OTEC Advanced Composite Cold Water Pipe:

Final Technical Report

Work sponsored by the US Department of Energy
and performed by Lockheed Martin Corporation
and its subcontractors under DoE / LM MS2
Cooperative Agreement # DE-FC36-08GO18172.
Includes other related work for completeness.

Dr. Alan K. Miller

LMSSC-ATC
Principal Investigator and CWP sub-system lead
Mr. Matt Ascari
LM MS2
Program Manager

Rev. L, Sept. 4, 2011

Status: All slides released,
LM PIRA #MAN201104003

copyright Lockheed Martin Corporation 2011 1

 Abstract and 1-page Executive Summary

Ocean Thermal Energy Conversion can exploit natural temperature gradients in the oceans to generate usable forms of energy (for example,
cost-competitive baseload electricity in tropical regions such as Hawaii) free from fossil fuel consumption and global warming emissions.

The #1 acknowledged challenge of constructing an OTEC plant is the Cold Water Pipe (CWP), which draws cold water from 1000m depths up to
the surface, to serve as the coolant for the OTEC Rankine cycle. For a commercial-scale plant, the CWP is on the order of 10m in diameter.
This report describes work done by LMSSC developing the CWP for LM MS2 New Ventures’ emerging OTEC business. The work started in early
2008 deciding on the minimum-cost CWP architecture, materials, and fabrication process. In order to eliminate what in previous OTEC work
had been a very large assembly/deployment risk, we took the innovative approach of building an integral CWP directly from the OTEC platform
and down into the water. During the latter half of 2008, we proceeded to a successful small-scale Proof-of-Principles validation of the new
fabrication process, at the Engineering Development Lab in Sunnyvale.

During 2009-10, under the Cooperative Agreement with the US Dept. of Energy, we have now successfully validated key elements of the
process and apparatus at a 4m diameter scale suitable for a future OTEC Pilot Plant. The validations include:

•Assembly of sandwich core rings from pre-pultruded hollow “planks,” holding final dimensions accurately
•Machine-based dispensing of overlapping strips of thick fiberglass fabric to form the lengthwise-continuous face sheets, holding accurate
overlap dimensions
•Initial testing of the fabric architecture, showing that the overlap splices develop adequate mechanical strength (work done under a parallel
US Naval Facilities Command program)
•Successful resin infusion/cure of 4m diameter workpieces, obtaining full wet-out and a non-discernable knitline between successive stepwise
infusions.

After an introductory section describing the background and initial work, the major sections of this report describe (mostly by graphical
means) the results of these validations at 4m scale. Within each section, the key results, evidence, and future improvements are summarized
concisely in tables.

The last section of this report describes future plans to complete the CWP risk retirement and implementation.

An extended Executive Summary at the end provides the key results from this work at the level of one photo montage and one table per
validation.
copyright Lockheed Martin Corporation 2011 2
 Contributors to this work
• CWP subsystem lead and Principal Investigator: Alan Miller
• LM MS2 CWP Program Manager: Matt Ascari
• LM SSC CWP Program Manager: Nat Shankar
• Design: Chad Nevills, Patrick Quigley, Bobby Lane
• Stress analysis: Matt Nahan, Doug Hillson, Matthew Nahan, Greg Cuzner
• Manufacturing engineering: Herman DeVliegher
• Engineering Development Lab: Ted Rosario (chief CWP technician), Mike Garcia, Jose
Tellez, Frank Phillips, Gene Evans, Gill Hennessee, Manny Martinez (lead)
• Control system wiring: Loretta Robinette, Tina Retana
• Janicki Industries (Paul VanSant, project manager)
• Owens Corning Technical Fabrics
• Glasforms, Inc.
• West Virginia Univ. (Prof. Hota Gangarao; David Dittenber)
• Initial strategic approach: Steve Bailey

copyright Lockheed Martin Corporation 2011 3

 Table of Contents
I. OTEC and the Cold Water Pipe
II. The LM Advanced Composite Cold Water Pipe
A. Selection of configuration, materials, and fabrication strategy
B. Selection of fabrication method and details
III. Analyses to date of CWP structural behavior
IV. Proof-of-Principles validation of the CWP fabrication process
V. Design configuration of the full CWP fabrication apparatus
VI. Validation of the CWP fabrication process at 4m diameter scale
A. Validation 1 – Core plank production and assembly into core rings
B. Validation 2 – Fabric architecture, production, overlap splice performance, and fabric dispensing process
1. Fabric architecture
2. Validation of overlap splice strength
3. Fabric dispensing validation
C. Validation 3 – Stepwise infusion molding process
1. Apparatus
a. Hard shells
b. Soft tools including Re-Usable Resin Distribution Line (RURDL)
c. Resin and resin handling elements
d. Acceptance testing at Janicki, transport to Sunnyvale, and installation in B/132 High Bay
e. Molding region control system
2. Initial validation workpiece
a. Layup and insertion into Molding Region
b. Infusion of workpiece
c. Results on infused workpiece
d. Validation of analytical tools for predicting infusion behavior
VII. Path forward
VIII. Executive Summary (including one-page graphical and tabular summaries of overall approach and each
validation)
copyright Lockheed Martin Corporation 2011 4
 I. OTEC and the Cold Water Pipe
• OTEC
• Major OTEC plant components
• The Cold Water Pipe is (literally) the single biggest challenge of OTEC
• The checquered past of Cold Water Pipes for OTEC projects
• Prior attempts at building OTEC plants have been bedeviled by disasters
constructing and deploying the Cold Water Pipe

copyright Lockheed Martin Corporation 2011 5

 OTEC

OTEC (Ocean Thermal Energy Conversion) is a renewable energy

technology that can provide large quantities of baseload
(24/7/365) electricity in tropical climates, for example in Hawaii
and at overseas Navy bases. The key benefits are:

•No fossil fuel is required

•No greenhouse gasses are emitted

copyright Lockheed Martin Corporation 2011

6
 Major OTEC plant components
In the OTEC
Rankine cycle,
warm surface
water boils the Heat exchangers,
ammonia working turbines, pumps,
fluid, and cold and generators
water from the are above and
depths condenses just under the
the ammonia surface
vapor back to
liquid.

The Cold Water Pipe

extends down to the
A floating
cold water layer
platform is
located at 1000m
generally the
depth, and carries
preferred
the cold water to the
configuration, to
surface
minimize the total
length of the Cold
http://coastalmanagement.noaa.
Water Pipe gov/programs/otec.html
7
copyright Lockheed Martin Corporation 2011
 The Cold Water Pipe is (literally) the single biggest challenge of OTEC

The CWP ........by 10m / 33 ft in diameter

for a full-
scale
plant is
1000m /
3300 ft
long…..

That is 2+
…and about the width of a modest
Empire
2-story house
State
Buildings….
.

http://www.trulia.com/property/1073310701-Single-Family-
http://2.bp.blogspot.com/_7tB79IR_qrQ/S Home-Cedarhurst-NY-11516
CVh6hwIVBI/AAAAAAAADn0/l5plCrhYsAk/
s400/empireStateBuilding.jpg

copyright Lockheed Martin Corporation 2011

 The Checquered Past of Cold Water
Pipes for OTEC projects
Date Diameter CWP material Results (CWP)

Claude – Cuba 1930 2m Corrugated steel sections 1st 3 attempts to deploy CWP
(on-shore plant) failed; 4th was successful

Claude – Brazil 1935 2.4m Steel sections CWP broke off in a storm during
(plant ship) deployment; plant was abandoned

Mini-OTEC 1979 0.7m HDPE (high density Successful

polyethylene)
DoE sub-scale 1983 2.4m x Fiberglass face sheet / foam Successful
CWP test (OTEC-1) 122m long core sandwich sections

Japan (Nauru; 1981 0.7m HDPE Built; operated; a hurricane wiped

shore-based demo) out the facility

India (Bay of 2003 1m HDPE CWP dropped twice; project

Bengal ) abandoned
Makai ONR study 2007 10m Steel sections Not built yet
Sea Solar Power HDPE – multiple tubes Not built yet
studies
copyright Lockheed Martin Corporation 2011 9
 Prior attempts at building OTEC plants have been bedeviled by
disasters constructing and deploying the Cold Water Pipe, starting
with Georges Claude’s OTEC plant in Cuba, 1930
“Deploying the 2,200-yard,
cold water intake tube in
Cuba, far left, proved the
undoing of Claude, standing,
left, inside the pipe. Storms,
waves, and logistical
problems encountered while
laying the unwieldy and
vulnerable pipe, above, sank
Claude’s dream to harvest
energy from the ocean.”

http://www.invitingdisaster.com/images/Ocean_Thermal_Energy_-_AHIT_Winter_09.pdf
copyright Lockheed Martin Corporation 2011 10
 II. The LM Advanced
Composite Cold Water
Pipe
• Summary of LM Advanced Composite Cold Water Pipe Approach
• II.A - Selection of configuration, materials, and fabrication
strategy
• II.B - Selection of fabrication method and major details

copyright Lockheed Martin Corporation 2011 11

 Pictorial summary of LM Advanced Composite Cold Water Pipe Approach
Fabrication off of the OTEC platform eliminates major
Integral one-piece sandwich-wall pipe with no joints deployment risks and enables affordable fabrication of
in the main load-bearing face sheets maximizes huge integral pipe
durability and reliability

Stepwise VARTM process enables fabrication of a Concept design of large-scale apparatus helped select
1000-m long one-piece CWP critical elements for validation at 4m diameter scale

Pultruded core
ring assembly

Inner and outer

fabric dispensers
inside vacuum
chamber

Molding
region

copyright Lockheed Martin Corporation 2011 12

 II. A - Selection of
configuration, materials, and
fabrication strategy
(Work done prior to the DoE program, under LM MS2 IRAD funding)

• Cold Water Pipe structural requirements

• CWP architecture
• A section of LM CWP at 4m diameter scale
• Fabrication strategy
• Primary basis for selecting the current CWP materials

copyright Lockheed Martin Corporation 2011 13

 Cold Water Pipe Structural Requirements:
Origins, loadings, resulting features, analysis status
Origin Loading on CWP Leads to feature (see Analysis status
next slide)
Wave-driven Axial fibers; integral OK
Fatigue strains
platform motions pipe without joints
Net external global Hoop fibers; sandwich OK
Internal suction
pressure wall
Grippers and Localized external Larger # of ribs in OK
bushings pressure sandwich wall
Clump weight, pipe OK
Axial strain Axial fibers
wet weight
Water currents Bending strain Axial fibers OK
1500 psi water
1000m length Vented hollow core
pressure at depth
made of composite OK
Mildly negative wet laminate
weight
30-year life, OK by material
Seawater corrosion Vinyl ester resin
immersed heritage
copyright Lockheed Martin Corporation 2011 14
 CWP architecture
Construction: Sandwich wall with hollow vented core

• Core: Pre-pultruded hollow “planks”

assembled into core rings

• Face sheets:
Longitudinally
continuous fabric
strips, applied over
assembled core rings

