More info a
Recent advances in Pulsed Eddy Current inspection
of corrosion under insulation near pipe flanges
Vincent Demers-Carpentier, Maxime Rochette, Joël Crépeau,
Florian Hardy, Marc Grenier, Charles Tremblay,
Marco Michele Sisto, Martin Turgeon
Presenter: Vincent Demers-Carpentier
Content
• Pulsed Eddy Current (PEC) Working Principle
• Signal representation and interpretation
• Limitations of PEC
• Compensated wall thickness evaluation
• Mass effect correction
• Conclusion and Future development
NDT in Canada 2017 Conference (June 6-8, 2017)
PEC Working Principle (1/3)
PEC consists in the analysis of the transient eddy current inside a conductive
component following a sharp electromagnetic transition. There are 3 phases:
1. The emission phase (the pulse) during which the probe injects magnetic fields
that penetrate and stabilize in the component thickness
Time
Excitation pulse
Time
Magnetic response
NDT in Canada 2017 Conference (June 6-8, 2017)
PEC Working Principle (2/3)
PEC consists in the analysis of the transient eddy current inside a conductive
component following a sharp electromagnetic transition. There are 3 phases:
2. The cut-off phase which induces strong eddy currents into the component
when the magnetic field emission is stopped abruptly
Time
Excitation pulse
Time
Magnetic response
NDT in Canada 2017 Conference (June 6-8, 2017)
PEC Working Principle (3/3)
PEC consists in the analysis of the transient eddy current inside a conductive
component following a sharp electromagnetic transition. There are 3 phases:
3. The reception phase during which magnetic sensors measure the decay of the
magnetic field as eddy currents diffuse into the material thickness
3
Time
Excitation pulse
3
Time
Magnetic response
NDT in Canada 2017 Conference (June 6-8, 2017)
A-Scan in Reception Phase
Thinner wall thicknesses change the shape of the A-scan
• Shorter eddy current diffusion time
• Quicker signal drop in a Log-Log scale
1
2
• Different slope in a Log-Lin scale 3
4
Log-Log Scale Log-Lin
Log-Lin ScaleScale
Voltage - Log scale (v)
Voltage - Log scale (v)
Voltage - Log scale (v)
0 50 100 150
Time - Linear scale (ms)
1 10 100 0 50 100 150
Time - Log scale (ms) Time - Linear scale (ms)
NDT in Canada 2017 Conference (June 6-8, 2017)
Applications Suitable for PEC
NDT in Canada 2017 Conference (June 6-8, 2017)
What Pulsed Eddy Currents do well
Detection of corrosion in presence of high lift-off which can come from:
• Protective coating
• Insulation
• Corrosion product
• Marine growth
• Concrete
• Repair wrap
Penetrates thick wall
Works through weather jacket and / or
thin metallic coating
NDT in Canada 2017 Conference (June 6-8, 2017)
Main Limitations
Minimum detectable volume loss is fairly large
• Current probe detect FP volume
volume loss that covers
about 15% of the
probe footprint Volume of smallest
detectable defect
Provide an average WT measurement within the footprint of the probe
• Depth sizing underestimation for small indication
NDT in Canada 2017 Conference (June 6-8, 2017)
Impact of the Average WT Measurement
Large corrosion Small corrosion
(larger than averaging area) (smaller than averaging area)
Average WT Average WT
Good sizing accuracy Undersizing of the flaw!
