ASM Metals HandBook Vol. 17 - Nondestructive Evaluation and Quality Control

Подождите немного. Документ загружается.

Reference test standards for calibrating and standardizing the bond inspection instruments are essential. Even with

adequate reference standards, inspection often produces conflicting information or a deviation of results. In these cases,

more than one inspection method should be used in order to improve the accuracy of the data and to define the deviating

condition. Under ideal conditions, the test standard with built-in defects of selected sizes should closely duplicate the

structure adherends to be inspected. The laminate skin should be of the same thickness, and the honeycomb core should

be of the same size and density. Other variations to be encountered in the test specimen, such as tapered core, chemically

milled skins and doublers, and so on, should be incorporated into the test standard.

Except for built-in defects, the test standard for adhesive-bonded structures should be fabricated in the same manner as

the production assembly. The void or defect should be introduced in the same bond line as that to be inspected in the

structure. The standard can be a series of simple test specimens composed of details identical to the several areas of the

assembly to be inspected and having a bond of good quality but with controlled or known defect locations and size.

Attempts to produce voids or controlled understrength bonds by the local application of grease or other foreign materials

have been found to be ineffective. All adhesives and primers vary and may not respond to the following suggested

methods of standard preparation. Therefore, the finished standards should be evaluated by various NDT methods prior to

validation of their use as an inspection standard. Neutron radiography is an excellent method for verification if voids,

porosity, or both have been generated in the standard. The standard must then be inspected by the test instrument chosen

for the particular part to confirm its ability to locate the flaw.

Metal-To-Metal Reference Standards. Various methods have been developed for producing simulated voids,

porosity, or unbonds in metal-to-metal bonded laminates. Numerous ways to produce reference standards for metal-to-

metal bonded joints or assemblies are described in Ref 10. Because of the variety of such standards, it is not possible to

describe all of them in this article, and it is recommended that the reader consult the references for additional information.

Figure 46 shows two methods for producing standards for simulated skin-to-adhesive unbonds. The method 1 or 2

standards are useful for all NDT application methods provided the Teflon or Mylar inserts do not bond to the adherends.

Fig. 46

Two methods of fabricating unbond reference standards. (a) Method 1 uses two pieces of aluminum

sheet that have been anodized and primed and then bonded together with a piece of adhesive that has various-

size cutouts filled with 0.13 mm (0.005 in.) thick Teflon disks. (b) Method 2 uses metal as in method 1, but the

voids are produced by placing a disk sandwich of Mylar and scrim cloth of desired size on the uncured adhesive.

The assembly is then bonded.

Void, Porosity, and Unbond Standards. Voids or unbonds of specified sizes are difficult to produce because free-

flowing adhesive may fill the anticipated void or unbond area. One researcher was successful in producing natural flaws

in the adhesive layer of metal-to-metal bond joints (Ref 1). By trial-and-error methods, defect conditions such as a void

produced by removing a section of adhesive, a void produced by inserting a 0.64 mm (0.025 in.) diam wire between

adherends, and porosity caused by inserting a 0.38 mm (0.015 in.) diam wire between adherends prior to bonding can be

produced at specific locations in a bonded joint and with the desired sizes to be detected in an x-ray radiograph.

The methods 3, 4, and 5 standards (Fig. 47) are useful for manual bond test methods (such as Fokker, Harmonic, and 210

bond test applications) and for pulse-echo ultrasonics. The holes at the edge and center of the method 3 standard are

required if pinch-off of adhesive occurs near the edge with thicker adhesive near the center of the standard. For maximum

sensitivity, a bond joint should be inspected from both sides, and method 3 and 4 standards are used. If the joint can be

inspected from only one side, the method 5 standard is used. Holes are drilled after the part is bonded.

