One or more of the Following Statements may affect this Document
• This document has been reproduced from the best copy furnished by the
organizational source. It is being released in the interest of making available as
much information as possible.
• This document may contain data, which exceeds the sheet parameters. It was
furnished in this condition by the organizational source and is the best copy
• This document may contain tone-on-tone or color graphs, charts and/or pictures,
which have been reproduced in black and white.
• This document is paginated as submitted by the original source.
• Portions of this document are not fully legible due to the historical nature of some
of the material. However, it is the best reproduction available from the original
Produced by the NASA Center for Aerospace Information (CASI)
(HiSA-Tfl-82499) fi^FOBfilSUHMIiT OF SBB
ALUHIMUa COHPONEliTS BX UALBUT BULL BLAST
MBHOVAL OF PBOIBCIIVE COATIAGS (BASA) 19 p
HC 402/SF A01 CSCL 21tt Uncias
REFURBISHIVENTOF SRB ALUMINUM COMPONENTS
BY WALNUT HULL BLAST REMOVAL OF PROTECTIVE
By Wendell R. Colberg, Gail H. Gordon,
and Charles H. Jackson ^
Materials and Processes Laboratory
George C. Marshall Space Flight Center
Marshall Space Flight Center, Alabama
1. aeaMT NO. t. SOVOINMiNT aocisa on no.
S. RCCINItNT*S CATAI.OO NO.
4. TITCi ANQ SUtTtTCt
Refui!^iBhment of SRB Aluminum Components by Walnut
Hull Blast Rmnoval of Protective Coatings
t. RtaORT DArt
a. MCaFOSMINO OR6AN1IATION CCOC
7. AUTNoais) Wendell R. Colberg. Gail H. G(»^n. and
Charles H. Jackson
•.RtarORMINO oaOANttATION NtRONr »
S. NtarONMINO OnaMilCATKHt N«.Mi MM AHOaCfl
(Jeorgo C. Marshall Space Flight Center
Marshr.ll Space Flight Center. Alabama 35812
10. 'NONR UNIT NO.
n. CONTRACT OR SR ANT NO.
1$, TVRf OR RtROR'. S RCRIOD COVKRCD
12 . sroNsoaiKO aoincv namc ano Aooacss
National Aeronautics and Space Administration
Washington, D.C. 20546
14. SRONtORINO ASCNCY CODE
1 IS. SUNIH.tMCNTANY NOTtS I
Prepared by Materi^s and Processes LaboratcMry. Science and Engineering Directorate. |
A test program was conducted to develop, optimise, and scale-up an abrasive
blasting procedure for refurtislanent oi speciflc SRB com^xmcnts ; Aft Skirt.
Forward Skirt. Frustrum. and painted piece parts.
Tes* specimens utilising i2l9 T87 aluminum substrate of varying tldckjtessos
were pre^mred and blasted at progressively increasing pressures (2.76 ^ 10® • 5.52
^ 10^ N/m*) with selected abrasives. Specimens were then analysed for material
response. The optimum blasting parameters were determined on panel specimens
and verified on a large cylindrical Integrated Teat Bed (ITB). This report
presents findings and conclusions of that study.
The euttiore express eppreetetlon to the ialloirinf personnel Ibr thehp euHMMPt
V. L. Hopper, BTS9
J. A. Cox, ET39
y. L. Cox, BT39
B. Petton, ET39
E. R. Reiech, EH13
C. F. WiUiemeon, EH13
k. H. Higgins, EH34
K. Jordan, CO-OP (EH43)
C. Wagner, CO-OP (BH43)
K. McCarty. CO-OP (BH43)
G. R. Marsh, EH22
C. Salmon, ET39
Dr. W. E. Hill. BH43
TABLK OF CONTENTS
SURFACK 1‘RKPARATION /BACKGROUND 1
PRELIMINARY TESTING 3
HARDWARE TESTED 4
S t ross 7
BLASTING OF INTEGRATED
BED (ITB) HARDWARE
RECOMMEND AT U)NS 12
LIST OF TABLES
Table Title Page
1. PhysUal Properties of Abrasives 3
2. Average Compressive Stress (40-80 psi Pressure Range) 8
3. Average Warpage (40 80 psi Pressure Ram^e) 8
4. Average Roughness (40 80 psi Pressure Range) 9
5. Optimum Blasting Parameters for Aluminum SRB Structures 10
6. Abrasive Cost and Availability 11
REFURBISHMENT OF SRB ALUMINUM COMPONENTS BY WALNUT
HULL BLAST REMOVAL OF PROTECTIVE COATINGS
The Solid Rocket Boosters (SRB) were designed and developed to be used with
Space Shuttle Main Engines to provide the ii^tial thrust to lift the Shuttle from the
launch pad to an altitude of 44 km. At that altitude, the SRBs will separate and
start their return to Earth. A parachute recovery subsystem provides for controlled
descent. The SRBs are recovered from the ocean and transferred to Kennedy Space
Center (KSC) for disassembly and refurbishment.
