Advanced Airframe Structural Materials A Primer and Cost Estimating Methodology Susan A. Resetar, J. Curt Rogers, Ronald W. Hess | Cea RadThe research reported here was sponsored by the United States Air Force under Contract F49620-91-C-0003. Further information may be obtained from the Long Range Planning and Doctrine Division, Directorate of Plans, Hq USAF. Library of Congress Cataloging in Publication Data Resear, Susan A, 1961~ ‘Advanced arfame stuctural materials: a primer and cost estimating methodology / Susan A. Resetar, J. Curt Rogers, Ronald W. Pp. om. "Prepared forthe United States Ais Force.” *RADIG-AP” Includes bibliographical references, ISBN 0-8330.1128-6 1. Airframes—Estimates. 1 Rogers, J. Curt, (James Cus), 1939. T, Hess, Ronald Wayne, 1946— ML Tite. TLGTI.6RA6S 1991 (629,134°31—e20 91-10855 ‘The RAND Publication Series: The Report is the principal publication doc- ‘umenting and transmitting RAND's major research findings and final research results, The RAND Note reports other outputs of sponsored research for general distribution, Publications of RAND do not necessarily reflect the opin- ions or policies of the sponsors of RAND research. Published 1991 by RAND. 1700 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138R-4016-AF Advanced Airframe Structural Materials A Primer and Cost Estimating Methodology - Susan A. Resetar, J. Curt Rogers, Ronald W. Hess A Project AIR FORCE Report prepared for the United States Air Force RAND ‘Approved for public release; detibuton mitesPREFACE ‘This report documents the results of a study addressing the implications of advanced materials on airframe structure cost. More specifically, an estimating methodology for overall airframe cost was developed that is sensitive to the material composition of the basic structure and suitable for use in a program's conceptual stage when little detailed design information is available. A primer describing and assessing new materials technology and related manufactar- ing processes is also presented. ‘This research was performed for the Deputy Assistant Secretary for Cost and Economics and was undertaken as part of a project entitled “Air Force Resource and Financial Issues” within the Resource Management and System Acquisition Program of Project AIR FORCE.SUMMARY ‘This report identifies, describes, and quantifies the cost effects of structural materials that are likely to be incorporated into aircraft becoming operational in the 1990s (aluminum, aluminum-lithium, steel, titanium, graphite/epoxy, graphite/bismaleimide, and graphite/ thermoplastic). ‘The first half of this report is a primer for advanced aircraft structural materi- als emphasizing polymer matrix composites. ‘The second half of the report contains both cost data and a cost estimating methodology sensitive to material mix. For each material type, separate cost factors are presented for two time frames, the late 1980s and the mid-1990s, end for the following cost elements: nonrecurring engineering, nonrecurring tooling, recurring engineering, recurring tooling, manufacturing labor, manufacturing material, and quality assurance, ‘These factors are based on data obtained from Boeing Airplane Company, General Dynamics Corporation, Grumman Aerospace Corporation, Lockheed Aerospace Systems Corporation—California Division and Georgia Division, LTV Aerospace and Defense Aircraft Group, McDonnell Douglas Corporation, Northrop Aircraft Division, and Rockwell Interna- tional Group. ‘The new materials and processes that are most likely to be utilized in aircraft structure vill be aluminum-lithium, superplastically formed (SPF)/diffusion bonded (DB) titanium, and polymer matrix composites (for example graphite/epoxy).! The other categories of composite ‘materials (ceramic matrix, metal matrix composites and carbon/carbon composites), powder ‘metallurgy, and titanium-aluminum alloys are not likely to constitute substantial proportions of 1990s aircraft. These materials either (1) offer properties that are too specialized for widespread application to aircraft structure or (2) are behind polymer matrix composites or other metals in terms of development. ‘Two important categories of polymer matrix composites are thermosets and thermoplas- tics. All composites used up to now on production aircraft have been thermosets. However, ‘continuous fiber reinforced thermoplastics have recently received considerable attention. The important distinction between the two is that thermoplastics, unlike thermosets, may be reheated and reshaped. This property offers enormous potential for processing ease. However, few enginsering data exist regarding thermoplastics, creating uncertainty in both the design and fabrication processes. Although composites clearly offer several performance advantages including reduced weight, reduced number of fasteners, increased corrosion resistance, reduced radar cross- section, and the potential for extended life, this increase in capability comes at a price. The cost data collected clearly show that composite materials are more expensive than aluminum on a cost-per-pound basis. In fact, in terms of overall recurring cost per pound,? composites are over two times more expensive than aluminum. A few of the reasons for the higher costs of ‘composites are: (1) properties of these engineered materials frequently need to be verified, (2) tooling requires sophisticated design and durability because of the severe thermal cycling ‘experienced in the autoclave, (3) current hand-layup techniques are very labor intensive, and (4) quality assurance techniques and processes are in their formative stages and are time con- suming. “aluminum, titanium, and steel will continue tobe important material in future ercaft structural epplications. eit ing otic recuring eeaeig eng toning, mating or, mel ent entProjections for the mid-1990s indicate that the hours per pound for both conventional ‘metal materials and composite materials will decrease. However, composites will still be on the order of two times more expensive than aluminum (per pound). These projections were based on a company’s ongoing development work and capital investment projections. Unfortunately, at the time the worksheets were distributed (1987), the outlook for the defense industry was substantially brighter than the current one, Thus, given current production stretchouts and cancellations, it is not clear that companies will be willing (or able) to make the capital invest- ‘ments necessary to achieve the projections reported here. In addition to the structural cost factors described above, an approach for estimating overall airframe costs was developed. The method, which is suitable for use in a program's ‘conceptual stage when little detailed design information is available, takes into account not ‘only the cost of the airframe structure but also the cost of the airframe subsystems (e., elec- trical, hydraulic, environmental) and the cost of final assembly/integration. Reduced to its simplest form, the method applies weighted material indexes to baseline cost estimating rela- tionships that utilize aircraft empty weight and maximum speed as the principal explanatory variables. ‘The method was applied to two hypothetical fighter aircraft to got some feel for the net effect of composites on overall airframe costs. One of the fighters weighed 13,000 Ib and had ‘an alhaluminum structure. ‘The other fighter weighed 11,700 Ib and was based on the assump- tion that only 5,200 Ib of graphite/epoxy would be required to replace 6,500 Ib of aluminum (a 20 percent weight savings). A comparison of the costs of the two vehicles indicated that the substitution of graphite/epoxy for half the aluminum in the structure would increase nonrecur- ring sirframe costs by about 3 percent and recurring airframe costs by roughly 35 percent? However, acquisition cost is by no means the only criterion for deciding whether or not to lize advanced materials such as composites. Operations and support costs must also be con- sidered and the net life-cycle cost effect then balanced against the performance advantages. this concusion ssumen tht we started with an alshuminum sireret and thatthe only weight savings was that ‘rom the meteril substitution. Different material mixes as well es potential aircraft essing would produce different realACKNOWLEDGMENTS Lieutenant Colonel Terry Luettinger, formerly of the Cost Programs Division of the Office of the Deputy Assistant Secretary for Cost and Economics, provided invaluable sponsor- ship of this work. John Birkler, a RAND colleague, originally proposed this work and was instrumental in determining the scope of the project. Robert Neff, formerly of the Aeronauti- cal Systems Division, and James Weathersbee, formerly of the Naval Air Systems Commend, assisted in the distribution of the worksheets to the major U.S. aircraft and helicopter com- panies, Arthur Strathman provided crucial insights into the costs of composite materials. RAND colleague John Matsumura and Fred Timson of Northrop provided timely and thorough reviews. Finally, the following individuals provided the data without which this analysis could never have been undertaken. Boeing Airplane Company Lockheed Aerospace Systems Henry Schwartz Corporation—Georgia John Tate Larry Palmer General Dynamics Corporation ohn Denheyer LITV Aerospace and Defense Aircraft ‘Tom Bronte Grow Cnt Mier Grumman Aerospace Corporation "iin Winding Steve Beckman Harold Quigley McDonnell Douglas Corporation Evelyn Crosman Rick Weis ‘Samuel Dastin Rick Baniak Lockheed Aerospace Systems Northrop Aircraft Division Corporation—California Fred Timson_ Division Fernando Heyer Pai Fae Rockwell International Group Boge ioharton George Bostter ore Dick Goven Jack WoodsCONTENTS PREFACE .. SUMMARY . ACKNOWLEDGMENTS . FIGURES .... Section 1 INTRODUCTION ...... Background .... 6... Approach noba I MATERIAL CHARACTERISTICS . Composite Materials . ‘Metal Alloys II DESIGN CONSIDERATIONS AND MANUFACTURING PROCESS TECHNOLOGIES u Composite Design and Analysis Considerations . rt Composite Fabrication Techniques 236 New Metal Processes . 51 IV, COST INFORMATION 55 Industry Worksheets... 55 Underlying Assumptions. 5 Summary of Worksheet Responses 8 Applying the Data : 7 er Summary oo...) SI 70 V, SUGGESTED METHODOLOGY FOR ASSESSING THE EFFECT OF STRUCTURAL MATERIAL COMPOSITION ON OVERALL AIRFRAME COST n Description of Approach n Application to Hypothetical Aircraft. 1 VI. CONCLUSIONS = Relative Material Coste 80 Assessing the Effect of Structural Material Comps Overall Airframe Cost a1 Future Work 82 Appendix ‘A. SAMPLE WORKSHEET ...... B. OTHER STUDIES OF COMPOSITE MATERIAL COSTS | C. COST ELEMENT DEFINITIONS ..... D. DEFINITION OF TERMS . BIBLIOGRAPHYBepneseepe . Advanced composite material usage vs. first flight date |. Tension comparison: Specific stiffness vs. specific strength for unidirectional 2. Matrix processing requirements: 3. Matching aircraft temperature requirements to material service temperature . "The ceramic mold process for producing complex shapes from powder . A typical superalloy hot isostatic process cycle FIGURES F-144 structural material distribution . F-15E structural material distribution F/A-18A structural material distribution B-1B structural material distribution AV-8B structural material distribution. ‘V-22 structural material distribution . Advanced composite classifications . . Reinforcement forms ..... an Compression comparison: Specific stifiness vs. specific strength for unidirectional prepreg... 2+. 2-.0ee0eceeeeeceeseeeeceeeeeeeees prepreg rmosets vs. thermop limits. . Effect of improved operations on manufacturing hours of composite stabilizer | skinsTABLES 1. Airframe functional categories considered. . A 2. Summary of historically achieved composite weight savings. : 3. Matrix material characteristics... .. pononoG 4, Material properties (tensile) 6. 6 . Process manufacturing requirements and costs . . Suitability of manufacturing processes to alternative ‘manufacturing forms 7. Sensitivity of NDI methods to different flaw types. 8. Breakdown of fabrication effort 9. Manufacturing composite stabilizer skins: Baseline and improved operations 10, Study participants . ; és 11, Structural materials list... - 6. eee 12, Nonrecurring engineering hours per pound ratios 13, Nonrecurring tooling hours per pound ratios 14. Recurring engineering hours/pound .. . . 15, Recurring tooling hours/pound . 16. Recurring manufacturing hours/pound . 17. Recurring quality assurance hours/pound 18. Material cost factors . 19. Cost improvement curve slopes and weight sizing factors 20. Industry average fully burdened labor rates. 21, Late 1980s: Recurring dollars per pound (industry average) « - 22, Mid-1990s: Recurring dollars per pound (industry average). 23. Overall recurring dollars per pound (Industry min/max) . 24, Part selection asa function of graphite composition for a generic fighter 25, Percentages of functional cost elements attributable to structure... .. \irframe CERS based on sample of 13 post-1960 aircraft 27. Airframe materials utilization 28. Relative material complenity factors . 29, Functional cost element improvement curve slopes. 80. Required imputs .......ssccceeeveeeteeeseeeceeeeeees 81. Effect of composite substitution on airframe cost... +. see ee eeee 32, Effect of projected manufacturing improvements on airframe cost . 33. Materials, costs, and time periods . a 34, Aluminum and graphite/epoxy cost comparison B.1. Summary of ASD materials survey .1. WBS categories included in RAND C2. RAND airframe cost elements . BRSSLSKRSSSSSSRSAES BLSSSsasaI, INTRODUCTION Military applications of advanced materials—composites and new metal alloys—are found in aircraft, satelite, and missile structures; engine parts; solid rocket motor casings; and pro- tective armor. ‘This study concentrated on the issues surrounding advanced material usage in military sireraf. Emphasis has been placed on composite materials because of their relative importance in such applications, although advanced metal alloys are also described. Composites can offer several advantages over conventional materials. Their superior Jntness, strength, and stiffness aid in improving aircraft: maneuverability and expanding flight envelopes. ‘They also have greater fatigue and corrosion resistance than conventional rmetals, Weight savings achieved through the use of new materials may increase payload-range capsbility, provide the opportunity to downsize subsystems at a constant performance level, or allow for better fuel efficiencies. Because composite parts are built up ply by ply, united designs, which may require fewer fabrication and assembly hours, are possible. Some compos- ites may also play a role in meeting low-observability requirements because they have desirable electrical properties Given the evolving state of materials technology, considerable uncertainty surrounds the acquisition cost of future aircraft structure. Factors that contribute to the uncertsinty inelude: ‘The material composition of future aircraft Speeifie properties of comm=-.:aity available materials. Matocinl price treads ‘Tre availability and m=cusiiy of material processing technologies. ‘The !evol of automation in aircraft design and manufacturing processes. ‘The incentives to invest in the capital equipment necessary to automate the manufée- turing process and work with certain categories of new materials given fewer new tary programs. ‘The materials of construction as well as the methods of design and production are charg- ing dramstically. As a result, the use of cost estimating relationships (CERs) based solely on historical data of the 1960s and 1970s to project the costs of airframes in the late 1980s and early to mid-1990s is a dubious proposition. Moreover, considerable amounts of design and production cost data for new materials are not likely to be available for some time, Conse- quently, this study was undertaken with the goal of providing cost analysts with an interim ‘method for assessing the costs of future aircraft systems. More specifically, the objectives of this study were to (1) prepare a technical primer on advanced structural materials and (2) develop a methodology, suitable for use in the early planning stages of a program, for estimat- ing airframe structure acquisition costs. rented tes Gr weeny impact ach se Spel fa) ote tny sey to per hoy, om iow enn inpacts Sacha roped is) nr ars thay Sosy to epee An ‘Senn they are efi new ochnlogy, tei ons (oma Gllarper pound ben) tend to be higher than cost of mare conventional ateBACKGROUND Because of the advantages they offer, advanced composite materials have increasingly been incorporated into aircraft. Figure 1 shows the trend of composite material incorporation into military aircraft as a percent of structural weight. As the figure shows, composites have become an increasingly important materiel for airframe structural applications.* Initislly, composites were incorporated into non-flght-rtical secondary structures such as access doors and flaps. Later, as data and experience were gained, the amount applied to aireraf: structures increased to the point of being incorporated into primary (load-bearing and Alightritial) structure and complex-shaped components such as wing skins, stabilizer skins, and longerons. For example, boron/epory was used in an F-4 rudder and F-L11A secondary structure before being introduced into the F-14A horizontal stabilizer (the F-14A was the first aireraf: to incorporate composite material in a flight-critical component). Figures 2 through 7 show the structural material distribution on the F-14A, F-15E, F/A-18A, B-1B, AV-8B, and V-22, They are arranged by date of first flight, and they illustrate the range of shapes, sizes, and locations of various airframe subessemblies where composites have been used. 50 10 ‘Composite percent of structural weight 6 6 72 76 80 a 88 2 Year of frst fight NOTE: n addon tothe V.22, several other aeration development ae projected Yo make ‘edonsive use of composite tne 12, ATF (VF-227YF-23), andthe 8-2. Fig. 1—Advanced composite material usage versus first flight date The incorporation of composites into commercial aircraft has lagged military applications, primarily beause the commercial vironment is more sensitive to cost considerations. As peints of comparison, only abot 1 percent ofthe Boeing 747 structure weight is composite, while the Boeing 767 has nly about 3 percent. However, the Beech Star hip 400 Th (structure weight) business aireraft with a top speed of 887 mph, ie 67 percent compote (45 percent ‘aphit/oposy and 22 percent Kevlar and Gbergias)a . me e wos OS memset oy Fas ct FM oy Se oa _ see BO women a aS = ys rT — eee maamnn fags 0 ae unm e wat" 2 ol, a a em er oS ee etn Bln AJVuOUlY asi-s Nmoopivaue ONIENLOVaNNVINMANUFACTURING BREAKDOWN FIA-18A AIRCRAFT i i i 5 Me als ae 1 js it K i . Sy : cy 4 ‘ 4! at ll é 4, ro &: 4) ilu “ la Pol @@ AX _ Sia yi Yar W Wi,B-1B MATERIALS MATERIAL ENTAGENMOG)V3YE ONIENLOVANNVN LAVHOUIY OWSN 88-AV enc ming cr a emresantya onnqunsip [eHow yeimonNs Ze-A—L Bh 108 J0 Asaun09 ounbiy6 To dite, the advanced composite materials of primary importance and usage are the epcx- its (Kevlar/epoxy, graphite/epoxy, and boron/epoxy). Greater application of toughened epex- , graphite/bismaleimide, and other graphite/polyimides is expected in the future. Much attention has also been paid to graphite/thermoplastic materials because of their potential pro- cessing ease, although at this time they are much less mature. According to the open litera ture, both graphite/thermoplastic and graphite/bismaleimide are being planned for the advanced tactical fighter (ATF) prototypes. On the advanced metals side, greater application of aluminum-lithium alloys and superplastically formed aluminum and titanium is expected. APPROACH ‘Typically, RAND studies collect actual data and perform statistical analyses to develop CERs.* In this instance we believed the many uncertainties surrounding the use of these new materials end the dynamic nature of the material technology and its associated manufacturing processes warranted an industry survey. More specifically, the reasons a survey approach was ‘chosen rather than a statistical analysis of historical data are as follows: * The current data (for which production experience is available) are limited in terms of: — The number of observations. ‘There are only a half dozen historical data points {military aircraft programs) encompassing all composite material type: —The range of material types. Some materials, such as aluminum-lithium and sraphite/thermoplastic, have not been incorporated into production aircraft; as a result, no historical data, except for data based on developmental experience, exist ‘or these materials. — The level of usage. Projected levels of usage are far beyond what has been attained hy existing production aircraft, ‘The manufacturing technology is rapidly evolving. ‘The worksheet, a copy of which is provided in App. A, consisted of five sections: car- porate history, material usage within an aircraft, material technical information, cost (both nonrecurring and recurring cost elements), and general questions. Information was collected by material type for materials that would potentially be used in airframe structure no later than the 1290s. Materials of interest are those either currently in production aircraft or just recently out of the laboratories. ‘Two time periods were considered: the late 1980s and the mid-1990s. Data for the late 1980s were to reflect a company’s current experience, while data for the mid-1990s were to reflect: a ccmpany’s best judgment regarding the evolution of the basic technical knowledge of ‘the materials as well as of design and manufacturing techniques. Summarized cost information is reported for each of the functional categories listed in Table 1. Improvement curves, weight sizing facters, and material buy-to-fly ratios are also reported. ‘Although this analysis has emphasized airframe structure costs, overall airframe costs include not only the basic structure, but also airframe subsystems (e.., electrical, hydraulic, environmental) and final assembly/integration. We have therefore developed a method for estimating total airframe costs by combining the results of this study with those of a previous Por example, soe Hess and Romanoff 1967; Birkler, Garfnke, and Marka, 1982.16 Table 1 AIRFRAME FUNCTIONAL CATEGORIES CONSIDERED onctional Categorie ‘Aadrened Engineering NRE Tooling NRE Manufacturing R Quality assurance R RAND study.‘ As a result, structure costs are placed in context and not considered in isola- tion Section I contains a description of the new materials for those unfamiliar with composite materials and advanced metal alloys. Section III describes the manufacturing processes associ- ated with advanced materials. The emphasis is on composite fabrication techniques. Section TV presents the nonrecurring and recurring cost information collected with our worksheets, and Sec. V describes our suggested methodology, with an example. The last section, Sec. VI, resents our conclusions. The appendixes contain a sample of the worksheet distributed to industry (App. A), brief synopses of other composites cost studies (App. B), definitions of cost elements (App. C), and definitions of terms (App. D). “Hes and Romanoff, 1967 ‘To compare this data collection effort and methodological approach with tho literature, see App. B, which briefly esriben other recent composites cost stdin.Il. MATERIAL CHARACTERISTICS ‘This section discusses new material composition, properties, and general characteristics. ‘The intent is to familiarize cost analysts with the new materials so that they may better under- stand them and be able to assess the uncertainties and risks involved, ‘Advanced materials are classified into the following categories: ‘* Advanced composite materials: ceramic matrix composites, polymer matrix compos- ites, and metal matrix composites. Composite materials, polymer matrix materials in particular, are the new materials of primary concern for future aircraft structure and as a result are emphasized in this report. Their strength, stiffness, and cost properties are particularly suited for aircraft structural applications, Metal and ceramic matrix composites are at a much earlier stage of development than polymer matrix com- posites. Metal matrix composites offer superior temperature capabilities but parts fabrication is still difficult. Ceramic matrix composites also have superior temperature capabilities but are brit- ‘tle and therefore not suitable for most structural applications. ‘Aluminum, titanium, and steel, the materials historically used in aircraft, will continue to be used. Titanium is used in military aircraft where strength, toughness, heat resistance, and high structural efficiency are required. The Air Force Materials Laboratory is studying powder metallurgy alloys, but they are still in a developmental stage and their incorporation into air- craft is further on the horizon. Consequently, they will not be discussed here in detail. Composite materials have two constituents, the fiber and the matrix. It is common practice to refer to a composite material first by fiber (or reinforcement) material type and then by the ‘matrix type (see Fig. 8). The matrix holds the fibers together, redistributes the loads between the fibers, and plays an important role in many of the operating characteristics of the composite material (service temperature, resistance to chemicals, damage tolerance, etc.). The fiber/matrix interface can also be an important. consideration since relative slippage or disbonding can occur under various conditions. The fiber gives the material its strength and stiffness properties and is thus the load bearing element. Fibers may be short (chopped) or long (continuous). The five fibers predominantly used are glass, Aramid (DuPont's Kevlar is inthis class), graphite (carbon), quartz (silica), and boron. Glass reinforced plastic, or fiberglass, was first used in the 1940s and 1950s in small boats, aircraft, radomes, and automobiles. It was the first composite material developed, but itis not considered an advanced composite because ofits low-value properties. Car- bon fiber, developed in the 1960s, is stronger and stiffer than glass, making it a primary candidate for structures with high loads. Aramid fibers lie somewhere in between glass and graphite fibers in terms of strength and stiffness. Aramid’s toughness makes it suitable for ballistic protection. COMPOSITE MATERIALS. Reinforcing Elements (Fibers) ‘The reinforcing elements provide the required stiffness and strength necessary to carry the primary structural loads, There are three forms of reinforcements: continuous fibers, short or long whiskers, and particulates. Continuous fibers may come in unidirectional tape or Ww6 ‘Advanced Composites Ceramic Matrix Composites | | Polymer Matrix Composites | | Metal Matrix Composites (Mc) (PMc) (MMC) ‘Thermosets ‘Thermoplastics + Epoxies + Polyimides + Bismaleimides Fig. 8—Advanced composite classifications (by matrix material) woven (fabric) form. ‘The various forms are shown in Fig. 9. Fiber reinforcements dominate itary applications and are the only reinforcements considered in this report. The effective- ness of any reinforcement is largely determined by volume fraction, orientation, and length. ‘The volume fraction of the reinforcement is the percentage of the total volume of the part filled with the reinforcing material. Thus, a volume fraction of 0.6, typical of many ‘graphite/epoxies, means that parts fabricated from these materials are composed of 60 percent fibers end 40 percent matrix material (by volume). In addition to such factors as the stiffness, strength, and the volume fraction of the fibers, the effect of fiber orientations on final part properties is also very important. The con- cept here is the relative alignment between the directions of the critical loads and the direc- tions of the fibers; the greater the alignment between the fibers and the loads, the greater the structural efficiency. As the fibers become more randomly distributed with respect to the applied loads, structural efficiency reduces substantially. Thus, randomly oriented fibers are not used in primary structures. Designers will strive to align the fibers and the loads to minimize the laminate thickness (and the weight of the part). However, in many practical parts, a certain degree of structural efficiency (weight) is traded for higher margins of safety, which accounts for the uncertainties associated with composites. ‘The optimum ply layup angles (or fiber orientations) for a complex composite part are not completely determined in advance. During the design process, the orientation of the fibers in a part may experience several iterative changes, for several reasons: a better understanding of the loads that will be applied to the structure, the incorporation of producibility concerns,1 Fabric ‘SOURCE: Now Structural MatorialeTochnclopls: Opportunii for ho Uso of Advanced Coramics and Composies, OTA, Septomber 1986, p. 43. Fig. 9—Reinforcement forms changes in the stress analysis, etc, Also, the relative newness and uncertainty associated with composites can also result in conservatism with respect to the laminate thickness and the layup angles. Finally, as the length-to-diameter ratio of the fibers increases, the reinforcing effect, increases. Therefore, long, continuous fibers contribute more stiffness than short fibers. Graphite filaments are composed of thousands of individual fibers. ‘The diameter of a single graphite fiber is about 0.0004 inches. Filaments are packaged in a variety of ways in ‘order to suit the needs of industry. Unidirectional tape and multidirectional febrics are most, ‘common. Unidirectional tape has the filaments aligned with the rolling direction of the tape. Febrics ere supplied with filaments running in two (or more) directions. Often, these fabrics ‘and tapes have been preimpregnated with a resin matrix material and are called “prepteg” tapes. Although fibers can be woven together in three or more directions, the maximum weight savings will generally be achieved when uniaxial tapes or biaxial fabrics are used; many struc- ‘tures are loaded predominantly in one direction, and a weight penalty must be paid if extra fibers are aligned at angles to the load. For most airframe applications, continuous graphite (carbon) fibers have been used as the ements. Such other materials as boron and Kevlar (Aramid) are also used, but technology has dominated because it provides fairly high performance at “reason- able” cost. Boron fibers surpass graphite in terms of strength, but the cost of boron (raw material) is about three times that of graphite. Boron fibers are also more difficult (and ‘expensive) to machine because they are harder and about ten times thicker than graphite. Such thick fibers are limited to parts having radii of curvature greater than 3 in. and thus are eliminated from many substructural applications (graphite fibers may be used in parts with0 radii of curvature as small as 1/16 in.’ Kevlar fibers have very good tensile strength proper- ties and are cheaper than graphite. However, they are considerably less stiff and have lower compressive strength. These weaknesses eliminate Kevlar from many critical applications. res of performance potential, “specific stiffness” and “specific strength,” are defined by dividing the stiffness (or modulus) and strength of the material by its density. The results are a measure of a candidate material’s efficiency and are a useful comparative tool in fan industry that often describes structural components as “stiffness critical” or “strength criti cal.” Although there are other issues to consider, a comparison of these normalized properties highlights the superior performance potential of the advanced composites (see Figs. 10 and 11). On a “per pound” basis, the advanced composites are inherently stiffer and stronger than metals, ‘The properties in Figs. 10 and 11 are given for both compression and tension to highlight the faet that some materials (Kevlar, boron, glass) perform very differently depending on their stress state? Usually, the compressive strength of the material is the weaker of the two; this is 1200 1000 _- Se eo 800 600 400 ‘Specific stitness (inches x 10°) ‘Specific strength (inches x 10°) Fig. 10—Compression comparison: Specific stiffness vs. specific strength for unidirectional prepreg DeD/NASA Advanced Composites Design Guide, 198 the diinction between simple tenile and compresive tres states can be made in the following way. If an cbjet ts undergoing relative elongation in the direction of che applied load and simultaneous “thinning” in the {tanrvene (or to-lond) cretion, it sin a tate of tension. Ifthe object is shortening in the direction of the fed and “bickering” inthe no-load direction ils in compresion. A simple but extreme example of this distinction ean be ‘made with putty, Pulling on «piece of eof pty from both ends produces avery noesble elongation and thinning to the pont of breakage. Squeezing the putty renuls ina shortening and thickening due to comprewion. Compressive Toad my reul fn a buckling instabllity long before theultimate compressive strength of the material is reached. ‘Thea 1200 1000 8 2 ia 8 Eiioas de Specific stiffness (inches x 106 ) 200 | Metals E-ciass/PMC a 8 Specific strength (inches x 10°) ‘Fig. 11—Tension comparison: Specific stiffness vs. specific strength for unidirectional prepreg true for Kevlar and glass. However, boron fibers are actually stronger in compression than in tension. Graphite fibers have approximately equal properties in tension and compression. Such responses must be considered when deciding which materials are appropriate for certain parts. Kevlar is very common in pressure vessel applications because the fibers are predom- ‘nantly in tension, and Kevlar is fairly cheap and damage tolerant. Graphite is more appropri- ‘ate for structures that are subject to substantial bending loads, which generate areas of both. tension and compression in the part. In Figs. 10 and 11, the rather large bores representing sraphite/epoxy reflect the general range of properties that can be provided by suppliers, depending on the specific application, Note that not every point in the graphite/epoxy boxes ‘can be achieved in practice; they reflect the approximate range the materials’ properties currently cover. Certain properties of fibers can be enhanced, but often that will be at the expense of other properties. For example, graphite fiber stiffness and strength are inversely related; as cone is enhanced, the other is decreased. Several varieties of graphite fibers are on the market, ranging from high strength graphite (with lower stiffness) to high stiffness graphite (with lower strength). ‘rednpoiten of a wtrctare to buckling instabilities can be mathematically calculated and is 9 strong function ofthe ‘Rroctare'spometey ea well as its material stiffens. For example, a log, thin stel rod may buckle at low comsres- eiittae "f'ahor, thick steel rod will probably never buckle end will simply fall (crush) when the compreaive ftreogth of the ste! ie reached.In summs © The fibers provide structural stiffness and strength (which can save weight). © Graphite fibers have the best combination of strength, stiffness, and cost and are therefore the most widely used fiber. + ‘The material formebility and machining characteristics are largely determined by the fiber selection. * Fibers are provided in many formats: unidirectional (tape), bidirectional (fabric), etc. + Final part properties are largely determined by fiber volume fraction, fiber orientation, and fer length. Matrix Materials The function of the matrix is to facilitate load transfer between the fibers, which plays an important role in the longevity and damage tolerance of a composite part. Damaged fibers ‘must be able to redistribute their loads through the matrix to the surrounding undamaged fibers ‘An intuitive example of the role of the matrix is the response of a composite part to compressive loads. Although the fibers resist most of the load, the matrix lends the necessary geometric support. If the matrix disintegrates or dissolves, the fibers cannot carry any ‘compressive loads and the structure collapses. ‘Thus, the condition of the matrix is extremely important with respect to compressive structural loads. Tensile loads, however, may still be carried by the fibers even if the matrix fails. ‘The matrix/fiber interface may also help to stop cracks that originate in the fiber from Propagating catastrophically. A local impact can break a few fibers, but crack propagation should be inhibited by a ductile matrix or a fiber/matrix debond. Matrix mechanical properties such as stiffness and strength are substantially lower than those of the fibers. Because of this relative weakness, the matrix determines the mechanical, thermal, and chemical limits and the relative moisture levels of the material system. In opera- tional terms, the matrix determines the environmental limite (temperature, chemical solvent exposure, humidity, etc.) of the composite part. Industry has been working to develop tougher and more damage resistant matrix systems, ‘but such advances have occurred more slowly than fiber stiffness and strength improvements. ‘Also, superior performance requirements often translate into additional processing difficulties ‘and hence costs increase (at least until the manufacturing processes become more mature). ‘The advanced composite material systems can be divided into three general categories: polymer matrix composites (PMC), metal matrix composites (MMC), and ceramic matrix com- posites (CMC). PMCs have prevailed in the airframe industry for more than 25 years because they strike an effective balance among performance, versatility, and cost. Metal matrix materials have superior high temperature properties but are very expensive to manufécture. Although several demonstration components have been produced from metal matrix materials, the demand for them is held low by cost considerations. The mechanical properties of ceramic ‘matrix composites are unsurpassed at elevated temperatures, but they are practically unused in airframe applications because of very low fracture toughness values. Current retearch ‘emphasizes the toughening of ceramic matrix composites. Each matrix type is discussed below in more detail Polymer Matrix Composites. Polymer matrices can be classified as “thermosets” or “thermoplastics,” depending upon their chemistry and the cycle that is required to processthem. Both have organic chemistries and service temperature limitations between 200 and 600°. ‘Thermosets have polymer chains that cross-link during the application of a prescribed schedule of temperature and pressure (cure cycle). This structure promotes dimensional stabil- ity, solvent resistance, and high temperature resistance, but itis also irreversible and slow to eure. For example, a typical epoxy part will require about 8 to 12 hours of processing time, and once cured, it cannot be reprocessed. Most epoxies require processing temperatures around 350°F. ‘The defense industry has been working with fiber-reinforced, thermosetting epoxies for 25 years and is thus fairly well aware of their strengths and weaknesses. Concern about the lim- ited service temperatures (275°F under hot/wet conditions) and low toughness of epoxies resulted in the development of bismaleimide and polyimide systems. Bismeleimides are actu- ally “addition” polyimides that have lower processing (and operating temperatures) than the rest of the polyimide family. Bismaleimide service temperatures can reach 450°F and can be processed with the same equipment (autoclaves) used for epoxies. Other polyimides can reach service temperatures of up to G00°F, but their processing temperature exceeds G50°F. This is an important manufacturing threshold; major modifications to manufacturing processes and equipment are required above 625-650°F. ‘Thus, polyimide parts are more expensive to pro- duce. Production parte made from these advanced thermoseta have begun to appear ( ‘raphite/oismaleimide parts on the AV-8B). Bpories are hampered by several concerns. Perhaps the primary concern is low tough- ness. This means that extensive internal damage (delamination), invisible to the naked eye, :ay result from such low energy impacts as the dropping of a tool. ‘This damage can have a severe degrading effect on the compressive properties of composite parts. Consequently, com- posites are penalized by the use.of low design strains in order to conservatively compensate for this weakness. Tougher matrices would allow designers to increase the composite design strains ard have higher confidence. Over the lifetime of a part, some epoxies gain weight because of their tendency to absorb water. Obviously, this undesirable feature reduces mechanical properties and vehicle perfor- mance factors. Additionally, vehicles that engage in prolonged supersonic or hypersonic flight will not be able to use epoxies (because of @ 275°F service temperature limit) in areas of extreme cerodynamic heating. In such areas, the alternatives might include polyimides, metal matrix materials, and carbon/carbon. Thermoplastes have been the subject of intense research because of their potential for increased toughness, short processing times, and reformability of parts. In 1982, polyethereth- cerketone (PEEK), was the first thermoplastic to have both reasonable high temperature perfor- ‘mance ard good solvent resistance. Until then, either of the two, but not both, of those characteristics could be achieved in a thermoplastic. Both, however, are necessary in most er- frame applications. ‘There are many thermoplastics on the market; some of the more common include polyetheretherketone (PEEK), polyphenylene sulfide (PPS), and polyetherketone (PEK). ‘Thermoplastics do not undergo any permanent chemical transformations (such as crosslinking) during the fabrication process. ‘This means that scrap and mishandled parts might be reformed by the application of heat and pressure. ‘Thermoplastics are more flexible than thermosets with respect to cure cycle perturbations; although small changes do not seem to bother thermoplastics, they can have substantial effects on thermoset parts. In addition, thermoplastic prepregs do not require refrigeration, and, perhaps most important, partPy fabrication times are measured in minutes rather than hours (thermosets). ‘Thermoplastics can ‘also be welded. ‘Thermoplastics are potentially more resistant to impact damage and cracking than ther- ‘mosets. This could have a considerable effect. on design strain levels and, therefore, weight savings. High temperature strength and solvent resistance are still fairly low. ‘These deficien- cies are being examined and there is evidence to suggest that improvements will be made to make them more competitive with thermosets. The long-term mechanical properties (fatigue ‘and creep) of reinforced thermoplastics are uncharacterized. Fatigue and creep tests are expen- sive and time consuming to perform; they are not often done in the early stages of material development. Confidence in the material is therefore reduced until these properties can be well characterized. Other deficiencies of thermoplastics include poor drapability and lack of tack. ‘This means that the prepreg plies are stiff, difficult to form, and do not tend to stick together after pressure has been applied. Heat must be applied to thermoplastic prepregs during the layup process in order to enhance ply formability. Therefore, some modifications will have to be ‘made ‘0 any automated layup procedures that have been designed for thermoset plies (which are naturally formable and sticky). High processing temperatures (650-800°F for thermoplastics vsraus 350°F for thermosets) fare alto a serious consideration. When processing temperatures exceed 650°F many conven- tional manufacturing procedures and equipment must be revised or replaced. For example, metal tools will degrade very quickly, rubber tools:cannot be used at all, and autoclaves are subject to extreme wear. Manufacturing processes that minimize autoclave use for thermoplas- tos are being developed (such aa pultrusion). In Fig. 12, thermoplastic and thermoset processing temperatures and times are compared. ‘An autoclave cure cycle for a thermoset part can require as long as 12 hours when heat-up and cool-down times are considered," but many thermoplastic parts will probably require less than cone hour of processing time. However, part consolidation for thermoplastics requires tempera- tures in excess of 650°. Such temperatures place heavy demands on both autoclaves and the conventional manufacturing techniques, which have been oriented toward the mild processing temperature (350°F) of thermosets, In fact, autoclave use tends to negate the advantages of thermeplastics because a substantial amount of time is required just to heat up (and coo! down) all ofthe air and mold mass within the autoclave to such high temperatures. ‘Therefore, other fabrication concepts are required and are being developed. However, autoclave use may bbe unavoidable for certain kinds of parts. According to a National Research Couneil report,* the following processes have been used to either consolidate or form continuous fiber reinforced thermoplastic parts: Autoclave lamination and molding. Continuous lamination. Filament winding, Pressure forming. Pultrusion, Roll forming. ‘Vacuum forming. *Thisincludes autoclave time only. Post curing is generally done in less expensive ovens. “Natinal Research Coun, 1987+1000 ‘New fabrication concepts required K Above o threshold < temperature : 2 Thermosets é & Thin Thick 200 laminates laminates ' 1 1 ° ° 2 4 6 8 a a eT) Processing time, hours Fig. 12—Matrix processing requirements: ‘Thermosets vs. thermoplastics Unreinforced thermoplastics as well as those reinforced with short, randomly oriented fibers have been used in low-temperature, secondary structural applications for several years. But advanced composite thermoplastics with continuous fibers are a new concept with little operational experience. Much emphasis is being placed on these materials in current design concepts, but the database is small when compared with fiber-reinforced thermoset systems epoxies). It will take time to realize the full advantage of the benefits offered by rein- forced thermoplastics and to overcome the technical problems that will surely arise. Some trephite/PEEK wing skin panels for the F/A-18 were installed and evaluated in 1988. Other demonstration components (a torque box, an ATF prototype wing, and a helicopter tal) have been made as well, but long-term, in-service observations have not yet been made. Current ‘thermoplastics should retain their integrity up to 200-350", which is comparable to thermoset temperature Metal Matrix Composites (MMC). Metal matrix composites are at « much earlier stage of development than PMCs. These materials offer superior operating temperatures (eg. 1,000°F for titanium) but are hampered by complex and immature fabrication procedures. Using fiber reinforcements with metal matrices yields superior temperature capabilities while 1aintaining strength-to-density properties greater than those achieved by superalloys. Meval matrix composites offer better compression strength than polymer or ceramic matrix compos- ites. Aluminum is the most common metal matrix; but other metals are also used, including aluminum, titanium, magnesium, copper, and superalloys. Some of the reinforcing materials include grephite, boron, silicon carbide, and aluminum oxide. Manufacturing processes includehot pressing, casting, hot isostatic pressing (HIP), pultrusion, forging, and superplastic forming (SPF)/diffusion bonding (DB). Temperatures for these processes can exceed 2000°F, and pres- ‘sures of up to 2000 psi may be required. ‘The ATF may offer the first opportunity to test the feasibility of primary MMC struc- tures in a production environment. Lockheed is in the early testing phase of a program intended to demonstrate the superior performance and cost competitiveness of MMC sireraft parts. They are fabricating and testing four vertical tail structural boxes that could be used in fn ATP design. ‘These structures have skins fabricated from silicon carbide reinforcements and an aluminum matrix. Previous Lockheed trade studies concluded that MMC fightr ai- craft components can reduce structural weight 15 to 30 percent and reduce production costs 5 to 50 percent, Testing of F-15 tail sections made out of metal matrix composites is also being carried out. ‘Most current applications for MMCs are in aircraft engines and missile structures. A ‘MeDonnell-Douglas study concluded that existing MMC materials could greatly improve the range (9 percent) and reduce the structural weight (20 percent) for certain mi ‘designs. ‘MMCs have also found applications in certain space structures (space telescope) where a zero coeffient of thermal expansion is desirable. They may also be necessary for future vehicles that travel for long periods at supersonic speeds. MMCs are also experiencing some use in the auto industry; Toyota and Mitsubishi are producing pressure-cast, aluminum matrix engine parts MMCs are currently limited by their high manufacturing cost, fiber/matrix interaction problems, and limited database. Ceramic Matrix Composites (CMC). Ceramic matrix materials are at an even earlier ‘stage of development than MMCs. These materials are neither organic nor metal. They have superior wear resistance, high temperature strength, and chemical stability compared with metals. Two obstacles to widespread use of ceramic matrix composites are (1) the development: of high strength, high modulus, small diameter, continuous fibers whose mechanical properties are no; drastically degraded by ceramic matrix processing or operating conditions and (2) fabri- cation processes that result in a uniform microstructure of nondegraded aligned fiber sur- ‘rounded by low porosity matrices.’ Poor fracture toughness properties make them susceptible to sudden, perhaps catastrophic, failure. This drawback removes ceramics from most primary structural applications. Research is being conducted to improve their fracture toughness, and demand for them will increase. Matrix materials for CMCs include carbon, glas, silicon car- bide, ond alumina. Manufacturing processes include sintering, hot isostatic pressing, hot press- ing, malt infiltration, and chemical vapor deposition. Military applications for ceramics are currently limited to auch structures aa radomes, bearings, and turbine engine parts. Carbon/carbon is more advanced than other CMCs and is finding uses in such high temperature environments as solid rocket motors and space shuttle leading edge. ‘Matrix Summary. The primary role of the matrix is to transfer loads between the fibers. It determines the following important properties of composites: resistance to crack pro- pagation (fracture toughness), damage tolerance, solvent resistance, processing temperatures, Processing times, and corrosion resistance. "Department ofthe Interior, 1888.n ‘Thermosets + Thermoset epoxies have dominated the airframe market. ‘© ‘Thermoset advantages include good solvent resistance, low processing temperatures and pressures, prepreg formability, and large experience and mechanical property data- bases. ‘© Thermoset disadvantages include long processing times, low toughness, high absorp- tion of water, short prepreg shelf life, permanent cure (parts cannot be reformed), end difficulty to repair. ‘Thermoplastics ‘* Fiber reinforced thermoplastics may greatly reduce manufacturing costs. Many demonstration components have been made but production run data is not yet avail- able. ‘© Thermoplastic advantages include very fast part consolidation, reform ‘toughness, and long shelf life. ‘© Thermoplastic disadvantages include small experience and mechanical databases, bigh ‘ew material costs, high processing temperatures, poor drapability, and low tack. Others ity, igh ‘© Metal matrix materials have excellent high temperature properties but are still extremely expensive to manufacture and are subject to wide variations in properties between batches. Demands for these materials will increase. © Ceramic matrix materials are still too brittle for current airframe applications. Composite Materials and Their Applicability Several factors normally enter into the selection of airframe structural materials, includ- ing strength, stiffness, density, fracture toughness, fatigue crack resistance, corrosion resis- tance, temperature limits, produetbility, repairability, and cost. However, for supersonic air- craft, operating temperature drives material suitability. Figure 13a indicates how aircraft skin temperature varies with Mach number.® Two temperatures are shown—that encountered on the nose, leading edges, and air inlets, and that encountered over the remaining skin surface. ‘The surface temperature for subsonic jets cruising in the stratosphere is just below 0° F? but inereases rapidly with speed. Additionally, temperatures at the nose, leading edges, and air inlets increase more quickly and reach higher levels than temperatures over the rest of the air- craft surface. In Fig. 13b, the service temperature limits of the materials addressed in this study are shown along with two metal matrix composites. The epoxies, thermoplastics, and bismaleimides are pretty much limited to aircraft flying in the Mach 2.0 to Mach 2.5 range. Sustained speeds in the Mach 3.0 to Mach 3.5 range will require titanium, steel, or a titaniam matrix composite. The valves shown are, of course, approximations since actual skin temperatures wil vary with alrlow conditions, ‘surface Git and atmospheric conditions. “in the cae of absonieeirrat the skin temperate requirement is determined not by the crise Mach mum at tude but rather by the ambient temperatures the aircraft encounters,35 rz z 30 25 20 18 1.0 og “1000-100 200-300 400-500 600 700 60000 1000 Temperature (degrees F) (A) Aircratt skin temperatures as a function of mach number (cruise at 36,089 feet) Material ‘Aluminum Metals Polymer Matrix Composites ‘Metal Matrix Composites Epoxy® ‘Thermoplastic Bismateimide ‘Aluminum matrix ‘Aluminum ithium Titanium Stool ‘Titanium matrix 0100-200 800—-400~—«500~—~G00~— 700-800-900 1000 Temperature (degrees F) * Mostepo%ies are good to 250 degrees Fahrenheit a lew are good to 950 depres. (8) Matrix material service temperature limits (sustained) Fig. 13—Matching aircraft temperature requirements to material service temperature limiteComposite Material Weight Reduction Several operational requirements are fueling the incorporation of composites, one of which is a constant desire for increased payload capability. ‘The most prominent advantage ‘composite materials provide is their potential for saving weight. Three characteristics of com- posites contribute to this, These are: ‘© The favorable ratios of strength to density and stiffness to density offered by the fiber (see Figs. 10 and 11). ‘© The elimination of fasteners through a unitized design and cocured parts. ‘© The efficiency created by aligning the fibers in the direction of load. ‘The weight saving potential of composites has two aspects: (1) part-for-part weight reduction and (2) resizing, or scaling down of other parts and subsystems. ‘Table 2 compares the weights of components fabricated from metals and their composite equivalents (part-for-part weight reduction). Experience shows weight savings between 15 and 31 percent have been achieved, with weight savings for primary structure (wings, fuselage see- tions, stebilizers, etc.) being slightly less than savings for secondary structure (flape, fairings, Table 2 SUMMARY OF HISTORICALLY ACHIEVED COMPOSITE WEIGHT SAVINGS Fie Fight Weight etal Composite ere Savings Weight ‘Weight Date Component ()() Metals db) Composites 1980 AV-AB Wing 10 om Ofe 1914 BA Horizontal Stabilizer «16,815. 