Materials Distributed overlap splices

• Low-cost X-Strand1 glass between adjacent fabric strips
develop full properties in
fibers having superior fatigue circumferential direction
resistance

• Derakane 80842 toughened vinyl ester resin matrix having very little water absorption
1. Product of Owens Corning
Technical Fabrics
Integral one-piece CWP with no joints in the main load-
2. Product of Ashland Chemicals bearing face sheets maximizes durability and reliability
copyright Lockheed Martin Corporation 2011 15
 A section of LM CWP at 4m diameter scale

Resin
distribution
medium (RDM)

Fiberglass face
sheet
Tongue-in-groove edges
of pultrusions Pultruded core

copyright Lockheed Martin Corporation 2011 16

 Fabrication strategy – Manufacture the CWP
directly from the floating platform
Overall Fabric on dispensing
process: rollers (continuous with
Stepwise fabric in molding region)
VARTM
infusion Hollow pultruded core
molding rings

Molding region

Gripper/translator

Cured CWP

US Patent Application Publication #

US2009/0309271A1, 2009;
European Patent Office Application
#PCT/US2009047388, 2009

Fabrication from the platform eliminates major deployment challenges, enables

integral fabrication of huge pipe, reduces transportation costs
copyright Lockheed Martin Corporation 2011 17
 Primary basis for selecting the current CWP materials: Recurring cost
(materials and fabrication) of minimum-cost design of 10m CWP for 100
MW plants, in materials not eliminated by technical show-stoppers
Relative Cost of hollow sandwich-wall CWP's for 100 MW plant
Cost of platform to support CWP / clump weight
Direct cost of CWP
$140,000,000
$120,000,000
$100,000,000
$80,000,000
$60,000,000

$40,000,000
$20,000,000
$0

Fiberglass Carbon-fiber Steel, current HDPE Steel, prior

composite work OTEC design

Carbon-
Requirements met: Fiberglass fiber Steel HDPE
composite
External pressure Yes Yes Yes Yes
WIM cyclic strain Yes Yes No Yes
WIM axial buckling Yes Yes Yes Yes
• WIM = Wave-induced VIV cyclic strain (without strakes) Yes Yes No Yes
motions (from platform) Streaming and clump weight axial Yes Yes Yes Yes
• VIV = Vortex-induced Platform rotation Yes Yes Yes Yes
vibration Manufacturable in low-cost configuration Yes Yes Yes No

Fiberglass and carbon-fiber composite are the two candidates with no show-stoppers. They have
similar net costs and can be fabricated by the same methods. Fiberglass was chosen to simplify
galvanic corrosion issues and to avoid near-term supply questions in the carbon fiber marketplace.
copyright Lockheed Martin Corporation 2011 18
 II.B - Selection of fabrication
method and major details
(Work done prior to the DoE program, under LM MS2 IRAD funding)

• Can we use an off-the-shelf composite fabrication process?

• Various possible liquid resin infusion processes as applied to vertical stepwise
molding of a CWP
• VARTM – A quick summary tutorial
• VARTM of a large wind turbine blade shell
• Recent applications of large-scale VARTM technology
• CWP processing steps using stepwise VARTM
• Tooling strategy
• Elements of VARTM that are “tweaked” for our application
• Holding, lowering, and supporting the CWP during on-the-water fabrication

copyright Lockheed Martin Corporation 2011 19

 Can we use an off-the-shelf composite fabrication
process?

Process Comments relative to usage for OTEC CWP

Spray-up (chopped Insufficient properties, VOC’s into environment
fibers)
Wet layup (fabric) Low fiber volume; VOC’s into environment
Prepreg layup/ Very expensive materials and labor; can only manufacture
autoclave cure finite-length segments in autoclave
Automated tape laying Very expensive equipment and materials
Filament winding Very difficult to incorporate a sandwich core; finite-length
manufacturable segments require joints
Pultrusion (of entire Required CWP cross-section is way too large – Pulling
CWP) force, fabric guidance machinery would be main issues
Matched die molding CWP much too large – Mold rigidity and weight issues
Thermoplastic forming Very expensive materials
Liquid resin infusion See next slide
(RTM, VARTM)

copyright Lockheed Martin Corporation 2011 20

 Various possible liquid resin infusion processes as
applied to vertical stepwise molding of a CWP

Process Mold Mold fill Can voids be

process controlled?
1. Resin Transfer Rigid fixed- Resin flows through the Yes. Apply high resin
Molding (“ordinary cavity mold fabric itself. Slow fill pressure during cure.
RTM)” surrounds part. for large parts. But requires very rigid
mold, prohibitive for
large parts

2. VARTM One-sided tool Resin flows through Yes. Pull good vacuum
(Vacuum-Assisted with vacuum high-permeability resin on fabric before infusion,
Resin Transfer bag as the distribution medium. to evacuate the air.
other side Rapid fill for large
Molding)
parts.

VARTM is the only composites fabrication process that meets the needs

copyright Lockheed Martin Corporation 2011 21

 VARTM* – A quick summary tutorial

Setup Process
•Lay up dry fabric on tool
VACUUM EXIT LINE (VEL)
•Apply PS, RDM, RDL, RIL, VDL, VEL,
(VDL)
VACUUM DISTRIBUTION LINE (VEL) and vacuum bag
VACUUM BAG •Open fill valve
•Vacuum sucks liquid resin from a
RESIN DISTRIBUTION MEDIUM (RDM) reservoir into the part
POROUS SEPARATOR (PS) •Resin spreads quickly in the resin
distribution medium (RDM, high
DRY FABRIC RESIN DISTRIBUTION permeability)
LINE (RDL) •Resin only has to travel through-
BRIDGING
CONTROL RESIN INLET thickness in the fabric (low permeability)
LINE (RIL) •Overall wetout time for large parts is
short and within the pot life of many
FILL VALVE resins
TOOL
OR CLAMP) •After cure, the porous separator (PS)
allows the RDM to be peeled off of the
LIQUID RESIN part. Or the PS can be omitted and the
RESERVOIR RDM left on the part.

*Vacuum-Assisted Resin Transfer Molding

copyright Lockheed Martin Corporation 2011 22
 VARTM of a
large wind
turbine blade
shell Resin
distribution
Vacuum lines
distribution Resin inlet
line lines
Vacuum exit
Resin
line
reservoirs
Vacuum bag
Vacuum Fabric
trap
Mold
copyright Lockheed Martin Corporation 2011 23
 Recent applications of large-scale VARTM technology
Modern wind turbine rotor Fiber composite highway bridge segments
blades (fiberglass shells) (integral deck & beams)
Size scale is similar to our CWP

www.compositeadvantage.com

3m (10 ft.) demonstration Deep Space reflector

antenna infused in one shot
http://www.compositesworld.com/articles/wind-blade-manufacturing-part-i-m-
and-p-innovations-optimize-production.aspx

•VARTM is currently being used

on some very large components

•There is a large business base of

VARTM-experienced personnel,
material suppliers, and
equipment fabricators

copyright Lockheed Martin Corporation 2011 LMSSC, Sunnyvale, CA 24

 CWP processing sub-steps (a-f) using stepwise VARTM
(Cross-section through left side of apparatus)
Pre-assembled ring of
pultruded core planks Fabric on rolls

Vacuum
chamber
Silicone
rubber soft
tools
OD ID Molding
region
hard shells
(a) Load materials: (b) Evacuate air (c) Infuse resin from bottom: (e) Release:
Maintain vacuum behind from fabric and Use liquid behind soft tools to Apply vacuum
soft tools. Move cured compact it, by counteract hydrostatic behind soft tools
CWP down, which pulls pulling vacuum pressure gradient in liquid (f) Post-cure:
new fabric, RDM, and on workpiece resin column Circulate hot air
core segments into and fabric rolls (d) Initial cure: Ambient or hot water next
molding region temperature to workpiece
Repeat
Stepwise VARTM process enables fabrication of a 10m diameter, 1000-m long one-piece CWP
copyright Lockheed Martin Corporation 2011 25
 Tooling strategy: Use the pultruded core as a “flyaway tool” that
defines the shape of the workpiece and becomes part of it. Silicone
rubber “soft tools” (re-usable vacuum bags) compact the face sheet
fabric against the outside and inside of the core to mold the CWP.
Outer hard shell
Outer soft tool

Outer face sheet

fabric (molded
Inner onto core)
soft Inner face sheet fabric
tool (molded onto core)
Sandwich core
Inner hard shell assembled from
pultruded “planks”

• Outer and inner soft tools pull away from workpiece to allow axial insertion of new workpiece
material (core and fabric), then press towards workpiece during molding operations.
• Outer and inner soft tools present smooth faces to part during cure, then release easily
•Outer and inner hard shells become containers which allow fill behind soft tools with a liquid,
enabling VARTM of very tall workpieces*.
*US Patent Application Publication
#12/951,239, Nov. 22, 2010 copyright Lockheed Martin Corporation 2011 26
Elements of VARTM that are “tweaked” for our application

•The process is applied multiple times in a stepwise

manner to a single workpiece
•The shape-defining element (assembled pultruded
core) becomes part of the product
•The vacuum bag (soft tools) can be pulled away from
the workpiece over the entire inner and outer surfaces
•The resin distribution lines are re-usable
•The resin distribution medium is left on the product
and doubles as a “gel-coat”

copyright Lockheed Martin Corporation 2011 27

 Holding, lowering, and supporting the CWP during on-the-water
fabrication
Work done by Makai Ocean Engg. under LM program N62583-09-C-0083
sponsored by the US Navy Facilities Command

Grippers
Translator

Guide
bushings A 21” scale model was tested
successfully on a scale-model
sandwich CWP.

Gripper and bushings have been designed by Makai

Ocean Engineering to manipulate and support the Makai Ocean Engineering
CWP during fabrication and subsequent service.
copyright Lockheed Martin Corporation 2011 28
 III. Analyses to date of CWP
structural behavior
(Work done outside of the DoE program, under LM MS2 IRAD or US Navy funding)

• External pressure buckling

• Localized buckling from gripper and guide bushing pressures
• Fatigue resistance after seawater conditioning to saturation
• Coupled hydrodynamic analysis of fatigue from platform motions

copyright Lockheed Martin Corporation 2011 29

 External pressure buckling
Eigenvalue 1 = 4.9
Buckling Pressure = 36.75 psi.

