Predicted microstructure in repair technology of Ti-6Al-4V...Qualitative microstructure prediction...
Transcript of Predicted microstructure in repair technology of Ti-6Al-4V...Qualitative microstructure prediction...
1
Predicted microstructure in repair technology of Ti-6Al-4V
TRAN Hoang-Son
TCHOUFANGchuindjang
PAYDAS Harkan
JARDIN Ruben
Raoul Carrus
Mertens Anne
HABRAKEN Anne Marie
28/03/2018 EMMC16 Nantes
University of Liège
+ Elena Eseva(master thesis)
2
²
2
Microstructure in Ti6Al4V parts built or repaired by Additive Manufacturing
What are still the scientific challenges?
Understanding all the out of balance phenomena due tohigh thermal and cyclic gradients
Developing accurate models to predict microstructures
This presentation:• Two different experiments enhancing different microstructures• A validated thermal model able to give reliable temperature
history qualitative prediction and analysis to progress toward quantitative prediction
3
Laser cladding, DMD, LMD, LENS, DED, LAM...
²
3
Bhattacharya & al. (2011)
Innovative
technology
Production of
dense parts
Multilayer metal
deposit
Very high cooling
rates
(ultrafine grain
microstructure) Sirris
Need of a thermal model:Study of processing parameters:
laser power
powder flow
preheating temperature (T°)
laser beam velocity
44
Temps [s]
Continuous Cooling diagram– Ahmed et al.1998
Tem
per
atu
re[°
C]
Microstructure
evolution
²Material: Ti-6Al-4V
α'
αm
α
Time [s]
5
Phenomenological Models of Solid-phase Transformations
²
Crespo, A.,2010 Modelling of Heat Transfer and Phase Transformations in the Rapid Manufacturing of
Titanium Components
6
Yes
Yes
Yes
Yes
Yes
Yes
Flowchart for microstructure model implementation
Phenomenological Models of Solid-phase Transformations
²
Diffusional transformation (α’→α+ β, β↔ α): temperature and time dependent
• Johnson-Mehl-Avarami-Kolmogorov Model + additivity rule 𝑓𝛼′(𝑇 + 𝑑𝑇)= 1 − [1-exp[−𝑘 𝑡𝑓 + ∆𝑡 𝑛]](1 − 𝑓𝛼′𝑒𝑞 )
where 𝑡𝑓 is fictious time which would resulted in fraction 𝑓𝛼′ 𝑇 at 𝑇 + 𝑑𝑇
Martensitic transformation (β→α’): temperature dependent only
• Koisitinen – Marburger Model : 𝑓𝛼′(𝑇)= 𝑓
𝛼′ (𝑇0)+(𝑓𝛽 (𝑇0) - 𝑓𝛽𝑟)[1-exp[−𝛾(𝑀𝑠 − 𝑇)]]
Master thesis2018Elena Esteva
7
Phenomenological Models of Solid-phase Transformations
²Martensitic transformation (β→α’): temperature dependent only
• Koisitinen – Marburger Model : 𝑓𝛼′(𝑇)= 𝑓
𝛼′ (𝑇0)+(𝑓𝛽 (𝑇0) - 𝑓𝛽𝑟)[1-exp[−𝛾(𝑀𝑠 − 𝑇)]]
Ms is a great variety in the literature
• (Tan et al. 2015) consider an Ms =1073 K
• (Kelly SM 2004) Use lower values such as Ms = 848 K
• (Charles Murgau C et al. 2012) Ms = 898 K
• (Crespo A, et al. 2010) Ms = 923 K
Ms point depend on the chemical composition of titanium
alloys and cooling rate
The completion of the martensitic reaction is almost achieved when the
temperature reaches the end point of the transformation temperature (Mf).
• Very few values are mentioned in the literature , very disparate
• (Crespo A, et al. 2010) Mf = 673 K
• (Jovanovic et al. 2006) Mf = 298 K
Ms and Mf → need further investigation
8
Experiments
²
8
Laser cladding as a repair technology for Ti6Al4V alloy: influence of incident energy and
building strategy on microstructure and hardness.
Paydas,et al. Materials and Design 2015.
