A SPREADSHEET-BASED INVERSE ANALYSIS PROCEDURE FOR FLEXURAL SPECIMENS

-STRAIN SOFTENING OR HARDENING SAMPLES-

1
 Contents
Page

Introduction ................................................................................................................... 3

Chapter

1 Model .......................................................................................................... 5
1.1 Derivation of the Moment-Curvature Relationship ................................. 5
1.2 Algorithm to Predict Load Deflection Response ..................................... 9
1.3 Determination of Deflection Response ................................................10
2 Flexural Modeling ........................................................................................13
2.1 Four Point Bending – Strain Hardening ...............................................13
2.2 Other Generated Data .......................................................................21
2.3 Four Point Bending – Strain Softening.................................................23
2.4 Other Generated Data .......................................................................31
3 Flexural and Tensile Modeling .......................................................................34
3.1 Flexural Bending with Tensile Data .....................................................34
3.2 Other Reported Data .........................................................................38
4 Flexural Modeling with Residual Strength Comparison .....................................40
4.1 Back Calculation with Residual Strength Comparison ............................40

References ..................................................................................................42

2
 Introduction

The uniqueness of the spreadsheet format is that adjustments to the curve fitting parameters are

updated in real time, reducing analysis time and allowing a more accurate simulation. Using a model

the parameters can be manipulated in a spreadsheet environment to model and best fit

an experimental load deflection response. Units are not necessary as long as they are consistent. For

English units ensure that all inputs are in inches, pounds force, or psi. For SI units ensure all inputs are in

millimeters, Newtons or MPa. This rule of consistent units applies to every entry (including the sample

dimensions) in the spreadsheet. The sample values that are given in the examples are only related to this

specific sample set, values for input parameters and material properties will vary with respect to the

shape of the experimental load deflection curve. This user guide uses the ASTM C-1609 (or 4-point bend

test) as an example, the 3-point bend test can also be simulated using this spreadsheet by changing the

test method option from 4 to 3 point bending (see figure 3).

There are 2 variations of this spreadsheet that have been developed the first includes the simultaneous

modeling of flexural and tensile experimental data. The second is the prediction of associative JCI-SF4,

RILEM TC 162-TDF and ASTM C-1609 standard residual strength parameters. The comparison of residual

strength parameters from different standards provides a far more accurate determination of the residual

strength of the FRC.

This user guide uses the following versions of the back-calculation spreadsheet. The sample data has

been pre-loaded into the spreadsheet in order familiarize the user with the effect of changing the model

parameters on the simulation curves. It can also be noted that parameters are interchangeable. If for

example; if a fit is done using the standard Load-Deflection spreadsheet (v1.7) and you need the residual

strength parameters for JCI, RILEM and ASTM, just input the parameters from v1.7 into v2.5 to generate

the residual strengths. Because the spreadsheet uses the simulated response to generate the data, the

experimental data is needed in this case, making the transition quick and straight forward.

3
Below are the associative spreadsheets included with this package;

Load-Deflection - FRC_Back_Calculation_Spreadsheet_LD_v1.7_(Dist).xlsx

Load-Deflection and Tension - FRC_Back_Calculation_Spreadsheet_TENSION_v1.5_(Dist).xlsx

Load-Deflection Standards - FRC_Back_Calculation_Spreadsheet_Std_v2.5_(Dist).xlsx

4
 CHAPTER 1

Model

1.1 Derivation of the Moment-Curvature Relationship

Figure 1 presents a constitutive model for homogenized strain softening and hardening fiber reinforced

concrete. As shown in Figure 1(a), the linear portion of an elastic-perfectly-plastic compressive stress-

strain response terminates at yield point (cy, cy) and remains constant at compressive yield stress cy

until the ultimate compressive strain εcr. The tension model in figure 1(b) is described by a trilinear

response with an elastic range defined by E, and then post cracking modulus Ecr. By setting Ecr to either a

negative or a positive value, the same model can be used to simulate strain softening or strain hardening

materials. The third region in the tensile response is a constant stress range defined with stress cst in

the post crack region. The constant stress level can be set to any value at the transition strain, resulting

in a continuous or discontinuous stress response. Two strain measures are used to define the first

cracking and transition strains (cr, trn). The tensile response terminates at the ultimate tensile strain

level of tu. The stress-strain relationship for compression and tension can be expressed as:

