Post on 06-Feb-2017
Polymer Engineering CenterUniversity of Wisconsin-Madison
Experimental and Numerical Studies of the Temperature Field in Selective Laser Sintering to Improve Shrinkage and Warpage Prediction
Prof. Dr.-Ing. Natalie RudolphPolymer Engineering Center
Department of Mechanical EngineeringUniversity of Wisconsin-Madison
1513 University AveMadison, WI 53706
Advanced Qualification of Additive Manufacturing Materials Workshop,July 20-21, 2015 in Santa Fe, NM
Polymer Engineering CenterUniversity of Wisconsin-Madison
Overview
Additive Manufacturing: SLS, FDM®, AFP
• Process characteristics
• Shrinkage and warpage effects
Investigations of impact factors:
• Study of crystallization
• Study of stiffness development during cooling
• Measurement of temperature field e.g. during coating
Outlook: Online-monitoring of crystallization
3D CAD model à STL format
1
2D Slicing
2
3D finished part
4
Additive building process, e.g. SLS
3
Polymer Engineering CenterUniversity of Wisconsin-Madison
Powder bed fusion: SLS Material extrusion: FDM Directed energy depos.: AFP
v v v
F
an AM process in which focused thermal energy is used to fuse materials by melting as they are being deposited: Automated Fiber Placement
an AM process in which material is selectively dispensed through a nozzle or orifice: Fused Deposition Modeling
an AM process in which thermal energy selectively fuses regions of a powder bed: Selective Laser Sintering
ASTM F2792, 2014
Additive Building Principles
Polymer Engineering CenterUniversity of Wisconsin-Madison
Powder bed fusion: SLS Material extrusion: FDM Directed energy depos.: AFP
v v v
F
q Local melting of continuous fiber reinforced polymer during deposition
q (mostly) laser heat sourceq Material itself keeps part
shapeq mostly 2D, curved shapes
q Heat conduction in the nozzle
q Deposition of molten material and local remelting
q Support material neededq 3D shapes
q Local melting of deposited powder
q Laser heat sourceq Surrounding solid powder
creates “mold” q Complex 3D shapes
Differences
Polymer Engineering CenterUniversity of Wisconsin-Madison 5
Prof. Lomov, KU Leuven
Challenges:- Always two layers with perpendicular orientation- Fiber ondulations reduce strength- Scrap of 50-150% depending on design- Draping over complex geometries with defined orientation
Fiber-Reinforced (Textile) Composites
Polymer Engineering CenterUniversity of Wisconsin-Madison 6
IVW Kaiserslautern
LCC TU Munich
Draping of Textiles
Polymer Engineering CenterUniversity of Wisconsin-Madison 7
Leibniz-‐Ins0tut Dresden, 2014
Tailored Fiber Placement
Polymer Engineering CenterUniversity of Wisconsin-Madison
Roller Cu?er Tape
Heat source
Mold
Tape feed F
Automated Fiber Placement
Polymer Engineering CenterUniversity of Wisconsin-Madison 9
0º [0]5 90º [0]10
σM (MPa)
εM (%) E (GPa) μ (-‐)
Mean value 1459.05 1.39 97.75 0.36
Standard dev. 48.68 0.02 2.33 0.01
σM (MPa)
εM (%) E (GPa)
Mean value 31.28 0.57 5.72
Standard dev. 1.52 0.03 0.04
ISO 527-‐5, 2010
Tensile tests 0º - 90º (PA6 CF)
Polymer Engineering CenterUniversity of Wisconsin-Madison
x
y
Top viewx
z
Front view
Laser/deposition path
x
z y
x
z y
T
T
10
Temperature Distribution during Building
The scale/temperature range depends on the process!