Fig. 47 Additional methods for fabricating unbond reference standards. (a) Me

thod 3 uses three flat,

unchamfered aluminum sheets of different thicknesses. This standard can be inspected from either side. Hole

diameters can vary from 6.4 to 13 mm ( to

in.) and are free of adhesive. (b) Method 4 uses two chamfered

aluminum sheets of different thicknesses with drilled holes 9.5 or 13 mm ( or in.)

in diameter (free of

adhesive) and edge voids made by inserting Teflon shims of the size shown. Shims must be removed after

bonding. This standard can be inspected from either side. (c) Method 5 is a flat step standard using a buildup of

various sheet thic

knesses that matches the structure to be inspected. There should be no adhesive at the base

of the holes. This standard should be inspected from the top side only. Dimensions given in inches

Standards containing porosity can be made by the wire insertion method or by bonding a large-area panel 0.6 × 0.6 m (2 ×

2 ft, minimum) at low pressure (275 kPa, or 40 psi, maximum). In bonding with a mat (nonwoven) carrier, the adhesive

pinches off at the outer edge, producing a picture frame of gross porosity close to the edge. Scattered porosity and some

small voids are produced throughout the rest of the panel. After ultrasonic C-scan or radiographic inspection,

representative specimens or small standards can be removed from the panel. The methods for producing voids and

porosity are useful for creating defects at specific locations in slow-cycle fatigue test specimens. This type of specimen is

used to study the effects of defects under load, cycle rate, or environment.

Burned-Adhesive Standards. Method 6 (Fig. 48) is used to fabricate a standard representing burned adhesive caused

by improper drilling or sawing of bonded panels. If improper drills, excessive drill speed and feed, or no coolant is used,

burning of the adhesive adjacent to the hole may occur.

Fig. 48 Additional methods for fabrication unbond reference standards. (a) Method 6 presents a burned-

adhesive standard in which (1) is a good-

quality hole, (2) is a slightly burned adhesive, (3) is a grossly burned

adhesive, and (4) is a grossly burned edge caused by abusive band sawing. (b) Method 7 becomes a variable-

quality standard by inserting tapered shims between face and back sheet before bonding. Three 13 mm (

in.)

wide shims are shown in place. Tapered shims cause increasing porosity going from right to left. Half of the

face sheet can be removed after inspection to reveal the condition of the adhesive. Dimensions given in inches

Variable-Quality Standards. Method 7 (Fig. 48) is a way to fabricate a variable-quality standard. Through the use of

tapered shims, good quality is obtained at the thin end and poor quality at the thick end. Figure 49 shows a positive

reproduction made from an x-ray negative of AF-55 adhesive in a tapered standard. For visual correlation, only the right-

hand upper skin need be removed from the panel, thus revealing the variable adhesive quality. The left-hand side is kept

intact to determine variations in NDT instrument response to different adhesive quality conditions.

Fig. 49 Positive print from x-ray negative of tapered-shim variable-quality standard shown as method 7 in

Fig.

Calibration of Bond Testers. When the size of the unbond or void is not of primary concern or to calibrate

instruments to a bond/no-bond response, the following standards prove adequate. These standards are especially useful for

resonant-type instruments such as the Fokker, Harmonic, Sondicator, and 210 bond testers. They relate well to the bonded

part when a 120 to 175 °C (250 to 350 °F) curing epoxy or epoxy phenolic resin system, which does not attenuate the

ultrasonic energy, is used. Basically, the standards are made by cutting 38 mm (1.5 in.) diam holes in a 13 mm ( in.)

thick plywood panel and then bonding 75 × 75 mm (3 × 3 in.) aluminum sheets, using room-temperature curing

adhesives, so that the center of the sheet is centered over the holes in the plywood. The thickness of the aluminum sheets

is chosen to match the thickness of the aluminum adherends of the test part. The number of sheets is governed by the

number of adherends in the test part. When the aluminum sheet is bonded to a wooden base, it does not resonate freely

and reacts similarly to unbonds in the part being inspected.

Honeycomb Reference Standards. Method 1 or 2 (Fig. 46) is used to create a void or unbond reference standard for

honeycomb. For honeycomb standards, the unbonds are created between the adhesive-to-skin and adhesive-to-core

interfaces. Skin-to-adhesive voids or unbonds are usually easier to detect than adhesive-to-core unbonds. Another way to

prepare a void or unbond standard for a honeycomb panel is illustrated in Fig. 50. This standard is useful when attempting

to detect unbonds on both sides of the panel if inspection can be performed from only one side.

