Parts consolidation key to lower cost composites

One-piece composite components are replacing multiple parts previously made both from composites and from metals or other materials. The motivations for the design and the challenges for manufacturing can differ as much as the end products do. But a common incentive for part consolidation is cost — composites frequently earn their way onto a project not just for their properties, but because reduction of part count makes it possible to mold and assemble what would otherwise be a much more expensive multipart structure at or near the cost expected for parts made with competing materials. Yet a great deal of design and engineering innovation may be required before those benefits can be reaped.

Accommodating sleek contours

That was certainly the case when Ford Motor Co. sought a way to package the rear-mounted, 500-hp engine in its new GT sports car. Ford needed a tight, aerodynamic "decklid" engine enclosure that would meet safety, weight and cost goals without compromising aesthetics. With that in mind, engineers looked for ways to reduce the cost and weight of the approximately 6-ft square decklid assembly, which consists of a four-part exterior paintable aluminum skin (including left- and right-side quarter panels, a header, and an engine vent cover), with an inner structure to stiffen the assembly.

For the team at Ford's Research and Advanced Engineering group, a one-piece composite inner structure was the ultimate solution, despite some technological challenges — challenges which Ford and its composites fabricating partner, Sparta Composites (San Diego, Calif.) were able to resolve.

"With composites, we were able to do a more aggressive design," notes Eric Kleven, Ford's GT composites specialist. "We had to make far fewer manufacturing compromises" than would have been required for a multi-piece approach, he says. The one-piece inner assembly stiffens the entire decklid by eliminating flex points where bonding and rivets would join multiple pieces.

A traditional stamped aluminum approach wasn't feasible: the part presented such complex curvatures, with reverse angles as well as angles of less than 90°, that stamping them would result in die lock. Manufacturing an aluminum inner structure would involve multiple parts that would require secondary assembly, as well — an option both cost- and weight-prohibitive. The remaining choice of materials, therefore, was between composites or superplastic-formed aluminum. (The superplastic forming process heats aluminum sheet stock to about 500°C/930°F, the point at which it can be plastically formed over a single-faced tool to create a part's shape.) Though lighter than stamped aluminum, superplastic-formed aluminum would cost more and still require four or more individual parts to make up the inner structure. However, the least expensive materials for a composite inner structure would cost nearly three times as much as the premium superplastic-formed aluminum product. A composite solution could go forward only if tooling and assembly costs could be reduced to make up the difference, says Kleven.

Ford briefly considered using composites only for the complex outer portions of the deck's inner structure assembly, which support the quarter panels. However, engineers quickly calculated that, in this case, the substantial difference in coefficient of thermal expansion (CTE) between the inner's aluminum and carbon/epoxy components would create unacceptable stresses at the joints.

Cost concerns notwithstanding, the Ford team requested a quote from Sparta Composites for a four-piece, all-composite design for the inner structure. However, Ford GT manufacturing manager Matthew Zaluzec reports, "Dimensionally, we could not effectively control the geometry of multiple pieces." With a producible manufacturing tolerance as variable as ±0.7 mm per piece, the overall dimensions of a multi-piece deck lid inner had the potential to vary from specification by as much as twice that amount. It was at this point that Kleven joined the GT team and suggested a one-piece composite component. "It seemed to me," Kleven recalls, "that instead of having to pay for four tools, this could be made in one piece, which would help us to meet our weight and cost targets, too."

