Modern Composite Aircraft Technology Ralph W. Kiger

By Hans D. Neubert 6051 Prado St.

13731 Magnolia Ave. Garden Grove, CA 92640

and

Anaheim, CA 92807

(All Rights Reserved by the Authors) PART IV. QUALITY ASSURANCE, REPAIR AND ENVIRONMENTAL CONSIDERATIONS

INTRODUCTION

Typically, general aviation aircraft today have service lifetimes in excess of 30 years. In homebuilt aircraft, one presumes that similar service is to be expected. Since most builders have expended considerable labor and financial resources to get his or her bird into the air, the premise that the ship has resale value at some later date is not intuitively obvious. The point we are making here is that the long term durability and in-service performance is controlled to some extent during the design phase. Serviceability, then, is a concept that must be given early consideration, rather than as an afterthought. The utilization of composite materials in primary aircraft structure is relatively new, and therefore very little background data is available to draw upon in order to project 30 years ahead. We are not totally clear on what the problems are going to be, let alone the answers. Based on what is known from the sailplane, power boat, building materials industries, we would like to give you some background on the long term durability aspects. One thing is clear, and that is that the airframe is likely to receive some damage during its useful life. A section in this Part IV deals with the repair of composite structure. Since the builder not only creates the assemblies comprising the aircraft, but also the basic structural material, he should have some understanding of quality control and nondestructive inspection. Finally, the environmental and some other mute (and still somewhat controversial) aspects are reviewed. INSPECTION OF MATERIALS AND ASSEMBLIES

Raw Materials: Once fabricated, it is exceedingly difficult to determine the quality of composite construction parts. Carefully controlled systematic inspection of raw materials must be made in accordance with rigid materials specifications, and the fabrication must be controlled by strict adherence to rigid process specifications. The responsibility for assuring the quality of raw materials for homebuilt applications rests with the material supplier as specified by the designer. Proof of fabrication quality becomes the responsibility of the homebuilder and is accomplished by conforming to the manufacturing processes specified by the designer of the aircraft. Oftentimes the only way of determining conformance to these process specifications is through use of nondestructive inspection methods. Nondestructive tests are an essential component of production processes, as well as end product inspection. If necessary, they can be applied to all processes, components and assemblies. Most nondestructive indications are quali-

tative, not quantitative, and their interpretation involves judgment based on considerable experience. The relations between discontinuities and the performance capabilities of materials and systems are critically dependent upon the intended service conditions and operating environments. The significance of any abnormality should be initially verified by destructive means by the designer to assist the homebuilder in later evaluation of similar indications. Cores: Specifications normally designate acceptable density ranges and minimum strength properties for core materials. The acceptance of material for use therefore depends on careful inspection for weight and strength consistency. Natural core materials, such as balsa and mahogany, vary over a wider weight range than synthetic core materials, making inspection for density conformance of prime consideration. Tensile, compressive, shear or other tests are sometimes made as part of the inspection. Their purpose is either to insure proper strength requirements or as a check on other properties, such as the proper cure of the resin in a resin-impregnated glass fabric or paper honeycomb core. Some of the common characteristics of typical core materials that require investigation by inspection are presented in Table 1. Facings: Materials employed for facings of fiberglass sandwich components present no unusual inspection problem. Before impregnation, glass fabric must be clean, properly treated and of a definite uniform weave pattern. TABLE 1 — PRINCIPAL CORE MATERIAL CHARACTERISTICS THAT REQUIRE INSPECTION

Core Material

Characteristics (in order of importance)

Balsa and Density, defects*, slope of grain, Mahogany moisture content. Paper Honeycomb Cure, configuration, bonding,

density. Metal Honeycomb Bonding, alloy, configuration, perforation, density. Glass Fabric Configuration, bonding, resin conHoneycomb tent, density, dielectric properties. Foamed-in-Place Foaming characteristics, uniformity of foam, density. *Common defects are knots, rot, wormholes, wane, checks and splits. SPORT AVIATION 55

Adhesives and Resins: Adhesives must first be evaluated for this suitability and performance in the type of application for which their use is intended. Some specifications require that bond tests be made at specific intervals to determine that there has not been any deterioration of the adhesive. Once a particular adhesive has been selected, inspection must be made to determine that the various batches as received are the same quality as the original samples, and that the adhesives are not used when the quality of their performance has been reduced by overaging or improper storage. Tests of bond strength have been widely used as a means of originally selecting the adhesive, determining the conditions under which the adhesive can be used and as a means of inspecting the uniformity of the adhesive. Other tests, such as physical appearance, pH, viscosity, specific gravity and solids content, have also been used to aid in inspecting the uniformity and storage characteristics of adhesives. Inspection of parts for size, prior to bonding, is extremely important in preventing voids in the finished product. Tolerances on bonding fixtures are equally important for the same reason. Completed Components: Sandwich parts are inspected for conformance to dimension, weight, configuration, uniformity and strength requirements of the applicable specification. The relative importance and tolerances allowable for each characteristic depend upon the application. For example, antenna housings require panels of uniform and exact thickness, while structural panels require, primarily, certification of adequate bond strength. Secondary structural parts are less critical, but must be of proper size and shape.

Detail Inspection: Detail inspection should include the following: * Thickness of laminate in representative areas * Dimensional check of periphery * Dimensional check of moldline * Visual check for defects Detail inspection, although simple and routine, can provide valuable information, as well as indications that significant changes have occurred in the fabrication process. Assuming that satisfactory parts have been produced on a given tool and baseline physical data established, small changes in laminate thickness might indicate improper resin flow and resulting degradation of mechanical properties. Physically marking such conditions would result in closer analysis by other inspection techniques to delineate defects that might otherwise go unobserved in normal evaluation. DELAMINATION

FIGURE 1 DIAGRAM OF TYPICAL DELAMINATION DEFECT (MAGNIFIED)

DEBOND AREA

FACE SHEE1

Structural parts must be critically inspected for

areas of questionable bond between facings and core. Areas having no bond are usually readily detectable by several of the common inspection methods, but areas having merely subnormal bond strength are exceedingly difficult to locate by inspection of nondestructive test methods. Testing of production samples or test coupons of various parts to destruction can be useful in providing information on manufacturing techniques and consistency, process control and structural integrity. Defect Types: The majority of defects that occur will fall into one of the following categories. A definition of the defect is given, along with most of its com-

mon causes. Delamination — A delamination is the separation of adjacent layers within a multilayer structure. It may result from gas pockets forming and being entrapped during the cure cycle, from the resin being too far advanced or partially cured prior to layup, or from contamination (see Figure 1). Debonds — A debond is the lack of a bond in a joint area between two separate details. It may result from improper fit of the details, failure of the adhesive bond

HONEYCOMB

FOAM

FIGURE 2 DIAGRAM Of TYPICAL DEBOND

VOIDS OR MINUTE TRAPPED ' QUANTITIES OF AIR OR GAS

pOOOOOQDOOQ 'OOOOOOOOOO

>oooooa~ ) QO Q Q Q Q O

FIGURE 3 DIAGRAM OF TYPICAL POROSITY CONDITION (MAGNIFIED)

NORMAL

or contamination of one of the faying surfaces (see

Figure 2). Porosity — Porosity is a condition of minute voids in a given area within a solid material. Porosity may be caused by insufficient flow of the resin during cure, localized excess heating of the resin during cure resulting in a foaming action, and resin contamination with a material that volatizes at cure temperatures (see Figure 3). Resin Variations — Both resin-rich and resin-starved areas are likely to occur under conditions of improperly controlled bleeding during cure. Either condition can have a significant effect on the mechanical properties of the cured laminate (see Figure 4).

