Scribd
Upload
130 views27 pages

Web073 Die Casting Modeling Capabilities2. Modeling Differences

The document discusses modeling capabilities for die casting, specifically comparing finite element and finite volume modeling approaches. It covers topics like the basic equations, differences between the approaches, considerations for mesh density and heat transfer coefficients, and comparing simulation results to experimental data.

Uploaded by

sinr
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
130 views27 pages

Web073 Die Casting Modeling Capabilities2. Modeling Differences

The document discusses modeling capabilities for die casting, specifically comparing finite element and finite volume modeling approaches. It covers topics like the basic equations, differences between the approaches, considerations for mesh density and heat transfer coefficients, and comparing simulation results to experimental data.

Uploaded by

sinr
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 27

Die casting modeling capabilities

2. Modeling Differences

By Charles Monroe
Charles Monroe, PhD

• Assistant Professor at the University of Alabama


at Birmingham (2012-present)
• Senior Research Engineer at Caterpillar Inc. in the
Materials Processing Group (2008-2011)
• PhD and MS work at University of Iowa on topic of
Hot Tearing (2003-2008)
• Undergrad at Penn State (1999-2003)
Multi-scale modeling exists because no
single model can predict all the behavior
Simulation is an approximation of
reality.
• Limited to questions asked.
• Three basic conservation equations
– Mass
– Momentum
– Energy
• Simulation tips
– Project contract
– Mesh density
– Etc.
Different approaches in simulation to
reach the same solution.
• The basic equations never change between
packages
• User interface, speed, support, and
presentation of results can change
– Need to understand what your company offers
• Most simulation packages blur the traditional
distinction of method:
– Finite Element or
– Finite Volume or Differencing
What is Finite Element?
Nodes Element
Displacements Force and Stress
Temperature Heat flow and Energy

Notes about finite elements


1. Unstructured (can be irregular in shape and connectivity)
2. At least four nodes are needed to define each element
(more for higher order elements)
3. Data is known on the boundaries not in the center of
elements
4. Best for continuous or smooth fields like stresses or
internal flows without boundaries
What is Finite Volume?
Nodes and Elements are the same.
Displacements Force and Stress
Temperature Heat flow and Energy

Notes about finite volumes


1. Structured (known neighbors)
2. One node/element and only first order within the element.
3. Data is known in the center of element and not on the boundaries
4. Best for discontinuous fields like solidification, flows with boundaries,
etc.
Be aware of the base methods of your package and
these limitations will guide you as you ask questions
about how your simulation can be improved.

Finite Volume Finite Element


 Not boundary fitted (Does not  Boundary fitted (matched
match shape) shape)
 Easy to get a quality mesh (more  Not easy to get good mesh
accurate internally) (not as accurate internally)
 Discontinuities handled easily  Discontinuities hard to handle
 Does not handle stresses well  Handles stress calculation

Takeaway: at similar mesh densities and proper setup both


methods will give the same or very similar results!
Memory requirements can be greater for
Finite Element than for Finite Volume.
If you match meshes between Finite Element and
Finite Volume you have 2 options
1. Finite Element nodes = Finite Volume elements
• Similar memory requirement
• Quality of the finite element result is not necessarily
good because there are less nodes in the bulk
2. Finite Element elements = Finite Volume elements
• Finite element requires much more memory ~4 times
• Quality of finite element solution is generally better but
due to higher refinement
Available packages as advertised at the
NADCA Congress
• Those with booths at the NADCA Congress
– EKK
– ESI Group with ProCast and Quickcast
– Flow3D, Flow Science Inc
– Magma
– Novacast
Modeling Package Differences
• Broadly categorized across spectrum of two
categories

Novacast Flow3D EKK


Magmasoft Procast

Finite Volume Finite Element

• All the packages solve the same fundamental


conservation equations
• Can be setup to be show the same basic results
Important questions to ask about your
modeling software.
• Is my management still
committed to supporting this
technology?
• How effectively am I able to get
buy-in from others in my
organization?
• How is my simulation team?
• Who is in the driver’s seat?
• What are our goals and how do
we measure our success?

A Nowak, 2000, Die Casting Engeer


Specific method related issues to
consider.
• Coarse mesh examples
• HTC differences
• Reading results
You should speak with your software provider to
ask about how these issues are addressed with
your package.
Coarse Meshes do not represent the
geometry well.
Geometry Finite Element Finite Volume

At nodes At cell centers

Temperatures
Heat transfer Coefficients
can differ between methods.

hFE hFV

q = h (T -T¥ ) hexp ¹ hFE ¹ hFV


Note: Heat transfer coefficient, h, has units of W/m2K and
depends on surface area and mesh refinement, quality of the
temperature results on the boundary.
Note: The current standard practice to measure heat transfer
coefficients is to measure temperatures and back out from
the simulation what the h should be for that process. The
standard h values as default in your code are useful rules of
thumb.
Reading results?
Why is reading results such a big deal?
• We have results everyone can agree with
– Temperature
– Velocity
– Pressure
– Displacement
– Anything that can be measured directly
• Results causing argument
– Any criteria functions, especially porosity
– Air entrainment
– Tracers or material tracking
Comparison with CT results
• Quantitative comparison of pore fraction between simulation and experiment
• Qualitative comparison favorable
• Pore fraction over predicted by simulation

3.5 Base Case

3
#177 width
x
Pore Area Fraction, gp (%)

#176
2.5

1.5

0.5

0
0 2 4 6
Distance, z (mm)

0
Simulations are modeling transient
process.
• Steady state produces one result
• Transient process produces many results
– Grab snapshots at instances of the process
• Velocity
• Temperature
– Final results or Criteria functions
• Maximum velocity or erosion (integrated quantity)
• Last place to fill
• These results inform each other (i.e. looking at
the temperature field shows how the porosity
formed)
Two reasons for simulation porosity
not to agree with real castings
• Simulation is not including something that
happened during the real process
– Oxides, venting, htc, mesh, setup, etc.
– Narrows our investigation to these issues
• Process is variable itself and the example is
not representative of the simulation
– Example following
– We can simulate this too!
Simulation are correlated to castings
and do not give exact agreement
Simulation Experimental
Result Results
Base Case Casting #177 Casting #176 Casting#189

gp,avg = 1.72% gp,avg = 0.50% gp,avg = 0.48% gp,avg = 0.23%


Xray versus Slice views should be
carefully used
• X-ray view comments
– Can be difficult to determine locations of defects
– only see the outside of the defect
– Validation is qualitative including size
– Geometric root cause may be easier to see
• Slice
– Location is easier
– Magnitude can be determined with a good slice
– Validation is easier when compared to cut sections
– Could miss indications
Xray and Slice Example

Slice Xray
Process variability can be modeled and
then used to improve the method

Modify the pressure profile


4 things to do
• Re-evaluate Nowak’s questions today
• List productivity issues that give you re-
occurring problems like in meshing, htc, etc.
and ask your simulation package provider.
(they probably know some good tricks)
• Check your meshes and htc properties, are
they appropriate for your problems?
• Are your results answering the questions
posed?
Attend the next webinar
1. Major simulation limitations
– Spray model
– Flow length
– Cooling lines
– Heating loads (steady state or cycle interruptions)
Questions
Name: Charles Monroe
Phone: 205-975-4128
Email: [email protected]

You might also like