Kutz M. Handbook of materials selection

Подождите немного. Документ загружается.

688 NONDESTRUCTIVE INSPECTION

Fig. 32 Crack nucleation and growth to fracture.

Fig. 33 Life extension by inspection and removal of cracks that initiate due to fatigue.

Structural integrity can be assured by periodic inspection to remove (replace/

repair) components that have developed cracks and thereby provide a quantiﬁed

continuing safe life. The inspection/maintenance interval is a function of the

service life and the capability of nondestructive inspection capabilities to reliably

detect components with cracks or cracklike ﬂaws.

Retirement for Cause. The concept of ‘‘retirement for cause’’ (RFC) is to

discard a component when the probability of its failure reaches unacceptable

levels. This scenario has enabled both knowledge-based service life extension

and signiﬁcant cost savings for many modern engineering systems and compo-

nents. The periodic inspection and removal of ﬂawed components concept is

shown schematically in Fig. 34.

The RFC philosophy means that a structure or ﬂeet of structures can be

managed by knowledge of crack initiation and growth characteristics. The phi-

losophy assumes that components that contain small cracks, at some size level,

have been removed at the time of production acceptance. Components/structures

that do enter service are subjected to cyclic loading (fatigue) during their service

lifetime. The mean time to initiation of a fatigue crack in service varies widely,

as shown in Fig. 33. As the components age in service, a few will be expected

to initiate and grow fatigue cracks earlier than the general population. Inspection

is then performed on the group of similar aged structures or components, and

those that contain size cracks above a certain size are removed from service.

8 CONSIDERATIONS FOR INSPECTABILITY IN MATERIALS SELECTION 689

No Defect

Return to Service

Defect Found

Retire from Service

Fig. 34 Retirement for cause life-cycle maintenance incorporating quantitative inspection.

The remaining components are returned to service and continue in useful life to

the next calculated inspection interval. The same group is reinspected at the next

interval, and those containing cracks above a certain size are removed with the

remainder again returned to service. By continuing this life-cycle management

process, the maximum useful life of all the components or ﬂeet of structures

may be realized. The key to this process is quantiﬁcation of the inspection

process to determine the ﬂaw size that can be reliably detected and removed

from service. The calculated inspection interval is based on probable growth of

any detected ﬂaws before the next inspection cycle. As an added safety factor,

the inspection interval is usually set at one-third to one-half of the calculated

time required for a ﬂaw to grow to a critical failure inducing size. In this manner,

structural integrity can be maintained and service life can be extended until the

component is retired for cause—crack initiation, wear, corrosion, dimensional

changes, etc.

Damage-tolerant design and life-cycle management are dependent on detec-

tion and removal of cracks and cracklike stress risers during production and

during life. Such detection must necessarily be nondestructive and quantiﬁcation

of detection capabilities are required.

Requirements for Quantitative Nondestructive Inspection

Nondestructive inspections/evaluations (NDI/E) have traditionally been rou-

tinely applied to a variety of engineering materials, components, structures, and

systems to provide conﬁdence in structural integrity. Development and imple-

mentation of NDI/E methods were often prompted by a catastrophic failure of

an important engineering system. This experience base of previous failures was

integrated into ‘‘engineering art,’’ and NDI/E provided a deterministic but un-

quantiﬁed measure of conﬁdence in structural integrity and continuing service.

The advent and implementation of fracture mechanics in engineering design

established the need for quantiﬁcation of the capabilities of applied NDI/E pro-

cedures. That need moved nondestructive inspection requirements from a trades-

man’s art to an integrated engineering technology.

Requirements to establish a reliable nondestructive inspection include the fol-

lowing:

690 NONDESTRUCTIVE INSPECTION

1. Access. Access to perform an inspection is essential and must be

considered as a primary engineering design parameter. An inspection can-

not be performed unless key areas of the component are accessible. This

may require disassembly in some cases and must be considered as a part

of the life-cycle maintenance procedures.

2. Capability. The inspection method must be capable of detecting the re-

quired ﬂaw sizes to a deﬁned reliability and conﬁdence levels. Selection

of the applicable inspection method is primarily the responsibility of the

NDI engineer. A design cannot be supported if the critical ﬂaw size is

below the detection capability level.

3. Reliable. The reliability of an inspection method is dependent, in part,

on the reproducibility of the measurements being made by the NDI pro-

cedure. The reproducibility of a procedure is, in turn, largely dependent

on the ‘‘calibration’’/setup used for the NDI procedure. Additionally, the

process must be repeatable, i.e., the NDI procedure is dependent on pro-

cess control rigor in application of the NDI method/process.