Refurbishment consists of cleaning, analysis, and repair of the SRB Structures
Subsystem Components and return to the inventory for the next flight. The Struc^
tures Subsystem, Thermal Protection System (TPS), will be completely removed after
each flight with an assessment of paint degradation and repair, as required, prior to
replacement of the TPS. After 3 to 5 flights the paint will be totally removed and new
new paint applied. This study was initiated to develop an abrasive blasting process
to strip paint from the aluminum components in the nost economical and expeditious
manner with minimal damage to the aluminuii substrate.
SURFACK PREPARATION /BACKGROUND
Bare aluminum, exposed to the elements, particularly ocean water and spray
will corrode rapidly, therefore the surface must be provided with a barrier to the
destructive forces in the form of a well Ixnided protective (paint) coating. Before
application of the p*’otective coating, the svibstrate surface must be prepared with
consideration to prt>cessing economics, harmful effects to the substrate and the ease of
operations. Available surface preparation methods for surfaces already coated with
oxidized or damaged paint inchKie chemical stripping which is slow and presents
effluent disposal problems, and mechanical metlx>ds such as wire brushing and hand
sandiiig. wiiich are slow and labor intensive.
Experience acquired at this Center and descriptions in the literature show the
most efficient and practical method for paint removal from SRB structures to be
abrasive blast cleaning. This is a process in which iwitaminants. old paint coatings,
etc. . are removed from metal surfaces by forceful impingement of an abrasive material
to provide a clean surface suitable for the application of protective coatings (the
abrasive blast cleaning proc*ess is not final in itself but is the initial stage of a sub
sequent three coat finish pn>cess . in the case of aluminum, insisting of a chrtimate
conversion coating, primer coat, and paint topcoat). The type and availability of
equipment to propel the abrasive particles was a consideration. There are basically
two ways of providing the energy source to the abrasive media in the blasting process.
One utilizes compressed air as the vehicle to propel the abrasive to the surface of the
work article, and the other is airless blasting whereby power of a hardened blast
wheel provides centrifugal force to propel the abrasive. The latter method was not
considered for this work, since correctly sized equipment was not available.
The compressed air blasting techniques use compressed air to propel a stream
of abrasive particles at high velocity onto a substrate surface. The expended energy
of these particles on impact has the effec' of breaking up surface contaminants and
coatings and effecting their removal with creation of a patterned or cratered surface
profile on the substrate surface. This surface pattern is often referred to as anchor
In order to achieve optimum conditions for maximum efficiency in the operation,
the abrasive cleaning equipment must have integrated design features. The equipment
components, such as. hoses, nozzles, metering valve, pressure controls, couplings,
etc. . must be properly sizes and configured with the appropriate relationship to the
selected abrasive and the total energy of the system. The kinetic energy the abrasive
transmits is computed according to the formula;
E ^ 1/2 mV^
or energy transnitted is proportional to the mass (or weight) of abrasive and to the
square of its ve ocity.
Selection of abrasives for evaluation was therefore governed by the following
characteristics of the abrasives and of the substrate/ process.
Abrasive Characteristics Substrate/Process Characteristics
Condition of surface to be cleaned
Surface finish required
Type of component
Given the intent (guidelines) to completely remove the SRB paint every 3 to 5
flights with no more than 9 mils total metal removal over the duration of 20 flights,
the rationale was to find a fast, successful, and realistically priced method that would
inflict minimal component substrate damage in terms of:
I nd uced
This document is the result of that effort.
Natural abrasives used for surface finishing include the diamond, emery, car-
borundum, sand, crushed garnet and quartz, trlpoli, and pumice. ArtifidUd abrasives
are mostly silicone carbide, aluminum oxide, boron carbide, or boron nitride marketed
under trade names . Apricot pit, pecan, black walnut, english walnut, and rice
shells (hulls) as well as corn robs also serve as natural organic abrasive materials
when properly crushed and graded with sdeves .
Successful processing depends on the uniformity of size of the abrasive employed.
