Assessment of Advanced Ultrasonic and Infrared Inspection Methods to Detect Delaminations and Water Ingress in Composite Honeycomb Materials David G. Moore.

Assessment of Advanced Ultrasonic and Infrared Inspection Methods to Detect Delaminations and Water Ingress in

Composite Honeycomb Materials

David G. Moore and Ciji L. Nelson Sandia National Laboratories

Nondestructive Evaluation and Experimental Mechanics Department

Post Office Box 5800 MS-0557Albuquerque, New Mexico 87185, USA

NDCM Conference May 20- 25, 2013Le Mans, France

Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. This presentation is declared a work of the U.S. Government and is not subject to copyright protection in the United States.

Sandia National Laboratories is studying composites due to their unique structure and potential for aerospace applications.

New composite manufacturing techniques are requiring us to implement advanced inspection methods.

Reference standard creation is key to the development of material characterization establishing inspection criteria limits and assessing equipment capabilities.

Introduction

Overview of Composites

• Generally four types of flaws in composite materials:

– Disbond– Core Damage – Delamination– Porosity

• These flaws can occur due to the following:– Impact damage – Lighting Strike (heat damage)– Manufactures Defect

Defects found in Composite Structures

Core Damage

Engineering Sample

The specimen is 20.3 by 20.3 cm composed of a Nomex™ honeycomb structure sandwiched between two carbon graphite laminate weave skins (3 plies thick). Five flaws were created in the specimen:

1) 2.54 cm diameter epoxy potted honeycomb core (full thickness);

2) 2.54 cm square Teflon shim (located between plies 2 and 3);

3) 12.7 mm diameter Teflon shim (located between plies 2 and 3);

4) 1.27 square disbond (located at adhesive bondline);

5) 2.54 cm diameter disbond (located at adhesive bondline).

Computed Tomography Baseline Inspection

The Computed Tomography (CT) technique collects penetrating radiation measurements from the composite sample’s x-ray opacity using an amorphous silicon digital detector array (flat panel).

The source and detector remain constant while the part is rotated. These slices are then collected and mapped together to create a three dimensional CT-density map.

The fraction of the x-ray beam that is attenuated will directly relate to the density and thickness of the material through which the beam has traveled.

Sh

allo

wer

Dee

per

CT Inspection Results

A Perkin Elmer (2048 by 2048 pixels) amorphous silicon flat panel detector with a 0.20 by 0.20 mm pixel pitch was used. The panel was scanned at a 1-to-1 ratio giving it a geometric resolution of 20 mils. The x-ray source operated at 160 Kilovolts and 4.4 mA

Top of the laminate showing the weave pattern at the

surface of the part.

Interface between the laminate and Nomex

honeycomb structure.

Sh

allo

wer

Dee

per

CT Inspection Results (continued)

Midway through the honeycomb core crushing

Honeycomb material (disbonds detected)

Side view of the panel showing an

area of core damage

Ultrasonic Properties

Material Sound Velocity v(m/sec)

Density ρ(kg/m3)

Acoustic ImpedanceZ = ρv (kg/m2 sec) x 106

Aluminum 6320 2700 17Water (20 °C) 1483 998 1.48Air (20 °C) 343 1.204 0.00041HysolTeflon

28501520

15802200

4.523.3

Composite- graphite-Epoxy

3070 1450 4.4

Ultrasonic InspectionUltrasonic transducers transmit sonic waves into a sample and measure the reflected responses.

Near surface resolution can be improved with a delay line tip, to provide a time delay between sound generation and reception of reflected energy.

Low-frequency Resonant, Ultrasonic Transducers transmit ultrasonic waves that penetrate through the laminate and enter the honeycomb cell wall at the node bond adhesive interface.

Front Surface Echo Back SurfaceEcho & Bondline

0.22 mm

Ultrasonic (UT) Setup and Results

The Composite Sample was UT inspected using the below equipment and setup,

- MAUS V™, (Mobile automated UT scanner) was used to acquire the images.

- A 5 MHz probe, 6.35 mm in diameter.

- Scanner resolution- 0.5 mm.

- The gate was set to monitor the backwall signal from the laminate.

AmplitudeGate Image

DepthGate Image

Resonance inspection requires a narrowband transducer that can be excited at its natural resonance frequency. A continuous standing wave is coupled into the material.

Resonant inspection can be used to inspect honeycomb materials and compliment conventional ultrasonic inspection.

The cursor tracks the shift in the signal phase (X) and amplitude (Y). Any changes in structural resonance (disbonds, or delaminations) is represented by changes in the Resonant Frequency at that point in the sample.

Resonance Basics

Shift in Resonant Frequency

Am

plitude

Phase

Resonant inspection was moderately successful on the sample at 110 KHz. Resonance probe could not reliably detect the Teflon inserts but could detect the epoxy filled honeycomb core.

The signal is attenuated around the perimeter of the Teflon inserts making the defect features hard to discern.