 Ec c 0   c   cy

 c  c    Ec cy  cy   c   cu

0
  c   cu
(1)

5
  E t 0   t   cr

E +E      cr   t   trn
 t  t    cr cr t cr
  E cr  trn   t   tu
0  t   tu
 (2)

where, c , t , c and t, are compressive and tensile stresses and strains respectively. In order to

derive the closed form solutions for moment-curvature response in non-dimensional forms, the material

parameters shown in Figure 1 (a)&(b) are defined as a combination of two intrinsic material parameters:

the first cracking tensile strain cr and tensile modulus E in addition to seven normalized parameters with

respect to E and cr as shown in Eqs (3)-(8).

 cy  trn  
 ;  ; tu  tu ; cu  cu
 cr  cr  cr  cr (3)

Ec Ecr (1   )
 ; = 
E E (  1) (4)

 cst

E cr (5)


crit 
3  1 (6)

6
Figure 1 - Material model for homogenized fiber reinforced concrete (a) compression model and (b)

tension model.

The normalized tensile strain at the bottom fiber and compressive strain at the top fiber are defined

as:

 tbot  ctop
 ; 
 cr  cr (7)

They are linearly related through the normalized neutral axis parameter, k.

 cr  cr k
 or   
kd d  kd 1 k (8)

Substitution of all normalized parameters defined in Eqs. (3) - (5) into Eq. (1)&(2) results in the following

normalized stress strain models:

7
  0   1
 0   
 c     t    1    1 1   
      cu 
E cr  E cr       tu
0 cu   0
 tu   (9)

In the derivation of moment curvature diagram for a rectangular cross section with a width b and depth

d, the Kirchhoff hypothesis of plane section remaining plane for flexural loading is applied. By assuming

linear strain distribution across the depth and ignoring shear deformation, the stress- strain relationships

in Figure 1(a)&(b) are used to obtain the stress distribution across the cross section at three stages of

imposed tensile strain: 0<<1, 1<<and  < < tu . For stage 2 and 3 there are two possible

scenarios: the compressive strain at top fiber is either elastic (0<  < or plastic ( <  < cu). The

neutral axis depth ratio k is then found by solving equilibrium of forces. Finally the moment capacity is

calculated by taking tension and compression forces around the neutral axis and the corresponding

curvature is obtained by dividing the top compressive strain with the neutral axis depth. Moment M and

curvature  are then normalized with their cracking moment Mcr and cracking curvature cr to obtained a

normalized moment M’ and curvature ’, respectively. Note that, from now on, the primed terms (M’ or

’ ) refer to the normalized quantities with respect to cracking moment Mcr or cracking curvature cr. For

example M’ = M/Mcr and ’ =/cr . The expressions for calculating the moment curvature and neutral

axis for all stages of applied strain are given in Eq. (10)&(11) and Table 1.

1
M ( , k ,  ,  )  M cr M '( , k ,  ,  ) M cr  bd 2 E cr (10)
6

2 cr
 ( , k ,  ,  )  cr  '( , k ,  ,  ) cr  (11)
d

8
Table 1 - Neutral axis parameter k, normalized moment M’ and normalized curvature ’ for each stage

of normalized tensile strain at bottom fiber (

For a given set of material parameters and dimension of the beam section, the moment curvature

diagram can be generated by substituting an incremental normalized top compressive strain  from zero

up to failure. Two possible moment curvature responses, deflection hardening (  > crit) and deflection

softening ( < crit), with the critical value for normalized post peak tensile strength is given in Eq. (6).

1.2 Algorithm to Predict Load Deflection Response

The load–deflection response of a beam can be obtained by using the moment–curvature response, crack

localization rules, and moment-area method as follows.

1. For a given cross section and material properties, the normalized tensile strain at the bottom fiber β is

incrementally imposed to generate the moment–curvature response using Eqs. (10), (11), and the

expressions given in Table 1. For each value of β in stage 2 and 3, the condition for compressive stress

9
λ < ω or λ > ω is verified in advance of moment–curvature calculation.

2. Since a moment curvature diagram determines the maximum load allowed on a beam section, the

discrete moments along the diagram are used to calculate the applied load vector P = 2M/S. Where S is a

spacing between the support and loading point, S=L/2 for three point bending and S=L/3 for four point

bending.