4. Human Factors. Since the reproducibility and repeatability of an NDI

procedure are dependent on rigor in process control, the skill of the hu-

man operator is an important factor in NDI application. Most NDI

methods/procedures require dexterity in application and detection/pattern

recognition skills for ‘‘interpretation’’ and decision processing. Most NDI

methods produce indirect measurements of a property or conﬁguration

and relates those measurement to a desired property or behavior. Proﬁ-

ciency in such relationships are a product of the knowledge and skill of

the NDI engineer and the knowledge and skill of the inspector/operator.

8.4 Quantiﬁcation of NDE Capabilities

Nondestructive inspection methods have a long history of application that range

from simple visual inspections to complex ultrasonic and eddy current inspec-

tions. Since most inspections were originally practiced as an art, the craft was

subject to a wide range of assumptions, exaggerations, misinformation, and mis-

applications. The requirement to quantify the detection capability has reduced

many of those deﬁciencies but has generated new barriers from practitioners

who resist change. Without discipline and quantiﬁcation, many of inspection

procedures must be regarded as ‘‘inspection ceremonies.’’ This applies to many

established procedures that were generated and perpetuated by ‘‘speciﬁcation.’’

The issue at hand is not ‘‘the smallest ﬂaw that has been detected, but the largest

ﬂaw that has been missed’’ by an inspection procedure or operator.

The methodology that is applied to quantiﬁcation of nondestructive

inspection/evaluation capability is known as a ‘‘probability of detection (POD)’’

assessment. The single-valued inspection capability parameter that is provided

by the POD method is a crack size that can be detected (reliably) by application

of an inspection procedure. That value is often quoted as the 90/95 or the a

NDE

(nondestructive evaluation) capability/parameter and is the value that is used in

engineering design for use of nondestructive inspection as an element of vali-

dating component structural integrity.

8 CONSIDERATIONS FOR INSPECTABILITY IN MATERIALS SELECTION 691

Fig. 35

8.5 Probability of Detection

The probability of detection (POD) methodology was generated to provide some

of the conﬁdence level in ﬂaw detection capability that is used in engineering

design. The basic guidelines and values for materials properties are available

from MIL-HDBK-5.

The number of parameters and variances inherent to the

application of an inspection procedure greatly exceed the materials properties

variances, and the POD method is intended to account for the end-to-end vari-

ances and to provide a detection capability as a function of ﬂaw size. Figure 35

is an example of a typical POD curve. The a

NDE

value is the point at which the

POD curve crosses the 90% POD threshold. This point has been selected by

convention and is the basis for most modern design applications.

A POD curve is generated by passing a large number of ﬂaws of varying size

through an inspection process/procedure and recording either the ‘‘hit/miss’’

(detection or nondetection) data or the scalar output of the procedure. The data

are then ﬁt to a mathematical model

to provide rigor and consistency in the

analysis and to produce the POD result. Typically 60 or more cracks of varying

sizes are required to produce the POD curve.

The POD method is speciﬁc to the ﬂaw/artifact variables, inspection method,

the inspection procedure, the inspection equipment, the inspection process, cal-

ibration, detection threshold used in the decision processing, test coupon/object

variables, geometry, inspection environment, and human factors and is thus an

end-to-end process measurement value. If any of the parameters change, the

POD process must be repeated for rigor in application. The importance of pro-

cess control in actual implementation is self-evident. Generation of a POD curve

is expensive. Methods of reducing cost are constant considerations. In engi-

neering application of the deterministic parameters, reference data are often used

as the starting point for analysis. The NDE Capabilities Data Book

provides a

handbook reference of POD curves that have been generated for various mate-

rials, inspection processes, and inspection conditions. This reference is a starting

point when considering incorporation of an inspection requirement into a design.

For damage-tolerant designs, it is important to consider the capabilities of

various nondestructive inspection methods for both purposes of cost of the in-

692 NONDESTRUCTIVE INSPECTION

spection and the cost/beneﬁts ratio of inspection capabilities on life-cycle man-

agement. For example, ﬂuorescent penetrant inspection is a widely used

inspection method due to its inherent capabilities, wide area coverage, and eco-

nomical cost of application. In general application, the capability of ﬂuorescent

penetrant processes is the detection of 0.100- to 0.150-in. long cracks. Fluores-

cent penetrant inspection is capable of detection cracks smaller than 0.100–0.150

in. in size but requires more discipline and control with corresponding increase

in application cost. The lower limit for practical penetrant inspection is 0.030

in. This level requires extraordinary process control and discipline and is there-

fore not applicable to most components. If detection of smaller cracks than this

is required, then it is necessary to apply alternate inspection methods such as

eddy current. Eddy current inspection is in turn a focused inspection method

and therefore has a higher cost. Regardless of the NDI method selected, it may

be necessary to assess and demonstrate detection capability using the POD

method (for fracture-critical applications). The demonstrated capability is spe-

ciﬁc to a facility, equipment, procedure, and operator. Only operators so qualiﬁed

may be allowed to perform inspections of fracture-critical components.

