Numerical and Experimental Investigation of Thermoplastics in Multi-Axis Forming Processes

Available online at www.sciencedirect.com Available online at www.sciencedirect.com

ScienceDirect ScienceDirect

Available online atonline www.sciencedirect.com Available at www.sciencedirect.com Procedia CIRP 00 (2019) 000–000 Procedia CIRP 00 (2019) 000–000

ScienceDirect ScienceDirect

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

Procedia CIRP 00 (2017) 000–000 Procedia CIRP 85 (2019) 96–101 www.elsevier.com/locate/procedia

2nd Conference on Composite Material Parts 2nd CIRP CIRP on Conference onMaterial Composite Material Parts Manufacturing Manufacturing 2nd CIRP Conference Composite Parts Manufacturing (CIRP-CCMPM 2019)

Numerical Investigation of Thermoplastics in Numerical and and Experimental Experimental Investigation Thermoplastics in Multi-Axis Multi-Axis 28th CIRP Design Conference, Mayof 2018, Nantes, France Forming Forming Processes Processes A new methodology to analyze the functional and physicala architecture of b a a a Klaus Dröder b, Bernd-Arno Behrensa, Florian Bohnea, Alexander Chugreeva, Henrik Schulzea, Klaus Dröder , Bernd-Arno Behrens , Florianoriented Bohne , Alexander Chugreev , Henrik Schulze , b product family identification existing products for an assembly Birk Wonnenbergb* Birk Wonnenberg * Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat

Institute of Forming Technology and Machines, Leibniz Universität Hannover, An der Universität 2, 30823 Garbsen, Germany Institute of Forming Technology and Machines, Leibniz Universität Hannover, An der Universität 2, 30823 Garbsen, Germany b Technische Universität Braunschweig, Institute of Machine Tools and Production Technology, Langer Kamp 19b, 38106 Braunschweig, Germany b Technische Universität Braunschweig, Institute of Machine Tools and Production Technology, Langer Kamp 19b, 38106 Braunschweig, Germany Écoleauthor. Nationale d’Arts et Métiers, Arts et Métiers ParisTech, [email protected] EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France * Corresponding Tel.:Supérieure +49-531-391-7178; fax: +49-531-391-5842. E-mail address: * Corresponding author. Tel.: +49-531-391-7178; fax: +49-531-391-5842. E-mail address: [email protected] a a

* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]

