Experimental investigations and numerical modelling of metal melt - - PowerPoint PPT Presentation

1

Experimental investigations and numerical modelling of metal melt flows in induction furnaces

Sino-German-Workshop October 11-12, 2004, Shanghai (China) Egbert Baake, Andrejs Umbrashko Institute for Electrothermal Processes University of Hannover

Contact:

• Prof. Dr.-Ing. Egbert Baake

Tel.: +49 511 762 3248 Institute for Electrothermal Processes Fax: +49 511 762 3275 University of Hannover e-mail: baake@ewh.uni-hannover.de Wilhelm-Busch-Str. 4 Internet: www.etp.uni-hannover.de D-30167 Hannover Germany

2

Introduction Characteristics of turbulent metal melt flow in induction furnaces

• low frequency oscillations
• 3D flow structure

Melt flow in cold crucible furnace

• experimental work
• 3D tansient numerical simulation

Conclusions

Contents

3

Industrial process requirements for melting in induction furnaces

Mixing and homogenisation

• f the entire melt

Homogenisation of the temperature, avoiding of local overheating, but realizing of sufficient superheating of the entire melt Intensive stirring at the melt surface (melting of small-sized scrap, carburization process) Avoiding of erosion and clogging

• f the ceramic lining

Avoiding of melt instabilities, splashing or pinching Intensive stirring for cleaning of the melt (zinc removing)

Optimisation of the heat

and mass exchange in the melt

4

Main features of the induction furnace metal melting processes

Heat and mass transfer and the temperature distribution in the melt are determined by 3D instationary turbulent melt flows Experimental investigations, e.g. measurements of the turbulent melt flows are very limited in industrial furnaces, experimental investigations are possible in model furnaces with model melts Optimal design and optimisation of the operation behaviour of induction furnaces needs 3D instationary numerical simulation of the turbulent melt flow and the heat and mass transfer using experimentally verified models

5

magnetic field

• distribution of power
• el. efficiency

geometry of melt

velocity field

homogenisation of melt

temperature field meniscus shape

• superheating
• heat flow
• crucible temperature

skull

liquid-solid-interface

Physical Correlations

• ptimization of design and operating parameters

6

Crucible wall Crucible bottom

Melt flow measurements in induction crucible furnace

Vmax ≈ 20 cm/s

1

max max ≈

′ v v

Baake, E. et.al.: 1994

7

• 25
• 20
• 15
• 10
• 5

5 10 15 20 25 10 20 30 40 50 Zeit in s Geschwindigkeit in cm/s

Measurement of local flow velocity (ICF): near the crucible wall between the main flow eddies

Low-frequency oscillations Oscillation period: 8...12 sec

8

crucible wall crucible bottom

Application of 2D and 3D RANS (k-ε) turbulence models:

Calculation results of turbulent characteristics are different from measurement results

9

RANS (k-ε model)

• Whole energy spectrum is modelled
• Relatively low mesh resolution

requirements

• Steady-state simulations

RANS (k-ε model)

• Whole energy spectrum is modelled
• Relatively low mesh resolution

requirements

• Steady-state simulations

DNS

• All scales are resolved directly
• Very high requirements for

computational resources

• Simulations of industrial installations

are impossible DNS

• All scales are resolved directly
• Very high requirements for

computational resources

• Simulations of industrial installations

are impossible LES

• Large scales are resolved directly while only small

scales are modelled

• Relatively high mesh resolution requirements
• Transient 3D simulations

LES

• Large scales are resolved directly while only small

scales are modelled

• Relatively high mesh resolution requirements
• Transient 3D simulations

CFD problem Re ≥ 104 CFD problem Re ≥ 104

10

Calculated local flow velocity: (3D transient LES)

Low-frequency oscillations Oscillation period: appr. 10 sec

• 0,25
• 0,2
• 0,15
• 0,1
• 0,05

0,05 0,1 0,15 0,2 0,25 10 20 30 40 50 60 Zeit in s Geschwindigkeit in m/s r = 0.155 r = 0.14 r = 0.075

11

Frequency (Hz) Amplitude

Fourier analysis of the measured and calculated

• scillations of the axial velocity components near

the crucible wall between the main flow eddies Measurement Calculation

12

Long-time period averaged velocity field in the melt of the ICF (3D transient LES)

velocity magnitude azimuthal velocity particles trajectories

13

3D hydrodynamic model of an industrial induction crucible furnace

P = 4540 KW Hind = 1.33 m Rcr = 0.49 m Filling level 90 %

14

Kinetic energy of the oscillations

Experiment (model furnace) Calculations (industrial furnace)

15

Time-averaged flow pattern [m/s]

16

Transient flow development [m/s]

17

Transient flow development (cross-section) [m/s]

18

Calculation of the particle tracing in the melt

• f the ICF (3D transient LES)

19

melt with meniscus shape crucible segment inductor bottom slit

skull current

melt flow EM-forces heat conduction radiation

(water cooled) (water cooled) (water cooled

Features of the Induction Furnace with Cold Crucible

slitted crucible to realize efficient electromagnetic transparency free melt surface, based

• n electromagnetic forces

water cooled bottom and crucible segments leads to solid layer (skull) heat losses by radiation and conduction depending

• n the meniscus shape

20

wall-skull bottom-skull contact point free surface crucible-bottom crucible inductor real ideal

Optimisation of electromagnetic and thermal parameters

Maximisation of the overheating temperature, which is the key parameter of the process Improvement of the total efficiency of process Reliable, reproducible and stable melting process

Cold crucible induction furnace

21

Experimental set-up for semi-levitation melting

Rc = 72.5 mm Hi = 208 mm 5 coil turns P = 200 kW f = 9.2 kHz T = 660-720°C Rc = 72.5 mm Hi = 208 mm 5 coil turns P = 200 kW f = 9.2 kHz T = 660-720°C

22

Melting process of Aluminium

23

Temperature measurements in Aluminium

24

Measured temperature field in Aluminium

25

Melt flow measurements in Aluminium with the electromagnetic velocity probe

6

35

14

stainless- steel holder stainless- steel case magnet core electrodes

coil

26

Flow pattern and temperature distribution simulated with 2D RNG k-ε turbulence model

27

3D LES-model for Aluminium melting

~3.8•106 elements Time step 10 ms Smagorinsky-Lilly subgrid viscosity model Parallel computations with FLUENT 6.1 software at the HLRN*-system ~3.8•106 elements Time step 10 ms Smagorinsky-Lilly subgrid viscosity model Parallel computations with FLUENT 6.1 software at the HLRN*-system

*HLRN – scientific supercomputer network of North Germany

28

[m/s] [m/s]

Time-averaged flow pattern Time-averaged flow pattern An intermediate flow pattern An intermediate flow pattern Results of 3D transient LES modeling

vm~40 cm/s

29

Results of 3D transient LES modeling

30

Time-averaged temperature distribution Time-averaged temperature distribution

ºC

Results of 3D transient LES modelling

Measured temperature distribution Measured temperature distribution

ºC

31

Results of 3D transient LES modelling

32

Results of 3D transient LES modelling

33

Conclusions

Heat and mass transfer processes in the melt of induction furnaces are significantly influenced by large scale low-frequency oscillations of the recirculating flow main eddies Comparison of the LES modelling results with experimental results show good agreement 3D-transient LES is a reliable numerical tool to simulate the turbulent melt flow with in-stationary low-frequency flow

• scillations in induction melting installations