faa william j. hughes technical center probability of detection studies to quantify flaw detection...

FAA William J. HughesTechnical Center

Probability of Detection Studies to

Quantify Flaw Detection inComposite Laminate Structures

Probability of Detection Studies to

Quantify Flaw Detection inComposite Laminate Structures

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company,for the United States Department of Energy’s National Nuclear Security Administration

under contract DE-AC04-94AL85000.

Dennis Roach & Tom RiceSandia National Labs

FAA Airworthiness Assurance Center

A340 HTP SkinA340 HTP Skin


ITG Team ParticipantsITG Team ParticipantsInspection Task Group Team ParticipantsInspection Task Group Team Participants

CACRC Inspection Task Group Members:

Wolfgang Bisle – AirbusChris Dragan – Polish Air Force Institute of TechnologyDon Duncan – US AirwaysJim Hofer – BoeingQuincy Howard – BoeingJeff Kollgaard – BoeingFrancois Landry – Bell HelicopterRobert Luiten – KLM AirlinesAlex Melton – Delta Air AirlinesEric Mitchell – American AirlinesStephen Neidigk – Sandia Labs AANCKeith Phillips – Airbus Tom Rice – Sandia Labs AANCDennis Roach – Sandia Labs AANC (Chair)Vilmar da Silva do Vale – EmbraerDennis von Seelen - Lufthansa Technik Darrell Thornton – UPSSam Tucker – United AirlinesRoy Wong – Bombardier

Rusty Jones, Larry Ilcewicz, Dave Galella, Paul Swindell – FAA


Composite Structures on Boeing 787 Aircraft

Carbon laminate

Carbon sandwich

Fiberglass

Aluminum

Aluminum/steel/titanium pylons

A380 Pressure Bulkhead

Composite Center Wing Box

Program Motivation - Extensive/increasing use of composites on commercial aircraft and increasing use of NDI to inspect them

Program Goals: Assess & Improve Flaw Detection Performance in Composite Aircraft Structure


One airline reports 8 composite damage events per aircraft (on avg.) with 87% from impact; cost = $200K/aircraft

Sources of Damage in Composite StructureSources of Damage in Composite Structure

Bird StrikeTowing Damage

Lightning Strike on

Thrust Reverser

Disbonding at skin-to-

honeycomb interface


Significant Internal Damage

Source: Carlos Bloom (Lufthansa) & S. Waite (EASA)

Inspection Challenge – Hidden Impact DamageInspection Challenge – Hidden Impact Damage

Backside fiber failure from ice impact

Visible Impact Damage – external skin fracture

Backside Damage – internal skin fracture & core crush

Extent of Visible Damage from Outside

Damage from ground vehicle


Cumulative POD Curves for the Woodpecker on All Panel Types

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.5 1 1.5 2 2.5 3

Flaw Size (Dia. in Inches)

Pro

bab

ility

of

Det

ecti

on

3 Ply Carbon 3 Ply Fiberglass 6 Ply Carbon 6 Ply Fiberglass 9 Ply Carbon 9 Ply Fiberglass

Experiment to Assess Flaw

Detection Performance

Composite Flaw Detection Experiment

Participation from over 25 airlines and maintenance depots

Industry-wide performance curves generated to quantify:

• How well current inspection techniques are able to reliably find flaws in composite honeycomb structure

• The degree of improvements possible through integrating more advanced NDI techniques and procedures.

Composite HoneycombFlaw Detection Experiment

Composite HoneycombFlaw Detection Experiment

Blind application of techniques to study hits, misses, false calls, and flaw sizing


Purpose• Determine in-service flaw detection capabilities: 1) conventional NDT

methods vs. 2) improvements through use of advanced NDT. • Optimize laminate inspection procedures.• Compare results from hand-held devices with results from scanning

systems (focus on A-scan vs. C-scan and human factors issues in large area coverage).

• Provide additional information on laminate inspections for the “Composite Repair NDT/NDI Handbook” (ARP 5089).