Abstract Abstract Process planning for multi-axis forming presses is a particular challenge. This process provides the option to actively influencing the material Abstract Process planning for multi-axis forming presses is a particular challenge. This process provides the option to actively influencing the material flow in the forming process by defining a six dimensional tool motion path and the tool velocity. By comprehending this interaction, it is possible flow in the forming process by defining a six dimensional tool motion path and the tool velocity. By comprehending this interaction, it is possible to today’s control and thereby tailor the induced local material properties the workpiece. Experiments were conducted a multi-axis press, which In environment, the trend towards more product of variety and customization is unbroken. Due to with this development, the need of to control business and thereby tailor the induced local material properties of the workpiece. Experiments were conducted with a multi-axis press, which is based on a Stewart platform. A simple plane workpiece geometry is chosenproducts to analyse the flow behaviour and the temperature evolution of the agile and reconfigurable production systems emerged to cope with various and product families. To design and optimize production is based on a Stewart platform. A simple plane workpiece geometry is chosen to analyse the flow behaviour and the temperature evolution of the glass mat (GMT) forming process.product Subsequently, numerical of the multi-axis forming is carried systems asthermoplastics well as to choose the during optimalthe product matches, analysisaamethods aresimulation needed. Indeed, most of the knownprocess methods aim to glass mat thermoplastics (GMT) during the forming process. Subsequently, numerical simulation of the multi-axis forming process is carried out andavalidated with the experimental data. The numerical analysis product focuses families, on the material modelling as largely well as in theterms prediction of the flow analyze product or one product family on the physical level. Different however, may differ of the number and out and validated with the experimental data. The numerical analysis focuses on the material modelling as well as the prediction of the flow characteristics. Regarding material modelling of GMT,comparison an extensive material characterization is performed tocombinations describe the for flowthe behaviour. A nature of components. This fact impedes an efficient and choice of appropriate product family production characteristics. Regarding material modelling of GMT, an extensive material characterization is performed to describe the flow behaviour. A prediction of the flow behaviour of GMT with reference to tool motion is enabled. For the FE simulation the element-free Galerkin method (EFG) system. A new methodology is proposed analyze existing products of their functional and physical architecture. The aim is to cluster prediction of the flow behaviour of GMT to with reference to tool motionin is view enabled. For the FE simulation the element-free Galerkin method (EFG) is applied for modelling the fluid structure interaction andfor adaptive procedures. these products in new assembly oriented product families the optimization of existing assembly lines and the creation of future reconfigurable is applied for modelling the fluid structure interaction and adaptive procedures. assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and ©functional 2020 The analysis 2019 Authors.isPublished by Elsevier B.V. a© performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the scientific CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the of the 2nd CIRPproduction Conferencesystem on Composite Partsdesigners. Manufacturing. similarity between product families by scientific providingcommittee design support to both, plannersMaterial and product An illustrative Peer-review under responsibility of the Peer-review under responsibility of the scientific committee committee of of the the 2nd 2nd CIRP CIRP Conference Conference on on Composite Composite Material Material Parts Parts Manufacturing. Manufacturing. example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of Keywords: Forming; Fiber reinforced plastic; Temperature; Kinematic; Finite element methode (FEM) thyssenkrupp Presta Fiber France is then plastic; carriedTemperature; out to give aKinematic; first industrial evaluation of the(FEM) proposed approach. Keywords: Forming; reinforced Finite element methode © 2017 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 28th CIRP Design Conference 2018.