Although CTE differential would not be a problem in a one-piece inner structure, CTE would be a factor when the inner structure was joined to the outer skin panels, which Ford fabricated from aluminum to ensure a Class A finish. Since the assembly would have to withstand heat buildup (from running the 500-hp engine at wide-open throttle) and stops on hot pavement, Ford addressed the likelihood of CTE mismatch by hem-rolling the aluminum outer and composite inner pieces together at the edges, using three different adhesives to join them, all currently supplied by The Dow Chemical Co. (Midland, Mich.). A two-part epoxy provides high-modulus bonding at the hems while a two-part urethane forms a pliable bond in the center of each panel, and a one-part urethane is used at the interface with the rear windshield. The bonding strategy successfully accommodated the mismatch. “It was not as big an issue as we though it would be,” Kleven admits. “It was all containable.” To prevent galvanic corrosion at the aluminum/carbon interface, Sparta Composites designed a thin glass fiber scrim with high resin content into the hem flange. While the adhesives themselves help to isolate carbon from aluminum, as well, Ford took the additional precaution of manufacturing the outer skin with electrocoated aluminum, with the result that the coating permits no direct contact between the composite and raw aluminum.

As Kleven had anticipated, the decision to mold the inner in one piece was the key to meeting cost targets. The unidirectional carbon/epoxy prepreg selected for the deck lid’s inner structure cost about $8/lb, expensive by comparison to aluminum for stamping ($1.30/lb) and superplastic-form aluminum ($2.50/lb to $3.50/lb).  But the prepreg contributed to overall cost-effective because less material was needed: an aluminum inner structure’s wall thickness would typically be 1.5 mm, but 1.2-mm thick the carbon/epoxy achieved the requisite stiffness at half the weight.  With only one tool to make for the composite component, rather than four stamping dies, Kleven notes, “It is relatively cheap.” For stamped aluminum, says Zaluzec, “We would probably be at about $1 million for the primary forming tools, and we would still have to trim and flange and so on.” By contrast, the price for the composite inner structure’s single Invar tool and associated fixtures was approximately $250,000.

Although unidirectional carbon/epoxy prepreg was specified, at less than half the price of other prepreg forms, Sparta and Ford decided to use traditional layup methods, rather than an automated process. By automotive standards, the specified production rate was low — only 9 to 10 parts per day. But for a traditional aerospace manufacturer like Sparta, Zaluzec recalls, “It was culture shock, because of the complexity and volume.”

The unidirectional prepreg would be difficult to work into the mold’s complex curvatures, explains Sparta’s Mike Vairo, division manager, composites product manufacturing. “A part like this would take days.” When he and Sparta’s program engineer J.T. Lyons told their comrades that they planned to lay up seven plies of unidirectional prepreg in the tool, “they laughed,” Vairo recalls. “They said it couldn’t be done.”

In order to adapt the traditional layup method to the speed and the repeatability required by Ford, Sparta streamlined the process with strategic pre-production planning and some material innovations. Sparta developed exact ply patterns, using FiberSIM CAD-integrated design/analysis software, from VISTAGY Inc. (Waltham, Mass.). “Without FiberSIM, I don’t know if we could have made the part out of uni.”

Patterns were cut on an Autometrix (Grass Valley, Calif.) automated CNC cutting table, using nesting software that optimized material usage. A laser projection ply templating system from Laser Projection Technologies (LPT, Londonderry, N.H.) identified the order in which to pick up the nested pieces from the cutting table to create a correctly sequenced kit — containing as many as 100 patterns cut from a single swath of prepreg. The same LPT system then identifies the ply location on the tool in the lamination step, a key to maintaining the repeatability, process control, and speed required for automotive applications. A software solution enabled the Autometrix cutting table to interact seamlessly with FiberSIM’s pattern export software and LPT’s templating system. Sparta Composites also developed a proprietary offline pre-plying step that would enable technicians to get a jump on the next inner structure while the previous layup was cycling through the autoclave on the single tool.

Finally, a quick-cure epoxy was used in the prepreg supplied by Toray Composites America Inc. (Tacoma, Wash.), so that cure could be completed in only 10 minutes at temperatures ranging from 140°C to 148°C (285°F to 300°F). “That really helps with cycle times,” says Kleven. With 4,500 units of the GT planned, Ford now has a patent pending for the entire deck lid assembly (the composite inner and the aluminum outer).

