Modeling of Cement Based Composite Laminates Barzin Mobasher Department of Civil & Environmental Engineering Arizona State University Tempe, Arizona, 85287-5306
HPFRCC-4, Ann Arbor, Michigan, June 16-18, 2003.
Scope of Presentation
Introduction Pultrusion Based Cement Composite Systems Theoretical Aspects of Formulation
Non-linear matrix response
Cracking Softening
Case Studies
Elastic response Traditional Composite laminate approach
Glass Fiber/ carbon fiber epoxy composites Cement based Glass fiber composite laminate Polypropylene fiber cement composites Fabric composites Comparison with Experimental results Repair and retrofit approach
Conclusions
Areas of Application for high performance fiber reinforced Materials
high tensile-toughness characteristics superior impact, earthquake, and fatigue characteristics. prefabricated Structural elements, thin sheets, panels, cladding members. structural repair and retrofit. I-beams, structural members. Sound abatement walls.
Scope of Presentation
Introduction Pultrusion Based Cement Composite Systems
Filament Winding
Fiber Spool
Fabric Reinforcing Methodology
Cement composites with 50 MPa Tensile Strength and more than 1% strain Capacity 50 Unidirectional
Stress, MPa
40 30
0/90/0 20 10
GFRC Mortar
0 0.000
0.004 0.008 0.012 Strain, mm/mm
0.016
Lamina stacking optimization for strength and toughness 60 50
Unidirectional
Stress, MPa
40
[0/-45/45/90]s
30 20 10
[45/-45]s
0 0.000
0.005
0.010 0.015 0.020 Strain, mm/mm
0.025
Flexural response of unidirectional and angle ply composite laminates
Equivalent Flexural Stress, MPa
40
30
Unidirectional
20
0-90-0
10 Paste 0
0
4
4.8% Continuous AR Glass Fibers
8 12 16 Deflection, mm
20
Polypropylene Fiber based Unidirectional Composites 20
Polypropylene Based cementitious laminates with 7% volume fraction of continuous fibers.
Stress (MPa)
16
PP #5
12 8 4 0 0.00 0.02 0.04 0.06 0.08 0.10 Strain (mm/mm)
Microcrack Toughening Mechanisms
Cracked Laminate
Distributed microcracking in unidirectional composite
Weak Interfaces Result in Strong & Tough Composites
Interfacial Microcracking in between plies is a toughening Mechanism
Various stages of cracking
Fabric Reinforced Cement composites 25
80 BT-GNSP21
Stress, MPa
60 15 40 10 20
5
0
0
0.01
0.02
Strain, mm/mm
0.03
0 0.04
Crack Spacing, mm
20
Crack Spacing for AR-Glass and Polyethylene Fabrics 100
Crack Spacing, mm
80
AR Glass Fabric Polyethylene
60 AR-Glass fabric
40 20 0
0
0.01 0.02 0.03 Strain, mm/mm
0.04 Polyethylene Fabric
Stiffness degradation and Crack Spacing Relationship 1000
Tangent Stiffness, MPa
