Eddy Current Inspection of Composites - Jentek Sensors … ASNT Spring... · Eddy Current Inspection of Composites Zachary Thomas, Neil Goldfine, Christopher Martin, Robert Lyons,

Slide 1 Copyright © 2013 JENTEK Sensors All Rights Reserved.

MWM sensors and MWM-Arrays are covered by several issued and pending patents, including, but not limited to: 8,237,433, 8,222,897,8,050,883, 7,994,781, 7,876,094, 7,812,601, 7,696,748, 7,589,526, 7,533,575, 7,528,598, 7,526,964, 7,518,360, 7,467,057, 7,451,657, 7,451,639,7,411,390, 7,385,392, 7,348,771, 7,289,913, 7,280,940, 7,230,421, 7,188,532, 7,183,764, 7,161,351, 7,161,350, 7,106,055, 7,095,224, 7,049,811,6,995,557, 6,992,482, 6,952,095, 6,798,198, 6,784,662, 6,781,387, 6,727,691, 6,657,429, 6,486,673, 6,433,542, 6,420,867, 6,380,747, 6,377,039,6,351,120, 6,198,279, 6,188,218, 6,144,206, 5,966,011, 5,793,206, 5,629,621, 5,990,677 and RE39,206 (other US/foreign patents issued and pending).

ASNT Spring Research Symposium 2013Memphis, TN

March 18-21, 2013

© JENTEK Sensors 2013

Eddy CurrentInspection of CompositesZachary Thomas, Neil Goldfine, Christopher Martin,

Robert Lyons, and Andrew Washabaugh

JENTEK Sensors, Inc., 110-1 Clematis Avenue, Waltham, MA 02453-7013 Phone: 781-642-9666; Email: [email protected]

Copyright © 2013 JENTEK Sensors All Rights Reserved.


Presentation Outline

• Background– Development program goals– MWM-Array® technology– Depth of penetration of sensing fields

• Basic model predictions and validation measurements– Eddy-current extension of micromechanical model– Orientation effects– Field spatial variations

• Example imaging applications– Volumetric imaging of damage– Composite Overwrapped Pressure Vessel imaging

• Summary


Development Program Approach

• Goals– To develop model-based methods for (primarily) carbon fiber composite NDT– To demonstrate high resolution damage and condition imaging for composites– To develop volumetric stress sensing magnetic stress gages for composites

• Approach– Focus on eddy current methods and sensor designs that are readily modeled.– MWM-Arrays uses a linear drive eddy current sensor array construct

• Can induce eddy currents in the linear fibers of carbon fiber composites• Use winding geometry changes to alter penetration depth and assess

material condition (e.g., damage and stress)• Funding

– NASA for micromechanical model development and application to composite overwrapped pressure vessels (COPVs)

– Army for rotorblade NDT– Navy for NDT of aircraft composites


MWM®-Array Technology• Eddy current array geometry designed to match (isotropic) models for responses• The voltage induced on sense element(s) is measured.• Measurement grid methods provide conversion of measured responses into physical

properties (e.g., conductivity, lift-off, permeability)

MeasurementGrid Methods

Magnetic field interactingwith test material


Example MWM®-Arrays• MWM-Array dimensions can be adjusted for the application

=> Drive-sense gap (spatial wavelength) affects penetration depth

FA1547 elements

medium wavelength

FA497 elements

several wavelengths

FA2837 elements

small wavelength

FA2437 elements

large wavelength


MWM-Array Depth of Penetration• Magnetic field decays exponentially with distance away from sensor

– Decay rate determined by skin depth at high frequencies and sensor dimensions at low frequency

increasing size

Skin depthlimited

0.1

2

3

456

1

2

3

456

Pen

etra

tion

Dep

th (

mm

)

103 104 105 106 107 108 109

Frequency (Hz)

456

10

2

3

456

100

2

3

Pen

etration D

epth (m

ils)

FA28

FA24 Electrical Conductivity

(kS/m)

10100 1

MWM-Array

FA154

0


Micromechanical Model: Eddy-Current Extension• Model considers fiber bundles as a composite cylinder assemblage

– Solve for field around a single fiber and extend to fiber bundle– Effective complex permeability and conductivity depend upon orientation with respect to fiber axis,

fiber density and fiber contact• For carbon fiber composites

– Graphite fibers: ~7 m diameter, nonmagnetic, ~20 kS/m (0.0344%IACS)– Radius << skin depth for typical eddy-current frequencies

• Indicates a strong orientation dependence of the properties– MWM-Arrays with linear drives can provide a measure of these orientation dependent responses

perp

Fiber par

perp

Radius a

perp

par

perp

* *par perp o par f fv 0perp

This first order model neglectsinterconnections (touching)

between fibers


Composite Measurements: Orientation Effect• Center element for FA28 MWM-Array• Strong response when aligned with fibers in individual plies

UnidirectionalLayup

0.35

0.34

0.33

0.32

0.31

0.30

0.29

Re

al P

art

(H/m

)

720630540450360270180900-80

-60

-40

-20

0

20

40

Imag

inary P

art (x10

-3 H/m

)

