Op#mizing)Pulsed)Eddy)Current …3)Orion)Sample)Descrip#on) NDT in Canada 2015 Conference 12 3.3 NAVAIR Sample Description This sample was acquired for testing purposes from the NAVAIR

Op#mizing Pulsed Eddy Current Probes for Inspec#on of CP-‐140

Aurora Lap-‐Joints

•  D.M. BuF, P.R. Underhill and T.W. Krause •  Royal Military College of Canada •  Kingston, ON •  [email protected]

NDT in Canada 2015 Conference

NDT in Canada 2015 Conference, June 15-17, 2015, Edmonton, AB (Canada) - www.ndt.net/app.NDTCanada2015

Agenda

• Background • Probe Design and Alignment • Sample Descriptions • Analysis and Results • Discussion • Conclusions


Background


Figure 1: CP-‐140 Aurora AircraS.

http://www.airforce-technology.com/projects/cp-140-aurora-maritime-surveillance-aircraft/cp-140-aurora-maritime-surveillance-aircraft5.html

Probe Design


Figure 2: Image of 5 mm (grey), 6 mm (blue) and 8 mm (red) ferrite core diameter probe configura#ons.

5 mm Ferrite Core

6 mm Ferrite Core

8 mm Ferrite Core

Probe Design


Figure 3: Detailed probe specifica#ons.

Parameter Probe 1 Probe 2 Probe 3 Probe 4

Driving Coil Inner Diameter

5 mm 6 mm 8 mm 8 mm

Average Differen#al Pair Spacing

12.2 mm 13.7 mm 16 mm 14.7 mm

# of Turns (Pick-‐up Coils)

350 400 400 400

Driving Coil Resistance

20.7Ω 14.0Ω 18.1Ω 15.1Ω

Probe Alignment


[5]

Figure 4: Probe with alignment guide and sample [1].

D ≈ 7 mm

Probe Alignment


Pick-up Coil

Pick-up Core

Bottom Layer

Crack

Top Layer

Ferrous Fastener

Pick-up Coil

Driving Coil

Driver Core

Image Courtesy of V. K. Babbar

Figure 5: COMSOL model of currents encircling ferrous fastener.

P-‐3 Orion Sample Descrip#on


12

3.3 NAVAIR Sample Description This sample was acquired for testing purposes from the NAVAIR depot in Jacksonville,

Florida and is shown in Figure 7 and Figure 8. The sample is based on the structure of

the Lockheed P-3 Orion, which has the same airframe as the CP-140 Aurora used by the

Royal Canadian Air Force (RCAF). The overall dimensions of the sample are 54 cm x

28 cm x 3 cm with two sections of aluminum plate 2.8 mm thick. The plates are joined

together by a row of ferrous fasteners in a lap-joint where the two plates partially overlap

as shown in Figure 7 and Figure 8. Each fastener has a length of 15 mm with a head

diameter of 7.0 mm and a shaft diameter of 4.5 mm [12]. The electrical conductivity of

the fasteners is 3.57 x 106 S/m with a relative magnetic permeability of 66 [3].

Figure 7: NAVAIR sample with notches, view from top

Figure 8: NAVAIR sample with view of fastener

Ferrous Fastener

Lap-Joint Edge

12

3.3 NAVAIR Sample Description This sample was acquired for testing purposes from the NAVAIR depot in Jacksonville,

Florida and is shown in Figure 7 and Figure 8. The sample is based on the structure of

the Lockheed P-3 Orion, which has the same airframe as the CP-140 Aurora used by the

Royal Canadian Air Force (RCAF). The overall dimensions of the sample are 54 cm x

28 cm x 3 cm with two sections of aluminum plate 2.8 mm thick. The plates are joined

together by a row of ferrous fasteners in a lap-joint where the two plates partially overlap

as shown in Figure 7 and Figure 8. Each fastener has a length of 15 mm with a head

diameter of 7.0 mm and a shaft diameter of 4.5 mm [12]. The electrical conductivity of

the fasteners is 3.57 x 106 S/m with a relative magnetic permeability of 66 [3].

