Differentiating between packaging material and geometry using the Syntouch Biotac

(1)

IN

DEGREE PROJECT MECHANICAL ENGINEERING, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2019,

Differentiating between packaging material and geometry using the Syntouch Biotac

HENRY ERIKSSON

KTH ROYAL INSTITUTE OF TECHNOLOGY SCHOOL OF ENGINEERING SCIENCES

(2)

(3)

Degree project in Solid Mechanics, second cycle

Differentiating between packaging material and geometry using the Syntouch Biotac

by

Henry Eriksson, henryer@kth.se June 2019

(4)

(5)

Abstract

Grip stiffness is an important property of carton board packaging. The structure of the packaging needs to withstand the handling of one or many consumers. A carton board packaging that feels stiff when handled conveys a sense of luxury to the consumer, one example of this is the packages containing new flagship smart phones. If a package is more or less stiff is at this moment subjectively interpreted by test panels. An objective method of measuring grip stiffness is sought after for repeat ability and speed. This master thesis investigates the influence of geometry and material parameters on the grip stiffness of a carton board packages. The purpose is to determine how the measured point loads differ between different geometries for one particular material.

Measurements are conducted using a tensile tester and a sensory device, Syntouch Biotac. Only two of the 19 sensors are analyzed in this thesis. Ther data analyzed in this thesis comes from experiments conducted by the author and experiments conducted during a bachelors thesis at ¨Orebro University. The same equipment was used for both experiments. The authors experiments were aimed at finding how a rotation of the packaging would affect the results whereas no rotations were made during experiments not conducted by the author. The authors own experiments were also made using an actual human finger, but only on packages that were not rotated. A finite element study was performed to validate the results from the Syntouch Biotac. X-ray computed tomography was used to investigate if the damage done to the carton board material by a human finger was similar to the damage done to the carton board material by the Syntouch Biotac.

’

Results from the Syntouch Biotac show that it is possible to tell where along and how close to the edge of the packaging the Syntouch Biotac is touching the packaging and that it is possible to discern between materials if the surface weights are different enough.

The X-ray computed tomography show that damages done to the carton material cone by either the Syntouch Biotac or a human finger, are not possible to tell apart with the method of analysis used in this thesis.

I den här masteruppsatsen undersöktes geometri- och materialparametrars p˚averkan p˚a kartongförpackningars greppstyvhet. Syftet var att avgöra hur de olika punktlasterna skiljer sig ˚at mellan olika geometrier för ett givet material.

(6)

Sammanfattning

Greppstyvhet är en viktig egenskap hos kartongförpackningar. Strukturen behöver kunna motst˚a hantering av en eller flera kunder. En kartongförpackning som upplevs som styv när den hanteras ger kunden ett intryck av lyx, ett exempel är förpackningar för nya smarttelefoner. Om en kartongförpackning är mer eller mindre greppstyv avgörs just nu subjektivt av en testpaneler. En objektiv och upprepningsbar metod för att mäta greppstyvhet snabbare behövs. I den här masteruppsatsen undersökts geometri- och materialparametrars p˚averkan p˚a kartongförpackningars greppstyvhet. Syftet var att avgöra hur de olika punktlasterna skiljer sig ˚at mellan olika geometrier för ett givet material.

M¨atningar gjordes med en dragprovare och en sensorisk m¨atenhet, Syntouch Biotac.

Endast tv˚a av 19 sensorer analyseras i det här arbetet. Datan som analyseras i det här arbetet kommer fr˚an experiment utförda av författaren samt experiment gjorda i samarbete med en kandidatuppsats vid Örebro Universitet. Samma utrustning användes för b˚ada experiment. Författarens experiment gjordes för att undersöka hur rotation av kartongförpackningarna p˚averkade resultaten. Ingen rotation utfördes under experi- menten som gjordes i samband med kandidatuppsatsen p˚a Örebro Universitet. Exper- iment utförda av författaren inneh˚aller även kompression av förpackningar gjorda med ett mänskligt finger, detta gjordes endast p˚a förpackningar som inte roterades. En FEM- studie utfördes för att validera resultaten fr˚an Syntouch Biotac. Datortomografi användes för att undersöka om skador som uppkommit p˚a kartongmaterialet av mänskligt finger var lika de som uppkommit fr˚an Syntouch Biotac.

Resultaten fr˚an Syntouch Biotac visar att det är möjligt att se var längs med eller hur nära kanten p˚a förpackningen som Syntouch Biotac trycker p˚a förpackningen, samt att det är möjligt att skilja mellan material om deras ytvikter skiljer sig tillräckligt mycket.

Datortomografin visar att skador som uppkommit p˚a kartong materialet av mänskligt finger eller Syntouch Biotac inte g˚ar att särskilja med den analysmetod som använts i det här arbetet.

(7)

Acknowledgements

First and foremost want to thank my supervisor, Christer Korin, associate Professor at Orebro University. Thank you for your support and sharing your knowledge in the areas¨ of carton board packaging and paper mechanics. I have had plenty of freedom to shape this thesis as I have seen fit and through this I have learned a lot.

Secondly I would like to extend a big thank you to Anton Jansson, PhD, at ¨Orebro University. Thank you for all your help with the CT imaging and for your patience when giving instructions on how to use the imaging software.

Lastly a thank you to my examiner, Professor S¨oren ¨Ostlund, at KTH. Your comments on my drafts have all been useful and the finished thesis is better because of them.

(8)

(9)

1 Introduction

Packaging made from cellulose based materials has been around for quite some time.

Perhaps the most familiar applications are thin paper bags or egg containers. Today fluids such as milk or fruit juice and luxury goods such as expensive whiskey and chocolate are commonly sold in carton board packaging. Primary packaging is the last piece of packaging separating the consumers hand from the product. Depending on the condition of the primary packaging the product will often be judged by the consumer and a damaged piece of primary packaging might be enough to discourage a purchase. Therefore, primary packaging must be viewed as integral to the product [1]. From a sustainability point of view carton packaging has the advantage over plastic packaging of being biodegradable over a shorter period of time [2] and a renewable resource. The largest part of the production of plastics in 2012 consisted of plastics aimed at the packaging industry. The production of plastics in 2012 reached 288 million tonnes [3]. This means that there is a lot of ground to be covered by carton board packaging. Since consumers decide for themselves what to buy and generally tend to choose less damaged packaging, it is important to understand what causes certain types of damages and how to prevent these in the future. The consumers attitude towards certain types of packaging and brands may also be affected by the appearance and performance of the packaging [4]. Further, disruption of the packaging industry by carton packaging and the move towards more biodegradable packaging material are some of the reasons behind the research carried out in the field of grip stiffness.

1.1 Aim of this thesis and reasoning behind it

This thesis aims to investigate the response of at least two of the electrodes of the Syn- touch Biotac that has not already been analyzed in [5]. Sensors already analyzed are mentioned in Section 2.1 and this thesis will focus on analyzing sensors E2 and E12. The reasons for choosing E2 and E12 to be analyzed were two fold. For the measurements that were made, the impedance graphs were consistent in shape which facilitated charac- terization of the behaviour of the sensors and made results far less difficult to compare to each other had this not been the case. The other reason also stems from measurements.