An Experiment to Assess Flaw Detection Performance in Composite Laminate Structures

An Experiment to Assess Flaw Detection Performance in Composite Laminate Structures

737 Composite Horiz. Stabilizer

A380 Fuselage Section 19


Approach

• Statistical design of flaws and other variables affecting NDI - range of types, sizes & depths of flaws

• Study factors influencing inspections including composite materials, flaw profiles, substructures, complex shapes, fasteners, secondary bonds, and environmental conditions

• POD and signal-to-noise data gathering

• NDI Ref. Stds. prepared to aid experiment

Flaw Detection in Solid Laminate CompositesFlaw Detection in Solid Laminate Composites

Expected Results - evaluate performance attributes 1) accuracy & sensitivity (hits, misses, false calls, sizing) 2) versatility, portability, complexity, inspection time (human factors) 3) produce guideline documents to improve inspections 4) introduce advanced NDI where warranted


Specimen Types – Solid laminate carbon (12 to 64 plies)• Contoured and tapered surfaces• Substructures – stringers, ribs, spars; honeycomb impediment• Bonded & sealed joints; fasteners• Large enough to warrant scanners; complex geometry to challenge

scanners• Carbon, uniaxial tape

Low Energy Impact

0

0

0

00

-45

-45

0

9090

+45

+45

Specimen Design - Flaw Detection in Solid Laminate Composites

Specimen Design - Flaw Detection in Solid Laminate Composites

Flaw Types - statistically relevant flaw distribution with range of sizes & depths (near front & back surfaces; in taper regions)

1) interply delaminations (“kissing” and air gap)2) substructure damage3) skin-to-stiffener disbonds4) simulated impact damage


Thick Laminate With Complex Taper - FabricationThick Laminate With Complex Taper - Fabrication

Flaw templates - ensure proper location of flaws

Flawsinserted


Thick Laminate With Complex Taper - FabricationThick Laminate With Complex Taper - Fabrication


Thick Laminate With Complex TaperThick Laminate With Complex Taper

20 PLIES

12 PLIES3.50"

4.75"

14.00".25" STEP PER PLY

STRINGER 2.00" X 2.00" X .125" THK

20 PLIES

12 PLIES

.25" STEP PER PLY

.50" STEPPER PLY

4.75"

38.00"8.25"

7.25"

.50" STEPPER PLY

19.00"

7.25"

7.25"

8.00"

2.00"

1.00"

COMPLEX TAPER NO. 1 FLAW LAYOUT10-14-04

DELAM 1 (4PL-PI) COMPLEX GEOMETRY

.50".75"1.00" .25"

DELAM 2 (GR-I)COMPLEX GEOMETRY

.50".75"1.00" .25"

1.50"

1.50"

.50".75"1.00" .25"1.50"

IMPACT (IM-T)TAPER

2.00"

2.00"

DELAM 1 (4PL-PI)FULL THK

DELAM 2 (GR-I)FULL THK

IMPACT (IM)FULL THK

.50".75"1.00" .25"1.50"2.00"

.50".75"1.00" .25"1.50"2.00"

.50".75"1.00" .25"1.50"2.00"

.50".75"1.00" .25"

DELAM 1 (FBH) SUBSTRUCTURE

.75"1.00"1.50"2.00"

DISBOND 1 (PT)SUBSTRUCTURE

1.00" .75" .50" .25"

DISBOND 2 (GR-I)SUBSTRUCTURE

A B C D E F

G H I J K L

M N O P Q R

M-2

P-1

R-2B-3

E-1

L-2

I-1

O-3

S T U V W X

Y Z AA BB CC DD

EE FF GG HH II

JJ KK LL MM

NN OOPP QQ

RR SS TT UU

FF-3

W-2

Y-2

BB-2

U-3

BB-3

Z-3

HH-3

X-2

T-2

GG-2

PP-2

TT-2

F-2

L-1

LL-3

38.00"

19.00"

Type I Specimen


0.75"

PI GR-I FBH

PI GR-I FBH

32 PLY

2.75" 10.50"

(B/W PLIES 16 & 17)

(B/W PLIES16 & 17)

(IN SUBSTR.) (B/W LAMINATE &SUBSTR.)

