Professional Documents
Culture Documents
Doctoral Thesis in Solid Mechanics (Impact of Paperboard Deformation and Damage Mechanisms On Packaging Perfomance)
Doctoral Thesis in Solid Mechanics (Impact of Paperboard Deformation and Damage Mechanisms On Packaging Perfomance)
OF TECHNOLOGY
Academic Dissertation which, with due permission of the KTH Royal Institute of Technology,
is submitted for public defence for the Degree of Doctor of Philosophy on Monday the 20th
March 2023, at 13:00 in F3, Lindstedtsvägen 26, Stockholm.
ISBN 978-91-8040-488-4
TRITA-SCI-FOU 2023:05
– Italian proverb
This page intentionally left blank
Abstract
All tests were performed at several levels of relative humidity (RH). Linear
relations between the mechanical properties normalized with their respective
value at 50 % RH and moisture ratio were found.
In Paper C, the master curve developed in Paper B was used to estimate material
input parameters for simulating a Box Compression Test (BCT) at different
moisture levels by using an orthotropic material model with a stress-based failure
criterion, i.e., a relatively simple material model with few input parameters. The
result showed that it was possible to accurately predict the load-compression
curve of a BCT when accounting for moisture. Furthermore, it was concluded
that modeling the creases’ mechanical properties is vital for capturing the stiffness
response of the package. Here, a measurable approach for reducing the creases’
mechanical properties was suggested, based on a folding test to obtain the relative
creasing strength (RCS) and a short-span tensile test to obtain the relative tensile
strength (RTS). It should be emphasized that the model does not include any
fitting parameters. All input data is based on measured values. Due to the
importance of creases, the RCS and RTS ratios were investigated further in
Paper D. When evaluated against normative shear strength during creasing, the
RCS and RTS values together formed a creasing window, where the RTS values
corresponded to in-plane cracks (upper limit) and the RCS values corresponded
to delamination damage (lower limit). It was observed that both the lower and
upper limits exhibit linear relations as functions of shear strain.
• Dragprov (i planet),
• Dragprov (ut ur planet),
• Korta kompressionsprov (SCT),
• Böjstyvhetsprov,
• Skjuvprofilsprov.
Samtliga prov utfördes vid flera olika nivåer av relativ fuktighet (RH). Linjära
relationer mellan mekanisk egenskap normerad med motsvarande värde vid
50 % RH och fukt noterades.
Sammanfattningsvis var det primära syftet med den här avhandlingen att utveckla
en lättanvänd modell med få materialparametrar som kan prediktera
last-deformationskurvor för olika lastfall. Syftet var inte att modellen skulle
användas för att simulera fysiska experiment så noggrant som möjligt, utan istället
att öka kunskapen om skademekanismer under provningsförfaranden. En tydlig
fördel med det här tillvägagångssättet är att modellen kan användas för att
antingen variera förpackningsgeometrin, eller göra en parameterstudie av de
ingående materialparametrarna. Den kan också användas för varje skikt separat,
vilket skulle hjälpa både konverterare och producenter. Utöver detta har det visats
att modellen kan ta hänsyn till fukt, om masterkurvan från Artikel B används.
Slutligen bör det förtydligas att modellen inte innehåller några
parameteranpassningar och att all ingångsdata bygger på mätbara värden.
Preface
The work presented in this thesis was carried out at the Research Institutes of
Sweden (RISE) and the Department of Engineering Mechanics, KTH Royal
Institute of Technology, Stockholm, Sweden, between June 2018 and January
2023. The work would not have been possible without the generous financial
support of STFI’s Association of Interested Parties (STFIs Intressentförening).
Finally, I would like to thank my family and friends for all their support. Especially
my wife Matilda, for supporting me through these years of strange working hours.
A final thanks to my sons Harald and Folke, who ensure I do not oversleep in the
mornings.
Gustav Marin
List of appended papers
Paper A: Stiffness and strength properties of five paperboards and their moisture
dependency
Paper C: Experimental and finite element simulated box compression tests on paperboard
packages at different moisture levels
Paper A: Principal author and performed all experimental work and analysis.
Marin, Nygårds and Östlund jointly evaluated and interpreted the results.
Paper B: Principal author and performed all experimental work and analysis.
Marin, Nygårds and Östlund jointly evaluated and interpreted the results.
Paper C: Principal author and performed all experimental work, simulations and
analysis. Srinivasa wrote the initial code that Marin developed for this study.
Marin, Nygårds, Srinivasa and Östlund jointly evaluated and interpreted the
results.
Paper D: Principal author and performed the experimental work (together with
Nygårds) and analysis. Marin, Nygårds and Östlund jointly evaluated and
interpreted the results.
In addition to the appended papers, the project has resulted in the following
presentations:
1 Introduction ............................................................................................................ 1
2 State-of-the-art........................................................................................................ 3
2.1 Packaging performance ................................................................................ 3
2.1.1 How packages break ................................................................................ 3
2.2 Corrugated board .......................................................................................... 7
2.3 Paperboard ..................................................................................................... 8
2.3.1 Converting a paperboard to a package ...............................................10
2.3.2 Structural properties of paperboard ....................................................12
2.3.3 Mechanical properties of paperboard .................................................13
2.4 Influence of moisture .................................................................................18
2.5 Finite element simulations .........................................................................24
3 Materials and Methods ........................................................................................27
3.1 Material characterization ............................................................................27
3.1.1 In-plane tensile test ................................................................................27
3.1.2 In-plane compression test .....................................................................28
3.1.3 Out-of-plane tension .............................................................................28
3.1.4 Out-of-plane shear test ..........................................................................28
3.1.5 Creasing and folding ..............................................................................29
3.2 Packaging testing .........................................................................................29
3.2.1 Box compression test ............................................................................30
3.2.2 Torsion test .............................................................................................30
3.3 Finite element model ..................................................................................31
4 Results and discussion .........................................................................................35
5 Conclusions ...........................................................................................................43
6 Future work...........................................................................................................45
7 Bibliography ..........................................................................................................47
1 Introduction
The world is facing innumerable environmentally-related challenges, and plastic
waste is one of them. More than 25 billion metric tons of plastic waste are
generated yearly in Europe only. Approximately 1.5 – 4 % of global plastics
production ends up in the ocean annually. These numbers are relevant for the
packaging industry since 59 % of plastic waste is due to packaging [1]. Partly
because of this, the European Union has been acting by banning single-use
plastics [2]. The required reduction of plastic use in future environmentally
friendly packages enables the huge potential for the paper industry since paper-
based material is relatively cheap but also a recyclable material. Regarding the
recyclability of paper, a recent study indicates that fibers may be recycled 25 times
without any reduction in the tensile index [3]. This shift in the usage of packaging
materials can be observed in a regular grocery store, where packages previously
made of plastics are now made of paper-based materials; an example is illustrated
in Figure 1.