It was at first believed that mirroring the rotation of the package would just mirror the response between E2 and E12.

(12)

2 Background

Research into damage appearance and mechanisms of damage initiation and propagation during application of concentrated loads and distributed loads has been ongoing. When applying concentrated loads to carton board packaging with spherical indenters it was found that spherical indenters with a small radius initiated both primary and secondary damage. Primary damage consisted of a vertical yield line and secondary damage was a parabolic yield line. When spherical indenters with larger radii were used no primary yield line could be detected. [6]. Dependence of collapse load on application of a concentrated load on a carton board package was investigated, and it was found that the collapse load of the a carton board package (PCL) was independent of where along a line parallel to the crease the concentrated force was applied as well as that the stiffness of the package is more influenced by geometry than material and that the visual appearance of the damage had little dependence on geometry on a macro scale [4]. Regarding distributed loads, an analytical model for approximating box compression resistance (BCR) [7] has been researched which works by dividing a package into its constituent parts such as panels and flaps. Only material parameters plays an integral part. The material parameters used are the short span compression test (SCT) values and bending resistance. A high degree of accuracy was achieved when the model was compare to experimental results regarding loading direction, geometries and grammages. Regarding the measuring device Syntouch Biotac an attempt of estimating normal forces, point of contact and torque has been made while assuming a cylindrical and spherical geometry of the Syntouch Biotac.

Furthermore an attempt at separating the torque and shear forces is made, but without success. However, pure tangential forces were successfully extracted through analytical means. [8]

2.1 Data from previous thesis

Previous measurements reported in [5] are further analyzed in this master thesis. The analysis is not the same that was made in [5] but pertains to the aim in this thesis.

The analysis in [5] involves Biotac sensors E7, E8, E9 and E10 and was performed on four different kinds of packages that differed in either geometry, grammage or material.

One kind of packaging had the same geometry, grammage and material as the packaging erected and measured on for the purpose of this thesis, namely the package designated M1G1 in Table 1. The four different types of packaging are presented below and are referred to through text or plots later on.

(13)

(a) Geometry 1 (G1) (b) Geometry 2 (G2)

Figure 1: Dimensions for geometry 1 and 2 , respectively.

The data for each individual type of packaging are given in Table 1 and are taken from [5].

Table 1: Dimensions and material data for geometries 1 and 2 in figure 1

Product Grammage, (gsm) Dimensions (mm) Designated title Quantity

BK White 290 78 x 50 x 110 M1G1 10

BK White 315 78 x 50 x 110 M2G1 11

BK White 340 78 x 50 x 110 M3G1 13

BK White 290 98 x 98 x 280 M1G2 10

Packages are compressed by the Syntouch Biotac at certain points along the packages.

Depending on geometry there are either 10 or 18 test points. For clarity these are shown in Figures 2 and 3 for geometry 1 (G1) and geometry 2 (G2), respectively.

Figure 2: Compression positions for geometry 1 (G1)

(14)

Figure 3: Compression positions for geometry 2 (G2)

Geometry 1 (G1) have two insert tabs on each short side. The insert tabs are not inserted under the same panel, meaning that if the package is positioned as in Figure 1b, one insert tab is facing up and the other insert tab is facing down. Geometry 2 (G2) have one opening with an insert tab and an envelope bottom. For clarity, an envelope bottom is shown in Figure 4. Layouts for geometry 1 (G1) and geometry (G2) are presented in Figures 5 and 6.

Figure 4: Envelope bottom

(15)

Figure 5: Layout of blank for geometry 1 (G1)

(16)

Figure 6: Layout of blank for geometry 2 (G2)

2.2 Assumptions and limitations

In order to simplify and delimit the necessary work done in this thesis, some assumptions and limitations have been made.

• Effects of creep and moisture on the results are not analyzed.

• The only parameters affecting the results are assumed to be orientation of the Syntouch Biotac relative to the packaging, packaging material and geometry.

(17)

3 Theory

3.1 Paperboard

The manufacturing of paper or paperboard begins with a furnish where the water content is above 99% and below 1% wood fibres. The furnish is sprayed from a headbox onto a moving wire which acts as strainer. Water is allowed to pass through while the fibres are trapped on top of the moving wire. The difference in velocity between the moving wire and the furnish sprayed onto it gives paper and paper board anisotropic material properties. The difference in velocity aligns most of the wood fibres in the moving direction of the wire, also commonly known as the ”machine direction” and will hereafter be denoted MD. Perpendicular to MD is the ”cross direction” which is abbreviated CD. Although paper and paper board comes in thin sheets it is still a three dimensional material. While MD and CD are in-plane directions, the out-of-plane direction is the thickness direction and will hereafter be denoted ZD, see Figure 7. Because of the alignment of the fibers along MD, paper and paper board possesses higher stiffness and strength in MD than in CD. Depending on the origin of the wood fibres and the manufacturing process, the stiffness in MD can be up to 3 and 300 times greater than in CD and ZD , respectively.

Formations is the name of the non-uniform in-plane mass variability. The fibers have a tendency, during drying of paper, to form bundles which may vary in size from a few centimeters down to a few millimeters. Because material properties such as stiffness and strength is in part dependent on density, these in-plane mass variations are the same as density variations if the thickness of the paper is assumed to be uniform and thus the mass variations can lead to in-plane variation in material properties. [9]

Figure 7: Illustration of the coordinate system used when referring to paper or paper board. [9]

Apart from the type of wood used for making the fibers and the manufacturing process, the type of pulp used for making the paper product will ultimately affect its mechanical properties. Two commonly used types of pulp are mechanical pulp and chemical pulp.

Mechanical and chemical refers to the way the wood fibers have been processed in order to be divided into smaller pieces. Mechanical pulp means that the wood chips have been mechanically processed by for example grinding wheels. Simultaneously as grinding occurs, heat and water is introduced and this will soften the lignin so that the fibers in the wood chips can be separated. This processed is repeated until satisfaction. Chemical pulp starts its journey as wood chips. The wood chips are cooked in several chemical solutions that will dissolve 80-90 % of the lignin that binds the fibers. This will allow for separation of the fibers. The yield of the mechanical pulping process and the chemical process is 95 % to 50-65 % , respectively. Because of the absence of mechanical process-

(18)

ing of chemical pulp, fibers in chemical pulp are longer and often result in a stronger paper board. The fibers have to bond well to give the end product sufficient strength.To make sure the fibers can bond well they enter a mechanical process named ”beating” or

”refining”. This mechanical process also makes the fibers more flexible. [9]

It is common for paper board to be made out of several plies. Each ply can be made out of chemical or mechanical pulp and thus possess different material properties such as stiffness and strength. To obtain the desired mechanical properties of the finished paper board the plies are placed closer to or further away from the neutral plane of the paper board. The paper board thus works like a sandwich structure [10] where the outer plies provide bending, compressive and tensile strength and stiffness and the middle plies provide the bulk which gives the outer plies their distance to the neutral plane.

