NASA

/ CR-2000-209847

Structural
Concepts

Design Methodology
of Uncertainty

K. Y. Lin, Jiaji Du, and David

Department
University

February

of Aeronautics
of Washington,

2000

Rusk

and Astronautics
Seattle,

Washington

Based

on

The NASA

STI Program

Office

CONFERENCE

Since its founding, NASA has been dedicated

to the advancement
of aeronautics
and space
science. The NASA Scientific and Technical

SPECIAL PUBLICATION.
Scientific,
technical, or historical information
from
NASA programs,
projects, and missions,
often concerned
with subjects having
substantial
public interest.

The NASA STI Program Office is operated by

Langley Research Center, the lead center for
NASA's scientific and technical information.
The NASA STI Program Office provides
access to the NASA STI Database, the largest
collection of aeronautical
and space science
STI in the world. The Program Office is also
NASA's institutional
mechanism
for

TECHNICAL
PUBLICATION.
Reports
of completed
research or a major
significant phase of research that
present the results of NASA programs
and include extensive data or theoretical
analysis. Includes compilations
of
significant scientific and technical data
and information
deemed to be of
continuing
reference value. NASA
counterpart
of peer-reviewed
formal
professional
papers, but having less
stringent limitations
on manuscript
length and extent of graphic
presentations.
TECHNICAL

MEMORANDUM.

Scientific and technical findings that are

preliminary
or of specialized
interest,
e.g., quick release reports, working
papers, and bibliographies
that contain
minimal annotation.
Does not contain
extensive

analysis.

CONTRACTOR

REPORT.

Scientific

technical findings by NASA-sponsored

contractors
and grantees.

and

PUBLICATION.

Collected papers from scientific and

technical conferences,
symposia,
seminars, or other meetings sponsored
or co-sponsored
by NASA.

Information
(STI) Program Office plays a key
part in helping NASA maintain this important
role.

disseminating
the results of its research and
development
activities. These results are
published
by NASA in the NASA STI Report
Series, which includes the following report
types:

... in Profile

TECHNICAL
TRANSLATION.
language translations
of foreign
scientific and technical material
pertinent

to NASA's

English-

mission.

Specialized
services that complement
the
STI Program Office's diverse offerings
include creating custom thesauri, building
customized
databases,
organizing
and
publishing
research results ... even
providing
videos.
For more information
Program

about

the NASA

STI

Office, see the following:

Access the NASA STI Program

Page at http://www.sti.nasa.gov

Home

E-mail your question

help@sti.nasa.gov

Fax your question to the NASA STI

Help Desk at (301) 621-0134

Phone the NASA STI Help Desk at (301)

621-0390

via the Internet

to

Write to:
NASA STI Help Desk
NASA Center for AeroSpace
7121 Standard Drive
Hanover,

MD 21076-1320

Information

NASA

/ CR-2000-209847

Structural
Concepts
K. Y. Lin,

Jiaji

Department

Space

and

David

of Aeronautics

University

National

Du,

Design Methodology
of Uncertainty

of Washington,

Aeronautics

and

Based

on

Rusk
Astronautics

Seattle,

Washington

and

Administration

Langley Research
Center
Hampton,
Virginia 23681-2199

February

2000

Prepared
for Langley Research
under Grant NAG-I-2055

Center

Available

from:

NASA Center for AeroSpace

7121 Standard Drive
Hanover,

MD 21076-1320

(301) 621-0390

Information

(CASI)

National Technical

Information

5285 Port Royal Road

Springfield,

VA 22161-2171

(703) 605-6000

Service

(NTIS)

FOREWORD

This

report

summarizes

September

30,

The principal
research

under

investigator

assistant.

researcher

and support

project
from

and Dr. Dave

the

Research

period

Center

was Dr. K. Y. Lin.

scientist

Backman

manager

from

of the Boeing

Stroud

of the FAA

of NASA,
are greatly

of May

Grant

David

West

is Dr. W. Jefferson

Dr. Jeff
Swartz

during

Langley

Du, a visiting
Dr. Bjorn

of this research

Ilcewicz

accomplished

the NASA

Dr. Jiaji

The NASA

work

of this program

for this project.

this project.

Dr. Larry

1999,

the

No.

Rusk

Virginia

Dr. Bjorn

1998

NAG-I-2055.

was the graduate

University,

Company
Stroud.

16,

was

the

also contributed

Invaluable
Backman

appreciated.

to

discussions
of Boeing,

ABSTRACT

The

principal

tolerant

goal

aircraft

past service

of this

structures

using

experience.
for a given

is defined

as the compliment

The cumulative
of Safety"

The

an existing

strength
load

in-service

and

of the

new

the

required

composite

sandwich

derived

data from
component,

structural
safety

sizing

parameters

approach

broad

potential

has

design,
level

above

definition

existing

aircraft,

(sizing).

The

the flaw

consists

design

with results

to the relative

safety

of the

application

to damage-tolerant

uncertainty.

showing

aircraft

event

the critical

following

analyses

was

design.

"Level

The

"Level

in the structure.

configuration

method

types,

than

the baseline

tolerance

damage

inspection

of the discrete

of the

establishing

structural

that incorporates

will not be detected.

at each location

damage

and determining

size larger

is the product

type present

conducting

for various

the structural

structure

for damage

of an equivalent

for a single

flaw

and that

process

of Safety"

of Safety"

exists,

of each damage

the

"Level

that a single

structure

a design

on the concept

"Level

for the entire

from

damage

panel

is based

The discrete

of Safety"

structural

of structural

process

of the

for each flaw

method

is to develop

of the probability

strength

"Level

values

design

collecting
for

structure.

size for residual

program

a definition

The design

of Safety"

flaw

research

steps:

safety

for residual
for a given

demonstrated
the sensitivity
"Level

structural

level

on

a
of

of Safety"
design

with

EXECUTIVE

There

are

at least

using

factors

apply

to aircraft

contain

fundamental

of safety

novel

measured

two

and

that

structural

measure

of various
of safety

throughout

the

aircraft.

damage

structures

past service

experience.

an approach

of an equivalent

"Level

inspection

is defined
flaw

the discrete
location

The

design

method
in-service

an existing

strength

of the

structural
new

To demonstrate

notch
strength

"Level

and

of the

for each

of safety

"Level

flaw

of each

with

and

based

efficiency
a design
"Level

of Safety"

exists,

of

for a single

flaw

size larger

and that the flaw

structure

damage

no

on the concept

that a single

for the entire

the relative

of structural

design

and

be easily

is to develop

a definition

of the structure

cannot

In addition,

program

to

systems,

to determine

level

of the probability

procedures

be difficult

material

aircraft.

discrete

of Safety"

may

reliability

structural

The

strength

values

derived

for each

safety

damage

above

design,
level

will

is the product

type

present

of

at each

aircraft,

conducting

consists
establishing

damage

and determining

of the

the baseline

tolerance

structural

following

steps:

safety

analyses

level

for residual

configuration

for a given

(sizing).

methodology

on a new

as a function
analysis

type.

definition

existing

component,

strength

A two-step

the

data from

structural

the design

for residual
damage.

from

damage

load and the required

analyzed

size for residual

new

it is not possible

aircraft

is studied.

as the compliment

of Safety"

of safety

using

design

in the structure.

collecting
for

of Safety"

use

of this research

to damage-tolerant

The cumulative

"Level

goal

aircraft

procedures

is a consistent

aircraft

that incorporates

not be detected.

there

tolerant

Safety"

than the critical

that

these

safety

principal

for

event

on the

The

process

In this report,

levels

As a result,

options

it is unlikely

First,

configurations,

Second,

component.
design

to traditional

factors.

unconventional

concepts.

for a structural

importance

shortcomings

knockdown

have

SUMMARY

model

The residual

structure,

of damage
was

used

strength

a composite
size

sandwich

for disbond,

to determine
vs. damage

panel

delamination

post-buckling
size

results

were

was
and

residual
used

to

demonstrateapplication of the "Level of Safety" design processesusing two example

problems. The influenceof the structuralsizing parametersonthe overall "Level of Safety"
was alsodemonstratedin the examples. Bayesianstatisticaltools are incorporatedinto the
designmethodto quantify the uncertaintyin the probability data,and to allow post-design
damagedata to be usedto updatethe "Level of Safety" values for the structure. Some
methods of obtaining in-service damage data for the current aircraft fleet have been
suggested. Concernsregarding the calculation of "Level of Safety" values for existing
aircraftcomponentshavealsobeendiscussed.
The definition of structural"Level of Safety",andthe designmethodologyderivedfrom it, is
an extensionof reliability theory andstatisticalanalysistools to the designandmaintenance
of damage-tolerantaircraft structures. The methodpresentsa unified approachto damage
tolerancethat allows a direct comparisonof relative safety betweenaircraft components
using different materials, construction techniques,loading or operational conditions. It
incorporatesplanningfor the serviceinspectionprograminto the designprocess.The useof
Bayesian statistical tools in the "Level of Safety" method provides a mechanismfor
validating the damageassumptionsmadeduring the designprocess,and for reducingthe
level of uncertaintyandrisk overthelife-cycle of the structure.

TABLE

INTRODUCTION
1.1

Background

1.2

Review

of existing

OBJECTIVES

EQUIVALENT

technologies

LEVEL

General

Approach

3.2

Defining

"Level

3.3

Establishing

3.4

Collection

3.5

Application

3.6

Damage

3.7

Discussion

3.8

Mathematical

of Safety".

a Baseline
Data

16

.....................................................

17
17

on Existing

17

......................................................................

Structures

to Advanced

22

..........................................................

Structural

Design

Concepts

23

.....................

................................................................................

Considerations

Introduction

4.2

Material

4.3

Tensile

4.4

Compressive

4.5

Residual

Strength

strength

Case

1: Panels

4.5.2

Case 2: Panels

4.5.3

Case 3: Panels

5.3

Examples

OF EXAMPLE

.............................
STRUCTURE

...........

35

...............................................................................

37

.......................................................................

Honeycomb

with a Disbond

Sandwich

Panels

...................................

...........................................................................

with a Delamination
with Notches

35

35

of the laminates

of Damaged

33

....................................................................

..............................................................................

39
39
40
47
51

...................................................................................................................

55

.....................................................................................................................

55

DEMONSTRATION

Outline

Formulation

.................................................................................

of the Laminates

4.5.1

5.2

DETERMINATION

and Properties

Strength

Introduction

of Safety

31

.................................................................................................................

Systems

5.1

in the Level

25
26

...................................................................................................................

4.1

RESULTS

APPROACH

of Safety

Schemes

STRENGTH

Summary

12

........................................................................................

Level

of Methodology

RESIDUAL

4.7

.................................................................................

OF SAFETY

Size Updating

Discussion

11

.......................................................................................................

of Flaw

4.6

11

.................................................................................................................

3.1

...........................................................................................................
.................................................................................................................

OF CONTENTS

OF DESIGN

METHOD

..............................................................

.................................................................................................................

of Design

Procedures

of Equivalent
AND

Safety

CONCLUSIONS

.....................................................................................
Based

Design

............................................................

.................................................................................

57
57
57
59
65

6.1 Benefitsof anEquivalentLevel of SafetyApproach..................................................

65
6.2 Limitationsof the CurrentFormulation......................................................................
65
6.3 Topicsfor FurtherResearch
........................................................................................
67
APPENDIX.....................................................................................................................
68
REFERENCES..............................................................................................................
123

TABLE

Figure

1. Flow-Chart

Figure

2. Prior and Posterior

Distributions
of Parameter
Alpha Updated
with Measured
Damage Sizes of 3,4,5 Inches ...............................................................................

70

3. Bayesian
Updating
of Detected
Damage Size Distribution
with Measured
Size of 3,4,5 Inches ...............................................................................................

71

Figure

Figure

of Developing

OF FIGURES

Equivalent

Safety

Aircraft

...........................................

4. Three Cases of Damage:

Case 1. Disbond;
Case 2. Delamination;
...............................................................................................................................

Case

Damage

3. Notches
2

Figure

5. Finite

Element

Mesh

for a Sandwich

Panel

with a Circular

Figure

6. Finite

Element

Mesh

for a Sandwich

Panel

with an Elliptical

Figure

7. Verification

Figure

8. Buckling

Figure

9. Case 1: Buckling
Load of a Face Sheet with a Circular Disbond (Variation
Thickness)
.............................................................................................................

Figure

of Finite
Load

10. Buckling

Figure

11. Case

Figure

12. Case

Sequence)

1: Buckling

Stacking
Figure

of a Face

Sequence)

13. Comparison
Isotropic

for Buckling

Disbond

Sheet

under

Uniform

with an Elliptical

Load

of a Face

Sheet

Load

of a Face

Sheet

Disbond

Damage

.................

74

Determination

..............

75

..........................

(Variation

76

in
77

in

Element

Plate with a Circular

Analysis
Open

Result

of a Sandwich

(Variation

1: Residual

in

Disbond

79
(Variation

in Compression

Panel

in Thickness)

Strength

Solution

for an

in Thickness)

Disbond

Loaded

..................................................................
Panel

81

with Analytical
Solution for an
.................................................
82

with a Circular

of a Sandwich

(Variation

with Analytical

in
80

Hole ............................................................

15. Residual

Compression

(Variation

...............................................................................................

of Finite

Strength

78

Disbond

with an Elliptical

Figure

Loaded

73

Pressure

with a Circular

14. Comparison
of Finite Element Analysis
Results
Isotropic Plate with an Elliptical
Through Notch

16. Case

....................

...............................................................................................

Figure

Figure

Load

Damage

.............................................................................................................

1. Buckling

Stacking

Model

of an Elliptical

Load

Thickness)

Element

69

with an Elliptical

in
83

Disbond

................................................

84

Figure17. ResidualStrengthof a SandwichPanelwith a CircularDisbondLoadedin

Compression(Variationin StackingSequence)...................................................
85
Figure18. ResidualStrengthof a SandwichPanelwith anElliptical DisbondLoadedin
Compression(Variationin StackingSequence)...................................................
86
Figure19. Case2: Buckling Load of aFaceSheetwith a CircularDelamination(Variation
in Thickness).........................................................................................................
87
Figure20. Case2: Buckling Load of a FaceSheetwith anElliptical Delamination
(Variationin Thickness)........................................................................................
88
Figure21. Case2: Buckling Load of aFaceSheetwith a CircularDelamination(Variation
in StackingSequence)
...........................................................................................
89
Figure22. Case2: Buckling Load of a FaceSheetwith anElliptical Delamination
(Variationin StackingSequence)
..........................................................................
90
Figure23. Case2: ResidualStrengthof a SandwichPanelwith a CircularDelamination
Loadedin Compression(Variationin Thickness)................................................
91
Figure24. Case2: ResidualStrengthof a SandwichPanelwith anElliptical Delamination
Loadedin Compression(Variationin Thickness)................................................
92
Figure25. Case2: ResidualStrengthof a SandwichPanelwith a CircularDelamination
Loadedin Compression(Variationin StackingSequence)..................................
93
Figure26. Case2: ResidualStrengthof a SandwichPanelwith anElliptical Delamination
Loadedin Compression(Variationin StackingSequence)..................................
94
Figure27. Case3: ResidualStrengthof a SandwichPanelwith a CircularHole on OneFace
SheetLoadedin Tension(Variationin Thickness)...............................................
95
Figure28. Case3: ResidualStrengthof a SandwichPanelwith anElliptical Notch on One
FaceSheetLoadedin Tension(Variationin Thickness)......................................
96
Figure29. Case3: ResidualStrengthof a SandwichPanelwith a CircularThrough-theThicknessHole Loadedin Tension(Variationin Thickness)...............................
97
Figure30. Case3: ResidualStrengthof a SandwichPanelwith anElliptical Through-theThicknessNotch Loadedin Tension(Variationin Thickness).............................
98
Figure31. Case3: ResidualStrengthof a SandwichPanelwith a CircularHole on OneFace
SheetLoadedin Compression(Variationin Thickness).......................................
99
Figure32. Case3: ResidualStrengthof a SandwichPanelwith anElliptical Notch on One
FaceSheetLoadedin Compression(Variationin Thickness)............................
100

Figure33. Case3: ResidualStrengthof a SandwichPanelwith a CircularThrough-theThicknessHole Loadedin Compression(Variationin Thickness)....................101

Figure34. Case3: ResidualStrengthof a SandwichPanelwith anElliptical Through-theThicknessNotch Loadedin Compression(Variationin Thickness)..................102
Figure35. Probabilityof DamageDetectionPD(a)
Figure36.

Probability

Figure37.

Level

Figure38.

Density

of Safety

Function

vs. Critical

Design

Load

Loaded

in Compression

vs. Level

Figure39.

Design
Loaded

Figure40.

with Various

for Detected
Damage

of Safety

Damage

Inspection

for a Sandwich

Panel

Inspection

103
104

Types ......... 105

with a Circular

Disbond

I) .......................................................

106

Load vs. Level

in Compression

of Safety for a Sandwich

Panel with an Elliptical
(Inspection
Type I) .......................................................

Disbond
107

Design

Load

of Safety

Loaded

in Compression

Figure41.

Design
Loaded

Load vs. Level

in Compression

Figure42.

Design

Load

Disbond

Figure43.

Design
Disbond

Figure44.
Figure45.

Figure47.
Figure48.
Figure49.

Load

(Inspection

Loaded

with a Circular

Disbond
108

of Safety for a Sandwich

Panel with an Elliptical
(Inspection
Type II) .....................................................

Disbond
109

of Failure

in Compression

vs. Probability

Type

Panel

II) .....................................................

vs. Probability

Loaded

for a Sandwich

(Inspection

of Failure

in Compression

for a Sandwich
Type

Type

with a Circular

I) ........................................

for a Sandwich

(Inspection

Panel

Panel

Design

Load

vs. Probability

Loaded

of Failure

in Compression

I) ........................................

for a Sandwich

(Inspection

Type

Panel

vs. Level of Safety for a Sandwich

Panel with a Circular
Loaded in Compression
(Inspection
Type I) ................................

Design

vs. Level

Load

Loaded

Design

vs. Level

Load

of Safety

for a Sandwich

in Compression
of Safety

(Inspection

for a Sandwich

in Compression

(Inspection

Panel
Type
Panel

112

113

114

with an Elliptical
I) ................................

115

with a Circular

Delamination

Loaded

II) ...............................

116

Design Load
Delamination

vs. Level of Safety for a Sandwich

Panel with an Elliptical
Loaded in Compression
(Inspection
Type II) ...............................

117

Type

111

with an Elliptical

II) .......................................

Design Load
Delamination

Delamination

110

with an Elliptical

Design Load vs. Probability

of Failure for a Sandwich
Panel with a Circular
Disbond Loaded in Compression
(Inspection
Type II) .......................................

Disbond

Figure46.

vs. Level

Type

...........

Size po(a) ..........................

