Structural Adhesive Behavior: Experimental and Numerical...

Aiman Shibli & Mohsen A. Issa

Structural Adhesive Behavior:Experimental and Numerical Study

Reliable Engineering Computing - 2014 conference May 27th ,2014

University of Illinois at ChicagoDepartment of Civil and Material Engineering

7/23/2014 2

Outline

Introduction

Objective

Mechanical Properties

Material Models

Material Model Validation

Summary & Conclusion

7/23/2014 3

Outline

Introduction

Objective


Material Models



Section failure

Shear out Splitting Bearing Failure

IntroductionUICUNIVERSITY OF ILLINOIS

AT CHICAGO

RECRELIABLE ENGINEERING

COMPUTING 2014

Structural Adhesive Applications:

Bolting create stress concentrations which lead to premature failure Adhesive joints for civil infrastructure

applications are in its infancy due to the lack of understanding bonded joint behavior and lack of design guidelines..

Adhesive joints are being increasingly used due to their unique characteristics and advantages. They have been widely used in the fields of aerospace, automotive and electronic packages.

Civil infrastructure applications have considered utilizing adhesive for joining load-bearing components as an excellent candidate for replacing the traditional joining methods such bolting and riveting, especially for scenarios involving joining Fiber Reinforced Polymers (FRP) composites.

7/23/2014 5Department of Civil and Materials Engineering

Introduction

High joint efficiency and dimensional stability

Adhesive distribute load across the entire joint area

Fewer stress concentrations

No need to dig holes and damage the adherends; holes are points where moisture ingress can occur, which can affect durability

Low structural weight

Low fabrication cost and improved damage tolerance

Rot resistance

Ease of use

Advantages:

UICUNIVERSITY OF ILLINOIS

AT CHICAGO


COMPUTING 2014


Introduction

Adhesive failure

Stock-break failureCohesive failure

Thin-layer Cohesive failure

Fiber-tear failure Light-fiber-tear failure

Failure Modes of Adhesive Joints:

The properties of the bulk adhesive The adhesion properties at the interface


AT CHICAGO


COMPUTING 2014

1- 2-

3- 4-

5- 6-

7/23/20147Department of Civil and Materials Engineering

IntroductionUICUNIVERSITY OF ILLINOIS

AT CHICAGO


COMPUTING 2014

Design Parameters for Adhesively Bonded Joints

Structural Adhesive: Mechanical Property: property of the bulk adhesive and adhesion at the interface Requirement for curing Resistance to environmental parameters

Joints: Geometry: surface area and bondline thickness Parameters concerning loading and environment

Fabrication: Preparation before bonding: surface treatment Curing and post-curing: clamping pressure and control environment

7/23/2014 8

Outline

Introduction

Objective


Material Models



7/23/2014 9

Characterize adhesive’s mechanical property under tensile and shear loadings, at low and high rates (static and dynamic)

Create FEA material model for the adhesive that can describe the adhesive behavior accurately

Validate material model experimentally and computationally at coupon level (coupon tension and shear) and sub-system level (four-point bend and dynamic ball drop)

Department of Civil and Materials Engineering

ObjectiveUICUNIVERSITY OF ILLINOIS

AT CHICAGO


COMPUTING 2014

7/23/2014 10

Outline

Introduction

Objective


Material Models



Adhesive Mechanical Properties

Tensile and Shear testing:

• It is important to prepare bulk specimens that is in a suitable size for testing, and can accurately represent the properties of the thin layer of the adhesive in a bonded joint.

Adhesives are viscoelastic materials, therefore their mechanical properties depend upon loading mode and rate.


