www.dot.state.pa.uswww.dot.state.pa.us

Moisture Sensitivity TestingPAPA Conference 2016

Neal Fannin&

Dennis Morian

www.dot.state.pa.us

Stripping

www.dot.state.pa.us

Indications of a Problem

• Had severe and obvious problems with the premature deterioration of some asphalt mix designs with certain aggregates.

• Districts with moisture damage issues were forced to require minimum amounts of liquid anti-strip additives to mitigate moisture damage problems.

• Mix designs in border areas with other states needed anti-strip when used in other states but not in PA.

• No mix designs ever seemed to fail PA modified AASHTO T283 testing needing anti-strip.

www.dot.state.pa.us

Research Project Started

• COST BENEFIT ANALYSIS OF ANTISTRIP-ADDITIVES IN HOT MIX ASPHALT WITH VARIOUS AGGREGATES research started 2011.

• Final report May 2015.

www.dot.state.pa.us

• Hand off to Denny

Q E S

Factors for Moisture Related DamageMoisture in pavementF-T cyclesTraffic loadingMoisture susceptible materialsAll of these factors can be found in PA

Q E S

Study Objectives

Evaluate accuracy of moisture resistance testingPerform LCCA and CBA on moisture

resistance testing and antistrip (AS) usage under different scenarios

Q E S

Study Phases

Part One: Introduction/Literature Review, Lab TestingPart Two: Life cycle cost analysis

(LCCA)Part Three: Cost benefit analysis (CBA)

and Conclusions

Q E S

PART ONE: INTRODUCTION/ LITERATURE REVIEW/ LABORATORY TESTING

Q E S

Test ProcedureNo. of States

Lottman (NCHRP 246) 3Tunnicliff & Root (ASTM D 4867) 6Modified Lottman (AASHTO T 283) 30

Immersion-Compression (ASHTO T 165) 5Hamburg Wheel Tracking 2

Table 1. HMA Moisture Resistance Tests in Use by State Agencies as of 2003 (Aschenbrenner, 2003).

Q E SFrom Hicks et al., Moisture Sensitivity in Asphalt Pavements—a National Seminar, 2003

Q E S

State Predominant Aggregate Type(s) Used in HMA

Verified Mandate Continues in 2012

Colorado Crushed gravel YesGeorgia Crushed granite YesIdaho Crushed gravel YesMississippi Crushed gravel YesMontana Crushed gravel YesNorth Carolina Crushed gravel and

crushed graniteYes

South Carolina Crushed granite YesSouth Dakota Crushed gravel South Dakota does not currently require

antistrip additive in all HMA mixesVirginia Crushed

limestone/dolomite (not in surface courses), granite, traprock and/or sandstone

Yes

Wyoming Crushed gravel Requirements for addition of hydrated lime are apparently done on a project-by-project basis; overall extent of mandated use is not clear

Table 5. States Requiring the Use of AS in HMA(Aschenbrenner, 2003; USGS, 2008).

Q E S

Why AASHTO T 283?

No moisture resistance test is highly accurate (say, 90 % or higher)Accuracy of new procedures not well

documentedAccuracy of T-283 is generally better

than most people think, and is well documented

Q E S

Significance of Moisture Resistance Testing ErrorsType I error

– “Good” mixes that fail testing– Cost of error is unnecessary AS

Type II error– “Bad” mixes that pass testing– Cost of error is high maintenance,

premature failure– Much costlier than type I error

Q E S

Materials: Sampling Matrix

District(s)Other

State(s)

Limestone, Dolomite,

Etc.Crushed Gravel

Diabase, Granite, Gneiss

Sandstone, Quartzite

1 OH 12, 9 7 3 23, 4 NY 1 3 3

5, 6, 8 MD, NJ, VA

3 1

10, 1112 WV 1

Q E S

Laboratory Testing Summary

Stripping Potential CategoryLow

SaturationHigh

SaturationPass Fail Pass Fail

L—Low 6 0 6 0M—Moderate 2 0 1 1H—High 8 0 2 6HA—High, tested with antistrip as designed/approved 10 0 9 1

MA—Moderate, tested with antistrip as designed/approved 1 0 1 0

U—Unknown, tested without antistrip as designed/approved 17 0 17 0

UA—Unknown, tested with antistrip as designed/approved 1 0 1 0

Q E S

Accuracy of High SaturationT 283 Testing (This project)Type I error rate: 0 %Type II error rate

– 25 % for highly susceptible aggregates– 50 % for moderately susceptible

aggregates

Q E S

Accuracy of High SaturationT 283 Testing: Overall, Including Results of From Similar Research

Type I error rate: 6 %Type II error rate

– 23 % for highly susceptible aggregates– 62 % for moderately susceptible

aggregates

Advanced Asphalt Technologies, LLC

“Engineering Services for the Asphalt Industry”

Cost Benefit Analysis Of Antistrip-Additives In HMA With Various Aggregates

Part 2: Life Cycle Cost Analysis

Dennis Morian, P.E.;

Guangming Wang, Ph.D, P.E.

