Composite Filament Winding Copyright © 2011, ASM International®

S.T. Peters, editor All rights reserved.

www.asminternational.org

CHAPTER 1

Filament Winding—
Introduction and Overview
S.T. Peters, Process Research, Consultants

Introduction significant number of papers on rocket motors/

pressure vessels have had South Korea as their
The objective of this book is the facilitation source.
of fabrication of useful composite shapes by the The advances in machine control give the
most efficient and least costly method: filament winders an opportunity to wind noncylindri-
winding. A previous book on filament wind- cal and nonsymmetric objects and allow for a
ing, Filament Winding; Composite Structure much more precise laydown of fiber. The higher
Fabrication (Ref 1.1), was published in 1994 capacities of computer memory that is now
by the Society for the Advancement of Material available allow the control of up to six axes of
and Process Engineering. It was aimed at the motion: mandrel rotation, cross feed, horizontal
large audience of filament winders who were carriage movement, vertical carriage movement,
using two- and three-axis machines with either wind eye rotation, and wind eye yaw (Fig. 1.1).
mechanical or computer controls. The audience Control of ancillary components is now com-
has changed. Now, most winders have computer mon so that there can be automated mandrel
controls and at least three axes. Winding with placement, automatic tie-on and cut-off, auto-
four axes is common because the shapes of the mated loading of wet fiber-covered mandrels into
products have evolved to include more complex- the curing oven, and finally, automated extrac-
ity. However, the automation used on the wind- tion. The new machine systems have allowed
ing machine and ancillary components does not pipe manufacturers the ability to wind rather
eliminate the need for proper fiber handling. than hand lay-up, spray-up, resin transfer mold,
This book is a primer on how to use the new or compression mold their elbows, reducers, and
machines, starting with an overview of what tees. Filament winding these components results
is available for automation, a look at what has in better mechanical properties through the use
been done, and a primer on handling the fibers of continuous fibers, with the accompanying
with minimum damage prior to laydown. lowering of manufacturing costs (Ref 1.2).

Computer and Computer Controls Composite Analysis

There has been great progress in machine Increasingly better methods for composite
automation during the last 16 years. Much of analysis have given designers a quicker, more
the progress can be attributed to the impact of reliable answer for closed-form analysis. There
computers, although much of this work has been are “B”-basis allowable data for fiber/resin com-
accomplished outside of the United States. For binations for 121 °C (250 °F)-curing prepreg for
example, most publications in the realm of rein- the aircraft composite component industry (e.g.,
forced piping have originated in Canada, and a AGATE) (Ref 1.3), but there is no comparable
2 / Composite Filament Winding

Fig. 1.1 Schematic illustration of the manufacture of a tapered box beam on a computer-controlled six-axis filament winder.
Courtesy of Mclean Anderson. Source: Ref 1.1, © S.T. Peters, 2009

database for filament-wound composites that use Bottle, enabling dome and skirt testing and opti-
prepreg tow and wet resin-impregnated tow or mization as well as providing information on the
roving. The principles of composite analysis and pressure vessel. These subjects are more fully
“B”-basis allowables are discussed more exten- elaborated on in Chapter 8.
sively in Chapters 8 and 7, respectively. Two new designs for small pressure vessels
have gained wider use. Isotensoid vessels (Fig.
1.2) are essentially composite pressure vessels
Structural Analysis with Isotensoid domes but without a cylindri-
cal section. They are proposed for use as liquid
Structural analysis has also progressed so propane gas containers and would be placed in
that structures can be modeled and the model the automobile spare wheel well. For automotive
transferred into any of several finite-element use, rigorous tests must be imposed. Additional
analysis programs, which can also incorporate information is presented in Chapter 8.
laminate analysis routines and libraries of com- Toroidal pressure vessels (Fig. 1.3) are in
posite micromechanical properties for fiber/ commercial use to contain air for breathing.
resin combinations. These tools can accomplish Advantages quoted for the QinetiQ, United-
in one day what normally took several months Kingdom-manufactured doughnut-shaped com-
in 1994, when Ref 1.1 was published. posites, are lighter weight, better use of volume,
and greater protection for the pressure regulator
Pressure Vessels (Ref 1.4, 1.5).