AT CHICAGO


COMPUSING 2014


Load Cell

Moving Crosshead

Holding Grip

Specimen

Holding Grip

F

F

F

F

100mm x 20mm x 2mm150mm x 20mm x 16mm

Adhesive:10mm x 20mm x 0.2mm

Tensile specimen Shear specimen

Determination of Tensile & Shear Behavior at Quasi-Static Rate


150mm

16mmAdhesive t =0.2mm

80mm

15mm

150mm long x 16mm thick x 20mm wide100mm long x 20mm wide x 2mm thick

0

1

2

3

4

5

6

7

8

9

0 0.02 0.04 0.06 0.08 0.1 0.12

Stre

ss [

MP

a]

Strain

Shear Test

0

1

2

3

4

5

6

7

8

9

0 0.2 0.4 0.6 0.8

Stre

ss [

MP

a]

Strain

Tensile Test


AT CHICAGO


COMPUTING 2014



QS HR-100/sec HR-1000/sec

Young Modulus (MPa) 1566 3960 5040

Stress @ yield (MPa) 5.2 19.6 22.4

Elongation @ yield (%) 1.2 1.2 1.16

Elongation @ fail (%) 79 33 42

0

5

10

15

20

25

0 0.2 0.4 0.6 0.8

Stre

ss [

MP

a]

Strain

Tensile Test

Quasi-Static Strain Rate 100/s Strain Rate 1000/s

0

10

20

30

40

50

60

70

0 0.02 0.04 0.06 0.08 0.1 0.12

Stre

ss [

MP

a]

Strain

Shear Test

Quasi-Static High Rate 2000/s High Rate 10000/s

QS HR- 2000/sec HR-10000/sec

Shear Modulus [Mpa] 630 756 1314

Max. Shear Stress [Mpa] 9 32 64

Shear Strain @ fail [%] 11 9 6

Both linear and non-linear properties of adhesives are dependent on the loading rate.


AT CHICAGO


COMPUTING 2014

Determination of Tensile & Shear Behavior at High Rates


Adherend Mechanical Properties

0

50

100

150

200

250

300

0 0.02 0.04 0.06 0.08 0.1 0.12St

ress

[M

pa]

Strain

Al-6061-T6

Modulus [GPa] Yield Stress [MPa]

67 271

Modulus [GPa] Yield Stress [MPa]

Quasi-static 2.9 50

High Rate 80/s 2.9 88

0

20

40

60

80

100

0 0.5 1 1.5 2 2.5 3

Stre

ss [

MP

a]

Strain

Quasi-Static High Rate 80/s

Polycarbonate


AT CHICAGO


COMPUTING 2014

Tensile Test for Adherends (Al & PC):


7/23/2014 15

Outline

Introduction

Objective


Material Models



Adhesive Material ModelMaterial Modeling approaches:

Linear Elastic

• Conversion of nominal stress-strain to true (Cauchy) stress-strain using relations:

nomtrue

nomnomtrue

1In

1

• Plasticity model assumed isotropic yield and hardening behavior.• Plastic data in terms of true stress, true plastic strain.• From strain decomposition, true plastic strain to be obtained from

• Elastic response fully specified by Young’s modulus, E, and Poisson’s ratio, ν

E

true

nom

p

e

true

t

true

p

true

1InIn

Linear Elastic – Plastic

Linear Elastic – Viscoelastic

, material coefficient are and


AT CHICAGO


COMPUSING 2014


Adhesive Material Model

Material Modeling approaches:

Linear Elastic Linear Elastic - Plastic Linear Elastic - Viscoelastic

A B C D E

*Elastic *Elastic *Elastic *Elastic *Elastic

E = 1566 MPa E = 5040 MPa E = 1566 MPa E = 5040 MPa E = 1566 MPa

*Plastic *Viscoelastic

True Stress –true strain:0.01/s, 100/s and 1000/s

The hardening curve data are required in the tabular form of yield stress with plastic strainwhere the first pair of numbers must correspond to the initial yield stress at zero plastic strain.