Ultimately, the big benefit of the deck lid inner’s part consolidation, Zaluzec believes,” is investment reduction.”

The bottom line for the upper deck

When Boeing Integrated Defense Systems St. Louis. Mo., U.S.A.) recently redesigned the forward pylon on the CH-47 Chinook military helicopter, the company teamed with V System Composites (Anaheim, Calif.) to drastically consolidate part count and reduce manufacturing costs by cocuring what were previously separate metal outer skin and stiffeners of the pylon’s upper deck, in a single composite part, using a streamlined VARTM process rather than traditional prepreg/autoclave methods.

Formerly, three of the five major sections of the existing pylon had been constructed from composites: the left- and right-hand work platforms and the aft fairing. But the forward fairing and upper deck had been traditional sheet-metal buildups. Manufacturing the upper deck in one-piece is expected to lower the pylon’s part count from more than 100 to 5.

"This part was redesigned for producibility,” says Mike Louderback, V System president and general manager. “For two years, we worked very closely with Boeing engineers in an integrated product development team.”

Though materials costs for the composite upper deck were much higher than for the metal version, the innovative use of VARTM enabled Boeing to bypass entirely the significant tooling and assembly costs for the metal and the multi-step layup and cure required with prepreg and autoclave methods. Co-infusing and cocuring the upper deck skin and stiffeners in a single operation resulted in a more structurally sound interface that eliminates most of the hundreds of fasteners that ordinarily would join the stiffeners to the skin. Compared to the number of fasteners in the original metal upper deck, there are “virtually none in the design we have now,” says Boeing engineer/scientist Lee Kitson.

Beyond labor and material savings for fasteners, cocuring the structure eliminated the labor intensive process of drilling holes for fasteners, a design aspect that V System VP of programs and business development Paul Oppenheim, says, “design engineers hate.”  And it vastly reduced the laminate buildup that is typically necessary around drill sites and, consequently, simplified the failure analysis that would be required on the drilled parts. Further, integration of the skin and stiffeners also eliminated the capital cost of an assembly jig, which is no longer necessary.

Because it must withstand significantly lower heat and pressure than an autoclavable tool, the production VARTM tool costs about 50 to 75 percent as much as tooling used for hand layup/autoclave processing, contributing to what Louderback estimates will be an overall 30 percent savings, compared to conventional prepreg layup and autoclave cure. Much of the tooling cost savings is attributed to V System’s simplified approach to VARTM — modifications that Louderback says, “applied RTM methodologies to VARTM.” Specifically, V System’s VARTM process infuses large parts using only one injection port, positioned in the tool rather than the vacuum bag. This, in turn, simplifies the vacuum bagging system. The result is a flat bagging operation and fewer consumables, such as infusion media. The flat bagging also means that creep will not affect part quality as it can with more complex bagging. The single-port configuration cuts bagging time by reducing  the incidence of leaks associated with complex resin transport systems, yet the system achieves wetout in less than 10 minutes. The resulting part exhibits less than 0.5 percent void content.

The VARTM’d upper deck was layed up with dry reinforcements, which eliminated the scrap cost associated with traditional prepreg. The process also allowed the use of an inexpensive resin system, EPON EPIKOTE 862, a low-viscosity bisphenol F epoxy from Resolution Performance Products (Houston, Texas), which met both Boeing’s standards and V System’s production needs for an infusable resin.

Boeing redesigned various aspects of the entire forward pylon and reconfigured the work platform/upper deck interface as well, making a direct comparison of the original and new upper decks misleading, but the overall improvements are impressive. While pylon weight is not expected to drop more than five percent, because much of the structure is already made from lightweight materials, Kitson estimates that manufacturing cost for the entire new pylon will decrease by 40 percent or more, compared to the original design.

