CHARACTERIZATION OF THE SHEAR BEHAVIOUR OF WOOD USING THE IOSIPESCU TEST

1
CHARACTERIZATION OF THE SHEAR BEHAVIOUR OF WOOD
USING THE IOSIPESCU TEST
University of Trás-os-Montes e Alto Douro Vila
Real, Portugal

• J. C. Xavier

Master of Science Thesis
LMPF seminary - 11/12/2003
2
Plan

• Introduction
• The Iosipescu test
• Numerical simulation of the Variable Span Method
• Numerical simulation of the Iosipescu test
• Experimental work
• Presentation and discussion of the experimental
  results
• General conclusions and future work

3
Introduction

• Wood modelling at the macroscopic level

• Shear properties

• Shear moduli GLR , GLT , GRT .
• Shear strengths SLR , SLT , SRT .

4

• Drawbacks of the standardized tests for the
  identification of the shear properties of wood
  3,4

• Give only the shear properties parallel to the
  fibres
• (shear moduli GLR, GLT and shear
  strengths SLR e SLT).

1. The variable span method 5,6 proposed for the
   determination of EL and GLR (or GLT), is not a
   fundamental test.

1. The failure of the specimen of the shear block
   test 7,8 proposed for the identification of SLR
   and SLT, occurs under stress concentrations.

3 Yoshihara et all.. Journal of Wood Science,
4415-20, 1998. 4 Rammer D.R. e L.A. Soltis.
Res. Pap. FPL-RP-527, FPL, 1994.
5 prEN 408. European Committee for
Standardization, 2000. 6 ASTM D198-94. American
Society for Testing and Materials, 1994.
7 ASTM D143-94. American Society for Testing
and Materials, 1994. 8 NP 623. Portuguese
Standard, 1973.
5

• Among different shear tests for orthotropic
  materials Iosipescu test
• (standard test for synthetic composite
  materials 9).

• Aim of this work

• Investigation of the applicability of the
  Iosipescu shear test for characterizing the shear
  behaviour of wood Pinus Pinaster Ait.

Justifications for the choice of the Iosipescu
test

1. simultaneous identification of the shear modulus
   and the shear strength, in a particular symmetry
   plane.

1. possible application of this test method for all
   the symmetry planes of wood thanks to the small
   size of the specimen.

9 ASTM D 5379-93. American Society for Testing
and Materials, 1993.
6
The Iosipescu test

• Iosipescu specimen 9

9 ASTM D 5379-93. American Society for Testing
and Materials, 1993.
7

• General view of the Iosipescu fixture 9

9 ASTM D 5379-93. American Society for Testing
and Materials, 1993.
8

• Data processing 9

Experimental information
e45º , e45º , P
Engineering shear strain
e6 e45º e 45º
Nominal shear stress
s6 P/A
Apparent shear modulus
Apparent shear strength
9 ASTM D 5379-93. American Society for Testing
and Materials, 1993.
9

• Aspects about the identification of G12

• For an orthotropic material the distributions
  of s6 and e6
• are not homogeneous 10,11.

The correction factors C e S are calculated
through finite element analyses.
10 Pierron F. e A. Vautrin. Composite Science
and Technology, 561-72, 1994. 11 Pierron F.
Journal of Composite Materials, 32(22)1986-2015,
1998.
10

• The distribution of e6 through the thickness of
  the specimen can be heterogeneous due to
  geometrical imperfections of its loading surfaces
  10,11.

• This effect is eliminated by considering e6 as
  the average of the shear strains measured over
  both lateral faces of the specimen 10,11.

11

• Aspects about the identification of S12

• The failure of the specimens occurs under a
  homogeneous stress state although both s6 and s2
  components exist 12-14.

• S12 should be determined through a failure
  criterion.

• s2 component should be calculated from finite
  element analysis, introducing in the model an
  suitable shear constitutive law.

12 Pierron F. e A. Vautrin. Composite Science
and Technology, 57(12)1653-1660, 1997. 13
Pierron F. e A. Vautrin. Journal of Composite
Materials, 31(9)889-895, 1997. 14 Odegard G. e
M. Kumosa. Journal of Composite Materials,
33(21)1981-2001, 1999.
12
Numerical simulation of the variable span method

• Aim

• Investigating the applicability of the variable
  span method 5,6 for validating the Iosipescu
  test.

