Академический Документы
Профессиональный Документы
Культура Документы
97-206310
Investigation Composite
of Springback Material
Discretionary
Associated
Fund
With
Component
Fabrication
Final Report,
Space
Flight
Center,
Alabama
November
1997
The NASA
STI Program
Office...in
Profile
Since its founding, NASA has been dedicated the advancement of aeronautics and space science. The NASA Scientific and Technical Information (STI) Program Office plays a key this important part in helping role. NASA maintain
to
CONFERENCE
PUBLICATION.
Collected
papers from scientific and technical conferences, symposia, seminars, or other meetings sponsored or cosponsored by NASA. SPECIAL or historical PUBLICATION. Scientific, technical,
Office is operated
by
Research Center, the lead center for scientific and technical information. The
projects, and mission, often concerned with subjects having substantial public interest. TECHNICAL TRANSLATION.
NASA STI Program Office provides access to the NASA STI Database, the largest collection of aeronautical and space science STI in the world. The Program Office is also NASA's institutional mechanism for disseminating the results of its research and development activities. These results are published by NASA in the NASA STI Report Series, which includes the following report types: TECHNICAL PUBLICATION. Reports of completed research or a major significant phase of research that present the results of NASA programs and include extensive data or theoretical analysis. Includes compilations of significant scientific and technical data and information deemed to be of continuing reference value. NASA's counterpart of peer-reviewed formal professional papers but has less stringent limitations on manuscript length and extent of graphic presentations. TECHNICAL MEMORANDUM. Scientific and technical findings that are preliminary or of specialized interest, e.g., quick release reports, working minimal analysis. CONTRACTOR contractors REPORT. Scientific and papers, and bibliographies that contain annotation. Does not contain extensive
English-language translations of foreign scientific and technical material pertinent to NASA's mission. Specialized services that complement the STI Program Office's diverse offerings include creating custom thesauri, building customized databases, organizing and publishing research results...even providing videos. about the NASA STI Program
Office, see the following: Access the NASA STI Program Home Page at http ://www.sti.nasa.gov E-mail your question help@sti.nasa.gov via the Internet to
Fax your question to the NASA Access Help Desk at (301) 621-0134 Telephone 621-0390 Write to: NASA Access Help Desk NASA Center for AeroSpace Information 800 Elkridge Landing Road Linthicum Heights, MD 21090-2934 the NASA Access Help Desk at (301)
NASA/TM-97-206310
Investigation Composite
of Springback Material
Discretionary
Associated
Fund
With
Component
Fabrication
Final Report,
Space
Flight
Center,
Alabama
and
Center
November
1997
Available
from:
NASA Center for AeroSpace Information 800 Elkridge Landing Road Linthicum Heights, MD 21090-2934 (301 ) 621-0390
National
Technical
Information
Service
ii
TABLE
OF CONTENTS
1. INTRODUCTION 2. EXPERIMENTAL
1 2
Designed Selection
2 4 6 7 7 8 10 12 14 14 16 18 22 24 27 28 29
Configuration Considerations
Bagging Procedures .......................................................................................................... Part Fabrication ................................................................................................................ Collection ................................................................................................................. Experiment OF RESULTS Problems Data ................................................................................................. ................................................................................................... Confirmation
2.7 Data
........................................................................................................ ............................................................................................................
3.3 Taguchi Analysis of the Data ............................................................................................ 3.4 Discussion of the Factors .................................................................................................. 3.5 Confirmation Experiment ................................................................................................. ...................................................................
AND RECOMMENDATIONS
.............................................................................................................................. ...................................................................................................................................
iii
LIST
OF FIGURES
tooling
angles
3 7 9 9 11 l1 12 -.............. 13 13 15 15 17
2.
vacuum
.........................................................................................
part ready
................................................................................................. ............................................................................................ .............................................................................. diagram .............................................................................. ............................................................................. tooling during layup ..........................................
tooling
configuration in female
of bridging springback
problem
data versus
run number
..................................................................
LIST
OF TABLES
test matrix
........................................................................ ................................................................
3 4 5 8 16 17 19 20 20 21 25 25 26 29 29 30 30 30 31 31 31 32 32 32 33 33
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.
of factors
......................................................................... ..................................................................
test matrix
S/N ratio calculation S/N response Mean Control ANOVA ANOVA response factor table table
........................................................................................
summary
Confirmation
run summary
1 ...........................................................................................................
Raw data for run 2 ........................................................................................................... Raw data for run 3 ........................................................................................................... Raw data for run 4 ........................................................................................................... Raw data for run 5 ........................................................................................................... Raw data for run 6 ........................................................................................................... Raw data for run 7 ........................................................................................................... Raw data for run 8 ........................................................................................................... Raw data for run 9 ........................................................................................................... Raw data for run 10 ......................................................................................................... Raw data for run ! 1 ......................................................................................................... Raw data for run 12 ......................................................................................................... Raw data for confirmation run .........................................................................................
vi
LIST
OF SYMBOLS
sample
counter
J
k
H
level of the factor runs counter number effective number sample of factors number including in the estimated mean
rt e
n)
r
runs with factor size in a single ratio for error for mean
S/N
SSe
SS m
ss,
Ve
total sum of squares error variance pooled factor sample average average data point i error variance
Ve_ xj
Yi
fl
Ve
Vep Vm
vT
total degrees
of freedom
vii
TECHNICAL MEMORANDUM
OF SPRINGBACK MATERIAL
ASSOCIATED
WITH
COMPONENT FUND
DISCRETIONARY
1. INTRODUCTION
As the National Single Stage to Orbit To accomplish properties good fatigue Polymer matrix
toward
is critical.
(lower matrix
used in some
structural examples
vehicles
in order
of such hardware
(intertank,
As the use of composite the designs fabricate caused spring-in even metalworking during increase. These on a controlled by residual zero spring. on thick to conform stress.
the complexity proven difficult was originally prepregs springback, on sharp It poses as easily more
of to a
designs
the action
composite springin,
primarily be forced
angles
Springin
or springback
to the rest of an assembly. of this research of composite associated springback parameters according will be selected and manufacturing project
The objective involved attempt processing hand layup measurement contributing to minimize and design techniques,
some
processing
and design
parameters and by
in the fabrication
components
in order
to obtain
a better panels
materials. fractional
Composite
angled
precision can be
equipment, factors
the fabricated
and a confirmation
polymer
2. EXPERIMENTAL
APPROACH
method followed
factors
in springback experiment
in
components
in this section.
of the Taguchi
Designed
Robust
design,
commonly
known
as Taguchi
Methods,
by Dr. Genichi Taguchi. Its purpose is to develop products and processes which perform consistently as intended under a wide range of user's conditions. This consistency is achieved by maximizing robustness; of factors meaning, which maximize the intended performance. fractional results 2 experiments is a specially to investigate designed the main effects and to main of a system while minimizing the impact tend to degrade Methods utilize
Taguchi interactions are distributed investigate effects making The scope are more
factorial Array
in a design.
array
in that interactions is its capability regarding for screening. effects, array. 4 This from
to all columns.
of this design
recommended factors,
array. 3 The conclusions it an excellent ordered fractional choice factorial not higher
confounding choice.
in this array,
is to look at individual
interaction
this array
approach drastically cuts down on the number 2 ! 1=2,048 trials, down to a total of 12 trials. Taguchi of factor included levels a range be designed techniques are intended against angles in order
optimum a larger
performance effects
through conditions.
environmental to provide
(noise).