Glass Fabrics 100
10
1 80
Polyethylene Fabric
60 40 20 Crack Spacing, mm
0
Scope of Presentation
Introduction Pultrusion Based Cement Composite Systems Theoretical Aspects of Formulation
Elastic response Traditional Composite laminate approach
Composite Elastic Properties
Rule of Mixtures for longitudinal stiffness:
E1( ω ) = E f V f + Em ( ω )( 1 − V f )
Halpin-Tsai equations for transverse modulus E2 , G12, and ν12
E2 ( ω ) =
Em ( ω )( 1 + ξ ηV f ) 1 − ηV f
η=
E f − Em ( ω ) E f + ξ Em ( ω )
Incremental Approach
Unidirectional approach
∆ε ij = S jki ∆σ k
( )
σ = S i k
i −1 jk
or in matrix form:
σ1 S11 σ k = σ 2 = S 21 τ12 i 0 S11 =
∆ε ij + σik−1
1 E1( ω )
S12 = −
S12 S 22 0 υ12 E1( ω )
−1
0 ∆ε1 σ1 0 ∆ε 2 + σ 2 S 66 ∆γ12 i τ12 i −1 S 22 =
1 E2 ( ω )
S 66 =
1 G12 ( ω )
Q11 Q12 0 Sij−1= Q21 Q22 0 0 0 Q66
Orientation Effects and Stress Transformation σ Cos 2θ σ 1 x 2θ σ = σ = T Sin ij y 2 τ − Sinθ Cosθ τ 12 xy
1 R = 0 0
0 0 1 0 0 2
Sin2θ Cos 2θ Sinθ Cosθ
− 2Sinθ Cosθ Cos 2θ −Sin2θ
2Sinθ Cosθ
σ x σ y τ xy
m=n θn
Qij = Tij R Qij R Tij−1 hm-1
n ) A = ∑ Qijm( h − h ij m m−1 m=1 1 n 2 − h2 ) Bij = ∑ Qijm( hm m−1 2 m=1 1 n D = ∑ Qijm( h3 − h3 ) ij 3 m m−1 m=1
hm
m=1 θ1
Incremental Approach to Laminate Formulation Piecewise linear approach: ΔN ΔM
=
A B
B D
0 Δε Δκ
n A = ∑ Qijm( h − h ) m m−1 ij m=1 1 n 2 − h2 ) B = ∑ Qijm( hm m−1 ij 2 m=1 1 n D = ∑ Qijm( h3 − h3 ) m m−1 ij 3 m=1
For uniaxial loads in a symmetric lamina (B=0) : For bending only in a symmetric lamina (B=0) :
N or M
ε or δ
ΔN = A Δε 0
ΔM = D Δκ
Modes of failure σ1t
σ1c
τ12
σ 2t
σ 2c
τ 23
Initial Failure Criteria
F ( σ1 , σ 2 , τ12 ) = 1
General Yield surface Strength criteria σ1 ≥ σ1fu Tsai-Wu Criteria
σ 2 ≥ σ 2fu
τ12 ≥ τ12fu
F ( σ1 , σ 2 , τ12 ) = F11σ1 + 2 F11σ1σ 2 + F22 σ 2 + F66 σ12 + F1σ1 + F2 σ 2 = 1 2
F1 =
R := −
1 1 − σ1t σ1c
F2 =
σt1 := 50
σc2 := 15
σt2 := 10
σ12 := 5
σc1 := 15
σ13 := 5
1 1 − σ 2t σ 2 c
2
F11 =
1 σ 1t σ 1c
σ23 := 8
1 1 2 2 σ − σ σ σ + σ + σ + 150 2 112500 1 2 150 1 30 2 750 1 7
1
1
2
F22 =
1 σ 2t σ 2c
F66 =
1 2 σ 6u
Scope of Presentation
Introduction Pultrusion Based Cement Composite Systems Theoretical Aspects of Formulation
Elastic response Traditional Composite laminate approach
Non-linear matrix response
Cracking Softening
Three zones of stress-strain response for the matrix
Elastic matrix
Cracking matrix
failure by means of the yield surface criterion, ultimate strength, Tsai-Wu reduced stiffness, model the stiffness degradation use a scalar damage variable stress-strain response obtained using Nemat-Nasser and Hori’s approach.
