EVALUATION OF THE SERVICE PERFORMANCE

OF AN INNOVATIVE PRECAST PRESTRESSED

CONCRETE PAVEMENT

A Thesis
presented to
the Faculty of the Graduate School
University of Missouri – Columbia

In Partial Fulfillment
of the Requirements for the Degree

Master of Science

by
GRANT C. LUCKENBILL, E.I.

Dr. Vellore S. Gopalaratnam, P.E., Thesis Advisor

July 2009
 The undersigned, appointed by the Dean of the Graduate School, have examined the
thesis entitled

EVALUATION OF THE SERVICE PERFORMANCE OF AN INNOVATIVE

PRECAST PRESTRESSED CONCRETE PAVEMENT
presented by

Grant C. Luckenbill, E.I.

A candidate for the degree of

Master of Science

and hereby certify that in their opinion it is worthy of acceptance.

____________________________________
Dr. Vellore S. Gopalaratnam, P.E.

____________________________________
Dr. Glenn Washer, P.E.

____________________________________
Dr. Sanjeev K. Khanna

________________________________________________________________________
ACKNOWLEDGEMENTS

I would like to extend my gratitude to the Missouri Department of Transportation

and Federal Highway Administration for jointly funding the project. Many dedicated

individuals from MoDOT and FHWA have demonstrated strong commitments to the

advancement of new technologies, specifically John Donahue of MoDOT and Sam Tyson

of FHWA. Without their hard work and dedication, pilot projects such as these would

not be undertaken. I would like to acknowledge the many professionals at the MoDOT

District office in Sikeston, MO who were very supportive of this project. Efforts to

accommodate the research team’s many requests from Eric Krapf and Michael Chasteen

were much appreciated. Jim Copeland took time aside to aid the research team with

surveying the pavement. Finally, from MoDOT, I would also like to acknowledge Terry

Fields, senior construction inspector, for his help with the research team.

I would also like to commend the work of the Transtec Group including David

Merritt, design engineer, for providing design expertise and an extensive background in

precast pavements. Additionally I would like to thank the professionals from Concrete

Products Incorporated (recently acquired by Prestress Services Industries) and Gaines

Construction Company who were very accommodating and aided the research team with

coordination efforts during fabrication and construction. Specifically, Andrew Maybee

and Jimmy Davis from CPI who were very helpful with the research team during

fabrication.

I would like to thank my advisor, Dr. V.S. Gopalaratnam. Dr. Gopal has provided

guidance and demanded superior performance during my undergraduate and graduate

________________________________________________________________________

Page ii
degree programs. His unwavering dedication has significantly contributed to my growth

as an engineering professional. I will proudly carry the lessons you have taught

throughout the rest of my career.

I would also like to thank my parents, family, and friends for aiding in my

success. Without their support and guidance, many of these opportunities would not have

been made possible.

________________________________________________________________________

Page iii
TABLE OF CONTENTS

ACKNOWLEDGEMENTS ............................................................................................ II
TABLE OF CONTENTS ............................................................................................... IV
LIST OF TABLES .......................................................................................................... VI
LIST OF FIGURES .......................................................................................................VII
NOMENCLATURE / LIST OF NOTATION .............................................................XII
ABSTRACT .................................................................................................................. XIII
1. INTRODUCTION........................................................................................................1
1.1. GENERAL INFORMATION AND PROJECT SCOPE........................................1
1.1.1. PPCP Project on Interstate 57 and Experimental Investigation ...................1
1.1.2. Research Objectives .....................................................................................2
1.2. OVERVIEW OF PAVEMENT DETAILS AND CONSTRUCTION
PRACTICES ...........................................................................................................3
1.2.1. Fabrication at the Casting Yard ...................................................................4
1.2.2. Construction of Precast Pavement ...............................................................5
1.2.3. Pavement Panel Designs ..............................................................................7
1.3. ORGANIZATION OF THESIS ...........................................................................11
2. BACKGROUND INFORMATION .........................................................................13
2.1. OVERVIEW OF PPCP TECHNOLOGY ............................................................13
2.1.1. Design Considerations for PPCP ...............................................................14
2.2. PPCP PROJECTS IN THE UNITED STATES ...................................................18
2.2.1. Iowa Approach Slab on Highway 60 .........................................................19
2.2.2. Interstate 10 in El Monte, CA ....................................................................20
2.2.3. Outer Road near Interstate 35 in Georgetown, TX ....................................20
2.3. FIELD INSTRUMENTATION OF CONCRETE PROJECTS ...........................22
3. EXPERIMENTAL PROGRAM ...............................................................................25
3.1. FIELD INSTRUMENTATION ............................................................................25
3.1.1. Types of Embedded Instrumentation ........................................................25
3.1.1.1. Strain-Gage Rebar .............................................................................25
3.1.1.2. Vibrating Wire Gage .........................................................................27
3.1.1.3. Vibrating Wire Strandmeters ............................................................28
3.1.1.4. Temperature Gages ...........................................................................29
3.1.2. General Design and Placement Considerations .........................................30
3.1.3. Instrumentation Layouts ............................................................................32
________________________________________________________________________

Page iv
 3.1.4. Data Acquisition System............................................................................33
3.1.5. Remote Monitoring ....................................................................................35
3.2. LABORATORY EXPERIMENTS ......................................................................36
3.2.1. Materials Testing .......................................................................................36
3.2.2. Thermal Investigation of Embedded Instrumentation ...............................37
3.3. CHALLENGES FOR REMOTE DATA ACQUISITION ...................................41
3.3.1. Excessive Heat Buildup Affecting Sensitive Hardware ............................42
3.3.2. Moisture Intrusion and Corrosion ..............................................................43
3.3.3. Lightning Protection ..................................................................................44
3.3.4. Snow Removal and Protective Plates ........................................................45
4. SERVICE PERFORMANCE OF PPCP SYSTEM ................................................47
4.1. GENERAL INFORMATION ...............................................................................47
4.2. PAVEMENT THERMAL PERFORMANCE ......................................................48
4.2.1. Temperature Variations .............................................................................48
4.2.2. Daily Thermal Loadings ............................................................................49
4.2.3. Weekly Thermal Behavior .........................................................................57
4.2.4. Seasonal Variations in Panel and Global Pavement Responses ................61
4.3. PAVEMENT RESPONSE DUE TO VEHICULAR LOADING .........................70
4.4. EFFECTIVE POST-TENSIONING STRESS DISTRIBUTIONS ......................73
4.4.1. Post-tensioning Stress Distributions affected by
Poor Transfer at Joints ...............................................................................73
4.4.2. Post-tensioning Stress Losses due to Friction ............................................76
4.5. TRANSVERSE AND LONGITUDINAL CRACKING ......................................77
4.5.1. Expert Task Group Meeting in Sikeston, MO ...........................................80
4.5.2. Visual Crack Surveys .................................................................................82
4.6. JOINT PANEL PERFORMANCE .......................................................................82
5. CONCLUSIONS ........................................................................................................87
5.1. PROJECT OBSERVATIONS ..............................................................................87
5.2. CONSTRUCTION CHALLENGES ....................................................................88
5.3. SERVICE PERFORMANCE ...............................................................................89
5.4. PAVEMENT LONGEVITY ................................................................................90
5.5. RECOMMENDATIONS FOR FUTURE WORK ...............................................90

REFERENCES .................................................................................................................92
APPENDIX A ...................................................................................................................94
APPENDIX B ...................................................................................................................95

________________________________________________________________________

Page v
LIST OF TABLES

TABLE 4.1 – INSTRUMENT LOCATIONS AND EVENT SUMMARY FOR JOINT PANEL A32

ON JULY 13, 2006..........................................................................................51

TABLE B.1 – LOCATIONS OF INSTRUMENTS USED IN PANEL C1 ......................................96

TABLE B.2 – LOCATIONS OF INSTRUMENTS USED IN PANEL B1 ......................................96

TABLE B.3 – LOCATIONS OF INSTRUMENTS USED IN PANEL B2 ......................................97

TABLE B.4 – LOCATIONS OF INSTRUMENTS USED IN PANEL B3 ......................................97

TABLE B.5 – LOCATIONS OF INSTRUMENTS USED IN PANEL B4 ......................................98

________________________________________________________________________

Page vi
LIST OF FIGURES

FIGURE 1.1 – OVERALL PPCP SECTION LAYOUT WITH DRIVING LANES SHOWN (25

PANELS PER SECTION; SECTION 3 IS HEAVILY INSTRUMENTED) ................4

FIGURE 1.2 – TYPICAL SECTION OF PPCP PANEL ASSEMBLY AND LAYOUT MODIFIED TO

REFLECT MISSOURI PROJECT (MERRITT, MCCULLOUGH ET AL. 2000)...4

FIGURE 1.3 – JOINT PANEL ON POLYPROPYLENE OVER ASPHALT, AND AGGREGATE BASE

(MISSOURI PROJECT) (NOTE: INSTRUMENTATION DATA CABLE EXITING

THE END OF THE PANEL) ..............................................................................6

FIGURE 1.4 – PLAN VIEW OF TYPICAL BASE PANEL ............................................................8

FIGURE 1.5 – SECTION OF BASE PANEL LOOKING PERPENDICULAR TO TRAFFIC

DIRECTION ...................................................................................................8

FIGURE 1.6 – LIFTING ANCHOR, CHAIRS, AND PRESTRESSING STRANDS ...........................8

FIGURE 1.7 – PLAN VIEW OF TYPICAL JOINT PANEL ........................................................10

FIGURE 1.8 – SECTION OF JOINT PANEL LOOKING PERPENDICULAR TO TRAFFIC ...........10

FIGURE 1.9 – JOINT PANEL CASTING (LEFT SIDE CURED FOR 1 NIGHT, RIGHT SIDE READY

FOR CASTING ON 2ND DAY) ........................................................................11

FIGURE 2.1 – ILLUSTRATION OF PAVEMENT SECTION SPANNING OVER VOID IN BASE

MATERIAL ..................................................................................................15

FIGURE 2.2 – SURFACE FINISHING OF A TYPICAL BASE PANEL AT THE

PRECASTING YARD ....................................................................................16

FIGURE 2.3 – PARTIAL WIDTH PANEL PLACEMENT ON GEORGETOWN FRONTAGE ROAD,

TX PPCP (MERRITT 2002) .......................................................................21

________________________________________________________________________

Page vii
FIGURE 3.1 – SCHEMATIC OF THE STRAIN GAGE CONFIGURATION ON THE STRAIN GAGE

REBAR (EATHERTON 1999).......................................................................26

FIGURE 3.2 – INSTRUMENTED REBAR SHOWING INSTALLED STRAIN GAGES ..................26

FIGURE 3.3 – MODEL 4200 VIBRATING WIRE GAGE FROM GEOKON INCORPORATED ....27

FIGURE 3.4 – MODEL 4410 VIBRATING WIRE STRANDMETER, UNSHEATHED .................29

FIGURE 3.5 – THREE THERMOCOUPLES ATTACHED TO A FIBER REBAR COUPLED TO

POST-TENSIONING AND PRE-TENSIONING STRANDS ..................................30

FIGURE 3.6 – TYPICAL INSTRUMENTED BASE OR ANCHOR PANEL ...................................32

FIGURE 3.7 – INSTRUMENTED JOINT PANEL A32 .............................................................32

FIGURE 3.8 – OVERALL VIEW OF TEST-SECTION AND LOCATION OF INSTRUMENTED

PANELS (A REFERS TO A JOINT PANEL, B REFERS TO A BASE PANEL,

AND C REFERS TO AN ANCHOR PANEL) .....................................................33

FIGURE 3.9 – SIGNAL CABINET WITH MAIN DATA-ACQUISITION EQUIPMENT INSTALLED

AT THE EDGE OF RIGHT OF WAY ................................................................34

FIGURE 3.10 - JUNCTION BOX INSTALLED IN BLOCKOUT CAST IN OUTSIDE SHOULDER OF

PRECAST PAVEMENT PANELS.....................................................................35

FIGURE 3.11 – UNRESTRAINED INSTRUMENTED REBAR AND VIBRATING WIRE GAGE IN

TEMPERATURE CONTROLLED OVEN..........................................................38

FIGURE 3.12 – (A) TEMPERATURE HISTORY (B) STRAIN HISTORY OF EMBEDDED AND

UNRESTRAINED REBAR INSTRUMENTS ......................................................40

FIGURE 3.13 – IDEALIZATION OF INSTRUMENT RESPONSE DUE TO AN INCREASE IN

TEMPERATURE, ΔT ...................................................................................41

________________________________________________________________________

Page viii
FIGURE 3.14 – SIGNAL CABINET PROTECTED DURING THE HEAT OF THE DAY BY A SHADE

ROOF ..........................................................................................................43

FIGURE 3.15 – CLOSE-UP OF CJC DAMAGED BY LIGHTNING...........................................45

FIGURE 4.1 – DAY AND NIGHT COOLING TRENDS ...........................................................49

FIGURE 4.2 – ONE DAY WINDOW FROM JULY 13, 2006 FOR PANEL A32 (A)