56 APRIL 1977

RESIN STARVED

RESIN RICH

FIGURE 4 DIAGRAM SHOWING EFFECT OF RESIN VARIATIONS

NON-DESTRUCTIVE INSPECTION (NDI)

It is important that designers and homebuilders working with composite materials be familiar with NDI for the following reasons: a. The designer requires knowledge that the material he uses for his product will meet material specifications (reliability). b. The fact that composites behave as structures and not as homogeneous materials means that the effect of any imperfection is not only a function of the imperfection itself, but also of its location within the geometry and stress state appropriate to its use. Thus, the decision as to whether an imperfection constitutes a defect cannot, except in gross cases or if acceptance standards are indicated on the design drawing, be decided immediately by the homebuilder. c. If a component structural failure should occur, NDI records render invaluable aid in determining the cause of failure. d. Insight in NDI techniques is valuable for future composite designs and acceptable field repair techniques. Visual Inspection: Visual examination, made immediately after a sandwich panel has been removed from the tool or bag, and if it was cured with heat, often reveals unbonded areas as blisters. These blisters remain extended for a short period only, or until the drop in panel temperature reduces the internal pressure of the panel. During this short interval in which the blisters are visible, they may be outlined with an approved marking pen for future location and possible repair. Visual inspection methods appear to have only limited usefulness. If a part has blisters, the presence of defective areas is demonstrated and the part can be rejected or marked for salvage immediately. If no blisters are visible, however, the absence of defective areas is not proved and the part must be subjected to further tests by a more dependable method. Glass fabric facings, particularly void-free laminates, permit visual inspection of the core and sometimes aid in detecting poor bonds. The use of lights has been tried for determining faulty areas of blisters in various types of sandwich construction, but mainly in panels with glass fabric facings. By varying the arrangement and angle of lighting, some blistered areas can be detected, but not with any degree of reliability. Poorly bonded areas cannot be detected by means of lights. Tapping: Tapping is one of the simplest and most effective methods in use for testing for voids in the adhesive bond between the facings and the core of a sandwich part. The only equipment necessary for this test is a small metal piece, such as a coin or a small light hammer. During inspection by tapping, parts should be freely supported, as on three padded points, to eliminate sound interference from the support. A well-bonded area will produce a clear tone, while an unbonded area usually produces a lower tone or a dull thud. This method has been found to be reasonably satisfactory for detecting areas where the facings of the sand-

wich are not firmly attached to the core. It has been found, however, that if there is intimate contact between the facing and the core, no difference in tone quality can be detected between these areas and those that are well bonded. Poorly bonded areas, therefore, cannot be differentiated from well-bonded areas by means of tapping. Tests have shown that very light tapping is more selective than are heavy blows. Considerable experience is required to locate defective areas consistently, because parts of different construction

give off different tones and the tones on a single part vary with the position on the part. Variation in tone is especially noticeable within a few inches of the edge. Dye Penetrants: In cases where the delaminations, etc., extend to the surface, they can be indicated by the use of dye penetrants. A possible disadvantage of dye penetrants is interaction between the solvent and the composite materials. However, the method is well established, inexpensive and suited for the rapid examination of large surfaces. Thermographic Inspection: Disturbance of expected, uniformly distributed, thermal conductivity of sandwich panels can be indicative of unbonded areas or inclusions. Detection of these areas is possible by use of thermocouple readings, infrared sensing photographic or television cameras, or color changes in liquid crystal coatings as uniformly distributed heating is applied to the opposite sandwich facing. Inversely, the frost pattern immediately formed when a cooled sandwich panel is brought into a warm moist atmosphere may also show unbonded areas. Detection Limits: Table 2 lists the NDI methods discussed above with the detection capability for each method. Specific detection limits are not provided, as they are a function of part configuration. REPAIR

With the application of fiber-glass epoxy, as with any other type of construction, it is inevitable that a certain amount of damage will occur. During the manufacturing stages, where hazards of dropped tools and equipment are encountered, serious damage to composite parts may be eliminated by protecting exposed corners and by using temporary protective covers. Proper precautions will minimize damages; but, when damage does result, acceptable methods of repair must be available. Principles of Repair: Repair procedures are developed with the objective of equaling, as nearly as possible, the strength of the original part, with a minimum of increase in weight or change in aerodynamic characteristics and electrical properties where applicable. This can only be accomplished by replacing damaged material with identical material or an approved substitute. In order to eliminate dangerous stress concentrations, abrupt changes in cross-sectional areas must be avoided whenever practicable by tapering joints, by making small patches round or oval-shaped instead of rectangular and by rounding corners of all large repairs. Smoothness of outside surfaces of aircraft is a necessity for proper performance and, consequently, patches that project above the original surface must be avoided if at all possible. When this is impossible, the edges must be generously tapered to fair the repair into the original contour. Classes of Repair: It is sometimes necessary or more convenient to install a temporary repair over minor damage and later replace it by a permanent repair. These temporary or emergency repairs are normally devised to fit the application and therefore are not considered here. The repairs shown for fractures completely through the part may be further subdivided according to accessibility of the part from both sides or from one side. Structural panels such as wing surfaces, sandwich floors and bulkheads must often be repaired in place, and sometimes by working from one side only. If both sides are accessible, the same procedure can be followed; however, additional optional techniques are shown for alternate use. For convenience in presentation and for efficiency in designating repairs to composite constructions, damages are divided into groups or classes, according to severity and possible effect upon the structure. The following

classes are used in presentation of the repair techniques in this article. SPORT AVIATION 57

TABLE 2

^"^--^^^

NDI DEFECT DETECTION CAPABILITY

BONDED ASSEMBLIES

DEFECT

DE LAMINATIONS

DEBONDS

POROSITY

NDI METHOC>^\^^

CORE DAMAGE

INCLUSIONS

MISLOCATED DETAILS

FIBER DEFECTS

SONICS (TAPPING)

A

A

C

B

B

C

D

THERMOGRAPHY

B

B

B

C

B

A

C

PENETRANTS

A*

A*

B*

D

D

D

D

LEGEND -

A= B= C= D =

GOOD CAPABILITY FAIR CAPABILITY POOR CAPABILITY NO CAPABILITY

NOTE -

* = EDGES OR SURFACES ONLY

Class 1: Dents, scars, scratches or erosion in the facings, not accompanied by a puncture or a fracture. Class 2: Punctures or fractures in one facing only, possibly accompanied by damage to the core, but without damage to the opposite facing. Class 3: Holes or damage extending completely through the sandwich, affecting both facings and the core. Class 4: Extensive damage requiring replacement of a complete sandwich part or parts.

DAMAGE TO FACING

DAMAGED MATERIAL REMOVtD »Y SANDING

REPAIR TECHNIQUES

Class 1 Repairs: Scars, scratches, surface abrasion or rain erosion may be repaired as follows. Apply one or more coats of resin, catalyzed to room temperature, to the abraded surface (number of coats depends upon the severity of the abrasion). Small fractures may be filled with a putty made from room temperature-setting resin and short glass fibers. Over this coated surface, apply a sheet of cellophane that extends 2 or 3 inches beyond the painted surface. After it is taped in place, work out all air bubbles and excessive resin with the hand or a rubber squeegee. The resin can then be allowed to cure at room temperature or, if necessary, the cure can be hastened by the use of infrared bulbs or hot sandbags. Occasionally, on small parts, the whole part can be put in an oven set at 100°C (212°F) to hasten the cure. After the resin has been cured, the cellophane is removed, the excess resin is sanded off and the whole repaired area is lightly sanded preparatory to refmishing. If the surface abrasions, scratches or scars are deep enough to seriously affect the strength of the facing (usually to more than the first ply of fabric), they should be repaired in the following manner. Sand the damaged

REPLACED LAMINATIONS ,SANDED TO CONTOUR

FIGURE 5 STEPS IN REPAIR OF CLASS 1 DAMAGES TO SANDWICH FACINGS

DAMAGE TO FACING AND CORE

DAMAGED CORE REMOVED FACING SCARFED FOR PATCH

area either by hand or with a flexible disk sander to a

smooth contour, as shown in Figure 5. Sand to a distance of at least 100 times the depth of material removed. Coat the sanded area with one coat of room temperature-setting resin and apply pieces of glass fabric

WET-LAMINATED PATCH

soaked in resin to a resin content of about 50 percent.