3. The beam is segmented into finite sections. For a given load step, static equilibrium is used to

calculate moment distribution along the beam and moment–curvature relationship along with crack

localization rules to identify the curvature.

4. The deflection at mid-span is calculated by numerical moment-area method of discrete curvature

between the support and mid-span. This procedure is applied at each load step to until a complete load

deflection response is obtained. A simplified procedure for direct calculation of the deflection is presented

in the next section.

1.3 Determination of Deflection Response

With the moment-area method, the curvature diagram of a half-model according to the internal

moment in Figure 2 (c-e) is divided into several areas. By taking the moment of the areas around the left

support, the mid-span deflection is obtained. A set of equations for calculating the mid-span deflection δ

of the three-point bending at the first cracking (δcr), at ultimate (δu) when material has μ>μcrit, and at

ultimate when material has μ<μcrit are presented in Eqs.12 (a-c).

1 2
 cr  L cr (12.a)
12

L2
2 
u   2M u2  M u M cr  M cr2  u   M u2  M u M cr  cr    crit (12.b)
24M u 

10
 u Lp
u 
8
 2L  L   M12M L  L  2L 
p
u cr
p   crit (12.c)
cr

Similarly, a set of equations for four point bending can be written as:

23 2
 cr  L cr (13.a)
216

L2
2 
u   23M u2  4M u M cr  4M cr2  u   4M u2  4M u M cr  cr    crit (13.b)
216M u 

5L2u M u L2cr
u     crit (13.c)
72 27 M cr

11
 P P/ P/
(a) 2
22
22
S S S 22 S S
L L
Mu Mu
Mcr
(b) Mcr

cr cr
(c)
  

S S S/2

u u

cr  cr 
(d)

   

M cr  M  M cr  M  S
S 1  cr  S S 1  cr  S
Mu  Mu  Mu  Mu  2

u u
Mu  Mu 
M cr cr M cr cr
(e)  
 

S- Lp/2 Lp/ S S/2

2
Figure 2: Three and four point bending test: (a) Experimental setup; (b) Moment distribution; (c)

Curvature distribution at first bilinear cracking; (d) Curvature distribution at ultimate moment for high

normalized post peak tensile strength (> crit); (e) Curvature distribution at ultimate moment for low

normalized post peak tensile strength (< crit).

12
 CHAPTER 2

Flexural Modeling

2.1 Four Point Bending – Strain Hardening

The four point bending test ASTM-1609 is carried out on standard beam dimensions with load applied at

third points along the span (see Figure 2a). The load deflection curve will show an increase in load

capacity after the linear elastic portion has ended. The hardening continues to increase as shown in

Figure 3 or it can plateau as the sample continues to carry load.

Step 1:

Paste reduced load deflection data (up to 1000 points) into the columns labeled “Load” and “Deflection”.

This will populate the experimental load deflection curve we are trying to model as shown in Figure 3.

13
Figure 3 – Experimental load deflection response, showing deflection hardening characteristics.

Step 2:

Input the sample dimensions and test method. Table 2 shows an example.

Input Definition Value

Test Method Type of test 4
b Width of sample 50
d Depth of sample 25
L Length of span 300

Table 2 – Dimensions and test method definitions.

14
 Model Parameters
Beam Size
Test Method 4 Point Bending
b= 50
d= 25
L= 300

Figure 3 – Beam size and test method inputs.

Step 3:

Determine the best fit for the Young’s modulus for the linear elastic phase. This done by increasing or

decreasing the value of the Elastic modulus (Young’s modulus) E. Fiber reinforced cement (FRC) has a

Young’s modulus in the range of 3000000-5000000 psi (20000-35000 MPa).

Figure 4 – Linear section of load deflection clearly modeled by E=25000 MPa.

15
Step 4:

Now εcr will need to be fitted to the approximate point where the linear elastic behavior ceases and non-

linear behavior begins. This parameter will depend on the type of FRC being tested; cement performance,

fiber type and fiber dosage all contribute to the point where cracking is observed. In this case a value of

0.00013 was used. Figure 5 shows how the simulated curve has changed in response to the change in εcr.

Figure 5– Fitting of εcr to load deflection curve

16
Step 5:

The post cracking slope for strain hardening is represented by E cr is related through the parameter η.