16.00"

1.50"

1.51"

1.50"

3.19"

ALL FLAWS NOT IN OR UNDER STRINGER ARE 1.00" IN DIAMETERAND FLAWS IN OR UNDER STRINGER (IN LAMINATE) ARE .75" IN DIAMETER

PI GR-I FBH PT(B/W PLIES 16 & 17)

(B/W PLIES16 & 17)

(IN SUBSTR.) (B/W LAMINATE &SUBSTR.)PI

GR-I

FBH

PI

GR-I

FBHPI GR-I FBH PT

1.375"

10-26-06

2.50" 2.75" 2.75"

STRINGER TYPE "A" (TYP)

STRINGER TYPE "B" (TYP)

= 1.00" DIAMETER FLAW

= 0.750" DIAMETER FLAW

STRUCTURE THICKNESS NO. OF PLIES

LAMINATE # 3 0.208" 32LAMINATE # 3 & STRINGER "A" 0.433" 67

0.283" 44LAMINATE # 3 & STRINGER "B"

TYP

0.75"TYP

STRINGER TYPE "A" = 0.225" THK (APPROX. 35 PLIES)

STRINGER TYPE "B" = 0.075" THK (APPROX. 12 PLIES)

1.25"

0.25"

2.50"

1.50"

2.50"

2.50"

3.25"

2.50"

2.50"

0.76"

0.60"

5.00"

4.50"

6.50"

13.25"

8.00"

1.00" 2.00" 2.00" 2.00"

1.25"

20 Ply Ref Std

32 Ply Ref Std

Reference Standards – Feedback PanelsReference Standards – Feedback Panels


Specimen Set - Flaw Detection in Solid Laminate Composites


Thickness Range:12 – 64 plies

Inspection Area:46 ft.2

Number of Flaws:202




Thickness Range:12 – 64 plies

Simple Tapers

Complex tapers

Substructure Flaws

Curved Surfaces

Array of flaw types

NDI Ref. Stds.


Solid Laminate Experiment Participants Solid Laminate Experiment Participants

http://www.movimed.com/index.html

http://www.olympus-ims.com/


Solid Laminate Flaw Detection Experiment Implementation

Solid Laminate Flaw Detection Experiment Implementation

PODs calculated for overall laminate, by thickness family, by substructure effects, by complex geometry effects, by flaw types, etc. 57 inspectors


POD Curves for 12-20 Ply Solid Laminate Family


False Calls: Constant thickness = 0.4/inspectorComplex Geometry = 4.0/inspector

34 ft.2 inspection area

Individual and Cumulative Comparisons

Flaw Size (Diameter in Inches)

Pro

bab

ilit

y o

f D

etec

tio

n

Overall: POD[90/95] = 1.29” dia.

Constant Thickness(12, 20, 28 plies): POD[90/95] = 0.86” dia.

Complex Geometry(tapered, curved, substructure, fasteners, honeycomb): POD[90/95] = 1.49” dia.


POD Improvements from Use of Methods to Ensure Proper Coverage

POD Improvements from Use of Methods to Ensure Proper Coverage


Pro

bab

ilit

y o

f D

etec

tio

n




False Calls: Constant thickness = 0.8/inspectorComplex Geometry = 0.3/inspector

12 ft.2 inspection area


Constant Thickness(32 plies): POD[90/95] = 0.74” dia.


Substructure: POD[90/95] = 1.50” dia.

Individual and Cumulative Comparisons


Pro

bab

ilit

y o

f D

etec

tio

n


Desire to Transition Inspectors from “Average” to “Good” to “Outstanding”Desire to Transition Inspectors from

“Average” to “Good” to “Outstanding”


Pro

bab

ilit

y o

f D

etec

tio

n


Overall Performance of Pulse-Echo UT forFlaw Detection in Composite Laminates

Overall Performance of Pulse-Echo UT forFlaw Detection in Composite Laminates


Pro

bab

ilit

y o

f D

etec

tio

n


Constant Thickness(32 plies): POD[90/95] = 0.80” dia.



Ramp Damage Check Experiment (RDCE) – “Spin Off” Subset of Solid Laminate Experiment

Ramp Damage Check Experiment (RDCE) – “Spin Off” Subset of Solid Laminate Experiment

Purpose: to assess new, ultrasonic-based “Go”/“No-Go” equipment that OEM’s plan to deploy at airports and other

non-scheduled maintenance depotsusing non-NDI personnel (e.g. A&Ps).

