Advanced Lost Foam Casting Technology PH

FINAL REPORT on
ADVANCED LOST FOAM CASTING TECHNOLOGY

Phase V
By
WANLIANG SUN
HARRY E. LITTLETON
And
CHARLES E. BATES
TO
THE DEPARTMENT OF ENERGY,
THE AMERICAN FOUNDRY SOCIETY,
and
AFS-DOE-LFC FOAM CASTING CONSORTIUM MEMBERS
DOE CONTRACT N0. DE-FC07-99ID13840
REPORT NO. 527985-2004 Project Final Report
February 2005
ADVANCED LOST FOAM CASTING TECHNOLOGY, PHASE V
Table of Contents
EXECUTIVE SUMMARY ··················································· 5
ACCOMPLISHMENTS
1. Computational Model and Data Base ································ 10
1.1 Pattern Material Degradation Properties Data Base ··············· 10

1.1.1 Characterization of Pattern Degradation through Foam Pyrolysis ·10
1.1.1.1 Heater Material ······································· 11
1.1.1.2 Heater Size and Shape ·································· 11
1.1.1.3 Gas Generation Regimes ································· 13
1.1.1.4 Gas Generation Rate ···································· 13
1.1.1.5 Temperature of Gaseous and Liquid Pyrolysis Products ····· 13
1.1.1.6 Resistance Pressure of Pattern Degradation ·············· 14
1.1.1.7 Molecular Weight of Foam Pyrolysis Products ············· 14
1.1.1.8 Effects of Molded and Sawed Surfaces ···················· 14
1.1.1.9 Effects of Foam Pattern Fusion Level ···················· 14
1.1.1.10 Degradation of Pattern Adhesives ······················ 15
1.1.2 Visualization of Metal Foam Replacement through Real Time X-Ray

Technology ··················································· 16
1.1.2.1 Real Time X-Ray Apparatus ······························ 16
1.1.2.2 Metal Filling of Lost Foam Aluminum Castings ············· 18
Experiment Matrix; Coating Perm and Pattern Fusion Level; Foam
Density; Coating Permeability; Glue Joint; Metal Filling and
Casting Quality; Damaged Coating; Sectional Metal Front
Profile
1.1.2.3 Interior Quality of Lost Foam Aluminum Castings ·········· 21
1.1.2.4 Effects of Pattern Tooling Types on Metal Filling ········ 21
1.1.2.5 Metal Filling of Lost Foam Iron Castings ················· 22
1.1.3 Pattern Degradation and Casting Defect Formation ················ 23

1.1.3.1 Effect of Down Sprue Design ······························ 23
1.1.3.2 Effect of Process Parameters on Defect Formation ·········· 23
1.1.3.3 Effect of Foam Materials (T180 and T185) ·················· 24
1.1.3.4 Effect of Alloy Chemistry ······························· 25
1.1.3.5 Effect of Superheat and Hydrogen Content ················· 26
1.1.3.6 Development of Defect Database ·························· 26
1.2 Computational Model ············································ 27

1.2.1 Problems of Existing Codes ··································· 27
1.2.2 Development of Physical Model ································ 28
1.2.3 Development of Algorithms ····································· 29
1.2.4 Achievement in Computational Model ····························· 29
1.2.5 Validation and Improvement with Real Time X-Ray ·············· 31
- 2 -
2. Casting Dimensional Accuracy ····································· 31
2.1 Identify Sources of Casting Distortion ························· 31
2.1.1 Case Study 1: Long Pipe ······································· 31

2.1.2 Case Study 2: Valve Body ······································ 31
2.1.3 Sources of Casting Distortion ································ 33
3. Pattern Production ··············································· 33
3.1 Pattern Bead Properties ········································ 33

3.1.1 Properties of Raw Beads ······································ 33
3.1.2 Pattern Density Measurement with Dielectric Sensors ·········· 34
3.1.3 Pattern Permeability ········································· 34
3.1.3.1 Apparatus ············································· 34
3.1.3.2 Pattern thickness and fusion level ····················· 34
3.1.3.3 Pattern permeability, metal filling and casting quality ·· 36
3.1.3.4 Threshold of air flow rate ······························ 37
3.1.4 Computational modeling of Bead Filling and Steaming ·········· 37
3.2 Pattern Tooling ················································ 37

3.2.1 Tooling design ················································ 37
3.2.2 Tooling Types ················································ 38
3.2.3 Filling and Bead Arrangement ································· 38
3.2.4 Pattern Molding Parameters ···································· 39
3.2.4.1 Experimental approach ································ 39
3.2.4.2 Pattern fusion level ··································· 39
3.2.4.3 Open porosity ········································· 40
3.2.4.4 Density Gradient ······································ 40
3.2.4.5 Foam pattern temperature and pressure ··················· 41
3.2.5 Pattern Properties and Casting Quality ························ 41
3.2.6 Pattern Quality Control Manual ······························· 42
4. Improved Pattern Materials and Processes ························ 42
4.1 Improved Pattern Materials ····································· 42

4.1.1 Foam Pattern Pre-coat ········································· 42
4.1.2 Development of “No Fold” Bead Materials ······················ 43
4.1.3 Evaluation of Alternative Foam Materials – Polyurethane ······ 43
4.1.4 Evaluation of Alternative Foam Materials – PMMA ··············· 44
4.2 Improved Pattern Degradation Process ··························· 44

4.2.1 Pattern Additives ············································· 44
4.2.2 Coating Additives ············································ 45
4.2.3 Effects of Pouring Atmosphere ·································· 45
4.2.4 Minimum Degradation Energy ····································· 45
- 3 -
5 Coating Control ··················································· 46
5.1 Coating Quality Control ········································ 46

5.1.1 Coating Control – Case Study ································· 46
5.1.2 Correlation between UAB and GM Coating Permeability Test ····· 46
5.1.3 Reorientation of Research Focus ······························· 47
5.2 Coating Consistency ············································ 47
5.3 Vacuum Assist ··················································· 47

5.3.1 Drawback of Vacuum Assist ···································· 47
5.3.2 Metal Filling of Lost Foam Iron with Vacuum Assist ··········· 48
6. In-Plant Case Studies and Technology Exchange ··················· 48
6.1 Case Study – Equipment Maintenance ····························· 48

6.2 Case Study – Gas Cavity and Cold Laps in Lost Foam Iron
Castings ··························································· 49
6.3 Case Study – Pattern Quality and Casting Quality ··············· 50
6.4 Case Study – Pattern Permeability Control on the Production Floor
······························································· 50
6.5 Case Study - “House keeping” on the production Floor ··········· 50
6.6 Case study – Leakage in Cylinder Head ·························· 51
7. Energy and Environmental Data ···································· 51
8. Technology Transfer ···············································51
Acknowledgements ····················································53
Use of This Report and Information Contained Herein ·················54
- 4 -
EXECUTIVE SUMMARY
Previous research, conducted under DOE Contracts #DE-FC07-89ID12869,

DE-FCO7-93ID12230 and DE-FC07-95ID113358 made significant advances in
understanding the Lost Foam Casting (LFC) Process and clearly identified
areas where additional developments were needed to improve the process
and make it more functional in industrial environments. The current project
focused on eight tasks listed as follows:
Task 1. Computational Model for the Process and Data Base to Support
the Model
Task 2. Casting Dimensional Accuracy
Task 3. Pattern Production
Task 4. Improved Pattern Materials
Task 5. Coating Control
Task 6. In-Plant Case Studies
Task 7. Energy and the Environmental Data
Task 8. Technology Transfer
This report summarizes the work done on all tasks in the period of
October 1, 1999 through September 30, 2004. The results obtained in each
task and subtask are summarized in this Executive Summary and details are
provided in subsequent sections of the report.
Task 1. Computational Model and Data Base. This task developed (1)
a data base on foam degradation and (2) a physical model of the foam
replacement process.
There were three subtasks that involved (1) conducting foam pyrolysis
experiments, (2) examining pattern replacement using a real time X-Ray
system, 3) observe defect formation in castings using the x-ray system,
and (4) pattern replacement modeling. This task resulted in much more
accurate computer simulations useful in minimizing defect incidence.
Foam pyrolysis experiments – These experiments measured parameters

associated with foam replacement by molten metal and provided data needed
to predict defect formation with mathematical models. The properties
measured included pyrolysis gas evolution rates, effluent gas temperatures,
gas pressures, and gas molecular weights. Degradation parameters
associated with patterns having various degrees of surface quality were
measured, and data on the degradation of pattern adhesives were gathered.
The effect of pyrolysis heater geometry on pattern degradation was examined
to be sure that the data being developed was accurate. It was determined
that the heater geometry did not have a significant effect on foam
degradation.
The database has been used in two computer codes to reduce scrap
and casting defects. General Motors Corporation has developed a
proprietary code to substantially reduce scrap and processing costs. These
- 5 -
data have also been provided to Flow 3D, MagmaSoft, and Procast, and two
of these codes have incorporated some of the data.
Real time X-Ray – An x-ray system was used to study pattern replacement
by molten metal. Aluminum and iron casting studies were conducted. Effects
of pattern fusion, pattern density, coating permeability, and glue joints
were determined.
The x-ray system has provided valuable data on foam replacement.

Pattern quality and consistency was found to be essential for controlling
defect formation in castings. Perhaps 60% of the scrap produced in
commercial foundries is associated with inconsistencies in patterns.
The x-ray system has also provided data on how folds and porosity
occur in castings, and the information has been used by participants to
improve castings quality. X-Ray data provides a benchmark for validating
computer models of flow, fill, and pattern replacement. Quantitative data
such as metal filling time, velocity and number of converging metal fronts
have been extracted from video files and related to the pattern quality
and the data used to improve computer codes.
Defect formation – Defect formation, especially fold and porosity

formation, and effects of sprue design, process parameters, foam materials,
alloy composition, metal superheat, hydrogen content, and visual
interpretation of defective surfaces were investigated. Results of these
evaluations are summarized as follows:
• Four sprue designs were evaluated without finding any significant

difference in metal velocity or temperature. However, castings
poured with foam sprues had more fold defects than those poured
with ceramic sprues.
• Foam density, degree of foam bead fusion, and coating permeability
effects on casting defect formation were evaluated in a statistical
matrix involving 24 castings. Patterns having a low degree of
fusion always filled at a high rate regardless of the coating type.
Patterns coated with low permeability wash materials had fewer
fold defects.
• Foam bead materials with and without bromine additives were
evaluated. The additives substantially reduced the incidence of
casting defects.
• The effects of aluminum composition on defect formation were
investigated. The silicon concentration had no significant effect
on the incidence of fold defects. However, castings produced with
aluminum alloy 206, had significantly more casting defects than
those produced with 319 or 356 alloys.
• The effect of metal superheat and hydrogen content on porosity
was examined. As the superheat increased the metal velocity
increased slightly, and fewer casting defects were produced.
• Data on root causes of defects in Lost Foam Casting were compiled.

The database correlates the appearance of casting defects such
- 6 -
as folds, blisters and internal porosity to the Auger analysis.
The defects were found to contain various amounts of carbon, oxygen,
aluminum and silicon that affect the appearance. Now, the
appearance can be related to the defect cause.
Computational model – Flow Science and ESI ProCast are the two leading
commercially available programs for simulating the Lost Foam casting
process. In order to improve the commercial codes, experiments were
conducted to develop a physical model of events occurring in the mold and
algorithms that describe these events. Both ESI Procast and Flow 3D have
incorporated some of the information and made significant progress in
describing pattern pyrolysis.
ESI Procast released a new version of their code for simulating the
lost foam process and provided UAB a copy for beta testing with sponsor
companies. The new version allows inputs for coating parameters including
coating thickness and permeability; pattern parameters including gas
fraction, and glue joint effects; and heat transfer at the metal-foam
interface to be varied and the effects calculated.
Flow 3D implemented a new pattern density model to simulate density

gradients in patterns. The real time X–Ray system validated this code.
Task 2. Casting Dimensional Accuracy. Two sponsors having dimensional

issues were examined to resolve dimensional variations. Previous work has
shown that dimensional variation can be caused by excessive compaction
table accelerations during the fill and compaction cycle and by sand thermal
expansion. Distortion in one of the foundries was found to be caused by
uneven sand filling in and around the pattern.
Three approaches are available to reduce distortion caused by uneven

sand filling including:
• Adjust the fill and compaction cycle to ensure internal cavities

in the pattern are filled at the same rate as the exterior to
balance the compaction pressures on the pattern.
• Re-orient the pattern to allow easier pattern cavity filling.
• Mask the fill screen to achieve a more uniform fill of the flask.
Task 3. Pattern Production. This task focused on determining the

effects of foam bead and pattern tooling on pattern permeability.
• A pattern permeability apparatus was developed as a tool for

