Autoclave Quality Outside The Autoclave?

Pioneers of out-of-autoclave processing in aerospace applications answer a qualified but enthusiastic Yes!
#northropgrumman #outofautoclave #space


Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

Historically, most designers of structural composite components for aerospace applications have taken for granted the need for an autoclave cure. Despite the fact that cost and time advantages of out-of-autoclave processing are well documented, a perception persists that such processes cannot produce the same mechanical performance that autoclave curing produces. Does this perception have any basis in fact?

The answer depends on what one means by "autoclave quality," says John Fish, VP of engineering at V System Composites (VSC, Anaheim, Calif.), one of a growing number of processors in the aerospace market that have begun to question the old assumptions. Fish admits that in some nonautoclave processes there's a knockdown factor for laminate properties, such as fiber volume, "because you don't have the full pressure that you have with the autoclave." But he's quick to point out that this knockdown can be very small. In his experience with vacuum-assisted resin transfer molding (VARTM) of aerospace composites, Fish reports that 60 percent fiber volumes have been achieved with unidirectional fabrics while 56 percent can be reached with wovens, compared to 62 and 58 percent, respectively, with autoclaved prepreg. A VARTM'd part's stiffness, proportional to fiber volume, would therefore experience a knockdown of only 3 to 4 percent compared to autoclave structures, he claims. More importantly, Fish contends that seeing product acceptability only in terms of per-formance knockdowns takes too narrow a view. Overall weight, part count, manu-facturability, processing time, surface quality and, of course, cost all figure into the equation, he maintains. In fact, taking the autoclave out of the part production equation introduces opportunities for gains in one or more of the above-mentioned categories, which can deliver benefits that, for the customer, overbalance modest knockdowns.

Alternatives to the autoclave

VARTM is one of three processing alternatives that proponents claim can achieve aerospace-grade results without resort to autoclave cure. VARTM denotes a variety of related resin infusion processes now commonly used in the marine, transportation and infrastructure markets. The processes differ radically from prepreg processing in that fiber reinforcements and core materials are layed up dry in a one-sided mold and vacuum bagged. Liquid resin then is introduced through one or more ports strategically placed in the mold, and drawn by vacuum through the reinforcements by means of a series of designed-in channels and/or carefully placed infusion media that facilitate fiber wetout. Unlike the autoclave, VARTM cure requires neither high heat nor high pressure. VARTM's comparatively low-cost tooling makes it possible to inexpensively produce large, complex parts in one shot.

The second alternative is resin transfer molding (RTM). Classic RTM requires a more expensive two-part closed mold, often made of metal. Again, dry reinforcement is placed in the bottom half of the RTM mold, the mold is closed and sealed, then resin is pumped into the mold under positive pressure through injection ports. The method uses very low-viscosity resins, and resin and catalyst are metered and mixed in automated dispensing equipment just before infusion — which permits the use of fast-curing resin systems. Molds are typically heated. Therefore, RTM can reduce cycle time from what can be several days required for hand layup of prepreg and autoclave cure to hours or even minutes. Moreover, the process yields dimensionally accurate, complex parts with good surface detail and low void content, and delivers a smooth finish on all exposed surfaces.

In RTM and some VARTM processes, reinforcements are built up outside the mold into a "preform," approximating the net shape of the finished part. Preforming permits the mold to be used more efficiently (no hand layup) and also presents opportunities to use automated preforming processes, such as braiding.

The simplest and newest method is oven curing of prepregged parts, using prepregs recently developed for out-of-autoclave processing. The most straightforward of the alternatives, this method offers the aerospace composite manufacturer the smallest departure from conventional practice. While VARTM and RTM for the aerospace market are greater departures from common aerospace practice, they offer the greater potential for significant savings in production cost and time. As the following examples illustrate, early results for all three methods are more than promising. In fact, many applications are already in production or well on the way.

Eliminating pre- & post-processing steps

EDO Fiber Innovations (Walpole, Mass.) recently combined VARTM and automated braid preforming to manufacture the fuselage of the Joint Air-to-Surface Standoff Missile (JASSM). Designed and fabricated for prime contractor Lockheed Martin Missiles and Fire Control (Orlando, Fla.), the 4.3-m/14-ft long missile body with 61-cm/2-ft wide trapezoidal cross-section is VARTM'd in three pieces, but begins as a one-piece, net-shape preform.

