1. Report No.

FHWA/LA.06/417 2. Government Accession No. 3. Recipient's

Catalog No.

4. Title and Subtitle 5. Report Date

Comparative Evaluation of Subgrade Resilient Modulus from August 2007

Non-destructive, In-situ, and Laboratory Methods
6. Performing Organization Code
03-3P

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

Louay N. Mohammad, Ph.D.; Kevin Gaspard, P.E.; Ananda Herath, Ph.D., 417
P.E.; and Munir Nazzal, Ph.D.
9. Performing Organization Name and Address 10. Work Unit No.

Louisiana Transportation Research Center

4101 Gourrier Avenue 11. Contract or Grant No.
Baton Rouge, LA 70808 03-3P
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
FHWA LADOTD Final Report
Office of Technology Application P. O. Box 94245 7/2003-3/2006
400 7th Street, SW Baton Rouge, LA 70804
Washington, DC 20590 14. Sponsoring Agency Code

15. Supplementary Notes

Conducted in cooperation with the U.S. Department of Transportation, Federal Highway Administration
16. Abstract

Field and laboratory testing programs were conducted to develop models that predict the resilient modulus of subgrade soils from
the test results of DCP, CIMCPT, FWD, Dynaflect, and soil properties. The field testing program included DCP, CIMCPT, FWD,
and Dynaflect testing, whereas the laboratory program included repeated load triaxial resilient modulus tests and physical
properties and compaction tests. Nine overlay rehabilitation pavement projects in Louisiana were selected. A total of four soil
types (A-4, A-6, A-7-5, and A-7-6) were considered at different moisture-dry unit weight levels. The results of the laboratory and
field testing programs were analyzed and critically evaluated. A comprehensive statistical analysis was conducted on the collected
data. The results showed a good agreement between the predicted and measured resilient modulus from the various field test
methods considered. The DCP and CIMCPT models were enhanced when the soil moisture content and dry unit weight were
incorporated. The results also showed that, among all backcalculated FWD moduli, those backcalculated using ELMOD 5.1.69
software had the best correlation with the measured Mr. Finally, the Mr values estimated using the approach currently adopted by
the LADOTD were found to correlate poorly with the measured Mr values.

17. Key Words 18. Distribution Statement

Resilient Modulus, Miniature Cone Penetration, FWD, Dynaflect, Dynamic Unrestricted. This document is available through
Cone Penetration, Subgrade Soils the National Technical Information Service,
Springfield, VA 21161.
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
69
 Comparative Evaluation of Subgrade Resilient Modulus from
Non-destructive, In-situ, and Laboratory Methods
(Final Report)

by

Louay N. Mohammad, Ph.D.

Professor of Civil and Environmental Engineering
Director, Engineering Materials Characterization Research Facility

Kevin Gaspard, P.E.

Research Engineer

Ananda Herath, Ph.D., P.E.

Postdoctoral Researcher

And
Munir D. Nazzal, Ph.D.
Research Associate

LTRC Project No. 03-3P

State Project No. 736-99-1121

conducted for

Louisiana Department of Transportation and Development

Louisiana Transportation Research Center

The contents of this report reflect the views of the authors/principal investigator who are
responsible for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the views or policies of the Louisiana Department of Transportation and
Development, the Federal Highway Administration, or the Louisiana Transportation
Research Center. This report does not constitute a standard, specification, or regulation.

August 2007
 ABSTRACT

The resilient modulus (Mr) is a fundamental engineering material property that describes the
non-linear, stress-strain behavior of pavement materials under repeated loading. Mr attribute
has been recognized widely for characterizing materials in pavement design and evaluation.
The 1986 AASHTO guide for design of pavement structures has incorporated the Mr of
subgrade material into the design process. Considerable attention has also been given to it in
the design and evaluation of pavement structures in the Mechanistic-Empirical Pavement
Design Guide (MEPDG).

Field and laboratory testing programs were conducted to develop models that predict the
resilient modulus of subgrade soils from the test results of various test devices, namely,
Falling Weight Deflectometer (FWD), Dynamic Deflection Determination (Dynaflect),
Continuous Intrusion Miniature Cone Penetrometer (CIMCPT), and Dynamic Cone
Penetrometer (DCP). The field testing program included DCP, CIMCPT, FWD, and
Dynaflect testing, whereas the laboratory program included repeated load triaxial resilient
modulus tests, and physical properties and compaction tests. Nine overlay rehabilitation
pavement projects in Louisiana were selected. A total of four soil types (A-4, A-6, A-7-5,
and A-7-6) were considered at different moisture-dry unit weight levels. The results of the
laboratory and field testing programs were analyzed and critically evaluated. Subsequently,
statistical models for predicting the resilient modulus were developed. The results showed a
good agreement between the predicted and measured resilient modulus from the various field
test methods considered. Two models were developed for the DCP and CIMCPT, namely, a
direct model that includes the measurements of these devices and a soil property model that
includes the measurements of these devices as well as the physical properties of tested soils.
It was noted that the soil property models had a better prediction than the direct models. The
results also showed that, among all backcalculated FWD moduli, those backcalculated using
ELMOD 5.1.69 software had the best correlation with the measured Mr. Finally, no
significant correlation was found between the Mr values estimated using the approach
currently adopted by the LADOTD and those measured in the laboratory.

iii
 
 ACKNOWLEDGMENTS

The U.S. Department of Transportation, Federal Highway Administration (FHWA), the

Louisiana Department of Transportation and Development (LADOTD), and the Louisiana
Transportation Research Center (LTRC) financially support this research project.

The effort of William T. Tierney, research specialist/LTRC in conducting the cone

penetration tests and soil sampling is appreciated. Shawn Elisar, Glen Gore, Gary Keel, and
Mitch Terrel are greatly appreciated for conducting the falling weight deflectometer and
Dynaflect tests. The assistance of Amar Raghavendra in getting the MTS system operational
for the resilient modulus tests is acknowledged. The help of the geotechnical laboratory staff
in conducting various soil tests is also appreciated.

v
 
 IMPLEMENTATION IN PAVEMENT DESIGN

This report presents the results of a study conducted to develop resilient modulus prediction
models of subgrade soils from different in-situ tests, including: FWD, Dynaflect, CIMCPT,
and DCP.

The devices considered in this study can be utilized for design, construction, maintenance,
research, quality control/quality assurance, and forensic analysis. Each device and method
has its assets and liabilities. Practically speaking, the DCP will probably be utilized more by
the design, maintenance, and construction sections, simply because of its cost (< $2,500),
versatility, maintenance, and ease of use. Currently, only the LADOTD research section,
LTRC, owns and operates an FWD, a Dynaflect, and a CIMCPT. It is noted, as of this
writing, that the purchasing of a brand new FWD, Dynaflect, and CIMCPT would cost
$250,000, $80,000, and $100,000, respectively. The following sections provide a description
of the possible implementation of the considered in-situ test devices in the pavement design
and analysis procedures.

Dynamic Cone Penetrometer (DCP)

1) Design of New and Rehabilitated Pavements. LADOTD currently utilizes the 1993
AASHTO method to design its pavement. One of the factors used to determine the
pavement thicknesses is the subgrade resilient modulus. Instead of using the current
method, which utilizes an average value for each parish, the subgrade resilient
modulus could be determined by testing with the DCP. This would assure that the
resilient modulus would be accurately represented for the project. Furthermore, the
new Mechanistic-Empirical Pavement Design Guide requires that testing be
conducted to utilize level II data for design.
2) Forensic Analysis of Pavement Failures. This tool can be utilized to determine the
in place soil conditions (resilient modulus or Dynamic Cone Penetrometer Index
(mm/blow)) in areas in which pavement failures have occurred. With this
information, the design, construction, or maintenance engineer can make an accurate
assessment of the soil conditions and develop an appropriate rehabilitation strategy.

Falling Weight Deflectometer (FWD)

The FWD can be utilized with confidence in the design of rehabilitated pavements, as well as
for forensic analysis, due to good correlation with laboratory tests provided by this study. It

vii
is not a good tool for quality control because it is subject to inaccuracies when testing is
conducted directly on soils or unbound base courses, such as stone. It does have the
advantage of being able to assess the pavement structure quickly without having to drill holes
through the pavement structure, as is required with the DCP.

Dynamic Deflection Determination (Dynaflect)

The Dynaflect can be utilized with confidence in the design of rehabilitated pavements,
forensic analysis, and quality control due to good correlation with laboratory tests provided
by this study. Unlike the FWD, it can be used for quality control, but the DCP would be a
better choice, for reasons previously mentioned.

Continuous Intrusion Miniature Cone Penetrometer (CIMCPT)

CIMCPT can be used in similar situations as the DCP. It is less labor intensive and quicker
than the DCP. It has the advantage of being able to go deeper (greater than 25 feet) into the
subgrade than the DCP. The CIMCPT is suitable for the site conditions that require a cut.
However, it is mounted to a vehicle and thus less versatile and more costly to purchase and
maintain than the DCP.

Implementation Presentation and Guidelines

An implementation presentation can be developed and presented to each district to

familiarize personnel with the capabilities of each tool. Furthermore, a pavement analysis
guideline can be published and distributed within LADOTD. It is recommended that the
Engineering Directives and Standard Memo (EDSM) and pavement design manual of
LADOTD be revised to incorporate the use of these devices.

viii
 TABLE OF CONTENTS

Abstract .......................................................................................................................................... iii

Acknowledgments............................................................................................................................v

Implementation Statement ............................................................................................................ vii

List of Tables ............................................................................................................................... xi

List of Figures ............................................................................................................................ xiii

Introduction .....................................................................................................................................1

Background .................................................................................................................................... 3

CIMCPT Test Device .......................................................................................... 4

FWD Test Device ................................................................................................. 5
Dynaflect Test ....................................................................................................... 7
DCP Test Device.................................................................................................... 8

Objective .......................................................................................................................................11

Scope .............................................................................................................................................13

Methodology .................................................................................................................................15
`
Field and Laboratory Testing Program ............................................................................. 15

Descriptions of Testing Sites ................................................................................ 15

Description of Field Tests ......................................................................................18
Laboratory Testing ................................................................................................ 22

Discussion of Results .....................................................................................................................27

A Field Representative Resilient Modulus Value of Subgrade Soils .............................. 28
Development of Mr Prediction Models for the DCP Test Results ................................... 28
Development of Mr Prediction Models for the CIMCPT Test Results ............................. 36
Development of Mr Prediction Models for the FWD Test Results .................................. 45

Results of ELMOD 5.1.69 Backcalculation ................................................................45

Results of MODULUS 6 Backcalculation ...................................................................45
Results of EVERCALC 5.0 .........................................................................................51

Results of Florida Equation ............................................................................................. 52

ix
 Development of Mr Prediction Models for the Dynaflect Test Results ........................... 52
Results of the LADOTD Method ..................................................................................... 56

Limitations of the Models ................................................................................................. 56

Conclusions ................................................................................................................................... 59

Recommendations ..........................................................................................................................61

References ..................................................................................................................................... 63

Appendix A ..................................................................................................................................66

x
 LIST OF TABLES

Table 1 Input levels for the M-E Design Guide ..............................................................................4

Table 2 Summary of CIMPT models developed by Mohammad et al [10, 11]..............................6

Table 3 Mr -DCP correlations reported in literature .....................................................................10

Table 4 Test factorial ....................................................................................................................16

Table 5 Soil classification test procedures ....................................................................................24

Table 6 Dry unit weights and moisture contents of soil tested .....................................................25

Table 7 Physical properties of soils tested ....................................................................................26

Table 8 DCP and laboratory Mr test results (this study) .............................................................29

Table 9 DCP and laboratory Mr test results [20] ..........................................................................30

Table 10 Ranges of variables of subgrade materials used in DCP model development ............30

Table 11 A correlation matrix for the DCP test results (p-value) .................................................32

Table 12 A correlation matrix for the DCP test results (r-value) .................................................32

Table 13 Summary of stepwise selection......................................................................................34

Table 14 Summary of multiple regression analysis for variable selection ...................................34

Table 15 Results of analysis of DCP- Soil Property Model .........................................................35

Table 16 CIMCPT and laboratory Mr test results for this study (this study) ...............................37

Table 17 CIMCPT and laboratory Mr test results [10] .................................................................38

Table 18 Ranges of variables of subgrade materials used in CIMCPT model development .......39

Table 19 A correlation matrix for the CIMCPT test results (p-value) ........................................41

Table 20 A correlation matrix for the CIMCPT test results (r-value) ..........................................41

Table 21 Results of the variable selection for CIMCPT- Mr model ............................................42

Table 22 Results of the multiple regression for CIMCPT- Mr model .........................................43

xi
Table 23 Results of FWD backcalculation using ELMOD software ............................................46

Table 24 Results of statistical analysis for Mr -FWD (ELMOD 5.1.69) model ...........................47

Table 25 Results of FWD backcalculation analysis using MODULUS 6 software .....................50

Table 26 Results of statistical analysis for Mr -FWD (MODULUS 6) model..............................51

Table 27 Results of FWD backcalculation using EVERCALC 5.0 and Florida equation .........53

Table 28 Dynaflect test results ...................................................................................................55

Table 29 Dynaflect statistical results ..........................................................................................58

xii
 LIST OF FIGURES

Figure 1 A typical friction cone penetrometer .................................................................................5

Figure 2 Continuous intrusion miniature cone penetration..............................................................5

Figure 3 Dynatest Model 8000 (FWD) ............................................................................................7

Figure 4 Typical DYNAFLECT deflection basin ............................................................................8

Figure 5 (a) The DCP test (b) A typical DCP profile ....................................................................10

Figure 6 Field-testing layout for each set ......................................................................................13

Figure 7 Locations of the pavement projects ................................................................................17

Figure 8 Pavement structures .........................................................................................................20

Figure 9 Shelby tube specimen location ........................................................................................21

Figure 10 MTS Triaxial Testing Machine .....................................................................................23

Figure 11 Variation of resilient modulus with DCPI .....................................................................31

Figure 12 Variation of resilient modulus with Log(DCPI) ............................................................31

Figure 13 Variation of resilient modulus with 1/DCPI ..................................................................31

Figure 14 Variation of resilient modulus with γd ...........................................................................31

Figure 15 Variation of resilient modulus with w ...........................................................................31

Figure 16 Variation of resilient modulus with γd/w .......................................................................31

Figure 17 Variation of laboratory measured Mr with 1/DCPI .......................................................34

Figure 18 Residuals from DCP-soil property model .....................................................................35

Figure 19 Laboratory measured Mr vs. values predicted from DCP-soil property model .............36

Figure 20 Variation of tip resistance with resilient modulus .........................................................40

Figure 21 Variation of sleeve friction with resilient modulus .......................................................40

Figure 22 Variation of resilient modulus with γd/w .......................................................................40

xiii
Figure 23 Variation of resilient modulus with γd ...........................................................................40

Figure 24 Variation of resilient modulus with w ...........................................................................40

Figure 25 Predictions from the CIMCPT-direct model .................................................................43

Figure 26 Residuals from CIMCPT-soil property model ..............................................................44

Figure 27 Predictions from the CIMCPT-soil property model ......................................................44

Figure 28 Mr versus FWD modulus backcalculated ELMOD 5.1.69 (7-sensor with no seed
values) ............................................................................................................................................47

Figure 29 Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (9-sensor with no seed
values) ............................................................................................................................................48

Figure 30 Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (9-sensor with seed
values) ............................................................................................................................................48

Figure 31 Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (calibration = 2) ...............49

Figure 32 Mr vs. FWD moduli backcalculated using MODULUS 6 (semi infinite subgrade
layer) ..............................................................................................................................................51

Figure 33 Mr vs. FWD moduli backcalculated using MODULUS 6 (finite depth) .......................52

Figure 34 Mr vs. FWD moduli backcalculated using EVERCALC 5.0 ........................................54

Figure 35 Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 Florida equation ...............54

Figure 36 Dynaflect statistical results ............................................................................................56

Figure 39 LADOTD method estimated resilient modulus .............................................................57

xiv
 INTRODUCTION

The resilient modulus of pavement materials and subgrades is an important input parameter
for the design of pavement structures. Therefore, an accurate measurement of Mr is needed to
ensure the efficiency and accuracy of the pavement design. Many studies that were
conducted to demonstrate the effects of pavement materials’ Mr on the design of pavements
showed that the input value of Mr has a dramatic effect on the designed thickness of the base
course and asphalt layers.