• Triangular core cell design works well for external pressure loadings
• FEA predicted buckling pressure of 36.75 psi exceeds operating pressure of 7.5
psi by a large Margin-of-Safety

Doug Hillson
copyright Lockheed Martin Corporation 2011 30
 Localized buckling from gripper and guide bushing pressures
Work done at LM MS2 under LM program N62583-09-C-0083
sponsored by the US Navy Facilities Command

Pressure
applied over 128”
axial length

Mode 1: Buckling Mode 2: Mode 3: Mode 4:

Load = 77.5 psi 133.5 psi 153.6 psi 161.7 psi

In the bushing and gripper regions,

the lower modes will be prevented
by the rigid circular rings of the
guides and grippers

Mode 5: 163.8 psi Detail View Deflections magnified by

2X

4m CWP has local buckling pressure of 164 psi, adequate for the localized
forces imposed by bushings operating at 50 to 75 psi local pressure. Tony Paulic
copyright Lockheed Martin Corporation 2011 31
 Fatigue resistance after seawater conditioning to
saturation: Selected CWP material shows only 6%
knockdown relative to dry material
X-Strand/vinyl ester hot press cured laminate tension-tension fatigue resistance at ambient
temperature - Dry, and wet after seawater conditioning to saturation
W.Va.U. tests: X-Strand 75% 0's filament-wound, hot press cured laminates in air (no conditioning)
Best fit through WVaU test data on unconditioned 75% 0's laminate tested dry
W.Va.U tests: X-Strand 75% 0's filament-wound, hot press cured laminate in seawater -- after RT/1700 psi seawater conditioning to saturation
Adjust WVaU pre-exponential constant to provide best fit through mean WVaU data on seawater conditioned 75% 0's laminate tested in seawater
Design minimum for 75% 0's material after seawater conditioning (10% below average strength per scatter in other data sets)
WVaU 75% 0's dry tensile test data (Strain amplitude calculated from UTS/Modulus)

0.008

0.007
Strain amplitude (R=0.2)

0.006

0.005

0.004

0.003
75% 0's in seawater - design minimum equation:
0.002
Strain amplitude = 0.007875 * Nf ^ (-0.0802)
or
0.001
Nf = (Strain amplitude/0.007875)^(1/(-0.0802))
0.000
1E-01 1E+00 1E+01 1E+02 1E+03 1E+04 1E+05 1E+06 1E+07
West
Virginia
* Saturated weight gains were about 0.1%
Fatigue life, Nf
Univ.
Seawater fatigue testing provides known material behavior for current work.
copyright Lockheed Martin Corporation 2011 32
Coupled hydrodynamic analysis of
fatigue from platform motions
Work done by John Halkyard and Associates and
Houston Offshore Engineering under LM program
N62583-09-C-0083 sponsored by the US Navy Facilities
Command. Work utilized design minimum curve from
the previous figure.

•Without a gimbal, the current 10m CWP design has a

slightly negative Margin-of-Safety
•Tapering does not help much
•A gimbal brings the MOS to a strongly positive level

Shan Shi, John Halkyard

copyright Lockheed Martin Corporation 2011 33
IV. Proof-of-Principles Validation
of the CWP fabrication process
(Work done prior to the DoE program, under LM MS2 IRAD funding)

• Proof-of-Principles apparatus
• First Proof-of-Principles workpiece
• CWP fabrication process elements validated by the Proof-of-Principles work
at 19” scale

copyright Lockheed Martin Corporation 2011 34

The CWP fabrication process was first validated at laboratory
scale (19” dia.) with a Proof-of-Principles apparatus

copyright Lockheed Martin Corporation 2011 35

35
 First Proof-of-Principles workpiece infused by VARTM

36 copyright Lockheed Martin Corporation 2011

36
 The PoP work validated behavior of the CWP fabrication
process in the following areas, and provided the confidence
to continue with the same approaches into the subsequent
DoE-sponsored work

• Mold action

• Fabric behavior

• Sealing ends of current workpiece / pulling

vacuum

• Infusion behavior

• Resin behavior

• Co-bonding behavior

copyright Lockheed Martin Corporation 2011 37

 Proof-of-Principles work validated the key elements at 19” scale (p. 1 of 2)

Element Tested? Works as Use same approach in

expected? Prototype and beyond?
Mold action
Expanding/contracting soft tools Yes Yes Yes
Good compaction of FS Yes Yes Yes
Good release of soft tools from part Yes Yes Yes
Fabric behavior
Accurate mechanized dry fabric placement Yes Yes Yes
No unacceptable wrinkling in outer FS Yes Yes Yes – with pre-stitched unit
fabric
Sealing / pulling vacuum
Lower inflatable seals Yes Yes Yes
Upper vacuum chamber over fabric rolls Yes Yes Yes
Infusion behavior
Good resin spreading (using VARTM) Yes Yes* Yes
Full wetout through-thickness Yes Yes Yes
Good wet-out at knit-line between steps Yes Yes Yes
*After changing to a 2nd generation resin spreading method (VARTM) based on results with the 1st generation method
copyright Lockheed Martin Corporation 2011 38
 Proof-of-Principles work validated the key elements at 19” scale (p. 2 of 2)

Element Tested? Works as Use same approach in

expected? Prototype and beyond?
Resin behavior
Resin mixing and handling without Yes Yes Yes – Using metering/
exotherm mixing machine
Rapid ambient temperature cure Yes Yes Yes
Good final properties
Low void content Yes Yes Yes
Uniform fiber volume In progress
Accurate fiber volume control In progress
Co-bonding behavior
Good bond between infused face Yes Yes Yes
sheets and prefabricated core
Good dimensional control of part
Circularity when using core “planks” In progress Yes
Uniformity of outer diameter In progress Yes
Shear key works OK Yes Did not Yes, or use staggered
engage fully planks instead

copyright Lockheed Martin Corporation 2011 39

 The PoP work identified the major improvements to be made in
the scaled-up apparatus
Element Inadequacy with PoP Improvement in Prototype
Soft tools Suspected pinhole may have led to Use double layer of silicone rubber with
water leakage onto part and monitoring between them for any water
incomplete cure of resin in run 3B leakage
Shear key Suspected lack of full engagement of More precise tolerancing and in-process
shear key on Run 3B checking to assure full engagement
Coverage with Excessive gaps between RDM strips Use overlap splices between RDM strips
RDM caused incomplete wet-out
Monitoring resin External sight tube did not work Use pre-calculated resin quantity for fill;
height during reliably due to air bubbles getting into possibly also use embedded fiber optic as
infusion it from resin inlet, and shooting to top cross-check
Resin distribution Non-reusable design requires Develop solvent-flushable unit that can be left
line units excessive fabrication labor in place
Fabric Some circumferential wrinkling Have fabric supplier pre-stitch fabrics together
configuration observed; one cause may be double into a single unit (standard industry
layer of fabric wound on supply rolls technique). Track down any other causes.
Lower vacuum Using Tygon tubing limited radial Use purchased SealMaster seals designed
seal travel with ample radial travel
Access to inner Lack of center access hole in Incorporate large center access hole by
resin distribution cylindrical vacuum chamber makes changing to toroidal vacuum chamber
line unit access to inner RDL awkward
Vacuum leaks Many joints between components Incorporate standard vacuum sealing methods
were sealed with sealant tape such as O-rings between components
copyright Lockheed Martin Corporation 2011 40
 V. Design configuration of the
full CWP fabrication apparatus

• Design configuration of the full CWP fabrication apparatus

• The apparatus will go inside a weather enclosure on the floating OTEC
platform

copyright Lockheed Martin Corporation 2011 41

 Design configuration of the full CWP fabrication
apparatus implements the process at large scale

Upper support structure

Core assembly region

Central tube supports
all lower inner
portions of apparatus

Vacuum chamber with

fabric dispensers inside

Molding region

Chad Nevills,
Cured CWP Pat Quigley
copyright Lockheed Martin Corporation 2011 42
 To fabricate the CWP for the Pilot Plant and commercial plants, the
apparatus will go inside a weather enclosure on the floating OTEC platform

Work done by Sound and Sea Technologies

under an LM program sponsored by the US
Navy Facilities Command

Weather enclosure
(cutaway view)
Core assembly region
Material handling lift
Vacuum chamber for fabric rolls and
with fabric core planks
dispensing rolls
inside
Molding region Fabric and core
planks received
inside shipping
containers

Floating platform

Cured CWP
Mark Brown,
Sound and Sea
copyright Lockheed Martin Corporation 2011 43
VI. Validation of the
CWP fabrication
process at 4m
diameter scale
Work sponsored by the US Department of Energy and
performed under DoE / LM MS2 Cooperative
Agreement # DE-FC36-08GO18172

copyright Lockheed Martin Corporation 2011 44

 Validation of the CWP
fabrication process at 4m
diameter scale
• Major apparatus and process elements selected for validation at 4m CWP
scale under the DoE AWPP
• VI.A - Validation 1 – Core plank production and assembly into core rings
• VI.B - Validation 2 – Fabric architecture, production, overlap splice
performance, and fabric dispensing process
• VI.C - Validation 3 – Stepwise infusion molding process

copyright Lockheed Martin Corporation 2011 45

Major apparatus and process elements selected for
validation at 4m CWP scale under the DoE AWPP

1. Core plank
production and
assembly into
core rings

2. Fabric
architecture,
fabric production,
and fabric
dispensing
mechanisms

3. Molding
region and
stepwise
infusion
validation
copyright Lockheed Martin Corporation 2011 46
 VI.A - Validation 1 – Core
plank production and
assembly into core rings
• Pictorial summary of core assembly validation
• Pultruded Core Planks
• Shear keys
• Shear key tooling
• Shear key infusion
• Core assembly adhesive and application equipment
• Representative core ring assembly operations
• Assembling core ring
• First assembled core ring (still showing internal cross-section)
• Staggered planks method of joining adjacent core rings
• Tabular summary of core production and assembly validation results at 4m
scale
copyright Lockheed Martin Corporation 2011 47
Pictorial summary of core assembly validation

copyright Lockheed Martin Corporation 2011 48

 Pultruded Core Planks
Core plank design Tongue and groove edges lock adjacent
core planks together radially

Core planks are made using the same fiberglass material,

Assembled core ring (28 planks) resin type, and fiber orientation distribution as face sheets
with shear key at bottom •Avoids any stresses due to mismatching elastic constants

Pultruded core plank

Planks manufactured by Glasforms, Inc, Birmingham, AL 49

copyright Lockheed Martin Corporation 2011
 Shear keys

Shear keys (split)

Shear keys manufactured by Janicki Industries, Sedro-Woolley, WA

copyright Lockheed Martin Corporation 2011 50
 Shear key tooling
(Putty over foam over wood base on steel frame)

Shear key tooling designed and manufactured by Janicki Industries, Sedro-Woolley, WA

copyright Lockheed Martin Corporation 2011 51
 Shear key infusion

Work done at Janicki Industries, Sedro-Woolley, WA

copyright Lockheed Martin Corporation 2011 52
 Core assembly adhesive and application equipment

A COTS thick putty adhesive was selected based on the following behavior *:
•Even a ½” bead stays in place vertically
•Squeezes out well to fill the bondlines between planks

A COTS mixing and dispensing pump designed for thick adhesives was used

*Major requirement was fabricability of large assembly, not strength, because in

the final CWP the face sheets hold the planks together.

copyright Lockheed Martin Corporation 2011 53

Representative core ring assembly operations

copyright Lockheed Martin Corporation 2011 54

Assembling core ring

copyright Lockheed Martin Corporation 2011 55

 First
assembled
core ring
(still showing
internal cross-
section)

October 27, 2010

copyright Lockheed Martin Corporation 2011 56

 Staggered planks method of joining adjacent core rings
(instead of shear keys) – p. 1 of 2
Advantages
•Direct transfer of core axial loads through plank edges
instead of through face sheets (zero stress concentration
in face sheets)
•No protruding elements on inside or outside of core
•Eliminates need to seal core planks to shear keys