Size with Various

(Inspection

Types

Figure50. DesignLoad vs.Probabilityof Failurefor a SandwichPanelwith a Circular

DelaminationLoadedin Compression(InspectionType I) ................................
118
Figure51. DesignLoad vs.Probabilityof Failurefor a SandwichPanelwith anElliptical
DelaminationLoadedin Compression(InspectionType I) ................................
119
Figure52. DesignLoad vs.Probabilityof Failurefor a SandwichPanelwith a Circular
DelaminationLoadedin Compression(InspectionType II) ...............................
120
Figure53. DesignLoad vs.Probabilityof Failurefor a SandwichPanelwith anElliptical
DelaminationLoadedin Compression(InspectionType II) ...............................
121
Figure54. ProbabilityDensityof Actual DamageSizep(a)
Inspection

Methods

Derived

from Two Different

.............................................................................................

10

122

1.1

INTRODUCTION

Background

Traditional
safety

design

procedures

for the loads

knockdown

for aircraft

and knockdown

factors

have

been

structures

factors

obtained

are based

on a combination

strength.

Both

past

decades

of design

traditional

design

for the
from

the

five

of factors

the factors

of

of safety
for

and
metal

aircraft.
There

are

First,

at least

because

materials,
apply

novel

that

structural

in essence,
condition

temperature,

unnecessary

weight

A second

that

throughout

the

improvement

structures

approach

for designing

moisture,

damage,

traditional

of various

of safety

safety.

design

procedures

available

during

of these
the design

to produce

the case
for

occurs

may

systems,

of composite
scatter

materials.

This approach,
for

results

and

and to damage

structures.

This

to

in composite

effects

simultaneously

etc.

metallic

be difficult

material

larger

composite

loading,

design

each

design

in substantial

a consistent

on the safety

can lead

to excessive

on the concept
In particular,

and

level

for the

of safety

11

final
and

measures

to determine

is a consistent

based

process

options

is that

that there

situation

problems.

procedures

it is not possible

design

it is unlikely

in overall

many

new

to environmental

scenario"

As a result,

That

overcome

use

to account

of composite

aircraft.

structural

procedures

configurations,

procedures

for example,

case

of

importance

with no measure

traditional

configurations,

a "worst

are not available.

the relative

conventional

procedures.

and

penalties.

shortcoming

reliability

for

these

Consider,

design

conservative

assumes

concepts,

concepts.

to these

developed

unconventional

and the sensitivity

led to a very

designer

have

shortcomings

were

structural

of traditional

properties

New

fundamental

procedures

and familiar

Adaptations

have

the

to aircraft

contain

two

design.
efficiency

(with

of the

aircraft.

level of safety
weight

of "design
measures

of

with

under

of safety
This

and

any precision)
In addition,
and efficiency

no corresponding

uncertainty"
and

information

throughout

safety

help to

reliability
allows

the aircraft

are
the
- no

unnecessaryover-designsin some areas. As a result, designerscan saveweight while

maintainingsafety. In addition, in designunderuncertaintyit is possibleto determinethe
sensitivityof safetyto designchangesthat can be linked to changesin cost. For the same
cost, aircraft can be made saferthan with traditional designapproaches,or, for the same
safetyandreliability, the aircraft canbe madeat a lower cost. Designunderuncertaintyalso
hasapplicationto the flight certificationprocess,asit allowsthe uncertaintyinherentin any
new design to be quantified. Thus, flight certification criteria can be establishedwhich
define the safety margins necessaryfor compliancebased on the level of uncertainty
associatedwith the design.
Basedon the aboveconsideration,a researchprogramwas establishedby the University of
Washingtonto studythe feasibility of developinga designprocedurebasedon conceptsof
uncertainty and of applying this procedureto the design of airframe structures for
commercialtransport. Theprogramis being sponsoredby NASA LangleyResearchCenter.
The new designprocedureis basedon the fact that designdata such as loading, material
properties,damage,etc. areof statisticalcharacter. Designproceduresbasedon uncertainty
havethe potentialfor reducingthe weight andcostof airframestructureswhile maintaining
prescribedlevel of safety. Theseprocedurescould alsohelp reducethe designcycle time,
particularlyfor unconventionalaircraft thatusenewmaterialsandnovel structuralconcepts.

1.2

The

Review

of existing

non-deterministic

disciplines

have

been

design

Integrated

combines
IPACS,
techniques

(Ref.1,

procedure

mechanical

are the

input

engineering

for composite

and
data

of

specific
physical
and

all

of the

This

current

design

for decades.

structures

structures
Composite

Structures
system

properties,

resin

12

data

are

emphases

in various

has

applied

been

In recent

as well.

(Ref. 5).

structure,

of these

research

methodology

applications

composite

Analysis

mechanics,

is one

2, 3, 4).

to aerospace

Probabilistic

physics,
fiber

approach

and electronics

applications

probabilistic
the

design

of engineering

civil, mechanical

technologies

Chamis

The research

years,

and

manufacturing.

and

the fiber

considered

random

has generated
The

properties,

there

developed

(IPACS).

concepts

to

procedure
In

placement

variables.

probabilisticlaminationtheoryis then establishedusing a micromechanicsapproach.This is

followedby a probabilisticfinite elementanalysisbasedon structuralmechanics.The output
of IPACS includesstructuralsizing,failure predictionandload limiting application. IPACS
doesnot includeoperationallifetime considerationssuchasmaterialdegradationandrandom
damageprocessesduring service.
Kan, et al., proposeda probabilisticmethodologyfor compositeairframecertification. The
original work focusedon probabilisticmodelsto characterizedatascatterin compositestatic
strength and fatigue-life tests (Ref. 6).

The goal was to evaluate structural testing

requirementsto achieve B-basis allowablesfor flight certification. Their methodswere

extendedto include data scatterin bondedand cocured structures,and to assessimpact
damage requirementsfor certification (Ref. 7).

The impact threat to aircraft was

characterizedusing a Weibull distributionof impact energy. A damagedetectionthreshold

of Barely Visible Impact Damage(BVID) was set for a dent depth ___
0.05 in. in thin
laminates. A methodwas presentedfor predictingpost-impactresidualstrengthof built-up
structureswhich incorporatesa statistical analysisof data scatterfrom compressiontest
specimenswith the impact threat distribution, to give an integratedprobabilistic reliability
analysisprocedure. This model was then modified to reduce the number of empirical
coefficientsandtestdatapointsneededfor ananalysis(Ref. 8).
Rouchon(Ref. 9) hasalsocontributedto compositestructuraldesign,primarily in two major
areas: 1) certification and compliance philosophy; 2) probabilistic inspection for fleet
reliability. Rouchon'sefforts in the areaof certification andcompliancephilosophyaddress
second source material qualification, conditions to simulate environmentaleffects, and
damagetolerancedemonstrationfor accidentalimpact damage. His work on probabilistic
inspectionis focusedon the needto detectimpactdamagein compositestructuresbeforethe
critical load level for catastrophicfailure is reached(Ref. 10). A simplified probabilistic
approachwaspresentedfor damagetoleranceevaluation,wherepost-impactresidualstrength
dataare combinedwith probabilistic assessments
of impact damagethreatsand flight load
factors to set inspectionintervals for maintaining failure probabilitiesbelow a threshold

13

value.This approachis being usedto certify the ATR72 andfuture generationsof Airbus
aircraft.

Gary andRiskallainvestigatedthe applicationof Northrop Grumman'sProbabilisticDesign

Model to determinestructuralreliability valuesfor a moderncompositeaircraft (Ref. 11).
The Northrop Grummandesignmodel is a Monte Carlo simulationin which the probability
distributionsof operatingstressandmaterial strengthare subjectedto lifetime risk drivers
suchasmaterialquality, manufacturingquality, thermalstress,gust,operatingenvironment,
andoperationalstructuraldamage.Failure probability is definedas the probability of stress
exceedingstrength. To validate the designmodel, site visits were conductedat airline
maintenancefacilities and Naval aviation depotsto gatherhistorical data on operational
damageincurred on compositestructures. This data was input into the designmodel to
assessthe structuralreliability of the LearFan 2100wing box.
The works reviewed here are only a small sampleof the range of researchdevotedto
probabilisticmethodsappliedto aerospacestructures,yet they provideimportantinsight into
how far thesemethodshave come, and illustrate areaswhere further efforts are needed.
Chamis's model is an important design tool for assessingthe variability in composite
manufacture,analysisandtesting,but it doesnot incorporatemeansfor evaluatingeffectsof
servicedamageonreliability. Kan's probabilisticwork is gearedtowardsflight certification,
anddoesnot directly addressdesign. The methodsobtainedonly apply to a specificdamage
mechanismin composites(impact), and do not incorporateprobabilities associatedwith
detectingimpactdamagein an aircraft fleet. Rouchon'swork alsois gearedtowardsflight
certification, and acknowledgesthe role inspection plays in maintaining the safety of
damage-tolerantstructures. He also addressessome of the limitations inherent in
probabilistic methodologies,namely that a large and detailed databaseis needed to
characterizeimpact damageprobabilities, and that the threshold approachto damage
detectionmay be inadequate.At the presenttime,however,he hasnot proposedany means
of incorporatingtheseconcernsinto his probabilistic methodologies.Northrop Grumman's
designmodelmay be oneof the mostrobustyet to apply probabilisticmethodologiesin the
designprocess,andhasbeendemonstratedon a modern,flight-certified compositeairframe.

14

However, it doesnot incorporatethe influence of damagedetection probabilities in its

reliability assessments.
All of the work reviewed thus far focuseson compositeimpact as the primary damage
mechanismdriving the use of probabilisticmethodsin damagetolerance. However, these
methodscanbe appliedequallyas well to otherdamagemechanisms,andfor othermaterial
systems.Most of this work hasbeenfocusedon specificareasandapplications,anddoesnot
take a broadoverviewof the structuraldesignprocessfor reliability. Therefore,thereexists
a need for a unified probabilistic approachto reliability that is independentof specific
materialsystemsandstructuralconfigurations,andthat canbe appliedto the entirelife-cycle
of a structure. It shouldtake into accountthe influenceof detectionprobabilitiesin setting
inspection intervals and defining critical damagethresholds,and also account for the
existenceof multiple damagesof different types in a structure. This is the thrust of the
currentresearcheffort outlinedin this report.

15

The

specific

goals

acceptable

structural

accumulated
methods
process

to detect
which

of Safety"

design

and

details,

procedures

The

will
for

structural

sizing
The

specific

sensitivities

result

data and operational

experience

and that can also be used

and structural

concepts

in future

of in-service

design

of non-destructive

will be used
of an existing

matches

resize

systems,

the "Level

designs.

16

will

that

of Safety"
associated

dissimilar

of

be

aircraft
structural

residual

used

and
allows

"Level

safety-based

analyses

and

data-driven

requirements,

explicit

requirements

methodology

the existing

between

load

a design

structure,

quantifiable,

to be made

procedures

design

aircraft

of

inspection

to develop

or exceeds

methodology,

the uncertainty

aircraft

ability

deterministic

design

to quantify

the

safety

safety-based

to quantify

assessments

material

The

will be a uniform

probabilistic

which

incorporating

to

on

is to be an objective,

different

structures.

definition

of Safety"

structure

identified

developed

a workable

definition

of relative

using
the

resulting

method

comparisons

Using

be

design

to establish

and

"Level

of a new

new

structures

etc.

properties

criteria.

the equivalent

in the design

that will allow

The

are

based

components,

damage.

evaluates

program

of Safety"

to aircraft

such

value.

components

research

"Level

damage

uses this value

process

of this

OBJECTIVES

resize
strength

to demonstrate

flight

certification

utilization

of the existing

of service
aircraft

with the use of new

fleet,

materials

3.1

General

EQUIVALENT

approach

is defined

in this chapter.

a probabilistic

"Level

from in-service
resulting
design

process

Safety"

data

upon.

Once

A flow-chart

Defining

enable

the

quantitative
and

type

tolerance
component
some
criteria

design

be detected
threshold.

is the joint

so that

used

This method

using

damage

design

efforts

concepts

baseline

data

in service,

value

to

The

meet.

A
the

the "Level

of

is maintained

inspection

in the design

is used

quantifies

and applies

safety

on

type.

must

which

based

collected

for each damage

applications,

the

of Safety"

or

and maintenance

process,

and to reduce

the

to developing

equivalent-safety

of the approach

is given

aircraft

structures

in the following

is given

in

context.

of Safety"

evaluation

a part

philosophies

of these

structure

with the structure.

is needed

of damage

structural

is built and placed

explanation

objective

method

and

the assumptions

approach

"Level

materials

"Level

probabilities
future

of an aircraft

3.2.

structures

for which

behavior

the structure

1. The detailed

baseline

of Safety"

in Section

with detection

structure

associated

of the

aircraft

tolerance

to validate

level of uncertainty

To

APPROACH

of structural

of existing

for new

the

definition
is derived

a safety

to size

"Level

approach

combined

is defined

can be used

3.2

of Safety"

establish

definition

Figure

tolerance

in the damage

improved

SAFETY

the equivalent

A mathematical

experience,

values

uncertainty

for determining

damage

the

OF

Approach

A general

evaluate

LEVEL

which

that

and repaired
A convenient
density

level

incorporates

is exposed

require

probability

of the

before
way

safety

design

to during
damage

of

its

data

to define

function

approach

limit.

17

along

aircraft

life.

during

the

of the component
the "Level

component,

with data on the amount

operational

accumulated

the strength

of an

Modem
service
is degraded

of Safety"

for damage

based

size and NDI

damagelife

of

beyond
on these
detection

This "Level of Safety" conceptwas initially proposedby Bjom Backmanof the Boeing
CommercialAircraft Group, and is definedin statisticalterms as: "The compliment
probability

that a flaw

the structure

There

size that is larger

is incurred,

are two random

Detection

state "d",

density

function

pg (d,a)

= pc(dla)

in which

Because

variables
which

random

is the

conditional

function
variable

Since

they

size "a",

which

are not independent,

d2 ....

damage

is not detected

probability

damage

density

of

is cominuous;
the joint

and 2.

probability

density

function

function

and

p(a)

is the

marginal

size.

"d" has only two possible

is detected

p(a)

values,

that is:

is the sum of two terms.

+ pg(d2,a)

= p_(dlla)p(a)
= [p_(dlla)

1. Damage

probability

of actual

damage

= pg(dl,a)

strength

(3.2-1)

d/ ....

p(a)

residual

p(a)

density

the marginal

size for

will not be detected."

involved:

is discrete.

flaw

pg (d,a) is:

pc(dla)

probability

and that the flaw

than the critical

of the

+ p_(d2la)p(a)
+ p_(d2la)]p(a)

(3.2-2)

Thus,
[pc(dlla)

When

+ p_(d21a)]

the conditional

= 1

probability

the probability

of detection

the probability

of detection

p_(dlla)

(3.2-3)

under

density

function

the condition

for damage

p_(dl a) is evaluated
that the damage

at d = dl, it stands

size is a.

for

Let PD(a) denote

size a, that is:

=PD(a)

(3.2-4)

18

Usingequation(3.2-3),we have
pc(d21a) = 1- PD(a)

Substituting

(3.2-4)

p(a)

= p(a)PD(a)

The

first

damage

term

in which,

+ p(a)[1

yields:
(3.2-6)

is proportional

or mathematically

to the

(Multiplication

probability

density

function

of detected

Rule):

= f(dl)g(aldl)

(3.2-7)

= PD(a), g(aldl)

= Cpo(a)

or

C is a normalizing

I p(a)da

into (3.2-2)

- PD(a)]

of (3.2-6)

p_(dlla)

p(a)PD(a)

where

and (3.2-5)

size po(a),

p(a)pc(dlla)

(3.2-5)

= po(a) and fldl)

p(a)

constant

is a constant.

Therefore,

= Cpo(a)/PD(a)

that is determined

(3.2-8)

by the condition

of:

(3.2-9)

= 1

Substituting

(3.2-8)

iC_a=l

into (3.2-9)

or

Therefore,

according

yields:

1/fP(a)

da

(3.2-10)

to (3.2-8):

p0(o)/ p0(o)do
P ( O) - _

Note

aop_)

that the actual

(3.2-11)

and detected

damage

size distributions

are two distinct

functions

in this

equation.
Finally,
where

"Level

of Safety"

a_ is the critical

is defined

damage

as 1- Probability(damage

size for failure

of the structure:

19

size

> a_ I not

detected

),

"Level

of

Safety"=l-

Ip(a)[1-P_(a)]da

(3.2-12)

tic

Using

equation

"Level

of

The above

(3.2-11)

Safety"=l-

definition
discrete-source

generally

not true.
of flaws

i P(a)[1-Pv(a)}ta

assumes

a single

number

forp(a):

a single

flaw

is another

inspection

is present

Sometimes

there

/=[P(a)

event

in the

is no flaw,

random

variable

da

at a fixed

structure.

times

must

point

In most

and other
that

(3.2-13)

there

be involved

in time,
real

and that only

structures,

are multiple
to define

this

flaws.
the

is

The

Level

of

Safety.

Assuming

that hazard

and there

is no interaction

"Level

of Safety"

"Level

in which

of

the

total

variable,

Safety"

is reduced

between

multiple

interaction
and

will

other,

between

for multiple

Safety"=

random

to each

due to an individual

/
1-

flaws

the

to the

a significant

not be addressed

of the multiple

should

flaws,

in the

is /t.

of flaws

further

the proper

flaws

definition

structure

According
increases.

dependent
stress

then

of other

of

be:

present

to this
Whether

zones

and modeling
here

in order

(3.2-14)

at the time

of evaluation

definition,

the

or not there

on the size and location

concentration

analytical
any

of the existence

Po (a)
Pz)(a)]cla/[--da,
//:oPz)(a)

mean

number

is highly

and relative

poses

flaws

of flaws

for which
as the

each

i P(a)[1
a Pz)(a_-

number

flaw is independent

"Level

challenge

to the

to simplify

the

of

is interaction

of the damages

in the structure.

is a

relative

Damage

structural
method

zone

designer,
as much

as

possible.

In the case that various

damage

type has its own probability

mechanisms

distributions.

exist
Thus,

20

simultaneously
the above

definition

in the structure,
should

each

be modified

flaw
to:

"Level

of

where

Safety":

i denotes

number

of

damage

damage

determined

by

known,

"Level

"Level

of

1-

FI/

type,

types

damage

--[l-Pv,(a)]cla
p'(a) [1

tolerance

in the
criteria.

structure

critical

size that can be tolerated

load and structural

geometry

a_, = F, (P,I),

where

and load,

P is load

into Equation

(3.2-16)

"Level

of

Safety"=Fo(P,1)

Equation

(3.2-18)

to a structure

further

Nr,

lower

Poi(a),

i and Nr

is the total

integration

PDi(a),

limit

the

(3.2-18).

ti, (i=1,2 .... ,Nr)

modified

are

/
by deterministic

materials,

structure.

structural
Hence,

analyses

as the

a_ is a function

of

that is:

(3.2-17)

sizing

laminate

dimension

such

in a sandwich

panel.

as sheet

metal

Substituting

gage

thickness,

Equations

or

(3.2-17)

load,

(3.2-18)

structural

structure's

Alternatively,

location

size and "Level

dimension

can be sized by solving

of multiple

aci is

of ac_, (i=1,2 ..... Nr), that is:

by the damaged

with specific

of type

yields:

relates

and

Equation

In the case

a_i can be found

and 1 is a structural

of a face-sheet

the structure

The

(3.2-15)

i = 1,2,...N r

the thickness

using

of flaws

as a function

acl,ac2,,acN

For a given
flaw

When

Safety"=F

number
structure.

can be expressed

da
[/"

ti is the mean

possible

of Safety"

/_[ P'(a)

when
Equation

damage

are given,

of Safety".
"Level

required

"Level

(3.2-18)

for 1.

in a structure,

to:

21

When

of Safety"
of Safety"

the definition

the load
can

applied

be evaluated

and load are given,

of "Level

of Safety"

is

wherej

denotes

number

of damage

In turn,

the "Probability

damage

"Pr obability

type, NL is the total number

types

of

at location

should

= 1 -"Level

of

Safety"

Pi',j(a)
P
(a)

3.3

example

Establishing

To establish
level

of safe operation

for

the

Equipment

aircraft

fleet

various

states

aircraft

to the

flaw

vehicle.