The use of Teflon inserts in a resonant inspection reference standard should be carefully considered. Teflon inserts may not be usable for reliable instrument calibration or establishing a reject criterion.

Amplitude (X) Image

Phase(Y) Image

Resonance Inspection Results

Thermal Material Properties

Conduction: energy transfer from a more energetic particles to less energetic particles within a material. Interactions between particles are due to a thermal gradient.

Fourier's law defines time rate of heat transfer through a material. The heat flux is proportional to the negative gradient in the temperature and to the area. The proportionality constant is the transport 𝑘property thermal conductivity W/(m °C).

Heat flux q" is the heat transfer rate in direction x per unit area perpendicular to the direction of transfer. Since heat transfer rate is a vector quantity it can be written in general of the conduction rate equation:

T1 > T2 T2

𝒒 𝒙=−𝒌𝒅𝑻𝒅𝒙

x

= q"

q"

Thermophysical Properties Thermophysical properties have two distinct categories: Transport and Thermodynamic.

Transport properties include the diffusion rate coefficients thermal conductivity and kinematic viscosity.

Thermodynamic properties are useful to define the state of equilibrium. Density – ρ (kg/ m3) Heat capacity – c (J/ kg °C) and volumetric heat capacity ρc (J/ m3 °C ). Solids can store large amounts of thermal energy when compared to gases.

The ability of a material to conduct thermal energy relative to it ability to store thermal energy is termed thermal diffusivity α (how fast the material temperature adapts to the surrounding temperature).

Thermal effusivity (ε) is a measure of a materials ability to exchange thermal energy with its surroundings

MaterialConductivity, 𝑘

W/(m °C) 

Specific Heat, cp

J/(kg °C) 

Density,ρ

kg/(m3 )

Diffusivity,α

m2/sec 1 x 10 -7

Effusivityε

J/(m2 °C)s 

Phenolic (resin pressed)

0.3766 1255 1380 2.174 807.667

Teflon 0.2510 1004 2170 1.152 739.6Carbon

Graphite167.36 707.1 2250 1052 16317.6

CFRP ParallelCarbon Fibers

7 1200 1600 36.45 3666.06

CFRP PerpendicularCarbon Fibers

0.8 1200 1600 4.167 1239.45

Epoxy (hysol) 0.1945 1172 1210 1.372 525.271Aluminum 2024 T3 121 875 2780 497.43 17156.1

 Copper 397.48 384.9 8940 1155. 36982.8

Stainless Steel 304 14.644 502.1 7920 36.83 7631.1

GRP ParallelGlass Fibers

0.38 1200 1900 16.67 930.81

GRP PerpendicularGlass Fibers

0.30 1200 1900 13.16 827.04

Thermal Properties

Infrared BasicsActive Thermography (AT) is a technique where a stimulus is applied to a surface and causes it to heat or cool in such a way to allow the surface characteristics to be observed by an infrared camera.

Lamp

IR camera

defect

PC

heat conduction

emitted IR

light

IR Image

Source: Thermal Wave Imaging

EchoTherm EquipmentThe EchoTherm System used the following components and settings:

- FLIR 6106 camera was ran at 60Hz at an image size of 640 X 512 pixels, with an InSb detector and 14 bit output.

- The hood contained two xenon flash lamps partially surrounded by parabolic reflectors and 5000 Joules per lamp.

- The collected images were processed with Thermographic Signal Reconstruction (TSR)™ technology in the Mosaiq software package, to provide a set of images representing the heat transfer.

Source: http://www.thermalwave.com

Bond separationDelaminations

Dents/crushed core: No greater than 76.2 mm (3.0”) diameter or deeper than 12.7 mm (0.050”) Holes and punctures: No greater than 25.4 mm (1.0”) diameter. Disbonds between core and skin: (core crush, split core) No greater than 30 cm (12.0”) Core bond separations: No greater then 45.2 square cm (7.0 square inches).Liquid Intrusion in cells: No more than 40 cells

Inspect to make sure water and moisture is within limits. No further water removal efforts are required provided entrapped water is within limits and leak path is found, repaired, and water drip test confirms repair.

“Damage” Limits for Composites

Infrared Data Analysis

Cool Warm

First derivative of reconstructed intensity per unit time.

First and second derivatives yield substantial contrast at earlier thermal decay times in the image sequence. The images will be sharper. (less lateral diffusion).

5.69

6.36

7.03

7.69

8.36

9.02

0.02 0.06 0.27 1.16 4.79 20.03

time (sec)

Epoxy

Teflon

Source: http://www.thermalwave.com

High Diffusivity (α )

Low Diffusivity (α )

High Effusivity (ε)

Low Effusivity (ε)

-0.40

-0.20

0.00

0.20

0.40

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

-0.90

-0.70

-0.50

-0.30

-0.10

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

5.69

6.36

7.03

7.69

8.36

9.02

0.02 0.06 0.27 1.16 4.79 20.03

time (sec)

IR Results (Epoxy and Teflon)

Intensity Time Plot 0.55 seconds after pulse

Log of reconstructed intensity per time.