Intentional in which
environment
be achieved. Robustness is a product insensitive to factors that are difficult range observed in composite angles hardware
that performs consistently on target and is relatively 4 The three different angles chosen cover the typical 60 , 90 , and 120 . The male configuration 1. A more detailed discussion giving of the tooling 2.7. in experimental from each panel, tooling is presented for each
showing in section
all three
is presented discussion
2.3. Three
data points
will be taken
test condition.
A thorough
is presented
Randomization
is the cornerstone
design. 4 Randomization of the trial run order protects the experimenter from any unknown and uncontrolled factors that may vary during the entire experiment and influence the results. This will prevent order a bias in the interpretation characteristic(s) to prevent any unintentional of which biasing factors and interactions cause a change in the average in random of the quality of interest. 5 The runs designed in the experiment. in this experiment were done
FIGURE
l.--Different
tooling
angles
as shown
on male
tools.
Taguchi and
test noise
matrix
designed
for
this
experiment
l. This as the
table
levels
conditions
for each
experimental
from order.
which The
be done
in random
on the right
for data
collection.
TABLE
1 .--LI2
orthogonal
array
Taguchi
Run/Factor 1 2 3 4 5 6 7 8 9 10 11 12
A 1 1 1 1 1 1 2 2 2 2 2 2
B 1 1 1 2 2 2 1 1 1 2 2 2
C 1 1 2 1 2 2 2 2 1 2 1 1
D 1 1 2 2 1 2 2 1 2 1 2 1
E 1 1 2 2 2 1 1 2 2 1 1 2
F 1 2 1 1 2 2 1 2 2 1 2 1
G 1 2 1 2 1 2 2 2 1 1 1 2
H 1 2 1 2 2 1 2 1 2 2 1 1
I 1 2 2 1 1 2 1 1 2 2 1 2
J 1 2 2 1 2 1 2 1 1 1 2 2
K 1 2 2 2 1 1 1 2 1 2 2 1
The factors
to be tested
in this experiment
were chosen
based
on a literature
review
experience gained from previous composite material programs worked at NASA Marshall Center. These factors and the associated levels included in the test matrix are summarized TABLE 2.--Summary
Factor CureTemperature Resin Flow Fiber Modulus Tool Material Radius Orientation Tool Radius No. Plies Layup Pressure Intensifier Resin Content Cure Method A B C D E F G H I J K
of factors
Level I 250 F Low Low Alum Male 0.25 Inch 8 (O/90)s Off Bleed Oven
of the factors
are directly
related
selection
(factors configuration
A, B, and C). The considerations often are a key drive and its related processing fabrication in
are presented
in section
2.2. Tooling
composite
The tooling
requirements
for factors
is the single
element
in producing in relation
processing-
processing-related
are outlined
Material composite
consist of two distinct components which work together to achieve The first of these two materials is the matrix binder, or resin. Many of which to use is based Epoxies on the application usage and/or range. environment resin which system, phenolic, are the most commonly used polymer Bismaleidies,
the resultant desired resins are available, the hardware primarily and polyimides because are
to be built
examples of other resin systems that are used for higher temperature applications such as leading edges, aerostructures, and nozzles. Similar to the fiber selection, only epoxy resins were selected for use in this experiment experiment There on the desired temperatures, in springback distinguish experiment. for consistency more transferable purposes. The selection percentage resins, resins of epoxy resins also makes hardware the results being from the to the largest formulations characteristics. mismatch. of composite built in industry. to use based cure play a key role A in the
of which
250 F and 350 E Thermal due to the thermal if the curing temperature
characteristics
at both of these
temperatures
will be used to
difference
has an effect
on springback.
are different
epoxies
temperatures, order
the flow characteristics The levels viscosity to satisfy Fiberite. chosen the difference
of for this
the resins.
requirements
for factors
the vendor,
summao'.
Common
fibers
used are graphite, have made for consistency the stress range fibers.
fiberglass, the
and
availability,
fibers
their usage
requirements a wide
graphite
fibers
to utilize.
The loads
the required
strength
by Herculus,
modulus
are very
dictated
by Toray,
is a commonly
fiber. These
being
dependent resin
to be used using
hardware: include
or fibers
preimpregnated
prepreg.
as prepregs
themselves
upon
the processing forms utilized, designed, of their tape is used a of each amount of of
to be used. Woven
tape are the two most common of roll widths. tow bundle
and both are produced on the application, Woven fabrics fibers different materials into desired
on rolls and available to be woven are the desired shapes aligned materials
size as well as with a particular involving thickness. have hand surfaces. Unidirectional
of tooling
to be discussed factors
in an identical unidirectional
woven
required
and lead time from the manufacturer. Therefore, a standard and fiber using set of processing areal weight,
the scope
These
2.3 Tooling The material Materials graphite, defining fabrication commonly rubber, the lowest process selected for tooling metals approach is dependent (invar, should
Configuration on the requirements aluminum), sand, around nonmetals the requirements tolerances, tow-cost, (foam, of the composite (composites, component. monolithic in
steel, center
wood),
and one-time
use materials
surface efficient
to design based
tooling.
on two key considerations---cost to be used The angles tooling for any subsequent run of parts. in this made in a be easily to be fabricated production could
and time. The parts to be made production, the tooling could easily should
for a large
bending fixture and supported with a frame for stability. Readily fabrication were 2219 Aluminum and 304 Stainless Steel. These difference--thermal experiment. A key element assembly. tooling Composite of the part for tolerance sometimes The tooling used. radius Female in composite components control. tooling part design are fabricated This design is more often common is the tolerance so that critical and easier expansion. The inclusion of this difference
materials at the time of have a key inherent property D in the cooling material
fit-up
of the part in the subsequent are on the tooling orientation tooling surface of the is also later. the radius which
interfaces to layup
consideration processing
determines problems,
presents
will be discussed
is factor
E in this experiment. part design is the radius in the angle to be fabricated. F. The radius process.
key consideration will include however, was designed panel. include utilized does
in composite is often dictated such that each The basic several surface concept
a tight radius,
0.25 inch,
and a shallow
radius,
result
in approximately to this basic Each piece of the layup 1/8-inch so that the bar or
angled
were
based
tool has two basic components: of composite was chosen surface. prepreg. Holes The
The layup
with dimensions
24 by 24 inches.
thin enough
a firm layup
around
perimeter of the plate are used to fasten the plate to the frame, and the frame also has a stabilizer across each side. These features help keep the tool rigid and prevent unwanted warpage, bending, thermal cycling
24 plies. Each
of approximately
in final panel
thicknesses industry
fall within
these thicknesses. of a composite is symmetric part is that the layup sequences so that no springback angle of each ply has fabrication; spring.