monitoring and controlling pattern permeability. The threshold
pattern permeability affecting the metal flow was found to be about
1.0 cc/sec-cm2.
• Validation experiments were conducted, and Arena-Flow-EPS was found
to accurately describe the flow of beads into pattern cavities. The
model has been found useful for tooling design and pattern defect
analysis. A density model was implemented in the code in this phase
- 7 -
of work to make the modeling more robust.
• Pressure transducers and thermocouples were placed in a tool to
measure the effects of molding parameters on pattern quality. It
was found that the time the pattern is above 100°C affects porosity
and the pattern density gradient, which in turn influences defect
formation in castings.
• Patterns molded under different conditions were poured to determine
effects of pattern variations on defect formation. The amount of
open porosity in the pattern was directly correlated with casting
scrap. This result was consistent with observations made with the
x-ray equipment.
• A manual was developed that outlines the parameters that must be
controlled during pattern molding and how these parameters can be
controlled to make consistent patterns. The manual is available at
AFS and UAB.
Task 4. Improved Pattern Materials and Processes. Modifying currently

available pattern bead materials was explored to make the process more
robust. Using a pattern pre-coat provided some guidance for recent research.
Test castings poured using this pattern pre-coat exhibited an average of
a 82% reduction in the number of surface folds and a reduction in average
fold length by 42%.
A new bead material, referred to as “no fold”, was introduced by

a participating company. The new bead stock reduced the fold defect
formation. Other foam stock including Polyurethane and PMMA were evaluated.
None of the materials evaluated produced acceptable castings.
Several approaches were used in an effort to improve the pattern

degradation and minimize casting defects. Coating and pattern additives
were explored, respectively, and it was found that certain additives
significantly reduced the incidence of casting defects. Pouring under
certain atmospheres was also found to be beneficial. The patentability
of these concepts is now being determined.
Task 5. Coating Control. Developments on coating consistency were

continued in participating company plants. A high scrap rate on an engine
block was associated with inconsistent permeability. Measurement
consistency was examined and values obtained at UAB were compared to values
obtained in a sponsor company. Correlations were made on two pieces of
equipment using ten coating materials commonly used for producing iron
and aluminum castings. Some effort was also directed at measuring system
permeability, which includes both coating permeability and pattern
permeability.
Pouring under the influence of vacuum was explored. The application

of a vacuum might help remove pyrolysis products, but the high fill rates
often increases the incidence of pyrolysis defects and metal penetration.
No effective way was found to control the carbon pick-up in steel castings
so work on this subtask was directed, with the agreement of consortium
sponsors, to in-plant trouble shooting and technology exchange.
- 8 -
Task 6. In-Plant Troubleshooting and Technology Exchange. Two levels
of training courses were developed and given at participating foundry
companies. “Entry Level” was designed for new employees, and the “Process
Level” course was designed for engineers and production personnel. Both
courses provided insight and understanding of process control needed to
minimize scrap.
Seven case studies conducted at participating foundries illustrate

how to control the process and produce quality castings. These case studies
not only helped improve the production yield, but also provided UAB with
insight into real issues of interest to operating companies.
Task 7. Energy and Environment Data. Energy consumption and emissions

data on the Lost Foam process were compiled from consortium members. In
general, the environmental burdens associated with the Lost Foam and
semi-permanent mold processes are similar. Overall, Lost Foam was found
to be the most environmentally friendly process among semi-permanent mold
and precision sand process.
Task 8. Technology Transfer. Technology transfer between UAB

personnel and sponsors occurred in the form of meetings at four month
intervals. Thirteen (13) sponsor meetings were held at UAB, at American
Foundry Society (AFS) Headquarters in Chicago and in plants of sponsor
companies. Additional meetings were held to review the research and examine
its applicability in sponsor foundries. Technology transfer was also
performed though conferences such as AFS Casting Congress, Lost Foam
International Conference and TMS Annual Meeting.
- 9 -
2000 - 2004 Final Report
OBJECTIVE
The purpose of this project was to advance the state of the art for
the Lost Foam (LF) Casting process and improve the competitiveness of the
U.S. metal casting industry. Activities to accomplish these objectives
are conducted through a Lost Foam Technology Center located at the
University of Alabama at Birmingham (UAB). The Center is focused on
developing and demonstrating advanced casting technology and transferring
this technology to project participants. Participants include casting
designers, foundry suppliers, equipment producers, producing foundries
and casting users. This project provides a means for designers,
manufacturers, and purchasers/users of castings to advance the science
of casting production. The project gives participants contact with leaders
in the industry and helps develop the best available technology.
ACCOMPLISHMENTS
1. Computational Model and Data Base
The objective of this task was to develop a data base of liquid/gas

ratio, energy of degradation, and power density for patterns used in
producing both aluminum and iron castings. The technology developed allows
new materials to be evaluated in a straightforward manner and with a minimum
of foundry casting experiments. The resulting data is used in a
computational model to describe pattern replacement. The second objective
was to understand and minimize defects associated with pattern replacement
(pyrolysis). The computational model predictions were compared to
experimental casting data to validate the model.
1.1 - Pattern Material Degradation Properties Data Base
Foam pyrolysis studies focused on mass and heat transfer in the

kinetic zone between molten metal and the pattern during pouring and
developed data for a mathematical model. This subtask was accomplished
through three tasks conducted at UAB (University of Alabama at Birmingham)
and UMR (University of Missouri - Rolla):
1. Pattern degradation during Foam Pyrolysis

2. Pattern replacement visualization using real time x-ray technology
3. Relating pattern degradation to casting defect formation
Achievements in these tasks are summarized as follows:
1.1.1 Characterization of Pattern Degradation during Foam Pyrolysis
- 10 -
1.1.1.1 Heater Material
An effort was made to extend the temperature capability of the foam

pyrolysis apparatus to iron and steel pouring temperatures (1600°C) using
silicon carbide (SiC) heaters. SiC offers resistance to oxidation and is
a relatively low cost material. A 0.95 cm diameter SiC heating element
was instrumented with platinum/rhodium thermocouples attached to the front
and back heater faces. Pyrolysis experiments were conducted at 1600°C on
EPS, PMMA, and a copolymer containing 30% PMMA/70% EPS covering the density
range from 19.4 to 21.6 kg/m3 (1.21 to 1.35 pcf)– at a velocity of 5.1
cm/s.
A thermal gradient of approximately 800°C developed in the heater

as PMMA was pyrolized. The gradient could be reduced by either using a
higher thermal conductivity heater or reducing the heater thickness by
a factor of 10. Reducing the heater thickness was not feasible because
of machining problems with SiC.
Carbon was found to be deposited on the heater front face. Carbon

buildup ranged in thickness from less than 1 mm for copolymer patterns
to 3 mm for EPS patterns. The carbon layer insulated the heater, and caused
overheating and burnout of the front thermocouple. The conclusions reached
were that SiC is not a feasible heater material because of its low
conductivity and carbon deposition caused the heater to fail.
1.1.1.2 Heater Size and Shape
Experiments were also conducted to evaluate the effect of the heater

size and shape on foam pyrolysis. Heater assembly of the foam pyrolysis
apparatus is shown in Fig. 1. The residence time of the polymer in the
kinetic zone should depend on the size and shape of the heater. Two sets
of experiments were conducted using both concave and convex heaters. The
experiments were videotaped to study mass and heat transfer in the kinetic
zone.
The heater shape was not found to have an effect on power density
or specific degradation energy. Gas fraction for the concave heater was
about 15% higher than for the convex shape. Based on the visual observations
and analysis of the data, the higher gas fraction produced by the concave
heater was probably a result of liquid degradation as pyrolysis products
built up on the heater face outside the kinetic zone.
Heater shape did affect the escape of liquid pyrolysis products from
the kinetic zone. The convex shape facilitated flow around the heater in
comparison to the concave heater. The heater width did not have a
significant effect on the gas fraction.
It was concluded that the effect of heater geometry on the residence

time of the polymer in the kinetic zone was too small to observe. This
implies that the channel thickness and metal front shape is not critical
in a computational model for pattern replacement by molten aluminum.
- 11 -
(a)
(b)
Figure 1. Heater assembly of foam pyrolysis apparatus. (a) Front view,

(b) rear view. TCs = thermocouples.
1.1.1.3 Gas Generation Regimes
- 12 -
New data was obtained using the 1 cm wide heater in the temperature
range of 550-800°C. The degradation data in this range is important for
modeling Lost Foam because this is the range of temperatures used to make
aluminum castings. Three gas generation regimes were found:
(a) Pattern melting, characterized by small gas fraction (Fgas < 10%)
and occurring at heater temperatures below 590-620°C;
(b) Gas film formation, at Theater > 680-710°C with Fgas > 30%;
(c) Transition from melting to gas film formation with gas fractions
varying from 10% to 30%.
The transition points between the gas generation modes appeared as

discontinuities in the gas fraction curves.
1.1.1.4 Gas Generation Rates
A gas generation (vaporization) rate (Rv, measured in grams of gas

produced per second from a cm2 of degrading pattern surface), was introduced
to describe the volume of gas produced in the kinetic zone during pyrolysis.
Recession velocity had no effects on gas generation rate at temperatures
below 650°C. At 650°C, there was an inflection, followed by a rate,
dependent on the recession velocity. Higher Rv values and steeper slopes
were associated with higher recession velocities.
The gas generation rate at recession velocities of 1.78 cm/s and

3.33 cm/s leveled off at temperatures between 1050°C and 1150°C. This
was a result of a thick gas film between the heater and the pattern which
reduced the heat transfer rate.
The gas generation rate data were analyzed using an Arhenius

expression, and it was found that the gas generation rate at temperatures
in the range of 550-670°C followed a straight line for different velocities.
The pyrolysis activation energy was found to be 131 kJ/mole. This value
was about 30% below the value found in the value of 192.6 kJ/mole reported
in the literature for polystyrene. The discrepancy was attributed to the
difference in temperature between the polymer and the heater surface caused
by the insulating gas film.
The gas generation rate data deviated from the straight line above
about 640°C at Vrec=1.78 cm/s and above 670°C at Vrec=3.33 cm/s. The deviation
points correspond to the discontinuities in the gas fraction curves and
indicate a change in the gas generation regime–i.e., the development of
an insulating gas film between the heater and the pattern.
1.1.1.5 Temperature of Gaseous and Liquid Pyrolysis Products
Experiments were conducted to measure the temperature of the

pyrolysis products in the kinetic zone and estimate the gap between the
metal and the pattern. The temperature of the gas escaping the kinetic
zone at foam recession rates of 1.78-3.33 cm/s increased from 450°C to
730°C as the heater temperature was raised from 620 to 1050°C. The
- 13 -
temperature of the liquid fraction was approximately 315°C at a heater
temperature of 600°C and 446°C at a heater temperature of 1000°C.
The gas temperature measurements did not indicate any effect of the
recession velocity on gas temperature. The average gas temperature was
found to be approximately 380°C and 700°C at a heater temperature of 600°C
and 1050°C, respectively. The standard error of the data about the fit
was 58°C.
1.1.1.6 Resistance Pressure of Pattern Degradation
A load cell was used to measure the force between the foam bar and
the pyrolysis heater. The force data was used to calculate the pattern
degradation resistance pressure and put in a computer model to predict
metal filling velocity.
Pyrolysis experiments were conducted to develop pattern degradation

resistance data for EPS. Patterns were pyrolized using heater temperatures
between 550-870°C using recession rates of 1.93 cm/s and 3.52 cm/s. The
pressure data exhibited a hyperbolic response at pressures near 150 g/cm2
(2.13 psi) at 555°C, 27 g/cm2(0.38 psi)at 720°C, and 6 g/cm2 (0.09 psi)
at 850°C. The pressure values at higher recession rate of 3.52 cm/s at
heater temperatures of 650-800°C were 6-7 times higher than those at the
lower recession rate. The data were in good agreement with data obtained
on pattern recession rates observed in 319 aluminum castings.
1.1.1.7 Molecular Weight of Foam Pyrolysis Products
Samples of EPS liquid residue were collected and the molecular weight
(Mw) of the residue samples determined. It was found that residue molecular
weight was about 25% of the Mw of virgin EPS (75,000 vs. about 300,000).
The molecular weight was independent of the heater temperature and pattern
recession rate.
1.1.1.8 Effects of Molded and Cut Pattern Surfaces
Several experiments were conducted to determine the heater/foam

interface using patterns with molded and cut surfaces. EPS patterns were
saw cut and tested at recession velocities of 1.78 cm/s and 3.33 cm/s in
the temperature range of 600-1000°C. The results were compared with data
from seven molded EPS patterns having an average density of 25 kg/m3.
Patterns recession rates were 1 cm/s, 1.8 cm/s, and 3.4 cm/s, using heater
temperatures of 600°C and 780°C at each velocity.
The heater/pattern interface on the molded bars was smoother than

the interface produced in saw cut patterns. Fewer gas bubbles broke through
the molded surfaces compared to cut surfaces. A molded pattern surface
provides less effective area for gas escape compared to a saw cut pattern.
1.1.1.9 Effects of Foam Pattern Fusion
Experiments were conducted to determine if gases produced during
- 14 -
pattern pyrolysis could escape the kinetic zone through a pattern. EPS
patterns were pyrolized and the events video taped to observe gas flow
through the pattern. The heater temperature varied from 660°C to 735°C.
Patterns having a low degree of fusion degraded with an average

recession velocity of 2.8 cm/s, about 25% faster than the normally fusion
patterns. Analysis of the pyrolysis videos showed that holes appeared in
the pattern near the heater. The holes appeared to be caused by hot gaseous
degradation products escaping through the pattern ahead of the kinetic
zone. Fewer holes were observed during pyrolysis of normally fused patterns.
This data indicates that pattern permeability is an important factor
affection gas escape from the metal-pattern interface, and it must be
controlled to achieve consistent casting quality.
1.1.1.10 Degradation of Pattern Adhesives
Degradation data is needed on pattern adhesives to serve as the