Q E S

What is a Life Cycle Cost Analysis?A means of comparing alternatives

including all costs anticipated within the defined analysis period. Often expressed as Present Worth-

accounts for the time value of money. For this project, EUAC format was

used.

Q E S

LCCA Approach and Scenarios LCCA Approach

1)1()1(

n

n

iiNPVEUAC

k

n

kk i

PMCICNPV)1(

11

Conversion between NPV and EUAC

Q E S

Fixed Input Values

Variable ValueDiscount Rate* 2%Analysis Period (Years) 24Assumed Project Length (Mile) 1Lane Width (Feet) 12HMA Density (lb/sy/in) 110

Asphalt Adjustment Multiplier (AAM)*(values from ECMS)

1.12

Q E S

Traffic Inputs Based on SRL Levels

SRL Level

Two-way ADT % Truck

Lanes in One

Direction

Divided Roadway

E 50,000 15 (20% S.U. and 80% Comb.) 2 Yes

H 15,000 12 (20% S.U. and 80% Comb.) 2 Yes

G 5,000 10 (20% S.U. and 80% Comb.) 1 No

M 2,000 8 (20% S.U. and 80% Comb.) 1 No

Q E S

Two Analysis ConceptsControl model- expected normal

performanceRealistic model

– Highly susceptible – Moderately susceptible– Each with and without antistrip

Optimistic model-similar matrix, but assumes less reduction in performance of susceptible mixes

Q E S

Why Does This Have to be So Complicated?B/C analysis and LCCA are both

potentially very sensitive to assumptionsWe have made a wide range of

assumptions to determine how these effect the results of the analysis If the results are highly sensitive to the

assumptions, they are possibly suspect…

Q E S

Performance Assumptions w/ ExperimentalDesign: Performance Cycle Assumptions

GeneralPerformanceAssumption forSusceptible Mixes Antistrip

Resistant Mixes Highly Susceptible Mixes Moderately Susceptible Mixes

Performance Cycles: Performance Cycles: Performance Cycles:

No. Duration Total No. Duration Total No. Duration Total

RealisticWithout 2 12 24 4 6 24 3 8 24

With N/A N/A N/A 3 8 24 2 12 24

OptimisticWithout 2 12 24 3 8 24 2 12 24

With N/A N/A N/A 2 12 24 2 12 24

Maintenance & Rehab. activities and cycles followed LCCA guidelines of Pub.242.

Each maintenance activity triggered a user cost, which is affected by ADT. An analysis period of 24 years was selected based on comments from the

project panel. A 2% discount rate was selected based on PennDOT Memo dated on February

10, 2014. All costs were converted to equivalent annual uniform cost (EAUC) values to

facilitate the comparison of alternatives

Q E S

LCCA Results for Realistic Scenario

Comparison of EAUC (Excluding User Cost) Among Realistic Scenarios

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000

$35,000EA

UC

($/L

ane-

mi.)

SRL-M $14,259 $27,100 $22,162 $21,469 $14,308SRL-G $14,750 $28,020 $22,887 $22,194 $14,799SRL-H $15,143 $28,757 $23,468 $22,775 $15,192SRL-E $15,535 $29,493 $24,048 $23,355 $15,584

C RHN RHS RMN RMS

C= no moisture damage; RHN= highly susceptible mix, no antistrip; RHS = highly susceptible mix with antistrip; RMN = moderately susceptible mix, no antistrip; RMS = moderately susceptible mix with antistrip.

Q E S

LCCA Results for Optimistic ScenarioComparison of EAUC (Excluding User Cost) Among Optimistic Scenarios

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000EA

UC

($/L

ane-

mi.)

SRL-M $14,259 $22,089 $14,980 $14,259 $14,308SRL-G $14,750 $22,815 $15,471 $14,750 $14,799SRL-H $15,143 $23,395 $15,863 $15,143 $15,192SRL-E $15,535 $23,976 $16,256 $15,535 $15,584

C OHN OHS OMN OMS

C= no moisture damage; RHN= highly susceptible mix, no antistrip; RHS = highly susceptible mix with antistrip; RMN = moderately susceptible mix, no antistrip; RMS = moderately susceptible mix with antistrip.