TEMPERATURE HISTORY (B) STRAIN HISTORY ..........................................52

FIGURE 4.3 – ONE DAY WINDOW FROM JULY 13, 2006 SHOWING ALL INSTRUMENTS FOR

PANEL A32 (A) TEMPERATURE HISTORY (B) STRAIN HISTORY ................54

FIGURE 4.4 – ONE DAY WINDOW FROM JULY 13, 2006 FOR PANEL B3 (A) TEMPERATURE

HISTORY (B) STRAIN HISTORY ...................................................................55

FIGURE 4.5 – JOINT PANEL A31 DURING MILD TEMPERATURES (LEFT) AND HOT

TEMPERATURES (RIGHT)...........................................................................56

FIGURE 4.6 – MEASURED CONCRETE STRAINS IN PAVEMENT AT A SHORT-TERMED

WINDOW (A) TEMPERATURE HISTORY (B) STRAIN HISTORY .....................58

FIGURE 4.7 – MEDIUM WINDOW INDICATING WEEKLY HEATING AND DRASTIC COLD

FRONT WITH ASSOCIATED CONCRETE STRAINS (A) TEMPERATURE

HISTORY (B) STRAIN HISTORY ...................................................................60

FIGURE 4.8 – WEEKLY INSTRUMENT HISTORY FOR PANEL B3 (A) TEMPERATURE

HISTORY (B) STRAIN HISTORY ...................................................................61

FIGURE 4.9 – SIX MONTH WINDOW SHOWING LONGITUDINAL CONCRETE STRAINS IN

PANEL B2 AT DIFFERENT LOCATIONS (R2 AND R5) (A) TEMPERATURE

HISTORIES (B) STRAIN HISTORIES..............................................................63

________________________________________________________________________

Page ix
FIGURE 4.10 – SIX MONTH WINDOW SHOWING LONGITUDINAL CONCRETE STRAINS AT

IDENTICAL PANEL LOCATION (R2) IN DIFFERENT INSTRUMENTED PANELS

(A) TEMPERATURE HISTORIES (B) STRAIN HISTORIES ..............................65

FIGURE 4.11 – SIX MONTH WINDOW SHOWING TRANSVERSE CONCRETE STRAINS AT

IDENTICAL PANEL LOCATION (R3) IN DIFFERENT INSTRUMENTED PANELS

(A) TEMPERATURE HISTORIES (B) STRAIN HISTORIES ..............................67

FIGURE 4.12 – TWO MONTH WINDOW SHOWING LONGITUDINAL CONCRETE STRAINS AT

IDENTICAL PANEL LOCATION (R1-V1) MEASURED USING INSTRUMENTED

REBAR R1 AND VIBRATING WIRE GAGE (V1) (A) TEMPERATURE HISTORY

(B) STRAIN HISTORY ...................................................................................68

FIGURE 4.13 – STRANDMETER RESPONSE AT CENTER OF 250’ TEST SECTION DURING

TYPICAL WINTER-TIME TEMPERATURE EXCURSION (A) TEMPERATURE

HISTORIES (B) STRAIN HISTORIES..............................................................70

FIGURE 4.14 - TRAFFIC STRAIN (REBAR RESPONSE) IN THE PAVEMENT AT CROWN .......71

FIGURE 4.15 – DURATION OF TRAFFIC RESPONSE THAT WAS VERIFIED VISUALLY .........72

FIGURE 4.16 – RESULTING CONCRETE RESPONSE FROM A TRACTOR TRAILER PASSING

OVER ..........................................................................................................73

FIGURE 4.17 – CONCRETE STRAIN FOR A TYPICAL BASE PANEL DURING POST-

TENSIONING OPERATIONS..........................................................................75

FIGURE 4.18 – CONCRETE STRAIN MEASURED AT THE R2 LOCATION (BEGINNING OF

OUTSIDE SHOULDER) .................................................................................76

FIGURE 4.19 – AVERAGE POST-TENSIONING CONCRETE STRAIN HISTORIES FOR

INSTRUMENTED PANELS ............................................................................77

________________________________________________________________________

Page x
FIGURE 4.20 - LONGITUDINAL CRACK IN DRIVER SIDE WHEEL LANE ..............................78

FIGURE 4.21 - SCHEMATIC OF ONE LONGITUDINAL CRACK ............................................79

FIGURE 4.22 – TYPICAL CRACK LOCATIONS OF A 4-PANEL SET OF THE 3RD TEST

SECTION ON MAY 9, 2007 ..........................................................................80

FIGURE 4.23 – FLEXIBLE JOINT COMPOUND SQUEEZING OUT ON A HOT DAY WITH

MINOR AMOUNT OF CHIPPING OF RIGID COMPOUND, JOINT PANEL A32

(JUNE 27, 2006)..........................................................................................84

FIGURE 4.24 – RIGID JOINT COMPOUND CHIPPED AWAY MORE EXTENSIVELY, JOINT

PANEL A32 (AUGUST 16, 2006) .................................................................84

FIGURE 4.25 – MODERATE DEGRADATION TO RIGID JOINT COMPOUND, JOINT PANEL

A32 (MAY 9, 2007) ....................................................................................85

FIGURE A.1 – ONE DAY WINDOW FROM 12/27/2006 FOR PANEL A32 (A) TEMPERATURE

HISTORY (B) STRAIN HISTORY ...................................................................94

FIGURE B.1 – CONVENTION FOR GAGE LOCATIONS .........................................................95

________________________________________________________________________

Page xi
NOMENCLATURE / LIST OF NOTATION

T - TEMPERATURE

t - Time

ε - Strain

R - Rebar Strain Reading

CTE - Coefficient of Thermal Expansion

PSI - Pounds per Square Inch

A - Area of Concrete for a Given Cross-Section

UPS - Uninterruptible Power Supply

CJC - Cold Junction Compensation (Circuit)

α - Coefficient of Thermal Expansion

BN - Instrumented Base Panel Label (Number ‘N’)

CN - Instrumented Anchor Panel Label (Number ‘N’)

A3N - Instrumented Joint Panel Label (Number ‘N’)

RN - Instrumented Rebar Label (Number ‘N’)

VN - Vibrating Wire Gage Label (Number ‘N’)

TN - Thermocouple Label (Number ‘N’)

N. A. - Neutral Axis

E - Modulus of Elasticity

ν - Poisson’s Ratio

________________________________________________________________________

Page xii
EVALUATION OF THE SERVICE PERFORMANCE OF AN
INNOVATIVE PRECAST PRESTRESSED CONCRETE
PAVEMENT

Grant C. Luckenbill, E.I.

Dr. Vellore S. Gopalaratnam, Thesis Advisor

ABSTRACT

Precast Prestressed Concrete Pavement (PPCP) has many advantages over

conventional roadway construction techniques. PPCP is the product of an optimization

of conventional materials coupled with economical fabrication and transportation means

to create a product that exceeds the performance and implementation of current pavement

rehabilitation methods. Pre-compressing concrete pavements results in a more efficient,

thinner section translating to material savings as well as improved long-term durability.

Precast pavement allows faster replacement and rehabilitation of existing roadways as

well as providing an economical alternative for new construction to minimize undesirable

traffic congestion that causes increased fuel consumption and lost productivity.

Decreased construction times are a significant advantage in locations where elevated

hazards pose additional risk to worker safety and construction seasons are limited. This

project, near Sikeston, MO on Interstate 57, explored feasibility and long-term

performance of precast roadway panels subjected to adverse ‘Midwest environment’

________________________________________________________________________

Page xiii
(extreme temperatures in summer accompanying deicing salts in winter) in addition to

evaluation of current construction methods.

The focuses of this thesis are to characterize the thermal behavior and evaluate the

overall service performance of the pavement system. Results of thermodynamic

experiments, to develop an understanding of the output of strain gage instrumented rebar

cast in concrete, are presented. Analysis of results from the investigation include: (a)

construction challenges that may affect long term durability (b) local and global prestress

distributions within PPCP (c) stress losses during post-tensioning operations (frictional

and stress transfer between panels), (d) daily thermal loadings, (e) weekly and seasonal

temperature variations and corresponding pavement behavior. Pavement response to

traffic loads is presented to contrast daily thermal loadings. Visual crack surveys

(longitudinal and transverse) and joint panel performance over the year long evaluation

period are discussed.

________________________________________________________________________

Page xiv
1. Introduction

1.1. General Information and Project Scope

In conjunction with the Federal Highway Administration, the Missouri

Department of Transportation committed to jointly fund and build a new Precast

Prestressed Concrete Pavement (PPCP). A section of Interstate 57 near Charleston, MO

was chosen for rehabilitation using the PPCP program. The project was completed in

December, 2005 and opened to traffic in mid January, 2006. The goal of the Missouri

project was to advance technologies developed in recently completed projects in

Georgetown, TX and El Monte, CA and evaluate the durability of PPCP in harsh

environmental conditions.

1.1.1. PPCP Project on Interstate 57 and Experimental Investigation

The Interstate 57 project was the first large scale PPCP project undertaken in

Missouri. Sufficient right of way and funding enabled the installation of asphalt

crossovers which aided the construction of the pavement project by relieving time

constraints on constructors. In turn, this enabled constructors to experiment and work out

the most efficient methods for construction. The site in southern Missouri was also

chosen to evaluate the performance of the PPCP technology subjected to harsh

environmental conditions. Missouri is known to have extreme seasonal temperature

variations. De-icing salts are commonly used on their roadways. The long-term

durability of the PPCP test section was “put to the test” with the combination of harsh

environmental conditions and heavy truck traffic (approximately 30% of ADT).

________________________________________________________________________

Page 1
1.1.2. Research Objectives

The charge of the University of Missouri – Columbia research team was to

evaluate the performance of the PPCP subjected to severe weather and traffic conditions

and develop performance data useful for future projects. The team heavily instrumented

several panels to quantify pavement performance and validate design assumptions. Due

to the broad scope of the research goals of the overall pavement project, this thesis is

accompanied by two companion theses by Cody Dailey “Instrumentation and Early Age

Performance of an Innovative Prestressed Precast Pavement System” (2006) and Brent

Davis “Evaluation of Prestress Losses in an Innovative Prestressed Precast Pavement

System” (2006). These reports, which focus on the instrumentation, materials testing,

fabrication, early-age behavior, and construction of the PPCP, will be referenced

throughout this thesis when overlapping topics are discussed. The project report

submitted to MoDOT by the research team, “Performance Evaluation of Precast

Prestressed Concrete Pavement, RI03-007,” will also be referenced throughout this paper.

This thesis presents thermal and strain gradient data in conjunction with laboratory

experiments which focus largely on the characterization of service performance of the

PPCP. This data will be helpful in quantifying the effectiveness of PPCP as a rapid

rehabilitation pavement alternative to conventional design practice characterized by the

following:

o Evaluation of construction methods on the behavior of prestressed system

to optimize design and aid in developing preferred practices to expedite

construction.

________________________________________________________________________

Page 2
 o Study of daily and seasonal temperature dependent effects on the concrete

panels and their interaction within the post-tensioned pavement system.

o Overall pavement performance with respect to longevity, durability and

how cracking and prestress losses may affect these characteristics.

1.2. Overview of Pavement Details and Construction Practices

The Precast Prestressed Concrete Pavement (PPCP) test section in Missouri

replaced a dilapidated 45 year old section of cast-in-place (CIP) concrete just west of

Charleston, MO on the northbound lanes of I-57. Three specific types of precast,

prestressed panels make up the PPCP system: base panels, joint panels, and anchor

panels. They are pre-tensioned in the transverse direction at the casting yard and post-

tensioned in the longitudinal direction (parallel to traffic). Each panel is 10’-0” x 38’-0”.

The 38 ft dimension is perpendicular to traffic. There is a 4’-0” inside shoulder, two 12’-

0” driving lanes and a 10’-0” outside shoulder. The panels were cast with a constant 2%

grade from the crown to ensure proper drainage.

A 1,010 ft section of conventional cast-in-place pavement was replaced with four

sections of post-tensioned precast pavement panels. A typical section consisted of an

anchor panel near the middle with eleven or twelve base panels on each side with a joint

panel at both ends. These joint panels were heavily reinforced since they contained the

post-tensioning blockouts and served to accommodate thermal expansion/contraction of

the 250 ft section. Figure 1.1 shows the layout of the four sections of pavement (the

highlighted section is heavily instrumented and will be discussed later in detail). Figure

1.2 shows the layout of panels within each individual section.

________________________________________________________________________

Page 3
Figure 1.1 – Overall PPCP section layout with driving lanes shown (25 panels per
section; Section 3 is heavily instrumented)

Figure 1.2 – Typical section of PPCP panel assembly and layout modified to reflect
Missouri Project (Merritt, McCullough et al. 2000)

1.2.1. Fabrication at the Casting Yard

All 101 precast panels were fabricated by Concrete Products Incorporated (CPI)

in Memphis, TN between mid-October and December 2005. They were cast two at a

time, crown up, in self-stressing steel casting beds outdoors. The panels were pre-

tensioned in the transverse direction using 0.5” uncoated, seven-wire low relaxation

strands (270 ksi). The pre-tensioning was incorporated largely to accommodate lifting

and transportation stresses. Post-tensioning parallel to the direction of traffic was

performed after placement. After casting, the panels were steam cured overnight. This

was done to minimize shrinkage and ensure proper curing of the panels. The following

________________________________________________________________________

Page 4
morning, they were de-tensioned and de-molded provided the concrete had reached an

initial compressive strength of 3500 psi. Later they were stacked in the storage location

of the casting yard to await transportation to the site near Charleston. This single day

casting/curing/de-molding of the precast panels enabled the fabricator to turn out two

typical base panels per day. The joint panels, which were more complicated in design to

fabricate, were cast in two days. An in-depth description of the manufacturing

procedures of the precast panels was detailed in (Dailey 2006).

1.2.2. Construction of Precast Pavement

The construction of the PPCP section was performed by Gaines Construction

from Wentzville, MO. Typical panel placement rates were between 8 – 25 panels per

day. The base, a very critical component of pavement construction, was made up of a 4”

asphalt treated base over a 4” permeable crushed stone filter layer. A layer of

polypropylene was used as a friction reduction layer and to ease the construction of the

pavement. This friction reduction layer is critical in post-tensioned pavement where sub-

grade friction reduces the efficiency of prestressing. Figure 1.3 shows an instrumented

joint panel on top of the prepared asphalt base with polypropylene sheet clearly visible.

________________________________________________________________________

Page 5
 PT Blockout

Joint Panel

Pavement Poly. Sheeting

Base

Figure 1.3 – Joint panel on polypropylene over asphalt, and aggregate base
(Missouri project) (Note: Instrumentation data cable exiting the end
of the panel)

Panels were taken directly from the trucks via crane and positioned at the end of

the PPCP section. The joints between the panels were sealed with a slow-curing epoxy,

intended to ease placement and seal the joints from water intrusion. Two post-tensioning

strands (0.6 inch seven-wire low relaxation strands) were fed through and stressed lightly

to recover gaps and slack in the pavement system. However, during complications of

feeding post-tensioning strands much of the epoxy placed on the joints was allowed to

harden. This resulted in an uneven surface at the joint which impaired uniform load

transfer between panels. Additional information regarding the implications of the

hardened epoxy layer is discussed in Chapter 4. Wooden and steel shims were placed

between panels on the southern edge to aid in recovering pavement misalignment. The

usage of rigid shims resulted in uneven distribution of post-tensioning stresses across the

panels. Construction challenges such as these were mitigated during construction of the

________________________________________________________________________

Page 6
PPCP pavement. However, adjustments made during construction affected the pre-

stressed pavement and complicated the monitored service performance.

1.2.3. Pavement Panel Designs

The Transtec Group from Austin, TX designed the PPCP system consisting of

base, anchor, and joint panels. The majority of the panels (92 of 101) were made up of

base panels, with 4 anchor and 5 joint panels.

Standard 60 ksi epoxy coated rebar bordered the edges of each base panel and is

not shown on the schematics. Its role was to provide typical edge reinforcement to curb

cracking and fragmentation of the corners. Each base panel contained eight pre-

tensioning strands as shown in Figure 1.4 and Figure 1.5. The pre-tensioning strands

located in the top half of the panel were draped to follow the slope of the crown and meet

cover requirements at the shoulders. Draping was accomplished by placing varying

heights of galvanized chairs under the strands at key locations (see Figure 1.6).

Pretensioned Strands
(8 @ 1’-3”)
38’

10’

Inside Outside
Post-Tensioning Ducts
Shoulder Shoulder
(18 @ 2’)

Figure 1.4 – Plan view of typical base panel

________________________________________________________________________

Page 7
 Pre-tensioning Strands

Outside
Shoulder 7” 10 7/8”

10 -0”
Figure 1.5 – Section of base panel looking perpendicular to traffic direction

Figure 1.6 –Lifting anchor, chairs, and prestressing strands

Anchor panels are similar to base panels with the addition of full depth holes near

the center. These panels are located at the midpoint of each PPCP section and anchor the

entire section globally providing a restrained thermal origin to minimize displacements at

joints. Reinforcing dowels were driven into the sub grade through the 4 in diameter

blocked-out anchor sleeves and grouted. The fabrication of the anchor panels were cast

intermittently with the similarly designed base panels.