Preform fabrication begins with braiding of the missile case's inner skin. Control of fiber volume in the preform is critical to achieving "autoclave quality" in this structure, notes EDO Fiber Innovations general manager Garrett Sharpless. "To achieve high fiber volume in braiding," he explains, "you have to be certain the fiber tows are placed in close proximity to one another. We achieve that by directly braiding over a mandrel to control the architecture in the braided structure." Braiding must be performed at relatively high tension to compact the preform on the mandrel, but also must be done in a manner that avoids abrading or breaking the fibers — especially challenging given the missile's unusual shape. Sharpless reports the resulting net-shape preform requires minimal debulking (on the order of 2 percent) once the preform is in the mold.

Tackified multiaxial noncrimp fabrics from SAERTEX USA LLC (Huntersville, N.C.) are layed up over the inner skin. In addition, thermally formed and CNC machined Rohacell foam details are placed in key areas of the preform and located with removable tooling pins. A second braided ply is then applied to serve as the outside laminate skin.

The completed preform is slit to separate the upper shell from the bottom forward and aft cover pieces. (The tackified fabrics help maintain the shape and relative fiber orientations of the braid.) The forward and aft cover preforms are removed from the mandrel and placed in separate VARTM tools. The upper shell preform remains on the mandrel until it has been placed in its VARTM tool, to maintain preform shape. Caul plates are selectively placed on portions of inside surfaces of the three components, to control critical thickness and achieve designed inside surface quality. Then the parts are vacuum bagged and infused with Huntsman Advanced Materials' (The Woodlands, Texas) one-part RenInfusion 8610 epoxy resin, designed especially for VARTM applications. The laminate is oven-cured at 99°C (210°F) and postcured at 149°C (300°F) to ensure maximum glass transition temperature.

EDO Fiber Innovations developed the fabrication process during initial phases of the JASSM missile program, which is now in full-rate production. Prior to the production contract, a process im-provement program, sponsored by the Manufacturing Technology Div. of the Air Force Research Laboratory's Materials and Manufacturing Directorate (Wright-Patterson AFB, Ohio), refined the automated temperature and pressure controls during resin infusion and dimensional control of the inner mold line, thus improving net edge molding. Therefore, the aerodynamic outer surface requires no machining and only minimal manual flash removal.

EDO Fiber Innovations used a similar process to create preforms for the RTM'd horizontal stabilizer on Bell Helicopter's Modular Affordable Product Line (MAPL) rotorcraft demonstrator. The stabilizer, a one-piece structure that passes through the rear section of the fuselage, spans 1.5m/5 ft and has a chord width of 20 cm/8 inches and chord depth of 3.8 cm/1.5 inches and is stiffened by four full-span box-beam spar elements. Spar preforms are Hexcel (Dublin, Calif.) AS4 carbon fiber, triaxially braided over four rectangularly cross-sectioned mandrels. The spar preforms were subsequently assembled and overwrapped with uni and woven biaxial fabric for the skins. The assembly was infused with Cytec 5250-4 one-part bismaleimide (BMI) resin, to handle the high-temperature operating environment created by the stabilizer's proximity to the jet engine's exhaust flow.

Unlike the metal stabilizer it replaced, the one-piece, hollow composite structure requires no secondary bonding. It also eliminates the potential for moisture absorption/accumulation that is typical of foam- or honeycomb-cored sandwich constructions. The component weighs only 7.7 kg/17 lb, a 50 percent reduction from the previous design. In static testing prior to flight tests, the part successfully endured 390 percent of design ultimate load in conditions of 93°C/200°F and high humidity.

Enabling cost-slashing "unitization"

VSC has VARTM'd a set of components for the forward pylon on the CH-47 Chinook helicopter, built by Boeing Integrated Defense Systems (St. Louis, Mo.). The parts replace the previous structure, which was a heavily riveted assembly built up from a number of hydroformed aluminum subcomponents. The replacement part is expected to be baselined by the U.S. Army in 2007 for new Chinooks, fabricated in 2008 and fielded in 2009.

The pylon's upper deck and the forward and aft fairings have been replaced with new carbon/epoxy components, reducing the overall part count of the forward pylon from 277 to 72, with a corresponding reduction in the number of fasteners, from 2,526 to 845. In the upper deck alone, part count has plummeted from 100 to 5 (see HPC January 2004, p. 16). This "unitization" was key to the Affordable Rotorcraft Secondary Structures program under which the new design was created, explains Will Tolotta, Boeing's program manager. "In order to reduce cost — our main purpose — we wanted to eliminate a lot of the assembly," he recalls.

