Asphalt Compaction Evaluation using GPR:TH 52 and TH 14 Field Trial

Dr. Kyle Hoegh, MnDOTDr. Shongtao Dai, MnDOT

Mn/DOTOffice of Materials and Road Research

Pavement density has great effects on performance. Lack of density --- localized failure 1989 – “Effect of Compaction on Asphalt Concrete Performance” (Wash.DOT)

Each 1% increase in air voids (over 7 percent) tends to produce ~10 percent loss in pavement life.

Core used to determine density At a particular location, does not represent the entire

pavement density. Need a way to obtain full coverage of the surface

GPR is a potential tool: Continuous profile Locate relative high or low density areas based on

dielectric map

Why?

Data Collection Video3D Radar and Rolling Density

Meter

Equipment

Wave propagation in solids

Provides full coverage

Principal

Summer Testing Objectives Selected TH52 (D6) and TH14 (D6)Validate calibration methodology on large scale

pavements.Make recommendation for feasibility of

implementation. i.e. when it can and can’t be used Assess repeatability of the method

Gathering data necessary for specification development.

Project #1 (Summer 2016) TH52 (D6):

~7 miles M&O: Mill 1.5” and overlay 2x1.5” 4 Test Sections (FHWA funding)

No added binder + 4 rollers (control) Added binder (+0.3%) + 4 rollers No added binder + 5 rollers Added binder (+0.3%) + 5 roller

The entire 7 mile project was scanned 30 scans per foot (10 scan-4 in. moving average used in analysis) 3 antenna measurements per pass Core calibrations along the entire project were used to

develop a model relating RDM measurements to air void measurements

1.5”1.5”

3” Exist

TH14 (D6): 14 miles M&O: Mill 2” and overlay 2” and1.5” 4 Test Sections:

¾” mix + 3 rollers (control) ¾” mix + 4 rollers ½” mix + 3 & 4 rollers ½” mix (Evotherm) + 3 rollers ¾” mix (Evotherm) + 3 rollers

Scanned 11 Miles on Top lift 30 scans per foot (10 scan-4 in. moving average used in analysis) 3 antenna measurements per pass Core calibrations along the entire project were used to develop a

model relating RDM measurements to air void measurements

1.5”2”

4-5” Exist

Project #2 (Summer 2016)

General Process On-Site Identification of

high and low levels of compaction

TH14

CurrentMnDOTRDM

Relating Dielectric Measurements to Air Void Content

y = 15.652e-1.013x

R² = 0.6887

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

4.50 4.70 4.90 5.10 5.30 5.50 5.70 5.90

Core

Mea

sure

d Ai

r Voi

ds

RDM Measured Dielectric

TH52

Relating Dielectric Measurements to Air Void Content

3D Radar

UMN MN RDM

UMNMaine RDM

UMN Nebraska RDM

Uncertainty in Core Measured Air Voids

Hall, K. D., F. T. Griffith, and S. G. Williams. TRB Record No. 9 1761, pp. 81‐85.

Highway 52 Findings - Histogram All Data Collected

Sampling Rate = 0.4 in/scan.

> 26 million measurements Analysis based on 4 in.

moving average Equivalent to >1 million

cores Summary Stats

6.78% Median Air Voids 97.5% locations

less than 9%air voids

TH 52 – All Test Sections RDM RDM Comparison of Test Sections

Section with added binder+5 rollers has highest density

Control (4R+No Added Binder)5 R on No Added Binder

4 R on Added Binder

5R on Added Binder

TH 52 – All Test Sections Cores Comparison of Test Sections

Section with added binder+5 rollers has highest density in both cores and RDM measurements

Insignificant differences otherwiseControl (4R+No Added Binder) 5 R on No Added Binder 4 R on Added Binder 5R on Added Binder

Bot

h La

nes

mean 5.97 6.31 6.5 5.895% CI 0.56 0.74 0.86 1.66STD 0.95 1.07 0.98 1.6997.5th % 7.83 8.42 8.42 9.112.5th % 4.12 4.21 4.58 2.49n Data 11 8 5 4

TH 52 -Sorted Histograms Top lift Mainline Left

and Right Lane Summary: 6.4% (R) and 6.6%

(L)air voids respectivelyDensity:

Right: 93.6% Left: 93.4%

STD: 1% and 0.9% 97.5% locations

below 8.4%air voids forboth lanes

Blue – Right Lane (4’-8’)Red – Left Lane (4’-8’)