The resilient modulus of pavement materials is typically determined using the RLT test.
However, this test requires well trained personnel and expensive laboratory equipment. In
addition, it is considered to be relatively time consuming. Therefore, highway agencies tried
to seek different alternatives. Various empirical correlations have been used to determine
resilient modulus in the last three decades. The resilient modulus of subgrade soils is related
to several parameters, such as the soil support value (SSV), the R-value, the California
bearing ratio (CBR), and the Texas triaxial classification value. However, these parameters
do not represent the dynamic load behavior under moving vehicles.

To overcome the disadvantages in the subgrade Mr estimation procedures, different in-situ

techniques were proposed to determine the Mr of different pavement materials. These
techniques are characterized by the ease of operation and their ability to assess the structural
integrity and estimate the elastic moduli of in situ pavement layers. They have an additional
advantage of being able to assess the pavement structure without destroying it.

This study was initiated to evaluate the use of different in situ testing devices as an
alternative for determining the pavement materials Mr through laboratory triaxial tests. For
this purpose, field and laboratory testing programs were performed. The field program
included conducting CIMCPT, FWD, Dynaflect, and DCP tests on nine pavement projects.
In addition to the laboratory repeated load triaxial resilient modulus, physical soil properties
tests were performed on samples from the tested sections. Statistical analyses were
performed to develop models that predict the resilient modulus measured in the laboratory
based on the results obtained from the different in situ testing devices considered.
 BACKGROUND

The resilient modulus is a fundamental engineering material property that describes the non-
linear stress-strain behavior of pavement materials under repeated loading. It is defined as the
ratio of the maximum cyclic stress (σcyc) to the recoverable resilient (elastic) strain (εr) in a
repeated dynamic loading. The American Association of State Highway Transportation
Officials (AASHTO) 1993 and the MEPDG have adopted the use of resilient modulus of
subgrade soils as a material property in characterizing pavements for their structural analysis
and design. The MEPDG provided three different levels of input as a means for obtaining the
resilient modulus of subgrade materials. The levels are presented in Table 1.

The Mr is typically determined in the laboratory through conducting the Repeated Load
Triaxial (RTL) test on representative material samples. Generally, the RLT test requires well
trained personnel and expensive laboratory equipment; it is also considered relatively time
consuming. Therefore, different state agencies were hesitant to conduct it, and instead used
different approaches to estimate the Mr. One of these approaches is the use of empirical
correlations with physical properties of tested soils. During the last three decades, various
empirical correlations have been proposed and used to predict Mr. Van Til et al [1] related Mr
of subgrade soils to the soil support value (SSV) employed in the earlier AASHTO design
equation. They also made a correlation chart in which the values of Mr can be determined by
the internal friction of the R-value, the CBR, and the Texas triaxial classification value.
Many other correlations between Mr, the CBR, the R-value, and soil support values were also
developed [2]. The Louisiana Department of Transportation and Development (LADOTD)
has historically estimated the Mr of subgrade soils based on the soil support value (SSV)
using the following equation:

2
⎛ ⎛ 53 ⎞ ⎞ ⎛ ⎛ 53 ⎞ ⎞
M r = 1500 + 450 ⎜ ⎜ ⎟
⎜⎝ 5 ⎠ (SSV - 2) ⎟ - 2.5 ⎜ ⎜
⎟ ⎜ ⎝ 5 ⎟⎠ (SSV - 2) ⎟
⎟
(1)
⎝ ⎠ ⎝ ⎠

where
Mr = resilient modulus and
SSV = soil support value.

The SSV is obtained from a database based on the parish system in Louisiana. Currently, the
LADOTD uses a typical Mr value for each parish instead of obtaining subgrade Mr values for
each project. This can lead to inaccuracies in the pavement design, since the subgrade Mr can

3
vary from site to site within the parish as well as seasonally. Thus, the use of Mr based on a
typical parish value can result in an under design of pavement structure leading to premature
pavement failures.

Table 1
Input levels for the M-E design guide [3]

Material Input Level 1 Input Level II Input Level III

Granular Materials Measured Mr in Estimated Mr from Default Mr
laboratory correlations
Cohesive Measured Mr in Estimated Mr from Default Mr
Materials laboratory correlations

Another alternative for estimating the Mr of subgrade soils is the use of in situ test devices.
Different devices have been proposed and used during the last few decades. The following
sections give a brief background of the in situ devices investigated in this study.

CIMCPT Test Device

The CIMCPT is a simple and economical test that provides rapid, continuous, and reliable
measurements of the soil physical and strength properties. As shown in Figures 1 and 2, the
CIMCPT device consists of a continuous push device, hydraulic motor, miniature cone
penetrometer, and data acquisition system. The cone is attached to a coiled push rod, which
allows a continuous penetration, and is mechanically straightened as the cone is pushed into
the soil. As the miniature cone penetrates into the ground, the tip resistance (qc) and sleeve
friction (fs) readings are recorded. The penetration resistance is related to the strength of the
soil. The tip resistance depends on the size of the cone tip, rate of penetration, types of soil,
density, and moisture content.

During the last few decades, the CIMCPT test has gained popularity among other in situ tests
in the characterization of subgrade soil, the construction control of embankments, the
assessment of the effectiveness of ground modification, and other shallow depth (upper 5 to
10 m) applications [4]. Mohammad et al. developed different models for predicting the
resilient modulus of coarse and fine soils from the CIMCPT test results [5-13]. A summary
of these models is presented in Table 2.

4
 Figure 1
A typical friction cone penetrometer

Figure 2
Continuous intrusion miniature cone penetration

FWD Test Device

Based on early work in France during the 1960s, the Technical University of Denmark, the
Danish Road Institute, and the Dynatest Group have gradually developed and employed the
FWD for use as nondestructive testing of highway and airfield pavements. The FWD is a
trailer mounted device that delivers an impulse load to the pavement, as shown in Figure 3.
The equipment automatically lifts a weight to a given height. The weight is dropped onto a
300 mm circular load plate with a thin rubber pad mounted underneath. A load cell measures
the force or load applied to the pavement under the plate, and the deflections caused by the
impulse load are measured by sensors placed at different distances from the center of the load
plate. Based on the measured load and deflections of the elastic moduli of the tested
pavement, layers can be backcalculated using one of the different softwares available, such
as MODULUS, ELMOD or EVERCALC.

Because of its versatility and ease of use, the FWD is becoming the device of choice of
highway agencies. The Florida Department of Transportation conducted a survey of the 50
states and three Canadian provinces to assess the current practices of using FWD [14]. Their

5
results indicate that 70 percent of the surveyed agencies use the modulus determined from the
FWD data to estimate subgrade strength.

The relation between the moduli obtained from FWD and the laboratory measured resilient
modulus was examined in previous studies. Rahim et al [15] suggested that, for different
types of cohesive and granular soils, the FWD moduli backcalculated using MODULUS 5.0
software was, on average, identical to the laboratory measured Mr.

Table 2
Summary of CIMPT models developed by Mohammad et al. [10, 11]

Correlation Comment
Mr 1 ⎛ f ⎞ γ Fine grained soil based on the
= ⎜ 31.79qc + 74.81 s ⎟ + 4.08 d in situ stresses
σc0.55 σv ⎜⎝ w ⎟⎠ γw

M (q σ ) f γ
r = 6.66 c b − 32.99 s + 0.52 d Coarse grained soil based on
σ 0 .55 σ 2 q ( wγ ) the in situ stresses
c v c w
M q f γ
r = 47.03 c + 170.40 s + 167 . d Fine grained soil based on the
σ 0.55 σ σ w γ traffic and in situ stresses
3 1 1 w
Mr qcσ b γ
0.55 = 18.95 + 0.41 d Coarse grained soil based on
σ3 σ12
γ ww the traffic and in situ stresses
Note:
Mr- resilient modulus (MPa),
σ3- minor principal stress (σc- confining) (kPa),
σ1- major principal stress (σv- vertical stress) (kPa),
qc - tip resistance(MPa),
f s- sleeve friction (MPa),
w- water content (as a decimal),
γd- dry unit weight (kN/m3), and
γw- unit weight of water (kN/m3)
σb - bulk stress

6
 Figure 3
Dynatest Model 8000 (FWD)

Dynaflect Test Device

The Dynamic Deflection Determination (Dynaflect) is an electromagnetic system for

measuring the dynamic deflection of a surface or structure caused by an oscillatory load.
Measurements are independent of a fixed surface reference. The deflections measured on
flexible pavements by the Dynaflect system have been correlated to those obtained by the
Benkleman Beam by a number of research groups in highway departments and universities.
The Dynaflect induces a dynamic load on the pavement and measures the resulting
deflections using geophones, usually five, spaced under the trailer at approximately 300 mm
(1 foot) intervals from the application of the load. The pavement is subjected to 4.45 kN
(1000 lbf) of dynamic load at a frequency of 8 Hz, which is produced by two counterrotating,
unbalanced flywheels. The cyclic force is transmitted vertically to the pavement through two
steel wheels, spaced 508 mm (20 inches) from center to center. The dynamic force during
each rotation of the flywheels varies from 4.9 to 9.3 kN (1100 to 2100 lbf).

7
 Figure 4 shows a typical Dynaflect deflection basin. The Dynaflect measures only
half of the deflection bowl, while the other half is assumed to be a mirror image of the
measured portion. In Figure 4, the measurement W1 is the maximum depth of the deflection
bowl and occurs near the force wheels. The terms W2, W3, W4, and W5 are the deflections at
geophones 2, 3, 4, and 5, respectively.

The maximum deflection, W1 provides an indication of the relative strength of the

total road section. The surface curvature index, SCI (W1-W2), provides an indication of the
relative strength of the upper (pavement) layers. The base curvature index, BCI (W4-W5),
and the fifth sensor value W5 provide a measure of the relative strength of the foundation.
For all four parameters, W1, SCI, BCI, and W5, lower values indicate greater strength.

To the knowledge of the authors, no research was conducted to correlate the Dynaflect test
measurements to the resilient modulus of subgrade soils.

Figure 4
Typical DYNAFLECT deflection basin

DCP Test Device

DCP is a portable instrument that consists of an 8 kg sliding hammer, an anvil, a pushing rod
(diameter 16 mm), and a steel cone tip, as shown in Figure 5a. The cone tip angle is 60
degrees, and its diameter is 20 mm. The diameter of the pushing rod is less than that of the
cone base. This design assists in reducing the frictional forces along the wall of the cone
penetrometer. The DCP test consists of pushing a conical tip, attached to the bottom of the
pushing rod, into the soil layer and measuring the resistance to penetration.

8
DCP tests are designed to estimate the structural capacity of pavement layers and
embankments. The DCP has the ability to verify both the level and the uniformity of
compaction, which makes it an excellent tool for the quality control of pavement
construction. In addition, it can also be used to determine the tested pavement’s layer
thickness.

During the past decades, the DCP measurement has been correlated to many engineering
properties, such as the CBR, shear strength, and elastic modulus. In addition, different
models were developed to predict the laboratory measured Mr using DCP test results. A
summary of these models is presented in Table 3. The MEPDG software also used the DCP
results to estimate the Mr values of different pavement layers by first computing the
California bearing ratio (CBR) using the CBR-DCP relation proposed by Webster [16]
(Equation(2)) and then predicting Mr based on the Mr-CBR relation suggested by Powell et
al. [17] (Equation(3)). However, since the CBR is estimated using a static test, these types of
correlations do not take into account the dynamic behavior of pavements under moving
vehicles.

292
CBR = (2)
DCPI1.12

M r = 17.58 (CBR)0.64 (3)

where
Mr = resilient modulus in MPa, and
DCPI = penetration index, mm/blow

9
 (a) (b)

Figure 5
(a) The DCP test (b) A typical DCP profile

Table 3
Mr-DCP correlations reported in Literature
Study Correlation Soil type Comment
Hasan [18] M r = 7013.065 − 2040.783ln(DCPI) Cohesive Mr in psi, DCPI in in/blow

M r = a o ( DCPI )
a1
(γ dr
a2
+ ( LL / w c )
a3
) Cohesive Mr in psi,
DCPI in in/blow;
Wc is moisture content;
LL is Liquid limit ;
George et al.
cu is coefficient of uniformity;
[19]
M r = a o (DCPI / log cu )a1 (w cr a 2 + γ dr a3 ) Granular wcr= field moisture ; and
optimum moisture
field γ d
γ dr =
maximum γ d
ao,a1,a2 and a3 model
coefficients.

10
 OBJECTIVE

The objective of this research is to develop models that predict the resilient modulus of
subgrade soils from the test results of various in situ test devices, namely, DCP, CIMCPT,
FWD, and Dynaflect, along with properties of tested soils. The study also evaluates the
advantages and limitations for the different in situ devices considered. The results of this
study will be used to develop guidelines for the implementation of the measurements of the
considered in situ test devices in pavement design procedures including the new Mechanistic-
Empirical pavement design method.

11
 SCOPE

Nine pavement projects in Louisiana were selected for field FWD, Dynaflect, CIMCPT, and
DCP tests. These projects were LA333, LA347, US171, LA991, LA22, LA28, LA344,
LA182, and LA652. Three sets (A, B, and C) of tests were conducted at each pavement
project site, as shown in Figure 6. Each testing set was approximately 500-ft apart, unless
field conditions dictated otherwise. Each set contained nine points (1 to 9). A total of four
soil types (classified as A-4, A-6, A-7-5, and A-7-6, according to the AASHTO soil
classification) were considered at different moisture-dry unit weight levels. The DCP tests
were performed at points 1, 4, and 7 in a set. The FWD and Dynaflect tests were performed
at all nine points in a set. The CIMCPT tests were performed at points 3, 6, and 9 in a set.
The field experimental program also included obtaining Shelby tube soil sampling at points
2, 5, and 8. Once testing was completed, subgrade material was augered out of points 2, 3, 5,
6, 8, and 9 and used to perform classification tests. The laboratory experimental program
consisted of repeated load triaxial resilient modulus on the Shelby tube specimens. In
addition, test results from recently completed research projects were also incorporated in the
model development [10,21].

Figure 6
Field-testing layout for each set

13
14
 METHODOLOGY

Field and laboratory testing programs were performed on soils of nine pavement projects in
Louisiana. Field testing consisted of conducting FWD, Dynaflect, CIMCPT, and DCP tests.
Furthermore, the laboratory program included conducting repeated load triaxial resilient
modulus tests and physical properties and compaction tests. Laboratory tests consisted of the
determination of resilient modulus and properties of investigated soils. A typical layout of the
field testing program is shown in Figure 6. Table 4 presents the test factorial of this study.

Field Testing Program

The following sections present a description of the sites considered in this study. A brief
description of the in situ tests and the testing procedures pursued in this study is also
provided.

Descriptions of Testing Sites

Both the LADOTD headquarters pavement and geotechnical design engineer and LADOTD
district design and water resources engineer sections were consulted to obtain the location of
projects that were currently in the design or construction process. These projects
encompassed various pavement typical sections and soil conditions and thus allowed
representative samples of the soils typically encountered in Louisiana highway construction
to be evaluated. Figure 7 presents the locations of each testing site, while the pavement for
the projects selected is shown in Figure 8. A brief description of each site is provided below.

Route LA 333. This project is located in Vermillion Parish, and testing was conducted in the
northbound lane. Site testing was conducted at locations with minimal cracking to reduce
errors in the data collection process, though such locations were difficult to locate. The
pavement typical section consisted of 6 inch thick asphalt concrete pavement, 8.5 inch thick
soil cement base course, and a clay embankment with a plastic index (PI) ranging from 22 to
26.

