part ii- crack thick multilayer structure inspectionimtt-usa.com/publications/2008/part ii-...

Innovative Materials Testing Technologies, Inc.

1

2008 NDT Symposium Sponsored by Lockheed Martin NDT Working Group March 24-26, 2008

FG RFEC Technique for Thick Multilayer Aircraft Structures

Inspection

Part II Crack Detection

Yushi SunInnovative Materials Testing Technologies, Inc.

3141 W. Torreys Peak DriveSuperior, CO 80027Tel: 303 554 8000Fax: 303 554 8001

Email. [email protected]


2


Part II – Deeply hidden crack detection

Topic 1 ChallengesTopic 2 Raster scan using a sliding probeTopic 3 Rotational scan using a rotary probe

minimizing noise from fasteners• 0.200” thru crack under 0.5” Al layer, Fe fastener, edge effect

Topic 4 Automated rotary scanner ensuring constant speed for online signal processing and crack identification

• 0.133” corner crack under 0.25” Al layer, Ti fastener, edge effect• 0.150” corner crack under 0.25” Al layer, Fe fastener, edge effect

Topic 5 Study # 1 – edge effectTopic 6 Study # 2 – edge + ferromagnetic fasteners


3


Topic 1 - Challenges in Thick & Multilayer Structure Crack Detection

1. Requirement of deep penetration in multilayer structure2. Structure variations around a fastener – background noise

comparable or greater than crack signal3. Material property variations, such as permeability variation of

a steel fastener.4. Extremely weak crack signal submerged in background noise

and structure variation signals5. Signal magnitude is not necessarily be the indication of

existence of a crack. Other parameters, signal phase angle or signal shape must be used for crack identification

6. Signal processing or/and pattern recognition is needed for crack identification and quantification



4


Topic 2 - Raster Scan Using a Sliding Probe

A. Photos of FG RFEC Sliding Probes for Crack Detection

RF4 V3Footprint: 0.85” x 2.15”

Coil Center-to-Center Distance, CCD = 1.15”

RF2 V3Footprint: 0.3” x 0.62”

Coil Center-to-Center Distance, CCD = 0.3”



5


Sliding Probe Example 1Detecting 2nd Layer Notches in 0.25” + 0.25” Thick

B-52 Wing Spar Structure

181716

Y

Xx5x1

14 Scan Area

15 19

Edge Cutx0

29 mmX – Scan direction

19.7 mm

Excitation coil

Pickup coilFastener

15 – 2nd layer notches (0.100” & 0.245”)16-17 – 1st layer notch18-19 – 2nd layer notch



6


Detecting Crack in 0.25” + 0.25” Thick B-52 Wing Spar Structure

15: 2nd layer vertical notches (0.100” & 0.245”)

16-17: 1st layer

Other holes

f = 100 Hz, EFD = 19.7 mm



7


Limitation of Sliding Probe Raster Scan -Strong Noise from Fasteners

Signal from deep crack submerged by noise



8



Topic 3 - Rotational Scan Using A Rotary Probe Minimizing Noise from Fastener

Ball-bearing rotation guide

Slip ring Connector

Probe carriage

Composite layer

Titanium layer under inspection

Slip ringProbe carriage

Probe head with centering pin


9


Rotational Scan – Minimizing noise from fastener


-+ -+

Diff Sensors

Fastener- head Excitation Coil

Diff Sensors

Fastener- head Excitation Coil

Probe 2 – No signal unless there is a crack

FastenerSensor

Drive coil

FastenerSensor

Drive coil

Probe 1 – Constant signal unless there is a crack


10


Success Example 1 Detecting Cracks in Raised-Head Fastener holes

Specimen

A pocket closely

matches fastener

head

FG RFEC Probe

Additional Guide for

Probe Rotation Test

Specimen

A round probe head serves as a guide for probe rotation



11


0.100” rivet head overhang

Rotorcraft - Detection of Second Layer Cracks Topside

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.91

0 0.02 0.04 0.06 0.08 0.1 0.12

Flaw Length (in.)

Prob

abili

ty o

f Det

ectio

n

Phasec/Walking Probe Insp. #1 Phasec/Walking Probe Insp. # 2Phasec/Walking Probe Insp. # 3 Foerster-Rivet Check w/Custom ProbeRemote Field Eddy Current Concave Probe Insp. #1Concave Probe Insp. #2

Comparison of POD with other NDI TechniquesCarried out by FAA AANC, Sandia Labs



12


Success Example 2 Detecting Lower Layer Cracks in Boeing 723 Lap-Joint



13


New Challenge in Deep Crack DetectionWith Edge Effect and Steel Fastener

1. Signal magnitude is no longer an indication of existence of a crack

2. Other parameters, such as phase angle and/or shape of impedance plane must be used for crack identification

3. Concept of pattern recognition is needed



14


Example Detecting 0.200” 2nd Layer EDM Notch

in A Specimen Simulating0.500” + 0.140” C-130 Wing Structure

Specimen provided by Canadian Air Force

Notch

Steel fasteners

0.50

0” th

ick

1stla

yer

2nd

laye

r

Front View Bottom View Side View



15


Topic 4- Rotational Scan: Result of the Example Crack free

Observation:In deep crack detection signal magnitude is no longer the indication of existence of a crack, other parameters, such as phase angle and/or shape of impedance plane must be used for crack identification. Concept of pattern recognition is needed.