TH 52 -Sorted Histograms Top lift Joint Unconfined

and Confined Summary: 8.69%(UCJ) and 7.29%(CJ)

air voids, respectivelyDensity:

UCJ: 91.3%; CJ:92.7%; R = 98.5%

STD: 1.8% (UCJ) and 1.2%(CJ)

97.5% locations below 11.99%(UCJ) and 10.2%(CJ) air voids, respectively

Blue – Conf. J (3” offset)Red – Unconf. J (3” offset)

TH 52 -Sorted Histograms Top lift Mainline and

Confined Joint Summary: 6.59% (ML) and 7.29%(CJ)

air voids, respectivelyDensity:

CJ=92.7%; ML=93.4%; R=99.3%

97.5% locations below 8.37% and 9.69% air voids, respectively

Blue – Conf. J (3” offset)Red – Mainline

TH 52 -Sorted Histograms Top lift Mainline vs Confined and

Unconfined Joints Summary: 6.59% (ML), 7.29%(CJ) and

8.69(UCJ) air voids, respectively Density:

CJ=92.7%; UCJ=91.3%; ML=93.4%

UCJ/ML=97.7% (ML-UCJ=2.1%); CJ/ML=99.3% (ML-CJ=0.7%)

97.5% locations below 8.11%(ML), 9.89%(CJ) and 10.7(UCJ) air voids, respectively

Blue – Conf. JRed – MainlineYellow – Unconf. J

TH 52 -Sorted PlotsBreak data down by location

TH 52:Comparison with other FactorsImport RDM data into Veta for comparison with IC and other data

9.72 9.725 9.73 9.735 9.74 9.745 9.75 9.755 9.76x 104

3

3.5

4

4.5

5

5.5

6

6.5

7

Stationing [ft]

Die

lect

ric [

]

Local Increase after Added Roller

Local decreases (blue) at unconfined edges

dielectric

[A] [B] [C]

TH 52 – Experimental Design Results in Left Lane MainlineNo Significant difference in control mix from 4 rollers (red) to 5

rollers (blue), except initial jump

TH 52 – Comparison with other Factors

Import RDM data into Veta for comparison with IC and other data

9.72 9.725 9.73 9.735 9.74 9.745 9.75 9.755 9.76x 104

3

3.5

4

4.5

5

5.5

6

6.5

7

Stationing [ft]

Die

lect

ric [

]

Local Increase after Added Roller

Local decreases (blue) at unconfined edges

dielectric

1070+50 1071+00 1071+50 1072+00 1072+50 1073+00 1073+50 1074+00 1074+50

-7.5-6.25

-5.0-3.75

3D Radar Relative Compaction (3 in. X 3 in. spacing)

Rolling Density Meter Relative Compaction (2 ft. Spacing)

• 3D Radar can provide better coverage and precision at a tighter spacing. However RDM has advantages too:

• Less Expensive• Requires less user expertise (ex. Antenna

Correction)• RDM is easier to operate close to joints

when adjacent lanes are open to traffic• Real time results (easier for providing on-

site feedback)

Lower Compaction Higher Compaction

GPR: Asphalt Relative Compaction Assessment

TH 14 – All Test Sections Comparison of Test Sections

Adding a roller: compaction density increase on this project. Adding Evotherm: not much difference on compaction. Mix B (3/4-) to A(1/2-): not much difference on compaction.

Red – ¾”mix + 3 rollers [Control (4’-8’)]

Green - ¾” mix (Ev) + 3 rollers

Yellow - ½” mix + 3 rollers

Blue – ¾” mix + 4 rollers

Red – ¾”mix+ 3 rollers [Control (4’-8’)]

Green – ½” mix (Ev) + 3 rollers

Yellow - ½” mix + 4 rollers

Blue – ¾” mix + 4 rollers

TH 14 – All Test Sections Comparison of Test Sections (Core data)

TH 14 – All Test Sections

Percent within limits (PWL) implications Good measure if enough data: takes into account magnitude and

spread in data < 10 data points for each QA Core assessment:

Core PWL st.dev > 10 percent

>150,000 data points for each RDM assessment

http://www.fhwa.dot.gov/publications/research/infrastructure/pavements/pccp/04046/04046.pdf

FHWA-HRT-04-046 Evaluation of Procedures for Quality Assurance Specifications

TH 14 -Sorted Histograms 3 Roller on ¾” Mix:

Confined Joint and Mainline.