2780 G/F, B/B, ‘Gi, Ti AL N/A Ba Vertical Stabilizer 8 oe WA PSE Flap ” oe NVA PAA Overwing Feiting Es G/B, Be 1368 Fd Horizontal Stabilizer 18 3/7, Al, Stel B15 PAS Speed Brake a C/E, ALT, Steel 1062727 Horaootal Stabilizer 22220, m of 182727 Elevator BT 189 G/B ribs, spars, skins 1965 DC-10 Verto! Stabler 221,005 ‘N/A 1716 DCO Upper AR Rudder 274 20 G/EGYE 961 LOI Alleron aan 7 Ofer, spars skins; Kevler/Nomex ‘rnling-edge wedge 1964 Le1011 Vertical Stabilzer 28, 22 N/A SOURCES: Decamp and Johnson, 198% Zweber, Bade, and Yue, 187; Destin, 1086; Warwick, 198,2» elevators, ailerons, fins, and rudders). This was consistent with our industry survey, which indicated weight savings in the 20 to 25 percent range. More advanced composites may yield greater savings. A 1986 OTA study indicated that in the mid-1990s polymer matrix composites ‘may save between 30 and 40 percent over the weight of comparable metal structure, assuming advanced fibers and modified epoxy and thermoplastic matrices.? ‘The discussion above has focused on the weight savings directly achievable by substitut- ing composite structure (which is more structurally. efficient) for metal structure, However, ‘even greater savings in aircraft weight are obtainable through a process known as “resizing.” ‘That is, because some parts are now made of more structurally efficient composite, the remain- ing structure has less of a load to support and can be downsized. Moreover, reducing the struc- ture will cause the aircraft to require less fuel to accomplish the same mission. Less fuel ‘means a smaller fuel system is required. Consequently, there is more structure savings because ‘there is less fuel volume to contain. And as the weight comes down, other systems, such as the landing gear and hydraulic system, can also be reduced in size; but the biggest savings comes in the engine(s). Because the aircraft weighs less, it does not require as much thrust to meet any ‘acceleration or specific excess power requirements. Reducing the thrust reduces the weight and volume of the engine(s) and another round of structure and subsystem downsizing takes place. ‘Whether the weight differences associated with resizing are substantial depends primarily on two factors: ‘© The aircraft's mission and performance requirements. High performance fighter air- craft usually exhibit much greater sens ticularly cargo and bomber aircraft ‘* The freedom to make design decisions. ‘The greater the number of design decisions ‘that have been made, the more limited the scope for resizing. By far the most impor- tant of these decisions is the engine scale. For example, when the engine scale is fixed and structure weight is removed, the resulting performance gain cannot be “sized out” to save weight (and cost). Over the long run of aircraft evolution, technology advances, such as the incorporation of composite materials, have generally been used to achieve higher performance, not lower weight. ‘The weight savings discussed above are short-run considerations. A specific aircraft develop- ‘ment project will have weight, performance, and cost requirements or goals. Designers will ‘consider many tradeoffs among these various requirements. ‘The process of resizing is used extensively in generating alternative design solutions to meet the requirements, particularly during the conceptual design phase. Because only one design alternative is carried through complete development, it might be more appropriate to identify the lower weight of the final design solution compared to those designs not selected as weight “avoidance” rather than ‘weight “savings.” The important point is that the performance capabilities of the final design would have been achieved only at higher weight without the use of the advanced technology. ‘Some aavings presented in the table were based on shuminum structure and others on titanium. ottce of Technology Assessment, 1988, pp. 142-142a Composite Materials Summary ‘Tables 3 and 4 have been provided for a quick reference. They summarize characteristics of commonly used matrices (Table 3) and reinforcements (Table 4) for airframe structural ele- ments, Within each category there may be a considerable range in the values beyond those indicated here, depending on the specific matrix and supplier. ‘Table 3 ‘MATRIX MATERIAL CHARACTERISTICS "Temperatare CF) Service Damage Matric Process Limit Tolerance Polymers ‘Bpoxiest aso 280 tow Polyimidee 550-650 400-550 low ‘Thermoplastic 570-800 380 high Metals ‘Anima 1000600. _very igh ‘Teaniom 16901000 very high Ceramics 1000-3800 3000-1000 ow ome eporiee have service temperature Kita of 350° but the rule of thumb i 250" F Table 4 MATERIAL PROPERTIES (TENSILE)* ‘iifingae Stengih Dens Fiber/Matarial = MSE KSI"Ibjeuin. Fibers ‘Graphite 20-100 30-1000 065, Boron oO sot Silicon carbide 60600, OT Glas 10-12 400-600 07 ‘Aramid (Kevlar) 10-30 500045 ‘Mal materials ‘Alorinom or 80 0 ‘Titaniom 1518 100-180 36 Steet 25-30 200-900 28 “Walues are in the Aber direction and therefore sepresent the maximum fer eapebility. MBI is milions of pounds per aquare inch, KI is thousands of pounds per square inch.2 METAL ALLOYS: Aluninum-Lithium Aluminum-lithium (AI-Li) alloys are about 10 percent lighter (and stiffer) and 5 percent stronger than current aluminum alloys. The fatigue properties of Al-Li are also better. Lithium is the lightest metal, and adding small amounts of it to aluminum can greatly enhance ‘mechanical properties. Unfortunately, lithium is highly reactive and can explode during the alloying process. Only recently have these problems been solved sufficiently to produce a com- mercially available AI-Li alloy. Problems with Al-Li include weldability and batch-to-batch consistency. Aluminum-lithium might be used as an alternative to composites in some applications (bulkheads, superplastically formed internal fuel tanks) or to save weight if it is substituted for aluminum. According to some sources Al-Li designs might result in 15 percent part weight reductions, In a demonstration program, McDonnell-Douglas replaced conventional aluminum wing skins on the F-15 with Al-Li and reduced weight by 10 pereent. Powder Metallurgy Powder metallurgical (P/M) processes can offer several benefits with respect to conven- tional cast or wrought procedures. P/M components can have superior mechanical and micro- structural characteristics; they also tend to be more compatible with net shape operations (such ‘8 superplastic forming/diffusion bonding and hot isostatic pressing (HIP), which are discussed in the next section) and are more responsive to nondestructive evaluation (NDE) and inspec- tion (NDI) techniques. ‘These processes are being applied to conventional titanium and alumi- ‘num alloys (for example the 7000 series), AI-Li alloys, and superalloys such as Rene 95 and IN 100. Alcoa introduced the first commercially available aluminum powder metallurgy alloys in 1980, Powder metallurgy materials are made, as the name implies, using standard materials such as aluminum and titanium in powder form. Small amounts of other materials are then added, most frequently zine, copper, magnesium, or cobalt to enhance properties. Because fine Particles of the materials are mixed, or alloyed, better properties can be achieved. For exam- ple, AI-Li may be made in two ways, using traditional ingot-metallurgical techniques or powder metallurgy. If aluminum-lithium alloys are produced by ingot metallurgy, where the alloying cloments are dissolved in molten aluminum and then cast into ingots, the process effectively limita lithium content to 3 percent. Higher lithium contents, and therefore lower alloy densi- ties (which lead to weight savings) may be possible with powder metallurgy. Because P/M ‘materials are alloyed using finer grain structures, a more homogeneous distribution of elements is achieved, making for an optimum combination of strength and toughness, Lithium contents of up to 6 percent instead of 3 percent may be achieved if P/M techniques are used. Another example is aluminum alloys. An article in Flight International stated that com- pared with ingot metallurgy alloys, aluminum P/M alloys 7090 and 7091 offer higher strength and superior fatigue crack initiation performance. It went on to state that Boeing has used ‘P/M 7090 in its 85 Ib landing gear forging for the 757.” These materials can be provided in traditional forms. For example, the first commercially available aluminum P/M alloys—7090 and 7091—were identical in form to a cast aluminum ingot, and, according to Alcoa, subse- ‘Warwick, 1985, pp. 20-22.quent fabrication steps are essentially alike." They can also be provided in powdered form for ‘use in near net shape fabrication process. ‘The powder can be loaded into a preshaped form or cast and then it is typically subjected to extremely high temperature and pressure. Because the cast shape is close to that of the finished product, wastage of material is minimized, as is the need for labor. The downside is that mold costs are high.!? ‘The primary use for P/M parts has been in jet engine components. ‘These applications have been fairly routine since the early to middle 1970s, More work has and is being formed on applying P/M techniques to other airframe components. Consequently, about six F/A-I8 aircraft are gathering operational experience on P/M titanium alloy engine mount sup- ports and arrestor hook supports. The original (metal forging) hook support weighs 260 1b whl the hot isostatic pressed P/M part weighs only 65 Ib. P/M airframe parts are till some- what new and in need of further development. ‘They should be considered wherever there is a need for high strength alloys Tia, "clark and Flemings, 1986, p. 54III. DESIGN CONSIDERATIONS AND MANUFACTURING PROCESS TECHNOLOGIES ils give a designer added flexibility, but this flexibility complicates the ‘The thermal and mechanical properties of composites are different from those of meials. In addition composites are processed differently. Most are not machined, milled, oF formed as metals are; shapes are created by placing a series of plies on a form and then per- manently setting them under intense heat and pressure for a specified period of time. COMPOSITE DESIGN AND ANALYSIS CONSIDERATIONS Composites are two-phase, orthotropic matarils. This means that, unlike homogeneous isotropic metals, their mechanical properties can be tailored to meet specific design require- ments, For example, itis usually optimal to align the fibers predominantly in the direction of the load. Or it may be desirable to design a structure with a coefficient of thermal expansion of zero in a given direction. Tailoring permits the most efficient use of the material. Unfortunately, the very two-phase composition that allows structural optimization also introduces the potential for other, less desirable complexities. In large past, these complexities result from the interaction of the two very different components, the fiber and the matrix. ‘The fibers are strong, stif, and thermally resistant; the matrix is generally much weaker, less stiff, end more susceptible to corrosion and solvents. The fiber-matrix interphase is also com- plex and can greatly affect the toughness and damage tolerance of the composite. The chemis- ties end properties of both fiber and matrix must be understood and tradeoffs must be con- sidered to maximize the properties of interest (toughness, stiffness, strength, solvent resistance, ete) In addition, much has been made ofthe fact that material standards have not been estab- lished. Material properties can vary from batch to batch and among suppliers. Guidelines relating material properties to final part properties are still being developed. Knowledge is lacking on how structures fabricated with these materials degrade under load. Standards as to the effects voids and cracks per square inch (or in absolute size) have on final part properties or load capability have yet to be fully determined. Designers must consider this uncertainty and potential variability during the design process and additional modelling may be required. Other important examples of the differences between composites and metals complicate the design process. For example, when a hole is drilled into a flat metal plate, the stress con- centration factor at the edge of the hole has a value of three, meaning that the stress at the ‘edge of the hole is three times the “far” fold stress (a few hole-diameters away). ‘This is true for any metal and any plate thickness, In a composite plate, the presence of a hole can create stress concentration factors with values anywhere between two and seven, depending upon the laminste configuration. Fiber-dominated layups (those with a high percentage of zero degree plies) tend to have higher stress concentration factors than matrix-dominated layups (those with a high percentage of angle plies). Obviously, such complex responses must be considered in the design of the structure's attachment points. ‘A delamination is a separation between plies. Under tensile loads, delaminations are not serious, but they can have a tremendous effect on the shear and compressive responses of36 composite structures. If a delamination of a critical length were to go undetected, it could propagate under sustained or cyclic loads and ultimately cause the collapse of the structure. Delaminations often occur during the manufacturing process; a part may be laid up incorrectly, the cure eycle (temperature and pressure) may not be implemented properly, etc. ‘Time and resources must be spent inspecting composite parts for the presence of any delami- nations, which, if found, must be assessed in analytical or experimental terms. If any are con- sidered “critica,” repairs must be performed and the integrity of the part verified or the part ‘must be scrapped. Analytical Models ‘The analytical tools developed to predict the behavior of composites need further develep- ment. Many of these tools are based upon research into homogeneous, isotropic metals and are not generally applicable to composites, ‘The analytical tools that have been developed to predict the damage growth behavior (and residual part strength) of composites are still in early development. They are hampered by the fact that there are no universally accepted theories relating to the damage growth and failure of compotite materials. While erack growth in metals is fairly well understood, the rules of delamination growth in composites still elude both industry and academia. ‘Thus, any analyti- cal results must be heavily supplemented with rigorous (and often expensive) testing programs that apply to the specific part design and loads of interest. Design ‘A cautious and conservative approach has been used in introducing composites into erii- cal structures. Because of the analytical inadequacies, wide scatter in batch-to-batch prepreg properties, and the variability between menufacturing processes, designers have imposed can- servative design strain values on the materials. Stress analysts impose high “factors of safety” to the stresses they predict for the part. ‘These conservative practices add weight to the part. With greater expertise, development of new analytical tools, and production of tougher materi- Als, confidence will increase and these weight penalties will be reduced. ‘The premise behind unitized design philosophy is that the most benefits (structural weight and cost savings, performance, range, payload, etc.) wll accrue when a weapon system is initially designed with composites in mind.’ In this way, there is maximum compatibility ‘among all parts, and maximum use of automated manufacturing processes is possible. A basic tenet of this philosophy is that composite designs will rarely be cost, competitive with metals unless the part and fastener count (and therefore tool count) can be drastically reduced. Unitized designs will tend to emphasize such techniques as cocuring and integral stif- fening. ‘As designs become more unitized, tradeofis begin to appear, On the one hand, repair and inspection procedures become more complicated; certain areas of the system will be more diff- cult to access, larger pieces will have to be handled and inspected, the expense of a mishandled autoclave run increases, etc. On the other hand, certain elements of risk have been reduced; if fewer holes are drilled for fasteners, fewer cracks will be initiated and cyclically grown. ‘There will also be fewer tolerances accounted for. "A wie ein fr the piosophy of desing and bung structure in one pic, hey ning the need for fasteners. ‘Traditionally thew arucres were manufactured in seperate pieces and then mechanically ‘aligned, shinmed, and fastened. The unitization of the place then cen save much ofthe alignment and assembly work ‘and makes the lace las Ikely fo contain undetected flaws associated with driling holes, etetis difficult to assess where the limit of unitized designe will occur. One limit may sim- ly be autoclave size; 8 ft (diameter) by 20 ft (length) autoclaves are common (although larger ‘ones exist). Another unknown is the effect of thermoplastics. If they live up to their potential, repairs should be much easier (and quicker) to perform in the field than for their thermoset ‘counterparts; this trend would encourage the production of larger, more integral parts. COMPOSITE FABRICATION TECHNIQUES * Fabrication costs are driven by the design requirements of the structure. ‘These require- ments determine a set of candidate materials that determine a set of fabrication processes. Generally speaking, superior performance requires tore costly materials and fabrication tech- niques, Primary structures, such as wing or stabilizer skins, are highly loaded, flight-critical structures that must be resistant to fatigue and environmental effects. Secondary structures do not cerry critical loads, and the consequences of failure are less dramatic. ‘Therefore, primary structures demand high performance materials and are expensive to make while secondary and ‘more moderately loaded parts may be able to take advantage of lesser materials and cheaper, quicker manufacturing processes. ‘Table 5 provides an overview of composite manufacturing process temperature and pres- sure control requirements. Also included are rough estimates of tooling, production, and material costs. Each process is applicable to a limited number of materials, and the require- ments of each material vary in terms of temperature and pressure controls. Part shape and size put additional constraints on the applicable fabrication technique employed, Table 6 presents information on the suitability of certain manufacturing techniques to various component forms. For example, medium to large planar panels such as wing and stabilizer skins, fuselage skins, and doors, because of their form can be fabricated using auto- lave curing, elastic reservoir molding, thermoforming (thermoplastics), hot stamping, ot rapid cure (thermoset). Primary structures meet strenuous stiffness and strength needs by controlling the orien- tation of the fibers with respect to the design loads. "This requirement: for carefully controlled fer alignment is perhaps the most important cost driver with respect to composite parts, has slowed the adoption of automation, and has forced the use of expensive manual labor for many critical parts, ‘Automated fabrication procedures have been extensively studied and developed in the last ten years. Some contractors have heen, or are, in the process of installing expensive aystems that they hope will beneficially affect their production rates and costs. ‘These systems are starting to come on line now, but it is difficult to predict what problems will arise and how ‘truly cost-effective such systems will be. “Implementing automation in this [sirframe] industry has not been easy given the limited production of aircraft, the high cost of automation equipment, and the complexity of typical aircraft composite parts." Complex operations requiring great dexterity will probably remain manual procedures. However, some operations, such as ply cutting and layup ((or certain Darts) will very likely be automated in the near future and the cost/time savings should be sub- stantial "Wace, 1987, p88.sermaojomeg,Table 6 ‘SUITABILITY OP MANUFACTURING PROCESSES TO ALTERNATIVE MANUFACTURING FORMS: (Bramples listed ae intended tobe illustrative) Form of Manufactured Component Large Highly Med/Large Integral, Contouped Plain.” Closed, Open Detailed Process Structure’ Parts” Panels? Sections Sections® —Partal Le Lea curing Blartic No No Yeo NoYes Pomible molding (ERM) ‘Thermo: No No Yeo No Yes No {forming thermo: plastics Injection No No No Ye Yen Yes molding Hot No No Ya No Yen Simple stamping ‘Brackets Rapid cure No No Yee Yes Yes Simple thermoset Brackets Pultrusion No No No No Ye No Filament == Yes Yew Roe ie ne Ne sinding ‘SOURCE: Mahon, personal communication, July 1889, ‘nce fuselage skin with stiffeners, wing skins with stitfener, and bulkheads. includes lending edges and feirings. “Includes wing skins, stable skins, fuselage skin, and doors inches cloed hat setion sifenes, duvets, and piping. ‘cludes stiffeners (*L? shaped end 2" shaped), beams ib, and frames, ‘Tnclads ftings and brackets, ‘The Conventional Fabrication Process for Laminates A simple and generic description of the fabrication steps for composite parts follows: ‘Tool and material preparation, Pattern (ply) cutis Material transfer and orientation (layup). Debulk. Precure and vacuum bag preparations. Cure and posteure.‘Tool and Material Preparation. Most tools used to produce composite parts are metallic. Long-term production tools are usually made of steel or nickel, but aluminum and ‘occasionally plastic tools have been successful as well. Plastic, composite, rubber, or aluminum ‘tools are often not durable enough to survive long-term production runs and mishandling. However, there are circumstances where composite tools may be necessary. Toeling costs for composite materials can be high partly because of the poor machinsbil- ity characteristics of composites. A metal part, by comparison, does not require as strict tool tolerances because machining away excess material is’ easier. Duplication of expensive tools because of long autoclave cure cycles is often necessary. Any thermoplastic parts that mus: be processed in an autoclave will incur high tooling costs because of the high temperatures (600°) the tools and the autoclave must endure. Ancther tooling cost driver is the mismatch between the thermal expansion of the com- posite part and that of the metal tool during the cure cycle. Steel, and sometimes aluminam, tools are often necessary because of strict processing temperature and tool durability require- ments. However, metal tools will tend to expand more than the composite part. As long as processing temperature and part tolerance requirements are not too high, the warping effect may not be critical and acceptable parts may be produced from metal tool. ‘The trend in composites is indeod toward materials with higher processing (and service) temperatares. Bismaleimides, polyimides, and thermoplastics all have higher manufacturing ‘temperature requirements than conventional epoxies. Therefore, the thermal mismetch between composite parts and metal tools becomes more important; excessive mismatch can damage both the tool and the part. ‘Thus, different (and probably more expensive) materala must be used for tool fabrication. However, to some extent, this problem can be alleviated by ‘more sophisticated tool designs and analytical models. These approaches can help minimize the deformation mismatch before tool product ‘Another alternative for parts that require high processing temperatures is tools fabriceted from composites. This concept has the advantage of « close thermal match between the tool and the part, which may be necessary for complex shaped structures. During the cure cycle, such took expand and contract in better harmony with the part since they are made of the same muterial® Unfortunately, composite tools suffer from a number of deficiencies: Their durability is in question, they are slow to heat up, and they are very expensive (one contractor estimated some of his composite tools to be 20 times more expensive than metal counterparts). One reason for this expense is that even more tools are required to make the composite tools. ‘Thermoplastic materials will further tax conventional metallic tool design; thermal mismatch effects would be even greater because of the higher processing temperatures required for thermoplastics. Conventional aluminum tools will probably not be possible since tempera ‘ures will exceed 600°F. The durability of steel tools would also be affected by the substan- tially higher processing temperatures. ‘Thus, research into composite tools is a necessary and expensive effort Long layup times can also be a tooling problem. Ideally, expensive production tools should be “cooking” nearly all the time; but delays can accrue if « large, complex part requires, perhaps # week to lay up. To minimize the effects of long layup times, some companies have several cheap plastic male tools for each female production tool. Several parts can be laid up on the male tools simultaneously. ‘Then these parts can be easily transferred, one at a time, to the femele tool for curing. In that way, the cumulative production time (over several units), or the number of expensive autoclave-capable tools, can be greatly reduced. “Another advantage to block graphite tools i thet they can be machined using numerically controlled machine tools jut an conventional machined metal parti formed.Pattern Cutting. Manual pattern (ply) cutting is a slow process, about 100 inches per ‘minute for a trained composite technician. This technique is most often used for small and medium sized parts with complex contours. Such parts are not yet amenable to automation. ‘Manual pattern cutting makes it difficult to cut multiple plies with accuracy, the process is slow and labor-intensive, and costly inspection procedures must be implemented. Many con- tractors have already purchased and are using automated pattern cutting facilities wherever possible. Automated pattern cutting procedures include Gerber knife machines, waterjets, lasers, ‘and chisel cutters. Cutting speeds are much higher than for manual labor: 300 inches/minute and multiple ply capability (four plies of graphite/epoxy) for Gerber machines; much higher limits are planned for the chisel cutters. All of these procedures are numerically controlled and highly accurate even for large plies, which are difficult to handle using manual techniques. Automated ply cutting should eliminate some of the costly inspections required for manual pro- ‘cedures. These systems should be able to handle thermoplastics and thermosets in exactly the same way. Automated ply cutting facilities are already common among the major composite Material Transfer and Orientation (Layup). ‘The layup process, by one estimate, representa 30 percent of all composite labor costs. As mentioned earlier, much of this is driven by requirements for fiber alignment. It is in the layup process that automation may have ite ‘most influence with respect to cost savings. ‘Hand layup is estimated to account for over 90 percent of all parts in the aircraft indus- ‘try. Hand or manual layup is often necessary because complex shapes are not amenable to automated tape layers. This means that many parts have contours that exceed the natural compliance of the prepreg tape. Hand layup procedures can be lengthy, error-prone (despite the attention of the technician), and costly. For example, the AV-8B wing skin, which is com- posed of about 200 individual plies spanning an area of nearly 300 square feet (10 x 30), requires about 160 Iabor-hours of hand layup. Research estimates that humans are likely to ‘make mistakes in the layup process if there are more than 20 separate plies to form. Thus, time and money must be spent monitoring hand layup parts and correcting (if possible) any mistakes. Automated layup can often be performed by tape laying machines for larger, mildly con- toured parts such as wing and stabilizer skins. For such parts, automated tape laying machines can operate many times faster than manual procedures. Another advantage is that plies laid up by machines often do not need separate debulking procedures because of the high compac- tion capabilities of the machines. Many parts have contours and ply designs that are too com- plex to be readily handled by these machines. However, there are new programming tools that will enhance compatibility between the parts and the tape laying machines, ‘Many new designs are benefiting from “natural path” programming. ‘These software rou- tines graphically display the natural paths that the plies will want to follow given the part geometry, the type, width, and thickness of the prepreg material. Designs can thea be evaluated and, if necessary, changed, to more fully comply with the use of automated tape layers. ‘Thermoplastics (and advanced thermosets such as some bismaleimides) have poorer drap- ability and tack than epoxies. Therefore, some additional heat during layup is required to enhance formability. In manual layup procedures, bismaleimides can be softened with heat ‘guns (90-190°F), However, thermoplastics require as much as 650°F for forming. Automated ‘tape laying machines for these materials use warm (bismaleimides) and hot (thermoplastics)a dispensing heads to lay up the plies. Such machines are new for thermoplastics and currently confined to producing flat laminates. This limitation is somewhat mitigated by the fact that a flat thermoplastic laminate (produced by an automated tape layer) can be reheated and recon- solidated against a mold of the required curvature more swiftly and accurately than if the entire layup process were done by hand. State of the art, numerically controlled tape laying ‘machines are estimated to cost between $2,000,000 and $3,000,000. A hot head tape layer, required for thermoplastics, would be at the higher end of the range, perhaps more. ‘Ply Management ia an issue of growing importance as automation and production rates increase. ‘Thousands of plies will have to be managed on a daily besis. Industry is investing in ‘automated, computer-managed systems that must cope with many issues: coding of plies as to ‘material type and fiber orientation, which part the plies belong to, sequence in the layup, ‘lapsed time from removal from the freezer (thermosets), kitting (bagging plies for freezer storage), ply retrieval, and preparation of layup books. ‘The Automated Ply Laminating System (APLS), being developed at McDonnell Douglas in St. Louis, will produce flat ply collations, which con then be manually formed about the cur- ing tool. Plies will be cut, labeled, kitted, stored, retrieved, and collated with minimal human assistance. Debulking. Debulk operations may be required before cure to compact the plies and to liminate any interply gaps or voids. Thick laminates may require on the order of 10 debulle cycles to properly seat the plies onto the tool. This time-consuming process can sometimes be climinated or reduced by the use of automated tape laying procedures that exert high pressures on the uncured part. Precure and Vacuum Bag Preparations. Matrix bleeder materials are required to censure that the proper amount of excess resin is bled out of the laminate during the cure cycle. ‘This is important because excess resin will only degrade the final part properties as well as add extra weight. Such materials es peel ply, porous teflon, nonporous teflon, and fiberglass bleeder plies are placed in intimate contact with the uncured laminate after it has been laid up fon the tool. The laminate assembly is then ready to be sealed in a vacuum beg. Bagging procedures are extremely important for autoclave cures. According to the DODINASA Structural Composites Fabrication Guide, improper bagging and sealing operations fare probebly the most prevalent cause of scrapped parts. If the vacuum seal is lost during the ccure cycle, the pressure pushing the laminate against the forming tool vanishes. Without the support of the tool, the plies will shift, and part tolerances (and contours) will not be met. ‘Also, any volatile gases generated during cure will not be drawn out of the laminate, which increases the possibility of undesirable voids and delaminations in the cured part. Therefore, ‘careful bagging and sealing procedures must be maintained and checked. Cure/Posteure/Consolidation. A cure cycle is a prescribed schedule of temperature and pressure that is required to process the thermosetting matrix material of the laminate to its final hardened state. This cycle may be specified by the material vendor, or the contractor ‘may use a cycle that has been developed in-house. ‘A typical cure cycle for graphite/epoxy lasts about 3-4 hours, reaches pressures of 50-100 psi and temperatures of 350°F, and requires a postcure of several hours. Graphite/polyimide laminates require a maximum pressure of about 200 psi, a maximum temperature around 650°F, and a postcure of nearly 16 hours. After curing, the laminate is removed from the tool and stripped of the matrix bleeder