Multi-Physics Process Simulation of StaticMagnetic Fields in High Power Laser Beam

Welding of Aluminum

Marcel Bachmann, Vjaceslav Avilov,

Andrey Gumenyuk, Michael Rethmeier

COMSOL Conference Milan 2012, October 10 – 12, 2012

Excerpt from the Proceedings of the 2012 COMSOL Conference in Milan

Overview

Introduction

Physical Principles

Numerical Modeling & Results

Experimental Observation

Conclusions

COMSOL Conference Milan 2012 Marcel Bachmann 1

Introduction

Applications deep-penetration laser beam welding

Ship-building industry

Reactor vessels & tanks

Power plant components

Aerospace industry

High-Speed Car FerriesAustan Ships

Aluminum Reactor VesselsMaulifab Engineers

NASA tank dome

COMSOL Conference Milan 2012 Marcel Bachmann 2

Introduction

Characteristics of the laser beam welding of thick metal parts

Related issues:

Intensive Marangoni convection

Buoyancy forces

Recoil pressure

Surface stability

Melt ejections

Non-uniform solidification

laser beam

COMSOL Conference Milan 2012 Marcel Bachmann 3

Introduction

Possibility to decelerate the melt? apply volumetric forces contact-less

all welding positions

COMSOL Conference Milan 2012 Marcel Bachmann 4

Physical Principles

Hartmann Effect – Flow Deceleration

u

j

B

FL

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Norm alized Velocity u/ umean

0.0

0.2

0.4

0.6

0.8

1.0

No

rma

lize

d c

ha

nn

el

wid

th

Ha2 = 0 analyt ic

Ha2 = 1 analyt ic

Ha2 = 25 analyt ic

Ha2 = 100 analyt ic

Ha2 = 0 sim ulat ion

Ha2 = 1 sim ulat ion

Ha2 = 25 sim ulat ion

Ha2 = 100 sim ulat ion

1.6

Liquid metal flow induces electric currents

Ratio of magnetic induced to viscous drag Ha2 = (BL)2σ/η

Resulting Lorentz force directed against melt movement

COMSOL Conference Milan 2012 Marcel Bachmann 5

Physical Principles

Permanent magnets applied to partial penetration welding

Mitigation of weld pool dynamics

Avoid turbulent flow pattern

Achieve parallel side walls

COMSOL Conference Milan 2012 Marcel Bachmann 6

Numerical Modelling

air

magnet

workpiece

50 mm

20 mm

50 mm

~21 mm

keyhole

Workpiece width: 40 mm

Magnet cross section:

50 mm × 50 mm

Vertical shift of the

magnet: 20 mm

Thermal

Simulation

Computational

Fluid Dynamics

Simulation

Electro-

magnetic

Simulation

Material

Properties

900,000 tetrahedral finite elements ⇒ 8,100,000 DoF

COMSOL Conference Milan 2012 Marcel Bachmann 7

Numerical Modelling

symmetry

WORKPIECE

inflowoutflow

adiabatic

free slip

Marangoni

adiabatic

free slip

Marangoni

adiabatic

free slip

free slip

adiabatic

Marangoni

adiabatic

free slip

Navier Stokes equations volume source term

F = −ρg − c1(1−fL)

2

f3L+ε

(u− uweld) + j×B

Buoyancy Solidification modelling Lorentz force

COMSOL Conference Milan 2012 Marcel Bachmann 8

Velocity distribution

B = 0 B = 0.5 T

B = 1.0 T B = 2.0 T

0.1

0

0.02

0.04

0.06

0.08

velo

city

, m

/s

Welding speed 0.5 m/min

3D flow ⇒ 2D flow

Buoyancy suppression

Marangoni flow suppression

COMSOL Conference Milan 2012 Marcel Bachmann 9

Velocity distribution

0.2 m/s

�5 0 5 10 15 20x, mm

40

42

44

46

48

50

z, m

m

0.2 m/s

�5 0 5 10 15 20x, mm

40

42

44

46

48

50

z, m

m

0.2 m/s

�5 0 5 10 15 20x, mm

40

42

44

46

48

50

z, m

m

0.2 m/s

�5 0 5 10 15 20x, mm

40

42

44

46

48

50

z, m

m

B=0 B=0.5 T

B=1 T B=2 T

COMSOL Conference Milan 2012 Marcel Bachmann 10

Temperature distribution

500

750

1000

1250

1500

1750

2000

2250

2500

Te m pe ra ture T, K

B = 0 B = 0.5 T B = 1 T B = 2 T

Marangoni convection suppressed at free surface

Weld bead shortening

Curvature in solidification line eliminated

COMSOL Conference Milan 2012 Marcel Bachmann 11

Macro Section Solidification line

0 2 4 6 8 10 12 14y, mm

25

30

35

40

45

50z

- pos

ition

, mm

B = 0B = 0.5 TB = 1.0 TB = 2.0 Ttheoretical case

Marangoni convection suppressed

Weld bead width decrease

Curvature in solidification line eliminated

COMSOL Conference Milan 2012 Marcel Bachmann 12

Local Hartmann Number

0.0 0.2 0.4 0.6 0.8 1.0magnetic flux density B, T

102

103

104

105

106

Har

tman

n n

um

ber

Ha2

laminar Hartmann number

turbulent Hartmann number

η = ηlam + ηturb

Jump in Ha2 = (BL)2σ/η

Once turbulence is damped

Ha2lam ≈ Ha2turb

Critical Ha2 for effective flow

control around B=0.4 TPosition: 5 mm behind keyhole and 5 mm below surface

COMSOL Conference Milan 2012 Marcel Bachmann 13

Experimental Observation

AlMg3 Welding Experiments with 16 kW disc laser system andpermanent magnets

magnet pole

magnet pole

Bmax around 0.5 T

Weld narrowing

Less surface waviness

Wine-glass shape ⇒ V shapeLaser power: 16 kWWelding velocity: 0.5 m/minFocus position: -4 mmShielding gas: 30 l Ar

COMSOL Conference Milan 2012 Marcel Bachmann 14

Experimental Observation

Reference B = 500 mT

⇒ Same tendency in experiments and simulations!COMSOL Conference Milan 2012 Marcel Bachmann 15

Experimental Observation

B = 500 mT

Same results with changed

orientation of the magnet field

Evidence for the occurance of

Hartmann effect

Laser power: 16 kWWelding velocity: 0.5 m/minFocus position: -4 mmShielding gas: 20 l Ar

COMSOL Conference Milan 2012 Marcel Bachmann 16

Conclusions

Influence of steady magnetic fields in partial penetration laser beamwelding of aluminium is shown

Significant changes in the velocity distributions

Maximal turbulence damping beginning with Ha2 ≈ 104

Weld pool geometry changes due to Hartmann effect with increasingmagnetic field from wine-glass shape to V shape

Changes in solidification line can influence mechanical andmetallurgical properties

COMSOL Conference Milan 2012 Marcel Bachmann 17

Thank you for your attention.

Grant No. GU 1211/2-1 Grant No. 17.625 N

COMSOL Conference Milan 2012, October 10 – 12, 2012