Detection of crack/corrosion

using active thermography

- Zishan Ali Khan
 Objective of the project

Designing a Non-Destructive Evaluation (NDE) method for

detection of de-bonding/corrosion in a multi-layered medium by
the use of active infrared techniques.
 Problem description
𝑟
  𝑞  ′ ′ Materials
1 – Paint
𝑧
 
2 – CFRP
𝑧=0
 
1 3 – GFRP
𝑧=𝑧 1
 
2 4 – Aluminum
𝑧=𝑧 2
 
5 – Primer
3
𝑧=𝑧 3
 
4  
𝑧=𝑧 4
  - Heat Input
5  𝑧=𝑧
5 - Corrosion/de-bonding
Half-sectional view of specimen
Flow chart of active infrared thermography (IRT)
Using infrared
camera

Specimen Thermogram is
Input - Heat temperature recorded at
rises regular time

Using halogen
lamps or
ultrasonic
Using Thermogram
transducer
discrete sequence
Fourier processed
transform
Image ref.- Amazon
Defect detection

Principle :
At the surface, defective
zones will be at different
temperatures with
respect to non-defective
zones due to difference
Defect detection from temperature profiles (sound area and
in thermal conductivities defective zone)

Pulsed phase thermography reviewed - C. Ibarra, X. Maldague- QIRT Journal -

2004
 Various available methods in active IRT
1. Pulsed Thermography
2. Pulsed Phase Thermography
3. Modulated (Lock-in) Thermography
4. Vibrothermography (Thermosonics)
LITERATURE REVIEW
 Pulsed thermography (PT)
● Input – Pulse heat using flash lamp
● Infrared camera measures the temperature of the
surface
● Subsurface discontinuities are resistance to heat
flow
● It produce abnormal temperature at the surface
● PT data is processed using mainly three techniques
1. Thermal contrast based technique
2. Pulsed phase thermography (PPT)
3. Thermographic signal reconstruction
 Experimental setup of PT/PPT

𝑧
 

“Active infrared thermography techniques for the nondestructive testing of materials”, C. Ibarra, X. Maldague - Ultrasonic and Advanced
Methods for Nondestructive Testing and Material Characterization, pp. 325-348 (2007)
●  The 1D solution of the Fourier equation for the propagation of a Dirac heat
pulse

Where –
= Energy absorbed by the surface (J/ m2)
= Initial temperature
= Thermal Conductivity (W/m -°C)
= Heat Capacity (J/kg -°C)
 Limitations of PT
● Sensitivity of the technique to edge effects
● High stress concentration leads to failure of a bonded repair at edges
● This leads to challenges in detecting the early stages of dis-bonding due
to the aforementioned edge effects
● Variation in thermal properties also effect end result
● PT raw data is also sensitive to uneven heating
 Pulsed phase thermography
●  
● In this, data is transformed from the time domain to the frequency domain
using the one-dimensional discrete Fourier transform (DFT)
  𝑁 −1
𝑖 2 𝜋 𝑛𝑘
𝐹 𝑛=Δ 𝑡 ∑ 𝑇 ( 𝑘 Δ𝑡 ) exp −
𝑘 =0
( 𝑁 )
=R e 𝑛 +𝑖 I m𝑛

= imaginary number ()
= frequency number ( = 0, 1, 2,….N)
= Real part of the transform
= Imaginary part of the transform
Δt = sampling Interval
 Lock-in thermography
● Heat Input - Periodical (sinusoidal) heating by an intensity modulated laser
beam/halogen lamp
● Thermal response is decomposed by a lock-in amplifier to extract the
amplitude and phase of the modulation
● Internal defects produces changes in amplitude and phase of the response
signal at the surface.
 Experimental setup for Lock-in thermography

 𝑧

“Active infrared thermography techniques for the non-destructive testing of materials”, C. Ibarra, X. Maldague - Ultrasonic and
Advanced Methods for Non-destructive Testing and Material Characterization, pp. 325-348 (2007)
 
The Fourier’s law 1D solution for a periodic thermal wave propagating through a semi-infinite
homogeneous material may be expressed as-

Where –
= Initial temperature
= 2f (rad/s), modulation frequency
= Thermal wavelength (2) (m)
µ =
= thermal diffusivity (m2/s)
= Frequency (Hz)
 ●  
I = Input Signal
S = Response

● Change in Amplitude (A) is

A= 

● Change in Phase ()

= arctan

“Active infrared thermography techniques for the

nondestructive testing of materials”, C. Ibarra, X.
Maldague - Ultrasonic and Advanced Methods for
Nondestructive Testing and Material Characterization, pp.
Only valid for Sinusoidal stimulation
325-348 (2007)
 Vibrothermography (VT)