• Finite element models

• 3D models of the three-point-bending test
  developed in ABAQUS 6.2-1.

• Wood was modelled as
• continuous
• homogeneous
• orthotropic
• linear elastic.

13

• Configuration of the specimens in the
  three-point-bending tests

• Geometrical model used in the finite element
  analyses

14

• Elastic properties used in the numerical models

EL(1) ER(1) ET(1) nLR(1) nLT(1) nRT(1) GLR(2) GLT(2) GRT(2)
(GPa) (GPa) (GPa) (GPa) (GPa) (GPa)
15,13 1,91 1,01 0,47 0,05 0,59 1,11 1,10 0,18

1. Pinus Pinaster Ait. 15.
2. Pinus Tarda L. 16.

• Calibration of the friction coefficient

• Mesh and boundary conditions of the model

15

• Numerical results

• Euler-Bernoulli beam theory

16

• Timoshenko beam theory

f1 f2 f3 f4
EL (GPa) 16,63 (9,9) 16,05 (6,1) 16,01 (5,8) 15,57 (2,9)
GLR (GPa) k 1,2 0,74 (33,6) 1,12 (0,6) 1,22 (9,6) 1,94 (74,6)
k 1,5 0,92 (17,0) 1,39 (25,8) 1,52 (37,0) 2,42 (118,3)
17

• Kinematical assumption of the Timoshenko beam
  theory

• The deflection is the same for each point
  belonging to the same vertical cross section,
  initially perpendicular to the neutral axis.

This assumption is not verified at midspan (AC )
A
B
C
18
Numerical simulation of the Iosipescu test

• Aims

• Determination of the stress and strain fields in
  the central region of the Iosipescu specimen of
  Pinus Pinaster Ait.

• Computation of the correction factors C and S.

• Finite element models

• 2D models developed in ANSYS 7.0 and ABAQUS
  6.2-1.

• The hypothesis and elastic properties are the
  same as the ones used in the numerical simulation
  of the variable span method.

19

• Nominal dimensions of the Iosipescu specimen

• Mesh of the finite elements models

20

• Boundary conditions 17-19

1. base

ANSYS

• iterative
• (LR plane)

1. with contact

ABAQUS
17 Pierron F. PhD, University of Lyon I,
1994. 18 Ho H. et al. Composite Science and
Technology, 46115-128, 1993. 19 Ho H. et al.
Composite Science and Technology, 50355-365,
1994.
21

• Numerical results

• Comparaison and validation of the boundary
  conditions

(LR plane)
22

• Stress and strain fields for the LR specimen

• Stress field in the central region of the
  specimen

• Stress distribution along the vertical line
  between notches

sLR /P/A
sRR /P/A
23

• Strain field over the strain gauge area

24

• Stress and strain fields for the LT specimen

• Stress field in the central region of the
  specimen

• Stress distribution along the vertical line
  between notches

sLT /P/A
sTT /P/A
25

• Strain field over the strain gauge area

26

• Stress and strain fields for the RT specimen

• Stress field in the central region of the
  specimen

• Stress distribution along the vertical line
  between notches

sRT /P/A
sTT /P/A
sRR /P/A
27

• Strain field over the strain gauge area

28

• Calculation of the correction factors C and S

Symmetry planes Correction factors Correction factors Correction factors
for wood C S CS
LR 0,97 0,99 0,95 (4,8)
LT 0,92 0,99 0,91 (8,6)
RT 1,04 0,97 1,01 (0,6)
29
Experimental work

• Preparation of the specimens

• Material wood of Pinus Pinaster Ait. (maritime
  pine), 74-year-old, from Viseu (Portugal).

30

• Iosipescu specimen

LR specimen
LT specimen
RT specimen

• Moisture content 9,5 12,1
• Density 0,537 0,623
• 0/90 strain gauge (CEA-06-125WT-350), bonded on
  both faces of the specimens, with the M-Bond
  AE-10 adhesive.

