University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Nebraska Department of Transportation Research
Nebraska LTAP
Reports

6-2011

Non-Nuclear Method for Density Measurements

Yong-Rak Kim
University of Nebraska-Lincoln, yong-rak.kim@unl.edu

Yong Cho
University of Nebraska - Lincoln

Follow this and additional works at: https://digitalcommons.unl.edu/ndor

Part of the Transportation Engineering Commons

Kim, Yong-Rak and Cho, Yong, "Non-Nuclear Method for Density Measurements" (2011). Nebraska Department of Transportation
Research Reports. 143.
https://digitalcommons.unl.edu/ndor/143

This Article is brought to you for free and open access by the Nebraska LTAP at DigitalCommons@University of Nebraska - Lincoln. It has been
accepted for inclusion in Nebraska Department of Transportation Research Reports by an authorized administrator of DigitalCommons@University of
Nebraska - Lincoln.
 Nebraska
Transportation
Center

Report # SPR-1(10)P335 Final Report

WBS# 26-1116-0050-001

NON-NUCLEAR METHOD FOR DENSITY MEASUREMENTS

Dr. Yong Cho

Assistant Professor
Construction Systems Engineering
University of Nebraska-Lincoln

Dr. Yong-Rak Kim

NTC
Nebraska Transportation Center
262 WHIT
2200 Vine Street
Lincoln, NE 68583-0851
(402) 472-1975

“This report was funded in part through grant[s] from the Federal Highway Administration [and Federal Transit Administration], U.S.
Department of Transportation. The views and opinions of the authors [or agency] expressed herein do not necessarily state or reflect those
of the U. S. Department of Transportation.”
 Non-Nuclear Method for Density Measurements

Principal Investigator Co-Principal Investigator

Dr. Yong Cho Dr. Yong-Rak Kim
Assistant Professor Associate Professor
Construction Systems Engineering Department of Civil Engineering
University of Nebraska-Lincoln University of Nebraska-Lincoln

Researchers
Koudous Kabassi
Ziqing Zhuang
Heejung Im
Chao Wang
Thaddaeus Bode

A Report on Research Sponsored By:

Nebraska Department of Roads

Lincoln, Nebraska

June 2011
1. Report No. 2. 3. Recipient‘s Accession No.
SPR1 (10) P335
4. Title and Subtitle 5. Report Date
Non-nuclear Method for Density Measurements June 2011

6.

7. Author(s) 8. Performing Organization Report No.

Yong Cho, Koudous Kabassi, Ziqing Zhuang, Heejung Im, Chao
Wang, Thaddaeus Bode, and Yong-Rak Kim.

9. Performing Organization Name and Address 10. Project/Task/Work Unit No.

The Charles W. Durham School
of Architectural Engineering & Construction
1110 S 67th St.
Omaha, NE 68182 11. Contract © or Grant (G) No.

12. Sponsoring Organization Name and Address 13. Type of Report and Period Covered
Nebraska Department of Roads (NDOR)
1400 Highway 2
PO BOX 94759
14. Sponsoring Agency Code
Lincoln, NE 68509
26-1116-0050-001

15. Supplementary Notes

16. Abstract
Quality control (QC) and quality assurance (QA) are necessary to ensure fulfillment and compliance to specifications, guidelines, manuals,
and programs which outline methods and requirements during construction. Density, an important part of quality control, can be used to
evaluate the quality of Hot Mix Asphalt (HMA) and soil compaction. This study investigated new technologies used for QC and QA by
comparing the Pavement Quality Indicator (PQI) model 301 with a nuclear gauge and core sample measurements for HMA. For soil QC and
QA, non-nuclear technologies—the Electrical Density Gauge (EDG), the Moisture Density Indicator (MDI), and the Light Weight
Deflectometer (LWD)—were also investigated against a nuclear gauge and traditional non-nuclear methods of measurement. Overall, the
nuclear gauge shows higher accuracy and higher correlation with cores than the non-nuclear gauges tested in this study. A thorough
investigation of calibration methods was also performed, both in the lab and on the field, to improve the accuracy of the PQI‘s results. Data
analyses showed that the accuracies of the non-nuclear soil gauges are somewhat lower than that of the nuclear gauge. With an improved
methodology to create soil models for the EDG and standardized ways to develop the LWD‘s target values, the EDG and LWD could have a
similar or better accuracy than the nuclear gauge. With the EDG and the Soil Density Gauge (SDG), both recently ASTM approved, non-
nuclear soil technology is the future. Furthermore, the non-nuclear gauges could be a better alternative to a nuclear gauge when the following
benefits are considered: (1) economic savings; (2) faster data measurement (PQI); (3) elimination of intense federal regulations and safety
concerns; (4) elimination of licensing and intense training.
17. Document Analysis/Descriptors 18. Availability Statement
Hot Mix Asphalt (HMA), soil, nuclear gauge, quality assurance, quality control

19. Security Class (this report) 20. Security Class (this page) 21. No. of Pages 22. Price
Unclassified Unclassified 75

ii
 Table of Contents

Chapter 1 Introduction .................................................................................................................... 1

Chapter 2 Hot-Mix Asphalt ............................................................................................................ 4
2.1 Methodology ......................................................................................................................... 4
2.2 Literature Reviews ................................................................................................................ 4
2.3 Testing Methods.................................................................................................................... 6
2.3.1 Core Method .................................................................................................................. 6
2.3.2 Nuclear Method ............................................................................................................. 7
2.3.3 Non-Nuclear Method (PQI) ........................................................................................... 8
2.3.4 Calibration...................................................................................................................... 9
2.4 Data Analysis ...................................................................................................................... 11
2.4.1 Outliers......................................................................................................................... 12
2.4.2 Poor Core Samples ....................................................................................................... 13
2.4.3 Average Difference ...................................................................................................... 14
2.4.4 Student T-Test .............................................................................................................. 14
2.4.5 Coefficient of Correlation ............................................................................................ 16
2.4.6 Coefficient of Determination ....................................................................................... 17
2.4.7 Coefficient of Determination based on average density from each site ...................... 18
2.4.8 Error of the Standard deviation .................................................................................... 19
2.4.9 Data Reliability ............................................................................................................ 20
2.4.10 Number of Cores for PQI Calibration ............................................................................ 24
2.5 Conclusions ......................................................................................................................... 26
Chapter 3 Soil ............................................................................................................................... 28
3.1 Methodology ....................................................................................................................... 28
3.2 Literature Review................................................................................................................ 28
3.3 Testing Methods.................................................................................................................. 30
3.3.1 Density-Drive Cylinder ................................................................................................ 30
3.3.2 Water Content of Soil by Oven Dry method ............................................................... 31
3.3.3 Nuclear Method ........................................................................................................... 31
3.3.4 The Electrical Density Gauge (EDG) .......................................................................... 32
3.3.5 The Moisture Density Indicator (MDI)........................................................................ 33
iii
 3.3.6 The Light Weight Deflectometer (LWD) .................................................................... 35
3.4 Test Procedures and Methodology...................................................................................... 36
3.5 Data Analysis ...................................................................................................................... 38
3.5.1 Outlier .......................................................................................................................... 38
3.5.2 Coefficients of Correlation (R) and Determination (R2) analyses ............................... 39
3.5.3 Average Difference and Error of Standard Deviation ................................................. 41
3.5.4 Site by Site Analysis .................................................................................................... 44
3.5.5 Test Status Analysis ..................................................................................................... 48
3.6 Conclusions ......................................................................................................................... 50
Chapter 4: Economic Analysis...................................................................................................... 51
4.1 Life Cycle Costs .................................................................................................................. 51
4.2. Analysis.............................................................................................................................. 53
Chapter 5 Conclusions and Recommendations............................................................................. 56
Chapter 6 NDOR Implementation Plan ........................................................................................ 57
References ..................................................................................................................................... 58
Appendix A: Site 1........................................................................................................................ 62
Appendix B: Site 2 ........................................................................................................................ 63
Appendix C: Site 3 ........................................................................................................................ 64
Appendix D: Site 4........................................................................................................................ 65
Appendix E: Site 5 ........................................................................................................................ 66
Appendix F: Site 6 ........................................................................................................................ 67
Appendix G: Site 7........................................................................................................................ 68
Appendix H: Site 8........................................................................................................................ 69
Appendix I: Site 9 ......................................................................................................................... 70
Appendix J: Site 10 ....................................................................................................................... 71
Appendix K: Site 11...................................................................................................................... 72
Appendix L: Site 12 ...................................................................................................................... 73
Appendix M: Site 13 ..................................................................................................................... 74
Appendix N: Calibration results of different numbers of cores .................................................... 75

iv
 List of Figures

Figure 2.1. Dry ice cools the hot pavement 7

Figure 2.2. Nuclear Gauge is shown measuring density 8

Figure 2.3. PQI model 301shown taking measurements 9

Figure 2.4. On-site set up for core measurements 11

Figure 2.5. Lab set up for core measurements 11

Figure 2.6. PQI‘s relationship with core samples before removing the outliers 12

Figure 2.7. PQI‘s relationship with core samples after removing the outliers 13

Figure 2.8. Example of rejected cores ...........................................................................................13

Figure 2.9. Example of an accepted core .......................................................................................14

Figure 2.10. PQI & Nuclear Gauge relationship vs. Core Samples ...............................................18

Figure 2.11. Relationship between each gauge and core samples .................................................19

Figure 2.12. Absolute Density Differential Variation for both gauges 20

Figure 2.13. PQI & Nuclear Gauge Data reliability 21

Figure 2.14. Comparison between traditional difference and calibration difference 26

Figure 3.1. Shelby Tube driven in the ground 31

Figure 3.2. Nuclear gauge taking soil measurements ....................................................................32

Figure 3.3. Darts driven into ground for EDG test 33

Figure 3.4. MDI connected to a laptop for density measurements 34

Figure 3.5. LWD is measuring stiffness of the soil 36

Figure 3.6. Highway 370 site 37

Figure 3.7. Platteview Intersection Site 37

Figure 3.8. Standard vs. Nuclear Gauge and EDG Density 39

v
Figure 3.9. Standard vs. Nuclear Gauge and EDG Moistures 40

Figure 3.10. Standard vs. Nuclear Gauge (before correction factors) and EDG Density 41

Figure 3.11. Average Density Difference of gauges compared to standard 42

Figure 3.12. Variation of Nuclear Gauge and EDG Density 42

Figure 3.13. Variation of Nuclear gauge and EDG Moisture Content (%) 43

Figure 3.14. Nuclear Gauge and EDG vs. Standard Density for Site 1 44

Figure 3.15. STVD Density Errors for Site 1 45

Figure 3.16. Nuclear Gauge and EDG vs. Standard Moisture Contents for Site 1 45

Figure 3.17. STDV Moisture Content (%) Errors for Site 1 46

Figure 3.18. Nuclear Gauge and EDG vs. Standard Density for Site 2 46

Figure 3.19. STVD Density Errors for Site 2 47

Figure 3.20. Nuclear Gauge and EDG vs. Standard Moisture Contents for Site 2 47

Figure 3.21. STDV Moisture Content(%) Errors for Site 2 48

Figure 4.1. Fifteen year break-even lifecycle cost comparison graph 55

vi
 List of Tables

Table 2.1. Average difference and T-test results between both gauges and core values 15