Initial test: Forming leak-tight longitudinal plank-to-plank

joints (1)

Results
•Plank-to-plank joints did not leak
•Air leakage through pultrusion wall prevented
demonstration of a leak-tight assembly
•Some lessons were learned on details of assembly and
methods of applying adhesive

Notes
(1) During face sheet infusion, interior of core will be vented to

atmosphere. Any air leaks through joints or planks introduces air

into fabric being infused with resin, with risk of voids.
copyright Lockheed Martin Corporation 2011 57
 Staggered planks method of joining adjacent core rings – p. 2 of 2

Second test: Demonstrate a leak-tight staggered-planks assembly

Procedures:
•Choose a plank with relatively high surface porosity (as a severe test)
•Apply inexpensive COTS liquid sealant (“slosh compound”) to the inside of the plank
•Cut plank into 4 portions
•Assemble two portions of the plank end-to-end, using same thick core assembly adhesive as earlier, incorporating
configurational improvements based on lessons learned from 1st test
•Add two continuous planks along edges, to form the staggered joint configuration
•Vacuum bag over inner and outer assembled plank surfaces, and vacuum drop test

Results:
1. Adhesive application and assembly operations worked well and required minimal labor
2. Initial test after single application of slosh compound: High leak rate
3. Apply slosh compound to cut ends of planks, and re-test: Lower leak rate
4. Apply second coat of slosh compound to insides of all plank portions:
• Still lower leak rate, about 6” Hg in 5 minutes (Close to an acceptable level for a VARTM tool)
• No air leaks were audible at the end-to-end plank joints

Conclusions:
•A configuration has been developed which enables straightforward adhesive-joining assembly of longitudinally
adjacent core planks into a staggered-plank joint, eliminating the shear keys
•The joints between planks are leak-free
•Attention should be given to production of leak-tight pultruded planks for future work. If necessary as a re-work
measure at the pultrusion vendor in order to pass acceptance tests, low-cost liquid sealant slosh compound can be
used to seal up any porosity in the planks.
copyright Lockheed Martin Corporation 2011 58
 Summary of validation results at 4m scale –
Core production and assembly
Component and operations Results and key challenges met Evidence Future Needs

Produce custom pultrusions Successful – Achieved proper Pultruded planks; good fit-up For fabrication of long
designed to meet all initially known fabric guidance, resin wet-out, and dimensional control in CWP on-the-water, need
loads, and meeting requirements of and straightness in complex assembled core ring to add more ribs to
stepwise infusion fabrication multi-cell hollow pultruded increase local buckling
process (at Glasforms) shape made from all-fabric with strength under pressure of
some sparse fiber directions gripper and bushings
Bond pultruded planks into core Successful – Achieved proper Results at multiple locations:
ring spacing to fit shear keys. +/- 1/8” variability in
Achieved good circularity and diameter over the 158”
diameter control diameter
Test vacuum tightness of as- Not successful – A random This was despite the plank Need to improve the leak-
received pultruded planks sample core plank was vacuum manufacturer’s water-tank tightness of the core
bagged on its outside. Drop testing of best and worst planks as-pultruded, or if
testing showed a 6” Hg drop in planks, with application of necessary with
5 minutes. resin on the outside of all to subsequent sealing
seal them.
Test vacuum tightness of plank-to- Successful – The assembly leak Same vacuum drop as the
plank joints (staggered planks tightness is as good as the planks themselves. No
approach) planks themselves audible evidence of air
leakage at longitudinal joints
Produce leak-tight core assembly Successful – Overall drop-test 6” Hg vacuum drop in 5 Improve the leak-tightness
by pre-applying a commercial “slosh leak rate of staggered-joint minutes of the core planks as-
sealing compound” inside each cell assembly is close to acceptable pultruded, and use slosh
of each plank, then assemble into level for VARTM tools. compound as a re-work
staggered-planks configuration, and measure if necessary to
vacuum test ensure leak-free
pultrusions.
copyright Lockheed Martin Corporation 2011 59
 VI.B - Validation 2 – Fabric architecture,
production, overlap splice performance, and
fabric dispensing process

• Pictorial summary of fabric dispensing validation

• VI.B.1 – Fabric architecture
• VI.B.2 - Validation of overlap splice strength
• VI.B.3– Fabric dispensing validation
• Tabular summary of fabric architecture and dispensing validation results at
4m scale –

copyright Lockheed Martin Corporation 2011 60

 Pictorial summary of Fabric Dispensing validation
Function: Fabric is dispensed by machine under constant tension maintained by servomotor-driven
rollers, and is guided by special “bibs” into its curved configuration.
Machine-based fabric placement eliminates hand layup labor and is the key to low recurring costs.
Status: Validation proved that the apparatus forms consistent overlap splices, well within required
tolerances. Outer face sheets
Note worker for scale

Magnitude and consistency of roll-to-roll fabric

overlap during five runs
(Outer dispensers; at top, middle, and bottom
positions)
1
Roll #1-#2

0
Roll #2-#3
0 2 4 6 8 10 12 14
Overlap of zeros, inches

-1
Roll #3-#4
-2
Roll #4-#5
-3
Roll #5-#6
-4

Roll #6-#1
-5

-6 Allowable limit
without gap

-7

Work done at Janicki Industries (Paul

VanSant, Project Manager) under
subcontract from LM-MS2 with
technical direction by LM-SSC.

August-Sept., 2010

Inner face sheets

copyright Lockheed Martin Corporation 2011
61
 VI.B.1 – Fabric architecture
• Custom pre-assembled 3-ply face sheet stitched preform
• Face sheet fabric architecture
• Photo of CWP segment including one fabric splice in each face sheet

copyright Lockheed Martin Corporation 2011 62

 Custom pre-assembled 3-ply face sheet stitched preform

Individual fabric plies at maximum thickness

capability of Owens Corning stitching machines

Face sheet fabric pre-form thickness = 0.120”

warp

0.040”
- 45
Q4100
90
45

warp
- 45
Q4100
90
45

warp
- 45
QM4105
90
0.5 oz/ft^2
Chop mat 45

Pre-assembling the fabric into two thick preforms simplifies the fabric dispensing apparatus
copyright Lockheed Martin Corporation 2011 63
 Face sheet fabric architecture
Each face sheet is divided into two major layers (preforms) of stitch-assembled face sheet fabric

Fabric for 4m CWP

Layer 1 •6 rolls at <100” wide
each cover the
circumference of the
CWP
•Each preform layer is
Layer 2 0.120” thick
•Preform fabric was
stitch-assembled by
Owens Corning from 3
plies of custom stitched
X-Strand fabric
•75% 0’s (axial)
•Balance is 90’s and 45’s
•0’s are left out of the
last 3” along each edge

•Architecture creates generous 6” wide overlap splices for 90’s and 45’s carrying circumferential loads
•Splices in each major layer are reinforced by continuous fabric in other major layer
•By leaving out the 75% zeros in the overlap splices, the thickness increase at the splices is minimized
copyright Lockheed Martin Corporation 2011 64
Photo of CWP segment including one
fabric splice in each face sheet

Thickness increase at face sheet fabric splices is barely detectable, and presents no
problem going through apparatus, gripper, and bushings

copyright Lockheed Martin Corporation 2011 65

 VI.B.2 - Validation of overlap
splice strength

Under a parallel LM OTEC program sponsored by the US Naval Facilities Command, sections of our CWP
design were fabricated and tested under local external pressure, to validate the predictions of its local
buckling resistance. The validation was successful, and is described fully in the Technical Development
Report of that program.

For one of the three fabricated specimens, we included overlap splices per our fabric architecture. As shown
in the next figure, that specimen failed at a location away from the splice (and at about the predicted failure
load) . This indicates that the splice regions held the required in-plane compressive load successfully, and
provides initial validation of our fabric architecture.

copyright Lockheed Martin Corporation 2011 66

 Validation of fabric overlap splice strength
Local pressure buckling test Testing done by Makai Ocean Engg. under LM
program N62583-09-C-0083 sponsored by the
on section of 4m CWP US Navy Facilities Command

Overlap splices

Makai
Ocean
Engg.

The specimen failed away from the splices, at about the expected pressure.
The fabric splices held the required in-plane compressive load. The concept is validated
copyright Lockheed Martin Corporation 2011 67
 VI.B.3– Fabric dispensing
validation
• Configuration of fabric and RDM rolls
• Apparatus during assembly
• Assembled apparatus
• Fabric dispensers computer control system
• Testing constant-payout-tension behavior
• Outer fabric dispensers (typical)
• Outer fabric dispensing, with formation of typical overlap splice
• Inner fabric dispensers (typical)
• Inner fabric dispensing, with formation of typical overlap splice
• Validation of fabric dispensing accuracy
• Tabular summary of fabric architecture and dispensing validation results at
4m scale
copyright Lockheed Martin Corporation 2011 68
 Configuration of fabric and RDM rolls

Outer
Inner

Face sheet
fabric rolls
(2 layers)

RDM rolls

Can put all rolls for Need to put all rolls

any one layer at the for any one layer on
same level two adjacent levels

copyright Lockheed Martin Corporation 2011 69

 Apparatus during assembly
Inner dispensers support frame Outer dispensers support frame

Janicki Industries photos

copyright Lockheed Martin Corporation 2011 70
 Assembled apparatus
(during validation of outer face sheet dispensing)

Inner dispensers frame

Inner face sheet fabric rolls go

here

Outer dispensers frame

Outer face sheet fabric rolls

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 71

 Fabric dispensers computer control system

Janicki
Industries
graphic
Unified control system allows operation of all rolls together or any rolls separately:
•Constant tension during payout (or backing up) with velocity controlled by gripper motions
•Controlled rotational speed if desired
•Zero motion (braking) at all other times, including loss of power

All specified behaviors have been validated during acceptance testing

copyright Lockheed Martin Corporation 2011 72
Testing constant-payout-tension behavior

copyright Lockheed Martin Corporation 2011 73

Outer fabric dispensers (typical)

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 74

Outer fabric dispensing, with formation of
typical overlap splice

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 75
Inner fabric dispensers (typical)

copyright Lockheed Martin Corporation 2011 Janicki Industries photo 76

Inner fabric dispensing, with formation of typical
overlap splice

copyright Lockheed Martin Corporation 2011 Janicki Industries photo 77

 Validation of fabric dispensing accuracy
Magnitude and consistency of roll-to-roll fabric
overlap during five runs
(Outer dispensers; at top, middle, and bottom
positions)
1
Roll #1-#2

0
Roll #2-#3
0 2 4 6 8 10 12 14
Overlap of zeros, inches

-1
Roll #3-#4
-2
Roll #4-#5
-3
Roll #5-#6
-4

Roll #6-#1
-5

-6 Allowable limit
without gap

-7

•The apparatus forms consistent overlap splices, well within required tolerances.
•The fabric dispensing apparatus is validated.
copyright Lockheed Martin Corporation 2011 78
 Summary of validation results at 4m scale – Fabric
architecture and dispensing
Component and operations Results and key challenges met Evidence Future Needs