H includes
stacking
is intentionally
an unbalanced however,
The vacuum material hardware. shapes. which, to complex be attained, pressure Additionally,
bagging
of a composite
to evenly
and consistent
of part thickness
and assistance
in core placement
is the weakest
and, therefore, key link. The main purpose of the resin is to bind the load-carrying fibers The even distribution of vacuum pressure aids in a more precise control of the fiber/resin In order to provide a consistent testing environment for this experiment,
together. ratio. 7
panels were fabricated using the same basic bagging techniques and the same bagging materials. This was an essential element in the processing of these parts to ensure that they were processed consistently for better table comparison. The bagging materials materials were used for the pro_e_sing from a single vendor of all the parts source (Airtech is summarized International, in 4. 8 All the bagging obtained
Inc.), and each type of material used came from only one manufacturing additional noise into the experimental environment. All materials were minimum, the highest temperature cure in the test matrix, 350 E
TABLE 4.--Bagging Material VacuumBag Breather loth C Solid Release Film Pressure intensifier Bleeder loth C PorousRelease Film Sealant ape T MoldRelease
materials
test matrix.
Designation IpplonDP1000 Ultraweave 1324 A4OOOR Airpad Bleeder easeC L A4OOORP GS-213 Release-All 30
Comments Nylon,O.O02-1nch Thickness, 90 F Usage 3 Nylon6-6 Nonwoven,3 oz/yd2,450 Usage t F O.O02-1nch Thickness, 00 FUsage 5 Uncured Nonsilicone Rubber Fiberglass, O.O09-1nch Thickness, 00 F Usage 8 O.045-1nch Holes,0.25-InchCenters, imilarto A4OOOR S 400 FUsage Liquid,500 F Usage
bagging technique
stackup
is depicted factor
3. Based
to more
were utilized,
intensifier
was to be used,
in order
in radiused
porosity, and the potential for delaminations. The (**) materials were used only when was to be used, factor J. Resin is bled out of the prepreg during cure to control the resin part. Vacuum paths ports were utilized on both sides part ready of the tooling in the breather in figure cloth 4. for air inside the bag. A bagged 2.6 Part Fabrication for cure is shown
in a composite as escape
Each processing
according
to the factors
conditions;
environmentally
laboratory
controls helped to eliminate any potential source of environmental noise that could enter and influence the data. It is not always desired to eliminate all noise from an experiment, controllable condition, experiment noise could factor achieve was designed in section robustness into this experiment. across a larger Three panels were made an envelope as described 2. l : 60 , 90 , and 120 . This provided set of operating
conditions.
VacuumBag Breather loth(x2) C Solid Release Film* Pressure Intensifier* Solid Release Film Bleeder Cloth(x2)** PorousRelease Film** PrepregLayup MoldRelease ToolingSurface
L '--
) .--
: ..............................
Sealant ape T vacuum bagging stackup.
FIGURE 3.--Basic
FIGURE 4.--Bagged
part ready
for cure.
Prior to the layup of the composite parts, the tools had to be prepared. The bolts a stable tool. Each tool was then cleaned with solvents to remove contaminates surface. The outside The remainder the resin perimeter of the layup surface was covered with liquid-based bonding with 2-inch-wide from being mold release. surface,
tape.
the sealant
with for
to the tooling
All prepreg process. templates. Teflon template Prior time in laboratory and at the exact
TM
materials conditions
must
be stored
in cold storage
to prevent
acceleration were
of the resin cure adequate size with then cut using are cut the same stock which and coated causes the on the
to the layup
was taken out of the freezer per the test matrix, made from
and allowed
to thaw. The plies required, to cut the plies ensures These templates were to prevent resin from
The use of templates required angles. served tape. The tape to become
that all plies for the parts thin aluminum to the template transferring
very tacky,
thus inhibiting
its efficient
laid up centered
tool, per specifications During there were a previous factor. vacuum the 4th, layup,
in the test matrix. the bulk factor wrinkles, or looseness Debulking of the layup or folds between was controlled. which could Extreme create care was used to ensure wrinkles layup or bridging to ensure cloth, during the bulk adequate and a that
was used extensively of a minimal performed ensure stack included nontacky, adequate
during
bag. Debulks
16th, and 24th plies for a minimum also done after layup, for a minimum Following a vacuum the completed
to section
The parts were cured matrix, variety factor K, dictated ways. control within of different
using
supplied
were to be cured.
Composite
The autgclave
5, is programmable
control
2.7
Data
Collection
to ensure 16 inches
consistent
Prior
to
was marked
for indexing
from this side of the tool. This procedure to one-twelfth locations parts of a degree to the desired were three in nine data points
protractor Recall
to be measured
made for each run of the per run. At the completion marking be mapped on the tool. This to the part. The
on the left half of the same side as the index the measurements from the tool could
by which
l0
==
i
FIGURE 5.--Autoclave
used
for processing.
FIGURE 6.--Oven
11
FIGURE 7.--Tool
measurement
procedure.
The part was removed were cycling grease which baseline. These measurements
from
surface
Identical
measurements including
may show
a small
factors,
of the measurement
Locations
pencil in the angle facing the tool side at distances of 2, 6, and 10 inches from the side of the part had been numbered. These locations allowed for three evenly spaced measurements across each location 8. from which between springback. in the analysis the data were collected the tool baseline A negative value on the tool and part map to each other and the part measurement that the panel sprang inward. indicates is shown These data in to the edge effects of a part was 2 inches, The angle in order to get a true angle measurement that as by edge of the panel. was then measured at each of these points
measurement
at each
..... of levels of the factors This is such as the be considered run for this and
a combination during
which important
to be significant
by the analysis,
must be run. The purpose the analysis phase. should experiment experiments,
of the
array,
columns,
preliminary experiment
by a confirmation
5 The confirmation
12
FIGURE
8.--Part
measurement
procedure.