In Abaqus G(t ) and K(t ) are specified in one of four different ways:

Prony series curve fit from experimental stress relaxation test data:

*VISCOELASTIC, TIME=RELAXATION TEST DATAProny series curve fit from experimental creep test data:*VISCOELASTIC, TIME=CREEP TEST DATAProny series coefficients specified directly by the user:*VISCOELASTIC, TIME=PRONYDefined from frequency-dependent cyclic test data:*VISCOELASTIC, TIME=FREQUENCY DATA

0

50

100

150

200

250

300

0 0.05 0.1 0.15 0.2

Tru

e S

tre

ss [

MP

a]

True Plastic Strain

QS HR - 100/s HR - 1000/s


AT CHICAGO


COMPUTING 2014


7/23/2014 18

Outline Introduction

Objective


Material Models


o Sandwich Structure

o Quasi-Static Bend Test

o Dynamic Bend Test

o FEM model

o Quasi-Static Bend Simulation

o Dynamic Bend Simulation



Sandwich structures were constructed where the adhesive was sandwiched between two substrates (Al and PC); two configurations were considered as shown in figure-1.

Strain Gauges

Plates: 70mmX20mmX1mm

Adhesive: 70mmX20mmX0.2mm

Top Span: 35mm

Bottom Span: 60mm

Configuration-1: Equal length adherends

Strain Gauges

Big Plate: 70mmX20mmX1mm

Small Plate: 20mmX20mmX1mm

Adhesive: 70mmX20mmX0.2mm

Top Span: 35mm

Bottom Span: 60mm

Configuration-2: Unequal length adherends


AT CHICAGO


COMPUTING 2014


Sandwich Structure

Loading pins

Supporting pinSupporting pin

Quasi-Static Bend Test

Dynamic Bend Test


Accelerometer

Two strain gauges for each sandwich structure were used. One at the center of the top plate and one at the center of the bottom plate

Accelerometer was placed at the top fixture in dynamic testing in order to extract acceleration and use it as input in the computational study

Loading pins



AT CHICAGO


COMPUTING 2014



Quasi-Static & Dynamic Testing:

0

50

100

150

200

250

0 2 4 6 8

Forc

e (

N)

Displacment (mm)

Force - Displacement

Spc1

Spc2

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0 50 100 150 200 250

Stra

in

Force (N)

Force vs Strain

Spc1- Top Strain Spc1- Bottom Strain

Spc2 - Top Strain Spc2 - Bottom Strain

0

20

40

60

80

100

120

0 2 4 6 8

Forc

e (N

)

Displacement (mm)

Force vs Displacement

Spc-1

Spc-2

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0.008

0.01

0.012

0 50 100 150

Stra

in

Force (N)

Force vs Strain

Spc 1 - Top Strain Spc 1 - Bottom Strain

Spc 2 - Top Strain Spc 2 - Bottom Strain


AT CHICAGO


COMPUTING 2014

Quasi-Static Test Results:


-0.006

-0.004

-0.002

-1E-17

0.002

0.004

0.006

0.008

0.01

0.012

0 0.005 0.01 0.015

Stra

in

Time (s)

Strain vs Time

Strain1-Spc1-15cm-2hit Strain2-Spc1-15cm-2hit

-300

-200

-100

0

100

200

0 0.002 0.004 0.006 0.008 0.01

Acc

ele

rati

on

, g

Time (s)

Acceleration vs Time

Spc1-15cm-1hit

-600

-500

-400

-300

-200

-100

0

100

200

0 0.002 0.004 0.006 0.008 0.01 0.012

Acc

eler

ati

on

Time

Acceleration vs Time

TEST - 35CM HEIGHT

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0.008

0.01

0.012

0 0.002 0.004 0.006 0.008 0.01 0.012

Str

ain

Time [sec]

Strain vs Time

TEST-35CM-Top Gauge TEST-35CM-Bottom Gauge

15 cm Ball Drop30 cm Ball Drop


AT CHICAGO


COMPUTING 2014

Dynamic Test Results:



Accelerometer


Accelerometer

Finite Element Analysis Model

A full three-dimensional finite element model of the sandwiched structure (Al plate, Adhesive and PC plate), and the supports of the four-point bend fixture were built using the modeling software ABAQUS. The Plates and the adhesive were modeled with full-integration brick elements. The two strain gauges were modeled as two membrane elements [0.8x0.8mm]. Supports of the four-point bend fixturewere modeled using analytical rigid surfaces.