V System has successfully fabricated a test deck and is now gearing up to fabricate additional decks for demonstration and testing. The test program will validate the structural characteristics of the deck, and provide necessary feedback to the integrated design process. Production implementation is scheduled for mid 2005. Currently, V System can build one upper deck in three days but expects production improvements to drive this time down to one deck per day. The quality, speed, and cost all add up to a very attractive fabricating process. “The allure for Boeing,” says Louderback, “is that it meets the quality production standards for aerospace with minimal consumables.”

One potential drawback to parts consolidation, Kitson notes, is repairability. “Previously our customer could replace one damaged component in an assembly. Now they will have to replace or repair the whole piece, so they’ll be more likely to repair it.” Even here, though, the process of unfastening and changing out a metal piece is not particularly less involved than performing a composite repair. Boeing therefore fully expects a satisfied customer — and a very satisfying bottom line.

The program is currently in the second of three phases, and also targets field maintenance issues. “For example,” explains Kitson, “we are looking at advanced materials that will provide improved impact resistance, reducing the chance of damage if maintenance personnel drop a tool on it.” Boeing may also make the aft portion removable to facilitate replacement of internal equipment.

Tailoring foot functionality

Parts consolidation has been the means to expand rather than redesign the College Park Industries (CPI, Fraser, Mich.) product line. The prosthetics manufacturer made its reputation 17 years ago, with its signature model, the TruStep, an intricately designed and engineered prosthetic foot that closely mimics human foot performance. It does so on the strength of a complex, 17-component mechanism that features three key structures molded from a proprietary composite material, which functionally replace the major bone structures of the human foot — the ankle, forefoot and heel. The forefoot features a bifurcated (forked) toe that imitates the independent suspension provided by the first and fifth metatarsals in the human forefoot, while variable “soft” components provide customizable shock absorption that optimizes wearer comfort by complying with uneven terrain and providing an impulse for maintaining forward momentum. While CPI would gain nothing from reducing the Trustep’s part count, part consolidation has paved the way for additional models that  meet special needs. A 15-component follow-on to the TruStep, the Venture features a composite heel and forefoot fabricated in one piece. Its wider, longer and flatter forefoot improves agility and provides greater dynamic response from the toe for the highly active adult. The simplified design reduced manufacturing costs, but its unique functionality commands a retail price one-third higher than the TruStep.

Recently, CPI set out to develop a lower-cost foot, the Tribute, to serve both the moderately active, primarily elderly domestic market segment and the more price-sensitive international market says CPI president Eric Robinson. The Tribute eliminates 5 of the Trustep’s 17 parts, borrowing concepts from the Venture but consolidating the ankle, heel and fore foot into a single composite component. While the more straightforward design simplifies the prosthetic’s range of responses, the elimination of joints, use of a wider forefoot and a lower heel yield a very stable foot that retails for about 25 percent of the price of the TruStep. Hand layup of a proprietary blend of reinforcements along with a customized resin is followed by compression molding and a postcure. Both the simplified design and CPI’s manufacturing advances make the low price possible.

The biggest manufacturing challenge has been tooling for the unitary composite piece, which integrates a donut-shaped ankle component. To produce this piece, the tool incorporates a sophisticated, proprietary modification to ease both layup and release of this complex piece.

Lessons

Among the “selling points” associated with composite construction, parts consolidation ranks high, headlined by lower costs (especially for tooling and elimination of fasteners and secondary bonding steps. But designs that reduce part count also can improve manufacturability and help achieve certain goals for part functionality, such as greater stiffness or stability. Obstacles to part consolidation remain. The subtle features of the TruStep prosthetic foot, for instance, may always demand a large part count. And automakers have yet to view composite surfaces as ideal substrates for Class A finishes. Nevertheless, the savings form parts consolidation can often cost-justify a wholesale change in manufacturing protocols, allowing OEMs new freedom to experiment with VARTM and other alternatives to traditional hand lay-up methods. That alone could put parts consolidation at the top of the composites designer’s priority list.