24inches
4
,l
Indexed Side of Tool 8 in. 12 in. 16 in.
2 in. I _
10 in. I
12 inches
FIGURE 9.--Data
point location
mapping
diagram.
13
3. DISCUSSION
OF RESULTS
collected from
a discussion
fabrication performed
wilt be presented,
and finally,
experiment
3.1 Fabrication The fabrication to process environmental of the panels noise which control for this experiment same conditions of a factor tooling
the exact
into the experiment. led to the elimination mentioned, difficulty intensifiers wrinkles is difficult include questionable female in getting wrinkles,
One problem,
during
from the test matrix. processing problems. These radius. Potential problems Additional problems or looseness tape that was in the material tape; they are more
As previously primarily debulks between form. stem from and pressure plies, creating
air bubbles,
which directions;
can lead to bridging the unidirectional a problem is limited inherent are used because
used in this experiment When part designs into these "workable" the rational
this experiment
to unidirectional
is presented
in section
This processing
limitation
was encountered.
tooling
configuration
in figure 10. Despite extreme care during layup and additional debulks to help aid compaction, in the female radius proved to be unavoidable. A closeup of a layup in a female tool is shown figure tooling 11. The wrinkles worse and unidirectional orientation and bridging prepreg in the radius of each from tape had been were evident encountered. during the layup process to become E, the radius with the inclusion subsequent ply. An unworkable This resulted Consequently, situation panels between
and continued
of the tooling,
fabricated
experiment. The analysis was still run as intended, but any information on this factor is lost. Recall, that the material form selection was the key the reasoning for the selection were of this form, during as presented the fabrication in section of the 2.2.
to this problem.
Also recall
or fabrication
anomalies
encountered
14
FIGURE 10.--Standard
female
tooling
configuration.
FIGURE 1 1.--Closeup
of bridging
problem
in female
tooling
during
layup.
15
3.2 The procedures procedures process established would for data collection a baseline unwanted
from which
be obtained
not induce
environmental
into the experiment_ in the appendix. over a range Table from 5 presents section 2.1, angles
The raw data for each run in the test matrix a summary the intent of the resultant in this experiment mean springback is to analyze
is presented springback
in order to provide a larger environment in whichrobustness could be achieved. in section 3.3 will treat the data as nine data points from a single source, rather from three different sources. However, prior to the analysis, springback the single used used becomes exception. is one-twelfth some more observations positive examination standard of the data. The measured 0.0934,
can be made
as the tooling
in 11 of the 12 runs, with run 9 being deviation, the accuracy run 9 is inside of the measurement the accuracy
to the other
the relative
in the tools. This accounts for run 9 not following the same trend as the other of the numbers and the accuracy of the device have masked the data for run 9. equipment had been available trend as the rest of the data. to use for data collection, run 9 probably
TABLE 5.--Summary Run 1 2 3 4 5 6 7 8 9 10 11 12 60" -1.3750 -2.2084 -1.5695 -2.3194 -0.4028 -1.8889 -1.7083 -2.9583 -0.0417 -4.0000 -4.1389 -1.4305 90" -1.t944 -1.4306 -0.8195 -1.7639 0.1250 -0.9861 -1.2083 -1.8472 0.0834 -2.4722 -3.2223 -1.1944
springback 120 -0.3750 -1.1111 -0.5972 -1.3750 0.8334 -0.9444 -0.8889 -1.2639 -0.0972 -1.1667 -1.9722 -0.9583
data.
Average StdDev -0.9815 -1.5833 -0.9954 -1.8194 0.1852 -1.2731 -1.2685 -2.0231 -0.0185 -2.5463 -3.1111 -1.1944 0.4935 0.5137 0.4470 0.4146 0.5489 0.4983 0.3591 0.7550 0.0934 1.3452 0.9488 0.2083
summary
mean
springback
data for each run on the individual seen in this graph; be the primary drivers
tools in the
runs 7, 9,
16
1.0
0.0
._
w CI
._
J_
"" cI,
-4.0
] 60 ool T
-5.0 0 1 1 I 2 1 3
RunNumber
FICURE 12.--Summary Additional springback magnitudes rankings for the average observations of each run using from
springback
data versus
by run number.
the data from all the tools for that run. Despite
the differing
and variabilities,
the relative
run to run.
TABLE 6.--Relative
Run 1 2 3 4 5 6 7 8 9 10 11 12 60 3 8 5 9 2 7 6 10 1 11 12 4
springback
90 5 8 3 9 2 4 7 10 1 11 12 6
rankings
120 2 8 3 11 4 6 5 10 1 9 12 7
(1 =lowest).
Average 3 8 4 9 2 7 6 10 1 11 12 5
17
3.3 The Taguchi factors optimize several robustness, replications There The three determine springback characteristic concept These is based factors
Taguchi
Analysis
of the Data (SIN) value. ratio to determine variability 5 being evaluated. is to The significant in order to
on the use of a signal-to-noise and levels are then chosen; to the desired the amount depending nominal the mean
to adjust
of variation
are several
This experiment
minimize
springback.
is better--springback evaluated
or zero,
is the goal.
Therefore,
SIN ratio is
S/N=-IOlog(Ve)
(1)
where variance.
for the data set. 5 This form of the SIN equation in another value, form but is a function this form to calculate by doing is
of the Since
Ve is calculated
a no-way
Ve=SSe/V
(2)
where
v e is the degrees
of freedom
associated
SS e can be obtained
SSe:SST-SS
(3)
is expressed
by
SST:i_I
y2
(4)
where
r is equal
to the number by
of repetitions
in a trial regardless
of noise
levels.
for
the mean
can be expressed
SS m =rx();)
(5)
18
of freedom
Ve,
can
also be obtained
by subtraction
V e = V T -- V m
(6)
of freedom of freedom
of repetitions
in a trial regardless
of noise
levels, to
The equation
Ve =
r-I
(7)
Combining
the terms
in the equations,
Ve can be simplified
to
y_y2-r(y)2
i=1
Ve=
r-1
(8)
calculations different
is presented tools,
in table
7. Also included
that contribute
data points
TABLE 7.--S/N
Run 1 2 3 4 5 6 7 8 9 10 11 12 Mean --0.9815 -1.5833 -0.9954 -1.8194 0.1852 -1.2731 -1.2685 -2.0231 -0.0185 -2.5463 -3.1111 -1.1944 SSI
ratio calculation
SSm 8.6696 22.5623 8.9171 29.7932 0.3087 14.5881 14.4820 36.8372 0.0031 58.3525 87.1124 12.8402
summary.