1. Introduction 1. Introduction

The research presented in this paper refers to the application The research presented in this paper refers to the application of glass mat reinforced thermoplastic composite (GMT) Keywords: Assembly; Design method; Family identification of glass mat reinforced thermoplastic composite (GMT) Fibre reinforced plastics (FRP) are commonly used to material, which is based on randomly aligned glass fibres and Fibre reinforced plastics (FRP) are commonly used to material, which is based on randomly aligned glass fibres and integrate multiple parts into one complex component and polyamide 6. A multi-axis forming process is investigated in integrate multiple parts into one complex component and polyamide 6. A multi-axis forming process is investigated in reduce part weight by substituting pure metal parts. FRP this research. Besides one axis in common pressing processes, part weight by substituting pure metal parts. FRP thisthe research. Besides one characteristics axis in commonmanufactured pressing processes, 1.reduce Introduction of range and and/or materials base on a polymeric matrix and fibre reinforcement. furtherproduct additional axes are used to realise a tool motion to materials base on a polymeric matrix and fibre reinforcement. further additional axes are used to realise a toolchallenge motion in to assembled in this material system. In thisThe context, the main Commonly applied fibre materials are carbon, glass and natural induce a specific flow. flow pattern can be defined Commonly applied fibre development materials are carbon, glassdomain and natural induce a specific materialisflow. The canwith be defined Dueandtoare the fast of modelling analysis nowbeforehand notflow onlypattern to cope single fibres typically available as short,inlongthe or endless fibres. for exampleand in the part design in order to align the fibres and are typically available as short, long or endless fibres. for example in the part design beforehand in order to align the communication and anof ongoing trend of material digitization products, a limited product range or existing product families, For the design process one specific global and and part glass fibres. For the designmanufacturing process of oneenterprises specific global material and part glass fibres. digitalization, are facing but also tothebealignment able to analyze to compare products define behaviour numerical simulations are conducted inimportant order to With of theand glass fibres through the to pressing behaviour numerical simulations are conducted in order to With the alignment thebeglass fibresthat through the pressing challenges today’s the market environments: continuing new product families. Itofcan observed analyse andinoptimise forming process. To asimulate FRP process anisotropic material behaviour can be classical obtained existing [1]. The analyse and optimise the forming process. To simulate FRP process families anisotropic behaviour canofbeclients obtained [1]. The tendency reduction of product development times and product are material regrouped in function materialstowards the mechanical properties of fibre and matrix material intentionally induced anisotropy can be used ortofeatures. enable materials the mechanical properties of fibre and matrix material intentionally induced anisotropy can be used to shortened product In addition, is an increasing However, assembly oriented product families toenable find. in the process needlifecycles. to be known. The flowthere behaviour is strongly advantageous mechanical properties in are thehardly context of in the process need to be known. The flow behaviour advantageous mechanical properties in the context of demand of customization, the same time inisastrongly global On the product level, products differforming mainly can in two temperature dependent andbeing hence at sets high requirements for the lightweight design.family Furthermore, multi-axis be temperature dependent and hence sets high requirements for the lightweight design. Furthermore, multi-axis forming can be competition competitors all over the This trend, main characteristics: the number ofhighly components and (ii) the process. All with interactions and influences mustworld. be considered for applied in order to(i) manufacture complex hybrid process. All interactions and influences must be considered for applied in order to manufacture highly complex hybrid which is description inducing the development from macro to micro type of components mechanical, electronical). a reliable of the forming behaviour and therefore the components of FRP (e.g. and metal, whichelectrical, seems to be a promising a reliable description of the forming behaviour and therefore the components of FRP and metal, which seems to be a promising markets, results in diminished lot sizes due to augmenting Classical considering production mainly single mechanical part behaviour. approach formethodologies an efficient lightweight [2,products 3]. The mechanical part behaviour. approach for an efficient lightweight production [2, 3]. The product varieties (high-volume to low-volume production) [1]. or solitary, product familiesand analyze the process can already reduce existing misalignment of inserts minimise process can reduce misalignment of inserts and minimise To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which identify potentials causes difficulties regarding an efficient definition and 2212-8271 possible © 2019 The optimization Authors. Published by Elsevier in B.V.the existing 2212-8271 © 2019 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee ofknowledge the 2nd CIRP Conference on Composite Material Parts Manufacturing. production system, it is important to have a precise comparison of different product families. Addressing this Peer-review under responsibility of the scientific committee of the 2nd CIRP Conference on Composite Material Parts Manufacturing. 2212-8271 © 2020 The Authors. Published by Elsevier B.V. This is an©open article Published under theby CC BY-NC-ND 2212-8271 2017access The Authors. Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of scientific the scientific committee theCIRP 2nd Design CIRP Conference on Composite Material Parts Manufacturing. Peer-review under responsibility of the committee of the of 28th Conference 2018. 10.1016/j.procir.2019.09.024

Klaus Dröder et al. / Procedia CIRP 85 (2019) 96–101 Author name / Procedia CIRP 00 (2019) 000–000

2. Material characterisation For composite materials there are several combinations regarding fibre type, orientation, composition and polymer material. Already in the design stage, the composite material is specified with regard to the forming process. With reference to the great number of FRP combinations, the material differs in its properties, e.g. flow behaviour. For a few composite materials rheometric tests can be used to describe the mechanical behaviour. Different scientific work exists for the characterisation of the flow behaviour for GMT material. Squeeze flow testing is the primarily applied parametrisation method [4, 5]. Several approaches for deriving material model parameters from squeeze flow data are presented in [6]. The primary chosen approach in literature for modelling the flow behaviour of GMT is the description as a non-Newtonian fluid. Herein, the flow behaviour is described with help of the viscosity, which itself is described depending on the flow regimes. During tool closure extensional flow as well as shear flow can be observed [7]. Both can be described with help of a power law model [8], which takes into account the influences of shear rate, strain rate and strain hardening effects [5]. In the context of continuums mechanics, a generalised Newtonian fluid with a degenerating viscosity can be described through an idealised plastic material model [6]. This approach makes it possible to derive a plastic material model based on the viscosity data [9].