This parameter is dependent on the values of α and μ through equation (4). By manipulating α and μ the

post cracking slope can be fitted to the load deflection curve. The parameter α adjusts the horizontal

location of the transition point, while μ adjusts the vertical position of the transition point and the vertical

position of the tail of the curve. Increasing μ will raise the post crack residual portion of the simulated

curve. In this case the values of α and μ are given in Figure 6. Figure 7 shows the change in the

simulated curve with respect to the parameter changes. The parameter η is automatically calculated per

equation (4) and should reflect the post crack slope Ecr as negative in the softening case and positive in

the hardening case (see Figure 6).

The relationship between the parameters α and μ has other effects on the shape of the simulated curve.

It should be noted that changes made in εcr will also require changes to the parameters α and μ to

realign the simulation curve. The value of Lp automatically populates based on units, crack localization

rules and the type of test method chosen.

17
 Model Parameters
Beam Size
Test Method 4 Point Bending
b= 50
d= 25
L= 300

Don't need Lp 1.00

Material Model
E= 25000
cr 0.00013
α 40
Tension
 0.95
η= 0.0187179
μ= 1.73
βtu= 30

Compression
 10.8
cu = 40

Figure 6 – Parameters α and μ are highlighted

18
 Figure 7 – Change in simulated curve with respect to α and μ.

Step 6:

Now that most of the curve has been fitted we need to determine the ending point to the simulated

curve, βtu is used as the normalized ending point parameter.

19
 Model Parameters
Beam Size
Test Method 4 Point Bending
b= 50
d= 25
L= 300

Don't need Lp 1.00

Material Model
E= 25000
cr 0.00013
α 40
Tension
 0.95
η= 0.0187179
μ= 1.73
βtu= 105

Compression
 10.8
cu = 40

Figure 8 – Parameter inputs with βtu updated to reflect end point.

Load Deflection Experimental

Simulation

1600

1400

1200

1000
Load

800

600

400

200

0
0 1 2 3 4 5 6 7

Deflection

Figure 9 – Load deflection curve response to change in βtu.

20
The parameter βtu is dependent on the user’s preference on when the ending point of the simulation

curve needs to be. In some cases the entire deflection curve is not necessary, through this parameter

adjustments can be made.

Step 7:

The determination of softening or hardening deflection behavior is determined by the parameter μcrit

which is the normalized post peak tensile strength; see equation (6). The parameter ω determines the

cross over point from deflection softening to hardening. In this case;

(10.8)
crit   0.344
(3(10.8)  1)

Since μ=1.73>μcrit we see deflection hardening behavior in the simulated curve. Again, we see the

interaction of the normalized parameters. By adjusting the value of ω (normalized compressive yield

strain), we adjust the value of μcrit which effects which deflection equation is used; see equation (13b-c).

2.2 Other Generated Data

There is a lot of data being modeled in this spreadsheet, with the simulated load deflection curve being

the product of other useful processes. The spreadsheet generates a simulated tensile and compressive

stress strain response under the stress strain worksheet; see Figure (10) and (11). The spreadsheet also

generates the simulated moment curvature response and the stand alone simulated load deflection

response under the moment curvature worksheet; see Figure (12) and (13).

21
 Tension Model Compression Model

6.0000 40

35

Compressive Stress
5.0000

30
Tensile Stress

4.0000
25

3.0000 20

15
2.0000
10
1.0000
5

0.0000 0
0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.000 0.001 0.002 0.003 0.004 0.005 0.006

Tensile Strain Compressive Strain

Figure 10 – Simulated tensile and compressive stress strain response for deflection hardening case.

Beam Size
Test Method 4 Point Bending b= 50.000 L= 300
Don't need Lp 1 d= 25.000 S= 100.0000

Material Model
E= 25000.00
cr = 0.0001300
α 40.00
Tension
γ 0.950 Strain 0.000000 Stress 0.0000
η 0.019 εcr 0.0001300 σcr 3.25
= 1.730 εtrn 0.0052000 σtrn 5.62
tu = 105.00 εtrn 0.0052000 σcst 5.62
εtu 0.0136500 σtu 5.62

Compression
= 10.80 Strain 0.000000 Stress 0.00000
cu = 40.0000 εcy 0.0014040 σcy 33.35
εcu 0.0052000 σcy 33.35

Critical Parameters
μ crit 0.343949045

Figure 11 – Stress strain screenshot of computed parameter values for hardening case.