baseline for the development of new materials. Two types of experiments
were conducted where glue joints were placed transverse to the test pattern
length. The joint thickness ranged from 0.17 to 0.42 mm and heater
temperatures of 600 and 700°C were used. The results are summarized as
follows:
• The movement of a glue joint into the hot zone caused a significant
decrease in the rate of pyrolysis. Glue joints with a thickness
of 0.16-0.28 mm, produced contact times of 0.2-0.33 s.
• Higher driving pressures and heater temperatures reduced the contact
time.
• Glue joints imposed an additional thermal load and reduced the heater
temperature by 30-50°C. The same event occurs in molds as metal
contacts glue joints.
The degradation properties of some protein and new hot melt adhesives
were also evaluated. The protein (gelatin) adhesives may be superior to
hot melt adhesives due to their lower working temperature (43-54°C) and
longer tack time.
A glue joint was put in EPS patterns transverse to the bar length.
The pyrolysis of two hot melt and two protein adhesives, designated as
HM-30-7176, HM-30-7177, P-500R-125, and P-600R-137, were examined. The
glue joints evaluated ranged in thickness from 0.16 to 0.43 mm.
The bars containing the glue joints were pyrolized under conditions
representing those encountered when pouring aluminum. The degradation
parameters measured were the contact time of heater with glue joints and
heater temperature change during contact.
The results are summarized as follows:
• Degradation of all hot melt glue joints occurred by melting and

vaporization and reduced the heater temperature about 33°C during
contact, which lasted about 0.29 s.
- 15 -
• No significant differences in degradation parameters were found
between different hot melt adhesives. Higher temperature drop was
produced by thicker joints as expected.
• Protein adhesives degraded by char formation. The char created some

thermal insulation and increased the contact time with the heater.
• Protein adhesives do not appear suitable for use at the present time.
Water solubility of modified proteins is another obstacle for the
use of these adhesives.
1.1.2 Examination of Pattern Replacement using Real Time X-Ray Technology
1.1.2.1 Real Time X-Ray Apparatus
The real-time x-ray system, used to examine events inside molds during
filling, is comprised of two X-Ray units. Overview of the X-Ray unit and
set up for metal pouring are shown in Fig. 2. The first unit consists of
a 320 kV tube having a source spot size of 0.8mm × 0.8mm and a 9” tri field
image intensifier with a Sony XC75CCD camera. The operation of this unit
is integrated by a microprocessor console.
The second unit consists of a 160 kV micro-focus X-Ray tube with

a focal spot size of 5 to 200 microns and an A-SI (amorphous silicon) digital
detector. The detector has an active area of 203 × 254mm and has a pixel
array that is 1997 × 2592 in the x and y axes and contains a total of 3.1
million pixels. Each pixel can provide 12 bit images or 4066 shades of
gray with a frame capture rate of 7/second. The X-Ray systems are located
in a room of steel-lead-steel construction. The x-ray control and data
acquisition systems are located in an adjacent room.
As the molten metal replaces foam patterns, X-Rays pass through the
flask and project onto a cesium-iodide (CsI) fluoroscopic screen which
produces visible images. The images are recorded using a 30 frame per second
camera connected to a VCR. Video images can then be transferred to a computer
controlled image processor for image processing, such as image enhancement
and color-coding.
The molding material used is usually silica sand with a grain fineness
number (GFN) of 55. Sand is rained into a flask around a pattern while
the flask is vibrated vertically or horizontally to achieve uniform
compaction. A pouring apparatus with remote control pours the liquid metal
inside the x-ray vault. The crucible is rotated by a motor-sprocket-chain
at rates of from 0.2 rpm to 10 rpm depending on the desired pouring rate.
Many of the castings poured were 6inch × 8inch × 0.3inch (152mm ×

203mm × 8mm) plates. The alloy poured was usually 356 Aluminum poured at
1450F. Triplicate castings were usually poured to ensure data
repeatability.
- 16 -
Lead
Room
X-Ray Tube
(a)
Image
Intensif
Pouring
cart
Lost
Foam
Mold
(b)
Fig. 2. Set-up for real time X-Ray observation of metal filling of lost
foam castings. (a) Overview of the real-time X-Ray facilities; (b) set
up for the pouring.
- 17 -
1.1.2.2 Metal Filling of Lost Foam Aluminum Castings
Experimental Matrix
Experiments were conducted to determine the effects of EPS pattern
density and degree of fusion (open porosity), coating permeability, and
pattern glue joints on pattern replacement and casting defect formation.
Most experiments were conducted on 8 mm thick plates. Pattern density maps
were obtained using a capacitive probe, an Industrial Analytics (IA) x-ray
technique, and Lixi, also an x-ray technique. Sequential maps of metal
front profiles made from x-rays were compared to pyrolysis traces on the
surface of the resulting castings.
Data Acquisition
Pattern property parameter - Pattern fusion: Two degrees of pattern
fusion were considered. One was referred to as “low fusion”, where beads
were in contact but not expanded sufficiently to eliminate voids between
beads, and the second was referred to as “normal fusion” where beads were
sufficiently fused to produce a smooth pattern surface.
Pattern property parameter - Foam density map: The patterns were

scanned at quarter inch intervals with a small diameter X-Ray beam and
the density of the pattern measured from the transmitted x-ray intensity.
From these maps, local densities and density gradients in patterns could
be defined.
Coating parameters - Dry coating weight on patterns and coating

permeability values were determined. Coating permeability was measured
using procedures developed by Littleton and published in 1997.
Metal front profiles - Metal front profile maps were made by printing
x-ray images made during filling at 6 frame intervals (0.2 second). These
maps allowed the metal front profiles to be examined without replaying
the video repeatedly. Changes in the metal front profile were clearly seen
in the maps made at 0.2 second intervals. Also, the instantaneous metal
fill velocities could be obtained from the maps. These velocities provide
accurate data for validating computer models.
Two characteristic parameters were developed from the maps of metal

front profiles: 1) Number of converging metal lines. 2) Average metal
filling velocity along the plate centerline. The average velocity along
the centerline of the plate was used to compare the effects of experimental
parameters on casting quality.
Pyrolysis traces and defects - The liquid pyrolysis products produced

during foam pattern replacement produce lines on casting surfaces. The
distribution of the lines and defects (folds and blisters) were recorded
using a digital camera and located on a coordinate grid. The number of
casting defects was taken as the characteristic parameter.
Coating Permeability and Pattern Fusion

The data indicates that coating permeability and degree of pattern
fusion(open porosity) have significant effects on the metal front shape,
and metal velocity into the mold, and defect formation. Metal replacement
- 18 -
of low fusion patterns was faster and more unstable than replacement of
normally fused patterns. Patterns having a low degree of fusion
consistently produced jagged metal fronts and fold defects, while normal
fusion patterns produced smooth (stable) metal fronts and much less fold
defects. Patterns having the same degree of fusion that were coated with
materials having a high permeability resulted in high metal velocities
and large amounts of liquid degradation products that produced defects
in castings. Some liquid pyrolysis products floated in the casting cavity
and others produced surface defects.
Foam Density
The nominal pattern density was not found to have any statistical
effect on the velocity for replacing patterns having a low degree of fusion.
High density areas in patterns of a low degree of fusion significantly
reduced metal velocities, produced merging metal fronts, and increased
casting defects. Effect of foam density gradient on metal filling and metal
filling on fold defect formation are shown in Fig. 3. However, high density
areas in normally fused patterns did not markedly affect metal velocity.
Coating Permeability
Coating permeability also had a significant effect on metal velocity
and the occurrence of defects. The coating must have proper permeability
in order to minimize converging metal fronts and the formation of bubbles
of liquid pattern pyrolysis products. A pre-coat sealer on the surface
of a normally fused pattern decreased the filling velocity by about 30%.
Glue Joint Effects

Section and perimeter glue joints were found to decrease the metal
velocity and increase the number of converging metal fronts. Perimeter
glue joints did not have any significant effect on the metal filling
velocity or the number of converging metal fronts.
Metal Filling and Casting Quality

Comparisons of metal front profiles with pyrolysis traces and casting
defects showed that converging metal fronts lines were related to pyrolysis
traces and fold defects. These results were consistent with the previous
studies that correlated fold defect occurrence to the pattern replacement
velocity.
Cracked Coatings
Effects of cracked coatings were determined by deliberately
scratching coated patterns. The scratches changed the metal front profile
from smooth to irregular and resulted in more casting defects. Cracks and
similar coating defects must be minimized.
Metal Front Profile

Metal front profiles were recorded using slices removed from low
and normally fused patterns. Metal front profiles obtained from normally
fused patterns were flat and produced the fewest casting defects. Metal
front profiles obtained using patterns with low degree of fusion were
concave and produced converging metal fronts, which tended to produce more
defects.
- 19 -
(a)
(b)
Fig.3. Quality of casting poured with low fusion pattern with 1.3 lb/ft3
density. (a) overlay of map of metal front profile on pattern density map;
(b) overlay of map of metal front profile on casting surface.
Pattern replacement with no pattern coating

Pattern replacement under these conditions resulted in high metal
velocities, irregular metal fronts, bubble formation, and gas holes. This
is the same type of fill pattern as when filling of patterns with high
permeability coatings.
- 20 -
Metal filling through hollow sprues
The average replacement velocities seen with patterns poured with
hollow sprues was higher than similar patterns poured with foam sprues.
1.1.2.3 Interior Quality of Lost Foam Aluminum Castings
The interior quality of the 6 inch x 8 inch UAB aluminum plate castings
was inspected using digital x-rays. Porosity bands were found in most
castings poured with normally fused patterns, except for some special
conditions. Patterns having a lower degree of fusion produced smaller bands
of defects. A reduced casting thickness was associated with pyrolysis
products entrapment between the coating and metal at converging metal
fronts.
Experiments were conducted where variations were made in plate

thickness, pouring temperature, and gating arrangement to determine causes
of the line of porosity in cast aluminum. Porosity bands were found in
plates with thicknesses of 4, 8, and 12mm, but the porosity was worse in
the thicker castings.
The effect of pattern pre-coating and glue joints on the occurrence

of porosity lines was examined. The porosity bands were reduced in severity
but still present in castings produced using low fusion patterns with a
pre-coat. The pre-coat reduced the metal velocity, generating less liquid
pyrolysis products. Glue joints in normally fused patterns also slowed
the metal fill rates and reduced the porosity band. Decreasing the amount
of glue similarly decreased the porosity.
Studies of the effect of coating permeability, thickness, and surface

condition on defect formation indicated that the metal front profile
changed from smooth to fingered as the coating permeability increased.
Smooth filling eliminated the porosity band.
The porosity defects were examined using a stereo microscope and

SEM, and most of the pores were found to be caused by gas. Some shrinkage
and hydrogen related pores were present, but most were caused by gas from
pattern pyrolysis.
The effect of top gating on porosity bands was determined using 8

and 24mm thick patterns having a density of 1.3pcf. Top gating produced
bands on one or both sides of test castings.
The porosity band appears to be the result of entrapped gas from

pattern pyrolysis products in the molten metal. The bubbles float toward
the top casting surface and agglomerate just below the solidified metal
along the top edge of the castings.
1.1.2.4 Effects of Pattern Tooling on Metal Filling
A series of patterns having dimensions of 10×200×180mm were produced

in two types of tools and the effect on defect formation was determined.
Both low and normally fused patterns were blown in both vented and ventless
tools. The permeability of each pattern was measured; each pattern was
- 21 -
coated, and then poured with aluminum alloy. Pattern replacement was viewed
using the x-ray system. Metal front profile maps were made and overlaid
on the permeability maps to determine the effects of the tooling on pattern
replacement and defect formation. Three patterns were examined under each
set of experimental conditions to ensure data repeatability.
Vented tools produced slow and smooth metal fronts when the patterns
were well fused, and fingered metal fronts when the patterns were not well
fused. Ventless tooling produced fingered metal fronts with both low and
high fused patterns.
Pattern permeability maps were then overlaid on metal front profile

maps. The permeability of patterns molded in vented and ventless tools
was found to be quite different. When metal replaced the pattern, the metal
followed the regions of higher permeability and produced erratic metal
filling behavior. The patterns produced in ventless tools generally had
higher density variations and more irregular filling.
Foam pyrolysis traces on casting surfaces were recorded as a measure

of surface quality. Pyrolysis traces on the back side of patterns were
more than on the fill gun side of the pattern. This was probably caused
by a non symmetrical permeability on the two sides of the patterns. The
back sides of patterns had higher permeability values compared to the fill
gun side. Castings poured with well fused patterns made in ventless tools
are cleaner than equivalently fused patterns molded in vented tools. This
suggests that the highly fused vent area in patterns made in vented tools
may be a site of liquid pyrolysis accumulation.
1.1.2.5 Metal Filling of Lost Foam Iron Castings
Several modifications were made to the x-ray vault, metal melting

and pouring equipment, and pattern coating procedures to examine pattern
replacement when pouring iron. The parameters examined included degree
of pattern fusion, pattern density, coating permeability and glue joints.
The UAB plate patterns were poured using bottom gating systems. Type K
thermocouples and pressure transducers were used to measure the iron
temperature and the amount of vacuum on the flask.
The degree of pattern fusion had a significant effect on metal filling.