Route LA 347. This project is located in St. Landry Parish, and testing was conducted in the
southbound lane. Site testing was conducted at locations with minimal cracking to reduce
errors in the data collection process. The typical pavement section consisted of 5 inch thick
asphaltic concrete pavement, 8.5 inch thick soil cement base course, and a clay subgrade with
a PI ranging from 27 to 38.

15
Route US 171. This project is located in Beauregard Parish, and testing was conducted in
the northbound lane. Since the wearing course was scheduled to be placed later, the typical
pavement section that was tested consisted of 5 inch thick asphaltic concrete binder course,
10-inch thick crushed stone base course, 12 inch thick cement treated subbase, and a clay
subgrade with a PI ranging from 12 to 29.

Table 4
Test Factorial

Project Site Lab. Mr FWD DCP CIMCPT Dynaflect Shelby

(test (test (test (test (test points) tubes
points) points) points) points) (test
points)
A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA333 B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA347 A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
US171 A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA991 A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA22 A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA28 B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA344 B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA182 B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
A 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
LA652 B 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
C 2,5,8 1 to 9 1,4,7 3,6,9 1 to 9 2,5,8
Legend: FWD- Falling weight deflectometer, DCP- Dynamic cone penetration, CIMCPT- Continuous
intrusion miniature cone penetration test, Lab. Mr -Laboratory measured resilient modulus

16
 Figure 7
Locations of the pavement projects

Route LA 991. This project is located in Iberville Parish, and testing was conducted on the
westbound lane. The typical pavement section consisted of 4 inch thick asphaltic concrete
pavement, 12 inch thick soil cement base course, and a clay subgrade with a PI ranging from
13 to 26.

Route LA 22. This project is located in Ascension parish, and testing was conducted on the
eastbound lane. The section selected for testing had received a maintenance overlay to repair
failed pavement areas. The typical pavement section varied. For site A, the asphaltic
concrete was 17inches thick. Sites B and C had an asphaltic concrete pavement thickness of
13 inches. The asphalt concrete thicknesses for each site includes the thickness of the
asphaltic concrete wearing, binder, and base course. Each site had a clay subgrade with a PI
ranging from 20 to 24.

Route LA 28. This project is located in Vernon Parish, and testing was conducted on the
eastbound outside shoulder. The pavement shoulder typical section consisted of 5 inch thick

17
asphaltic concrete pavement and 10.75 inch thick crushed stone. Each site had a clay
subgrade with a PI ranging from 43 to 61.

Route LA 344. This project is located in Iberia Parish, and testing was conducted on the
eastbound lane. The pavement section consisted of 7.25 to 6 inch thick asphaltic concrete
pavement and 7.5 to 7.0 inch thick soil cement base course. Sites A and B had a heavy clay
subgrade with a PI ranging from 34 to 39, and Site C had a lean silt subgrade.

Route LA 182. This project is located in Lafourche Parish, and testing was conducted on the
eastbound shoulder. The shoulder section was less than two years old and showed no signs of
distress. The asphalt pavement thickness varied from 2 to 3 inches, and soil cement base
course varied from 8 to 8.25 inches. Each site had a lean clay subgrade with an average PI of
23.

Route LA 652. This project is located in Lafourche Parish, and testing was conducted on the
eastbound lane. The asphalt pavement ranged from 3.3 to 3.9 inches, and the soil cement
base course ranged from 8.9 to 9.4 inches. Each site had a heavy clay subgrade with a PI
ranging from 46 to 50.

Description of Field Tests

A visual survey of each of the tested sites was conducted prior to performing the different
field tests. Based on this survey, a testing layout was established. The field testing included
using different in situ test devices. A brief description of those tests is presented in the
following sections.

FWD Tests

FWD tests were conducted on all nine points for each testing set, as presented in Figure 6.
The Dynatest Model 8000 was used in this study to conduct all FWD tests. This device
applies a transient load (approximately a half-sinusoidal wave with a loading time between
25 and 40 milliseconds) to the pavement layer by dropping a weight from a specified height
on a 300 mm circular loading plate with a thin rubber pad mounted underneath. Different
load magnitudes can be generated by varying the mass of weight and drop height. A 9,000-
pound load level was used in this study. The pavement deformation induced by the applied
load is obtained using sensors (geophones) located at different distances from the center of
the load plate. In this study, the deformation was obtained using nine sensors. Based on the
measured load and deflections, the elastic moduli of the different tested pavement layers
were backcalculated using the different softwares and methods described below.

18
Florida Equation. The Florida Department of Transportation (FDOT) developed the
following equation, known as Florida equation, to determine the subgrade resilient modulus
[21]:
0.898
⎛ P⎞
E FWD = 0.03764 ⎜⎜ ⎟⎟ , (4)
⎝ dr ⎠

where
EFWD = subgrade resilient modulus estimated from the FWD results (psi),
P = applied load (pounds), and
dr = sensor deflection at 36 inches from the load plate [thousands of an inch (mils)].

ELMOD software version 5.1.69 [22]. This software was developed by Dynatest
International, and it uses the Microsoft Access database for storing data from the field
acquisition and backcalculation results. Different input values influence the backcalculated
layer moduli values; these include: layer thickness, seed values, max depth to rigid layer,
linear, non-linear, radius of curvature fit, and deflection basin fit.

MODULUS software version 6.0 [23].This software was developed by the Texas
Transportation Institute (TTI). It is a friendly program that has built in references to assist in
the backcalculation process. The backcalculations were performed with semi-infinite
subgrade and finite subgrade depths to bedrock models.

EVERCALC software version 5.0 [24]. This software was developed by the Washington
Department of Transportation (WSDOT). The program uses the WESLEA layered elastic
analysis program for forward analysis and a modified Augmented Gauss-Newton algorithm
for optimization. It can handle up to 5 layers, 10 sensors, and 12 drops per station.

Dynaflect Tests. Dynaflect tests were conducted at each of the nine points of each tested
site. Since the Dynaflect deflections should be corrected for the temperature as well as for
other variables, the procedure for determining Dynaflect deflection correction factors,
developed by Southgate [25], was utilized to adjust the Dynaflect deflections to a standard
temperature of 60O F. The fact that the applicability of the procedure used to the conditions
and construction materials in Louisiana was verified in a previous study is worth noting [26].

DCP Tests. DCP tests were conducted on three points in each testing set, as presented in
Figure 6. To perform the DCP tests, a one inch diameter hole was first drilled through the
asphalt concrete pavement and base course with a Dewalt Rotary hammer drill. The DCP

19
cone was then lowered through the hole and placed on the subgrade. The depth of penetration
into the subgrade varied from approximately 24 to 36 inches, depending on site conditions.
The field DCP tests were performed according to the American Society for Testing and
Materials (ASTM) test procedure, D6951. During a typical DCP test, the penetration depth of
DCP for each hammer drop (blow) was recorded and used to plot the DCP profile (blows vs.
depth) for the tested soil. The DCPI value was then determined as the slope of that profile.

LA 333- Pavement LA 347- Pavement US 171- Pavement

6 in.- Asphalt concrete 5 in.- Asphalt concrete 5 in.- Asphalt concrete

8.5 in.- Soil cement base 8.5 in.- Soil cement base 10 in.- Stone base

A-6/ A-7-6 Clay A-7-5 Clay 12 in.- Cement-treated soil

A-6/ A-7-5 Clay

(a) (b) (c)

LA 991- Pavement LA 22- Pavement LA28- Pavement

4 in.- Asphalt concrete Asphalt concrete 5 in.- Asphalt concrete

(17 in.-for site A)
12 in.- Soil cement base 10.75 in. - Stone base
(13 in.- for site Band C)

A-6/ A-7-6 Clay A-6/ A-7-6 Clay A-7-6/ A-7-5 Clay

(d) (e) (f)

LA 344- Pavement LA 182- Pavement LA652- Pavement

7.25 in.- Asphalt concrete 2.5 in.- Asphalt concrete 3.9 in.- Asphalt concrete

7 in.- Soil cement base 8 in.- Soil cement base 9 in.- Soil cement base

A-7-6 Clay A-7-6 Clay A-7-5 Clay

(g) (h) (i)

Figure 8
Pavement structures
CIMCPT Tests. CIMCPT tests were conducted on three points in each testing set, as
illustrated in Figure 6. The miniature cone penetrometer used in this study had a cross

20
sectional area of 2 cm2, a friction sleeve area of 40 cm2, and a cone apex angle of 60 degrees
and was attached to a coiled push rod, which replaces the segmental push rods in the standard
cones. Prior to conducting the CIMCPT tests, a six inch diameter hole was augured through
the asphaltic concrete pavement and base course with a core rig. The six inch diameter hole
was augured approximately six inches into the subgrade to ensure that any loose aggregate
from the asphaltic concrete or base course was removed from the hole. Once the hole was
augered, the cone was advanced into the ground at a rate of 2 cm/sec to a depth of
approximately nine feet below the base course with continuous measurements of the tip
resistance (qc) and sleeve friction (fs).

Shelby Tube Samples. Shelby tube samples were obtained at three points for each test
section, as shown in Figure 6. To obtain Shelby tube samples, a six inch diameter hole was
first augured with a core rig through the asphaltic concrete layer, and the base course layer
and six inches into the subgrade. The core rig was then used to shove the three inch diameter
Shelby tube into the subgrade. Although the Shelby tubes were 30 inches long and were fully
pushed into the subgrade, only a 5.8-inch long specimen could be obtained from the tube.
The obtained specimen was representative of the subgrade soil layer within 6 to 18 inches
from the base course, as shown in Figure 9.

Once the tube was removed from the ground, the soil specimen was extracted from the tube
using the extrusion device mounted on the truck. The soil specimens were then trimmed and
wrapped in plastic and aluminum foil. They were then stored in Styrofoam containers and
transported to the LTRC laboratory. The samples were kept in a 95 percent relative humidity-
controlled room until they were tested.

6”
AC Pavement
and base course

Augered hole 6”

Shelby tube 6” to 12”

specimen

3”

Figure 9
Shelby tube specimen location

21
Laboratory Testing Program
The laboratory testing program in this study consisted of conducting RLT resilient modulus
tests and tests to determine the physical properties of tested soils. The following sections
provide a description of these tests.

RLT Resilient Modulus Test

RLT Mr tests were conducted on the 5.6 inches high and 2.8 inches wide specimens obtained
from Shelby tube samples collected in the field. All tests were performed using the Material
Testing System (MTS) 810 machine with a closed loop and a servo hydraulic loading system.
The applied load was measured using a load cell installed inside the triaxial cell. Placing the
load cell inside the triaxial chamber eliminates the push-rod seal friction and pressure area
errors and results in a reduction in the testing equipment error. An external load cell is
affected by changes in confining pressure and load rod friction, and the internal load cell,
therefore, gives more accurate readings. The capacity of the load cell used was ± 22.25 kN
(±5000 lbf.). The axial displacement measurements were made using two linearly variable
differential transducers (LVDT) placed between the top platen and base of the cell to reduce
the amount of extraneous axial deformation measured compared to external LVDTs. Air was
used as the confining fluid to the specimens. Figure 10 depicts a picture of the testing setup
used in this study.

Resilient modulus tests were performed in accordance with AASHTO procedure T 294-94
[27] standard method. In this test method, the samples are first conditioned by applying
1,000 load cycles to remove most irregularities on the top and bottom surfaces of the test
sample and to suppress most of the initial stage of permanent deformation. The conditioning
of the samples is followed by a series of steps consisting of different levels of cyclic
deviatoric stress, such that the resilient modulus is measured at varying normal and shear
stress levels. The cyclic loading consists of repeated cycles of a haversine shaped load pulse.
These load pulses have a 0.1 sec load duration and a 0.9 sec rest period.

Results obtained from the resilient modulus test were used to determine the non-linear elastic
coefficients of the generalized constitutive model shown in Equation 5, which were used to
determine the resilient modulus values at a field representative stress state.
k2 k3
Mr ⎛ θ ⎞ ⎛ τoct ⎞
= k1 ⎜⎜ ⎟⎟ ⎜⎜ + 1⎟⎟ , (5)
Pa ⎝ Pa ⎠ ⎝ Pa ⎠

22
where
M r = resilient modulus,
θ = σ 1 + σ 2 + σ 3 = bulk stress,
σ 1 = major principal stress,
σ 2 = intermediate principal stress,
σ 3 = minor principal stress/ confining pressure,
τ oct = 1 (σ 1 − σ 2 ) 2 + (σ 1 − σ 3 ) 2 + (σ 2 − σ 3 ) 2 ,
3
Pa = normalizing stress (atmospheric pressure) = 101.35 kPa (14.7 psi), and
k1, k2, k3 = material constants.

Physical Property Tests

Soil property tests were also performed on the Shelby tube samples in accordance with the
AASHTO and LADOTD standard test procedures. The tests included: determining moisture-
unit weight (standard Proctor curve), Atterberg limits, hydrometer, sieve analysis, and soil
classification of soils tested. Table 5 presents a summary of the designation of standard tests
that were performed. The in situ dry unit weight ( γ d ) and moisture content (w) of tested soils
are presented in Table 6. Table 7 shows the physical properties of the soils.

LVDTs
Clamps

Load Cell

Figure 10
MTS Triaxial Testing Machine

23
 Table 5
Soil classification test procedures

Test LADOTD AASHTO

Sample Preparation TR 411M/411-95 T87-86
Hydrometer TR 407-89 T88-00
Atterberg Limits TR 428-67 T89-02, T90-00
Moisture/Density Curves TR 418-93 T-99-01
Sieve Analysis TR 113-75 T88-00
Organic Content TR 413-71 T194-97
Moisture Content TR 403-92 T 265

24
 Table 6
Dry unit weights and moisture contents of soil tested

Project Site/Soil Test γd w Project Site/Soil Test γd w

ID Point (pcf) (%) ID Point (pcf) (%)
2 97.0 23.3 2 86.3 31.7
A 5 94.5 25.2 A 5 85.7 32.4
8 102.1 17.8 8 84.4 36.2
LA333 2 96.4 21.7 2 88.8 30.1
B 5 85.7 32.5 B 5 88.8 30.6
8 83.8 34.8 LA347 8 87.0 32.1
2 93.9 25.0 2 80.7 35.9
C 5 90.7 23.0 C 5 79.4 35.9
8 104.6 17.6 8 76.9 36.6
2 93.9 33.6 2 101.4 26.6
A 5 95.1 30.9 A 5 102.1 24.2
8 101.4 21.8 8 102.7 25.3
2 97.7 25.1 2 102.7 25.3
B 5 102.7 24.7 B 5 102.7 26.1
US171 8 99.6 27.3 LA991 8 102.7 25.2
2 114.7 16.9 2 102.7 25.1
C 5 107.1 16.9 C 5 102.1 25.4
8 112.8 15.5 8 102.1 25.6
2 104.0 25.4 2 80.0 32.6
A 5 110.3 21.2 A 5 66.2 47.2
8 103.3 24.3 8 72.5 31.7
2 102.7 24.5 2 76.2 30.9
B 5 107.7 20.5 LA182 B 5 112.8 30.8
LA22 8 104.0 25.3 8 66.2 57.7
2 110.3 19.1 2 78.1 31.7
C 5 104.6 21.8 C 5 94.5 28.3
8 107.7 21.0 8 58.0 58.0
2 95.1 23.2 2 80.0 32.6
A 5 94.5 27.3 A 5 66.3 47.2
8 99.6 24.8 8 72.3 31.7
LA344 2 80.7 30.8 LA652 2 91.0 29.8
B 5 95.8 31.0 B 5 66.0 48.3
8 84.4 32.6 8 63.1 49.8
2 104.6 24.8 2 94.4 28.3
C 5 94.5 27.3 C 5 61.5 54.9
8 87.0 33.0 8 56.8 60.3
2 106.5 22.0
A 5 97.7 21.3
LA28 8 107.1 21.6
2 102.1 21.3 Legend: w - moisture content, γd - dry unit weight
B 5 102.1 21.2
8 102.1 20.7