Blue: Conf J. side(3”offset)

Red: Mainline 4-8’

Summary Stats 5.23 (ML) and

5.06(CJ) dielectric, respectively

STD: 0.11 and 0.11, respectively

3.6% higher dielectric in mainline CJ/ML=96.7%

TH 14 -Sorted Histograms

Blue: conf.J (3”offset)

Red: Mainline (4’to8’)

Summary Stats 5.22 (ML) and

5.20(CJ) dielectric, respectively

STD: 0.10 and 0.09, respectively

0.4% higher dielectric in mainline CJ/ML=99.6%

3 Roller on ¾”mix( Evotherm):Confined Joint and Mainline.

TH 14 -Sorted Histograms4 Roller on ¾” Mix: Confined Joint and Mainline.

Blue: Conf.J side(3”offset)

Red: Mainline 4-8’

Summary Stats 5.40 (ML) and

5.29(CJ) dielectric, respectively

STD: 0.10 and 0.12, respectively

2.1% higher dielectric in mainline CL/ML=97.9%

TH 14 -Sorted Histograms

Blue: Conf.J (3”offset)

4 Roller on ½” Mix:Confined Joint and Mainline.

Red: Mainline 4-8’

Summary Stats 5.38 (ML) and 5.35

(CJ) dielectric, respectively

STD: 0.09 and 0.09, respectively

0.6% higher dielectric in mainline CJ/ML=99.4%

Implementation Recommendations:

Longitudinal Joint (RDM)

Recommendation #1 - LJ: Require dielectric distribution readings from RDM per 500ft.

Ex: Require E of 5.31 >= 90% density? E of 5.31 includes > 95% data Take cores at E=5.31, Then measure density

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.9

95

.03

5.0

75

.11

5.1

55

.19

5.2

35

.27

5.3

15

.35

5.3

95

.43

5.4

75

.51

5.5

55

.59

5.6

35

.67

5.7

15

.75

5.7

95

.83

5.8

75

.91

5.9

5

% o

f to

tal m

ea

sure

me

nts

in e

ach

dat

ase

t

Dielectric

Recommendation #2 - LJ: Require RDM readings at the logitudinal Joint and X distance away from the Joint, use % difference of dielectric. (No cores required)

Ex: Require Joint E >= 92% of mainline E ?

4

4.5

5

5.5

6

6.5

0 20 40 60 80 100 120

Relat

ive Pe

rmitt

ivity,

e

distance, ft

S5 -3.0 ft EB

S5 -3.0 ft EB

S5 -1.0 ft EB (unconfined)

S5 -1.0 ft EB (unconfined)

Core Location

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.8 5 5.2 5.4 5.6 5.8 6 6.2

S5 -1.0 ft EB

S5 -3.0 ft EB

Unconfined Joint Median

3 ft from Joint Median

dEJoint 3' away (mainline)

91.45% different

Implementation Recommendations:

Mainline – 3D GPR

Recommendation #3A - M(3DGPR)Survey the whole project surfaceRequire dielectric distribution from 3DGPR every 500ft.Example A: Require E of 5.2 = 92% density?

E of 5.2 includes > 95% datatake cores at E=5.2, then measure density

Surface Arrival Amplitude

Longitudinal Distance from Starting Point, feet

Tru

ck L

ane

<--

----

----

--->

Cen

ter

Lane

Tra

nsve

rse

Dis

tanc

e fr

om th

e Lo

ngitu

dina

l Joi

nt, f

eet Dielectric Map

47 94 141 188 235 282 329 376

-2.25

-1.75

-1.25

-0.75

-0.25

0.25

0.75

1.25

1.75

2.25 4

4.2

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8

6

DielectricConstant

Recommendation #3B - Mainline(3DGPR):Require a test strip to establish the dielectric histogramand establish E and density relationshipExample B: E of 5.2 = 92% density

E of 5.2 includes > 95% datatake cores at E=5.2, then measure density

Require to construct test strip where material changesThen use established histogram for project acceptance: dielectric

distribution on mainline should be similar. 95% data should > 5.2. (No cores required) No OK

Summary GPR can provide a continuous coverage of the relative

compaction levels (higher dielectric = higher compaction)

Histograms and general statistics can be used to give a complete assessments of the in-place compaction

Potential Uses: Assess compaction uniformity for QC/QA. Provide on-site feedback to contractor of high and low compaction

locations that they can cross-check with differences in mix or paving strategies in those locations to determine optimal construction procedures

Identification of trends in the air void content maps that can be cross-checked with IC and other data to determine the most critical factors in achieving higher density