NDT in Canada 2017 Conference (June 6-8, 2017)
Example of Flaw Undersizing
Lab mockup sample – Flat bottom holes
Plate WT 0.5’’ (12.7 mm)
A B
Insulation height 2’’ (50.8 mm)
Average
Defect Diameter Real WT
WT
A 3’’ 66% 89.5% C
B 6’’ 33% 66.8%
C 3’’ 33% 85.7%
NDT in Canada 2017 Conference (June 6-8, 2017)
A-scan from Defect Smaller than Footprint
Actual Signal
Signal - Log Scale (V)
2 12 22 32 42 52
Time - Lin Scale (ms)
NDT in Canada 2017 Conference (June 6-8, 2017)
A-scan from Defect Smaller than Footprint
Actual Signal = contribution
Actual Signal of Nominal
Signal - Log Scale (V)
2 12 22 32 42 52
Time - Lin Scale (ms)
NDT in Canada 2017 Conference (June 6-8, 2017)
A-scan from Defect Smaller than Footprint
Actual Signal = contribution
Actual Signal of Nominal + Defect
Signal - Log Scale (V)
2 12 22 32 42 52
Time - Lin Scale (ms)
NDT in Canada 2017 Conference (June 6-8, 2017)
Compensated Wall Thickness
Goal: Isolate the defect contribution from the signal
How it works:
• Analyze a defective region rather than a single point
• Fit an analytical equation on
each data point
• Find a defect contribution ratio
• Calculate a compensated WT
NDT in Canada 2017 Conference (June 6-8, 2017)
Results of CWT
Lab mockup sample – Flat bottom holes
A B
Plate WT 0.5’’ (12.7 mm)
Insulation height 2’’ (50.8 mm)
C
Real Average Comp.
Defect Diam.
WT WT WT
A 3’’ 66% 89.5% 67.1%
B 6’’ 33% 66.8% 36.7%
C 3’’ 33% 85.7% 39.8%
NDT in Canada 2017 Conference (June 6-8, 2017)
Mass effect
Mass effect is a strong signal contribution from large metallic masses near the probe
• Pipe saddles • T-pipes
• Supports • Nozzles
• Welded I-beams
NDT in Canada 2017 Conference (June 6-8, 2017)
Mass effect
When the probe approaches a mass, the strong and slow mass signal hides the signal from the pipe
1. Pipe wall: Far + Thin
→ Weak, fast-decaying signals
2. Flange: Close + Thick Insulation
→ Strong, slow-decaying signals
3. Within one FP distance, flange 2
signal hides pipe signal
Flange
1
Pipe wall
This results in an increasing
measured wall thickness.
3
NDT in Canada 2017 Conference (June 6-8, 2017)
A-scan from mass effect
New dominant contribution from mass
Signal - Log Scale (V)
CWT is blinded by mass
signal unless a proper
correction is applied
2 12 22 32 42 52
Time - Log Scale (ms)
NDT in Canada 2017 Conference (June 6-8, 2017)
CWT with mass effect correction
Concept of mass effect correction:
1. Identify a “reference line” mainly influenced by nominal
and mass
2. Analysis of this reference line allows to estimate the
contribution of the mass only
3. Apply CWT technique by including the mass contribution
in the fitting procedure
NDT in Canada 2017 Conference (June 6-8, 2017)
Example: small defect near flange
7in OD, STD pipe
with small defect milled at 1in from flange
Flange
• Average measured WT before compensation is
heavily undersized as this defect is much
smaller than the probe footprint
• Uncorrected CWT is blinded by mass
Defect barely visible here
• Correction allows for accurate defect sizing
Compensated Compensated
Insulation Defect Real
Defect length Average WT WT without WT with
thickness width WT
correction correction
2 inch 2 inch 0.75 inch 24% 96% Blinded 26.1%
NDT in Canada 2017 Conference (June 6-8, 2017)
Conclusions and future developments
PEC is an efficient screening technique to detect corrosion in several CUI
applications.
The Compensated WT algorithm addresses the main weakness of PEC:
undersizing of small flaws
The new correction for mass effects enables CWT near metallic
masses like flanges, nozzles etc.
Future development:
• Reduce footprint size
• Improve A-scan analysis to further enhance sizing of defects smaller than
footprint
• Further improve robustness and accuracy of CWT in presence of mass effects.
NDT in Canada 2017 Conference (June 6-8, 2017)