Figure 1. Two applications (straw and cover) that were made of plastic previously.
The function and shape of paper-based packaging material may vary. For
example, it can be used as wrapping or converted into a box or container. The
latter function is typically based on a corrugated board or paperboard. A
corrugated board combines thin flat papers (liners) with a wavy-shaped paper
(fluting) in between. A paperboard is typically thick paper with high basis weight,
or grammage. In this thesis, the focus will primarily be on paperboard.
1
production, different material properties in different directions, often referred to
as machine direction (MD), cross direction (CD), and thickness direction (ZD).
With these drawbacks as a starting point, this thesis aims to evaluate how a
package’s performance is affected by its material properties, geometry, or
construction on the material level. Physical testing and finite element (FE)
simulations have been performed to succeed with the project’s goal. The work
performed for this thesis complements the literature by focusing on the packaging
level of paperboard packages. The approach uses a basic material model based on
measurable input parameters and FEM for varying design, geometry, and load
case. From this, it is possible to gain knowledge about the material parameters
affecting the behavior of the package, and deformation mechanisms in general,
instead of focusing on simulating a specific load case as accurately as possible.
Specific issues on the packaging performance that the thesis addresses are:
2
2 State-of-the-art
• Primary package
• Secondary package
• Tertiary package
where the primary package refers to the package in contact with the product, for
instance, a milk package. Several primary packages are grouped into a secondary
package to facilitate logistics. Secondary packages are after that grouped into a
larger tertiary package. The tertiary package is dimensioned to be placed on a
pallet and is typically used for transportation and warehouse storage. These
different categories naturally have different purposes, which is why the design of
the different packages differs. However, there are different requirements for
different types of packages within the same category as well. Therefore, even if a
milk package and a shoe box categorize as primary packages, they have different
requirements and are designed differently.
The purpose of the package is, therefore, essential for the developing process,
and several aspects need to be considered. From a mechanical point of view,
external loads and grip stiffness, for example, are essential to consider in the
design process. Furthermore, visual aspects and printability are vital as well. For
packages that contain more luxurious products, 3D effects, such as deep drawing,
in the package are a common method to visualize the exclusivity of the product.
Different and appropriate properties must be evaluated with different methods
to analyze whether a package fulfills its purpose.
3
desirable for several reasons. For instance, cracks in a milk package may lead to
leakage, and the function criterion is consequently not fulfilled. However, even if
the function is not affected, the customer easily deprioritizes a damaged primary
package such as, e.g., the one in Figure 2.
Packages are exposed to different situations that may damage them. They should,
for instance, endure a long-time stacking and transportation and ideally also
survive dropping to the ground. It is, therefore, essential to ensure that one
designs a package and mechanically optimizes it for its purpose, i.e., it is necessary
to analyze the packaging performance. To succeed with this, well‑developed
methods, such as physical testing, analytical methods, or finite element (FE)
simulations, are required.
One of the most analyzed load cases for paper-based packages is the stacking
performance, which can be evaluated by performing a Box Compression Test
(BCT), where a testing apparatus compresses a package between two plates and
records a load-compression curve. The test is established and well described in
the literature, where Frank [7] summarizes the background of the test method and
presents challenges with a focus on corrugated boxes. The box compression
strength, defined by the BCT value, is obtained from the test procedure, i.e., the
maximum force a box can carry before failure. Here, it is essential to emphasize
that this critical force is evaluated at the first local maximum from the load-
compression curve, not necessarily the ultimate peak of the curve. If the global
maximum occurs after the first peak, the testing apparatus might compress the
package so that any potential content would be damaged. Figure 3 shows a typical
load-compression curve for a BCT. As shown, an initial slack behavior is present
due to the presence of horizontal creases.
4
Figure 3. Typical load-compression curve from a box compression test (BCT) [8].
Analytical models for predicting the BCT value, such as McKee or Grangård,
typically depend on physical testing. For example, if the package design is
changed, the parameter 𝑘 in Eq. (1) might change, and additional testing must be
performed before these models are applicable.
The drawbacks of physical experiments and analytical methods requires the usage
of new development tools when the packaging industry faces a demanding future.
5
One possible way forward is to use finite element analysis (FEA) in the
development process. One can use FEA in the package design process to account
for other variables besides strength. For BCT, several studies investigate this on
packages of corrugated board. Nordstrand et al. [12] developed a FE method for
predicting strength for corrugated board containers, including a failure criterion
based on the Tsai-Wu tensor failure theory [13], and compared the predicted
failure load to experimental data. In the same year, Biancolini and Brutti [14]
accurately predicted initial buckling during box compression and estimated the
critical load by an eigenvalue buckling analysis. A clear advantage of
FE simulations is the possibility to evaluate different designs, which is utilized by,
for instance, Han and Park [15], who simulate corrugated boxes with ventilation
and hand holes. Fadiji et al. [16] have similar motivations and report the
differences in results when changing the vent’s number, orientation, and shape.
Jiménez-Caballero et al. [17] also illustrate the advantages of FEM by simulating
BCT for different geometries. Garbowski et al. have recently developed analytical-
numerical approaches for predicting BCT of corrugated boxes with various
openings [18], similar to the vents, and for corrugated boxes with various
perforations [19].
For paperboard packages, Beldie and Sandberg [20] were early in the field,
analyzing different parts of a package during a BCT, and concluded that the
horizontal creases affect the stiffness of a paperboard package. Later, similar
studies on BCT of paperboard packages were performed with a more accurate
shape of the load-displacement curve. Luong et al. [21] accurately predicted a
paperboard package’s strength and stiffness response during a BCT by explicitly
implementing an orthotropic elastoplastic model for in-plane stresses in Abaqus
developed by Mäkelä and Östlund [22]. However, this method requires several
parameters determined by fitting them to experimental data. Luong [23] also notes
that the creases are crucial for the stiffness response of the package, which is why
he reduces the stiffness of the creases. Zaheer et al. [24] investigate critical
buckling loads and consider different geometries and influences of creases.
Using FEA allows the studying of stress and strain fields that arise during loading,
contributing to a better understanding of the crucial mechanisms activated in the
loaded package. In addition, it is possible to isolate and modify different design
parameters, such as the package’s perimeter and the packaging material’s
mechanical properties and see how they affect the behavior of the package. In the
end, this may lead to better knowledge and the development of improved
sustainable packages and paperboard materials. FEA also enables the possibility
of investigating varying load cases. Analyses, where FE simulations have been
performed and compared with experimental observations, have also been done
on a macroscopic level for 3D forming [25], brim forming of cups [26], loading
of cigarette packages [27], and drop testing [28]. Common for all these analyses
was that the authors used relatively simple material models to make accurate
predictions. When less advanced material models are used in the analysis, one can
reduce time and focus on activated mechanisms during testing.