Figure 7: NAVAIR sample with notches, view from top

Figure 8: NAVAIR sample with view of fastener

Ferrous Fastener

Lap-Joint Edge

Figure 6: P-‐3 Orion sample with notches, view from top and side.

P-‐3 Orion Sample Descrip#on


NAVAIR Sample Crack Summary Fastener Size mm (in) Orienta#on

1 3.30 (0.130) 135°

2 1.90 (0.075) 270°

3 1.78 (0.070) 225°

4 0.89 (0.035) 90°

5 5.46 (0.210) 45°

6 3.30/0.89 (0.130/0.035) 90°/315°

7 2.03 (0.080) 270°

9 1.52 (0.060) 225°

10 2.79 (0.110) 135°

11 0.89 (0.035) 90°

12 5.46/3.30 (0.215/0.130) 45°/225°

13 1.52 (0.060) 270°

14 5.08 (0.200) 315°

19 2.79 (0.110) 225°

21 0.89/2.79 (0.035/0.110) 45°

Figure 7: P-‐3 Orion sample notch characteris#cs.

37

Figure 19: NAVAIR sample with notches, view from top.

Figure 20: NAVAIR sample with view of fastener.

The NAVAIR sample has electric discharge machined (EDM) notches located in 15 of the fastener

bore holes, and eight with no notches (8, 15, 16, 17, 18, 20, 22 and 23). Each notch is cut at a 45°

angle to the edge of the bore hole, giving the notch a 1:1 aspect ratio. The notch orientations are

based on the schematic shown in Figure 21. This sample contains fasteners with notches in the top

layer, bottom layer, and three that contain top and bottom layer notches. The notch lengths and

orientations are located in Table 4, where NN stands for no notch.

Figure 21: View from top. Notch orientation diagram.

A 135°

B 90°

E 270°

F 315°

C 45°

D 225°

0° 180°

Top Edge of Sample

CP-‐140 Sample Descrip#on


Figure 8: Photo of sec#on of CP-‐140-‐TT-‐1B test sample.

Differen#al Pick-‐up Coil Response


Figure 9: Raw PEC differen#al pick-‐up coil response showing the front and back end transient response and signal gate.

Signal Gate

Time (ms x 10-‐2)

Amplitu

de (V

)

Back End

Front End

Square Pulse Input

Transient Response

Faraday’s Law

Principal Components Analysis


0

1

2

0 0.05 0.10 0.15 0.20

Experimental DataVector 1Vectors 1, 2 and 3

file: e:\research\nsercdata.epfile: e:\research\nsercdata.ep

Time [ms]

Volta

ge [V

]

-0.2

-0.1

0

0.1

0.2

0 0.05 0.10 0.15 0.20file: e:\research\nsercvectors.epfile: e:\research\nsercvectors.ep

S3

S1

S2

Time [ms]

Eige

nvec

tors

[V]

Figure 10: PEC original signal with first eigenvector reproduc#on and eigenvectors 1-‐3 reproduc#on. Insert shows the three eigenvectors used for the reproduc#on.

V1

V2

V3

Principal Components Analysis


Figure 11: 3-‐D view of P-‐3 Orion data PCA scores S2, 3 & 4.

68

Figure 51: 3-D view of NAVAIR data PCA scores S2, 3 & 4.

In order to better distinguish between notches and blanks, the Mahalanobis Distance is calculated

for each fastener. It is apparent from Figure 51 that PCA scores S2, S3 and S4 could be used to

calculate the MD, as they appear to provide good separation for some fasteners with notches.

However, five scores (S1-S5) are used to calculate the MD as some information is also located in S1

and S5 as discussed earlier in Section 5.8.2. Using scores S1-S5 increases crack detection and lowers

false call rates. The MD is then compared to the threshold, calculated using equation 2.50. The

threshold calculated for this experiment was 4.2 at 99% confidence and 3.5 for 95% confidence.

