, 2016

8. Mandal, T., Hanz, A. Bahia, H., “Challenges in using the DCT test to Determine the Role of Asphalt Mix Design Variables in Cracking Resistance at Low Temperatures.” International Journal of Engineering, Taylor & Francis 2017. ISSN: 1029-8436.

MTE Experiences with Performance TestingDiscussion PointsPerformance Testing Experiences - MTEDisc Shaped Compact Tension (DCT) Test�Thermal Cracking ResistanceTest Implementation - DCT�Procedure and SpecificationTest Implementation – DCT�Evaluation #1 – Effect of Mix Design FactorsTest Implementation – DCT�Evaluation #1 – Effect of Mix Design FactorsTest Implementation – DCT�Evaluation #2 – Aging and Aggregate TypeTest Implementation – DCT�Evaluation #2 – Aging and Aggregate TypeTest Implementation – DCT�Evaluation #3 – MnDOT ReportTest Implementation – DCT�Observations and DiscussionSCB�Intermediate Temperature Cracking Test SCB - LSU�Adjusting for a Northern ClimateSCB – D8044�ILS 1424 – Phase 1SCB – D8044�ILS 1424 – Phase 1SCB – D8044�ILS 1424 – Phase 1SCB-IFIT�Initial Evaluations SCB - IFIT�RAP/RAS Content & VolumetricsSCB I-FIT�Effect of PBR - Flexibility IndexSCB I-Fit�Potential Benefits to Monitor ProductionSCB I-Fit�Potential Benefits to Monitor ProductionSCB I-FIT Concerns�Repeatability – Single LabSCB I-FIT Concerns�AgingSCB IFIT Concerns�Effects of Loading Rate and AgingChallenges�StandardizationChallenges�Implementation Approach and Value AddedChallenges�One Size Does Not Fit All – Setting LimitsNext Steps�Ideal ProcessNext Steps�Quality Assessment - GapsRemarks/Discussion PointsThank You!ReferencesReferences