«MacroClad» et «Decrease Track Length (DTL)»
strategy«MacroClad» et «Constant Track Length
(CTL)» strategy
99
Microstructure of Constant Track Length (CTL)
Experimental data
Final microstructure in CTL
979μm
522μm
Paydas,et al. Materials and Design 2015
+ Tran et al. Materials & Design 128 2017
1010
Results of indentation campaign
Z
X
(b)
1mm
A B C D E F G H I J K L M N O P Q R S T U
Z
X
A B C D E F G H I J K L M N O P Q R S T U HV10
Z
X
(e)A B C D E F G H I J K L M N O P Q R S T U V W
1mm
A B C D E F G H I J K L M N O P Q R S T U V W HV10
(a) (c)
(d)
(f)
Hardness maps and corresponding hardness indentation grids (Paydas et al.2015)
Decrease Track Length (DTL)
Constant Track Length (CTL)
1111
Applied flux
Applied flux
Active
elementNewly active
element
Inactive element
Convection and
radiation element
Laser used by Sirris
Power density
concentration
1.4 mm
1.6 mm
Top-hat profile of the laser beam
energy
Birth element technique
Numerical model
²
At each time step → Updated boundary conditions
1212
Model Validation
²
Geometry of the machined substrate Path of laser beam (7 tracks/layer)
Odd layer Even layer
100 200 300 400 500 600200
300
400
500
600
700
800
900
1000
1100
Time (s)
Te
mp
era
ture
(K
)
Deposit cycles with constant track
length and time duration (10 Layers)
Laser off
Temperature measured at the thermocouple
Temperature measurement
13
Thermal history at the thermocouple due to the first five layers
Sensitivity analysis: T° at thermocouple
Constant Track Lenght strategy
0 50 100 150 200300
400
500
600
700
800
900
1000
1100
1200
Time (s)
Te
mp
era
ture
(K
)Time-Temperature
Set 1 Set 2
Set 3 Experimental
ΔT end_set1=169K
ΔT end_set2=4K
ΔT end_set3=184K
Tran et al.
Materials and Design 2017
h: 5, 52, or 100 W/m2K
7350 Solid Elements
for Substrate
9274 Nodes
14
0 100 200 300 400 500300
400
500
600
700
800
900
1000
1100
Time (s)
Te
mp
era
ture
(K
)Time-Temperature
Simulation
Experiment
Comparison between the simulated and the experimental
thermal history from the thermocouple
Validation at thermocouple for 10 layers
Constant Track Lenght strategy
Tran et al.
Materials and Design 2017
15
Constant Track Lenght strategy
hDL
hHAZ-β
Fusion zone (FZ) and the heat-affected zone
(HAZ)
2D view of thermal field within HAZ
Depth Layer 1 Layer 2 Layer 3 Layer 4 Layer 5
Simu
result
hDL (μm) 508 688 709 730 793
hHAZ (μm)
HAZβ+ HAZα+β
1618 1864 2174 2377 2605
Measured
hDL (μm) 450
Not accessible, different zones cannot be
recoveredhHAZ (μm)
HAZβ+ HAZα+β
1501
Depth of Fusion zone
16
Decrease Track Length strategy
0 100 200 300 400 500300
400
500
600
700
800
900
1000
1100
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Simulation
Experiment
Validation at thermocouple for 10 layers
17
Point 1 – Section A
0 100 200 300 400 500 600
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 1
Constant Track Lenght strategy
Liquid
β
α+β
Ms
18
Point 1 – Section A
0 100 200 300 400 500 6000
50
100
150
200
250
300
350
400
450
500
550
Time-Cooling rate
Time (s)
Co
olin
g r
ate
(K
/s)
1.5K/s
20K/s
410K/s
525K/s
Constant Track Lenght strategy
High T or Low 𝑻 → Ms
19
Point 2 – Section A
0 100 200 300 400 500 600
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 2
Liquid
β
α+β
Ms
Constant Track Lenght strategy
2020
Hardness map – CTL – Paydas et al.2015
𝛼 𝛼
Constant Track Lenght strategy
21
Point 1 – Section A
0 100 200 300 400 500
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 1
Liquid
β
α+β
Ms
Decrease Track Length strategy
Ms → Yes
22
Point 2 – Section A
0 100 200 300 400 500
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 2
Liquid
β
α+β
Ms
Decrease Track Length strategy
Less Ms than point 1
23
Point 3 – Section A
0 100 200 300 400 500
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 3
Liquid
β
α+β
Ms
Decrease Track Length strategy
Less Ms than points 1 and 2
24
Point 4 – Section A
0 100 200 300 400 500
500
1000
1500
2000
2500
3000
3500
Time (s)
Te
mp
era
ture
(K
)
Time-Temperature
Tβ
TS
MS_Kelly
Mf
MS_Krespo
Point 4
Liquid
β
α+β
Ms
Decrease Track Length strategy
Ms
25
Decrease Track Length strategy
Hardness measurement (Hakan et al. 2015)
α′ martensite
26
Conclusion & Perspectives
²Qualitative microstructure prediction // experienceMs 800°C and Ms for αm higherHAZs size within substrate Melt pool size
Done
basket-wave Widmanstätten
structure
Prediction in Constant Track
Length:
• Homogeneous T° history
• 𝑇𝑎𝑣𝑒𝑟𝑎𝑔𝑒 > Ms
• 𝑇at the end low
basket-wave Widmanstätten
structure + α’ Martensite
Prediction in Decrease Track
Length:
• Heterogenous T° history
• At some location : 𝑇𝑎𝑣𝑒𝑟𝑎𝑔𝑒< Ms
and 𝑇 high
Tran et al. Materials & Design 128 (2017)
27
Conclusion & Perspectives
²Done
Couple thermo-metallurgical analysis to predict %
phase and hardness map. (Esteva)
Couple thermo-mechanical analysis to predict residual
stresses (Jardin, Son)
Need Mf to get quantitative MS through KM
However Module + Experiment available …
On going
Qualitative microstructure prediction // experienceMs 800°C and Ms for αm higherHAZs size within substrate Melt pool size
END
28