Coarse or oversize particles cause deep scratches which are difficult to remove, land an
excess of finer particles will slow production. Abrasive grains are graded on a series
of screening and grading devices and should not be allowed to become mixed or to
come in contact with oil or oily dust . Recirculation of abrasives eventually renders
tlwm ineffective, since continued inpact causes grain cracking, rounding of edges, and
loss as dust.
TABLE 1. PHYSICAL PROPERTIES OF ABRASIVES 14]
3.5 - 3.9
Si0 2 (.quartz)
2.653 - 2.66
3CaO . Fe gO 2 . 3S iO 2
3.64 - 3.9
6.5 - 7.0
Al 20 g (corundum)
3.97 - 4.10
At the beginning of the study several test samples were prepared and blasted.
This served as a mechanism through which operators were trained in proper use and
maintenance requirements of test instruments. This initial analysis was also used to
standardize the angle and distance of gun from substrate that would produce optimum
results. It was concluded that a 90 deg angle and 0.127 m (5 in.) from the sub-
strate gave best results. Samples for preliminary testing were prepared in the same
way as the study test specimens
Since the aft skirt, forward skirt, and frustum, the major SRB components to
be refurbished by NASA are built of Aluminum 2219-T87, it was chosen as the sub-
strate to be used in the development of an abrasive blasting process for the removal
of the protective coating. Tho test specimens were prepared as rectangular panels
-2 2 2
with an exposed area of 9.29 x 10 m (144 in. ) from three different thickn^ses:
1.57 X 10'^ m (0.062 in.), 3.18 x lO"^ m (0.125 in.), 6.35 x lo’^ m (0.250 in.). The
test specimens were prepared using the procedure specified in Marshall Specification
Fiostik white epoxy top coat
Because of the physical characteristics of this substrate and past experience with
abrasive blasting processes, different abrasives and sizes were deliberately selected
to inflict minimal substrate damage.
The equipment selected was a Pauli & Griffin Type I Dry Honer model DH48
(Fig. 1). This is a self-contained abrasive blast unit. The abrasives are suspended
and propelled by a high velocity air strean. After striking the work surface, abra-
sives fall to the cabinet hopper and are conveyed to the cyclone separator. The dust
and light weight abrasives that do not settle down are directed to a dust collector for
disposal. The heavier abrasives that pass through the screen are returned to the
storage hoppar and then back to the blasting system: Model DH48 has a 4.54 kg
(10 lb) cleaning powder capacity and the blasting gun is a suction type with a 4.7<^
- 3 - 3
X 10 m (3/16 in.) air jet and 9.53 x lO m (3/8 in.) nozzle. Figure 2 shows a
front view of the gun, nozzle, and hose. The nozzle is considered the most important
component of the system. The amount of energy dissipated and cleaning speed are
functions of the nozzle distance to the work surface. The nozzle diameter is selected
in accordance to the air power supply available and system design. Another factor
in the selection of the nozzle is its length. A long nozzle will provide a high velocity
with increased concentration of abrasive on impact, while a relatively short nozzle
provides a wider area of impact with the abrasive being spread and decreased
m . •
V ^ VI
V' ' ^
T-iffure 2. Gun, nozzle, and hose.
Non-destructive techniques were utilized to measure residual atrMS and surface
roughness. The X-ray method was used to measure the induced stress in the alumi-
num specimen; it was possible to use this method oecauae of the elasticity of the
aluminum . A Brush Instruments Surf Indicator was used to measure tlM average
surface roughness. Test specimen's warpage and thickness were measured with a
Vernier Caliper and a straight edge. All readings used are an average of data
A Dermitron thickness measuring instrument was used to determine the paint
thickness. Test specimens were weighed before and after blasting and the metal loss
was found to be negligible.
In abrasive blasting, particles are propelled toward the material being blasted
with velocity which causes surface indentations. These indentations resttlt in local
plastic yielding. As the expansion of an affected area occurs, adjacent material, not
plastically affected, restrains this imposed expansion. The plastically deformed layer,
being dimensionally deformed and yet restrained from compensating expansion into
adjacent space, is compressively stressed during the operation and retains a certain
amount of residual stress. Residual stresses may be defined as stresses that would
remain in an elastic solid body if all external loads were removed. One important
consideration was to find an abrasive that would induce minimum compressive stress
in the surface being cleaned.
Eight different abrasives were selected based on past experience and desired
results on the substrate:
1) Silica sand 40/70
2) Silica sand 80/90
3) Garnet 25
4) Garnet 80
5) Silicon carbide 30/60
6) Aluminum oxide 36
7) Aluminum oxide 80
8) Walnut hulls 12/20.