The stress-strain relationship is heavily influenced by moisture content (MC) and the relative humidity. That this is true can be demonstrated by soaking a piece of paper or paper board into water or taking the same piece into a steamy sauna. The more moisture that enters the piece of paper, the softer it gets and eventually the young’s modulus in any direction will decrease towards zero. The decrease in Young’s modulus for different RH and MC is shown in Figure 8.

3.1.1 Relative humidity, RH

Relative humidity, abbreviated RH, is a term used for describing the amount of water vapour in the air compared to how much water the air could at most hold at a given temperature and pressure. Relative humidity is given in percent. As an example, if the relative humidity is said to be 40 % it means that the air where the relative humidity was measured currently holds 40 % of the amount of water it could potentially hold at that given temperature. [11]

3.1.2 Moisture content, mc

The moisture content is the ratio between the mass of moisture in the paper and the total mass of the paper including the moisture, it can be written as:

mc = Mmoist

M_tot = Mtot− Mdry

M_tot (1)

where mc is the moisture content, M_moist is the mass of the moisture contained within the paper, M_dry is the mass of the paper without any moisture and M_tot is the mass of the dry paper and the moisture combined. Moisture content is measured by weighing a pice of paper, drying it and then weighing it again. On comparison there will be a mass difference, this mass difference is said to be due to loss of moisture and this mass is taken as the moisture content. [12]

(19)

Figure 8: Stress-strain curves for MD and CD. The relative humidity is 40 % and 95 % respectively and moisture contents are 6.6 % and 20 % respectively. [9]

3.2 Multiple plies

Packages made out of carton board are often manufactured to have multiple plies. In- creased number of plies usually means a thicker board which in turn implies larger bending stiffness. Bending stiffness is an important property of carton board because it it will decrease compliance. Different types of materials are used depending on where in the cross section it is to be used. The plies closest to the neutral plane are made out of mechanical pulp because of it provides bulk and thereby thickness. The outer plies are made out of chemical pulp which have higher in-plane tensile stiffness. The outer plies contributes to the bending stiffness of the carton board and takes almost all of the normal stress in the cross section while the bulky middle plies carry the shear stress. The carton board thus acts a sandwich plate where outer plies are the faces of the sandwich plate.

For simplicity one can look at the flextural rigidity of a sandwich beam to explain the effect of multiple plies.

D = Z

Ez²dz = E_ft³_f

6 + E_ft_fd²

2 + E_ct³_c

12 = 2D_f + D₀+ D_c (2) Where E_f and E_c are the youngs modulus for the faces (chemical pulp) and the core (mechanical pulp), respectively. t_f and t_c are the thicknesses for the faces and the core , respectively and d is the distance between the netural layers of the faces. The build up of the crossection results in three contributions to the total flextural rigidity D, these are:

• Df, the flextural rigidity of the faces alone when bending around their own neutral axes (black dashed lines, Figure 9).

• D0, the stiffness of the faces when bending about the neutral axes of the entire cross section (green dashed line, Figure 9).

• Dc, the flextural rigidity of the core when bending around its own neutral axes.

Simplifications to the total flextural rigidity can be made if the faces can be considered thin in comparison to the thickness of the core [10]. Figure 9 provides a visual explanation of the crossection along with the previous text.

(20)

Figure 9: Crossection of a sandwich beam. Darker color represent chemical pulp while lighter color represents mechanical pulp.

3.3 Crease

A crease is an intentional and permanent deformation created along a straight line in for example carton board. The crease facilitates folding of a blank into a package with well defined corners. The facilitation of nicely folded corners comes from the type of damage that is done to the material during creasing. As stated before, carton board can be made up of multiple layers, each attached to the adjacent layer. During creasing the carton board material is placed between a male die and a female die, the dies are often made out of a material that is rigid in comparison to to carton board material. The male die is then moved closer to and pushed into the female die. Often times the creasing operation is performed along multiple lines at once and the pattern of female dies laid out is then referred to as a matrix and has a corresponding pattern of male dies pushing the carton board material into it.

When the creasing operation is finished the expectation is that the carton board material will be easy to fold along the crease line. This can be expected since during the creasing operation delaminaton of the many different layers as well as micro cracks form due to shear stresses and compression. A piece of carton board that is folded without being creased creates an irregular folding line due to the adhesion of adjacent layers in the material. A piece of carton board that has been creased, meaning the adjacent layers are separated from each other, will allow folding along a well defined line as previously stated. Adjacent layers are allowed to slide in relation to each other and compression of the bottom layer is avoided. Besides a well defined folding line the goal is to decrease the bending resistance and thereby achieve a hinge effect, i.e a folding line with zero bending resistance. A crease is often made along either MD or CD. A crease line/folding line perpendicular to MD is referred to as an MD-crease and a crease line/folding line perpendicular to CD is referred to as a CD-crease, see Figure 10.

(21)

Figure 10: Illustration of an MD crease. [13]

Creasing geometry is important when performing a creasing operation. Crease geometry is influenced by the geometry of the male die, i.e creasing rule width and female geometry die, i.e. matrix height, channel width and channel depth. Creasing depth, i.e the depth with which the male die is pushed into the carton board, is also included in the crease geometry. To achieve a good enough crease the creasing depth has to be large enough so that the right amount of delamination is induced in the material. The channel width has to be equal to two times the carton board thickness plus the width of the creasing rule, otherwise the creasing rule runs the risk of pinching the carton board and breaking it [14]. If the channel depth is too large the crease runs the risk of losing definition which might cause irregular folding lines when folding the blank.

3.4 Grip Stiffness

Grip stiffness is a term frequently used despite not being well defined. It is used between manufacturers of paper board packaging and paper board material and their customers to describe the perceived stiffness of a product packaging when being handled by a costumer [15]. Grip stiffness can be explained by the following example:

Imagine buying orange juice at the grocery store. Normally it is no problem handling the package due to the geometry of the package and the ability of the material to maintain the geometry during handling. The package can be said to have a certain grip stiffness.

Now imagine buying the same orange juice but in a bag-like container. Everyone familiar with for example plastic bags know that they do not maintain their initial geometry when being handled. Trying to pick up a container of orange juice in this type of packaging would not be easy, assuming the container is not over filled or has some internal pressure, since one would simply displace the liquid inside, making it difficult to get a firm grip.

This type of packaging can be said to lack grip stiffness. Other examples of packaging that can be said to possess a high grip stiffness is packaging for expensive mobile phones, liquor and luxury foods.

The quest to define, to objectively measure and ultimately to predict grip stiffness is of importance because it has been shown that consumers are willing to pay more for the same product if the packaging feels good in their hands [16] and signals high quality and luxury. Thus, grip stiffness is important both to the vendor and consumer.

(22)

3.5 Industrial X-ray computed tomography

X-ray computed tomography is a non destructive testing method from which a generated 3D model of an object can be obtained through piecing together many 2D cross sectional images of said object. The goal is to gain knowledge of internal features of the object without taking it apart or otherwise damaging it. As recent as 2005 the hardware needed to make this method accurate, more available and still affordable, was developed. X-ray computed tomography, hereafter abbreviated CT, can be used for predicting material performance, investigating porosity of for example casted components, metrology and failure analysis.