Lay these pieces of fabric in place in the sanded depression, as shown in Figure 5. Tape cellophane in place

over the repair and work out excess resin. After the

resin has cured, the surface of the repair is sanded down to the original surface of the facing. Class 2 Repairs: Damages that extend completely through one facing of the sandwich and into the core

require removal of the damaged core and replacement

of the damaged facing in such a manner that normal

58 APRIL 1977

CORE REPLACED

FIGURE 6

STEPS IN THE REPAIR OF CLASS 2 DAMAGES

stresses can be carried over the area. Figure 6 shows

one method for accomplishing this type of repair. The damaged portion is carefully trimmed out to a circular or oval shape and the core removed completely to the opposite facing. Caution must be exercised not to damage the opposite facing or to start delaminations between that facing and the core around the damage. The damaged facing around the trimmed hole is then scarfed back carefully by using a flexible disk sander, a belt or rotating pad sander, or by hand to a distance of at least 100 times the facing thickness. This scarfing operation must be done accurately to a uniform taper and usually takes a little practice before acceptable scarfs are obtained. Contour lines produced by the individual plies of fabric in the sanding operation can be used to judge the accuracy of the scarfed surface. WARNING

The sanding operation on laminates reinforced with glass fabric gives off a fine dust that may cause skin irritation. In addition, breathing an excessive amount of this dust may be injurious. Therefore, precautions as to skin and respiration protection must be observed.

The details of the scarfed method are shown in Figure

7. They consist of trimming out the damaged portion

and proceeding as with Class 2 damage, except that the

opposite side of the sandwich is provided with a temporary mold or block to hold the core in place during the first step. After the first facing repair is cured completely, the mold and the shim (which temporarily replaced the facing on the opposite side) are removed and the repair is completed by repeating the procedure used in the first step. Typical steps in making this type of repair are shown in Figure 7. By using the stepped-joint method shown in Figure 8. the damaged portion is trimmed as before to a round or oval shape, or to a rectangular or square shape (prefer having rounded corners). The thicknesses of the individual plies in the facings are determined, for choice of replacement fabrics, from the portion removed; the total overlap of the stepped joint is computed from the number of plies in one facing, minus one, times IVz inches. This overall size is then marked on the sandwich. The marking can be done with cellophane tape or by lightly scratching the surface. The outer layer of fabric only is then cut with a sharp knife or a specially prepared cutter along these lines. CAUTION

A piece of replacement core material (or a suitable substitute) equal in thickness to the original core material is cut to fit snugly in the trimmed hole. The glass fabric laminations for the facing repair are then prepared, with the largest piece being cut to the exact shape of the outside of the scarfed area. The smallest piece is cut so that it overlaps the scarfed area by its proportionate amount, depending on the number of plies in the repair, and the intermediate pieces are cut to have equal overlaps. A convenient means of preparing these pieces is to brush-spread the resin on the pieces of fabric and sandwich the spread fabric between two sheets of colored cellophane. The pieces are then cut to shape without the usual fraying at the edges. The resin content of the fabric should be about 50 percent. When all of the pieces are ready for assembly, the opening from which the damaged core was removed is coated on all sides and bottom with room temperaturesetting resin. The piece of core that is to be inserted is likewise coated on all sides, including top and bottom surface, and inserted in the hole. The pieces of fabric are then laid in place by first removing the cellophane sheet from one side of the fabric, placing the exposed fabric in position on the repair and then removing the second cellophane sheet. The whole area is then covered by a piece of cellophane and carefully worked down to remove as much excess resin and air as possible. Light pressure is applied to the cellophane by means of sandbags, taping (if the repair is on a convex area) or a vacuum blanket, if facilities permit. After the repair has cured, it is lightly sanded to contour it to the original shape and it is then ready for refinishing.

An alternate method that may be used for Class 2 repairs is the "stepped-joint" method, described under Class 3 repairs. Class 3 Repairs: Damages that are completely through the sandwich may be repaired by two methods: (1) The scarf-joint method (similar to that described for Class 2 damage); and (2) the stepped-joint method described later. The scarfed method is normally used on small punctures up to 3 or 4 inches in maximum dimension, and in facings made of thin fabrics (which are difficult to peel). The stepped-joint method is often employed on larger repairs to facings composed of thick

fabrics.

Do not cut through more than one layer. If the layer of fabric underneath is scratched, the strength of the repair will suffer. Using a knife blade, the outer fabric layer can be lifted and carefully peeled away from the layer underneath until the entire sheet is removed. This leaves a clean-cut step round the area. The process is then repeated, with the cut being made at a distance of 1V6 inches inside the original step, as shown in Figure 8. Each consecutive layer of glass fabric lamination shall be removed in this manner except the last one (bonded to the core), which is exposed for an area approximately l'/2 inches wide around the trimmed hole. This surface is then lightly sanded. A piece of core material of identical thickness to that in the sandwich is prepared of the same material (or an approved substitute) and of a size to provide a snug fit in the trimmed hole. Glass fabric sheets of appropriate thickness are then cut slightly too large (approximately an inch or two oversize) for each of the steps in the repair. A mold and shim combination is now prepared for the opposite side of the sandwich to preserve the contour while the first facing is being repaired. After the mold and shim have been temporarily secured in place by clamping, propping or lashing, the damaged area is ready for the first step in rebuilding. The replacement core piece is coated with resin on all edges and the top surface only, leaving the bottom surface (next to the temporary shim) uncoated. It is then inserted in place above the temporary shim. The glass fabric sheets for repairing the facing are now impregnated with resin to a resin content of about 50 percent, and the smallest one is laid in place over the replaced core. It is then trimmed with scissors to the exact shape of the trimmed hole. After the trimmed portion has been removed, successive plies of glass fabric are laid in place

and trimmed, just as was done for the first ply. An extra layer of 112 cloth is then applied over the repair and trimmed so that it laps about Vz inch over the undamaged facing. The area is then covered with cellophane, and the excess resin and air are worked out as described earlier. Class 3 Repairs (Access Not Available to Both Facings): When access is not available to the inner facing SPORT AVIATION 59

HOLE THRU SANDWICH

HOLE THRU SANDWICH

HOLE TRIMMED, CORE REMOVED, AND FACING "STEP-PEELED"