Test specimens were analyzed with a Faststress Analyzer (automated X-ray diffraction)
This is a non -destructive technique used to determine induced surface stress on the
substrate: strains are measured only at the surface where the stress is relieved in
the normal direction. Table 2 shows the average stress v:^iues recorded from the
analyzer. No surface stress was detected in penels blasted with walnut hulls when
analyzed with Fastress Analyzer.
TABLE 2. AVERAGE COMPRESSIVE STRESS (40-80 psi PRESSURE RANGE)
Average Compressive Stress (ksi)
Silicon Carbide 30/60
Aluminum Oxide 36
Silica Sand 40/70
Aluminum OxiJe 80
Silica Sand 80/90
Walnut Hulls 12720
no stress detected by analyzer
When a specimen is blasted, re.«'idual compressive stress induces convex curva-
ture (warpage) on the peened side. Warpage is dependent upon the amount of abra-
sive striking the surface (the amount is proportional to blasting time) and to abrasive
particle size, speed, direction, hardness, and rheological properties.
Table 3 represents average warpage induced by each abrasive for pressures in
5 5 2
the range 2.76 x 10 to 5.52 x lo N/m (40 to 80 psi). As shown, blasting with
walnut hulls gives the least amount of convex curvature. The abrasive producing
the next smallest warpage was Aluminum Oxide 80.
TABLE 3. AVERAGE WARPAGE (40- 80 psi PRESSURE RANGE)
1.14 > 10'^ m (0.4506 in.)
Silicon Carbide 30/60
1. 18 ' 10'^ m (0.4636 in. )
Aluminum Oxide 36
9.41 X 10'^ m (0.3706 in. )
Silica Sand 40/70
1. 3 X 10'^ m (0.5120 in.)
8.08 > 10'^ m (0.3183 in. )
Aluminum Oxide 80
6.15 X ic'^ m (0.2421 in. )
Silica Sand 80/90
7.98 X 10'^ m (0.3142 in. )
Walnut Hulls 12/20
4.90 X lO"* m (0.0193 in.)
of POOR QOALlTf.
Surface finish of the test specimens blasted was modified by the pee ting action
of the abrasive gi'ains. Different factors such as abrasive si ?e, hardness, speed,
shape, and impact angle provide a wide range of irregularities. This surface irregu-
larity is generally called anchor pattern or surface roughness. A Surf Indicator was
used to measure surface finish of blasted panels. This instrument measures height
irregularities created by abrasive impacts. An arithmetic average of the irregularities
yields an average roughness value for a particular area being measured. Table 4
shows average roughness results for each abrasive. These values clearly indicate
that panels blasted with walnut hulls exhibited the least amount of surface roughness.
TABLE 4. AVERAGE ROUGHNESS (40-80 psi PRESSURE RANGE)
Average Roughness (ym)
5.05 (199 yin.)
Silicon Carbide 307 60
4. 19 (165 yin.)
Aluminum Oxide 36
4. 17 (164.33 yin.)
Silica Sand 40/70
3. 55 (140 yin.)
3.11 (122.30 yin.)
Aluminum Oxide 80
2.06 (81.33 yin.)
Silica Sand 807 90
2.03 (79,99 yin.)
Walnut Mulls 12/20
1.06 (41.70 yin.)
Several test panels were randomly selected to simulate hardware I’efurbishment
after being exposed to a salt water environment for seven days (it has been estimated
that it would take no more than one week to recover SRB hardware in the worst
possible conditions). The aluminum panels (2219 and 6061 alloys) wei*e blasted with
walnut hulls to remove the Bostik topcoat and primer, acetone wiped, and repainted
with the Bostik system. The test specimens were then prepareil for and placed in a
5 percent salt spray chamber for 4032 hr. Tape adhesion tests were then performed
on each panel.
Test results indicated that the refurbished Bostik coating system performance
was. indeed, very similar to the Bostik touting system as originally applied to the
substrate. Based on the test results, it is recommendeti that the chromate conversion
coating be thoroughly examined after walnut hull blasting, and if bare aluminum is
exposed, a re application of the chromate conversion coating is required prior to
re upplication of the Bostik coating system.
BLASTING OF INTEGRATED TEST BED (1TB) HARDWARE
A relatively large cylindrical ITB segrnem was blasted with the selected abrasive,
walnut hulls 12/20. 'I ' results of this blasting were used as the basis for scale-up
estimations and recommendatiotjs.