Normally fused patterns filled smoothly. Patterns having a low degree
of fusion filled with "w" shaped metal fronts and lustrous carbon defects
formed at the converging metal fronts. Patterns with a low degree of fusion
produced higher fill velocities larger casting defects. The foam density
had a mixed effect with higher density foams producing lower metal
velocities and smaller defects. Pattern density had no significant effect
on metal velocity or defect area at low degrees of fusion.
Two coatings having permeabilities of 42.7 cm3/(cm2.s)and 37.0

cm /(cm2.s) were used to examine the effects on filling of iron castings,
3
but no statistical effect on metal velocity was found in this range. However,
high permeability coatings reduced the incidence of lustrous carbon defects.
Glue joints parallel to the metal filling direction distorted the metal
front, and defects were reduced in low fusion patterns but increased in
- 22 -
normally fused patterns having glue joints.
1.1.3 Pattern Degradation and Casting Defect Formation
1.1.3.1 Effects of Sprue Design
Twenty-four (24) flange castings were poured to determine effects

of metal velocity and analyze the defects produced. The sprue types
included an un-tapered hollow sprue, a tapered hollow sprue, a solid foam
sprue, and a "hollow" foam sprue. Castings were poured in duplicate with
a filter, with an aluminum plug, or with both a filter and a plug using
each type of sprue.
The pyrolysis-related defects in the castings were analyzed and a

data base on defects constructed which included the frequency of occurrence
and location within the casting. The defect colors varied, and may be
related to the composition and/or defect origin. As a consequence, the
defect colors were mapped and correlated to the casting parameters.
Selected defects were subjected to Auger and SEM analysis.
It was found that the metal velocities were slightly higher when
hollow sprues were used compared to solid foam sprues. Straight hollow
sprues produced the highest velocities. However, differences in metal
velocity were small, and the sprue type had a very small effect on casting
quality. Low permeability coatings were used so few blisters or pores
were found in the castings. However, folds were found in all castings,
and more folds were found in castings produced with foam sprues (either
solid or hollow) than in the castings produced with ceramic hollow sprues.
The largest number of folds was found in castings produced using hollow
foam sprues.
The smallest number of folds was found using hollow tapered ceramic
sprues, which produced fewer liquid pyrolysis products and less turbulence
during sprue filling. The effect of various filter/plug combinations on
fold formation was inconsistent. Folds tended to be concentrated in the
last half of the casting filled with metal, while blisters normally formed
in the first half of the casting to fill.
1.1.3.2 Effect of Process Parameters on Defect Formation
A matrix of 24 flange castings was poured using two foam densities

(1.1 pcf and 1.6 pcf), high and low degrees of fusion, and high and low
coating permeability. Castings were poured at 800°C using 319 aluminum.
Coating permeability ranged from 5 to 100, as determined by the GM

Screen Test procedure. In general, the 1.6 pcf, patterns having a low degree
of fusion filled at higher rates irrespective of the coating permeability.
Un-diluted low permeability coatings produced the highest metal
velocities.
Defect analyses showed that the low permeability coatings produced

fewer defects. The castings had lower blister areas and fewer fold defects.
- 23 -
Auger analyses were made on selected defects, and high carbon
concentrations were found in areas with gold/brown discolorations. Less
carbon was present on silver-colored surfaces. The darkness of the surface
may depend on the amount of carbon present, i.e. darker brown or gold
surfaces have thicker carbon layers compared to the lighter colored areas.
An Auger analysis made on a “brown smear” on a casting had a thick carbon
layer. A sample of low permeability, un-diluted coating was examined using
an SEM both before and after pouring a casting, but no significant
difference in the structure could be found.
1.1.3.3 Effect of Foam Materials (T180 and T185)
T180 and T185 bead stock are identical EPS beads in molecular weight
and size, but T185 “no fold” beads contain an additive to reduce the
molecular weight of EPS pyrolysis products. A box pattern was used, and
the effects of four factors including foam density (1.3 pcf and 1.6 pcf),
degree of fusion (high and low), bead type (T180 and T185), and coating
type (low perm/low liquid absorption and low perm/high liquid absorption)
were examined.
The first task examined the effect of the sprue type. A 2×2 full
factorial design was used to examine sprue and pattern density effects.
The sprues were tapered ceramic, and the pattern densities were 1.1 and
1.86 pcf.
It was found that metal velocity did not significantly change with
foam density when using the small tapered sprue; however, the pattern
density did affect the metal velocity when the large tapered sprue was
used, probably because the large sprue could be filled faster and kept
full during pouring. Small tapered ceramic sprues were used in subsequent
experiments.
Two box patterns were attached to a common sprue and poured with
356 aluminum at 760±20°C. Ninety-six (96) castings were produced and
evaluated. It was found that the use of T185 beads reduced the metal velocity
and the number and size of casting defects formed. The use of a low
permeability/high liquid absorption coating also decreased the metal
velocity. Both normally fused and high density patterns decreased the metal
velocity.
The defect size was smaller when using normally fused patterns made
from T185 beads and applying a low permeability/high liquid absorption
coating. As the pattern density increased, the size of the defects decreased
in castings made with low fusion patterns, but increased with density when
using normal fusion patterns. Regardless of the type of coating used,
patterns produced using the T185 beads produced castings with smaller
defects.
Using a statistical analysis, the parameters having the greatest

effect were the bead type (36.0%) and degree of fusion (35.3%). The use
of T185 decreased the total defect area from 30 mm^2 to about 12 mm^2.
The highly fused pattern also decreased the total casting defect area.
- 24 -
1.1.3.4 Effect of Alloy Composition
The current model describing pyrolysis defect formation suggests

that the freezing range of the alloy contributes to the types of defects
produced. This study was conducted using a Al-Si alloys containing 1%,
6%, 11%, and 12.3%Si. Pure aluminum and an Al-50%Si master alloy were used
as charge materials. All castings were poured at approximately 125°C above
the respective alloy liquidus temperatures. Two flange castings (see Fig.
4) of each composition were produced, velocities measured, and defects
examined.
Figure 4 Top view of the flange pattern with a cross-sectional view

on the right.
The 1% Si alloy having the longest freezing range produced the highest
metal velocity. Higher silicon concentrations produced similar but lower
fill velocities. The 6%Si alloy had a lower liquidus as well as a shorter
freezing range. The 11%Si and 12.3%Si alloys were near-eutectic alloys
with the lowest liquidus and shortest freezing range.
Folds, blisters, and pores were found in most castings. Folds and
pores were concentrated opposite the ingate where the metal temperature
was near the liquidus temperature, and blisters were concentrated near
the gate. The silicon content did not have a significant effect on fold
defect severity. However, blisters were severe in the 1%Si alloy and were
near zero at the highest silicon concentration.
No internal pores were observed in the 1%Si and 6%Si alloys, but
a large number of pores were present in the near-eutectic alloys. These
results do suggest that the solidification sequence affects the nature
of the pyrolysis defect.
- 25 -
A designed experiment was also conducted to determine if there were
different kinds of defects produced by different casting alloys including
206, 319 and 356. Flange patterns (see Fig. 4) were used with densities
of 1.3 and 1.6 pcf and prepared with T180 and T185 (no fold) beads.
The results again confirmed that castings produced using T185

patterns have significantly fewer casting defects. Castings poured using
alloy 206 exhibited significantly more casting defects than those poured
with 319 and 356 alloys. The high density T185 patterns produced a higher
metal fill velocities compared to those produced with T180. This indicates
that the additive in T185 beads reduces the molecular weight of the liquid
pyrolysis products.
1.1.3.5 Effect of Superheat and Hydrogen Content
Experiments were conducted to study the effects of hydrogen content

and superheat of the melt on defect formation in castings. Eighteen (18)
flange patterns described above were cast at three superheat values and
three different contents. The superheats were 75°C, 150°C, and 195°C. The
hydrogen contents ranged between 2 and 20 percent.
A good correlation was found between superheat and defect formation

but not with the hydrogen concentration. As the superheat increased, the
metal velocities increased by about 0.25 cm/s, and this increased both
the number of folds and internal pores. The best results were obtained
with a superheat of 75°C and 150°C.
Higher superheat values also increased the number of blisters. This

was probably a result of the fact that more superheat allowed more of the
pyrolysis products to float to the surface and be entrapped just beneath
the metal skin. The folds, blisters, and internal porosity occurred
randomly in the casting. Hydrogen pores were also randomly distributed
in the castings irrespective of the hydrogen concentration.
1.1.3.6 Defect Database
Samples of different defects were collected and analyzed by Auger

(AES) and Scanning Electron (SEM) and the results are put in the defect
data base. One of the databases is demonstrated in Fig. 5. Auger analysis
indicated that black pores were coated with a thick layer of carbon and
contained very little aluminum oxide. The carbon layer thickness may
increase in larger pores. SEM analysis also suggested the presence of
carbonized pyrolysis products having a smooth skin.
- 26 -
100
90
Weight Percent Carbon (%) Brown surface
80 discoloration
70
60
50
40
30
20
10
0
0 20 40 60 80 100 120 140 160 180
Sputter Tim e (m in)
4 - 4 Cast n
i g Sur f ace St an dar d N/ A B2 - 6 Fr act ur e Sur f ace St an dar d N/ A 11A I n t er n alPor e Black 2 3 B I n t er n al Por e Black 2 3 A I n t er n alPor e Black 12 C I n t er n alPor e Black
4 D I n t er n al Por e Black 8 A Blsi t er Gold 9 A Blsi t er Sivl er 2 - B Blsi t er Sivl er B2 - 6 M si c Fold Sivl er 4 - 4 M si c Fold Sivl er
10 A M si c Fold Sivl er 9 B M si c Fold Sivl er 4 B T op En d Fold Sivl er 2 B T op En d Fold Sivl er 3 C T op En d Fold Sivl er 3 A T op En d Fold Sivl er
2 A Bot t om En d Fold Sivl er 3 B Bot t om En d Fold Sivl er 4 A Fold Sivl er 10 H T / C Fold Gold 10 E T op En d Fold Gold 10 D Bot t om En d Fold Sivl er
1A Blsi t er Gold 2 A Blsi t er Gold 2 B Blsi t er Sivl er 2 C Blsi t er Gold 2 D Blsi t er Gold 3 D Blsi t er Gold
1B En d Fold Gold 1C En d Fold Gold 1D En d Fold Gold 2 G En d Fold Gold 3 A En d Fold Sivl er 4 A En d Fold Sivl er
5 A En d Fold Sivl er 5 B En d Fold Sivl er 5 C En d Fold Sivl er 1E M si c Fold Sivl er 1F M si c Fold Gold 3 B M si c Fold Sivl er
3 C M si c Fold Sivl er 4 B M si c Fold Sivl er 2 A Blsi t er Gold 1A I n t er n al Por e Sivl er 1- 3 M si c Fold Light Br own 5 A Sur f ace Discolor at o
i n Br own
Fig. 5. Comprehensive AES depth profile showing weight percent carbon

versus sputter time.
Folds exhibited wrinkled surfaces appearance under both optical and

SEM microscopes. Some of the folds had a gold color, and Auger analysis
found that the fold contained a thick layer of carbon. In “silver” colored
folds, the carbon layer was thinner, and the carbon content appeared to
be lower. The carbon layer suggested that these folds may be composed
primarily of aluminum oxide, with only adventitious carbon present. An
oxide layer was found beneath the carbon layer on both types of folds.
1.2 Computational Model
The objective of this work was to develop a fill and solidification

code that accurately described the defect forming events of metal/pattern
exchange as described in 1.1. The initial plan was to develop the algorithms
based on parameters controlling the metal/pattern replacement and then
put these algorithms into a commercial code. A decision was made, however,
to simplify the modeling by using experimental data to describe the
metal/pattern replacement using data from the Pattern Pyrolysis Apparatus.
The use of the experimental data significantly shortens the computation
time. The x-ray system described in 1.1.2.1 was used to verify the code
predictions.
1.2.1 Existing Codes
Flow-3D and ProCast are the two leading commercially available

programs for simulating Lost Foam casting. ProCast assumes the foam is
- 27 -
converted to gas at the metal-foam interface and recently incorporated
gas escape through the coating and sand. Flow-3D considers transient liquid
formation but does not provide for gas escape or pressure in the pyrolysis
zone. Flow 3D does consider heat transfer at the liquid metal surface and
tracks the metal free surface.
The features of flow and fill have been examined in a study of a