22
 Normalized Moment Curvature Diagram
Moment Curvature Diagram
5.00 8.E+04

Normalized Moment, M'

4.50 7.E+04
4.00
6.E+04
3.50

Moment, M
3.00 5.E+04
2.50 4.E+04
2.00 3.E+04
1.50
2.E+04
1.00
0.50
1.E+04
0.00 0.E+00
0.00 20.00 40.00 60.00 80.00 0.E+00 2.E-04 4.E-04 6.E-04 8.E-04
Normalized Curvature, ' Curvature, 

Normalized Moment Curvature Moment Curvature

(a) (b)

Figure 12 – Strain hardening (a) normalized moment curvature and (b) moment curvature.

Load Deflection Response

1,600
1,400
1,200
1,000
Load, P

800
600
400
200
0
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000
Deflection, 

Figure 13 – Simulated load deflection response

2.3 Four Point Bending – Strain Softening

The four point bending test ASTM-1609 is carried out on standard beam dimensions with load applied at

third points along the span (see Figure 2a). The load deflection curve will show a decrease in load

23
capacity after the linear elastic portion has ended. The softening continues as shown in Figure 14 or it

can plateau as the sample continues to sustain load as deflection continues.

STEP 1:

Paste reduced load deflection data (up to 1000 points) into the columns labeled “Load” and “Deflection”.

This will populate the experimental load deflection curve we are trying to model as shown in Figure 14.

Load Deflection

Experimental
100000

90000

80000

70000

60000

50000
Load

40000

30000

20000

10000

0
0 1 2 3 4 5 6 7

Deflection

Figure 14 - Experimental load deflection response, showing deflection softening characteristics.

Step 2:

Input the sample dimensions and test method. Table 3 shows an example.

Input Definition Value

Test Method Type of Test 4
b Width of sample 150
d Depth of sample 150
L Length of span 450

24
 Table 3 – Dimensions and test method definitions.

Beam Size
Test Method 4 Point Bending
b= 150
d= 150
L= 450

Figure 15 – Beam size and test method inputs.

Step 3:

Determine the best fit for the Young’s modulus for the linear elastic phase. This done by increasing or

decreasing the value of the Elastic modulus (Young’s modulus) E. Fiber reinforced cement (FRC) has a

Young’s modulus in the range of 3000000-5000000 psi (20000-35000 MPa).

25
 Load Deflection

Experimental simulated
100000

90000

80000

70000

60000
Load

50000

40000

30000

20000

10000

0
0 1 2 3 4 5 6 7

Deflection

Figure 16 – Linear section of load deflection clearly modeled by E=20000 MPa.

Step 4:

Now εcr will need to be fitted to the approximate point where the linear elastic behavior ceases and non-

linear behavior begins. This parameter will depend on the type of FRC being tested; cement performance,

fiber type and fiber dosage all contribute to the point where cracking is observed. In this case a value of

0.00013 was used. Figure 17 shows how the simulated curve has changed in response to the change in

εcr.

26
 Load Deflection

Experimental simulated
180000

160000

140000

120000

100000
Load

80000

60000

40000

20000

0
0 1 2 3 4 5 6 7

Deflection

Figure 17– Fitting of εcr to load deflection curve

Step 5:

The post cracking slope for strain hardening is represented by Ecr is related through the parameter η.

This parameter is dependent on the values of α and μ through equation (4). By manipulating α and μ the

post cracking slope can be fitted to the load deflection curve. The parameter α adjusts the horizontal

location of the transition point, while μ adjusts the vertical position of the transition point and the vertical

position of the tail of the curve. Increasing μ will raise the post crack residual portion of the simulated

curve. In this case the values of α and μ are given in Figure 18. Figure 19 shows the change in the

simulated curve with respect to the parameter changes. The parameter η is automatically calculated per

equation (4) and should reflect the post crack slope Ecr of a negative value in the softening case (see

Figure 18).

27
The relationship between the parameters α and μ has other effects on the shape of the simulated curve.

It should be noted that changes made in εcr will also require changes to the parameters α and μ to

realign the simulation curve. The value of Lp automatically populates based on units, crack localization

rules and the type of test method chosen.