Before there are X-rays there is a metal filament , often made from Tungsten. A current is passed through the filament which at some point will start emitting electrons.

A strong electric potential is used together with magnetic lenses to focus the electrons emitted onto a target material. X-rays are emitted from the target material via two phenomena, either electron absorption or deceleration/deflection of electrons. The lat- ter phenomena, deceleration/deflection of electrons rely on the fact that when electrons impact the target material they loose energy. The energy that is lost by the electrons is emitted as X-rays. This way of creating X-rays produces a continuous X-ray spectrum and is called ”Bremsstrahlung”, a word stemming from the two German words ”brems”

and ”strahlung” meaning ”brake” and ”radiation”, respectively. The incident electrons, target material and emitted X-rays are shown in Figure 11.

Figure 11: Electron beam interaction with the target material. Image courtesy of Anton Jansson [13].

X-rays have the ability to penetrate material. When x-rays penetrate an object and comes out the other end, the intensity of the x-rays has decreased an amount depending on material properties such as density. Another word for the decrease of intensity in this case is attenuation and it is the materials attenuation properties which will enable the subsequent reconstruction of the 3D volume. The attenuation of the x-rays through the material roughly follows the Lambert-Beer law which predicts an exponential decrease of intensity of an x-ray through the material.

(23)

where I is the intensity of the x-rays that have passed through the material, I₀ is the intensity of the x-rays before entering the material, µ(E) is the attenuation coefficient which is material dependent and is a function of the energy of the x-ray, x is the thickness of the material that a given x-ray had to penetrate to get to the other side. The attenuation is used to produce many 2D cross sectional images of the material, also called projections. If many projections of the same sample is taken from many viewing angles, it is possible to generate a digital 3D volume from these.

A CT system contains at least an x-ray source, a rotating plate called a stage relative to which the sample must stay completely still and a detector, as shown in Figure 12.

Figure 12: A cone beam X-ray CT setup. Image courtesy of Anton Jansson [13]

The rotating plate is placed between the x-ray source and the detector. There are different types of x-ray beams in use today, in this thesis only a beam type called cone beam is used and hence is the only one referred to, shown in Figure 13c. CT systems that use a cone shaped beam illuminate the entire sample once per angle increment as shown in Figure 13.

Figure 13: Different CT setups. X-ray sources, rotating samples and detectors are shown for each. a) Pencil beam setup b) Fan beam setup c) Cone beam setup d) Parallel beam setup. Image courtesy of Anton Jansson [13].

(24)

When a full revolution of the sample has been completed, enough images should have been obtained so that a generation of a complete 3D volume is possible. A schematic of a reconstruction is shown in Figure 14.

Figure 14: A reconstruction is shown schematically. Blue fences represent the detector at different angles when the sample rotates. Image courtesy of Anton Jansson [13].

3.6 Syntouch Biotac

The Syntough Biotac is a tactile sensor capable of detecting sensory information that a real finger can detect. This sensory information consist of temperature, forces and vibrations. Sensors which facilitate the detection of this sensory information are embedded in a rigid core and are different in kind. The different type of sensors and what they can detect are:

• 19 sensing electrodes evenly spread out over the rigid core capable of detecting changes in impedance as the incompressible fluid is deformed during contact between an object and the Syntouch Biotac. Sensors are denoted E1 - E19. These act together with four excitation electrodes, X1 - X4. When an electrode is sampled a voltage is passed into the excitation electrodes. Four pathways of conduction from each excitation electrode to each sensing electrodes guarantees that the change in

(25)

between the elastomeric skin and the sensor increases, and increases as the distance decreases [8]. This means that the impedance of the each given sensor can be interpreted as a measurement of the distance between the sensor and the skin, and with the information of the impedance for each sensor one could perhaps imagine the shape of the Biotac during measurement. The layout of sensors E1 to E19 and excitation electrodes is shown in Figure 15

• A hydro-acoustic pressure transducer that can detect vibrations that are created as the Syntouch Biotac slides different surfaces.

• A thermistor located at the tip of the Syntouch Biotac that detects changes in temperature due to thermal gradients that arise during contact between different materials and the Syntouch Biotac.

Figure 15: Layout of sensors E1-E19 and excitation electrodes on the Syntouch Biotac.

(26)

4 Method & Materials

In this section it is described how the equipment to perform the relevant measurements was used. The Syntouch Biotac was used together with the Lloyd LR5K tensile tester to perform the measurements pertaining to grip stiffness and the x-ray tomograph SKYSCAN 1272 from BRUKER was used to perform the x-ray of the paper board packaging, needed to inspect the damage done during measurements with Syntouch Biotac and Lloyd LR5K tensile tester.

4.1 Handling of the paperboard packaging

The packaging used to conduct experiments in this thesis is made out of BillerudKorsn¨as White with surface weight 290 gsm. The dimensions of the packaging are shown in Figure 16. The packaging was erected from templates on Feb 4 2019 and Feb 8 2019. Prior to using the packaging for any measurements the packaging was relaxed for at least 72 hours after being erected.

Figure 16: Package geometry

4.2 Testing with Lloyd LR5K tensile tester & Syntouch Biotac

The package was placed on the steel table belonging to the Lloyd LR5K tensile tester as shown in Figure 17a. Black lines were drawn using the packaging as template so that packages could be switched out between measurements and still maintain the same position and rotation relative to the Syntouch Biotac. The middle of the panel was used as a pivot line and lines representing 0°, +30°and -30°. The pivot line can be seen in Figures 17a and 17b along with lines parallel to 0°and -30°. In Figures 18a and 18b lines parallel to 0°, -30°and 30°are shown along with the pivot line seen from above as a red dot. The intersection of these three lines is the point around which the red pivot line in Figures 17a and 17b was rotated.

(27)

(a) Orientation 1, rotation is 0° (b) Orientation 2, rotation is -30° Figure 17: Illustration of the positioning of the packaging during measurements.

(a) Lines drawn for different orientations

(b) Illustration of the different orientations

Figure 18: Illustration of the positioning of the packaging during measurements.

When positioning the package parallel to the 0°-line as in Figure 17a, the package was placed so that the first point of contact with the Syntouch Biotac occured 10 mm from the crease. This is shown in Figure 19a. Also, the package was positioned so that the first point of contact would occur in the middle of the length of the crease. This was to avoid as much of the load bearing effects of the corners as possible, since corners are capable of carrying more of the applied load. Also, the corner marked ”crease” in Figure 19a was always positioned so that it would be diagonally opposite to the glue flap, this was to avoid the influence of the glue flap on the stiffness [4], i.e the slope of the force-displacement curve.

(28)

(a) Point of contact

(b) Glue flap

Figure 19: Illustration of first point of contact as well as positioning of the glue flap in relation to first point of contact.

Deciding upon the zero point for displacement was done by approacing the package with the Syntough Biotac attached to the Lloyd LR5K Tensile Tester until the force reading from the Lloyd LR5K read 0.2 N ± 0.02 N. Upon reaching this force value the displacement and force was tared, setting both to zero, and the Syntouch Biotac was elevated an arbitrary amount but preferably just so a small gap between the Syntouch Biotac and the package was visible.