V. 0.2435 0.2639 0.1998 0.1719 0.3013 0.2483 0.1289 0.5700 0.0087 1.8097 0.9002 0.0434 S/N 6.1344 5.7854 6.9931 7.6480 5.2104 6.0498 8.8961 2.4416 20.5898 -2.5760 0.4566 13.6257
10.6179 24.6736 10.5158 31.1682 2.7189 16.5747 15.5135 41.3969 0.0729 72.8300 94.3140 13.1873
The response
tables
using shows
table the
for the S/N ratio will be generated. levels. Second, the response table mean and the associated levels.
factors
This table
19
factor
is considered
separately
to create
these Let
tables. by
table
to perform
this calculation.
X be any factor
table is calculated
1 [S/N]xj
nj
=--_i[S/N]k nj
(9)
where j = the level (1 or 2), k= runs in which factor X is set at level j, and nj =the number of runs where factor X is set at levelj (6 for every factor except E, which is 12, since there is only one level). The S/N response factor control indicate factors the strongest table is presented which deltas in table reduce 8. The largest differences factors between the levels one-half for each
factors
variability.
As a general
rule, about
of the
are to be selected.
9 The strongest
are B, C, D, I, and K.
TABLE Lev./Fac. 1 2 Delta A 6.3035 7.2390 0.9355 B 8.4734 5.0691 3.4043 C 9.0400 4.5025 4.5375 D 5.1036 8.4389 3.3353
table. G 6.1347 7.4078 1.2731 H 5.9502 7.5923 1.6421 [ I 5.1312 8.4113 3.2801 J 6.8279 0.1133 K 3.4581 6.6263
6.7146 10.0844
The response
in the y response
table is calculated
similar
to the entries
in the S/N
table,
nj
The factor
); response
in table adjust
differences
between factors
the levels
for each
indicate
the strongest
The strongest
are A, B, H, and K.
TABLE 9.--Mean
Lev./Fac. 1 2 Delta A -1.0779 -1.6937 0.6157 B -1.1451 -1.6265 0.4815 C -1.4514 -1.3202 0.1312 D -1.3572 -1.4144 0.0571 E -1.3858 0.0000
response
F -1.4676 -1.3040 0.1636
table.
G H I J -1.4437 -1.3279 0.1157 K -0.7585 -2.0131 1.2546
20
generated,
the control
factors
may be put
y, and variation,
S/N
levels value.
to reduce
variation classes
and class
of the control
is presented
factor
Affect S/N &_
Affect
S/N
Bl CI D2 E1 F2 G2 H2 I2 J2 K]
of each factor
level is necessary
The factors
will be addressed
Class
I Factors--These levels
factors is placed
affect
both average,
focus
on determining Factor
on the variation. S/N ratio. Level 1 also has the more desirable y
B--Level
response. Therefore, this selection is easy, B I. Factor K--Level 1 has the same characteristics the choice is K1. factors affect variation, S/N ratio. S/N ratio.
as in factor
B. Therefore,
Class -
II Factors--These C--Level
S/N, only. The mean Again, effect is not significant. effect Choose C 1. the mean is not significant. level 2 is the more
Factor
Factor I--Level 2 has the higher S/N ratio. desirable level for the mean. Choose I2.
Even though
not significant,
21
ClassIII Factors--Thesefactorsaffectvariation,S/N,only. - FactorA--Level 1hasthe meancloserto the targetvalue.Variationis not significant. Choose 1. A - FactorH--Level 2 hasthemoredesirable mean.Eventhoughnot significant,level 2 is the moredesirablelevel for the variation.ChooseH2. ClassIV Factors--Thesefactorsmaybe setatthe mosteconomicallevel sincenothing is affected. - FactorE--Recall, femaletools wereremovedfrom theexperiment,leavingonly onelevel for this factor,El. - FactorF--This factorhasvery little significanceto thevariationor mean.Level 2 is chosen because is easierto fabricate,F2. it - FactorG--Based on the datain the S/Ntable,level 2 is chosen.Enoughmaterialwas readily availableto fabricateconfirmationpanels, 2. G - FactorJ--Level 2 is moredesirablefor boththevarianceandmean.This factoris easy to processateitherlevel.ChooseJ2. The "PaperChampion"canbe established now thatthefactorandlevel analysisis complete. The "PaperChampion"is the optimaldesign,on paper,based the factoranalysisdoneto determine on the significantfactorsthatcontributeto the varianceandmean.This designwill be usedasa confirmationexperimentasdescribedin section2.8,with thepurposeof validatingthe conclusions drawn duringthe analysisphase. he confirmationexperimentwill bepresented section3.5. T in The "PaperChampion"for this experimentis A1B1C ID2EIF2G2H212J2K
3.4 This section These facts will discuss Discussion of the Factors in the test matrix in relation to the observed Recall, results. the The use I.
are important
in understanding is to minimize
experiment.
objective of tables
of this experiment
a wide operating
environment.
8 and 9 will help in the evaluation Factor A was the curing curing temperature factor.
of each factor. of the epoxy resin. expected, The mean given springback was lower in the temperature; for the using
temperature
resin.
thermal
robustness, temperature
Therefore,
resin.
It is expected results.
that a lower
and produce
mean
and variability
22
FactorC wasthefiber modulus.Lower modulusfibersaretypically easierto process sincethey arelessbrittle thanhighermodulusfibers.This wasa significanteffectin the observed variability in the experiment. he mean,however,wasnot effectedby the choiceof fiber modulus. herefore,the lower T T modulusfiber wasselected. FactorD wasthe tooling material.Steelhasa thermalexpansion coefficientof half thatfor aluminum;thennaturally,the steelwouldbeexpected performbetter.This wasconfirmedby the to significantobserved S/N ratio.The tooling materialselectionwasnot significantto the mean;therefore, steeltooling wasselected. FactorE wastheradiusorientationof the tooling. As discussed section3.l, this factorwas in droppedfrom thetestmatrix.Male tooling wasusedfor theremainderof the experiment. FactorF wastheradiusof thetooling. This factorwasnot foundto havea significanteffect on the meanor variability. Themoreshallowradiuswaschosenfor the confirmationrun because it is easierto fabricateandhasa betterchance producinga higherquality part. of FactorG wasthethicknessof the part.Thickerpartsprovidemorestability afterthecureof the resinis completethana thinnerpart.The analysisconfirmedthat thicker parts are more robust and the
mean was closer to the desired since target. However, these facts were not found to be significant. required A thicker layup was chosen, available. Factor significance layer shrinkage the analysis did lean in that direction and the material was readily
configuration lamination
of the parts.