Table 2.2. Coefficient of Correlation and R-squared between both gauge densities vs. Core

density 17

Table 2.3. Correlation between Nuclear Gauge and Cores 19

Table 2.4. Range of data within ± 1 and ± 2 standard deviation 21

Table 2.5. Core Sample Density vs. MTD values 22

Table 2.6. PQI and Nuclear gauges Density vs. MTD values 23

Table 2.7. Level of confidence comparison in specific range 24

Table 3.1. Outliers removed for Density 38

Table 3.2. Outliers removed for Moisture Content 38

Table 3.3. Summary of R and R squared values 40

Table 3.4. Summary of STDV and average differences for gauges 44

Table 3.5. Test status Analysis of all gauges 49

Table 4.1. Costs associated with owning and operating a nuclear gauge 52

Table 4.2. Costs of owning and operating the non-nuclear gauges 52

Table 4.3. Cumulative Cost Combination of PQI and the soil gauge 54

vii
 List of Abbreviations

Electrical Density Gauge (EDG)

Dynamic Cone Penetrator (DCP)

Hot Mix Asphalt (HMA)

Light Weight Deflectometer (LWD)

Moisture Density Indicator (MDI)

Pavement Quality Indicator (PQI)

Quality Assurance (QA)

Quality Control (QC)

Soil Density Gauge (SDG)

Time Domain Reflectometry (TDR)

viii
 Acknowledgements

The authors thank the Nebraska Department of Roads (NDOR) for the financial support

needed for this study. The NDOR Technical Advisory Committee (TAC) comprised of Robert

Rea, Omar Qudus, Curt Mueting, Moe Jamshidi, Gary Thayer, Mick Syslo, Tim Krason, and

Brandon Varilek, was very helpful in providing technical support and guidance. Next thanks go

to Matt Beran, Mark Lindemann, Douglas Churchwell, Gary Mangen, Jeffrey Stolp, Joe Kuehn,

and Bob Rankin from the NDOR for their help throughout this project. The authors would also

like to thank Jeff Boettcher from Constructors who was very helpful in helping and allowing the

research team on jobsites. Everyone from Constructors on-site was very helpful and contributed

to this project. Thanks also go to Lindsay Fincher from Mid-America Transportation Center for

her help on this report.

ix
 Abstract

Quality control (QC) and quality assurance (QA) are necessary to ensure

fulfillment and compliance to specifications, guidelines, manuals, and programs which

outline methods and requirements during construction. Density, an important part of

quality control, can be used to evaluate the quality of Hot Mix Asphalt (HMA) and soil

compaction. This study investigated new technologies used for QC and QA by comparing

the Pavement Quality Indicator (PQI) model 301 with a nuclear gauge and core sample

measurements for HMA. For soil QC and QA, non-nuclear technologies—the Electrical

Density Gauge (EDG), the Moisture Density Indicator (MDI), and the Light Weight

Deflectometer (LWD)—were also investigated against a nuclear gauge and traditional

non-nuclear methods of measurement. Overall, the nuclear gauge shows higher accuracy

and higher correlation with cores than the non-nuclear gauges tested in this study. A

thorough investigation of calibration methods was also performed, both in the lab and on

the field, to improve the accuracy of the PQI‘s results. Data analyses showed that the

accuracies of the non-nuclear soil gauges are somewhat lower than that of the nuclear

gauge. With an improved methodology to create soil models for the EDG and

standardized ways to develop the LWD‘s target values, the EDG and LWD could have a

similar or better accuracy than the nuclear gauge. With the EDG and the Soil Density

Gauge (SDG), both recently ASTM approved, non-nuclear soil technology is the future.

Furthermore, the non-nuclear gauges could be a better alternative to a nuclear gauge

when the following benefits are considered: (1) economic savings; (2) faster data

measurement (PQI); (3) elimination of intense federal regulations and safety concerns; (4)

elimination of licensing and intense training.

x
 Chapter 1 Introduction

Quality Assurance (QA) of Hot Mix Asphalts (HMA) pavements was first introduced in

1986 (Andrewski 2003) to validate selected variables‘ accuracy and conformity to standards and

regulations. For HMA, density measurement is a general quality control (QC) and QA method

which uses either nuclear gauge readings or core density measurements. Density is measured as

part of the quality control process by paving contractors and for quality assurance by the

Nebraska Department of Roads (NDOR). Core density measurement is done in accordance with

the American Association of State Highway and Transportation Officials (AASHTO) procedure

AASHTO T 166. However, the destructive coring process creates holes in the new pavement,

though they are later patched. Nevertheless, this creates an imperfection in the pavements and

could cause long-term issues such as cracks and potholes. Furthermore, measuring core density

generally takes time. Core results are not typically available until the next day in order to allow

for corrections to the paving process and compaction to be completed. The required use of some

laboratory equipment adds an additional cost factor which must be considered. A minimum of

one full-time lab technician is usually required to run all the tests. Only a small number of cores

(usually less than ten) are used to gauge the values for several miles of pavement; therefore, the

coring process does not always provide solid results as some loose particles can be lost and affect

the density. This can lead to inefficient gathering of information which in turn can affect the

quality of HMA pavements.

Nuclear gauge technology offers a faster method of determining in-place HMA density,

and has been used successfully to replace and/or complement most coring in many states.

Depending on the specifications, Nebraska uses the coring method solely, or a combination of

the coring system with nuclear gauges. With nuclear gauges, come many advantages and

1
disadvantages. Nuclear gauges operate with the use of radioactive materials that may be

hazardous to the health and well-being of the operators. Therefore, proper precautions and care

need to be taken during operation. All users must have received radiation safety training and be

aware of the applicable safety procedures and regulations. The use of dosimeters or film badges

is also required for personal monitoring during use. Along with operation guidelines, routine

procedures such as source leak tests and annual calibration are recommended to properly

maintain the gauges. Strict licensing and re-licensing, record-keeping, and storage of the gauges

are all added to the complications of nuclear gauge technology. Finally, transporting radioactive

materials also requires rules and regulations. Consequently, there is a high demand for a device

that is accurate, easy to use, quick, non-destructive, and non-radioactive. The PQI seems one of

the gauges to overcome many or all of the problems posed by the core method and nuclear

gauges.

The quality of pavement foundation is affected by the properties of its sub-grade and

compaction conditions (Hancher et al. 2003). To ensure appropriate backfill, soil is compacted to

achieve its minimum physical properties. The foundation materials are therefore usually

compacted at different moisture conditions to identify moisture and density maximum values

(ASTM D 698/AASHTO T99) that will be used later for quality assurance. For those reasons,

density and moisture content are the common factors used to evaluate soil compaction. The

density in-place or in situ density is the general method used for QA. Like HMA, nuclear gauges

can measure in-place soil density and moisture content (ASTM D6938- 10, the Standard Test

Method for In-place Density and Water Content of Soil and Soil-aggregate by Nuclear Methods.)

which can be compared to the soil‘s maximum dry density and optimum moisture content for

quality control purposes. Other means of obtaining in situ density are the Standard Test Method

2
for Density and Unit Weight of Soil In-Place by the Sand-Cone Method (ASTM D1556-07), the

Standard Test Method for Density and Unit Weight of Soil In-Place by the Rubber Balloon

Method (ASTM D2167-08), or Standard Test Method for Density of Soil In-Place by the Drive-

Cylinder Method (ASTM D2937-10). When these lengthy and destructive traditional methods

are combined with the high costs, intense regulations, safety concerns (to just name few

problems with nuclear gauges), non-nuclear technology standardization for QA and QC seems

logical. To do so, the efficiency of these non-nuclear devices needs to be proven.

The main goal of this study is to assist the Nebraska Department of Roads (NDOR) with

supporting data in order to adopt non-nuclear gauges as a test modality to assess HMA pavement

and in-place soil. As a step towards this goal, the objectives of this research are:

(1) To assess the effects of a considerable number of factors potentially affecting the density and

moisture measurements generated by non-nuclear gauges on HMA compared to the nuclear

gauge and core samples through intensive field and lab tests; and

(2) To find the most effective method to assess soil compaction through field and lab tests; and

(3) To conduct economic analyses for the best alternative.

To be adopted for all QC and QA purposes, the performance of non-nuclear technology

must initially be proven adequate. Although a number of studies have showed non-nuclear

devices‘ capability, there has been disagreement in test results and recommendations for their

uses. In order to be accepted and adopted as standards, the accuracy and repeatability of non-

nuclear methods should be equivalent or better than the nuclear and other traditional methods. In

order to evaluate this claim, the PQI model 301 was tested against the nuclear gauge and core

samples for HMA QC. Similarly, the EDG, MDI, and LWD were all tested and compared

against the nuclear gauge as well as a selected traditional method for soil QC and QA.

3
 Chapter 2 Hot-Mix Asphalt

2.1 Methodology

The first objective of this research is to measure the effectiveness of the PQI model 301

which was compared to a nuclear gauge in terms of accuracy. The project examined the

determination of field density of HMA mixtures, and first examined the PQI as a possible new

way to gather real-time quality control data. After that part was established, a strategy for the

evaluation of the PQI was developed. The traditional core sampling method was selected as

standard, and both the nuclear gauge and PQI density measurements were compared against it.

The next step was then to find innovative ways to improve the data accuracy by coming up with

various calibration methods along with different techniques of measurement.

2.2 Literature Reviews

Different studies have been done to measure the effectiveness of nuclear and non-nuclear

gauges. In 1999, a Humboldt nuclear gauge was compared to the first model of the PQI for

variation in compaction and density variables (Rogge and Jackson 1999). Both gauges were

tested at forty-five different locations for six site visits. Both gauges were compared to cores that

were taken at each test area, and findings revealed that neither density values correlated well

with core densities (Rogge and Jackson 1999).

Sully-Miller Contracting Company also compared a nuclear gauge to the PQI in order to

study variance (Miller and Sully 2000). Standard deviations of the PQI were much lower and

different as compared to the nuclear gauge‘s standard deviations. The difference in surface

texture caused the nuclear gauge to show bigger variations, which appeared to have no impact on

the PQI. It was concluded that the PQI was accurate for HMA density measurements (Miller and

Sully 2000).

4
 Conversely, Henault evaluated the effectiveness of the PQI model 300 for quality

assurance testing in his study (Henault 2001). The calibration method of five core offset was

used on the ten different sites tested. The nuclear gauge results were much more correlated than

that of the PQI and, consequently, it was not recommended for quality assurance tests (Henault

2001).

Prowell and Dudley conducted a similar study in 2002 and reported that the nuclear

gauge showed better correlations with cores than the PQI (Prowell and Dudley 2002). Allen,

Schultz, and Willet also compared a nuclear gauge‘s density measurements to that of a non-

nuclear gauge. The five core average offset calibration method was used to improve the PQI‘s

density values. Findings validated the use of the PQI for quality control, but not quality

assurance (Allen and al. 2003). After improvements have been made to better non-nuclear

gauges, Hurley, Prowell and Cooley compared the newer PQI in 2004 to the nuclear gauge. A

total of twenty site visits were made and while the PQI had improved, it was still inferior to the

nuclear gauge for density measurements (Hurley et al. 2004). Schmitt, Rao, and Von Quitos did

a study in 2006 to compare the PQI model 300, model 301, and Pave Tracker 2701-B to the

nuclear gauge. To start, no calibration was made to the gauges to observe the results, and data

revealed that nuclear gauges‘ values were much greater than the non-nuclear gauges‘ values.