Polymer Engineering CenterUniversity of Wisconsin-Madison
q Shrinkage is the difference between the part dimensions in the molten and the solid state due to the volume contraction during cooling
q Residual stresses are formed during cooling due to rapid quenching and shrinkage inhibition
q Warpage is the change of the part shape (e.g. spring-in at corners) due to non-symmetric residual stress distributions. It is caused by:Ø Inhomogeneous shrinkage over the part cross-section (e.g. due to
differences in temperature on the part surface)Ø Local shrinkage differences within the part (e.g. due to varying wall
thicknesses)Ø Anisotropy of shrinkage (e.g. due to the orientation of molecules or fibers)
Residual stress model without phase change effects (derived from dimensional analysis)
11
Shrinkage and Warpage
σ = − 2
3Eβ1−υ
(Ts −Tf ) ⋅! ! =σ= residual stress, E= Young’s modulus, b= thermal expansion coefficent, ν= Poisson’s ratio, Ts, Tf= solidification and final temperature, geometric factor
Osswald/Menges, 2012
Polymer Engineering CenterUniversity of Wisconsin-Madison
Hea
t flo
w
Temperature T
DSC-Results for PA 12 (PA 2200® powder)
Heating 10K/min
Cooling 10K/min
Thermal Process Window for SLSIsothermal crystallization
Hea
t flow
Temperature T
DSC-Results for PA 12 (PA 2200® powder)
Heating 10K/min
Cooling 10K/min
Crystallization
Based on: DIN EN ISO 11357-1
Part bed temperature
Start meltingat Ti,m
Endo
Exo
Start crystallization at Ti,c
W/g
°C
Melting
Heating 10K/minCooling 10K/minCooling 5K/min
Schematic DSC-results for PA 12
Polymer Engineering CenterUniversity of Wisconsin-Madison
5 mm
Warpage of Layers
Polymer Engineering CenterUniversity of Wisconsin-Madison
CurlingDisplacement and defect
Warpage in the first section of
a part
Displacement of the part position
Polymer Engineering CenterUniversity of Wisconsin-Madison
Warpage in FDMCooling effects in massive parts
Print parameters: 230ºC nozzle temperature, 120ºC bed temperature, 40mm/s travel rate, 50% infill, 2 shells
Pfeifer, 2015
4”
Polymer Engineering CenterUniversity of Wisconsin-Madison
Layup on cold mold Layup on heated mold
16
F tape
mold
laser F
Heated mold
Warpage in AFP Heated vs. cold mold
Zenker, 2014
Polymer Engineering CenterUniversity of Wisconsin-Madison
Polyamide 12 (PA 650, ALM)
Isothermal crystallizationDSC
Polymer Engineering CenterUniversity of Wisconsin-Madison
Isothermal crystallizationAvrami model
Polymer Engineering CenterUniversity of Wisconsin-Madison
• Cone-plate rheometer• Modulus development at 0.1 Hz• G’ and G’’
Elastic ModulusRheology: Measurement in oscillation
Polymer Engineering CenterUniversity of Wisconsin-Madison
Prediction of residual stressesCorrelation of modulus and crystallization
Polymer Engineering CenterUniversity of Wisconsin-Madison
• Prediction of shrinkage and wa rpage us ing advanced computing simulations
• Application to thermal processes
Simulation
Polymer Engineering CenterUniversity of Wisconsin-Madison
Shrinkage and warpageFirst results with Matlab
BUT…what does the temperature field look like?
