Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor In Situ Monitoring , Measurement and control of Direct Digital Additive Manufacturing Jyoti Mazumder* University of Michigan January 9th, 2013 *Robert H Lurie Professor of Engineering @ University of Michigan
34
Embed
In Situ Monitoring , Measurement and control of Direct Digital Additive Manufacturing
In Situ Monitoring , Measurement and control of Direct Digital Additive Manufacturing. Jyoti Mazumder * University of Michigan January 9th, 2013. * Robert H Lurie Professor of Engineering @ University of Michigan. Outline. Background History of DMD Introduction DMD System Overview - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
In Situ Monitoring , Measurement and control of Direct Digital
Additive Manufacturing
Jyoti Mazumder*
University of Michigan
January 9th, 2013
*Robert H Lurie Professor of Engineering @ University of Michigan
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Outline
• Background History of DMD • Introduction
– DMD System Overview• Advances in DMD System
– Geometry Control– Temperature Control– Composition Prediction– Microstructure Prediction– Modeling
• Summary
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Vision: Part on Order Anywhere
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Running to Moon: Mold & Mirrors
0.5 mm wall thickness in
steel
0.5 mm wall thickness in
steel
Polished to 40 Angstroms!
Polished to 40 Angstroms!
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
• Simulation:• Weight on control: 100000000• Prediction horizon: 30• Control horizon: 5• Tfilter = [1 -0.8]
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
(a) (b)
(c) (d)
Pictures of the deposition at (a) 10th layer, (b) 20th layer, (c) 30th layer and (d) 40th layer Cladding height at different layers
Molten Pool Temperature Control
0 10 20 30 400
2
4
6
8
10
Cladding layer number
Cla
dd
ing
he
igh
t (m
m)
With control, aWith control, bNo control, aNo control, b
baSubstrate
Cladding
y
zA
x3mm step
One Inch Cube Cladding with Temperature Control
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Composition Prediction
Substrate
Collecting lens bead
Laser beam
Signal processing
unit
spectrometer
Hopper1
Hopper2
Alloys without Phase Transformation
Cr-Fe
Ni-Fe
Alloys With Phase Transformation
Ti-Fe
Ni-Al
Ni-Ti
Experimental Setup
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Composition Prediction: Cr-Fe Alloy
• Calibration Curve
0 10 20 30 400.4
0.6
0.8
1
Cr-
I 42
8.97
2nm
/Fe-
I 43
0.79
01nm
0 10 20 30 400.5
0.6
0.7
0.8
0.9
1
Cr-
I 42
8.97
2nm
/Fe-
I 43
2.57
61nm
0 10 20 30 400.2
0.3
0.4
0.5
0.6
0.7
Cr-
I 43
4.45
1nm
/Fe-
I 43
0.79
01nm
Cr weight percentage0 10 20 30 40
0.2
0.3
0.4
0.5
0.6
0.7
Cr-
I 43
4.45
1nm
/Fe-
I 43
2.57
61nm
Cr weight percentage5 10 15 20 25 30 35 40
3700
3750
3800
3850
3900
3950
4000
4050
plas
am t
empe
ratu
re (
K)
Cr weight percentage
Line Intensity Plasma Temperature
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Prediction of Cr% in the Alloy
• Composition Variation < 5%
5 10 15 20 25 30 35 40-40
-20
0
20
40
60
80
100
Cr weight ratio percentage
com
positio
n v
ariation (
%)
from single line ratiofrom temperature
from electron density
from four averaged line ratio
from seven averaged line ratiofrom seven averaged line ratio and electron density
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Ni-Al Alloy Phase Transformation and Line Intensity Ratio (Patent Pending)
60 70 80 900
20
40
60
80
Al-I
394
.4nm
/Ni-I
349
.296
nm
Atomic percent Ni60 70 80 90
0
10