A

Tensile test machine

A1

Tool system

B Lower tool

Upper tool

h(t)

Spacer

Position of sample

In order to ensure accurate input data for the flow behaviour of the used material, an experimental set up has been created and press rheometry tests have been carried out for the applied GMT material. Fig. 1 shows the experimental set up (A, A1) and the expected flow behaviour in the press gap (B). After heating to a homogenous temperature the specimens are pressed with a velocity of 0.5 mm/s. The spacer limits the minimum gap height to 2 mm. An anti-sticking foil is inserted in order to reduce friction. In Fig. 2 the measured force is plotted over the gap height. In order to exploit the sophisticated plastic material model formulations in LS-DYNA, a flow curve is derived from the experimental data for each test temperature. The measured forces over height and the derived flow curves are plotted in Fig. 2. For the temperature range from 230 °C up to 290 °C the force as well as flow stress curves show identical characteristics regarding the curve shape. The measured forces at a test temperature of 230 °C are higher due to the increasing matrix stiffness, caused by the proximity to the solidification temperature (219°C). Additionally, an increase in force is observed in a gap height range from 2.5 mm to 2.0 mm for all temperatures. This is caused by the compaction of the fibre bed. An averaged flowcurve over the investigated test temperatures is computed (Fig. 2 right, blue curve). 50

230 250 270 290

40 30

C C C C

20 10

0

2

5 4 3 Gap height in mm

15

10

230 C 250 C 270 C 290 C Aver.

5 0 0.0 0.5 1.0 1.5 Effective plastic strain

Fig. 2. Force over gap height (left) and derived flow curves (right)

3. Experimental characterization of the multi-axis forming process

A1

Computer

97

Flowstress in MPa

porosity within the boundary zone of adjacent materials through induced material flow. In the scope of the here presented work, a numerical simulation is carried out in order to predict the material flow. Subsequently, the numerical results are compared to the experiments. Finally, conclusions of the presented work are drawn.

Force in kN

2

Gap Flow height direction

Fig. 1. Experimental press rheometry set up and sketch of material flow (𝜇𝜇coefficient of friction)

For the experiments a forming press with parallel kinematic was used. This kinematic is based on a hydraulic actuated Stewart platform, which can move the upper tool in all six axes. The machine was intentionally built to reduce the mechanical complexity of orbital forming [10]. The machine with a forming tool and forming material are displayed in Fig. 3. The workspace of this machine is designed cylindrical with a height of 50 mm (z-axis) and a diameter of 100 mm (x- and y-axis). In this workspace, the upper tool can be rotated by 5 ° around the vertical axis (C-rotation-axis) and by 15 ° around the other axis (A- and B-rotation-axis). For the forming process, the machine applies a force up to 50 kN in z-axis and 5 kN in xand y-axis. For the experiments, a plain lower die and a curved upper tool were used. With these tools, the forming material is pressed to a plate with a rolling movement of the upper tool. The lower tool is equipped with thermocouples on the surface to measure the surface temperature of the compound in the forming process.

Klaus Dröder et al. / Procedia CIRP 85 (2019) 96–101 Author name / Procedia CIRP 00 (2019) 000–000

98

a)

3

To compare the experiments to the simulation results, the process was observed sideways in an open tool setup. A camera captures the tool movement and the material position and shape. With this information, the expanse of the flow front can be calculated. An image processing algorithm searches for the furthest extend of the forming material in the image. The positon of this flow front is measured relatively to the forming tool in the image. The results from the image processing and the forming material, coloured in blue, are displayed in Fig. 5. The farthest left (1) and right (2) flow front are marked with a dashed line.

b)

c)

z y

Rotatory pressing movement

Fig. 3. a) image of the experimental machine; upper part: six hydraulic actors with glass scales and load cells; b) upper and lower tool; c) position of the thermocouples in the process

200

0.20

0

0.15

-200

0.10

-400

0.05

-600

x y (≈x) z A B C (≈B)