The data recording for the Syntouch Biotac was then started and subsequently the test sequence for the Lloyd LR5K was initiated. The Lloyd LR5K would move past the previously set zero point and continue down 10 mm into the package, thereby compressing it. Upon reaching 10 mm the Lloyd LR5K would return to the zero point and thereby completing one test. The Lloyd LR5K was elevated slightly above the package and the package was removed, marked with the appropriate marking and stored away.

Furhtermore, the package was also positioned on the steel table so that the Syntouch Biotac would touch down as far away as possible from the short sides of the package.

The crease closest to the Syntouch Biotac, denoted ”crease” in Figure 19a is the crease located the furthest away from the glue flap of the package, the position of which is pointed out in Figure 19b. This was done on purpose since it is known that corners take most of the load during box compression and that the orientation of the package so that force application is directly on to the glue flap strongly influences the force-displacement curves [4], [7].

(29)

4.3 X-ray tomography

The damage done to the package by the Syntouch Biotac was inspected by eye. A piece containing the entirety of the damages material was cut out from the package. A pair of scissors was used to remove the peripheral material around the damage so that only the damaged piece of carton remained. This is shown in Figure 20. The size of the carton board piece after removing excess material varied depending on the size of the damage that could be observed by looking at the package.

Figure 20: Two damaged carton board pieces. One with excess material (left) and one without excess material (right).

The damaged pieces of carton were placed into a holder with a capacity of four pieces per holder, a 3D-model of the holder is shown in Figure 22. The carton pieces were fixed in place by applying Loctite super glue to the tracks of the holder prior to mounting the carton board pieces. The holder, together with the four carton board pieces attached to it were mounted inside the X-ray tomograph into to specimen mount that is shown in Figure 21.

Figure 21: Mount for CT specimen inside the CT scanner.

Prior to starting the x-ray procedure the holder, while being mounted inside the x-ray tomograph, was left alone for five hours. This was so that any residual stresses in the glue or glue-carton interface could relax. The size of each voxel, i.e the resolution of the

(30)

scan, was approximately 5 µm . This means that any movement beyond 5 µm would create a blurry and useless scan.

After completing one scan, the holder was removed from the x-ray tomograph and stored away. The data was then transferred to a computer where all pictures taken during the scan was processed and the resulting 3D volume was treated with the analysis software for industrial computed tomography, VG Studio MAX 3.0 .

Figure 22: Holder used for holding up to four carton board CT specimen.

4.4 Gathering and analyzing data from Biotac

4.4.1 Biotac GUI ver 3

Gathering of experimental data from Biotac was done through a Biotac GUI provided by Syntouch Inc, Biotac GUI ver.3. The software records data for pressure vibration, temperature heat transfer and impedance. Impedance is recorded for each of the sensors denoted E1 through E19. The software was freely available from Syntouch webpage.

4.4.2 Grippy

The tensile tester Lloyd LR5K and the Syntouch Biotac runs on independent clocks.

To use the data gathered from the Biotac along with data from the Lloyd LR5K, their independent clocks needed to be synchronized. This was done through a proprietery and freely available software written by Eriksson ([4], [6]). Grippy outputs a .csv-file containing force- and displacement data from the Lloyd LR5K as well as all the data gathered from the Syntouch Biotac. A common time is also included for the synchronized data.

(31)

4.4.3 Analyzing Biotac data

The data that was analyzed from the Syntouch Biotac was only the impedance for sensors E2 and E12. The data for these two sensors was gathered both from the authors own measurements and from measurements done by Abdulkareem and Al-Radi [5].

Analysis of the data was carried out by first graphing each sensor data and then performing a visual inspection. An example of such curves is shown in Figure 23. Based on that the impedance for each sensor decreases from its initial value when the skin above it is deformed so that less fluid is between it and the sensor and that the opposite is true for when the distance between the sensor and the skin increases [8] one can deduce if the Biotac is compressed on one side more than on the other by looking at impedance graphs for sensors positioned on opposite sides of the Biotac, as sensors E2 and E12 are.

Figure 23: Representative curves of Impedance vs Time for sensors E2 and E12 for when the Lloyd LR5K was displaced 10 mm into the package, thus collapsing the package.

For measurements made by the authors of [5] two values were sought after. One was the drop in impedance from its starting value to the first minimum before the peak impedance value, the other value was the impedance difference between the first minimum before peak value and the actual peak impedance value. These values are clarified in Figures 24 and 25, respectively.

Starting value for the impedance, impedance at first minimum and peak impedance were all identified automatically and their differences were computed. The following example will enlighten the reader on how the tables in the results subsections shown in Appendix A.1, A.2, A.3 and A.4 were created.

For example, packages made by material one with geometry 1 (abbreviated M1G1) were all tested along the same side. Each package was tested at a total of 10 points arranged in a 2 by 5 grid. For M1G1 there was a total of 10 packages tested and thus for each position, 1A for example, there would be a total of 10 curves. The same applies to

(32)

all other positions. For each position at a time, the relevant values would be computed and stored. The mean value would be computed for each of these relevant values as well as their standard deviations. The computed mean values and corresponding standard deviations, for each position, would then be put in a table and also plotted using the built in function for error bar plots in MATLAB R2018a. Tables for the mean value and standard deviations of impedance difference between starting impedance and impedance at first minimum, as well as tables for the mean value and standard deviations of impedance difference between first minimum and impedance at peak impedance are found in the subsections A.1, A.2, A.3 and A.4 in the Appendix.

Figure 24: Impedance vs time curve with arrows pointing to the points on the curve between which the difference between starting impedance and impedance at first minimum is measured. This distance is abbreviated SIFM.

(33)

Figure 25: Impedance vs time curve with arrows pointing to the points on the curve between which the difference between impedance at first minimum and peak impedance is measured. This distance is abbreviated PIFM.

4.4.4 Measured global stiffness

In [5] maximum global stiffness as measured by the Lloyd LR5K Tensile tester was measured and the data was saved. This data is simply taken from the data provided by the authors of the article, the same data that is further analyzed in this thesis.

4.5 Numerical simulation of carton board sheet

Pushing down on a carton board sheet was simulated with Comsol Multiphysics 5.4. The composite module was used and layered shells were employed. BK White 290, the same material used for M1G1 and M1G2, was approximated using a three-layer composite.

Each layer was assumed to be linear elastic and orthotropic. The outer layers were given material properties for chemical board while the middle layer was given material properties for mechanical board, material properties are found in Table 2. The layup is unidirectional, meaning that each layer is oriented as the layer in Figure 27. A prescribed load of 60 N was applied onto a circular area with its center at the same positions as described in [5] when testing the packages with Syntouch Biotac, the circular area is shown in Figure 26 along with direction of movement (blue arrow) for the prescribed load between each simulation. Boundary conditions for the composite plate are shown in Figure 26. Long sides (side 2) are locked in translation in x, y and z-directions while the short sides (side 1) are locked in translation in z-direction. Sides 1 and 2 are clarified in Figure 27.