Appropriate contact definitions were defined between the rollers and plates surfaces. The bottom support was constrained in all directions in order to prevent rigid body motion. The top support was also constrained in every direction except the vertical direction.


AT CHICAGO


COMPUTING 2014


Al

PC

Adhesive

Membrane element (Strain gauge)

Membrane element (Strain gauge)

Analytical body

Quasi-Static Bend Simulation

0

20

40

60

80

100

120

140

0 1 2 3 4 5 6 7

Fo

rce

(N)

Displacement (mm)


Test Sim-Elastic Sim-ElasticPlastic

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0 20 40 60 80 100 120 140

Str

ain

Force (N)

Force - Strain

Test - Top Gauge Test - Bottom Gauge

Sim-Elastic-Bottom Gauge Sim-Elastic-Top Gauge

Sim-ElasticPlastic-Top Gauge Sim-ElasticPlastic-Bottom Gauge

0

50

100

150

200

250

0 2 4 6 8

Forc

e (N

)

Displacement (mm)


Test Sim-Elastic Sim-ElasticPlastic

-0.01

-0.005

0

0.005

0.01

0.015

0.02

0.025

0.03

0 50 100 150 200

Str

ain

Force (N)

Force - Strain

Test - Top Gauge Test - Bottom Gauge

Sim-Elastic-Top Gauge Sim-Elastic-Bottom Gauge

Sim-ElasticPlastic-Top Gauge Sim-ElasticPlastic-Bottom Gauge


AT CHICAGO


COMPUTING 2014


Dynamic Bend Simulation

-0.01

-0.005

0

0.005

0.01

0.015

0 0.002 0.004 0.006 0.008 0.01 0.012

Str

ain

Time [sec]

Strain vs Time

SIM-35CM-ElasticPlastic-Bottom Gauge SIM-35CM-ElasticPlastic-Top Gauge

SIM-35CM-Elastic-Bottom Gauge SIM-35CM-Elastic-Top Gauge

TEST-35CM-Top Gauge TEST-35CM-Bottom Gauge


AT CHICAGO


COMPUTING 2014



7/23/2014 26

Outline

Introduction

Objective


Material Models



7/23/2014 27

The sandwich structure with unequal-length plates is able to capture the sensitivity to the adhesive material modeling approaches. However, the sandwich structure with equal-length plates is insensitive to the adhesive layer and failed as a configuration for capturing or investigating adhesive behavior and properties

Force-displacement and force-strain results in quasi-static bend test show that the linear-elasticmaterial model will over predict the stiffness of the sandwich structure where the elastic-plastic material model mimics the experimental results much better.

Dynamic bend simulations with unequal-length-plate sandwich structure show that the overall behavior of the structure is sensitive to the adhesive material modeling approach. The elastic-plastic material modeling approach predicts the testing results very well, unlike the elastic material modeling approaches which correlate poorly with experimental results because they do not account for plasticity.


Summary & Conclusion UICUNIVERSITY OF ILLINOIS

AT CHICAGO


COMPUTING 2014

7/23/2014 28

Adhesive material property is sensitive to rate of loading, stiffness increases with the increase of the loading rate even in the elastic region.

Determination of adhesive property over wide range of strain rate is needed to enable predictions under dynamic loading.

Although both of the elastic and plastic phases of this specific adhesive are loading rate dependent, plastic behavior of the adhesive has bigger impact than its elastic behavior on the overall sandwich structure behavior. This is due to the fact that the adhesive will start deforming while the adhered substrates are still in their elastic regions.

The accuracy of the computational calculations is dependent on the validity of the material models used in the analysis to describe the material behavior and the availability of suitable test data for these models.



AT CHICAGO


COMPUTING 2014Summary & Conclusion

Structural Adhesive Behavior: Experimental and Numerical...

Documents

Transcript of Structural Adhesive Behavior: Experimental and Numerical...