Aquite, 2014
Polymer Engineering CenterUniversity of Wisconsin-Madison
Example: temperature field during coatingMajor Components of a SLS System (DTM 2500)
80°C80°C 166°C
Part(melt)
Part bed(powder)
Part bed heater
Overflow
Feederheater
Coating powder
Powderreservoir
Roller
CO2- Laser
xy
z
Based on: VDI - Guideline 3404 [1]
Polymer Engineering CenterUniversity of Wisconsin-Madison
Data acquisition equipment:
• NI-9211/NI-9214®
• 11 K-type thermocouples
• LabVIEW 2013®
Process settings:
• PA 12 powder (PA 650® )
• Tpart bed = 166 °C
• Tfeeder = 80 °C
• vRoller = 75 / 125 / 175 mm/s
Data acquisitionTemperature measurements
Sensor P1
Sensors atposition P1 – P11
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Data acquisitionRoller temperature for different roller speeds
P8
P1
P11
Rol
ler
°C
Tfeeder
x
yz
P8
P1
P11
Rol
ler
Part
bed
20
40
60
80
100
120
0 50 100 150 200 250 300
Tem
pera
ture
Number of layers
75 mm/s - P1
75 mm/s - P8
75 mm/s - P11
175 mm/s - P1
175 mm/s - P8
175 mm/s - P11
°C
Feed
er
Steady state
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Data acquisitionTemperature distribution along the roller
[°C]Number of layers
20
40
60
80
100
120
-200 -100 0 100 200 300
Tem
pera
ture
Position along the roller in x-direction
1000 s / 0
2000 s / 16
3000 s / 36
4000 s / 56
5000 s / 76
6000 s / 96
7000 s / 116
8000 s / 136
9000 s / 158
mm
°C
ΔT
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Subdivision inGeometrisches modellMaterialmodellThermisches ModellMotion modellMesh
Modeling and SimulationTask Preprocessing – Model setup
Material modele.g. definition of• density• spec. heat capacity• thermal conductivity
Geometrical modele.g. dimensions of• roller• part bed• powder wave
Based on: Experimental data / Literature data & assumptions / machine data
Kinematic modele.g. definition of• moving domains• points in time for
boundary conditions
Thermal modele.g. definition of initial and boundary conditions• temperatures • heat fluxes
Thermal simulation
Polymer Engineering CenterUniversity of Wisconsin-Madison
Modeling and SimulationKinematic model
Temperature:Tpowder (0<t<tend) = 80°C
Temperature:Tpart bed (0<t<t1) = 80°CTpart bed (t1<t<tend) = 166°C vroller
166°C
tend
80°C166°C
t1
80°C166°C
t0
250 mm
200 mm
xy
z
Velocity field for all domains underneath the roller and the coating powder:Translation movement in negative y-direction with vroller
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Rroller
Modeling and SimulationGeometrical and thermal model
Temperature for t0 < t < tend:TBC1 = Tfeeder = 80°C
Coating powder
New coated layer
Previous layer (melt)
Part bed (powder)
Heat fluxTRoller= 60 °Ch = 1000 W/(m2 K)
Temperature for t0 < t < t1:TBC2 = Tfeeder = 80°C
Temperature for t1 < t < tend:TBC2 = Tpart bed = 166°C
For all domainsInitial temperature for t0:TIC1 = Tfeeder = 80°C
BC: Boundary condition IC: Initial condition
xy
z
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Evaluation Temperature distribution along the part bed
80°C166°C
t1200 mm
166°C
For tend
Form t1 to tend
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Evaluation Temperature distribution along the part bed
166°C
Coating powder
Previous layerNew layer
Part bed
For tend
xy
z
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
50
70
90
110
130
150
0 5 10 15 20 25 30
Tem
pera
ture
Position in X-direction
[°C]
Evaluation Temperature distribution in the new layer
z = 0 mm
z = -0.05 mm
z = -0.1 mmz = 0.1 mm
z = 0.05 mm
Delay due to low thermal conductivity
mm
°C
50
70
90
110
130
150
170
0 5 10 15 20 25 30 35
Tem
pera
ture
Position in X-direction
Temperature distribution along the part bed
z= 0.0 mm - v = 75 mm/s z= 0.1 - v = 75 mm/s
z= -0.1 - v = 75 mm/s z= -0.05 mm - v = 75 mm/s
z= -0.05 mm - v = 75 mm/s
[°C]
[mm]
z = 0.1 mm z = 0.05 mm
Time delay due to low thermal conductivity
Direct contact with roller surface