6
converted to a package. It is vital to design the material appropriately to avoid
these cracks; therefore, detailed knowledge of the material’s construction is
required.
One main advantage that makes paper-based materials suitable for packaging is
the stiffness-weight ratio; paper generally has high stiffness relative to its low
weight. In paper physics, the literature commonly refers to bending stiffness. It is the
board’s resistance against the bending deformation, expressed as a function of the
elastic modulus, 𝐸, and the area moment of inertia, 𝐼. By modifying the board’s
construction, e.g., the ZD density profile, its area moment of inertia can be
altered. This allows for optimization for increased bending stiffness.
A corrugated board visualizes this theory well since it follows the design of a
sandwich construction and the concept of an I-beam: outer facings (liner) with
high stiffnesses and strengths, separated by a wave-shaped core (fluting),
according to Figure 4. The ingoing materials of a corrugated board, i.e., the liner
and the fluting, are called container boards and typically glued together with a
starch-based adhesive.
There are different profiles of corrugated boards. A profile specifies the geometry
of the flute (i.e., the wavelength and height). A letter categorizes the profile, where
profile A is the highest flute (4.0-4.8 mm) with the longest wavelength (8.3-10
mm). Since a larger flute increases the distance between the liners, the
manufacturers use this when high bending stiffness is required. On the other
hand, small flutes are commonly used when a high-quality print is preferred.
7
2.3 Paperboard
8
Figure 5 consists of outer plies with higher density, stiffness, and strength than
the middle ply, which generally consists of fibers with low density. Changing each
ply’s furnish enables modification of the mechanical properties. For example,
fibers from different wood sources have different tensile indexes (corresponding
to specific strength, i.e., tensile strength divided by density), which Figure 6 (a)
illustrates.
(a) (b)
Figure 6. Tensile indexes for different fibers, redrawn from (a) [30] and (b) [31, p. 34].
The reader should also note that fibers from different sources have different
lengths. The dimensions of typical wood fibers are 20-30 µm in width and 1 mm
in length for hardwood and up to 3 mm for softwood (commonly pine and
spruce), respectively. The fiber length may affect the mechanical properties as well
as the printability of a paper [32, 33]. Nygårds [34] recently introduced an
approach based on material characterization of split plies, where he expresses the
in-plane stiffness and strength as a linear equation of density and fiber length.
Hence, it is possible to adjust the mechanical properties of each ply by modifying
the orientation, the density, and the fiber length. Depending on whether
mechanical or chemically based methods are used for separating the fibers in the
pulp manufacturing process, they will also achieve different properties. The main
difference from a mechanical point of view is that mechanical pulp generally has
lower strength and stiffness since a significant fraction of the fibers is
damaged during the mechanical process [35, 29, p. 21]. This difference is
illustrated in Figure 6 (b), where groundwood (GWD) and mechanical pulp (MP)
have the lowest tensile indexes. Termo-mechanical pulp (TMP) and chemi-
thermomechanical pulp are also mechanical pulps treated with heat and
chemicals. In a more recent review paper, Petterson et al. [36] publish similar
results as in Figure 6 (b), but the tensile index for CTMP reaches 50 kN/kg.
However, when manufacturing the CTMP after preheating to a temperature
above the softening temperature of lignin (HTCTMP), the tensile index reaches
the same levels as in Figure 6 (b). SBK is an acronym for semi-bleached kraft
pulp, and UBS is unbleached kraft pulp. For example, a typical standard folding
9
box board (FBB) comprises two outer plies with chemical pulp (kraft pulp) and a
middle ply consisting of mechanical pulp, such as CTMP.
During the creasing procedure, a ruler is punched into a die with the paperboard
in between. By doing this, shear strains develop in the material and cause
delamination, as illustrated in Figure 7. Hence, the delamination procedure is
associated with the breaking of fiber-fiber joints. Furthermore, the structure of
the paperboard is crucial for increasing the extent of delamination in the creased
zone, which Carlsson et al. [39] conclude. A study by Nygårds et al. [41] points out
that an ideal structure for enabling delamination and facilitating the plies to
deform plastically in shear internally is a paperboard with well-defined plies and
weak interfaces.
(a) (b)
Figure 7. Before (a) and during (b) a creasing procedure.
10
To illustrate the importance of the creasing procedure, Figure 8 shows
cross-section pictures of folded specimens at different creasing depths, d.
In Figure 8 (a), where the specimen was uncreased, one can observe how the
inner surface, or the bottom ply, has been compressed. Cavlin [42] noted this in
1988 and concluded that compression failure occurs at the peak value of the
curve.
In the literature, several authors present different approaches to define the shear
strains caused by the creasing procedure, for instance, Halladay and Ulm [40],
Donaldson [44], and Nagasawa et al. [45, 46]. All approaches interpret geometrical
relations between parameters such as the creasing depth, the channel width, and
the ruler’s width. The definition of a shear measure is further developed by Coffin
and Panek [47] and Panek et al. [48], where the paperboard draw is introduced to
define a creasing window. The draw defines how much paperboard draws into the
die during the creasing operation. With this approach, creasing windows ensure
that the relative crease strength (RCS), also referred to as the bending resistance
ratio, is sufficiently low, i.e., avoiding cracking and securing good crease quality.
The relative crease strength (RCS) is the ratio between the maximum bending
creased
force achieved during the folding of a creased board, Ff , and an uncreased
uncreased
board, Ff , hence
𝐹𝑓creased
𝑅𝐶𝑆 = . (3)
𝐹𝑓uncreased
The folding operation is evaluated by different methods, where one common way
of evaluating the procedure is two-point bending, which provides a force-
deflection curve. From this curve, the bending stiffness can be evaluated
according to the Euler-Bernoulli beam theory, and the peak value characterizes
the maximum bending force discussed in Eq. (3). The bending stiffness at two-
point bending is given by [49]:
19.1𝐹𝐿2 (4)
𝑆=
Φ𝑏
where Φ is the angular deflection, 𝑏, is the width and 𝐿 is the length of the
specimen. The force, 𝐹, refers to the value measured at 5°, where paperboards
generally remain in the folding curve’s elastic region. Figure 9 shows typical force-
deflection curves in MD from two-point bending and visualizes the effect of
creasing on the maximum bending force and the bending stiffness as a
complement to Figure 8.