Joint panel fabrication began mid-December. Due to the complexities of

retooling, amount of reinforcement, and functional geometry of the panel, each joint

________________________________________________________________________

Page 8
panel required to be cast in two separate halves. Figure 1.7 shows a plan view of a

typical joint panel. Block-outs toward the center of the panel allow access to the post-

tensioning ducts, which were filled with grout after post-tensioning. Joint panels have 12

pre-tensioned strands instead of 8 (see Figure 1.8). The top strands are draped with the

slope of the crown while the bottom strands are straight. Each half is connected by

smooth dowels that provide shear transfer between sections (not shown in schematic). A

cold-joint between sections was accomplished by using temporary bulkhead that fastened

to the bottom of the bed. The cold-joint was needed to ensure that the joint panel

“opened up” during post-tensioning operations. Figure 1.9 shows two halves of a joint

panel, one side has cured for one night and the other is ready for casting the following

afternoon. Five total joint panels were fabricated for the 1,010 stretch of pavement. The

panels at the ends of the PPCP section were cast without post-tensioning block-outs on

one half. The side of the panels adjoining to cast-in-place concrete pavement was

dowelled in with conventional rebar reinforcement.

Pretensioned Strands (4 @ 6” from Edges; Post-Tensioning

2 @ 1’ from Outside Edge) Block-Outs
38’

Traffic Direction

10’

Inside Outside
Post-Tensioning Ducts
Shoulder Shoulder
(18 @ 2’)

Figure 1.7 – Plan view of typical joint panel

________________________________________________________________________

Page 9
 Prestressing Strands Post-tensioning Blockouts

10 7/8”
7”

10’-0”
Figure 1.8 – Section of joint panel looking perpendicular to traffic

Figure 1.9 – Joint panel casting (left side cured for 1 night, right side ready for
casting on 2nd day)

1.3. Organization of Thesis

Brief overviews of the chapters contained in this thesis are presented in the

following paragraphs.

Chapter 1 describes the goals, motivation, and an overview of the design for the

Precast Prestressed Concrete Pavement project on I-57.

________________________________________________________________________

Page 10
 Chapter 2 provides a literary review of past concrete pavement projects.

Discussions on construction designs, prestressed concrete, and instrumentation projects

are also presented.

An overview of the experimental program which includes the instrumentation,

design and placement, and analyses of the thermal behavior of embedded instrumentation

are presented in Chapter 3.

An in-depth look at the service performance of the pavement system is presented

in Chapter 4. Thermal and strain gradients are presented for time windows which

facilitate discussion on characterization of the factors that affect pavement performance.

________________________________________________________________________

Page 11
 This page intentionally left blank.

________________________________________________________________________

Page 12
2. Background Information

2.1. Overview of Precast, Prestressed Concrete Pavement Technology

Precast, Prestressed Concrete Pavement technology is a new approach to

pavement design in a field that has seemingly tried all of the possible permutations for

optimization of pavement design. PPCP is specifically designed to address some of the

problematic areas of conventional cast in place pavement. Current design practices and

construction of precast pavement technology have largely been adapted from other

applications of prestressed and precast concrete in bridge design. The allure of PPCP

technology has benefited from decreased shipping costs and more significant political

and monetary benefits from expediting timely pavement projects in high traffic volume

areas. Although other prestressed technologies used frequently in bridge deck and

girders have proved useful in aiding with the design of PPCP, means of construction and

reliability are challenged by clients seeking to employ precast pavement. Pilot PPCP

projects have been sought by several DOT’s for experimentation to acclimate contractors

and evaluate the overall effectiveness as a design alternative to conventional pavement.

The pilot project using PPCP in Missouri was performed to address several issues

previous projects have neglected to address. The Missouri Project would evaluate the

effectiveness of PPCP as a design alternative in more intimidating climate and subjected

to heavy traffic volumes. The previously completed pavement projects have been located

in milder climates where the pavement was not subjected to rapid freezing and thawing

________________________________________________________________________

Page 13
and usage of de-icing salts. The performance of PPCP in a harsh environment will serve

as a testing platform for further evaluation.

PPCP allows constructors to perform rapid rehabilitation of dilapidated roadways

during off peak travel times and place the roadways back into service very rapidly.

Traffic congestion due to the presence of construction activities results in, among many

other variables, increased fuel consumption and lost work time, or user costs and safety

issues related to construction (Merritt 2001). By avoiding peak travel times for

construction, safety for workers and travelers is improved by limiting exposure to

construction areas. PPCP pavement design alternatives combined with improved safety

and timely construction present a clear set of benefits that can be utilized by project

managers to decide on the appropriate usage of PPCP. Descriptions of recent PPCP

projects are included in the following sections.

2.1.1. Design Considerations for PPCP

PPCP, like many other prestressed concrete applications, utilizes a pre-

compressive force to minimize amounts and strength of materials, which results in a more

economical design. Constructing an 8” prestressed pavement instead of a 12”

conventional pavement results in a savings of over 770 cubic yards of concrete per lane

mile. Other advantages are inherent with prestressed systems such as the ability to span

voids that develop underneath pavement due to many factors. These voids reduce the

support of conventional pavement and create highly stressed localized areas which may

reduce the life of the pavement under repetitive wheel loading (See Figure 2.1). Simply

________________________________________________________________________

Page 14
increasing the prestressing force will help the panel act like a thicker pavement (Merrrit,

McCullough et al. 2000).

Prestressing
Tendo ns

Precast Slab

Base with
Void Shown

Figure 2.1 – Illustration of pavement section spanning over void in base material

Another inherent benefit from precast pavement is the ability to control the

quality of the finished product. Timeliness is held paramount most often when placing

conventional pavement due to traffic, workers, and high equipment costs and other

constraints. These influences can weigh heavily on builders and can compromise the

quality of a finished product. When the pavement is fabricated in advance, more

effective quality control, lower tolerances in size and shape, and selective production

schedules can be used under controlled conditions since fabricators are not under these

constraints (See Figure 2.2).

________________________________________________________________________

Page 15
Figure 2.2 – Surface finishing of a typical base panel at the precasting yard

The controlled environment of a precasting yard also enables a more economical

means for controlling the delicate curing process. The panels can be cured in a number

of ways to minimize shrinkage effects and residual stresses. Further investigations on

casting procedures and curing effects are discussed in (Dailey 2006).

PPCP is very effective in mitigating serviceability problems such as cracking and

load transfer. Cracks can spall, fault, and allow water to penetrate the base creating voids

under the pavement and facilitate freeze-thaw damage. Crack widths are kept closed by

the elastic behavior of the prestressed system. In conventionally reinforced pavement

these cracks would open up wider with each successive freeze/thaw cycle and eventually

expose the reinforcement and base to water and de-icing salts. This quickly results in

degradation of the pavement or creates load-transfer problems meriting repair or

replacement. The pre-compression forces in PPCP serve to keep cracks from opening up

wide enough to create the aforementioned problems. The shear friction alone, provided

________________________________________________________________________

Page 16
by the pre-compression in a prestressed pavement, provides optimal load transfer across

joints and cracks (Merritt, McCullough et al. 2000).

Concrete poured on a base course will tend to have a rough underside, because it

takes the shape of the base course, thus increasing the friction of the bottom surface of

the pavement. Additionally, concrete poured onto the base will be restrained by a larger

mechanical means in which small slivers of concrete are allowed to seep in between the

aggregates causing small surface irregularities or “fingers” tying the concrete to the base.

When the concrete shrinks during curing, the restraint against the base causes residual

stresses in the concrete. These additional stresses are not typically included in design

calculations since the geometric constraints are not easily quantified. PPCP has two

advantages over CIP concrete since the majority of shrinkage will have occurred at the

casting yard and precast panels will have a smooth underside limiting mechanical friction

with the base. Fewer shrinkage cracks result in decreased maintenance and repair costs

and extend the lifespan of the pavement.

Base preparation is critical to the effectiveness of prestressing and ride quality

achieved after installation of PPCP. Polyethylene sheeting has been used to minimize

frictional losses between pavement slabs and the base. Not only do slight elevation

differences between panels create additional friction by slight miss-alignment of the post-

tensioning ducts but can also create an undesirable audible “bump” at panel joints.

Diamond grinding at select locations or the application of a smooth leveling course is

often required. Shear keys are cast into the joints of the pavement to ensure proper

vertical alignment and load transfer.

________________________________________________________________________

Page 17
 The performance of PPCP in this and other pilot projects has proven it to be a

viable substitute for repair and replacement as well as for new construction of pavement

systems. PPCP possesses several design features that make it attractive for

implementation on a wide variety of projects such as interchanges, approach slabs,

ramps, weigh-in-motion scales, un-bonded overlays, and temporary pavement crossovers.

PPCP sections for a ‘standard roadway crossover’ can be stockpiled and used at multiple

locations. Precast pavement panels can also be configured to accommodate unique

geometry, typical crowns, cross-slopes and super-elevation transitions. Similar types of

unique panel designs and alternative means for pavement system construction methods

were well received by precasters and contractors at the Missouri DOT/FHWA

showcasing workshop in August of 2006 following the completion of the Missouri Pilot

Project.

2.2. PPCP Projects in the United States

The development of PPCP for rapid rehabilitation projects began in the mid

1980’s. Projects utilizing CIP prestressed concrete in Texas and South Dakota have

proven effective. Since the pavement performed well in these states coupled with

advancements in precasting and shipping means facilitated the use of precast slabs. More

recently in Georgetown, Texas a frontage road along Interstate 35 was replaced using

precast prestressed concrete panels. The experience from the Texas project allowed

transportation officials to showcase the benefits of precast construction and demonstrate

the advantages of precast panels will serve in urban pavement replacement projects. The

following sections describe the recent PPCP projects undertaken throughout the country.

________________________________________________________________________

Page 18
2.2.1. Iowa Approach Slab on Highway 60

A challenging Iowa project successfully incorporated precast prestressed concrete

pavement for bridge approach slabs in September 2006. The charge of this project was

for Iowa DOT to refine design and construction details of PPCP bridge approach slabs.

This project was located on Highway 60 near Sheldon, Iowa. The bridge approach slab

tied into a 30 degree skewed, integral abutment for the northbound bridge crossing the

Floyd River. The twin southbound bridge was constructed with conventional cast in

place approaches to provide a direct comparison in performance during service. The

pavement panels were typically 14 ft x 20 ft x 12 inches with panels adjacent to the

abutment cast to match the skew. They were assembled in “lane-by-lane” construction

on a closed roadway for the new bridge. This project offered a unique bridge approach

slab design for IADOT by tying the approach slab into the abutment and moving the

expansion joint out to the end of the approach slab. Research by IADOT has shown that

removing the expansion joint near the abutment limited water infiltration and erosion of

embankment material around the abutment and minimized settlement.

Due to the success of the pilot project, IADOT has scheduled the replacement of

failed bridge approach slabs at either end of twin bridges under traffic with PPCP. The

end result will be a further understanding and confidence of this pavement system as an

alternative means of rehabilitation of bridge approach slabs and replacement of

deteriorated or poorly constructed paving notches on high volume roadways and bridges.

________________________________________________________________________

Page 19
2.2.2. I-10 in El Monte, CA

In April 2004 the California Department of Transportation (CalTrans) completed

a pilot project using PPCP on Interstate 10 near El Monte, CA. This project involved a

little more complexity compared with the Texas project by requiring varying cross-slopes

cast into the panels, and nighttime construction operation (Tyson and Merritt 2005). The

total length of roadway replaced was 248 ft and consisted of two driving lanes and a 10 ft

shoulder that were part of a widening project on I-10. The panels were cast with a flat

bottom and a variable depth to maintain a desired cross-slope to match the roadway

profile. The pavement was placed on a level 6 inch lean concrete base on top of 8 ½ inch

aggregate base. A total of 31 panels were fabricated for the project. The panels were

installed at a rate of 15 panels per 3 hours (Transtec 2009). The panels were prestressed

transverse to the direction of traffic and post-tensioned in two 124 ft sections longitudinal

to traffic (Tyson and Merritt 2005). The pavement widening project was completed

successfully and has generated interest for future applications.

2.2.3. Outer road near I-35 in Georgetown, Texas

The first large scale implementation of precast, prestressed pavement was

installed on an outer road near I-35 in a jointly funded Texas DOT and FHWA project.

The frontage road was located just north of Georgetown, TX. Full and partial width

panels were used for this project to test the feasibility of the two panel types. Both types

of panels were post-tensioned longitudinally while the partial width panels had additional

post-tensioning ducts in the transverse direction (See Figure 2.3). A total of 339 panels

were fabricated, of which 123 were full width and 216 were partial width. The full width

________________________________________________________________________

Page 20
panels were wide enough to accommodate two twelve foot lanes an eight foot outside

shoulder and a four foot inside shoulder. The partial width panels were 16 feet and 20

feet in width respectively. When placed, the centerline of the roadway matched with the

joint between the two panels. The Georgetown pilot project presented many challenges

for precast pavement implementation (Merritt 2001). Among the successes were

demonstrating that the match-casting is not necessary due to the rugged formwork used in

fabrication and geometric mechanisms such as the mating connections cast into each

panel that provide for alignment of the post-tensioning ducts in the field. Successful

implementation of PPCP technology is dependent on the constructability and flexibility

of contractors to develop new practices where standard details have not been developed.

Figure 2.3 – Partial width panel placement on Georgetown Frontage Road, TX

PPCP (Merritt 2002)

________________________________________________________________________

Page 21
 The pavement in Texas has been in service since March 2002 and no maintenance

related issues have been reported (Transtec 2009). Careful planning and willingness to

explore innovative means of precast implementation by the DOT, precast supplier, and

contractor contributed to the overall success of the project in Georgetown, Texas which

opened doors for pilot projects in California and Missouri.

2.3. Field Instrumentation of Concrete Projects

The relatively new age of Precast, Prestressed Concrete Pavement technology

compared with that of traditional pavement techniques creates a demand for the

demonstration of performance. The caveats that affect the design and performance of

various traditional pavement types are well known by designers through experience and

testing which dictate the designer’s evaluation for pavement design. Field measurement

and rigorous evaluation of pilot projects can supply quantifiable information regarding

PPCP performance. Strain, temperature, deflection, and durability monitoring are

effective means for evaluation and characterization of pavement performance and

validation of design. These types of investigations can prove useful for design engineers

in the following areas:

• Understanding thermal and frictional response to describe base and sub-

base interactions to evaluate proper base materials, thicknesses, loss of

support, and usage of bond breakers or friction reducers such as

polyethylene sheeting to maximize prestressing efficiency.

________________________________________________________________________

Page 22
 • Validate in-service performance of structural design by analyzing strain

response of the pavement under traffic loading as well as investigate

expected fatigue life and long-term durability.

• Provide an evaluation of stressing operations by monitoring prestressing

strands and concrete strain. This data can be used to monitor pavement

behavior during and immediately after construction to improve design and

construction methods.

________________________________________________________________________

Page 23
 This page intentionally left blank.

________________________________________________________________________

Page 24
3. Experimental Program

3.1. Field Instrumentation

The following sections provide an overview of the instrumentation and data

acquisition system used for the project. Design, specifications, construction and

calibration of custom instrumentation used for the research project are detailed in-depth

in (Dailey 2006).