HOLE TRIMMED, CORE REMOVED AND FACINGS SCARFED

WET-LAMINATED PATCH

TEMPORARY BLOCK OR MOLD

TEMPORARY BLOCK OR MOLD

TEMPORARY SHIM CORE REPLACED

s

x

TEMPORARY SHIM

CORE REPLACED

COMPLETE BY SCARFING AND WET-LAMINATING OPPOSITE FACING IN A SIMILAR MANNER

FIGURE 8 FIGURE 7

DETAILS OF STEPPED-JOINT METHOD OF REPAIR OF CLASS 3 DAMAGES

STEPS IN THE SCARFED TYPE OF REPAIR TO CLASS 3 DAMAGES

of a sandwich panel, the following procedures are suggested: 1) Remove the damaged facings and core in the form of an oval area, using the general methods described on preceding page. 2) Fabricate glass fabric backing plate and drill necessary holes in plate and sandwich panel (Figure 9) for self-tapping screws to be used in applying pressure. 3) Clean the surface of the backing plate and the area around the cutout on the back of the repair. 4) Apply adhesive to the backing plate and position in place on the inner facing of the sandwich part. 5) Fabricate replacement core patch and place in the core cavity over the backing plate. 6) Apply cellophane sheet, plywood caul and self-tapping screws, as shown in Step 1 of Figure 9. 7) After the adhesive has cured, the screws and caul can be removed and the screw holes filled with a viscous mixture of epoxy resin. 8) The upper overlap facing is then bonded in place using screw pressure. Heat for curing the adhesive can be applied to the outer bond line by the use of heat blankets, hot sandbags or heat lamps, but it is difficult to obtain heat on the inner bond line, and therefore adhesives should be

used there that cure adequately at room conditions. Voids and Delaminations: Foam sandwiches sometimes develop voids or delaminations at the interface between the facing and the foamed-in-place core, caused by a blow from a blunt object. These voids may be repaired by injecting a small quantity of room temperaturecatalyzed resin by means of a hypodermic needle and syringe. Two 1/32-inch holes are drilled in the facing

at diametrically opposite sides of the void area. The resin is injected into one hole until it starts to extrude

from the other. Slight pressure is then applied by means

of a hot sandbag over the void to remove excess resin,

and to insure contact between the facing and the core while the resin is curing.

60 APRIL 1977

ENVIRONMENTAL CONSIDERATIONS

Durability: An acceptable composite structure, in addition to possessing specified initial physical properties, must be relatively unaffected by continued exposure to conditions imposed by severe service. Obviously, with composite combinations involving new and untried materials, it is impossible to provide the assurance of adequate service tests, and reliance must therefore be placed on tests conducted after artificial aging under conditions thought to be typical of severe service. Artificial aging tests are time-consuming, and often do not include all the possible combinations of exposure conditions that might be encountered in actual service. Weathering: The service history of sandwich of reinforced plastic materials combined with simulated weather aging has developed data showing wide scatter, but analysis of the information showed that service life reduced strength more than did shelf life. After 3 years exposure at several outdoor sites, the salt air exposure appeared to be the most severe and produced a somewhat eroded surface on laminates, with exposure of some glass fibers on the surface. A summary of U.S. Air Force service experience showed sandwich with reinforced plastic facings had no basic weakness except for rapid rain erosion. A variety of solvent materials can be successfully resisted by epoxy resin matrix composites for exposures of 30 days, except exposure to paint removers, which degrade strength drastically.

The most serious and difficult service problem is that of water entry into the sandwich cores. Water accumulation can result in damage to the panel from freezing, which can rupture the core or panel itself. Fluid

entry can also contribute to dynamic unbalance and

weight gain in such parts as control surfaces. In several instances, foam core materials have failed due to vibration and have accumulated moisture, with a resultant loss of the bond. In production, or during fabrication and

finishing operations, moisture may enter if panels are

immersed or sprayed with water or other fluids. Also, panels with adhesive bonds with poor moisture resistance, or bonds of a porous nature and skins or edges of laminated plastic or other porous materials, allow fluids to enter by transmission directly through the material, though the quantity depends on the degree of porosity. Difficulty may result even with a fully sealed panel because of one or more of the following conditions: * High levels of sonic vibration cause sealant breakdown * Structural deflection * Differential pressure * Inadequate overlap in bonded joint * External accidental damage to panel * Inadequate design for sealing * Thermal stresses due to steam cleaning cause sealing or bonded joints to fail * Aging and deterioration of sealing materials * Inadequate material properties The techniques and materials used to provide effective sealing must take the above conditions into consideration. The quality of the adhesive bond — i.e., degree of porosity and adhesion qualities — has a direct bearing on the sealing characteristics. Polysulfide and filled epoxy compounds are the most frequently used materials for sealing joints between edge members, relief cutouts and tool or vent holes. For laminated facings, the use of void-free processing techniques and nylon or epoxy sealer coatings is recommended. Erosion and Impact Hazards: Rain and Sand — Water (rain) and sand particle impact subject a material to combined tensile, compressive and shear loads at rates significantly greater than in static mechanical property tests. It appears that, in general, those materials in a given class having highest hardness and tensile or yield strength are the most resistant to solid or liquid

particle erosion. In addition, material properties such as ductility, modulus, shear and fatigue strength enter into the erosion process of a material. The amount of erosion can be related to particle size, quantity and velocity, and depends on the response of the material to dynamic impact and the time of exposure. Conventional paint finishes can generally be selected or designed so as to provide the desired level of erosion protection for most composite structures (other than leading edges), such as control surfaces, fixed trailing edges, doors, etc. Greater care is required to provide erosion protection for leading edge structures, as defined in Figure 10. Materials such as neoprene or polyurethane have proven satisfactory in the protection of fiber-glass structures against erosion. Impact Hazards — Through both flight and ground servicing environments, aircraft components can suffer damage due to foreign object impact (e.g., hailstones, runaway debris, dropped tools, personnel contact, etc.). A synergism of design features and material systems is necessary to provide adequate resistance to such hazards and is best accomplished at component inception, as opposed to afterthought. Judicious selection of material type and form, lamina stacking order, thickness and properties can mitigate the consequences of many common hazards. Impact damage typically manifests itself through delaminations or disbonds up to the point at which fibers fracture and the laminate (or facing) is completely penetrated. It is preferable then to have materials with high elongation and high interlaminar shear strength to absorb the energy induced by the impactor S-glass fibers, for instance, have been shown to have better impact strength than E-glass due to their higher strain-to-failure, but are much costlier. Rubber-

FILLED SCREW HOLES IN COVER PLATE .(HOLES USED IN BONDING COVER PLATE) SCREW HOLES UNDER COVER PLATE THROUGH PANEL AND BACKING PLATE

'OUTER EDGE Of COVER AND BACKING PLATE

WOOD OR SELF-TAPPING METAL SCREWS PLYWOOD CAUL SHEET

STEP 1 RELEASE FILM

BACKING PLATE (REPLACEMENT FACE SHEET), AND REPLACEMENT CORE BEING BONDED / INTO PLACE / REPLACEMENT CORE STEP 2 ........... TAPERED COVER PLATE (REPLACEMENT FACE SHEET)

ADHESIVE SQUEEZE-OUT NON-RIGID FOAM CUSHION

fl^-^r.-j )' -,,/,"•"-

FILM

BEING BONDED INTO PLACE / /

( TAPERED COVER; \PLATE OVERLAP \ ______I——

125 T

COMPLETED REPAIR FILLED SCREW HOLES

SECTIONS THROUGH A - A SHOWING STEPS IN MAKING REPAIR FIGURE 9

METHOD OF REPAIRING CLASS 3 DAMAGE WHENl PANEL IS ACCESSIBLE FROM ONLY ONE SIDE

SPORT AVIATION 61

ZONE Of MAXIMUM EROSION TANGENT TO AIRFOIL , AIRFOIL CHORD PLANE

There may not be enough historical evidence, experiments or tests available to draw a clear cut conclusion to questions as "What happens to the structural integrity of my composite aircraft primary structure if ... ?"