11
Figure 9. Two-point bending curves of uncreased and creased specimens [50].
12
Figure 10. Typical surface profile for a paperboard.
As seen in Eq. (5), knowing the in- and out-of-plane stiffnesses is essential for
characterizing the material. Moreover, it is vital to know the strength of the
material, in tension, compression, and shear, for the in-plane and out-of-plane
directions, respectively. Noteworthy is that these properties are not always
published in data sheets for commercial use. For paperboard, other properties
such as bending stiffness, are commonly presented. Some applications or material
models require additional parameters as well. However, the stiffness properties
from Eq. (5), in combination with the strength for the different material
directions, are a good starting point for material characterization.
13
2.3.3.1 In-plane properties of paperboard
The in-plane properties of paperboard, especially the strength and stiffness in
tension, have been thoroughly investigated in the literature since they are vital for
analyzing the material’s performance, whereas Niskanen [57] reviewed the field
until 1993. Niskanen concludes that two factors govern the strength of paper: the
strength of fiber-to-fiber bonds and the formation of paper, i.e., a uniform paper
tolerates higher stresses than non-uniform paper. Fiber breaks are not considered
the critical factor in tension unless for specific scenarios when the bonds are very
strong. The explanation behind bond failure is not discussed further in this thesis.
However, it is well analyzed in the literature, where Cox [58] in 1952 discusses a
shear-lag mechanism where the stress is transmitted from one fiber to another
through the bonds. However, on a board level, the pure in-plane tensile properties
are obtained from uniaxial tensile tests, where different standard procedures are
established (such as ISO [59] and ASTM [60]), and different types of equipment
following the standards have been developed.
Even if this thesis primarily discusses the stiffness and strength properties of
paperboard, it is essential to mention the size dependency of in-plane tensile tests.
The characterization of tensile fracture and critical length is discussed by Cavlin
[61] in 1974, where the brittleness of a material is related to the critical length of
a specimen, i.e., the shortest length for which the load-elongation curve rapidly
drops from peak load to zero [62]. This is explained by the ratio of incrementally
released elastic energy and fracture energy within the specimen. In large
specimens, the rate of released elastic energy is sufficiently large to cause an
unstable fracture of the specimen, i.e., a brittle and rapid fracture. Figure 11 shows
the length dependency for a specimen, where specimens of the same paperboard
with the same widths but different lengths were exposed to tensile load. For
shorter lengths, a stable reduction of the load-elongation curve is observed since
the released rate of elastic energy is lower than the rate of required fracture energy
needed for progressing the fracture. Tryding [63] analyzes this concept further,
and Tryding et al. [64] characterize the crack growth in paperboard by introducing
stress-widening curves for analyzing the post-peak cohesive behavior. Hagman
and Nygårds [65] further investigate the sample size effects with an alternative
approach, using tensile test and speckle analysis. They conclude that there is a
difference in strain behavior depending on the length-to-width ratio and that this
was dependent on the board composition.
14
Figure 11. Tensile test with different specimen lengths (width = 15 mm) [66].
The magnitude of the properties in paperboard varies with density [34]. However,
to introduce the reader to representative numbers of in-plane tensile properties,
an early work by Persson [67, p. 26] is emphasized. The study was performed on
a three-plied commercial duplex board, where Persson reports the in-plane
stiffnesses to be 5.4 GPa (MD) and 1.9 GPa (CD). The basis weight of the board
was 240 g/m2 and a thickness of 0.374 mm, i.e., a density of around 640 kg/m3.
The tensile strengths were 46 MPa in MD and 22 MPa in CD. These are
representative paperboard numbers confirmed in other studies [68, 69].
15
sheet strength is increasing the elastic properties of the board. They also highlight
the effect of the clamping procedure and the effect it has on the delamination.
Regarding numbers, representative values of the compressive strength for a multi-
ply board with a thickness of 0.390 mm and a density of 740 kg/m3 are 15.2 MPa
in MD and 12.5 MPa in CD, respectively [75].
According to Fellers and Donner [76], the causes of compression failure are
elastic buckling of fiber segments (low-density sheets) or shear dislocations in the
fiber walls (high-density sheets). In 1983, Habeger and Whitsitt [77] developed a
mathematical theory describing compressive failure related to in-plane and out-
of-plane shear stiffnesses. Lately, Hagman et al. [75] have investigated the cause of
compression failure on multi-plied commercial boards. Hagman et al. state that
shearing the interfaces, combined with the onset of plasticity in the loading
direction, initiates failure during SCT. They further concluded that the main
parameters determining compressive strength were the elastic modulus, local
shear strength, and out-of-plane strength. Hagman and Nygårds [78] also show
that the short-span compression test is governed by in-plane stiffness and
through-thickness delamination. Hence, the in-plane compression is also related
to the out-of-plane properties.
The final in-plane property discussed here is the in-plane shear modulus, which is
important when describing paperboard as an orthotropic material. Persson [67]
estimated it by performing in-plane tensile loading at 45° to the MD and CD,
page 16. Baum et al. [79] derived the expression
(𝐸𝑥 𝐸𝑦 )2
1
(6)
𝐺𝑥𝑦 = 1
2(1+(𝑣𝑥 𝑣𝑦 )2 )
for the in-plane shear modulus in 1981, where the in-plane Poisson’s ratio was
assumed to follow:
𝐸 (9)
𝑣𝑥𝑦 = 0.293√𝐸𝑥
𝑦
and Poisson’s ratios in the out-of-plane directions were set to zero [80]. These
assumptions are used in several studies in the literature [69, 75].
16
combined normal and shear through-thickness loadings. The tensile strength in
the thickness direction varies in the literature, depending on the paper-based
material investigated. For handsheets, Van Liew reports 0.55 and 0.95 MPa for
two different handsheets, and Girlanda and Fellers report 0.2-1.75 MPa,
depending on the furnish. Persson measured the out-of-plane strength of the
paperboard to 0.33 MPa, which is similar to the values reported for the middle
ply by Nygårds: estimated to be 1.3 MPa for the top ply, 0.25 MPa for the middle
ply and 0.95 MPa for the bottom ply, respectively.
17
separated with a loading direction oriented in MD or CD, with a prescribed
deformation. The shear testing procedure results in a load-deformation curve,
from which the strength and stiffness properties can be obtained. Fellers [93]
published data (not commercially produced board) on these rigid block tests in 1977
with an out-of-plane shear modulus of 70 MPa and a shear strength of 1.1 MPa.