3.1.1. Types of Embedded Instrumentation

Several types of instrumentation were used on this project to quantify and

characterize concrete behavior and develop an understanding of the prestressed pavement

system at different stages of construction. Direct measurement of concrete strain and

prestressed strand strain can be accomplished by embedded strain gages inside of the

concrete matrix and attached directly to prestressing strands during casting. The

following sections describe the instrumentation used to monitor the performance of PPCP

test sections.

3.1.1.1. Strain Gage Rebar

Typical #4, Grade 60 rebar was used to fabricate sensitive strain gage

instrumentation for embedment in the concrete pavement panels. Approximately 24 inch

long sections were cut and machined smooth in the center to accommodate a

temperature-compensating full-bridge circuit. Two gages were installed longitudinally

on the rebar and two were installed along the circumference. A schematic of the circuit

________________________________________________________________________

Page 25
used is illustrated in Figure 3.1. The ends of the strain gage rebars were threaded to

accommodate gripping for calibration. A completed strain gage rebar is shown in Figure

3.2. Strain gage instrumented rebars are capable of measuring very dynamic events such

as stress transfer from strand cutting and traffic loads while in-service. The instruments

are also very robust and have a good track record from previous projects completed at the

University of Missouri – Columbia.

S+
S- P+
P- P+

R1 R2

R2
R1

Eo S-
S+

R3 R4
4 V D.C.

R3 R4 Smooth Machined
Surface
P-

Ribs on
Reinforcing Bar

Figure 3.1 – Schematic of the strain gage configuration on the strain gage rebar
(Eatherton 1999)

Figure 3.2 – Instrumented Rebar showing installed strain gages

________________________________________________________________________

Page 26
3.1.1.2. Vibrating Wire Gages

Commercially available vibrating wire embedment type strain gages were used to

complement the dynamic strain gage instrumented rebars (Model 4200, Geokon Inc.).

Vibrating wire gages are reliable for long-term strain measurements due to the nature of

its design which does not depend on electrical resistance like a traditional strain gage.

The gage consists of a wire stretched between two flanges, an electromagnetic plucking

device, and a thermistor used for temperature compensation. The gage operation relies

on the change in resonant frequency of the wire based on its length. When one flange

displaces relative to the other, the wire is elongated resulting in a change in resonant

frequency. This change in resonant frequency can then be related to strain by simple

mechanics. The 6 in. gage is depicted in Figure 3.3.

The vibrating wire gages are very useful for long-term strain measurements;

however, dynamic events cannot be measured due to settling time of the stretched wire.

After embedding the gage in concrete, a zero reading can be taken. At any time the zero

reading can be referenced, and the state of strain of the concrete can be determined based

on changes in frequency from the base state (zero reading).

Figure 3.3 – Model 4200 vibrating wire gage from Geokon Incorporated

________________________________________________________________________

Page 27
3.1.1.3. Vibrating Wire Strandmeters

Model 4410 Vibrating Wire strandmeters were acquired from Geokon

Incorporated for use in the test panels to monitor prestressing strand strain. The gage

operates on the same principles as the model 4200 discussed above. However, clamps at

either end accommodate fixing to a prestressing strand. All strandmeters were

individually calibrated in-house to ensure predictable performance in the field. The

gages were encased in a PVC tube filled with grease in order isolate the gage from the

surrounding concrete and only measure strain in the post-tensioning strand. Blockouts

were cast into the test panels to accommodate the installation of strandmeters prior to

stressing, see Figure 3.4.

The strandmeters primary function was to measure the strain of the post-

tensioning strands during the stressing operations as well as maintain a record of strand

behavior during service. Strand stresses can be correlated with frictional and time-

dependent concrete losses to characterize pavement behavior and validate intended

design.

Figure 3.4 – Model 4410 vibrating wire Strandmeter, unsheathed

________________________________________________________________________

Page 28
3.1.1.4. Temperature Gages

Type T thermocouples utilizing a copper-constantan connection were used for

concrete temperature measurement. The specified temperature range was -328o to 663o F

(-200° to 900° C). The thermocouples were cut to length, welded using thermocouple

welders, and coated in epoxy at the University of Missouri – Columbia. This type of

temperature measuring device is very advantageous due to its robustness, ease of use, and

accuracy (+/- 0.1o C). Gages were positioned throughout the cross sectional depth, as

shown in Figure 3.5, and at various locations across the panel width and length.

TC

TC

TC

Figure 3.5 – Three thermocouples attached to fiber rebar coupled to post-tensioning

and pre-tensioning strands

iButtons manufactured by Dallas Semiconductors were also used to measure

temperature (Model DS1922L.) An array of twelve iButtons was used within one of the

precast panels as a pilot experiment to validate the welded wire thermocouples and to

provide measurement for higher resolution temperature performance through the depth of

the pavement. The iButtons store time and temperature logs in self contained memory

units and requires only a single lead wire to communicate with a computer or other data

________________________________________________________________________

Page 29
logging device. Lead wires were attached and the entire iButton coated in epoxy to

protect from corrosion and grounding inside of the wet concrete.

3.1.2. General Design and Placement Considerations

The precast panels used in the project are identified by two different methods.

The identification system used by CPI and Gaines Construction used letters and numbers

to signify the different panel types. An “A” panel was a joint panel, a “B” panel was a

base panel, and a “C” panel was an anchor panel. Since three different types of joint

panels were used, a number after the “A” differentiated the joint panels. Labels “A1” and

“A2” represented the joint panels at the north and south limits of the overall pavement

test section respectively. The symbol “A3” was used for the three intermediate joint

panels in the project.

To differentiate the instrumented panels from the non-instrumented panels the

MU research team added a number after the symbols used by the contractors. The panel

numbering increased from south to north. For example the four base panels were labeled

B1, B2, B3 and B4. The southern-most base panel was B1 and the northern-most was

B4. The single instrumented anchor panel was marked C1, and the joint panels were

marked A31 and A32 respectively.

The gages within the panels were further identified by their type and location.

Vibrating wire gages were marked with a V, instrumented rebar with an R,

thermocouples with a T, and strandmeters with an S. The location of the gage was

identified by a number after the type of gage. The seven instrumented panels

incorporated five different devices to measure strain and temperature of the concrete

________________________________________________________________________

Page 30
along with strain in the post-tensioning strands. Figure 3.6 depicts typical

instrumentation in a base or anchor panel and Figure 3.7 shows the instrumentation

locations in joint panel A32. Concrete strain was monitored using the strain gage rebars

and vibrating wire gages mentioned previously. Post-tensioning strand strain history was

measured by vibrating wire strandmeters. Temperature measurements were observed by

thermocouples and iButtons (Maxim) embedded in the concrete along with thermistors

incorporated inside of the vibrating wire gages.

Pretensioned Strands (8@1’-3”) Blockout for Strandmeter Junction Box

38’-0”

10’-0”
X3
X3

Inside Shoulder Post-Tensioning Ducts (18@2’-0”) Outside Shoulder

Instrumented Rebar VWG Thermocouple

Figure 3.6 – Typical instrumented base or anchor panel

________________________________________________________________________

Page 31
 Pretensioned Strands (3@0”-6” T&B) Junction Box
38’-0”

X2 X3 X3

10’-0”

Inside Shoulder Outside Shoulder

Post-Tensioning Ducts (18@2’)
Instrumented Rebar VWG Thermocouple

Figure 3.7 – Instrumented joint panel A32

3.1.3. Instrumentation Locations

The pilot project encompassed 1,000 feet of roadway rehabilitation and consisted

of four, 250 ft long post-tensioned sections. The primary goal of the research program

was evaluate the performance of the PPCP with regard to temperature, loading, local

strains, and joint displacements. To accomplish this, the research team decided to focus

on a single 250 ft section and instrument panels within this section. Section 3 of the 4

sections along the traffic direction was chosen. It was selected based on its proximity to

an AC power source and to limit possible transition effects from conventional concrete

pavements adjacent to the PPCP. Four base panels, two joint panels, and one anchor

panel were instrumented to characterize the performance of the pavement system under

environmental and vehicle loading. Figure 3.8 shows the location of the instrumented

panels within the chosen section. The panel marked B4 in Figure 3.8 lies outside the

third section and was instrumented for redundancy purposes.

________________________________________________________________________

Page 32
 A31 C1 B1 B2 B3 A32 B4

38’-0” Direction of
Traffic

Figure 3.8 – Overall view of test-section and location of instrumented panels (A

refers to a joint panel, B refers to a base panel, and C refers to a
anchor panel)

3.1.4. Data Acquisition System

A custom data-acquisition system was assembled and used for the monitoring of

embedded instrumentation. The data acquisition system consisted of a Campbell

Scientific CR10X data logger, (3)-32 differential AM416 relay multiplexers, 110V AC to

12V DC power supply, two AVW1 vibrating wire interfaces, and an NL100 network link

interface for remote communication. A centralized location at the test site was chosen to

minimize cable lengths and power requirements where the system was housed in an all-

weather signal cabinet typically used to house electronics at traffic signals (Figure 3.9).

________________________________________________________________________

Page 33
Figure 3.9 - Signal cabinet with main data-acquisition equipment installed at the
edge of right of way

Junction boxes were cast into the shoulder of instrumented panels to provide

quick-connect after panel shipment, house a cold-junction compensation circuit needed

for thermocouples and voltage regulation circuit for regulating power used by

instrumented rebars (Figure 3.10). Voltage regulation at the pavement was necessary due

to the voltage loss in the long lengths of cabling to connect instrumentation required due

to the large foot print of the project. Multi-pair, stranded wires were run from the

junction boxes of each panel back to the data acquisition system for signal transmission.

________________________________________________________________________

Page 34
Figure 3.10 - Junction box installed in blockout cast in outside shoulder of precast
pavement panels

3.1.5. Remote Monitoring

Proximity to a nearby exit on Interstate-57 where businesses were located enabled

the research team to gain access to high speed DSL and electricity. An inexpensive spur

line to get closer to the test section was installed by the utility companies for the research

team. Luxuries such as electricity and internet service are not commonly available on

remote instrumentation sites. Reliable electricity and commercial internet service

improved the reliability and simplified powering the data-acquisition systems and

instrumentation. Similar long-term monitoring projects performed by the University of

Missouri-Columbia located in remote areas required sophisticated, custom power saving

circuits accompanying solar panels and back-up generators due to the lack of utilities.

These challenges are best avoided if conditions allow and proper planning is allotted

during the early phases of site selection.

________________________________________________________________________

Page 35
 High-speed DSL enabled the research team to perform real-time data monitoring

from the laboratory. Researchers were able to upload customized programs to perform

specific tests catered to gather timeline and event specific data on selective

instrumentation. Typical programs were designed to perform instrumentation

diagnostics, high-frequency readings for wheel load response, monitor prestress levels,

and evaluate temperature-strain response during short-term (hours), medium-term (days

to weeks), and seasonal weather events. Customizable programs also benefited

instrumentation life and electricity savings by utilizing only the gages or thermocouples

needed for specific long-term experiments and powering down the remaining

instrumentation.

3.2. Laboratory Experiments

3.2.1. Materials Testing

A series of material tests were performed to gain a complete understanding of the

composition and behavior of the concrete and materials used in the PPCP system. This

data was used to understand pavement behavior resulting from temperature and vehicle

loading as well as generate long-term loss models by characterizing shrink, creep, and

relaxation to verify results measured in the field. Additionally, samples were prepared to

measure the chloride ion resistance which was used to predict pavement durability with

respect to salt penetration. Laboratory material testing and results were presented in

(Davis 2006).

The following is a summary of the specific tests that were performed.

• Uni-axial Compressive Strength Tests @ 7, 28, and 56 days

________________________________________________________________________

Page 36
 • Unrestrained Creep & Shrinkage

• Chloride Permeability Tests

• Freeze Thaw Tests

• Three-Point Flexure Tests

3.2.2. Thermal Investigation of Embedded Instrumentation

Fundamental understanding of how the instrumented rebars behave embedded in a

hardened concrete matrix is useful and necessary to analyze results from service

measurements.

In an effort to minimize temperature dependent effects, self temperature

compensating full bridge circuits were used on the instrumented rebars. In order to

understand how the embedded instrumented rebar system responds to thermal and

mechanical loads from restraint, laboratory experiments were performed using both

unrestrained instrumented rebars as well as instrumented rebars embedded in concrete.

One instrumented rebar and one vibrating wire gage were cast into a 6 inch x 6 inch x 24

inch (15.24cm x 15.24cm x 60.96 cm) long concrete specimen in order to duplicate the

response of an instrumented rebar embedded in the pavement. Another set of similar

instrumentation was supported by metal wires so as to eliminate any restraint to their free

movement. Both sets (embedded and unrestrained) of instrumentation were put in an

oven and subjected to programmed temperature histories. Figure 3.11 shows the

unrestrained instruments suspended by thin wire in the oven.

________________________________________________________________________

Page 37
Figure 3.11 – Unrestrained instrumented rebar and vibrating wire gage in
temperature controlled oven

Figure 3.12 shows the responses from the instrumented rebars during the heating

of a 14-day old concrete specimen to the temperature history shown in Figure 3.12a. The

instantaneous strain (approximately 1 µstrain/ºC) induced in the unrestrained instruments

are likely due to temperature dependent non-uniform strain gradients. After several days,

the output returned to zero, wherein the rebar had reached a uniform temperature.

Elevated temperatures were maintained for nearly 12 days to ensure the entire concrete

prism had reached a uniform temperature. The embedded rebar in Figure 3.12 took

several days before reaching expected magnitudes of strain based on theoretical

predictions of the 52ºC temperature excursion. The theoretical rebar strain was

calculated by multiplying the change in temperature (ΔT) by the difference in CTE of

concrete and steel, which equals -6.2 µstrain/ºC. The embedded rebar strain was

compressive in nature with an increase in temperature, which supports the logic used to

describe the embedded instrumentation performance. The magnitude of strain in the

embedded rebar reached higher values than computed from ΔT. This larger magnitude is

attributed to drying shrinkage, which was exaggerated by desiccation due to the high

________________________________________________________________________

Page 38
temperatures in the oven (for a relatively green concrete, 14 days old when experimented

was started). This was confirmed after the heat had been turned off. The resulting

magnitude of residual strain measured by the embedded rebar was equal to the difference

between the theoretical embedded rebar strain and actual embedded rebar stain

(approximately 100 µstrain of drying shrinkage strain). Concrete strain was calculated by

multiplying the scalar (-6 / 6.2), which was calculated from the difference in CTE’s of

concrete and the rebar. It can be seen in Figure 3.12 that after 11 days, the temperature

measured by the thermocouple (wrapped up with the embedded rebar, R17), had reached

room temperature rather quickly yet there was still strain recovery over the following

days that mimicked the unrestrained rebar signal.

________________________________________________________________________

Page 39
Figure 3.12 – (a) Temperature history (b) Strain history of embedded and
unrestrained rebar instruments

The concrete strain measurements during service are largely dictated by thermal

excursions from everyday heating and cooling. These analyses have led to the following

idealization and subsequent equation that was used to interpret the service performance

data of the instrumented panel sections.

o The concrete (6 µstrain/ºC) has a CTE of roughly half that of the embedded steel
(12.2 µstrain/ºC) instrument.
o The system experiences a +ΔT
o The concrete expands Δ, and the steel wants to expand 2Δ. Yet the steel is
restrained by the concrete to only Δ.