Usage Scenario: the equipment will be deployed whenever visual clues or other events occur which warrant closer scrutiny

of a composite laminate structure; ground personnel, with appropriate training on such equipment, will set up the

equipment in accordance with OEM-supplied procedures, and then make an assessment of the region in question.

Limitations: such “Go”/“No-Go” UT equipment is intended to be used to assess local indications or regions only, not

for wide-area inspections; equipment operators are directed to very distinct locations.


Ramp Damage Check Experiment – Ultrasonic Devices with “Go”/“No-Go” Capabilities

Ramp Damage Check Experiment – Ultrasonic Devices with “Go”/“No-Go” Capabilities

Olympus –“Ramp Damage Checker”

General Electric –“Bondtracer”


Ramp Damage Check Experiment – POD Curves for Composite LaminatesRamp Damage Check Experiment –

POD Curves for Composite Laminates


Pro

bab

ilit

y o

f D

etec

tio

n

• Overall: POD[90/95] = 0.78” dia.

• Similar results for GE Bondtracer and Olympus Ramp Damage Checker

• Similar performance from both inspectors and A&P mechanics

• Less spray in individual performances

False Calls = 0.6/inspector14 ft.2 inspection area


Thermography

MAUSSystem

SAM System

Shearography

Wide Area and C-Scan Inspection MethodsWide Area and C-Scan Inspection Methods

PE Phased Array UT UT Wheel Array

UltraImage Scanner


Solid Laminate Experiment – Advanced NDI Testing Evaluations

Solid Laminate Experiment – Advanced NDI Testing Evaluations

• Pulsed Thermography – Thermal Wave Imaging• Phased Array UT – Olympus Omniscan • Phased Array UT – Toshiba Matrix Eye • Phased Array UT – Sonatest RapidScan• Phased Array UT – Boeing MAUS V• Phased Array UT – GE RotoArray & Phasor• Microwave – Evisive• Digital Acoustic Video – Imperium Acoustocam• Ultrasonic Video - DolphiCam• Line Infrared – Mistras• Shearography – LTI• Shearography – Dantec• Backscatter X-ray – Scannex• Acousto Ultrasonics – Physical Acoustic Corp. T-SCOUT• Locked-In Infrared – MovieTherm• Laser UT – iPhoton


PA-U

T, Toshiba M

atrixEye

PA-U

T, Olympus O

mniScan

Laser U

T, iPhoton

PA-U

T - NDTS M

AUS V

DAV, I

mperium A

coustoCAM

PE-U

T, Industry

Flash IR

- TWI

Shearo

graphy - Dantec

Line IR

- Mistra

s

Lock-In

IR, M

oviTHERM

POD[90/95] 0.69 0.72 0.85 0.88 1.12 1.13 2.3 >3 >3 >3False Calls 1 0 0 0 0 5.5 0 1 NA 26

Sample of POD Results for Composite Flaw Detection Performance of Advanced NDI

Sample of POD Results for Composite Flaw Detection Performance of Advanced NDI

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Pro

bab

ility

of

Det

ecti

on


Individual & Cumulative POD[90] Curve Comparison for the Combined 12-20 and 20-32 Ply Specimen Sets for All Flaws in the Constant Thickness & Complex Geometry Regions

9 - Advanced NDI Methods

PA-UT, Toshiba MatrixEye

PA-UT, Olympus OmniScan

Laser UT, iPhoton

PA-UT, NDTS MAUS V

DAV, Imperium AcoustoCAM

PE-UT, Industry Cumulative

Flash IR, TWI

Shearography, Dantec

Line IR, Mistras

Lock-In IR, MoviTHERM


Pro

bab

ilit

y o

f D

etec

tio

n


Survey of Industry Composite NDI TrainingSurvey of Industry Composite NDI Training

Only 25% of responders currently have special composite NDI training in

place


Question 5 – Do inspectors also receive general composite training to

understand composite materials, plies, lay-ups, scarfed repairs, composite

design, composite processing, etc.?