materials. Thermoset parts are often postcured (tempera- ‘ture only) in ovens. ‘The purpose of the postcure is to achieve maximum crosslinking and strength in the thermosetting matrix.a Past programs simply recorded the temperature and pressure levels as a function of time, within (for example) the autoclave to show compliance with the recommended cure cycle. Future emphasis will be placed on monitoring the key physical and chemical parameters of the ‘actual matrix material. This should help to eliminate some of the wide scatter in the mechani- cal properties of composites since it is the state of the matrix that needs to be properly moni- tored, not the pressure and temperature levels in the autoclave, Ultimately, if the sestter is reduced, confidence will be increased and design penalties for composites reduced. ‘Thermoplastic materials are said to consolidate rather than to cure. ‘This process is much ‘quicker (minutes), but temperatures can reach 800°F, depending upon the material. Also, ther- ‘moplastics do not require a postcure cycle. These materials will require more capable and dur- able manufacturing equipment (autoclaves, ovens, presses, ete.) than those currently used for epoxies. Autoclave curing is the most common, state-of-the-art method used for producing high quality composite parts with thermosetting resin matrices. Autoclaves are expensive; a single autoclave 20 feet (diameter) by 50 feet (length) can cost $7,000,000 (for use with thermosets) or $11,000,000 (for use with thermoplastics). (As autoclaves get larger, they cost more.) Assembly. Mechanical assembly procedures include drilling holes, trimming, cutting, ‘sanding, bonding, cocuring (practically speaking, this is part of the cure cycle), fitup, mechani- cally fastening, ete. ‘Some of these operations are more difficult (and more costly) with composites because of their brittle nature and historically low matrix toughness (of thermosets). This makes compos- ite parts very sensitive to low-energy impacts (such as the accidental dropping of a tool). Although there may be no visible damage on the surface, delamination and matrix cracking ‘may exist in the interior of the laminate. Unseen internal damage can have an enormous negative effect on compressive strengths. Composites are also sensitive to the drilling of high tolerance holes; matrix cracks and fiber damage can be induced around the edges of holes. Current research emphasizes the toughening of existing thermosetting matrices, the exploita- tion of the inherently tougher thermoplastics, and the development of appropriate thermoset/thermoplastic hybrids. These efforts should yield beneficial results with respect to decreased levels of inspection, increased tolerance to low-energy impacts, and reduced mainte- nance requirements. Assembly can be a considerable portion of the total part cost; estimates range as high as 40 percent, One reference concluded that drilling and countersinking holes between szaphite/epoxy parts can be four times (per hole) as expensive as between aluminum parts. Fastener costs and installation can be between one and five times as expensive as for alumi- num perts. The expense of joining graphite/epoxy to aluminum structures is largely driven by the galvanic reaction that occurs when aluminum and graphite are in intimate contact, Corro- sion is the consequence of any such contact, and thus more expensive titanium fasteners are used instead of aluminum. Other concepts for avoiding this galvanic link include the layup of 4 single fiberglass ply, which acts as insulator between the incompatible materials. After all of the machining is done, the separate parts are assembled in the fitup pro- cedure. Fitup is a good test of the fabrication system since it must account for and integrate all of the accumulated tolerances, mistakes, and defects. Fitup and assembly costs can be as much as 15 percent of total manufacturing cost, Cecuring is an operation in which multiple parts may be both cured and adhesively bonded to one another during a single, often autoclave, run,6 Composites compete more effectively, in an economic sense, with metals if designers can reduce the number of parts in an assembly. This “integral structure” approach reduces not only the otal part count but also the number of holes that must be drilled and the number of fasteners that must be installed, expensive operations for composites. Generally weight is also reduced. Detailed parts can be cocured, but major assemblies are mechanically fastened. Many subassemblies that are currently secondarily bonded will instead be cocured by the mid-1990e. Examples of cocuring indicate that manufacturing costs can be reduced by as much 8 40 per- cont. Secondary structures, where the structural consequences of failure are fairly low, are rea- sonable candidates for cocuring. Heavily loaded primary structures will probably not be cocured because of inspection requirements. In primary structures, it is usual and prudent to nondestrictivély inspect all parts in detail before assembly. This is not possible with current ‘cocuring techniques. ‘Thus, primary structures are likely to use secondary bonding operations ‘or mechanical fastening methods. A recent example of cocuring is the AV-8B fuel tanks. ‘The trend toward cocuring is motivated by the desire to reduce the number of fasteners and assembly operations. However, some do see risks associated with this strategy: “Large multi-component assemblies will increase the tool size and the number of details. Bonding fix- tures with 1,000 separate details will generate an accountability, storage, and replacement prob- lem. More integral structures will affect supportability as well. Reliable, on-aircraft repir techniques will be required since these larger, unitized parts will not be as easy to remove, repair, or replace. Advantages of Cocuring Lower part count. ‘Smaller inventory. Fewer holes and fasteners. Lower manufacturing costs. Disadvantages of Cocuring Reduced access for inspection. Increased complexity for tool design and integration. Increased risk if autoclave performs improperly. Increased expectations on field-level, on-aircraft repair skills. Inspection. According to the DOD/NASA Fabrication Guide (Meade, 1982), “Many defect types can affect the quality of a composite structure and no single nondestructive test can find and isolate all of them.” Common defects include delaminations, foreign matter (inclusions), high porosity (too little matrix material), honeycomb core damage, moisture, fiber breaks, and matrix cracks. These flaws may originate during the manufacturing of the part, a8 f result of accidental mishandling during normal airframe use and maintenance, or as a conse- quence of battle damage. ‘There are many methods of nondestructive inspection, but each has its limitations. Table 17 describes the sensitivity of several NDI methods to different flaw types. An entry of “VG” indicates good sensitivity and reliability between the particular flaw type and NDI method, An entry of “G” indicates less reliability or limited applicability, and an entry of “L” indicates “Spinks, 1988, p. 425.Table 7 SENSITIVITY OF NDI METHODS TO DIFFERENT FLAW TYPES, eee eer NDI Method Xoray Neutron ‘Thermal Ueasonic | Radiog: Radiog- Infre- Tep Bay Flaw Type Tranamission* raphy raphy Laser red Test Acoustic” Current Visual nr eg eee Porosity = = SS Fereign material §—-VG/G ee Dalarinaton vo Matrix racks i OS 8 2 ee = Fiber breaks = Oe ce pact damage on. - = G 6 @ L = 4 Skin/skindiabond —-VG/G ve) ove Skin/core dsbond —VG/G LoL wow ¢ @ © = Core damage VG VG ee ee = CeevOme te ‘Water intrasion L aE Ee eee ee SOURCE: Meads, 1988, “Urrasonic trasmision includes four types of testa: through tranamsision, pulse echo, angle, and reso "Acoustic inctdes two types of txts: emission and ultrasonic. not applicable VG: Good sensitivity and reliability: good candidate for primary method. Lees reliability or Knited eppicbility; may be good beckup method 1: Limited applicability bay provide some usefl information. Aimited applicability. Every flaw must be well characterized so that engineering decisions with respect to repair or part replacement can be made. ‘The inspection of composite structures is complicated in part because composites have so many different flaw types (compared with metals, which are mainly inspected only for cracks or corrosion). Several of these flaw categories are identified in Table 7. Naturally the seriousness of the flaw depends on its size; however, the more serious flaw categories are delaminations, fiber breaks, porosity, and crushed core (included under core damage). Additionally, each inspection technique (also shown in. Table 7) is not sensitive, or has limited sensitivity, to each category of flaw. ‘Thus, several techniques must be used on individual parts depending on their specific design and criticality. ‘Levels of inspection are determined by such factors as part type (primary, secondary, nonstrue- tural); stress analysis; part complexity; and part history. As the materials, manufacturing Processes, and analytical tools develop, the number of costly inspection procedures should be reduced. ‘Ustrasonic transmission methods are the most commonly used method with epoxy lam- inates. A quick scan of Table 7 shows that nine out of the ten flaw types listed are reasonably sensitive to at least one of the ultrasonic procedures Currently, there is an evolving interest in thermal IR procedures displaying varying part quality in terms of colors. Magnetic resonance imaging (MRD, which is finding increasing use in medical fields, is also likely to be applied to inspection of aircraft parts. Many contractors now have automated inspection stations.Specialized Fabrication Processes ‘The traditional and most common method of composite fabrication in the aircraft indus- ‘try encompasses manual layup in combination with autoclave cure. However, there are special- ied layup, cure processes (thermosets), and consolidation techniques (thermoplastics) that are ‘used when possible. Below we describe the following techniques: Filament winding (layup). Braiding (layup). ‘Vacuum bag curing. ‘Thermal expansion curing. Hot roll forming. Injection molding. Compression molding. Hot stamping. ‘Thermoforming. Pultrusion. Hydroforming. Elastic reservoir molding. Filament winding has primarily been used for large, axisymmetric structures such as solid rocket, motor cases, pressure vessels, and helicopter blades. However, other shapes, such as ‘beams and channel elements, are now being considered; filament wound fuselages might also become practical. During filament winding, individual fibers or narrow prepreg tapes are dispensed from a translating head to a rotating mandrel. Fibers are pulled through a resin bath just before con- tact with the tool. Angles of the fiber relative to the rotation axis of the tool are controlled. Both the density and mechanical properties of the part can be affected by a predetermined ten- sion in the fibers or tape. ‘This process has been performed with thermoplastic materials, but the high temperature processing requirements pose some manufacturing problems. Parts formed in this manner can either be autoclave cured or room temperature cured. ‘Braiding also hes been used to produce long and continuous lengths of simple (can be tapered) cross sections. Preforms of the reinforcing fibers are produced and then impregnated with the matrix material. Multidimensional braiding through the thickness direction confers greater damage tolerance characteristics to the part. This technology is still labor-intensive. ‘Vacuum bag curing is very similar to autoclave curing. ‘The maximum external pressure is limited to 14.7 psi (atmospheric). Heat is usually supplied by an oven. Vacuum bag curing produces parts of lesser density (and cost) than those cured in autoclaves. ‘Thermal expansion curing is another fairly cheap method of creating the necessary curing pressures by operating on the principle of differential thermal expansion. Uncured plies are ‘wrapped around rubber molds and then placed inside a metal cavity. As the oven temperature increases, the rubber will try to expand more than the metal cavity. This restrained thermal growth creates a pressure that is transferred from the rubber to the metal cavity through the Taminate. This process is cheap, but care must be taken in the design of the fixtures #0 that the proper pressures are created. Hot rolling of thermoplastics is related to roll forming in metals. Roll forming runs flat ‘metal sheet through a series of rolling dies that gradually shape the metal into the desired cross-section. This process is intended for the production of long and continuous members of« constant cross-sectional shape. Thermoplastic components can be produced by heating the dies sufficiently to soften the resin and form the part. Injection molding produces parts by injecting a measured quantity of resin and chopped fiber into a molding die cavity that defines the shape of the part. Consolidation occurs under the application of heat and pressure. This process is adaptable to both thermosets and ther- moplastics. Molds are expensive, so parts that are required in fairly high numbers would take the greatest advantage of this procedure. Current parts can be complex but generally small. Research with respect to larger parts is ongoing. One problem is the prediction and control of short fiber orientations; lack of control can lead to unequal distributions of mechanical proper- ties, Current cycle times are generally less than five minutes. Compression molding has been used with short fibers and thermosetting matrices. It is being extended to include thermoplastics. Parts are usually simpler in shape and larger in size than those produced by injection molding. ‘The raw materials are placed in a heated mold cav- ity and then squeezed in a hydraulic press. Cycle times are in hours, as opposed to injection molding cycle times of minutes, ‘Hot stamping is related to stamping of parts in metal sheets. To adapt this process to thermoplastic materials, heat (up to 700°F for PEEK) must be supplied. Cycle times are gen- erally less than five minutes. Stamping pressures might reach 5,000 psi for some parts. This ‘process is suitable for chopped fiber reinforcements, Thermoforming is related to hot stamping but is used when the fibers are required to ‘maintain specific orientations. Again, the basic process is that the laminate is preheated and transferred to a mold in a press where the final part shape is determined. ‘Typical presoures range from 15-100 psi. Pudtrusion is useful for producing lengths of members with constant cross-sections. It ‘uses guides, shape dies, and heat to produce long parts with constant cross-sections, which are then automatically cut to desired lengths. ‘The process consists of pulling dry fibers through a resin bath and then through steel dies that define the shape of the part and control the ammount, of resin in it. Heat must be supplied to cure the part. This process is economical but is ccurrentiy limited to straight pieces. With more development, components with varying cross- sections should be achievable. ‘Thermoset parts can be reasonably produced using pultrusion, and it is readily adaptable to thermoplastic materials. However, experience with pultrusion for thermoplastics is still limited. ‘Hydroforming fiber-reinforced thermoplastics is also based upon experience with metals. The procedure uses a hydraulic press fitted with a flexible pressurizing medium on the upper platen and a rigid form block on the lower platen. ‘The composite part is preheated to the Drone ming tempaatre (perhaps in wn oven) and is traafred tothe form Mock The ic closed and the laminate is formed into the desired shape. While the pressure is toplid the rann ool below the forming temperature and thus wil ain the proper ape when removed. Press time should be on the order of half an hour or less. Elastic reservoir molding (ERM) produces sandwich components consisting of rigid, polyurethane, foam core with reinforcing facesheets. Graphite, Kevlar, and glass fibers have ‘often been used in combination. ‘The foam is preimpregnated with a known amount of resin, which is transferred to the facesheets under pressure. The resins used are thermoset epoxies or polyesters. This process can use existing hydraulic presses. Rigid molding dies are attached to the upper and lower platens of the press. The foam core (including resin) and the reinfore- ing facesheets are placed between the dies. The molding dies are closed by platen pressure and ‘the resin is squeezed through the fabric facesheets. Excess resin and gases are ventilated from« the dies. The application of molding pressure and heat cures the part. ERM can also be used with prepreg tapes or fabrics. Automation of the Manufacturing Process ‘This section discusses where, and to what extent, factory automation will penetrate compos- ite part fabrication in the airframe industry. As little as ten years ago, all aspects of the production of composite parts were dominated by manual procedures.” Since that time automation has made, and should continue to make, great progress in some areas of the manufacturing process (eg., material cutting, ply management, and nondestructive testing). However, progress is likely to remain limited in other production activities, particularly those confined to operations on small or ‘complex parts (as is often the case in layup procedures). Advances in automation seem to be largely driven by individual computer systems so that information can be quickly and accurately disseminated. Our research found that most manufacturers believe automated procedures will play an ever increasing role in their ability to competitively produce advanced systems. However, ‘there is a range of opinions and expectations as to the extent these automated procedures will pre- vail in the 1996 timeframe. Using the cost information reported in Tables 14 through 18, the cost breakdown for produc- ing a composite part is approximately 10 percent for engineering, 10 percent for tooling, 60-65 per- cent for manufacturing, 10 percent for quality assurance, and 10 percent for material costs. Although the other cost elements are important, manufacturing offers the most leverage for auto- ‘mation (assuming automation is equally feasible across cost elements). Consequently, much of this discussion (and the discussion in the open literature) centers on automating manufacturing ‘The steps in the fabrication process, as listed in the previous section, are cutting, material transfer and orientation (layup), debulk, procure and vacuum bag operations, cure, and postcure, Material transfer and orientation by far is the dominant contributor to fabrication cost. This can be seen in Table 8, which lists the distribution of the fabrication effort for par- ticular components. ‘Automated Layup. The accuracy of fiber orientation, which may change between plies or series of plies, is crucial to part performance.” Small, highly contoured parts may always ‘Table 8 [BREAKDOWN OF FABRICATION EFFORT (Perea of fabrication hours) Fighter Stabilizer Skint Wing Skin? Gating 7 a ‘Material tranfer and orientation (layup) o 1 Debalke 1 7 Procure and vacuum bag operations ” 3 Gare and posteure M 10 “endo, 1982 Prato, 1985. Vacar, 1887, p. 88.«6 equite the dexterity of the human hand. Compound contours also add complexity, and automated tape layers are restricted to large, slightly contoured components using unidirec- tions] material. Automated layup of, for example, wing skins is generally limited to unidiree- tionel tape widths of less than a foot (often only three inches); fabrics are not currently used in thie process. Also, the automated layup procedure for such a part is really combining the separate steps of ply cutting and layup; the machine, in essence, makes its own plies as it moves about the tool, automatically unwinding and cutting the roll of tape according to its preprogrammed instructions. According to a McDonnell Douglas study, automated tape layers are precluded from ‘pplication to sections smaller than 4 by 4 feet.© Several studies have found that setap time and diseconomies with having to cut the tape and change direction are major reasons why tape layers are not cost-effective on both small parts and parts with irregular shapes that require the tape layer to lay short courses. “The principal factor limiting more widespread use of tape layers, however, is that most composite components have greater contour than the compliance of the tape, which ideally should butt edge-to-edge across a mold surface.”” For whatever rea- son, we observed that automated tape layers were in limited use during our plant tours. Our observations are confirmed by an article in American Machinist and Automated Manufacturing, which stated, “Despite the availability of tape layers, however, manual layup prevails in the sirereft industry. .. . About 95 percent of all layup jobs ... are performed by workers by hand oF using rollers and other aids, mainly because the complex contours of most parts are not amenable to tape layers, And it’s a long and tedious task.”® Future applications in this area may include the use of machine vision systems. ‘They will ke able to determine if plies have been oriented correctly and if any unwanted foreign ‘matter has attached itself to the pos, For smaller parts (Iess than 4 x 4 feet), or those with compound curvatures, flat ply colla- tion combined with hand layup may be an alternative to automated tape laying.? And in fact, MeDennell Douglas Aircraft found that on their AV-8B program, flat ply collation combined with hand layup was more cost-effective than an automated tape layer. With flat ply collation either tape or fabric may be used. The limitation on flat ply collation is a radius of curvature no smaller than 10 inches. Pultrusion and filament winding are other promising methods of automating the layup process. Both processes “create” their own plies by directly orienting and placing individual fibers. Pultrusion is limited to those components with constant cross-section, and filament winding is limited to axisymmetric sections, Thermoplastics may be more promising for pul- ‘trading than thermosets because once pultruded the thermoplastic piece may be heated and reformed into more complex shapes by twisting, bending, or flanging the component." Fil ‘ment winding, while cheaper than hand layup, is not cost-effective for smaller quantities where setup costs becomes a driver." If layup is automated, debulking is not necessary since pressure is applied during the layup process. Automated Ply Cutting and Ply Management Systems. The cutting of prepreg fabrice and tapes into the constituent plies of a component has, for the most part, been i, 1887, p88, Sid, p88 Huon, 1986. Kin, 1887, pp. 94-95. "Bid, p. 97,automated. In fact, every airframe manufacturer’s plant we visited had one or more Gerber cutters. ‘This operation lends itself more readily to automation because every cut can be per- formed on a flat surface. Therefore, no consideration has to be made for curvatures and con- tours even though the plies may ultimately be laid up on very complex tools. The advantages are further enhanced if the individual plies tend to be large in area since manual operations can sometimes be limited by the reach, strength, and accuracy of human hands, depending upon ply design and fabric/tape width. Also if the plies are large, the automated machines spend less time stopping, starting, cutting, and changing directions. Ultimately, such two- dimensional operations as ply cutting are often considerably easier to automate than such ‘three-dimensional operations as layup. Ply management systems “instruct” the cutters on the marking and nesting of the material and is another area where efforts to automate are making progress. Once cut, plies are marked with information such as the material type, fiber orientation, sequence in layup, ‘and out time. The plies are then collated and grouped into “trays” ready for layup onto the tool. Automated ply management systems are currently being implemented on the factory floor. Other Areas. Net part operations, trimming, drilling, etc., are also making progress. Fully automated ultrasonic inspection procedures already exist. X-ray, acoustic emission, and nuclear magnetic resonance imaging are also being examined with repect to their potential for futomaticn Development of such postcure operations must consider the inherent damage tolerance of cured composite parts. As mentioned earlier, composites tend to be brittle and have low interlaminer properties. ‘Therefore, they cannot tolerate mishandling as well as metallic parts during trim and drill operations. Potential Savings. Meade, 1982 lustrates the potential savings from automation, par- ticulariy in layup, and the associated improved material properties. Tatimates of fabrication time and cost were made forthe composite stabilizer skins ofa fighter aircraft using two pro- cedures. ‘The bateline case assumes predominantly manual operations, and the second cave includes the effects of automation and advances in the raw material characteristics. Specifi- cally, the second fabrication procedure included a robotic material transfer and orientation (layup) system, reusable vacuum bags, a low resin content prepreg material, and an automated tape dispensing system. ‘Comparisons were made between the largely manual procedures of the baseline case and ‘the automated techniques used in the second case. The modifications to the operations are listed in Table 9 and the resultant cost reductions are illustrated in Fig. 14. The total fabrica- tion hours required per shipeet decreased from 25 to 10 hours. The larget single savings in this pertizular case was an 80 percent reduction (time) in the material transfer and orientation (layup) process. Savings were also achieved in debulking, pattern cutting, and bagging. The analysis also showed a 45 percent decrease in part cost (assuming 900 unite) from the baseline case to the second ease. And, although labor accounted for 45 percent of the cost in the base- line case, it accounted for only 17 percent in the second case. As the fabrication processes become more automated, the raw materials become the main cost driver. ‘Tool and material preparation times as well as curing and net part operations times were not affected in this case. However, these parameters would also have been affected if a thermoplastic material, instead of an improved (low resin content) thermosetting material, had been used inTable 9 ‘MANUFACTURING COMPOSITE STABILIZER SKINS, [BASELINE AND IMPROVED OPERATIONS. Operation Baseline improved ‘Material 0" wide 0" wide ‘Tape dispense Manoa! Automated Py cutting Gerber knife Gerber knife Layup ‘Manual Robot Number of debuiks 13 None Bleeder plies With Without Bag material Dispomble Reusable Postcure With Without® SOURCE: Meade, 1062. “Relaxation of production requirement. + Hours reduced by 70% + Layup improvement is 50% of total hours reduction Ta Total hours: 35.0 Total hours: 10.0 ‘SOURCE: Meade, 1982. Fig. 14—EBffect of improved operations on manufacturing hours of composite stabilizer skinsa the second case, For example, freezers would not have been required to store the uncured prepreg material. The curing time would have dropped from hours to minutes; however, the autocleve requirements for temperature and pressure would have been more severe. A posteure would not have been required. Tool degradation would also have been aggravated by the shorter but hotier cure cycle. ‘A projection between manual and automated fabrication procedures was found for a fuselage section, This structure required the layup of inner skins, outer skins, longerons, and stiffeners. Total fabrication hours per shipset were projected to decrease from 630 to 130 hours. ‘Total layup time decreased from 360 to 110 hours. Although total fabrication hours ‘and total layup time both decreased by about 60 percent, layup costs still remain about 50 per- cent of the total using either method.!? ‘Automation Conclusions. Introduction of automation and mechanization on the fac- tory floor will be more evolutionary than revolutionary. As disussed above, certain aspects of ‘manufacturing, such as ply cutting and management, are generally more amenable to automa- tion than others, such as layup. ‘With respect to the cost of automation, we can safely say that any investment in cepital equipment will require quantities sufficient to support it. Many companies’ projections indi- cate future quantities will not warrant substantial levels of investment in automation (although ‘moderate levels are planned). This is typified in LITV's decision to postpone construction on their Integrated Composites Center. Some respondents have indicated that automation in metal manufacturing will outpace automation of composites manufacturing, at least by 1995. ‘While companies have made great strides to automate composites manufacturing, auto- mation is certainly not a given. Even though processes have been proven in development, the transition to production is not necessarily straightforward.!® To enter a process into produc tion there must be sufficient data and confidence and acceptable yields must be attained. In summary, the extent of automation on the factory floor will be defined by at least some of the following constraints: Part sizes, contours, and complexities. ‘Material celection. ‘Quantities (low in this industry). Funds available for capital investment. NEW METAL PROCESSES Superplastic Forming/Diffusion Bonding Superplastic forming and diffusion bonding (SPF and DB) are metal forming techniques. Both processes take advantage of the ability of some metals to endure elongations of eeveral hundred percent under certain conditions (microstructure and process temperature) in order to both form (SPF) and bond together (DB) complex metallic shapes. The potential payoffs are reduced part and fastener count (unitized designs), excellent bond strength, and ultimately lower prozurement and life cycle costs. ars, (986, p. 401. 