Under

current

Limit

Detailed

design

to evaluate

are meant

than

and

the

inspection

critical
strength

and certification

(3.2-20)

_]

designs

to replace,

using

the probability

the normal

I]"ij

_da

This process

of interest,

size becomes

is lower

for new

be benchmarked.

under

a flaw

they

(a)

5.

needed

aircraft

/)Poij

design

needed

data

from

the

this residual

service
Original

flaw in the aircraft

Damage-tolerance

the residual

same

of Safety"

collecting

that a critical

to maintain

practices,

the

the "Level

involves

regime.

when

to maintain

strength
safe

design

of a structure

operation

strength

value

of the
is at the

Load for the structure.

analyses

mechanisms

of Safety"

types

(OEM)

that

exposed

design

of Safety

must

will go undetected

philosophy

Level

as the current

Manufacturer

structure

in Chapter

"Level

, ,1,

[1- rDi'ta) J_a/Jo

is given

a Baseline

the equivalent

of the current
data

problem

and Nr: is the total

be:

An illustrative

locations,

j.

of Failure"

Failure"

of damage

must

and failure

be performed
modes

Safety"

value

based

damage

types

in each component,

for each

that the parts

on the formula

derived
and these

aircraft

component

are vulnerable
in Section
values

22

to ensure

to are accounted

3.2 must

be calculated

will be unique

that
for.

all damage
A "Level

of

for the various

to that component

in that

particular application. In addition, reliable and repeatablemethodsof detecting service

damageneedto be identified andquantifiedfor the existing structures.Much researchhas
beendonein the areaof definingProbabilityof Detection(POD) curvesfor variousforms of
Non-DestructiveEvaluation(NDE), andmost of this work hasbeenconcentratedin the area
of crack detectionin stress-skinnedmetal airframes(Ref. 12, 13, 14). However, not all
damagemechanismsinherentin a componenthave POD curvesreadily available for the
specificapplicationof interest. Many factors,suchas part geometry,part location,the skill
of the NDE equipmentoperator,or the environmentin which the inspectiontakesplacewill
significantly affect POD results,andcan shift the curve dramaticallyoneway or the other.
All of thesevariablesneedto be takeninto accountin the evaluationof equivalent"Level of
Safety", asthey canhavean importanceequalto or greaterthan the deterministicaspectsof
residualstrengthcalculation. Extensivetestingandanalysismay be necessaryto verify that
the assumedvaluesof the PODcurvesareactuallybeingachievedduring serviceinspections.
Another essentialelement to the "Level of Safety" calculationsis the availability of inservicedetecteddamagedatafor the specificstructuresof interest. The datais necessaryto
defineprobability densityfunctionsof detecteddamagesizes,andto characterizethe amount
of damageaccumulatedin a given time. A moredetaileddiscussionof this topic is presented
in Section3.4.

Establishingthe baseline"Level of Safety"for an existingaircraft componentmay be oneof

the most difficult andlabor-intensivestepsin the designunderuncertaintyprocess,but it is
necessaryto establishexactly what the vulnerabilitiesof existing structuresare to serviceinduceddamage. The results of this stepare then usedto define the minimum allowable
"Level of Safety" for future structuraldesignsthat sharesimilar functionsand operational
requirements.

3.4

Collection

A critical
detected

of Flaw

component
damage

size

in the

Data on Existing

determination

distributions

po(a)

Structures

of "Level
based

23

of Safety"

on inspection

data

is the

characterization

for an existing

structure.

of

Inspection

data to be obtained

frequency

of occurrence

detection.

A histogram

be constructed.

for each part must

over

a given

for the flaw

Although

for the U.S. commercial

regularly

submit

maintenance

Difficulty

Reports

Service
accessible

to the

instances

of detection

aircraft

designers

curves
the

However,

may

of this research

under
make

further

aircraft

From

on a regular

the

and

much

basis.

aircraft

to the

database

is not

a comprehensive

to develop

operational
to utilize

consistent

conditions.

the

data

Small

for

"Level

of this specific

is already
facilities

in the

form

of

(SDRS)

that

is

archive

of the

of

can then

repair

FAA

in a database

authorities

period

Licensed

much

method

of this data

collected

it can provide

the

service

are

into the details

this data,

structure.

utility

cracks

estimate

the method

a given
task,

of flaws,

it easier

the possible

data for discrete-source

a first-order

fleet

which

real-world

investigation

within

location

size of damage,

data

of all

necessary

for

damage

size distribution

changes

in the format

of

Safety"

application

of

calculations.

is beyond

the scope

effort.

To demonstrate

database.

repair

damage

on commercial

Although

and certification

for structures

system

aircraft

(SDR's),

and

the type of damage,

to be a monumental

actions

public.

period,

size distribution

this appears

collected

time

include

of such

in the fuselage

a Weibull

of detection,

be noted

hence

has

of several

function

that the sizes

to define

succeeded

through

detected

the detection

using

the data

of detected

of these

in collecting

of its aircraft

can be fitted

density

the need

Boeing

skins

distribution

of the probability

It should

a tool,

crack

the SDRS

points

to give

sizes po(a)

cracks

method

raw

in an

are a function

for each

data

of

point

utilized.

Other

methods

Riskalla

used

evaluate

the

have
on-site

design

utilized,

it should

detected

damage

available

visits

damage

probabilistic

incomplete

also been

picture
data

used

to airline

characteristics
methodologies

be apparent
over

the

of the

to obtain

lifetime
true

set is reduced,

on in-service

facilities

and

of composite

airframe

structures

11).

Regardless

it is virtually

and

of damage
more

uncertainty

24

of the

impossible

of a component,

distribution
more

data

maintenance

(Ref.
that

damage

so we

Naval

data

aviation

depots

to

work

on

their

collection

forced

in a structure.
creeps

Gary

for

to record
are

aircraft.

into the

every

methods
instance

to deal
As the

and

size

estimation

with

of
an

of the
of the

probability

density

functions

functions.

associated

calculations

with limited

to these

3.5

Application

With

a baseline
and

structures

with higher

most

techniques.

must

level,

from

incorporated
while

of "Level

of Safety"

regarding

their

way

and extensive

how

the

onto

alloy,

upon

in-service

structure

current
through

experience,

stressed-skin

aircraft

construction
concepts

designs,
will

new

replacement

primarily

and structural

new

advanced

of

or improving

with new materials

to find

in place,

design

is ensured

aluminum

Concepts

structure

the

aircraft

process

traditional

into

Design

maintaining

of existing

experience

system

design

damage

mechanisms
for service,

"Level

block

Along

into

of

up through

makes

owing

to the

perform

under

effect

on residual

to the operator.

methodology
the

design

PD(a),

component

the designer
data,

before

an

By carefully

25

the

at

level

to

of the

symptoms

of detecting

the concept

the

choosing

distributions

experience

the nature

starts

of

the various

can be declared

with this process.

allows

can quantify

about

and how

means

associated

the probability

test

Reliable

This process

and sub-structure

to in service,

previously

process.

system.

are made

and put into place

defined

and their

the component

assumptions

are also uncertainties

that define

probability

or structural

the way,

be identified

and there

mechanisms

will be exposed

themselves
must

functions

Use

works

the structure

uncertainties

damage

for any new material

level.

of Safety"

and detection

approach,

and gradually

will manifest

concept.

be

in the design

derived

be identified

damage

statistical

and testing

building

environment

these

of an existing

can

applications

damage

The

of density

environment.

the material

ready

of Safety"

integrity

of uncertainty

a traditional

the final

the sensitivity

Structural

of performance,

structural

for these

operational

strength

concepts

has been

amount

the uncertainty

to Advanced

"Level

The lack of service

it difficult

Using

the

levels

analysis

of which

large

for

structural

The

deterministic

to characterize

data sets, and to explore

of Methodology

value

levels.

is ongoing

effects.

materials

safety

Research

engineer

the parameters

for detected
the relative

with

similar

to incorporate

damage

of the
size po(a)

amount

of risk in the

concepts

in different

applications,empiricalevidenceandengineeringjudgementareall tools that canbe usedto

definethelevel of risk inherentin the newdesign.
With the uncertaintiesin the damagetolerancecharacteristicsof the new conceptaccounted
for, the "Level of Safety" definition is then usedto generatecurves of "Level of Safety"
valuesvs. flaw sizefor eachtype of damagemechanismpresentin the structure. Using the
baseline"Level of Safety" value as an allowable,critical damagesizescan be identified,
either individually or in combination,that will give a safetylevel equivalentto previous
designs. Deterministicanalysismethodsarethen usedto generatecontourplots of residual
strengthvs. flaw size for eachdamagemechanism,as a function of the structuralsizing
parameterschosenbeforehand. With theserelationshipsdefined,the structurecan now be
sizedfor a given load level to yield a relative damage-tolerance
safetyequalto, or greater
thanthe existing structure.

3.6

Damage

Size

Updating

a new

structure

has

Once

performing

under

maintenance
damage

tolerance

inherent

in those

methods

(Ref.

revised

curves,

that the value

earlier

"Level

of Safety"

Weibull

manufactured
be found
probability

made

during

products.

to approximate
density

into

be utilized

in the design

frequent

is a very
Failure
Weibull

function

service,

available

values

process,

curves

detection

the

well

model

known

mechanisms

in many

distributions,

for the

26

inspection

the level

sizes.

structure.

systems

mechanical

on the

damage

period,

and

than
to
the

reliability

systems

will be chosen

in the "Level

Based

can be revised

to predict

different

statistical

If it is found

frequent

program

about

of uncertainty

of Bayesian

or more

and

assumptions

time

used

size po(a)

it is actually

in a given

so this model

damage

initial

damage

of larger

how

scheduled

by the use

for detected

opportunities
value.

on

and to reduce

and maintenance

to its design

of detected

through

can be recalculated

the inspection

data

to validate

due to the accumulation

can be returned

relation

can

put

becomes

the distribution

design,

or more

and

This can be accomplished

of Safety"

has decreased

provide

data

assumptions.

"Level

assumed

The

behavior

built

conditions

The

15) to modify

initially

been

operational

actions.

Schemes

can often

to represent

of Safety"

in

the

method.

Other statisticalmodels, such as Normal, Gamma or Log-Normal distributions may be

appropriate,dependingonthe behaviorof the damagemechanismof interest. For simplicity,
we will only be concernedwith the Weibull distribution here, however the concepts
discussedare equally applicableto any statisticalmodel of a continuousrandom variable
distribution. The probabilitydensityfunctionof the Weibull modelis:

Po(a)

fl_ a _'-* exp

Assume

that the initial

= 4.

In Bayesian

variables.
initial

detected

analyses,

However,

detected

damage

damage

parameter
deviation

fo (o0-

l
0.082;-o__lt2)
,_5_

24

probability
al,a2,...,an,

fl is assumed

random

to be a constant

to follow

a Gamma

distribution

with its mean

4 for

of 2, and a

(
O_
exp _,- 0.08

size

Let fflal,a2

parameter

are considered

of 0.4, that is:

an, the

detected

scale

function

o_= 2 and fl

(3.6-2)

= 0.

actual

in the density

parameters

for o_> 0, otherwisefo(OO

detected

the

has Weibull

size distribution:

o_ is assumed

standard

size distribution

the parameters

for simplicity,

Po (a I oO = _ a_'-* exp
46

The shape

(3.6-1)

damage
damage

sizes,

..... an, o:) denote

density

function

When

(3.6-3)

new detected

distribution

damge

can be updated

al, a2 ..... an, are random

joint
of

probability
o_ under

density
condition

is:

27

size data
using

are obtained

the new

as al, a2...,

information.

The

variables.

function
of given

of al,a2 ..... an and

detected

damage

0_.
size

The
data,

f,(oC lal,a2,...,an)

f;

(al,

a 2,...,

an,

0_)

(3.6-4)

If;(al,a2,...,an,_)d_

Let g(al,a2
condition

..... anl o9 denote

of given

= g(al,a2,...,a

(3.6-5)

g(al,a2,...,a

density

function

of al,a2 ..... an under

variables

(3.6-5)

yields:

= _ g(al'a2"'"an
I g(al'

random

probability

n l a)fo(a)

into (3.6-4)

f,,(o_lal,a2,...,an)

Since

joint

o_. Then:

f ;(al,a2,...,an,a)

Substituting

conditional

a2,''''

la)fo(a)

(3.6-6)

an' I OOfo (oOdo_

al,a2 ..... an are mutually

n I o0 = lZIPo(ai

independent,

we have:

I o0

(3.6-7)

i=l

Substituting

(3.6-7)

into (3.6-6)

yields:

I-[ Po ( a _ IoOfo (o0

f,(alal,a2,...,an)

i=1

(3.6-8)

Il-[
__

That is, the probability

and the posterior
mean

density

distributions

of the updated

random

de = i of,(oe

lal,a2,...,an)doe

The updated

distribution

alpha

Po(ai

for the Weibull

] oOf (Od_

i=1

function

of o_ under

of o_ are plotted
variable

condition

in Figure

2.

of given
The Bayesian

al,a2

.....

an. The prior

estimate

of o_ is the

o_,that is:

(3.6-9)

of detected

distribution

damage

of detected

size is obtained
flaw

28

size:

using

the Bayesian

estimate

of

Po (al&)=

4--2

The initial

exp

and updated

to reach

an updated

function

ofpo(aloO

Po (a) =

8).

When

(3.6-9).

(3.6-2),

when

are given

distribution

density

detected

estimate

An alternative

is to use the marginal

_-

both Weibull

function

damage

probability

parameters

simultaneously.

probability

probability

under

density

(3.6-11)

of o_, which

sizes

are al = 3 inch,

as 2.1831

density

is given

function

using

in Equation
a2 = 4 inch,

equations

(3.6a3 = 5

(3.6-2),

of the detected

damage

density

o_ and fl are treated

Let us assume

function

as random

that initial

variables,

these

(3.6size is

parameters

o_ and fl are independent.

the

(3.6-12)

to the derivation
the condition

Then,

can

of o_ and fl is:

fo(a,fi) = fo(a)fo(fi)

Similar

approach

f. (a')da'

of o_ is calculated

The updated

3.

that is:

a _-1 exp

newly

in Figure

(3.6-10).

be updated
joint

size

probability

n = 3, the Bayesian

equation

flaw

(a')da'

is the updated

3), (3.6-8),

size distributions

of Equation

As an example,

inch,

flaw

detected

Po (a I a')f.

wheref,(o0

(3.6-10)

of (3.6-8),

of measured

the updated

flaw

joint

probability

sizes al,a2 ..... an, is obtained

density

function

of o_ and ,6,

from:

1-Ipo(a_
f,(o&fllal,a2,...,a,)

Ia', fl)fo

(a', fl)
(3.6-13)

i=1
n

....

The marginal

probability

i=1

density

functions

of o_ or fl are then derived

29

through

integration

as:

f_, (oe l al,a2,..

.,a,)

= i f, (o& fl l al,a2,..

f fl,,(fl lal,a2,'",a_)

The

Bayesian

(3.6-14)

.,a,)dfl

(3.6-15)

= i f,,(O& fl lal,a2,'",a_)doe

estimate

of

o_ or fl is the

of the

mean

updated

random

variable

o_ or /3,

respectively.

= iof_(oe

lal,a2,...,an)doe

(3.6-16)

fl = i _f &(fl

lal,a2,...,an)d_

(3.6-17)

The updated

distribution

and fl for the Weibull

marginal

size is obtained

using

Bayesian

estimates

of o_

(3.6-18)
exp

the updated

density

damage

distribution:

po (a l & fl) = --z-z-a

fl_

Alternatively,

of detected

distribution

function

through

of detected

integration

damage

size can also be obtained

using

the

as follows:

P (a) = I I p (a l & fl) f_ (& fl l al'a2'''''a')d_lfl

(3.6-19)

= I_ I- -;a -lexp
--_--_

It should

be pointed

distribution
probability
damage

fl

using

can be performed

(__)_]
a

f,(o_,fllal,a2,...,a,)do_d

out that the above

later

of detection.
size must

obtained
When

be modified
accordingly

Bayesian

damage

different
before

approach

size

used

as:

30

is to update

measurements

probabilities
being

fl

that

of detection

in Bayesian

detected
must

have

are employed,

updating.

Such

damage
the

size
same

the detected
a modification

G2(a2)

(3.6-20)

a2 = a1
fDl(al)

where

al and

a2 are

PDI(a)

and

PD2(a)

PD2(a)

are

available

damage

sizes

obtained

are corresponding

by inspection

probabilities

and al is detected,

type

of detection.

1 and
When

then a2 Can be determined

2, respectively
functions

by solving

PDda)

Equation

and
and
(3.6-

20), and vice versa.

An alternative
distribution

of combining

instead

distribution

p(a)

detection

If p(a)

way

of

new

detected

can be estimated

using

Equation

is selected

and old information

damage
from

size

detected

is to update

distribution.
damage

The

"actual"
"actual"

size distribution

damage
damage

size
size

and probability

of

(3.6-11).

to be the basic

variable,

the definitive

equation

for "Level

of Safety"

(3.2-

14) becomes:

"Level

of

Safety"=

1-

pij(a)[1-PDq(a)]da

(3.6-21)

acij

and the "Probability

"Probability

of

of Failure"

Failure"=

should

1- ]-1

be:

1-

(3.6-22)
ipij (a)[1 - P_q (a)lda]

"_

ac_j

The

same

damage

3.7

Bayesian

method

size distribution,

as that

developed

for po(a)

can be applied

to update

"actual"

p(a).

Discussion

The definition
an extension

of structural
of reliability

"Level
theory

of Safety",

and the design

and statistical

analysis

31

tools

methodology
to the design

derived

from

it, is

and maintenance

of damage-tolerantaircraft structures. This methodologyis one of the first attemptsto

developa unified probabilistic approachto damagetolerancethat encompasses
all of the
parametersspecificto the designof aircraft structures,but thatis flexible andgeneralenough
to be applicableto anytype of materialor structuralconfiguration. Severalcharacteristicsof
this approachareuniquefrom otherreliability-baseddamagetolerancemethodsthat precede
it.