Beam Size
Test Method 4 Point Bending
b= 150
d= 150
L= 450

Don't need Lp 1.00

Material Model
E= 20000
cr 0.00026
α 105
Tension
 0.95
η= -0.008365
μ= 0.13
βtu= 105

Compression
 10.8
cu = 40

Figure 18 – Parameters α and μ are highlighted

28
 Load Deflection

Experimental simulated
100000

90000

80000

70000

60000
Load

50000

40000

30000

20000

10000

0
0 1 2 3 4 5 6 7

Deflection

Figure 19 – Change in simulated curve with respect to α and μ.

Step 6:

Now that most of the curve has been fitted we need to determine the ending point to the simulated

curve, βtu is used as the normalized ending point parameter.

29
 Model Parameters
Beam Size
Test Method 4 Point Bending
b= 150
d= 150
L= 450

Don't need Lp 1.00

Material Model
E= 20000
cr 0.00026
α 105
Tension
 0.95
η= -0.008365
μ= 0.13
βtu= 235

Compression
 10.8
cu = 40

Figure 20 – Parameter inputs with βtu updated to reflect end point.

Load Deflection

Experimental simulated
100000

90000

80000

70000

60000
Load

50000

40000

30000

20000

10000

0
0 1 2 3 4 5 6 7

Deflection

Figure 21 – Load deflection curve response to change in βtu.

30
The parameter βtu is dependent on the user’s preference on when the ending point of the simulation

curve needs to be. In some cases the entire deflection curve is not necessary, through this parameter

adjustments can be made.

Step 7:

The determination of softening or hardening deflection behavior is determined by the parameter μcrit

which is the normalized post peak tensile strength; see equation (6). The parameter ω determines the

cross over point from deflection softening to hardening. In this case;

(10.8)
crit   0.344
(3(10.8)  1)

Since μ=0.13<μcrit we see deflection softening behavior in the simulated curve. Again, we see the

interaction of the normalized parameters. By adjusting the value of ω (normalized compressive yield

strain), we adjust the value of μcrit which effects which deflection equation is used; see equation (13b-c).

2.4 Other Generated Data

There is a lot of data being modeled in this spreadsheet, with the simulated load deflection curve being

the product of other useful processes. The spreadsheet generates a simulated tensile and compressive

stress strain response under the stress strain worksheet; see Figure (22) and (23). The spreadsheet also

generates the simulated moment curvature response and the stand alone simulated load deflection

response under the moment curvature worksheet; see Figure (24) and (25).

31
 Tension Model Compression Model
6.0000
60

Compressive Stress
5.0000 50
Tensile Stress

4.0000 40

3.0000 30

2.0000 20

1.0000 10

0.0000 0
0.000 0.010 0.020 0.030 0.040 0.050 0.060 0.070 0.000 0.002 0.004 0.006 0.008 0.010 0.012

Tensile Strain Compressive Strain

Figure 22 – Simulated tensile and compressive stress strain response for deflection softening case.

Beam Size
Test Method 4 Point Bending b= 150.000 L= 450
Don't need Lp 1 d= 150.000 S= 150.0000

Material Model
E= 20000.00
cr = 0.0002600
α 105.00
Tension
γ 0.950 Strain 0.000000 Stress 0.0000
η -0.008 εcr 0.0002600 σcr 5.20
= 0.130 εtrn 0.0273000 σtrn 0.68
tu = 235.00 εtrn 0.0273000 σcst 0.68
εtu 0.0611000 σtu 0.68

Compression
= 10.80 Strain 0.000000 Stress 0.00000
cu = 40.0000 εcy 0.0028080 σcy 53.35
εcu 0.0104000 σcy 53.35

Critical Parameters
μ crit 0.343949045

Figure 23 – Stress strain screenshot of computed parameter values for softening case.

32
 Normalized Moment Curvature Diagram
Moment Curvature Diagram
2.50 7.E+06

Normalized Moment, M'

6.E+06
2.00
5.E+06

Moment, M
1.50 4.E+06
3.E+06
1.00
2.E+06
0.50 1.E+06

0.00 0.E+00
0.00 50.00 100.00 150.00 0.E+001.E-042.E-043.E-044.E-045.E-04

Normalized Curvature, ' Curvature, 

Normalized Moment Curvature Moment Curvature
(a) (b)

Figure 24 – Strain softening (a) normalized moment curvature and (b) moment curvature.