127 Beech Starship fuselage (600 Ib) was originally conceived of as a single, flament-wound structure. ‘The tol. ing for this approach cov: about $6 million, Beech stated that they had difcuties achieving the dosied level of ‘chanieal property In te flament-wound composites and thatthe manufecturng cot was very high. Ultinaaly, {hey decided to hand layup the fuselage in two pieces that reattached primarily by bonded wurfaces,‘The generic process of superplastic forming consists of placing flat sheet stock (titanium, aluminum) over a die of the desired part shape. The stock is heated to the processing tempera- ‘ture (exceeding 1,500°F) and a burst of inert gas forces the flowing material into the die. ‘The ‘metal stock assumes the required part shape and is held under temperature and pressure for a short time before cooling. Diffusion bonding requires that the mating surfaces be pressed together with pressures that locally exceed the yield stress of the material.. The bonds result from the diffusion of atoms across the mating surfaces at elevated temperatures. ‘The strength of these bonds ‘approsches the strength of the parent metal. Not all metals may be diffusion bonded (alumi- ‘num cannot). Examples of diffusion bonded titanium parts include the wing carry through of the B-1B bomber and the space shuttle main engine mountings. SPF and DB processes require very similar processing environments and can be per- formed sequentially (with the same equipment); two or three parts may be first superplastically formed and then bonded together. Since the processes are so compatible, considerable reduc- tions in manufacturing time and costs are possible. SPF/DB are primarily used to manufacture titanium parts. Titanium is nearly indispens- able in airframe applications for several reasons: high temperature capability (1,000°F), low density, high strength and stiffness, high damage tolerance, and corrosion resistance. ‘Titanium is difficult to machine, but its microstructure is adaptable to SPF/DB processes, thus avoiding much expensive machining processes. Research is also being performed on SPF aluminum and aluminum-lithium. ‘SPF/DB titanium parts have demonstrated 10-50 percent weight savings, but they cannot compete with the cost of conventional aluminum parts, Raw material costs alone exceed aluminum by a factor of 8. Therefore, for the foreseeable future, titanium SPF/DB parts will be confined to those areas requiring exceptional performance. ‘There are many examples of SPF/DB demonstration components (usually related to cenginet), but only one production part was found: @ nacelle beam frame for the B-1B. With conventional techniques, this frame has eight detail parts and 96 fasteners. With the SPF/DB Processes, these numbers were reduced to one detail part and zero fasteners. Structural weight ‘was recuced by 30 percent. SPF/DB is readily adaptable for producing reinforced sheet strue- ‘tures, internally stiffened structures (two sheets), and sandwich structures (three sheets). Powder Metallurgy Powder metallurgy procedures use, as the name implies, very rapid solidification rates of powder forms. Hot isostatic pressing (HIP) is the consolidation technique of primary interest with respect to airframe parts. ‘The P/M process for such structures can be broken down, very roughly, into three operations * Powder production. + Containerization. + Hot isostatic pressing. ‘The first step is to produce “clean spherical powder free of contaminants.” Vacuum induction melting followed by inert gas atomization is a common process, The second step consists of loading the powder into metalic or ceramic molds that define the shape of the part (inctuding any desired tolerances). The smaller the tolerance envelope ia, che more efficient, or Dili, 19863 near net shape, the operation becomes, The production of these molds can be a complex matter and there is the possibility of introducing contaminants from the mold into the part. ‘The mold is then placed inside a steel container; any remaining volume in the container is filled with a granular ceramic medium that transfers external pressure to the mold and part. ‘The steel container is then welded, outgassed, and sealed (see Fig. 15). ‘The third step, which is the HIP phase, consists of the application of a specified cycle of heat and pressure. Cycle times can surpass 8 hours, maximum temperatures and pressures can exceed 2,000°F and 25,000 psi (see Fig. 16). ‘ Further development of P/M processes must occur in several areas: ‘© Better analytical modeling and quantitative descriptions of the processes are required. ‘* Sensors that provide feedback on key microstructural parameters of the part need to be developed and incorporated into manufacturing procedures. ‘+ Nonmetallic inclusions that are introduced in the manufacturing procedures can sub- stantially reduce the part's low-cycle fatigue (LCF) properties. There has been consid- erable concern about the reliability of LCF properties in P/M parts. Some speculation ‘about the cause of an F-18 crash in 1980 centers around the fatigue life of some of the P/M engine components. The true cause remains unknown. LCF properties in P/M perts can be very good when adequate attention is paid to the manufacturing secondary metal can with secondary pressing media and evacuate Finished part Hot isostatic press ‘SOURCE: £1. Duk, “Near Net Shape Procass Using HIP ot Aloy Powder Paris,” Not Shape “Technology m Aeronautical Stuctre, Vo Il, Natonal Academy Press, 1986. Fig. 15—The ceramic mold process for producing complex shapes from powderPressure (ksi) 2200 1800 1400 1000 (Jo) esrmesedue, 600 200 ime (hours) SOURCE: £4. Du, "Near Net Shape Process Using HIP of Aloy Powdor Parts,” Not Shapo Technology lt Aotonautal Stuctres, VoL. Ii, Nalonal Academy Pross, 1906, Fig. 16—A typical superalloy hot isostatic process cycle procedures, but verification tests should be performed. This problem may diminish as nonceramic processes evolve. '* According to one paper, considerable emphasis has been placed on developing the strength and environmental resistance of the alloys in P/M parts while the metalwork- ing needs of the fabricator have been ignored* ‘The restrictions on metalworking ‘temperatures may inhibit the ability of the alloys to undergo superplastic deformations and thereby negate the possibilities of many efficient net shape procedures. MDs, 1986,IV. COST INFORMATION Costs by material type are presented and discussed in this section. These results were derived from worksheets distributed to the airframe industry. After the initial reporting phase of the project, participants were asked to verify that their response was consistent with our for- ‘mat and to respond to a series of questions clarifying the results. ‘This additional information ‘was incorporated into the report. INDUSTRY WORKSHEETS We considered an industry survey to be the best approach to collect cost information because (1) the number of historical data points is limited, (2) we were interested in materials not yet incorporated into production aircraft, and (3) we wished to capture the increased level of usage in aircraft. Worksheets were designed to provide information that was material sensi- tive and could be used to build up to an estimate of total airframe structure. A list of the par- ticipating companies is provided in Table 10. Information was requested for the materials listed in Table 11 for late-1980 and mid-1990 time periods. In general, we found that companies were willing to provide cost information for materials with which they had actual production experience. They were reluctant, however, to provide projections for materials still in the developmental stage because of the sensitivity and ‘uncertainty of the data. Consequently, the list of materials for which we actually obtained suf- ficient data is considerably smaller: Conventional aluminum Aluminum-lithium Conventional titanium Conventional steel Graphite/epoxy Graphite /bismaleimide Graphite/thermoplastic Nevertheless, despite its reduced scope, the list still encompasses the composite materials currently considered most important to the next generation of aircraft: graphite/epoxy, greph- ite/ bismaleimide, and graphite/thermoplastic. Table 10 ‘STUDY PARTICIPANTS ‘Bosing Miltary Aifplane Company TUITV Aerospace and Defense Aireraft Group General Dynamics Corporation ‘McDonnell Dougis Corporation Grumman Avrospace Corporation [Northrop Aieraft Division Lockheed Asroopace System Corporstion— Rockwell Ineruatonal Group California Divison Lockheed Aerorpace Systems Corporation — (Georgia Division 55‘Table 11 ‘STRUCTURAL MATERIALS LIST ‘Materials "Polymer matrix pony ‘raphe /epoxy Kevlar/epory oron/epoxy polyimide ‘raphe /bismaleimide ‘ws/bismaleimie thermoplastic raphie/thermoplatic Metal mattix sraphite/aluminom boron/aluminum (ceramic matrix Carbon/carbon Whisker filed composites ‘luminum with slicon carbide whiskers Glan fberreinfored plastic: Advanced Metals and Ale ‘Atuminam conventional aluinom sluminamlithium alloys powdor metallurgy aluminum ‘Tianiam ‘conventional titania follow: 1 0. m1. Vv. v. ‘The corporate experience section requested previous and planned experience with advanced ‘materials in major subassemblies. Section II requested information on where a material might be incorporated into an airframe (wing, fuselage, or empennage). The third section requested background information on material characteristics and the aircraft application. For example, the work sheet included questions on material strength and stiffness, material form (tape or broadgood), type of aircraft and its performance (subsonic or supersonic), the design concept, A sample worksheet is provided in App. A.a ‘and level of automation in manufacturing. ‘The purpose of this section was to set the stage for subsequent cost-related questions. Section IV requested cost data for each of the functional elements listed in Table 1” as well as information concerning cost improvement curves, weight sizing factors, and material buy-to-fly ratios. ‘The last section of the worksheet contained questions on funding profiles, tooling concepts, ete. Generally speaking, responses for the late-1980s time period reflect a company’s cumula- tive production experience with a given material type though that time. In contrast, responses for the mid-1990s reflect a company’s projection of what costs for these materials will be bated on: ongoing development work and an assumed level of capital investment. Unfortunately, when the worksheets were distributed in 1987, the outlook for the defense industry was con- siderably brighter than is the current case. Given current production stretchouts and céncella- tions, itis not clear that companies will be willing (or able) to make the capital investments necessary to achieve the projections. Consequently, the mid-1990 values are undoubtedly on the optimistic side. UNDERLYING ASSUMPTIONS 1. Al costs are in FY90 dollars. 2. The labor hours and material cost factors include all structural fabrication and all assembly up through the airframe group level (wing, fuselage, and empennage). ‘The values do not include final assembly and checkout nor any subsystem installation done before final assembly and checkout. 3, All labor hour and material cost factors represent cumulative average values for a quantity of 100 aircraft and a finished material weight of 1000 Ib.? We used the log- linear cumulative average form because of its (1) simplicity, particularly in a program's planning stages or when one is dealing with large numbers of aircraft, end (2) compatibility with a previous RAND methodology.‘ 4, All responses assumed some mix of material form. The form used will vary from part to part and will depend on a specific parts contour or strength and stiffness require- ments. In an aircraft many parts will require broedgoods, tape, or both. 5. Most responses assumed cocuring of smaller parts and mechanical fastening of large substructures, 6, All responses assumed some mix between manual and semi-automated production facilities (for composites this is characterized somewhat in Sec. 11). No one assumed 1 completely automated system either in the late 1980s or the mid-1990s. "Nonzecuring engineering, nonrecurring tooling, recurring engineering, recurring tooling, recurring manufactring labor, recurting manufacturing smaeral, ad recuring quality assurance. Definitions of these eost elements appear in Apc Sal recurring cost valus obtsined en part ofthe industry survey represented cumulative average hours (or dors) per pound for 100 unit Not all the values reflected the sume finished material weight, Consequenty, we adjusted Sich contractor's values toa finshed material weight of 1000 I using the weight-izing fctore sugested by that cn ‘ens and Romanoff, 1967. ‘The loglinear unit eurve formulation may be more epproprate when one is dealing vith mal quantities or sogmanted curves, or when actuals on easier lots are avaiable.8 SUMMARY OF WORKSHEET RESPONSES This section presents the results of our industry survey beginning with the nonrecurring elements and followed by the recurring cost elements. For each material type and time period, both the average and range of company responses are reported. We determined the everage values for each material type by averaging each firm's minimum and maximum values and then averaging across firms. The range of values presented for a given material type represents the minimum and maximum values reported regardless of firm. Ideally, the range of intercompany responses would be fairly narrow. Unfortunately, this is frequently not the case. However, plausible explanations for such intercompany differences do exist: © Type of aircraft produced by company. Generally speaking, combat aircraft involve ‘more complex parts and utilize higher-temperature materials than noncombat aircraft. Companies producing combat aircraft will tend to have higher per-pound costs than Producers of noncombat aircraft. ‘© Corporate experience with specific material types. It is not unreasonable to expect that as a firm gains production experience with a given material type, its costs for that. ‘material will fall. Firms with limited new material production experience are likely to hhave higher than average costs. ‘+ Level of factory automation. Everything else being equal, firms with higher levels of ‘automation are likely to have lower per-unit labor charges (but higher per-unit over- head charges). ‘Nonrecurring Cost Elements Since nonrecurring costs are often a function of specific program characteristics, simple ratios for an all-aluminum aircraft were requested on the worksheets distributed to industry, particularly for nonrecurring engineering and nonrecurring tooling. ‘The ratios reflect the non- recurring effort required for a component made of a particular material relative to the effort that would be expended for that same component to be made of late-1960s aluminum beseline. ‘These ‘actors apply to airframe structure design and exclude subsystem development. Nonrecurring Engineering. Nonrecurring engineering includes the engineering hours spent designing the airframe. Specifically it includes hours expended for (1) design, consisting of studies, stress analysis, aerodynamics, weight and balance analyses, and integration; (2) wind-tunnel models and mockup wubsystems, and static ‘and fatigue articles; (4) preparation and release of drawings; and (5) process and materials specification. Excluded are engineering hours not directly attributable to the airframe: flight testing, ground handling equipment, spares, and training equipment. ‘The industry average, minimum, and maximum ratios for nonrecurring engineering hours per pound are presented in Table 12. On average, nonrecurring engineering hours per pound in the late 1980s are estimated to be 40 to 70 percent higher for composites than for metals. Only one response suggested the effort required would be less, ‘The specific reasons given for additional nonrecurring engineer- ing hours for composite materials are: * Additional effort is required to manage the individual composite plies. Composite designers have to specify the number and orientation of the plies for each part to achieve the desired strength and stiffness characteristics,Table 12 [NONRECURRING ENGINEERING HOURS PER POUND RATIOS (Late 1960 sluminue = 1.00) Tate 15805 ‘Mid 19808 Material Type __ Average Min/Max Average Min/Max ‘Auminom i too 10 ~—oaio Ablthium 11 tos ko ogna Thanium 11 «Moston Steel li ags toga Graphite /epory 1409251200 Graphitefoumaleimide 150926 «13RD Graphitehermoplatic 1.7 09/80 ka_——O.772S © Engineers are not familiar with the materials and have little experience designing with them, Additional effort is required to develop the analytic tools necessary to model ‘and test them. © There is a lack of material standards and universally accepted safety margins since many properties of these materials are unknown or unproven. New material proper- ties have much greater variability than metals. ‘Therefore, additional modeling of material properties and behavior under load conditions is required. © Achieving high tolerances when required by performance considerations is difficult. ‘The main argument given for fewer engineering hours is: ‘+ Design unitization will reduce the part count and simplify the overall design process. ‘One company suggested that in total, engineering hours are insensitive to material type. ‘That is, « specific part or subassembly will take X amount of engineering hours regardless of the material composition. However, because composites weigh less, a factor must be applied to the hours per pound to compensate. This is in fact the approach NAVAIR took in their study on composites, which relates initial engineering hours/pound to weight savings.® Presumebly greater levels of weight savings will be achieved either by strict design controls and novel tech- niques or increased use of the new materials. The NAVAIR factors for nonrecurring engineer- ing are a function of estimated weight savings. For a weight savings of 10 percent, engineering hours/pound would increase 11 percent; for a 20 percent savings, hours would increase by 25 percent; and for a 30 percent savings in weight, hours would increase by 43 percent (or 1-30). ‘Overall, engineering hours are expected to decrease between the late 1980s and the mid- 1990s, For the metals, the decrease was on the order of 10 percent and attributable primarily to the greater use of computer aided design. For the composites, decreases were estimated to be slightly greater—between 10 percent and 25 percent—because of the combined effect of increased computer aided design and greater familiarity with the materials. ‘Nonrecurring Tooling. ‘Tooling refers to the tools designed solely for use on a particu- lar airframe program, and includes layup tools, autoclave tools, assembly tools, dies, jigs, *Weathersboe, 1988,Cy fixtures, work platforms, and test and checkout equipment. General purpose tools or ‘machinery—such as Gerber cutters, milling machines, presses, routers, lathes, X-ray and ultra- sonic equipment, and autoclaves—are considered capital equipment. Tooling hours include all effort expended in tool and production planning, design, fabrication, assembly, installation, ‘modification, maintenance, rework, and programming and preparation of tapes for numerically controlled machines. Nonrecurring tooling refers to the costs of the initial set of tools and all duplicate tools produced to attain a specified production rate. Industry ratios for nonrecurring tooling hours per pound are presented in Table 13. Non- recurring tooling for composite products is, and will be, substantially higher than nonrecurring tooling for aluminum products. The reasons given to support this conclusion are: * Exposure to high temperatures and pressures in the autoclave. — Increased tool design effort. For example, in the autoclave the tools will expand because of exposure to heat. Tool designers must consider the relationship of the coefficients of thermal expansion between the tool and the material being processed to ensure the desired final part shape is attained, — Higher cost tools. ‘The tools will have to be able to withstand the high temperatures and pressures of the autoclave, thereby limiting the use of aluminum tools and requiring the use of steel, graphite, and electroplated nickel tools. In generel, fin- ished tools made of these materials are more expensive than those made of alumi- num. ‘© The need for highly accurate tools since finished parts can not be reworked once auto- claved, as is possible with metals. * The need for more tools because of the long processing times in the autoclave. Some have said that tools will now be required to make the production tools (if the tools themselves are made of composites). ‘One respondent believes tooling costs may be decreased if a sufficiently unitized design ‘can reduce the overall quantity of tools required. Although individual tools may be very expen- sive, a unitized design will mean fewer parts and smaller tool quantities The differences in nonrecurring tooling costs among the three composites are largely attributable to differences in processing temperature. Currently, processing temperatures above 525° to 650°F are difficult to manage. It was estimated that processing temperatures Table 13, NONRECURRING TOOLING HOURS PER POUND RATIOS. (Late 1980s, aluminum ~ 1.00) Tate 19608 "Mi 10900 ‘Material Type Average Min/Max Average Min/Max 10 oso 10 oso 12 nT os 14 0987 14 aya pi mm 18 0725 14 aspn0 17 O7ms 1s asa Graphitethermoplesic 20 07/3018 05/25 ura above this threshold can add 60 percent to the cost of an individual tool. Newer composites such as polyimides and thermoplastics require processing well above this mark. ‘Tooling costs are expected to decrease by the mid-1990s. One reason given was that computer aided design will help in the planning of the tools. Better material standards, famil- iarity with the manufacturing processes, and improvements therein are other likely reasons for the decrease. In the case of titanium, one company believed tooling costs would rise in the 1990s if SPF/DB processes are used. : ‘The previous discussion and the cost information presented centered on tooling costs per pound fer the airframe. From part to part, tooling costs can vary considerably based on con- tour complexity, number of attachment points, tolerance requirements, autoclave processing temperature and pressure requirements, and size. Nonrecurring Cost Elements Summary. Indications are that nonrecurring costs for airframe structure incorporating new materials will be higher than those for a conventional air- frame, in some cases much higher. As expertise is gained and historical data on new material properties increase, these values will decrease but probably not to the levels of conventional materials. Recurring Cost Elements ‘This section provides cost-per-pound factors for each material type for each of the five recurring cost categories: engineering, tooling, manufacturing labor, manufacturing materials, and quality assurance. Once again, both the average and range of company responses are precented. Unfortunately, the range of responses for the recurring elements is frequently very large. However there was a tendency for the majority of responses to cluster in narrow ranges. Consequently, we also document the cluster range and recommend its use for general sensi- tivity analyses rather than the larger, unrestricted ranges. Recurring Engineering. Industry responses for recurring (or sustaining) engineering are summarized in Table 14. Recurring engineering clearly varies by material type. Unfor- tunately, we have not yet been able to establish a logical relationship between material type ‘and sustaining engineering. In fact, two companies indicated that new materials will not affect sustaining engineering. Moreover, a third company declined to provide any response for this cost category, stating that sustaining engineering hours are a function of many variables (government mandates, funding levels, program phase, and aircraft mission), none of which are related to material type. Tn any case, for the firms that did respond, sustaining engineering hours per pound for srephite/epoxy are expected to be twice the hours of aluminum in both the late 1960s and ‘mid-199%e (with graphite/biameleimide and graphite/thermoplastic somewhat higher). By the ‘id-1996, graphite/epoxy is projected to be roughly twice as expensive per pound as alumi- ‘num; however, the bismaleimides’ and thermoplastics’ sustaining engineering hours per pound will be roughly equivalent to those of the epoxies. Reeurring Tooling. Recurring (or sustaining) tooling refers to the effort required to ‘maintain and repair production tools. These hours are reported to be greater for structure fabricated with new materials than with aluminum, The values shown in Table 16 are over ‘twice thceo of aluminum.® Some of the reasons for higher recurring tooling costs for composite materials are: A NAVAIR stady (Weatherebee, 1983) reported approximately a 30 percent increese in sustaining tooling for rephite/eory,lse than the value shown here.Table 14 RECURRING ENGINEERING HOURS/POUND (Cumulative average hour for 100 units for 1000 Ib of tructare) ‘Min/Max Min/Max ‘Time Peviod/ Average Valve in Value in Material Type Value. Cluster Sample Tate 19808 ‘Aluminum 10 aso oan ‘Aviesiam Li aga 04 ‘Tianiun M4 Orn oss Stee 11 ogo asa Graphite/epoxy 1808s Oana Graphiteimmaleimide 2113/28 OA Grephite/thermoplastic 28 09/82 (O/T. Mi. 19008 Aluminum 09 oso oan Ablchiam, 19 alo oan ‘Tianism 12 ams asa Steel Ll oso ass Graphite/epoxy 15 oso asia Grophite/bismalsimide 18 = 10723 aae. Graphivethermoplatic 14 08722_03/86 Table 15 RECURRING TOOLING HOURS/POUND (Cumulative average hous for 160 units for 1000 Ib of structure) ‘Min/Max Min/Max ‘Time Peviod/ Average Valuein Value fn Material Type Value Chaster Sample ite 1808 ‘Aluminum 1s asf as/52 Avian 17 Osa asyas ‘Titanium 300529 os/ar Steel, 23065 Ow 73 Graphite/epoxy a8 ow/e7 08/93 Graphitebismaleimide 3.7 = O8/BB 08/93 Graphite/thermoplastic © 89 O/T. OT/IOS Mia. 18908 ‘Aluminom 15 ogy osyas ‘ALthiam 17 ose ogyas ‘Titanium 28 068 08/82 Steal 2308/23/73 Graphite /epoxy 32 04/80 04/86 Graphite/bismaleimide 8305/81 (04/85 Graphite/thermoplastic 38 06/70 04/105‘* Tools must be cleaned after going through the autoclave, a time consuming process. ‘+ Because of thermal cycling in the autoclave, tools undergo more wear and need to be replaced more frequently. ‘© Composite tools are less durable than metal tools since nicks and scratches cannot necessarily be machined away. ‘© Storing composite tools is more demanding. ‘The saphite/bismaleimide and graphite/thermoplastic values in Table 15 reflect the slightly higher temperatures and longer cycles required for an autoclave cure over those required to process graphite/epoxy. If manufacturing processes that replace autoclave cure are developed for thermoplastics (those taking advantage of thermoplastics’ reformability, such as the stamping out of parts), some respondents indicated that tooling hours may actually decrease below the levels reported for graphite/epory. Recurring Manufacturing. Recurring manufacturing hours/pound are presented in Table 16. In general, the minimum values are based on larger, fairly simple subsonic aircraft, whereas the maximum values are based on aircraft utilizing more complex, contoured struc- tures, Manufacturing hours per pound clearly vary with material type. Aluminum-lithium and steel are only slightly more expensive on an hours per pound basis than aluminum, while titanium is noticeably more expensive, because titanium is particularly difficult to machine and is highly resistant to chemical milling. Finally, composites are much more expensive than aluminum ona dollar per pound basis. Graphite/epoxy and graphite/bismaleimide are approximately twice as, ‘expensive es aluminum in the late 1980s. Informed judgments voiced during our visits were that ‘saphite/bismaleimide requires 10 percent to 20 percent more hours per pound than Table 16 [RECURRING MANUFACTURING HOURS/POUND (Cunmultive average houre for 100 units for 1000 I of structure) Min/Max Min/Max ‘Time Period/ ——Average-Value in—_Value in Material Type Value Chater Sample Tate 19608 ‘Alumina a1 72 ase Avithiom 14 A078 ‘Titaniom ua ant 7/a08 Steet 9 BAT 5/228 Graphive/epoxy 16219720 28/408 CGraphive/bismeleimide 10.4 9m 50/89 Grephite/thermoplastic 168 10/25 5.0/2868 ‘Mid-19908 ‘Aluminum so gto agst09 ‘avithism afin ‘Titanium 2s 78 58/288 Steet os as aes Graphite/epory ms fie (2a Grephie/biemsleimide 1658/20 40/403 Graphitethermoplanic 145 9/214. 2/255‘sraphite/epoxy because (1) it has less tack and poorer drapability characteristics, making it more difficlt to lay up; and (2) it has a slightly longer cure time. Although many believe thermoplastics have the potential to be treated more like motels—stamped, pultruded, welded, and machined easier; reformed; processed quicker; etc.