25
 Table 7
Physical properties of soils tested

Passing Soil Classification

Silt Clay LL PI γdmax wopt
Site #200 Soil
(%) (%) (%) (%) (pcf) (%) USCS AASHTO
(%) Type
LA333 Lean
95 63 32 37 15 105.0 20.0 CL A-6
A clay
LA333 Lean
97 58 39 42 20 101.5 20.5 CL A-7-6
B clay
LA333 Lean
94 55 39 41 15 109.0 18.5 CL A-7-6
C clay
LA347 Heavy
96 53 43 69 38 86.0 28.0 CH A-7-5
A clay
LA347 Heavy
93 62 31 52 27 93.5 20.2 CH A-7-5
B clay
LA347 Heavy
95 58 37 67 37 92.0 24.0 CH A-7-5
C clay
US171 Lean
72 18 54 46 28 114.0 19.0 CL A-7-5
A clay
US171 Lean
84 57 27 46 29 115.0 18.5 CL A-7-5
B clay
US171 119.0 Lean
53 30 23 27 12 12.0 CL A-6
C clay
LA991 Lean
80 72 8 38 13 105.0 22.0 CL A-6
A clay
LA991 Lean
89 59 30 39 16 104.0 20.0 CL A-6
B clay
LA991 Lean
68 24 44 51 26 100.0 21.0 CL A-7-6
C clay
LA22 Lean
80 50 30 40 23 110.0 17.5 CL A-6
A clay
LA22 Lean
82 50 32 43 24 109.0 17.0 CL A-7-6
B clay
LA22 Lean
87 55 32 39 20 109.0 17.0 CL A-6
C clay
LA28 Heavy
76 23 53 62 43 104.3 21.0 CH A-7-6
A clay
LA28 Heavy
95 9 86 98 61 94.2 27.0 CH A-7-5
B clay
LA344 Heavy
93 45 48 57 34 97.7 22.7 CH A-7-6
A clay
LA344 Heavy
95 47 48 52 39 98.5 22.1 CH A-7-6
B clay
LA344 Lean
94 56 38 20 3 101.3 21.8 ML A-4
C silt
LA182 Lean
86 52 34 41 23 105.4 19.1 CL A-7-6
A clay
LA182 Lean
83 47 36 42 23 107.3 17.1 CL A-7-6
B clay
LA182 Lean
93 53 40 46 22 104.3 18.4 CL A-7-6
C clay
LA652 Heavy
95 15 80 99 49 86.4 32.8 CH A-7-5
A clay
LA652 Heavy
96 24 72 91 46 78.5 36.7 CH A-7-5
B clay
LA652 Heavy
97 15 82 87 50 76.0 36.5 CH A-7-5
C clay
Legend: AASHTO- American Association of State Highway and Transportation Officials, LL- Liquid limit,
PI- Plastic index, USCS- Unified soil classification system, wopt-Optimum moisture content, γdmax -Maximum
dry unit weight

26
 DISCUSSION OF RESULTS

The main focus of this study was to develop models that predict the resilient modulus of
subgrade soils from the results of the CIMCPT, DCP, FWD, and Dynaflect test data and
predict the physical properties of soil tested. Prior to the development of models, a field
representative Mr value was defined.

A comprehensive statistical analysis was conducted using the Statistical Analysis System
(SAS) program to develop models that predict the resilient modulus of subgrade soils from
the results of various in situ tests devices considered in this study (CIMCPT, DCP, FWD, and
Dynaflect test). Direct models that only consider the results from the different types of test
devices were developed. In addition, multiple regression models were used to correlate Mr
with the measurements obtained from each DCP and CIMCPT test and to determine the
physical properties of tested soils.

The development of multiple regression models includes several steps. In the first step,
scatter plots between the dependent variable and the independent variables are examined for
possible linear correlations. The significance of the linear correlations between any two
variables is measured using the Pearson product-moment coefficient of correlation (r). If the
value of r is zero or near zero, such indicates that no evidence of an apparent linear
correlation is present. If the value of r is positive or negative one, a perfect linear correlation
does exist. Based on the results of this step, all possible variables that showed good linear
correlation with the dependent variable are examined.

The second step of the development of multiple regression models includes choosing the best
model with least number of dependent variables. Different methods are available in selecting
the best model. In this study, the stepwise selection method was used. This method fits all
possible simple linear models and chooses the best one with the largest F-test statistical
value. Then, all possible two-variable models that include the first variable are compared,
and so on. The significance of each variable included is rechecked at each step along the way
and removed if it falls below the significance threshold.

Based on the results of the variable selection analysis, multiple regression analysis is
conducted on the best model selected. To check for its adequacy, examine the significance
of independent variables, and detect any multicolinearity (possible correlations among the
independent variables ) or heteroscedasticity (unequal error variance) problems. The
adequacy of the model is assessed using the F-test. The probability associated with the F-test
is designated as Pr> F or p-value. A small p-value (less than 0.05) implies that the model is

27
significant in explaining the variation in the dependent variable. The t-test is utilized to
examine the significance of each of the independent variables used in the model. Similar to
that of the F-test, the probability associated with the t-test is designated with a p-value. A p-
value that is less than 0.05 indicates that, at a 95 percent confidence level, the independent
variable is significant in explaining the variation of the dependent variable. The
multicolinearity is detected using the variance inflation factor (VIF). A VIF factor greater
than 10 indicates that weak dependencies may be starting to affect the regression estimates.
Finally, the residual plot is used to check for heteroscedasticity by examining whether the
data has a certain pattern.

A Field Representative Resilient Modulus Value of Subgrade Soils

A field representative stress condition for subgrade soils consisted of a vertical stress level of
41.3 kPa (6 lbf/in.2) that included a cyclic stress level of 37.2 kPa (5.4 lbf/in.2) and a contact
stress level of 4.1 kPa (0.6 lbf/in.2). A confining stress level of 14.0 kPa (2 lbf/in.2) was also
considered. These stress levels were selected based on a stress analysis conducted to compute
a field representative stress condition in the subgrade layer [15,18]. The interpolated Mr was
considered as the laboratory measured Mr from the repeated load triaxial test. This stress
level also corresponds to the “resilient modulus at the break point” proposed by Thompson et
al. [28].

Development of Mr Prediction Models for DCP Test Results

Tables 8 and 9 present the combined DCP and Mr results that were used in developing
regression models that predict the laboratory measured Mr from the DCP test results. The fact
that Table 9 includes DCP test results from a recently completed project at the LTRC is noted
[20]. The ranges of variables used in the regression analysis are presented in Table 10. In
order to determine the independent variables that should be included in the multiple
regression analysis, possible linear correlations between the dependent variable Mr and
DCPI, Log (DCPI), 1/DCPI, dry unit weight (γd), water content (w), and γd/w were first
considered. Figures 11 through 16 present the scatter plots between the dependent variable
and independent variables. The fact that as the Mr decreases the DCPI increases is noted.
Such implies that soil stiffness decreases as the DCPI increases. Therefore, there may be a
good linear correlation between the inverse of DCPI and Mr. Figures 14 and 15 demonstrate
that the laboratory measured Mr increases with the increase in the dry unit weight and the
decrease in the water content. Finally, Figure 16 shows the variation of Mr with the γd/w. The
fact that Mr increases with a decreasing slope as the γd/w increases is noted.
Tables 11 and 12 present the correlation coefficient matrix of all variables for this study. The

28
fact that the best correlation was found between the Mr and 1/DCPI (r = 0.87, p-value
<0.001)is noted. In addition, γd, w, and γd/w were also found to have a significant relation to
Mr. Based on this result, the 1/DCPI, γd, w, and γd/w variables were further used in the
stepwise selection analysis.
Table 8
DCP and laboratory Mr test results (this study)
Project Site/Soil Test Lab. Mr DCPI Project Site/Soil Test Lab. Mr DCPI
ID Point (ksi) (mm/blow) ID Point (ksi) (mm/blow)
2 6.3 18.8 2 9.0 13.7
A 5 4.5 21.5 A 5 12.7 9.9
8 5.8 20.7 8 9.1 12.5
LA333 2 5.7 21.0 2 12.0 11.0
B 5 3.8 24.4 B 5 10.5 12.0
8 2.7 21.6 LA347 8 10.7 11.6
2 3.9 20.0 2 8.1 14.0
C 5 3.3 24.4 C 5 7.6 17.8
8 6.0 18.9 8 8.4 13.9
2 2.2 34.4 2 4.4 27.2
A 5 3.4 30.5 A 5 4.3 27.9
8 3.5 30.8 8 4.4 24.8
2 3.5 30.0 2 4.3 25.9
B 5 7.2 17.2 B 5 4.5 26.0
US171 8 4.5 26.8 LA991 8 4.5 26.0
2 13.3 9.6 2 3.8 22.0
C 5 10.2 12.1 C 5 3.7 26.9
8 9.3 12.9 8 3.5 23.0
2 5.8 20.0 2 3.8 34.1
A 5 5.7 19.0 A 5 3.6 38.0
8 5.6 23.0 8 4.6 28.9
2 5.7 18.0 2 3.8 30.1
B 5 7.8 14.9 LA182 B 5 5.1 23.4
LA22 8 8.6 13.0 8 4.1 36.8
2 5.6 21.0 2 2.8 30.0
C 5 5.9 20.0 C 5 3.4 35.1
8 5.6 23.0 8 2.7 53.3
2 4.4 21.0 2 1.9 53.4
A 5 4.2 24.5 A 5 1.1 65.2
8 4.3 24.5 8 2.6 47.0
2 4.5 18.9 2 3.1 40.0
B 5 4.6 21.4 B 5 2.7 30.0
LA344 8 4.6 31.3 LA652 8 5.6 28.1
2 5.7 18.2 2 1.6 60.0
C 5 5.5 19.3 C 5 2.6 42.3
8 6.0 18.6 8 2.2 46.0
2 4.8 35.3
A 5 4.0 41.0 Legend: DCPI- DCP penetration index, Lab. Mr –
8 4.9 37.0 Laboratory resilient modulus measured at a cyclic
2 12.6 9.0 stress level of 37.2 kPa (5.4 lbf/in.2), contact stress
LA28 B 5 10.3 12.0 level of 4.1 kPa (0.6 lbf/in.2), and confining pressure
8 10.5 13.0 of 14 kPa (2 lbf/in.2)

29
 Table 9
DCP and laboratory Mr test results [20]

Type of γd W Lab. Mr DCPI

Soil ID Location
Material (pcf) (%) (ksi) (mm/blow)
Clay-1 Lab 110.9 11.0 10.4 17.0
Clay-2 Lab 117.8 12.5 12.0 16.7
Clay-3 Lab 104.6 14.6 8.3 23.0
Clay Clay-4 Lab 117.2 13.9 12.1 13.0
Clay-5 Lab 95.8 8.4 9.7 18.4
Clay-6 Lab 106.5 9.4 10.1 15.0
Clay-7 Lab 109.6 13.3 10.2 22.5
Clayey Silt-1 Lab 101.4 19.0 7.0 26.1
Clayey Clayey Silt-2 Lab 100.2 15.4 9.7 18.8
Silt Clayey Silt-3 Lab 100.8 20.1 7.2 27.0
Clayey
Field 104.0 18.5 6.2 29.0
Silt(ALF)
LA-182 Field 100.2 21.1 5.6 36.0
Clay
US-61 Field 100.8 15.6 9.0 10.2
*Clay ALF 4 Field 102.1 23.6 5.3 24.2
Legend: DCPI- DCP penetration index, Lab. Mr – Laboratory resilient modulus measured at a cyclic stress
level of 37.2 kPa (5.4 lbf/in.2), contact stress level of 4.1 kPa (0.6 lbf/in.2), and confining pressure of 14 kPa
(2 lbf/in.2), w - moisture content, γd - dry unit weig

Table 10
Ranges of variables of subgrade materials used in DCP model development

Property Range for Range for Range for Range for

A-4 soils A-6 soils A-7-5 soils A-7-6 soils
No. of samples 6 26 45 15
Mr (ksi) 5-10 4-14 1-14 3-9
DCPI (mm/blow) 19-36 10-28 9-65 13-41
PI (%) 4-6 12-23 27-61 15-43
γd (pcf) 100-104 96-118 57-113 84-108
w (%) 15-24 8-27 21-60 18-35
LL (%) 22-28 27-40 46-98 41-62
Sand (%) 7-58 11-35 4-28 3-32
Silt (%) 28-72 37-72 9-62 23-58
Clay (%) 14-23 8-32 27-86 32-53
Passing sieve #200
42-93 65-89 72-96 68-97
(%)
Legend: Mr – Resilient modulus, DCPI- DCP penetration index, PI- Plasticity index, w- Water content, LL-
Liquid limit, Silt- Percentage of silt, Clay- Percentage of clay, γd- Dry unit weight

30
DCPI (mm/blow

Log DCPI (mm/blow

80 2.0
60 1.5
40
1.0
20
0.5
0
0.0
0 5 10 15
0 5 10 15
Mr (ksi)
Mr (ksi)

Figure 11 Figure 12
Variation of Mr with DCPI Variation of Mr with Log (DCPI)
1/ DCPI (mm/blow

0.15 150
⎯ d (pcf)
0.10 100
0.05 50
0.00 0
0 5 10 15 0 5 10 15
Mr (ksi) Mr (ksi)
Figure 13 Figure 14
Variation of Mr with 1/DCPI Variation of Mr with γd
Water content (%)

80 15
⎯ d /w (pcf/%)

60 10
40
5
20
0
0
0 5 10 15
0 5 10 15
Mr (ksi) Mr (ksi)

Figure 15 Variation of Mr with water content Figure 16 Variation of Mr with γd/w

31
 Table 11
A correlation matrix for the DCP test results (p-value)

γd w
Mr
DCPI γd #200 %Silt %Clay LL PI Log
1/DCPI
/w (DCPI)
γd - <0.001 <0.001 <0.001 <0.001 <0.001 0.32 <0.001 <0.001 <0.001 <0.001 <0.001

w <0.001 - <0.001 <0.001 <0.001 <0.001 0.28 <0.001 <0.001 <0.001 <0.001 <0.001

Mr <0.001 <0.001 - <0.001 <0.001 0.24 0.44 0.009 0.09 0.21 <0.001 <0.001

DCPI <0.001 <0.001 <0.001 - <0.001 0.15 0.98 0.40 0.05 0.004 <0.001 <0.001
γd <0.001 <0.001 <0.001 <0.001 - <0.001 0.81 <0.001 <0.001 <0.001 <0.001 <0.001
/w
-#
<0.001 <0.001 0.24 0.15 <0.001 - 0.006 <0.001 <0.001 <0.001 0.19 0.22
200
%Silt 0.32 0.28 0.44 0.98 0.81 0.006 - <0.001 <0.001 <0.001 0.03 0.38

%Clay <0.001 <0.001 0.009 0.40 <0.001 <0.001 <0.001 - <0.001 <0.001 0.003 0.10

LL <0.001 <0.001 0.09 0.05 <0.001 <0.001 <0.001 <0.001 - <0.001 0.03 0.042

PI <0.001 <0.001 0.21 0.004 <0.001 <0.001 <0.001 <0.001 <0.001 - 0.10 0.68
Log
<0.001 <0.001 <0.001 <0.001 <0.001 0.19 0.03 0.003 0.03 0.10 - <0.001
(DCPI)
1/DCPI <0.001 <0.001 <0.001 <0.001 <0.001 0.22 0.38 0.10 0.42 0.68 <0.001 -

Legend: DCPI- Dynamic cone penetration index, γd- Dry unit weight, w- water content, PI- Plasticity index,
LL- Liquid limit, #200- Percent passing #200 sieve, %Silt- Percentage of silt, and %Clay- Percentage of clay

Table 12
A correlation matrix for the DCP test results (r-value)

γd w
Mr
DCPI γd #200 %Silt %Clay LL PI Log
1/DCPI
/w (DCPI)
γd 1.00 -0.89 0.42 -0.49 0.75 -0.52 0.10 -0.45 -0.49 -0.42 -0.43 0.34

w -0.89 1.00 -0.48 0.50 -0.86 0.49 -0.11 0.44 0.48 0.43 0.45 0.36

Mr 0.42 -0.48 1.00 -0.76 0.56 -0.14 0.08 -0.27 -0.18 -0.13 -0.85 0.87

DCPI -0.49 0.50 -0.76 1.00 -0.42 0.15 -0.004 -0.10 -0.24 0.29 0.96 -0.85
γd 0.75 -0.86 0.56 -0.42 1.00 -0.62 -0.03 -0.40 -0.47 -0.42 -0.39 0.33
/w
-#
-0.52 0.49 -0.14 0.15 -0.62 1.00 0.29 0.40 0.46 0.37 0.14 -0.13
200
%Silt 0.10 -0.11 0.08 -0.004 -0.03 0.29 1.00 -0.76 -0.60 -0.64 -0.22 0.09

%Clay -0.45 0.44 -0.27 -0.10 -0.40 0.40 -0.76 1.00 0.88 0.86 -0.31 -0.17

LL -0.49 0.48 -0.18 -0.24 -0.47 0.46 -0.60 0.88 1.00 0.95 0.23 -0.09

PI -0.42 0.43 -0.13 0.29 -0.42 0.37 -0.64 0.86 0.95 1.00 0.17 -0.04
Log
-0.43 0.45 -0.85 0.96 -0.39 0.14 -0.22 0.31 0.23 0.17 1.00 -0.97
(DCPI)
1/DCPI 0.34 0.36 0.87 -0.85 0.33 -0.13 0.09 -0.17 -0.09 -0.04 -0.97 1.00

Legend: DCPI- Dynamic cone penetration index, γd- Dry unit weight, w- water content, PI- Plasticity index,
LL- Liquid limit, #200- Percent passing #200 sieve, %Silt- Percentage of silt, and %Clay- Percentage of clay

32
Table 13 presents a summary of the results of the analysis. The fact that the best prediction
model should include only 1/DCPI and γd/w variables can be noted. In addition, the 1/DCPI
variable had a much higher partial R-square than the γd/w variable, which suggests that it has
a greater influence on the model prediction. In an effort to demonstrate the effectiveness of
the selection analysis, a multiple regression analysis was conducted on a model that includes
1/DCPI, γd, w, and γd/w as independent variables. Table 14 presents the results of this
analysis. The fact that the 1/DCPI and γd/w are the only significant variables (Pt<0.05); these
are compatible with the results of the variable selection analysis can be noted.