Moisture Absorption and Ultraviolet Degradation:

AIR FLOW ASSUMED PARALLEL TO AIRFOIL CHORD PLANE

FIGURE

10

EROSION PROTECTION AREA

modified epoxy resins are tougher than conventional nitrile epoxies, phenolics or polyesters, but their choice should not sacrifice interlaminar strength. Woven fiberglass epoxy fabrics possess inherently better resistance to delamination from impact than do impregnated unidirectional tapes due to the mechanically interlocked nature of the weave. Multidirectional^ oriented laminates resist damage from impact better than unidirectionally oriented laminates, since energy dissipation tends to follow the path of the stiffer constituent (fiber) of the composite and, hence, the more fibers in more directions the better. It stands to reason that, for a given impactor shape and energy, that a thicker laminate will suffer less damage than a thinner one, and thereby retain more of its original strength (up to the point of penetrating damage). As general practice, the primary load-carrying layers of a laminate (usually designated as the 0° direction) should be placed within the laminate (not on the extreme surfaces) so as to prevent damage to them due to more mundane hazards such as abrasion and scratches. Core selection should be based first on structural strength (and cost or fabricability) considerations and then that selection modified, if necessary, to provide adequate damage resistance to the sandwich. The core of a sandwich panel will absorb energy from impact by compressing (crushing) and flexing. Thus, it is best to have a thick, but compressible, core that has sufficient shear strength to carry flight loads, yet not impose a severe weight penalty. Continuous cores (e.g., foam, balsa) help reduce facing damage more than cellular cores (e.g., hexagonal honeycomb), since the latter provides only intermittent support at the nodes (cell walls) and the facing experiences high local bending strains between the nodes which can cause fiber failure at lower energies. The addition of extra layers of material (sometimes called "sacrificial plies") at the surface enhance the durability of composite laminates. Such concepts as elastomeric films (polyethylene or mylar), aluminum foil, ballistic nylon and Kevlar fabric/epoxy have been employed to reduce he deleterious effects of erosion, abrasion, impact and moisture penetration. Complete impact protection is, of course, impossible. The designer of composite aircraft should, consequently, take precautions to design for inspectability, accessibility, repairability, replaceability and avoid minimum gage designs where durability is a consideration.

ENVIRONMENTAL RELATED DESIGN FACTORS Throughout the course of these four articles, we've alluded to a number of factors which relate the initial design to long term effects. These effects, summarized in Table 3, are important if the aircraft structure is to be utilitarian over a long service life and retain its market value. The conditions or effects listed are understood by the composite technological community at various levels of competence and agreement. In a number of cases, the level of agreement is minimal, primarily due to the fact that little or no reliable data base exists. 62 APRIL 1977

There are two phenomena which were described in Part I, July SPORT AVIATION. The absorption of moisture from the air into the resin is a process that can be computed from the one-dimensional diffusion equation. For thin laminates, this process takes days to weeks, depending on the relative humidity-temperature environment and the chemical nature of the cured resin. The effect is to reduce the glass-transition temperature of the resin, which in turn reduces the temperature at which the composite is a viable structural material. In general, it is a process which cannot be prevented and, at best, can only be retarded. Moisture absorption is considered in the design process by using an empirical strength "knock down" factor or by using strength allowables generated at the "hot-humid" test condition. Ultraviolet radiation serves to sever the chemical chain comprising epoxy resin systems. This process can be prevented by protecting the structural composite by surface gel coats, paint, barrier films, additives to the resin, etc. Since UV degradation is a long term (years) process, and since it may be readily avoided, it is not considered directly in the design process except for a note on the drawings. Abrasion/Impact Damage: Invariably, some time during the lifetime of the aircraft structure, damage will occur. This article deals with this condition (as well as the erosion problem) and offers suggestions for its prevention and repair. Fatigue Life: In comparison to metals, which have fatigue-safe stress levels of one-third their ultimate static strength, composites have fatigue-safe stress levels from two-thirds to one-fourth their ultimate static strengths. This wide range is fiber orientation-to-load direction dependent. A unidirectional laminate will have substantially differing ultimate static strength to fatigue runout strength (107 cycles) ratios when loaded in the fiber direction and in the direction transverse to the fibers. In multi-oriented laminates, the presence of internal laminate shear influences the fatigue strength. To date, there is no good reliable analytical model which successfully predicts composite fatigue strength, and therefore one must rely heavily on test data taken at the specimen and component level. Hot and Cold Environment: Epoxy resin, being a thermosetting plastic, loses its strength and stiffness characteristics with increasing temperature. If the aircraft structure is expected to sustain design loads at the high temperature extreme (130 F to 160 F), then the structural analyst/designer must check that condition. The reduction of room temperature strength allowables at 150 F for ambient cured epoxies may be as much as 25%. To some extent, this substantial reduction in strength may be reduced by post-curing the structure at the high temperature extreme (say 160 F). This may be done with heat lamps, a box or tent connected to the wife's clothes dryer, etc. Another approach is to minimize the heating due to solar radiation. This is

why all composite sailplanes are white. In any event, the degree to which the hot environment design condition influences the design process depends on the expected service environment. In cold environments, the strength of fiber-glass epoxy composite laminates increases somewhat. At -40 F, this increase in strength would not normally be greater than 10% over the room temperature values. At the

cold condition, differential expansion between the composite and surrounding structure may cause problems.

Fiber-glass composites have thermal expansions of about 6 microinches per degree F. Wood, steel, and titanium match the composite thermal expansion quite closely; e.g., wood, 2 to 7 microinches/F; steel, 6 to 9 microinches/ F; titanium, 5 to 6 microinches/F. Materials which do not match as closely include: aluminum, 13 microinches/ F; acrylics (Plexiglass/Swedcast), 30 to 50 microinches/ F; and Dynel/epoxy, 40 to 50 microinches/F (estimated). Materials having different expansivities may be structurally joined provided the engagement is not too great and the thermal extremes not too large, since the induced stresses between the two materials is a function of AF.

Creep/Relaxation: A composite material could exhibit an appreciable amount of creep, depending on the state of stress and the temperature. Viscoelastic flow in the matrix, as well as internal flaw formation and Table 3. CONDITION OR EFFECT

MOISTURE ABSORPTION BY RESIN

growth, are normally the main sources of creep. Creep, then, can be observed as a measurable change in dimension, deflection or rigging. Except for unidirectional specimens loaded parallel (in-plane) to the fibers, considerable nonlinearity has been observed for multioriented composites. Creep for multi-oriented composites is a short-term process, where measurable changes can be detected within minutes to hours, depending on the stress level. In cases where the load is cyclic or reversible, relaxation back to the original state or position is observed. In cases where the load produces very high constant stress, creep rupture may occur. That is, if the applied load produces sufficiently high enough stresses, the creep will be sufficient to cause failure. A timedeflection history of an end-loaded cantilever beam might be a) initial static tip deflection due to the end load at time = 0 hours; b) continued tip deflection due to creep at the inboard end between time = 0 plus hours and time = XX hours; and c) at time = XX plus hours, the beam fails at the inboard end. For a beam where the stress is 90^ of the ultimate static strength, failure

Summary of Environmental and Configuration Related Design Factors

CONSEQUENCE

PLASTIZATION OF

RESIN, REDUCTION OF STRENGTH ALLOWABLES

HOW OBSERVED

NOT OBSERVED

ULTRAVIOLET DEGRADATION OF RESIN

ABRASION/IMPACT (HANGER RASH)

REDUCTION OF FATIGUE

STRENGTH

REDUCTION OF STRENGTH BUILT INTO DESIGN ALLOWABLES (I.E. KNOCKDOWN FACTOR)

DUE TO DISRUPTION OF BOND BTN FIBER & RESIN CHALKING OF SURFACE;