(34)

Displacement data for every position was extracted from Comsol and treated in MAT- LAB where a high-degree polynomial was fitted to the displacement data points. The discrete data points and the fitted polynomial is shown in Figure 28. What is being sought after is the absolute value of the gradient of the displacement curve at the positions of sensors E2 and E12. Where the slope is extracted for sensor E2 and E12 is shown in Figure 29. The values for each position for both M1G1 and M1G2 are collected and plotted in results section 5.6.1 in Figures 57 through 60.

Table 2: Material properties of bottom, middle and top ply of the simulated paper board.

Material properties are taken from [17]

Bottom ply Middle ply Top ply

E₁[M P a] 8800 3200 5900

E2[M P a] 130 160 230

E₃[M P a] 3000 1200 2700

G₁₂[M P a] 68 30 60

G₁₃[M P a] 1600 640 1400

G23[M P a] 68 30 60

ν₁ [-] 0 0 0

ν₂ [-] 0.51 0.47 0

ν₃ [-] 0 0 0

Thickness [µm] 20 380 20

Density [kg/m³] 690 690 690

Table 3: Board dimensions for M1G1 och M1G2 in Figure 27 Side 1 [mm] Side 2 [mm]

M1G1 78 110

M1G2 98 280

Figure 26: Application of force (red arrow) and direction of movement (blue arrow) of applied force between positions. Boundary conditions are also shown.

(35)

Figure 27: Layup of unidirectional composite.

Figure 28: Displacement data from COMSOL (blue circles) and polynomial fitted to displacement data (orange curve). Part of the curve in red is the area of interest for the fit, here the fit should be perfect.

(36)

Figure 29: Displacement curve and corresponding first derivative (slope) at position 5B for M1G1 together with positions for E2 (green vertical line) and E12 (light blue vertical line) where the value of the slope. Touchcenter (black vertical line) is where the Biotac first touches the package.

(37)

5 Results

Results in sections 5.1 and 5.2 pertain to packages tested by the author and rotated

±30°and 0°, respectively. These packages are identical in both material and geometry to packages designated M1G1, which were tested by Abdulkareem and Al-Radi [5].

5.1 Packages rotated ± 30 °

For packages turned to +30°and -30°according to Figure 18b representative results are shown. Package 28 is chosen to represent -30°while package 39 is chosen to represent +30°.

A visual inspection of the curves reveal that when the packages are turned -30°thereby placing sensor E2 further away from the crease, sensor E2 displays a flatter top while sensor E12 displays a narrower top. When packages are turned -30°, sensor E12 is closer to the crease than at 0°or +30°.

Regarding the top of the curves for sensors E2 and E12 the roles are switched when turning the packages to +30°. At +30°sensor E2 is closer to the crease and sensor E12 is furhter away from the crease. As stated before it is possible to see a change in flatness of the top of both curves for +30°when compared to -30°. The top of the curve for sensor E2 is more pronounced while the top of the curve for sensor E12 is flatter.

Upon further inspection the curves for sensor E2 display a larger drop from its initial value before increasing. This first drop in impedance from the initial value looks the same for all E2 sensors independent of turning the packages +30°or -30°. The same magnitude in first impedance drop from the initial value cannot be observed from the curves for sensor E12.

All curves can be seen independently and in a common plot in the Appendix.

(a) Impedance vs time, sensor E2 (b) Impedance vs time, sensor E12

Figure 30: Representative impedance vs time curves for sensors E2 and E12 for packages turned to -30° and compressed 10 mm by the Lloyd LR5K.

(38)

Figure 31: Representative impedance vs time curves for sensors E2 and E12 for packages turned to +30° and compressed 10 mm by the Lloyd LR5K.

5.2 Packages rotated 0 °

For packages turned 0°according to Figure 18b, curves in Figure 32 are representative.

Similiar flatness of the top of the curves can be observed but the initial drop in impedance from the initial value is greater for sensor E2 than sensor E12. This behavior was consistent for all curves, which are shown independently and together in the appendix.

Figure 32: Representative impedance vs time curves for sensors E2 and E12 for packages turned to 0° and compressed 10 mm by the Lloyd LR5K.

5.3 Tomography

X-ray computed tomography scans of packages in subsection 5.2 are presented in this section. Results regard whether or not the visible damage is symmetric around the initial damage which is a vertical yield line or not. Figures in this subsection show the outside of the material. To view Figures that show the inside of the material the reader is referred to the appendix. Only representative images with explanation are shown here. Subsections named ”Package” refer to damages done to the material by the Syntouch Biotac while subsections named ”Manual” refer to damages done to the material by the authors right index finger. Textboxes with ”E2 side” and ”E12” side points to what side of the initial

(39)

through 40. The side of the sample where the vertical yield line is pointing too is the blue, vertical, side of the package which during the experiment is parallel to the loading direction.

5.3.1 Package 1

Figure 33: The vertical yield line is shown together with the secondary damage, the parabolic yield lines. The parabolic yield lines on each side of the vertical yield line are different in appearance. The parabolic yield line on the E12-side is not as visible. The damage done on package 1 looks to be asymmetric on either side of the vertical yield line, i.e at the force application point. The smaller arc is on the ”E12 side” and the larger arc is on the ”E2 side”.

(40)

5.3.2 Package 14

Figure 34: The initial damage, the vertical yield line, is shown together with the secondary damage. The parabolic yield lines are referred to here as ”smaller arc” and ”larger arc”.

The two parabolic yield lines do not appear the same on either side of the vertical yield line, i.e the damage is asymmetric around the vertical yield line. The smaller arc is on the ”E12 side” and the larger arc is on the ”E2 side”.

5.3.3 Package 16

Figure 35: The vertical yield line is shown together with the parabolic yield lines. When viewing the convex side of the sample the parabolic yield lines show up clearly and are unevenly distributed on either side of the vertical yield line, i.e the secondary damage is

(41)

5.3.4 Package 18

Figure 36: The initial and secondary damage are shown together. The many yellow arrows point to the parabolic yield lines which are distributed asymmetrically around the initial damage. The smaller arc is on the ”E2 side” and the larger arc is on the ”E12 side”.

(42)

5.3.5 Manual 1

Figure 37: The initial and secondary damage are shown together. The two parabolic yield lines are denoted ”smaller arc” and ”larger arc” and are distributed asymmetrically about the initial damage. The smaller arc is towards the ”E2 side” and the larger arc is towards the ”E12 side”. Since this package has been compressed manually, text boxes indicating ”E2 side” or ”E12 side” should be taken as the sides where the sensors would be placed if the Syntouch Biotac was performing the compression instead of an actual finger.

(43)

5.3.6 Manual 3

Figure 38: Initial and secondary damage are shown together. The two parabolic yield lines are denoted ”smaller arc” and ”larger arc” and are distributed asymmetrically about the initial damage. The smaller arc is towards the ”E12 side” and the larger arc is towards the ”E2 side”. Since this package has been compressed manually, text boxes indicating

”E2 side” or ”E12 side” should be taken as the sides where the sensors would be placed if the Syntouch Biotac was performing the compression instead of an actual finger.