False conﬁdence in crack detection capabilities is the norm for many inspec-

tion facilities and operations. Since crack-to-crack variance is one of the POD

process parameters, small cracks may be detected and larger cracks missed.

Unfortunately, statistical variance in ‘‘calibration’’ and process control is often

a cause for variance in detection. This fact reinforces the requirement to assure

a stable NDI procedure (calibration and process control) before an assessment

by the POD method is initiated.

The POD method of assessment has proven useful, not only as tool to meet

design engineering requirements but also as a tool for a quantiﬁable comparison

of the performance of various NDI procedures, to qualify inspection facilities

and personnel, and to validate critical procedures in difﬁcult inspection appli-

cations.

Lessons Learned from POD Assessments

A large number of POD assessments have been completed and results docu-

mented in various applications. The primary requirement is to have a stable

calibration and process controls before a POD assessment is initiated. Rigid

calibration and process controls are necessary in application to provide conﬁ-

dence that the same level of detection has been maintained. Reliable application

demands reproducibility and repeatability to maintain a capability. The POD

process is sensitive to a wide range of parameters. These include:

●

Flaw (artifact) variables

●

Test object variables

●

Inspection method variables

●

Materials variables

●

Equipment variables

●

Inspection procedure variables

●

Inspection process variables

●

Calibration variables

9 SUMMARY 693

●

Acceptance criteria/decision variables

●

Human factors

Individual NDI procedures and applications considerations and identiﬁable

sources of variance supplement this list. Variance in a single parameter will vary

the POD and thus the capability of the NDI procedure.

8.6 General Process Control Is Required in All NDI Applications

Inspection for fracture control is a specialized application and reﬂects the full

capabilities and beneﬁts of quantitative NDI. Most engineering components are

not fracture critical and do not require the rigor of demonstrating ﬂaw detection

capabilities. Rigorous quality control in these applications is, however, required

due to the dependence of NDI capabilities on general structural integrity and

life management. Cost should not be an issue since it costs no more to perform

in an undisciplined (unreliable) manner than it does to perform in a discipline

(reliable) manner. The labor may cost a small amount more, but the materials

and equipment are the same. The results can vary signiﬁcantly in terms of legal

and economic exposure.

Care must be exercised in establishing engineering requirements to assure

that the required level of detection is applied. Specifying a small ﬂaw size does

not necessarily provide a superior or applicable inspection, but may add unnec-

essarily to the inspection cost. A common practice in the past has been to specify

‘‘no defects/no cracks/or no ﬂaws.’’ The practice arose for fracture-critical com-

ponents where the engineer wanted to know about every ﬂaw detected and thus

be able to review and disposition all ﬂaws found. This is an unacceptable practice

since it leaves the level of inspection to the NDI engineer and no qualitative

value is provided. An acceptable practice is to specify a critical ﬂaw size and a

reportable ﬂaw size. This imposes the requirement to demonstrate detection

capability at the reportable ﬂaw size level and to reject at the critical ﬂaw size

level. This method provides a quantiﬁed margin for the inspection process and

provides a higher conﬁdence in results. The reportable ﬂaw size level is espe-

cially useful for automated inspection processes since it provides a mechanism

for reassessment and trend monitoring of inspection data for purposes of com-

ponent system performance assessments. The melding of engineering design and

NDI to provide quantitative analysis, speciﬁcation, measurement, and reporting

has supported engineering efﬁciency and conﬁdence levels that are the hallmark

of modern engineering materials, components, structures, and systems.

9 SUMMARY

The intent of this chapter is to provide the reader with a basic understanding of

the fundamentals of nondestructive inspection both from the fundamental science

and reliability standpoints. This latter point is important since the design process

often requires the engineer to specify a reliability of a component. While all

designers realize that a material or manufacturing process speciﬁcation signiﬁ-

cantly affects the reliability of a design, the inclusion of an inspection procedure

is also a signiﬁcant factor in reliability. In choosing a particular material, the

designer automatically limits the inspection choices. For example, only ferro-

magnetic alloys such as cast iron and most steels are inspectable with magnetic

694 NONDESTRUCTIVE INSPECTION

particle inspection. Another example might be the limited ability to detect deeply

buried ﬂaws using thermal inspection methods on highly conductive aluminum

components. Therefore, the design engineer should be cognizant of the limita-

tions that material choice places on the applicable inspection methods. In these

days of product liability law suits, an understanding of inspection technology

has become another critical component of the designer’s required skill set. The

authors hope that this chapter provides a ﬁrst step to acquainting design and

process engineers with some of the basic considerations necessary to achieve

reliable, cost-effective products. If additional information is needed, the reader

is referred to the references or to Internet sites cited.