Load Deflection Response

100,000
90,000
80,000
70,000
60,000
Load, P

50,000
40,000
30,000
20,000
10,000
0
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000
Deflection, 

Figure 25 – Simulated load deflection response

33
 CHAPTER 3

Flexural and Tensile Modeling

3.1 – Flexural Bending with Tensile Data

By utilizing the same modeling concepts as outlined in section 2.1 the inclusion of tensile

data does not affect the functionality of the way the spreadsheet simulates the load deflection

response. The following example uses the ASTM C-1609 test data and tensile data in the fitting

of both curves.

STEP 1:

Upload flexural load-deflection and tensile stress strain data into the appropriate columns under

the “RAW DATA” worksheet tab.

Units are: in, lbs, psi

or: mm, N, Mpa
Please Use Units Consistently

Experimental Data
Tensile
Deflection Load Strain Stress
0 0 0 0.065102
0.005929 3014 8.58E-06 0.390614
0.01186 6029 8.58E-06 0.651024
0.01779 9043 3.43E-05 1.04164
0.02371 12060 4.29E-05 1.32375
0.02964 15070 4.29E-05 1.62756
0.03557 18090 4.29E-05 1.99647
0.0415 21100 6.01E-05 2.31113
0.04743 24120 8.58E-05 2.6475
0.05336 27130 0.000103 2.87535

Figure 3.1 – Raw data input

34
STEP 2:

Input flexural test method and dimensions of flexural sample. The user definition of the post

cracking slope η, can be input here by placing a “Y” in the appropriate cell, a “N” will source a

derived η value from α and μ (see equation 4).

Model Parameters

User Defined η (Y or N)?

Y
User Defined η = 0.015

Beam Size
Test Method 4 Point Bending
b= 150
d= 150
L= 450

Figure 3.2 – Flexural test method, sample dimensions and user definition of parameter η.

35
STEP 3:

Manipulate Young’s Modulus to simulate the initial slope of the load deflection curve.

Care must be taken to observe the effect on the simulated stress-strain curve also.

Figure 3.3 shows the dual effect of simultaneously modeling the two curves.

120000
Load Deflection
Experimental
100000
Simulation
80000
Load

60000

40000

20000

0
0 0.5 1 1.5 2 2.5

Deflection
Tension Model
Experimental 6

Simulation

5
Tensile Stress

0
0.000 0.002 0.004 0.006 0.008 0.010 0.012

Tensile Strain

Figure 3.3 – Simultaneous load-deflection and tensile stress-strain response.

36
STEP 4:

Changing the values of α and μ will alter the post peak orientation of both curves (as described

in chapter 2). However care must be taken to ensure a close enough fit for both curves.

120000
Load Deflection
Experimental
100000
Simulation
80000
Load

60000

40000

20000

0
0 0.5 1 1.5 2 2.5

Deflection

Tension Model
Experimental 7

Simulation
6
Tensile Stress

0
0.000 0.002 0.004 0.006 0.008 0.010 0.012

Tensile Strain

Figure 3.4 – Final curve fit for both load-deflection and tensile stress-strain response.

The simulated responses shown in Figure 3.4 are determined after considering how close to

each experimental curve to go. Experimental tensile data has previously proved to under

estimate the actual tensile response, consequently some judgment decisions must be made

37
when determining how close to fit each curve. Figure 3.5 shows the parameters (in SI) used to

produce the simulated load-deflection and stress-strain responses in Figure 3.4.

Model Parameters

User Defined η (Y or N)?

Y
User Defined η = 0.015

Beam Size
Test Method 4 Point Bending
b= 150
d= 150
L= 450

Don't need Lp 1.00

Material Model
E= 23000
cr 0.0002
α 27
Tension
 0.95
η= 0.015
μ= 0.95
βtu= 50

Compression
 100
cu = 100

Figure 3.5 – Model parameters for the dual load-deflection and tensile stress-strain response.

3.2 Other Reported Data

In addition to load-deflection, stress-strain and moment curvature plots, the spreadsheet

produces experimental data analysis. Back calculation parameters are also produced and shown

in Table 3.1. The spreadsheet automatically detects the units (SI or English) based on the span

length of the sample, if L > 150 then it takes the units to be SI (flexural samples are rarely 150

inches long).