FA28, Channel 18, 10 MHzUniaxial specimen

Real part Imaginary part

0.35

0.34

0.33

0.32

0.31

0.30

0.29

Re

al P

art

(H/m

)

720630540450360270180900

Orientation Angle (deg)

-80

-60

-40

-20

0

20

40

Imag

. Part (x10

-3 H/m

)FA28, Channel 18, 10 MHzLMT Bend Panel 1

Real part Imaginary part

Quasi-isotropic Layup(alternating layers at

-45º, 0º, 45º, 90º)


Simple layered-Media Composite Representation• Layup for quasi-isotropic test panel

– Uniaxial properties for each layer• MWM-Array sensitive to composite layers with fibers

oriented parallel to drive windings• Composite layer considered insulating if fibers NOT

within several degrees• This visualization indicates that each sensor

orientation is only sensitive to a subset of plies at varying depths within the composite.

MWM-Arrayh

0 DEGREES45 DEGREES90 DEGREES-45 DEGREES

eff

eff

eff

eff

eff

eff

eff

eff

MWM-Arrayh

MWM-Arrayh

eff

eff

eff

eff

eff

eff

eff

eff

eff

effeff

eff

effeff

eff

eff

eff

eff

eff

effeff

eff

eff

eff

MWM-Arrayh

MWM-Arrayh

MWM-Arrayh

MWM-Arrayh

MWM-Arrayh

MWM-Arrayh

0º orientation 45º orientation 90º orientation -45º orientation

(note that the angle is relativeto the fibers at the surface)


Layered-Media Composite Grid Example• Conductivity/lift-off measurement grids

assuming quasi-isotropic layup– Non-zero conductivity only for aligned layers

in each orientation• Primarily observe response shift as

effective lift-off changes with orientation• General agreement with measurement

data in each orientation– Data is below the grids for the deep plies (0º

and 90º), so other factors need to be considered

0 45

90

-45

45º

90º

0º

-45º

Quasi-isotropic Layup(note that the angle here refers

to the panel orientation)

FA49, Far channel

Increasingconductivity

Increasinglift-off


Example COPV Layup

• Representative layup for Composite Overwrapped Pressure Vessels (COPV)

• MWM-Array sensitive to composite layers with fibers oriented parallel to drive windings

• This indicates that the sensor orientation is important for assessing the fiber properties

non-fiber orientationAl liner

MWM-Array

h

Al liner

MWM-Array

h

eff/2

eff/2

eff/2

±17º or ±18º orientationAl liner

MWM-Array

h

Al liner

MWM-Array

h

eff

eff

±90º orientationAl liner

MWM-Array

h

Al liner

MWM-Array

h

eff/2

±60º orientationAl liner

MWM-Array

h

Al liner

MWM-Array

h


Simpl Layered-Media Model with Measurements

0.35

0.34

0.33

0.32

0.31

0.30

Rea

l Par

t of T

rans

indu

ctan

ce (H

/m)

600550500450400350300250200150100500


-60

-40

-20

0

20

Imag. P

art of Transinductance (x10-3 H

/m)

FA28, Ch 18, 10 MHz, LMT08Real Imag

MeasModel

0.35

0.34

0.33

0.32

0.31

0.30

0.29

Rea

l Par

t of T

rans

indu

ctan

ce (H

/m)

600550500450400350300250200150100500


-60

-40

-20

0

20

40 Imag. P


/m)

FA28, Ch 18, 10 MHz, COPVReal Imag

MeasModel

0.10

0.05

0.00

-0.05

Rea

l Par

t of T

rans

indu

ctan

ce (H

/m)

600550500450400350300250200150100500


-60

-40

-20

0

20

40

60 Imag. Part of Transinductance (x10

-3 H/m

)

FA24, Ch 18, 10 MHz, COPVReal Imag

MeasModel

FA24

Qua

si-is

otro

pic

CO

PV

FA28

• Gaussian distribution model assumed for fiber conductivity in each layer• FA28 shows reasonable agreement with rotation measurements• FA24 shows broader response and background variation consistent with low-level

background conductivity (i.e., fiber touching should be considered)

0.10

0.05

0.00

-0.05

Rea

l Par

t of T

rans

indu

ctan

ce (H

/m)

600550500450400350300250200150100500


-60

-40

-20

0

20

40

60 Imag. P


/m)

FA24, Ch 18, 10 MHz, LMT08Real Imag

MeasModel


Model Response: Field Variations for Composite• 3D model for field distribution around drive• Plots show normal field component along nominal

position of sense elements in rectangular loops– Drive winding aligned with uniaxial composite fibers

• Field varies with position along sense element array and degree of anisotropy Composite material

(anisotropic properties)

Air [insulating dielectric](isotropic properties)

z

Plane for MWM-Array

-10

-8

-6

-4

-2

0

2

4

6

Rea

l par

t of B

z (x1

0-9 T

)

-0.4 -0.2 0.0 0.2 0.4

y position (m) [x value at center of sense element array]