V6 engine block. The results predicted areas of defect formation, but some
adjustments of the heat transfer coefficient were necessary to provide
metal velocities that matched experimental values.
1.2.2 Development of Physical Model
Previous research has shown that patterns are degraded into gas and
liquid components in a kinetic zone between the metal front and the receding
pattern. The air in the pattern and the pyrolysis gases were thought to
exit the casting cavity through the coating in the kinetic zone. The
remaining liquid components were thought to flow into the metal/coating
interface by a combination of wetting, wicking and pressure driven flow.
These liquid components were heated by the metal and converted to gas which
passed into the sand after the metal front had passed.
Burke, Gustafson and Zhao proposed an alternate mechanism where the

liquid pyrolysis products trapped between the metal and coating formed
liquid carbonaceous islands that degrade to gas which flows through the
coating. Depending on local temperatures and pressures, the gas could break
the metal surface tension and enter the metal to produce porosity and folds.
Heat transfer conditions at the metal/coating interface favor this
mechanism. The liquid does not initially wet the coating, as previously
assumed, but instead slowly degrades into gas in contact with the hot metal.
Polystyrene vaporization creates an insulating gas film around the pools
which keeps the polymer at a relatively low temperature (180-400°C). The
area covered by the pools decreases with time (or distance from the metal
front) as the polymer degrades.
The vaporized liquid either passes through the coating or enters

the metal, depending on local pressures. EPS degradation products and air
from the pattern escape from regions between the metal and the coating
through areas not covered by the liquid pools of pyrolysis products.
The proposed escape path for gases was challenged since this path
did not accommodate pattern permeability. Patterns segments were pushed
against heated blocks at typical aluminum temperatures and then quenched
in water. A surface layer of liquid polystyrene with multiple holes was
found on the degraded surface, followed by a layer of collapsed beads and
then the virgin pattern beads. These observations indicate that the gap
between the metal front and liquid layer is functioning as a heat transfer
gap, and the gap behind the liquid layer is functioning as a gas exit path.
The holes in the liquid layer allow gas access to the pattern. These results
have been incorporated into the physical model of metal/pattern exchange.
The transfer of the gaseous pyrolysis products from the casting cavity
during metal filling was examined using a screen window on the wall of
- 28 -
a casting flask. Considerable amounts of gas were visible at and behind
the advancing metal front. The movement of liquid pyrolysis products was
examined by pouring onto the foam patterns. Bubbles containing gaseous
and/or liquid pyrolysis products floated through the metal and burst at
the top. These observations indicate that entrapment of gaseous and liquid
pyrolysis products occurs during and after metal filling occurs.
1.2.3 Development of Algorithms
Data was needed on the size of the kinetic zone (KZ) between molten
metal and receding patterns, the temperature distribution in this zone,
and the viscosity of gaseous and liquid degradation products. These data
cannot be directly measured so a heat and mass transfer model was developed.
The model was based on principles of mass/heat transfer, lubrication
theories, experimental pattern degradation data, and video recordings of
pyrolysis experiments.
The KZ was assumed to have a uniform thickness and to consist of

a mixture of liquid and gaseous degradation products with average and
uniform viscosity, heat capacity, and thermal conductivity. The
temperature in the KZ was assumed to vary from the heater temperature to
the EPS “melting” temperature (150°C). Heat transfer in the KZ was assumed
to occur by conduction, convection, and radiation. The model used data
including heater power density, gas fraction, and resistance pressure.
The KZ parameters were calculated for aluminum casting conditions

using degradation data for 25 kg/m3 EPS developed for the range of 600-790°C
at foam recession rates of 1, 3, and 4.5 cm/s. At recession rates of 3
cm/s and 4.5 cm/s, the KZ thickness did not vary with the heater temperature
and was estimated at 0.06 mm and 0.04 mm, respectively.
At a recession rate of 1 cm/s, the predicted KZ thickness increased

from 0.16 to 0.2 mm as the heater temperature increased from 600°C to 790°C.
The average KZ temperature did not depend on the recession rate and was
estimated to be between 335°C and 405°C for a heater temperature range
of 600-790°C.
The average density of the mixture of gaseous and liquid pyrolysis

products in the KZ was calculated using the ideal gas law and the rule
of mixtures. In the heater temperature range of 600-790°C, the KZ density
was estimated at 12-10.7 kg/m3 at a recession velocity of 1 cm/s, 15.9-11.2
kg/m3 at a recession velocity of 3 cm/s, and 23.5-11.9 kg/m3 at a recession
velocity of 4.5 cm/s. The estimated viscosity of the mixture of the liquid
and gaseous degradation products decreased exponentially from about 0.02
to 0.0013 poise as the average KZ temperature increased from 330°C to 405°C.
The recession velocity was not affected by the viscosity.
The model developed through data must be extended to describe

heat/mass transfer in the KZ under casting conditions. The main issues
include: (a) the effect of pattern properties and geometry on the metal
front shape and (b) the escape of pyrolysis products through the coating.
1.2.4 Computational Model
- 29 -
Flow Science incorporated the physical model to predict defect sites
in a box pattern and a marine engine block. The simulations used UAB
velocity data to calculate the heat transfer coefficients, and predicted
defect locations agreed well with the experimentally determined locations.
Although progress is being made, it is still necessary to have correct
heat transfer coefficients as a function of metal velocity.
UES Software has updated the Pro-CAST Lost Foam module to include
the effects of gas pressure produced by pattern degradation; however, the
values of permeability are artificially high. Although the improvement
is significant, additional development remains to track liquid prolysis
products.
FLOW-3D® has demonstrated their capability to locate probable defects

in castings, and provide insight into the origin of defects (see Fig. 6).
This information can then be used to make process improvements. For example,
the computational model suggested that a certain gate modification would
reduce defect formation. This was indeed found to be the case.
Figure 6. Comparison between experimentally observed defects (left),

computational predictions in bottom (middle) and slightly deeper into
bottom surface (right). Box pattern was used for this work.
New capabilities were incorporated into FLOW-3D® to enable users

to describe pattern density gradients. The utility of the predictions was
evaluated using instrumented castings and x-ray examinations of both simple
and complex castings.
The model predicted reduced porosity in a new engine block design,

and correctly predicted reduced porosity with a gate change. It was also
found that the metal temperature distribution at the end of the mold filling
was not uniform, and this had an important effect on shrinkage. The FLOW-3D®
- 30 -
shrinkage model predicted the general location of the porosity, but the
location was biased slightly toward the cylinder walls.
1.2.5 Validation with Real Time X-Ray
Real time x-ray results indicated that local pattern permeability

had an important effect on the shape and velocity of the metal front. High
pattern permeability results in higher metal front velocities. Pattern
permeability differences were simulated by assigning pattern densities
values to different pattern regions based on permeability measurements.
The working hypothesis was that higher permeability regions would have
higher foam replacement rates. By segmenting patterns into regions with
density inversely related to permeability, the dynamics of pattern
replacement were simulated without multi-phase modeling.
Simulation results gave a much better correlation with metal front

shapes when permeability was incorporated. Moreover, when an inverse
exponential functionality was used between density and permeability, the
metal front shape was nearly identical to that observed in x-ray
observations. The correlations have prompted the use of this modeling
technique to address production casting issues.
2. Casting Dimensional Accuracy
The objective of this task was to identify sources of casting

distortion. These studies included dimensional studies on blown patterns,
glued pattern assemblies, coated and dried assemblies, and distortion
caused by the compaction and pouring processes. Control measures were
implemented in the foundries.
2.1 Sources of casting Distortion
2.1.1 Case Study 1: long pipe
An opportunity arose to study distortion of eight foot tall flasks.

These flasks amplify pattern distortion because of the pattern length.
Pipe extensions ranging from 3 to 6 feet long are cast in a four-on
configuration. Molds are prepared on a vertical compactor. Some bending
and length reduction were being experienced. Experimental work identified
bending to a result of non-uniform sand filling in and around the pipe
extensions. The sand force on the outside of the pipe pattern caused the
pattern to bend towards the unsupported interior. Since the pattern was
unsupported at the top, the cumulative effect was a distorted pattern and
casting. The distortion was minimized by masking the sand rain gate to
produce uniform filling in and around the pattern. Additional dimensional
accuracy was obtained by suspending the cluster from the sand hopper.
2.1.2 Case Study 2: Valve body
The distortion associated with the production of a valve was also

conducted by applying strain gages to patterns, calibrating the strain
gages, and measuring the deformation during sand filling and compaction.
The experimental set up is shown in Fig.7, and a typical pattern distortion
- 31 -
curve is demonstrated in Fig.8. The data indicated that sand was being
compacted around the valve before the interior cavities were filled. This
caused lateral movement of the pattern walls toward the unsupported
interior. The distortion could be significantly reduced by simply rotating
the valves 90 degrees in the flask.
Strain gage for temp

compensation
Working strain
gages
To Data
Acquisition system
Bridge completion
resistor
Resistor for bridge
balance adjustment
10V DC
(a)
clam p 1 clam p 2
flask
ch 1 ch 4
ch 6 end 1 en d 2 ch 5
ch 2 ch 3
ch 7 ch 1 0
V ib ratio n
d irectio n en d 3 en d 4 ch 1 1
ch1 2
D ow n
S p ru e ch 8 ch 9
(b)
Fig.7. Experiment Set-up for 6in Valve. (a) Electric circuit of the testing
apparatus. There are 12 channels in the apparatus. Each channel has the
same circuit. (b) Location of the strain gages.
- 32 -
7_END3_L 8_END3_R 9_END4_L 10END4_R 11EN4_LO 12EN3_LO
300
250 Cycle start 2nd Table Cycle

started
200 1st Table Cycle
Sand filling
150 stopped(16s) finished (88s)
100
Pattern Deflection (0.001 in.)
50
0
-10 10 30 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 350 370
-50
60 sec @ 2800 rpm
-100 0% open sand
-150
0 sec rain
-200
-250
-300
-350
-400
10 sec @ 3000 rpm
-450 8 sec @ 2200 rpm 100% open sand
50% sand open 2 sec @ 2800 rpm
-500 2 sec rain
8 sec rain 0% sand open
0 sec rain Cycle Tim e(s)
4 sec @1800 rpm 4 sec @ 2000 rpm
50% open sand 50% open sand
4 sec rain 4 sec rain
Fig.8. Foam pattern dimensional change during sand filling and compaction.
2.1.3 Sources of casting distortion
Table acceleration and sand angularity were two parameters having

a significant effect on pattern distortion during flask filling and
compaction. Higher table accelerations and more angular sands produced
more distortion with constant table accelerations.
Sand thermal expansion has also been identified as an important factor.

Silica sand, which has a phase change around 1000°F, expands more than
other sands such as olivine, chromite, and synthetic mullite. Synthetic
mullite has a low coefficient of thermal expansion and rounded grains that
flow easier at low table accelerations. Several foundries are now using
synthetic mullite to improve casting precision.
3. Pattern Production
The objective of this effort was to understand parameters that control

foam pattern dimensional stability including the effects of pre-expansion,
bead conditioning, blowing parameters, steaming, and pattern aging effects.
A data base containing properties of polymeric materials currently used
in Lost Foam patterns was also developed. These properties are required
for computational codes to accurately describe the metal/ pattern
replacement process and predict defect formation. Another part of this
work was to understand the role of vented and ventless pattern tooling
on the bead filling and steaming of patterns.
3.1 Pattern Bead Properties
3.1.1 Properties of Raw Beads
- 33 -
Research during an earlier phase determined the specific heat and
degradation energy of Polystyrene (PS), Polymethylmethacralate (PMMA)
CoPolymer (a blend of PS and PMMA)using Thermogravimetric Analysis (TGA),
and Differential Scanning Calorimetry (DSC). The differences in properties
were small, and no further efforts with DSC or TGA are anticipated because
the pyrolysis properties are better measured using procedures described
previously in section 1.1, pattern material degradation properties
database.
3.1.2 Pattern Density Measurement with Dielectric Sensors
A new approach to nondestructive pattern density measurements, based

on a dielectric constant sensor, was evaluated. The dielectric properties
of foam samples were determined after the samples were soaked in olive
oil to fill the open pores. The bulk density and open porosity was then
measured using a buoyancy technique.
A linear relation was found between the dielectric constant, the

dielectric values from oil saturated samples and the open porosity values.
Both correlations agreed with a theoretical model. In-plant measurements
on patterns, as they were removed from the molding machine, indicated that
this method is an effective on-line process control tool. Development
work is needed in more plants to confirm its utility.
3.1.3 Pattern permeability
3.1.3.1 Apparatus
A permeability apparatus (see Fig.9) has been developed to measure

pattern permeability. The degree of fusion, which is reflected by
permeability measurements, is a major factor affecting casting defect
formation. This permeability technique uses a vacuum source with a suction
cup applied to patterns. Pattern permeability can vary by a factor of 100
depending on the degree of fusion, and the pattern permeability is as high
or higher than coatings used to make aluminum castings. A ‘threshold’ value
of the steaming pressure has been found that produces a fused pattern
surface with low permeability. This discovery led to a better understanding
of the escape path for gas during pattern decomposition. A portable version
of this apparatus has been developed for use at or near pattern blowing
machines. The repeatability of this apparatus is demonstrated in Fig.10.
The Pattern Permeability Apparatus is currently being evaluated in a
sponsor foundry to improve pattern consistency and casting quality.
3.1.3.2 Pattern Thickness and Degree of Fusion
A computational model of gas flow into the suction cup was developed
to understand the effects of pattern thickness and degree of fusion.
Permeability constants for low and normally fused patterns were used in
the calculations. Air compressibility and inertial flow effects were
neglected for simplicity. Air flow was modeled using uncoated patterns
with thicknesses of 0.4 cm, 0.8 cm, 1.6 cm, and 2.4 cm. The suction cup
gaskets dimensions were 1.27 cm ID and 0.48 cm wide. The vacuum level was
- 2 psi. Uniform and non-uniform parabolic pattern permeability variations
- 34 -
through the plate thickness were considered. Patterns with a non-uniform
permeability were described by the ratio of the pattern center to surface
permeability constants. Patterns with permeability constant ratios of 2,
5, and 10 were modeled. Permeability constants (Bo) of 10-8 and 10-9 cm2
were used in calculations.
Pattern thickness had a significant effect on the air flow rates,