Mechanical Waves travel

transducer directly Heat is generated at through specimen
simulates internal the cracks and transmitted
defects back to surface

Mechanical
excitation
IR camera detects Transducer picks up
the defect signature reflected signal
Lock-in Vibrothermography Burst Vibrothermography
● Also known as amplitude modulated VT ● It is is analog to PT approach
● It is analog to the LT approach

𝑧  𝑧 

“Active infrared thermography techniques for the nondestructive testing of materials”, C. Ibarra, X. Maldague - Ultrasonic and Advanced
Methods for Nondestructive Testing and Material Characterization, pp. 325-348 (2007)
 Advantages
● Ultrasonic waves are ideal for NDT since they are not audible to humans
● Defect detection is independent from of its orientation inside the specimen
● Both internal and open surface defects can be detected
 Techniques we are going to use and Why
1. Pulsed phase thermography (PPT)
● It uses 1-D discrete Fourier transform (DFT) to extract amplitude and phase
transformation from time-temperature data
● It helps in de-noising the signal
● It is fast and easy to deploy method
● Less sensitive to the non-uniformity of heating, surface irregularities and
environmental reflections
 Techniques we are going to use and Why

2. Vibrothermography/Thermosonics
• It is extremely fast
• It is most appropriate technique to detect micro defects
• There is only minimal heating of the inspected specimen since energy is
usually dissipated mostly at the defective areas
 Inspection of aerospace materials by
pulsed thermography (PPT), lock-in
Thermography (LT) and
vibrothermography(VT): A comparative
study
Clemente Ibarra-Castanedo, Marc Genest, Stéphane Guibert, Jean-
Marc Piau, Xavier P. V. Maldague and Abdelhakim Bendada

doi: 10.1117/12.720097
Experimental configurations-optical excitations
Experimental configurations-mechanical excitations

● Ultrasonic waves are

produced by
transducer
● Coupling medium is
necessary
Geometry and defect distribution of specimen - HONEYCOMB1108

● In Zone I, 20 Teflon
inserts are placed
between CFRP plies
● In Zone II, 6 Teflon
inserts are placed
between
1. adhesive and core(top
row)
2. facesheet and
adhesive(bottom row)
● In Zone III, 8 cells are
filled with water
 Results- Zone I – PPT and LT
PPT
• Possible to detect almost all
the defects from two
phasegrams
• It provides better contrast to
some defects
  phase at = 0.082 Hz
PPT   phase at = 0.74 Hz
PPT

LT
• LT detects almost all the defects from a single
phasegram
• Provides better overall detection

 LT phase at = 0.001 Hz

 Results- Zone I - VT

VT
• It only detects some of the thicker inserts (t = 0.33
mm, two bottom lines)
• Thicker inserts dissipate mechanical energy more
effectively
• Heat conduction from thinner inserts is practically
VT result
undetectable
 Results- Zone II – PPT and VT
PPT
● Slight presence of inserts(t = 0.16 mm) can be seen
● Fiber structure is visible
● Fiber structure is contributing to hide the defect signature

PPT
  phase at = 0.041 Hz

VT
• It shows the four largest inserts
• It has advantage over optical excitation in detecting
deeper defects

VT result
 Results- Zone III – PPT, LT and VT

● Water ingress is successfully detected by PT

and LT but not by VT
● Water detection by VT is difficult due to very
high transmission loss through aluminum
honeycomb core
 Synthetic PPT phase at = 0.07 Hz
Geometry and defect distribution of specimen - Glare
● Glare – Glass-reinforced fiber metal
laminate
● It contains 3 multilayer glass fiber
structures between 4 Aluminum
layers
● Different material and sizes inserts
are embedded between second and
third ply
 Results
PT
● All defects can be seen using PT
● Data is processed using thermographic signal
reconstruction (TSR) with first time derivative
● Aluminum inserts have a distinctive signature
PT using TSR first derivative at t=0.6 s
(inverted sign)

VT
• Most of the defects can be seen using VT
• But it is more difficult to differentiate between
material using VT

VT result
 Conclusion
● PT is fast and easy to deploy
● Optical techniques are affected by non-uniform heating
● But it can be reduced using processing algorithms, e.g. PPT, TSR
● LT is generally slower than PT
● Energy requirement of LT is low
● VT is useful to detect deeper delaminations
● VT is more suitable for smaller objects
● VT is useful in detecting micro defects
 VALIDATION
• Finite difference method (FDM) will be used for determining
temperature distribution
• For validating FDM, we will match it with analytical solution of heat
diffusion equation
 Finite difference method
Forward in time and central in space (FTCS)
●  
 