The first is the differentiationin the method betweenprobability distributionsof

damage

size po(a)

and

overlooked

in previous

a structure

is subject

damage

size

detected

damage

necessary
The

data

never

distributions
in use
which

a means
"Level

of Safety"

risk factors

that

unique

These

effects

does

actual
there

detected

and

uncertainty

PD(a).
actual

can then

that

of damage

structure,

be incorporated

functions

for

This is
size.

in a structure

can

associated

with the

analysis

distributions

statistical

probability

the use of post-design

of the

Thus,

methods

are the

same,

results.

with the

the designer

size

in

of zero

probabilistic
size

used.

a flaw

be uncertainty

damage

often

size goes to zero.

is the use of Bayesian

allow

density

can detect

Many

reliability

method

probability

will always

associated

and to enable

on the safety

by

distinction
size distributions

of the inspection

distribution

probability

The tools

data on damage

not exist

or unconservative

the

important

go to zero as the damage

because

the

to collect

represented

to this methodology

method,

may have

be

the

and detection

assumptions.

knock-down

Another

characterized,

This is a very

of detection

method

are that

to characterize

probabilistic

priori.

an inspection

assume

aspect

attempt

distributions

may lead to erroneous

A second

Any

generally

size whose

of po(a)

today

size p(a).

to the probability

of this

be exactly

damage

methods.

should

because

implications

actual

detected

inspection

having

into the design

to provide

distributions

the effects
specific

without

in the

data to validate

to investigate

without

tools

data

the

unknown
available

resorting

to arbitrary

factors.

unique

aspect

is the incorporation

for a structure.

Although

Safety"

has not been

formulas

the

addition

of inspection
of this

accomplished

variable

intervals
to the

to the reliability
derivations

to date, it will be included

32

estimates

of the

in future

"Level

of

iterations

of

the "Level of Safety"methodology. This ultimately will allow the importantparametersof

an inspection and maintenanceprogram to be included as essentialvariables in the
preliminary designprocess,where maximum benefit can be realized in the sizing of the
structureto obtainthe bestperformancefor the lowestlife-cycle cost.

3.8

Mathematical

As mentioned

Considerations

in Section

probabilities

of detected

derivations

have

can assume.

the detected

damage

damage

size p(a)

Cpo(a)

= p(a)PD(a)

size distribution

for

approaches

either

zero

or some

detected

to the actual

be finite
the

zero.

These

pose

significant

of Equation

over

product

damage

and

Formulation

methodology

actual

damage

implications

the first form

of Equation
dependent

differentiates
size

numerical

po(a) is highly

detection

One assumption

a probability

Therefore,

of Safety

p(a).

on the

between

The
shapes

(3.2-8),

resulting
that

these

it can be seen that

on the distributions

of actual

PD(a).
(3.2-8)

PDF

must

and

Level

of Safety"

po(a)

probability

of the

p(a)

size

By examining

form

assign

"Level

statistical

and detection

will go to zero.

the

damage

important

distributions

The

3.7,

in the

minimum

size distribution

characteristics
problems

should

detection

in the statement

distribution

of damage

of damage

actual

damage

po(a),

must

such

and

that

as the

than

detection

zero,

is that we can

This implies
and zero

probability

also go to zero as the damage

for the formulation

that

elsewhere.

distributions,
size approaches

of Equation

constant

size

of detection

(3.2-8)

sizes in a structure.

of the normalizing

damage

the probability

of Equation

size greater
size

are inconsequential
in the calculation

be

threshold,

implicit

any interval
of the

probability

(3.2-8),

but

C if it is in the form

(3.2-10).

(3.2-10)

In this form,
zero.

Plotting

C is an improper
the integrand

integral
as a function

with the value

of a would

33

0/0 at the lower

show

integration

that it approaches

limit

infinity

of

as a

goesto zero. Sincethe integrandis proportionalto p(a),

will also go to infinity
depending

on

proportional
finite

the

as a goes

shape

of

to the function

on the interval

distributions
generally

so the

Numerical

provided

roundoff

limits

redefined

as:

1/J

for po(a)

available,

integral,

the

for values

of

and Po(a),

order

Assume

of a close

of

singularity

can

is rapid

must

integration

routine

are

of this type

can be integrated

the

integrand

In order

expression
the

1.

For the types

cannot

the relative

to be evaluated

be

of

is not

determined

convergence

satisfactorily

Equation

is
to be

for the integrand

function

encountered.

function

for the integral

be less than

to check

enough

size distribution

to zero.

of

be used

damage

that

a closed-form

convergence

the

integrals

of singularity/3

integration

that

Improper

distribution.

[0, oo], the order

assumed

analytically.

to zero.

the actual

of the
before

(3.2-10)

the

can

be

OO

lim C=1/a-_ [ P(a)

c--->O
r (a)

da

(3.8-1)

where

the limit is approximated

numerical

convergence

convergence
actually

The parameters
themselves,
integral

that define

quantify
structure.
probability

the initial
is violated,

probability

Failure
density

particularly

chosen

if the order

the integral

of singularity

of po(a)

and PD(a),

for

any

discussions
damage

in distributions

functions.

34

size

appears

of actual

absolute

to converge,

but

is near one.

and in fact the distributions

to a finite

value.

to the actual

of the distributions

emphasize
data

Note that relative

to guarantee

a probability

used for the parameters

of these

zero.

converges

that we can assign

The results

to do so can result

where

from

condition

so that the integral

assumption

of detection

of 8 away

insufficient

may be cases

and the values

for the function.

the

but

the distributions

be carefully

diverges,

size distribution
admissible

There

in the limit,

must

small value

is a necessary,

of the integral.

diverges

by some

the need

accumulated

damage

If the
damage
are not

to carefully
on

a given

size that are not valid

4.1

RESIDUAL

DETERMINATION
STRUCTURE

OF EXAMPLE

Introduction

To

demonstrate

structural
as

STRENGTH

the

concept,

an

equivalent

a graphite-epoxy

example

material

mechanisms

inherent

many

of research

years

(flaps,

ailerons,

aircraft

fleets.

"Level

for

in composite

elevators,

Safety"

honeycomb

system

and

of

the

rudders,

Service-induced

there

data

a metal

have

applied

to

an

was

selected

structure.

been

are many

currently

damage

as

composite

of

structures

Also,
etc.)

sandwich

redesign

sandwich

testing.

approach

advanced
to be used

The

damage

well-characterized

composite

in use

throughout

is therefore

available

through

sandwich
the

civil

structures
and

military

for a variety

of these

applications.

As part

of the process

deterministic

structural

of applying
analysis

of the

sandwich

results

were then used as inputs

4.2

The

Material

selected

material

core

Three

varies

densities

laminate

with

different

different

and

types

safety

methodology

for the characterization

of damage.

into the probabilistic

and

system

of the laminate

sequence.

is developed

under

Systems

graphite-epoxy
stiffness

panels

the equivalent

The

design

structure,

of the residual

established

damage

strengths
tolerance

methodology.

Properties

is a honeycomb
the

were

core

created

is made

systems

density,

and

panel,

of Nomex.

by changing

lamination

honeycomb

sandwich

Variations

the number

are used
Nomex

in which

of graphite/epoxy

and the honeycomb

35

of plies

in this research.
honeycomb

were used in the study.

The ply properties

to a new

core are:

cores

the face-sheet
in the
and
The

strength
their

stiffness

of three

is a
and

stacking
of the
different

Tension:

E1= 22.9mpsi,E2

= 1.34 mpsi,

0.34,

V12----

G12 = G13 = G23 = 0.29 mpsi,

Strength

Compression:

in fiber direction,

Maximum

longitudinal

Maximum

transverse

Maximum

engineering

E1 = 22.0 mpsi,

X = 350 ksi.

strain,
strain,

eL = 15300
eT = 5680

shear

strain,

E2 = 1.34 mpsi,

_te

_te

_:

= 21000

_te

0.34,

V12----

G12 = G13 = G23 = 0.29 mpsi,

Strength

Core(1):

in fiber direction,

Maximum

longitudinal

Maximum

transverse

Maximum

engineering

X' = 295 ksi.

strain,
strain,

eL' = 13500
eT' = 5680

shear

strain,

_T'

_te

_te
= 21000

_te

E1 = E2 = 200 psi, E3 = 20 ksi, v12-- 0.5, v13 -- v23-- .01,

G12 = 20 psi, G13 =7 ksi, G23= 3.5 ksi.

Core

(2):

E1 = E2 = 600 psi, E3 = 60 ksi,

G12 = 60 psi,

Core

(3):

The thickness

of the lamina

lay-up

0.5,

V13

---- V23----

.01,

V13

---- V23----

.01,

G13 =13 ksi, G23 = 6 ksi.

E1 = E2 = 900 psi, E3 = 90 ksi,

G12 = 90 psi,

V12----

V12----

0.5,

G13 =17 ksi, G23 = 9 ksi.

is 0.005

sequences

inch and the thickness

of the core is 1.0 inch.

The three

laminate

are:

Laminate

(1):

[-45/0/45/0/90]s

laminate

thickness

= 0.05 inch;

Laminate

(2):

[-45/0/45/9012s

laminate

thickness

= 0.08 inch;

Laminate

(3):

[-45/0/45]019012s

laminate

thickness

= 0.10 inch.

36

4.3

Tensile

The tensile
property

Strength

strength

of the Laminates

of the laminates

data for Laminate

-0.6198

(1), [-45/0/45/0/90]s

0.1220

When

The solution
Nx = -3311

I_y =

1721

Using

lamina

,we have:

mpsi

the maximum
stress

versus

transverse
strain

relation

strain

of the lamina,

is obtained

e22 = 5680

as:

equation

is:

lb/in

[.te

22.9 mpsi,

increases
E2 = v12

-0.6063

further,
=

to stress

the stiffness

G12 = 0. Then,

0.1174

0.11740.4010
0

According

theory.

0.40270

of the above

As the load

[A(2)] =

The resultant

0.1220

failure

0.1451

strain reaches

[,te, the 90 ply fails.

the longitudinal

by first-ply

[A(1)I= 0.1220 0.4027

0

is determined

resultant

the stiffness

0
0

properties

mpsi

0.1359

versus

strain

relation:

37

matrix

of the

90 ply lamina

of the laminate

change

becomes:

to E1 =

IN!2
t[0.6063
0.1174
= 0.1_74

The strain

components

ex = 1.5609

ey

in terms

Nx (2)

of load intensity

are expressed

Mohr's

circle,

strain

components

strain

failure

Nx (2) [18

e 45 = 0.5520

Nx (2) ge

Y45 = 2.0179

Nx (2) ge

the maximum

in +45 degree

criteria,

the failure

directions

load intensity

For the 0 ply lamina:

15300
1.5609

N_

as:

[18

e45 = 0.5520

Using

o.1359/Ly_

Nx (2) ge

-0.4570

Using

0.40100

- 9802

lb/in

longitudinal

direction

transverse

direction

(2)

5680

-12429

lb/in

0.4570

For the 45 ply lamina:

38

are determined

as:

can be determined

as:

15300

- 27717

0.5520
5680

lb/in

-10289

longitudinal

lb/in

direction

_ansverse

direction

0.5520
21000
-10406

2.0179
Comparing

these

load.

failure

The

tremendously
Laminate

Laminate

results,

(1) is 9802

the

the

load

drops,

Following

lb/in

0 ply

intensity

which

The

same approach

compressive

first

is 9802

means

procedure,

(3), [-45/0/45/0/9012s,

Compressive

fails

the

failure

in its longitudinal

lb/in.

After

laminate

fails.

that,

the

direction

with

strength

of the

Therefore,

the

the

tensile

strength

are determined

strength

of Laminate

as 10490

(2),

The results

the tensile

strength

strength

= 7435

lb/in

Laminate

(2), [-45/0/45]]9012s,

Compressive

strength

= 8930

lb/in

Laminate

(3), [-45/0/45]019012s,

Compressive

strength

= 14870

types

in Figure
considered.
panel

are

according

of possible

4, were

18 inches

of Damaged

damage

analyzed.

In order

of

to model
and

to the damage

and

respectively.

was

also used

for

are:
Compressive

Strength

lb/in,

of the laminates

(1), [-45/0/45/0/90]s,

Three

strength

[-45/0/45]]9012s,

lb/in and 19604

Laminate

Residual

laminate

of the laminates

used to determine

strength.

4.5

tensile

increasing

lb/in.

same

4.4

shear

Honeycomb

commonly
For each

found

damage

the skin between

24 inches,
configuration

Sandwich

in honeycomb
type,

circular

two stringers,

respectively.
as follows:

39

lb]in

Analysis

Panels

sandwich

panels,

and elliptical
the length

flaw

as shown
areas

and the width

is categorized

into three

were
of the
cases

Case1. A honeycombpanelwith variable sizesof circular or elliptical disbondsbetween

face sheetandcore. The maximumsizeof the circular disbondis 8 inchesin diameter. The
maximumsize of the elliptical disbondis 12 inchesin major axis. The aspectratio of the
ellipseis 3.
Case2. A honeycombpanelwith variablesizesof circular or elliptical delaminationsin one
face sheet.The maximum size of the circular delaminationis 8 inches in diameter. The
maximumsizeof the elliptical delaminationis 12 inchesin major axis. The aspectratio of
the ellipseis 3.
Case3. A honeycombpanelwith variablesizesof circular or elliptical notchesthrough one
face sheetor throughthe thicknessof the panel.The maximumsizeof the circular notchis 8
inchesin diameter.The maximumsizeof the elliptical notchis 12inchesin majoraxis. The
aspectratio of theellipseis 3.
Loading conditionsin the analysisinclude in-plane tension or compressionapplied at the
edgesof the face-sheetsof the sandwichpanel. Thedistributionof the load alongthe edgeis
assumedto be uniform.

4.5.1

Case

1:

It is observed
sheets

with a Disbond.

that the weakest

and

the

manufacturing
little

Panels

core.

The

defects.

strength

causes

When

degradation

from the core having

portion

of the

the panel

since

a much

the

smaller

However,

the

disbond

significantly

sandwich

panel

because

of local

the tensile

residual

strength

of a honeycomb

is loaded

load

may

the

of the face

sandwich

panel

40

be either

mostly

stiffness

reduces

is the bonding

in in-plane

is carried

in-plane

buckling

of the

damage

panel

between

impact

tension,

from

the local

by the face-sheets.

the face
outside

disbond
This

or
has

results

than the face-sheet.

in-plane
sheet

compressive
in the disbond

is assumed

to be the

strength
area.
same

of

the

Therefore,
as the total

tensilestrengthof the two face-sheets,andonly sandwichpanelsthat areloadedin in-plane

compressionareanalyzedto determinetheir residualstrengthdueto disbond.

4.5.1.1

Tensile

Based

Residual

on the above

disbond

loaded

thickness.
panel

Strength

discussion,

in in-plane

Since

the residual

tension

the tensile

has two face sheets,

of a Sandwich

Panel

strengths

of honeycomb

are independent

strength

of disbond

of the laminates

the residual

strengths

with a Disbond.

size

are already

sandwich

panels

and core

properties

obtained

of the sandwiches

should

Residual

tensile

strength

of the sandwich

with laminate

(1) = 19604

lb/in,

Residual

tensile

strength

of the sandwich

with laminate

(2) = 20980

lb/in,

Residual

tensile

strength

of the Sandwich

with laminate

(3) = 39208

lb/in.

4.5.1.2

Compressive

The

residual

compression

strength

The

of

is actually

the maximum
area.

Residual

load

value

the

the

sandwich

A novel

analysis

two-step

critical

sandwich
buckling

the lamina

panels

load,

to determine

Step 1. Determination

when

post-buckling

residual

strength.

41

be:

loaded

in

strength

here

post-buckling

occurs

on the size and

The

buckling

load

panel

to determine

the critical

buckling

and the sandwich

in the

shape

in-plane
means
disbond

of the disbond

of the core.

utilized

buckling.

and

with a Disbond.

post-buckling

depends

has been

to local

and ii) using

of critical

strength

The

and the stiffness

approach
due

can take

Panel

sandwich

strength.

panel

of the post-buckling
of the face-sheet

of a Sandwich

disbond-damaged

its post-buckling

area, the stiffness

damaged

Strength

with a

two

steps

post-buckling
are:

load and maximum

strength

i) determination
strain

criterion

of
of
of

Finite

element

model

Due to the complexity

method
code

must

be used to obtain

ABAQUS

laminate

connecting

to the

the

disbond.

Symmetric

element

model.

finite

element

model.

uniform.

Pressure

one

of the spring
stiffness

(1) Prescribe
sheet,

(2) Apply
amount

spring
depends

is comprised

a uniform

is applied

along

stiffness

and

the face
and

is modeled

layered

of the core material.

damage

area.

and

element

in Figure

the

loading

mesh

6 for

the left and lower

the top edges

sheet

by elastic

the corresponding

The finite

along

element

of the face

used in

an elliptical
edges

of the

sheet

of the

mesh

is not

Stiffness

on the local

of the damage
to each
density

the following

node

of the

of the mesh.

method

on all the nodes

area, the finite

face

element
sheet,

In order

the

equivalent

to determine

the

is used:

of the finite

element

model

of a face

of shell elements.

pressure

of the pressure

are prescribed

is connected

constraint

disbond

finite

freedom

core

disbond

to be analyzed.

at the edge

of each spring,

out-of-plane

which

Spring

of nodal

within

of the

element

In this analysis,

honeycomb

the springs

a finite

commercial

compressive

5 for a circular

load

with a disbond,

16).

degrees

The

geometry

needs

conditions

concentration

When

equivalent

in Figure

of Equivalent

Due to the stress

stiffness

of the panel

(Ref.

six

by disconnecting

boundary

finite

with

with equivalent

the

panel

The general-purpose

property.

of the panel,

is shown

Determination

sheets

is modeled

only a quarter

analysis

elements

as the material

symmetry

condition,

the solution.

shell

two face

disbond

of the sandwich

as the tool for this task

by

elasticity

The local

Due

is adopted

is modeled

orthotropic
springs

of the geometry

on the top surface

is equal

of the face-sheet

to the out-of-plane

42

stiffness

finite

element

of the honeycomb

model.
core.

The

(3) PerformABAQUS static stressanalysisto obtainreactionforces at all the nodes. The

reactionforce at eachnodeis the equivalentstiffnessof the springthat will be connected
to the samenode.
(4) Convert the ABAQUS output data format into its input data format for these spring
stiffness'througha Fortranprogram.
(5) Alterationsof the corestiffness,asrequiredin theparametricstudy,only needto multiply
the ratio of the core stiffnessto the known equivalentspring stiffnessthrough another
Fortranprogram.
Sincethe disbond sizesare variable,the aboveprocedureswere used oncefor the finite
elementmodelassociatedwith eachdisbondsize.

Verification

of Finite

In order

to obtain

solutions

of buckling

panel

rectangular
Timoshenko's

computation
of 0.005

solution.

Results

of Critical

disbond

of varying
in Figure

Load

constraint
The

The

8.