— few companies have experience working with them, and the processes for manufacturing them are still in development. Other studies have estimated thermoplastic materials hours per pound to be twice those of aluminum.” We have been told that if thermoplastics are processed. ‘as thermosets (Iaid up and autoclaved), then they will never be cost-effective. Laying up ther- moplastics is more difficult because they have no tack and poor drapability characteristics at room temperature, necessitating external heat. If autoclaved, thermoplastics cure temperature requirements are at the limits of the existing capability in most manufacturing facilities® Achieving the values reported here appear to be contingent on the successful development of processing technologies for thermoplastics and the willingness to invest in the capital equip- ‘ment necessary to use them. Every company that did projections assumed automation would reduce manufécturing hhours per pound to some extent. Some believe that in mass production, with automation, sraphite/bismaleimide will be indistinguishable from graphite/epoxy in terms of manufacturing hours per pound. Examples taken from the DOD/NASA Structural Composites Fabrication Guide and reporied in Sec. II outline possible improvements in the process that may greatly reduce ‘manufacturing hours. The Guide states that manual layup accounts for 50 percent of spar’s labor cost and 90 percent of a cover’s labor cost. Industry's ability to develop and incorporate automated techniques, with a special emphasis on potential payoffs to automating layup, will determine the amount of savings achievable. The Guide also states that as much as half of the total fabrication cost of structural composite torque boxes can be attributed to mechanical fastering operations.’ It follows that another essential ingredient to reducing costs and weight is greater use of cocured assemblies. It is difficult to separate the effects of low observability requirements from the material effects on hours per pound. Some companies believe that this requirement could add 15 per- cent to 30 percent to recurring manufacturing hours (some of the higher values in our range as shown in Table 16 may reflect this complexity). Recurring Quality Assurance. Industry average, minimum, and maximum recurring quality assurance hours per pound are shown in Table 17. Some believe quality assurance hhours across all material types will be higher in general because of the customer requirement. for an improved product. A NAVAIR study did not quantify but stated that quality assurance hours for composites will be much higher than those for aluminum. Because the materials are ‘new and unproven, more testing is required. Also, composite materials quality assurance hours ‘may be higher because of added requirements driven by the materials themselves and their associated processes. For example, these materials must be inspected for internal delamina- tions and bond integrity, both of which are not possible to do visually and neither of which is a concem in metal fabrication. In addition, test results must be well characterized so that ‘engineering decisions with respect to defect severity and repair (or scrap) procedures can be made. To compound these problems, testing procedures and guidelines are still evolving. inally, the type and extent of the quality assurance procedure used will depend on the "Pingel, 1987, Sealing would alo be affected. Meads, 1982‘Table 17 RECURRING QUALITY ASSURANCE HOURS/POUND (Cumulative everage hours for 100 units for 1000 Ib of structure) Min/Max Min/Mex ‘Time Period/ Average Valu in Value in Matarial Type Valve Custer Sample Tate 19808 “Aluminum a7 onan sya ‘Avia 18 0a/28 04/35 ‘Titanium 27 10/4d 05/60 Steet 24 09/89 08/53 Graphite/epory 41 cata o7fog Graphitefoimalsimide © 4308/78 OS/LLB Graphitefthermoplastic 44 «10/78 08/108 Mi19908 ‘Aluminum 15 onze oan Abithiom LT 0/26 04/33 ‘Tianium 24 09/39/82 Steet 24 09/39 08/83 Graphite /epoxy 3108/88 05/92 Graphitefoismaleimide 3606/66 O6/104 Grephitethermoplastic 34 07/81 08/9. criticality of the component to the overall system. Some components may require several dif- ferent tests. The values reported here reflect composite usage between 30 and 50 percent of structural weight, some of which will occur in structural and load-bearing components. Material Costs. Three elements determine total material cost: raw materisl cost, bay- to-fly ratio, and material burden rate. ‘The buy-to-fly ratio is the amount of material pur- chased to complete a pound of finished part that “flies” away. Since in our context buy-to-fly ratios are used to calculate the total material bill, they should include material lost in poor material kandling processes, the machining or cutting processes, the fabrication of parts that are eventually scrapped because of flaws or deficiencies, and other steps in the manufacturing process that cause material to be lost.!? ‘The costs reported here represent a particular material’s average dollars per pound for the entire airframe structure, implicitly incorporating a typical mix of material forms. A material's specific costs and buy-to-fly ratio may vary widely depending on the form (sheet, plate, billet, tape, broadgood) or precise grade of material. For part level analysis, other detailed informa- tion regarding the specific material form would be more appropriate." ‘The material burden rate captures the costs of purchasing, handling, and inspection. The tates reported to us varied from company to company, reflecting differing accounting THT manufecturing labor improvement curve slope may aio be affected by the scrap rate. We have been tld that duting the very ealy stages of production, seap rates aa high ea 200 percent are not uncommon. Once prod tion is enterd, however, these problem wil have been resolved and parte will only rarely be discarded. ‘Therefor, terly manufoeturing labor improvement curve slopes tay be stegper than thece experienced later in production. ‘for example, limited information soggets that buy-to-y rtie for composite broadgoods are 25 to 75 percent higher than those for composite tape. Based on this limited information, the values in Table 18 more closely resemble composite tape ration. Buy-tofy ratios oun alto be function of how clove the raw material i to the final form. A {ower buy-o‘ly rato should corespond to higher material costs since the taterial would have required aditonalpro> ‘sing (to ptf close to final form).Procedures. We have been informed these procedures are being revised to ensure burden rates uniformly represent value added. The industry average of 15 percent was used in subsequent calculations (the range was 0 to 33 percent). ‘Table 18 summarizes the information we have received on raw material costs and buy-to- fly ratios. Composite raw material costs are expected to decrease in the mid-1990s approxi- ‘mately 20 percent because of greater supplier economies of scale. Buy-to-fly ratios are also estimated to decrease. At the time these projections, were made, the mid- to late 1980s, indus- ‘try expected that the demand for composite materials was increasing and therefore economy- of-scale benefits would be realized. However, the current situation is less certain with the potential cancellation or stretchout of major aircraft programs likely to reduce projected bene- fits. Additionally, increased energy prices may adversely affect petroleum based materials’ costs. ‘Advanced material costs have been of great concern to manufacturers because they are 60 ‘much more expensive than traditional materials. Some of these additional costs have been offset by lower buy-to-fly ratios. Moreover, material costs constitute a small proportion of ‘overall recurring dollars (in the range of 2 to 9 percent as shown in Tables 21 and 22). There- fore, even though composite materials can cost an order of magnitude more than stock metal ‘material, the effect on total cost is emall. Cost-Improvement Curve Slopes and Weight-Sizing Factors. Suggested cost- improvement curve slopes for adjusting the recurring cost values to quantities other than 100 Table 18, MATERIAL COST FACTORS: Raw Buy: Material Materai® ‘Time Period/ to Fly $b fb Material Type Ratio (FYB08) _(FY908) Tate 19808 ‘Aluninurs ae 2 ‘Avlghiom 42007 2 ‘Tianium 30% 6 Steed ee 1B Graphie/epory 2 8 Graphite/bismaleimide mW (Graphite/thermoplaatic ae. ‘Mid. 18008 ‘Akuminom 22 Attia 278 ‘Titanium 30m Steel 28 Graphie/epory 13st Graphitefbismalsimide 18 61 Graphitethermoplastic 18 66 "ho burden added. ‘The average industry burden is 5 percent.a are provided in Table 19 along with weight-sizing factors for adjusting finished-material weights to values other than 1000 Ib.!? Recurring Cost Element Summary. As the data in Tables 14 through 18 show, all recurring cost elements are expected to be greater for composite materials than for metals. ‘The support categories are higher for composites than for metals and the manufacturing hours are greater as well. Some reduction of all values in the 1990s is expected. Automation, general improvement of the procestes, and greater material supply are expected to be the important factors in the reduction of advanced materials’ total recurring costs per pound. Weight savings achieved through the use of these new materials will help mitigate the additional costs. APPLYING THE DATA ‘This section develops estimates of total recurring dollars per pound by material type and time period. Industry average labor hours are converted to dollars by means of the fully bur- dened Isbor rates shown in Table 20.” Material costs are then burdened and added to lsbor costa to got an overall recurring dollars per pound. ‘The results are shown for the late 19808 and mid-1990e in Tables 21 and 22, respectively. ‘The following conclusions can be drawn from these two tables: ‘Table 19 COST-IMPROVEMENT CURVE SLOPES AND WEIGHT SIZING FACTORS. (Percent) ‘Average Minimum ‘Maximo Improvement ave slope" ‘Recurring engineering cy 58 a5 Recurring tooling B 6 1 Manufacturing labor 8 n 1 Manufacturing meteral 87 ” 97 Quality assurance ® 6 100 ‘Weight sizing factors? Tabor a5 % %0 Material 6 & 100 "RAND Aircraft Airame Database (proprietary database consisting of com deta obtained ffom major aircraft fle, eithr directly from their records or inaireey through standard Department of Defense reports such asthe Contac: tor Cort Data Reporting (CDR) sytem) ites indurtry survey (eee App. A, Workahost IV) "Ws par of the industry survey, we requested improvement curve slopes by material type and time-frame forthe smanftctarng labor cost element (end only the manufacturing labor cost element). However, a lack of quantitative ‘eponoes prevented from establishing a relationship between slope and material type/time frame, Neverthelee, ‘tnhough urable to quantify the effect, several respondents believed composites would have steeper improvement crv Slopes than metals bechuse soany of the composite fabrication processes were dominated by manual operations and ‘were new aod therefore continually being improved. Another respondent suggected tat slopes were more sensitive to {Sutomation levels than material tye. "OTe labor rates represent industry averages and includ al cost for direct labor, faite, capital cost of mcney, noel and adminiatatve expenses, and other direct charges (computer time, support services, travel, overtime, Hinge Senet prenum, and factored labor). Profit is excudedTable 20 INDUSTRY AVERAGE FULLY BURDENED LABOR RATES (in 1990 dolar per hour) ‘Bngineering labor 3080 ‘Tooling labor 7080 Mancfactuing bor 66.10 Quality amurance labor 65.60 ‘SOURCE: These rates are besed on an Informal survey of six major alreratt Girne that requested thei identities not be dieloeed. Table 21 [LATE 1980: RECURRING DOLLARS PER POUND (INDUSTRY AVERAGE) (Cumulative average dollars for 100 units for 1000 Ib ofstrsceure in 1980 dllary) Gost Bloment ‘ALALLI Ti Steel Ge/B Gy/BMI_Gr/TP Manutactring aber” 6080859772 YO a8a aT Raw material a 8820150168180 ‘Total manufacturing cost 6M 768 1,065 7421201 14311810 ‘Support labor m2 998 4990s ‘Total reeuring cost 995.1004 1,558 1183 1806 2144 2.108, “inches fbviention and sesembly labor. “includes material burden of 16 percent. “Includes sustaining engineering, sustaining tooling, and quality asurance Table 22 ‘MID.3990 RECURRING DOLLARS PER POUND (INDUSTRY AVERAGE) (Camlative average dllars for 100 unite for 1000 Ib of structare in 1990 dolar) cerBiooent ALAM Stet Grit Ga/BMGurTP Manfocung bo” 629 6a 0 nia aRD ae aa Rw naeralé 2 '6 ‘we i ior ‘ota manutoturing cont 554 G8 912708888218 L085 ‘Support labor® 27 S14 440 4155087 606 ‘Toa curing cont N92 188219 AKG LAL 1.70L “Incas fibiation and evenly abo "includes material burden of 15 percent. *Yucludes sustaining engineering sustaining tooling, and quality assurance‘© On average, total recurring costs for composites are about twice those of aluminum and steel and 20 to 30 percent more than those of titanium. © Recurring costs for both metals and composites are projected to decrease between the lete 1980s and the mid-1990s. However, the improvement in composite costs (appraxi- mately 20 percent) is projected to be roughly twice the improvement in metal costs. ‘© Raw material costs are a small proportion of total recurring costs, typically less than 10 percent of the total and sometimes as little as 2 percent. ‘© Support labor costs account for 30 to 40 percent of total recurring costs. ‘Table 23 presents the range in overall recurring dollars per pound using the cluster ranges provided in earlier tables, ‘Two observations regarding these ranges stand out. First, compos: ite costs are subject to greater uncertainty than metal costs; the upper and lower bounds for the metals vary by a factor of two, and those for composites vary by a factor of three. Hew- ever, the larger range expressed for composites is consistent with the fact that, compared with metals, they are a newer technology with considerably less design and production experience. ‘The second observation concerns the absolute magnitude of the ranges. As stated earlier, part of the range can be explained by corporate-specific factors such as cumulative experience with 1 given material type and the level of factory automation. However, we believe that most of the range can be explained by part complexity. In general, for a given manufacturing tech- nigue, simple parts (large flat surfaces and nonload-bearing structure) will cost less per pound than complex parts (surfaces with compound curves, sine-wave spars, and load-bearing stroc- ture). Additionally, simple parts lend themselves to labor-saving automation in both the fabri- ‘cation and inspection processes, whereas it is considerably more difficult to automate these processes for complex parts. Part complexity can be reflected in either of two ways: ‘© Type of aircraft. Generally speaking, combat aircraft, because of their smaller size (which reduces the probability that automation will be cost-effective) and greater emphasis on performance (for example, higher speeds and greater maneuverability requirements, leading to more complex contoured structural elements) have greater part complexity than noncombat aircraft. ‘© Level of material usage. Generally speaking, as the level of usage of a given material type in a specific application increases, the part complexity also increases. This point is ilustrated by the example provided in Table 24, which shows where composizes would probably be incorporated into the structure of a generic fighter as the level of Table 23 OVERALL RECURRING DOLLARS PER POUND (INDUSTRY MIN/MAX) (Cumulative average dollars fr 100 units for 1000 Ib of structure in 1990 doles) Cont Bioment _AL_ALLI Ti ‘Steel Ge/B Ge/BMI_Gi/TP Tate 19806 ‘Minimum rcurting 614740774689 970955 1,048 ‘Maximum recurring 1202 1249 1826 1857 2682 2972 3123 ‘Minimum recurring 521 612,703 eB] 622812889 Maximum recurring 1046 1.207 1.649.811 2199 2490 2.507Table 24 PART SELECTION AS A FUNCTION OF GRAPHITE COMPOSITION "FOR A GENERIC FIGHTER Graphite Pereet of Structure Component Type w. 8 8 6S Nonstrctural acces door/panels -X—=X XXX Structural accew doors/panels” = «X =X =X XX, Vertical stabilizer skins Bex x, x) ax Hlorisontalvabilizer eins Mx x x) ar Wing shine Ro oe ee Control surtces x xX x x x Speed brake eet x cs ‘Landing getr doors ee x Xe xe x ‘Additional doors/panels KX xX x Xx ‘Additional control surfeces oe eH Spar, ibe eek ex x ‘Shear webs ekin panels K x x Longerons xx Frames, formers oe x Bolkheads x ‘SOURCE: Aircraft contactor. ‘usage rises. It clearly demonstrates that as the level of substitution of composites for ‘metals rises, the complexity of the application also rises. ‘SUMMARY Industry responses to a deta collection worksheet have been summarized in this section. ‘The average, minimum, and maximum values for nonrecurring and recurring costs have been reported by material type. Responses show that costs do vary by material type and thet on @ per pcund basis composite materials are more expensive than metal materials. Weight savings can mitigate the additional recurring costs. However, the values reported indicate it is unlikely the savings will be great enough to offset the additional costs.V. SUGGESTED METHODOLOGY FOR ASSESSING THE EFFECT OF STRUCTURAL MATERIAL COMPOSITION ON OVERALL AIRFRAME COST This section suggests an approach for aggregating the previously developed material cost factors into an overall structure cost. However, structure-related costs are only a part of overall airframe cost and are not normally considered in isolation.’ As indicated in Table 25, (on average, structure “accounts” for less than half of engineering (nonrecurring and recurring) and for 60 to 70 percent of manufacturing labor, manufacturing material, and quality assurance. Consequently, the approach that is outlined is one that considers overall airframe cost—that is, one that takes into account not only the cost of the airframe structure but also the cost of the airframe subsystems and final assembly/integration. ‘The basic inputs that the user must provide to exercise the methodology are as follows: ‘Aircraft empty weight (Ib). © Maximum speed (kn). ‘© Number of flight test aircraft. (Input list continues on next page.) Table 25 PERCENTAGES OF FUNCTIONAL COST ELEMENTS ATTRIBUTABLE TO STRUCTURE, Recurting Recurring Recurring Nonrecutring Nonreeurring Recurring Recurring Manufacturing Manufacturing Qualiy Aira Engineering Tooling Eapnering Tooling ‘Labor ‘Matai Asurace x 7 = a = = at a o 6 @ = = Mo & = = a = = Bae Pr = 2 = = cs. e ® @ % a a ctw a = a Py a % cua 8 Py 8 8 8 Fa 51 a 51 ou ® = Pie oe mu fo a s o Ps a a = 7 Pas 2 eo 8 ® 16 = Kos 88 309 fw = = Avene a 2 @ Ey e ‘SOURCE: Proprietary RAND Aireraft Airframe Groop Manhour and Cost ing oF incomplete data. The trm airframe cott refers tothe cost of the assembled structural and aerodynamic component ofthe air vehi- cle eat aipporesubaystema sasentil to 8 particular mission. Tt Includes not only the baie structure (wing, fumlage, fmpennege, and nacelles, but lao tho al induction system, starters, exhaust, fuel control system, inlet contr ae fen alighing gar (ties, tubes, whale, brakes, hydraulics, ee), eecondary power, furnishing (argo, penenger, 00D, ste, engine controls, instruments (Tight navigation, engine, etc), environmental contol, racks, moun, inerastem ‘able ane ditibution boxen ste, iabereat to and inseparable from the aasembled structure, dysamic systems, and ‘other equipment homogeneous tothe airframe. Airframe coats also encompass Ue inogration and intalltion ofthe propulsion avionice, and armament subaystemns into the airframe but not those efforts diecly related to theit develop ‘ent and manufacture. (For additonal clarification, see Department of Defense, MIL-STD-881, Work Breakdown, Structures jor Defense Material Items, 25 Apel 1875) nn + Type of aircraft (cargo or noncargo). ‘+ Total structure weight? by material type. «+ Percentages of functional cost elements attributable to structure. Default values are provided for all other necessary inputs, DESCRIPTION OF APPROACH Described below is a method for estimating the development and production coste of air- craft eirframes that is: sensitive to the material composition of the structure and suitable for use in a program's conceptual stage when little detailed information is available. Reduced to its simplest form, the method applies weighted material indexes to baseline CERs assumed to be representative of all-aluminum aircraft. It provides separate CERs and material indexes for the following major cost elements:” Nonrecurring engineering. Nonrecurring tooling. Development support. Flight test. Recurring engineering. Recurring tooling. Recurring manufacturing labor. Recurring manufacturing material. Recurring quality assurance. ‘The method does not, however, address the following: ‘© Costs of engines, avionics equipment, and armament. ‘© Training, support equipment, data, and spares. Baseline CERs ‘The baseline CERs are listed in Table 26 and are identical to those documented in -8255-AF* with the following exceptions: ‘© The engineering and tooling equations have been split into separate nonrecurring and recurring components. ‘* The manufacturing material, development support, and flight test CERs have been updated from FY77 dollars to FY90 dollars. ‘They were derived from a database consisting of 13 military aircraft with first fight dates ranging from 1960 to 1978: A-6, A-7, A-l0, C-5, C-ld1, F-4, F-14, F-15, F-16, F-18, Fel, S-3A, and T-39, Empty weights for the sample sireraft range from under 10,000 Ib to over 300,000 Ib, while speeds range from 400 kn to over 1,300 kn. "Por ype ofthis methodology, stricture weigh nclue the weight of turlage, empeanag, sae sad ar edon grup Ts oe onde du et ofthe algae teat ey or addon) Cane, ot ‘Dob, MIl-STD-1StiA: Woh and Balance Date Reporting Porm for Atte Cont slementdefinicone are provide in App. C- “He and Rommanof, 1967Table 26 AIRFRAME CERS BASED ON SAMPLE OF 13 POST-1960 AIRCRAFT Bauation ws OF ON ent a ‘NRENGR ~ 0168 EW" SP’ 7 eS wD (000) (018) NRTOOL ~ 186 EW2)” sp5™ a) (000) 003), 3 ~ 0563 Ew sp! “oom 6 (018) (012) Fr 154 BW" sp restac!! 33 BoB (082) (037) (010) ENGR) - ommne ew sp? ss st 7 (206) 088) TOO; = ors EWT™ se gam (00) (007) LABR,g)~ 141 EWS spo a (000) (018) MAT yp = 540 EWS?! sp a 0 8 8 (000) (003) {16 LABRygg (80) - - - 2 TABOLABR (oon) TR cosfficient of determination, SEE = standard error of estimate \ogathon); F = Festatiaticy N ~ sample size. Numbers in parentheses are Sigfeance levels of indvidon! variables. ‘DS = Development support cost (thousands of 1990 dollars) ENGRjop ~ Corulative reeusing engineering hours for 100 sirraf, (thousands "EW = Aircraft empty weight (b) FT ~ Flight eet cot (thoutands of 1990 dollars) LAB = Qulative reuring anhtrng ar bors fr 10 sireratt (housand) MATL = Cumulative recuring manufacturing material comt fr 300 sireraft (iousands of 1990 dolar) 'NRENGR ~ Nonrecurrin engineering hours (thousands) NRTOOL ~ Nonreeutring tooling hours (Chousands) @Aog ~ Cumulative recurring quality ateurance hours for 100 aircraft (ehoustade) ‘SP = Maximum speed (kn) "TRSTAC = Nomber of fight-tstsiterat TOOL yg = Cumulative recurring tooling hours for 100 aircraft (chousands? cs Application of weighted material indexes to these CERs requires that the following key assumptions be made: © That the CERs are representative of the baseline material (aluminum). © That the CERs are representative of late 1980s manufacturing technology. Both assumptions are admittedly heroic. For example, as ‘num content of at least three of the fighters in the database shown in Table 27, the alumi- only 60 percent. And substan- tial improvements in manufacturing technology have been made over the last three decades.™ Table 27 ‘AIRFRAME MATERIALS UTILIZATION Percentage of Airframe Structure Weight ‘Alneraf (ret Hight date) F4 Pu Pas. Pas P16 PAB Material (1961) (1967) i970) (1972) (976) A878) a er Viele | 0) et melo isl Steet wo 8 2 5 4 15 Composites to 2 6 4 Other se 1 0 For example, if the F-4 had been produced using F-15 manufacturing techniques, the unit cost of the 165th aircraft would have been 12-1/2 percent lower than it actually was.’ Unfortu- nately, there is not much that can be done to alleviate these difficulties.° To the extent that ‘the CERs actually embody more advanced material types and older manufacturing techniques than we have claimed, the-bias is clear: Applying late 1980s material cost factors to the CERs will result in somewhat higher estimates than we would otherwise have obtained.” Weighted Material Indexes ‘The weighted material indexes applied to each cost element are calculated as follows: Wwcr, - sTRFRAG, x[Z RMCR, x (GTRWTVE strwny] + [1 - srrFrag | where WMCF, = Weighted material cost factor for cost element ‘ STRFRAC, ~ Portion of cost element {attributable to structure (Table 25) RMCF,, = Material‘? structural cost index for cost element ‘ STRWT, Material ‘? structural weight ‘The relative material cost factors (RMCFs) are provided in Table 28. They have been calcu- lated on the basis of the average values presented in Sec. IV, For each cost element, the Inte- 1980s aluminum value was considered the baseline. The complexity factors represent the ratios of cumulative average values for 100 airframes normalized to a weight of 1000 Ib. Thus, ‘Stain, 1985, p. 428, One posiility would be to try to incorporate « material index dneetly into the CERa as part of the repeaion snalyis. However, this i «highly uncertain proposition since it would require the development of histor! material ott factrg, an effort that we believe would have it chance of succes. “Another concer with respect to the applicblty of the CERs to future aircraft has to do with the nonaractral sleframe subeystems (eg, electrical, hydraulic, environmental contre, and fol syatem). Imply, we asnane that there wil be no change. However, as lower radar crom-aaction end greater tansuverability are accght, many of the ontructralsubayrtme willbe pubed fo thir Liite, While an examination of auch elects was beyond Une tope of {his Ray, we alee the direction of change is unambiguour:Subeyeer costs wil incest Flav to the database to dave ouré a @ som domane # a spre raed i a " odode) z 8 z peg Mg ra = roy 60 Hy 066 PN : 2 en i : spreoqesg/aidst i ‘xode/engdeny vt 4 = st i ‘ nen = ars Am, MES scp = sr 201 samney Tomson, Resse z = boa Tosi wa (oort ~ ema so961 1) SUOLOV ALIXTTAWOO TWIEGLVW BALLV TE 82 196,6 to the extent the user believes these ratios vary with either the quantity or the weight, they should be adjusted accordingly.® Steps in Applying the Methodology Assess Applicability. ‘The first step in applying this method is to ensure that the empty weight and speed of the proposed aircraft lie within the range of the estimating sample: Aircraft empty weight (Ib): 9,753 to 820,085 Speed (kn): 400 to 1300+ Additionally, the estimator should ensure that the proposed aircraft does not differ substan- tially from the estimating sample along such other key dimensions as: ‘* The nature and stringency of the design requirements (e.g., radar signature require- ‘ments may be such that the tolerances, contours, and surface costings of the proposed aircraft do not reflect the estimating sample). * The type of programmatic strategies to be employed (eg., the classification of a pro- posed aircraft program as “special access” would clearly set it apart from the programs in the estimating sample). Apply Basic CERs. Once applicability has been assessed, values for each of the CERs listed in Table 26 should be determined. For those CERs estimated in manhours, the following. fully burdened labor rates (in FY90 dollars) can be used to convert hours to equivalent dollars: Engineering 80.80 Tooling 70.60 Manufacturing 66.10 Quality assurance 65.60 Calculate Adjustment Factors and Apply to Basic Values. Finally, the weighted ‘material cost factors should be calculated for each cost element, Default (average) values for the cost element structural fractions and relative material complexity factors are provided in ‘Tables 25 and 28, respectively. However, if the estimator has information that he ot she believes will more accurately characterize the proposed program, then thet information should ‘be used. For example, consider the wide variability in the nonrecurring engineering structural fractiors presented in Table 25; values run from a low of 12 percent (F-16) to a high of 63 per- cent (C-5). In percentage terms, structure-related nonrecurring engineering was five times as important on the C-5 program as on the F-16 program. However, we know the F-16 develop- ‘ment program was preceded by fairly “hi-idelity” prototypes, whereas the C-5 program was & very challenging, completely new development without prototypes. Moreover, the F-16 incor- porated an engine that was already in service (the F100) while the C-5 used the TF-39, which ‘was not only new but the highest thrust engine in the world at the time. Many such observa- tions can be made regarding the values in Table 25. The estimator should not disregard such observations when deciding on appropriate structural fractions? "oweer, we found no evidence in the industry survey responses to auggtet that these ratios do var with quantity or weight ‘Siar arguments can be made with rspect tothe relative material complenity factors presented in Table 28. For example, hese ration might be adjusted upward if the estimator beloves Ut a company's experience with an advanced ‘material type is lose Uhan the industry averege-n Adjusting Values to Quantities Other Than 100. To adjust recurring values to ‘quantities other than 100, improvement curve slopes are suggested as shown in Table 29. APPLICATION TO HYPOTHETICAL AIRCRAFT ‘The just-described approach is now applied to two hypothetical aircraft. As indicated in ‘Table 20, both planes have been defined such that their weight and speed are within the limits of the estimating sample and both have been assumed to be conventional acquisitions with respect to design and programmatic considerations. The only differences between the two craft are in the weights. The structure of Aircraft 1 is all aluminum and weighs 13,000 Ib, while the structure of Aircraft 2 is 55 percent aluminum and 45 percent graphite and weighs only 11,700 Ib. We assumed that only 5,200 Ib of graphite/epoxy would be required to replace 6,500 Ib of aluminum, a 20 percent weight savings.!