They also reported that the difference in nuclear and non-nuclear densities increased when the

pavement thickness increased. A mandatory calibration on each site test was then recommended

before measurements could be taken. A ten-core calibration was used and showed improvements

in the data. However, PQI‘s practicality was questioned (Schmitt et al. 2006).

In 2007, Kvasnak (et al. 2007) and a group of researchers) also compared the PQI and

Pave Tracker to the nuclear gauge to study factors that affect non-nuclear gauges. It was found

5
that roller pass, pavement moisture condition, and aggregate were among some of the factors that

affected density measurements. Another important finding was the need to study a test strip or

bed for calibration purposes (Kvasnak et al 2007).

2.3 Testing Methods

2.3.1 Core Method

Cores need to be extracted from the area where the nuclear gauge and PQI have been

used. Cores are taken soon after the pavement has been laid down and the roller passes. The

cores are usually very hot and therefore not easily drilled out. To facilitate the coring process, the

research team used dry ice (CO2) as a method to cool down the asphalt, as shown below in

Figure 2.1. Dry ice cools down the surface and leaves no trace of water, which helps with the

density measurements done on site for calibration purposes. Important care needs to be taken

when drilling to ensure underlying layers are not included in the sample. Drilling depth is usually

dictated by the bituminous layers. The results could be affected if the cores are tested with

excessive layers. After the cores have been drilled out, their bulk specific gravity measurements

are computed using the saturated surface dry method as specified in AASHTO 166 or by similar

guidelines. This measure of density has been adopted as the standard for such research. Nuclear

gauge density and PQI density are both compared to this density to measure accuracy. However,

biases occur in taking core density measurements because this method is not totally accurate and

can be offset by human errors, core debris left in holes, and many other factors including mix

types and hot weather temperatures.

6
 Figure 2.1. Dry ice cools the hot pavement

2.3.2 Nuclear Method

Nuclear gauges emit gamma rays from a radioactive source to measure density. The

emitted rays go through the compacted materials and use a count system that, combined with

other variables, are used to read the density. The research team performed nuclear readings on

HMA pavements using the American Society for Testing and Materials (ASTM) standard D

2950. The first five cores taken were used to calibrate both the nuclear and non-nuclear gauges.

Furthermore, the difference between the average of the first five nuclear gauge density

measurements and the average of the first five core measurements was used to offset the nuclear

gauge for the remaining measurements, as advised by Troxler 3440 operating manuals and

specifications. Figure 2.2 shows the Troxler 3440 nuclear gauge used for this study. The results

are then compared to the PQI‘s and documented for later analyses.

7
 Figure 2.2. Nuclear Gauge is shown measuring density

2.3.3 Non-Nuclear Method (PQI)

The PQI model 301, manufactured by Transtech Systems Inc., was used as a non-nuclear

alternative to measure density for the project. The PQI estimates density by measuring the

change in an electromagnetic field when a current is sent through the compacted material. A

dielectric constant proportional to the pavement‘s density is measured when the electrical current

is transmitted. The PQI model 301 is shown in Figure 2.3. The PQI is also calibrated and offset

using the average of the first 5 core density measurements, and by also following the manual and

operation specifications. Different measurement modes can also be used to improve the accuracy

of the results. The average mode, for example, automatically calculates an average of all the

densities at the measured spot, as long as they are within close proximity to each other (about 1

ft).

8
 Figure 2.3. PQI model 301shown taking measurements

2.3.4 Calibration

To improve the accuracy of the results, the gauges need to be properly calibrated. Density

measurements are relative measures of compaction, and are adjusted to be very close to the core

measurement. Several methods can be used for calibration. The AASHTO TP 68 standard

advises the users to record density measurements after each series of rollers pass through. Once

the density no longer increases, it is accepted and used to calibrate the devices. The AASHTO

TP 68 also recommends using the average of up to five core calibration densities to offset the

gauges. ASTM has also recommended similar methods of calibration. TransTech suggests a core

calibration using a minimum of five gauge readings at each location. ASTM has also published

numerous standards to recommend how electromagnetic devices should be calibrated (Kvasnak

et al. 2007). The research team started to calibrate the PQI by taking five single measurements at

a location, averaging the densities, and adjusting the results with the core measurements. To

9
improve the results, the readings are taken using an average mode of five to read a single

location.

The nuclear gauge reading is also done in both directions (parallel and perpendicular to

the pavement), and the average is computed for calibration. Dry ice, as introduced earlier, served

as a method to quickly cool down the pavement before coring. Dry ice also allowed the research

team to take cores without using water—allowing the cores to be measured right on site. All

cores are also measured later in the laboratory after a drying period of at least 24 hours. Both

measurements are compared, and adjustments were made to improve the results‘ accuracy.

Figures 2.4 and 2.5 show the cores while being measured both on-site and later in the lab. The

calibration method adopted by the research team conforms to the recommendations of both

manufacturers as well as those recommendations in the AASHTO TP 68 ―Standard Method of

Test for Density of In-Place Hot-Mix Asphalt (HMA) Pavement by Electronic Surface Contact

Devices.‖ Ideally, a calibration method will reconcile the differences between dissimilar

measures of the same property. However, in this case due to the unpredictability of the gauges

and other biases, perfect agreements are not always present, and regressions are used in analyses

to adjust one method to the others.

10
 Figure 2.4. On-site set up for core measurements

Figure 2.5. Lab set up for core measurements

2.4 Data Analysis

As noted earlier, this study has set out to compare measured differences obtained in the

field from both nuclear and non-nuclear density gauges. Both gauges were compared separately

11
against the study‘s control density measurement: laboratory tested core samples from the same

location. The underlying hypothesis of this study is that a proportional increase in measured core

density should linearly equate to a proportional increase in non-destructive density gauge

readings in the field. Unfortunately, due to external variables inherent to the paving and coring

process, data collected onsite does not follow an easily identifiable trend. Due to the external

variables, each data point was accepted or rejected based on a few key criteria.

2.4.1 Outliers

Generally, an outlier is identified as all values above the mean, plus or minus three

standard deviations (Los Alamos 2000). Initially, PQI density and core density correlation was

found to be extremely low at 4.21% for site number five (Figure 2.6). However, as Figure 2.7

illustrates, when outliers are excluded from the dataset, the correlation between readings from

the PQI and tested core samples increases dramatically to 56%. Outliners were taken out of the

data pool to improve the results for this study.

Figure 2.6. PQI‘s relationship with core samples before removing the outliers

12
 Figure 2.7. PQI‘s relationship with core samples after removing the outliers

2.4.2 Poor Core Samples

Extreme care should be taken to avoid altering and damaging cores during and after

coring. In this study, core samples that exhibited qualities of a poor specimen according to

AASHTO T166-05 were not included within the data pool for analysis. Figures 2.8 and 2.9

illustrate the kinds of cores that were accepted and rejected.

Figure 2.8. Example of rejected cores

13
 Figure 2.9. Example of an accepted core

2.4.3 Average Difference

After the appropriate filters were applied to the data pool, the average difference between

the core density and gauge density was found to be the most understandable method of

assessment to observe the differences among each gauge (Romero 2002). However, the average

difference or the t-test cannot assume that the gauge ‗trend‘ changes in the core density. To

highlight this point, Table 2.1 and Figure 2.10 describe data trends that were discovered through

an analysis of data collected onsite. When the difference is calculated, the PQI is 1.89 lb/ft3

lower than the cores, while the nuclear gauge‘s difference is 1.07 lb/ft3 higher than the cores.

However, Figure 2.10 shows that both gauges follow trends similar to that of the core sample

densities. If these gauges were evaluated based on the difference, the nuclear gauge would result

in closer values to core samples than the PQI‘s.

2.4.4 Student T-Test

To test for statistically significant differences between core samples and pavement

gauges, student T-tests are a sound analysis. In this analysis, the hypothesis is that the difference

between the core and gauge density readings is zero. In other words, if the t-test value is greater

14
than the t-value (95% confidence interval) using a probability t-value table, it can be concluded

that there is a statistical difference between gauge density and the core density (Romero 2002).

Table 2.1. Average difference and T-test results between both gauges and core values

Number Difference(lb/ft3) T-test

Site
of cores PQI Nuclear PQI Nuclear

1 9 3.972 0.6609 Reject Accept

2 10 3.2428 0.6428 Reject Accept

3 10 0.4195 1.4555 Accept Accept

4 9 1.074 1.331 Accept Reject

5 9 0.9281 0.7169 Reject Reject

6 9 2.098 0.1467 Reject Accept

7 9 1.608 2.058 Reject Accept

8 10 2.752 0.873 Reject Accept

9 9 1.3477 0.181 Reject Accept

10 15 1.784 0.45 Reject Accept

11 9 0.9613 2.3992 Accept Accept

12 20 2.3858 2.781 Reject Reject

13 10 2.0013 0.2137 Accept Accept

Average 1.89 1.07

For sites 3, 11, and 13, the statistical difference between each gauge and the cores is

greater than 95%. Both gauges therefore displayed density values that are very close to that of

15
the core. For the majority of the remaining sites, the nuclear gauge shows closer values to the

core‘s according to the student t-test analysis.

2.4.5 Coefficient of Correlation

The coefficient of correlation analysis is another method of evaluating the applicability of

a new gauge to measure density (Romero 2002). This analysis is used to decide if a statistically

significant linear relationship exists between the gauges when compared against core samples

(TransTech Systems 2004). The values of the coefficient of correlation range between +1 and -1.

If the value is close to +1, it indicates that there is significant correlation between gauge density

and core density.

Coefficients of correlations values for the nuclear gauges were higher than the PQI‘s for

most of the sites. This shows that the cores are better explained by the nuclear gauge, compared

to the PQI. It should be noted that there were few instances when the PQI‘s showed better

correlation (sites 2, 7, 9).

16
 Table 2.2. Coefficient of Correlation and R-squared between

both gauge densities vs. Core density

Coefficient of Coefficient of

Site Correlation( R ) Determination( R2 )

PQI Nuclear Gauge PQI Nuclear Gauge

1 0.198 0.6128 0.0392 0.3755

2 0.5046 0.064 0.2546 0.0041

3 0.2052 0.8211 0.0421 0.6742

4 0.7356 0.8901 0.5411 0.7922

5 0.7235 0.8295 0.5235 0.6881

6 0.746 0.9577 0.5565 0.9172

7 0.6476 0 0.4194 0.0025

8 0 0 0.2351 0.0082

9 0.7922 0.7185 0.6275 0.5163

10 0.138 0 0.019 0

11 0 0 0.1232 0.0006

12 0 0 0.0297 0.0006

13 0 0.5877 0.1681 0.3454

Average 0.252 0.407 0.275 0.333

2.4.6 Coefficient of Determination

Figure 2.10 indicates a weak correlation between both gauges individual densities as

compared to core density results—this is indicated by the low R2 values for both gauges.

17
Nonetheless, as shown in Table 2.2, four sites out of thirteen show a 50%+ relationship between

PQI density and core density. Additionally, five out of thirteen sites indicated that there is a

50%+ relationship between nuclear gauge densities and core densities.