A more recent study on commercial multiply boards reports that the middle ply
has an out-of-plane modulus of 30 MPa [25]. As an alternative approach,
Nygårds et al. [68, 94] developed a new way of evaluating shear strength: The
notched shear test (NTS). A tensile specimen is notched on the top and the
bottom (with an in-plane distance between the notches) and is then exposed to a
tensile load. Hence, a shear zone is created at the depth where the notches meet.
This method does not evaluate the out-of-plane shear modulus but is useful when
the shear strength profile is of interest [95, 96, 97, 98]. An alternative method for
evaluating the shear properties is the Arcan device, designed for combined
loadings in normal and shear. This method was used by Stenberg [99] to measure
the out-of-plane shear properties during high compressive loads.
Figure 12. Differences in water uptake for hemicellulose, cellulose, and lignin. Redrawn by
Salmén from data by [101].
18
Hence, complete softening may not be achieved until the fibers are immersed in
water. This is illustrated in the model by Salmén in Figure 13. In Figure 13, the
humid state corresponds to a relative humidity level above 80 %.
ellulose
microfibril
crystalline
isordered
one glass glass soft
atri
emicellulose, glass soft soft
lignin and
cellulose chain ends
Figure 13. Illustration of microfibril with surrounding matrix during different environmental
conditions, redrawn from [100].
There are, however, several studies on how different pulps interact with moisture,
but this thesis focuses on research on primarily commercial paperboard. The
paper-water interaction could either be through contact between paperboard and
liquid, i.e., hydrosorption, or by changing the relative humidity (RH): hygrosorption. In
this thesis, the effect of moisture will be analyzed through the perspective of
changes in RH, and hydrosorption will not be discussed further, but the reader
interested in hydrosorption is referred to the work of Larsson [102].
The relation between the amount of moisture in the material and the surrounding
environment is non-linear and irreversible. However, when the relative humidity
changes, the paper-based material will either gain water (adsorption) or lose water
(desorption). This history dependence is referred to as a hysteresis effect and is
not unique to paper. It is, however, well discussed in the literature of paper
materials. Even though the moisture sorption in paper is well investigated, there
is still a lack of consensus regarding the origin of the hysteresis phenomenon
[103]. A conceptual view of the effect is shown in Figure 14.
19
oisture measure
The effect on the in-plane mechanical properties from testing at different climates
is well documented, where the stiffness and strength decrease with increased
RH [103]. Benson [105] noted in 1971 that the in-plane strength and stiffness are
reduced by 50 % when the relative humidity changes from ca 20 % to 90 % RH.
This is valid for both MD and CD. Typical in-plane stress-strain curves are shown
in Figure 15, where the observations from Benson are verified. Similar curves as
those in Figure 15 were evaluated by Mäkelä [106] in 2010, although he also
investigated the impact of moisture on in-plane compression. Tensile tests at
different moisture levels have also been performed and simulated by
van der Sman et al. [107].
(a) (b)
Figure 15. In-plane stress-strain curves at different levels of RH for MD (a) and CD (b)
[108].
20
Out-of-plane properties are also affected by moisture but are not reported in the
literature to the same extent as the in-plane properties. Tryding and Ristinmaa
[109] published curves on tensile test in ZD at two different moisture levels in
2016 where the cohesive post-peak behavior was present, just like the curves in
Figure 16. More recently, Spiewak et al. [110] performed out-of-plane tensile tests
at different moisture contents when modeling the crack growth behavior of filter
paper.
21
(a) (b)
Figure 18. Folding at different levels of RH [111].
In the literature, mechanical properties such as stress and stiffness are evaluated
against RH, as in Figure 15 -Figure 18. This is relevant for some applications, but
the relative humidity primarily gives information regarding the testing
environment, not the amount of moisture within the material. Therefore,
methods for describing this have been developed in the literature. Perhaps the
most common way of evaluating the moisture in paper is the moisture content
[113], which is the ratio between the water and the total mass, i.e.:
𝑚𝑤𝑎𝑡𝑒𝑟 (10)
𝑚𝑐 =
𝑚𝑡𝑜𝑡𝑎𝑙
An alternative method is using the moisture ratio:
𝑚𝑤𝑎𝑡𝑒𝑟 (11)
𝑚𝑟 =
𝑚𝑑𝑟𝑦
This method does not have an ISO standard but has the clear advantage that it is
strictly linear to the mass of water. It has also been shown in the literature that
mechanical properties are linear to moisture ratio (within a limited range). Fellers
shows the linearity for strength and stiffness [31, p. 41 and 141]. Further,
Mäkelä [106] reports that the in-plane tensile strain at break and the compression
stiffness and strength are also linear to moisture ratio. The linear interval for in-
plane stiffness was studied by Mäkelä [114], which is shown in the initial regions
of the curves in Figure 19. It should be noted that the study was a drying study
performed on wet paper specimens. Hence, the material was not exposed to
moisture through hygrosorption, i.e., changing the relative humidity, which
explains the huge values of the moisture ratio.
22
Figure 19. Tensile stiffness index as a function of moisture ratio, redrawn from [114].
Linear relations to the moisture ratio are also valid for tensile strength in ZD,
according to measurements by Fellers [115]. In the same study, Fellers notes that
if the relative strength is plotted against the relative moisture ratio (i.e., strength and
moisture ratio normalized with the corresponding values at 50 % RH), a linear
relation is established independent of pulp. Other authors also report relative
values of the mechanical properties. Popil [116] studies relative values of SCT
when describing the relation between compression strength and moisture
content, and Back [117], studies the difference in reduction of in-plane tensile and
compression strength.
23
(a) (b)
Figure 20. Impact of temperature on in-plane tension for MD (a) and CD (b) [66].
24
criterion [126] may be used, which originally was developed in 1948 to describe
anisotropic metals. ill’s model suggests a linear, but not anisotropic, hardening.
For ductile materials, a power-law relation was introduced by Ramberg-Osgood
[127] in 1943 and was then applied to paper through an orthotropic elastic-plastic
model by Mäkelä and Östlund [22] in 2003. The model accurately predicts the in-
plane stress-strain response of paperboard. The most simplified way of describing
the elastic-plastic behavior of paperboard is a bilinear model, i.e., a model with
two straight lines connected in the yield point. This approach has been
investigated in the literature and is accurate for several cellulose-based materials
[128]. Its simplicity makes it, however, sensitive to the determination of the yield
point, which is discussed by Erkkilä et al. [129], where different empirical plasticity
models are reviewed. Commonly, the offset strain is set to be 0.1 or 0.2 % for
metals but may differ for paper materials. Several approaches to describing a
paperboard from a continuum perspective have been performed in the literature.