Overlap splices in face sheets Successful – Splice regions are as Local buckling
strong as the base laminate specimen failed
away from the
splices
Design and construct complex Successful Apparatus Core chucks which do not require periodic
apparatus (w. Janicki) re-charging with air pressure
Acceptance testing of Successful - Maintained constant Acceptance
apparatus (at Janicki) payout tension despite testing COC,
variations in “demand” payout photos, videos
velocity imposed at free end of
fabric
Validate fabric dispensing Successful - Variations in width Validation
operations (at Janicki) of fabric overlap are well below testing report,
allowable photos, videos

copyright Lockheed Martin Corporation 2011 79

 VI.C - Validation 3 –
Stepwise infusion
molding process
• Pictorial summary of stepwise infusion molding validation
• VI.C.1 – Apparatus
• VI.C.2 - Initial validation workpiece
• Tabular summary of stepwise infusion molding validation results at 4m scale
• Sidebar: Effects of axial forces on large seals

copyright Lockheed Martin Corporation 2011 80

 Pictorial summary of stepwise infusion molding validation
Hanging outer face sheet fabric on test core
Workpiece (fabric not Inner and outer hard shells,
shown) soft tools, and upper seals
•
Upper inflatable centering
ring and vacuum seal

Outer and inner soft tools

Outer and inner hard

shells
Resin
Re-Usable Resin
inlet
Distribution Line (RURDL)
Removing infused and cured
Bottom inflatable vacuum workpiece from mold
seals

On-line resin metering and mixing pump with Molding Region showing outer
auto-level control allows use of a small mixed resin resin inlet line (RIL) and solenoids
reservoir, avoiding exotherm concerns. controlling various subsystems

Infusion of
workpiece

•About 1000 lbs. of resin

was mixed and infused in 3
shots

•The fill time was within

about 20% of predicted
(after correcting the resin Key result: The knit-line between infusion steps is indistinguishable
viscosity to the low
temperatures in the from the base laminate. The stepwise infusion process is validated!
16-channel video display
unheated High-Bay) tracks resin flow fronts
inside the mold
•The specialized apparatus
enables infusion to be
done by a 2-person crew
Manual control panel with computer display of step-by-step
instructions and place to record all data and observations
copyright Lockheed Martin Corporation 2011 81
 VI.C.1 – Apparatus

• Molding Region apparatus elements and subsystems

• VI.C.1.a -- Hard shells (outer and inner)
• VI.C.1.b - Soft tools including Re-Usable Resin Distribution Line (RURDL)
• VI.C.1.c - Resin and resin handling elements
• VI.C.1.d - Acceptance testing at Janicki, transport to Sunnyvale, and
installation in B/132 High Bay
• VI.C.1.e - Molding region control system

copyright Lockheed Martin Corporation 2011 82

 Molding Region apparatus elements
and subsystems
Workpiece (fabric not
shown)

Upper inflatable centering

ring and vacuum seal

Outer and inner soft tools

Outer and inner hard

shells
Resin
Re-Usable Resin
inlet
Distribution Line (RURDL)

Bottom inflatable vacuum

seals

Janicki Industries graphic

copyright Lockheed Martin Corporation 2011 83
VI.C.1.a - Hard shells (outer and
inner)

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 84
 VI.C.1.b - Soft tools including
Re-Usable Resin Distribution
Line (RURDL)

• Achieving good sealing at the bottom of the apparatus

• Non-linear FEA was used to design key elements of the soft tools
• Re-usable Resin Distribution Line (RURDL)
• RURDL action
• Installing inner soft tool and RURDL around inner hard shell
• Photos of other elements of the Molding Region
• Durability of soft tool material

copyright Lockheed Martin Corporation 2011 85

 Achieving good sealing at the bottom of the apparatus

Approach:
Mechanical ingredients that seat the soft tool firmly against the
protruding cured end of the workpiece

Results:
•Robust sealing from the start
•Apparatus vacuum level rises to source vacuum level (to within gauge
accuracy) without any corrective actions
•Excellent vacuum is the key to achieving a non-discernable
knitline* during stepwise infusion
•No resin leakage whatsoever

*The knitline is the boundary between the resin cured in successive steps

copyright Lockheed Martin Corporation 2011 86

 Non-linear FEA was used to design key elements of the
soft tools
Main bottom seals analysis is shown here

Stage 1 – Seal Inflation Stage 2 – Fluid Pressure

Head Pressure 16.9 psi

[Absolute Pressure applied to

seal surface above contact
datum.]
Seal Contact Datum

Seal-to-Part
Friction
(µ = 1.0)
Internal pressure
22 psi

Internal Pressure
22 psi

x
Subsequently the pressure
was raised to 32 psi

Janicki Industries subcontractor analysis

copyright Lockheed Martin Corporation 2011 87
 Re-usable Resin Distribution Line (RURDL)
Inflate tubes during infusion Collapse tubes during cure
Fabric
Soft tool
Core

Internal Pressure
+15 psi gage

Bridge-Tool Bond
Rigid “Bridges”
Excess resin is
Resin flow passage squeezed out of
flow passage

Cavity Vacuum
-15 psi gage

Design of RURDL creates large passage for resin flow when air tubes are inflated,
but passage collapses completely when air tubes are deflated
copyright Lockheed Martin Corporation 2011

US Patent Application Publication #13/070,618, March 24, 2011 Janicki Industries graphic
copyright Lockheed Martin Corporation 2011 88
 RURDL action

Tubes inflated to 15 psi; RURDL

passage (on other side of black
Tubes not inflated
“bridges”) opens up as planned

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 89
Installing inner soft tool and RURDL around inner hard shell

Janicki Industries photos

copyright Lockheed Martin Corporation 2011 90
 Photos of other elements of the Molding Region
Inflatable seals Flow front monitoring system
LED lighting video camera port

Inner Soft tools Outer RURDL

Janicki Industries photos

copyright Lockheed Martin Corporation 2011 91
“Flying” the soft tools prove-out test core (foam rings held together with
fiberglass shear keys) into the Molding Region

Janicki Industries photos

copyright Lockheed Martin Corporation 2011 92
 Soft tools drawn back by vacuum

Top seals Top seals engaged with test core

Pressurized
lightly

Janicki Industries photos

copyright Lockheed Martin Corporation 2011 93
 Tests to evaluate durability of soft tool material

Purpose
• The soft tools are re-usable vacuum bags against which the resin cures during each step of the
fabrication process
• If the length per fabrication step is 39 ft. (current baseline), one Cold Water Pipe requires 80 steps
• Replacement of the soft tools during CWP fabrication is difficult
• Therefore tests were run in which samples of the soft tool material (silicone rubber) were
repeatedly cured against laminates made with the Derakane 8084 vinyl ester resin of the CWP
• After each cure, the silicone rubber was peeled away from the cured laminate while the release
force was measured.
• The purpose is to ascertain whether or not one set of soft tools is likely to survive for one CWP
fabrication.

Various possible cure cycles were investigated

• Until seawater fatigue testing for various cure cycles is completed, it is not definite which of the
following cure cycles typical for 8084 will be adequate:
1. Ambient-temperature cure only
2. Ambient-temperature cure + elevated-temperature (175F) post-cure (most likely cycle)
3. Elevated-temperature only (220F)
• Accordingly, the repeated release behavior was evaluated for these three cure cycles.
• The 220F cure produced release forces much higher than the others, and was discontinued.
• Behavior for the other cure cycles is shown in the following two slides

copyright Lockheed Martin Corporation 2011 94

 Durability of soft tool material
Repeated release tests on soft tool material
•Perform repeated 25
RT cure for >12 hrs + Post cure@175F for 6 hrs
layup/cure/release on RT cure for 12 HRS
samples of soft tool 20
Start new RT cure samples at Cycle #29

material

Release force, oz.

15
•Compare:
o Room temperature
10
(RT) cure + 175F post-
cure
vs. 5

oRT cure only 0

0 10 20 30 40 50 60 70 80 90 100

Cycle number

•After 100 cycles, release forces in baseline samples are no higher than at the start
•See next slide for interpretation of anomalous upswing in release forces over last 15 cycles
•Maximum release force with RT cure+175F post-cure is very similar to RT cure only
•Conclusion: We can use RT cure + 175F post cure for on-the-water CWP
fabrication with no detrimental effect on soft tool life compared to RT cure only
copyright Lockheed Martin Corporation 2011 95
 Anomalous behavior seen during last 15 cycles
Room temperature cure only - Baseline specimens Room temperature cure only - New specimens
25
RT cure for 12 HRS - Average started at cycle 29
Coupon 1 [86 releases 40
RT cure for 12 HRS - Average
20
Coupon 2 to this point] 35 Coupon 1
Coupon 3 Coupon 2
Release force, oz.

Release force, oz.

Coupon 4 30
[59 releases
15
25 to this point]
20
10
15

10
5

0 0
70 75 80 85 90 95 100 105 70 75 80 85 90 95 100

Cycle number Cycle number

RT cure + 175F Post-cure

25
•The upswing in release force starts at
cycle 87 of the overall test program
RT cure + 175F Post-cure - Average
Coupon 1
[86 releases to this point]
20
Coupon 2
Coupon 3
•Same point for specimens that have
Release force, oz.

Coupon 4
seen 59 releases or 86 releases
15
•Same point for RT only as for RT+post
10 cure
•These results suggest that the upswing
5
was an artifact of some uncontrolled test
0
variable, and is not an inherent property
70 75 80 85

Cycle number
90 95 100 105
of the soft tool material /cure cycle
copyright Lockheed Martin Corporation 2011 96
 VI.C.1.c - Resin and resin
handling elements
• Resin system
• Resin Inlet Line (RIL) and actuators
• Resin mixing system
• Catalyzed resin supply system
• Catalyzed resin reservoir

copyright Lockheed Martin Corporation 2011 97

 Resin system

http://www.derakane.com/downloadServlet?docPath=DPAPP1.asco.ashland.com%5CData%5Cascc%5Cecomdocsasc.nsf%5C51897
1E5A3AD564485256FCE004AA0B6%5C%24FILE%5C8084.pdf

copyright Lockheed Martin Corporation 2011 98

 Resin Inlet Line (RIL) and actuators

Show photos of
apparatus and RIL

•The only non-reusable element is the RIL itself

•Inexpensive leak-tight fittings and mechanical valves are used
•Shape of the fitting at the end of the RIL causes it to break off at the workpiece.
•Air cylinders press end of RIL against workpiece during cure, to ensure no stub is left.
copyright Lockheed Martin Corporation 2011 99
 Resin mixing system

http://www.mvpind.com/products/specialty-systems/infusion-system.html

copyright Lockheed Martin Corporation 2011 100

 Catalyzed resin supply system
During infusion During cure
From metering pump At bottom
Master Catalyst of molding Master
batch Catalyst
Hard plumbing region batch
Shut-off
Polyflow tubing valves

Clean out
Pump Turbulent turbulent mixer
“Autolevel”
flow mixer elsewhere and
control re-use. Replace
sensor (air RIL tubing and
pressure fittings after
polyflow tubing each step.
actuated)
Thin diaphragm keeps liquid at
ambient pressure but prevents
styrene fumes from escaping
Device for securing lines

5-gallon plastic pail

On lower platform On lower platform

Design of resin supply system allows huge workpiece to be filled quickly without
requiring large batches of catalyzed resin that might be prone to exotherm. About
1000 lbs. of resin was mixed and infused in 3 shots, with no problems.
copyright Lockheed Martin Corporation 2011 101
 Catalyzed resin reservoir

Mixing gun

Swirl tube

Air line for autolevel control

At the end of the run, we bring the

liquid level down to minimize the
leftover catalyzed resin

copyright Lockheed Martin Corporation 2011 102

 VI.C.1.d - Acceptance testing at
Janicki, transport to Sunnyvale,
and installation in B/132 High Bay
• Molding Region during acceptance test at Janicki Industries
• Molding region arrives in Sunnyvale for installation in B/132
• Getting a 16’5” wide part through a 16’ wide rollup door
• “Lab supports” for inner and outer molding regions
• Installing outer lab supports
• Lifting inner mold back into outer mold
• Installing inner lab supports

copyright Lockheed Martin Corporation 2011 103

Molding Region during acceptance test at Janicki Industries
(8-11-2010)
Final half of test was
conducted outside to
avoid possible
damage to other work
in Janicki’s Hamilton
facility just in case any
water leaked out. It
didn’t.