Q E S

LCCA Results for Sensitivity Analysis

Impact of Discount Rate on EAUC (Excluding User Cost)-SRL(H)

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000

$35,000

$40,000EA

UC

($/L

ane-

mi.)

D1T1 $15,143 $28,757 $23,468 $22,775 $15,192D2T1 $17,559 $31,061 $25,553 $24,910 $17,615D3T1 $20,179 $33,474 $27,805 $27,214 $20,242

C RHN RHS RMN RMS

C= no moisture damage; RHN= highly susceptible mix, no antistrip; RHS = highly susceptible mix with antistrip; RMN = moderately susceptible mix, no antistrip; RMS = moderately susceptible mix with antistrip.

Q E S

LCCA Results for Sensitivity Analysis

Impact of Traffic Growth Rate on EAUC (Excluding User Cost)-SRL(H)

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000

$35,000EA

UC

($/L

ane-

mi.)

D1T1 $15,143 $28,757 $23,468 $22,775 $15,192D1T2 $15,143 $28,757 $23,468 $22,775 $15,192D1T3 $15,143 $28,757 $23,468 $22,775 $15,192

C RHN RHS RMN RMS

C= no moisture damage; RHN= highly susceptible mix, no antistrip; RHS = highly susceptible mix with antistrip; RMN = moderately susceptible mix, no antistrip; RMS = moderately susceptible mix with antistrip.

Q E S

PART THREE: COST BENEFIT ANALYSIS AND CONCLUSIONS

Q E S

Cost Benefit AnalysisAnalyze all costs and benefits of a

possible decisionUse net present value (NPV) or

equivalent annual uniform cost (EAUC)B/C = 1 at break evenB/C > 1 indicates economical outcomeB/C < 1 indicates uneconomical

Q E S

CBA of Moisture Resistance Testing and Antistrip UsageCost = cost of testing and cost of AS;

relatively smallBenefit = increased pavement life, lower

maintenance costsCBA must consider costs, benefits and

accuracy of testing

Q E S

Conditional vs. Mandatory Antistrip UsageConditional use means AS only used

when mix fails test without itMandatory use means AS is used in all

mixes, regardless of test outcome– Mixes still need to pass moisture

resistance testing

Q E S

CBA Sensitivity Analysis

Discount rateTraffic growth ratePercentage of susceptible aggregatesRelative performance of poor

aggregates Improvement resulting from antistripMandatory vs. conditional AS usageWith and without user delay costs

Q E S

Accuracy of High SaturationT 283 Testing (Overall)Type I error rate: 6 %Type II error rate

– 23 % for highly susceptible aggregates– 62 % for moderately susceptible

aggregates

Q E S

Susceptibility of Aggregate to Moisture DamageUSGS Yearbook, 201390 % of aggregate production for HMA

in PA is crushed stone10 % is crushed gravelOf Crushed stone

– 65 % is limestone/dolomite– 35 % granite, diabase, sandstone, etc.

Q E S

Final Estimates of Resistant/Susceptible Aggregates70 % resistant

– 80 % limestone/dolomite/other resistant aggregates

– Reduced by blending for skid resistance20 % moderately susceptible, other

crushed stone and aggregate blends10 % highly susceptible, all crushed

gravel, some other aggregates

Q E S

B/C Ratio, without User Delay Costs, Realistic Scenario

Q E S

B/C Ratio, with User Delay Costs, Realistic Scenario

Q E S

B/C Ratio, without User Delay Costs, Optimistic Scenario

Q E S

B/C Ratio, with User Delay Costs, Optimistic Scenario

Q E S

B/C Ratio, w/o Delay Costs, Realistic Scenario, Discount Rates

Q E S

Benefit/Cost vs. Overall Savings

Benefit/cost ratio should only be used to determine whether or not a given course of action is economical—B/C ratios should not be used to decide among viable optionsAmong viable options (B/C >> 1),

selection should be based on net savings or similar criteria

Q E S

Saving from Antistrip Usage, without User Delay Costs

Scenario/Savings% Susceptible Aggregates

40 20 10Realistic/Conditional $8,000,000 $4,000,000 $1,900,000Realistic/Mandatory $14,700,000 $7,200,000 $3,400,000Savings Mand./Cond. $6,700,000 $3,200,000 $1,500,000Percent 6.0 3.2 1.6Optimistic/Conditional $6,600,000 $3,300,000 $1,600,000Optimistic/Mandatory $8,500,000 $4,100,000 $1,800,000Savings Mand./Cond. $1,800,000 $800,000 $300,000Percent 1.9 0.8 0.3