° Table 29 FUNCTIONAL COST ELEMENT IMPROVEMENT ‘CURVE SLOPES ‘Gamat Total Cont Element Slope (%)* Exponent ‘Recurring engineering 10 485 Recurring tooling M6, ae ‘Mamufecturing labor 156, sa ‘Manufacturing material 1m 790 Quality assurance 16 sa "quivalent camulative average slopes may be found by dividing the cumblative-total slopes by two. Table 30 [REQUIRED INPUTS Case & Case 1: 86% Aluminum/ Al-Aluminom 45% Graphite/epoxy Input Structure ‘Structure impty weight (8) 77000 353700 Speed (kn) 1300 1,300 [Number of tert aircraft » 2 ‘Structure weight (b) “Aluminum 1800 Graphite/-poxy ° Alrerat (ype Fighter (Cont factor time frame Late 18608 ‘Ununval design requirements ‘None ‘nunual programmatic considerations None “Ppotenally, even greater weight savings than suggested her are posible through the process of sireraft “resting” ‘As discus in See. I, the potential for resizing depends on the alrcraft requirements and the “freedom” to make‘Tre results are presented in Table 31. The most important observation with respect to ‘these cases is that based on today’s manufacturing technology, the substitution of ‘sraphite/epoxy for half the aluminum in the structure will increase nonrecurring cos:s by about 4 percent and recurring costs by roughly 36 pereent.1" Finally, the effect of projected manufacturing improvements on airframe cost is shown in ‘Table $2. As indicated, relative to the late 1980s, nonrecurring costs in the mid-1990s are pro- jected to decrease about 3 percent while recurring costs are projected to drop about 9 percent. ‘Table 31 [EFFECT OF COMPOSITE SUBSTITUTION ON AIRFRAME COST (Milions of FY9O dollars) Case case 1 85% Aluminom/ AlLAluminum 48% Graphite/epoxy Structure Strocrure as 95 am 385, 389 378 Ey ae 675 288, 1837 1078 ey 3882, 5008 108 Gesigndeisions (particularly with reapect to the engine). As result, the resizing of an aizcraR is « complex press that does aot led itself to simple rule of thumb Since our example is strictly illustrative, we have chosen 204 to ‘dross this eapct. However, in conducting real-world tradeofls, the cost analyst should consulta qualified engineer ‘egnrding the resizing inoue. ‘This conclusion axsumes that we stated with an al-sluminum aireraft and thatthe only weight saving resulted lzectly fom the material wubetitution, Dilfeent material mixes aa well o consideration of aireret resning sfeta would produce diferent rvults,Table 32 [EFFECT OF PROJECTED MANUFACTURING IMPROVEMENTS, ON AIRFRAME COST (alone of FY90 dollars) Case 2 Case 5 55% Aluminom/ 55% Aluminum/ 45% Geaphite/epory 45% Graphite/poxy Structure; Structure, Late 19805 Mie 18006 396, ez 385, 300 a8 578 ea 68 6 2168 os ae 358 328 1887 1378 1078 290 234 198 3682 a e108 saaVI. CONCLUSIONS This report identifies, describes, and quantifies the cost effects of structural materials likely to be incorporated into aircraft becoming operational in the 1990s. More specifically, it ‘summarizes cost information obtained from aircraft prime contractors for the materials, func- tional cost elements, and time periods given in Table 33. RELATIVE MATERIAL COSTS ‘The data indicate that both nonrecurring and recurring costs per pound are higher for composites than for metals. In most cases, these costs are considerably higher. Nonrecurring ‘engineering for composite structure is between 40 and 70 percent higher than for aluminum structure, largely because of the “newness” of composites (properties are not standardized and analytic tools for design are still being developed). Similarly, nonrecurring tooling for compos- ites is between 60 and 100 percent greater than for aluminum, largely because of the materials ‘used to make the tools, the complexity of part shapes (e.g, compound curvatures), and the required part accuracy (toleran On the production side, the overall recurring cost per pound’ of composite is about twice that of aluminum and steel and 20 to 30 percent more than that of titanium. Average hours- per-pound factors for aluminum and graphite/epoxy (the most technologically mature of the composites) for the late 1980s time frame are as given in Table 34. With one exception, the ‘grephite/epoxy factors are all roughly twice the aluminum values. Moreover, the effect of the single exception (raw material costs, which vary by a factor of almost 5) is eubstantially mi gated by the fact that material costs typically constitute less than 10 percent of total recurring costs, ‘The data also indicate that recurring costs for both metals and composites can be expected to decrease between the late 1980s and the mid-1900s. However, the improvement in Table 33, MATERIALS, COSTS, AND TIME PERIODS Material Types Coat Blements Tins Periods ‘Metals Nonreeurting Tate 19605 ‘Alumina Bngineering Mid. 19900 ‘Akuminom-ehium Tooling ‘Tianiom Recurring Stee Engincering Composites ‘Tooling (Gapite/epoxy ‘Manufacturing labor Graphitefoismeleimide Quality auturance Graphite/thermoplastic Manufacturing material ‘Normalized for weight (1000 Ib) and quantity (100)a ‘Table 34 ALUMINUM AND GRAPHITE/EPOXY COST COMPARISON Graphite Cont Bement Avuminom —Epoay Ratio ‘Recurring engineering (h/1D) 10 ie 19 ‘Recurring tooling (he) 1s 3622 ‘Recurring manufacturing labor (h/Ib) el e218 ‘Recurring quality assurance (h/Ib) 7 a1 24 Recurring manufacturing material (Sb) 27 048 composite costs (approximately 20 percent) is projected to be roughly twice that of metal ccsts, ‘an outcome we believe is consistent with the “newness” of composites technology and its larger scope for improvement. However, these projections were developed during a more optimistic time period (1987). In the current environment, where programs are being stretched ou or canceled, companies may be less willing (or able) to make the capital investments necessary to achieve these projections. ASSESSING THE EFFECT OF STRUCTURAL MATERIAL COMPOSITION ON OVERALL AIRFRAME COST Structure-related costs are only a part of overall airframe costs and are not normally con- sidered in isolation. Consequently, we have outlined an approach that considers overall sir frame cost—that is, a method that takes into account not only the cost of the airframe struc- ture but also the cost of the airframe subsystems and final assembly/integration. It is suitable for use in a program's conceptual stage when litle detailed information is available. Reduced to its simplest form, the method applies weighted material indexes to baseline CERs that are assumed to be representative of all-aluminum aircraft. The CERs, which are Virtually identical to those documented in R-8255-AF, were derived from a database consisting of 18 post-1960 military aircraft.” Empty weights for the sample aircraft, range from under 10,000 Ib to over 80,000 Ib, while speeds range from 400 kn to over 1,900 kn. The recom- mended material indexes were calculated on the basis of the average values presented in Sec. IV. For each cost element, the late 1980s aluminum value was considered the baseline. ‘The method was applied to two hypothetical aircraft, one having an all-aluminum struc- ture weighing 13,000 Ib and the other having a structure that was 85 percent aluminum ané 45, percent graphite/epoxy and weighing 11,700 lb.° The result, which is based on late 1980s ‘manufacturing technology, shows that the substitution of graphite epoxy for half the aluminum in the structure will increase nonrecurring costs by about 8 percent and recurring costs by roughly 35 percent. Other material variations will produce different results. The A-6, AT, A-10,C-5, C141, Pe, Fels, P16, P16, P18, FoI, $3, and 7-9. We anumed that only 5200 Ib of raphite/epoxy would be required to replace 6800 Ib of aluminum, « 20 perent weight savingsFUTURE WORK No study of the type described here is ever complete. However, at some point, itis neces- sary to call a halt, present the results, and move on to something else. Nevertheless, there are several avenues that might profitably be pursued in subsequent efforts. Updste of Material Cost Factors . As aircraft with higher percentages of composite usage begin to enter service, additional design and production cost data will be accumulated. Presumably, these additional data could hhelp to reduce the large uncertainty now associated with the costing of composites. Addi- tionally, other more advanced materials are currently under development for the National Aerospace Plane (NASP),* including carbon-carbon, titanium-aluminide, and silicon- carbide/titanium. Consequently, at some point in the future, it would undoubtedly be benefi- cial to update the factors presented herein as well as to collect cost data for the more tachno- logically advanced NASP materi Consideration of O&S Factors If weight is saved, composite aircraft will have lower fuel costs than their metal caunter- parts. However, to date, the iilitary services have accumulated only very limited composite ‘maintenance and repair experience. Moreover, that experience has generally been obtained on aircraft that use small amounts of composites in secondary structures. Future aircraft promise to use much greater amounts of composites and in more critical areas, Further complicating the ise are the trends toward unitized design (larger, more integral, parts) and stealth enhancement, including tighter tolerance requirements and the incorporation of radar- absorbing materials/structure. Unfortunately, repair concepts, inspection techniques, sparing philosophy, and required personnel skills for the support of sophisticated composite airoraft are not well defined, At some point, however, the costs associated with maintaining altemative structural materials will need to be addressed. “Teaperatures on NASP's nove and lading edges will excoed 3000°F, while temperatures over most ofthe remin- ing surfice structure will be on the order of 1500°FAppendix A SAMPLE WORKSHEET This appendix presents the worksheet upon which the industry survey was based. The five principal sections are: I. Corporate material experience II Material usage (subassemblies where material used) IIL. Material technical characteristics IV. Cost data 'V. General information Nonproprictary qualitative data requested in worksheet sections J, Il, Il, and IV were incorporated into the narrative throughout the report. Simple averages (across companies) of the cost data are presented in Sec. IV.|. CORPORATE MATERIAL EXPERIENCE |. CORPORATE HISTORY Please list both previous and planned experience with major subassemblies using composite materials or advanced metal alloys. (Corporate brochures welcome.) PRODUCTION PROGRAMS “Type of Weight (by “Advanced ‘Quansgy ‘Yeu of Airerate Subassembly per Shipset Material Used Produced | First Application RESEARCH AND DEVELOPMENT PROGRAMS Type of ‘Advanced Quantiy Year of Alera Subasembly Weight (bo) Material Used | Gtapplicabley| Application “If for security easons the aircraft type or any item, cannot be designated, we would sill appreciate a response to the items that you ean complete It is not necessary to restit responses to military aizrat programs. Any programs that may provide wseful information for tn airerat program ae appropiateMATERIAL USAGE (subassemblies where material used) MATERIAL NAME: Check the boxes that characterize the material's use inthe subassemblies listed below, Place a “C” inthe box ifthe material ‘currently is being used in the subassembly. Place a “P” in the box if the material is projected to be used. Subassembly ‘TYPICAL USES OF COMPOSITES/METAL ALLOYS WITHIN MAJOR AIRFRAME SUBASSEMBLIES. ‘Nonstructural | Primary Structure | Secondary Structure Subassembly | Simple | Complex | Simple | Complex | Simple | Complex Empennage Nacelle Landing Gear ‘Axe there other factors you believe affect a part's complexity (e., shape, size, number of interfaces)?Wl, MATERIAL TECHNICAL CHARACTERISTICS Please complet the following questions for the production aircraft that provide the basis for the cost information on the sub- sequent pages. MATERIAL NAME: (Gly tens sigan) MATERIAL TECHNICAL CHARACTERISTICS. Density (9 ‘Strength (psi x 108): ‘Modulas (psi x 104: Strain (%): ‘Maximum operating temperature (°F): Alloys—composition: RAW MATERIAL FORM: C]compostre: tape _____Brosdgood- ___Other (peat speci) Cimeravatvoy: sheer Piste ___ittst. __Forged __Exinsion —— Powder ___Other (please specify) rypcoraincrarT [Jrighievanack CJtranspon C)pomter Clietiopier Dicommerciat Ccenerat aviton [loser (pease speity) Penrormance: C]subsonic [] supersonic pesicn concert: [Jrratitonst meta desion [composite design Cloter (pease specity) LEVEL OF AUTOMATION IN MANUFACTURING: []Manva FC] semi-automated [C] Fatty Automated ‘Again ifthe type of aircraft cannot be designated for security reasons, we stil would appreciate responses to any cuestions that you8 IV. COST DATA MATERIAL NAMI COST INFORMATION For the requested cost information, please assume a notional aircraft the following material breakdown: % Material named above % Aluminum % Titanium % Other Estimated airframe unit weight (b) ASSUMPTIONS FY 87S (as of 1 April 1987). Ratios relative to conventional aluminum airframe. Present (1987) Projection (1995) Low High Low High NONRECURRING RATIOS frames (eg, materi supply and demand level offactory automaton, processing technolegy, ‘design tochnology, He.) “Ratios relative toa conventions aluminum airframe (e.g, Al = 1.00). **The dolar valu of new capital equipment needed for the advanced material (e.,autoclaves, Xray test equipment, lament wind- ‘ng machines, tape lay-up machines, ee)IV. COST DATA (continued) MATERIAL NAME: ASSUMPTIONS (Cumulative average costs for 100 units. FY 87S (as of 1 April 1987), Present (1987) Projection (1995) RECURRING VALUES* Engineering labor hours ‘Manufacturing raw materi $b buysto-y tio weight sing factor (%)** labor hours — materia —fabrication abor hours fabrication labor improve- ‘ment curve slope () -—sssembly lor hours —1ssembly labor improve meat curve slope (%) _—integration bbor hours> “integration labor improve ‘ment curve slope (%2) Sastaining tooling labor —boursb Quality assurance —labor hoursb Principal reasons for diferences between tine frames (eg, material supply and dematd, level of factory automation, processing technology, ‘esign technology, et.) Lute “Estimated values for finished airframe. Please identity actual values with an “A” and estimated values with an “E” after the entry. “ARCO factor,V. GENERAL INFORMATION* How do tooling requirements change with the advent of composite materials and new alloys? What are the factors in the decision o use meta tools versus composite tools? ‘Ave there any new materials or manufacturing technologies not mentioned above that may substantially alter cost expocta- tions in the foture? Briefly describe them. ‘Are there componenis of total weapon life cycle cost that you believe may be substantially altered (up or down) by composite rmaterialynew alloys (¢.g., development, production, operations, and support costs)? How and why? ‘Asa result of composite materials or new metal alloys do you believe there will be a shift in the typical funding profile for an aircraft program? Additional commenis. “Please atach addtional sheets if necessary.Appendix B OTHER STUDIES OF COMPOSITE MATERIAL COSTS Because of the underlying differences in the design and manufacturing technology of com- posite materials and conventional metals, historical cost databases and models based on metals technology may not be appropriate indicators of the cost of future aircraft systems. We have ‘identified four studies conducted in the 1980s that addressed the cost of composite materials in aircraft structure. They concentrated on recurring manufacturing costs at the total airframe level, probably because of the limitations of the data. Only one has attempted to go below this level. The studies are: + Scott Woo and William Ditto, Cost Estimating Relationships for Army Helicopter Com- posite Airframes, Army Aviation Systems Command, 1983, ‘+ Major Gordon Kage II Estimating the Cost of Composite Material Airframes Using the RAND Corporation Development and Procurement Costs of Aircraft Parametric Model, Air Force Institute of Technology, 1983. ‘+ Walt Pingel, Aerospace Industry Composites Survey, Air Force Aeronautical Systems Division, 1986. © Kirk Hoy et al., Advanced Composites Airframe Cost Model, MCR, Ine., 1987. Generally speaking, the two studies performed in 1983 estimated that the substitution of ‘composite materials for metals would reduce cost. ‘The two later studies, however, that the substitution of composite materials would increase costs. Material categorization also varies between these studies; some simply separate metals from composites, and others collect information by the specific material types. ESTIMATING THE COST OF COMPOSITE MATERIAL AIRFRAMES USING THE RAND CORPORATION DEVELOPMENT AND PROCUREMENT COSTS OF AIRCRAFT PARAMETRIC MODEL (DAPCA II) This study was conducted in 1983 by Major Gordon Kage II as an Air Force Institute of ‘Technology master's thesis. His objective was to develop a series of indices that reflect the differences in manufacturing cost for composite materials and metals for use with the RAND Development and Procurement Cost of Aircraft Model version III. Indexes were developed for nonrecarring tooling hours, recurring manufacturing hours, and material dollars, ‘These indices were developed using the ICAM and FACET models. ICAM provided information for metal parts, and FACET provided information for composite parts. Both ‘models require detailed information and estimate at the part level. Descriptive data for 30 detailed parts ranging from complex fuselage structure to simple spars and panels were col- lected. ‘The descriptive data were input into the ICAM and FACET models and the resulting estimates were compared to develop the indices. The indices were all less than 1.0 when com- Posites are incorporated, contrary to the data presented in the RAND, ASD, and MC&R reports,COST ESTIMATING RELATIONSHIPS (CERS) FOR ARMY HELICOPTER COMPOSITE AIRFRAMES ‘This report was issued in 1983 by the U.S. Army Aviation Systems Command and was prepared by Scott Woo and William Ditto. A methodology was developed that uses indices to ‘adjust a helicopter recurring CER for advanced technology in the 1990s. The indices were based on 1980 RAND data for V/STOL sircraft, and 1960 and 1983 Grumman data used in their Modular Life Cycle Cost Model (MLCCM). These indices are to be used in conjunction with a recurring manufacturing CER developed for the Army Time Phased Parametric Life (Cycle Cost Model. Data are presented that cover historical helicopter recurring manufacturing costs and material distribution. ‘This report also presents data on historical helicopter recurring manufacturing costs derived from various program office baseline cost estimates, learning curve slopes for old and new technology helicopters, and airframe material composition. The helicopter data include: ‘AH-1G, UH-1H, UH-60A, OH-60A, OH-58A, CH-47C, CH-54B, AH-G4A, CH-47D, AHIP, and ACAP. ‘Woo and Ditto concluded that raw composite material costa were much higher than tredi- tional metal material costs, and composite material labor costs were lower than metal labor costs. According to their calculations, overall composite structure is less expensive than metal structure because the weight savings and lower labor cost more than offset the higher material costa, AEROSPACE INDUSTRY COMPOSITES SURVEY ‘This study was performed by Walt Pingel of the Air Force Aeronautical Systems Division in 1986. The two objectives of this study were to derive labor indices for selected composite materials and to establish general trends in composite material usage in the aerospace industry. Data were collected by means of a survey for manufacturing, tooling, and quality assurance, for cach of 11 material types. ‘The labor indices represent the ratio of the level of effort required for the advanced material relative to aluminum. ‘These ratios represent: a mixture of the non- recurring and recurring portions of the cost elements. Industry average values are reported. ‘Teble B.1 summarizes the results. These factors must be used in conjunction with another estimating methodology, although no method was suggested. ADVANCED COMPOSITES AIRFRAME COST MODEL ‘This study was prepared by Kirk Hoy et al., of MCR, Inc., and was originally conducted for the Navy in 1987 and subsequently updated for the Air Force in 1988. The objective of this study was to collect aircraft production data and develop CERs. Recurring manufacturing labor hours and material dollars for conventional metal and composite aircraft were collected. Aireraft included in the database are A-6E, A-7A/D, A-10A, AV-8B, F-4B, F-6E, F-14A, F-16A, F-1GA, F/A-18, F-101C, F-111A, B-IB, the Boeing 737, and the Lockheed L-1011. Data were available at several levels for these programs. Most had total production hours available for a small subset of the work breakdown structure. A few had a breakdown of fabrication and assembly hours for particular elements in the work breakdown structure (WBS). Cost estimating relationships were developed for airframe, fuselage, wing, wing/stabilizer, and empennage labor hours, and airframe material using regression analysis. Because of data‘Table BL ‘SUMMARY OF ASD MATERIALS SURVEY (Andastry average of ratio of level of effort for advanced ‘material relative to arin) (Cost Blement Materia Manafecturing Tooling Quality Assurance Boroa/epoxy 2a 18 rr Boron/polyimide 2a 131 139 Boron/aluminum 28 200 158 Kevlar 1.80 La 130 Carbon/eatbon. 3am 242 1 ‘Advanced sluminom 1st 121 18 Graphite /epoxy 169 136 189 Graphite/atuninum 228 188 13 Graphite nickel 253 200 ras (Graphite /tsmalimie 192 1s 148 Graphite/thermoplastic 2.00 175 186 limitations, several of these CERs were not recommended for use.’ For each subassembly ele- ment the CERs took the linear form. Inputs to each relationship are metal weight and com- Posite weight. As an example, the airframe estimating relationship is presented below: ‘Manufacturing labor hours for unit 100 = 6480.2 + 5.01 x (metal weight) + 14.46 x (composite weight) Some information on support labor costs* and learning curve slopes is also presented. “in perticular, the wing, empennage, and srframe material relationships, *Susaining engineering, sustaining tooling, and quality assurance.Appendix C COST ELEMENT DEFINITIONS Work breakdown structure categories included in the RAND airframe cost structure are shown ir Table C.1.! A matrix that maps contractor cost-data reporting (CCDR) categories? and relevant WBS categories into specific RAND airframe cost elements is provided in Table C2. NONRECURRING ENGINEERING In general, the nonrecurring engineering cost element encompasses the hours expended in the study, analysis, design, development, and evaluation of the basic airframe. More specifi- cally, it includes engineering for design studies and integration; for wind-tunnel models, crop model, mockups, and propulsion-system tests; for laboratory testing of components, subsys- ‘tems, and static and fatigue articles; for preparation and maintenance of drawings and provess ‘and materials specifications; and for reliability, Engineering hours not directly attributable to the airframe itself (those charged to ground-handling equipment, spares, and training equip- ment) are not included. Engineering hours expended as part of the tool and production- planning function are included with the cost element tooling; those expended os part of the flight-test planning and evaluation effort are included in the flight-test cost element, Materiel, Purchased parts, and test equipment required to accomplish the engineering function are assumed to be included in the fully burdened engineering labor rate. ‘RECURRING ENGINEERING ‘Recurring engineering covers such things as customer support/liaison, identifying ways to correct operationally revealed deficiencies, and suggesting possible system improvements. Although the break between nonrecurring and recurring engineering is subjective, the point of segregation is frequently at some specified point such as “design freeze” or “OO percent engineering drawing release.” NONRECURRING TOOLING ‘Nonrecurring tooling encompasses the cost of the initial set of tools and all duplicate tools produced to permit a specific rate of production. In general, tooling refers only to those tools designed for use on a specific program—i.e., assembly tools, dies, jigs, fixtures, master forms, gauges, handling equipment, load bars, work platforms, and test and checkout equip- ment. General purpose tools such as milling machines, presses, routers, and lathes (except for the cutting instruments) are considered capital equipment. If such equipment is owned by the contractor (much of it is government-owned), an allowance for depreciation is included in the "See MIL-STD-861, Work Breakdown Structure for Defense Materiel lems See AFLCP 00-15, Contractor Cost Data Reporting Sytem.Table C1 WBS CATEGORIES INCLUDED IN RAND AIRFRAME COST ELEMENTS. WBS Category Tchad? “ir vehicle ‘Airrame" . Yeo Propulsiga' No ‘Avionics No ‘Armament/weapon delivery” No ‘Trining No Peculiar aupport equipment No Systema tet and evaluation Development test, ‘Techalealeveaation Operational evaluation Mockupe ‘Tot facilities Other STE System /project management Data ‘Engineering/management data Yee TLS deta tech ordere and manuals) No Operationa/site activation No No No Initial spares and repair parts No fight control system), the air induction system, eear- tary exhaust, tho fuel contol system, Inlet control sytem, lighting gear (Ure, tubes, wheels, brakes, Hydraulics, ete), secondary power, furnishings (cargo, ‘paasenget, troop, etc), engine control, ‘ntrunente (fight navigation, engine, ete), environ: ‘mental contro, rocks, mounts, and. internyetm, ‘ables and distribution bores, et, that are inherent fo and inseparable from the ascemblod structure, Aynamie systems, and other equipment homogeneous 1a the airframe.” All efforts directly related 1 pro Dulsion, avionics, and armament are excluded. loatallation ‘of ‘the propulsion, avionics, and ‘rmanent subsystems ia ‘ecounted for in the ai frame category, the design integration “efforts Included in system/project' management; and some teeting is included in eystem test and evaluation. overhead account. Tooling hours include all effort expended in tool and production planning, design, fabrication, assembly, and installation as well as the programming and preparation of tapes for numerically controlled machine ‘The cost of the material used in the manufacture Of the dies, jigs, fixtures, etc. is assumed to be included in the fully burdened tooling labor rate.Table C.2 [RAND AIRFRAME COST ELEMENTS (GODR Categories ‘Manufacturing Manufscturing Quality Engineering Tooling ‘Labor Material Control WBS Category NR OR NR R NR OR NR RNR R ‘ir vehide ‘Aicfame NBngr REngr NTool RTool DS Labe DS Mat QC QC Sytams test and evalation Development tests Neo — DS — DS — DS — DS — ‘Technial erabation i ee Opertiona evaluation ro PR RS Meckups NEnr = DS ee ey re ‘Tent elton Neg — DS — DS — DS — DS — Other STAE, Nes — DS — DS — DS — DS — Syutem/project management NEngr REngr DS RTool DS Labr DS Mal DS QC Engineering/manegement data NEngr REngt ~~ — Sn = gemecaring B — peng: NE ~ nonrecurring tooling, Roo! Eula DS = Gveoent sappors PT "fight ane Qe ually contra RECURRING TOOLING Recurring tooling includes tool maintenance, modification, rework, and replacement. MANUFACTURING LABOR Manufacturing labor is all the direct labor necessary to machine, process, fabricate, and assemble the major structure of an aircraft and to install purchased parts and equipment, engines, avionics, and ordnance items, whether contractor-furnished or government-furnished. ‘Manufacturing manhours include the labor component of off-site manufactured assemblies and effort on those parts that because of their configuration or weight characteristics are design- controlled for the basic aircraft. These parts normally represent considerable proportions of airframe weight and of the manufacturing effort and are included regardless of their method of acquisition. Such parts specifically include actuating hydraulic cylinders, radomes, canopies, ducts, pastenger and crew seats, and fixed external tanks. Manhours required to fabricate pur- chased parts and materials are excluded from the cost element. Nonrecurring labor undertaken in support of engineering during the development phase is included in the development support cost element. MANUFACTURING MATERIAL ‘Manufacturing material includes raw and semifabricated materials plus purchased parts (standard hardware items such as electrical fittings, valves, and hydraulic fixtures) used in the manufacture of the airframe. This category also includes purchased equipment (i., motors, generators, batteries, landing gear, air conditioning equipment, instruments, and hydraulic andPneumatic pumps), whether procured by the contractor or furnished by the government. ‘Where such equipment is designed specifically for a particular aircraft, itis considered as sub- contmcted, not as purchased equipment, and is therefore included in the manufacturing labor cost element. Nonrecurring material used in support of engineering during the development phase is included in the development-support cost element. DEVELOPMENT SUPPORT Development support is the nonrecurring manufacturing effort undertaken in support of ‘engineering during the development phase of an aircraft program. It is intended to include the ‘manhours and material required to produce mockups, models, test