IMBRITTLEMENT OF RESIN LEADING TO MICRO-CRACKING

COMMENTS

HOW MINIMIZED

NOT A SERIOUS PROBLEM WHEN TAKEN INTO CONSIDERATION FOR HOMEBUILT DESIGNS

RESISTANT MATERIAL

U-V DEGRADATION OF RESIN IS A LONG TERM PROCESS, BUT MUST

COAT, ALYLIC ENAMEL PAINT U-V BARRIER FILMS, WAX ETC

PROCESS

EXTERIOR SURFACES COATED WITH U-V (I.E. POLYESTER SURFACE

LOSS OF AERODYNAMIC PERFORMANCE, POSSIBLE

PITS, DEEP SCRATCHES, DINGS; RESIN APPEARS

IMPACT RESISTANT RESIN, THIN METALLIC OVERLAYS,

STRUCTURAL FAILURE INITIATION POINT

MILKY AT IMPACT POINT

(EG. .005 ALUM OR CRES FOIL BONDED TO LEADING

BE MINIMIZED SINCE IT IS A NON-REVERSIBLE

PROTECTION CONTRIBUTES TO LOWER LONG TERM MAINTENANCE

EDGES OR OTHER DAMAGE PRONE AREAS)

LONG TERM FATIGUE LIFE

USEFUL LIFE OF

STRUCTURE REDUCED

MICROCRACKING,

VARIATIONS OF RETURN

ON ACCOUSTIC TEST (COIN TAPING)

CONSIDERATION OF FATIGUE ASPECTS DURING DESIGN/ ANALYSIS PHASE

FATIGUE OF FG/EPOXY LAMINATES VERY GOOD. FATIGUE OF LAMINATE

TO-FOAM CORE BOND AND FOAM CORE IN SHEAR UNKNOWN

HOT AND COLD ENVIRONMENTAL EXTREMES (-40F TO 160F)

CREEP/RELAXATION

LOSS OF STRENGTH PROPERTIES WHEN HOT. POTENTIAL BOND FAILURES DUE TO MATERIALS THERMAL EXPANSION MISMATCH

VISCOELASTIC FLOW IN RESIN 1 INTERNAL FLAW

FORMATION & GROWTH

MICROCRACKING; RESIN CRAZING AT LOCATIONS WITH THERMAL MISMATCH IN MATERIALS

STRIVE TO UTILIZE MATERIALS HAVING

EXPANSIONS CLOSE TO LAMINATE EXPANSION

(EG. 4 130 STEEL, WOOD,

DON'T OVERSTRESS STRUCTURE WHILE HOT.

OBSERVE AREAS

OF POSSIBLE CONCERN DURING PRE-FLIGHT

AND TITANIUM MATCHES FG/EPOXY CLOSELY)

CHANGES IN DIMENSION, RIGGING, AND SHAPE

DESIGN STRUCTURE SO THAT NORMALIZED

OCTAHEDRAL SHEAR STRESS /^/T^OFALL

LAYERS IN LAMINATE IS 2000 PSI OR LESS FOR 1G CONDITION. STORE IN

HIGH LOADS PRODUCING HIGH STRESSES, CAUSES, CHANGES IN SHAPE

SINCE RESIN IS A VISCOELASTIC MATERIAL. PHENOMENA IS TEMPERATURE DEPENDENT

COOL ENVIRENMENT.

RESIN DAMAGE

DEGRADATION OF STRUCTURAL INTEGRITY

SOFTENING/DISOLVEMENT OF RESIN

DO NOT USE EPOXY PAINT REMOVERS, CHLORINATED OILS, S. STRONG ACIDS ON EPOXY SURFACES

HYDRAULIC FLUID, SOLVENTS (MEK, ACETONE, ETC), FUEL AND OTHER CHEMICAL AGENTS NORMALLY FOUND IN

VICINITY OF AEROSPACE STRUCTURES DO NOT NORMALLY CAUSE PROBLEMS RAIN/HAIL EROSION

LOSS OF AERODYNAMIC PERFORMANCE

PITTING OF SURFACES

SCALE EFFECT

REDUCTION IN STRUCTURAL MARGINS OF SAFETY

PREMATURE FAILURE DURING LOAD TESTS

(SEE ABRASION/IMPACT)

TESTING IS NECESSARY AT (1) MATERIAL, (2) STRUCTURAL ELEMENT, (3) SUBCOMPONENT, AND (4) COMPONENT LEVELS.

BONDED JOINTS HAVE DEMONSTRATED SIGNIFICANT SCALE

EFFECTS. SCALE

EFFECT DIFFERS FOR

FATIGUE i STATIC

STRENGTH S. DIFFERS

WITH CONFIGURATION LAMINATE DISCONTINUITIES

STRESS CONCENTRATION

INITIATION OF DAMAGE

AVOID HOLES i CUTOUTS IF POSSIBLE. REINFORCE WITH ADDITIONAL MATERIAL (DOILIES)

FRACTURE BEHAVIOR CONTAINING DISCONTINUITY DEPENDS 1. ANISOTROPY 2. EFFECT OF BOUNDARIES 3. SIZE 4. ORIENTATION RELATIVE TO LAMINATE.

SPORT AVIATION 63

might occur in minutes to hours, depending on the material, orientation, environment, prior fatigue history, etc. Aircraft are not normally subjected to that type of stress level but, if they were, the load time would usually occur over a few seconds, not minutes or hours. But the combined effects of moisture, temperature, fatigue and creep are not understood at this time, and therefore the design process shall be intentionally conservative until these effects can be properly treated and judicial recommendations offered. Resin Damage: Typical liquid agents found at airports, such as aircraft fuels, hydraulic fluid, cleaning agents, etc. do not affect epoxy resins. However, strong acids (sulfuric nitric, aqua regea and others), chlorinated oils (methylene chloride) and epoxy paint removers can and will seriously damage the resin. These materials should not be used on resin matrix composites. If an aircraft requires repainting, then the old paint must be removed and stripped by mechanical means, such as sanding. Rain/Hail Erosion: Erosion of aerodynamic stagnation points can occur due to rain and/or hail impact. Since the impact energy is a function of velocity squared, the higher speed aircraft would be expected to receive severe damage. If the primary structure of the wing consisted of only the "D box", or if torsional integrity only depended on the exterior skin, then protection of the leading edge becomes paramount. This protection could be in the form of additional sacrificial layers of material, thin metallic overlays and/or neoprene boots. Electromagnetic Compatibility: When fuel is pumped from storage into an aircraft, static change can be either accumulated in, or dissipated from, the fuel as it moves through various elements in the fuel delivery system. Typically, charging occurs near surfaces where ionic materials provide an electrical double layer. The flowing fuel sweeps away charged ions of one polarity, leaving the opposite charges to flow to ground. Then, the charge in the fuel dissipates as the fuel is exposed to electrical paths to ground, and the charged ions combine with available opposite charges. Obviously, charge generation and charge dissipation occur simultaneously throughout the entire fuel delivery system. Local system characteristics determine if the net effect is to charge or discharge the fuel. Charge generation is a function of the area of the fuel-to-surface interface, properties of the fuel and surface, and the relative velocity of the fuel past the surface (flow rate). Charge generation can be either positive or negative, and may change from one part of the system to the next. Charge dissipation is principally a function of the electrical conductivity of the fuel (one conductivity unit = 1 CU = 1 picosemen/cm) and time. If the conductivity of the fuel is high, charge dissipation (or neutralization) occurs almost instantaneously and very little charged fuel may be found even a short distance from the chargegenerating surface. If the conductivity of the fuel is low, the fuel itself acts as an insulator and the charge remains for a considerable length of time.