Critical to part reduction, stiffeners and frame members are fabricated integral to each upper deck or fairing in a single VARTM infusion, using an approach VSC aptly describes as Affordable Feature Integration (AFI). Autoclave processing likely would have involved secondary bonding of stiffeners, and Tolotta notes that VARTM's compliant, drapable preforms were better suited for the complex shapes than prepregs. Boeing and VSC worked closely on tooling design and concepts, which were vital to success, Tolotta says. "We needed a complex yet affordable tool set, one that could accommodate removable traps, which apply pressure and achieve compaction in tight areas, minimize resin pooling and porosity, and allow breakdown to extract the part." The bagged side of the components incorporated mandrel details to ensure spatial control of stiffener locations, critical thicknesses and the like.

VSC used its patented HyPerVARTM process, which needs no separate infusion medium. Instead the resin distribution system is incorporated into the process' proprietary tooling, and enables VSC to propagate the resin both in-plane and out-of-plane relative to the tool surface with a high degree of control. The result, claims VSC executive VP of programs and business development Paul Oppenheim, is consistent quality from skins to stiffeners, despite their different fiber forms and geometries.

Preforms for the pylon components are made from plain-weave carbon fabric with a diamond-hatch pattern of RS-11 tackifying agent. Supplied by YLA Advanced Composite Materials (Benicia, Calif.), the fabric maintains drapability but enables stable ply stacking, Tolotta reports. Paul Chernitsky, Boeing's lead designer for the new pylon, worked onsite at VSC to augment part producibility. For example, while ensuring that the design was not compromised, he relocated material splits within the preforms for optimal drapability and material usage.

Parts were infused with EPON 862 epoxy resin with W hardener from Hexion Specialty Chemical (Houston, Texas), Tolotta reports, because it achieves the damage-tolerance properties needed in a structure for which repair-related tool drops are one of the most important concerns.

Qualifying processes for primary structures

Although the pylon's upper deck and fairings are secondary structures, VSC has produced demonstrators of substructure similar to the spars and ribs used in composite airframes for Lockheed Martin Aeronautics Co. — Advanced Development Programs (Palmdale, Calif.) to illustrate the suitability of HyPerVARTM and AFI for primary structures. The trusses are created using modular tools, which can be reconfigured depending on the design demands for a particular truss. Using IM7 unidirectional fabric from Textile Products Inc. (Anaheim, Calif.) and a bidirectional bias weft fabric from Hexcel, demonstration truss structures infused with EPON 862 resin achieved about 1 percent void content with 60 percent fiber volume for unidirectional fabrics and 53 percent for bidirectional fabrics — reportedly approaching the properties of autoclaved prepreg.

Such demonstrations are needed, Oppenheim says, to overcome a misconception among aerospace designers that VARTM and RTM are not viable for primary structures. That misconception stems, in part, from concern about the qualification process. With a prepreg, Oppenheim explains, most of the qualification process can be done at the prepreg manufacturer's level on what will become a "standard" product available to all molders. "Anyone who uses that prepreg in the autoclave will get the same properties," he notes. In infusion processes, however, resin and reinforce-ment are combined in the mold, which sets up a much more difficult qualification scenario in which the burden of proof, as it were, is on the fabricator. One solution is a much more narrowly focused qualification program for a particular structure, which must be completed by the fabricator. However, Oppenheim contends that data from particular part programs could become part of a database that identifies the resin, fiber, fabric and pro-cessing parameters so that a more traditional laminate qualification program might be undertaken.

Radius Engineering Inc. (Salt Lake City, Utah) claims to have bypassed requalification issues by developing an RTM process that uses the same prepregs that prime contractors employ in autoclaved components. Called SQRTM (Same Qualified Resin Transfer Molding), it eliminates the autoclave yet avoids the qualifications required to add a new manufacturing process to an aerospace program. "We have created a robust process for prepreg matched-die molding," says Dimitrije Milovich, Radius Engineering president.

The challenge in RTM'ing a prepreg layup, Milovich notes, is establishing hydrostatic pressure within the mold sufficient to consolidate the layup and suppress voids that occur as gases evolve during cure. In the autoclave, this state is created as pressurized nitrogen applies external force to the vacuum bag. In a closed RTM tool, Milovich explains, pressure is applied externally, but directly to liquid resin, to force it into and throughout the mold cavity. For the SQRTM process, Radius Engineering discovered that in a mold cavity already filled by a prepreg layup, sufficient hydrostatic pressure could be created simply by forcing relatively a small amount of additional resin into the cavity through ports strategically positioned around the part perimeter. Because the tool is sized properly for the preform, this additional resin does not pool or create resin-rich areas.