Question 15 - If experience level is a factor in determining qualification to perform certain

inspections, do you use some sort of apprentice program to expose newer inspectors to such

inspections?


• Results can be used by OEMs and airlines to: 1) define detectability for various flaws/damage, 2) guide NDI deployment & training; used by FAA to produce guidance documents.

• Flaw Detection - Conventional, manually-deployed PE-UT: POD[90/95] = 1.12” dia.; skin flaw detection is higher, flaw detection in substructure is more challenging (POD[90/95] = 1.34” dia.)

• False Call rates were extremely low: 1 false call per 17 ft.2 (flaws ≥ 0.25 in.2)

• Optimum inspection rates = 2 ft.2 /hour• NDI Performance Obstacles – attenuation, complex signal reflections,

confounding presence of signal harmonics, rapid variations produced by changing/complex geometry, optimum deployment and difficulty with inspecting large areas

• Controlled and Proper Use of Ramp Damage Check “Go”/”No-Go” equipment can produce good performance (POD[90/95] = 0.77” dia.) Caution – proper calibration area must be used; requires good

schematics (available?) Cannot be used in taper or changing geometry regions

Conclusions – Inspection of Solid Laminate StructuresConclusions – Inspection of Solid Laminate Structures


Recommendations – How to move inspections from “average” to “good” to “outstanding”

Recommendations – How to move inspections from “average” to “good” to “outstanding”

• Increased exposure to representative composite inspections – common industry NDI Feedback Specimens

• Increased, focused composite NDI training (ARP covering 1st two bullets)

• Use of NDI and composite shop apprenticeships (OJT, awareness training, formal/uniform use of this tool)

• Enhanced NDI procedures – deployment, signal interpretation, clearer schematics showing structural configuration

• Use of inspection coverage aids should be required

• Divide large area inspections into a number of smaller regions

• Add audible alarms & detection lights to probe in “Go”/”No-Go” devices

• Prepare additional industry guidance to address training, use of NDI Reference and Proficiency specimens, procedures, composite construction awareness – produced by joint effort of OEMs, Airlines, FAA & industry groups


2013 Airlines for America NDT Forum

POD Studies to Quantify Flaw Detection in Composite Laminate Structures

Dennis RoachTom Rice

FAA Airworthiness Assurance CenterSandia National Laboratories

Composites have many advantages for use as aircraft structural materials including their high specific strength and stiffness, resistance to damage by fatigue loading and resistance to corrosion. The aircraft industry continues to increase its use of composite materials, most noteworthy in the arena of principle structural elements. This expanded use, coupled with difficulties associated with damage tolerance analysis of composites, has placed greater emphasis on the application of accurate nondestructive inspection (NDI) methods. Traditionally, a few ultrasonic-based inspection methods have been used to inspect solid laminate structures. Recent developments in more advanced NDI techniques have produced a number of new inspection options. Many of these methods can be categorized as wide area techniques that produce two-dimensional flaw maps of the structure. An experiment was developed to assess the ability of both conventional and advanced NDI techniques to detect voids, disbonds, delaminations, and impact damage in adhesively bonded composite aircraft structures. A series of solid laminate, carbon composite specimens with statistically relevant flaw profiles were inspected using conventional, hand-held pulse echo UT and resonance, as well as, new NDI methods that have recently been introduced to improve sensitivity and repeatability of inspections. The primary factors affecting flaw detection in laminates were included in this study: material type, flaw profiles, presence of complex geometries like taper and substructure elements, presence of fasteners, secondarily bonded joints, and environmental conditions. This experiment utilized airline personnel to study Probability of Detection (POD) in the field and to formulate improvements to existing inspection techniques. In addition, advanced NDI methods for laminate inspections – such as thermography, shearography, scanning pulse-echo UT, ultrasonic spectroscopy, laminography, microwave and phased array UT – were applied to quantify the improvements achievable through the use of more sophisticated NDI. This paper presents the experiment design used to evaluate applicable inspection techniques and the key results from the POD testing conducted at multiple airline facilities.

faa william j. hughes technical center probability of detection studies to quantify flaw detection...

Documents

faa william

experiment flaw detection

flaw sizing slide

composite damage events

flaw profiles

composite materials

composite horiz

simulated impact damage