It gets more complicated if the model should also account for damage
mechanisms, such as delamination. Since delamination is crucial for converting
operations, damage mechanisms must be considered when modeling paperboard;
consequently, some damage criterion might be needed within the model. This can
be introduced in different ways; one method is to introduce cohesive interfaces
between the plies of the paperboard, which for instance, is done by Xia et al. [130]
in 2002, Nygårds et al. [41] in 2009. Another method is to introduce continuum
damage. This approach was used by Isaksson et al. [131] in 2004, who included a
gradient-enhanced damage model in the constitutive relations when describing
the material behavior of test liner and kraft liner. Borgqvist et al. [132] investigate
paperboard material through a continuum approach, accounting for non-linear
kinematics and large strains. To account for anisotropy, they introduce three
structural vectors corresponding to the principal directions of a paperboard. By
letting the in-plane vectors deform as line segments and the out-of-plane direction
deform as a normal vector, they enable the decoupling of the in-plane and the
out-of-plane responses in shearing. Stenberg presents an elastic-plastic material
model for the out-of-plane mechanical behavior, which enables simulations under
high compressive loads in ZD. Stenberg [133] uses the peak load to identify
critical stress due to difficulties determining the load level where the deformation
changes from elastic to elastic-plastic. It should be noted that these models are
quite complex and require a large number of material input parameters.
Additionally, the complexity and the number of required parameters increase
further if moisture and temperature are included [120, 134].
25
approach was used by Huang and Nygårds when simulating creasing [69] and
folding [137], or the combined scenario, which was evaluated by Huang et al. [138].
Beex and Peerlings [139] analyzed creasing and folding by an orthotropic elasto-
plastic continuum model and a delamination model between the plies in order to
enable an opening behavior between the plies, and Giampieri et al. [140]
investigated the mechanical response of creasing and folding. In their study, they
developed a constitutive model where a damage variable is introduced, related to
the development of delamination inside the creased region. Another scenario that
has been investigated is deep drawing of paperboard. Wallmeier et al. [141] use an
anisotropic material model that accounts for temperature. Furthermore, they
proposed a plasticity model with a rectangular yield surface as a starting point.
However, since the in-plane shear stress reduces the size of the yield surface, they
introduce four equations that reduce the rectangle when increasing the yield stress
in shear.
One of the problems with advanced FE models describing the material behavior
in detail is that the model might be too complex to manage for other users. During
Simon’s review of elasto-plasticity models, he notes in-plane models with up to 28
non-zero parameters [142, p. 2414]. Hence, pursuing good accuracy for the
simulations compared to experimental data might be counterproductive. An
alternative approach is to use as straightforward material models as possible,
based on measurable input parameters, and use FEM for varying design,
geometry, and load case. From this, it is possible to gain knowledge about the
material parameters affecting the behavior of the package, and deformation
mechanisms in general, instead of focusing on simulating a specific load case as
accurately as possible.
26
3 Materials and Methods
To investigate the relations between material properties and their impact on the
performance of packages, the procedure of this thesis is categorized into three
different main areas:
𝜎=
𝐹
= ,
𝐹 (12)
𝐴 𝑏𝑡
and
𝛿 (13)
𝜀=𝐿,
where t is the thickness of the board and δ is the elongation of the specimen. The
elastic modulus was calculated following ISO standard [59].
27
Figure 21. The clamps used for short-span in-plane tensile tests.
28
Specimen
Figure 22. Conceptual illustration of rigid block shear test.
29
(a) (b)
Figure 23. Drawing of the blank (a) and mounted package (b).
30
in Figure 25 (b). The package was then placed within the square. After the glue
set, the specimen was fixed into the torsion tester, as shown in Figure 25 (a).
The mounted package was exposed to a rotation with a loading rate of 1°/s. The
torque was measured by the load cell of the equipment. To maintain zero axial
force, avoid axial deformation, and achieve a state of loading approaching simple
shear; axial load control was used. Thus, as the package was clamped at both ends
and became susceptible to compressive buckling under torsional loading, a tensile
load was applied to negate this such that the net axial force on the box was zero.
A geometry was created using LS-PrePost [147]. The same geometry was used for
BCT in Paper C and the torsion test in Paper E and followed the measures and
material directions from Figure 23 with a sheet thickness, t. A transparent view of
the FE model is shown in Figure 26, where the modeled creases are green, the
panels yellow, and the flaps orange, analogous to Figure 23. At the top and
bottom, the three small flaps are visible under one large flap. The manufacturer’s
joint is seen at the back of the right front panel.
31
Figure 26. Transparent view of the FE model.
32
Table 1. Material parameters for Paperboard A.
Property Paperboard A
Density [kg/m3] 764
EMD [MPa] 5810
ECD [MPa] 2980
σfMD [MPa] 59
σfCD [MPa] 35
σf,cMD [MPa] 59
σf,cCD [MPa] 35
GMD/ZD [MPa] 16.50
GCD/ZD [MPa] 16.50
GMD/CD [MPa] 1440
As mentioned, the creases are essential for simulating packages. Here, the
properties of the creases were reduced by multiplying all stiffness parameters with
the relative crease strength (RCS), previously described in section 2.3.1.
Additionally, a similar measure was used to reduce the tensile properties of the
creases: RTS, i.e., the ratio between the tensile strength of the creased and
uncreased board. Hence, the properties of the creases were reduced with measured
values. No fitting parameters were used.
33
This page intentionally left blank
34
4 Results and discussion
In Paper A, a thorough material characterization was performed on five
commercial paperboards with different basis weights but from the same
paperboard series. The purpose was primarily to achieve material data for the FE
simulations, but the effect of moisture was also investigated. The material
properties were normalized with the value at standard climate (50 % RH, 23 °C),
and linear relations between normalized mechanical properties and moisture ratio
were observed. Since all the paperboards in the study were from the same
manufacturer, extending the study to several manufacturers was natural. This was
done in Paper B, where paperboards with significantly different furnish and
construction were consciously chosen to examine how their in-plane mechanical
properties were affected by moisture. The result for the normalized strength is
shown in Figure 27, where each color represents a manufacturer.
Figure 27. Normalized strength as a function of moisture ratio for paperboards from four
different paperboards.
However, when both the material properties and the moisture ratio were
normalized, all paperboards coincided with the same master curve, shown in
Figure 28, where data for both MD and CD are included.
35
Figure 28. Normalized tensile strength as a function of normalized moisture ratio for 15
paperboards in both MD and CD.
The physical interpretation from the results in Figure 28 is that if any paperboard’s
moisture is doubled, the strength would be reduced by approximately 35 %. Thus,
it was possible to express any of the investigated mechanical properties as a linear
equation:
𝑃𝑟𝑜𝑝𝑒𝑟𝑡𝑦 𝑚𝑟 (15)
= 𝑎𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦 + 𝑏𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦 .
𝑃𝑟𝑜𝑝𝑒𝑟𝑡𝑦50% RH 𝑚𝑟 50% RH
Analogously, this was also performed for the in-plane stiffness and the SCT
values, which were the mechanical properties investigated in Paper B. The
parameters a and b for the linear relations are summarized in Table 2.
Table 2. Parameters for the linear relation between normalized property and normalized
moisture ratio.
Normalized property a b R2
Strength -0.35 1.36 0.96
Stiffness -0.37 1.40 0.96
SCT -0.51 1.51 0.96
These results are significant since it means that if an orthotropic material model
based on the in-plane stiffness and strength is used in FEM, it is also possible to
predict the package response at different moisture levels. This was shown in
36
Paper C, where physical BCT experiments were compared to FE simulations at
different moisture levels. The simulations were based on measured properties at
standard climate, and Eq. (15) and Table 2 when simulating the different moisture
levels. The result is shown in Figure 29.
Figure 29. Comparison of experiments and simulations for different moisture levels.
There are several things interesting to note here. First, the BCT values are
accurately predicted using a simple material model with few input parameters. An
overprediction is expected since FEM does not account for structural and
material imperfections.
37
Additionally, when BCT was analyzed further in Paper E, the material was
considered a three-ply laminated structure to investigate the effect of bending
stiffness. This effect was shown to be very small, illustrated by Figure 31, where
FE model 1 is the homogeneous model, calibrated for tensile stiffness and FE
model 2 is a three-ply laminated structure, calibrated for tensile and bending
stiffness.
Figure 30. Comparison of BCT curves for the experiments and the two FE models plotted up
to the first peak load for Paperboard A.
From this, two conclusions were drawn. First, the bending stiffness did not
significantly impact the critical force during BCT. Instead, it was shown that the
strength of the material had a more considerable impact. Secondly, determining
the mechanical properties of the separate plies, i.e., by calculating them from
Eq. (14), was representative. This is an essential observation since material
characterization on the ply level is quite complex.
All these observations are relevant if the purpose of the simulations is to predict
BCT for different moisture levels. However, the advantage of FEM is predicting
the compression strength of packages and observing damage mechanisms
during the procedure. Paper C results show that the maximum stresses coincide
with the occurrence of yield lines in the packages, see Figure 31. The results
support the theory that the package’s failure can be explained by local in-plane
failure in the elements close to the yield line, as analytically described by
Ristinmaa et al. [11].
38
(a) (b)
Figure 31. Comparison between FE simulation (a) and physical experiment (b).
(a) (b)
(c) (d)
Figure 32. Close-up photos of the damaged package at peak load during BCT.
39
In Figure 32 (a), the parabolic yield line is seen. This yield line was investigated
further by cutting the package, as the red line indicates in Figure 33. A cross-
section photo showed that the yield line is caused by delamination in the interface
between the bottom and the middle ply. Similar delamination behavior has
previously been seen by Eriksson et al. [150].
Figure 33. Cross-section photo of the yield line at peak load during BCT.
40
Figure 35. 3D scanned images of a package at different stages of twist.
(a) (b)
Figure 36. Stresses in MD (a) and CD (b) for Paperboard A expressed in MPa.
For both load cases, failure occurs in in-plane tension, not in in-plane
compression or in-plane shear. However, it should be emphasized that the model
does not account for out-of-plane properties in the failure criterion. When the
out-of-plane shear stresses are analyzed in the torsion test, the stress at maximum
torque reaches values around 1 MPa. This is in the same range as the measured
out-of-plane shear strength in Paper A. Thus, it would be interesting to further
investigate the effect of the transverse shear properties.
However, Johansson et al. [152] have recently observed that paperboard deforms
heavily in the thickness direction before in-plane failure, contributing to bond
failure in the paperboard, i.e., out-of-plane deformations are activated during
tensile loading.
41
This page intentionally left blank
42
5 Conclusions
This thesis investigated the impact of material parameters on packaging
performance. The analysis has been divided into material characterization,
packaging testing, and finite element simulations. From the initial investigation of
the material characterization, it was concluded that different mechanical
properties normalized with the properties at standard climate decrease linearly
when the moisture ratio of the paperboard is increased. The study was extended
to include paperboards from different manufacturers. It was concluded that when
the moisture ratio is normalized with the moisture ratio at standard climate, the
results from all investigated paperboards to a good approximation, followed the
same linear relation. This made it possible to estimate a mechanical property for
any board at different moisture levels by knowing only the following:
• the mechanical property for standard climate (50 % RH and 23 °C), and
• the parameters a and b that describe the linear relation.
This linear relation was used to estimate material input parameters for simulating
a Box Compression Test (BCT) at different moisture levels using an orthotropic
material model with a failure criterion, i.e., a quite simple material model with few
input parameters. The result showed that it was possible to accurately predict the
load-compression curve of a BCT when accounting for moisture. Furthermore,
it was concluded that modeling the creases’ mechanical properties is vital for
capturing the stiffness response of the package. Here, a measurable approach for
reducing the creases’ mechanical properties was suggested, based on a folding test
to obtain the relative creasing strength (RCS) and a short-span tensile test to
obtain the (RTS).
The effect of bending stiffness was investigated for both load cases by developing
two FE models. Model 1 treated the paperboard as a homogeneous material, and
Model 2 considered the paperboard to be a three-ply laminate structure. No
significant effect was noted, and it was concluded that the strength has a more
significant effect on the BCT than the bending stiffness.
43
This page intentionally left blank
44
6 Future work
The main advantage of the approach used in this thesis is the model’s simplicity.
However, for future studies, it would be of interest to explicitly include the out-
of-plane shear properties and investigate their impact on the package’s
performance. This will require modification in the material description since the
current material model does not account for out-of-plane shear or strength.
epending on the user’s needs, additional modifications might be interesting to
investigate, such as imperfections of the geometry and the introduction of
plasticity. From a paperboard manufacturer’s perspective, the laminate model,
developed in Paper E, is highly relevant since it will be possible to change the
properties of the separate plies individually. It is simultaneously easy to change
the package design, making this a valuable tool for evaluating the impact of
material properties and packaging performance. From a converting point of view,
it would be interesting to run complementing package simulations with different
reduction ratios of the creases based on the creasing window developed in
Paper D. A final suggestion is to investigate the impact of temperature. This has
not been included in this thesis, but it would be interesting to evaluate packaging
performance at different temperatures for some applications.
45
This page intentionally left blank
46
7 Bibliography
[1] European commission, "A European Strategy for Plastics in a Circular
Economy," 2018. [Online]. Available: https://eur-
lex.europa.eu/resource.html?uri=cellar:2df5d1d2-fac7-11e7-b8f5-
01aa75ed71a1.0001.02/DOC_1&format=PDF. [Accessed 20 September
2020].
