Effect of slab curling and warping on pavement layer backcalculation

1
Effect of slab curling and warping on pavement
layer backcalculation
2007 Falling Weight Deflectometer (FWD) Users
Group Conference Des Moines, Iowa September 29
October 02, 2007
M. Birkan Bayrak, Ph.D. Candidate and Halil
Ceylan, Ph.D. - Assistant Professor
2
Acknowledgements

• The support provided by the Iowa DOT for the
  Nondestructive Evaluation of Iowa Pavements
  research project is greatly appreciated.

3
Outline

• Introduction
• Background
• Curling and Warping in PCC Pavements
• Procedure of the Developed Approach
• ANN-Based Backcalculation Models
• Comparison with the MLR Analysis
• Application to the Actual Field Data
• Summary and Conclusions

4
Introduction

http//en.wikipedia.org/wiki/National_Highway_Syst
em

• about 3,000,000 miles rural highways
• about 900,000 miles urban highways

www.ljlindhurst.com/highway.html
5
Introduction

• Increased traffic volume
• Heavier loads
• Budget constraints
• Aging of existing systems

• To facilitate managing demanding task and to
  efficiently allocate resources fast
  and efficient way to determine structural
  condition of pavements

6
Background

• Concrete curling and warping of concrete slabs
  can influence the long-term performance and
  service life of the PCC pavements.
• The goal of the study is to develop a rapid
  methodology for backcalculating the Total
  Effective Linear Temperature Difference (TELTD)
  in jointed plain concrete pavements (JPCP) from
  the FWD deflection basins and the thickness of
  the concrete pavement layer.

7
Background

• Previously developed ANN-Based Backcalculation
  Models
• Concrete pavement layer modulus (EPCC) of the PCC
  layer,
• Coefficient of subgrade reaction (ks) of the
  pavement foundation,
• Tensile stresses at the bottom of the PCC layer
  (sPCC),
• Radius of Relative stiffness (RRS).

8
Background
9
Background
10
Background
11
Curling and Warping in PCC Pavements
The total amount of curling in a slab due to a
combination of the five factors
Temperature Gradient, DTtg

Moisture Gradient, DTmg

Built-In Temperature Gradient, DTbi

or
Drying Shrinkage, DTshr

Creep, DTcrp
12
Curling and Warping in PCC Pavements
TELTD The total effective linear temperature
difference
EBITD Effective built-in temperature difference
TELTD DTtgDTmgDTbiDTshrDTcrp
EBITD DTmgDTbiDTshrDTcrp
TELTD DTtg EBITD
13
Curling and Warping in PCC Pavements
Typical day-time curling
Typical night-time curling
Top of the PCC layer Warmer Bottom of the PCC
layer Cooler
Top of the PCC layer Cooler Bottom of the PCC
layer Warmer
14
ISLAB2000 Solution Model
ISLAB2000 FE model meshing for the six-slab JPCP
assembly
15
ISLAB2000 Solution Model
Deflections
Stressesin X-dir.
Stressesin Y-dir.
PrincipalStresses
Traffic Load Only
16
ISLAB2000 Solution Model
Deflections
Stressesin X-dir.
Stressesin Y-dir.
PrincipalStresses
Traffic Load Environmental Loading
17
Curling and Warping in PCC Pavements
The effect of the temperature-introduced FWD
deflection basins
18
Training Set Generation
EPCC
Structural Model
D0
hPCC
D8
ks

D60
LTE
TELTD
ANN Model
D0
D8

TELTD
D60
hPCC
LTE
19
ANN-Based Backcalculation Models
D0
D8
D12
D18
D24
D36
D48
D60
4-Deflection ModelsD0,D12,D24,D36 6-Deflection
ModelsD0,D12,D24,D36,D48,D60 7-Deflection
ModelsD0,D8,D12,D18,D24,D36,D60 8-Deflection
ModelsD0,D8,D12,D18,D24,D36,D48,D60
20
ANN-Based Backcalculation Models
Corner Deflection Basin ONLY
Center Deflection Basin ONLY
21
ANN-Based Backcalculation Models

• Load Transfer Efficiency (LTE) was taken into
  account in the analysis.

• Both center and corner deflection basins were
  used in the developed models.

22
Material Property/Model Inputs for ISLAB2000
Solutions
The ranges of the input parameters (ISLAB2000
analyses)
23
ANN-Based Backcalculation Models
Architecture of the ANN-based backcalculation
models
24
ANN-Based Backcalculation Models
BCM-TELTD-(4) Model
25
ANN-Based Backcalculation Models
BCM-TELTD-(6) Model
26
ANN-Based Backcalculation Models
BCM-TELTD-(7) Model
27
ANN-Based Backcalculation Models
BCM-TELTD-(8) Model
28
Noise-Introduced ANN-Based Backcalculation Models
29
Comparison with MLR Analysis
Schematic View of the Multiple Linear
Regression (MLR) Analysis
Schematic View of the Artificial Neural Network
(ANN) Analysis
30
Comparison with MLR Analysis
MLR Model Predictions
ANN Model Predictions
31
Field Data Analysis

• Challenges in the Curling / Warping Problem
• Flat condition of the slab is unknown
• Voids under slab corner and center
• Time of the FWD test
• Location of the FWD test
• Number of the FWD tests per slab
• Effect of the moisture on the slab behavior

32
Field Data Analysis
Slab 1 Slab Center
Slab 1 Slab Corner
Day
Day Night
Night
Slab 2 Slab Center
Slab 2 Slab Corner
Day
Day Night
Night
Slab 3 Slab Center
Slab 3 Slab Corner
Day
Day Night
Night
33
Field Data Analysis
Slab CENTER Testings
Slab CORNER Testings
Normalized slab center and corner D0 deflections
34
Field Data Analysis
SLAB 1
SLAB 2
SLAB 3
35
Field Data Analysis
SLAB 1
SLAB 2
SLAB 3
36
Summary and Conclusions

• Pavement engineers designers provided with the
  state- of-the-art finite element (ISLAB
  2000) solutions
• Sixteen ANN-based backcalculation models have
  been developed using some 8,000 ISLAB2000 finite
  element solutions
• No seed moduli input required
• No large computer resources needed
• Rapid predictions of Total Effective Linear
  Temperature Difference (TELTD)
• A simpler and easier backcalculation method for
  State DOTs for network and project level
  studies

37
Summary and Conclusions

• Backcalculated moduli of pavement layers based
  on flat-slab condition assumptions by using the
  FWD data may be unrealistic
• TELTD can be used as an adjustment parameter in
  the backcalculated moduli of pavement layers in
  the future studies
• TELTD in concrete slabs can be successfully
  backcalculated from FWD deflection basins with
  the developed ANN-based models
• To backcalculate the EBITD by using this approach
  is also possible if the actual field DTtg is
  measured in the field with additional methods

38
Summary and Conclusions

• Both center and corner FWD tests are needed for
  successful prediction of TELTD
• The thickness of the slab plays an important
  factor that affects the total amount of curling
  and warping in concrete pavements
• Load transfer efficiency is another significant
  factor that affects the FWD deflection readings
• Moisture gradients should be also taken into
  account in the EBITD calculations as it changes
  throughout the day

39
Questions
Thank You!