Radius Engineering developed a SQRTM application now used by Vought Aircraft Industries Inc. (Dallas, Texas) on the Enhanced Wing program for the U.S. Air Force Next Generation RQ-4B Global Hawk unmanned aerial vehicle (UAV). In earlier Global Hawk designs, the wing tip — the outer 3.3m/130 inches of the wing structure — consisted of 12 to 14 autoclaved parts. Working with Vought and prime contractor Northrop Grumman (Palmdale, Calif.), Radius Engineering developed a design that integrates parts, simplifies the structure, eliminates the autoclave and, thus, reduces costs.

As with the Chinook forward pylon, unitization prompted the effort. But in this case, Radius Engineering unified both the wing-tip components and the tool set, notes Radius Engineering's Global Hawk wing tip program manager Richard Nord. Each wing tip in the new design consists of three major components: a torque box, an inboard rib that helps attach the wing tip to the main wing, and a tip-cap closeout. Each torque box integrates six solid-laminate spars, leading and trailing edges and outboard rib in a unified structure. By comparison, the original design employed two honeycomb-and-skin spars with multiple ribs in each torque box. "Sandwich construction is struc-turally efficient but costly because fabrication involves several cures," says Nord. The integrated spars enabled Northrop to attain the required skin stiffness without honeycomb.

The tool set design permits fabrication of all parts for both left and right wing tips in only three tools. The left torque box is laid up in the first tool, the right torque box in the second, while the third contains molds for the remaining parts: left and right inboard ribs, left and right tip-cap closeouts, and two small panels that cover access holes in the bottoms of the two tip-cap closeouts. Each tool also has a cavity for process coupons for quality control monitoring purposes. All the parts are layed up with the same prepreg used in the autoclaved version, Cytec Engineered Materials Inc. (Tempe, Ariz.) M46J, which combines Cytec's 7714A epoxy resin and Hexcel AS-4 or Toray T650 fabric. Cure is performed under 120 psi of fiber bed pressure, supplied by a Radius Engineering pneumatic press. Radius Engineering's RTM 5000 flow-controlled injection system is used to inject the small amount of resin and provide 90 psi of hydrostatic pressure during the 121°C/250°F cure.

Nord credits the matched metal tool and RTM process for fiber volume of 58 percent and void content of 0.5 percent or less. The part meets or exceeds all of Northrop Grumman's performance requirements and weighs 5 percent less than the original, contributing to the Air Force's goal of enabling the Global Hawk to carry bigger payloads.

More important, says Nord, is the aerodynamic smoothness of the new design, which achieved unparalleled profile tolerance of ±0.2 mm (±0.0075 inch). Moreover, the one-piece design significantly reduced labor. "In the autoclaved parts, the bondline between the upper and lower skins had to be filled and faired, involving a lot of handwork," he says. The resin that is injected to create pressure within the SQRTM mold results only in a thin flash on the continuous leading edge.

Controlling part geometry

Elsewhere, conventional RTM is producing net-shape parts not possible via autoclaved prepreg, without extensive post processing. For the F-22 Raptor, Matrix Composites Inc. (Rockledge, Fla.) RTMs carbon/ bismaleimide (BMI) structures with C-channel profiles that require high repeatability and multiple finished surfaces within a complex geometry, reports company president David Nesbitt. The supports feature deep draws and abrupt corners, and must have surfaces that accommodate secondary bonding to the fairings. "In an autoclave, you'd typically get one finished surface, but the opposite side would require post-machining to achieve a nominal laminate thickness and smooth bonding surfaces," he explains.

The deep draws and abrupt corners present a special challenge for some out-of-autoclave processes. Nesbitt explains that multiple plies laid into female corners tend to "bridge" across the area creating porosity, blemishes and other undesirables. "The positive pressures associated with autoclaving tend to overcome these bridging anomalies," he notes, "but the lower the pressure of an out-of-autoclave process, the more critical the layup technique becomes." To circumvent bridging and similar problems, Nesbitt says, "Our approach is to rely on very robust tools and relatively high injection pressures." He also points out that RTM by its nature almost always eliminates bridging. "The matched tooling approach tends to drive the fibers where they need to be," he explains. "These more robust processing conditions result in reduced process variability and an overall reduced dependency on layup technique."

The RTM'd F-22 fairing supports routinely achieve fiber volumes of 55 to 60 percent and void contents of less than 0.5 percent, Nesbitt reports.

RTM-ing large components

Historically, RTM has not been ideal for large, relatively flat aerodynamic surfaces. Because infusion gates and vents usually are placed around part edges infusing resin across large areas without forming "islands" of trapped air and volatiles may require a degree of control over resin flow unachievable without placement of vents in aerodynamically functional surfaces. The resulting surface mars that these devices create must be machined out, significantly increasing labor cost and production time. To address this limitation, North Coast Tool & Mold Corp. (NCTM, Cleveland, Ohio) has developed DRIV (Direct Resin Injection and Venting), a proprietary device that enables venting on a finished surface. Unlike a conventional vent, DRIV inserts produce minimal flashing that is easily removed, explains NCTM's RTM technology manager Dan Davenport.