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
140
145
150
155
160
165
170
0.0 0.1 0.2 0.3 0.4 0.5
Tem
pera
ture
Time
60 °C - z= -0.05 mm
80 °C - z= -0.05 mm
100 °C - z= -0.05 mm
110 °C - z= -0.05 mm
Evaluation Influence of the roller temperature
°C
s
Risk of premature crystallization for Troller<100 °C
For vRoller = 75 mm/s
Temperature Ticfor start of crystallization(depends on cooling rate)
Part bed
New layerPrevious layerat z = -0.05 mm
Troller = 110 °C
Troller = 60 °C
Troller = 100 °CTroller = 80 °C
Temperature Tfc for recrystallization
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
140
145
150
155
160
165
170
0.0 0.1 0.2 0.3 0.4 0.5
Tem
pera
ture
Time
z= -0.05 mm - v = 75 mm/s
z= -0.05 mm - v = 125 mm/s
z= -0.05 mm - v = 175 mm/s
Evaluation Temperature vs. time in the scanned layer
Temperature Ticfor start of crystallization(depends on cooling rate)
°C
s
For TRoller = 60 °C
Part bed
New layerPrevious layerat z = -0.05 mm
Temperature Tfc for recrystallization
Higher risk of premature crystallization for vroller = 75 mm/s
Vroller = 175 mm/s
Vroller = 75 mm/s
Vroller = 125 mm/s
Lassmann, 2015
Polymer Engineering CenterUniversity of Wisconsin-Madison
Online-Monitoring of Phase Transitionsin Thermoplastics
Dielectric Analysis on PA6 and PPS
Polymer Engineering CenterUniversity of Wisconsin-Madison
DSC
Tg, ΔH Rheology η*
mechanical testing
DMA
Tg, E‘, E‘‘, tan δ
DEA ε*, σ*, tan δ, (Tg)
thermal analysis (mesoscopic)
mechanical testing (macroscopic)
electrical analysis under temperature (microscopic)
online measurem
ent laboratory m
easurement
Relevance of DEA
Polymer Engineering CenterUniversity of Wisconsin-Madison
dipole orientation
ion conductivity
+ - + -
+ -
+ -
+ - + - + -
+ - + -
TheoryParallel Plate Capacitor
C* = ε0ε r* Ad
C* = ε0ε 'Ad− iε0ε "
Ad
εr* = ε '− iε "
E!"
C: capacitanceϵ0: permittivity of free space ϵr: relative permittivity of the dielectric sampleA: area of the capacitor platesd: distance between the platesϵ′: relative permittivity ϵ′′: loss factor
Polymer Engineering CenterUniversity of Wisconsin-Madison
Tested material
electric field electrodes
sensor substrate
interdigitated electrodes (one-sided measurements)
parallel plate capacitor
DEA sensor
disposable sensor reusable tool mountable sensor
sensor design:
Theory in ApplicationDEA
Polymer Engineering CenterUniversity of Wisconsin-Madison
Material tested
electric field electrodes
sensor substrate
interdigitated electrodes (one-sided measurements)
CF
electrical conductive
short circuit
insulating / covering layer between electrodes and material
Theory in ApplicationModifications for conductive material, e.g. CF
Polymer Engineering CenterUniversity of Wisconsin-Madison
100 200-2,0
-1,8
-1,6
-1,4
-1,2
-1,0
-0,8
-0,6
-0,4
-0,2
0,0he
at fl
ow [m
W/m
g]
temperature [°C]
DSC PA6 neat first cooling DSC PA6-GF60 f irst cooling DSC PA6-CF60 f irst cooling
2
4
6
8
10
12
14
16
18
20
Dielectric PA6 neat first cooling Dielectric PA6-GF60 first cooling Dielectric PA6-CF60 first cooling
ε' [-]
Correlation of DSC and DEAPA6, Ticona PA6-GF & Ticona PA6-CF
Polymer Engineering CenterUniversity of Wisconsin-Madison
50 100 150 200 250-1,2
-1,0
-0,8
-0,6
-0,4
-0,2
0,0
0,2
heat
flow
[mW
/mg]
temperature [°C]
cooling DSC
5
10
15
20
25
30
cooling dielectric
ε' [-]
50 100 150 200 250-1,2
-1,0
-0,8
-0,6
-0,4
-0,2
0,0
0,2
heat
flow
[mW
/mg]
temperature [°C]
cooling DSC
0
1
cooling dielectric
dε'/d
T [1
/K]
Correlation of DSC and DEACrystallization of Ticona PA6-CF60
Chaloupka [5]
Polymer Engineering CenterUniversity of Wisconsin-Madison
160 180 200
0,0
0,5
1,0fra
ction
al c
ryst
alliz
ation
[-]
temperature [°C]
DSC dielectric
degree of crystallization is correlatable to DSC-results
Correlation of DSC and DEADegree of Crystallization of Ticona PA6-CF60
Polymer Engineering CenterUniversity of Wisconsin-Madison
50 100 150 200 250 300
rel.