Nuclear Gauge and PQI vs.

Core samples Density
PQI vs Core
148
y = 0.3253x + 93.95
PQI and Nuclear Gauge (pcf)

146
144 R² = 0.1857
142 PQI vs. Core
140 Nuclear vs. Core
138
136 Nuclear vs. Core
134 y = 0.6224x + 53.166
132
130
R² = 0.4252
130 135 140 145 150
Core Samples (pcf)

Figure 2.10. PQI & Nuclear Gauge relationship vs. Core Samples

2.4.7 Coefficient of Determination based on average density from each site

Table 2.3 shows a correlation between both gauges densities compared against individual

core locations and the site averages overall. Analysis results indicated that the measurements are

better explained by both gauges when considering site averages rather than individual locations.

18
 Table 2.3. Correlation between Nuclear Gauge and Cores

Individual Samples Site Averages

PQI Nuclear PQI Nuclear

R Squared Gauge Gauge

0.19 0.43 0.40 0.78

PQI and Nuclear Gauge vs. Cores

Density
148.00
146.00
Cores vs. Nuclear Gauge
PQI & Nuclear (in pcf)

144.00 y = 0.9908x + 1.7817

R² = 0.7817
142.00
140.00
PQI
138.00
Nuclear
136.00 Cores vs. PQI
y = 0.6494x + 48.787
134.00
R² = 0.3964
132.00
134.00 136.00 138.00 140.00 142.00 144.00 146.00
Core Samples (in pcf)

Figure 2.11. Relationship between each gauge and core samples

2.4.8 Error of the Standard deviation

Figure 2.12 illustrates the absolute density differential variation for both gauges. When

taken as a whole, the average difference between both gauges is very similar, varying by only

0.04 lb/ft3.

19
 Variation in Nuclear Gauge and PQI data

Nuclear SD: 2.00

abs pqi
error

PQI SD: 1.96

nuclear
error

0.00 2.00 4.00 6.00 8.00 10.00 12.00

Figure 2.12. Absolute Density Differential Variation for both gauges

2.4.9 Data Reliability

When comparing both gauges, it is important not to look only at how each gauge trends

as compared to the project‘s benchmark points (core samples), but to also look at an overall

tolerance. A clear grouping of PQI readings can be seen in Figure 2.13 where the nuclear gauge

data are spread more evenly throughout the plus or minus one and two standard deviation

boundaries. Table 2.4 demonstrates that 80% of the time the PQI data typically fall within 1

standard deviation of a core sample, as compared to 67% of the time for the nuclear gauge.

Results are even better, 99% of the time, for the PQI when using the 2 standard deviation range.

20
 PQI & Nuclear gauge
148
Mean ± 2 SD
146 of Core
144 Densities

142 Mean of core Sample

140 Density PQI

138 Nuclear
Mean ± 1 SD
136 of Core
Densities
134

132

130
PQI Nuclear

Figure 2.13. PQI & Nuclear Gauge Data reliability

Table 2.4. Range of data within ± 1 and ± 2 standard deviation

± 1 SD ± 2 SD

PQI data within

acceptable range (%) 79.86 99.28

Nuclear data within

acceptable range (%) 66.91 96.40

Describing individual gauge readings as compared to a mean of all collected benchmark

data is integral to creating and showcasing very simply how both gauges perform overall, but it

does not directly express to what extent each gauge reading can be trusted when compared to

their paired core samples.

21
 Table 2.5 shows the distribution of when exactly it is appropriate to reasonably accept

gauge readings. It was discovered that when core sample density results fall between 89% and 93%

of the maximum theoretical density (MTD) value of the mix design, both gauges can be assumed

to provide readings within the targeted 70% of a normally distributed bell curve. When applying

this finding to the PQI‘s previously collected readings, an average density difference of 0.59

lb/ft3 was found between the corresponding core samples and initial PQI readings. Thus, it is

recommended to select the core samples which range between 89% and 93% of MTD for

the PQI calibration at site.

Table 2.5. Core Sample Density vs. MTD values

Difference

Core sample density Number of % of the =|gauge-core|

compared to the MTD (%) Samples cores PQI Nuclear

86% 4 3% 5.79 5.01

87% 8 6% 4.68 2.96

88% 11 8% 3.48 3.33

89% 11 8% 1.96 2.49

90% 16 12% 0.71 0.77

91% 26 19% 0.78 0.96

92% 21 15% 0.70 0.36

93% 24 17% 1.14 0.63

94% 14 10% 2.02 0.20

95% 4 3% 4.89 0.14

22
 From Table 2.6 it can be concluded that when a core sample value falls between 90% and

94% of the maximum theoretical density, the PQI would give a very accurate comparison to the

traditional coring method. In this mentioned range, 72% of all collected data can be found and

considered accurate. Within the same accuracy range, the average difference between PQI

readings and corresponding core densities was 0.13 lb/ft3. The nuclear gauge, on the other hand,

does not provide a convenient range using the MTD values. When applying the same concept,

the nuclear gauge data would be deemed reliable when readings fall between 88% and 90%, and

between 93% and 94% of the MTD values.

Table 2.6. PQI and Nuclear gauges Density vs. MTD values

PQI and Nuclear Gauge

Number of Difference Number of Difference
densities compared to
Samples =|PQI-Core| samples =|Nuclear –Core|
the MTD value (%)

86~87% 1 6.32 3 2.56

87~88% 2 8.79 7 4.36

88%~89% 12 0.65 8 0.09

89% ~90% 15 1.52 13 0.19

90~91 45 0.41 27 1.13

91~92% 27 0.67 19 0.41

92~93% 18 0.58 27 1.7

93%~94% 10 0.68 17 0.75

94%~100 9 0.02 18 1.79

23
 Based on the results shown in Table 2.7, it is apparent that a tremendous improvement in

the level of confidence is achieved when operating both devices within a range of 89% to 93% of

the MTD value of the mix. Improvements are also significant when looking at the correlation

coefficient within the stated range. What this illustrates is that when ignoring collected data not

obtained within the recommended MTD range, the correlation improved by17%. This

demonstrates that if a core sample were to be taken at that location, the linear dependence

between what the PQI reads and what the core sample tests at can be trusted with 17% more

assurance.

Table 2.7. Level of confidence comparison in specific range

89% to 93% of Core samples of MTD Whole data

Core vs. PQI Core vs. Nuclear Core Vs Core vs

(98 data) (98 data) PQI Nuclear (139

(139 data) data)

Correlation 42% 56% 25% 41%

2.4.10 Number of Cores for PQI Calibration

This part of the project investigated a new method to determine the ideal number of cores

for the PQI calibration and improve the accuracy of PQI data. Traditionally, the offset is used to

decrease the difference between PQI data and core densities. In order to compare the differences

between the traditional and new method; three, five, eight and ten cores calibration were

investigated separately in this study.

24
 First, the traditional method was adopted to calibrate the PQI densities. Three (or five,

eight, or ten) cores are chosen randomly out of all data. The chosen set of cores is used to

calibrate the PQI. The difference between the calibrated PQI densities and core densities is

described as follows:

TD = ︱C1 – P2︱

Where,

C1-Core densities

P2-Calibrated PQI densities

Next, a linear regression was developed to obtain the difference. The calibrated PQI

densities (P2) are assumed to be an independent variable with TD as the dependent variable. A

linear regression equation Y= a*X + b was set up. While there are considerable combinations to

choose from, only the combination with the closest average R-square was adopted for further

calibration. Matlab™ was used to obtain the closest average R square value for this part of the

research. After substituting the calibrated PQI densities P2 for X, the adjusted difference Y2 was

acquired. Adjusted PQI value (AP) and linear regression difference (LD) were calculated as

follows:

AP = Y2 + P2

LD = ︱AP – C1︱

Where,

AP–Adjusted PQI value

LD–linear regression calibration difference

The results are attached in the appendix, and Figure 2.14 shows the TD and LD value.

25
 Comparison between TD and LD
3.00000

2.50000

2.00000

1.50000
TD
1.00000 LD
0.50000

0.00000
3 5 8 10
Cores

Figure 2.14. Comparison between traditional difference and calibration difference

As can be seen in Figure 2.14, both differences tend to trend lower with the increase of

the number of cores. When 8 or 10 cores were chosen, the linear regression differences were less

than the traditional differences. The linear regression difference is lowest when choosing 8 cores.

2.5 Conclusions

In conclusion, it was determined that the nuclear gauge has a slightly higher correlation

value than the PQI when compared to the core samples. Its average difference between the

nuclear gauge and the PQI was not significant (only a 0.82 lb/ft3 difference) when the research

team considered all the data pool. The PQI, however, shows more consistent results than the

nuclear gauge, in consequence of a smaller standard deviation. Furthermore, when cores and PQI

have higher densities than 90% of MTD, the PQI is statistically more accurate and has a much

better coefficient of correlation than the nuclear gauge. To determine better ways to calibrate the

PQI and eventually improve its accuracy, a trial model calibration was run.

26
 Different combinations were tried to determine the best statistical way to improve the

PQI‘s accuracy. It was concluded that the PQI performs at its best when 8 cores are used for

calibration. Other analyses revealed that the cores with a density that falls within 89% and 93%

of the MTD value should be used for calibration. Moreover, any PQI density that was greater

than 90% of the MTD value was proven as reliable information. In the event that the density

measured by the PQI is less than 90%, a core must be taken for density measurement.

27
 Chapter 3 Soil

3.1 Methodology

To accomplish the objective of assessing soil compaction, a comparison study of

usability and performance was conducted between a nuclear gauge (Troxler 3440) and three non-

nuclear gauge alternatives, including the Electrical Density Gauge (EDG) and the Moisture

Density (MDI) and the Light Weight Deflectometer (LWD). The EDG and the MDI were tested

for in-place moisture and density. The LWD, a stiffness-strength based criterion for evaluating

the QA/QC of a material, was also tested. The nuclear gauge was utilized to measure the in-situ

dry density and moisture content. Finally, the previously mentioned measurements were all

compared to a standard, the field dry unit weight measurement, which was determined by taking

a sample representative of each measurement area either with a Shelby tube or other method for

lab testing.

3.2 Literature Review

In 2000, McCook and Shanklin (2000) compared the accuracy of the nuclear gauge with

various traditional methods including sand cone, density-drive cylinder and rubber balloon.

Density test results from the nuclear gauge and sand cone were very similar. The few problems

that were identified in this study included the following issue: that some errors could be observed

with the sand cone due to the change in operating personnel and variation of density

measurement readings. The drive-cylinder method was the most consistent of the traditional

methods to measure density.

A similar study was done by Norrrany et al. (2000) to compare the sand cone method, the

drive-cylinder method and a nuclear gauge on various compacted soil types. Both the sand cone

28
and drive-cylinder methods resulted in a wider range of variability with the nuclear gauge, but

with the sand cone being the least variable.

Studies have been done to introduce stiffness and modulus methods as replacement

quality control methods of soil compaction. Livneh and Goldberg determined in their work in

2001 that the current unit weight quality control at the time was slow, hazardous, labor intensive,

and of an uncertain accuracy.