and flow rates through the cup decreased as the plate thickness increased.
When the permeability was uniform with a value of 10-8 the normalized flow
rates decreased from 28.5 cm3/(cm2.s) to 19 cm3/(cm2.s) as plate thickness
increased from 0.4 cm to 2.4 cm. When patterns had the same thickness but
a non-uniform permeability, the flow rates decreased from 24.2-27.2 cm3/cm2s
to 10.3-15 cm3/cm2s, depending on the center-to-surface permeability ratio.
Fig.9. Apparatus for Measuring Pattern Permeability
- 35 -
S8 S8
S7 S7
S6 S6 7.00-8.00
7.00-8.00 6.00-7.00
6.00-7.00 5.00-6.00
S5 S5
5.00-6.00 4.00-5.00
4.00-5.00 3.00-4.00
S4 3.00-4.00 S4 2.00-3.00
2.00-3.00 1.00-2.00
1.00-2.00 0.00-1.00
S3 0.00-1.00 S3
S2 S2
S1 S1
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
Fig.10. Repeatability of foam pattern permeability.
Air flow through the rear of the plate accounted for about 90% of the total
flow through the cup in plates having a thickness of 0.4 cm. The fraction
of air coming through the rear side of the plate decreased to 26-40% as
the plate thickness increased to 2.4 cm. Most of the flow occurred within
2-3 mm near the boundary of the gaskets on the front pattern face.
The measured air flow rates through patterns having different thicknesses
cannot be compared. The utility of this apparatus will be realized by
comparing air flow rates at specific locations on production patterns.
3.1.3.3 Pattern Permeability, Metal Filling and Casting Quality
Fig. 11. Overlay of metal filling profile and foam pattern permeability showing
effect of pattern permeability on metal filling behavior. The interval of the
metal filling profile is 0.4 second, and the unit of the pattern permeability
is: cm3/(cm2.s)
- 36 -
Real time x-ray results have shown that pattern permeability
differences have a dramatic effect on the shape and velocity of the metal
front during pattern replacement (see Fig. 11). Locally high permeability
values increase metal front velocities which results in multiple metal
fronts that converge to produce carbonaceous fold defects.
3.1.3.4 Pattern Permeability Threshold
A threshold gas flow rate that causes melt front instability was
defined to develop permeability acceptance criteria. Metal filling
profiles from different patterns were overlaid, and the minimum
permeability value affecting filling behavior was found to be about
1.0-cc/sec-cm2 in 10mm thick plates.
Work was continued on plates having different thickness values and

on commercial patterns. High permeability (low fusion) results in erratic
metal fronts, higher metal velocities, and a higher probability of folds
forming. Local pattern air flow rates above 0.5 cc/sec.-cm2 lead to
perturbations in metal front shapes that could produce folds at merging
metal fronts. The uniformity of local air flow rates may be more important
than the absolute values.
3.1.4 Computational Models for Bead Filling and Steaming
Arena-flow LLC., took the lead in this task. Arena-flow developed

software, Arena-flow-EPS, to simulate the bead blowing process and assist
tool design and blowing parameter selection.
The model was improved and tested at three sites: UAB, Arena-flow
LLC., and General Motors. Arena-flow was responsible for developing
physical and mathematical models of bead packing and local density
gradients. Twenty-five (25) foam patterns were selected and maps of the
permeability were made. These the patterns were supplied to Walford
Technologies for density mapping using an X-Ray scan technique. These data
were given to Arena-flow to further develop and validate blowing and
steaming models.
When the improvements were incorporated, Arena-flow-EPS was
successfully used for production application. A loosely packed low density
area in the combustion chamber area of cylinder head pattern was predicted.
Based on the modeling results, the process parameters were adjusted, and
the defect area in the foam pattern was eliminated. However, modeling failed
to simulate two of the most important pattern properties – pattern
permeability and density gradients.
3.2 Pattern Tooling
3.2.1 Pattern Cooling and Density
It has been found that the wall thickness of a pattern tool has little
effect on the rate of pattern cooling. It was originally thought that
density gradients within a pattern were due to uneven cooling across a
section, and that changing the tool wall thickness could eliminate the
gradients. Cooling relies on conduction, and the conductivity of EPS is
- 37 -
four orders of magnitude lower than the aluminum mold. Thus, no matter
how quickly the mold is cooled, the pattern can only cool at a rate permitted
by conduction within the pattern. Computer models incorporating an aluminum
mold and solid foam patterns have proven this to be true. Water vapor in
the pattern is the main heat conductor within a pattern.
Additional studies were made to simulate cooling during pattern

molding. The open porosity and water content were varied to determine
the effects on cooling rates. The resulting data were inconsistent with
experimental data from molding experiments. The models did show, however,
that patterns with higher open porosity cooled quicker because of the
higher thermal conductivity below 100° C. This was attributed to more
entrapped water in these foams.
3.2.2 Tooling Types and Permeability
Vented and ventless tools provide very different bead filling and
steam boundary conditions. A series of 10mm thick plate patterns was made
with both low and normal degrees of fusion and blown in both vented and
ventless tools. The two types of tools produces significant differences
in metal velocities, metal front shape and defect occurrence.
Patterns produced in ventless tools have higher permeability compared

to those molded in vented tools. Lower permeability gradients, i.e. more
uniform patterns, were observed in patterns produced in vented tools, and
the back sides of the patterns had relatively uniform and lower permeability
values. Higher permeability gradients were found in patterns molded in
ventless tools.
The distribution of high permeability regions was compared. Most

of the high permeability regions were repeatable from pattern to pattern.
However, the high permeability islands on the front and back faces of the
patterns prepared in vented tools were not symmetrical. The patterns blown
in ventless tools were symmetrical on both pattern surfaces. This indicates
that vents change the bead filling and steaming behavior.
The distribution of high permeability regions in patterns having

high and low degrees of fusion molded by the same tool was different. Only
one high permeability region appeared on both of the high and low fusion
patterns, while many others only show up either in low fusion or in high
fusion foam patterns produced in vented tools. One explanation might be
that longer or higher pressure steaming not only cured the high permeability
area, but also produced new high permeability regions by moving or
collapsing the beads. However, in ventless tools, most of the high
permeability regions seen on low fusion patterns are also seen on high
fusion patterns. This could be explained by: 1) The degree of fusion in
“high” fusion patterns is not as high as expected; 2) Bead filling may
be basic cause of the permeability islands or; 3) Bead movement is limited
during steaming. Additional study is required to understand the controlling
mechanism.
3.2.3 Filling and Bead Arrangement
- 38 -
The working hypothesis about the location of density gradients is
that the gradients develop when the beads are blown by the fill guns. Beads
in a pyramidal stacking arrangement with a density of 1.4 pcf can
theoretically have a density as low as 1.0 pcf if stacked in a simple cubic
arrangement. If the stacking arrangement could be controlled, the density
gradients might be also controlled.
It had previously been thought that the prepuff density and the bead
filling process controlled the pattern density. However, this is only true
if the bead packing arrangement is consistent. A literature review showed
that inconsistent bead packing could produce areas with dramatically
different densities and void fractions (open porosity/permeability). More
studies is required to understand bead filling and arrangement during
pattern molding.
3.2.4 Pattern Molding Parameters
3.2.4.1 Experimental approach
A full factorial matrix was conducted including: pre-expanded bead

density, pre-expanded bead volatile content, steaming time, steaming
pressure, vacuum assisted cooling, and bead filling pressure. The residual
volatiles consist of pentane and moisture at the time of molding. The
volatile content was controlled by letting the beads age from 24 to 72
hours. The pre-expanded bead density was changed in the pre-expander by
varying the steaming time.
Flange tooling was instrumented with thermocouples and pressure

transducers to record events in the molding cycle. The temperature of the
tool wall, center of the pattern cross-section, and the surface of the
pattern was recorded. The pressures inside the steam chest and inside the
pattern cavity were also recorded.
The degree of fusion and pattern density were measured using a low
power x-ray absorption technique by Industrial Analytics Corporation. The
fusion measurement produces a qualitative number that is a function of
the pattern surface and pattern shape. The open porosity, density, and
density gradient of patterns were measured using the olive oil absorption
technique.
3.2.4.2 Degree of pattern fusion
The fusion index varies as a function of the pre-expanded bead density

and the bead volatile content, with the volatile content having the greatest
effect. The cross-sectional density gradient was affected by the volatiles,
pre-expanded bead density, and the location in the pattern. The pattern
volatile content also had the largest effect on the density gradient.
The fusion index and the cross-sectional density gradient correlates

with higher values associated with higher cross-sectional density
gradients. Open porosity was related only to the fill gun location.
- 39 -
Higher bead densities and/or bead volatile contents increased the
fusion index and the cross-sectional density gradient. The degree of
pattern fusion and the cross-sectional density gradient, based on x-ray
measurements, were especially sensitive to residual volatiles in
pre-expanded beads. The fact that the fusion index and density gradient
were both affected by volatile content indicates that the driving forces
behind fusion and density gradients could be the same.
3.2.4.3 Open porosity
Patterns that were steamed with the highest temperatures and times
had low values of open porosity and did not appear to be substantially
affected by other molding variables. Patterns that were steamed at lower
temperatures and times had higher values of open porosity and were affected
by other molding variables. Open porosity varied with the time that the
pattern was above the Glass Transition Temperature (100°C). The open
porosity decreased with time until after about 25 seconds. Some increase
then occurred, perhaps because of bead collapse.
SEM analysis was made on patterns suspected to have typical open

porosity and open porosity caused by bead collapse. The patterns suspected
to have bead collapse had many holes in the bead cell skin. These holes
would cause the measured open porosity to be higher because of olive oil
penetration. By comparison, the remaining patterns had very few holes in
the beads.
The cross-sectional density gradient increased as the time above

100°C increased and as the steam pressure increased. The gradient was most
sensitive to the molding variables at intermediate power levels.
3.2.4.4 Density Gradient
Cross-sectional density gradients ranged from 1.0 (no gradient) to

1.2, a substantial gradient. The gradient increased as the time above 100°C
increased and as the pressure on the pattern increased. The gradient was
most sensitive to the molding variables at intermediate power levels.
The density values measured using two techniques had a correlation

coefficient of 0.94, and the open porosity measurements had a correlation
coefficient of 0.54. A linear trend existed between the two porosity values
with the x-ray technique providing a lower average values compared to oil
immersion.
Higher pressures in the pattern and the time that the pattern was
above 100°C affected the density gradient proportionally. Sections have
larger gradients if the pattern remains above the glass transition
temperature (100°C) for longer periods of time.
3.2.4.5 Foam pattern temperature and pressure
The patterns were classified according to the steaming energy applied:

low, medium/low, medium/high, and high. The steaming energy associated
- 40 -
with the low value was 1.1 bars of steam pressure for 6 seconds; the
medium/low values were 1.6 bars for 6 seconds; the medium/high values were
1.1 bars for 10 seconds, and high values were 1.6 bars for 10 seconds.
In general, higher power inputs produced larger temperature gradients and
higher cavity pressures. The use of vacuum-assisted cooling and the bead
volatile content had the most significant effect on the temperature and
pressure.
Thermocouple and pressure data from molding experiments generally

supported the working hypothesis. As the input energy was increased, the
temperature and pressure gradient in the pattern increased. At both the
lowest and the highest power input values, the pressures and temperatures
were not substantially affected by the volatile content. The temperature
and pressure readings were sensitive to all molding variables at
intermediate power input values.
The temperature and pressure data proved that pattern properties

are sensitive to the molding parameters. At high power input values, the
process was robust, and the steam cycle overwhelmed all other variables.
At low power values, variations in the bead volatile content had a much
larger effect.
3.2.5 Pattern Properties and Casting Quality
It was hypothesized that an arrow-shaped metal front would be the

most desirable during pattern replacement. Such a front shape should help
drive the liquid by-products from the cavity. A large cross-sectional
density gradient and/or a pattern open porosity should encourage the arrow
shape. The density gradient was produced during the cooling cycle where
the center of foam pattern continued to expand while expansion in the
surface layers slowed because of the cooling.
Hopper first reported a relationship between bead fusion and casting

defects (UAB – 2001/2002). The degree of pattern fusion was controlled
by the bead exposure time above the glass transition temperature. He also
found a relationship between the degree of bead fusion and pattern open
porosity. Casting results indicated that casting defects were more
frequently associated with higher levels of pattern open porosity.
Casting experiments were conducted with patterns having wide

variations in bead density, steaming pressure, and steaming time. These
variations produced large differences in open porosity and cross-sectional
density gradients.
Two mechanisms were found to control the metal fill velocity. The
metal front velocity for 1.2 pcf patterns appeared to be controlled by
density gradients while at higher pattern densities, the velocity was
controlled by pattern open porosity. The velocity at the lower density
appears to be controlled by removal of liquid pyrolysis products, and the
velocity at higher densities by gaseous by-product removal. A large
cross-sectional density gradient produced a relatively flat metal front.
Patterns with the smallest density gradients produced arrow-shaped metal
fronts. These results proved the hypothesis to be wrong.
- 41 -
Defect occurrence was not sensitive to the metal front shape. Patterns
that produced a relatively flat front produced the fewest number of folds.
Patterns having a cross-sectional density gradient and high open porosity
produced a metal velocity of 1.8 cm/sec, a nearly ideal fill velocity.
In general, patterns producing higher or lower velocities produced more
casting defects. Bead collapse may have been responsible for the defect
reduction. Ruptured cells may collapse faster and increase the removal
of liquid pyrolysis products, thus reducing fold defects.
3.2.6 Pattern Quality Control Manual
A Pattern Quality Manual was developed and distributed to the sponsors.