σ 1
2 3
Softening matrix, distributed cracking
strain-softening response, Karihaloo, and Hori’s softening Model
ε
Cracking Matrix- Zone 2 Stress
Damage B
1
ω A
ω O
0
0 Matrix parallel cracking
Strain
Damage Evolution Law-Zone 2 ω0 ω = β ω α ε ε + ( − ) um 1 0 k i
∀ f (σ 1 , σ 2 ,τ 12 ) < 1 , 0 ≤ ω i ≤ 1 ∀ f (σ 1 , σ 2 ,τ 12 ) = 1 1.00
Em =
Em 1+
16 ω i (1 - υ m2 ) 3
β = 0.8
0.80 E( ω) / E0
Karihaloo and Fu, 1990 α= 0.16, β= 2.3 εum= ultimate strain at failure for uniaxial conditions
β = 0.6
0.60
0.40
0.000 Nemat Nasser and Hori, Micromechanics: overall properties of Heterogeneous Materials, 1993
β = 0.4
0.002 0.004 Axial Strain, mm/mm
0.006
Computations Within a load step Compute overall stiffness Calculate the strains, stresses, curvature, and elongation Check for the failure of the entire lamina using ultimate strength update the structure
Check for the failure in the components of the lamina using the failure criteria (Tsai-Wu or ultimate strength Theory) update the lamina properties
Initialize all the geometrical and historical parameters for each lamina. Impose the strain distribution, update the stresses in this increment
∆N A( Q( ω )) ∆M = B ( Q( ω ))
B ( Q( ω )) D ( Q( ω ))
∆ε 0 ∆κ
[ ]
N xj ,i = N xj ,i −1+∆N xj ,i = N xj ,i −1+ [A( Q( ω ))] i ∆ε 0
M xj ,i =M xj ,i −1+∆M xj ,i =M xj ,i −1+ [D (Q(ω ))] i [∆κ ]
Ultimate Failure & Strain Softening Response 1.0
σfu is ultimate strength of fiber •Softening Response Post peak as a fraction of ultimate matrix stress, Model by Karihaloo and Fu, 1990
σ/ f t
•Ultimate FailureFailure Criteria for each lamina: σun = max(VfσfuCos2θm, σt2)
H = 0.25 gage length (in meters) E0= 30000 MPa εB= 200e-6 εA= 0.3* ft/ E0 ω0= εB * H Cd0=.16*(1-εA/ εB ) 2.3 w0=16*Cd0 / 9
0.8
0.6
0.4
0.2 0.0000
0.0001
0.0002 0.0003 Crack Opening,
0.0004
Scope of Presentation
Introduction Pultrusion Based Cement Composite Systems Theoretical Aspects of Formulation
Non-linear matrix response
Elastic response Traditional Composite laminate approach Cracking Softening
Case Studies
Glass Fiber/ carbon fiber epoxy composites Cement based Glass fiber composite laminate Polypropylene fiber cement composites Fabric composites Comparison with Experimental results Repair and retrofit approach
Glass-epoxy Composites
Unidirectional Glass-Epoxy and Woven Carbon-Epoxy Composites Vf (%)
Strength (MPa)
Elastic Modulus (GPa)
Glass Fiber epoxy, 0º, 90º
45
σfu = 1654 σmu = 70
Ef =72, υf = 0.2 Em= 4.0, υm=.18
Woven Carbon Textile
50
σt1=35-70
Ef = 100, υm=.25 Em= 6.0, υm=.18
Glass-Epoxy composites
Woven Carbon
Experiment
Simulation
60 Theory
40
Load, KN
Load, KN
0 degree
40
Sample 1
Vf =.5 ξ =0.2 Em= 6000. MPa Ef= 100000. MPa υm=.18 υf =0.25
20
20 90 degree Theory Experiment
0 0.00