In 2009, a study done in Thailand compared the sand cone with a nuclear gauge, the

Dynamic Cone Penetrator (DCP), and the Soil Density gauge (SDG). It was concluded that non-

nuclear technology had good potential with further development and research to be implemented

for construction quality control of earthwork (Wacharanon et al. 2009).

In 2007, another comparison study was done among the nuclear gauge, the MDI and the

EDG. Results showed good correlations of densities between both non-nuclear gauges and the

nuclear gauge. Moisture content showed a big variability between gauges. It was also suggested

that other in-place measurements should be done because the nuclear gauge data could not be

entirely trusted and used as standards (Brown 2007).

The Minnesota Department of Transportation (Mn/DOT) did various studies where they

compared the DCP, the LWD, a Percometer and Trident moisture content. It was concluded in

2006 that the LWD displayed a good level of accuracy close to the DCP and they also offered

suggestions to improve the LWD (Davich et al. 2006). Since then, different techniques and

methods were developed to estimate the LWD deflection target values for soil to assess the

compaction state of a soil (Siekmeier 2011).

In 2007, the New Jersey Department of Transportation (NJDOT) compared the MDI with

the nuclear gauge and concluded that very similar moisture contents measurements could be

29
observed between both gauges (Jackson 2007). High differences existed, nevertheless, with dry

density measurements. Improved calibration constants and methodology were recommended.

3.3 Testing Methods

3.3.1 Density-Drive Cylinder

The Standard Test Method for Density of Soil in Place by the Drive-Cylinder Method

(ASTM D2937-10) involves obtaining a relatively undisturbed soil sample by driving a cylinder

open at both ends in the ground (Figure 3.1). Once flush, the material around the cylinder is then

excavated. With the empty volume of the cylinder already known, the unit weight of the soil in

the cylinder can then be calculated in the lab. While in the lab, a sample of the soil can be dried

to provide a dry density of the material. This method was preferred over the sand cone test

(ASTM D1556) which consists of determining the in-place density and unit weight of soils using

a sand cone apparatus because of the inconsistency of density results. Similar to the sand cone

method, the rubber balloon method (ASTM D2177) consists of excavating a sample of soil and

measuring the volume of the hole dug out with a rubber-balloon apparatus. This method also

provides variable results depending on the users and pressure applied to the apparatuses while

filling the holes. A higher force applied on apparatuses will show a greater displacement. All

these inconsistencies in readings and data have led the research team to adopt the density-drive

cylinder as standard to measure in-place density.

30
 Figure 3.1. Shelby Tube driven in the ground

3.3.2 Water Content of Soil by Oven Dry method

While previously mentioned methods only determine the in-place density, soil bulk

density is determined by the weight of the soil per unit volume that is found by using an oven

maintained at a temperature between 105C and 115C. This method (ASTM D2216) consists of

drying a wet sample of soil in the oven for about 24 hours, and determining the weight of

moisture. This method was used as the standard and baseline of comparison for moisture content

measurement.

3.3.3 Nuclear Method

As with HMA, nuclear gauges emit gamma rays to measure density and moisture content.

Measurements were done according to ASTM D6938-10, Standard Test Method for In-Place

Density and Water Content of Soil and Soil Aggregate in Place by Nuclear Methods. Unlike the

HMA measurement, the gauge probe was driven into the ground to take measurements at 4, 6,

31
and 8 inches. Figure 3.2 shows a nuclear gauge taking soil density and moisture content

measurements.

Figure 3.2. Nuclear gauge taking soil measurements

3.3.4 The Electrical Density Gauge (EDG)

The EDG measures the electrical dielectric properties, along with moisture levels of the

material‘s compacted soil to determine its density and moisture content. The EDG does so by

measuring the radio-frequency current between four darts driven in the ground, as shown in

Figure 3.3. In order to measure the in-place physical properties of the soil, a soil model or

calibration process needs to have taken place in the lab. A sample representative of the soil to be

tested needs to be excavated and tested in the lab with the EDG at different moisture and

compaction levels. ASTM D7698-11 (Standard Test Method for In-Place Estimation of Density

32
and Water Content of Soil and Aggregate by Correlation with Complex Impedance Method) was

applied for EDG testing. A minimum of three lab tests are recommended by EDG‘s

manufacturer to have a good soil fit. The research team conducted nine lab tests to develop soil

models.

Figure 3.3. Darts driven into ground for EDG test

3.3.5 The Moisture Density Indicator (MDI)

The MDI uses the Time Domain Reflectometry (TDR) methodology, which measures the

travel time of an electromagnetic step pulse produced by the TDR pulse generator through spikes

driven in the ground (Brown, 2007). A personal digital assistant (PDA) or a laptop is then used

with the manufacturer-provided software to analyze the signal sent by the spikes (Figure 3.4).

The apparent dielectric constant and electrical conductivity of the soil are derived from the MDI

to estimate the soil‘s density and moisture content. Just like the EDG, soil models were also

33
required to determine the dry density of the soil. The MDI required a typical moisture densities

curve using the Standard Test Methods for Laboratory Compaction Characteristics of Soil Using

the Standard Effort (ASTM D698-07) or Standard Test Methods for Laboratory Compaction

Characteristics of Soil Using the Modified Effort (ASTM D1557-09). Once more than 4 points

have developed, a soil compaction curve will then indicate the maximum density and moisture

content for that material.

Figure 3.4. MDI connected to a laptop for density measurements

The MDI was acquired by the research team, but unfortunately failed to work during

testing. Multiple extensive efforts were made by the MDI‘s manufacturer to assist the research to

make the device function. During testing, systematic errors were consistent and did not allow the

34
team to take sufficient measurements. For such reasons, the analysis report does not include the

MDI data.

3.3.6 The Light Weight Deflectometer (LWD)

The LWD (Figure 3.5) consists of measuring a surface deflection as a result of applying

an impulse load to it by using ASTM E2583-07, the Standard Test Method for Measuring

Deflections with a Light Weight Deflectometer (LWD). The LWD consist of a light mass, an

accelerometer and a data collection unit (Siekmeier et al. 2009). Because the LWD measures the

deflection and modulus of elasticity of the soil, there was no direct relationship or method to

compare its measurements with the other gauges being tested in this study. The research team

therefore used a quality assurance procedure developed by the Mn/DOT (Siekmeier et al. 2009)

along with their specifications for excavation and embankment (Minnesota 2010) to determine

whether a soil area has been properly compacted. Based on a pass/fail criterion, comparisons can

then be made with other gauges. More details about this methodology are discussed in the

analysis section.

35
 Figure 3.5. LWD is measuring stiffness of the soil

3.4 Test Procedures and Methodology

Two sites composed of brown dirt and peorian loess soils were tested for this research.

The team first collected representative samples from each site to develop soil curves by the EDG,

MDI and the Standard Proctor Method. The results were then used to calibrate the nuclear gauge,

and determine in-place measurements for the EDG. Figure 3.6 and 3.7 are sites from Highway

370 near Gretna, NE and the Platteview Intersection site near Plattsmouth, NE, respectively.

36
 Figure 3.6. Highway 370 site

Figure 3.7. Platteview Intersection Site

37
 Once a spot was selected, all gauges were operated and their variables recorded at said

location. The different densities and moisture contents of the EDG and nuclear gauge were then

compared against the standard baselines of measurement methods mentioned above. Next, a

pass/fail analysis of all the methods was developed according to the Standard Specifications for

Construction in NDOR (Nebraska 2007). This analysis would give a better idea of what method

correlates most closely with the LWD. For better accuracy, other important analyses were also

run to compare the gauges.

3.5 Data Analysis

3.5.1 Outlier

Similarly to HMA data, outliers were also removed from the pool of data in order to

better analyze soil measurements. Outliers were removed when the difference between the

standard density and moisture was considerably greater or lower than the gauges‘ data—that is, a

standard deviation plus or minus 3. Tables 3.1 and 3.2 respectively show the set of outliners

removed from the density and moisture content measurements.

Table 3.1. Outliers removed for Density

Standard Nuclear Gauge EDG

110.4 pcf 75.63448 pcf 103.88 pcf

Table 3.2. Outliers removed for Moisture Content

Nuclear Gauge Standard EDG

74.30% 20.44% 18.30%

20.43% 29.87% 20.40%

38
3.5.2 Coefficients of Correlation (R) and Determination (R2) analyses

To observe a linear relationship between the gauges and the standard measure, the whole

pool of data were analyzed after removing the outliers (Figure 3.8 for density and Figure 3.9 for

moisture). Table 3.3 summarizes the coefficients that were observed.

Standard vs. Nuclear Gauge and EDG Density

115.00
y = 0.8693x + 11.614
110.00
Nuke, EDG Density(in pcf)

R² = 0.4836
105.00
100.00
95.00
90.00
Nuclear Gauge vs. Standard
85.00
EDG vs. Standard
80.00
y = 0.9112x - 0.8105
75.00 R² = 0.2413
70.00
85.00 90.00 95.00 100.00 105.00 110.00 115.00
Standard Density

Figure 3.8. Standard vs. Nuclear Gauge and EDG Density

39
 Standard vs. Nuclear Gauge and EDG Moisture
32
y = 0.9443x + 1.3275
27 R² = 0.7687
Nuclear Gauge, EDG (in %)

22

Nuclear Gauge vs. Standard

17
EDG vs. Standard

12
y = 1.0959x - 3.6821
R² = 0.4848
7
8 13 18 23 28
Standard (in %)

Figure 3.9. Standard vs. Nuclear Gauge and EDG Moistures

Table 3.3. Summary of R and R squared values

Density Moisture

Nuclear vs. EDG vs. Nuclear vs. EDG vs.

Standard Standard Standard Standard

Coefficient of Correlation (R) 0.695 0.491 0.876 0.695

Coefficient of Determination (R2) 0.483 0.241 0.768 0.484

The nuclear gauge has a higher R2 than the EDG‘s, and also correlates better with the

standard measurement. This could be due to the fact that the nuclear gauge data have been

corrected using the density and moisture corrections factors, as required by the Nebraska

Department of Roads new Standard Test Method for Nuclear Density Testing for Soils (NDOR,

2011). When using the initial data, that is before adjusting the nuclear gauge density values, the

40
coefficient of determination of the nuclear gauge and the standard is only 0.21 (Figure 3.10), a

little lower than that of the EDG. There are no current recommended methods that allow for

ways to improve and correct the EDG‘s data. The EDG and nuclear gauge have very similar

results when unmodified and direct data are considered, but the nuclear gauge performs better

when correction factors are applied.

Standard vs. Nuclear Gauge (before) and EDG

Density
115
y = 0.3602x + 65.924
Nuclear Gauge, EDG Density (in pcf)

110 R² = 0.2134
105
100
95
90 Nuclear Gauge vs. Standard
85 EDG vs. Standard
80
75 y = 0.9314x - 2.8281
70 R² = 0.2568
85 90 95 100 105 110 115
Standard Density (in pcf)

Figure 3.10. Standard vs. Nuclear Gauge (before correction factors) and EDG Density

For moisture content, both gauges have a high coefficient of correlations, but the nuclear

gauge has more data close to the standard.

3.5.3 Average Difference and Error of Standard Deviation

To determine how both gauges vary within the lab data standard deviations (STTV), error

of standard deviation and average difference analyses were conducted (Figures 3.11 , 3.12, and

3.13).

