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