The manual provides procedures for measuring pattern density, open porosity,
and volatiles as well as techniques for measuring pre-expanded bead density,
and volatile contents. Equipment, procedures, and data processing
techniques are described.
4. Pattern Materials and Processes
The objective of this task was to improve the pattern degradation

properties and reduce casting defects. Degradation of EPS, Co-polymer and
PMMA were addressed in Subtask 1.1. The property data has helped 1)
understand current pattern materials, and 2) to identify alternate
materials. EPS having a range of molecular weights was included.
4.1 Improved Pattern Materials
Development of new polymers is a lengthy process. The costs far exceed

the potential return on investment for the limited lost foam market, and
as a consequence, new pattern materials have not been developed. The
approach taken in this project is that efforts should concentrate on
enhancing currently available pattern materials. Developing a pattern
precoat was the first effort conducted.
4.1.1 Foam Pattern Pre-coat
The use of a pre-coat applied to flange patterns, prior to the

refractory coating, was found to reduce surface fold defects. Confirmation
experiments were performed using a box pattern. The box pattern is complex,
and provides a variety of metal flow characteristics. Eight castings were
poured, four of which had been pre-coated with a proprietary material.
There was an 82% decrease in the number of folds and a reduction

in average fold length of 42%, although some blisters and cold shuts
occurred. The mechanism for fold defect reduction is not known, but it
appears that the fill velocity is lower and more uniform. This infers that
the pre-coat reduced the metal velocity by sealing the pattern surface
porosity.
4.1.2 Development of “No Fold” Beads
- 42 -
A bromine additive was added to T180 beads to produce a T185 “no
fold” bead stock, and the effects of pattern density, fusion, and coating
type were examined. Castings poured using T185 beads replaced the pattern
more slowly, and exhibited fewer and smaller defects (see Fig. 12). The
use of a low permeability/high liquid absorption coating also decreased
the metal velocity and reduced the number of defects.
Fig. 12. Comparison of the resulting casting surface between the

T180 and T185 patterns. T180 = no additive foam beads, T185 = with
additives foam beads.
The difference in metal velocity as high density T180 and T185

patterns were replaced was negligible. However, the degree of pattern
fusion had a significant effect on defect size with T180 but not with T185
beads.
4.1.3 Evaluation of Alternative Foam Materials – Polyurethane
Pattern degradation research done in Subtask 1.1 indicated that the

liquid pyrolysis products accounted for 60–70% of the pattern weight at
aluminum pouring temperatures and 30–40% at iron pouring temperatures.
Liquid products are the root cause of most casting defects. The ultimate
solution to eliminate defects is to develop patterns that degrade entirely
into gaseous products.
Three types of Polyurethane (PU) foam, identified as High

Functionality, Low Functionality, and Taztol were evaluated as alternative
pattern materials. PU foam patterns produced smoother surfaces than EPS
which would benefit fatigue properties. With the exception of the Taztol
foam, the PU foams performed satisfactorily during coating, drying,
compaction and pouring. The Taztol foam collapsed as the coating was dried.
Both aluminum and iron castings were poured, but quality of the castings
was unacceptable because the foam density was too high. A lower density
- 43 -
should provide a significant improvement in casting quality.
A comparison of PU and EPS foams indicates that PU foam might be

suitable if the material has sufficient stiffness. However, the literature
indicates that CO and HCN are produced during pyrolysis, and this could
be objectionable in production environments.
4.1.4 Evaluation of Alternative Foam Materials - PMMA
Several observations were made based on pattern degradation in

Section 1.1). First, EPS degrades to about 20 – 30 % gas (70 – 80 % liquid)
at aluminum pouring temperatures while PMMA degrades to about 70 – 80 %
gas (20 – 30 % liquid). Since liquid pyrolysis products are the root cause
of many casting defects, PMMA could be a candidate pattern material.
PMMA patters were prepared using a box pattern, coated and poured
through different sprues. Casting surfaces were cleaner than those poured
with EPS patterns; however a significant number of folds were formed.
Castings poured with EPS patterns using a high permeability coating with
foam sprues and castings produced with low permeability coatings and
ceramic sprues produced some fold defects.
Castings poured using PMMA patterns, high permeability coatings and

ceramic sprues and those poured using low permeability coatings and foam
sprues were worse. The combination of a high permeable coating and ceramic
sprues probably produced turbulent metal flow and folds, while low
permeability coatings and foam sprues reduced the velocity and the
formation of fold defects.
Metal replacement of PMMA patterns was examined using the real time
X-Ray system. Metal bubbling was observed at the metal front. The gas
reduced the pattern replacement rate, and some parts did not fill completely.
This study also provided information about mass transfer during foam
replacement.
A similar study was made using 10mm thick plate patterns prepared
with PMMA and EPS. There was a significant increase in the number of folds
with PMMA compared to EPS patterns. Auger analysis found carbon and high
oxygen concentrations in folds associated with PMMA patterns. The oxygen
present in PMMA apparently aggravated fold formation.
4.2 Improved Pattern Degradation Process
4.2.1 Pattern Additives
Some additives to foam patterns show promise for facilitating polymer

degradation. Several experiments were conducted to explore their effects.
Beads with additives were prepared by Styrochem, blown, coated, and
castings poured. The results were presented in section 1.1.3.3.
4.2.2 Coating Additives
The addition of tri-calcium-phosphate (TCP) to coatings has shown
- 44 -
promise in reducing pyrolysis residue on aluminum castings. Although the
mechanism is not understood, the compound appears to release oxygen to
accelerate pyrolysis.
TCP additions did not eliminate surface pyrolysis traces, but they
were reduced and the coatings pealed cleanly from the casting surfaces.
TCP significantly reduced the lustrous carbon defects on iron castings.
The literature suggests that TCP may react with carbon to remove it from
the casting surface.
4.2.3 Effects of Pouring Atmosphere
A technique to pyrolize patterns at high heating rates was developed

by Walford Technology. This technique uses a laser spot as a heat source
and an infrared camera to monitor pattern degradation. Pyrolysis is done
in a sealed container, and pyrolysis products can be collected for analysis.
Using this technique, T185 no fold bead was found to disintegrate faster,
and the presence of Helium significantly increased the rate of degradation.
The possibility of introducing some helium into patterns was also

examined. Reducing the air contained in beads might reduce the severity
of oxide films in castings. The air content in foams can be reduced by
vacuum treatment and infiltration with an inert gas. However, the patterns
should be used quickly to prevent air reentry. The kinetics of air reentry
and the potential benefits of this approach are not yet established.
Walford Technology has questioned whether pattern fusion rather than

permeability might be critical feature affecting the production of good
castings. This hypothesis was explored using radiation imaging where
patterns were laser heated and the response recorded with an infra-red
camera. The internal bead fusion correlated with oil permeability data.
Highly fused regions of patterns had higher radiation heat transfer factors
compared to regions having a lower degree of fusion. This issue requires
additional study.
4.2.4 Pattern Degradation Energy
A study of EPS pattern degradation between 300 0C and 700 0C was

conducted to separate pattern melting and gas formation effects. The
Pattern Pyrolysis Apparatus (see section 1.1.1) was used with a pre-heated
block to simulate molten metal. A constant pressure of 1.5 psi drove
patterns into the block, and pressure and displacement were recorded. The
pattern recession velocity was a linear function of metal temperature up
to 4800C. The velocity decreased over a short temperature interval and
then became linear at about 500 0C. These results indicate that very little
pyrolysis occurs below 480 0C, and at these low temperatures, patterns
are degraded primarily by melting. Above 4800C, pyrolysis occurs and a
gas layer between the metal and receding pattern reduces the replacement
velocity because of the insulating gas effect on heat transfer.
5 Coating Control and Vacuum Assist
The objective of this investigation was to improve the procedures
- 45 -
for measuring permeability and liquid absorption of coatings. A faster
response to changes on the foundry floor was needed along with some
improvements in precision. Another objective was to understand the
relationship between coating permeability and extend lost foam casting
technology to produce steel castings.
5.1 Coating Quality Control
5.1.1 Coating Control – Case Study
The Coating Quality Control Procedures were used at a foundry

producing aluminum engines. Inconsistent permeability data was traced to
inconsistent drying procedures.
5.1.2 Correlation between UAB and GM Coating Permeability Procedures
A correlation was developed between the UAB permeability values and

the values obtained by GM. This correlation was necessary to assist UAB
personnel in troubleshooting efforts. Correlations between the values were
examined using ten typical coatings with permeabilities covering the range
commercially used for producing both aluminum and iron castings. Although
theory predicts a linear relationship between data obtained with both
techniques, a linear relationship was not obtained.
It was found that the GM procedure measures coating permeability

as it is influenced by both permeability and coating thickness as affected
by viscosity. The UAB technique measures permeability in a way that the
data can be used in classic equations to predict flow rates at various
pressures and coating thicknesses.
A geometry index “GI” was developed to normalize GM permeability

values for coating thickness and cross-sectional area. The correlation
between the UAB test and the normalized GM data does have a linear
relationship (see Fig. 13). The thickness of the GM coating disc must be
determined and recorded to correlate the two permeability values.
The normalized GM data were used to predict permeability under the

same condition as the UAB procedure. The predicted flow rates were higher
than the measured data, and the difference was greater for higher
permeability coatings.
The literature on various gas flow regimes indicates that

permeability measurements obtained using the GM procedure should be higher
than measured because of gas compressibility and the mean free path slippage
at low pressures. For a high permeability coating, the slippage increases
because of the low test pressure. This contributes to the larger difference
between the predicted results and measured data.
- 46 -
100
Flow(cm 3/(cm 2.sec))
PUAB = 10.188PGM + 1.8127
UAB Normalized
80
R2 = 0.9881
60
40
20
0
0 2 4 6 8 10
GM Perm (Normalized)(cm^4/(g.min))
Fig. 13. Plot of Correlation between UAB and GM normalized Data. The
normalized data fits a near perfect linear relationship.
5.1.3 Reorientation of Research Focus
Based on metal front shape and velocity measurements data obtained

in x-ray studies of pouring GM plates, an effort was begun to evaluate
effects of pattern open porosity on the ‘effective’ permeability of
combined pattern/coating systems. Preliminary examinations indicated that
the pattern open porosity made a significant contribution to the
‘effective’ system permeability. This may explain why previous research
on modeling gas removal failed to balance the heat transfer and the gas
removal equations. Based on above observations, additional research was
conducted on pattern quality control procedures.
5.2 Coating Consistency
The goal of this study was to develop procedures to monitor pattern

coating thickness. The Coating Quality Control Manual was revised and
issued to participating companies. Current coatings are dramatically more
consistent using these procedures.
5.3 Vacuum Assisted Pouring
The goal of this effort was to determine interactions between

externally applied vacuum to assist in the removal of pattern pyrolysis
products.
5.3.1 Drawback of Vacuum Assist
Earlier research indicated that pyrolysis products could be removed
- 47 -
from the casting cavity using high vacuum levels. Unfortunately, the vacuum
produced high metal flow rates and increased the incidence of defects
including burned on sand and porosity.
5.3.2 Steel Casting
The goal of this investigation was to determine the feasibility of

using the lost foam process to produce steel castings with a minimum of
carbon pick-up.
With current pattern materials, this task is nearly impossible.