________________________________________________________________________

Page 40
 o Hence, the strain in the instrumented rebar is measured as a compressive Δ to a
heating of the system.

D
Steel Rebar

2D
Concrete

D
 
 × (R1 − R0 )
CTEconcrete
∆ε Concrete = 
 CTEconcrete − CTE rebar 

Figure 3.13 – Idealization of instrument response due to an increase in temperature,

ΔT

3.3. Challenges for Remote Data-Acquisition

Many challenges had to be addressed with respect to the service performance of

the instrumentation and data acquisition system developed for the precast project. These

challenges sometimes resulted in delays because of the need to undertake several repair

visits after severe weather-related events. In spite of more than adequate planning and

installation of safety systems, such extensive instrumentation with a network of electrical

conductors spanning over 9,000 square feet serves as an easy sink for electrical activity

during thunder storms. A summary of the various challenges and appropriate remedies to

mitigate each problem are listed here so as to be helpful for future projects of a similar

nature.

________________________________________________________________________

Page 41
3.3.1. Excessive Heat Build-up affecting Sensitive Hardware

The signal cabinet that housed the power supply, communications and data

acquisition system was located in a field just beyond the shoulder alongside the

instrumented pavement section. This box was exposed to fairly high ambient

temperatures (build-up of temperature in excess of 160°F during peak summer days was

measured). These excessive ambient temperatures and resistive heat build up in the

voltage regulators resulted in malfunctioning of the voltage regulation circuitry that

supplied DC power to the instrumented rebars. A “belt-and-suspenders” approach helped

alleviate this problem. First a roof was built to protect the signal box from direct sun

exposure (Figure 3.14). Secondly, two heavy-duty equipment fans were installed in the

signal cabinet that were on at all times and allowed flow of air through the cabinet.

Third, all the voltage regulators were replaced with military grade regulators that were

specified for higher operating temperatures. And fourth, all the regulators (those in the

signal cabinet as well as the individual panel terminal boxes) were provided with larger

aluminum heat dissipation fins. Collectively all of these four upgrades essentially

eliminated problems associated with excessive heat build-up and associated electronic

instabilities.

________________________________________________________________________

Page 42
Figure 3.14 – Signal cabinet protected during the heat of the day by a shade roof

3.3.2. Moisture Intrusion and Corrosion

For an instrumentation project such as this, it is necessary to build-in sufficient

moisture protection for the electronic components and associated circuitry. While the

terminal boxes embedded in each of the instrumented panels were specified to be

hermetically sealed, the holes that allowed instrumentation cables into the box, also

allowed moist air even while the holes were sealed with silicone. The circuit boards were

also mounted with sufficient clearance from the bottom of the terminal boxes using

spacer legs to avoid standing water from interfering with their intended operations. The

circuit boards were additionally sprayed with a non-conducting urethane spray to water-

proof them. The terminal boxes also contained desiccants in cloth bags. With adverse

weather, a significant amount of precipitation and infiltration of moist air, the electronics

in the terminal boxes were exposed to moisture and on rare occasions (in some panels) a

small amount of standing water (inside the terminal boxes). On a few occasions the

moisture shorted the printed circuit boards in these terminal boxes despite all the

protective actions and had to be repaired or replaced. Continued monitoring of these

________________________________________________________________________

Page 43
terminal boxes during inspection visits, cleaning, caulking and replacement of desiccant

bags helped mitigate the moisture problem.

3.3.3. Lightning Protection

Even while all cables were well shielded and grounded, lightening strikes tripped

the protective circuit breakers, damaged the uninterruptible power supply (UPS), several

voltage regulators and cold junction compensation (CJC) circuitry. While studying the

“as-implemented” circuitry to fix the problem, it came to light that while the circuit was

well protected against a lightning strike on the power pole side of the system, there was

little protection against voltage surges on the instrumentation side of the circuitry. Close-

up of a CJC circuit board damaged by a lightning strike is shown in Figure 3.15. Diodes,

which prevent the reverse flow of electricity, were employed in all of the instrumentation

lines and across all of the voltage regulators to ensure that any electrical surge would be

discharged to the earth ground. After observing that the most viable method to properly

ground the system for lightning affected the magnitudes of the outputs from the

instruments due to a ground loop differential, a spark gap was employed to prevent

adversely affecting the system during normal operating conditions and still provide

lightning protection when needed.

________________________________________________________________________

Page 44
Figure 3.15 – Close-up of CJC damaged by lightning

3.3.4. Snow Removal and Protective Plates

During the first winter (January 2006), snow plows sheared off the bolts securing

the protective covers of the terminal boxes that were slightly above the surface of the

pavement. This allowed water to get in to the terminal boxes. Use of thinner (but yet

sturdy to withstand traffic loads) protective cover plates with counter-sunk recesses

allowed enough clearance so that the bolt heads could be flush with the top surface of the

protective plates. This mechanical upgrade ensured that no repeat of such damage

occurred during the second winter (January 2007).

________________________________________________________________________

Page 45
 This page intentionally left blank.

________________________________________________________________________

Page 46
4. Service Performance of PPCP System

4.1. General Information

A summary of service performance results from the year long monitoring of the

instrumented pavement panels is presented in this chapter. Typical daily, short-term (few

days) and longer term seasonal temperature and strain excursions are presented and

discussed. Vehicular loading events were measured and are presented. An evaluation of

the causes and characterization of prestress losses are presented and discussed. Also

included are observations for visual inspections carried out during the regular field

inspections of the pavement test sections. Since the more consistent data from the

instrumented rebars were available for the majority of the project duration, they are used

primarily for figures and discussions. The vibrating wire data, when available, had been

used to confirm magnitudes of strain excursions measured using the instrumented rebars.

However plots for vibrating wire strain data which are typically noisy and intermittent

due to frequent malfunctions and over-range chipping are not included. As a

consequence these results are not presented in this chapter.

It should be noted that strain data presented here includes the combined effects of

thermal loading, vehicular loading, viscous loading due to creep, shrinkage and

relaxation, and loads due to sub grade movement. However, given the time windows of

interest and sample rates of data acquisition, the dominant influence is due to thermal

loading. The effect of vehicular loading is isolated in one plot where a significantly

higher data acquisition rate was intentionally used to highlight this effect.

________________________________________________________________________

Page 47
 Strain measurements were recorded using instrumented rebars and vibrating wire

gages. Due to adverse weather events, 5 of 12 vibrating wire gages were damaged during

the storms of spring 2006 and operated intermittently. The remaining vibrating wire

gages have operated consistently throughout the duration of the project. More details on

the challenges faced with the data acquisition system can be found in Section 3.3. It

should be kept in mind that electrical drift can affect long term measurements using

resistance type strain gages. Careful planning and design of the instruments and data

acquisition system kept electrical complications at a minimum compared to true strain

signal output. Daily outputs of strain gages were also subject to voltage fluctuations

caused by temperature changes and were adjusted.

In looking at the strain histories generally presented in this thesis it is important to

recognize that the zero strain reference at the start of each plot doesn’t represent actual

“zero strain” value but is a reference for incremental excursions shown. In other words,

negative strains do not necessarily mean compression but are merely less tension. This is

typical for strain-gage based transducers where “zeroing” long term measurements to

study incremental events is more important than studying actual strain magnitudes.

4.2. Pavement Thermal Behavior

4.2.1. Temperature Variations

Thermal loads constitute the single most important influence on pavement strains

observed on a daily basis. However, since the pavement does not have a uniform cross-

section, heating, heat-retention and cooling occur at different rates for different cross-

sections resulting in gradient effects. Proximity within the concrete matrix to both air
________________________________________________________________________

Page 48
 and ground work as heat sources and sinks. This enables the thinner sections to heat and

cool more quickly than the thicker section around the crown. Figure 4.1 highlights the

potential for differential heating and cooling using a typical summer day and night. Point

A on the pavement panel is at the outer edge of the thinner section, and during the

daytime it heats quicker than the rest of the panel that is exposed to the air due to the

thicker cross section and proximity to exposed surfaces. Point B is the last portion of the

concrete section to heat up during the day, and during the night is the last section to cool

down (Point C). Similar trends in reverse are anticipated during cooling cycles during

nights or winter related seasonal cold fronts.

2%
A D
Pavement Pavement
Cross-Section B C Cross-Section

Sub-Grade Sub-Grade

Figure 4.1 – Day and Night Cooling Trends

4.2.2. Daily Thermal Loadings

Pavement response, calculated from individual instruments, have a thermal

component plus effects of restraint due to sub-grade movements, thermal gradients, and

eccentric prestressing.

________________________________________________________________________

Page 49
 The magnitude of recorded concrete temperature and rate of thermal loading is

affected by proximity of the particular section to exposed surfaces and thickness at that

location. Figure 4.2b indicates the daily concrete response at three instrumented rebar

locations (R1, R3 and R4 see inset) that are parallel to the traffic direction. Temperatures

recorded close to the strain measurement locations are reported in Figure 4.2a.

Theoretical concrete strain history was predicted by averaging the three

temperature change measurements and multiplying by the CTE of concrete assumed to be

6 µstrain/ ºC. The locations of the instruments and thicknesses of corresponding

pavement are shown in Table 4.1 (depth of instrument from driving surface). The

magnitudes of the recorded temperatures are reflective of both the thicknesses of the

pavement and proximity to the nearest exposed surface. A32_V3t at R7 is closer to two

surfaces than R1 or R3. This may explain why the temperature excursion at R7 is the

highest. Higher temperature swings produced larger strain excursions, even at localized

panel locations. R7 has a ΔT of 14.4ºC but exhibits a strain differential of 107 µstrain.

If the response were attributed only to the temperature change, R7 would only indicate

86.5 µstrain. The difference between the expected thermal strain and the observed strain

excursion is 20.5 µstrain and is consistent with the strain excursions of R1 and R3 shown

in Figure 4.2. This difference can be attributed to the several additional constraints

described earlier.

________________________________________________________________________

Page 50
Table 4.1 – Instrument Locations and Event Summary for Joint Panel A32 on July 13,
2006

Pavement Instrument Temperature Daily Strain

Instrument
Thickness Depth Change Excursion
(inches) (inches) (celsius) (µstrain)

Rebar 1 8.2 4.2 8.8 83

Rebar 3 10.9 5.5 10.4 82
Rebar 7 9.3 4.8 14.4 107

The duration that the strain level is sustained at peak levels during temperature

extremes is correlated well with pavement thickness at the location. In Figure 4.2 , R1

peaks at 5 hours and starts to indicate less tension faster than R3 and R7. R3 and R7

exhibit similar thickness dependent strain response.

________________________________________________________________________

Page 51
 0 4 8 12 16 20 24
50
110
A32_V1t @ R1
A32_T5 @ R3
45 A32_V3t @ R7
103
Average (Ref. for Theo.)

Temperaturel, T (°F)
Temperature, T (°C)
Ambient Temp
40
96

35 89

30 82

25 75
Instrumentation Plan
R8 R5 R2
V4 R6
V2

20 T7-8 68
Inside Shoulder

Outside Shoulder
Traffic
100 T4-6
R1
Direction
R7 V3 R3 V1 T1-3
R4 A32_R1
Vibrating Thermocouple
A32_R3
Instrumented
Rebar Wire Gage A32_R7
Thermal Theo.

ΔR3 = 82
50
Strain, ε (µstrain)

ΔR1 = 83

ΔTH = 65
0

ΔR7 = 106

-50
0 4 8 12 16 20 24
Time, t (Hours)

Figure 4.2 – One day window from July 13, 2006 for Panel A32 (a) temperature
history (b) strain history

The relative strain response from longitudinal and transverse concrete strains was

also studied in relation to the daily thermal history. Figure 4.3a and Figure 4.3b shows

the response from instrumented rebars during a 24-hour window for a typical summer

day. Strains proportional to the local temperatures of 35 to 60 µstrain can be observed

for the rebars aligned along the traffic direction (longitudinal rebars R1, R3, and R7). R5

________________________________________________________________________

Page 52
which is located transverse to traffic indicates much larger tensile strains on the order of

90 µstrain. The tensile strains of R5 in Figure 4.3b are also sustained at the high

magnitudes of strain for much longer than peak strains sustained by the longitudinal

rebars. The higher tensile strains and durations are also illustrated for the transverse

rebar, R3 in panel B3 (Figure 4.4b). Similar responses for the transverse strain behavior

were recorded for a typical winter day and presented in Appendix A (See Figure A.1),

highlighting the fact that this behavior is not unique to significantly different average

ambient temperatures (30ºC for the summer day shown versus 8ºC for the winter day).

________________________________________________________________________

Page 53
 0 4 8 12 16 20 24
50
A32_T1 A32_T2
A32_T3 A32_T4 108
45 A32_T5 A32_T6
A32_T7 A32_T8 103
A32_V1t A32_V3t

Temperaturel, T (°F)
Temperature, T (°C)
A32_V4t Ambient Temp
40 98

93
35
88

30 83

78
25
Instrumentation Plan
R8
73
R6 R5 R2
V4 V2

20 T7-8 68
Inside Shoulder

Outside Shoulder
Traffic
125 T4-6 Direction
R7 R1 V1 T1-3
V3 R3 R4

Instrumented Vibrating Thermocouple

100 Rebar Wire Gage

75
Strain, ε (µstrain)

50

25

0
A32_R1 A32_R3
A32_R4 A32_R5
-25
A32_R7 A32_V1c
A32_V3c Theoretical
-50
0 4 8 12 16 20 24
Time, t (Hours)

Figure 4.3 – One day window from July 13, 2006 showing all instruments for Panel
A32 (a) temperature history (b) strain history

________________________________________________________________________

Page 54
 0 4 8 12 16 20 24
50
B3_T1 B3_T2
B3_T3 B3_T4
45 B3_T6 B3_T7 50
B3_T8 B3_V1t
Temperature, T (°C) B3_V2t Ambient Temp.

Temperature, T (°F)
40 30

35 10

30 -10

25 -30

Instrumentation Plan
20 -50

Outside Shoulder
Inside Shoulder

Strandmeter
R5 R2 R1
R4 R3
125 T7-8
V2
V1
Traffic
T1-3
B3_R2 T4-6 Direction

B3_R3
100 B3_R4 Instrumented Vibrating Thermocouple
Rebar Wire
B3_V1c
Theoretical
75
Strain, ε (µstrain)

50

25

-25

-50
0 4 8 12 16 20 24
Time, t (Hours)

Figure 4.4 – One day window from July 13, 2006 for Panel B3 (a) temperature history
(b) strain history

The larger magnitude of temperature excursions and slower rate of recovering in

the transverse direction can be attributed to the different levels of restraint provided by

the surrounding concrete with respect to the asphalt stabilized base. In the longitudinal

direction, the pavement is restrained globally by the adjacent post-tensioned panels. The

strain measured in the middle of the panel will be largely dictated by pure thermal

________________________________________________________________________

Page 55
 behavior of the concrete sections due to the high level of restraint from the heavy sections

adjacent to the panel. Whereas in the transverse direction, the level of restraint from the

surrounding concrete with respect to the ground is lower and can take on behavior more

indicative of the response from the asphalt concrete base. The CTE of the asphalt

concrete base is higher than that of concrete and likely retains its’ heat longer than the

concrete pavement due to the proximity to the ground and insulation from the pavement.