First derivative of reconstructed intensity per time. Second derivative of reconstructed

intensity per time.

-0.40

-0.20

0.00

0.20

0.40

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

-0.90

-0.70

-0.50

-0.30

-0.10

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

5.66

6.38

7.09

7.81

8.52

9.24

0.02 0.06 0.27 1.16 4.79 20.03

time (sec)

IR Results (Epoxy and Disbond)

Intensity Time Plot 1.1 seconds after pulse

Log of reconstructed intensity per time.

First derivative of reconstructed intensity per time.

Second derivative of reconstructed intensity per time.

-0.40

-0.20

0.00

0.20

0.40

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

-0.90

-0.70

-0.50

-0.30

-0.10

0.05 0.16 0.54 1.79 5.98 20.03

time (sec)

5.66

6.38

7.09

7.81

8.52

9.24

0.02 0.06 0.27 1.16 4.79 20.03

time (sec)

IR Inspection Results (continued)

Intensity Time Plot 4.6 seconds after pulse

Log of reconstructed intensity per time.

First derivative of reconstructed intensity per time.

Second derivative of reconstructed intensity per g time.

Polymer PropertiesHydrophilic polymers characterized by different swelling factors were evaluated by: Bogomil YOCHEV, Svetoslav KUTZAROV, Damyan GANCHEV, Krasimir STAYKOV, Technical University – Sofia, Bulgaria for their ultrasonic properties.

- Preliminary experiments show that polymers having water content above 85% can not be used as solid dry-couplants for practical purposes because of their low mechanical strength. Might be used as a reference standard for water ingress into composite materials.

- When the water content increases the longitudinal ultrasonic velocity decreases and nears that of water – 1480 m/s.

- Hydrophilic polymers have longitudinal ultrasonic velocity significantly lower than the velocity of PMMA (type-H) – 2750 m/s, polyizoprene rubber (type-F) – 1840m/s and is comparable to that of Aqualene (type-G) – 1580m/s

Reference: Bogomil YOCHEV et.al

Cut from Aircraft Sample The aircraft sample was cut from a secondary load bearing surface of a general use commercial aircraft. It has naturally occurring skin to core separation and ply damage.

To simulate water ingress an elastomer (Aqualene) was placed at the ply to honeycomb surface. This elastic polymer is designed specifically for ultrasonic probes. The acoustic impedance is nearly the same as water. Its attenuation coefficient is lower than most elastomers and plastics.

0.10 mm skin thickness

Aqualene placement

15.2 cm

22.9 cm

Elastomer Experiment

Log of reconstructed intensity per time (2.61 seconds)

First derivative of reconstructed intensity per time (2.61 seconds).

Water Ingress Experiment

Cells were saturated with water and inspected with active IR. This technique works well for development of the technique but cannot be used for a long term reference standard.

Water Ingress Experiment (continued)

Log of reconstructed intensity per time (2.56 seconds).

First derivative of reconstructed intensity per time (2.56 seconds).

Water Ingress Experiment (continued)

First derivative of reconstructed intensity per time. These six images were taken at different soak times from left to right (0, 5, 18 minutes, top row and 53, 74, and 178 minutes, bottom row). This allows the water to migrate through the composite.

SummaryCore potting is an excellent way to produce a reference standard that will represent splices or honeycomb repairs. All inspection techniques were able to detect this manufacturing process. Resonant inspection and computed tomography could not reliability detect the embedded inserts between the plies. Ultrasonic resonant inspection can detect near surface disbonds in the laminate bonded to honeycomb but may not be able to establish a reference standard threshold. High frequency ultrasonics can detect the embedded inserts in the plies with a high signal-to-noise ratio.

Computed tomography was a valuable assessment tool for evaluating through the thickness of the honeycomb material as well as the ply to ply variations, but is limited on the size of part that can be inspected.

ConclusionsHoneycomb removal at the bondline to create a core disbond was identifiable with computed tomography and infrared inspection. The thermal properties of the epoxy and air are significantly different from the honeycomb and carbon graphite plies, however thermal properties of Teflon are not. The use of signal processing algorithms greatly enhances the detectability of flaws that are hidden in the raw infrared image. The TSR helps with the evaluation and characterization of indications in the sample. Flashed thermography can identify flaws within the first three plies and the bondline interface. The detectability of a Teflon insert embedded into the plies will depend on its size, depth from the surface and the degree to which its thermal properties differ from the surroundings. The use of representative test samples allows the signal process software to characterize fluid ingress into honeycomb cells and track the migration path. The elastomer material has material properties that almost simulate the thermophysical behavior of water ingress into honeycomb cell walls. The key difference in the thermal behavior is found in the first and second derivatives.

Questions?