OMR lab supports

were used to help
stabilize the molding
region on the bed of
the low boy truck.

Janicki Industries photo

copyright Lockheed Martin Corporation 2011 104
Molding region arrives in Sunnyvale for installation in B/132

copyright Lockheed Martin Corporation 2011 105

Getting a 16’5” wide part through a 16’ wide rollup door

copyright Lockheed Martin Corporation 2011 106

“Lab supports” for inner
and outer molding
regions

Janicki Industries graphic

copyright Lockheed Martin Corporation 2011 107
Installing outer lab supports

copyright Lockheed Martin Corporation 2011 108

Lifting inner mold back into
outer mold

copyright Lockheed Martin Corporation 2011 109

Installing inner lab supports

copyright Lockheed Martin Corporation 2011 110

 VI.C.1.e - Molding region
control system
• Control system approach for current program
• Control system components
• Resin flow-front height visual sensing system
• Installation of control system elements
• Control system - Installed

copyright Lockheed Martin Corporation 2011 111

 Control system approach for
current program
• Control all principal subsystems from a centralized location using clear, easy-
to-follow step-by-step settings
• Install and use actuators and sensors that will become part of the full
integrated CWP fabrication apparatus in the future
• Control these actuators manually by simple and inexpensive toggle switches
mounted in a panel. Read out the sensors by simple and inexpensive analog
voltmeters.
• Provide step-by-step instructions for operation of the control system by a
computer next to the control panel
• The graphical and macro-based features of this Excel program were written by Matt Ascari
• Use the same computer for logging in of data, results, and comments during
the runs

The alternative would have been a full computer control system for the Molding Region, and this was
judged to be out-of-scope and not needed for the current program, in which flexibility was preferred
over automation.
When the Molding Region is combined with the Fabric Dispensing System, we may expand the latter’s
computer control system to also cover the molding operations

copyright Lockheed Martin Corporation 2011 112

 Control system components
Vacuum regulators Air
Pressure or pressure
Solenoid valves vacuum sensors regulators

1/2”
NPT

2”
NPT
Panel meters
Toggle switches

Platform
scale

copyright Lockheed Martin Corporation 2011 113

 Resin flow-front height visual
sensing system

16-channel video display

and recorder
Sight glasses built into 16 Sony HD230 “bullet” cameras
hard shells of molding with wide-angle f2.1 lens
region

copyright Lockheed Martin Corporation 2011 114

Installation of control system elements

copyright Lockheed Martin Corporation 2011 115

Control system - Installed

copyright Lockheed Martin Corporation 2011 116

 VI.C.2 - Initial validation
workpiece
• VI.C.2.a - Layup and insertion into Molding Region
• VI.C.2.b - Infusion of workpiece
• VI.C.2.c - Results on infused workpiece
• VI.C.2.d - Validation of analytical tools for predicting infusion behavior

copyright Lockheed Martin Corporation 2011 117

 VI.C.2.a - Layup and insertion
into Molding Region

• Manual fabric layup on foam core

• Inner fabric layup
• Resin distribution medium
• Adding Resin Distribution Medium (RDM)
• Lift workpiece into mold
• Installing core stoppers and temporary upper seal-flip-prevention devices
• Install lid

copyright Lockheed Martin Corporation 2011 118

 Initial validation workpiece -- Manual fabric layup on foam core*

Hang fabric plies (over nylon release film - allows re-use of foam core just in case)
Outer fabric layup

*Because of delays receiving acceptable pultruded planks, the foam core that had been used in Molding Region
acceptance testing was used for the first Molding Region infusion trials instead of the pultruded plank core ring
copyright Lockheed Martin Corporation 2011 119
Inner fabric layup

copyright Lockheed Martin Corporation 2011 120

 Resin distribution medium –
Colbond EnkaRain Drain 9714

Active portion of RDM is removed wherever we want the resin to stop

copyright Lockheed Martin Corporation 2011 121
Add Resin Distribution Medium (RDM)

Top of active outer RDM was intentionally

“castellated” to check that resin stops controllably

copyright Lockheed Martin Corporation 2011 122

Lift workpiece into mold

copyright Lockheed Martin Corporation 2011 123

Installing 4x4 core stoppers and temporary upper seal-flip-prevention
devices (foam blocks and foam strips) on top of workpiece

copyright Lockheed Martin Corporation 2011 124

 Install lid

copyright Lockheed Martin Corporation 2011 125

 VI.C.2.b - Infusion of workpiece
• Labor requirements for infusion of workpieces
• Video monitor, regulators and gages, control panel, and computer that
provides step-by-step running procedures and data note-taking
• Vacuum levels achieved
• Configuration of the three infusions
• Photos concerning the infusion process
• Infusion #1: Inner facesheet – Cured lower portion
• Infusion #2: Inner facesheet - Preparation to infuse remainder
• Infusion #3: Outer facesheet, full height - Monitoring flow front position

copyright Lockheed Martin Corporation 2011 126

 Labor
requirements
for infusion of
workpieces

•The specialized apparatus

enables infusion to be
done by a 2-person crew

•About 1000 lbs. of resin

was mixed and infused in 3
shots

copyright Lockheed Martin Corporation 2011 127

 Video monitor, regulators and gages, control panel, and
computer that provides step-by-step running procedures and
data note-taking

copyright Lockheed Martin Corporation 2011 128

 Vacuum levels achieved

Source vacuum gage reads 26” Hg Vacuum chamber gage reads 27.5” Hg

Excellent vacuum achieved is the key to high-quality VARTM with no visible knitline
copyright Lockheed Martin Corporation 2011 129
Configuration of the three infusions
(only one side is shown)

copyright Lockheed Martin Corporation 2011 130

Photos concerning the infusion
process

copyright Lockheed Martin Corporation 2011 131

 Infusion #1: Inner facesheet –
Cured lower portion
•This edge
becomes the
“knitline” after
the rest of the
inner facesheet
is infused

•Face sheet
filled to a
uniform height
all around

copyright Lockheed Martin Corporation 2011 132

 Infusion #2: Inner facesheet -
Preparation to infuse remainder
Insertion of extra RDM to carry resin from the RURDL up to the not-yet infused region

copyright Lockheed Martin Corporation 2011 133

 Infusion #3: Outer facesheet, full height -
Monitoring flow front position
0 deg completing 180 deg completing
Before infusion fill at top fill at top

copyright Lockheed Martin Corporation 2011 134

 VI.C.2.c - Results on infused
workpiece
• Removing infused and cured workpiece from mold
• Inner face sheet showing uniform wet-out height
• Control of wet-out height
• Observations concerning longitudinal wrinkling in outer face sheet (only) of
4m validation workpiece fabricated on foam core
• Understanding wrinkling observed in outer face sheet of 4m workpiece, and
path forward that should eliminate it
• Inner face sheet shows successful operation of RURDL and RIL mechanism
• Samples cut from finished 4m diameter validation workpiece
• Results at knitline

copyright Lockheed Martin Corporation 2011 135

Removing infused and cured workpiece from mold

copyright Lockheed Martin Corporation 2011 136

 Inner face sheet showing uniform wet-out height
•As individual fabric tows
continue to soak up resin
after fill valve is closed, the
resin in the RDM drops to a
common liquid level, but the
structural fabric stays filled
at the maximum resin height.
The RDM appearance can
probably be improved by
keeping the fill valve open
longer than we did on this
run.

•Ring patterns are from shear

keys holding the foam core
rings together, and overlaps
of the double RDM. These
are not part of the CWP
design

•The flow front stops at a uniform height all around the circumference
•This validates our method of controlling the flow front height for stepwise infusion
copyright Lockheed Martin Corporation 2011 137
 Control of wet-out height
Boundary of wet-out in the fabric

As individual fabric
tows continue to soak
up resin after fill valve
is closed, the resin in
the RDM drops to a
common liquid level,
but the structural
fabric stays filled (see
next slide). The
appearance can
probably be improved
by keeping the fill
valve open longer
than we did on this
run.

The outer fabric wet out to a uniform

copyright Lockheedmargin of2011
Martin Corporation about 4 inches above the
intentionally placed “castellations” at the top of the active RDM, again
indicating good uniformity and control of wet-out height within the apparatus
138
copyright Lockheed Martin Corporation 2011
 Observations concerning longitudinal wrinkling in outer face
sheet (only) of 4m validation workpiece fabricated on foam core

Proof-of-principles workpiece: 4m validation workpiece:

No longitudinal wrinkling Got longitudinal wrinkling

copyright Lockheed Martin Corporation 2011 139

 Understanding wrinkling observed in outer face sheet of 4m
workpiece, and path forward that should eliminate it
Component Prior and Details of prior or current operations Interpretation Path forward
and current
operations results
Outer fabric 19” Proof- 1. Bottom edges of all fabric rolls were 1. Individual fabric strips
dispensing of-Principles attached to a “pull ring” and were started out very close to
and molding apparatus: pulled into the molding region axially their final diameter, and
No from fabric rolls and directly in contact were free to slide over
longitudinal with the core. No tackifier was needed each other
wrinkles or used. circumferentially as they
debulked onto the core
when vacuum was
pulled.
Outer fabric 4m Molding 1. We used an available foam core because 1. The breather debulks 1. There will not be any
support Region the pultrusions were late. substantially when breather fabric with
materials tested by 2. To be sure we didn’t crush the weak vacuum bag pressure is regular pultruded core.
and layup itself: foam core by possible resin shrinkage applied, forming 2. Fabric dispensing bibs
Got effects, we put breather cloth under the wrinkles in the overlaid should keep fabric
longitudinal outer face sheet. This inadvertently fabric and RDM. against core as it is
wrinkles in created excess circumference in the 2. The billowing from dispensed vertically, and
outer face fabric and RDM. draping probably also tension in dispensing
sheet 3. Because this was a standalone contributed to wrinkling mechanisms should
workpiece, individual fabric strips were 3. Circumference changes tend to keep it from
draped over the top of the foam core. at shear keys may have wrinkling.
They billowed out, creating additional contributed to wrinkling 3. We plan to get rid of
excess circumference. 4. The tackifier may have shear keys by going to
4. Fabric also hung away from workpiece in prevented the fabric staggered core planks.
between the intermediate shear keys, strips from sliding over This will get rid of
creating additional excess circumference each other at their excess circumference
5. Some sprayable tackifier was used to edges, contributing to there.
help hold the fabric strips in place. wrinkling. 4. No tackifier will be used
copyright Lockheed Martin Corporation 2011 140
 Inner face sheet shows successful operation of
RURDL and RIL mechanism

Key results:
•Insignificant resin stub left on cured workpiece at RIL
•Barely discernable pattern left on cured workpiece at RURDL.