2.00 4.00 6.00 Elongation, mm
8.00
0 0.00
Sample 2
0.20
0.40
0.60
0.80
Clip Gage Elongation, mm
1.00
Unidirectional and 0/90/0 fiber compositeseffect of fiber volume fraction
Vf = 6%
400 Vf = 0%
σt1= 10 MPa σt2= 5 MPa σc1= 40 MPa σc2= 40 MPa τ12= 5 MPa τ23= 5 MPa E m= 30000 ν m= 0.18
200
500
400
0/90/0 Glass Fiber Composites Vf = 6%
Nominal Load, N/mm
Nominal Force, N/mm
600
Unidirectional Glass Fiber Composites
300 Vf = 4%
σt1= 10 MPa σt2= 5 MPa σc1= 40 MPa σc2= 40 MPa τ12= 5 MPa τ23= 5 MPa E m= 30000 ν m= 0.18
200 Vf = 2%
Vf = 2%
100 Vf =0%
Vf = 0%
0 0.000
0.001 0.002 0.003 0.004 Axial Strain, mm/mm
0
0.005
0.000
0.001 0.002 0.003 Axial Strain, mm/mm
0.004
Response of a 6 stack 0/90/0 lamina Strain Distribution
150 Nominal Stress, MPa
10 8 6
z,mm
4 2 0 -2
100
50
-4 -6 -8 -10 0
0.5
1
1.5
2
2.5
3
0
1.5
2
2.5
3
3.5 x 10
10
Transverse Stress
6
8 6 4
z,mm
4 2
z,mm
1
-3
x 10
8
0
2 0
-2
-2 -4
-4
-6
-6
-8
-8
-10 0
0.5
-3
mm/mm x
Stress 10 Distribution
0
3.5
-10 2
4
6
MPa
8
10
12
0
0.5
1
1.5
2
MPa
2.5
3
3.5
4
Response of a 6 stack 0/90/0 lamina
Response of an 8 stack [0/45/-45/90/90]s lamina Strain Distribution 10 8 6
140
2 0
120
-2 -4 -6 -8 -10 0
0.5
1
1.5
2
mm/mm x Stress Distribution 10
2.5
3
3.5 x 10
-3
8 6
Nominal stress, MPa
z,mm
4
100
80
60
40
z,mm
4
20
2 0
0
-2
0
0.5
1
1.5
2
2.5
-4
Nominal strain, mm/mm
-6 -8 -10 0
5
10
15
MPa
20
25
3
3.5 -3
x 10
Response of an 8 stack [0/45/-45/90/90]s lamina
[0/45/-45/90/90]s Composites - Effect of Vf 200 [0/45/-45/90/90]s Vf =6%
Nominal Load, N/mm
160 120
Vf = 4% Vf =2%
80
σt1= 10 MPa σt2= 5 MPa σc1= 40 MPa σc2= 40 MPa τ12= 5 MPa τ23= 5 MPa E m= 30000 ν m= 0.18
Vf = 0%
40 0 0.000
0.001 0.002 0.003 Axial Strain, mm/mm
0.004
Polypropylene Fiber Composites Effect of Vf
2000
σt1= 5 MPa σt2= 5 MPa σc1= 40 MPa σc2= 40 MPa τ12= 5 MPa τ23= 5 MPa
40 Unidrectional Polypropylene Fiber Composites
Vf = 20%
Vf = 15%
1000
Vf = 10%
0 0.00
ω 0= 3.5e-4 Em = 30000 MPa Ef = 12000 MPa νm = 0.18 νf = 0.25
0.02 0.04 Axial Strain, mm/mm
0.06
Nominal Stress, MPa
Nominal Load, N/mm
3000
Effect of Lamina Orientation
30
ω 0= 3.5e-4 Em = 30000 MPa Ef = 8000 MPa νm = 0.18 νf = 0.25 σt1 = 5 MPa
Vf = 6%
unidirectional
20 0/90/0
10 90/0/90
0 0.00
0.02 0.04 Axial Strain, mm/mm
0.06
Comparison With Experimental Results of unidirectional and 0/90/0 composites 60 50 40 Stress, 30 MPa 20
Unidirectional Experiment Theory
Theory
[0/90]s
Experiment Em = 30000 Ef = 70000 Vf = 5% νm = 0.18
10 0 0.000
σt1= 10 MPa σt2= 5 MPa σc1= 40 MPa σc2= 40 MPa τ12= 5 MPa τ23= 5 MPa
0.005 Strain, mm/mm
0.010
Comparison of PPFRC with Experiments
Model Simulation
Stress, MPa
15
Damage Evolution Law
ω ik = ω 0 + α (ε1 − ε um ) β ω 0 = 0.,α = 5., β = 0.8
10
Experiments, Pivacek, Haupt, and Mobasher, 1998