-50

-40

-30

-20

-10

0

10 Imaginary part of B

z (x10-9 T)

span of drivespan of sense element array

fields outside drive

field varies across array

FA24, 10 MHzperp/par Real Imag1.00.0050.0010.00010.00001

y

Double-D Drive Winding


Uniaxial Specimens: FA28 parallel to fibers• Channel-to-channel variation consistent with field variation along drive winding

“High” conductivity composite

“Low” conductivity composite

“High” conductivity composite

“Low” conductivity composite


Field Mapping Setup• Excite drive of a FA24 and scan with sense elements from

a second array • Sense element responses proportional to normal

component of field– The calibration led to the responses having a negative polarity

compared to the previous results

CH.1

CH.37 (FA24)0.100” SPACING

CH.37 (FA28)0.040” SPACING

5.071”

Calibration(sense elements parallel to drive)

Scanning(sense elements perpendicular to drive)

FA24


Field Mapping: C-Scans• Greater field variation for carbon fiber composite

– Significant field extending beyond the drive winding in the horizontal direction– Within the drive, the imaginary part goes though a maximum near the center

Aluminum Alloy

Rea

l Par

tIm

agin

ary

part

Uniaxial Composite


Field Mapping: B-scan, composite• One channel passing along the center of the drive

Rea

l Par

tIm

agin

ary

part

Horizontal Position

field variation acrosssense element location

Fiel

ds e

xten

dou

tsid

e dr

ive


Field Mapping: B-scan, aluminum alloy• One channel passing along the center of the drive

Rea

l Par

tIm

agin

ary

part

Horizontal Position

fields relatively uniformacross sense element location

Fiel

ds d

o no

t ext

end

outs

ide

driv

e


Volumetric Property Imaging Approach

• Combination of sensor orientation and geometry can isolate depth and region of damage– sensor orientation determines plies– sensor geometry determines depth of sensitivity– spatial extent of damage determined from scan image

MWM-Arrayh

Damage regionFA26

0ºFA2690º

FA280º

FA2890º

Quasi-isotropic panelwith impact damage


Volumetric Imaging of Composite Impact Damage

Representative MWM-Array Scan Image Sample provided courtesy of Lockheed Martin


Representative Quasi-isotropic Panel Scan ImagesSubtractedAfter ImpactBefore Impact Subtracted and Smoothed

5 M

Hz

10 M

Hz

15.8

MH

z


Summary of Scans• Individual scans combined together to create composite cross-sectional view


Cross Sectional Images: Panel 1, Low Impact Level

Cross Sectional View along X-axis

Cross Sectional View along Y-axis

MWM-Array FA28 Data for 0.085-in. thick panels

0.000

-0.0850 5Horizontal distance (inches)

Dep

th (i

nche

s)

0.000


Dep

th (i

nche

s)


Cross Sectional Images: Panel 2, Medium Impact Level




0.000


Dep

th (i

nche

s)

0.000


Dep

th (i

nche

s)


Cross Sectional Images: Panel 3, High Impact Level




0.000


Dep

th (i

nche

s)

0.000


Dep

th (i

nche

s)


Composite Overwrapped Pressure Vessels (COPVs)• Helical wraps at several different orientations with a metallic liner• Eddy current scans can image both liner and composite properties


COPV: Low Frequency Inspection• 50 kHz• 90° drive orientation with 0.066-in. thick overwrap • At this frequency the sensor responds primarily to the liner• Effective lift-off images show dents in liner• Higher impact energy results in larger dents in the aluminum liner

Preliminary FilteringBefore Impact Damage After Impact Damage Baseline Subtracted

axial

angu

lar 30 ft-lbs

~46 mils

15 ft-lbs~22 mils 10 ft-lbs

~6 mils

20 ft-lbs~26 mils


COPV: High Frequency Inspection• 5 MHz

– At this frequency more of the signal related to the composite overwrap properties• 90° drive orientation with 0.066-in. thick overwrap• The conductivity images show significant spatial variations in the overwrap properties• Changes in the effective conductivity images highlight the damage

axial

Preliminary FilteringBefore Impact Damage After Impact Damage Baseline Subtracted

30 ft-lbs

angu

lar

20 ft-lbs

15 ft-lbs 10 ft-lbs


Summary

An eddy current extension to a micromechanical model has been developed for conducting fiber composites

Layered-media models have been developed to account for anisotropic properties

Measurements have verified the orientation dependence of the sensor response and field spatial variations in the vicinity of the sensor

Ongoing work is aimed at refining the models and correlating electrical properties to the properties of interest, such as:

- Fiber density variations within a ply- Stress dependent changes in the electrical properties- Volumetric assessment of stress and damage conditions


Questions?

Phone: 781-642-9666Email: [email protected]

www.jenteksensors.com

The views and opinions expressed in this presentation are those of the authors and do not necessarily represent official policy or position of JENTEK Sensors, Inc.,

or any Department of the U.S. Government.

Copyright © 2013 JENTEK Sensors All Rights Reserved.

Documents

Eddy Current Inspection of Composites - Jentek Sensors … ASNT Spring... · Eddy Current Inspection of Composites Zachary Thomas, Neil Goldfine, Christopher Martin, Robert Lyons,