16


Step motor inside

Motor controller inside

Probe carriage

Probe head

Connectors

Probe carriage

Rotation guide with suction base

Ball bearing rotation guide with suction base

Probe head

Sealing rubber tube

Vacuum area

Topic 4 - Automated Rotary ScannerEnsuring Constant Speed for Online Signal

Processing & Crack Identification



17


Computerized SSEC InstrumentEnabling Online Signal Processing

InstrumentSSEC II-S

Rotary Scanner

LaptopSpecimen

under inspection



18


Topic 5 - Study # IEdge effect



19


Example 1 – Simulating Boeing 707 Wing Structure Detection 2nd layer vertical cracks

with Ti fasteners, crack very close to 2nd layer edgeSpecimen provided by Boeing Phantom Work

A AB BC CD D0B C D

Upper Row

1st layer - 0.250” Aluminum

2nd layer - 0.313” Aluminum

0.75

0”

2nd layer edge

Ti Fasteners



20


No EDM Notch

Increase of EDM Notch Size

0.200”X0.267” 0.220”X0.293” 0.250”X0.293” 0.300”X0.313”0.170”X0.227”0.150”X0.200”0.133”X0.173”0.100”X0.133”

Example 1 - Impedance plane & ellipse fitting2nd Layer Crack near 2nd Layer Edge Non-ferromagnetic Fastener

Observation: without signal processing reliably detected EDM notch size > 0.250”(L)



21


* These are post-processing data. Real-time processing is also working, but more verification is needed.

Example 1 - Signal Rotation Angle2nd Layer Crack near 2nd Layer Edge Non-ferromagnetic Fastener

0 5 1 0 1 5-2 0

0

2 0

4 0

6 0

8 0

f a s te n e rs

angl

e (d

egre

e) S N R a t C : 3 .5 3ED M : 0 .1 5 0 " ( W ) x 0 .2 0 0 " ( D )

R o w : T L

A -

0.10

0”(L

)X0.

133”

(D)

B -

0.13

3”(L

)X0.

173”

(D)

C -

0.15

0”(L

)X0.

200”

(D)

D -

0.17

0”(L

)X0.

227”

(D)

E -0

.200

”(L)

X0.2

67”(

D)

F -0

.220

”(L)

X0.2

93”(

D)

G -

0.25

0”(L

)X0.

313”

(D)

H -

0.30

0”(L

)X0.

313”

(D)

AB

–N

o ED

M N

otch

D0

–N

o ED

M N

otch

EF –

No

EDM

Not

ch

FG –

No

EDM

Not

ch

GH

–N

o ED

M N

otch

BC

–N

o ED

M N

otch

CD

–N

o ED

M N

otch

H0

–N

o ED

M N

otch

Observation: with signal processingreliably detected EDM notch size ≥ 0.133”(L)



22


Topic 5 - Study # 2

Edge + Ferromagnetic Fastener



23


Example 2 – Simulating Boeing 707 Wing Structure Detection 2nd layer vertical cracks

with Steel fasteners, crack very close to 2nd layer edge

A AB BC CD D0B C D

Upper Row

1st layer - 0.250” Aluminum

2nd layer - 0.313” Aluminum

0.75

0”

2nd layer edge

Steel Fasteners



24


Example 2Detection 2nd layer vertically aligned cracks

with steel fasteners & crack close to 2nd layer edge

Challenges:1. Unlike the case of Example 1 where the fasteners are made of

non-ferromagnetic material, in Example 3 the fastener is made of steel. In this particular case crack signal DECREASES with increase of notch size as shown in next page.

2. The underline physics of this phenomenon so far remains unknown.

3. A notch can not be identified by any single parameter of detected signal. A Shape Factor P, which is a combination of a few features of the shape of the impedance plane, is used for identification of a notch.



25


Example 2 - Signals from Fastener Holes w/ and w/o a NotchSignal decreases with increase of notch size

and varies in its size and shape


AB

P=0.66

BC

P=0.92

CD

P=0.57

EF

P=0.71

FG

P=0.58

H0

P=0.91

A

P=0.66

B

P=0.74

C

P=2.51

F

P=7.21

G

P=4.58

H

P=100

D0

P=1.0

D

P=3.04

E

P=6.11

No notch signals

Notch Signals

Increase of notch size


26


0.2” EDM notch

Shape Factor P=6.11

without notch

Shape Factor P=0.57

Example 2 - Signals from Fastener Holes w/ and w/o a NotchSignal decreases with increase of notch size

and varies in its size and shape



27


0123456789

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Notched fastenersGood fasteners

C 0.15”

D 0.17”

E 0.20”

F 0.22”

G 0.25”

H 0.30”

P

B 0.13”

A 0.10”

AB

0.00”

BC

0.00”

CD

0.00”

D0 0.00”

EF 0.00”

FG 0.00”

H0 0.00”

Example 2Shape Factor P is used for Notch Identification



28


Summary on Study #21. Signal obtained in detection 2nd layer vertically

oriented crack that is close to a 2nd layer edge has abnormal behavior. The signal decreases with the increase of 2nd layer crack size.

2. It is impossible to identify a crack using a conventional approach of signal processing and crack identification.

3. The concept of pattern recognition is utilized in this case.

4. Automated rotary RF_RFEC probe provides highly repeatable signal that enables the use of online signal processing and pattern recognition.


part ii- crack thick multilayer structure inspectionimtt-usa.com/publications/2008/part ii-...

Documents