(44)

5.3.7 Manual 7

Figure 39: Initial and secondary damage are seen together. The two parabolic yield lines are denoted ”smaller arc” and ”larger arc” and are distributed asymmetrically about the initial damage. The smaller arc is towards the ”E2 side” and the larger arc is towards the ”E12 side”. Since this package has been compressed manually, text boxes indicating

(45)

5.3.8 Manual 8

Figure 40: Initial and secondary damage are shown together. The two parabolic yield lines are denoted ”smaller arc” and ”larger arc” and are distributed asymmetrically about the initial damage. The smaller arc is towards the ”E2 side” and the larger arc is towards the ”E12 side”. Since this package has been compressed manually, text boxes indicating

(46)

5.4 Analysis of provided measurement data

This section holds results for the further analysis that was made of the measurement data collected in [5]. Tables containing statistics for both PIFM and SIFM for sensors E2 and E12 are shown all together in subsections B.1 and B.2, respectively. This is to show the dependency of PIFM and SIFM on the board wuality and package geometry. Error bar plots for each sensor are grouped together for the same reason. Error bar plots are shown first for position A followed by position B.

5.5 PIFM E2 & E12

Error bar plots for PIFM are presented in this subsection. First four error bar plots show PIFM for position A, the last four error bar plots show PIFM for position B. Both sensors E2 and E12 are plotted in each plot for comparison between the two. Figures 41, 42 and 43 show PIFM at position A for geometry 1 (G1) as board quality increases. Figure 44 show PIFM for the lowest board quality for geometry 2 (G2) at position A.

Figures 45, 46 and 47 show PIFM at position B for geometry 1 (G1) as board quality increases. Figure 48 show PIFM for the lowest board quality for geometry 2 (G2) at position B.

For error bar plots pertaining to position A and geometry 1 (G1) error bars become more separated as the board quality increases. The same does not hold true for plots pertaining to position B and geometry 1 (G1). For board qualities 315 gsm (M2) and 340 gsm (M3) , plots pertaining to position A shown that sensor E12 senses a larger PIFM than sensor 2, for plots pertaining to position B this is not true all the time. For geometry 2 (G2), error bars are more separated for position A than for position B.

For error bar plots pertaining to position B, Geometry 2 (G2) in Figure 48, they switch places as the Syntouch Biotac moves from position 1B to 9B. This means that sensor E2 senses a larger PIFM than sensor E12 at position 1B and the opposite is true for position 9B. The same does nor entirely hold for error bar plots in Figure 44 where the error bars at positions 1A and 9A coincide.

(47)

Figure 41: Error bar plot for PIFM, position A, Sensor E2 (blue) and E12 (orange), M1G1.

(48)

(49)

Figure 45: Error bar plot for PIFM, position B, Sensor E2 (blue) and E12 (orange), M1G1.

(50)

(51)

5.6 SIFM E2 & E12

Error bar plots for SIFM are presented in this subsection. First four error bar plots show SIFM for position A, the last four error bar plots show SIFM for position B. Both sensors E2 and E12 are plotted in each plot for comparison between the two. Figures 49, 50 and 51 show SIFM at position A for geometry 1 (G1) as board quality increases. Figure 52 show SIFM for the lowest board quality for geometry 2 (G2) at position A.

Figures 53, 54 and 55 show SIFM at position B for geometry 1 (G1) as board quality increases. Figure 56 show SIFM for the lowest board quality for geometry 2 (G2) at position B.

For position A, geometry 1 (G1), error bars are separate at position 1A. Moving towards position 5A they close in on each other more, even coinciding for M1G1 in Figure 49. For M2G1 and M3G1 however the error bars only decrease their relative distance with the minimum relative distance reached at position 5A. For position A, geometry 2 (G2) in Figure 52, error bars switch place between position 1A and 9A. Sensor E2 senses a larger SIFM at position 1A than sensor E12. The opposite is true for position 9A. The relative distance between the error bars for both sensors are smaller at position 1A than at position 9A where the error bars have moved further apart.

For position B, geometry 1 (G1), error bars are separated at position 1B and decreas- ing their relative distance when moving towards position 5B. For M1G1 in Figure 53 the error bars even coincide past position 2B.

For geometry 2 (G2) in Figure 56 error bars are separated at position 1B with sensor E2 detecting a larger SIFM than sensor E12. At position 9B error bars for sensors E2 and E12 have switched places and are now at a larger distance relative to each other than at position 1B. Looking at Figure 56 one can see that this has been a gradual movement as the Syntouch Biotac progressed from position 1B to 9B.

(52)

Figure 49: Error bar plot for SIFM, position A, Sensor E2 (blue) and E12 (orange), M1G1.

(53)

(54)

Figure 53: Error bar plot for SIFM, position B, Sensor E2 (blue) and E12 (orange), M1G1.

(55)

(56)

5.6.1 Numerical simulation of carton board sheet

Results for numerical simulations are shown in this subsection. Figures 57 and 58 show results for M1G1 and Figures 59 and 60 show results for M1G2. For all results, the last position is a mirror image of the first position. Positions in between show that the slopes for both sensor E2 and E12 are similar to each other and that they might be difficult to tell apart.

Figure 57: Absolute values for slope of displacement for positions 1A - 5A for M1G1.

(57)

Figure 58: Absolute values for slope of displacement for positions 1B - 5B for M1G1.

Figure 59: Absolute values for slope of displacement for positions 1A - 9A for M1G2.

(58)

Figure 60: Absolute values for slope of displacement for positions 1B - 9B for M1G2.

5.6.2 Measured global stiffness

Table 4: Measured global stiffness at positions 1A, 1B, 5A and 5B for one representative package of type M1G1, as measured by the Lloyd LR5K.

Position, M1G1 Global stiffness [N/m]

1A 6078.2

1B 5867.2

5A 4819.3

5B 1201.8

1A 5997.4

1B 6359.2

5A 5106.9

5B 2097.7

(59)

1A 7900.6

1B 5237.8

5A 7437.1

5B 5043.4

1A 6531.6

1B 4236.6

9A 5115.1

9B 3889.4

(60)

6 Discussion

Results are discussed in this section. A brief overview is given of what has been done.

Results of SIFM and tomography are then discussed. When using the term ”error bar plots” in this discussion it pertains only to SIFM error bar plots. Results for PIFM are not brought up because PIFM could not discern between geometries and/or materials sufficiently well.

Packages identical to M1G1 were tested by the author. Syntouch Biotac was positioned in the middle along the crease and displaced 10 mm down into the package. The impedance curves for sensors E2 and E12 were not identical even though the Syntouch Biotac was positioned in the middle along the crease. To investigate if the impedance curves of sensors E2 and E12 could be in practice interchanged by rotation, the packages were rotated ±30°. What was mirrored was the area just around the global maximum as can be seen in section 5.1. What was not mirrored, however, was the SIFM for sensors E2 and E12. Representative impedance curves in Figures 30 and 31 show what SIFM remains larger for sensor E2 independent of rotation.

Results for SIFM in subsection 5.6 show that sensor E2 almost always retains a larger SIFM than sensor E12 for packages with geometry 1. The exception is for material 1 (M1) at positions 3, 4 and 5. Packages tested by the author for this thesis as well as packages M1G1, M2G1, M3G1 tested in [5] all had identical geometry and were oriented just the same, i.e 0°.