Fig. 50

Construction details for making a honeycomb standard for voids and unbonds. (a) Back skin and core

cut out to desired diameter. Cutting through face sheet adhesive shows skin-to-

adhesive unbond; cutting to the

adhesive shows adhesive-to-core unbond. (b) Cutting through back skin first to the adhesive to show skin-to-

adhesive unbond and then through the adhesive to show adhesive-to-core unbond. Note that Tefl

on disks can

be used to help separate adhesive from skins at cutouts. All tests are to be made from the face skin side.

Honeycomb crushed core standards are made by inserting 25 mm (1 in.) diam, 0.13 and 0.25 mm (0.005 and

0.010 in.) thick metal disks between the skin and core and then applying pressure. The disks are removed and the panel is

bonded. The 0.13 mm (0.005 in.) crushing is very subtle and is difficult to detect radiographically. The 0.25 mm (0.010

in.) crushing is easily detectable.

Water intrusion standards are produced by hot-bonding one skin to the core, adding various quantities of water into

the cells, and then cold-bonding the second facing sheet in place.

Core-splice standards are made by cutting the edges of the core with a taper and then applying foaming adhesive to

one side of the splice before bonding. Voids in the foaming adhesive are detectable by low-kilovoltage radiography. Core-

to-closure voids are made by eliminating foaming adhesive at small areas along the joint.

Separator sheet standards are produced by bonding-in a desired size of the material at the skin-to-adhesive and

adhesive-to core interfaces.

A blown core standard can be made by mechanically damaging the core prior to bonding.

Corroded core or core-to-adhesive corrosion standards can be made by bonding only one skin to the core

(open-face honeycomb) and then alternately spraying the honeycomb side in a salt spray system. The degree of corrosion

can be monitored visually. If it is desired to close up the panel, the second facing sheet can be cold-bonded to the other

side of the core.

Substitute Standards. When the ideal standards are not available, other standards of construction similar to that of the

test part can be used. However, honeycomb standards should not be used for metal-to-metal laminate structure, and vice

versa.

Inspection Without Standards. When standards are not available, a known undamaged area can be used as the

standard. Repaired areas are inspected by comparing them to an unrepaired area. The instrument reading may change

because of structural changes resulting from the repair. Repeated inspection scans must be conducted under various

instrument settings, and the inspection should be verified by other methods and/or instruments. A knowledge of the part

configuration (for example, number of bond lines, adhesive thickness, skin thickness, and so on) is essential for

interpreting the test results.

References cited in this section

1. M.T. Clark, "Definition and Non-Destructive Detection of Critical Adhesive Bond-Line Flaws," AFML-

TR-78-108, U.S. Air Force Materials Laboratory, 1978

P. Bijlmer and R.J. Schliekelmann, The Relation of Surface Condition After Pretreatment to Bondability of

Aluminum Alloys, SAMPE Q., Oct 1973

10.

D. Hagemaier and R. Fassbender, Nondestructive Testing of Adhesive Bonded Structures, SAMPE Q.,

Vol

9 (No. 4), 1978

32.

"Adhesive Bonding (Structural) for Aerospace Systems, Requirements for," MIL-A-83377

33.

"Adhesive Bonded Aluminum Honeycomb Sandwich Structure, Acceptance Criteria," MIL-A-83376

34.

"Inspection Requirements, Nondestructive: For Aircraft Materials and Parts," MIL-I-6870 (ASG)

35.

"Nondestructive Testing Personnel Qualification and Certification," MIL-STD-410

Nondestructive Inspection of Adhesive-Bonded Joints

Donald J. Hagemaier, Douglas Aircraft Company, McDonnell Douglas Corporation

Evaluation and Correlation of Inspection Results

The experience, confidence, and effectiveness of NDI technicians and engineering personnel are gained by correlating

NDI results with destructive testing. Comparisons can be made between the NDI results and the actual size, shape,

location, and type of defect. Mistakes will occur in the early stages of the bonding program, but if accurate records,

photographs, and sketches are kept, they can serve as effective training devices. Sections of defective parts should be kept

for training purposes. All these items add to the confidence of the inspector in evaluating and accurately reporting defects.

