Multiphysics Modeling of the Graphite
Electrode Joint in Electric Arc Furnaces
for Scrap Steel Recycling
Nathan May
James Glickstein
Maren Waldner
1
Problem Statement
• The Electric Arc Furnace
(EAF) uses graphite rods
as electrodes to strike a
powerful arc to melt
scrap steel
• GrafTech International is
a world leader in
supplying graphite
electrodes for the EAF
process
2
Problem Statement
• The Electric Arc Furnace (EAF) is a massive
industrial process, converting scrap steel into new
raw material (millions of tons per year worldwide)
• Less greenhouse gas emissions versus Basic Oxygen
Furnace steelmaking process (BOF, “Blast Furnace”)
[1]
• BOF: 1.6 tonnes CO2/tonne steel
• EAF: 1.08 tonnes CO2/tonne steel
[1] US EPA, AVAILABLE AND EMERGING TECHNOLOGIES FOR REDUCING GREENHOUSE GAS EMISSIONS
FROM THE IRON AND STEEL INDUSTRY, Sept 2012, [Link]
12/documents/[Link]
3
Problem Statement
• Previous finite element models have been very valuable
to GrafTech’s electrode business
• Previous models were relatively simple and coarse; new
software and hardware enable next-generation of FEM
• Develop a multiphysics model of the electrode joint,
including:
• Structural mechanics and contact physics
• Heat transfer
• Electric currents
• Validate model through comparison with real data
• Better understand process/performance relationships
• Reduce design cycle time
4
Problem Statement
Power Cables
Electrode Clamp
Furnace Lid
Graphite Electrode
Electrode Joint
Steel Pile
5
Problem Statement
Power Cables
Electrode Clamp
Furnace Lid
Graphite Electrode
Electrode Joint
Arc
Steel Melt
6
Failure Mechanism: Stub Loss
Internal
Threaded
Pin
High
Thermal
Stresses
Stub Loss
7
Failure Mechanism: Oxidation
Graphite reacts with
Some materials/furnace
oxygen at high temp
conditions cause faster
to form CO2 Characteristic oxidation
“oxidation
profile” taper
8
Failure Mechanism: Column Break
Internal Internal
Threaded Threaded
Pin Pin
Bending force Bending force
from scrap pile from scrap pile
9
Problem Statement
• Graphite Electrode consumption is a significant cost driver
of EAF, in particular:
• Stub loss: the tip of the electrode can crack and cause chunks of
electrode to fall into the melt
• Oxidation: furnace temperature is well over 1000 degC, and
graphite oxidizes into CO2 rapidly at this temperature
• Column breaks: “scrap cave-ins” or poor electrode addition exert
bending stress on column, leading to breaks high up and major
process disruption
• Successive electrodes are connected at a “joint,” which
becomes a critical point for reducing failure
• Because of the extreme environment, direct measurements
are often difficult or impossible
• Graphite is a non-homogenous, anisotropic material
• Multiscale (>8 meters electrode and <10 microns defects)
and multiphysics situation
10
Approach
• Make a model that considers all three failure
mechanisms!
• Used COMSOL Multiphysics v5.3 (model developed in
v5.0)
• Structural Mechanics module
• Heat Transfer module
• Electric Current module
• Moving Mesh module
• 2D Axisymmetric model:
• Stationary pre-tensioning simulation
• Transient multiphysics simulation of firing cycle
• 3D model:
• Bending stress simulation
• Transient multiphysics simulation of non-axisymmetric modes
11
Model Scheme Electric terminal
Flowing water
w/phase change
Contact physics for
solid mechanics, heat
transfer, and electric
currents
Radiation and powerful
convection from hot
furnace
Radiation from
electric arc 12
Model Validation
• Results were used on a comparative
basis, in particular comparing the
stress states of known failure
location across different cases
• Used a variety of stress metrics,
focusing on tresca and maximum
principle stress criteria
• Focus on joint area as most
common failure location
• Completed hundreds of simulations
of different cases, grouped into
about ten study areas
13
Stub Loss versus Firing Cycle
• EAF process has a
characteristic firing cycle
• Involves heating and cooling
segments
• Causes significant thermal
stresses from temperature
and material gradients
14
Stub Loss versus Firing Cycle
• EAF process has a
characteristic firing cycle
• Involves heating and cooling
segments
• Causes significant thermal
stresses from temperature
and material gradients
15
Stub Loss versus Firing Cycle
• EAF process has a
characteristic firing cycle
• Involves heating and cooling
segments
• Causes significant thermal
stresses from temperature
and material gradients
16
Stub Loss versus Firing Cycle
• EAF process has a
characteristic firing cycle
• Involves heating and cooling
segments
• Causes significant thermal
stresses from temperature
and material gradients
• Our analysis involves
comparing these stresses at
different times and places
17
Oxidation
• Water flows from
clamps down the
exterior of the electrode
• Reduces temperature
• Provides barrier to
oxygen migration
• Too much water flow
can shift energy balance
unfavorably
18
Oxidation
• Dependent on several factors:
• Furnace type
• Furnace oxygen/inert gas usage
• Furnace power settings
• Water cooling
• Deep multiphysics coupling is
required
• Simulate water as a phase-
change domain (built-in to
COMSOL), but with constant
velocity field
19
Oxidation
No Water Flow With Reference
Water Flow
20
Oxidation
• Use in-house testing to
develop temp vs oxidation
rate curves
• Integrate vs temperature
and time
• Simulate total oxidation
• Test different materials’
oxidation rates, simulate
consumption rate savings
• Predicted oxidation rates
match reality quite well With Reference
Water Flow
21
Electrode Bending
Bending force from
scrap pile
22
3D Cases
• 3D geometry is required for
bending and other stress
states
• Results make the model
even more powerful,
though require increasingly
demanding hardware
• Model DoF/memory can
exceed 2e7/350 Gb
• Results are very useful in
understanding column
breaks
23
Future Work
• Already a useful design and
engineering tool for GrafTech
• More 3D simulations!
• Deeper sensitivity analysis vs
furnace conditions and
materials
• Crack propagation and failure
at high temperature
• Cohesive zone debonding
• Physical testing/comparison
with DIC
24
Future Work
• Already a useful design and
engineering tool for GrafTech
• More 3D simulations!
• Deeper sensitivity analysis vs
furnace conditions and
materials
• Crack propagation and failure
at high temperature
• Cohesive zone debonding
• Physical testing/comparison
with DIC
25
Thanks to Josh Thomas of
AltaSim for help setting up
early model version!
¿Questions?
Thank you for listening!
26
Model Validation
27
Oxidation
• Oxidation is the baseline driver of
graphite electrode consumption
rates
• Around half of electrode baseline
consumption is wasted as oxidation
• Improved technology has steadily
reduced oxidation rates, but more
can be achieved!
• Water cooling is one key reducer of
oxidation
• Modeling can help us understand
key operating parameters to control
and limit oxidation
28
