In the mid-1990s, the Air Force Research Laboratory (AFRL) recognized that, despite the potential to drastically reduce aircraft structural weights compared to conventional metal structures, the aircraft industry was reluctant to implement advanced composites in new aircraft. For example, although composites previously were used in relatively small percentages on the F-15, F-16, and F-18, and despite early projections that the F-22’s airframe would be 50 percent composite by weight, the F-22’s actual fraction of composites is 25 percent. Reasons cited for the fallback included the perceived risk of this “new” technology and — as the F-22 faced cost scrutiny at the drawdown of the Cold War — the high costs of composites compared to metals.
To address these concerns, AFRL launched the Composites Affordability Initiative (CAI) in 1996. A team comprising representatives of the AFRL Materials and Manufacturing Directorate and Air Vehicles Directorate, the Office of Naval Research, Bell Helicopter, Boeing, Lockheed Martin and Northrop Grumman participated in a $152 million (USD), 11-year effort, seeking ways to make composites more affordable and more widely used on aircraft.
After investigation, the CAI team found that the key to affordability in composite structures was to reduce assembly costs. At that time, state-of-the-art aircraft structures (metal substructures with either metal or composite skins) had thousands of parts assembled with hundreds of thousands of fasteners. Drilling holes and installing fasteners is a major source of labor and rework in aircraft structures. The team reasoned that if composite parts could be integrated and/or bonded, then structural assembly costs could be drastically reduced.
From that point, CAI’s objective was to establish a fundamental change in the way aircraft structures are built — one that would give operators the confidence to fly large integrated and bonded structures. This required a multidisciplinary approach: maturation of materials and processes; an understanding of the structural behavior of bonded joints; quality assurance and nondestructive evaluation to ensure bonded joints remain bonded throughout an aircraft’s service life; and the sign-off of U.S. Department of Defense (DoD) aircraft certification authorities.
CAI’s technical achievements
The CAI approach involved the full spectrum of airframe design and manufacturing disciplines — unusual breadth for a research program, but crucial to program success. The results of CAI’s efforts are outlined here and will continue in Part II, in the HPC May 2007 issue (see “Editor's Pick's,” at left).
Vacuum-assisted resin transfer molding (VARTM): To facilitate structure integration, this out-of-autoclave process, developed initially to mold large yacht hulls, was transitioned to the aerospace industry. The VARTM process uses lower-than-atmospheric pressure (typically full vacuum) to pull liquid resin into a dry fiber reinforcement layup. There are two key advantages of VARTM over conventional autoclave curing. First, an autoclave is not needed, resulting in reduced capital equipment costs and a much larger supplier base for part fabrication. Second, epoxy resins are now available for VARTM cure at temperatures low enough to enable the use of inexpensive tooling, such as medium-density fiberboard, rather than the typical Invar tooling used for 177°C-/350°F-cure epoxy prepregs.
While the aerospace industry has dabbled in VARTM over the years, CAI has demonstrated its viability as a production method for large aerospace parts. The process has worked success- fully with several low-viscosity, aerospace-grade resins developed for VARTM processing: EX-1510 cyanate ester, formulated by Bryte Technologies (Morgan Hill, Calif.); SI-ZG-5A, a blend of oxirane-based resins developed by A.T.A.R.D. Laboratories, div. of Shade Inc. (Lincoln, Neb.) and HexFlow VRM 34, a two-part, amine-cured epoxy system from Hexcel (Dublin, Calif.). Wider use of the VARTM process could be enabled by low-viscosity toughened resins with properties similar to CYCOM 977-3 resin from Cytec Engineered Materials (Tempe, Ariz.). Such resins are currently in development but have yet to be demonstrated. CAI has demonstrated VARTM on parts as large as 50 ft²/4.65m² and its efforts have resulted in fiber volumes and per-ply thickness comparable to typical autoclave-cured aerospace composite parts. Further, the process has demonstrated versatility, indicated by the broad range of parts already fabricated, and has drastically reduced part and fastener counts with a corresponding reduction in fabrication costs. For example, each part in the photos above was originally a subassembly of, on average, 20 parts and hundreds of fasteners. Through these efforts, CAI has demonstrated that VARTM is an aerospace-ready process.
Adhesively bonded structures: While bonded structural joints are currently in service on DoD aircraft, including the F-18 and Global Hawk, there continues to be legitimate concern in the DoD airframe certification community about broadening their use. The inability of nondestructive evaluation techniques to discriminate between a good bond and a “kissing” bond (intimate contact between adhesive and structure without adhesion) has been the key roadblock. To ensure structural integrity, current certification policy requires a traveler coupon or a proof test (each is expensive) with every bonded joint during the manufacturing process. There is no method to ensure the bonded joint maintains integrity during its service life. In addition, there are many potential sources for problems in a faying surface bond (e.g., the top of an I-beam glued to a skin). To achieve a good faying surface bond, the adhesive must be spread over the spar/rib caps prior to assembly. Especially with large parts, this can lead to out-time issues, due to “skimming” (premature cure) of the adhesive surface, thus compromising mechanical properties. They also may require several verifilm cycles to ensure that joint tolerances permit the adhesive thickness required by the design. (Verifilm is a class of removable adhesive product that permits temporary bonds between parts for the purpose of verifying proper fit.) Despite the current unease, bonded structures have tremendous potential for aircraft structures. If designed correctly, bonded aircraft structures can greatly reduce part and fastener counts and structural assembly times.
Pi joint: CAI’s bonded-structures work centered on the “pi” joint. This stiffener, shaped like the Greek letter π, can be cocured or cobonded to the skin. The technique is analogous to blade-and-clevis joints typically used in woodworking. The pi joint has several advantages. First, it provides structural redundancy. The pi joint acts as a double lap-shear joint, increasing the surface area for bonding. When CAI tested this joining technique, using Hysol EA 9394 adhesive supplied by Henkel Corp. (Bay Point, Calif.), the pi joint took advantage of the adhesive’s inherently excellent shear properties. Pi joints also paved the way for reduced assembly times by eliminating verifilm checks. Also, they have eliminated the out-time issues present with faying surface bonds by injecting the adhesive into the gaps between the blade and the pi clevis walls after the stiffener blade is inserted.
Testing has shown that the joint is robust and exhibits predictable performance. A key finding from the CAI pi joint studies is that the pi joint bonded with room-temperature paste has five times more strength than the cocured joint of the pi stiffener to the skin. Thus, the pi joint will not be the weak link in a primary structural application. Also, the pi joint is tolerant of several defects: thick bond lines; a blade that is not parallel to the upright legs of the pi; a blade skewed to one side of the clevis; and typical manufacturing defects, such as voids and peel plies that were not removed prior to bonding. This robustness was proved by a series of tests, ranging from coupons to full-scale airframe components.
Structural demonstrations included a compact inlet duct with bonded frames; an F-35-like forward fuselage, wing, and vertical tail; an X-45A-like fuel bay and wing carry-through; and an X-45C-like wing. The X-45A wing carry-through and X-45C wing were structurally tested to design limit load, two lifetimes of fatigue with damage, design ultimate load and, finally, to failure. Both articles failed just above design ultimate load. In addition, the F- 35-like wing successfully demonstrated ballistic survivability of a bonded structure. These structural and ballistic tests show that bonded structures can meet structural requirements for military aircraft. CAI studies also showed that assembly times for bonding can be reduced from 50 to 80 percent compared to typical fastened structure, depending on the article, translating to cost savings of 20 to 50 percent.
In Part II, Russell reviews enabling tools for modeling bond joints, gauging bond durability and damage tolerance and ensuring bonded part quality and subsequent certification.