A simple linear regression analysis was conducted in an effort to develop a model that
directly predicts the laboratory measured Mr from the 1/DCPI value. The results of this
analysis yielded the model shown in Equation 6, which will be referred to as the direct
model. The model had a coefficient of determination, R2, value of 0.91 and root square error,
RMSE, value of 0.88 ksi. Figure 17 illustrates the results of regression analysis. The fact that
the proposed model fits the data may be observed. Figure 17 also shows the 95 percent
prediction interval. The 95 percent prediction interval is considered as a measure of the
accuracy of the Mr values predicted using the model developed. The fact that 95 percent of
the data points fall within the boundaries of this interval may be noted.

151.8
Mr = (6)
( DCPI )
1.096

where
Mr = Resilient modulus (ksi), and
DCPI = Dynamic cone penetration index (mm/blow).

In the absence of uniform soil properties along a soil layer, a direct relationship between the
resilient modulus and DCPI is useful. A correlation among resilient modulus, soil properties,
and DCPI may also be useful in examining the effect of soil properties on the DCPI predicted
Mr values. Therefore, a multiple regression analysis was also conducted to develop a model
that predicts laboratory measured Mr from the 1/DCPI and the physical properties of the
tested soils, which will hereafter be referred to as the soil-property model. The independent
variables that were used in the multiple regression analysis were 1/DCPI and γd/w, which
were selected based on the stepwise selection analysis (Table 13). Table 15 shows the results
of the multiple regression analysis. It is noted that both variables (1/DCPI and γd/w ) are
significant at a 95 percent confidence level. In addition, those variables have a VIF value
close to 1, which indicates that these variables are not collinear. Figure 18 presents the

33
residual plot of the DCP- soil property model. There is no distinct pattern among the
residuals; this rules out the model heteroscedasticity.

Table 13
Summary of stepwise selection
Variable Variable Number of Partial Model
Entered Removed Variables In R-Square R-Square F Value Pr > F
1/DCPI 1 0.794 0.794 338.98 <.0001
γd/w 2 0.082 0.876 56.74 <.0001

Table 14
Summary of multiple regression analysis for variable selection
Parameter
Variable t Value Pr > |t|
Estimate
Intercept 0.62 0.27 0.7857
1/ DCPI 220.63 21.30 <.0001
γd 0.024 -1.48 0.1422
w -0.027 0.93 0.3528
γd/w 0.66 6.57 <.0001

20

DCP - direct model

Mr = 151.8 (1 / DCPI1.096)
16
l

R2 = 0.9
ve
Le
n
tio
ic
ed
Pr
%
95
er

12
Measured Mr (ksi)

pp
U

l
ve
Le
n
io
ct
e di
Pr
%
8 95
r
we
Lo

0
0 0.04 0.08 0.12
1/DCPI1.096

Figure 17
Variation of laboratory measured Mr with 1/DCPI

34
 Table 15
Results of Analysis of DCP- Soil Property Model

Parameter Standardized
Variable DF t Value Pr > |t| VIF
Estimate Estimate
1/DCPI1.147 1 165.5 17.56 <.0001 0.77 1.12
⎛ γd ⎞
⎜w⎟ 1 0.0966 6.89 <.0001 0.30 1.12
⎝ ⎠
⎛ 1 ⎞ ⎛ γd ⎞
M r = 165.5 ⎜ ⎟ + 0.0966 ⎜ w ⎟
⎝ DCPI1.147 ⎠ ⎝ ⎠
where,
Mr –Resilient modulus (ksi),
DCPI – Dynamic cone penetration index (mm/blow),
γd –Dry unit weight (pcf), and
w – Water content (%).

5
DCP-Soil Property Model
4

2
Residuals (ksi)

1
Mr (ksi)
0
0 5 10 15
-1

-2

-3

-4

-5
Figure 18
Residuals from DCP-Soil Property Model

35
Figure 19 shows the Mr predicted by the DCP soil property model versus the Mr measured in
the laboratory. The fact that a good agreement was obtained between the predicted and
measured values with (R2=0.92 and RMSE=0.86) may be observed. Furthermore, the model
was able to provide a good prediction of the data obtained from a study reported by George
et al. [11] (Appendix A, Table A1) that was not used in the development of the model.

16

DCP - soil property model

Data used in model development

12 verification
predicted resilient modulus (ksi)

0
0 4 8 12 16
measured resilient modulus (ksi)

Figure 19
Laboratory measured Mr vs. values predicted from DCP-soil property model

Development of Mr Prediction Models for CIMCPT Test Results

A statistical analysis was performed on the CIMCPT and Mr test results shown in Tables 16
and 17 to develop models that predict the Mr from the CIMCPT test results. The models were
developed for fine grained soils using test results of LA333, LA347, US71, LA991, LA22,
LA28, LA344 and data from a previous LTRC project [10]. The CIMCPT and Mr test results
from the field test were used to develop the models. The ranges of variables are presented in
Table 18. The variation of the dependent variable Mr and tip resistance (qc), sleeve friction
(fs), γd, w, γd/w, plasticity index (PI), liquid limit (LL), percent passing #200 sieve (#200),
percentage of silt (%Silt), and percentage of clay (%Clay) are presented in figures 22 through
26.

36
 Table 16
CIMCPT and Laboratory Mr test results for this study (this study)

Project Site Test Lab. qc fs Project Site Test Lab. qc fs

Point Mr (ksi) (ksi) Point Mr (ksi) (ksi)
(ksi) (ksi)
2 6.3 0.7025 0.0022 2 9.0 1.5791 0.0421
A 5 4.5 1.0450 0.0058 A 5 12.7 1.2322 0.0464
8 5.8 0.9289 0.0131 8 9.1 1.3803 0.0377
LA333 2 5.7 0.2525 0.0102 2 12.0 2.6372 0.0058
B 5 3.8 0.2308 0.0029 B 5 10.5 1.0726 0.0639
8 2.7 0.4340 0.0087 LA347 8 10.7 1.4020 0..0276
2 3.9 0.5225 0.0169 2 8.1 1.3208 0.0581
C 5 3.3 0.3324 0.0203 C 5 7.6 1.4804 0.0377
8 6.0 0.6894 0.0305 8 8.4 1.5530 0.0479
2 2.2 0.1829 0.0131 2 4.4 0.0871 0.0087
A 5 3.4 0.2322 0.0087 A 5 4.3 0.0929 0.0087
8 3.5 0.2119 0.0160 8 4.4 0.1248 0.0087
2 3.5 0.2627 0.0160 2 4.3 0.1176 0.0087
B 5 7.2 0.2671 0.0160 B 5 4.5 0.1205 0.0102
US171 8 4.5 0.2656 0.0174 LA991 8 4.5 0.1089 0.0102
2 13.3 1.9013 0.0581 2 3.8 0.1176 0.0116
C 5 10.2 1.4340 0.0377 C 5 3.7 0.0987 0.0102
8 9.3 1.2627 0.0348 8 3.5 0.1350 0.0131
2 5.8 0.4296 0.0189 2 4.4 0.3512 0.0290
A 5 5.7 0.6168 0.0145 A 5 4.2 0.3672 0.0290
8 5.6 0.7983 0.0203 8 4.3 0.3643 0.0174
2 5.7 0.8520 0.0247 2 4.5 0.3614 0.0363
B 5 7.8 1.0015 0.0348 B 5 4.6 0.4165 0.0203
LA22 8 8.6 1.3716 0.0435 LA344 8 4.6 0.6430 0.0261
2 5.6 0.4296 0.0189 2 5.7 0.2743 0.0392
C 5 5.9 1.0552 0.0160 C 5 5.5 0.8665 0.0290
8 5.6 0.7663 0.0174 8 6.0 1.1248 0.0044
2 4.8 0.5065 0.0174 Legend: fs- Sleeve friction, Lab. Mr – Laboratory
A 5 4.0 0.4049 0.0116 resilient modulus measured at a cyclic stress level
8 4.9 0.4383 0.0145 of 37.2 kPa (5.4 lbf/in.2), contact stress level of 4.1
2 12.6 1.3077 0.0305 kPa (0.6 lbf/in.2), and confining pressure of 14 kPa
LA28 B 5 10.3 1.3077 0.0305 (2 lbf/in.2), qc- Tip resistance
8 10.5 1.7605 0.0421

37
 Table 17
CIMCPT and Laboratory Mr test results [10]

Site Soil ID γd W Lab. Mr qc fs

(pcf) (%) (ksi) (ksi) (ksi)
1 100.2 25.4 4.0 0.3628 0.0096
PRF-silty 2 104.0 23.0 4.3 0.4644 0.0104
clay 3 105.9 20.8 4.4 0.3904 0.0131
4 106.5 23.2 4.5 0.4093 0.0106
5 107.1 21.5 5.5 0.4572 0.0134
1 62.4 61.6 0.6 0.0406 0.0027
PRF-heavy 2 62.4 65.1 0.6 0.0450 0.0029
clay 3 64.3 60.4 0.8 0.0464 0.0033
4 63.0 62.5 1.5 0.0581 0.0033
5 64.3 59.0 0.9 0.0566 0.0027
6 64.9 59.5 1.4 0.0552 0.0026
1 106.5 21.5 4.2 0.3019 0.0151
I-10/ 2 108.4 19.6 3.4 0.2729 0.0163
LA-42 clay 3 104.0 23.0 1.9 0.1640 0.0081
4 102.7 21.4 2.9 0.2917 0.0173
5 105.9 20.8 3.4 0.2642 0.0137
6 103.3 22.5 1.8 0.1800 0.0090
1 109.0 24.1 6.9 0.4136 0.0219
2 102.1 23.0 4.7 0.3019 0.0166
LA-15 clay 3 105.9 28.4 6.5 0.3004 0.0179
4 96.4 27.3 5.2 0.3106 0.0140
5 112.2 18.8 8.8 0.4456 0.0195
6 96.8 31.4 3.6 0.2975 0.0159
1 114.0 24.9 4.8 0.2525 0.0144
LA-89 clay 2 101.4 26.8 2.3 0.1974 0.0156
3 100.2 28.6 1.4 0.0726 0.0090
4 107.7 24.6 2.8 0.2598 0.0151
1 115.3 9.5 8.5 0.4499 0.0180
Siegen 2 107.7 22.5 3.9 0.1916 0.0226
Lane clay 3 107.7 16.7 10.3 0.4877 0.0165
4 97.0 23.1 3.6 0.2337 0.0152
Legend: fs- Sleeve friction, Lab. Mr – Laboratory resilient modulus measured at a cyclic stress level of 37.2 kPa
(5.4 lbf/in.2), contact stress level of 4.1 kPa (0.6 lbf/in.2), and confining pressure of 14 kPa (2 lbf/in.2), qc- Tip
resistance, w - moisture content, γd - dry unit weight

38
 Table 18
Ranges of variables of subgrade materials used in CIMCPT model development

Property Range for A-4 Range for A-6 Range for A-7-5 Range for A-7-6
soils soils soils soils

No. of samples 8 26 18 39
Lab. Mr (ksi) 6-8 2-14 2-14 1-11
qc (ksi) 0.4-0.5 0.1-1.9 0.2-2.6 0.04-1.4
fs (ksi) 0.0096-0.0134 0.0022-0.0581 0.0058-0.0639 0.0026-0.0435
PI (%) <6 11-23 27-61 15-66
γd (pcf) 100-107 94-115 77-103 62-112
w (%) 21-25 9-29 21-37 18-65
LL (%) 28 27-40 46-98 41-93
Sand (%) 7 11-35 4-28 2-32
Silt (%) 70 30-72 9-62 14-58
Clay (%) 23 8-32 27-86 32-84
Passing sieve
93 65-89 72-96 68-98
#200 (%)
Legend: Lab. Mr – Laboratory resilient modulus measured at a cyclic stress level of 37.2 kPa (5.4 lbf/in.2),
contact stress level of 4.1 kPa (0.6 lbf/in.2), and confining pressure of 14 kPa (2 lbf/in.2), PI- Plasticity index, w-
Water content, LL- Liquid limit, Silt- Percentage of silt, Clay- Percentage of clay, γd- Dry unit weight, qc - Tip
resistance, fs - Sleeve friction

Figures 20 and 21 show the variation of Mr with the tip resistance and sleeve friction,
respectively. As the tip resistance and sleeve friction increase, the resilient modulus of
subgrade soils increases. This implies that soil stiffness increases as the tip resistance and
sleeve friction increase. Furthermore, this also indicates that there may be a good correlation
between Mr and both the tip resistance and sleeve friction. Figure 22 shows the variation of
Mr with the γd/w. As the γd/w increases, the Mr increases with a decreasing slope. Therefore,
there may be a correlation between the γd/w and Mr.

The correlation coefficient matrix of different variables is presented in Tables 19 and 20. A
good linear correlation between Mr and (qc) tip resistance and Mr and (fs) sleeve friction is
observed with r = 0.82 and r = 0.70, respectively. Such is expected, as the cone’s tip
resistance and sleeve friction measure the shear strength and frictional resistance of soils,
respectively, both of which are known to significantly affect the soil stiffness.