The inclusion
some the
the mean,
45 plies was selected. Factor intensifier better pressure mean used for factor I was the use of a pressure to reduce B can be used here; results. intensifier in the bagging stack effect for cure. the mean. included The use of the A similar argument and produce
was shown
in less erratic
and variability
intensifier. Factor J was the resin content or variability. robust resin and the mean was flow. As confirmed bagging sequence vessel. of the finished showed closer slightly in factor part. This factor that the parts to the desired was not found in which target. no resin Also, to have a significant bagging springback.
effect more
The analysis
a no-bleed
B, restricting
the resin
the no bleed
was selected. provides pressure on the part during stresses was found resin to be the most parts.
The autoclave
adds internal
residual
of being factor
the mean and variability. due to the residual was cured in the oven.
23
3.5 This section the conclusions confirmation factorial will outline during
Confirmation
Experiment experiment 3.3. Recall, low-resolution, in order from section to validate 2.8, the
drawn experiment
the analysis
is particularly
small fractional-
experiments, In section
are utilized. and the significant factors and the optimum experiment
levels
were
selected.
was A 1B 1C 1D2E IF2G2H212J2K Next, and interactions of the factorial then good Given allow the estimated effects.
combination mean
of the levels
is based to another
of additivity factors
can be added
additivity
in the design
effects.
on the difference
observed
#=2xj-(,,-I)xY
where estimate, n = the number andj of factors to avoid included in the estimate of the mean, to be included. these factors X is the factor Nonsignificant falling into the above included factors equation in the are not gives, is the chosen level of each of the factors overestimating.
(11)
11 Therefore,
only factors
(12)
Inserting
the y values
from table
9, this equation
becomes,
145 I-1.4514-1.4144-1.
1751-1.2685-0.7585)-(6)X(-1.3858) (13)
experiment
results
cannot
be expected experiment
agree
test matrix
replications.
for the confirmation experiment is one-twelfth the result is close to the estimate. Confidence The confidence interval
the size of the initial matrix. It is important, intervals are used for this purpose. 12 experiment is presented in reference
for a confirmatory
12 as,
(14)
24
error
variance, Vep is the degrees degrees of freedom considered If the actual and result experiment factors
of freedom for the in the calculation in the confidence reliable. added every error
pooled of/_,
variance, ne= total and r=- the sample the results for reproducin the
in the
experiment. error
recognized,
observed table
in the
factor, term
is presented
in table this
in table
12. It is from
confirmation
experiment.
TABLE
l 1.--ANOVA SS 10.2366 6.2595 0.4646 0.0881 0.0000 0.7225 2.1532 4.7924 1.4855 0.3616 42.5011 264.5185 333.5836
table
initial V
matrix.
10.2366 6.2595 0.4646 0.0881 0.0000 0.7225 2.1532 4.7924 1.4855 0.3616 42.5011 2.7270
table dof 1 1 1 1
1 1 1 100 107
25
size for the confirmation data points results is alpha, a. Most literature in 95-percent any additional found
trial will be the same as a run in the initial the only missing in the results, proper 4, results interval ingredient for experiments confidence in reference runs without of this type typically controls could select
= O.0239 + _/(3.9467)
x (2.6353)
108/'(1+7 )
, 1]
(15) experiment of (16)
-1.3639<_<1.4117 The raw data for the confirmation of this data and its Taguchi analysis experiment
is presented 13.
is presented
in table
TABLE 13.--Cor_rmation
Location 2 Inches 6 Inches 10 Inches Taguchi Analysis
run summary
60 -1.1667 -1.2500 -0.9583 Mean -0.9769 90 -0.9167 -0.8333 -0.8333
Me
and Taguchi
120 -0.958 -0.9166 -0.9587 S/N 16.9877
analysis.
o.o2db
experiment
mean,
the reproducibility
error recognized,
as reliable.
interpolation of the F tables, the observed mean value falls within the interval at an alpha value all the way up to _x =0.175. This places substantial weight on the validation results of this experiment. Several from discussion masked other observations run is much in section of the Taguchi lower analysis should be highlighted. matrix The observed except of the device variance have There-
of the confirmation
than any of the runs in the initial of the numbers of the measuring device
fore, the selection of class I and II factors to reduce the variability performed as desired. Also, the observed mean is one of the closest to zero from all the runs that were performed. This also confirms the selection of class II and III factors to adjust the mean to the target value performed matrix as desired. except one, run 9.
Finally, the observed S/N ratio is much higher than any of the runs in the initial A similar deduction can also be made about run 9 in this comparison.
26
4.
CONCLUSIONS
AND
RECOMMENDATIONS
yielded
several
results.
validated of the
the
effects,
error recognized,
as reliable. controls
The degree
to which
validated
and process
It also shows
Efforts designing
as a lesson
parts.
it was unfortunate
this factor
in the experiment,
used in the design as expected. viscosity factors material and design directly
of tooling questions.
consideration
when induce
with resin flow, however, must be dealt of the part for cure, selection,
with up front
springback; These
affect design,
material
The orthogonal effects of the factors. array it an excellent and the analysis,
effects,
effects,
Given
of this experiment
of using this array was achieved. efforts a classical should include an investigation those of the interaction involving These effects of the efforts may
to be significant accurate
in particular, analysis
help to develop
prediction
for springback.
27
REFERENCES
1. 2. 3. 4. 5. 6.
Composites, American
1993.
Workshop,
Graphs,
1987.
Montgomery, Ross,
Techniques Composite
Quality
edition, NASA
Silverman, 1996.
Spacecraft
Design
Contractor
Report
4708,
7. 8. 9. 10.
Basic
Guide
to Vacuum
Bagging,
Airtech Airtech
Group, Group,
Vacuum Design
Bagging
and Materials,
of Experiment
and Engineering
International,
Ridgard, C.: Accuracy and Distortion SME #EM93-113, 1992. Taguchi, G.: Introduction Processes, 1986. Taguchi, G.;Yokoyama, to Quality
of Composite
11.
Engineering--Designing
Quality
Into Products
and
12.
Y.;andWu,
Y.: TaguchiMethods--DesignofExperiments,
vol. 4, 1993.
28
APPENDIX
raw
data
for this
experiment
in tables confirmation
14-26. run.
First,
the data
for the
initial
test
be presented,
followed
by the data
data
presented tooling
below the
were
collected was
as described taken 9. The to the part after from. "Tool ]ayup. the
2.7.
The
"Tool"
reprethe data
angled
measurement
in figure prior
angle the
curing
part.
The
(Avg)" help
of the
"Tool
"Tool on the
(Post)" resultant
columns. data.