-800

-1000 -1200 -1400

0

5

10

15 Time in s

20

25

Fig. 4. Upper tool motion in the forming process

30

0.00

Orientation in rad

Position in m

The process is subdivided into three phases: Heating of the forming material in an oven, insertion of the material into the press and the pressing and rolling phase. The material was melted in a convection oven for 20 min. at 260 °C and inserted into the press after the matrix was molten. It is placed on one side of the lower tool and pressed to the other side in three forming steps. The tilted upper tool is moved at a velocity of 30 mm/s along the vertical movement (z-axis) to the starting position of the rolling movement. With this movement, the compound is pressed between the resulting nip. In the next forming step, the material is formed with a rolling movement of the upper tool. The upper tool rotates (A-rotation axis) around the tool centre point without a relative movement to the resulting part surface. After rolling, the upper tool moves along z-axis to end the forming process. The motion path of the tool centre point at the tip of the upper tool is displayed in Fig. 4. The tool motion in tests 3 and 5 is defined by two positions (start and end position) which are connected by linear interpolation. With this motion the height of the resulting plate is formed. Tests 1, 2 and 4 have an additional third position in the middle of the forming process with a short resting time in the forming process to obtain a reference point in the material flow. This reference point can be used to synchronize the measurements afterwards as well as to compensated possible deviations in the forming movement, which can occur in a multi-axis forming process without spacers.

x Flow direction 1

2 Lower tool centre

Fig. 5. Forming process of the molten GMT in an open die with an over layer of the image processing colouring the compound in blue; the farthest left and right edge of the compound are marked (1, 2) and the distance to the tool middle is determined

The results of the image processing and the temperatures from the forming experiments are presented in the following section. The results focus on the left flow front in these experiments, which are measured optically. All measurements are synchronised afterwards by setting the first steep rise of the signal as fifth second of process time. After this, the forming process takes another 15 s. Five tests were conducted and altered by the insertion point of the forming material. In addition, the shape of the material varies. The material cannot be kept in a certain shape after heating because of the tacky surface and high porosity. The position and shape of the compound are approximated by using an ellipse. The parameters of this shape can be determined by drawing the smallest ellipse including every apart of the inserted material and the biggest ellipse including only the inserted material. The averages of theses parameters were calculated and displayed for each test in Table 1. In addition, the percentages of the area of the ellipses under the punch were determined. The ellipses are also displayed in Fig. 6. Table 1: Initial position of the compound approximated by an ellipse in mm; x and z are the position of the ellipse middle to the tool middle; a and b are semi-major and semi-minor axis; percentage of the area of the ellipse under the upper tool before pressing

-0.05 -0.10

Test

x

z

a

b

Percentage under the punch

-0.15

1

45

13

60

18

0.58

-0.20

2

38

14

61

17

0.63

3

34

15

58

21

0.65

4

27

25

58

28

0.70

5

10

14

61

19

0.81

Klaus Dröder et al. / Procedia CIRP 85 (2019) 96–101 Author name / Procedia CIRP 00 (2019) 000–000

100

Initial position and shape

60 40 20

110 100 90 70

5

10 15 20 Time in s

60 40 20 0

25

100

100

Temperature in °C

d) 120

80 60 40 20 0

0

5

10 15 20 Time in s

25

0

5

10 15 20 Time in s

25

0

5

10 15 20 Time in s

25

0

5

10 15 20 Time in s

25

80 60 40 20 0

e) 120 100 80 60 40 20 0

80

0

80

c) 120

Temperature in °C

In Fig. 7, the progress of the left flow front in the forming process is displayed. The flow front proceeds rapidly while the upper tool moves along z-axis and slower to the maximum expanse, while the tool rotates over the surface. Test 1, 2 and 4 show a small peak at 12 s. This peak correlates with the extra point in the path description for the upper tool movement. The initial position of the flow front, values under 5 s, correlates to the initial position of the material. Distance from the tool middle in mm