They also add to the engineer's confidence in the inspector and in the data reported by the inspector.

Procedure

Before new adhesive-bonded assemblies are fabricated on a mass production basis, the first assembly is evaluated by

numerous methods. Inspections are performed and NDT results are evaluated to detect any variations in the part that are

indicative of discontinuities. The part is then cut up, and special precautions are taken to section through the

discontinuities or to separate the joint at the discontinuity to identify and verify the NDI results. To determine which NDT

method(s) detected the discontinuity, the NDI technicians use different color markings--for example, blue for x-ray, green

for ultrasonic, red for a bond tester, and so on. These markings can be affixed to the surface of the part or placed on a

transparent overlay of the part. When the part is sectioned, the results will identify which NDT method(s) correctly

identified the location, size, shape, and type of discontinuity. Although it is expensive, cutting the adhesive-bonded

assemblies is sometimes the only way to determine the nature of the discontinuity or to verify its existence. All inspection

results should be carefully documented for future reference.

Once a highly confident inspection technique is established, the remaining defective assemblies can be used for

developing repair procedures. The original NDI techniques can then be used for determining if the repairs are effective. In

some cases, defective parts are consistently produced, and the material, production process, or design must be changed to

correct the condition. This condition is usually evident when the same type of defect occurs in the same location in all

manufactured parts.

Reliable feedback from manufacturing personnel is beneficial to the NDI technician. This information can include such

items as too high or too low an oven temperature, incorrect heat-up rate, leaks in the vacuum bag, additional layers of

adhesive added at a particular spot in the assembly, and so on. With this information, the NDI technician is better

prepared to inspect a particular group of parts.

Destructive Correlations of NDI Results

Following is a typical example of correlating NDI results by destructive testing. In this example, a metal-to-metal joint

containing a bonded doubler and longerons and frames was rejected for numerous voids and porosity. The initial

inspection was performed using the Fokker bond tester, and the defective areas were marked on the surface of the part, as

shown in Fig. 51(a). The inspection results were verified using the 210 bond tester. The part was then sent out for thermal

neutron radiography. There was excellent correlation between the Fokker bond test results and the neutrograph (Fig. 51b).

Fig. 51 Correlation of NDI res

ults. The same specimen was inspected by (a) Fokker bond tester and (b)

neutron radiography. See text for discussion.

The panel was then ultrasonically C-scanned using the through transmission technique. Again, the correlation was

excellent, as shown in Fig. 52. A plastic overlay was made to record the Fokker bond tester results.

Fig. 52 Ultrasonic C-scan recording of panel in Fig. 51 showing voids and porosity.

Through transmission

method was used with low-gain (31 dB).

The part was then subjected to chemical milling to dissolve the aluminum adherends, leaving the adhesive layer intact.

The adhesive thickness was measured and recorded. The destructive to nondestructive correlation was again excellent and

revealed that the adhesive thickness in and around the defects was greater (0.33 to 0.46 mm, or 0.013 to 0.018 in.) than in

the nondefective areas (0.18 to 0.28 mm, or 0.007 to 0.011 in.). In addition to the correlations noted above, lap shear

specimens could have been made through the different defect conditions, and variations in bond strength could have been

determined.

Because of the variety of defects that can occur in honeycomb assemblies, defects detected by one test method should be

verified by an additional, but similar test method. Time and facilities permitting, the assembly should also be x-rayed in

the defective area.

Ultrasonic Wave Interference Effects

In the ultrasonic inspection of adhesive-bonded joints, the following two conditions may cause ultrasonic wave

interference (Fig. 24) effects:

• When inspecting bond lines under tapered metal doublers

• When inspecting close to the edge of a bonded laminate and the adhesive is thin due to pinch-off (

Ref

10)