The difference in magnitudes of the transverse and longitudinal strains can also be

attributed in part to “curling” resultant from differential heating/cooling between the top

and bottom of the concrete pavement. Curling in the transverse direction is likely to be

more than that in the longitudinal direction again due to levels of restraint.

The effect of daily temperature excursions can also be observed by looking at the

performance of a typical joint panel during the day. Figure 4.5 (Left) shows the silicone

based joint compound receding below the pavement surface during lower temperatures.

During the hottest times of the day the joint compound squeezes above the pavement

surface and appears to be damaged by vehicles passing over it (Right).

12 inches

Figure 4.5 – Joint Panel A31 during mild temperatures (Left) and hot temperatures
(Right)
________________________________________________________________________

Page 56
4.2.3. Weekly Thermal Behavior

Temperature and strain variations from a medium sized window were analyzed to

correlate pavement behavior with extended temperature excursions. Figure 4.6 illustrates

the temperature history and associated strain response for a five-day window of a base

panel in late September (2006). As expected, with increasing temperatures, the pavement

exhibits tensile strains. Cooling produces compressive strains. The responses of the

individual rebars are similar to the theoretical values. The temperatures for the

theoretical predictions are averaged values between two thermocouples closest to the

instrumented rebars. The magnitude of strain recorded by Rebar 2 (thickness of

pavement = 8.2 inches) are larger than that of Rebar 4 (thickness of pavement = 11

inches), in part due to Rebar 2 being located in a thinner cross-section of concrete. The

concrete strain behavior mimics what is to be expected for a moderate heating trend for

the entire duration. The day-to-day localized strain behavior is predictable with having

temperature histories.

The rate at which the temperature in the pavement increased or decreased was

observed to largely be a function of the proximity of the specific location to an exposed

surface and location-specific thickness of the pavement, as seen in Figure 4.6a.

Thermocouples 1 and 4 (T1 & T4) had a faster rate of heating and cooling since they

were located approximately 2 inches from the top surface of the pavement. It is for this

reason that they measured the hottest and most cool temperatures from day to day. The

rest of the thermocouples were located at mid-depth or at the bottom 1/3 of the cross-

section. As the location of the temperature measurement gets closer to the crown and

deeper in the pavement, smaller magnitudes of changes in temperature were observed as

________________________________________________________________________

Page 57
 shown in Figure 4.6a. The convention used was as follows: T1 closest to the top surface,

T2 in the middle, T3 at the bottom, T4 at top, T5 close to bottom. The exact locations

and thicknesses of the pavement at specific instrument locations are presented in

Appendix B. Concrete strain, as denoted by the individual devices (R2, R4) is largely

proportional to the magnitude of temperature at that cross-section of pavement.

0 20 40 60 80 100 120 140

45
C1_T1 C1_T2 110
C1_T3 C1_T4
40 C1_T5 C1_St
C1_V1t Ambient Temp
100

35
Temperature, T (°C)

Temperature, T (°F)
90
30

80
25

70
20

15 Instrumentation Plan 60
Outside Shoulder
Inside Shoulder

Strandmeter
R5 R2 R1
10 R4 R3 50
T4-5 V1
V2 Traffic
T1-3 Direction

80 Instrumented Vibrating Thermocouple

Rebar Wire

60 C1_R2
C1_R4
40
Theoretical
Strain, ε (µstrain)

20

-20

-40

-60

-80
0 20 40 60 80 100 120 140
Time, t (Hrs)

Figure 4.6 – Measured concrete strains in pavement at a short-termed window (a)

temperature history (b) strain history
________________________________________________________________________

Page 58
 Analysis of results using time windows of different lengths facilitated study of

seasonal variations and associated performance of the pavement section. Figure 4.7a

shows a moderate heating trend in mid July, including the movement of a cooling front

on Day 11 where the mean temperature drops approximately 5ºC (9ºF). It is readily

apparent in Figure 4.7a that the pavement temperatures are higher than the ambient air

temperature. The temperature in the pavement stays well above the low temperature at

night due to the ground retaining much of its heat. This effect is clearly noted on Day 11

in Figure 4.7 due to the cold front moving and the lowest temperature in the pavement

was still above the hottest air temperature of the previous day. Much larger strains were

measured by the instrumented rebars compared to the theoretical values, which were

validated by comparing with the strain response from the vibrating wire instruments.

This is in part due to the only location that temperature was measured in the panel was

located near the thickest portion, where the change in temperature is the lowest

throughout the day. It was from this temperature measurement the theoretical response

was predicted. Larger temperature variations are prevalent in the thinner sections which

resulted in larger changes in strain over the individual days. This relationship is indicated

in Figure 4.8.

________________________________________________________________________

Page 59
 0 4 8 12 16
45
B1_V2t 109

40 Ambient Temp

99

Temperature, T (°F)
Temperature, T (°C)
35

89
30

79
25

20 69
Instrumentation Plan

Outside Shoulder
Inside Shoulder

Strandmeter

15 R5
R4 R3
R2 R1 59
T4-5 V1
V2 Traffic
T1-3 Direction
140
B1_R2 Instrumented Vibrating Thermocouple
120 B1_R3 Rebar Wire

B1_R4
100 B1_R5
B1_V2c
80 Theoretical
Strain, ε (µstrain)

60

40

20

-20

-40

-60
0 4 8 12 16
Time, t (Days)

Figure 4.7 – Medium window indicating weekly heating and drastic cold front with
associated concrete strains (a) temperature history (b) strain history

________________________________________________________________________

Page 60
 0 24 48 72 96 120 144
40
120
Temp @ R2
Temp @ R4
35 Ambient Temp. 110

100
Temperature, T (°C)

Temperature, T (°F)
30

90
25
80
20
70

15
60
Instrumentation Plan

Outside Shoulder
Inside Shoulder

Strandmeter
10 R5
R4 R3
R2 R1 50
T7-8 V1
V2 Traffic
T1-3 Direction
T4-6
60
Instrumented Vibrating Thermocouple
Rebar Wire

30
Strain, ε (µstrain)

-30
B3_R2
B3_R4
B3_V1c
Theoretical
-60
0 24 48 72 96 120 144
Time, t (Hrs)

Figure 4.8 – Weekly instrument history for Panel B3 (a) temperature history (b)
strain history

4.2.4. Seasonal Variations in Panel and Global Pavement Responses

Analyses were performed to understand the effect of seasonal variations in

addition to the earlier discussed daily and weekly temperature fluctuations. By filtering

out hourly variations in temperature and strain it was possible to highlight long term and

seasonal warming/cooling trends. This was accomplished by comparing mean daily

________________________________________________________________________

Page 61
temperature and strain values (average of 24 hourly data points during each day for each

transducer). Figure 4.9 shows a six month window from early October of 2006 to the end

of April, 2007. This window of time represents the longest duration without significant

interruptions in the data acquisition system. Other windows of time show comparable

trends, even if there were frequent weather-related or equipment-related outages and

changes in the data acquisition programs to monitor different sets of instrumentation. The

plots in the following figures highlight typical winter cooling (October through February)

and typical spring heating (February through April) trends and associated strain histories.

There are no data for a small period in late November, when the data acquisition system

was down due to a power outage. The temperature history in Figure 4.9a includes both

mean daily ambient temperature as well as mean daily pavement temperature for Panel

B2. It can be observed that the excursion of mean pavement temperature is smaller than

that observed for the ambient temperature reflecting the time delay in heating and cooling

the concrete pavement and sub-grade mass. As in earlier discussions, it should be noted

that the strain plot provides incremental strain history during the time window of interest

and does not represent actual strain magnitude (which is less important to the discussions

here).

In Figure 4.9 an overall compressive trend (reducing strain magnitude) can be

observed as the mean daily temperature drops during winter and similarly a tensile trend

(reduced compression) can be observed when the mean temperatures rise during spring.

The theoretical concrete strain (αΔT) is calculated assuming a CTE of 6 µstrain/ºC (using

the average change in temperature recorded by all thermocouples in the panel).

________________________________________________________________________

Page 62
 10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
40
Mean Daily Ambient Temp
96
B2 Mean Daily Temp
30

76
Temperature, T (°C)

Temperature, T (°F)
20

56
10

36
0

16
-10 Instrumentation Plan

Outside Shoulder
Inside Shoulder

Strandmeter
R5 R2 R1
R4 R3
V1
-20 T4-5
V2 Traffic -4
T1-3 Direction

Instrumented
`
Vibrating Thermocouple
Rebar Wire
100

50 Mean Daily Concrete Strain from B2_R2

Mean Daily Concrete Strain from B2_R5
0 Theoretical
Strain, ε (µstrain)

-50

-100

-150

-200

-250

-300

-350
10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
Time, t (Days)

Figure 4.9 – Six month window showing longitudinal concrete strains in Panel B2 at
different locations (R2 and R5) (a) temperature histories (b) strain
histories

Rebars 2 and 5 (see panel inset in Figure 4.9 for rebar locations) both exhibit

expected trends in strain histories given the thermal loading history. Both of these

instrumented rebars are located where the pavement thickness is comparable (8.5 inches),

and are located at similar heights (4 inches from the bottom). However the magnitudes of

________________________________________________________________________

Page 63
strains at the two locations are significantly different for the same mean daily temperature

drop of approximately 26°C (peak strain differential of approximately 140 µstrain for

Rebar 5 as opposed to 280 µstrain for Rebar R2 from October 2006 to February 2007).

Local sub-grade friction and panel-specific in-plane bending effect due to use of steel

wedges along the outer shoulders (and resultant non-uniform panel to panel contact) are

likely reasons for variations in strain magnitudes between locations R2 and R5.

While simplistic and idealized this prediction captures the essence of trends in

strains from thermal loads. However magnitudes of strains are predicted inadequately. It

should be noted that in addition to thermal loads, the strain histories in plots like that

shown in Figure 4.9 are also influenced, in a location specific manner, by several other

factors including: elastic (modulus) and thermal (CTE) mismatch between pavement and

sub-grade and resultant sub-grade friction, restraint due to an improperly performing

joint, local thermal variations (differential thermal gradients due to differences in local

exposure/dissipation conditions and due to different pavement thicknesses) and changes

in prestressing force due to thermal effects. Relatively negligible influences can also be

attributed to creep and shrinkage of concrete, relaxation of prestressing steel, traffic loads

and strain gradients from bending. Figure 4.10 shows longitudinal strain histories

measured from different instrumented panels using instrumented rebars along the

passenger side wheel path of the right lane (see inset showing measurement location R2).

________________________________________________________________________

Page 64
 10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
40.0
Mean Daily Ambient Temp 96
Mean Daily Concrete Temp @ R2 Locations
30.0

76
Temperature, T (°C)

Temperature, T (°F)
20.0

56
10.0

36
0.0

Instrumentation Plan 16
-10.0

Outside Shoulder
Inside Shoulder

Strandmeter
R5 R2 R1
R4 R3
T4-5 V1
-20.0 V2 T1-3
Traffic
Direction
-4

Instrumented Vibrating Thermocouple

Rebar Wire
100
Mean Daily Concrete Strain from B2_R2
Mean Daily Concrete Strain from B3_R2
50
Mean Daily Concrete Strain from B4_R2
Mean Daily Concrete Strain from A32_R1*
0 Theoretical
Strain, ε (µstrain)

-50

-100

-150

-200

-250
* - Location R1 in A32 is
the same as Location
-300 R2 in the other panels

-350
10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
Time, t (Days)

Figure 4.10 – Six month window showing longitudinal concrete strains at identical
panel location (R2) in different instrumented panels (a) temperature
histories (b) strain histories

It should be noted that location R1 in joint panel A32 is identical to location R2 in

all base panels (B2, B3, and B4). It is interesting to observe that strain magnitudes

monitored in Figure 4.10b progressively decrease from panel B2 to B3 to B4 and to A32

(from the base panels in the middle of the test section to the joint panel at the end of the

test section). While not conclusive, the consistent and progressive reduction of peak
________________________________________________________________________

Page 65
strain magnitudes (around February 2007) suggests that sub-grade friction may have

some influence on this behavior. It is also important to observe from this figure, that

when temperature returns back to the initial value after approximately 6 months, the

differences in strain magnitudes in the various panels do not vanish, suggesting the effect

that causes peak strains in these panels to be different is not elastic (friction is an inelastic

phenomenon, unlike small thermal expansion/contraction due to seasonal temperature

excursions).

Figure 4.11 shows transverse strains measured from different instrumented panels

using instrumented rebars near the crown (see inset showing measurement location R3).

It should be noted that locations R4 and R5 in joint panel A32 are identical in the

transverse plane to location R3 in the base panel. The transverse strain histories across

the joint in the joint panel (R4 versus R5) are very similar. The magnitudes (150 – 250

µstrain) of peak strain events for transverse direction are comparable to those in the

longitudinal direction (Figure 4.9 and Figure 4.10). Again the magnitudes of

compressive strains during the coldest portion of the year are slightly larger than

predicted by only thermal behavior suggesting an accompanying change in prestressing

force or external frictional characteristic takes place during seasonal variations. The

difference in strain magnitude is largest during the middle of winter (February) and

decreases as the mean temperatures rise during early spring (April) back to levels similar

to those recorded in fall (October). However inconclusive, this strain recovery is more

elastic than recorded in the longitudinal direction suggesting it may be due to changing

prestress levels and eccentric bending effects caused by uneven stress distributions rather

than sub-grade friction.

________________________________________________________________________

Page 66
 10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
40.0
Mean Daily Ambient Temp 96
Mean Daily Concrete Temp @ R3 Locations
30.0

76
Temperature, T (°C)

Temperature, T (°F)
20.0

56
10.0

36
0.0

Instrumentation Plan 16
-10.0

Outside Shoulder
Inside Shoulder

Strandmeter
R5 R2 R1
R4 R3
T4-5 V1
-20.0 V2 T1-3
Traffic
Direction
-4

Instrumented Vibrating Thermocouple

Rebar Wire
100
Mean Daily Concrete Strain from B3_R3
Mean Daily Concrete Strain from A32_R4*
50
Mean Daily Concrete Strain from A32_R5*

0 Theoretical
Strain, ε (µstrain)

-50

-100

-150

-200
* - Location R4, R5 in A32
is the same as Location
-250 R3 in Panel B3

-300
10/2/06 11/2/06 12/3/06 1/3/07 2/3/07 3/6/07 4/6/07
Time, t (Days)

Figure 4.11 – Six month window showing transverse concrete strains at identical
panel location (R3) in different instrumented panels (a) temperature
histories (b) strain histories

Even while there was initial speculation that long-term drift in strain readings

from instrumented rebars might significantly affect strain histories, it is clear from Figure

4.12 that this is not the case. Strain history from instrumented rebar at location R1 in the

joint panel A32 is compared with similar history from the vibrating wire gage at the same
________________________________________________________________________

Page 67
 location for a two month window during December 2006 – February 2007. The strain

histories are nearly identical.