Behavior of the apparatus is validated for repeated infusions

with minimal expendable material replacement (only the RIL
itself) between infusions
141
copyright Lockheed Martin Corporation 2011
 Samples cut from finished 4m diameter validation workpiece

RDM side Core side

Top of last infusion

Laminate is fully wet-out everywhere, including regions

where resin level dropped in the RDM during cure

copyright Lockheed Martin Corporation 2011 142

 Results at knitline
Key result: The knit-line between infusion steps is
indistinguishable from the base laminate.

A slight color change on the external surface of the laminate

(due to different catalyst % on two infusions due to lab
temperature differences) serendipitously reveals where the
knitline is.

The stepwise infusion molding process is validated

copyright Lockheed Martin Corporation 2011 143
 VI.C.2.d - Validation of analytical tools for
predicting infusion behavior
(fill time and resin consumption)

• Predicting resin quantity and fill time including lines and valves
• Fill pattern and resin consumption
• Fill time corrected for viscosity vs. temperature, compared to prediction

copyright Lockheed Martin Corporation 2011 144

 Predicting resin quantity and fill time including lines and valves
Input from LMSSC flat panel VARTM model
Number of panels 1 Mixture, in^3Mixture, cc Calculate acceptable size of resin inlet line, subfeeds, and fill valve
Expected fill time without
Panel length, in 82.00 64" of fabric + RDM, + anot Total resin Total resin lines and valves Avg flow rate
Panel width, in 501.39936 cc. g min. sec. g/sec.
Fabric thickness, in 0.237 172575 176053 9.1667 550.002 320.0959675 Fill time without feed, subfeeds, or valve (VART
Margin in bottom of reservoir: 1/2" depth x 2"
diameter: in^3 0.0 0 Fluid density 1.020E+06 g/m^3 8084
RIL inside diam., in. 1.000 gravitational constant 1.00E+02 g/(Pa-sec^2-m)
RIL inside area, in^2 0.785
RIL length, in. 184 Pressure drop through main feed tube aninches meters
RIL volume, in^3, cc 144.4 2366.9 Tube length 184 4.67360
RDL length, ID, in. 501.40 wall thickness, outer diam 0.0675
RDL ID, in 1.316 Inside diameter, inches "=D9-2*C9" 1 0.02540
RDL area (in^2) 1.360 Flow area 0.7850 5.06E-04
RDL vol in^3 , cc 681.9 11174.3 Fluid viscosity 0.37 Pa-sec 8084 at 77F
Active RDM thickness, in 0.12 Average mass flow rate 320.0959675 g/sec Fluid velocity 0.48634606 m/sec
Active RDM porosity 0.88 Reynolds number 340.6000932 Per Shepard
Active RDM mixture quantity, in^3 4347.3 71239.0 Friction factor f 0.187903648 Per Marks p. Friction factor prime (=f/4) 0.046976 Per Marks p. 4-66
Outer RDM thickness, in 0.02 Estimated pressure drop for feed tube 4.17E+04 Pa
Outer RDM porosity 0.67 Feed tube mass (g), flow 301.57 767.36 0.39
Outer RDM mixture quantity, in^3 550.9 9028.2
VDM mixture quantity, in^3 0.0 0 Pressure drop through fill valve
Fabric mixture content 0.49328 Measure laminate fiber v/o Length of flow passage inside valve body 1 0.02540 was =2*0.75*D20/0.375
Fabric mixture volume, in^3, cc 4806.6 78766.19093 Tube nominal size, OD, inches
Resin in gaps between planks, in^3 0.0 Flow passage inside diameter, inches 0.75 0.01905
VDL length, in., vol in^3 0 0.0 0 Flow area 0.4416 2.85E-04
VEL's: How many? 0 Fluid viscosity 0.37 Pa-sec 8084 at 77F
VEL inside diam, in. 0.170 Average mass flow rate 320.0959675 g/sec Fluid velocity 8.646E-01 m/sec
Total VEL length, volume 48 0.0 0 Reynolds number 4.54E+02 Per Shepard
Fill time, Upper reservoir volume, fabric fraction un 0 0.0 0% Friction factor f 0.140927736 Per Marks p. Friction factor prime (=f/4) 0.035232 Per Marks p. 4-66
Total calculated mixture: in^3, cc 10531.2 172575 Estimated pressure drop for valve 7.17E+02 Pa

Flushing margin and mixture required: 0% 0.0 0 Pressure drop through RDL inches meters
Contingency and mixture required: 0% 0.0 0 Number of sub-feed tubes 2
Total mixture to be mixed, lb, in^3, cc 380.1 10531.2 172575 Tube length 250.69968 6.36777
wall thickness, outer diam 0 1.31590339
Catalyst PHMB by wgt; Catalyst PHMB by vol 1.986097319 1.899705173 Inside diameter, inches "=D31-2*C31" 1.316 0.03342
Density(g/cc)-Master batch, avg, hardener 1.01305008 1.020 1.039 Flow area 1.360 8.77E-04
Master batch (g) 169214 Catalyst (g) 3361 Fluid viscosity 0.37 Pa-sec 8084 at 77F

Master batch (lb) 373 Catalyst (lb) 7.403 Average mass flow rate 160.0479838 g/sec Fluid velocity 1.404E-01 m/sec
Master batch (cc) 167034 Catalyst (cc) 3235 Reynolds number 1.29E+02 Per Shepard
Master batch (gallons) 44.1 Catalyst (ga 0.9 Friction factor f 0.494526095 Per Marks p. Friction factor prime (=f/4) 0.123632 Per Marks p. 4-66
Estimated pressure drop for subfeed 9.48E+03 Pa
subfeed tube mass (g), fl 1992 625.29 3.19
Total pressure drop for valve, feed, subfeeds 5.19E+04 Pa
Atmospheric pressure 1.00E+05 Pa
Total Pressure to fill part in same time 1.52E+05 Pa

Corrected fill time for 1 atm avail pressure 13.92 min. 835 sec.
Estimated pot life of resin 114 min.

copyright Lockheed Martin Corporation 2011 145

 Fill pattern and resin consumption
Inner face sheet, upper portion
FF position, 0 degree side
FF position, 180 degree side
Resin weight consumed, lbs
100 400
90 350

Resin weight consumed, lbs.

Flow Front position, inches

80
300
70
60 250
50 200
40 150
30
100
20
10 50
0 0
0:00:00 0:15:00 0:30:00 0:45:00
Time since opening fill valve, H:M:S
•Actual resin consumed (360 lb) agrees almost exactly with predicted (373 lb)
•Rapid fill at 180 degrees (247 inches) away from RIL indicates that RURDL worked as planned
copyright Lockheed Martin Corporation 2011 146
 Fill time corrected for viscosity vs.
temperature, compared to prediction
Used generic
data on
viscosity of
vinyl ester
2400 cp at 65F resins vs.
temperature.
1300 cp at 77F
Derakane 8084
is a 40%
styrene resin
65F 77F
http://scholar.lib.vt.edu/th
eses/available/etd-42198-
113329/unrestricted/ch3-
3.pdf
Derakane 8084 has 40% styrene

•Predicted fill time at 77F = 14 minutes (LMSSC VARTM model; panel data)
•Fill time increase due to viscosity increase at 65F temperature = 2400/1300 = 1.85
•Predicted fill time corrected to 65F = 1.85 x 14 = 25.9 minutes
•Actual fill time = 31 minutes
Predicted fill time (corrected to actual temperature) is within 20% of measured fill time
copyright Lockheed Martin Corporation 2011 147
 Tabular summary of validation results at 4m scale -
Stepwise infusion molding
Component and operations Results and key challenges met Evidence

Vacuum testing of hard shells Successful - Achieved good vacuum in assembly after Acceptance testing COC
(at Janicki) welding quadrants together
Acceptance testing of basic Successful - Achieved good vacuum; operated RURDL’s; Acceptance testing COC, photos, videos
apparatus (at Janicki) filled fully with pressurizing fluid (water)
Resin height monitoring system Successful – Can see resin through soft tools Photos and clips from 16-channel video
(at LMSSC) recorder
Control system (at LMSSC) Successful – Operates all sub-systems from a central control Notes within control system software
panel records
Resin supply, on-line mixing, Successful – Mixes and feeds large quantities without Complete fill and cure of workpiece as
and resin feed system (at exotherm planned, negligible temperature rise in
LMSSC) curing workpiece
Functional test of RURDL and Successful - RURDL spreads resin circumferentially, then 2-step infusion filled as planned, leaving
RIL (at LMSSC) collapses. RIL breaks off as planned. negligible trace of excess cured resin on
workpiece
Pull good vacuum on Successful – Double lower seals arrangement was Vacuum gage readings at apparatus are
workpiece (at LMSSC) extremely effective in avoiding any air leakage past “cured very close to supply vacuum
CWP”
Infuse large workpiece (at Successful – Infused about 1000 lbs. of resin in 3 “shots”. Overall visual inspection of workpiece.
LMSSC) Got full wetout of fabric everywhere planned, in about Data on flow front position vs. time.
the planned time (correcting for temperature). Flow front
stopped at expected locations
Cure shrinkage behavior of Successful – Inner face sheet stays in close contact with Small cut-out samples
large workpiece core and does not pull away
Stepwise infusion molding (at Successful – “Knit-line” between infusion steps is not Knit-line cut-out sample
LMSSC) discernable visually.
copyright Lockheed Martin Corporation 2011 148
 Sidebar: Effects of axial forces on large seals
Component Results Evidence and diagnosis Future Needs
and operations
3. Infusion
molding

Large inflatable •Outer bottom Outer seals flipped out of position axially, causing major Janicki and LMSCC convened a
seals under seal leaked while water leakage. Both outer bottom seals were “flipped” over telecon including all individuals
axial forces filling with water about 90 degrees of arc. Outer RURDL disbonded from outer familiar with the work. All agreed
from water to the top (before soft tool. that the root cause was the
column and second step of LMSSC preliminary diagnosis (concurred with by Janicki) is instability in the large inflatable
vacuum (at inner stepwise that outer soft tool slipped downward on the RDM of the seals under strong axial forces,
LMSSC) infusion). workpiece under the large axial forces induced by the water and that an improvement was
It had held just column, in a way that did not happen during acceptance needed to stabilize them.
fine during first testing at Janicki because during that test, the soft tools were Two possibilities were identified,
step, which was to pressing directly on the foam test core, not the more slippery both designed to prevent axial
a lower water RDM. movement of their non-anchored
elevation. Downward slippage of the outer soft tool tore it from the sides:
It had held just outer RURDL, which was supported by straps from the outer a. Add diagonal straps over
fine during Janicki hard shell. seals, connected to hard
acceptance Upper “centering ring” seals, also under axial forces because shells
testing. of 1 atm vacuum pressure differential across them, did not b. Mold 2nd-generation axial
•Outer RURDL flip out because of foam blocks inserted above them to seals to a triangular shape
disbonded from protect against such an event. (Note: One upper seal had
outer soft tool. flipped out at Janicki before foam blocks were inserted.)