[5] McKinsey & Company, "Pulp, paper, and packaging in the next decade:
Transformational change," 7 August 2019. [Online]. Available:
https://www.mckinsey.com/industries/paper-forest-products-and-
packaging/our-insights/pulp-paper-and-packaging-in-the-next-decade-
transformational-change. [Accessed 11 January 2023].
[6] Metsä Group, "The global packaging market," 2022. [Online]. Available:
https://www.metsagroup.com/metsaboard/investors/operating-
environment/global-packaging-market/. [Accessed 11 January 2023].
[13] S. W. Tsai and E. M. Wu, "A general theory of strength for anisotropic
materials," Journal of Composite Materials, vol. 5, no. 1, pp. 58-80, 1971.
47
[14] M. E. Biancolini and C. Brutti, "Numerical and Experimental
Investigation of the Strength of Corrugated Board Packages," Packaging
Technology and Science, vol. 16, pp. 47-60, 2003.
[26] M. Upadhyaya and M. Nygårds, "A finite element model to simulate brim
forming of paperboard," 28th IAPRI Symposium on packaging 9–12 May,
2017, Lausanne, Switzerland, pp. 395-408, 2017.
48
[27] J. E. Gustafsson and M. Nygårds, "Loading and deformation of cigarette
packages—experimental and numerical comparison of paperboards," 28th
IAPRI Symposium on packaging, 9–12 May, 2017, Lausanne, Switzerland, 2017.
[31] C. Fellers, Pulp and Paper Chemistry and Technology: Volume 4 Paper
Products Physics and Technology, M. Ek, G. Gellerstedt and G.
Henriksson, Eds., Göttingen, Germany: Walter de Gruyter GmbH & CO,
2009.
[37] D. Coffin and M. Nygårds, "Creasing and folding," Advances in Paper Science
and Technology: Trans. 16th Fundamental Research Symposium Oxford 2017, pp.
69-136, 2018.
49
[41] M. Nygårds, M. Just and J. Tryding, "Experimental and numerical studies
of creasing of paperboard," International Journal of Solids and Structures, vol.
46, pp. 2493-2505, 2009.
[49] Paper and board - Determination of bending stiffness by stiffness by static methods -
General principles, ISO5628:1990(E).
[53] O. Schultz-Eklund, C. Fellers and P.-Å. Johansson, "Method for the local
determination of the thickness and density of paper," Nordic Pulp & Paper
Research Journal, vol. 7, no. 3, pp. 133-139, 1992.
50
[57] K. Niskanen, ”Strength and fracture of paper,” Trans. of the Xth Fund. Res.
Symp. Oxford, pp. 641-725, 1993.
[60] Standard Test Method for Tensile Properties of Paper and Paperboard Using
Constant-Rate-of-Elongation Apparatus, ASTM(1997): ASTM D 828.
[69] H. Huang and M. Nygårds, "A simplified material model for finite element
analysis of paperboard creasing," Nordic Pulp & Paper Research Journal, 2010.
[70] C. Fellers and S. Cavlin, "A new method for measuring the edgewise
compression properties of paper," Svensk Papperstidning, vol. 78, no. 9, pp.
329-332, 1975.
[72] Paper and board - Compressive strength - Short span test, ISO 9896:1989.
51
[74] A. Brandberg and A. Kulachenko, "Compression failure in dense non-
woven fiber networks," Cellulose, vol. 27, pp. 6065-6082, 2020.
[81] G. P. Van Liew, ”The deformation of paper in the Z-direction,” PhD thesis,
The institute of Paper Science and Technology, Atlanta, Georgia, 1973.
52
[89] H.-J. Schaffrath and L. Göttschling, "The behaviour of paper under
compression in z-direction," Proceedings of the 1991 International Paper Physics
Conference, Hawaii, 1991.
[101] J. Takamura, Japan Wood Res. Soc., vol. 14, pp. 75-79, 1968.
[104] Paper, board and pulps — Standard atmosphere for conditioning and testing and
procedure for monitoring the atmosphere and conditioning of samples, ISO 187:2022.
53
[105] R. E. Benson, ”Effect of relative humidity and temperature on tensile
stress-strain properties of kraft linerboards,” Tappi Journal, vol. 54, nr 4,
pp. 170-177, 1971.
[107] C. G. van der Sman, E. Bosco and R. H. Peerlings, "A model for moisture-
induced dimensional instability in printing paper," Nordic Pulp & Paper
Research Journal, vol. 31, no. 4, pp. 676-683, 2016.
[111] G. Marin, M. Nygårds and S. Östlund, "Raw data from: Stiffness and
strength properties of five paperboards and their moisture dependency,"
Tappi Journal, vol. 19, no. 2, pp. 71-85, 2020.
[113] Paper and board – Determination of moisture content of a lot – Oven-drying method,
ISO 287:2017.
[114] P. äkelä, ”Analysis of the paper web during its passage through the
drying section,” International Paper Physics Conference, pp. 151-155, 2003.
54
[121] S. Borodulina, A. Kulachenko, S. Galland and M. Nygårds, "Stress-strain
curve of paper revisited," Nordic Pulp and Paper Research Journal, vol. 27, no.
2, pp. 318-328, 2012.
[126] R. ill, ”A theory of the yielding and plastic flow of anisotropic metals,”
The hydrodynamics of non-Newtonian fluids, pp. 281-297, 1948.
[130] Q. S. Xia, M. C. Boyce and D. M. Parks, "A constitutive model for the
anisotropic elastic-plastic deformation of paper and paperboard,"
International Journal of Solids and Structures, vol. 39, pp. 4053-4071, 2002.
55
dependent and temperature-dependent constitutive model," Packaging
technology and science, vol. 29, no. 3, pp. 145-160, 2016.
[140] A. Giampieri, U. Perego and R. Borsari, "A constitutive model for the
mechanical response of the folding of creased paperboard," International
Journal of Solids and Structures, vol. 48, pp. 2275-2287, 2011.
[143] Lorentzen & Wettre, Lorentzen & Wettre, ABB, Stockholm, Sweden.
[144] MTS System Corp, MTS System Corp.; Eden Praire, MN, US.
[146] Zwick GmbH & Co. KG, August-Nagel-Straße 11, 89079 Ulm, Germany.
56
[150] . Eriksson, . Korin och F. Thuvander, ” amage to arton Board
Packages Subjected to oncentrated Loads,” Proceedings of the 19th IAPRI
World Conference on Packaging, pp. 172-182, 2014.
57