This The
"Part"
represents is the
location. the
The observed
"Spring" springback
difference at the
"Tool
(Avg)"
column.
in the part
location.
TABLE 14.--Raw
data for
run 1. Part 57.7500 58.0833 58.3333 88.6667 88.6667 88.6667 119.3333 119,5000 119.6667
Tool 60
Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
Tool (Pre) 59.5000 59.3333 59.3333 89.9167 89.9167 89.8333 119.8333 119,9167 120.0000
Tool (Post) 59.5000 59.4167 59.5000 89.8333 89.8333 89.8333 119.7500 119.7500 120.0000
Tool (Avg) 59.5000 59.3750 59.4167 89.8750 89.8750 89.8333 119.7917 119.8334 120.0000
Spring -1.7500 -1.2917 -1.0834 -1.2083 -1.2083 -1,1666 -0.4583 -0.3333 -0.3333
90
120
TABLE Tool 60 Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
15.--Raw
data for
run
2. Part 58.2500 58.3333 58.3333 88.8333 89.6667 89.5000 119.1667 119.1667 119.1667
Tooi(Pre) 60.0833 60.6667 60.5833 90.2500 90.8333 91.0000 120.0833 120.1667 120.2500
Tool (Post) 60.1667 60.8333 60.7500 90.4167 91.0833 91.0000 120.4t67 120.4167 120.3333
Tool (Avg) 60.1250 60.7500 60.6667 90.3334 90.9583 91.0000 120.2500 120.2917 120.2917
Spring -1.8750 -2.4167 -2.3334 -1.5001 -1.2916 -1.5000 -1.0833 -1.1250 -1.1250
90
120
29
TABLE Tool 60 Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
16.--Raw
data for
run
3. Pad 59.6667 59.5833 59.5833 89.0833 89.2500 89.0833 119.6667 119.7500 119.5833 Spring -1.6250 -1.5417 -1.5417 -1.9167 -0.7083 -0.8334 -0.6666 -0.5000 -0.6250
Tool(Pre) 61.0833 61.0000 60.9167 90.0000 89.9167 89.9167 120.3333 120.2500 120.1667
Tool (Post) 61.5000 61.2500 61.3333 90.0000 90.0000 89.9167 120.3333 120.2500 120.2500
ToolAvg) (
61.2917 61.1250 61.1250 90.0000 89.9584 89.9167 120.3333 120.2500 120.2084
90
120
TABLE Tool 60 Location 2Inches 6 Inches lOInches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
17.--Raw
data for Tool(Post) 61.0833 61.1667 61.0833 90.0000 90.0000 89.9167 120.5000 120.4167 120.3333
run 4. Tool (Avg) 61.0417 61.0834 61.0000 90.0000 90.0000 89.9584 120.4584 120,3750 120.2917 Part 58.7500 58.7500 58.6667 88.1667 88.2500 88.2500 119.0000 119.0000 119.0000 Spring -2.2917 -2.3334 -2.3333 -1.8333 -1.7500 -1.7084 -1.4584 -1.3750 -1.2917
Tool(Pre) 61.0000 61.0000 60.9167 90,0000 90.0000 90.0000 120.4167 120.3333 120.2500
90
120
TABLE Tool 60 Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
1 8.--Raw
data for
run 5. Tool (Avg) 60.1667 60.6667 60.6250 90.3750 90.9583 90.9584 120.3750 120.3334 120.3750 Pad 59,7500 60.2500 60.2500 90.2500 91.2500 91.1667 121.2500 121.1667 121.1667 Spring -0.4167 -0.4167 -0.3750 -0.1250 0.2917 0.2084 0.8750 0.8334 0.7917
Tool (Pre) 60.1667 60.8333 60.7500 90.4167 91.0833 91.0000 120.4167 120.4167 120.3333
Tool (Post) 60.1667 60.5000 60.5000 90.3333 90.8333 90.9167 120.3333 120.2500 120.4167
90
120
30
TABLE Tool 60 Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
19.--Raw
data Tool (Post) 61.4167 61.3333 60.6667 89.5000 89.6667 89.6667 121.2500 121.4167 121.3333
for
run 6. Tool (Avg) 61.2084 61.1667 60.6250 89.5417 89.625O 89.7084 121.2084 121.4584 121.3333 Part 59.0833 59.1667 59.0833 88.3333 88.6667 88.9167 120.2500 120.5000 120.4167 Spring -2.1250 -2.0000 -1.5417 -1.2084 -0.9583 -0.7916 -0.9583 -0.9583 -0.9166
Tool(Pre) 61.0000 61.0000 60.5833 89.5833 89.5833 89.7500 121.1667 121.5000 121.3333
90
120
TABLE Tool 60 Localion 2Inches 6Inches lOInches 2Inches 6 Inches lOInches 2 Inches 6Inches 10 Inches
20.--Raw
data Tool(Post) 61.0833 61.0000 60.9167 90.0000 89.9167 89.9167 120.3333 120.2500 120.1667
for
run
7. Part 59.2500 59.1667 59.1667 88.7500 88.6667 88.6667 119.4167 119.3333 119.3333 Spring -1.7083 -1.7500 -1.6667 -1.2084 -1.2083 -1.2083 -0.9166 -0.9167 -0.8334
Tool(Pre) 60.8333 60.8333 60.7500 89.9167 89.8333 89.8333 120.3333 120.2500 120.1667
Tool (Avg) 60.9583 60.9167 60.8334 89.9584 89.8750 89.8750 120.3333 120.2500 120.1667
90
120
TABLE Tool 60 Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
21 .--Raw
data for
run 8. Tool (Avg) 6O.2084 60.5417 60.6250 90.2917 90.8333 91.0833 120.2084 120.3333 120.3333 Part 57.3333 57.6667 57.5000 88.4167 89.1667 89.0833 118.9167 119.1667 119.0000 Spring -2.8750 -2.8750 -3.1250 -1.8750 -1.6666 -2.0000 -1.2917 -1.1666 -1.3333
Tool (Pre) 60.1667 60.5000 60.5833 90.3333 90.8333 91.0833 120.I667 120.3333 120.3333
Tool (Post) 60.2500 60.5833 60.6667 90.2500 90.8333 91.0833 120.2500 120,3333 120.3333
90
120
31
TABLE
22.--Raw
data for
run
9.