Equipment used for blasting the ITB segment included:
1) Comprei jr: Schramm model JD18A. serial No. UDP26662 512
2) Helper: Clemco model SCFW 24 52. serial No. 1*^067
3) Nozzle: 1.27 ^ lO’^ m (0.5 in.) I.D,
4) Hose: 5.08 x lo'^ m (2.0 in.) O.D.
The selected blasting seqi^nce cleaned an area of 6.233 m (67.09 ft ) and took
-3 2 2 .
2040 sec (34 min) thus producing a removal rate of 3.055 ^10 m /sec (1.97 ft /min)
Complete paint removal was accomplished without harm to the underlying chromate
conversion coating and without detectable substrate damage. Presuming the removal
rate is constant, we can estimate the cleaning time for flight hardware:
Area = 34.36 m
Time = 34.36 m^
( sec \ / 1 hr \
o ARC ,A 3 2 / 1 3600 sec /
3.055 > 10 m / \ /
2) Forward Skirt :
Time = 3.48 hr
Time = 2.5 hr
3. 13 hr
The optimum blasting parameters determined for aluminum SRB structures are listed
in Table 5.
TABLE 5. OPTIMUM BLASTING PARAMETERS FOR
ALUMINUM SRB STRUCTURES
5.52 ^ 10 N/r.r (80 psi)
Distance to substrate
0.3048 to 0.6096 m (12 to 24 in.)
Angle to substrate
45 to 90 deg
In house refurbishment of the ITB segment proved no major reconversion coating
of the aluminum substrate was required. The ITB segment was sectionally divided
and blasted with walnut hulls. Two blasted sections were reprimed and repainted.
Two unblasted sections were hand sanded and repainted, and one section was only
All five sections were tested using Adhesion (wet) Tape Test as per aSTM 6301.1
. No paint was removed in either section.
It was determined from data collected and test results that Walnut Hulls 12720 is
the best abrasive among those tested to clean aluminum SRB components while inducing
minimum damage to the substrate.
Walnut Hulls were assessed as the appropriate ahi*f^.sive for SRB refurbishment
1) No major reconversion coating of aluminum is required with Walnut Hiills
(complete reconversion coating is reauired with other abrasives tested in this study).
2) Walnut Hulls produced the least amount of surface roughness.
3) No compressive stress is induced by Walnut Hulls.
4) Negligible warpage is created by Walnut Hulls.
5) Walnut Hulls are 100 percent biodegradable; therefore, no pollution is asso-
ciated with their use.
6) When purchased in large quantities Walnut Hulls are the least expensive
abrasive tested (Table 6),
TABLE 6. ABRASIVE COST AND AVAILABILITY
Cost, $/kg (453.59 kg)
Silicon Carbide 30/60
Aluminum Oxide 30
Silica Sand 30/90
Aluminum Oxide 36
Silica Sand 40/70
Walnut Hulls 12720
a. Special Order
b. Prices subject to change.
Based on findings and conclusions of this study, the following are recommenda-
tions for scale-up of an abrasive blasting syston:
Abrasive: Walnut Hulls 12720
Pressure: 5.52 x lo N/m (80 psi)
Angle to substrate: 45 to 90 deg
Distance to substrate: 0.3048 to 0.6096 m (12 to 24 in.)
Equipment: Abrasive blast unit equivalent to that employed in ITB blasting
with added recirculation capability:
Compressor: Schramm model JD18A, serid No. UDP 2S662 512
Hqpper: Clemco model SCFW 2452, serial No. 10067
Nozzle: 1.27 x lO ^ m (0.5 in.) I.D.
Hose: 5.08 x lO ^ m (2.0 in.) O.D.
1. Brady and Clauser: Materials Handbook. 11th edition.
2. Abrasive Grain and Powders: Specifications and Recommendations; The Carborun-
dum Company, 1954.
3. Military Specification M1L-G-5634C; Grain, Abru^ive, Soft, for Carbon Removal.
5 June 1970.
4. Metal Finishing: Guidebook and Directory. 1973.
5. Almen and Black: Residual Stresses and Fatigue in Metals. McGraw Hill Book
Company, Inc., New York, 1963.
6. Heat Treating, Cleaning and Finishing Metals Handbook; 8th edition, Vol. 2.
7. Baumeister; Mark's Mechanical Engineering Handbook. McGraw Hill Company,
6th edition, 1958.
8. ASTM Method 3301. 1: Adhesion (Wet) Tape Test. September 1, 1965.