31

• Experimental procedure

• EMSE fixture 20

Tightness of the wedges with a dynamometrical key
1 Nm
20 Pierron F. Ecole des Mines de Saint-Etienne,
No. 940125, 1994.
32

• Experimental equipment

Controlled displacement rate of 1 mm/min
5 kN Load cell
Data acquisition system HBM SPIDER 8
INSTRON 1125 Universal test machine with the
capacity of 100 kN
Temperature of 23ºC (?1ºC) and relative humidity
of 45 (?5)
33
Presentation and discussion of the experimental
results

• LR specimens

• Typical experimental data measured for the LR
  specimens

Load (N)
Front face (A) Front face (A) Back face (B) Back
face (B)
Linear deformation measured with strain gauges
(me)
34

• Apparent average sLR eLR curves

Average shear stress (MPa)
Average engineering shear strain (me)

• The response of the specimens contain some
  variability.

• The curves are nonlinear the source of such
  nonlinearity can be attributed to 19,21
• (1) the nonlinear behaviour of the
  material
• (2) the geometric nonlinearity
• (3) the nonlinearity due to the contact
  conditions specimen/fixture.

35

• Dispersion of the shear moduli values (GLR , GLR
  and GLR )

a, B
a
a, A
Shear moduli (GPa)
Specimens
GLR GLR GLR
Mean (GPa) 1,41 0,151 1,54 0,181 1,48 0,121
C.V.2 () 14,1 15,0 10,3
a, B
a
a, A
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).
36

• Moisture content (u), density (d) and shear
  moduli (GLR, GLR)

a
c
a
c
Specimens u () d GLR (GPa) GLR (GPa)
1 11,9 0,561 1,33 1,27
2 11,8 0,607 1,59 1,52
3 12,1 0,612 1,57 1,50
4 11,8 0,605 1,62 1,54
5 12,1 0,615 1,48 1,42
6 10,3 0,538 1,32 1,23
7 10,0 0,537 1,22 1,16
8 10,4 0,609 1,53 1,46
9 9,4 0,614 1,65 1,58
Mean 11,1 0,589 1,48 0,121 1,41 0,111
C.V.2 () 9,1 5,6 10,3 10,3
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• Applying the t test for equality of means between
  two samples it is concluded that GLR and GLR
  belongs to the same population, at a 95
  confidence level.

a
c
37

• GLR shear modulus identified by the Iosipescu and
  offaxis 22 tests

Test method Test method Test method Test method
Iosipescu Iosipescu Off-axis Off-axis
d GLR (GPa) d GLR (GPa)
Mean 0,589 1,41 0,111 0,582 1,11 0,041
C.V.2 () 5,6 10,3 4,0 7,0
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• The dispersion of the GLR values are of the same
  order of magnitude.

• Applying the t test of equality of means, it is
  concluded that the GLR values from both tests
  lead to different proprieties, at 95 confidence
  level.

38

• sLR time curves

Shear stress (MPa)
Time (s)
39

• Shear stresses identified in the LR specimens

1f
Specimens sLR sLR
1 14,4 14,9
2 12,6 16,3
3 17,2 18,6
4 16,2 17,9
5 19,1 19,1
6 13,2 13,8
7 14,9 15,0
8 15,9 16,8
9 19,5 19,5
Mean (MPa) 15,9 1,91 16,9 1,61
C.V.2 () 15,2 12,1
ult
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• It is not possible to identify SLR using a
  suitable failure criterion, since the nonlinear
  shear constitutive law of wood Pinus Pinaster
  Ait. is not known.

40

• Shear stresses identified by the Iosipescu and
  offaxis 22 tests

Test method Test method Test method Test method
Iosipescu Iosipescu Off-axis Off-axis
sLR sLR sLR SLR1
Mean (MPa) 15,9 1,91 16,9 1,61 14,1 0,91 16,5 1,51
C.V.3 () 15,2 12,1 12,1 16,7
1f
ult
ult
(1) Shear strength determined using the Tsai
Hill failure criterion (2) Confidence intervals
at 95 confidence level (3) Coefficient of
variation (C.V.).