41
 Average difference Density compared to
standard
12
9.86 pcf
10

6 Nuclear gauge

4 EDG

1.71 pcf
2

0
Nuclear gauge EDG

Figure 3.11. Average Density Difference of gauges compared to standard

Variation of Nuclear Gauge and EDG Density

Error of
EDG
(STDV:
6.37 pcf)

Error of
Nuclear
Gauge
(STDV:
2.49 pcf)
0.00 5.00 10.00 15.00 20.00 25.00 30.00

Figure 3.12. Variation of Nuclear Gauge and EDG Density

42
 Variation of Nuclear Gauge and EDG Moisture
Error of
EDG
(STDV:
2.90)
Error of
Nuclear
Gauge
(STDV:
1.08)
0 2 4 6 8 10 12 14 16
Error of Nuclear gauge moisture Error of EDG moisture

Figure 3.13. Variation of Nuclear gauge and EDG Moisture Content (%)

Table 3.4 summarizes the STDV and the average differences of both gauges. Average

differences of 1.71 pcf for the nuke density data and 0.22% for moisture content compared to

9.86 pcf and 1.66%, respectively, for the EDG density and moisture content not only support the

coefficient of determination analyses, but also show a high variation among the EDG data. This

could be due to the fact that the soil model range used for the EDG might be too wide, which

could in turn widen the range of the EDG‘s measured data. Site by site analyses can also reveal

some other information about both the sites and soil tested.

43
 Table 3.4. Summary of STDV and average differences for gauges

Density Moisture

Nuclear Nuclear

Gauge EDG Gauge EDG

Average
1.71 9.86 0.22 1.66
Difference

STDV 2.49 6.37 1.08 2.90

3.5.4 Site by Site Analysis

Coefficient of correlations, determinations, average difference and standard deviation

analyses were all run with information derived from each site. Figures 3.14 through 3.21

illustrate this data through graphing.

Site 1: Nuclear Gauge and EDG vs. Standard

Dry Density
115.00
Nuclear gauge, EDG Density

Nuclear gauge vs Lab

105.00 EDG vs Lab
y = 1.1283x - 15.734
y = 1.3459x - 44.919
95.00 R² = 0.7634
R² = 0.3956
85.00 Nuclear gauge Density
75.00 EDG Density
65.00
85.00 90.00 95.00 100.00 105.00 110.00 115.00
Standard density

Figure 3.14. Nuclear Gauge and EDG vs. Standard Density for Site 1

44
 Site 1: STVD Density Error

0.00 5.00 10.00 15.00 20.00 25.00 30.00

Number of data

Figure 3.15. STVD Density Errors for Site 1

Site 1: Nuclear Gauge and EDG vs. Standard

Moisture
30
28 y = 0.7789x + 4.908
26 R² = 0.4054
Nuke, EDG(in %)

24
22
20
18
16 Nuke vs Standard
14 y = -0.0209x + 20.918
12 EDG vs Standard
R² = 0.0002
10
15.00 17.00 19.00 21.00 23.00 25.00
Standard Moisture

Figure 3.16. Nuclear Gauge and EDG vs. Standard Moisture Contents for Site 1

45
 Site 1: STDV moisture content error

Error for EDG

(STDV: 2.12)

Error for Nuclear

gauge
(STDV : 0.84)

0.00 2.00 4.00 6.00 8.00 10.00

Figure 3.17. STDV Moisture Content (%) Errors for Site 1

Site 2: Nuclear Gauge and EDG Density vs.

Standard Density
115
Nuclear gauge vs Standard
Nuclear gauge, EDG Density

110 y = 0.2806x + 73.455

R² = 0.085
105
Nuclear gauge vs Standard
100

95 EDG vs. Standard

90 EDG vs Standard
y = 0.0701x + 85.181
85 R² = 0.0038
80
90 95 100 105 110 115
Standard Density

Figure 3.18. Nuclear Gauge and EDG vs. Standard Density for Site 2

46
 Site 2: STDV Density Error

Error of EDG
(STDV : 5.41)

Error of Error of Nuclear gauge

Nuclear Gauge
Error of EDG
(STDV : 2.50)

0.00 10.00 20.00 30.00

Num of data

Figure 3.19. STVD Density Errors for Site 2

Site 2: Nuclear Gauge and EDG vs. Standard

Moisture Content (M)
21 Nuclear vs Standard
Nuclear gauge, EDG Moisture

19 y = 0.6122x + 6.3269
17 R² = 0.4619
15
13 Nuclear gauge vs
11 EDG vs Standard Standard Moisture
9 y = 0.4948x + 4.8889
EDG vs Standard
7 R² = 0.2411
Moisture
5
10 15 20 25
Standard Moisture (in %)

Figure 3.20. Nuclear Gauge and EDG vs. Standard Moisture Contents for Site 2

47
 Site 2: STDV Moisture Error

EDG
(STDV: 1.72)

Nuclear gauge
(STDV : 0.80)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Figure 3.21. STDV Moisture Content(%) Errors for Site 2

Figures 3.14 through 3.21 show consistent observations regardless of the site:

- The coefficient of determination of the nuclear gauge density is always higher than the

EDG;

- Site 2 had a very low correlation for both the nuclear gauge and the EDG in density

measurement.

In summary, site 1 showed better results than site 2. This may be due to the fact that more

data measurements were taken on site 1 (63 vs. 40 for site 2). Site 2 testing area may have also

had different soil types, which could have altered the results. In order to utilize the data gathered

by the LWD, a test status using NDOR‘s current quality assurance was used.

3.5.5 Test Status Analysis

To meet the compaction requirements, a test is deemed passed or failed when the

measured density is within 95% of the maximum density determined by the soil curve, and also

within the moisture content requirements (NDOR 2007). The research team only took random
48
measurements at various spots on both sites to compare all gauges. Some measurements were

taken at areas that were not previously compacted; therefore, some measurement spots would fail

the quality assurance test. As a consequence, the LWD, which measures soil deflection and

elastic modulus, could not be then directly compared with the nuclear gauge and the EDG. The

Mn/DOT has developed an excavation and embankment specification (Minnesota 2006) that

allowed the research team to know when the LWD passed or failed the testing. A pass or fail test

status comparison was made to view the relationship of each gauge with the standard. A

successful relationship would be one in which a gauge would pass when the standard passes, and

would fail when the standard method fails. The whole data used to develop this analysis is

included in the appendix section. Table 3.5 below summarizes the test status comparison.

Table 3.5. Test status Analysis of all gauges

Test Status Relationship with Standard Method Site 1 Site 2 Average

Nuclear Gauge 80.62% 65% 72.81%

LWD 41.24% 67.50% 54.37%

EDG 41% 37.50% 39.80%

The nuclear gauge and LWD were the most correlated with the standard method when

using a pass or fail quality assurance method of analysis. The quality assurance method used for

the LWD is not yet a standard, and is currently being improved by the Mn/DOT. A better way to

estimate the target value of the LWD might improve its correlation with the standard method.

Furthermore, when the nuclear gauge data did not apply the correction factor, its test status

49
relationship was only 63%, which is closer to the LWD. It means that the nuclear gauge raw data

had similar results with the LWD.

3.6 Conclusions

A direct density comparison between the nuclear gauge and the EDG revealed that the

nuclear gauge had a better correlation to the standard method. The EDG had similar results with

the nuclear gauge before the nuclear gauge correlation factors were used to improve the gauge.

Many researches are ongoing to find a way to improve the measured EDG‘s data, which could

then perform similarly or better than the nuclear gauge. Different mold shapes and compaction

methods are currently being tested by the manufacturer to improve the EDG‘s soil model. The

nuclear gauge has been used much longer, so methods of improvement have been developed for

better results.

The LWD, which measures the soil deflection, also displayed similar results with the

nuclear gauge when using raw data. However, the nuclear gauge had better correlation with the

standard method when the data was corrected. Different methods to estimate the LWD‘s target

values can also researched and tested.

50
 Chapter 4: Economic Analysis

4.1 Life Cycle Costs

Various techniques can be used to predict and analyze how much equipments would cost

over time. A lifecycle cost analysis considers all the costs associated with owning, operating, and

maintaining equipment for the duration of their useful life. For the lifecycle analysis done in this

case, costs such as maintenance and any other non-directly measurable costs were estimated

using previous data, quotes, and manufacturers‘ recommendations. Initial costs were those

received from retailers when acquiring the gauges.

Tables 4.1 and 4.2 summarize the costs associated with possessing the nuclear and non-

nuclear gauges.

51
Table 4.1. Costs associated with owning and operating a nuclear gauge

Cost of nuclear gauge $6,950

Radiation safety & Certification

$750
Class

Safety training $179

HAZMAT certification $99

RSO training $395

TLD Badge monitoring $140/yr

Life of source capsule integrity 15 yr

Maintenance & Recalibration $500/year

Leak test $15

Shipping $120

Radioactive Materials License $1,600

License Renewal $1500/ year

Reciprocity $750

Table 4.2. Costs of owning and operating the non-nuclear gauges

Annual

Initial Costs Maintenance

EDG $9,000 $0

LWD $8,675 $0

PQI $8,200 $500

52
4.2. Analysis

A basic analysis done by adding costs incurred over the gauges‘ life expectancies show

that a nuclear gauge always costs more than any combination of the PQI and non-nuclear soil

gauge as shown in Table 4.3 and Figure 4.1. The analysis is done using the lesser of the gauges‘

life expectancies, which is equivalent to 15 years.

In order to view the current benefit of using non-nuclear gauges, a net present worth cost

of all gauges can be computed as explained below:

• Net Present Worth of Costs (NPW) = Initial Costs + Yearly Costs (P/A, 15 yrs, 10%)

• NPW of Nuclear Gauge = $10,873 + $2,155 (P/A, 15yrs, 10%)

• NPW of PQI + Average Soil Gauge = $17,038 + $500 (P/A, 15yrs, 10%)

• NPW of Nuclear Gauge = $27,264

• NPW of PQI + Average Soil Gauge = $20,840

Despite the high initial cost of the non-nuclear gauges, they still hold an economic advantage

over the nuclear gauge when maintenance and operating costs are included. Figure 4.1 can

help to calculate the break-even point for both investments.

53
Table 4.3. Cumulative Cost Combination of PQI and a soil gauge

Nuclear

Year PQI + EDG PQI + LWD Gauge

0 $17,200 $16,875 $10,873

1 $17,700 $17,375 $13,028

2 $18,200 $17,875 $15,183

3 $18,700 $18,375 $17,338

4 $19,200 $18,875 $19,493

5 $19,700 $19,375 $21,648

6 $20,200 $19,875 $23,803

7 $20,700 $20,375 $25,958

8 $21,200 $20,875 $28,113

9 $21,700 $21,375 $30,268

10 $22,200 $21,875 $32,423

11 $22,700 $22,375 $34,578

12 $23,200 $22,875 $36,733

13 $23,700 $23,375 $38,888

14 $24,200 $23,875 $41,043

15 $24,700 $24,375 $43,198

54
 Nuclear Gauge Costs V. PQI+ Soil Non-nuclear
Gauge
$45,000
y = 2155x + 10873
$40,000

$35,000

$30,000
Costs

y = 500x + 17038 Nuclear Gauge

$25,000
PQI + Average Soil Gauge
$20,000

$15,000

$10,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Year
Break-even point=3.73 years

Figure 4.1. Fifteen year break-even lifecycle cost comparison graph

Although the initial costs of investing in the non-nuclear technology are higher than the

initial cost of the nuclear gauge, the yearly maintenance, along with savings from maintaining

and operating radioactive equipment, make the investment for non-nuclear gauges very

profitable in the long term. With a break-even point of 3.73 years, the investment in non-nuclear

technology makes sense. The EDG and the LWD cost about the same over their lifecycles. Costs

can therefore be a negligible factor in selecting one of the two gauges for soil QC and QA.