Despite this fairly major failure of one mechanical element of the apparatus, we were able to recover and complete
the molding region validation. The inner soft tool and inner RURDL were undamaged, and we validated the RURDL
behavior and stepwise infusion process on the inner face sheet. To complete the outer face sheet molding validation,
we installed an ad-hoc disposable RDL consisting of several turns of standard “spiral-wrap” RDL, with some extra
spiral wrap at the RIL to form a large open interface. Infusion of the outer face sheet was then successful.
copyright Lockheed Martin Corporation 2011 149
It’s time to declare victory for this phase of the CWP development

Nat Shankar, Matt Ascari, Alan Miller, Ted Rosario

copyright Lockheed Martin Corporation 2011 150
 VII. Path forward

The next figure summarizes the overall CWP roadmap in terms of past and recently
completed steps, and future planned steps.

copyright Lockheed Martin Corporation 2011 151

 LM OTEC CWP development and implementation roadmap

CWP Trade study Proof-of-Principle 4m Prototype Scale-up Pilot Plant 100 MW Plant
demo (PoP) CWP
Complete
Complete
Jan-May 2008 June-Dec. 2008 Jan 2009 to Dec. 2010 TBD TBD TBD
Activity: Select •Small-scale (18” •Large-scale (4m ID) •Large-scale •1000m CWP •1000m CWP
•Material ID) validation of validation of critical validation of attached to attached to 100
•Configuration fabrication process elements integrated process TBD platform MW platform
•Fabrication process •Complete •Collect recurring
labor cost data
Basis and size: 0.5m ID x 34” per 4m ID x 7.5 ft per step 4m ID x 19 ft per 4m ID x 19 or 10m ID x 19 or
10m ID x 1000m long step step, 3 or more 39 ft. per step 39 ft. per step
CWP for 100 MW successive steps

Location: LMSSC-ATC LMSSC-ATC and LMSSC Sunnyvale TBD Hawaii Hawaii

and Sunnyvale Sunnyvale B/132
Funding: MS2 LTTI IRAD AWPP-DoE/MS2 TBD TBD Commercial
152
MS2 LTTI IRAD CRADA copyright Lockheed Martin Corporation 2011 project
 VIII. Executive Summary
Critical elements of the LM OTEC CWP design and
fabrication apparatus and process have been validated at a
4m diameter scale. These elements include:

•Core plank production and assembly into core rings

•Fabric architecture, fabric dispensing, and overlap
splice performance
•Stepwise infusion molding

The following slides summarize the major results and

conclusions, at the level of one photo montage and one
table per validation element.
copyright Lockheed Martin Corporation 2011 153
 Pictorial summary of LM Advanced Composite Cold Water Pipe Approach
Fabrication off of the OTEC platform eliminates major
Integral one-piece sandwich-wall pipe with no joints deployment risks and enables affordable fabrication of
in the main load-bearing face sheets maximizes huge integral pipe
durability and reliability

Stepwise VARTM process enables fabrication of a Concept design of large-scale apparatus helped select
1000-m long one-piece CWP critical elements for validation at 4m diameter scale

Pultruded core
ring assembly

Inner and outer

fabric dispensers
inside vacuum
chamber

Molding
region

copyright Lockheed Martin Corporation 2011 154

Pictorial summary of core assembly validation

copyright Lockheed Martin Corporation 2011 155

 Tabular summary of validation results at 4m scale –
Core production and assembly
Component and operations Results and key challenges met Evidence Future Needs

Produce custom pultrusions Successful – Achieved proper Pultruded planks; good fit-up For fabrication of long
designed to meet all initially known fabric guidance, resin wet-out, and dimensional control in CWP on-the-water, need
loads, and meeting requirements of and straightness in complex assembled core ring to add more ribs to
stepwise infusion fabrication multi-cell hollow pultruded increase local buckling
process (at Glasforms) shape made from all-fabric with strength under pressure of
some sparse fiber directions gripper and bushings
Bond pultruded planks into core Successful – Achieved proper Results at multiple locations:
ring spacing to fit shear keys. +/- 1/8” variability in
Achieved good circularity and diameter over the 158”
diameter control diameter
Test vacuum tightness of as- Not successful – A random This was despite the plank Need to improve the leak-
received pultruded planks sample core plank was vacuum manufacturer’s water-tank tightness of the core
bagged on its outside. Drop testing of best and worst planks as-pultruded, or if
testing showed a 6” Hg drop in planks, with application of necessary with
5 minutes. resin on the outside of all to subsequent sealing
seal them.
Test vacuum tightness of plank-to- Successful – The assembly leak Same vacuum drop as the
plank joints (staggered planks tightness is as good as the planks themselves. No
approach) planks themselves audible evidence of air
leakage at longitudinal joints
Produce leak-tight core assembly Successful – Overall drop-test 6” Hg vacuum drop in 5 Improve the leak-tightness
by pre-applying a commercial “slosh leak rate of staggered-joint minutes of the core planks as-
sealing compound” inside each cell assembly is close to acceptable pultruded, and use slosh
of each plank, then assemble into level for VARTM tools. compound as a re-work
staggered-planks configuration, and measure if necessary to
vacuum test ensure leak-free
pultrusions.
copyright Lockheed Martin Corporation 2011 156
 Pictorial summary of Fabric Dispensing validation
Function: Fabric is dispensed by machine under constant tension maintained by servomotor-driven
rollers, and is guided by special “bibs” into its curved configuration.
Machine-based fabric placement eliminates hand layup labor and is the key to low recurring costs.
Status: Validation proved that the apparatus forms consistent overlap splices, well within required
tolerances. Outer face sheets
Note worker for scale

Magnitude and consistency of roll-to-roll fabric

overlap during five runs
(Outer dispensers; at top, middle, and bottom
positions)
1
Roll #1-#2

0
Roll #2-#3
0 2 4 6 8 10 12 14
Overlap of zeros, inches

-1
Roll #3-#4
-2
Roll #4-#5
-3
Roll #5-#6
-4

Roll #6-#1
-5

-6 Allowable limit
without gap

-7

Work done at Janicki Industries (Paul

VanSant, Project Manager) under
subcontract from LM-MS2 with
technical direction by LM-SSC.

August-Sept., 2010

Inner face sheets

copyright Lockheed Martin Corporation 2011 157
 Tabular summary of validation results at 4m scale –
Fabric architecture and dispensing
Component and operations Results and key challenges met Evidence Future Needs

Overlap splices in face sheets Successful – Splice regions are as Local buckling
strong as the base laminate specimen failed
away from the
splices
Design and construct complex Successful Apparatus Core chucks which do not require periodic
apparatus (w. Janicki) re-charging with air pressure
Acceptance testing of Successful - Maintained constant Acceptance
apparatus (at Janicki) payout tension despite testing COC,
variations in “demand” payout photos, videos
velocity imposed at free end of
fabric
Validate fabric dispensing Successful - Variations in width Validation
operations (at Janicki) of fabric overlap are well below testing report,
allowable photos, videos

copyright Lockheed Martin Corporation 2011 158

 Pictorial summary of stepwise infusion molding validation
Hanging outer face sheet fabric on test core
Workpiece (fabric not Inner and outer soft tools
shown) and upper seals
•
Upper inflatable centering
ring and vacuum seal

Outer and inner soft tools

Outer and inner hard

shells
Resin
Re-Usable Resin
inlet
Distribution Line (RURDL)
Removing infused and cured
Bottom inflatable vacuum workpiece from mold
seals

On-line resin metering and mixing pump with Molding Region showing outer
auto-level control allows use of a small mixed resin resin inlet line (RIL) and solenoids
reservoir, avoiding exotherm concerns. controlling various subsystems

Infusion of
workpiece

•About 1000 lbs. of resin

was mixed and infused in 3
shots

•The fill time was within

about 20% of predicted
(after correcting the resin Key result: The knit-line between infusion steps is indistinguishable
viscosity to the low
temperatures in the from the base laminate. The stepwise infusion process is validated!
16-channel video display
unheated High-Bay) tracks resin flow fronts
inside the mold
•The specialized apparatus
enables infusion to be
done by a 2-person crew
Manual control panel with computer display of step-by-step
instructions and place to record all data and observations
copyright Lockheed Martin Corporation 2011 159
 Tabular summary of validation results at 4m scale -
Stepwise infusion molding
Component and operations Results and key challenges met Evidence

Vacuum testing of hard shells Successful - Achieved good vacuum in assembly after Acceptance testing COC
(at Janicki) welding quadrants together
Acceptance testing of basic Successful - Achieved good vacuum; operated RURDL’s; Acceptance testing COC, photos, videos
apparatus (at Janicki) filled fully with pressurizing fluid (water)
Resin height monitoring system Successful – Can see resin through soft tools Photos and clips from 16-channel video
(at LMSSC) recorder
Control system (at LMSSC) Successful – Operates all sub-systems from a central control Notes within control system software
panel records
Resin supply, on-line mixing, Successful – Mixes and feeds large quantities without Complete fill and cure of workpiece as
and resin feed system (at exotherm planned, negligible temperature rise in
LMSSC) curing workpiece
Functional test of RURDL and Successful - RURDL spreads resin circumferentially, then 2-step infusion filled as planned, leaving
RIL (at LMSSC) collapses. RIL breaks off as planned. negligible trace of excess cured resin on
workpiece
Pull good vacuum on Successful – Double lower seals arrangement was Vacuum gage readings at apparatus are
workpiece (at LMSSC) extremely effective in avoiding any air leakage past “cured very close to supply vacuum
CWP”
Infuse large workpiece (at Successful – Infused about 1000 lbs. of resin in 3 “shots”. Overall visual inspection of workpiece.
LMSSC) Got full wetout of fabric everywhere planned, in about Data on flow front position vs. time.
the planned time (correcting for temperature). Flow front
stopped at expected locations
Cure shrinkage behavior of Successful – Inner face sheet stays in close contact with Small cut-out samples
large workpiece core and does not pull away
Stepwise infusion molding (at Successful – “Knit-line” between infusion steps is not Knit-line cut-out sample
LMSSC) discernable visually.
copyright Lockheed Martin Corporation 2011 160