There are no units associated with the MD, as it is a relative measure of distance from the centroid

of the cluster of blanks, in terms of standard deviations. The MD was plotted for each fastener and

the plot is shown in Figure 52, along with the decision thresholds, while results of this test are

shown in Table 13. It is worth noting that the MD associated with fasteners 5 and 14 corresponds to

the biggest second layer notches, at 5.46 mm and 5.08 mm, respectively. As the notches decrease in

size, so does the MD as indicated by MD values for fasteners 4 and 11. This indicates that there is a

correlation of MD with notch size, which will be investigated further in Section 5.8.5.

-0.8

-0.6

-0.4

-0.2

0

0.2

S2

-0.04-0.020

0.020.04

S3

-0.01

0.01

0.03

S4

No-notch

0.89 mm notch

MD

Differen#al Pick-‐up Coil Spacing


1 2 3 4 5 6 7 9 10 11 12 13 14 19 21

Probe 3 (8 mm) 96% 88% 56% 52% 100% 64% 60% 36% 40% 56% 100% 84% 100% 32% 40%

Probe 4 (8 mm) 100% 100% 100% 84% 100% 100% 76% 100% 100% 84% 100% 100% 100% 100% 100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100% % Detec=o

n

Fastener #

Detec=on Results 8 mm Probes

Figure 12: Side-‐by-‐side comparison of notch detec#on results for Probes 3 & 4.

Drive Coil Core Diameter


Figure 13: Side-‐by-‐side comparison of crack detec#on results for Probes 1, 2 & 4.

1 2 3 4 5 6 7 9 10 11 12 13 14 19 21

5 mm Ferrite Core 100% 100% 92% 76% 100% 96% 88% 44% 100% 60% 100% 60% 100% 100% 96%

6 mm Ferrite Core 100% 100% 100% 72% 100% 100% 68% 84% 100% 84% 100% 80% 100% 100% 100%

8 mm Ferrite Core 100% 100% 100% 84% 100% 100% 76% 100% 100% 84% 100% 100% 100% 100% 100%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

100% % Detec=o

n

Fastener #

Combined Detec=on Results for 5mm, 6mm and 8 mm Probes

DAQ System Noise Reduc#on


Figure 14: Comparison of eigenvectors resul#ng from measurements taken with Darlington Pair and OP Amp circuit boards.

-‐7.00E-‐01

-‐6.00E-‐01

-‐5.00E-‐01

-‐4.00E-‐01

-‐3.00E-‐01

-‐2.00E-‐01

-‐1.00E-‐01

0.00E+00

1.00E-‐01

2.00E-‐01

3.00E-‐01

1 31 61 91

Scaled

Amplitu

de

(arbitrary un

its)

Time (ms x 10-‐2)

Eigenvectors 1-‐5 (OP Amp)

Vec 1

Vec 2

Vec 3

Vec 4

Vec 5

-‐4.00E-‐01

-‐3.00E-‐01

-‐2.00E-‐01

-‐1.00E-‐01

0.00E+00

1.00E-‐01

2.00E-‐01

3.00E-‐01

4.00E-‐01

5.00E-‐01

1 31 61 91

Scaled

Amplitu

de

(arbitrary un

its)

Time (ms x 10-‐2)

Eigenvectors 1-‐5 (Darlington Pair)

Vec 1

Vec 2

Vec 3

Vec 4

Vec 5

Improved Detec#on Results


0

2

4

6

8

10

12

14

16

18

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Mah

alan

obis Distance

(arbitrary un

its)

Fastener #

MD 6mm Probe (OP Amp Board)

Figure 15: Mahalanobis Distance versus fastener # for measurements taken with improved circuit board.

MD Threshold

Improved Detec#on Results


0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

100%

Drive Coil Diameter

% Detec=o

n Detec=on Results for CP-‐140 Sample

5 mm

6 mm

8 mm

Figure 16: Overall notch detec#on results for CP-‐140 sample with 7.5 mm fastener head diameter.

Discussion of Improved Results


Figure 17: Finite element model showing flux penetra#on through a ferrous fastener ac#ng as a flux conduit [7].

8

Figure 3: Front view of a solved 3D finite element half-model showing flux penetration through the ferrous fastener (CDDP Probe design) [25].