Where
= discretizations in -direction
= discretizations in -direction
= discretizations in -direction
 Slab in rectangular co-ordinate system (1D)
●  Governing differential equation

𝑇
 
∞ 𝑇
 
∞
● Boundary conditions

0
 
𝑥
 

2𝐿 
 Slab in rectangular co-ordinate system (1D)
●  Analytical Solution
● Characteristic Equation
● FDM defined parameters
1. CFL = 0.5
2. Time step = 1s
● Calculated parameters
1. Space step ( ) = 0.0027 m
● Geometrical parameters
1. m
● Experimental parameters
1. Time of experiment = 3000s
 Slab in rectangular co-ordinate system (1D)
●  FDM solution

Where
= Discretization in
= Discretization in
Slab in rectangular co-ordinate system (1D)
 
Conclusion
• Analytical and finite
difference solution are
exactly matching
Material Properties
• Material = Stainless Steel
• = 14.2 W/m-K
• J/kg-K
• kg/m3

𝐿
 
Slab in rectangular co-ordinate system (1D)

Conclusion
• Analytical and finite
difference solution are
exactly matching
 Infinite cylinder in polar co-ordinate system (1D)
●  Governing differential equation

𝑇
 
∞ 𝑇
 
∞
● Boundary conditions

 
(0,0) 𝑟
 

𝑅
 
 Infinite cylinder in polar co-ordinate system (1D)
 
●  Analytical solution ● FDM defined parameters
1. CFL = 0.5
2. Time step = 1s
● Calculated parameters
1. Space step ( ) = 0.0027 m
● Geometrical parameters
1. m

 
● Experimental parameters
● Characteristic equation 1. Time of experiment = 3000s
Bi
Infinite cylinder in polar co-ordinate system (1D)
●  FDM Solution

Where
= Discretization in
= Discretization in
Infinite cylinder in polar co-ordinate system (1D)
 
Conclusion
• Analytical and finite
difference solution (FDM)
are exactly matching

Material Properties
• Material = Stainless Steel
• = 14.2 W/m-K
• J/kg-K
• kg/m3
𝑅
 
Infinite cylinder in polar co-ordinate system (1D)

Conclusion
• Analytical and finite
difference solution
(FDM) are exactly
matching
 Slab in rectangular co-ordinate system (2D)
●  Governing differential equation

● Boundary conditions
𝑇
  =𝑇 ∞

2W
𝑥 

2W
 
• Initially, body is at
• Face, is instantaneously brought to
• Face, is maintained at
 Slab in rectangular co-ordinate system (2D)
● Analytical Solution

where

 ´
𝑥=
𝑥
,´
𝑦=
𝑦
,´
𝑡=
𝛼𝑡  ´ 𝑇 −𝑇 ∞
𝑊 𝑊 2
𝑊
𝑇 =
𝑇𝑖− 𝑇∞
 Slab in rectangular co-ordinate system (2D)
●  FDM Solution

Where
= Discretization in
= Discretization in
= Discretization in
 Slab in rectangular co-ordinate system (2D)
●  Defined FDM parameters ● Geometrical parameters
1. CFL0.5 1. W = 0.02 m
2. Time step = 1s ● Experimental parameters
● Calculated FDM parameters 1. t = 30 s
1. Space step in = 8.42 X 10-4 m
2. Space step in = 8.42 X 10-4 m
Slab in rectangular co-ordinate system (2D)
Analytical FDM
Slab in rectangular co-ordinate system (2D)
Analytical FDM
 Temperature distribution in multilayer
medium with and without crack using FDM
Without crack With crack
𝑟
 
𝑟
 

𝑞′
  ′   - Heat Input 𝑞  ′ ′
- Crack
𝑧
 
𝑧
 
  𝑧=0 𝑧=0
 
1
𝑧=𝑧 1
  𝑧=𝑧 1
 
2
𝑧=𝑧 2
  𝑧=𝑧 2
 
3
𝑧=𝑧 3
  𝑧=𝑧 3
 
4
  𝑧=𝑧 4 𝑧=𝑧 4
 
Materials specifications Simulation parameters
 
S. No Materials Thickness (mm) • Defined FDM parameters
1 CFRP 2 1. CFL = 0.1
2 GFRP 0.5 2. Space step in direction,
3 Aluminum 3.5
3. Space step indirection,
4 S. Steel 100
• Calculated FDM parameters
1. Time step,
Crack • Geometrical parameters
  1. Radius = 4 mm
Crack dimensions in mm
2. Total thickness = 106 mm
Crack location
Without crack With crack
Without crack With crack
 Without crack
Side view
With crack