APPENDIX A: ULTRASONIC PROPERTIES OF COMMON MATERIALS

Liquids

Liquid

(20⬚C unless noted)

Longitudinal

Wave Speed

(⫻10

cm/s)

Density

(g/cm

)

Acetic acid 1.173 1.049

Acetone 1.192 0.792

Amyl acetate (26

⬚C) 1.168 0.879

Aniline 0.656 1.022

Benzene 1.326 0.879

Blood (horse) (37

⬚C) 1.571 —

Bromoform 0.928 2.890

n-Buytl alcohol 1.268 0.810

Caprylic acid 1.331 0.910

Carbon disulﬁde 1.158 1.263

Carbon tetrachloride 0.938 1.595

Cloroform 1.005 1.498

Formaldehyde (25

⬚C) 1.587 0.815

Gasoline (34

⬚C) 1.25 0.803

Glycerin 1.923 1.261

Kerosene (25

⬚C) 1.315 0.82

Mercury 1.451 13.546

Methyl alcohol 1.123 0.796

Oils

Campor (25

⬚C) 1.390 —

Castor 1.500 0.969

Condenser 1.432 —

Olive (22

⬚C) 1.440 0.918

SAE 20 1.74 0.87

Sperm (32

⬚C) 1.411 —

Transformer 1.38 0.92

Oleic acid 1.333 0.873

n-Pentane 1.044 —

Silicon tetrachloride (30

⬚C) 0.766 1.483

Toluene 1.328 0.67

Water (distilled) 1.482 1.00

m-Xylene 1.340 0.864

APPENDIX A: ULTRASONIC PROPERTIES OF COMMON MATERIALS 695

Solids: Metals

Metal

(20⬚C unless noted)

Longitudinal

Wave Speed

(⫻10

cm/s)

Transverse

Wave Speed

(⫻10

cm/s)

Density

(g/cm

)

Aluminum

Al (1100) 6.31 3.08 2.71

Al (2014) 6.37 3.07 2.80

Al (2024-T4) 6.37 3.16 2.77

Al (2117-T4) 6.50 3.12 2.80

Al (6061-T6) 6.31 3.14 2.70

Bearing babbit 2.30 — 10.1

Beryllium 12.890 8.880 1.82

Bismuth 2.18 1.10 9.80

Brass (70% Cu & 30% Zn) 4.37 2.10 8.50

Brass (naval) 4.43 2.12 8.42

Bronze (5% P) 3.53 2.32 8.86

Cadmium 2.78 1.50 8.64

Cerium 2.424 1.415 6.77

Chromium 6.608 4.005 7.20

Cobalt 5.88 3.10 8.90

Columbium 4.92 2.10 8.57

Constantan 5.177 2.625 8.88

Copper 4.759 2.325 8.93

Copper (110) 4.70 2.26 8.9

Dysprosium 2.296 1.733 8.53

Erbium 2.064 1.807 9.06

Europium 1.931 1.237 5.17

Gadolinium 2.927 1.677 7.89

Germanium 5.18 3.10 5.47

Gold 3.24 1.20 19.32

Hafnium 2.84 — 13.3

Hastelloy X 5.79 2.74 8.23

Hastelloy C 5.84 2.90 8.94

Holmium 3.089 1.729 8.80

Indium 2.56 0.74 7.30

Invar 4.657 2.658 —

Lanthanum 2.362 1.486 6.16

Lead 2.160 0.700 11.34

Lead (96% Pb & 6% Sb) 2.16 0.81 10.88

Lutetium 2.765 1.574 9.85

Magnesium 5.823 3.163 1.74

AM-35 5.79 3.10 1.74

FS-1 5.47 3.03 1.69

J-1 5.67 3.01 1.70

M1A 5.74 3.10 1.76

O-1 5.80 3.04 1.72

Manganese 4.66 2.35 7.39

Manganin 4.66 2.35 8.40

696 NONDESTRUCTIVE INSPECTION

Metal

(20⬚C unless noted)

Longitudinal

Wave Speed

(⫻10

cm/s)

Transverse

Wave Speed

(⫻10

cm/s)

Density

(g/cm

)