20
30
40
Al-I
394
.4nm
/Ni-I
352
.454
nm
Atomic percent Ni
60 70 80 900
20
40
60
80
100
Al-I
396
.15n
m/N
i-I 3
49.2
96nm
Atomic percent Ni60 70 80 90
0
10
20
30
40
50
Al-I
396
.15n
m/N
i-I 3
52.4
54nm
Atomic percent Ni
10um
? Ti67.5Fe25.5
B2 Ni65Al35
Gamma Prime Ni65Al35
B2 Ni65Al35
Gamma Prime Ni80Al20
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
XRD Pattern of Ni80Al20 Sample as Deposited
(111)
(200)
(220) (311)
(100)
20 30 40 50 60 70 80 90 100 110 120
Two-Theta (deg)
0
2500
5000
7500
Inte
nsi
ty(C
ou
nts
)
[Z02639.raw] NI80AL20
03-065-3245> AlNi 3 - Aluminum Nickel
(110)
(210)(211) (300)
(222) (400)(321)(320)
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Mathematical Modeling
• Process modeling of DMD to develop quantitative relationships between parameters for improved process control
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
0
ut
x
pu
Kuu
t
u
l
l
u
Continuity equation:
Momentum equation:
Energy equation: t
TCf
t
fTkTC
t
TC psspl
p
Lu
Solute equation: uu slsl ccfccDcDct
c
)(
Convection term Diffusion term Darcy term
Convection term Conduction term Phase change term at S/L interface
Phase diffusion term Phase motion term
Modeling: Governing Equation
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Heat transfer / Fluid flowFlow chart & Governing Eqns.
0
l
l
l
l
l
l
l
l s l s
hh k T h h
tu p
u u ut K x
u pv v v
t K y
u pw w w g T T
t K z
cc D c D c c f c c
t
u u
u
u
u
u u
• Governing Equations [1, 2]
Flux due to relative phase motion
Darcy term
Buoyancy term Nov, 2010
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Solute Transport: Advection dominant
Pectlet numberC: 3.15x104
Ni:1.69x105
Inside melting pool, Advective transport >> Diffusion transport
Pe Re ScLv
D
Ni concentration C concentration
Nov, 2010
Nominal composition in 4340 steel:1.75% Ni 0.4%
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
X
Z
YThe computation domain is not symmetric along laser moving direction
Start
Direct
ion o
f tra
vel
in th
e fir
st p
ass
Direct
ion o
f tra
vel
in th
e se
cond p
ass
FinishOverlap
TransitionSca
nning
leng
th
Beam size
Scanning width
Multiple Track Deposition Model
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Composition and Liquid Velocity Distribution
X (mm)
01
23Y
(mm
)
0.5
1
Z(m
m)
-0.3
-0.2
-0.1
0
Y
X
Z
COMP: 1 2 3 4 5
1 m/s
X (mm)
0
1
2
3
Y(m
m)
0
0.5
1
Z(m
m)
-0.2
0
0.2
0.4
0.6
COMP: 1 2 3 4 5
1 m/s
Computed chromium concentration profile:
x-z surface and x-y surface
y-z surface
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Thermo-physicalMaterial
Properties
Initialize / Heat Source
Heat, Mass And MomentumCalculation
Is it RapidSolidification?
Non-EquilibriumPartition Coefficient
Calculate the Composition and Phase
No
Is PhaseDetection from
Phase TransformationSensor Same asCalculated One?
No
Product Stop
Yes
Flow Chart
Yes
Change ProcessParameter From
Calculated Co-relations
1
2
3
4
5
6.
7.
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor
Summary and Conclusion
• Process Model – Simulate melt pool temperature, velocity, fluid interface
thermal cycle, and composition evolution and distribution• Process Sensor and Control Design, Optimization and
Implementation– Geometry Control– Melt pool temperature dynamics and control– Composition sensor– Microstructure sensor
• First time in the world one will have the capability to predict the microstructure during the process from plasma, leading to considerable cost and lead time saving
Center for Laser Aided Intelligent Manufacturing University of Michigan, Ann Arbor