12/21/06 1/21/07 2/21/07

40
Mean Daily Ambient Temp
96
A32 Mean Daily Temp
30

76
Temperature, T (°C)

Temperature, T (°F)
20

56
10

36
0

Instrumentation Plan 16
-10
R8 R5 R2
V4 R6
V2

T7-8
Inside Shoulder

Outside Shoulder
-20 Traffic
-4
T4-6 Direction
R7 R1 V1 T1-3
V3 R3 R4

Instrumented Vibrating Thermocouple

150 Rebar Wire Gage

Mean Daily Concrete Strain from A32_R1

100
Mean Daily Concrete Strain from A32_V1
Theoretical
50
Strain, ε (µstrain)

-50

-100

-150

-200
12/21/06 1/21/07 2/21/07
Time, t (Days)

Figure 4.12 – Two month window showing longitudinal concrete strains at identical
panel location (R1-V1) measured using instrumented rebar R1 and
vibrating wire gage (V1) (a) temperature history (b) strain history

________________________________________________________________________

Page 68
 The effect of service temperatures on prestressing force is also of interest from a

performance point of view. Figure 4.13 includes a plot of temperature (ambient and

pavement temperature at crown at mid-height where the post-tensioning strandmeter

monitored is located) and associated strandmeter strain history recorded in Panel C1. If

the post-tensioning strand was un-bonded, one would expect strand strain to decrease

with a decrease in pavement temperature due to elastic shortening of the pavement

section. However, since the post-tensioned strands are grouted, they behave as if they

were bonded, with a decrease in temperature producing tensile strains in the strand

instead (Figure 4.13b, due to prestressing steel which has a higher CTE being restrained

by concrete with a lower CTE – thus producing compression in concrete and tension in

steel for the incremental temperature event). Notice that Figure 4.13b shows actual

strandmeter strain magnitudes (i.e. uses the actual zero strain reference from the start of

the post-tensioning operations, rather than a dummy “zero strain reference” to highlight

effect of the temperature event alone). The loss in prestress force from when the post-

tensioning operations were completed includes losses due to initial elastic shortening,

friction, creep, shrinkage and relaxation. These losses were quantified in Brent Davis’s

companion thesis.

________________________________________________________________________

Page 69
 10/2/06 11/2/06 12/3/06
30.0 86
Mean Daily Ambient Temp
Mean Daily Concrete Temp @ Crown Mid-depth
74
Temperature, T (°C) 20.0

Temperature, T (°F)
62

10.0 50

38

0.0
26

Instrumentation Plan

Outside Shoulder
Inside Shoulder

Strandmeter
-10.0 14
R5 R2 R1
R4 R3
T4-5 V1
V2 Traffic
T1-3 Direction

6250
Instrumented Vibrating Thermocouple
Rebar Wire

6150
Strain, ε (µstrain)

6050

5950

5850
Mean Daily Strand Strain from SM_C1

5750
10/2/06 11/2/06 12/3/06
Time, t (Days)

Figure 4.13 – Strandmeter response at center of 250’ test section during typical
winter-time temperature excursion (a) temperature histories (b) strain
histories

4.3. Pavement Response due to Vehicular Loading

A data acquisition program was designed to isolate vehicular traffic strain from

other long-term influences such as temperature, creep, shrinkage and restraint effects.

Data was acquired at significantly higher acquisition rates (12 Hz per channel). This

________________________________________________________________________

Page 70
 gives a least count of 0.08 seconds since a large amount of data is acquired in a short

time, the total acquisition window was reduced to approximately 30 minutes.

Simultaneous to automated acquisition of data from the instrumented rebars, visual

observation of the traffic history was also recorded so that correlations could be made of

strain peaks in the response. An unrelated lane closure (right, outside lane) facilitated

visual monitoring of traffic. Traffic speeds were limited to 55 mph as a result of this lane

closure. The rebar strain response due to traffic loads on the driver-side wheel path of the

inside, left lane is illustrated in Figure 4.14. Since this experiment was performed in the

afternoon, the overall compressive trend (negative slope in the global response) seen in

Figure 4.14 is a result of the heating of the pavement.

Traffic Response
Instrumentation Plan

Outside Shoulder
3
Inside Shoulder

T7
R5 R2
See Figure 4.15 R4 R3 R1
V1
2 T4-6
V2 T1-3
Traffic
Direction

1 Instrumented Vibrating Thermocouple

Strain (mstrain)

Rebar Wire
0
-1
Strain from
-2
-0.75 ºC ΔT
-3 (4.5 µstrain)
-4
-5
Truck traffic loadings
-6 (1 - 3 µstrain)
-7
0 10 20 30
B4_Rebar 5 Time (Minutes)

Figure 4.14 - Traffic strain (rebar response) in the pavement at crown

Passenger vehicles were undetectable with respect to the +/ - 0.35 µstrain level of

noise within the signal of the instrumented rebars. Figure 4.15 displays the concrete

response at the crown of the pavement for a selected duration at which visual vehicle

________________________________________________________________________

Page 71
 count was also undertaken. Strains induced by tractor trailers on the pavement, which

make up approximately 1/3 of the vehicles on I-57, can be seen in Figure 4.15.

Traffic Response

1.5
111 seconds 25 Tractor trailers passed
during this duration (25 peaks)
1

0.5
Strain (mstrain)

-0.5

-1

-1.5
B4_Rebar 5
-2
1.5 3 4.5 6 7.5
Time (Minutes)

Figure 4.15 – Duration of traffic response that was verified visually

Figure 4.16 illustrates the individual tractor trailer response on the pavement at

the crown. The tractor trailer that passed the precast panel being recorded had four axles;

one at the front of the tractor, two at the back of the tractor, and one at the rear of the

trailer. The three compressive peaks at 111 seconds suggest a correlation to the

individual axles passing over the instrumented rebar. The two, rear tractor axles were

likely encompassed in the second compressive peak due to their proximity to each other.

The tensile peaks are the result of stress caused by the approaching axles that are not

quite directly over the instrumented rebar. It is useful to note that strain magnitudes from

truck traffic are typically under ± 3 µstrain, compared to strains of ± 6 µstrain for a ±1 ºC

excursion in pavement temperature. It is important to note that these strain readings were

taken near the neutral axis of the prestressed panels. Assuming a linear gradient, surface

________________________________________________________________________

Page 72
 strains (extreme fiber) caused by traffic loading are likely less than 25 µstrain which is a

small portion of the total daily strain behavior characterized largely by thermal

expansion/contraction.

Traffic Response

1.5
1
Strain (mstrain)

0.5
0
-0.5
-1
-1.5
B4_Rebar 5
-2
108 110 112 114
Time (Seconds)

Figure 4.16 – Resulting concrete response from a tractor trailer passing over

4.4. Effective Post-tensioning Stress Distributions

4.4.1. Post-tensioning Stress Distributions Affected by Poor Transfer at Joints

Prestressing provides additional bending resistance in critical areas for service

loadings, confinement and crack control. Post-tensioning also locks the panels together

globally, helping with load transfer and increasing the stiffness of the pavement sections.

This section describes the level of post-tensioning stresses and distribution as indicated

by the embedded instrumented rebars. Figures of strain distributions across the width of

the panels will aid in investigation of poor prestress transfer.

Comparisons throughout the section can be readily made between the various

panels because Rebar 1, 2, 4, and 5 are in the same cross-sectional locations for Panels

________________________________________________________________________

Page 73
C1, B1, B2, and B3. The predictions are shown with expected steps from each pair of

strands that are stressed. These steps are also visible in most of the measured data.

Figure 4.17 illustrates concrete behavior of a typical base panel near the center of

the 250 ft test section during post-tensioning. The theoretical response predicted was

used to compare with the average concrete strain and not the individual rebar responses,

whose magnitude is affected by the thickness of the panel at the measurement. By

comparing local strain magnitudes at the four locations along the width of the panel, an

idealized sense of where force is transferred can be developed. Lowest strains were

measured at Rebar 2, then slightly higher at Rebar 4, with the largest stress transfer

taking place at the inner and outside shoulders. The poor prestress distribution of this

interior panel is likely resultant of several factors that occurred during fabrication and

construction as follows.

1. Wooden and Steel shims were used during construction along the outside
shoulder (near R1) to correct global alignment problems.
2. The epoxy used to lubricate the edges during the placement and seal off the joint
was allowed to harden prior to post-tensioning.
3. During construction, feeding the post-tensioning strands was difficult because the
PT ducts were thought to have sagged during fabrication. Due to localized
frictional effects, the concrete prestress force near those sections would be
affected.
4. Alignment of the post-tensioning ducts may have been difficult to obtain due to
no transverse “keyway” or other physical system to align the panels.

Overall post-tensioning levels are obtainable, as represented by the average

concrete strain vs. the theoretical strain behavior in Figure 4.17, which was derived from

a simple mechanics approach. However, the durability and susceptibility of cracks

________________________________________________________________________

Page 74
 opening in localized concrete regions with lower than desired prestressing may come into

question in the future. Pre-tensioning will aid in the flexural capacity of these regions

and the resistance of surrounding stiffer concrete should minimize any serious

susceptibility to crack openings. More often than not for the instrumented base panels

near the center of the test section, the post-tensioning strain levels near the R2 location

were less than desired, as shown in Figure 4.18. Panel B3 indicates expected prestress

levels since it is the closest to the passive jacking end and does not incur the effects of

cumulative prestress loss along the length due to the aforementioned problems and also

frictional losses, which are discussed in the following section.

20

0
Strain, ε (µstrain)

-20

-40 B1_R1
B1_R2
B1_R4
B1_R5
-60 B1 Average
Theoretical

R4, S R2
R5 R1
-80
10 9 8 7 6 5 4 1 2 2 3 4 5 6 7 8 9 10
Post-Tensioning Stress Transfer Sequence and Rebar Locations
-100
0 10 20 30 40 50 60 70 80
Time, t (Min)
Figure 4.17 – Concrete strain for a typical base panel during post-tensioning
operations

________________________________________________________________________

Page 75
 20
Section 3
10 1
A31 C1 B1 B2 B3 A32 B4
Strain, ε (µstrain) 0 2
3
4
-10 5
6
-20
C1_R2 7
B1_R2
-30 8
B2_R2
B3_R2 9
Theoretical 10
-40
R4, S R2
R5 R1
-50
10 9 8 7 6 5 4 1 2 2 3 4 5 6 7 8 9 10
Post-Tensioning Stress Transfer Sequence and Rebar Locations
-60
0 10 20 30 40 50 60 70 80
Elapsed Time from (Min)
Figure 4.18 – Concrete strain measured at the R2 location (beginning of outside
shoulder)

4.4.2. Post-tensioning Stress Losses due to Friction

Measured levels of post-tensioning strains were also affected by several frictional

phenomena. It is difficult to extract trends that could be allocated to each type of

characteristic loss, but an overall behavior is understood. Levels of post-tensioning

stresses transferred to the panels are inhibited by sub-grade friction and friction within

the post-tensioning ducts. This is where alignment and possible sagging of the post-

tensioning ducts become critical to the effectiveness of prestressing.

Figure 4.19 illustrates the average concrete strain response of the four

measurements in each panel for the five panels measured in Section 3 during post-

tensioning. There does not exist a nice trend that coincides with frictional losses,

primarily because of localized frictional characteristics with the sub-grade (partial regions

that were filled with sand where the crane had left ruts provide different resistance than
________________________________________________________________________

Page 76
 the asphalt base), localized stress distribution inhibitors such as the shims that were

placed between the panels, and the result of only having taken a small sample (four)

measurements in each panel. Most importantly Figure 4.19 reinforces that post-

tensioning forces are possible to obtain with the implemented design save for working

out a few of the minor kinks in the construction and fabrication. Other frictional prestress

loss analyses were performed in (Davis 2006).

10
Section 3
0 A31 C1 B1 B2 B3 A32 B4

-10
Strain, ε (µstrain)

-20
C1 Average
B1 Average
B2 Average
-30 B3 Average
A32 Average
Theoretical
-40
R4, S R2
R5 R1
-50
10 9 8 7 6 5 4 1 2 2 3 4 5 6 7 8 9 10
Post-Tensioning Stress Transfer Sequence and Rebar Locations
-60
0 10 20 30 40 50 60 70 80
Time, t (Min)
Figure 4.19 – Average Post-tensioning concrete strain histories for instrumented
panels

4.5. Transverse and Longitudinal Cracking

Cracking and visual degradation surveys were performed during inspections in the

latter half of the 18-month service monitoring. Both longitudinal and transverse cracks

have developed in the precast panels. Figure 4.20 and Figure 4.21 show a longitudinal

________________________________________________________________________

Page 77
 crack that originated at a corner of a strandmeter block-out. The crack spans several

panels suggesting possible influence from external loading such as traffic loads and sub-

grade movements. The crack has likely propagated since that lane receives the load from

the drivers side wheels of the majority of vehicles and has followed the post-tensioning

duct where a reduced cross-sectional area exists. Figure 4.20 (right) shows the outline of

the crack that traverses several base panels.

Path of the longitudinal crack

12 inches

Figure 4.20 - Longitudinal crack in driver side wheel lane

________________________________________________________________________

Page 78
 B2 BASE BASE BASE BASE

Blockout

TRAFFIC DIRECTION

Figure 4.21 - Schematic of one longitudinal crack

Longitudinal cracking have also been documented along the shoulders of the

precast panels. Some transverse cracks are located on an observed schematic in Figure

4.22. Longitudinal cracking along the shoulders may also be exaggerated by shims used

between panels to correct alignment during construction and limited by perimeter rebars

designed for edge reinforcement. Transverse cracking has primarily occurred at the mid-

section of the panels. The scaling noted in Figure 4.22 is located on the inside shoulder

where the pavement was not ground smooth. An official crack survey has been

conducted at several times of the study by MoDOT to locate, measure the length, and

record the width of the cracks. It was found that the majority of the cracks are “hair line”

cracks.

________________________________________________________________________

Page 79
 Scaling Inside Shoulder

Grout
Ports

Outside
Shoulder

Figure 4.22 – Typical crack locations of a 4-panel set of the 3rd test section on
May 9, 2007

4.5.1. Expert Task Group Meeting in Sikeston, MO

An FHWA Expert Task Group Meeting was held in Sikeston, MO in August of

2006. Departments of Transportation for seven states, the FHWA, many researchers

from universities in the region, and members of the precast and prestressed industries

were represented. Presentations were given by David Nichols (MoDOT), Tommy Beatty

(FHWA), Sam Tyson (FHWA), Eric Krapf (MoDOT), David Merritt (Transtec Group,

Inc.), Vellore Gopalaratnam (UMC), Andrew Maybee (CPI Concrete Products), and John

Donahue (MoDOT). Much discussion was focused on the transfer of PPCP technology

and ways to facilitate practical implementation. The overall purpose of the ETG meeting

was to discuss and receive feedback from all parties involved in the conception, design,

fabrication, and construction of PPCP systems. Several in attendance noted that design

standards would facilitate more rapid use of the technology.

________________________________________________________________________

Page 80
 The difficult issues with PPCP are in the actual implementation of the

sophisticated designs. Designs using small post-tensioning ducts are used and may sound

appealing for designers, but the manipulation and placement of large, heavy concrete

panels that are relatively thin becomes complicated when tight tolerances are required.

The higher initial costs associated with PPCP will be offset somewhat by reduced costs

due to more rapid construction and potentially reduced maintenance. In any case, it is

likely that PPCP may be initially reserved for high volume urban areas where lane

closures are both expensive to end users and dangerous for workers and travelers.