Molybdenum 6.29 3.35 10.2

Nickel

Pure 5.63 2.96 8.88

Inconel 5.82 3.02 8.5

Inconel (X-750) 5.94 3.12 8.3

Inconel (wrought) 7.82 3.02 8.25

Monel 5.35 2.72 8.83

Monel (wrought) 6.02 2.72 8.83

Silver–nickel (18%) 4.62 2.32 8.75

German silver 4.76 2.16 8.40

Neodymium 2.751 1.502 7.10

Platinum 3.96 1.67 21.4

Potassium 2.47 1.22 0.862

Praseodymium 2.639 1.437 6.75

Samarium 2.875 1.618 7.48

Silver 3.60 1.59 10.5

Sodium 3.03 1.70 0.97

Steel

1020 5.89 3.24 7.71

1095 5.90 3.19 7.80

4150, Rc14 5.86 2.79 7.84

4150, Rc 18 5.88 3.18 7.82

4150, Rc 43 5.87 3.20 7.81

4150, Rc 64 5.83 2.77 7.80

4340 5.85 3.24 7.80

52100 Annealed 5.99 3.27 7.83

52100 Hardened 5.89 3.20 7.8

D6 Tool Steel Annealed 6.14 3.31 7.7

Stainless Steels

302 5.66 3.12 7.9

304L 5.64 3.07 7.9

347 5.74 3.10 7.91

410 5.39 2.99 7.67

430 6.01 3.36 7.7

Tantalum 4.10 2.90 16.6

Thorium 2.94 1.56 11.3

Thulium 3.009 1.809 9.29

Tin 3.32 1.67 7.29

Titanium (Ti-6-4) 6.18 3.29 4.50

Tungsten

Annealed 5.221 2.887 19.25

Drawn 5.410 2.640 19.25

Uranium 3.37 1.98 18.7

Vanadium 6.023 2.774 6.03

Ytterbium 1.946 1.193 6.99

Yttrium 4.10 2.38 4.34

Zinc 4.187 2.421 7.10

Zirconium 4.65 2.25 6.48

APPENDIX A: ULTRASONIC PROPERTIES OF COMMON MATERIALS 697

Solids: Ceramics

Ceramic

(20⬚C unless noted)

Longitudinal

Wave Speed

(⫻10

cm/s)

Transverse

Wave Speed

(⫻10

cm/s)

Density

(g/cm

)

Aluminum oxide 10.84 6.36 3.98

Barium nitrate 4.12 2.28 3.24

Barium titanate 5.65 3.03 5.5

Bone (human tibia) 4.00 1.97 1.7–2.0

Cobalt oxide 6.56 3.32 6.39

Concrete 4.25–5.25 — 2.60

Glass

Crown 5.66 3.42 2.50

Flint 4.26 2.56 3.60

Lead 3.76 2.22 4.6

Plate 5.77 3.43 2.51

Pyrex 5.57 3.44 2.23

Soft 5.40 — 2.40

Granite 3.95 — 2.75

Graphite 4.21 2.03 2.25

Ice (

⫺16⬚C) 3.83 1.92 0.94

Indium antimonide 3.59 1.91 —

Lead nitrate 3.28 1.47 4.53

Lithium ﬂuoride 6.56 3.84 2.64

Magnesium oxide 9.32 5.76 3.58

Manganese oxide 6.68 3.59 5.37

Nickel oxide 6.60 3.68 6.79

Porcelain 5.34 3.12 2.41

Quartz

Crystalline 5.73 — 2.65

Fused 5.57 3.52 2.60

Polycrystalline 5.75 3.72 2.65

Rock salt 4.60 2.71 2.17

Titanium dioxide (Rutile) 8.72 4.44 4.26

Sandstone 2.92 1.84 2.2–2.4

Sapphire (c-axis) 11.91 7.66 3.97

Slate 4.50 — 2.6–3.3

Titanium carbide 8.27 5.16 5.15

Tourmaline (Z-cut) 7.54 — 3.10

Tungsten carbide 6.66 3.98 10.15

Yttrium iron garnet 7.29 4.41 5.17

Zinc sulﬁde 5.17 2.42 4.02

Zinc oxide 6.00 2.84 5.61

Solids: Polymers

Polymer

(20⬚C Unless Noted)

Longitudinal

Wave Speed

(⫻10

cm/s)

Transverse

Wave Speed

(⫻10

cm/s)

Density

(g/cm

)

Acrylic resin 2.67 1.12 1.18

Bakelite 2.59 — 1.40

Buytl rubber 1.99 — 1.13

Cellulose acetate 2.45 — 1.30