38
 Output Data - Experimental
Flexural Toughness = 142240 (N-mm)
Bending Strength = 14.04608 (Mpa)
Flexural Stiffness = 584711.5 (Mpa)

Load @ 1st Crack = 34057.62 (N)

Defl @ 1st Crack = 0.058247 (mm)
Max Flex Load = 105345.6 (N)
Defl. @ Max Load = 0.947497 (mm)
Deflection Capacity = 1.604958 (mm)

Output Data - Back-Calc Parameters

Tensile Toughness = 0.043884 (Mpa)
εcr (μstr) = 200
E = 23000 (Mpa)
σcr = 4.60 (Mpa)
μ= 0.95
α= 27
εtrn (μstr) = 5400
εtu (μstr) = 10000
μσcr = 4.37 (Mpa)

Table 3.1 – Summary of experimental analysis and back-calculation parameters.

39
 CHAPTER 4

Flexural Modeling with Residual Strength Comparison

4.1 Back Calculation with Residual Strength Comparison

It may be necessary for the residual strength from the JCI-SF4, RILEM TC 164-TDF or

ASTM C-1609 to be compared to the described model in an attempt to take into account the

shifting neutral axis of a cracking sample. For a set of samples, the residual strength

parameters can be plotted and the linear relationship developed as shown in [4].

STEP 1:

The process for simulating the experimental load deflection data is identical to the previous

methods. Upload the experimental load-deflection data under the “RAW DATA” worksheet tab.

Then follow the process of manipulating the simulated curve until a satisfactory simulation is

accomplished.

STEP 2:

The difference in this version of the spreadsheet is under the “STANDARDS” worksheet tab.

This is where the different residual strength parameters for each of the outlined standards is

calculated and compared with the ASU model. Figure 4.1 shows the residual strength

parameters in appropriate units. A summary of these results are shown in figure 4.2. This

summary table makes copying large sets of analysis data into tables easier. It is interesting to

note that the RILEM residual strength is calculated in SI, then converted if needed into English

units.

40
Total Tensile Toughness (Gf) 0.048254 (Mpa)
Total Flexural Toughness (GF) 113310.3 (N-mm) RILEM (SI) JCI - SF4
L/150=δtb 1.293098 (mm)
Prop. Limit Load (FL)= 34057.62 (N) Rilem Tuff = 113340.0356 (N-mm) Tb= 113310.3 (N-mm)
δL= 0.058247 (mm) Rilem δ3 = 2.708246878 (mm)
Governing δ3= 1.084850881 (mm)

Defl Capac 1.084851 mm Df(BZ,3) 107239.5177 (N-mm)

feq,3 13.1802468 (MPa) σb = 11.6836 (Mpa)
1911.632681 (psi)

ASTM -1399 (ARS) Residual Strength - ASU

Geometry Defl. (mm) ~Defl. (mm) Load (N)
L= 450 (mm) Defl. Step 1 0.50 0.50 93974.89 μσcr = 4.37 MPa
b= 150 (mm) Defl. Step 2 0.75 0.75 103769.85
h= 150 (mm) Defl. Step 3 1.00 1 96744.27
Defl. Step 4 1.25 1.25 94014.91
Test Factor= 1
Average Load= 97125.98146 (N)
ARS= 12.95013086 (Mpa)

ASTM-1609
Error limit 0.005 L/150=δtb 1.293097759 (mm)
P150 93761.50672 (N)

f150 12.50153423 (MPa)

Figure 4.1 – Determination of residual strength parameters for JCI, RILEM and ASTM standards.

REPORTING
ASTM-1609 ASTM-1399 RILEM JCI-SF4 Proposed
f150 ARS feq,3 σb μσcr
12.502 12.950 13.180 11.684 4.37
Figure 4.2 – Summary table of calculated residual strengths.

41
References

1 Soranakom, C., and Mobasher, B., “Moment-Curvature Response of Strain Softening and Strain

Hardening Cement Based Composites,” Cement and Concrete Composites, Volume 30, Issue 6,

Pages 465-477, July 2008

2 Soranakom C, Mobasher B. Closed-form moment-curvature expressions for homogenized fiber-

reinforced concrete. ACI Material Journal 2007; 104(4):351-9.

3 Soranakom C, Mobasher B. Closed form solutions for flexural response of fiber reinforced

concrete beams. Journal of Engineering Mechanics 2007;133(8):933-41.

4 Bakhshi M, Barsby C, Mobasher B, “Comparative Evaluation of Early Age Toughness Parameters

in Fiber Reinforced Concrete”, Submitted to Materials and Structures, 2011.

42