55
 Chapter 5 Conclusions and Recommendations

Intense regulation and destruction of materials all call for a new method for HMA and

soil QA and QC. For HMA, the PQI offers a rapid measure of measurement, and is much more

economical than the nuclear gauge. Test results showed that the PQI can have similar and better

results than other alternatives. When the PQI reading is over 90% of the MTD value, the density

measured is completely reliable and better than the nuclear gauge. Coring of the pavement

should be done when the PQI gives values less than 90% of the MTD value. In that instance, the

research team recommends using cores that have measured densities within 89 to 93% of the

MTD value to calibrate the PQI.

Density and stiffness were introduced as ways to evaluate the compaction of soil. Testing

was done with non-nuclear technologies that were not as accurate as the nuclear gauge. Nuclear

technology has been around for so long, and various researches have been done to improve initial

means of measurement. These improvements have resulted in proven ways to improve the

nuclear gauge‘s accuracy, which has been adopted as a standard by some states and agencies.

The EDG data was very comparable to the nuclear gauge before correction factors were applied.

Numerous researches are ongoing to improve ways to develop soil models, which will in turn

improve the EDG‘s correlation with the standard. The LWD also showed better correlation with

the nuclear gauge when the initial data was used. Other methodologies to find target values and

properly assess soil compaction will lead to better results.

Generally, the tested non-nuclear gauges (PQI, EDG, and LWD) take much less time to

record measurements. Their initial costs are higher than the nuclear gauge, but have a greater

return on investments; namely, some manufacturers, like Transtech, offer a trade-in credit for the

PQI, for example.

56
 Chapter 6 NDOR Implementation Plan

From the findings of this research, the NDOR will consider future implementation of the

non-nuclear device for HMA acceptance for in-place density. The specification will utilize

similar parameters found in the research testing, such as the final acceptance standard, being the

actual roadway core for any tests under the 90% of Gmm. This is commonly the current practice

adopted for low density results when using the current acceptance devices, i.e. nuclear gauges or

cores. This will allow the contracting industry to utilize non-nuclear equipment with no

permitting, fewer regulatory standards and less equipment safety training and/or documentation

requirements, at an overall lower total ownership and operating cost.

The NDOR is going to expand on the initial findings of the non-nuclear soil density

testing through 'in-house' research with the Soils and Geotechnical Sections. This expanded

testing is an effort to test, quantify and accept soil compaction based on soil stiffness and

modulus values in conjunction with the Nebraska Soils Index system by utilizing the LWD

equipment. This will be used along with field moisture tests to create a new acceptance system

for in-place engineered grading and fill. The goal of this research is to identify an improved

testing and acceptance system utilizing non-nuclear devices and have much lower costs by less

permitting and regulatory requirements, less accounting and documentation systems, as well as

lower storage and maintenance costs on the equipment and personnel.

Robert C. Rea, P.E.

Assistant Materials and Research Engineer

Nebraska Department of Roads

57
 References

Allen, D. L., D. B. Schultz, Jr., and D. A. Willett. 2003. ―Evaluation of Non-Nuclear Density

Gauges.‖ Research Report KTC-03-24/FR115-01-1F. Kentucky Transportation Center,

University of Kentucky.

American Association of State Highway and Transportation Officials. 2010. Standard

Specifications for Transportation Materials and Methods of Sampling and Testing.

Washington D. C.

American Society for Testing Materials. 2009. Annual Book of ASTM Standards. West

Conshohocken, PA: ASTM International.

Andrewski, Andrew. 2003. ―Density Measurements of Hot Mix Asphalt (HMA) Pavements.‖

Final Report. Indiana Department of Transportation and the Federal Highway

Administration.

Bollen, K. A., and R. W. Jackman. 1985. ―Regression Diagnostics: An Expository Treatment of

Outliers and Influential Cases.‖ Sociological Methods & Research 13.4: 510-542.

Brown, Jeff. 2007. ―Non-Nuclear Compaction Gauge Comparison Study.‖ Final Report.

Vermont Agency of Transportation Materials and Research Section.

Davich, Peter, Felipe Camargo, Brett Larsen, Ruth Roberson, and John Siekmeier. 2006.

―Validation of DCP and LWD Moisture Specifications for Granular Materials.‖ Final

Report. Minnesota Department of Transportation.

Henault, J. W. 2001. ―Field Evaluation of a Non-Nuclear Density Pavement Quality Indicator.‖

Report No. 2227-F-01-3. Connecticut Department of Transportation, Bureau of

Engineering and Highway Operations, Division of Research.

Hurley, G. C., B. D. Prowell, and L. A. Cooley, Jr. 2004. ―Evaluation of Non-Nuclear

58
 Measurement Devices for Determination of In-Place Pavement Density.‖ Transportation

Research Record #1900, Transportation Research Board, Washington, D.C.

Jackson, Hudson. 2007. ―Assessment of the moisture density indicator for the construction

quality control of compacted dense graded aggregate base layers.‖

http://www.humboldtmfg.com/pdf/edg/RutgersU-report-

MDI_Nuclear_Gauge%20Comparison.pdf

Kvasnak, A. N., R. C. Williams, H. Ceylan, and K. Gopalakrishnan. 2007. ―Investigation of

Electromagnetic Gauges for Determining In-Place HMA Density.‖ Final Report, CTRE

Project 05-233. Center for Transportation Research and Education, Iowa State

University.

Larsen, D. A., and J. W. Henault. 2006. ―Quantifying Segregation in HMA Pavements Using

Non-nuclear Density Devices.‖ Data Collection Final Report for Connecticut, SPR-2238.

Connecticut Department of Transportation.

Los Alamos National Laboratory. 2000. Memorandum. Environment, Safety, and Health

Division.

McCook, D. K., and D. W. Shanklin. 2000. ―NRCS experience with field density test methods

including the sand-cone, nuclear gage, rubber balloon, drive-cylinder, and clod test.‖

Constructing and controlling compaction of earth fills STP 1384: 72-92.

Minnesota Department of Transportation. 2010. Draft Specifications for Excavation and

Embankment. Minnesota Department of Roads, St Paul, MN.

Minnesota Department of Transportation. 2011. ―Pilot Light Weight Deflectometer (LWD)

59
 Deflection method.‖ 2105 or 2106 Excavation and Embankment.

http://www.dot.state.mn.us/materials/gradingandbasedocs/LWD/2105-

06_LWD_051111.pdf.

Nebraska Department of Roads. 2011. ―Materials Management Standard Operating Procedures

Soil Moisture and Density.‖ http://www.dor.state.ne.us/mat-n-tests/pdfs-

docs/Materials_Management_Guidance/Soils/Final_MSOP_for_Soil_Moisture_and_Den

sity_Templates_March_2011.pdf

Nebraska Department of Roads. 2007. Standard Specifications for Highway Construction.

Lincoln: Nebraska Department of Roads.

Noorany, I., W. S. Gardner, D. J. Corley, and J. L. Brown. 2000. ―Variability in field density

tests.‖ Constructing and controlling compaction of earth fills. American Society for

Testing and Materials, West Conshohocken, PA, 58–71.

Prowell, B. D., and M. C. Dudley. 2002. ―Evaluation of Measurement Techniques for Asphalt

Pavement Density and Permeability.‖ Transportation Research Record #1789,

Transportation Research Board, Washington, D.C.

Rogge, D. F., and M. A. Jackson. 1999. ―Compaction and Measurement of Field Density for

Oregon Open-Graded (F-Mix) Asphalt Pavement.‖ Final Report, SPR 386. Oregon State

University, Corvallis, Oregon.

Romero, Pedro. 2002. ―Evaluation of Non-Nuclear Gauges to Measure Density of Hot-Mix

Asphalt Pavements.‖ The University of Utah, Salt Lake City, UT.

Schmitt, R. L., C. Rao, and H. L. Von Quintas. 2006. ―Non-Nuclear Density Testing Devices and

Systems to Evaluate In-Place Asphalt Pavement Density.‖ Report # 06-12; SPR# 0092-

05-10. Wisconsin Department of Transportation, Madison, Wisconsin.

Siekmeier, John. 2011. ―Unsaturated Soil Mechanics Implementation During Pavement

60
 Construction Quality Assurance.‖ 59th Annual Geotechnical Engineering Conference.

Siekmeier, J., M. Beyer, F. Camargo, P. Davich, J. Jensen, S. Merth, and C. Pinta. 2009. ―Using

the Dynamic Cone Penetrometer and Light Weight Deflectometer for Construction

Quality Assurance.‖ No. 2009-12. Minnesota Department of Transportation.

Sully Miller Contracting Company. 2000. ―A Summary of Operational differences Between

Nuclear and Non-Nuclear Density Measuring Instruments.‖ Quality Control Department.

TransTech Systems, Inc. 2004. ―Assessment of Field Asphalt Density Gauge Data When

Compared With Cores Processed According to AASHTO T-166.‖ Technical Report.

Von Quintus, H. L. 2009. ―NDT Technology for Quality Assurance of HMA Pavement

Construction.‖ Transportation Research Board, Washington, D. C., 33.

Wacharanon, Visan, Suppakorn Wachiraporn, and Auckpath Sawangsuriya. 2009. ―Innovative

Tools for Highway Construction Quality Control.‖

http://www.transtechsys.com/pdf/DOH%20article-english%20version.pdf.