80

0

Fig. 6. Insert position and shape of the compound approximated by an ellipse, dark blue line test 1, dark blue dotted line test 2, blue line test 3, blue dotted line test 4 and light blue line test 5

Temperature in °C

b) 120

100

Temperature in °C

Punch

99

a) 120 Temperature in °C

4

60 Fig. 8. Temperatures in test 1 to 5 at the sensor 1 to 4 are displayed in the diagrams a) to e)

50 40

Test 1 Test 2 Test 3 Test 4 Test 5

Initial position and shape

30 20 10 0

0

5

10

15 Time in s

20

25

30

Fig. 7. Movement of the flow front to the left side measured optically from the centre of the lower tool

The temperatures in the process were measured in the boundary layer for all five tests and displayed for all four sensors (see positions in Fig. 3). In Fig. 8, the temperature for at theses sensors in each test are displayed and test 3 (c) is exemplary described. The other tests show a comparable qualitative progression to each other. The surface temperature of the compound reaches the lowest peak value for the insertion position, here sensor 1 and 2. In the progress, the compound contacts the other sensors. These sensors reach higher peak values than the first two. Sensor 4 is not always in contact with the compound and in test 1, 2 and 4 the peak value is not reached within process time.

With the presented experimental setup and the measurement technique, the forming process and the flow of the material can be observed. The measured temperature in the process indicates that the material flows from the inside of the compound to the outside while the material surface shows no relative movement to the tool. Experimental observations can be compared to the results of a finite element analysis. The results of the flow front progress depend strongly on the flow behaviour of the material, the initial location of the material and the correct execution of the tool movement. 4. Numerical model In order to analyse the flow behaviour a numerical model is created in LS-DYNA and is displayed in Fig. 9. For simulation, only a section of the process has been selected in order to avoid long computation times. The initial shape of the GMT is idealized as a rectangular geometry. Geometric deviations are neglected. The lower and the upper tool are modelled with help of rigid shell structures. The lower tool is fixed in space. In order to realise the particular complex forming motion of the upper tool the centre of gravity is artificially set outwards of the structure itself using the command *PART_INERTIA. Hereby, a rolling motion along a line 3 mm above the lower tool surface can be realised by the upper tool. The motion of the upper tool is depicted in Fig. 10. In the first phase, the upper tool translates downwards. In a

Klaus Dröder et al. / Procedia CIRP 85 (2019) 96–101 Author name / Procedia CIRP 00 (2019) 000–000

100

second phase the rolling motion starts. This rolling motion is accomplished by overlapping a spatial as well as angular displacement boundary condition according to Figure 4. In order to obtain a time efficient model, mass scaling is applied to increase the effective time step size to 0.01 ms. Additionally, a time scaling procedure is performed in which the downwards motion is realised within 0.1 s and the subsequent rolling motion within 1 s. It is assumed that the temperature decrease only takes place in a small boundary layer adjacent to the tool surface and that the temperature of the core has a temperature above the solidification temperature. The main deformation takes place in the core region. Consequently, the temperature decrease in the boundary layer has an insignificant influence on the overall forming behavior. Therefore, a quasi-isothermal approach for the simulation model is applied by using the averaged flow curve as input for the simulation model. Section planes (uy=0)

z

5

5. Results and validation The material flow predicted by the simulation model is shown in Fig. 11. In the upsetting phase, the GMT is squeezed sideways. During the subsequent rolling phase, the material continues to expand sideways until the forming phase is over.

Upsetting phase

Rolling phase

Center of gravity Upper tool

y

z y

z y x

s

Fig. 11. Predicted material flow by numerical model

x Lower tool

GMT

Fig. 9. Simulation model of the forming process

Due to the predominant low forming velocities in the forming process the strain rate dependencies are neglected and only the strain hardening effects are taken into account. The material model *MAT_PIECEWISE_LINEAR _PLASTICITY is applied. In order to account for strong mesh distortion of the GMT, an element-free Galerkin approach (EFG) in combination with remeshing has been selected for discretisation.