The reason behind the difference in SIFM between E2 and E12 seems to be dependent on geometry. Packages tested by the author for which results are given in sections 5.1 and 5.2 and packages designated M1G1, M2G1 and M3G1 all have identical geometry. The Syntouch Biotac was at all times oriented relative to the packages so that sensor E2 was directed towards the side of the package with the insert tab crease facing up and sensor E12 was directed towards the side of the package with the insert tab facing up. Positions 1A and 1B are closest to the corner where the crease and insert tab crease meet while positions 5A and 5B are closest to the corner where the crease and insert tab meet.

FEM results show a greater slope at the position for sensor E2 than position E12 at starting positions 1A and 1B. At the end positions, 5A, 5B, 9A or 9B, the slope at the position for sensor E12 is greater. This makes one expect the SIFM to decrease for sensor E2 and increase for sensor E12 when moving from start position to end position on the package since SIFM for one sensor is a measurement of the distance between the skin-like material and that sensor because impedance decreases as displacement of the fluid close to the sensor increases. This is exactly what is shown in the results in section 5.6. The FEM results also show that if symmetric boundary conditions are imposed, results should be symmetrical. This is however not the case and the reason for this is attributed to the asymmetric boundary conditions of the package.

Figures 49 through 51, pertaining to position A, display error bar plots for geometry 1 (G1) with material 1, 2 and 3 , respectively. Error bars for M1G1 start to coincide already at position 3, meaning that sensors E2 and E12 become impossible to tell apart

(61)

M3G1, error bars are consistently separated and sensor E2 displays a larger SIFM than sensor E12 at all positions. For M2G1 and M3G1 it is even possible to see if the Biotac is at position 1 or 5. The increase in grammage when going from material 1 (M1) to material 3 (M3) seems to compensate for the loss in global stiffness when going from positon 1 to position 5. Higher grammage also seems to have an effect on the size of the standard variation, error bars for M2G1 and M3G1 are in general shorter than error bars for M1G1.

Figures 53 through 55, pertaining to position B, display error bar plots for geometry 1 (G1) with material 1, 2 and 3 , respectively. Error bars for M1G1 start coinciding already at position 3, meaning that sensors E2 and E12 become impossible to tell apart starting from position 3 going to position 5. It is not known why, for M1G1, error bars for E2 and E12 are much further apart at position 2B than they were at position 2A.

One explanation could be that since position 2B is further away from the crease that is perpendicular to the Biotac and that at all B-positions sensors E2 and E12 are closer to the crease. The crease gives a more rigid support than the panel itself leading to a greater difference in SIFM. Error bar plots for M2G1 and M3G1 show similar results for position B as they did for position A, with the exception that error bars at position 5 now coincide. This could be attributed to the significant decrease in stiffness from position 5A to 5B. The increase in grammage going from material 1 (M1) to material 3 (M3) does not seem to compensate enough for the loss in structural stiffness at position 5B to keep the error bars apart.

Error bar plots for M1G2 in Figures 52 and 56 show that error bars not only coincide but also switch relative positions when going from position 1 to position 9. For position A in Figure 52 error bars at positions 2, 3, and 4 coincide. For position B in Figure 56 error bars at positions 2, and 3 coincide. Since the only difference between these two plots is the Biotacs distance from the crease perpendicular to the Biotac, all differences can be attributed to this. The fact that the error bars switch relative positions is attributed to the larger size of geometry 2 (G2). The short side boundary on one side has little influence on the result when measurements were made on the opposite side.

CT imaging performed on packages tested by the author have not yet shown anything that can be related to the other results. CT images displayed an asymmetry of the damage however, but this asymmetry did not maintain its appearance in the sense that the smaller or larger parabolic yield line was always towards the same side of the package.

The packages that were CT scanned were however compressed at the middle along the long side crease of the package, i.e furthest away from the short side boundaries. Testing at a maximum distance from both short side boundaries, i.e minimizing their influence on the result, should increase the influence of both material and crease quality on the result. It is too difficult at this stage to say how large of an effect either of these have had on the damage appearance or on other results. It is however the authors qualified guess that if the Biotac is moved closer to either short side crease, the damage will still be asymmetric around the vertical yield line but have a consistent appearance in terms on which side of the vertical yield line the smaller or larger parabolic yield line is. The appearance of the damage done to the material by the Syntouch Biotac or the authors finger were similar and with the method of analysis in this thesis manual damage and damage done by the Syntouch Biotac could not be separated.

(62)

7 Conclusion

• With the work done in [5] and in this thesis, all sensors from the tip of the Syntouch Biotac down to and including E2 and E12 have been characterized.

• Under certain circumstances, the Syntouch Biotac and SIFM can be used to discern between grammages for packages of a given geometry.

• From CT imaging of damaged packages it is concluded that the Syntouch Biotac sufficiently well can mimic the damage done to a package by a human finger.

(63)

Appendices

A Impedance curves of provided measurement data

A.1 Impedance curves, M1G1

Results for M1G1 in Table 1 are presented in this appendix section. Figures A.1 through A.5 show impedance curves for sensors E2 and E12 for positions 1A to 5A for all packages where tests went well. Likewise Figures A.6 through A.10 show impedance curves for sensors E2 and E12 for positions 1B to 5B. These curves can be read together with their corresponding error bar plots in subsection 5.5 and subsection 5.6. The error bar plots can help confirm the shape changes of the impedance curves one can see when going through Figures A.1 through A.10.

(a) M1G1, position 1A, E2 (b) M1G1, position 1A, E12

Figure A.1: Impedance curves for E2 and E12. Position 1A on package type M1G1.

(a) M1G1, position 2A, E2 (b) M1G1, position 2A, E12

Figure A.2: Impedance curves for E2 and E12. Position 2A on package type M1G1.

Differentiating between packaging material and geometry using the Syntouch Biotac

Differentiating between packaging material and geometry using the Syntouch Biotac

HENRY ERIKSSON

Differentiating between packaging material and geometry using the Syntouch Biotac

Abstract

Sammanfattning

Acknowledgements

Contents

1 Introduction

1.1 Aim of this thesis and reasoning behind it

2 Background

2.1 Data from previous thesis

2.2 Assumptions and limitations

3 Theory

3.1 Paperboard

3.2 Multiple plies

3.3 Crease

3.4 Grip Stiffness

3.5 Industrial X-ray computed tomography

3.6 Syntouch Biotac

4 Method & Materials

4.1 Handling of the paperboard packaging

4.2 Testing with Lloyd LR5K tensile tester & Syntouch Biotac

4.3 X-ray tomography

4.4 Gathering and analyzing data from Biotac

4.5 Numerical simulation of carton board sheet

5 Results

5.1 Packages rotated ± 30 °

5.2 Packages rotated 0 °

5.3 Tomography

5.4 Analysis of provided measurement data

5.5 PIFM E2 & E12

5.6 SIFM E2 & E12

6 Discussion

7 Conclusion

Appendices

A Impedance curves of provided measurement data

A.1 Impedance curves, M1G1