• The Iosipescu test gives a good estimation of SLR
  for wood Pinus Pinaster Ait.

41

• LT specimen

• Typical experimental data measured for the LT
  specimens

Load (N)
Front face (A) Front face (A) Back face (B) Back
face (B)
Linear deformation measured with strain gauges
(me)
42

• Apparent average sLT eLT curves

Average shear stress (MPa)
Average engineering shear strain (me)

• The response of the specimens contain same
  variability.

• The curves are nonlinear the source of the
  nonlinearity can be attributed to 19,21
• (1) the nonlinear behaviour of the
  material
• (2) the geometric nonlinearity
• (3) the nonlinearity due to the contact
  conditions specimen/fixture.

43

• Dispersion of the shear moduli values (GLT , GLT
  and GLT )

a, B
a
a, A
Shear moduli (GPa)
Specimens
GLT GLT GLT
Mean (GPa) 1,33 0,121 1,34 0,101 1,34 0,081
C.V.2 () 12,2 10,3 8,5
a, B
a
a, A
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).
44
a

• Moisture content (u), density (d) and shear
  moduli (GLT, GLT)

c
a
c
Provetes u () d GLT (GPa) GLT (GPa)
1 11,7 0,603 1,38 1,26
2 11,7 0,595 1,41 1,29
3 11,7 0,590 1,43 1,31
4 11,5 0,599 1,55 1,42
5 11,4 0,592 1,29 1,17
6 10,8 0,581 1,19 1,09
7 10,6 0,606 1,38 1,26
8 11,3 0,556 1,27 1,16
9 10,8 0,574 1,21 1,11
10 10,5 0,593 1,25 1,15
Média 11,2 0,589 1,34 0,081 1,22 0,071
C.V.2 () 4,5 2,6 8,5 8,5
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• Applying the t test for equality of means between
  two samples it is concluded that GLT e GLT
  belongs to the same population, at a 99
  confidence level.

a
c
45

• GLT shear modulus identified in the Iosipescu and
  offaxis 22 tests

Test method Test method Test method Test method
Iosipescu Iosipescu Off-axis Off-axis
d GLT (GPa) d GLT (GPa)
Mean 0,589 1,22 0,071 0,538 1,04 0,051
C.V.2 () 2,6 8,5 4,0 8,1
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• The dispersion of the GLT values are of the same
  order of magnitude.

• Applying the t test of equality of means, it is
  concluded that the GLT values from both tests
  lead to different proprieties, at 95 confidence
  level.

46

• sLT time curves

Shear stress (MPa)
Time (s)
47

• Shear stresses identified in the LT specimens

1f
Specimens sLT sLT
1 15,4 16,6
2 14,7 19,0
3 15,5 17,3
4 14,5 18,6
5 16,1 17,3
6 16,5 18,1
7 19,1 20,6
8 16,0 17,5
9 14,7 18,1
10 16,1 18,1
Média (MPa) 15,9 1 18,1 0,82
C.V.3 () 8,4 6,1
ult
(1) These values does not follow a Normal
distribution (Shapiro-Wilk test) (2) Confidence
intervals at 95 confidence level (3)
Coefficient of variation (C.V.).

• It is not possible to identify SLT using a
  suitable failure criterion, since the nonlinear
  shear constitutive law of wood Pinus Pinaster
  Ait. is not known.

48

• Shear stresses identified in the Iosipescu and
  offaxis 22 tests

Test method Test method Test method Test method
Iosipescu Iosipescu Off-axis Off-axis
sLT sLT sLT SLT1
Mean (MPa) 15,9 18,1 0,81 14,0 0,81 16,6 1,01
C.V.3 () 8,4 6,1 9,5 10,9
1f
ult
ult
(1) Shear strength determined using the Tsai
Hill failure criterion (2) Confidence intervals
at 95 confidence level (3) Coefficient of
variation (C.V.).

• The Iosipescu test gives a good estimation of SLT
  for wood Pinus Pinaster Ait.

49

• RT specimen

• Typical experimental data measured for the RT
  specimens

Load (N)
Front face (A) Front face (A) Back face (B) Back
face (B)
Linear deformation measured with strain gauges
(me)
50

• Apparent average sRT eRT curves

Average shear stress (MPa)
Average engineering shear strain (me)

• The response of the specimens contains some
  variability.