Fuel with a high charging tendency, low electrical conductivity and processed at high flow rates can easily deliver enough electrical energy to an aircraft fuel tank

to produce incendiary sparking. The method by which that energy is handled as it enters the tank determines if an incendiary spark will occur. Arcing will not occur if the energy is dissipated uniformly at the rate it enters the tank. Arcing is likely if the energy is allowed to accumulate. If the accumulation of energy occurs in a conductive material that can give up significant amounts of energy instantaneously, the spark can contain a high enough concentration of energy to be incendiary. 64 APRIL 1977

The electrical conductivity of the tank must be considered in evaluating the ability of a particular design to safely dissipate expected levels of charge energy. Comparing idealized conductive and nonconductive tanks helps one understand the impact of tank conductivity on the dissipation of incoming charge energy, and to separate this effect from that of detail tank design. If the tank material is now considered perfectly insulating, the idealized spark conditions for the nonconductive tank may be deduced and several conclusions become apparent. Charged fuel in the idealized nonconductive tank has no path to the ground; hence, the fuel cannot discharge either in a controlled manner or by a spark. The tank/fuel combination is analogous to a charged capacitor. A path to ground must be available for sparking to occur. In the nontrivial case of an actual composite tank, ground paths of this type are probably common. They occur in the form of detail tank design such as systems, plumbing or splices to conductive structure. The sparking situation is created by detail tank design. Remember, a low fuel conductivity is required for an arcing voltage to develop between the fuel surface and ground in the conductive tank. However, this low fuel conductivity inhibits quick energy release. If a spark occurs, it is generally of low energy. The perfectly nonconductive tank, on the other hand, does not discharge the fuel. An accumulation of charge in a high conductivity fuel is possible. Incendiary sparking could be possible if a path-to-ground becomes available as the tank fills. Note, however, that the possibility of unbonded conductors is more likely for the nonconductive tank. Metal is typically used for splice plates, clips, etc. These items would naturally be grounded if they were installed in a metal tank. Special precautions might be necessary to assure grounding in a nonconductive tank. Also, the nonconductive tank would maintain the full incoming fuel charge level, exposing any charge collector to a higher charge level than would probably occur in the identical situation with a conductive tank. Fuel charging characteristics, fuel conductivity, fuel delivery equipment and fuel delivery rates all impact the sparking tendency within a given tank. Assuming, however, that all tanks will be exposed to a wide variation of these parameters, the following conclusions concerning idealized conductive and nonconductive tanks may be drawn. 1. Sparking in all tanks is primarily a function of detail tank design. 2. Extensive use of insulating materials may increase the possibility of a detail design deficiency. Many conductive clips, splices, fasteners, etc. will be used, and a conductive charge collector will be present if they are not adequately bonded. 3. The idealized nonconductive tank maintains the incoming charge in the fuel and presents this higher charge level to any sparking situation present. This characteristic is more critical if highly charged conductive fuel is delivered to the tank.

Factors affecting the consideration of actual composite fuel tanks are: 1. The extent to which the composite tank approaches the idealized nonconductive situation, -TS measured by the energy dissipating capability. 2. The necessity of special design considerations to assure bonding of all conductive materials used within the tank. 3. The capability to deliver highly charged and highly conductive fuel to the tank. To put all this into perspective, one must recognize

that a fiber-glass fuel tank is nonconductive, that the

design of the tank is very important, and that the grounding clip normally attached to the exhaust stack during

fueling may not offer any safety refuge for situations of a high charged fuel, low conductance fuel being pumped into the tank. Additional study and testing is required to clarify the problem and to assert constructive design guidelines. Electromagnetic Vulnerability: The vulnerability

in observing this type of discharge. However, these strokes are considered to be less severe than the cloudto-ground strokes, so 200.000A can be assumed to be

fiber-glass construction, this is still a controversial subject, since an inadequate understanding of the problem exists. For the time being, flight of primary structure fiber-glass aircraft in regions of lightning activity is not suggested. A natural lightning discharge initiates from a charge center in a cloud and propagates as a step leader toward another charge center or toward earth. Aircraft have triggered discharges, but usually the stroke is merely diverted from its path through the vehicle. An intense voltage exists between the step leader and the aircraft before attachment, and this causes streamers to form on all external surfaces of the aircraft. When one of these streamers contact the step leader, the aircraft is raised to the discharge potential and streamers extend off the opposite extremities of the structure. This step leader continues until another charge center or earth is contacted. An ionized path then exists. Conduction of a large surge current begins and the ionization wave travels back up the stroke path. This wave, referred to as the "return stroke", consists of a high-current surge with a fast wave front. The stroke may be considered stationary with the aircraft moving relative to it. This sweeps the ionized channel along the aircraft and permits contact at points aft of the attachment. This phenomenon is called "swept stroke" and is frequently accompanied by a "restrike". The restrike is a moderately high current surge caused by other charge centers within a cloud discharging down the ionized path. Note that the discharge only passes through the aircraft and is neither the termination or initiation point. Even large aircraft cannot store enough energy to appear as a sufficient charge center for lightning discharges to occur. Damage caused by lightning attachment to the metal skin of an aircraft is caused by the metal's inability to carry the discharge current without heating to a melting or vaporization temperature. Aluminum has good thermal and electrical conductivity, so aircraft skins have seldom shown severe damage. Since composites are less conductive, vulnerability will be greater, even to lower currents. Attachment studies on aircraft models indicate certain areas of high probability stroke contact. Actual strike data confirms that surfaces on the vehicle can be grouped into "zones", according to probability of direct attachment. From this, the zones can be defined as follows: * Zone 1 — Surfaces of the aircraft for which there is a high probability of direct lightning stroke attachment. * Zone 2 — Surfaces of the aircraft for which there is a high probability of a stroke being swept aft from a Zone

extremities (nose cone, wing/empennage tips, fairings, etc.). For an all fiber-glass composite aircraft, intuition might suggest that a nonconducting structure would survive, just as insulators are used on high voltage power lines. Not so; the total power (voltage times current) of a nominal strike is higher, the lightning strike phenomena is a dynamic (propagation) occurrence, and the characteristics of the material change in the presence of an ionized plasma. Severe damage has been reported (in laboratory tests) between metallic fasteners and the composite material, as might be in a typical structural joint. Fiber-glass sailplane damage (or nondamage) has not been adequately documented and reported, so as to lead one to prophesize blanket safety coverage for powered aircraft in all situations.

of composites to lightning strike is discussed here. For

1 attachment point. * Zone 3 — Areas other than those covered by Zones 1 and 2 which have a low probability of direct or swept stroke attachments. Data on the current amplitudes of a lightning strike is limited. Estimates have placed 200,000 amps as the realistic maximum for cloud-to-ground strokes, with this magnitude occurring only about 0.5% of the time. The 30,OOOA stroke is an average value. Cloud-to-cloud amplitudes are relatively unknown due to the difficulty

the worst case.