The material

is reached

Disbond

43

Poisson's
between

Size.

with

with

out-of-plane

to the face sheet is on

buckling

loads

computed

of an isotropic

and compared

with existing

properties

with

rectangular

solutions

in Yin

used in the verification

ratio v = 0.3, and a plate

the present

available

load of a sandwich

case of an isotropic

size is also checked

E = 10 Msi.,

Versus

7.

verified

of the critical

sizes were

in Figure

was

only a face-sheet

area.

of varying

17) as shown

Buckling

solutions

the disbond

agreement

model

The out-of-plane

(Ref.

Reasonable

Determination

not available,

disbond

modulus

Load

element

closed-form

within

18, 19) as shown

known

finite

are generally

with a circular

are Young's
inch.

the

Since

the nodes

solution

(Ref.

for Buckling

in the verification.

with an elliptical

and Jane

results,

buckling

except
plate

Models

analyses.

is considered

all the nodes

plate

reliable

due to local

constraint

Element

results

and

thickness
the well-

The

critical

disbond

buckling

loaded

load

of a face-sheet

in compression

above.

The results

Figure

9 and Figure

was

of the critical

computed

buckling

disbond.

Parametric

studies

conducted

thickness.

In Figure
the

buckling

thickness

10, laminate

When

versus

between

the

increase

thickness

Step 2. Determination

and shown

parameter
of plies

To provide

in each

critical

damage

buckling

size

in critical

is small,

buckling

load.

for the other

load

and

This

the

damage.

than

a small

or laminate

sequence.

data points,

the

Even
critical

In Figure

an intermediate

two laminates.

laminate

increase

in

10 and Figure

orientation,

disbond

described

are shown

sequence

is stacking

enough

sequence

model

in Figure

stacking

in the case of elliptical

stacking

element

or an elliptical

size relationship

of lamina

percentage

is varied.

a different

disbond,

disbond

finite

damage

12, the varying

same

a circular

the

load versus

in terms

especially

the disbond

in a large

stiffness

the

thickness

using

relationship

observed.
results

have

are different,

is plotted

nonlinear

also

11 and Figure

laminates

loads

and Figure

were

with

using

11 for the case of a circular

12 for the case of an elliptical

though

laminate

thickness

in laminate

is consistent

with

is

thickness
the

bending

relation.

of Post-Buckling

Residual

Strength

due to Local

Buckling.

Background

In general,

post-buckling

determined
criterion
thrust

through
(Ref.

of

this

residual

strength

a geometrically

18, 19, 20, 21).

research

of a laminate

nonlinear

finite

This type of analysis

effort.

Thus,

an

face-sheet

element

analysis

is extensive

efficient

with local damage

and

a material

and is presently

engineering

approach

can be
failure

beyond
needed

the

to be

developed.

In the
assumes
elastic

earlier

work

of Dost,

that the effect

modulus

et al. (Ref.

of the buckled

for the inclusion

is used

22),

area

a simple

model

on the laminate

in determining

44

was

proposed.

This

is a soft inclusion.

the residual

strength

model

A reduced

of the damaged

laminate.

Obviously,

laminates,

and is not suitable

Model

for post-buckling

According
carried

by the buckled

carries

normal

strain

strength

a constant
in the

loading

Other

of the laminate

location

additional
reaches

The

load

additional

portion.

This

is carried

loaded

structure

portion

of the laminate.

on these

assumptions,

area.

location

by the

load.

load.
is the
The

laminate

may exist.

buckled

load.

area

This

but

load

causes

no

stress

of the buckled

area.

the total

sum of the

laminate

strain

fails

when

excluding

the

material

laminate

of the notch

strain

Pcr and Pa## denote

The additional

by the

the total

strain

of the

same

strain

the undamaged

of

notch

as buckled

is the notch-tip.

residual

strength,

Ppo,t should

be:

the critical

buckling

by maximum

load

and

normal

the

strain

additional

criterion,

load,

respectively.

that is:

(Po_ + Pa_ K)eo = em,x

K is the notch-tip

buckled

(4.5.1-1)

load is determined

in which

at the

caused

with a through
are the

(4.5.1-2)
strain
laminate

concentration
in the loading

45

factor,

eo and

direction

due

em,x represent
to unit

load

or strain

Ppost = Pot + Pa,_,_

where

of the

ply lamina.

The size and the shape

the post-buckling

the

However,

as the original

location

load can still be

of deflection

laminate,

of the edge

area

is modeled

The critical

Quasi-isotropic

are made.

the

additional

buckling

level

buckling

limit of the zero degree

load

and

a constant

of the laminate,

throughout

of the buckled

and the critical

at the buckled

edge

can carry

occurs,

assumptions

that is, the critical

at the critical

strain

buckling

but an uncertain

three

direction

of the edge

located

Based

even

the maximum

when

to the

load,

occurs

portion

theory,

is applied

plates

determination

the following

load

for isotropic

study.

of the structure,

concentration

critical

portion

As an increasing

be used

for the present

stability

to this concept,

laminate

can only

residual

to linear-elastic

According

this model

the
and

uniform

maximum

strain the material can endure,respectively. As listed in the lamina property section,the
maximumcompressivestrain of the presentlamina is Cm,x
=13500ge. eois derived in the
laminate property determinationproceduresand the results are that eo = 1.8158 ge for
Laminate (1), eo = 1.5117 ge for Laminate (2), and eo = 0.9079 ge for laminate (3),
respectively. Since K
unknowns,

can

be determined

Ppo_t and P,_dd in Equation

by

finite

element

(4.5.1-1)

and

analysis,

Equation

there

(4.5.1-2).

are

only

Solving

two

for Ppo_t

yields:

,./1

eP"= e'+Leo
Equation
panels

(4.5.1-3)
with either

K is calculated
results,

the

reduction

method

or an elliptical

a finite

element

analysis

of the

Element

element

disbond

solution,

material

the finite

element

analysis

Figure

13 and Figure

loading
finite

direction
element

evaluation
disbond

Results

is selected

of notch-tip
and elliptical

of Post-Buckling

using

The

ABAQUS.

element

strength

model

strain

concentration

In order

as well

of the sandwich
factor

to obtain

as the

reliable

subsequent

data

for Notch-Tip

are

property

In order

of isotropic

elements

elasticity

to make
is used.

to the notch-tip.
along

Analysis

as that for the buckling

removed.

is extrapolated

Strength

analysis

except

that

comparisons

with

the

computed

by

The

stress

The computed

the line of symmetry

stress

are used.

values

from

A second

order

to fit the data.

14 show

and the
model

disbond.

were constructed
area

to the notch-tip

residual

in advance.

Models

models

classical

polynomial

finite

must be examined

in the

closest

post-buckling

disbond

of Finite

elements

to calculate

a circular

by

The same finite

three

used

appropriateness

Verification

the

is then

the comparisons

analytical

is obtained.
strain
disbond

of the computed

solution
The

to determine

(Ref.
same
the

23).

Acceptable

extrapolation
strain

Strength

Versus

concentration

46

Disbond

stress

accuracy

procedures

situations.

Residual

normal

Size

factors

are

distribution

in

of the

simple

used

in the

for both

circular

The post-bucklingresidual strengthof the sandwichpanel with a circular disbond or an

elliptical disbondon a face-sheetlaminateloadedin compressionwas computedusing the
methoddescribedabove. The resultsof post-bucklingresidualstrengthversusdamagesize
relationshipareshownin Figure 15andFigure 17for the caseof circular disbond,andshown
in Figure16 andFigure18for the caseof elliptical disbond.
Parametricstudieswere also conductedin terms of lamina stackingsequenceor laminate
thickness.In Figure 17 andFigure 18, the varying parameteris stackingsequence.Thereis
no significant differencein the post-bucklingresidual strengthas a result of variation in
stackingsequence.In Figure 15 andFigure 16,the varying parameteris laminatethickness.
An almostlinearrelationshipbetweenthe critical bucklingload andthe laminatethicknessis
observed.Whenthe damagesizeis small,the residualstrengthis limited by the compressive
strengthof the sandwichpanel, which is the total compressivestrengthof the two facesheets.This resultmeansthatthe sandwichpanelcantoleratesmallareasof disbondwithout
reductionof compressivestrength.

4.5.2

Case

Another

2: Panels

commonly

the face-sheet

observed

laminate.

or manufacturing
general,

a variable

the delamination
analysis

must

this research
is selected

The

laminate

effort,

case

analyzed

mechanism

Delaminations

defects.

When

level

of strength

be conducted

can be assumed

damage

do not have

to show

simple

with a Delamination

and

same

as acting

in-plane

extensive

since

Since

in the plane

occurs

there

is not intended

in the plane

of the

be, in
sides

of

mechanics

For the purposes

and only a simple

of
case

data trends.

of symmetry

carried

should

fracture

are required.

of

from outside

on the opposite

Consequently,

the load is primarily

47

damage

tension,

the materials

and the relative

of symmetry

is the delamination

impact

in in-plane

computations

of the approach

panels

by either

strain.

and computation

is that the delamination

in the study.

is loaded

degradation

such an analysis
the feasibility

may be caused

the panel

the

in sandwich

of the

by the face-sheet,

laminate.

In this

symmetric
the load

situation,

the

materialon two sidesof the delaminationwill still havethe samein-planestrainandthe load

carryingcapacityof the laminateis not affected.

However, the delaminationsignificantly reducesthe in-plane compressivestrengthof the

sandwich panel becauseof local buckling of the face sheet in the delaminatedarea.
Therefore,the tensileresidualstrengthof the sandwichpanel is assumedto be the sameas
the total tensile strengthof the two face-sheets,andonly sandwichpanelsthat areloadedin
in-planecompressionwereanalyzedto determinetheir residualstrengthdueto delamination.

4.5.2.1

Based

Tensile

Residual

on the above

delaminations
properties

discussion,

loaded

panel

the

in in-plane

and thickness.

the sandwich

Strength

Since

of a Sandwich

residual
tension

has two face sheets,

strengths
are

the tensile

Panel

of honeycomb

independent

strength

the residual

with a Delamination

sandwich

of delamination

of the laminates
strengths

of the sandwich

tensile

strength

of the sandwich

with laminate

(1) = 19604

lb/in,

Residual

tensile

strength

of the sandwich

with laminate

(2) = 20980

lb/in,

Residual

tensile

strength

of the Sandwich

with laminate

(3) = 39208

lb/in.

4.5.2.2

Compressive

The

residual

compression
the

load

the

of a Sandwich

delamination

damaged

its post-buckling

strength.

sandwich

panel

area.

The value

in the delamination

and shape
stiffness

of the

is actually

maximum

occurs

strength

Strength

of the delamination

can

take

Panel

sandwich

of the core.

48

core

obtained

and

be:

with a Delamination

panel

loaded

post-buckling

strength

when

post-buckling

of the

strength

of the sub-laminate

with

and

should

The

of the post-buckling

area, the stiffness

size

are already

Residual

Residual

panels

depends

in in-plane
here

means

sub-laminate
on the size

of the face-sheet

and the

The two-stepanalysisapproachto the residualstrengthof sandwichpanelsdue to disbond

wasalsousedto determinepost-bucklingresidualstrengthof damagedsandwichpanelsdue
to delamination.

Step1. Determinationof Critical Buckling Load

Finite

element

In order
bonded

to model
layers

disbond.

nodal

sheets

point.
with

plane
shown

equivalent

of the

the geometry

in Figure

The same

of element

core is modeled

panel

sheet

was used

by elastic

stiffness

of

the

layers

two

is represented

the

by two

as for the panel

by rigid beam
springs
core

elements

material.

with a
at each

connecting

of shell

of the finite

element

layer

of the face-sheet

finite

element

model

of Finite

used

damage

the two
The

elements

with a disbond

and in Figure

local

within

for Case

1 were

of the springs

the

is always

bonded

to a spring

Models

of damage
to the core.

regardless

for Buckling

49

The

were used,

along
along

of the
same

in-

which

are

delamination,

the left

and

the top edges

lower
of the

model.

used for this case.

are independent

symmetry

6 for an elliptical

load is also applied

as in the disbond

to the
condition.

and pressure

Stiffness

Element

loading

are also prescribed

Spring

is connected

owing

conditions

model

laminate

and the

of a panel

delamination

model,

of Equivalent

procedures

to be analyzed,

as in the analysis

boundary

element

needs

delamination

meshes

this case the stiffness

Verification

sandwich

Symmetric

Determination

face

are connected

disconnecting

5 for a circular

of the finite

face sheet

type

compressive

by

of the

element

respectively.
edges

The honeycomb

delaminated

area.

a quarter

finite

The same

of the shell elements

is modeled

corresponding

panel,

the locally

of shell elements.

delamination

Only

the delamination,

The two layers

in-plane
face

model

Load

The only difference

size.
Every

The reason
node

is that in
is that one

of this layer

of the size of the delamination.

Determination

of the

The

verification

of the

provided

in the

damage.

Hence,

Results

of Critical

The

critical

elliptical

Buckling

above.

thickness.

In Figure
the

buckling

were

loads

also

have

To

different

stacking

sequence

between

the critical

buckling

size

is small,

buckling

load.

provide

the

In Figure
data

than

laminate
was
buckling

of circular

of delamination

delamination

the finite

damage

or an

element

model

size relationship

delamination,

and shown

parameter

19 and Figure

the

stacking

of plies

other

two

in each

are

in Figure

Residual

stiffness

due to Local

using

the damage

increase

thickness

critical

relationship

When

in a large

Even

is laminate

A nonlinear

versus

Strength

the

is plotted

is observed.

results

sequence.

parameter

thickness

laminates.

or laminate

orientation,

20, the varying

thickness

thickness

sequence

is stacking

an intermediate

with the bending

of Post-Buckling

using

of lamina

percentage

points,

for

a circular

load versus

in terms

in laminate

This is consistent

with

computed

load and the laminate

increase

Step 2. Determination

case

determination

Size

22, the varying

same

enough

a small

load

delamination.

conducted

are different.

thickness.

for the

21 for the case

21 and Figure

laminates

is still meaningful

Delamination

22 for the case of elliptical

studies

though

Versus

of the critical

19 and Figure

Parametric

buckling

in compression

The results

in Figure

critical

of a face-sheet

loaded

for

here.

Load

load

analysis

damage

it is not repeated

buckling

20 and Figure

element

of disbond

delamination

described
shown

section

finite

in critical

relation.

Buckling

Background
The

method

for determination

described

in the section

exception

is that the assumed

buckled
thickness.

layer

of post-buckling

of disbond
notch

damage,

residual
is still valid

in the model

of the

sub-laminate

of the

The same

extrapolation

procedures

strength

as those

5O

for the case

described

face-sheet

due

laminate
used

to local

of delamination.

in the last section

instead

buckling,

of being

as
The

is only on the
through-the-

for the case of disbond

damages

are used in the evaluation

of notch-tip

the case of delamination

damages.

Results

Residual

of Post-Buckling

The post-buckling
elliptical

residual

delamination

the method

and shown

are shown

were

In Figure

no significant
stacking

The

results

sequence.

conducted

between

observed.

When

the damage

size is small,

sheets.

This result

4.5.3

Case

Severe

impact

3:

convenience.
damaged
this

behavior
from

were

occurring

open

Notch

the stress

hole

data

assumed

strength

and many

have

in a face
can then

parameter

panel

load

strength

can tolerate

delamination,

sequence

There

is laminate

is limited

is

of the variation

and the laminate

in

thickness.
thickness

is

by the compressive

strength

small

or laminate

sequence.

as a result

compressive

sheet

laminate

be modeled

to be applicable
The notched
factor

material

of the laminate.

theories

total

of the two

face-

areas of delamination.

been

cracks,

to both

strength

open

stress

Extensive
developed

hole

strength

concentration

research

to describe

51

treated

as an open hole

and so notch

of the laminate

and the unnotched

in the

is often

as a through-the-thickness

can also be used to model

concentration

of the laminate

the residual

damage

delamination.

stacking

strength

buckling

is the

versus

using

with Notches

of the sandwich.

strength
global

damage

which

strength

is stacking

24, the varying

the critical

that the sandwich

Panels

The

area.

study

means

residual

for

or an

was computed

25 for the case of a circular

parameter

23 and Figure

panel,

residual

26, the varying

factors

delamination

in compression

of lamina

relationship

sandwich

with a circular

in terms

linear

of the

loaded

concentration

Size

26 for the case of an elliptical

An almost

strength

panel

of post-buckling

in the post-buckling

In Figure

the strain

Delamination

23 and Figure

25 and Figure

difference

Versus

laminate

in Figure

also

to determine

of the sandwich

24 and Figure

studies

thickness.

strength

above.

in Figure

Parametric

Strength

on a face-sheet

described

size relationship

strain

has been

and

notch

analyses

cannot

be directly

of the laminate.
area

is usually

conducted

this phenomenon

for

crack
calculated

In fact,

higher

in past

in the

results

through-thickness

for

than

the
the

two decades

(Ref. 23, 24, 25, 26, 27,

28). However,dueto the complicatednatureof the compositestructuresubjectto damage,

up to now there is still no meansto determinenotchedstrengthwithout testing a notched
laminatespecimen.
Theresidualstrengthof a sandwichpanelwith a notchon its oneor two face-sheetlaminates
or through its thicknessis primarily determinedby the residualstrengthof the constitutive
laminates.Therefore,testdataon theresidualstrengthof a laminatewereused. The dataare
characterizedby the well-acceptedMar-Lin model (Ref. 25).

However, the Mar-Lin

parametersare for laminatesthat are infinitely large and havedifferent lamina tensile and
compressivestrengthsfrom the materialusedhere. In orderto usethe test datato determine
residual strengthof the sandwichpanels,three-dimensionalfinite element analyseswere
performedto makethe correction.

Mar-Lin

Model

The fracture
_

CN

Parameters

model

proposed

by Mar and Lin is expressed

in the following

Hc

form:

(4.5.3-1)

(2C)'Y

in which:
ey = = remote
m

= order

strain
of singularity

Hc = composite

fracture

= half crack

length

= geometry

The basic

toughness

factor

material

properties

and the Mar-Lin

model

In tension:
Ply strength

in fiber direction,

X = 250 ksi,

52

parameters

used

in this study are:

Laminate(1) [-45/0/45/0/90]s

Hc = 0.00667

in"',

m = 0.225,

Laminate(2) [-45/0/45/901es

Hc = 0.00667

in"',

m = 0.203,

Laminate(3) [-45/0/45/0/90]es

H_ = 0.00667

in"',

m = 0.225,

In compression:
Ply strength

in fiber direction,

X' = 220 ksi,

Laminate

(1) [-45/0/45/0/90]s

H_ = 0.0034

in"',

m = 0.35,

Laminate

(2) [-45/0/45/9012s

H_ = 0.0043

in"',

m = 0.26,

Laminate

(3) [-45/0/45/0/90]

H_ = 0.0034

in"',

m = 0.35,

The remote

load intensity

2s

is then calculated

by:

PN = eN EN TL

in which

(4.5.3-2)

EN is the in-plane

thickness

of the laminate.

elastic

modulus

For the laminate

of the laminate
used here,

in loading

these parameters

direction

and Tc is the

are:

In tension:
Laminate

(1) [-45/0/45/0/90]s

EN = 11.653

mpsi,

Tc = 0.05 inch,

Laminate

(2) [-45/0/45/901es

EN = 8.841

mpsi,

Tc = 0.08 inch,

Laminate

(3) [-45/0/45/0/90]

EN =11.653

mpsi,

Tc = 0.10 inch,

es

In compression:
Laminate

(1) [-45/0/45/0/90]s

EN = 11.233

mpsi,

Tc = 0.05 inch,

Laminate

(2) [-45/0/45/901es

EN = 8.541

mpsi,

Tc = 0.08 inch,

Laminate

(3) [-45/0/45/0/90]

EN = 11.233

mpsi,

Tc = 0.10 inch,

The
Section

elastic

modulus

4.2 using

lamination

es
values

are

calculated

theory.