Recent work performed by Horan [26] addressed the issue of PEC inspection for stress corrosion

cracking (SCC), through carbon fiber reinforced polymer (CFRP). The conductivity for CFRP is

essentially zero. Cracks emanated from around ferrous fasteners, and travelled span wise, fastener to

fastener, in the inner wing spar of the CF-188. The inner wing spar consists of a layer of 7.5 - 20

mm (0.3 - 0.8 inch) thick CFRP, where the spar is attached underneath, using ferrous and non-

ferrous fasteners. SCC in the inner wing spar occurs between fasteners, which are located

approximately 25 mm apart. Horan [26] [27] successfully detected cracks at large lift-off using coil-

based probes, which utilized a central driving coil wound around a ferrite core, and Giant Magneto-

resistive (GMR) sensors. GMR and coil sensor signals were analyzed using PCA.

1.3.3 PEC Signal Analysis

The response signals in PEC techniques provide information about the presence of potential

defects. Three commonly used methods of characterizing signals are shown below in Figure 4: time-

to-peak, peak amplitude and zero-crossing time [28]. Time-to-peak amplitude has provided defect

depth information in multi-layered structures [12]. He et al. [18] conducted experiments using

differential coils and Hall probes, showing that the peak amplitude of the response signal yields

information about the defect volume and that the zero crossing time yields information about the

Ferrite Core

Driving Coil

Ferrous Fastener

Pick-Up Coil

Conducting Plates

Crack

•  More flux induces larger currents in the surrounding structure

•  New electrical system produces a higher signal-‐to-‐noise ra#o

Conclusions

•  Notch detec#on improved as pick-‐up coil spacing was reduced

•  Notch detec#on improved with increasing drive coil diameter

•  100% of notches in the P-‐3 Orion sample could be detected with improved electronics

•  Improved detec#on results were further validated using CP-‐140 sample


References [1] C. A. StoF, P. R. Underhill and T. W. Krause, “Pulsed Eddy Current Detec#on of Cracks in Mul#layer Aluminum Lap Joints,” IEEE Sensors Journal, vol. 15, pp. 956-‐962, 2015. [2] D. R. Desjardins, G. Vallieres, P. P. Whalen and T. W. Krause, “Advances in Transient (Pulsed) Eddy Current for Inspec#on of Mul#-‐Layered Aluminum Structures in the Presence of Ferrous Fasteners,” Review of Progress in Quan#ta#ve Nondestruc#ve Evalua#on, vol. 31, American Ins#tute of Physics, 2012. [3] C. A. StoF, “Pulsed Eddy Current Inspec#on of Second Layer Wing Structure,” MASc. thesis, Dept. of Chem. & Chem. Eng., RMC, ON, 2014. [4] J. Larn, J. D. Carroll and P. E. Green, “Analyzing Mul#variate Data,” Pacific Grove: Brooks/Cole, 2003, pp. 4, 83-‐123, 264-‐275. [5] P. Horan, P. Underhill and T. W. Krause, "Pulsed Eddy Current Detec#on of Cracks in F/A-‐18 Inner Wing Spar Without Wing Skin Removal Using Modified Principal Components Analysis," NDT&E Interna#onal, vol. 55, pp. 21-‐27, 2013. [6] V. K. Babbar, P. R. Underhill, C. StoF and T. W. Krause, "Finite Element Modeling of Second Layer Crack Detec#on in AircraS Bolt Holes with Ferrous Fasteners Present,” NDT&E Interna#onal, vol. 65, pp. 64-‐71, 2014. [7] V. K. Babbar, P. P. Whalen and T. W. Krause, "Pulsed Eddy Current Probe Development to Detect Inner Layer Cracks near Ferrous Fasteners Using COMSOL Modeling SoSware,” COMSOL Conference, Boston, 2012.


Documents

Op#mizing)Pulsed)Eddy)Current …3)Orion)Sample)Descrip#on) NDT in Canada 2015 Conference 12 3.3 NAVAIR Sample Description This sample was acquired for testing purposes from the NAVAIR