Pressure vessel designers can benefit from the Pipes

analysis from netting, closed-form, and finite-
element analysis routines incorporated in the Pipe manufacturers have seen that multiangle
Composite Design and Analysis Code. There filament-wound pipe can confer some desirable
are several case dome configurations shown in properties, such as addition of higher bending
Chapter 8. Design has also benefited from the strength for a pipe simply supported at the ends,
wider use of the Standard Test and Evaluation hoop layers to improve buckling strength for
 Chapter 1: Filament Winding—Introduction and Overview / 3

Fig. 1.2 Composite Isotensoid pressure vessel. Courtesy

of Advanced Lightweight Engineering B.V.

Fig. 1.3 Toroidal pressure vessel. Courtesy of M. Hinton

and QinetiQ. Source: Ref 1.5

buried pipe, or more axial strength with low-

angle helicals for a vertical pipe (Ref 1.6).
Fig. 1.4 Proton-M launch vehicle. Interstage just above
Grid Structures strapons. Courtesy of Khrunichev State Research
and Production Space Center

The commercialization of grid structures has

been realized on the Proton-M launch vehicle,
which uses a composite Anisogrid structure for
interstage (first to second stage is open, and sec-
ond to third stage is overwound with aramid/
epoxy), payload attachment fittings, and com-
posite fairing (Fig. 1.4). A 15% weight savings
is reported for the lattice adapter compared to
other adapters (Ref 1.7, 1.8).
The Isotruss, developed at Brigham Young
University, is a three-dimensional configuration
that can be filament wound but, in the more com-
plicated designs, is more aptly fabricated with a
technology that looks more like weaving.
One of the initial Isotruss applications was a
composite bicycle frame weighing less than 3 lb
(1.36 kg) (Fig. 1.5) (Ref 1.9). Fig. 1.5 Isotruss bicycle frame
4 / Composite Filament Winding

Deep Sea Oil Platform Drill Risers which have a high fiber volume with a virtually
unlimited number of repeatable and accurate
Composite drill and production riser devel- fiber angles. As an alternate to woven fabric and
opment began almost 30 years ago, based on braiding, Storage Tek Composites (Louisville,
perceptions that they would reduce deck loads CO; now part of Oracle, Inc.) used dry high-
and provide several other benefits. The National strength carbon fiber, a prepreg delivery head,
Institute of Standards and Technology was the and a polyester scrim cloth that is melted at the
funding agent to initiate these programs. The ends of the mandrel to hold the fibers in place
principal composite contractors are Spencer for resin transfer molding (Ref 1.20).
Composites Corp. (Lincoln, CA), which is work-
ing with Kvaerner (Oslo, Norway), and Northrop
Grumman (Sunnyvale, CA), which is working REFERENCES
with ABB Vetco Gray (Houston, TX). The key
technologies are the metal-to-composite inter- 1.1 S.T. Peters, W.D. Humphrey, and R.F.
face (MCI) and the liner. The Spencer MCI uses Foral, Filament Winding; Composite
a proprietary traplock design, a 0.12 in. (3 mm) Structure Fabrication, 2nd ed., SAMPE
titanium liner, and titanium flanges, which make Publishers, Covina, CA, 1994, ©S.T.
it interchangeable with the standard titanium Peters, 2009
drilling riser currently used on the Heidrun ten- 1.2 J. Gilchrist, Automated Filament Wind-
sion leg platform in Norway. The riser has been ing of Large, Complex Shapes, SAMPE
in service since 2001. The Northrop Grumman International Symposium, May 2000
MCI is the convex version of the wound-in-place (Long Beach, CA), p 1452
hyperboloid joint invented at Westinghouse 1.3 AGATE, E-765 Approved Databases,
for the lower joint of the Tomahawk Missile Fibercote Industries, Waterbury, CT
Launcher. The Northrop Grumman riser has 1.4 J. Cook and J. Chambers, Toroidal Pres-
been slated for sea trial by Petrobras, Brazil sure Vessels for Breathing Apparatus,
since 2003 but has not seen service yet. Some SAMPE Paris Conference, April 22–24,
problems remain; for example, the composite 1998, p 125–138
drill riser is reported to be approximately 30% 1.5 S. Li and J. Cook, An Analysis of Fila-
to 3 times the cost of the steel riser, and the ment Overwound Toroidal Pressure Ves-
composite wall is thicker, which may negatively sels and Optimum Design of Such Struc-
influence hydraulics (Ref 1.10–1.16). tures, J. Pressure Vessel Technol., Vol
124, May 2002, p 215–222
1.6 R. Lea and C. Yang, “Improving the
Mechanical Properties of Composite
High-Speed Rotors Pipe Using Multi-Angle Filament Wind-
ing,” Paper 458, NACE Corrosion ‘98
The design of high-speed rotors now benefits
1.7 Y. Bakhvalov, V. Molochev, S. Petrok-
from the use of very high-strength carbon fibers
ovskii, V. Barynin, V. Vasiliev, and A.
such as Toray T 800 and T 1000 (Toray Com-
Razin, “Proton-M Composite Interstage
posites America, Tacoma WA). The research
Structures: Design, Manufacturing and
into high-speed rotor design started with fila-
Performance,” European Conference
ment-wound monolithic rings made with con-
for Aerospace Sciences (EUCASS), July
ventional epoxies, then liquid polyurethane elas-
2005 (Moscow)
tomers were introduced; continuous curing was
1.8 V. Vasiliev, A. Razin, G. Totaro, and F.
effected, and finally, multiring composite rotors
De Nicola, “Aniosgrid Conical Adapt-
have been selected for further study. Each itera-
ers for Commercial Space Applications,”
tion had a positive effect on the upper limit of
AIAA paper 2005-3440, p 2099–2107
rotational speed (Ref 1.17–1.19).
1.9 D. Jensen, A Glimpse into the World
of Innovative Composite Isotruss Grid
Structures, SAMPE J., Vol 36 (No. 5),
Filament-Wound Preforms Sept–Oct, 2000, p 8–16
1.10 J. Price, The “State of the Art” in Com-
The filament-winding process is a low-cost posite Material Development and Appli-
method of manufacturing composite preforms, cations for the Oil and Gas Industry,
 Chapter 1: Filament Winding—Introduction and Overview / 5