Members of the ETG spent time discussing the potential causes of cracking found

in the pavement described in the previous section. Some of the more important causes

for the cracking observed in the pavement system were identified as;

• Thermal shock due to fabrication of the panels during the winter time in

an outdoor bed where panels were steam cured and subsequently exposed

to sub-freezing temperatures. Even while only some cracks were visible

in the panel prior to placement in the field, it is speculated that residual

tensile stresses may have reduced intended levels of prestress.

• The varying thickness of the panel resulted in local non-uniform stresses

from prestressing, thermal and restraint loads.

• The effects of epoxy applied between panels that cured well before post-

tensioning operations could be completed results in unintended effects.

• Grouting operations used a much larger amount of grout than anticipated.

Since no grout was observed exiting the pavement, much of it could have

________________________________________________________________________

Page 81
 gone underneath the panels causing upward pressures and uneven stress

distributions to the base.

4.5.2. Visual Crack Surveys

The pavement sections were visually inspected each time the research team

visited the site. A total of 8 visits were undertaken during the 18-month service

performance monitoring of the pavement to inspect the pavement, survey joint

performance and troubleshoot instrumentation following weather-related damage to the

electronic circuitry. The overall performance of the pavement system has been very

good. Two aspects that could use more attention in future projects include “as

constructed” joint performance and ways to mitigate pavement-cracking. Both of these

issues have not, and are not likely to pose performance problems in the future. However,

experience from this project can serve to facilitate design, fabrication and construction

improvements where these aspects can be better addressed.

4.6. Joint Panel Performance

The joint panels were designed to open at the pre-engineered construction joint at

the center of the panel during post-tensioning operations. However due to difficulties in

threading the post-tensioning strands and resulting delays, the contractor decided to place

all panels in all four 250 ft sections prior to post-tensioning. This may have affected

performance at the joint panels as the original plan was to post-tension one 250 ft section

before placing panels for the adjacent section. In addition, the cold joint in the two-step

casting of the joint panels had better bonding than anticipated across the joint. As a result
________________________________________________________________________

Page 82
the joint panel between sections 3 and 4 (the more heavily instrumented joint panel –

A32) did not open up as designed. However, all other joint panels have performed well

during this study. In the contractor’s effort to open the A32 joint using jacks, the

concrete fractured adjacent to the intended joint. A patch mix was used to repair the

loose concrete and later filled as intended with joint compound. The damaged joint

operated fine until the summer of its first year in service. It can be seen in Figure 4.23

that by mid-June with temperatures reaching new highs, the joint compound had started

to chip away due to the restrained joint operation. The following photographs (Figure

4.23 – Figure 4.25) illustrate problems with the joint compound in Panel A32. Initially,

when the joint compound was squeezed out of the joint it was expected that the joint may

deteriorate significantly in time and may exhibit poor performance. However, since the

joint compound chipped away due to normal traffic wear, the recorded concrete strain in

the panel were similar to that of the base panels located in the middle of the 250 ft

section. The joint appeared to be performing well in the most recent inspection in May

2007.

________________________________________________________________________

Page 83
Figure 4.23 – Flexible joint compound squeezing out on a hot day with minor amount
of chipping of rigid compound, Joint Panel A32 (June 27, 2006)

Figure 4.24 – Rigid joint compound chipped away more extensively, Joint Panel A32
(August 16, 2006)

________________________________________________________________________

Page 84
 12 inches

Figure 4.25 – Moderate degradation to rigid joint compound, Joint Panel A32 (May 9,
2007)

________________________________________________________________________

Page 85
 This page intentionally left blank.

________________________________________________________________________

Page 86
5. Conclusions

5.1. Project Observations

The 1,010 ft precast prestressed pavement system on I-57 in south-eastern

Missouri is performing well since being opened to traffic in January, 2006. The

pavement surface and prestressed system are in good condition aside from some visible

cracking. These cracks, which are present in the transverse and longitudinal directions,

are not expected to present problems with regard to the durability or long-term

performance of the pavement system. Possible reasons for the cracking observed under

service conditions have been detailed in Section 4.5.1

The pilot project was successful and has provided useful insights into improving

fabrication and construction techniques that will benefit future projects. The structural

design of the panels and their economical implementation will only improve with the

experience gained in this and other similar projects.

The following sections outline the summary conclusions based on the evaluation

of results from the instrumentation program, companion laboratory tests, observations

from construction, and visual inspections of the performance of the precast prestressed

pavement system.

________________________________________________________________________

Page 87
5.2. Construction Challenges

The first two challenges are interrelated and stem from the decision by the

contractor to lay all 101 panels before installing and post-tensioning strands in each of the

250 feet sections.

• Threading of the post-tensioning strands through the ducts for the 250 feet long

sections proved difficult. Suspected impediments to strand installation were

cumulative variations in panel-to-panel alignment, misaligned ducts, and possible

ice buildup in some of the ducts.

• The usage of epoxy between precast panels to facilitate proper alignment of panel

edges hardened due to construction delays with threading the post-tensioning

strands. The hardened epoxy bonded the panels together and resulted in the

sections behaving as a monolithic 250 ft. unit prior to stressing instead of 25 10-ft.

panels. In hindsight, it would have been beneficial to require sequential post-

tensioning of the four 250 ft. pavement sections as originally planned.

• Wooden and steel wedges (shims) were inserted between panels along the outer

shoulder edge of pavement to correct the global pavement misalignment during

placement. The use of these wedges affected the stress distribution during post-

tensioning operations as documented earlier in Chapter 4. Use of wedges

(particularly stiff high-modulus steel wedges) in such a post-tensioned pavement

system should be disallowed for this reason.

________________________________________________________________________

Page 88
5.3. Service Performance

• It has been demonstrated that with appropriate data acquisition sampling

rates, monitoring of related embedded instrumentation, and methods of

analysis, it was possible to isolate and measure strains from traffic loads,

post-tensioning operations as well as thermal loads from daily, weekly,

and seasonal trends.

• Pavement strains due to temperature change are significantly larger than

all other types of loading (viscous effects such as creep, shrinkage and

relaxation, or vehicular loads). Daily longitudinal strains excursions

ranging from 50-100 µstrain were observed for cool or mild days. Hot

summer days appear to produce strain excursions ranging from 125-200

µstrain.

• Larger magnitude of strain in the transverse direction than the longitudinal

direction, while inconclusive, indicates lower levels of restraint

transversely.

• Additionally, thermal gradients were observed due to varying pavement

thicknesses and exposure conditions (top versus bottom pavement

surface).

• Measured vehicular loadings produced approximately 1-2 µstrain in the

precast pavement at the location of measured strain. This constitutes only

1-2% of the total strain produced from daily thermal loading.

________________________________________________________________________

Page 89
5.4. Pavement Longevity

The future performance of the precast prestressed pavement system is expected to

be without significant problems. The precast prestressed technology that has been

adapted for use as a pavement system has been proof-tested in previous decades of bridge

projects. However it would be naïve to say that among the multitude of pavement

problems that exist, none will arise. It is certain that many of the known significant

challenges, addressed earlier, have been targeted in the design of the prestressed

pavement system which will enable it to perform as well as or better than its traditional

pavement counterparts. The areas that are expected to meet or exceed design

expectations are the following:

• Ability of the pavement system to span voids that may form in the base

material due to erosion or settlement.

• Due to the built in compressive stresses in the prestressed pavement,

cracks are expected to stay closed. The smaller crack widths compared to

a non prestressed pavement are likely to minimize damage due to water

intrusion and resultant freeze-thaw.

• Increased material durability due to improved quality control at a

precasting yard.

5.5. Recommendations for Future Work

Future crack surveys should be performed to develop an understanding of the

extent of crack growth. If the extent of cracking is increasing, then potential causes could

be predicted with higher certainty by studying the problematic areas. Although the
________________________________________________________________________

Page 90
cracking currently present does not appear to have been or become detrimental to the

pavement system, it was not a desirable outcome. Further inspection and sampling on a

multi-seasonal level could provide useful data indicating the severity of the cracking and

possible extent of moisture intrusion. Useful information on the expected performance

can allow preventative measures to be performed to minimize detriments.

In-situ chloride permeability tests should also be performed after the pavement

has been subjected to deicing salts to determine the rate of chloride ingress in the field.

These results can then be compared to baseline readings taken on virgin specimens and

compared to field results of other pavement types. This testing may provide useful

knowledge on material specific performance unique to the pavement system.

________________________________________________________________________

Page 91
REFERENCES

ASTM (1998). Standard Practice for Capping Cylindrical Concrete Specimens,

ASTM:5.

ASTM (2002). Standard Test Method for Flexural Strength of Concrete (Using Simple
Beam with Third-Point Loading, ASTM: 3.

ASTM (2002). Standard Test Method for Fundamental Transverse, Longitudinal, and
Torsional Resonant Frequencies of Concrete Specimens, ASTM: 7.

ASTM (2003). Standard Test Method for Resistance of Concrete to Rapid Freezing and
Thawing, ASTM: 6.

ASTM (2005). Standard Test Method for Compressive Strength of Cylindrical Concrete
Specimens, ASTM: 7.

ASTM (2005). Standard Test Method for Electrical Indication of Concrete's Ability to
Resist Chloride Ion Penetration, ASTM: 6.

ACPA (2004). Concrete Types, American Concrete Pavement

Dailey, C. (2006). Instrumentation and Early Performance of an Innovative Prestressed

Precast Pavement System. Civil and Environmental Engineering. Columbia,
University of Missouri – Columbia.

Davis, B. M. (2006). Evaluation of Prestress Losses in an Innovative Prestressed Precast

Pavement System. Civil and Environmental Engineering. Columbia, University of
Missouri – Columbia.

Degussa (2006). Degussa Admixtures. 2006.

Eatherton, M. (1999). Instrumentation and Monitoring of High Performance Concrete

Prestressed Girders. Civil Engineering. Columbia, University of Missouri -
Columbia.

Geokon (1996). Instruction Manual Model VCE - 4200, Geokon, Inc.

Gopalaratnam, V. S., B. M. Davis, et al. (2007). Performance Evaluation of Precast

Prestressed Concrete Pavement, RI03-007. Columbia, University of Missouri –
Columbia.

Maxim (2007). iButton Instruction Manual Model DS1922L, Maxim IC, Inc.

________________________________________________________________________

Page 92
Merritt, D., F. B. McCullough, et al. (2000). The Feasibility of Using Precast Concrete
Panels to Expedite Highway Pavement Construction. Austin, TX.

Merritt, D. K., F. B. McCullough, et al. (2001). Feasibility of Precast Prestressed

Concrete Pavements. 7th International Conference on Concrete Pavements.

Merritt, D. K., B. F. McCullough, et al. (2002). Construction and Preliminary Monitoring

of the Georgetown, Texas Precast Prestressed Pavement, Center for
Transportation Research; University of Texas at Austin: 140.

Transtec (2005). Personal Correspondence. Precast Panel Design Drawings.

Transtec (2009). Precast Pavement. www.precastpavement.com.

Tyson, S. S. and D. K. Merritt (2005). Pushing the Boundaries, Federal Highway

Administration. 2006.

________________________________________________________________________

Page 93
APPENDIX A

0 4 8 12 16 20 24
16 60
A32_T1 A32_T2
A32_T3 A32_T4 54
14 A32_T5 A32_T6
A32_T7 A32_T8
A32_V1t A32_V3t 48
12 A32_V4t Ambient Temp

Temperaturel, T (°F)
Temperature, T (°C)

42
10
36
8 30

6 24

18
4
12
2 Instrumentation Plan
R8
R6 R5 R2 6
V4 V2

T7-8
0 0
Inside Shoulder

Outside Shoulder

Traffic
T4-6 Direction
R7 R1 V1 T1-3
V3 R3 R4
60
Instrumented Vibrating Thermocouple
Rebar Wire Gage

45
Strain, ε (µstrain)

30

15

0
A32_R1 A32_R3
A32_R4 A32_R5
A32_V1c A32_V2c
-15 Theoretical

0 4 8 12 16 20 24
Time, t (Hours)

Figure A.1 – One day window from 12/27/2006 for Panel A32 (a) temperature
history (b) strain history

________________________________________________________________________

Page 94
APPENDIX B

The exact locations of instruments (rebars, thermocouples, vibrating wire) used in this

project are tabulated below.

Figure B.1 – Convention for gage locations

________________________________________________________________________

Page 95
 Panel C1 X Y Z Depth X
(in) (in) (in) Section (in)
R1 31 45 3 6.3
R2 142 48 4 8.5
R3 264 52 7 11.0
R4 264 44 4 11.0
R5 384 62 4 8.5

T1 126 70 2 8.2
T2 126 70 4 8.2
T3 126 70 7 8.2
T4 364 52 2 8.9
T5 364 52 7 8.9
Table B.1 – Locations of instruments used in Panel C1

Panel
B1 X Y Z Depth X
(in) (in) (in) Section (in)
R1 29 60 2.5 6.2
R2 137 62 4 8.4
R3 252 55 4 10.8
R4 256 60 4 10.8
R5 378 62 4 8.6

T1 121 54 1 8.1
T2 121 54 3 5.6
T3 121 54 6 5.6
T4 366 68 3 8.9
T5 366 68 7 5.6

V1 137 62 4 8.4
V2 256 60 4 10.8
Table B.2 – Locations of instruments used in Panel B1

________________________________________________________________________

Page 96
 Panel
B2 X Y Z Depth X
(in) (in) (in) Section (in)
R1 27 60 4 6.2
R2 142 60 4 8.5
R3 254 56 4 10.8
R4 260 62 4 10.9
R5 392 47 4 8.3

T1 142 67 1 8.5
T2 142 67 3 8.5
T3 142 67 7 8.5
T4 364 52 1 8.9
T5 364 52 6 8.9

V1 142 60 4 8.5
V2 260 62 4 10.9
Table B.3 – Locations of instruments used in Panel B2

Panel
B3 X Y Z Depth X
(in) (in (in) Section (in)
R1 20 46 3 6.0
R2 138 62 4 8.4
R3 251 54 2 10.7
R4 262 59 4 11.0
R5 379 60 4 8.6

T1 145 57 2 8.6
T2 145 57 4 8.6
T3 145 57 6 8.6
T4 289 58 3 10.5
T5 289 58 6 10.5
T6 289 58 8 10.5
T7 361 67 2 9.0
T8 361 67 4 9.0

V1 138 62 4 8.4
V2 262 59 4 11.0
Table B.4 – Locations of instruments used in Panel B3

________________________________________________________________________

Page 97
 Panel
B4 X Y Z Depth X
(in) (in) (in) Section (in)
R1 31 61 3 6.3
R2 140 63 4 8.5
R3 255 54 2.5 10.8
R4 264 61 4 11.0
R5 375 62 4 8.7

T1 145 54 1 8.6
T2 145 54 4 8.6
T3 145 54 7 8.6
T4 355 54 2 9.1
T5 355 54 4 9.1
T6 355 54 7 9.1

V1 140 63 4 8.5
V2 264 61 4 11.0
Table B.5 – Locations of instruments used in Panel B4

________________________________________________________________________

Page 98