Q E S

Conclusions

B/C of low saturation testing is zeroB/C of high saturation testing with AS

usage always much greater than oneB/C for both conditional and mandatory

AS usage both much greater than one

Q E S

Conclusions

A critical issue affecting results of this analysis is the effectiveness of antistrip additives in actual pavements, as compared to laboratory tests…

www.dot.state.pa.us

Test result that told us we had a problem

Test ResultMoisture Resistance of Aggregates in Mix

Good Moderate Poor

Passed 3 1 5

Failed 0 0 0

Error Rates Type I Type II

0 % 100 % 100 %

www.dot.state.pa.us

Overall Accuracy of Modified Lottman Procedure, Level 2 Severity as Reported in Literature

Test ResultStripping Potential of Aggregates in Mix

Low Moderate High

Passed 18 8 5.5

Failed 1 5 17.5

Error Rates Type I Type II

5 % 61 % 25 %

www.dot.state.pa.us

Action Taken

• Letter to all producers of bituminous mixtures dated October 20,2014.– Requires all

SRL – E and H, 9.5 mm NMAS mixes must be revaluated before being used in the 2015 construction season.– Any new mix designs submitted must also be evaluated

under the new requirements.– Districts that currently require minimum anti-strip

amounts for certain aggregate types will continue to require them.

www.dot.state.pa.us

2016 & 2017 Requirements

• All wearing and binder mixes approved in 2016 must be reevaluated using the revised moisture susceptibility criteria in order to be approved in 2016.

• All Base mixes approved in 2017 must be reevaluated using the revised moisture susceptibility criteria in order to be approved in 2017.

www.dot.state.pa.us

Anti-strip Research Implementation

• Current Specials / Requirements

– Districts 1, 2, 4, & 9 Using special that requires:

• 0.25% if coarse aggregate used is gravel, sandstone, siltstone, calcareous sandstone.

• 0.25% if fine aggregate used is gravel, sandstone, siltstone, calcareous sandstone.

• 0.5% if both fine and coarse aggregate used is gravel, sandstone, siltstone, calcareous sandstone.

– Warm mix requires 0.25% anti-strip.

www.dot.state.pa.us

Anti-strip Research Implementation

• Proposed change– Require 0.25% (or minimum required by manufacturer

of anti-strip) in all asphalt mixes.• Will not require additional testing of mixtures.

– Foamed warm mix would not require additional testing but would contain the same anti-strip as the hot mix parent.

www.dot.state.pa.us

Anti-strip Research Implementation

• Proposed change for high stripping potential aggregates.

– gravel, sandstone, siltstone, slag, quartz, or shale

– Producers may add higher dosage and avoid testing (typically 0.5%).

www.dot.state.pa.us

Anti-strip Research Implementation

– Producer may choose additional AASHTO T283 testing for mixes with both coarse and fine aggregates that are gravel, sandstone, siltstone, slag, quartz, or shale.

– If all of the following are true the mix must have the higher dosage rate. (0.5%)

• The TSR of the higher dosage mixture is greater than the minimum dosage mixture.

• The unconditioned tensile strength of all AASHTO T 283 tests are above the minimum tensile strength requirement in section 11.3

• The unconditioned tensile strength of the higher dosage mixture is higher or within 5 psi below the unconditioned tensile strength of the lower dosage mixture.

www.dot.state.pa.us

Anti-strip Research Implementation

• Proposed change

– Computation of anti-strip dosage.• Based on virgin asphalt for mixes with RBR of 0.15 or less.

(Tier 1 of Appendix H)• Based on total asphalt for mixes over 0.15 RBR. (Tier 2 or

3 of Appendix H)

www.dot.state.pa.us

Anti-strip Research Implementation

• Please comment on the CT 1.

• Comments due - 2/4/2016

www.dot.state.pa.us

Sources

• C, Ivan Harnish, ArrMaz Custom Chemicals, 2/3/2010 PowerPoint

• Kevin Gnegy, District 9-0, Garth Bridenbaugh, Q.A., John Swalligan, District 2-0.

• Don Christensen, Advanced Asphalt Technologies, LLC

• Dennis Morian, Quality Engineering Solutions, Inc.

www.dot.state.pa.us

Questions?