End End End of

of End of of second
first second End of third layer first layer End of third layer
layer layer layer

• Simulation in MATLAB is successful in detecting crack

Property Measurement
  
Thermal conductivity ()

Steady state method Transient methods

Steady state method
 Heat flow method (ASTM C158)
Cold plate

Specimen
Heat flux transducer (HFT)
Hot plate

Specimen thickness and temperature range

• For 10-40 °C , up to 250 mm thickness
• For -195-540 °C, 25 mm thickness
Precision
• ± 1 % of the mean value
 Procedure

Verify thermal
Sampling of Make temp.
equilibrium by
specimen & its difference across
measuring emf
conditioning specimen (> 10 K)
output of HFT

  Make observations
Calculate until its
of mean temp. of
value falls within
specimen and temp.
0.5% of mean value
drop every 10 min
 Reference equation
●  
where
= Calibration factor of heat flux transducer, (W/m2)/V
= Heat flux transducer output, V
= Separation b/w hot and cold plates, m
= Temperature difference across specimen, K
Transient method
Hot disc method (ISO 22007-2)

Image ref. - ASTM standard

Image ref. - ASTM standard
 Hot disc method
●  
● Suitable for material having thermal conductivity = 0.01 W/mK < < 500 W/mK
● Principle

1. Specimen contains embedded hot disc probe with negligible heat capacity

2. Heat pulse produced by an electrical current produces dynamic

temperature field

3. Increase in temperature is measured as function of time

 Procedure
Assemble the
Place the probe Balance the electric
system in constant
between two half of bridge prior to the
temperature
specimen test
chamber

Measurement time Apply the heat

Calculations should be less than pulse and record
1s the temperature
 Reference equations
●  
● Temperature increase

where
● = total resistance of the leads, ohm
● series resistance , ohm
● initial resistance of the probe , ohm
● current through the sensor at the start of the transient, A
● voltage readings, V
● temperature coefficient of resistance (TCR) of the probe
 Reference equations
●  
● Temperature increase consist of two parts

Where
 is the increase in temperature over the insulating layers of the probe
 is the increase in the temperature of the specimen surface

● Solution of thermal conductivity equation is

 Reference equations
●  
● Solution of thermal conductivity equation is

where
● power output of the probe, W
● radius of the outermost ring source, mm
● thermal conductivity of the specimen material, W/m-K
● where
● Thermal diffusivity, mm2/s
● dimensionless specific time function
Heat capacity
Using differential scanning calorimetry (ASTM E1269)
● Apparatus Sample pan Reference pan

1. Differential scanning
calorimeter (DSC)
2. Weighing balance
(100mg or greater)
● Reference Material
1. Synthetic sapphire Heater
 Procedure
Put specimen & Heat the DSC at
Purge DSC with dry 20℃/min to the
reference material
nitrogen initial temperature
holder in DSC

Record steady state

Heat the DSC at Hold the DSC
isothermal baseline For
20℃/min from isothermally for at equilibrium
at upper temp.
initial to final temp least 4 min
limit

Repeat the Repeat the

procedure with procedure again Calculate heat
sapphire in the with specimen in the capacity
holder holder
 Calculations
 
where
= Heat capacity of the specimen
= Heat capacity of Sapphire
= vertical displacement between the
specimen holder and the specimen
DSC thermal curves at a given
temperature, mW
= vertical displacement between the
specimen holder and the sapphire DSC
thermal curves at a given temperature,
mW
= mass of the specimen, mg
= mass of the sapphire, mg
  Coefficient of thermal expansion ()
 Using thermomechanical analysis (ASTM E831)
● Apparatus
1. Thermomechanical Analyzer (TMA)
2. Cooling capability
3. Micrometer
● Specimen
1. Length – 2 to 10 mm
2. Lateral dimensions – less than 10 mm

Image ref. - Hitachi

 Procedure

Measure the specimen Place the temp. Move the furnace to

at 20 to 25 ℃ with sensor in contact enclose the
±25μm precision with specimen specimen holder

 
Record the change in Heat the specimen at
Calculate
length 5 ℃/min
 Reference equation
●  
Where
= specimen length
= Change in specimen length
= calibration coefficient
= change in temperature
THANK YOU
 Various methods
●
●  Steady state methods
1. Guarded hot plate method (ASTM C177)
2. Heat flow method (ASTM C158)
3. Calorimeter method (ASTM, 1993)
4. 3method
● Transient methods
1. Hot method (ISO 22007-2)
2. Laser flash technique (ASTM E1461)