39
 3
0.1
2 0.075

fs (ksi)
qc (ksi)

0.05
1
0.025
0 0
0 5 10 15 0 5 10 15
Mr (ksi) Mr (ksi)

Figure 20 Figure 21
Variation of Mr with qc Variation of Mr with fs

15 15.0
M r (ksi) 10.0
10
M r (k s i)

5 5.0

0 0.0
0 2 4 6 8 10 12 14 50 75 100 125
γd/w (pcf/%) γd (pcf)

Figure 22 Figure 23
Variation of Mr with γd/w Variation of Mr with γd

15.0
M r (ksi)

10.0

5.0

0.0
0 10 20 30 40
w (%)

Figure 24
Variation of Mr with w

40
 Table 19
A correlation matrix for the CIMCPT test results (p-value)

γd w Mr qc fs γd/w #200 %Silt %Clay LL PI

γd - <0.0001 0.01 0.95 0.42 <0.0001 <0.0001 0.0013 <0.0001 <0.0001 <0.0001
w <0.0001 - 0.0001 0.11 0.04 <0.0001 0.0016 <0.0001 <0.0001 <0.0001 <0.0001
Mr 0.01 0.0001 - <0.0001 <0.0001 0.001 0.26 0.39 0.12 0.92 0.38
qc 0.95 0.11 <0.0001 - <0.0001 0.46 0.38 0.91 0.53 0.48 0.89
fs 0.42 0.04 <0.0001 <0.0001 - 0.26 0.23 0.93 0.42 0.64 0.91
γd/w <0.0001 <0.0001 0.001 0.46 0.26 - 0.002 0.05 <0.0001 <0.0001 <0.0001
-#200 <0.0001 0.0016 0.26 0.38 0.23 0.002 - 0.004 0.008 0.006 0.06
%Silt 0.0013 <0.0001 0.39 0.91 0.93 0.05 0.004 - <0.0001 <0.0001 <0.0001
%Clay <0.0001 <0.0001 0.12 0.53 0.42 <0.0001 0.008 <0.0001 - <0.0001 <0.0001
LL <0.0001 <0.0001 0.92 0.48 0.64 <0.0001 0.006 <0.0001 <0.0001 - <0.0001
PI <0.0001 <0.0001 0.38 0.89 0.91 <0.0001 0.06 <0.0001 <0.0001 <0.0001 -
Legend: qc- Tip resistance, fs- Sleeve friction, γd- Dry unit weight, w- water content, PI- Plasticity index, LL-
Liquid limit, #200- Percent passing #200 sieve, %Silt- Percentage of silt, and %Clay- Percentage of clay
Table 20
A correlation matrix for the CIMCPT test results (r-value)

γd w Mr qc fs γd/w #200 %Silt %Clay LL PI

γd 1.00 -0.93 0.27 0.007 0.09 0.83 -0.40 0.33 -0.57 -0.63 -0.62

w -0.92 1.00 -0.39 -0.17 -0.22 -0.83 0.33 -0.40 0.60 0.63 0.63

Mr 0.27 -0.39 1.00 0.82 0.70 0.33 -0.12 0.09 -0.16 -0.01 -0.09

qc 0.007 -0.17 0.82 1.00 0.63 0.08 -0.09 0.01 -0.07 0.07 0.01
fs 0.09 -0.22 0.70 0.63 1.00 0.12 -0.13 0.01 -0.09 0.05 0.01
γd/w 0.83 -0.83 0.33 0.08 0.12 1.00 -0.32 0.21 -0.40 -0.48 -0.47
-#200 -0.40 0.33 -0.12 -0.09 -0.13 -0.32 1.00 0.31 0.28 0.29 0.20
%Silt 0.33 -0.40 0.09 0.01 0.01 0.21 0.31 1.00 -0.83 -0.69 -0.75
%Clay -0.57 0.60 -0.16 -0.07 -0.09 -0.40 0.28 -0.83 1.00 0.87 0.88
LL -0.63 0.63 -0.01 0.07 0.05 -0.48 0.29 -0.69 0.87 1.00 0.97

PI -0.62 0.63 -0.09 0.01 0.01 -0.47 0.20 -0.75 0.88 0.97 1.00

Legend: qc- Tip resistance, fs- Sleeve friction, γd- Dry unit weight, w- water content, PI- Plasticity index, LL-
Liquid limit, #200- Percent passing #200 sieve, %Silt- Percentage of silt, and %Clay- Percentage of clay

Tables 19 and 20 also show that qc, fs, γd, w, and γd/w are the only variables that have a
significant relation to Mr, and hence, they should be included in the variable stepwise
selection analysis. Table 21 presents a summary of the results of the stepwise selection
analysis. The fact that the best model includes qc, fs, and γd/w can be noted. The fs variable
had the greatest influence on the prediction of the model, as is indicated by the partial R2.

Regression analyses were conducted on the CIMCPT-Mr data to develop two models. The
first model, the direct model, relates the laboratory measured Mr directly to the fs and qc,
while the second model, the soil-property model, predicts laboratory measured Mr from fs, qc,
and the physical properties of the tested soil. The results of the first regression analysis
yielded the direct model shown in Equation 7. The direct model had R2 and RMSE values of

41
0.77 and 1.34, respectively. Figure 25 shows the variation of Mr predicted by the direct
model and the Mr measured in the laboratory. The results indicate that the model was
effective in predicting the Mr of subgrade soils from the results of the CIMCPT.

M r =2.12+ 3.44q c +63.15f s (7)

where
Mr = resilient modulus (ksi),
qc = tip resistance (ksi), and
fs=sleeve friction (ksi)

Table 22 presents the results of regression analyses that were conducted to develop the soil-
property model. The results show that the model had R2 of 0.86 and an RMSE of 0.96.
Furthermore, qc and γd/w had a more significant effect on the prediction of the model than fs,
as is indicated by the t-value. In addition, all three variables have VIF values less than five,
which indicates that these variables are not collinear. To test for any possible
heteroscedasticity of the CIMCPT soil-property model, the residuals are plotted against the
resilient modulus value as shown in Figure 26. The figure illustrates very little evidence of
heteroscedasticity in the model.

Figure 27 shows Mr predicted by the CIMCPT soil-property model and those measured in the
laboratory. It is observed that the model predicted Mr values were comparable with Mr
measured values. Such indicates that the model was effective in predicting the Mr values for
the soil tested.

Typical variation of tip resistance, sleeve friction, and predicted Mr with depth is presented in
Appendix A, Figure A1.
Table 21
Results of the Variable selection for CIMCPT-Mr model
Variable Variable Number Partial Model
Entered Removed Vars In R-Square R-Square C(p) F Value Pr > F
qc 1 0.6745 0.675 47.4290 184.44 <.0001
γd/w 2 0.0760 0.751 18.0526 26.79 <.0001
fs 3 0.0412 0.792 3.0173 17.23 <.0001

42
 Table 22
Results of the Multiple Regression for CIMCPT-Mr model
Parameter Standardized Variance
Variable DF Estimate t-value Pr > |t| Estimate Inflation
fs 1 3.547 13.19 <.0001 0.47 3.52
qc 1 52.886 5.15 <.0001 0.21 4.74
γd / w 1 0.517 12.33 <.0001 0.38 2.74
γd
M r = 3.55q c + 52.88f s + 0.52( )
w
where,
Mr –Resilient modulus (ksi),
qc –Tip resistance (ksi),
fs – Sleeve friction (ksi),
γd –Dry unit weight (pcf), and
w – Water content (%).

16

CIMCPT - direct model

12
predicted resilient modulus (ksi)

0
0 4 8 12 16
measured resilient modulus (ksi)

Figure 25
Predictions from the CIMCPT-Direct Model

43
 5

4 CIMCPT-Soil Property Model

Residuals (ksi) 1 Mr (ksi)

0
0 5 10 15
-1

-2

-3

-4

-5

Figure 26
Residuals from CIMCPT-Soil Property Model
16

CIMCPT - soil property model

12
predicted resilient modulus (ksi)

0
0 4 8 12 16
measured resilient modulus (ksi)

Figure 27
Predictions from the CIMCPT-Soil Property Model

44
Development of Mr Prediction Models for FWD Test Results
Three backcalculation software packages were used to interpret the FWD data, namely,
ELMOD 5.1.69, MODULUS 6, and EVERCALC 5.0. The Florida equation was also used
for comparison. During the testing process, there were three readings taken at a load of 9,000
lbs. The results used in the statistical analysis reflect the averages of the three readings. Since
MODULUS 6 only uses readings from seven sensors and the data were collected with nine
sensors, the files were modified to accommodate the MODULUS 6 software.

Results of ELMOD 5.1.69 Software Backcalculation

Linear backcalculation models were used in this study. Seed values refer to modulus input
values for layers prior to the beginning of the backcalculation process. The seed values used
for this study were taken from a previous study [29].

Four types of linear backcalculation models were used to backcalculate the FWD moduli.
The first two models used seven and nine sensors with no seed values. The third model used
nine sensors by inputting seed values in the backcalculation process. Finally, the fourth
model used was the one recommended by Dynatest Consulting, Inc. Further information on
the models used can be found in the ELMOD 5.1.69 manual. The fact that, in all
backcalculation analyses, the maximum depth of the rigid layer was fixed at 240 inches is
worth noting. The results of the FWD moduli backcalculation analyses using the four models
considered are presented in Table 23.

Linear regression analyses were conducted to develop models that predict the laboratory
measured Mr from the FWD moduli that were backcalculated using the previously mentioned
analyses. The results of the regression analyses yielded the models shown in Table 24.
Figures 28 through 31 illustrate the prediction of these models. The fact that among the four
backcalculation models evaluated in this study, models without seed values had better
correlation (R2=0.71 and RMSE=1.32ksi), while the model recommended by the Dynatest
had lower R2 value of 0.61 and higher RMSE value of 1.53 ksi, is noted.

Results of MODULUS 6 Backcalculation

MODULUS 6 backcalculation analyses were performed using semi-infinite and finite depth
to bedrock models. For the finite depth to bedrock model, the software provides a ratio called
E4/stiff layer to account for the stiffness relationship between the subgrade and bedrock
layers. In most cases, the software recommends the use of 100 for the E4/stiff layer ratio;
however, for a stiff subgrade layer, a value of five or less should be considered. Therefore,
three

45
 Table 23
Results of FWD Backcalculation Using ELMOD Software

No No No No
seed Cal=2 seed Cal=2
Site/ seed seed Site/ seed seed
Test 9- 9- Test 9- 9-
Project Soil 7- 9- Project Soil 7- 9-
Point sensor sensor Point sensor sensor
ID sensor sensor ID sensor sensor
Backcalculated Mr (ksi) Backcalculated Mr (ksi)
2 14.8 14.8 14.7 12.1 2 15.1 14.9 14.8 12.1
5 14.1 14.1 13.9 11.4 5 14.9 14.7 14.4 11.9
A A
8 14.2 14.2 14.0 11.4 8 14.6 14.7 14.6 11.9
2 10.4 8.5 10.1 8.4 2 16.0 16.5 15.7 13.3
5 11.6 9.0 11.1 8.9 5 15.0 14.9 14.8 12.3
B B
8 12.2 10.6 13.0 10.3 8 14.9 15.0 14.8 12.2
LA333
2 11.9 11.9 11.9 9.8 LA347 2 14.9 15.0 14.6 12.1
5 12.3 12.6 12.8 10.4 5 15.0 15.2 15.0 12.3
C C
8 11.2 11.2 11.3 9.0 8 15.6 15.5 15.4 12.8
2 11.9 11.9 12.1 9.3 2 9.7 9.4 9.6 7.6
5 11.2 11.7 11.5 8.9 5 8.6 8.5 8.6 6.7
A A
8 11.5 11.6 11.6 8.9 8 7.8 7.8 7.8 5.9
ksi
2 12.1 11.8 11.8 9.2 2 7.8 7.7 7.6 6.2
5 12.8 12.8 12.7 10.1 5 7.9 7.9 7.8 6.5
B B
8 12.7 12.8 12.8 10.0 8 9.4 9.4 9.3 7.6
US171 2 24.1 23.9 23.5 18.0 LA991 2 9.8 9.8 9.3 8.0
5 24.7 25.3 24.4 18.7 5 9.7 9.7 9.7 7.9
C C
8 27.2 27.6 26.3 20.3 8 10.4 10.5 10.1 8.4
2 15.0 14.7 14.6 15.7 2 6.9 7.0 6.9 5.4
5 15.4 15.6 15.4 16.4 5 7.2 7.3 7.3 5.7
A A
8 14.4 14.7 14.7 15.3 8 7.8 8.0 7.9 6.3
2 16.5 16.5 16.5 17.5 2 7.7 8.0 8.0 6.7
5 18.4 18.8 18.7 19.4 5 7.4 7.5 7.3 6.3
B B
8 17.8 17.6 17.7 18.6 8 7.8 7.8 8.0 6.5
2 14.8 14.8 14.8 15.6 LA182 2 8.4 8.7 9.2 7.3
LA22
5 14.6 14.7 14.5 15.5 5 8.5 8.5 9.0 7.1
C C
8 16.2 16.2 16.3 16.9 8 8.4 8.7 8.8 7.0
2 8.6 8.7 8.8 7.1 2 4.2 4.1 4.0 3.5
5 8.9 8.9 9.0 7.4 5 4.2 4.1 4.2 3.5
A A
8 9.1 9.1 9.4 7.5 8 4.5 4.5 4.5 3.8
2 10.2 10.2 10.2 8.5 2 6.7 6.7 6.8 5.5
5 6.7 6.6 6.4 5.5 5 4.9 4.9 5.0 4.1
B B
8 6.0 5.9 5.8 4.7 8 4.2 4.2 4.2 3.5
LA344 2 10.7 10.8 10.8 8.7 LA652 2 4.6 4.6 4.5 3.5
5 11.4 11.3 11.5 9.3 5 4.8 4.8 4.8 3.9
C C
8 11.0 11.1 11.5 9.2 8 4.6 4.5 4.5 3.7
2 14.9 15.6 16.3 12.9 Legend: Cal- Calibration, Mr –Resilient modulus
5 13.7 14.0 13.8 11.2
A
8 12.6 12.9 13.1 10.6
2 25.0 26.1 26.1 20.8
B 5 25.1 26.2 26.2 20.8
LA28
8 26.2 27.1 27.0 21.9

46
 Table 24
Results of statistical analysis for Mr-FWD (ELMOD 5.1.69) model

ELMOD 5.1.69 Model R2 RMSE

7-sensor (no seed) M r = 0.40 E fwd + 0.49 0.71 1.32
9-sensor (no seed) M r = 0.39 E fwd + 0.64 0.70 1.32
9-sensor (seed) M r = 0.39 E fwd + 0.61 0.69 1.36
9-sensor (Cal=2) M r = 0.40 E fwd + 1.13 0.61 1.53
2
Legend: Efwd- Backcalculated modulus from FWD (ksi), Mr- Resilient modulus (ksi), R - Coefficient of
determination, RMSE- Root mean square for error

16

ELMOD 5.1.69 (7 sensor - no seed)

Mr= 0.40 Efwd + 0.49
12 R2 = 0.70 l
ve
Le
ion
ict
P red
measured resilient modulus (ksi)

5%
r9
pe
Up
8
l
ve
Le
ion
d ict
r e
P
5%
4 r9
we
Lo

-4
0 5 10 15 20 25 30
modulus from FWD (ksi)

Figure 28
Mr versus FWD modulus backcalculated ELMOD 5.1.69 (7-sensor with no seed values)

47
 16

ELMOD 5.1.69 (9 sensor no seed)

Mr = 0.39 Efwd + 0.64
l
R2 = 0.7 ve
Le
n
12 ti o
e dic
Pr

measured resilient modulus (ksi)

5%
r9
pe
Up

el
L ev
t ion
ic
ed
Pr
%
95
w er
4 Lo

0
0 5 10 15 20 25 30
EFWD (ksi)

Figure 29
Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (9-sensor with no seed values)
16

ELMOD 5.1.69 (9 sensor seed)

Mr = 0.39 Efwd + 0.61
l
R2 = 0.69 ve
Le
n
12 ti o
e dic
Pr
measured resilient modulus (ksi)

5%
r9
pe
Up

el
L ev
t ion
ic
ed
Pr
5%
er9
w
4 Lo

0
0 5 10 15 20 25 30
EFWD (ksi)

Figure 30
Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (9-sensor with seed values)

48
 16

ELMOD 5.1.69 (Cal=2)

Mr = 0.4 Efwd + 1.31
R2 = 0.61
el
ev
12 nL
ic tio
ed
Pr
measured resilient modulus (ksi)
5%
r9
pe
Up

el
ev
nL
ic t io
ed
Pr
%
4
r 95
we
Lo

0
0 4 8 12 16 20 24
EFWD (ksi)

Figure 31
Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 (Calibration = 2)

MODULUS 6 backcalculation analyses were conducted using finite depth to bedrock models
for E4/stiff layer ratio values of 100, 5, and 3. Based on the results of these analyses, the
FWD backcalculated moduli values closest to the laboratory measured Mr were selected.