• The curves are nonlinear.

51

• Dispersion of the shear moduli values (GRT , GRT
  and GRT )

a, B
a
a, A
Shear moduli (GPa)
Specimens
GRT GRT GRT
Mean (GPa) 0,278 0,0631 0,286 0,0301 0,282 0,0381
C.V.2 () 27,2 12,4 16,2
a, B
a
a, A
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).
52

• Moisture content (u), density (d) and shear
  moduli (GRT, GRT)

a
c
a
c
Specimens u () d GRT (GPa) GRT (GPa)
1 11,3 0,542 0,221 0,216
2 11,6 0,551 0,254 0,259
3 11,7 0,559 0,341 0,348
4 11,7 0,556 0,271 0,276
5 12,1 0,548 0,249 0,254
6 10,2 0,622 0,338 0,345
7 9,8 0,622 0,311 0,318
8 11,4 0,623 0,280 0,285
Média 11,2 0,578 0,282 0,0381 0,288 0,0391
C.V.2 () 7,2 6,5 16,2 16,2
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• Applying the t test for equality of means between
  two samples it is concluded that GRT e GRT
  belongs to the same population, at a 95
  confidence level.

a
c
53

• GRT shear modulus identified by the Iosipescu and
  Arcan 23 tests

Ensaio de corte Ensaio de corte Ensaio de corte Ensaio de corte
Iosipescu Iosipescu Arcan Arcan
d GRT (GPa) d GRT (GPa)
Média 0,578 0,288 0,0391 0,650 0,229 0,0351
C.V.2 () 6,5 16,2 5,9 24,0
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• The dispersion of the GRT values is slightly
  greater in the Arcan tests.

• Applying the t test of equality of means, it is
  concluded that the GRT values from both tests
  lead to different proprieties, at 95 confidence
  level.

54

• sRT time curves

Shear stress (MPa)
Time (s)
55

• Shear stresses identified in the RT specimens

1f
Specimens sRT sRT
1 2,38 3,29
2 2,76 3,88
3 0,97 4,16
4 2,86 3,36
5 4,65 4,65
6 1,01 4,62
7 1,16 5,63
8 3,27 5,18
Mean (MPa) 2,38 1,08 2 4,35 0,702
C.V.2 () 54,3 19,2
ult
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).
56

• Shear stresses identified by the Iosipescu and
  Arcan 23 tests

Test method Test method
Iosipescu Arcan
sRT sRT
Mean (MPa) 4,35 0,701 4,54 0,311
C.V.2 () 19,2 12,1
ult
ult
(1) Confidence intervals at 95 confidence
level (2) Coefficient of variation (C.V.).

• It was found a good agreement between sRT values
  identified in both tests. However, as the failure
  of the Iosipescu RT specimens does not correspond
  to shear, it is not possible to say that the
  Iosipescu test gives a good estimation for sRT to
  wood Pinus Pinaster Ait.

ult
57

• Comparison between the LR and LT specimens

• Applying the t test of equality of means between
  two samples it is concluded, for a 95 confidence
  level, that
• (1) GLR and GLT are different properties, with
  GLR gt GLT .
• (2) sLR and sLT are equal properties,
  suggesting that for wood Pinus Pinaster Ait. SLR
  SLT .

ult
ult
58
General conclusions and future work

• The GLR, GLT and GRT shear moduli identified by
  the Iosipescu test are greater than the ones
  obtained by the off-axis and Arcan tests by 26,
  17 e 20, respectively, leading to different
  properties at a 95 confidence level.

• Although it is not possible to directly identify
  the shear strengths SLR, SLT and SRT using the
  Iosipescu test, it was proved that this test
  gives a good estimation of these properties, at
  least for the LR and LT planes.

59

• Perpectives

• The use of identification technics, based on
  optical measurements and heterogenous fields, in
  order to identify several mechanical properties
  from only one test method.

• The use of a micro/macro approach that allows the
  estimation of the macroscopic behaviour of wood
  through the characterization of its
  micro-structure.