perm
. [-]
temperature [°C]
cooling DSC cooling dielectr ic
-0,4
-0,2
0,0
0,2
DSC
[mW
/mg]
tool mountable (reusable) sensor
50 100 150 200 250 300
first heating DSC first heating dielectr ic
rel.
perm
. [-]
temperature [°C]
0,0
0,1
0,2
0,3
DSC
[mW
/mg]
first
hea
ting
cool
ing
Glass transition
Onset [°C] End [°C]
DSC 77,5 88,6
Dielectric 85,5 106,4
Melting Tm
Onset [°C] End [°C]
DSC 269,5 290,1
Dielectric 276,2 284,2
Crystallization Tc
Onset [°C] End [°C]
DSC 239,4 249,4
Dielectric 246,8 260,8
Glass transition Tg, Tm and TcTicona PPS-CF67
Polymer Engineering CenterUniversity of Wisconsin-Madison
tool mountable (reusable) sensor
first
hea
ting
Melting
Onset [°C] End [°C]
DSC 269,5 290,1
Dielectric 275,1 284,8
Cold crystallization
Onset [°C] End [°C]
DSC 110,1 123,0
Dielectric 109,2 151,5
50 100 150 200 250 3001E-4
1E-3
0,01
0,1
1
10
loss
fact
or [-
]
temperature [°C]
first heating die lectric
0,01
0,1
first heating DSC
DSC
[mW
/mg]
Cold CrystallizationTicona PPS-CF67
Polymer Engineering CenterUniversity of Wisconsin-Madison
• Local melting and remelting of polymer material leads to more complex shrinkage and warpage cases
• Development of Young’s modulus as a function of temperature and solidification is significant for part warpage
• Numerical simulation of warpage successful
• Error is decreased by taking inhomogeneity of the temperature field into account
• All local deposition and melting processes are governed by these same phenomena.
• DEA is suitable for online- measurements of phase transitions in composite manufacturing
Conclusions Summary
Polymer Engineering CenterUniversity of Wisconsin-Madison
Prof. Tim OsswaldWilliam Aquite
Wolfgang LassmannThomas Pfeifer
Thomas Zenker (Fraunhofer ICT)and all PEC graduate students
Acknowledgements
Polymer Engineering CenterUniversity of Wisconsin-Madison
Thank you for your attention!
Contact
Prof. Dr.-Ing. Natalie RudolphUniversity of Wisconsin-MadisonPolymer Engineering Center
1513 University Ave1035 Mechanical Engineering building
natalie.rudolph@wisc.edu
Dielectric AnalysisDipl.-Phys. Alexander ChaloupkaFraunhofer ICTBranch Functional Lightweight Design FIL
Am Technologiezentrum 286159 Augsburg
alexander.chaloupka@ict.fraunhofer.de
Polymer Engineering CenterUniversity of Wisconsin-Madison
[1] VDI - Guideline 3404. Additive manufacturing – Basics, definitions, processes. Verein Deutscher Ingenieure, Berlin 2014
[2] D. Rietzel: Werkstoffverhalten und Prozessanalyse beim Laser-Sintern von Thermoplasten. Dissertation, University Erlangen-Nuremberg, 2011.
[3] DIN EN ISO 11357-1 – Kunststoffe – Dynamische Differenz-Thermoanalyse (DSC) – Teil 1: Allgemeine Grundlagen (ISO 11357-1:2009); Deutsche Fassung EN ISO 11357-1:2009, in B.V. GmbH (Ed.)
[4] T. Osswald, G. Menges: Materials Science of Polymers for Engineers, 3rd ed., Hanser Publishers, Cincinatti, 2012
[5] A. Chaloupka, A. Wedel, I. Taha, N. Rudolph, K. Drechsler: Phase change detection in neat and fiber reinforced polyamide 6 using dielectric analysis, Materials Science Forum Vols 825-826 (2015) pp 944-951, Trans Tech Publications
References