Regression analyses were conducted to correlate the laboratory measured Mr from the FWD
moduli backcalculated using the semi-infinite and finite depth analyses shown in Table 25.
Based on the results of the regression analyses, the models shown in Table 26 were
developed. Figures 32 and 33 illustrate the two models, respectively. The fact that the
regression model developed using the semi-infinite analysis was better than that developed
using the finite depth analyses that were obtained when using the FWD moduli
backcalculated from an analysis that did not utilize seed values is noted. However, both
models had a relatively low R2 (0.46 and 0.54) and high RMSE value (1.7 ksi and 1 ksi),
which indicates that a poor correlation exists between the Mr and the FWD moduli
backcalculated using MODULUS 6 software. Such is also observed in Figures 34 and 35,
where data points were widely scattered about the model line.

49
 Table 25
Results of FWD backcalculation analysis using MODULUS 6 software

Project Site/Soil Test Semi-infinite Finite Project Site/Soil Test Semi- Finite
ID Point Depth ID Point infinite Depth
Backcalculated Mr (ksi) Backcalculated Mr (ksi)
2 18.4 11.5 2 17.3 8.3
A 5 17.5 10.8 A 5 17.5 11.2
8 17.0 10.9 8 17.1 10
LA333 2 13.3 6.6 2 18.1 12.3
B 5 15.3 7.1 B 5 17.0 9.8
8 16.6 8 LA347 8 17.4 11.2
2 13.9 9.8 2 17.7 9.8
C 5 15.1 10.4 C 5 17.7 9.8
8 14.4 9 8 16.9 10.3
2 14.8 7.8 2 12.9 7.6
A 5 13.5 7.1 A 5 12.5 6.7
8 13.8 7.3 8 12.5 5.9
2 13.7 7.9 2 9.1 6.2
B 5 16.4 8.1 B 5 9.5 6.5
US171 8 17.0 8.4 LA991 8 11.3 7.6
2 28.0 14 2 11.5 8
C 5 29.4 14.9 C 5 11.7 7.9
8 31.4 15.9 8 12.3 8.4
2 26.1 15.7 2 11.9 5.4
A 5 27.9 16.4 A 5 12.1 5.7
8 28.3 15.3 8 11.4 6.3
2 27.4 17.5 2 9.4 6.7
B 5 27.3 19.4 LA182 B 5 8.2 6.3
LA22 8 25.9 18.6 8 10.6 6.5
2 24.1 15.6 2 10.7 7.3
C 5 24.6 15.5 C 5 10.9 7.1
8 24.6 16.9 8 11.3 7
2 12.5 6 2 9.2 3.5
A 5 12.0 6.2 A 5 10.0 3.5
8 14.0 6.4 8 10.6 3.8
2 12.8 6.3 2 16.4 5.5
B 5 12.8 8.5 B 5 12.1 4.1
LA344 8 11.0 5.5 LA652 8 11.1 3.5
2 14.8 8.2 2 8.4 3.5
C 5 15.8 8.7 C 5 7.5 3.9
8 14.9 8.6 8 7.3 3.7
2 15.9 11.7
A 5 14.4 11.1 Legend: Mr –Resilient modulus, SL- Stiff layer
8 13.5 10.3
2 26.2 18.5
LA28 B 5 26.6 18.8
8 27.5 19.6

50
 Table 26
Results of statistical analysis for Mr-FWD (MODULUS 6) model

MODULUS 6 Model R2 RMSE

Semi Infinite M r = 0.25E fwd + 1.02 0.54 1.38
Finite Depth M r = 0.40E fwd + 0.90 0.46 1.7
2
Efwd- Backcalculated modulus from FWD (ksi), Mr- Resilient modulus (ksi), R - Coefficient of
determination, RMSE- Root mean square for error (ksi)

16

MODULUS 6 (Semi-infinite)
Mr =0.25Efwd + 1.02 l
eve
nL
12 R2 = 0.54 ictio
P re d
er 95%
Upp
measured resilient modulus (ksi)

el
4 Lev
ion
dict
% Pre
e r 95
Low

-4
5 10 15 20 25 30 35
EFWD (ksi)

Figure 32
Mr vs. FWD moduli backcalculated using MODULUS 6 (semi infinite subgrade layer)

Results of EVERCALC 5.0

Table 27 shows the results of the FWD moduli backcalculation using EVERCALC 5.0
software. Regression analysis was performed on the Mr and the FWD moduli backcalculated
using EVERCALC 5.0 software. The results of this analysis yielded the model shown in
Equation 19. The model had R2 and RMSE values of 0.51 and 1.62, respectively. Figure 36
presents the results from the statistical analysis. The fact that poor correlation exists between
the FWD moduli backcalculated using EVERCALC 5.0 and the Mr measured in the
laboratory is noted.

M r = 0.26Efwd + 1.19 (8)

where
Mr = resilient modulus (ksi), Efwd= backcalculated modulus from FWD (ksi).

51
 16

MODULUS 6 (Finite Depth)

Mr = 0.4 Efwd + 0.9
R2 = 0.61
12 el
n L ev
ctio
ed i

measured resilient modulus (ksi)

Pr
95%
p er
Up

l
L ev e
4 ion
ict
P r ed
5%
er 9
Upp

0
4 8 12 16 20
EFWD (ksi)

Figure 33
Mr vs. FWD moduli backcalculated using MODULUS 6 (finite depth)

Results of Florida Equation

The FWD moduli backcalculated using the Florida equation is shown in Table 27. Equation 9
presents the correlation between the FWD moduli backcalculated using the Florida equation
and Mr measured in the laboratory. While Figure 35 illustrates this correlation, The fact that
the correlation is poor and has a low R2 value of 0.49 is noted.

M r = 0.24Efwd + 0.94 (9)

where
Mr = Resilient modulus (ksi),
Efwd= Backcalculated modulus from FWD (ksi).

Development of Mr Prediction Models for Dynaflect Test Results

LADOTD developed a chart to determine the subgrade modulus and structural number based
upon deflection readings taken with the Dynaflect. This chart was used to obtain the
subgrade moduli Ed from the Dynaflect test results (Table 28). Equation 10 and Figure 36
present the result of the regression analysis that was conducted to correlate the
backcalculation results with the laboratory measured Mr. The correlation had an R2 value of
0.73 and an RMSE value of 1.46.
M r = 0.41Ed + 2.26 (10)

52
where
Mr = Resilient modulus (ksi),
Efwd= Backcalculated modulus from FWD (ksi).

Table 27
Results of FWD backcalculation using EVERCALC 5.0 and Florida equation

Project Site ID Lab Ever Fl Project Site ID Lab Ever Fl

Mr (ksi) Mr (ksi)
2 6.3 18.3 19.4 2 9.0 15.8 19.0
A 5 4.5 17.4 18.8 A 5 12.7 15.8 18.3
8 5.8 16.9 18.3 8 9.1 16.2 18.2
LA333 2 5.7 13.0 14.2 2 12.0 17.2 19.4
B 5 3.8 15.2 16.8 B 5 10.5 16.3 18.7
8 2.7 16.3 18.4 LA347 8 10.7 16.6 18.6
2 3.9 13.6 15.2 2 8.1 17.1 18.4
C 5 3.3 14.9 16.5 C 5 7.6 16.7 18.4
8 6.0 14.0 15.9 8 8.4 16.5 19.5
2 2.2 14.5 16.9 2 4.4 12.5 14.5
A 5 3.4 13.4 15.4 A 5 4.3 12.3 15.7
8 3.5 13.4 15.5 8 4.4 12.7 16.5
2 3.5 13.5 14.9 2 4.3 8.9 10.5
B 5 7.2 15.4 17.0 B 5 4.5 9.4 11.1
US171 8 4.5 15.6 17.1 LA991 8 4.5 11.0 13.4
2 13.3 27.0 29.8 2 3.8 11.0 13.5
C 5 10.2 28.8 32.5 C 5 3.7 11.5 14.0
8 9.3 29.9 33.7 8 3.5 11.7 14.5
2 5.8 26.2 27.8 2 3.8 13.0 15.0
A 5 5.7 27.9 30.2 A 5 3.6 12.7 15.1
8 5.6 28.6 32.5 8 4.6 11.3 13.6
2 5.7 27.8 28.6 2 3.8 9.1 10.3
B 5 7.8 26.9 27.6 LA182 B 5 5.1 8.0 9.4
LA22 8 8.6 25.5 25.9 8 4.1 10.1 12.0
2 5.6 24.6 24.4 2 2.8 10.3 12.1
C 5 5.9 25.2 25.3 C 5 3.4 10.6 12.5
8 5.6 24.8 25.8 8 2.7 11.0 13.1
2 4.4 13.0 14.4 2 1.9 5.1 6.3
A 5 4.2 12.4 13.4 A 5 1.1 5.7 7.2
8 4.3 14.1 16.5 8 2.6 6.1 7.5
2 4.5 12.0 13.1 2 3.1 9.8 11.5
B 5 4.6 9.7 8.5 LA652 B 5 2.7 7.1 8.7
LA344 8 4.6 8.0 7.9 8 5.6 6.6 8.6
2 5.7 14.9 16.2 2 1.6 8.1 11.6
C 5 5.5 15.6 17.7 C 5 2.6 7.5 9.5
8 6.0 14.9 16.4 8 2.2 7.4 9.5
2 4.8 15.6 17.3 Legend: Elm- Ever- EVERCALC, Fl- Florida
A 5 4.0 14.4 16.1 equation, Lab- Laboratory,
8 4.9 13.6 15.0
2 12.6 25.0 27.7
LA28 B 5 10.3 25.0 28.0
8 10.5 25.0 29.5

53
 16

Evercalc 5.0
Mr = 0.26 Efwd + 1.19
12 R2 = 0.51 el
Lev
ion
dict
Pre
er 95%
Upp

measured resilient modulus (ksi) 8

4 v el
n Le
ictio
Pred
5%
er 9
L ow

-4
5 10 15 20 25 30
modulus from FWD (ksi)

Figure 34
Mr vs. FWD moduli backcalculated using EVERCALC 5.0
16

Florida equation e l
L ev
Mr = 0.24 Efwd + 0.94
ion
R2 = 0.49 ic t
12 ed
Pr
5%
r9
pe
Up
meaured resilient modulus (ksi)

l
L eve
ti on
di c
4 Pre
5%
er 9
Low

-4
0 10 20 30 40
modulus from FWD (ksi)
Figure 35
Mr vs. FWD moduli backcalculated using ELMOD 5.1.69 Florida equation

54
 Table 28
Dynaflect test results

Project Site/Soil Test Lab. Mr Dynaflect Project Site/Soil Test Lab-Mr Dynaflect
ID Point (ksi) moduli ID Point (ksi) moduli
(ksi) (ksi)
2 6.3 8.2 2 9.0 19.0
A 5 4.5 7.7 A 5 12.7 18.3
8 5.8 7.9 8 9.1 18.2
LA333 2 5.7 7.1 2 12.0 19.4
B 5 3.8 7.8 B 5 10.5 18.7
8 2.7 8.7 LA347 8 10.7 18.6
2 3.9 5.8 2 8.1 18.4
C 5 3.3 5.9 C 5 7.6 18.4
8 6.0 5.6 8 8.4 19.5
2 2.2 7.0 2 4.4 4.2
A 5 3.4 6.5 A 5 4.3 4.1
8 3.5 6.5 8 4.4 4.2
2 3.5 6.7 2 4.3 3.5
B 5 7.2 7.6 B 5 4.5 3.7
US171 8 4.5 7.5 LA991 8 4.5 4.0
2 13.3 16.7 2 3.8 3.8
C 5 10.2 15.8 C 5 3.7 3.7
8 9.3 14.7 8 3.5 3.8
2 5.8 6.9 2 3.8 4.2
A 5 5.7 7.0 A 5 3.6 4.3
8 5.6 7.3 8 4.6 4.1
2 5.7 8.0 2 3.8 3.9
B 5 7.8 8.4 LA182 B 5 5.1 4.0
LA22 8 8.6 7.8 8 4.1 3.8
2 5.6 6.2 2 2.8 3.8
C 5 5.9 6.2 C 5 3.4 4.1
8 5.6 6.3 8 2.7 4.1
2 4.4 3.8 2 1.9 2.4
A 5 4.2 4.0 A 5 1.1 2.7
8 4.3 4.3 8 2.6 2.9
2 4.5 4.3 2 3.1 4.2
B 5 4.6 3.2 B 5 2.7 2.7
LA344 8 4.6 3.3 LA652 8 5.6 2.4
2 5.7 4.3 2 1.6 3.7
C 5 5.5 4.4 C 5 2.6 3.2
8 6.0 4.3 8 2.2 3.3
2 4.8 9.0
A 5 4.0 9.7 Legend: Lab. Mr – Laboratory resilient modulus
8 4.9 9.8 measured at a cyclic stress level of 37.2 kPa (5.4
2 12.6 23.5 lbf/in.2), contact stress level of 4.1 kPa (0.6 lbf/in.2),
LA28 B 5 10.3 23.5 and confining pressure of 14 kPa (2 lbf/in.2)
8 10.5 24.0

55
Results of the LADOTD Method
Figure 37 shows the results of comparing the LADOTD resilient modulus values obtained
from the soil support values (SSV) that are assigned for each parish (Appendix A, Table A2)
with those obtained from laboratory testing. The range of resilient modulus values for the
locations tested using the LADOTD method was from 7.6 to 9.2 ksi, while the laboratory
resilient modulus values ranged from 1 to 14 ksi. Most of the LADOTD method estimated
that Mr values are not comparable with the laboratory measured values. These results are
acceptable, as the LADOTD uses a typical average SSV value for the emitter parish;
however, the Mr value can vary from site to site within the parish.

Limitations of the Models

The prediction models developed in this study are valid for cohesive subgrade soils with Mr
values from 1 to 14 ksi, PI values from 3 to 66 percent, LL values from 20 to 99, and other
soil properties, as presented in Table 7.