parts, static and fatigue test items, and other hardware items (excluding complete flight-test aircraft) needed for airframe development. FLIGHT TEST Flight test includes all costs incurred by the contractor in the conduct of flight testing excep: production of the test aircraft. Engineering planning, data reduction, manufacturing support, instrumentation, all other materials, fuel and oil, pilot’s pay, facilities, rental, and insurence costs are included.. Flight-test costs incurred by the Air Force, Army, or Navy are excluded. QUALITY CONTROL Quality control refers to the hours expended to ensure that prescribed standards are met. It inciudes such tasks as receiving inspection; in-process and final inspection of tools, parts, subaseemblies, and complete assemblies; and reliability testing and failure-report reviewing. ‘The preparation of reports relating to these tasks is considered direct quality-control effort.Appendix D DEFINITION OF TERMS ‘The following definitions have been taken from New Structural Materials Technologies: Opportunities for the Use of Advanced Ceramic nd DOD/NASA Advanced Composites Design Guide, Volumes 1-4, Flight Dynamics 1s ‘and Composites, QTA-TM-E-32, September Laboratory, Air Force Wright Aeronautical Laboratories, 1983. adhesive advanced filaments alloy anelasticity autoclave autoclave molding brittle fracture carbon/graphite A substance capable of holding two materials together by surface ‘attachment. In the Guide, the term is used specifically to designate structural adhesives, which produce attachments capable of transmit- ting substantial structural loads. Continuous filaments made from high strength, high modulus materials for use as a constituent of advanced composites. ‘A material having metallic properties and consisting of two or more elements. A characteristic exhibited by certain materials in which strain is a function of both stress and time, such that while no permanent deformations are involved, a finite time is required to establish equilibrium between stress and strain in both the loading and unloading directions. Showing different physical or mechanical properties in different directions. Lightweight polyaromatic amide fibers having excellent high tem- perature, flame resistance, and electrical properties. These fibers are used as high strength reinforcement in composites. it is sometimes considered to be the ratio of the load direction dimension to the transverse dimension. Also, in fiber micro-mechanics, it is referred to as the ratio of length to diameter. ‘A closed vessel for producing an environment of fluid pressure, with ‘or without heat, to an enclosed object while undergoing a chemical reaction or other operation. ‘A process similar to the pressure bag technique. The layup is, covered by a pressure bag, and the entire assembly is placed in en autoclave capable of providing heat and pressure for curing the part. ‘The pressure bag is normally vented to the outside. A break in a brittle material due to the propagation of cracks origi- nating at flaws. ‘These fibers, which are the dominant reinforcement in “advanced” ‘composites, are produced by pryrolysis of an organic precureor—e., 87ceramic chemically bonded ceramics ‘compressive stress consolidation of parts ‘continuous fiber ‘continuous filament yarn coupling agent crazing creep cross-linking croseply exystel cure stress, polyacrylonitrile (PAN), or petroleum pitch—in an inert. atmosphere. Depending on the process temperature, fibers having high strength or high elastic modulus may be produced. An inorganic, nonmetallic solid. Used here to distinguish advanced cements and coneretes, which are consolidated through chemical reactions at ambient temperatures (generally involving uptake of water), from high-performance ceram- ies, such as silicon nitride and silicon carbide, which are densified at high temperatures. ‘A stress that causes an elastic body to shorten in the direction of the applied force. Integration of formerly discrete parts into a single part that encom- passes several functions, a key advantage of engineered materials such as ceramics and composites. A reinforcing fiber in a composite that has a length comparable to the dimensions of the structure. Yam formed by twisting two or more continuous filaments into a single, continuous strand. ‘Any chemical substance designed to react with both the reinforce- ‘ment and matrix phases of a composite material to form or promote a stronger bond at the interface. Coupling agents are applied to the reinforcement phase from aqueous or organic solution, from the gas phase, or added to the matrix as an integral blend, ‘The development of a multitude of very fine cracks in the matrix material. A time-dependent strain ofa solid, caused by stress. ‘The formation of chemical bonds between formerly separate polymer chains. Any filamentary laminate that is not uniaxial. Same as angleply. In some references, the term crossply is used to designate only those laminates in which the laminae are at right angles to one another, while the term angleply is used for all others. In the Guide, the two terms are used synonymously. With the advent of the Guide lam- inate orientation code, the reservation of a separate terminology for only one of several basic orientations is unwarranted. ‘A homogeneous solid in which the atoms or molecules are arranged in a regularly repeating pattern. To change the properties of a thermosetting resin irreversibly by chemical reaction—i.e,, condensation, ring closure, or addition Cure may be accomplished by addition of curing (cross-linking) agents, with or without catalyst, and with or without heat. ‘A residual internal stress produced during the curing cycle of com- posite structures. Normally, these stresses originate when different components of a wet layup have different thermal coefficients of expansion.debond deformation, plastic deformation delamination dielectric disbond extrusion fabric failure fiber fiber content fiber count A deliberate separation of a bonded joint or interface, usually for repair or rework purposes. [Any alteration of shape or dimensions of a body caused by stresses, ‘thermal expansion or contraction, chemical or metallurgical transfor- ‘ations, or shrinkage and expansions due to moisture change. ‘The separation of the layers of material in a laminate. ‘This may be local or may cover a large area of the laminate, It may occur at any time in the cure oF subsequent life of the laminate and may arise from a wide variety of causes. ‘A material that is an electrical insulator or in which an electric field can be sustained with a minimum dissipation of power. ‘An area within a bonded interface between two adherends in which ‘an adhesion failure or separation has occurred. It may occur at any time during the life of the structure and may arise from a wide variety of causes. Also, colloquially, an area of separation between ‘two laminae in a finished laminate (where “delamination” is pre- ferred). ‘The ease in which a material may be formed into a complex con- toured shape without undesirable features (i. folds, wrinkles, ec.) ‘The ability of a material to be plastically deformed by elongation without fracture. A borosilicate glass most used for glass fibers in reinforced plastics. ‘The property whereby a solid material deforms under stress but recovers its original configuration when the stress is removed. ‘An individual warp yarn, thread, monofilament, ot roving. ‘A process in which a hot or cold semisoft solid material, such as metal or plastic, is forced through the orifice of a die to produce continuously formed piece in the shape of the desired product. ‘A generic material construction consisting of interlaced yarns or fibers, usually @ planar structure. Specifically, a cloth woven in an established weave pattem from advanced fiber yarns and used as the fibrous constituent in an advanced composite lamina. In a fabric lamina, the warp direction is considered the longitudinal (L) ditec- tion, analogous to the filament direction in a filamentary lamina. Collapse, breakage, or bending of a structure or structural element ‘such that it can no longer fulfill its purpose. Failure of a material by cracking resulting from repeated or cyclic stress. ‘A single homogeneous strand of material, essentially one- dimensional in the macro-behavior sense, used as a principal consti- tuent in advanced composites because of its high axial strength and modulus. ‘The amount of fiber present in a composite, usually expressed a a percentage volume fraction or weight fraction of the composite. ‘The number of fibers per unit width of ply present in a specified see- tion of a composite.100 fiber direction fiber system filamentary composites filament winding filler finish flame-sprayed tape flash flexure fracture stress fugitive binder ‘The orientation or alignment of the longitudinal axis of the fiber with respect to a stated reference axis. The type and arrangement of fibrous material that constituies the fiber constituent of an advanced composite. Examples of fiber sye- ‘tems are collimated filaments or filament yarns, woven fabric, ran- domly oriented short-fiber ribbons, random fiber mats, whiskers, ec, A variety of fibers characterized by extreme length, such that there are normally no filament ends within a part except at geometric discontinuities. Filaments are used in filamentary composites and are also used in filament winding processes, which require long con- tinuous strands, A major form of advanced composites in which the fiber cons‘ituent consists of continuous filaments. Specifically, @ filamentary compos- ite is a laminate composed of several laminae, each of which consists of a nonwoven, parallel, uniaxial, planar array of filaments (or fila- ‘ment yarns) imbedded in the selected matrix material, Individual laminae are directionally oriented and combined into specific mul- tiaxial laminates for application to specific envelopes of strength and stiffness requirements. ‘An automated process in which continuous filament (or tape) is treated with resin and wound on a removable mandrel in a pattern, Yam oriented at right angles to the warp in a woven fabric. ‘A second material added to a material to alter its physical, mechani- cal, thermal, or electrical properties. Sometimes used specifically to ‘mean particulate additives. A material with which filaments are treated containing a coupling ‘agent to improve the bond between the filament surface and the resin matrix in a composite material. In addition, finishes often con- tain ingredients that provide lubricity to the filament surface, preventing abrasive damage during handling, and a binder that pro- ‘motes strand integrity and facilitates packing of the filaments. A form of metal matrix preply in which the fiber system is held in place on a foil sheet of matrix alloy by a metallic flame-spray deposit. Bach flame-sprayed preply is usually combined in the layup stack with a metal cover foil or additional metal powder to insure complete encapsulation of the fibers. During consolidation, all the metallic constituents are coalesced into a homogeneous matri ‘Excess material that forms at the parting line of a mold or die, or is extruded from a closed mold. Any bending deformation of an elastic body in which the points orig- inally lying on any straight line are displaced to form a plane curve. ‘The minimum stress that will cause fracture, also known as frecture strength. ‘A resinous material used in the fabrication of metal matrix green tape (which see) preplies to hold the fiber system in place on the ‘metallic foil sheet during shipping, storage, handling, and layup. During laminate consolidation, the fugitive binder decomposes and the products completely vaporize.gel coat lass slass-ceramic ‘green tape hardness heat treatment heterogeneous homogeneous horizontal shear hot isostatic pressing hot pressing hybrid impact strength inclusion 101 A quick-setting resin used in molding processes to provide an improved surface for the composite; itis the first resin applied to the mold after the mold-release agent. ‘A state of matter that is amorphous or disordered like liquid in structure, hence capable of continuous composition variation and lacking a true melting point, but softening gradually with increasing ‘temperature. : Solid material, partly crystalline and partly glassy, formed by the controlled crystallization of certain glasses. A form of metal matrix preply in which the fiber system is held in place on a foil sheet of matrix alloy by a resinous fugitive binder (which see). Bach green tape preply is usually combined in the layup stack with a metal cover foil to insure complete encapsulation of the fibers. During consolidation, the fugitive binder decomposes and the products completely vaporize. Resistance of a material to indentation, scratching, abrasion, or cut~ ting. Heating and cooling of a material to obtain desired properties or conditions. Descriptive term for a material consisting of dissimilar constituents separately identifiable; a medium consisting of regions of unlike properties separated by internal boundaries. (Note that all nonho- ‘mogeneous materials are not necessarily heterogeneous.) ‘A technique for recording and later reconstructing the amplitude end phase distributions of a wave disturbance, Descriptive term for a material of uniform composition throughout; a ‘medium thet has no internal physical boundaries; a material whose properties are constant at every point—with respect to spatial coor- dinates (but not necessarily with respect to directional coordinates). ‘The term “horizontal shear” is not approved nomenclature for the Guide, but it is included here for information only because of occa sional encounters in the literature. Same as “interlaminar shear.” ‘A form of ceramic, powder metallurgical, or ingot metallurgical form- ing or compaction process in which the mold is flexible and pressure is applied hydrostatically or pneumatically from all sides. Forming a metal powder compact or a ceramic shape by applying pressure and heat simultaneously at temperatures high enough for sintering to occur. ‘A composite laminate composed of laminae of two or more composite ‘material systems. Ability of a material to resist shock loading. A physical and mechanical discontinuity occurring within a material or part, usually consisting of solid, encapsulated foreign material. Inclusions are often capable of transmitting some structural stresses and energy fields, but in a noticeably different (and less desirable) degree from the parent materialwe injection molding integral composite structure interlaminar interleminar shear internal stress, residual stress interphase, interface intralaminar isotropic lamina laminate laminate orientation layup load mandrel metal Forming metal, plastic, or ceramic shapes by injecting « measured ‘quantity of the material into shaped molds. Composite structure in which several structural elements that would conventionally be assembled by bonding or mechanical fasteners after separate fabrication are instead laid up and cured as a single, complex, continuous structure—e.g., spars, ribs, and one stiffened cover of a wing box fabricated as a single integral part. The term is sometimes applied more loosely to any composite structure not assembled by mechanical fasteners. Descriptive term pertaining to some object (e.., voids), event (e.g., fracture), oF potential field (eg, shear stress) referenced as existing or occurring between two or more adjacent laminae. Shearing force tending to produce a relative displacement between ‘two laminae in a laminate along the plane of their interface. A stress system within a solid (eg, thermal stresses resulting from rapid cooling from a high temperature) that is not dependent on external forces. ‘The boundary layer between the matrix and a fiber, whisker, or par- ticle in a composite. Descriptive term pertaining to some object (e4., voids), event (e.g., temperature gradient) existing single lamina without reference to any adjacent Having uniform properties in all directions. The measured proper- ties of an isotropic material are independent of the axis of testing. ‘A single ply or layer in a laminate made of a series of layers. A product made by bonding together two or more layers or laminae of material or materials. ‘The configuration of a crossplied composite laminate with regard to ‘the angles of crossplying, the number of laminae at each angle, and the exact sequence of the lamina layup. A process for fabricating composite structures involving placement of sequential layers of matrix-impregnated fibers on a mold surface. ‘The weight that is supported by a structure, or mechanical force that is applied to a body. ‘A form mixture or male mold used for the base in the production of ‘a part by layup or filament winding. ‘The composite constituent that binds the reinforcement together and transmits loads between reinforcing fibers. ‘An opaque material with good electrical and thermal conductivities, ductility, and reflectivity; properties are related to the structure in which the positively charged nuclei are bonded through a field of mobile electrons that surrounds them, forming a close-packed struc- ture.microstructure modulus of elasticity ‘monolithic nnear-net-shape nondestructive testing, evaluation orthotropic phase plasticity polymer pore, porosity prepreg proof test pultrasion radiography radome refractory Sela ‘The internal structure of a solid viewed on a distance scale on the order of micrometers, The microstructure is controlled by processing and determines the performance characteristics of the structure. ‘A parameter characterizing the stiffness of a material, or its resis- tance to deformation under stress. For example, steel has a fairly high modulus, while Jello has a low modulus. Constructed from a single type of material ‘The original formation of a part to a shape that is as close to the desired final shape as possible, requiring as few finishing operations as possible. Any testing method that does not involve damaging or destroying the ‘test sample; includes use of X-rays, ultrasonics, magnetic flux, etc. Showing symmetry of properties in two orthogonal planes but dif- ferent properties in the third plane. ‘A region of a material that is physically distinct and is homogeneous in chemical composition. ‘The property of a solid body whereby it undergoes a permanent ‘change in shape or size when subjected to a stress exceeding a partic- ular value, called the yield value. Substance made of giant molecules formed by the union of simple molecules (monomers); for example, polymerization of ethylene forms a polyethylene chain. Flaw involving unfilled space inside a material that frequently limits the material strength. Fiber reinforcement form (usually tape, fabric, or broadgoods) that hhas been preimpregnated with a liquid thermosetting resin and cured to a viscous second stage. Thermoplastic prepregs are also available. A predetermined test load, greater than the intended service load, to which a specimen is subjected before acceptance for use. ‘A fabrication process that uses guides, shape dies, and heat to pro- duce long parts with constant cross sections, which are then auto- matically cut to desired lengths. The process consists of pulling dry fibers through a resin bath and then through steel dies that define the shape of the part and control the amount of resin in it. Heat ‘must be supplied to cure the part. ‘The technique of producing a photographic image of an opaque specimen by transmitting a beam of X-rays or gamma rays through it onto an adjacent photographic film; the transmitted intensity reflects variations in thickness, density, and chemical composition of the specimen. A strong, thin shell made from a dielectric material, used to house a radar antenna. Capable of enduring high-temperature conditions. ‘A magnesia-alumine-silicate glass that provides very high tensile strength fiber reinforcement. Often regarded as the reinforcement fiber dividing “advanced” composites from reinforced plastics.106 shearing stress slip, slurry stress structural materials or assembly substrate tack tensile strength, ulti- ‘mate tensile strength thermal conductivity ‘thermoplastic resin thermosetting resin tribolory ultrasonic testing unibody unitized design A stress in which the material on one side of a surface pushes on the material on the other side of the surface with a force parallel to the surface. ‘Method for the consolidation and densification of metal or ceramic powders by heating without melting. ‘A forming process in the. manufacture of shaped refractories, cer- mets, and other materials in which sip is poured into porous plaster molds. Slip or slurry is a suspension of ceramic particles in water with a cresmy consistency. Change in length of an object in response to an applied stress, divided by undistorted length. ‘The force acting across a unit area in a solid material in resisting the ‘separation, compacting, or sliding that tends to be induced by exter- nal forces. Those parts of # system that support most of the loading on the whole system. Base surface on which a material adheres, for example a surface to be coated. Stickiness of a prepreg. ‘The maximum stress a material subjected to a stretching losd can withstand without breaking. ‘The rate of heat flow under steady conditions through unit area per unit temperature in the direction perpendicular to the area—the abil- ity of a material to conduct heat. ‘A material containing discrete polymer molecules that will repeatedly soften when heated and harden when cooled; for example, polyethylenes, vinyls, nylons, and fluorocarbon: ‘A matrix material initially having low viscosity that hardens because of the formation of chemical bonds between polymer chain. Once ccured, the material cannot be melted or remolded without destroying its original characteristics; examples are epoxies, phenolics, and polyimides. A parameter measuring the amount of energy required to fracture a ‘material in the presence of flaws. ‘The study of the phenomena and mechanisms of friction, lubrication, and wear of surfaces in relative motion. A nondestructive test method that employs high-frequency mechani- cal vibration energy to detect and locate structural discontinuities or differences and to measure thickness of a variety of materials. Integrated structure containing the chassis as well as elements of the body of an automobile. A unitized design refers to the philosophy of designing and building a structure in one piece, thereby eliminating the need for fasteners. ‘Traditionally these structures were manufactured in separate piecesviscoelasticity wettability yield strength 105 ‘and then mechanically aligned, shimmed, and fastened. The unitiza- tion of the piece then can save much of the alignment and assembly work and makes the piece less likely to contain undetected flaws associated with drilling holes, et. Property of a material that is viscous but also exhibits certain elastic properties such as the ability to store energy of deformation, and in which the application of stress gives rise to a strain that approaches its equilibrium value slowiy. Deterioration of a surface due to material removal caused by relative ‘motion between it and another material ‘The ability of any solid surface to be wetted when in contact with a liquid. ‘The lowest stress at which a material undergoes plastic deformation. Below this stress, the material is elastic.BIBLIOGRAPHY ‘Air Force Materials Laboratory, Advanced Composites Fabrication Guide, Air Force Materials Laboratory, Air Force Wright Aeronautical Laboratories, December 1976. Air Force Materials Laboratory, Air Force Technical Objective Document: FY 87, Air Force ‘Materials Laboratory, Air Force Wright Aeronautical Laboratories, AFWAL-TR-85-4000, December 1985. Bates, William F., et al, Preliminary Design Concepts for Metal Matrix Composite Aircraft Structures, AFWAL-'TR-85-123, Lockheed-Georgia, April 1988. Birkler, J. L,, J. B. Garfinkle, and K. E. Marks, Development and Production Cost Estimating Relationships for Aircraft Turbine Engines, RAND, N-1882-AP, October 1982. Booing Military Airplane Company, Seattle, Damage Tolerance of Composites, interim report no. 8 for period 1 March 1986-81 August 1986, Bray, Donald, Reinforced Alumina Composites, Paper given to Fabricating Composites ‘88 Conference, Aluminum Company of America, September 1988. (Clark, Joel P., and Merton C. Flemings, “Advanced Materials and the Economy,” Seieniifie ‘Anaerican, October 1988. Dastin, Samuel, Advanced Composite Materials and Structures, Seminar briefing charts, August 1986. Decamp, Major Ronald, and Major David Johnson, A Study of Composite Materials Effects on Aircraft Costs, Project 81-009-DAY, 11 January 1982. DoD/NASA. Advanced Composites Design Guide, Vols. 1-4, Flight Dynamics Laboratory, Air Force Wright Aeronautical Laboratories, Dayton, Ohio, 1983. Dostal, Cyril A., et al. (eds.), Engineered Materials Handbook, Vol. 1, Composites, ASM Inter- national Committee, Metals Park, Ohio, 1988. Dulis, E. J., “Near Net Shape Process Using HIP of Alloy Powder Particles,” Net Shape Tech- nology in Aeronautical Structures, Vol. III, National Academy Press, Washington, DC, 1986. Hoss, R. W., and H. P. Romanoff, Aircraft Airframe Cost Estimating Relationships: Study “Approach and Conclusions, RAND, R-3255-AF, December 1987. Hoy, Kirk, ASD Advanced Materials Cost Research, Final Project Briefing, MC & R, 9 March 1988, Hoy, Kirk, Scott M. Allard, Bruce J. McKinney, and T. Bemard Fox, Advanced Composite Air- frame Cost Model, MC&R, TR-8692-1, 27 September 1987. Huttrop, Marie, “Cost Reduction Through Design for Automation,” McDonnell Douglas Cor- poration, in J. David Oetting et al., Proceedings of the Seventh Conference on Fibrous Composites in Structural Design, Air Force Wright Aeronautical Laboraties, AFWAL- TR-86-3094, June 1985, Kage, Major Gordon D., Estimating the Cost of Composite Material Airframes Using the RAND Corzoration Development and Procurement Costs of Aircraft Parametric Model (DAPCA ID, Aix Force Institute of Technology, Wright-Patterson. AFB, September 1983. Klein, Allen J., “Composites: Moving Toward Automation,” Advanced Composites, March/ April 1987, pp. 93-99. LITV Aerospace and Defense Company, Briefing Charts: Carbon-Carbon Cost Data, LTV Aerespace and Defense Company, August 21, 1986. 1108 Mahon, J., “Advanced Materials Processing Concepts,” AFWAL: ‘TR-84-4063, Grumman ‘Aerospace Corporation, April 1984. Mar, J. W., “Fracture, Longevity, and Damage Tolerance of Graphite Epoxy Filamentary Com- Posite Material,” ATAA, J. Aireraft, Vol. 21, No. 1, January 1984, Marx, W., “Composites Assembly,” Near Net Shape Technology in Aeronautical Structures, Vol. W,, National Academy Press, Washington, D.C., 1988. Materials Pace ATF Design, Aerospace America, April 1987. Maugh, Tomas H., TI, “Composites—the Lightweight Champs of Aircraft Industry” Los “Angeles Times, March 9, 1987, Part I, p. 3. Meade, L. E., Briefing to Composites Supportability Workshop, September 16, 1988, Lockheed ‘Aeronautical Systems Company. Meade, L. E., DoD/NASA Structural Composites Fabrication Guide, 34 ed., Vol. I~Material, Fabrication, Tooling, Quality Assurance, Cost, Producibility, Equipment, and Experience, DTIC Report, AFWAL-TR-85-4107, 1982. National Research Council, Net Shape Technology in Aerospace Structures: Workshop Proceed: ings, September 9-12, 1985, Volumes 1-4, 1986, National Research Council, The Place for Thermoplastic Composites in Structural Components, National Materials Advisory Board, Commission on Engineering and Technical Systems, NMAB-434, Washington, D.C., December 1, 1987. Nicolai, Leland M., Fundamentals of Aircraft Design, METS, Inc., Xenia, Ohio, 1975. Office of Technology Assessment, New Structural Materials Technologies: Opportunities for the Use of Advanced Ceramics and Composites, Washington, D.C., September 1986. Office of Technology Assessment, Advanced Materials by Design, Washington, D.C, June 1988, Performance Materials, Aerospace and Defense Group, McGraw-Hill, Inc., New York, May 4, 1987, p. 2. Performance Materials, Aerospace and Defense Group, McGraw-Hill, Ine., New York, February 18, 1990, p. 1. Pingel, Walter D., Aerospace Industry Composites Survey, Aeronautical Systems Division, Day- ton, Ohio, 1987. Pitstic, Jeffrey L., ASD/Industry Cost Workshop Wrap Rate Study, Cost Division, Aeronauti- cal Systems Division, United States Air Force, Proprietary, 1985. Rigdon, Mark A., Definition Study for the Advanced Tactical Fighter Program Office, The Ans. Iytie Sciences Corporation, Dayton, Ohio, TR-5306-9-1, April 20, 1987. Spinks, D. F,, “Composite Assembly Proposals,” Net Shape Technology in Aerospace Structures: Vol IV, Future Composite Manufacturing Technology, National Academy Press, Washing ton, D.C., 1986. Stekler, H. 0. “Technological Change in the Military Aircraft Industry,” Technological Fore- casting and Social Change, July 1985. United States Department of the Interior, Mineral Commodity Profiles, Bureau of Mines, Pre- print bulletin 675, Washington, D.C., 1985 Edition. Vaccari, John A., “Automating Composites Fabrication,” American Machinist and Automated ‘Manufacturing, November 1987, pp. 88-98. Warwick, Graham, “Aluminum Fights Beck,” Flight International, 11 April 1987, p. 11. Weathersbee, James A., Cost Estimating Techniques for Composite Materials, Naval Air Sys- tems Command, Washington, D.C., 1983.Woo, Scott $.Y. and William J. Ditto, Cost Estimating Relationships (CERs) for Army Heli- ‘copter Composite Airframes, U.S. Army Aviation Systems Command, December 1983. ‘Zwveben, Carl, Michael Bader, and Alfred Yue, Introduction to Composite Materials Class Brief- ing Charts, UCLA Extension, September 15 to 16, 1987.RAND/R-4016-AF