Lockheed Martin Aeronautics Co. (Fort Worth, Texas) recently confirmed the viability of DRIV-assisted RTM for fabrication of large parts by producing three demonstration vertical tail structures on tooling designed and built by NCTM. The carbon/BMI tail components, which measure 4m/13 ft long and 1.5m/5 ft wide, include two external skins surrounding 14 hollow torque tubes. For the torque tubes, North Coast created 14 interlocking, tapered mandrels, from 229 cm to 254 cm (90 to 100 inches) in length, over which are pulled triaxial braided socks from A&P Technology (Cincinnati, Ohio), which were tapered by adjusting braid angles via numerical controllers on A&P's MegaBraider machinery. Fabrics and uni tapes make up the external skins, with Cytec 5250-4 BMI resin. Special clamping, also developed by North Coast, allowed Lockheed Martin to avoid the multimillion-dollar expense of a press originally specified to clamp the two outer mold cavities together.

"We've molded up to 68 percent fiber volume, and 62 to 65 percent is very doable," Davenport contends, adding that void content falls below 1 percent.


Moving from autoclave to oven

The high-quality aerodynamic surfaces on exterior fairings of The New Piper Aircraft Inc.'s (Vero Beach, Fla.) Meridian and Malibu aircraft are the result of use of a carbon/epoxy prepreg specifically developed by Advanced Composite Group Inc. (ACG, Tulsa, Okla.) to achieve autoclave-quality finish via oven cure, without large secondary finishing costs. Piper wanted to reduce weight and improve the surface finish of the fairings, which previously were made using wet layup. Oven curing met the program's cost targets, recalls Nesbitt, whose company, Matrix Composites, designed and now manufactures the new fairings for Piper. ACG's LTM24ST single-side film (partially impregnated) prepreg combines Toray T-300 carbon 2x2 twill fabric with epoxy resin, and cures at low temperatures (50°C/122°F) in just one hour. Because much more expensive bagging materials are used in the autoclave to reduce risk of bag failure, savings in materials and autoclave time netted an estimated 25 percent decrease in overall part price, Nesbitt says. "The tooling is easier and less expensive to fabricate as well."

"We are achieving as-molded surface finishes comparable to those achieved using an autoclave," he claims. "Surface porosity has been eliminated, reducing the post operation time required to produce a high-quality painted surface."

While LTM24ST meets the per-formance demands for these lightly loaded fairings, ACG's aerospace market sector manager Chris Ridgard admits that this prepreg was not designed for highly loaded structures. But the company's new generation of oven-curable prepregs are, he reports. In its development efforts for high-performance applications, ACG learned lessons about the differences between prepregs targeted for the oven versus the autoclave. "We mistakenly believed in the early days that low viscosity was a good thing for oven cure," Ridgard recalls. But with low viscosity, the resin can flow prematurely and lock off escape paths for entrapped air and volatiles, which are critical for out-of-autoclave processing. By comparison, these gases will dissolve into the resin under autoclave pressures. These distinctions mean that the degree of impregnation is far more critical in oven-cure than in autoclave prepregs, and impregnation must be more carefully controlled.

ACG's latest, medium-temperature oven-curable prepregs, MTM 45-1 and MTM 46 are producing fiber volumes equivalent to autoclave prepregs and near-zero void content, Ridgard reports. Both product lines offer a minimum cure temperature of 82°C/180°F. MTM 45-1 is designed for maximum wet-service temperature of 121°C/250°F while the less-expensive MTM 46 has a maximum wet service temperature of 82°C/180°F. Under a contract with the Air Force Research Laboratory (Wright-Patterson AFB, Ohio) and working with the NASA/FAA National Center for Advanced Materials Performance (NCAMP), ACG is developing B-basis values for these two products, using an approach similar to the U.S. Federal Aviation Admin.'s AGATE (Advanced General Aviation Technology Experiment) methodology.

Faster, cheaper … and often better

Out-of-autoclave methods are not only faster and less expensive, but fabricators also have optimized them for greater manufacturability and eliminated most secondary processes, yet molded parts of greater complexity.

"These processes allow you to do something that you couldn't do before," VSC's Fish sums up. "When the design itself can be optimized to a point beyond what is possible with autoclaved prepreg, then it opens up many opportunities." The next step? According to Fish, that's simple: "Spread the word."