In order to compare the numerical results to the experimental results, the distance s which is measured from the tool middle to the outer right corner of the sample is plotted over time in Fig. 12. During the upsetting phase, the material is quickly squeezed to the sides, which results in a steep rise in the plot. Followed by a less steep increase of the distance s caused by the continuous material flow during rolling phase. In the end of the process, the flow velocity decreases until the forming phase has completely terminated. This flow pattern was also observed in the experimental tests (see for instance Test 4). The quantitative deviations are in the range of the experimental results.

Fig. 12. Comparison of the experimentally and numerically predicted material flow Fig. 10. Kinematic of the upper tool

6

Klaus Dröder et al. / Procedia CIRP 85 (2019) 96–101 Author name / Procedia CIRP 00 (2019) 000–000

6. Summary and conclusion At first, a material characterisation of the used GMT has been carried out in order to describe the material behaviour. A set of experiments was conducted with a pressing and rolling movement. A six degree of freedom kinematics was used with an open die to observe the material flow in the process. The material flow depends on the insertion position and the initial shape of the GMT. Based on the material data a numerical model was set up in LS-DYNA, incorporating the particular difficult tool motion, consisting of an upsetting and a rolling phase. The predicted flow pattern shows good accordance with the experimental results for identical insertion positions. Minor deviations can be explained by the shape and amount of the material under the punch in the experiments. Acknowledgment This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – 318620418. References [1] Radtke, A.: Steifigkeitsberechnung von diskontinuierlich faserverstärkten Thermoplasten auf der Basis von Faserorientierungs- und Faserlängenverteilungen (Fraunhofer IRB-Verl., Stuttgart, 2009) [2] Behrens, B.-A., Hübner, S., Grbic, N., Micke-Camuz, M., Wehrhane, T., Neumann, A.: Forming and Joining of Carbon-Fiber-Reinforced Thermoplastics and Sheet Metal in One Step, 17th International Conference on Sheet Metal, SHEMET 2017, University of Palermo, Palermo, Italy; 10 April 2017 through 12 April 2017. [3] Behrens, B.-A., Raatz, A., Hübner, S.,Bonk, C.,Bohne, F., Bruns, C., Micke-Camuz, M.: Automated Stamp Forming of Continuous Fiber Reinforced Thermoplastics for Complex Shell Geometries, 1st CIRP Conference on Composite Materials Parts Manufacturing, CIRP CCMPM 2017, Karlsruhe, Germany, 8 june through 9 june 2017. [4] Kau, H.-T., Hagerman, E. M.: Experimental and analytical Procedures for Flow Dynamics of Sheet Molding Compound (SMC) in Compression Molding, Polymer Composites, Vol. 8, No. 3, 1987. [5] Collyer, A. A., Clegg D.W.: Rheological Measurement, Springer Science+Business Media Dordrecht, 1998 [6] Engmann, J., Servais, C., Burbidge, A. S.: Squeeze flow theory and applications to rheometry: A review, Journal of Non-Newtonian Fluid Mechanics, 132, 2005, pp. 1-27 [7] Vahlund, C. F., Gebart, B. R.: Thermoplastic (GMT) in Large Tools and at High Closing Velocities, Intern. Polymer Processing XVII, 2002 [8] Kotsikos, G., Bland, J. H., Gibson, A. G: Rheological Characterization of Commercial Glass Mat Thermoplastics (GMTs) by Squeeze Flow Testing, Polymer Composites, Vol 20, 1999 [9] Behrens, B.-A.; Moritz, J.; Chugreev, A.; Bohne, F.; Schulze, H.: Finite Element Analysis for sheet metal reinforced hybrid structures produced via non-kinematical constraint manufacturing processes , 18th European Conference on Composite Materials, 24th - 28th june 2018, Athens, Greece [10] Hesselbach, J., Behrens, B.-A., Dietrich, F., Rathmann, S., Poelmeyer, J.: Flexible forming with hexapods, Prod. Eng. Res. Devel., 2007, pp. 429-436

101