Tool 60
Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
Tool (Pre) 61.3333 61.1667 60.6667 89.1667 89.1667 89.0000 120.6667 120.8333 120.7500
Tool (Post) 61.3333 61.0833 60.6667 89.0833 89.1667 89.0833 120.5833 120.9167 120.6667
Tool(Avg) 61.3333 61.1250 60.6667 89.1250 89.1667 89.0417 120.6250 120.8750 120.7084
Part 61.2500 61.0833 60.6667 89.2500 89.1667 89.1667 120.5000 120.7500 120.6667
Spring -0.0833 -0.0417 0.0000 0.1250 0.0000 0.1251 -0.1250 -0.1250 -0.0416
90
120
TABLE
23.--Raw
data
for
run
10. Part 55.0833 56.0000 54.9167 87.2500 88.0833 86.5000 118.3333 I19.1667 118.4167
Tool 60
Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
Tool (Pre) 59.5000 59.4167 59.5000 89.8333 89.8333 89.8333 119.7500 119.7500 120.0000
Tool (Post) 59.1667 59.1667 59.2500 89.7500 89.6667 89.5833 119.7500 119.8333 119.7500
ToolAvg) (
59.3334 59.2917 59.3750 89.7917 89.7500 89.7083 119.7500 119.7917 119.8750
Spring -4.2500 -3.2917 -4.4583 -2.5417 -1.6667 -3.2083 -1.4167 -0.6250 -1.4583
90
120
TABLE Tool 60 Location 2Inches 6Inches 10 Inches 2Inches 6 Inches 10 Inches 21nches 6Inches 10 Inches
24.--Raw
data
for
run
ll. Pad 57.0833 57.2500 56.5833 86.0833 86.3333 86.0000 119.0833 119.2500 118.9167 Spring -4.2917 -4.04t7 -4.0834 -3.2501 -3.0834 -3.3334 -1.9167 -1.8750 -2.1250
Tool(Pre) 61.4167 61.3333 60.6667 89.5000 89.6667 89.6667 121.2500 121.4167 121.3333
Tool (Post) 61.3333 61.2500 6O.6667 89.1667 89.1667 89.0000 120.7500 120,8333 120.7500
Toot(Avg) 61.3750 61.2917 60.6667 89.3334 89.4167 89.3334 121.0000 121.1250 121.0417
90
120
32
TABLE
25.--Raw
data for
run
12.
Tool 60
Location 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches 2 Inches 6 Inches 10 Inches
Tool (Pre) 59.1667 59.1667 59.2500 89.7500 89.6667 89.5833 119.7500 119.8333 119.7500
Tool (Post) 59.0833 59.0000 59.0833 89.5833 89.5000 89.4167 119.6667 119.7500 119.8333
ToolAvg) (
59.1250 59.0834 59.1667 89.6667 89.5834 89.5000 119.7084 119.7917 119.7917
Part 57.7500 57.6667 57.6667 88.4167 88.4167 88.3333 118.7500 118.8333 118.8333
Spring -1.3750 -1.4167 -1.5000 -1.2500 -1.1666 -1.1667 -0.9583 -0.9583 -0.9583
90
120
TABLE Tool 60 Location 2Inches 6Inches lOInches 2Inches 6Inches lOInches 2 Inches 6 Inches 10 Inches
26.--Raw Tool(Pre) 60.8333 61.0000 60.7500 89.5000 89.5833 89.6667 120.5000 120.7500 120.7500
data for
confirmation
run. Pad 59.7500 59.6667 59.7500 88.5000 88.6667 88.7500 119.6667 119.9167 119.8330 Spring -1.1667 -1.2500 -0.9583 -0.9167 -0.8333 -0.8333 -0.9583 -0.9166 -0.9587
Tool (Post) 61.0000 60.8333 60.6667 89.3333 89.4167 89.5000 120.7500 120.9167 120.8333
Toot(Avg) 60.9167 60.9167 60.7084 89.4167 89.5000 89.5834 120.6250 120.8334 120.7917
90
120
33
APPROVAL
COMPONENT
DISCRETIONARY
M.A.
Benzie
has been
reviewed energy
content. or programs
Review
Department Classification
has been
in its entirety,
REPORT
DOCUMENTATION
PAGE
Form
OMBNo.0704-0188
Approved
Public reporting burden for this collection of information is estimated to average I hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of inlormation. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operation and Reports, 1215 Jefferson Davis Highway, Suite 1204. Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503 1. AGENCY USE ONLY (Leave Blank) 2. REPORT DATE 3. REP'oRT TYPE AND DATES COVERED
November
4. TITLE AND SUBTITLE
1997
Technical
Memorandum
5. FUNDING NUMBERS
Investigation of the Springback Associated With Composite Material Component Fabrication /MSFC Center Director's Discretionary Fund Final Report, Project No. 94--09)
6. AUTHORS
M.A. Benzie
7. PERFORMING ORGANIZATION NAMES(S) AND ADDRESS(ES) 8. PERFORMING REPORT ORGANIZATION NUMBER
George C. Marshall Space Flight Center Marshall Space Flight Center, Alabama 35812 9.SPONSORING/MONITORING AGENCY NAME(S) ADDRESS(ES) AND National Aeronautics and Space Administration Washington, DC 20546-0001
11. SUPPLEMEWrARY NOTES
10.
M-842
SPONSORING/MONITORING AGENCY REPORT NUMBER
NAS A/TM-97-206310
Prepared
12a.
by Materials
and Processes
STATEMENT
Laboratory,
Science
and Engineering
12b.
Directorate
DISTRIBUTION CODE
DISTRIBUTION/AVAILABILITY
.._f
200 words)
The objective of this research project was to examine processing and design parameters in the fabrication of composite components to obtain a better understanding and attempt to minimize springback associated with composite materials. To accomplish this, both processing and design parameters were included in a Taguchi-designed experiment. Composite angled panels were fabricated, by hand layup techniques, and the fabricated panels were inspected for springback effects. This experiment yielded several significant results. The confirmation experiment validated the reproducibility of the factorial effects, error recognized, and experiment as reliable. The material used in the design of tooling needs to be a major consideration when fabricating composite components, as expected. The factors dealing with resin flow, however, raise several potentially serious material and design questions. These questions must be dealt with up front in order to minimize springback: viscosity of the resin, vacuum bagging of the part for cure, and the curing method selected. These factors directly affect design, material selection, and processing methods.
14. SUBJECT TERMS 15. NUMBER OF PAGES
44 composite
1'7. sEcuRITY
materials,
CLASSIFICATION
springback,
18.
Taguchi
methods
19. SECURITY CLASSIFICATION
16.
PRICE
CODE
A03
20. LIMITATION OF ABSTRACT
SECURITY
CLASSII:iC'ATION
oFREPORT Unclassified
NSN 7540-01-280-5500
OF _S PAGE Unclassified
OFABSTRACT Unclassified
i,i i Standard Form 298 (Rev. 2-89) Prescn_bed ANSI Std, 239-18 by 298-102
Unlimited