5
Vf = 6% Em = 30000 MPa Polypropylene Ef = 8000 MPa Fiber Composites νm = 0.18 νf = 0.25 σt1 = 5 MPa w0= 3.5e-4 Softening Coefficient
0.000
0.005
0.010
Strain, mm/mm
0.015
Flexural Response of Glass Fiber Composite Laminates Nominal Moment, N-mm/mm
1600
1200
σt1= 5 MPa σt2= 5 MPa σc1= 10 MPa σc2= 10 MPa τ12= 5 MPa τ23= 5 MPa
Vf = 6%
Vf = 4%
800
Vf = 2%
400
0 0.0000
Vf =0%
Em = 30000 Ef = 70000 νm = 0.18
0.0001 0.0002 0.0003 Curvature, 1/mm
0.0004
Response of an 8 stack unidirectional lamina to Flexure Strain Distribution
a b
10 8
c
-2
c
-4 -6 -8 -10-3
b
4
a
b
c
a
6 z,mm
4 2 0
2 0
c
σxx
-2 -4 -6
a
b
-8
-2
-1
0 1 mm/mm
-10 -10
2 3 x 10-3
-5
0
10 5 MPa
15
20
25
10000 8
b
a
6
Moment/unit width, N mm /mm
10
c
4 z,mm
z,mm
10 8 6
2 0 -2 -4
c
σyy
b
-6
a
-8 -10
-4
-3
-2
-1
0
MPa
1
2
3
4
5
9000 8000
c
7000
b
6000 5000 4000 3000 2000 1000 0
a 0
0.5 1.5 1 Curvature, 1/mm
2
2.5
3
3.5 x 10-4
Flexural Load-Deflection Strain Distribution
sigmax
0.25
0.25
0.2
0.15
0.1
0.1
0.05
0.05
z,mm
0.15
z,mm
Stress Distribution 8 layers 0 degree
0.2
0
0 -0.05
-0.05
Strain Distribution
-0.1 -0.15
-0.1 -0.15 -0.2
-0.2 -0.25 -1
-0.5
0
0.5 1 mm/mm
1.5
2
-0.25 -4000
2.5
-3000
-2000
-1000
0
1000
2000
3000
psi
-3
x 10
sigmax 0.25
250
0.2 0.15
200
0 degree
0.1 0.05
force, lbs
z,mm
150
0
Stress Distribution 8 layers [02/902]s
-0.05
100
[02/902]s
50
-0.1 -0.15 -0.2 -0.25 -4000
0
0
1
2
3
4 5 deflection, in
6
7
8
9 -3
x 10
-3000
-2000
-1000
0 psi
1000
2000
3000
Composite Retrofit - Reinforced Concrete beamGlass Epoxy Bonded Laminate Concrete Steel-concrete layer Concrete Glass-epoxy material
Young’s Modulus psi
Posisson’s ratio
Compressive Strength, psi
Tensile strength, Psi
Coefficient of Thermal expansion
epoxy
6.e5
0.28
20000
6000
2.6e-6
glass fibers
11.5e6
0.25
200000
200000
1.e-6
paste
3.e6
0.17
8000.
500.
1.e-6
steel
29.e6
0.3
36000
36000
6.e-6
Stress Distribution in Flexural Sample
Vs= 0.12 tgl-ep
Moment Curvature computations of a Retrofitted RC beam 60000
Composite Retrofit
Vg= 0.4
Moment, Lb-in
Vs= 0.12
tgl-ep= 0.2"
12”
40000
tgl-ep= 0.1"
10” tgl-ep= 0.0"
20000
tgl-ep 0
0
0.0001
0.0002
Curvature, 1/in
0.0003
0.0004
Modeling of Fabrics
Crack Spacing
Debonding Length
Pullout Slip response
Conclusions
Theoretical models based on composite laminate Theory can be used to predict the mechanical response of various cement based composite systems. Combination of damage mechanics based methods and composite laminate theory can result in a useful method to evaluate the response of various fiber matrix, geometry, and loading combinations. The formulation is applicable to a wide range of materials studied including unidirectional, angle ply, sandwich, fabric reinforced, and also retrofit composite systems.