Proceedings of the 12th (2002) Interna- posium, May 2000 (Long Beach, CA), p
tional Offshore and Polar Engineering 2136–2198
Conference (Kitakyushu, Japan), 2002, p 1.16 B. Spencer, personal communication,
125–130 March 4, 2010
1.11 M. Jellison, R.B. Chandler, G. Prideco, 1.17 C. Gabrys and C. Bakis, Design and
M. Payne, and J. Shepard, “Drillstring Manufacturing of Filament Wound Elas-
Technology Vanguard for World-Class tomeric Matrix Composite Flywheels, J.
Extended-Reach Drilling,” OTC 18512, Reinf. Plast. Compos., Vol 16 (No. 6),
2007 Offshore Technology Conference, 1997, p 488–502
April–May 2007 (Houston, TX) 1.18 C. Gabrys and C. Bakis, Filament Wind-
1.12 D. Johnson, Demand for Composites ing of Thick Composite Rings with High
in Deepwater Oil Production, SAMPE Radial Accretion Rates, Proceedings of
International Symposium, May 2000 the International Exposition, The Soci-
(Long Beach, CA), p 2105–2117 ety of the Plastics Industry, New York,
1.13 Current Composite Drilling Riser Proj- 1998, p 1904–1915
ects, Compos. World, Jan. 2, 2002 1.19 S. Ha, H. Kim, and S. Han, Effects of
1.14 G. Galle and J. Whitehead, A Success- Rotor Sizes and Epoxy System on the
ful Program for Composites in the Oil Process-Induced Residual Strains within
and Gas Industry, SAMPE International Multi-Ring Composite Rotors, J. Com-
Symposium, May 2000 (Long Beach, pos. Mater., Vol 38 (No. 10), 2004, p
CA), p 2128–2135 871–884
1.15 D. Johnson, K. Lo, and H. Wu, Develop- 1.20 D. Howell and F. Roundy, Filament Wound
ment of Rigid Composite Risers—A Sta- Preforms for RTM, SAMPE J., Vol 36
tus Report, SAMPE International Sym- (No. 2), March–April 2000, p 41–46