53

from

lamina

properties

given

in

Finite Element

Two

finite

notches.

Model

element
One

models

were

notch.

as that used in the analysis

by eight-node
face

sheet

point

solid

elements

and the core

of the 3-D finite

with

is built

loading

same reason
the panel

needs

the

same

as those

in the

conditions

are also prescribed

Geometry

Factors

The

geometry

sandwich
element
60 times

Results

panel

the notch

of Residual

Strength

in both tension

compression

laminate

connection

positioned

previous

case

of one

of the sandwich

of both

cases.

is the

at every

each

in-plane

nodal

panel

and its

panel.

the mid-plane

finite

sheets

between

to be analyzed.

in-plane

is for the

the face-sheet

element
The

panel,

meshes

boundary

face-sheet
notch

case is that the laminate

the

having

of

having

and

the

the

case

a
are

loading

cases.

a notch

of variable

core

in this analysis

conditions

as in the previous

to the

only one eighth

and
used

Due

and

size were

of the

obtained

by finite

size is 90 in 90 in, which

is at least

size.

that the Mar-Lin

face-sheet

The

with

core is modeled

of the panel

a through-the-thickness

The baseline

The honeycomb

only a quarter

in the same manner

for the

having

analysis.

Assuming
strength

factors

The

of the two face

of the notched

case

panels

and the other

the geometry

for the

notch.

sandwich

a notch,

about

to be analyzed

through-the-thickness
also

having

model

elements

of the sandwich

allow

and the symmetry

element

properties.

rigid beam

model

condition

elastic

of the

is in one face-sheet

with a disbond.

linear

with

element

The finite

of a panel

In the case of only one face-sheet

associated

for the analysis

of them is for the case that a notch

case of a through-the-thickness
same

constructed

Versus

model

Notch

parameter

and compression,

having

was calculated

Size

a circular
according

Hc is proportional

the residual
notch

strength

or an elliptical

to the formula:

54

to the corresponding
of the sandwich
notch

loaded

lamina

panel

with a

in tension

or in

2HcENT L S
(2c)'Y
SO

PR =

in which
fiber

S denotes

direction

(4.5.3-1),

laminates

carrying

results

Figure

results

the

test.

and So the ply strength

Equation

assumption

(4.5.3-3)

on Hc and

in the

is derived

using

consideration

of two

residual

strength

versus

notch,

shown

were

also conducted

relationship

shown

results

in Figure

32 and Figure

size

and

The

are shown

in Figure

studies

notch,

respectively.

size relationship

and

damage

in Figure

shown

34 for the case

in

28 and Figure

of the compressive

31 and Figure

in terms

are

residual

33 for the case of

of an elliptical

of laminate

notch,

thickness

and the

in the figures.

Discussion

found

laminate,
factor

that

the

the

stiffer

for selecting

material,

the

stiffness

must

core

stiffness

cannot

the core

must

be to avoid

core

stiffer

core

stiffness
will

be determined

is the

be an independent

weight

generally

global

buckling

of the

be heavier.

for each face-sheet

variable.

The

of the face

sandwich

panel.

Therefore,

the

thicker

sheet.

For the
best

core

the

Another
same

core

density

or

laminate.

Summary

Methods
or notch

in the

29 for the case of a circular

Parametric

are given

It was

tensile

damage

notch,

respectively.

used

(4.5.3-2),

of an elliptical

versus

a circular

specimen

in the fiber direction

load.

27 and Figure

strength

4.7

test

ply strength

Equation

of the

30 for the case

4.6

the current

of the

Equation

The

(4.5.3-3)

for residual
have

Appropriate
(disbond,

been

finite

strength

analyses

of sandwich

panels

with a disbond,

delamination

developed.

element

delamination

models

and notch)

for the residual

have

been

55

strength

constructed.

analyses

of the

three

cases

Resultson critical buckling loads andpost-buckling strengthfor sandwichpanelswith

disbondor delaminationhavebeenobtained.

Results

on residual

a through

the thickness

Parametric
sheets

strength

studies

and

have

or a notch.

Both

and

analysis

been

post-buckling

delamination

tension

notch

for sandwich
have

conducted
residual

compression

for all the cases

been

loading

panels

with a single

notch

on one face

sheet

or

obtained.

to determine
strength

conditions

studied.

56

of

the

critical

sandwich

are included

buckling

panels

load

with

in the damage

of face
disbond,

tolerance

5.1

DEMONSTRATION

METHOD

Introduction

The general
been

method

defined

concept,

in Section

behavior

methods

outlined

versus

flaw

Using

this

determine

data,

Detailed

"Level

the

"Level

are

chosen

the

in the

safety

in the following

approach

to an advanced

have

structural

The

deterministic

strength

is applied

the design

design

for analysis.

using

shown

methodology

assumptions

in Section

damage
analysis

of a sandwich

panel

figures

of the

Appendix.

to the

sandwich

panel

criteria

for a given

to

load level.

section.

inherent

3.2.
process

The

in the

overall

"Level

of Safety"

method

here to illustrate

how

of

method

application

the design

have

been

is restated

and

procedure

progresses

examples.

Evaluation

The "Level

of Safety"

1.

damage

data

Structure

structure

on the distributions

methods

Estimate

the probability
structure,

on Existing

of an existing

detection

existing

was

of the residual

of damage

to meet

as applied

investigated

charts

of Safety"

of Safety"

Procedures

into a step-by-step

of Safety"

Collect

been

are described

and

previously

types

system

has

Design

"Level

the process

material

necessary

of Design

the given

4.

for various

of this example

condensed

To demonstrate

of the composite

derivations

defined

for the equivalent

honeycomb

in Section

size

Outline

using

3.2.

the size of panel

The details

5.2

and equations

a composite

tolerance

2.

OF DESIGN

for an existing

using

can be determined

through

of flaw

types,

structure,

of detection
the assumptions

sizes,

as described

curves

for each

described

57

flaw

in Section

type

in Section

the following

number

of flaws

and

3.4.

of damage
3.3.

steps:

detected

on the

Using

the data

the structure
structure
value

"Level

Use

from

as a function

using

the

of "Level

of Safety"

accomplished

through

5.

7.

Assume

distributions

from

Steps

the allowable

Bayesian

data

value

the "Level

3.2.

of Safety"

This result

for the

strength

sizing

damage

design

becomes

of new

formulas

of the new

parameters

size po(a)

and engineering

strength

of

of the existing
the

allowable

structures

can be

structure

as a function

of

of interest.

and probability

of detection

curves

judgement.

derived

the relation

of "Level

equal to or greater

structure

has

during

in Section

between

3.2,

structural

combined
geometry

with

the

data

and "Level

of

statistical

of Safety"

been

built

of Safety"

from

Step 3, size the structure

to give

than the allowable.

and

put

into

service,

collect

data

on the

damage

operation.

methods

can be used to update

the "Level

on the residual

load levels.

of Safety"

the

structural

5, determine

Using

approach

the residual

of Safety"

4 and

in Section

data

steps:

of detected

on available

for given

accumulated

to determine

Safety"

After

derived

the following

"Level

size, determine

of Safety"

size and the various

"Level

8.

"Level

damage

the

design

for Design

analysis

Using

with

for new designs.

Perform

PD(a) based

formulas

Approach

equivalent

1 and 2, coupled
of damage

of Safety"

of the

4.

Steps

defined

the assumed

in Section
probability

of the structure.

58

3.6, combined
distributions

with the data from

Step

of Step 5, and to recalculate

8,

10.If the "Level of Safety" falls below the allowablevalue,the inspectionandmaintenance

programcan be updatedto provide earlier or more frequentdetectionopportunitiesto
returnthe "Level of Safety"to its designvalue.

5.3

Examples

Some

illustrative

safety

design

of Equivalent

examples

approach.

will at a minimum
allowable

Level

Probability

For

the

sake

damage

The

of Safety,

damage
size

same

Level

of Safety

taken

from

the metal

consider

types
for

Gamma

po(a)

distribution

parameters

application

sandwich

structure,

of the

aluminum

is 99.9%.

damage

35.

associated
the

II.

it.

(a)

The

type

Using
above,

equation
the

numerically
FORTRAN.

diameter

(3.2-10)

normalizing
using

that

one.

The
into a

different

(disbonding),
types

of detection

detection

of disbond

probabilities

k = 2.1,

damage

0 = 1.3

Type

detected

is assumed

for

damage
I and

as a function

type has a different

detected

one

of inspection

1 - e -a14/L.2 for Inspection

inspection
The

parameters

0 = 1.5 for Inspection

PDI

damage
Two

The probability

Each

with

one

structure.

to be

Type

in Figure

Inspection

of

damage

to follow
Type

I, and

Type II:

(5.3-1)

probabilities

to be the disbond

panel

This translates

po(o) - 0kr(k-----S
Detected

equivalent

compression

as an existing

of only

the damage.

are assumed

with

k = 2.2,

a case

in the sandwich

Inspection

size are plotted

distribution

the

the

and one flaw

-a14/2

to demonstrate

of 0.1%.

for the two inspection

1--e

here

Design

a composite

are to be used to detect

PvII (a)=

Based

task is to design

of simplicity,

location

methods

are used

maintain

of Failure

Safety

are plotted

in inches

from

the SLATEC

For each iteration,

36.

The damage

size variable

with the probability

distributions

a is taken

for each function.

Section

constant

in Figure

3.2, along
for

DQAGI
relative

the

"Level

Gaussian
and absolute

59

of

Safety"

quadrature

calculation

subroutine

error for convergence

is

assumed
integrated

in double-precision
of the subroutine

was set at 10 12.

To illustrate

integral

are listed

solutions

Table

the

in Table

limit of integration

ofNormalizing

distributions,

c approaches

Constant

TypeI

l/C

Inspection

Type

10 4

1.80730322

2.22955533

10 5

1.80859168

2.23020679

10 6

1.80884877

2.23031005

10

1.80890007

2.23032641

1.80891030

2.23032900

error

was

shown,

the integral

distribution

for the

ofp(a)

Equation

detected

is assumed

to be a valid

probability

function

of the

80%

panel

plotted

by

sizing

Alternatively,
and

4.

in Figures
in the
plots

37.

are shown

calculated

residual

of laminate
38 through
figures,

as these
load versus

in Figures

design

Data

size

the most
behavior

types,

and the

and

points

"Probability
45.

the "Level

size and inspection

in an Excel
from

versus

load

critical

and "Level
stacking

outside

It is more

type.

of Safety

damage

size

of Safety",

as a

sequence,

are

of Safety"

the design

of Failure"

of

spreadsheet,

the Level

with a "Level

are generally

42 through

60

scheme

strength

thickness

41.

SO

on the

previously,

of disbond

of disbond

between

Based

9,

for both inspection

a trapezoidal

Elimination
the

at 10 8.

of 10

lI

function.

constants

using

Relations

of design

value

as a function

substituting

parameters

are not shown

also be obtained,

numerically

in Section

and shown

is calculated

in Figure

be performed
obtained

established

and the normalizing

at c values

are listed

to a finite

are integrated

relations

constants

to converge

(3.2-13)

with the results

can

normalizing

subroutine

is assumed

of the sandwich

The equations

by the integration

the

zero.

2.22544611

estimates

interest.

assumed

1.80084780

reliable

than

of the

10 3

Roundoff

plots

Table

Inspection

10 8

Safety"

properties

1 as the lower

1. Convergence

Using

convergence

lower

range

relationships

practical

then

of
can

to use these

curvesratherthanthe "Level of Safety"plots since"Level of Safety"is so closeto unity for

thetypical designrangeof interest.
For a given limit load P/ira
sequence

necessary

the charts
0.1%,

Inspection

a "Level

42 through

Inspection

Disbond

45.

these

Disbond

thickness

the metal

and

structure

"Probability

stacking

is found

from

of Failure"

PF,_llow of

type

a strong

(-45/0/45/90)2s
(-45/0/45/90)2s

min. laminate

Disbond

results,

it can be seen

detection

probability

Inspection

the

Type
largest

of Safety"
or some

than

laminate

above

that

combination

may be utilized,

different

inspection

detection

method,

aircraft

components

accept

the

lower

trade
types.

structure.
Type

used

thickness

1 and

substantial

subject

In general,

detection

and

metal

2, a smaller

61

with

This means

that

the method

if
the
the

Type

1 is

(-45/0/45/90)2s

for the design.

considerations
increase

of the

for

a specific

to the inspection

of actual

standpoint,

Inspection

for

to maintain

Inspection

thickness

and operational
probabilities

program

be used

laminate

from a life-cycle

associated

35.

If only

costs

in applying

lower

must

to the

as detection

a somewhat

maintenance

structure.

have

2 has

in Figure

and cost savings

Therefore,

probability

inspection

Type

weight

the cost and difficulties

also increase.

Inspection

of (-45/0/45/0/90)2s

of the existing

is also

of the inspection

1, as is shown

in the

of Types

thus realizing

of design

of the

2 were

(-45/0/45/90)2s

that the parameters

size

Inspection

(-45/0/45]0190)2s

- min. laminate

on the resulting

This type

above

At an allowable

- min. laminate

influence

used,

laminate

Type II

Elliptical

"Level

of Safety"

min. laminate

Disbond

Circular

structure,

the minimum

Type I

Elliptical

only

lbs./in.,

are:

Circular

From

to provide

of Figures

the results

of 8000

Type

it may be better
2 and

use

to

a larger

laminatethicknessin the final design. The "Level of Safety"methodology,andthe datathat

results from it, allows the designer to make these types of cost-benefit trades on a
quantitativebasisearlyin thedesignprocess.
Once the structureis in service,the data collected on the damagesizesdetectedduring
operationalinspectionsis usedwith the Bayesianstatisticalmethodto updatethe assumed
damagesizedistributions. Sincean exampleof this processis describedin Section3.6, it
will not berepeatedhere. The reviseddamagesizedistributioncanbeusedto recalculatethe
"Level of Safety" of the structure,thus validating the assumptionsmade in the design
process.
A more detailed exampleis also shown to illustrate the equivalent safety methodusing
multiple damagetypes in different locations. The composite sandwichpanel discussed
previouslywill be redesignedfor a load concentrationat a particularlocation. The far-field
loading areaon the panel is referredto as Location 1, andthe load concentrationareais
referredto asLocation 2. Location 1 usesa laminatestackingsequenceof (-45/0/45/90)2s,
and Location 2 usesa (-45/0/45/0/90)2sstackingsequence. Location 1 is subjectto two
flaws: a circular disbondandanelliptical delamination.Location2 is subjectto anelliptical
disbondanda circular delamination.Each flaw type at eachlocation hasa uniquedetected
damagesize distribution po(a)
nomenclature

of each curve

at each damage

Location

is listed

detection
below,

probability

where

PD(a)

t is an array

associated

with

of the mean

number

location:

1
Circular
Elliptical

Location

and

Disbond

(a = diameter)

Delamination

(a = major

axis)

poll(a),

PDll(a),

tll = 1

po21(a),

Pml(a),

ll21 = 1

po12(a),

PD12(a),

1112 = 1

2
Elliptical

Disbond

(a = major

axis)

62

it.

The

of flaws

CircularDelamination(a = diameter)

For

simplicity,

it is assumed

curves

of Equation

Type

I is to be used

the same as those

(5.3-1).

limit load

= 17,500

lbs./in,

structure

is 99.9%.

PFll

PF21 = 0.0.

5.

Elliptical

Location

distributions

for each

= Pol(a)

poll(a)

= Po2da)

= pozz(a)

PD12(a)

= PD22(a)

= PDI(a)

PDll(a)

= PD21(a)

= PDII(a)

lbs./in,

is applied

2.

The

disbond

of Failure"

into a "Level

disbond

at Location

For the

circular

sandwich

53 is used

tolerance
limit

load

2, Figure

"Level

1, Figure

at a 6000
of LSll

of Safety"

lb./in,

Inspection
type

are

for the entire

44 can be used to
load.

= 0.9999.

The result

analysis

shows

for any

of the damage

of Failure"

that a panel

at Location

7. These

results

2, Figure

are listed

sizes

PFll

= lxlO

4, LSll

= 0.9999

Delamination

PF21 = 0.0,

Disbond

PF12 = 6x10 5, LS12 = 0.99994

63

investigated.
of Failure"

50 is used

in tabular

LS21 = 1.0

of

with a circular

43 is used to find a "Probability

delamination

is

For the elliptical

to find a "Probability

2
Elliptical

1, and

the

1, and a limit load of Plim 2

at Location

of Safety"

1, Figure

of PF12 = 5x10

to Location

allowable

damage

for the

the damage

Disbond

follow

inspection

1
Circular

t22 = 1

at Location

probabilities

= Po22(a)

at Location

of Failure"

size

II is to be used

Po12(a)

to Location

translates

damage

PD22(a),

example:

will not fail at the given

= 6x10

"Probability

Type

The detection

For the circular

This is because

For the elliptical

Location

is applied

damage

delamination

2.

of Plim 1 = 6000

4, which

delamination

PF12

at Location

the "Probability

= lxl0

all of the detected

Inspection

in the previous

A design

determine

that

Po22(a),

form

as:

to find

of
a

CircularDelamination

Using

Equation
LS

For the
value

(3.2-19),

= LSll.

allowable

exceeds

design

has a higher

The "Level

accuracy
based

These

allowable

provides

detected

damage

sizes

structure

depending

distributions

can be determined
= 99.98%

the calculated
damage

as:

"Level

profile

of Safety"

analyzed,

the

new

it will replace.

of the actual

damage

on the inspection

in the formula.
and

0.9999995

of the structure

for the

an estimate

will vary

LS22

(1.0)(0.99994)(0.9999995)

Thus,

safety than the existing

estimates

7,

of LSa_ow = 99.9%,

by 0.02%.

method

= 5XI0

of Safety"

= (0.9999)

of Safety"

of the probability
on the

"Level

LSI2. LS2_. LS22

of Safety"

a structure.

the overall

"Level

the

PF22

detection

The

actual

probabilities

size distribution

method

utilized

damage
can

be

in

and the

size distribution
calculated

by

the

equation:

P(a) /_f P(a) da

-pD _/ao
p_
)

P(a)

The results
plotted

actual

of this calculation

in Figure

indicating

that

damage

significantly,
the

distribution

infinity

(5.3-2)

using

54 for the two

the

distributions

distributions
however,
of p(a)

as the damage

inspection
assumed

for both

for various

the distributions

in the

inspection

inspection

as discussed

types.

assumed
The

example
types.

types.

in Section

size a goes to zero.