All pattern materials pyrolize to produce carbon which is absorbed into
the steel. Permission was given by the sponsors to discontinue work on
this task. The effort was redirected to Task 6, In Plant Troubleshooting
and Technology Exchange.
6. In-Plant Case Studies
The objective of this task was to assist sponsors solve in-plant

problems associated with lost foam casting production. This process also
helps UAB understand real production issues.
6.1 Case Study – Equipment Maintenance
A significant technology transfer effort occurred in an aluminum

engine foundry. This foundry experienced high levels of scrap that was
interrupting the flow of castings to the machining center. Process changes
had been made without understanding the downstream effects, and equipment
maintenance had been neglected. A process review found two areas in pattern
production where significant improvements could be made. New procedures
were implemented to measure bead density and to establish a time period
for using the beads.
Procedures for measuring coating permeability and viscosity were

implemented in both the foundry at the coating supplier. Within three months
scrap rates had decreased from 40% to below 10%. A training seminar was
then held to provide guidance for further improvements.
6.2 Case Study – Gas Cavity and Cold Laps in Iron Castings
Several experiments were conducted in an iron foundry to investigate

the formation of a gas cavity in pipe flanges. SEM examination found that
the defect was caused by entrapped pattern pyrolysis products. A group
of foam patterns was wired with metal position probes and thermocouples
(see Fig. 14) to understand the formation of the gas cavity defect. Metal
filling profile, metal velocity and temperature near the flange were
measured, and it was found that pipe sections did not fill evenly (see
Fig. 15). It was also found that enlarging the bottom of the sprue decreased
the filling time and produced more uniform filling.
- 48 -
TC1 TC2 9~12
Section d 5~8
Probes 4,8,12
TC6
TC8
Section c
Probes 3,7,11 CASTING
CASTING
Section b E
Probes 2,6,10
5~8 1~4 F
1~4
TC5 9~12
Section a TC7
Probes 1,5,9 sprue
1~4 TC1
5~8
9~12 1~4
TC3
Ground Probe CASTING CASTING
A B
Steel TC2
Flask TC4
5~8 9~12
THERMALCOUPLES VELOCITY PROBES

(a) (b)
Fig. 14. Location of metal position probes and thermocouples. Fifty-two

position probes and eight pairs of thermocouples were inserted into foam
pattern before cluster was put into flask: a) side view; b) top view.
Fig. 15. Metal front profiles in different sections. Time that metal reached
position probes is filling time. The longer the filling time, the lower the metal
level is: a) casting A; b) casting B; c) casting E; d) casting F.
In addition to the gas cavity defect, surface defects (cold laps)

were also experienced. Temperature measurements in the mold found that
the metal was only slightly above the liquidus as the top flange was filled.
Experiments were conducted to determine metal and sand temperatures. It
was found that the defects could be eliminated by increasing the sand
- 49 -
temperature by 25°F. The slightly elevated sand temperature provided time
for liquid pyrolysis to escape and produce a smooth metal fill profile.
The scrap total scrap rate fell to below 5%.
6.3 Case Study – Pattern Quality and Casting Quality
A ductile iron foundry experienced periods of high scrap rates

associated with different batches of patterns. The ‘good’ batch produced
castings with less than 5% scrap while the ‘bad’ batch produced 25% scrap.
Pattern density and porosity measurements found that scrap was associated
with high density and open porosity gradients even though the average
pattern densities were identical.
The polymer molecular weight (Mw) and the glass transition

temperature (Tg) were determined and no difference was found between either
values in the ‘good’ and ‘bad’ batches of patterns. Other parameters
including coating permeability and viscosity, and metal temperature were
within normal limits. The density variations which were at the root of
the problem were thought to be caused by the pattern blowing operation,
but data could not be obtained from the pattern shop to prove it.
6.4 Case Study – Pattern Permeability Control on the Production Floor
Designed experiments were conducted at a sponsor foundry to determine

the effect of pattern permeability on the occurrence of casting defects.
The goal was to determine the effect of pattern permeability by blowing
patterns having different pentane contents. It was found that:
1) Pentane contents from 1.7 % to 2.4 % resulted in a wide range

of pattern permeabilities with the lower pentane levels producing
the higher values.
2) Casting scrap decreased with increasing pentane contents and lower

pattern permeabilities; and,
3) A large permeability range was found in pattern slices where

porosity defects occurred. The permeability range was related to
the location of fill guns.
Real time x-rays of a section of a cylinder head found that gas holes
was originated from lighteners in a pattern slice. The tool was modified
to eliminate the blind holes and the two fill guns added. The acceptable
pentane at during blowing was changed, and scrap decreased to less than
3%.
6.5 Case Study – Shot Blast Carryover
Shot blasting materials were found to be the cause of casting defects

in one plant. Steel shot were found using SEM and EDX analysis in the defect.
These steel balls probably came from remelted parts that had been shot
blasted.
- 50 -
6.6 Case study – Pore Defect in a Lost Foam Iron Casting
Experiments were conducted to identify causes of a pore defect. The

casting sample was x-rayed and additional interior pores found. The casting
was fractured and the surfaces were analyzed with SEM and EDX. Coating
debris, as well as carbon was found. This indicates that the defect was
caused by foam pyrolysis, but not by slag as thought. Experiments are
underway to eliminate the problem.
7. Energy and Environmental Data
The available data from sponsors was received and reviewed. The
most comprehensive report received was “Life Cycle Assessment of Aluminum
Casting Processes” prepared by GM Corporation.
The Life Cycle report compares Lost Foam to other casting processes
including semi-permanent mold and precision sand processes. The
environmental burdens associated with raw material, energy consumption,
gaseous emissions, and waste generation are analyzed. In general, the
environmental burdens associated with the Lost Foam and semi-permanent
mold processes were similar and lower than burdens associated with
precision sand process. Overall, Lost Foam was determined to be the most
environmentally friendly process.
8. Technology Transfer
The objective of this work was to assure timely and cost effective
project performance, disseminate technical information to project
participants, the U.S. metals casting industry, associated industries,
and the general public.
Technology transfer between UAB personnel and sponsors occurs in

the form of meetings at four month intervals and individual contact as
requested by the sponsors. These meetings provide reviews of the research
achievements and input from the sponsors concerning future directions.
Two levels of training courses have been prepared and used. These
are referred to as (1) Entry level for new employees and (2) Process
Understanding for Engineers and Key Production Personnel. These courses
have increased the understanding of features that control the Lost Foam
Process. The information has been presented to sponsors in seminars and
the manuals distributed to sponsor companies. These courses were directed
at process control personnel including supervisors, engineers, and
managers. The important parameters controlling the process are presented
along with measurement and data analysis procedures.
Follow-up visits to foundries have been encouraging. The increased

understanding of the controlling features of the process has significantly
increased the attention to details on the production floor.
The following are details of the sponsor meetings given in the past
years:
- 51 -
No. Meeting location and date Participant
man-hours
1 Chicago, IL on October 20 -21, 1999 756
2 Birmingham, AL on February 16-17, 2000 648
3 Chicago, IL on June 20 -21, 2000 568
4 Chicago, IL on October 17 -18, 2000 492
6 Chicago, IL on June 20 -21, 2001 456
7 Saginaw, MI on Oct. 22, 2001 344
9 Frankenmuth, MI on June 19-20, 2002 432
10 Saginaw, MI on Oct. 23-24, 2002 648
12 Knoxville, TN on June 5, 2003 216
13 Chicago, IL on Oct. 23, 2003 256
Interaction with the sponsors at these meetings allows sponsors to

respond to the project progress and provide direction for further research.
Based on sponsor directions the research focus has generally dealt with
(1) developing a computational models that describe mold filling and metal
solidification, (2) reducing casting defects to metal properties and
conserve energy, (3) improve casting accuracy and precision, and (4)
determining environmental emissions.
A survey was conducted among sponsors in February 2003 to determine

the areas of most interest. The results of the survey are presented below
with higher rating numbers indicating high priority research topics.
Summary of Research Priorities as Established by

Project Participants
Research Topic Rating
A. Computational Model 98
B. Aluminum Defect Reduction 96
C. Aluminum Property Data 94
D. Pattern and Bead Variability 90
E. Improved Coatings 81
F. Alternate Pattern Materials 80
G. Casting Dimensions 73
H. Glue Related Defects 69
I. Iron Defect Reduction 62
J. Iron Property Data 56
K. Ductile Iron Property Data 54
L. Emissions 52
M. Low Volume Products 43
These priorities fall into five general categories including (1)

development of a computational model that will describe mold filling and
metal solidification (Item A), (2) reduced casting defects that will
improve metal properties and conserve energy (Items B, C, D, E, F, H, I,
- 52 -
J, K), (3) improved casting accuracy and precision (Item G), and (4)
environmental emissions (Item L).
Results of the research are also presented at national and/or

international conferences such as those held by the American Foundry
Society and the Materials Society (TMS).
ACKNOWLEDGMENTS
We wish to express grateful acknowledgment and deep appreciation

to the many companies and individuals who participated in the project by
providing advice, descriptions of particular difficulties, defects, and
guidance regarding the most fruitful areas for investigation.
A special word of thanks goes to Joe Santner for guidance throughout

the project. We also wish to express our appreciation to the Department
of Energy for technical assistance and partial funding of the project.
Special thanks go to Bob Trimberger, Ehr Ping Huangfu and Dibyajyoti
Aichbhaumik. Matching funds from DOE were provided under Cooperative
Agreement DE-FC07-99ID13840 with Amendments. Matching funds were also
provided to the University of Missouri at Rolla by the Missouri Research
and Training Center.
Our hope is that the information contained in this report will be

of assistance to companies throughout the United States in advancing the
technology of Lost Foam Casting.
____________________________ ________________________
CHARLES E. BATES HARRY E. LITTLETON
____________________________
WANLIANG SUN
- 53 -
USE OF THIS REPORT AND INFORMATION
CONTAINED THEREIN
Limitation of Liability
The faculty and staff of UAB associated with this project have used
their professional experience and best professional efforts in performing
this work. However, UAB DOEs not represent, warrant or guarantee that
its research results, or product produced therefrom, are merchantable or
satisfactory for any particular purpose, and there are no warranties,
express or implied, to such effect. Acceptance, reliance on, or use of
such results shall be at the sole risk of Sponsor. In connection with
this work, UAB shall in no event be responsible or liable in contract or
in tort for any special, indirect, incidental or consequential damages,
such as, but not limited to, loss of product, profits or revenues, damage
or loss from operation or nonoperation of plant, or claims of customers
of Sponsor.
Report No.: 527985-2004 Project Final Report
To: Department of Energy; The American Foundry Society; and

the AFS/DOE/LFC Foam Casting Consortium Member Companies
Date: February 22, 2005
- 54 -

Advanced Lost Foam Casting Technology PH

Uploaded by

Copyright:

Available Formats

Advanced Lost Foam Casting Technology PH

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advanced Lost Foam Casting Technology PH

Uploaded by

Copyright:

Available Formats

FINAL REPORT on

ADVANCED LOST FOAM CASTING TECHNOLOGY

THE DEPARTMENT OF ENERGY,

THE AMERICAN FOUNDRY SOCIETY,

AFS-DOE-LFC FOAM CASTING CONSORTIUM MEMBERS

DOE CONTRACT N0. DE-FC07-99ID13840

REPORT NO. 527985-2004 Project Final Report

EXECUTIVE SUMMARY ··················································· 5

1. Computational Model and Data Base ································ 10

1.1 Pattern Material Degradation Properties Data Base ··············· 10

1.1.2 Visualization of Metal Foam Replacement through Real Time X-Ray

1.1.3 Pattern Degradation and Casting Defect Formation ················ 23

1.2 Computational Model ············································ 27

2.1 Identify Sources of Casting Distortion ························· 31

2.1.1 Case Study 1: Long Pipe ······································· 31

3. Pattern Production ··············································· 33

3.1 Pattern Bead Properties ········································ 33

3.2 Pattern Tooling ················································ 37

4. Improved Pattern Materials and Processes ························ 42

4.1 Improved Pattern Materials ····································· 42

4.2 Improved Pattern Degradation Process ··························· 44

5.1 Coating Quality Control ········································ 46

5.2 Coating Consistency ············································ 47

5.3 Vacuum Assist ··················································· 47

6. In-Plant Case Studies and Technology Exchange ··················· 48

6.1 Case Study – Equipment Maintenance ····························· 48

7. Energy and Environmental Data ···································· 51

8. Technology Transfer ···············································51

Use of This Report and Information Contained Herein ·················54

Previous research, conducted under DOE Contracts #DE-FC07-89ID12869,

Foam pyrolysis experiments – These experiments measured parameters

The x-ray system has provided valuable data on foam replacement.

Defect formation – Defect formation, especially fold and porosity

• Four sprue designs were evaluated without finding any significant

• Data on root causes of defects in Lost Foam Casting were compiled.

Flow 3D implemented a new pattern density model to simulate density

Task 2. Casting Dimensional Accuracy. Two sponsors having dimensional

Three approaches are available to reduce distortion caused by uneven

• Adjust the fill and compaction cycle to ensure internal cavities

Task 3. Pattern Production. This task focused on determining the

• A pattern permeability apparatus was developed as a tool for

Task 4. Improved Pattern Materials and Processes. Modifying currently

A new bead material, referred to as “no fold”, was introduced by

Several approaches were used in an effort to improve the pattern

Task 5. Coating Control. Developments on coating consistency were

Pouring under the influence of vacuum was explored. The application

Seven case studies conducted at participating foundries illustrate

Task 7. Energy and Environment Data. Energy consumption and emissions

Task 8. Technology Transfer. Technology transfer between UAB

The objective of this task was to develop a data base of liquid/gas

1.1 - Pattern Material Degradation Properties Data Base

Foam pyrolysis studies focused on mass and heat transfer in the

1. Pattern degradation during Foam Pyrolysis

Achievements in these tasks are summarized as follows:

1.1.1 Characterization of Pattern Degradation during Foam Pyrolysis

An effort was made to extend the temperature capability of the foam

A thermal gradient of approximately 800°C developed in the heater

Carbon was found to be deposited on the heater front face. Carbon

1.1.1.2 Heater Size and Shape