16

Dynaflect
Mr = 0.41 Efwd + 2.26
R2 = 0.73
l
ve
12 Le
ion
ict
ed
measured resilient modulus (ksi)

Pr
%
r 95
pe
Up

8
el
L ev
on
i ct i
ed
Pr
5%
r9
we
Lo

0
0 5 10 15 20 25
modulus from FWD (ksi)

Figure 36
Dynaflect statistical results

56
 15
LADOTD-Method

LADOTD method estimated Mr (ksi)

10

0
0 5 10 15
Measured Mr (ksi)

Figure 37
LADOTD method estimated resilient modulus

57
 Table 29
Summary of the analysis

Method Model R2 RMSE Comments

(ksi)
DCP-Soil ⎛ 1 ⎞ ⎛ γd ⎞ Subgrade soils:
Property Model M r = 165.5 ⎜ ⎟ + 0.0966 ⎜ w ⎟ 0.92 0.88 1<Mr<14 ksi
⎝ DCPI1.147 ⎠ ⎝ ⎠
DCP-Direct 151.8 Subgrade soils:
Model Mr = 0.91 0.88 1<Mr<14 ksi
( DCPI )1.096

CIMCPT- Soil γd Subgrade soils:

Property Model M r = 3.55q c + 52.88f s + 0.52( ) 0.86 0.96 1<Mr<14 ksi
w
CIMCPT- Mr=2.12+ 3.44qc+63.15fs 0.77 1.34 Subgrade soils:
Direct Model 1<Mr<14 ksi
Dynaflect M r = 0.41Ed + 2.26 0.73 1.46 Nomographs and
temperature correction
ELMOD 5.1.69 M r = 0.40 E fwd + 0.49 0.71 1.32 7-Sensor no seed value

MODULUS 6 M r = 0.27 E fwd + 0.82 0.52 1.60 Semi infinite subgrade

EVERCALC M r = 0.26 E fwd + 1.19 0.51 1.62 Subgrade soils

5.0 1<Mr<14 ksi
Florida M r = 0.24 E fwd + 0.94 0.49 1.65 Subgrade soils
Equation 1<Mr<14 ksi
LADOTD No correlation established N/A N/A N/A
Method
Legend: DCPI – Dynamic cone penetration index (mm/blow), Ed- Modulus from Dynaflect (ksi), Efwd- Modulus
from FWD (ksi), LADOTD- Louisiana Department of Transportation and Development, N/A- Not applicable,
Mr –Resilient modulus (ksi) at a cyclic stress level of 37.2 kPa (5.4 lbf/in.2), contact stress level of 4.1 kPa (0.6
lbf/in.2), and confining pressure of 14 kPa (2 lbf/in.2), qc –Tip resistance (ksi), fs – Sleeve friction (ksi), RMSE-
Root mean square for error (ksi), γd –Dry unit weight (pcf), w – Water content (%)

58
 CONCLUSIONS

This report presents the development of models in an effort to predict the resilient
modulus of subgrade soils from the test results of DCP, CIMCPT, FWD, Dynaflect, and
soil properties of subgrade soils. Field and laboratory testing programs were conducted.
The field testing program included DCP, CIMCPT, FWD, and Dynaflect testing, whereas
the laboratory program included repeated load triaxial resilient modulus tests and
physical properties and compaction tests. Comprehensive statistical analyses were
conducted on the laboratory and field test results of subgrade soils. Based on the results
of this study, the following conclusions can be drawn:

• The DCP soil-property model ranked the best for the prediction of resilient modulus
of subgrade soils, followed by the DCP direct model, the CIMCPT soil-property
model, the CIMCPT direct model, Dynaflect, ELMOD 5.1.69, MODULUS 6,
EVERCALC 5.0, the Florida equation, and the current DOTD method.

• A good agreement was obtained between the Mr predicted using DCPI and those
measured using repeated load triaxial tests.

• The predicted Mr values obtained from the CIMCPT direct model, which included
CIMCPT tip resistance and sleeve friction as independent variables, matched the
measured Mr values. This demonstrates the applicability of the CIMCPT test results
in predicting the Mr of pavement subgrade cohesive soils.

• The DCP and CIMCPT test results are influenced by the soil properties, and the two
models were enhanced when moisture content and dry unit were incorporated.

• Among all backcalculated FWD moduli, those backcalculated using ELMOD 5.1.69
software had the best correlation with Mr measured in the laboratory repeated loading
triaxial tests.

• From a practical standpoint, the subgrade modulus, as determined from the DCP-soil
property model, DCP-direct model, CIMCPT soil-property model, CIMCPT direct
model, Dynaflect, or FWD utilizing ELMOD 5.1.69 backcalculation software, may be
used with the same confidence, considering the ranges of the coefficient of
determination.

59
 • The coefficients of determination (R2) for models predicting Mr of subgrade soils
using the MODULUS 6, EVERCALC 5.0, and the Florida equation were the lowest
among the models developed.

• The Mr values estimated using the approach currently adopted by the LADOTD
were found to correlate poorly with the laboratory measured Mr values.

60
 RECOMMENDATIONS

This report presents the results of a study conducted in an effort to develop resilient modulus
prediction models of subgrade soils from different in situ tests such as FWD, Dynaflect,
CIMCPT, and DCP. The approach of predicting Mr used in this study is an improvement
over the current procedure used by LADOTD in pavement design application. The fact that
the models are mainly applicable to cohesive soils with PI values from 3 to 66 percent, LL
values from 20 to 99, and other soil properties, as presented in Table 7 is noted.

The following initiatives are recommended in order to facilitate the implementation of this
study:

1) Implement the results of this study into the design manual for use by LADOTD engineers.

2) Establish an implementation and verification study through field projects. Selected

projects should incorporate various types of cohesive soils.

3) The proposed study should incorporate granular soils in order to facilitate the development
of generalized Mr prediction models for all soils encountered during construction of
roadways in Louisiana, as the models in this study were developed for cohesive soils and
may not be capable of predicting Mr values of granular soils.

61
 REFERENCES

1. Van Til, M.; McCullough, B.; Vallerga, B.; and Hicks, R.; Evaluation of AASHTO
Interim Guides for Design of Pavement Structures, Report NCHRP 128, Highway
Research Board, 1972.

2. Mohammad, L.N.; Puppala, A. J.; and Alavilli, P.; Investigation of the Use of
Resilient Modulus for Louisiana Soils in the Design of Pavements, Final Report
FHWA/LA-94/283, Louisiana Transportation Research Center, Baton Rouge, 1994.

3. NCHRP Project 1-28 A.; Harmonized Test Methods for Laboratory Determination of
Resilient Modulus For Flexible Pavement Design. 2003.

4. Tumay, M.T.; and Kurupp, P.U.; and Boggess, R.L; “A Continuous Intrusion
Electronic Miniature Cone Penetration Test System for Site Characterization,”
Geotechnical Site Characterization, Proc. 1st International Conf. On site
characterization-ISC’98, Atlanta, Vol. 1, 1998, pp. 1183-1188.

5. Herath, A., “A Study of the Applicability of Intrusion Technology for Evaluating

Resilient Modulus of Subgrade Soil,” Ph.D. Dissertation, Department of Civil and
Environmental Engineering, Louisiana State University, 2001.

6. Mohammad, L.N., and Herath, A. Resilient and Permanent Deformation Properties

of Untreated and Treated Unbound Pavement Materials. Interim Report, ALF
Experiment No. 4, Louisiana Transportation Research Center, Baton Rouge, LA,
2005.

7. Mohammad L.N.; Huang, B.; Puppala, A.; and Alen A; “A Regression Model for
Resilient Modulus of Subgrade Soils,” In Transportation Research Record No. 1687
TRB, Natinal Resrach Council, Washington D.C., 1999, pp. 47-54.

8. Mohammad, L.N.; Titi, H.H.; and Herath. A., “Evaluation of Resilient Modulus of
Subgrade Soil by Cone Penetration Test Results,” Seventh International Conference
on Low-Volume Roads, Vol. 1, Baton Rouge, Louisiana, May 1999, pp.236-245.

9. Mohammad, L.N,; Titi H.H.; and Herath. A.; “Intrusion Technology: An Innovative
Approach to Evaluate Soil Resilient Characteristics.” ASCE annual convention,
Boston, 1998, pp.39-58.

10. Mohammad, L.N; Titi H.H.; and Herath A.; Investigation of the Applicability of
Intrusion Technology to Estimate the Resilient Modulus of Subgrade Soil. Final
Report No. 332, Louisiana Transportation Research Center, 2000.

11. Mohammad, L.N; Titi H.H.; and Herath, A.; Effect of Moisture Content and Dry Unit
Weight on the Resilient Modulus of Subgrade Soils Predicted by Cone Penetration

63
 Test, Final Report No. 355, Louisiana Transportation Research Center, Baton Rouge,
Louisiana, U.S.A., , 2002.

12. Mohammad, L.N; Titi H.; and Herath A.; “Determination of Resilient Modulus of
Cohesive Soils Using the Continuous Intrusion Miniature Cone Penetration Test.”
ASTM Special Technical Publication, No. 1437, 2003, pp. 233-251.

13. Mohammad, L.N; Herath, A.; and Gudishala, R.; Development of Models to Estimate
the subgrade and subbase Layers Resilient Modulus from In-situ Devices Test Results
for Construction Control, Final Report No. 406, Louisiana Transportation Research
Center, Baton Rouge, Louisiana, U.S.A., 2007.

14. Nazef , A., and Choubane, B., “Survey of Current Practices of Using Falling Weight
Deflectometers (FWD),”Proceeding of Pavement Evaluation Conference,
Gainesville, 2002.

15. Rahim, A.M., and George, K.P., “Subgrade Soil Index Properties to Estimate
Resilient Modulus,” CD-ROM of Transportation Research Board Annual Meeting,
2004.

16. Webster, S.L.; Brown, R.W.; and Porter, J.R, Force Projection Site Evaluation Using
the Electric Cone Penetrometer and the Dynamic Cone Penetrometer, Report GL-94-
17U.S., Waterways Experimental Station, 1994.

17. Powell, W.D.; Potler, J.F.; Mayhew, H.C.; and Nunn, M.E., 1084. The Structural
Design of Bituminous Roads. TRRL, Report LR 1,132, 62 pp., 1990.

18. Hassan, A., “The Effect of Material Parameters on Dynamic Cone Penetrometer
Results for Fine-grained Soils and Granular Materials,” Ph.D Dissertation, Oklahama
State University, Stillwater, 1996.

19. George, K.P. and Uddin, W.; Subgrade Characterization for Highway Pavement
Design. Final Report, MS-DOT-RD-00-131, 2000.

20. Abu-Farsakh, M.Y.; Alshibli, K.; Nazzal, M. D.; and Seyman, E., Assessment of In-
Situ Test Technology for Construction Control of Base Courses and Embankments,
Final Report No. 385, Louisiana Transportation Research Center, Baton Rouge,
Louisiana, 2004.

21. Choubane, B., and McNamara, R.L., A Practical approach to predicting Flexible
Pavement embankment moduli using Falling Weight Deflectometer (FWD) data,
Research Report FL/DOT/SMO/00-442, Florida Department of Transportation, State
Materials Office, 2000.

22. Backcalculation Software ELMOD version 5.169, Dynatest Consulting, Inc., Ojai,
California 93023, 5.1.69.

64
23. Backcalculation Software MODULUS version 6.0, Texas Transportation Institute
(TTI), Texas Department of Transportation.

24. Backcalculation Software Evercalc version 5.0, Washington Department of

Transportation.

25. Southgate, H.F., “An Evaluation of Temperature Distribution within Asphalt

Pavements and its Relationship to Pavement Deflections,” Commonwealth of
Kentucky, Department of Transportation, Bureau of Highways, Division of Research,
April 1968.

26. Kinchen, R. W., and Temple, W. H., Asphaltic Concrete Overlays of Rigid and
Flexible Pavements. Report FHWA/LA-80/147, Louisiana Department of
Transportation and Development, Baton Rouge, LA, 1980.

27. AASHTO. “Standard Method of Test for Determining the Resilient Modulus of Soils
an Aggregate Materials”, American Association of State Highway and Transportation
Officials, 1993. T 307-99, 2003.

28. Thompson, M.R., and Robnett, Q.L., “Resilient Properties of Subgrade Soil.” ASCE
Transportation Engineering Journal, 1979, pp.1-89.

29. Rada, G.R.; Rabinow, S.D.; and Witczak M.W.; and Richter C.A., “Strategic
Highway Research Program Falling Weight Deflectometer Quality Assurance
Software.” Nondestructive Deflection Testing and Backcalculation for Pavements,
Proceedings of a Symposium, Transportation Research Record, Journal of the
Transportation Research Board, No. 1377, 1992, pp. 36-44.

65
 Appendix A

Figure A1: Typical Profile of Tip Resistance (qc), Sleeve Friction (fs), and Predicted Mr
(LA333 Site, Test Point C8)

Table A1: Test Results for Verification of DCP Models

Table A2: Mr Estimated From LADOTD Method

66
 Table A1
Test Results for Verification of DCP Models

Location Test γd w Lab. Mr DCPI

and Site Point (pcf) (%) (ksi) (mm/blow)

1591+00 117.2 14.6 7.4 23.3

1347+00 117.2 15.4 9.1 16.7
1595+00 107.1 18.2 10.2 12.3
1596+00 110.3 16.1 9.9 13.3
88+00 108.4 14.8 12.3 10.6
90+00 110.9 17.8 10.6 13.6
94+00 113.4 16.8 11.2 13.6
Mississippi 96+00 115.9 15.1 11.9 11.0
(George et al. [19]) 108+00 108.4 18.1 9.3 15.0
114+00 106.5 22.0 4.1 27.3
116+00 107.7 18.9 5.5 25.2
172+00 115.3 16.2 9.1 12.7
176+00 115.9 17.3 5.2 29.0
178+00 109.6 20.7 6.2 20.6
262+62 104.0 19.1 9.7 12.9
264+50 103.3 17.2 10.4 12.1
266+00 110.3 18.5 11.9 11.5
670+00 108.4 15.8 10.6 11.9
Legend: DCPI- DCP Index, V- Verification data from another study [19], w- Moisture content, γd- Dry unit
weight, Lab. Mr – Laboratory resilient modulus measured at a cyclic stress level of 37.2 kPa
(5.4 lbf/in.2) and confining pressure of 14 kPa (2 lbf/in.2)

67
 Table A2
Mr Estimated From LADOTD Method

Parish Soil Support Resilient Parish Soil support Resilient

Value Modulus (psi) value Modulus (psi)
Acadia 3.7 8797 Madison 3.8 9176
Allen 3.6 8413 Morehouse 3.8 9176
Ascension 3.6 8413 Natchitoches 4.0 9916
Assumption 3.5 8023 Orleans 3.4 7627
Avoyelles 3.8 9176 Ouachita 4.0 9916
Beauregard 3.7 8797 Plaquemines 4.0 9916
Bienville 4.0 9916 Pointe Coupee 3.8 9176
Bossier 3.7 8797 Rapides 4.0 9916
Caddo 4.1 10278 Red River 4.1 10278
Calcasieu 3.8 9176 Richland 3.9 9549
Caldwell 4.0 9916 Sabine 3.9 9549
Cameron 3.8 9176 St. Bernard 3.5 8023
Catahoula 3.7 8797 St. Charles 3.4 7627
Claiborne 4.1 10278 St. Helena 3.9 9549
Concordia 3.6 8413 St. James 3.5 8023
Desoto 3.8 9176 St. John 3.4 7627
East Baton Rouge 3.6 8413 St. Landry 3.8 9176
East Carroll 3.8 9176 St. Martin 3.5 8023
East Feliciana 4.4 11330 St. Mary 3.7 8797
Evangeline 3.9 9549 St. Tammany 3.8 9176
Franklin 4.0 9916 Tangipahoa 4.2 10634
Grant 4.0 9916 Tensas 3.8 9176
Iberia 3.8 9176 Terrebonne 3.7 8797
Iberville 3.6 8413 Union 4.1 10278
Jackson 3.8 9176 Vermillion 3.4 7627
Jefferson 3.5 6023 Vernon 3.7 8797
Jefferson Davis 3.6 8413 Washington 3.8 9176
Lafayette 4.0 9916 Webster 3.9 9549
Lafourche 3.8 9176 West Baton Rouge 3.8 9176
Lasalle 3.8 9176 West Carroll 3.9 9549
Lincoln 4.1 10278 West Feliciana 4.2 10634
Livingston 3.9 9549 Winn 4.0 9916

68
 Figure A1
Typical Profile of Tip Resistance (qc), Sleeve Friction (fs), and Predicted Mr
(LA333 Site, Test Point C8)

0.0 0.0 0.0

1.0 1.0 1.0

Depth, m

2.0 2.0 2.0

3.0 3.0 3.0

0 1 2 0.00 0.03 0.05 0 2 4 6 8

Tip Resistance (ksi) Sleeve Friction (ksi) Predicted Mr (ksi)

69
This public document is published at a total cost of $1,635.80.
370 copies of this public document were published in this first
printing at a cost of $1,020.80. The total cost of all printings of
this document including reprints is $1,635.80. This document
was published by Louisiana State University, Graphic Services,
3555 River Road, Baton Rouge, Louisiana 70802, and Loui-
siana Transportation Research Center, to report and publish
research findings for the Louisiana Transportation Research
Center as required in R.S. 48:105. This material was duplicat-
ed in accordance with standards for printing by state agencies
established pursuant to R.S. 43:31. Printing of this material
was purchased in accordance with the provisions of Title 43 of
the Louisiana Revised Statutes.