61
 Appendix A: Site 1

Rice value: 151.00 (pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in Density Diff in Density

ABS (Core – ABS (Core –
PQI) Nuclear)
1 139.36 145.5 141.75 6.14 2.39
2 141.29 144.3 142.8 3.01 1.51
3 139.97 145.5 140.2 5.53 0.23
4 141.28 146.1 141.15 4.82 0.13
5 140.85 143.3 141.1 2.45 0.25
6 140.51 143.8 141.55 3.29 1.04
7 138.47 143.3 139.75 4.83 1.28
8 142.23 144.6 141.85 2.37 0.38
9 140.18 143.5 139.95 3.32 0.23
Average 140.46 144.43 141.12 3.97 0.66
Average difference (pcf) 3.97 0.81

62
 Appendix B: Site 2

Rice value: 154.40(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core –
PQI) Nuclear)
1 141.3 140.6 139.85 0.7 1.45
2 143.36 138.6 139.75 4.76 3.61
3 141.7 138.4 140.75 3.30 0.95
4 140.39 137.2 140.35 3.19 0.04
5 142.35 139.7 142.1 2.65 0.25
6 141.68 137.9 140.15 3.78 1.53
7 141.56 138.3 142.4 3.26 0.84
8 143.3 139.6 142.7 3.70 0.60
9 141.81 138.1 141.05 3.71 0.76
10 140.58 137.2 142.5 3.38 1.92
Average of difference (pcf) 3.24 1.20

63
 Appendix C: Site 3

Rice value: 154.75(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core –
PQI) Nuclear)
1 144.70 144 145.85 0.70 -1.15
2 144.69 144.1 145.35 0.59 -0.66
3 143.05 143.9 145.75 -0.85 -2.70
4 143.96 143.8 146 0.16 -2.04
5 145.61 142.4 146.3 3.21 -0.69
6 145.21 144.4 145.6 0.81 -0.39
7 145.25 143.9 146.15 1.35 -0.90
8 142.07 143.4 144.1 -1.33 -2.03
9 142.19 143.1 144.6 -0.91 -2.41
10 143.37 144 144.95 0.47 -1.58
Average of difference (pcf) -0.42 -1.46

64
 Appendix D: Site 4

Rice value: 150.76(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core –
PQI) Nuclear)
1 135.86 137.2 136.4 -1.34 -0.54
2 135.35 137 138.9 -1.65 -3.55
3 142.73 138.8 143.05 3.93 -0.32
4 136.59 137 136.9 -0.41 -0.31
5 141.46 139.2 143.1 2.26 -1.64
6 141.62 138.7 143.2 2.92 -1.58
7 143.49 137.8 143.55 5.69 -0.06
8 138.32 136.8 139.25 1.52 -0.93
9 135.41 137.6 139.8 -2.19 -4.39
Average of difference (pcf)
1.19 -1.48

65
 Appendix E: Site 5

Rice value: 151.88(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 138.84 141.4 138.55 -2.56 0.29
2 137.75 141.1 138.6 -3.35 -0.85
3 141.85 141.5 140.8 0.35 1.05
4 141.60 141.8 141.05 -0.20 0.55
5 141.70 142.3 140.25 -0.60 1.45
6 141.80 141.5 140.15 0.30 1.65
7 140.62 141.1 139.3 -0.48 1.32
8 140.41 141.7 139.85 -1.29 0.56
9 140.05 141.5 138.9 -1.45 1.15
Average of difference (pcf) -1.03 0.80

66
 Appendix F: Site 6

Rice value: 151.6(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – PQI) ABS (Core – Nuclear)
1 139.46 139 139 0.46 0.46
2 141.59 139.4 141.6 2.19 -0.01
3 144.43 139.1 143.75 5.33 0.68
4 139.69 138.2 139.85 1.49 -0.16
5 141.09 139.1 141.2 1.99 -0.11
6 144.7 140.1 144.45 4.60 0.25
7 142 139.5 143.2 2.50 -1.20
8 138.9 138.6 139.5 0.30 -0.60
9 141.23 139.1 141.85 2.13 -0.62
Average of difference (pcf) 2.33 -0.15

67
 Appendix G: Site 7

Rice value: 153.4(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core –
PQI) Nuclear)
1 135.58 136.5 136.8 -0.92 -1.22
2 133.92 137.7 135.25 -3.78 -1.33
3 139.65 137.8 134.55 1.85 5.10
4 138.97 139.5 139.75 -0.53 -0.78
5 136.4 138.2 139 -1.80 -2.60
6 136.79 138.7 140.2 -1.91 -3.41
7 132.77 136.5 139.3 -3.73 -6.53
8 136.4 138.4 139.1 -2.00 -2.70
9 136.55 138.2 141.6 -1.65 -5.05
Average of difference (pcf) -1.61 -2.06

68
 Appendix H: Site 8

Rice value: 152.82(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in Diff in density

density
ABS (Core ABS (Core – Nuclear)
– PQI)
1 137.2105 139.6 141.6 -2.39 -4.39
2 142.8013 139.8 143.2 3.00 -0.40
3 142.2401 139.4 141.25 2.84 0.99
4 143.3141 138.9 144.5 4.41 -1.19
5 136.4442 139.6 143.65 -3.16 -7.21
6 141.8341 137.3 142.4 4.53 -0.57
7 142.3126 138.2 142.5 4.11 -0.19
8 141.8765 138.4 141.45 3.48 0.43
9 143.3574 137.6 139.95 5.76 3.41
10 143.3295 138.4 142.95 4.93 0.38
Average of difference (pcf) 2.75 -0.87

69
 Appendix I: Site 9

Rice value: 153.4(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 139.5378 140.6 140.45 -1.06 -0.91
2 136.8946 138.9 136.6 -2.01 0.29
3 137.2703 139.3 137.15 -2.03 0.12
4 138.0902 140.1 139.75 -2.01 -1.66
5 137.6734 138.3 138.8 -0.63 -1.13
6 140.6103 140.1 138.6 0.51 2.01
7 142.3972 141.7 140.25 0.70 2.15
8 135.5922 139.2 137.25 -3.61 -1.66
9 139.3049 141.3 140.15 -2.00 -0.85
Average of difference (pcf) -1.35 -0.18

70
 Appendix J: Site 10

Rice value: 151.63(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 136.229 137.19 132.65 -0.96 3.58
2 135.724 138.29 136.75 -2.57 -1.03
3 135.416 137.59 135.3 -2.17 0.12
4 138.0347 138.09 138.6 -0.06 -0.57
5 138.274 137.29 133.25 0.98 5.02
6 137.14 137.29 134.65 -0.15 2.49
7 134.4266 137.29 137.15 -2.86 -2.72
8 137.0279 137.09 136.85 -0.06 0.18
9 137.2721 137.09 137 0.18 0.27
10 134.1401 137.09 139.15 -2.95 -5.01
11 133.2356 137.69 136.85 -4.45 -3.61
12 132.2507 137.39 135.1 -5.14 -2.85
13 132.662 137.09 135.9 -4.43 -3.24
14 132.477 136.79 134.6 -4.31 -2.12
15 139.283 137.09 136.55 2.19 2.73
Average of difference (pcf) -1.78 -0.45

71
 Appendix K: Site 11

Rice value: 152.94(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 135.34 134.77 134.6929 0.57 0.65
2 135.84 134.97 132.5612 0.87 3.28
3 134.94 133.92 132.1983 1.02 2.74
4 136.04 135.77 138.5644 0.27 -2.52
5 135.54 133.47 138.7553 2.07 -3.22
6 135.24 132.97 139.3517 2.27 -4.11
7 134.54 133.57 142.4331 0.97 -7.89
8 135.74 132.72 131.7215 3.02 4.02
9 135.54 134.62 138.0946 0.92 -2.55
Average of difference (pcf) 1.33 -1.07

72
 Appendix L: Site 12

Rice value: 153.19(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 139.2268 137.757 139.8241 1.47 -0.60
2 139.3268 138.807 138.0621 0.52 1.26
3 137.9268 139.007 137.1564 -1.08 0.77
4 139.0268 138.257 138.5112 0.77 0.52
5 139.0268 136.957 139.0936 2.07 -0.07
6 139.4268 139.957 140.0971 -0.53 -0.67
7 140.8268 140.707 138.8858 0.12 1.94
8 138.8268 139.257 139.5808 -0.43 -0.75
9 140.6268 139.357 135.0415 1.27 5.59
10 139.7268 138.207 132.98 1.52 6.75
11 139.2268 141.757 139.1269 -2.53 0.10
12 140.7268 141.657 134.7572 -0.93 5.97
13 138.7268 137.157 139.9493 1.57 -1.22
14 139.1268 142.557 141.1577 -3.43 -2.03
15 141.1268 142.157 131.9402 -1.03 9.19
16 139.4268 139.307 132.222 0.12 7.20
17 140.7268 141.357 134.831 -0.63 5.90
18 137.7268 141.457 134.1427 -3.73 3.58
19 141.1268 144.707 139.7142 -3.58 1.41
20 137.8268 137.257 134.9493 0.57 2.88
Average of difference (pcf) -0.40 2.39

73
 Appendix M: Site 13

Rice value: 153.19(pcf)

Num Core(pcf) PQI(pcf) Nuclear(pcf) Diff in density Diff in density

ABS (Core – ABS (Core – Nuclear)
PQI)
1 137.76 138.17 138.22 -0.41 -0.46
2 140.06 138.37 139.22 1.69 0.84
3 144.36 134.67 141.17 9.70 3.20
4 138.73 135.77 134.32 2.96 4.41
5 137.94 134.97 140.57 2.97 -2.63
6 141.28 135.67 145.97 5.61 -4.69
7 139.39 133.07 136.67 6.32 2.72
8 135.60 135.07 137.02 0.53 -1.42
9 135.41 137.17 133.12 -1.76 2.29
10 132.61 137.67 136.17 -5.06 -3.56
11 136.44 138.97 139.27 -2.53 -2.83
Average of difference (pcf) 1.82 -0.19

74
 Appendix N: Calibration results of different numbers of cores

N 3 cores 5 cores 8 cores 10 cores

O Date R TD LD R TD LD R TD LD R TD LD
- 0.7663 - 0.7391
1 8.28.09 -0.760 2.53269 1.35023 0.973 1.86626 0.80506 0.902 1.86626 3 0.948 1.86626 7
0.9845 0.7650
2 9.10.09 0.994 1.81987 0.90231 0.906 1.85501 0.85382 0.837 1.69901 1 0.839 1.71616 1
1.0936 0.9876
3 9.17.09 0.938 1.21021 1.79222 0.074 0.87958 0.90645 0.320 1.02632 9 0.394 0.87798 2
10.20.0 1.6943 1.6693
4 9 0.993 4.25486 2.37950 0.779 2.51261 6.68436 0.733 2.44728 0 0.606 2.43110 7
10.26.0 1.1591 1.1605
5 9 0.614 1.23372 1.12386 0.614 1.23372 1.12386 0.515 1.21085 0 0.599 1.21907 1

75
10.27.0 - 2.1723 3.0780
6 9 -0.680 2.61162 2.76696 0.680 2.61162 2.76696 0.242 1.74819 0 0.578 1.64344 8
2.7771 2.5197
7 5.27.10 0.845 2.41408 6.53573 0.584 2.41948 2.95257 0.328 2.53447 6 0.045 2.49281 9
1.7788 1.6092
8 6.24.10 0.391 2.61909 2.17532 0.764 2.38895 1.74972 0.619 2.55776 9 0.767 2.39969 0
2.7933 2.8669
9 6.30.10 0.799 2.82318 3.01438 0.121 2.72540 2.80460 0.082 2.78722 1 0.562 2.73893 5
1.6007 2.2454
10 8.6.10 0.944 2.16671 2.07993 0.831 2.08830 1.63053 0.576 2.59676 7 0.719 2.34962 2
- 3.6783 - 3.3633
11 8.25.10 0.715 3.38246 3.58584 0.050 3.55134 3.59592 -0.474 3.38246 4 0.152 3.38246 7
1.5576 1.3275
12 9.16.10 0.701 1.46234 1.95816 0.390 1.31725 1.46138 0.035 1.55122 0 0.090 1.31514 3
- 2.2289 - 2.2944
13 9.17.10 -1.000 3.02991 4.17139 0.744 3.00478 2.79913 -0.613 3.01491 8 0.577 3.15433 0
1.8681 1.8943
N/A
Avg 2.42775 2.60276 N/A 2.18879 2.31803 N/A 2.18636 0 N/A 2.12208 4
NOTE: Where, TD= Traditional difference, and LD= Linear regression difference