64

curves

are very

problem
Results

Figure
3.8,

in the example
close

closely

using

to each

are
other,

approximate

actual

54 also illustrates

in that

problems

it asymptotically

data

may

the
vary

the behavior
approaches

of

6.1

The

Benefits

benefits

varied.

of an Equivalent

of an equivalent

By taking

of safety

safety

on

equivalent

a data-driven
with engineering

the design
where

process.

safety

operating
Safety"

have

to be made

between

Also

certification

Safety"

the

meaning

However,

that the "Level

than

the

instant

provides

damage

to fully

The existing
event

tolerance
into

system

comparison

structures

for

and over

into the

"Level

made

airlines

a range

of

in different

assumptions

tool

in the

value

does

structure

experience

65

currently
the

only describes

point

just calculated

Past

not

characterize

derivation

at a fixed

of damage

of calculation.

along

of

during

and risk over the life-cycle

a valuable

methodology

needed

of Safety"

data,

is

of the

and

flight

of aircraft

types.

Formulation

safety

inspection

the state

the

maintaining

or material

tools

the

program

an objective

statistical

risk in a fleet of aircraft,

issues

structure.

for validating

method

equivalent

tolerance

for a single

variable.

the

of Bayesian

frees

methodology

inspection

or similar

and

factors

with the damage

and allows

structures,

still

or "soft"

of structure

the level of uncertainty

to manage

of a particular

of Safety"

incorporation

of the Current

here,

of damage

dissimilar

type

This

of Safety"

of empirical,

to any

for arbitrary

while

for the service

characterized,

a mechanism

important,

Limitations

range

been

design,

associated

planning

are many

reduced.

The "Level

the uncertainty

is applicable

and for reducing

authorities

As presented

The

provides

process,

structure.

other

mechanisms

method

structures.

It incorporates

method

in the

design

the need

is substantially

for the inclusion

to quantify

to aircraft

failure,

efficiencies

allows

The

environments.

the design

6.2

judgement,

approach

practices

than existing

also

Approach

to structural

structural

of a structure.

damage

relative

that

of Safety"

design

to, or better

CONCLUSIONS

of Safety

approach

maximum

process

characteristics

"Level

traditional

to pursue

AND

Level

a probabilistic

based

designer

RESULTS

in time,

and

is continually
is invalid
with

encompass

operational

"Level

of

a structure's

"Level

is essentially

a state

changing

at any other

mechanical

the

with
point

systems

time,

in time

holds

that

structuresin an operationalenvironmentaccumulatedamageover a period of time, and

previouslyaccumulateddamagemay grow due to cyclical loading. However,the current
formulationdoesnot accountfor this. As damageis accumulated,the "Level of Safety" of
the structurewill decreaseover time, until the damageis detectedandrepairedat the next
scheduledinspectionevent. Undetecteddamagewill continueto grow until reachinga size
that canbe detectedat somefuture time. Therefore,it is also necessaryto incorporatethe
schedulingof multiple inspectionandrepair eventsinto the methodto periodicallyreturnthe
"Level of Safety"to a point aboveits critical designvalue.
Interactionsbetweendifferent damagetypes,andmultiple damageeventsarealsoneglected
in the currentformulation. The location of the damageplaysa strongrole in theseeffects,
andcanbe modeledpartially by assumingthe worst-caselocationfor damageaccumulation
in the analysisof the residualstrengthvs. damagesizefor the structure. However,this may
leadto overly-conservativeresidualstrengthresults. To alleviatethis, detaileddatamustbe
collectedon the locations of damagerelative to the zonesof stressconcentration,which
couldprovedifficult to obtain.
The behaviorof the detectionprobability anddetecteddamagesizedistributionsneedsto be
investigatedusing experimentaldatafor a rangeof structuralconfigurations. It is unknown
how difficult it will be for probability distributions based on actual data to meet the
conditions discussedin Section 3.8. The criteria that the probability density function
estimateof actualdamagemust be integrableover the entire domain of damagesizesmy
prove to be so difficult to satisfy for the current "Level of Safety" formulation that the
methodmayneedto bereformulatedto be of anypracticaluse.
Anotherunquantifiedvariablein the methodis the effect that finite data setshave on the
confidencelevel of the equivalentsafetyestimates.Generally,the more datapointsthat are
used,the higher the confidencelevel will be. With a finite numberof data and a required
confidencelevel, the valid "Level of Safety" can only be the lower limit of the one-side
confidenceinterval of the quantity. This may unnecessarilypenalizethe structurein the
designphase,whendetaileddamagedatais mostdifficult to comeby.

66

6.3

Topics

The

following

the method
broader

,,

for Further

topics

to some

range

are recommended
of the issues

of damage

Behavior

Research

mentioned

tolerance

of probability

for further

research

previously,

to investigate

and to extend

the

sensitivity

the method

of

to cover

issues:

distributions

derived

from

actual

damage

data

for a variety

of

structures.

,,

,,

Extension

of present

inspection

intervals.

Demonstration
fuselages,

of

method

method

empennage,

Investigation

of finite

,,

Incorporation

of damage

,,

Extension
loading,

,,

material

Characterization
composite

on

built-up

rates

and

of damage

monolithic

accumulation,

structures,

and multiple

including

wings,

etc.

,,

of method

to incorporate

data set effects

and confidence

interaction

and growth

to incorporate

probabilistic

properties,

manufacturing

of probability

level determination.

effects.

assessments

variation,

of detection

structures.

67

curves

of other

variables

such

as

etc.

for

service

inspection

methods

of

APPENDIX

68

Start

Probability
of detected

density
flaw size

for each flaw

each location

type in

Determination
residual

of design

of

strength

of damaged
structures
FEM

using

process

Probability
of total
number
of flaws for

Probability
detection

each
each

flaw

size for each

type

in each

flaw type
location

Evaluation
Safety

of Level

the definition

Level

of Safety

and

critical

damage

size

of

of the structure

using

updating

in

In design

of

process

location

Accepted

Level

of Safety

taken

from

Selecting
the local

structure
load

existing

airplanes

During

size for

operation

to satisfy

Update

the

inspection

required
Level of Safety
via critical
damage
size

maintenance

data
from

in-service

Manufacture

and operation

inspection

Figure

1. Flow-Chart

of Developing

69

of

flaw

of

damage
Bayesian
data

Damage
collected

of flaw
as a function

Equivalent

Safety

Aircraft

and
plan

t.==
t-

_l!sua(]

_l!l!qeqoJd

=.

_J
t-

=
.==

?0

tr_

c_
N

e_

C_

O
=

.
e_

CO

eiim

c_
e_

N
iim

CO

=
e_

e_

C_
c_
e_

O
L_3
C_
0

C_
0

L_3

_====

_====

d
_l!suap

L_3
0
0

O
e_

_l!l!qeq0Jcl

=
e_

7]

j=

_l

_ -_
AI

04f_

o_

o_

VI

O
_,_

04f_

_,_

_,_

jv_

>

o_

,ml

,ml

,ml

,ml

,ml

,ml

,ml

,ml

,ml

,ml

,ml

,ml

o_

72

\
\,

\
\

\
\

o_

o_

\
\

o_

o_
o_

o_

73

\\

\
I

\
\

\
\\\

\\

o_

o_
I
I

\
\

o_

o_

o_

o_

o_

74

<-_

t t

C)
c
..c
_0

.E_ w

HI

I
co

l_

J
e_

e_

"0
"0
e_

0
.Q
(n

e_

e_

cJ

o_

co

(u!/ql) _I!suaIu! peo-I

75

J
t_

=
J
t_

t_

=
tJ

=
=
=

,i

I
I

"0
c
0
.Q
u_
"0

,i

,i

0
u_
X
0

=
I

.
J

;=
J
t_

76

oJ

GO

e"

me

G)
s_
I

CO
V

"0
C
0
..Q
me

"0
0
_r_
L_

G)
G)

(u!/ql)/_l!suelu!

peol

77

J
.
J=

"T--

I
I

J=

J
J
J=
J

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

"1"-"

"1"-"

"1"-"

"1"-"

"1"-"

0
0
0

0
0
0

0
0
0

0
0
0

J
=

(u!/ql) l!sualu!

peo-I

78

=
=
0

=
A

O0

e|i

=
LO

LO

"0

CO

e-

0
..D
(/)

,=
i

q,_

"0
_-

E
m
|i

=
I

0
0
0
0
CO

0
0
LO
C_

0
0
0
C_

0
0
LO
_-

(u!/ql) A]!sue]u!

0
0
0
_-

0
0
LO

peo-I
=

79

=
=

e"
|m

0
I

11
G)
L_
11
"0
cO
tm

I
I

_=

|m

"0
0
|m

L_

X
11
L_

0
11

_J

=
I

=
0
0
0
0

0
0
_

0
0
0

(u!/ql)

0
0
_

_l!SUalU!

0
0
0

peo-I

0
0

_4
=

8O

=.

A
p-

e_

e_

=
I

e_

LO

e_

C
o
.E_ w
I- LL

{"
me
I

o..

(
co

"O
_)

;;_

e_
J

C
i

O
(-

e_

O
s_

N_

u
{00
u)

,4mJ

e_

e_

me

_J

_r_

L_
O
O
r_

0
(0

(!sd)

0
LO

uo!loaJ!p

0
_J-

0
O0

6u!peol

0
OJ

0
",--

u! ssaJ:lS

=
e_

8]

_L
p--

rJl

(D

_')
()

*_

(
Ix)

(
o

c-

e-

(
o

.E_ w
I- u_

Q.

tO

O
e-

@J
O
C

(
C
C

j
_

o
co

_
I

.m

T'--

T--"

T--"

T--"

(!sd)

uoBoaJ!p

fiu!peol

u! ssaJtS

82

o
C,l

._
I

e_

o
0
w-e_

00
o

o_
_

00_

r_,,,

LO

LO

LO

LO

LO

LO

.i

O0

/i

_--

In

"

o
e_

!.--

co

rI

"o

_,_

e_
e_

e_
e_

0,1

/f

C)
0
0
0
LO
03

0
0
0
0
03

0
0
0
LO
(kl

(u!/ql)

0
0
0
0
(kl

0
0
0
LO
_r-

_l!sualu!

0
0
0
0
_r-

0
0
0
LO

peo-I
=
e_

83

LO
C_

C_

os

o
e"

mm

I
I

L_

0
._

0
LO

X
s._

0
ee

0
0
0
LO

0
0
0
0
_

0
0
0
LO
CJ

cu!/ql)

0
0
0
0
CJ

0
0
0
LO
",--

Al!sualu!

0
0
0
0
",--

peo-I

84

0
0
0
LO

c_

*i

O0

O0

0
O_

LO

cO

eo=
G)

0
LO

*i

ii

LO

o=
I

*i

CO

"0
cO
.Q

*i

ii

*i

)
"0
*i
L__
*i

ca
*i

c_
ii

c_

c_

0
0
0
0

0
0
0

0
0
0

0
0
0

TII

,TII

,TII

,TII

0
0
0

0
0
0

0
0
0

0
0
0

r_

(u!/ql) _l!suelu!

peol
c_
*i

85

J
t_

_J
o_

0_

0_

0
O_

0
O_

LO

"_" 0
LO

LO

==

P._

.==

q,_

-==
o_

t_

y
0
0
0

0
0
0

J=
J
t_

0
0
0
0

0
0
0

0
0
0

(u!/ql) AI!sueIu!

0
0
0

0
0
0

peo-I

0
0
0

0
J

J
t_
o_

86

=
.
J=
=
=

C)
I-

C)

CJ

C)

c)

o_

c)

_=

t.O
!

t.O
!

t.O

l-

!
i

=
l_

_J
J=

"0

J
J

J=
J

!
=
i

0
0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
J

_J

(u!/ql) _]!sua]u!

peo-I

J
I-

87

.
o_

J=

L_
o_

o_

o_

o_

o_

00_

o_

o_
I
I

J=
o_

_._' _.1 q.)

,_
X

0
0
0
U3
C_

0
0
0
0
C_

0
0
0
U3
v--

0
0
0
0
v--

(u!/ql) _!sua_u!

0
0
0
U3

J
J

J=

o_
I

pe0-1
J

o_

88

e_

e_

e_
e_

e_

00

e_

.i
J

"0
e_

J
e_

J
J

C_J

,i

e_
i

0
0
0
I._
OJ

0
0
0
OJ

0
0
I._
",--

(u!/ql) _!sue_u!

0
0
0
",--

0
0
I._

0
J

peo-I
J

e_

89

o_

o_

o_

o_

U3

O0

O0

o_

0
03

(_

,_

o_

(
I

U3

G)
I

o_

o_
I
I

o_

j
J=
J

o_
I

0
0
0
0
_J"

0
0
U3
03

0
0
0
03

0
0
U3
C_

0
0
0
C_

(u!/ql) l!sualu!

0
0
U3
v-

0
0
0
v-

0
0
U3

0
J

peo-I
J

o_

9O

_,_

.i

c_

.i

c_

0
.i

IZ)

IZ)

IZ)

IZ)

"qF

A
11F

.i

_I,

11
i

<

.i

+. ,+;
.i

'_
i

1'i.
_11

0
0
0
L_3
03

0
0
0
0
03

0
0
0
L_3
C_

(u!/ql)

0
0
0
0
C_

0
0
0
L_3
v-

l!sualu!

.i

WF

O3

?./'

.i

0
0
0
0
v-

0
0
0
L_3

pe01
.i

9]

.i

(/)

OJ

tllm

i
,,,1

i
I
v

I
v

41

",-

I)
')

"ill

0
(-

E
()

41 I

_1
0
_m

,III

.,.im
0

/,

++

I,=

,,

I,=

o
0
0
0
LO
09

0
0
0
0
09

0
0
0
LO
C_

0
0
0
0
C_

0
0
0
LO
_--

0
0
0
0
_--

(u!/ql) _l!suolu!

0
0
0
LO

peo-I
I,=

91

_,

o_

o_

0
o_

00

00

03

tO
_-

tO
_"
tO

.c_
o_

0
tO
o_

o_

o_

0
o_

o_

o
0
0
0
CO

0
0
0
_-

0
0
0
C_

0
0
0
0

0
0
0
O0

(u!/ql) l!suelu!

0
0
0
CO

0
0
0
_-

0
0
0
C_

pe0-1
o_

93

o_

o_

LD
o_

00
0
0
O_
0

00

0
O_

e=

im

O=

uO

t_

"_" 0
uO
I

O=

uO

"C3

_')

CD
C

E
o_

///'
o_

0
0
0
0
(.0

0
0
0
_J"

0
0
0
Od

0
0
0
0

0
0
0
O0

(u!/ql) AI!sueIu!

0
0
0
(.0

0
0
0
_J"

0
0
0
Od

peo-I
o_

94

Ckl
0

II,!

0_
Ckl

o_ _-

o_

00_
t.O t.O

0
t.O

0
t.O

0
t.O

0
t.O

!
v

!
v

CO

r_

ITI

!
v

l)

CO

*_

L_

e_

r[I
/

._

i'r

|m

J_

C_

kr

*_

/
0
0
0
0
LO

0
0
0
0
_

0
0
0
LO
CO

0
0
0
0
CO

0
0
0
LO
Ckl

0
0
0
0
Ckl

(u!/ql) _l!sualu!

0
0
0
LO
.T--

peo-I

95

0
0
0
0
.T--

0
0
0
LO

J
J

J=
J
CJ
J

LO
J=

Z
i

O_

o_

.=_

ii

LO

j_

o_

_ o_=_
o._

U_

o_

LO

_'_

CO

_.[',,_

LO

e_

0
0
0
0
LO
_"

0
0
0
0
_"

0
0
0
LO
CO

0
0
0
0
CO

0
0
0
LO
CJ

0
0
0
0
CJ

(u!/ql) l!sualu!

0
0
0
LO
_-

0
0
0
0
_-

0
0
0
LO

peo-I
e.l

e_

96

0
L_

0_
L_

0
L_

L_

L_

L_

!
v

!
v

co
!

_
A

o_

_m

,.=E-

o_

am

'=

ca

o_

o_

I
o_

e_

0
0
0
0

0
0
0
LC)
C_

0
0
0
0
C_

0
0
0
LC)
04

0
0
0
0
04

(u!/ql) _l!suelu!

0
0
0
LC)
_

0
0
0
0

0
0
0
LC)

peo-I
(",1

e_

97

_=

e_

0
0
0
LC)

J=
cJ

Z
i
.
cJ
o_
J=

Od

&

(D
Cq
00d
0

J=

LO
0
0"_

T
J=

0
l_
J=

O)

rfI

im
V

i
_i

e_
i

o_

LO
r_

"
o
0

co

I
o_

o_

ul

i=. i

im

o_

_=;_
cJ
o== I_
=_t

03

LO

_i't
l_
I

0
o_

0
0
0
U3

0
0
0
0

0
0
0
U3

0
0
0
0

0
0
0
U3

0
0
0
0

0
0
0
U3

0
0
0
0

0
0
0
U3

(u!/ql) i(l!su01u!

peo-I
i
l_
o_

98

==

J=

O0

:=

ee_

t.O
e_

e_

_=

e_

E
_=

in

c_

=
e_

o
0
0
0
Lt3
O3

0
0
0
0
O3

0
0
0
_'3
C_

0
0
0
0
C_

0
0
0
_'3
v-

(u!/ql) _l!suelu!

0
0
0
0
v-

peo-I

0
0
0
Lt3

ee)

=
e_

99

_J
_J

_J
J

LI3

_J

O0
Oq

Z
00"J

LO

LO

LO

LO

LO

LO

e_

|i

e_

r-

.C
e_

_I'

"1

q)

0
I

l=

o-=

J
'=

l=

,,_
0

-_'_

0
|i

LI3

t_

ugh

=
e_

0
0
0
I.I3
0'3

,J

0
0
0
0
0'3

0
0
0
I.I3
0,,I

0
0
0
0
0
0,,I

0
0
0
I.I3
",--

(u!/ql) Al!sualu!

0
0
0
0
",--

0
0
0
I.I3

e4

peo-I
=
e_

100

o
0
oJ

C_
_"

0
oJ

_o _

_.=.

il
J

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

CO

CO

C_

C_

v-

v-

0
0
0

0
eo

(u!/ql) _l!sualu!

peo-I
J

101

J=

Z
J

.
o_

J=

L_

&
J=
J=
C_
J=
I

00J

LO

LO

LO

LO

LO

LO

o_

t-

"_
V

o_
I

o_

-_

o_

o_

o_

o_

0
0
0
0
LO
O0

0
0
0
0
O0

0
0
0
LO
(NI

0
0
0
0
(NI

0
0
0
LO
_--

(u!/ql) l!sualu!