To our knowledge, the problems and test data associated with lightning strikes have been for aircraft

with aluminum construction and fiber-glass or graphite

SUMMARY

In this part of the series, we've attempted to surface a number of not frequently discussed issues on the use of composites for aircraft, as well as common techniques for their repair should damage occur. For the builders, as well as the designers, we hope this part of the series was informative. SUBSEQUENT ARTICLE

Fabrication of the outer wing static test article is proceeding on schedule. The final installment will be on the fabrication methods and test results of that component. REFERENCES

We've arranged our references in a bibliographical form for your convenience. Materials 1. Ferro Fiberglass Data Sheet, "Ferro S Glass Filament Winding Roving Prices", 7 January 1971. 2. Technical Bulletin, "Ferro S-1014 Ultra-High Strength Glass Fiber Products", FG-101. 3. Technical Bulletin, "Ferro S-1014 Ultra-High Strength Zero Twist Fabrics (with Data Sheets)", FG-102. 4. Owens-Corning Fiberglass Corporation Publication No. 5-ASP-6869, "We've been where you've been, we're going where j'ou're going". 5. 3M Data Sheet SP-RNB(1031)BE, "Scotchply Reinforced Plastics". 6. Owens-Corning Fiberglass Corporation Publication No. 5-ASP-6870, "Comparative Data; E, S and S2 Glass". 7. Burlington Glass Fabrics Company, "Industrial Woven Glass Fabrics (A Guide to Specifications and Prices)". 8. Owens-Corning Fiberglass Corporation Publication No. l-PL-6305, "An Introduction to FiberglassReinforced Plastics/Composites", March 1974. 9. Owens-Corning Fiberglass Corporation Publication No. 5-GT-5442-A, "Textile Fibers for Industry", March 1972. 10. J. H. Flickinger, "Elastomer Modified Epoxy Resins", 22nd SPI Technical Conference. 11. Dow Chemical Company Technical Bulletin, "High Peel Strength Liquid Epoxy Resin, SC-7575.02, for SPORT AVIATION 65

Improved Peel Strength, Tensile Shear Strength and Toughness in Epoxy Adhesives". 12. F. J. McGarry, "Crack Propagation in Fiber-Reinforced Plastic Composites", Fundamental Aspects of Fiber-Reinforced Plastic Composites, Wiley and Sons, 1968. 13. C. A. Harper (Editor-in-Chief), "Handbook of Plastics and Elastomers", McGraw-Hill, 1975. 14. Hexcel Aerospace Bulletin TSB120, "Mechanical Properties of Hexcel Honeycomb Properties", August 1971. 15. Hastings Plastics Company TNB 11-1-3, "Hathane 1680 Polyurethane Foam Data Sheet", 15 July 1973. 16. MIL-HDBK-17 and 17A, "Plastics for Aerospace Vehicles", Part 1, January 1971. 17. MIL-HDBK-5B, "Metallic Materials and Elements for Aerospace Vehicle Structures", September 1971. 18. Uniglass Industries, "Uniglass Fabrics Handbook", 1967. 19. "Modern Plastics Encyclopedia", Vol. 53, 10A, McGraw-Hill, 1976/1977. 20. Shell Chemical Company Technical Bulletins SC: 72-60, SC: 72-25, SC: 71-14, SC: 71-1, SC: 71-12. Design and Analysis 21. A. G. H. Dietz, "Composite Engineering Laminates", MIT Press, 1969.

22. MIL-HDBK-23, "Structural Sandwich Composites", December 1968. 23. P. Kuhn, "Stresses in Aircraft and Shell Structures", McGraw-Hill, 1956. 24. E. E. Sechler and L. G. Dunn, "Airplane Structural Analysis and Design", Wiley and Sons, 1942. 25. A. Gomza and P. Siede, NACA TN 1710, "Minimum Weight Design of Simply Supported Transversely Stiffened Plates Under Compression", Sept. 1948. 26. B. W. Rosen, NACA TN 3633, "Analysis of the Ultimate Strength and Optimum Proportion of Multiweb Wing Structures", March 1956. 27. E. E. Sechler, "Elasticity in Engineering", Wiley and Sons, 1952. 28. G. Gerard, "Minimum Weight Analysis of Compression Structures", NYU Press, 1956. 29. G. Gerard, "Optimum Number of Webs Required for a Multicell Box Under Bending", Journal of Aeronautical Sciences, January 1948, pp. 53-56. 30. W. C. Paulsen, "Finite Element Stress Analysis", Machine Design, Part 1, September 30, 1971, pp. 46-52; Part 2, October 14, 1971, pp. 146-150; Part 3, October 28, 1971, pp. 90-94. 31. K. H. Huebner, "Finite Element Method — Stress Analysis and Much More", Machine Design, January 10, 1974, pp. 92-99. 32. R. M. Jones, "Mechanics of Composite Materials", McGraw Hill, 1975. 33. D. H. Emero and L. Spunt, "Optimization of Multirib and Multiweb Wing Wing Box Structures Under Shear and Moment Loads", 4th AIAA/ASME Struc-

38. S. Timoshenko and S. Woinowsky-Krieger, "Theory of Plates and Shells", 4th Edition, McGraw-Hill, 1957. 39. S. A. Ambartsumyan, "Theory of Anisotropic Plates", Technomic Publishing Co., 1970. 40. J. E. Ashton and J. M. Whitney, "Theory of Laminated Plates", Technomic Publishing Co., 1970. 41. G. N. Savin, NASA TT F-607, "Stress Distribution Around Holes", November 1970. 42. S. G. Lekhnitskiy, FTD-HT-23-608-67 (AD 683218), "Anisotropic Plates", March 1968. Environmental 43. D. J. Birmingham, M.D., "Contact Dermatitis from Synthetic Resins", Regional SPI Technical Conference, NY Section, "Plastics in the Medical Sciences", September 1967. 44. J. R. Stahmenn, "Model Studies of Strike Probabilities to Selected Points on Aerospace Vehicles", Paper No. 700915, 1970 Lightning and Static Electricity Conference, 1970. 45. J. F. Shaeffer, "Aircraft Initiation of Lightning", AFAL-TR-72-325, 12-15 December 1972, pp. 192200. 46. J. A. Plumer, "Data from the Airlines Lightning Strikes", AFAL-TR-72-325, 12-15 December 1972, pp. 282-289. 47. N. Cianos and E. T. Pierce, Stanford Research Institute Technical Report 1, "A Ground Lightning Environment for Engineering Usage", August 1972. 48. SAE Committee AE-4 Special Task F, "Aerospace Recommended Practice: Lightning Effects Tests on Aerospace Vehicles and Hardware", 13 September 1975. 49. R. O. Brick, C. H. King and J. T. Quinlivan, "Coatings for Lightning Protection of Structural Reinforced Plastics", AFML-TR-70-303, Part II, February 1972. 50. C. D. Skouby, McDonnell Aircraft Company, "Electromagnetic Effects of Advanced Composites", January 1975. Other 51. J. E. Dougherty, Jr., "Certification Requirements for Civil Composite Commercial Airframe Structures", 3rd NASA/USAF Conference on Fibrous Composites in Flight Vehicle Design, Williamsburg, Virginia, 4-6 November 1975. 52. FAR, Part 23, Changes 1 through 15.

tures, Dynamics and Materials Conference. 34. F. R. Shanley, "Principles of Structural Design for Minimum Weight", Journal of Aeronautical Sciences, March 1949. 35. W. R. Micks, "Method of Estimating the Compressive Strength of Optimum Sheet-Stiffener Panels for Arbitrary Material Properties, Skin Thickness and Stiffener Shapes", Journal of Aeronautical Sciences, October 1953, pp. 705-715. 36. G. Gerard, "Minimum Weight Analysis of Orthotropic Plates Under Compressive Loading", Journal of Aeronautical Sciences, January 1960.

37. S. Timoshenko, "Theory of Elastic Stability", McGraw-Hill, 1956.

A VP-2 recently completed by D. H. D. Lowry (EAA

17348), 34 Bradgate Rd., Belleville, Ontario. It has since had a long turtle deck added to eliminate some rudder

buffet. The tires are 8.00 x 6.00 on 6.00 x 6.00 wheels. 66 APRIL 1977