0
0
0
0
_--

0
0
0
LO

0
J

pe0-1
J

o_

102

O0

m
m

O3
EL

O3
EL
=

tO

tO

C.3
O3
EL
Gr3
t-

C.3
O3
EL
Gr3
t-

c,.O

=
=

|m

,=

N
im

(/)
=

Cl
%

\
%

C)
C)

CJ

c_

(e) a d

c_

_l!l!qeqoJd

c_

c_

uo!loalaQ

103

o_

c=
A

c=

|m

LO

N
iim

0")

o_

E
a

o_

o_
I
o_

c=

C_

0
CO

Lr)

C_I

Lr)

._--

Lr)

Od

"-

{el od _|!l!qeqoJcl

a6euJeo pa|oa|aO

104

o_

o_

im

=
o_

N
LO

|m

>
o_

o_

t_

C_
m

t_
t_
CO

|m

o_
o_

>

Cd

|m

C_

>

0
=
L_

LO

0"_

LO

_leJes

jo leAe3

105

LO

r_

LO

o_

0
0
0
0
cO

00

00_

kO

kO

kO

kO

kO

kO

0
0
0
L._

0
0
0
0

I=
im

im
m
i

im

0
0
0
L._

J
f

I=
I=
m

"0

0
0
0
0

im

_I

I
i

0
_I

0
0
0
140

I
I

0
0

O_
O_
0

co
O_
0

_..
O_
0

cO
O_
0

140
O_
0

_
O_
0

cO
O_
0

0_I
O_
0
e_

_leJes jo leAe-I

106

.E

L_3
0
0

_-

"I"-"

L_3
03
(33

03
03
0

6
Kle_es

L_3
O0
(33

O0
03
0

L_3
_
(33

03
0

t_

jo leAel

t_

107

0
0
__
Or)

00 _

00J

o_

0
0
0"3

0
0
0
LO
Od

LO

LO

LO

LO

LO

LO

o_

!
V

!
V

o_

Od

o_
m
I

im

0
0
0
LO

C
o_

o_

C
m

"D

,q

0
0
0

[]

0
_1
.4-,

o_

"7:
im

,q

--I

11

I
I
I

0
0
0
LO

.4

o_
::::::::::::::::

Od
0

0"3

o_ o_ d
o d

_leles

CO

CO

::::::::::::::::::

0o 0o 0o 0o
d d d d

_0 leAe-I

108

o_

//

_K
/
0
C_
m

j,,./

_4
i

0
C_

i o
I

--I

mm

0
0
0
LC_

!
!
!
!

o_

,-

o_ o_

ddd

o_

o_

o_

ddd

o_

o_ o_

ddd

KleJes jo leAe-I

109

o_

o_

o_

0
I

III
0

o_

m
o_

c_
0

o
u)

ILl
0

v
Z

Z
m_

L_
0I
ILl
z

Or)
0
O_

04
0

00_

0
O_

.D

.D

iii

o.

n.

4_
I

0
m_

o_

m_

o
0

IJ.

I
o_

(3")
0
I

0
0
0
0

0
0
0
L_
_

0
0
0
0
_

0
0
0
L_
,r--

0
0
0
0
,r--

0
0
0
L_

III
0
o_

('u!/sql) _1!suelul
peo7 l!m!-I
o_

110

o
o_

u')

u')

I
V

O'3

=_

I
o_

III

II

III
0
0
0
0
0"3

0
0
0
_
OJ

0
0
0
0
OJ

0
0
0
_
",-

('u!/sql) _l!sualul

0
0
0
0
",-

0
0
0

0
o_

peo-I 1!_!-I
o_

lll

o_

e_

0
e_

<5
i

e_

am

<5 _

e_

o,,

e_

d
O0
m

O'3

0
0
0

..Q
"Q
0
_L

I
0

i
_

0
0
0
0
0

e_
i
e_

0
!

III
0
0
0
0
03

0
0
0
L.O
OJ

0
0
0
0
OJ

0
0
0
L.O
"_

('u_/sa_) _|!sua|ul

0
0
0
0
"_

0
0
0
L.O

0
e_

peo-I |!m!-I
=
e_

112

_._

-_

..]

0
0
0
0
CO

('u!/sql) _l!sualul

peo-I l!LU!-I

,,g

o_

113

O
O
O
O
0'3

O
O
O
L.O
O4

O
O
O
O
O4

.=_"
a= ,'--m

_K

41,

tf
mm

r"

t_
C

O
O
O
L.O

or}

O
O
O
O

_'.-4

O
,--I

r=
mm

03

,--I

14'3 14'3 14'3

0

O
O
O
L.O

14'3 14'3 14'3

!

03
03

00
03

I'_
03

_3
03

u')
03

_I"
03

03
03

04
03

_le=l.eSjo leAel

114

,,&

_,_

o_

0
0
0
CO
o_

0
0
0
0
cO

0
A
v

o_

I
V

0
0
0
L._

o_

e"
am

o_
I
I

e_

0
0
0
0

am
I

eo_

0
0
0
L._

e"
m
_

"0
0
.J

0
0
0
0

am

.J

!
0
0
0
u3

0
0
0

o_

03
0

leJes

03
0

03
0

jo leAel
o_

115

o_

0
0
0
U3
CO

O0

0
0
tO

03
tO

0
tO

tO

tO

o_

0
0
0
0
CO

o_

0
0
0
U3
C_

E
I

0
tO

_r _r

0
0
0
0
C_

,..I"
/
AW"
iF

e"
=m

o_

e_
m
V

=m

e"J

0
0
0
U3

...a
/
/

era

o_

0
.J

/
/

0
0
0
0

0
0
0
U3

=m

E
=m

.J

_E

0
C_
0

03
0

03
0

03
0

03
0

00
0

00
0

00
0

00
0

o_

Kle_es jo leAel
o_

116

ol

0
0
__

C_

_.

0
OJ

C_
_

C
OJ

03

LO

LO

LO

U_

U_

U_

ol

0
0
0
0
o9

__

ol

0
0

II

cJ

ol

ol
I
I

e_
I

0
0

"1

0
0,1

ol
i

ol

/
/

)
)

'_

_-

0
_1

ol

0
0

t_

,1

t_

iN

_-

) i

..1

0
0
0

)
)
)
)
--

0
ol

0
_

03
0

03
0

03
0

03
0

03
0

03
0

03
0

03
0

03
0

K|e_es _o leAel

t--

ol

117

r_

o_

o_

i
I

III
0
i

o_

,l

ill
m

c,O

x_

o_
I

III
0

ill
Ill

i..(")
0
I

_,

III

ii -i

m
Im

l_l

__

II
0
o_
im
m
im

0"_

i
o_

0
I

c_

1.1

III

o_

o_

...............................................................................................
.__.............................
I .._..I
.............................................
III
I

0
0
0
t.,O
CO

0
0
0
0
CO

0
0
0
t.,O
_

0
0
0
0
_

(lU!/Sql) _l!sualul

0
0
0
t.,O
_

0
0
0
0

0
0
0
t.,O

0
o_

peo-I l!lu!-I
o_

118

6
I

0
0

I
I

o
m

L.O
0

W
,--

_:_
O

I"0

mm

LI.

_+

II_,I

OJ
0

(,/'j

W
0

O
mm

,0

W
,-

O
,12.

o_
o_
I

0
0
0
LO
CO

0
0
0
0
CO

0
0
0
LO
C_

0
0
0
0
C_

('u!/sql) _l!sualul

0
0
0
LO
v-

0
0
0
0
v-

0
0
0
LO

peo-I l!tu!-I

119

_'_

- "i
ip

_,I

o
m

,_I,.

II 'I

II _,I

m+
0

,--_"

U'J

_I,
0

O3

II _'I

0
0

_+
0

!
V

_I, II
"lTr+

0
0
0
U3
03

I _'I

0
0

i
V

!
V

,lJ,

_=_

t,",J
_

= _.,
_E

_
"'_
m.._

"I

..........P?TP

..........rPir

......... ]rl]

...........I]PI

0
0
0
0
03

0
0
0
U3
CJ

0
0
0
0
CJ

0
0
0
U3
"_-

('u!/sql) _|!sua|ul

..........P?Tr

0
0
0
0
"_-

......... Tl?r

0
0
0
U3

peo-I :l.!m!-I

120

LU
_"

=
--_

e4

I
I

0
m

0
0
0

ii"

0b

_-,-

,0
0
0
0

I.I.

"_

LO

LO

LO

e_

03
0

e_
0
,12.

LU
I
v

I
v

I
v

'4+

IPI

0
I

LU
0
0
0
LO

0
0
0
0

0
0
0
LO

0
0
0
0

('u!/sql)/q!sualul

0
0
0
LO

0
0
0
0

peol

121

0
0
0
LO

l!tu!l

0
>,

0
E

==.

II _

_
0

am

o_

=.
00
m
m

C1)

C1)

>.,

>.,

(.C)

(.C)

O
C1)

O
C1)

c/)
(.-

c/)
(.-

.=

CO
C_

d
nm

C_

_J
o_

t_J

N
iim

C/)

_J
N

o_

03

Cq

C/3
_J

=
J
,_
C_
o_

_J

o_
I
o_

CO

I.(3

"_-

03

Cq

,v-

c;

c;

m)d _3Jl!qeqoJd

C_
t_

eOeuJeo lem,0V
_J
t_

=
o_

122

1.

Cruse, Thomas
York, 1997.

2.

Sundararajan,
1995.

3.

Ditlevsen,
O. and
Inc., 1996.

4.

Soares,
C. Guedes,
Publishers,
1997.

5.

6.

Chamis,

C.C.,

R.S.,

H.P.,

Damage
11.

Gary,

Methods

for

R.

Structures,

Northrop

R.S.,

Advanced

Prediction

Jean,

"How

P.M.,

and

Methodology

Federal

Aviation

12. Rummel,

Ward

Capabilities

Data

(NTIAC),

for

and

Book,
Texas

Kluwer

Academic

Testing

and

1997.

E.S.,

Certification

Testing

Aircraft

Division,

Methodology

for

1997.

Methodology

for

Impact

Damaged

1998.

Composite
Structures,
Recent
Progress
Proceedings
of 17 th ICAS Congress,

the Situation

of the No-Growth

Application

to

ICAF'

97 Composite

M.G.,

DOT/FAA/AR-95/17,

Composite

Administration,
D,

Wiley

1990.

to Address

Riskalla,

& Sons,

Certification

No. DOT/FAA/AR-97/79,

Approach?:

with Composites,"

John

Corporation

No. DOT/FAA/AR-96/111,

Report

Probabilistic

Saether,

New

& Hall,

Materials:

pp. 23-42,

Inc.,

Champman

Design,

Composite

and

Dekker,

Methods,

Structural

Design,"

Cordero,

Reliability

pp. 1439-1447,

Design

Center

Reliability

J., "Certification
of Large Airplane
Trends
in Compliance
Philosophy,"

Rouchon,
with

Report

Enhanced

Stockholm,
10.

H.P.,

Marcel

Handbook,

Structural

R. and Whitehead,

Structures,

Mechanics

Composite

for Composite
CA, October
1986.

Structures,

Rouchon,
and New

H.O.,

Design,

STP 1242, ed. by S. J. Hooper,

Kan,

H. P., Cordero,

Kan,

Structural

"Probabilistic

Whitehead,

Kan,

Mechanical

Probabilistic

13, ASTM

Composite
.

Madsen,

Design_, Vol.

Composite
8.

Reliability-based

C., Probabilistic

Methodology
Hawthorne,
7.

A.,

REFERENCES

Matzkanin,
Third
Research

Workshop,

Structures,

Washington,

Edition,
Institute

U.S.

A.,

123

Impact

17 June

1997.

of Probabilistic
of Transportation,

1997.

Nondestructive

Nondestructive
Austin,

Edinburgh,

Department

in Fatigue,

Accidental

Development

DC, August

George

Concept

Low-Velocity

Testing

Inc., Austin,

Evaluation
Information

TX, November

(NDE)
Analysis
1997.

13. Spencer,Floyd W., DOT/FAA/AR-96/65,

Benchmark

Inspections,

Administration,
14.

Spencer,

Rivet

Deparmaent
December
15. Ang,

DC, October

Using

H.S.,

Karlsson

17. Timoshenko,

Wilson

& Sorensen,

and

Jane,

K.C.,

Aviation

John Wiley

Reliability

Stability,

Concepts
& Sons,

20.

K.C.,

and

Yin,

W.L.,

"Refined

2nd Edition,

Buckling

and

Buckling

Dimensional
International

Delaminations,
II: Results
for
Journal of Solids and Structures,

Whitcomb,

J.D.,

Delamination,"

"Three-Dimensional

Journal

of Composite

and

Small

Washington,

DC,

in Engineering

Manual,

McGraw-Hill,

Postbuckling

Analysis

of Two-

Journal

of Solids

Analysis

of Two-

International

Postbuckling

24.

Awerbuch,

J.,
and

Composites,

Vol.4,

Tan,

Seng

Company,
25.

and

Predictions

Madhukar,

Anisotropic
Laminates
and Conclusion,"
Vol.29, No.5, pp. 611-639,
1992.

Analysis
Materials,

of

Vol.23,

Postbuckled

pp. 862-889,

pp.3-159,

C., Stress
Inc.,

Embedded

1988.

and Fracture
of Elliptical
Delaminations
Structures,
Vol.9, pp. 139-159,
1988.

p. 354-363,
M.S.,

Experiments:

5.7,

1961.

Dost, E.F., Ilcewicz,

L.B. and Gosse, J.H., "Sublaminate
Stability
Based
Impact-Damaged
Composite
Laminates,"
Proceedings.
of the 3 rd Technical

23.

Planning

Version

22.

for Composites,

Cracks
U.S.

Kassapoglou,
C., "Buckling,
Post-Buckling
Laminates
Under Compression,"
Composite

Society

on

Aviation

Techniques,

21.

American

Report

1975.
User's

Dimensional
Delaminations,
I: Analysis
and Validation,"
and Structures,
Vol.29, No.5, pp. 591-610,
1992.
19. Jane,

for

Inspection

ABAQUS/Standard

"Refined

Project
Federal

Administration,

H., Probability

Inc.,

S.P., Theory_ of Elastic

W.L.,

Detection
Nondestructive

Federal

Principles",

Research

Transportation,

1996.

Eddy-Current

and Tang,

Vol. 1: "Basic

Inspection
of

DOT/FAA/AR-97/73,

Heads

Alfredo

16. Hibbitt,
1997.

W.,

Visual

Department

of Transportation,
1998.

and Design,

18. Yin,

Washington,

Floyd

Beneath

U.S.

in

Modeling
of
Conference,

1988.

"Notched
Review,"

Strength
Journal

of
of

Composite

Reinforced

Laminates:
Plastics

and

1985.

Concentrations

in Laminated

Composites,

Technomic

Publishing

1994.

Mar,
J.W.
and
Lin,
Discontinuities,"
Journal

K.Y.,
"Fracture
of
of Composite
Materials,

124

Boron]Aluminum
Composites
Vol. 11, pp. 405-421,
1977.

with

26. Whitney, J.M. andNuismer,R.J., "StressFractureCriteria for LaminatedComposites

ContainingStressConcentrations,"Journal of Composite Materials, Vol. 8, pp. 253-265,
1974.
27.

Waddoups,
Mechanics
pp. 446-454,

28.

M.E.,

Eiswmann,

of Advanced

and

Kaminski,

Materials,"

Journal

B.E.,

"Macroscopic

of Composite

Fracture

Materials,

Vol.5,

1971.

Chang,
F.K.
and Lessard,
Containing
an Open
Hole
Composite

J.R.

Composite

Materials,

Vol.25,

L.B.,
"Damage
and Subjected
1991.

125

to

Tolerance
of
Compressive

Laminated
Loadings,"

Composites
Journal
of

REPORT

DOCUMENTATION

PAGE

Form Approved
OMB No. 0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data
sources, gathering and maintaining the data needed, and completing and reviewing the collection of information.
Send comments regarding this burden estimate or any other
aspect of this collection of information, including suggestions for reducing this burden, to Washington
Headquarters
Services, Directorate for Information Operations and
Reports, 1215 Jefferson
Washington, DC 20503.
1. AGENCY

USE

Davis Highway,

ONLY

(Leave

Suite 1204, Arlington,

blank)

VA 22202-4302,

2. REPORT

February
4. TITLE

AND

and to the Office of Management

DATE

and Budget, Paperwork

3. REPORT

2000

SUBTITLE

Structural

TYPE

Contractor

AND

Reduction

DATES

Methodology

Based on Concepts

(0704-0188),

COVERED

Report
5. FUNDING

Design

Project

NUMBERS

of Uncertainty
522-31-71-02

6. AUTHOR(S)
K. Y. Lin, Jiaji Du, and David Rusk

7. PERFORMING

ORGANIZATION

NAME(S)

Department
of Aeronautics
University of Washington
Box 352400
Seattle,

WA

AND

8. PERFORMING

ADDRESS(ES)

REPORT

and Astronautics

98195-2400

9. SPONSORING/MONITORING

AGENCY

NAME(S)

AND

10. SPONSORING/MONITORING

ADDRESS(ES)

AGENCY

National Aeronautics
and Space Administration
Langley Research Center
Hampton, VA 23681-2199

11.

SUPPLEMENTARY

REPORT

NUMBER

NASA/CR-2000-209847

NOTES

Langley Technical
Final Report
12a.

ORGANIZATION

NUMBER

Monitor:

DISTRIBUTION/AVAILABILITY

W. Jefferson

Stroud

STATEMENT

12b.

DISTRIBUTION

CODE

Unclassified-Unlimited
Subject Category
39
Distribution:
Availability:
NASA CASI (301) 621-0390
13. ABSTRACT

(Maximum

200

Standard

words)

In tiffs report, an approach to damage-tolerant

aircraft structural design is proposed based on the concept of an
equivalent "Level of Safety" that incorporates
past service experience in the design of new structures.
The
discrete "Level of Safety" for a single inspection event is defined as the compliment of the probability that a
single flaw size larger than the critical flaw size for residual strength of the structure exists, and that the flaw will
not be detected.
The cumulative "Level of Safety" for the entire structure is the product of the discrete "Level of
Safety" values for each flaw of each damage type present at each location in the structure.
Based on the definition of "Level of Safety", a design procedure was identified and demonstrated
on a composite
sandwich panel for various damage types, with results showing the sensitivity of the structural sizing parameters
to the relative safety of the design. The "Level of Safety" approach has broad potential application to damagetolerant aircraft structural design with uncertainty.

14. SUBJECT

TERMS

damage

tolerance,

reliability,
17. SECURITY
OF

damage
CLASSIFICATION

REPORT

Unclassified
NSN

7540-01-280-5500

15. NUMBER

probability,
detection,

Bayesian
composite

18. SECURITY
OF THIS

statistics,
sandwich,

CLASSIFICATION
PAGE

Unclassified

inspection,

aircraft structures
19. SECURITY

CLASSIFICATION

OF ABSTRACT

OF

PAGES

130

uncertainty

16. PRICE

CODE

A07
20.

LIMITATION
OF

ABSTRACT

Unclassified
Standard

Form

Prescribed
298-102

by ANSI

298 (Rev.
Std.

2-89)

Z-39-18