SIMULTANEOUS MEASUREMENT OF MATERIAL PROPERTIES AND THICKNESS OF CARBON STEEL PLATES USING PULSED EDDY CURRENTS

V.O. de Haan 1 and P. de Jong 2 1 BonPhysics Research and Investigations BV, Putterhoek, The Netherlands; 2 Röntgen Technische Dienst bv, Rotterdam, The Netherlands V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 1

Outline

• Introduction to RTD-INCOTEST © • Principe of new method • Measurements and results • Discussion • Conclusion V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 2

RTD-INCOTEST

© • Insulated COmponent TEST • Corrosion survey method • Ferrous objects covered with insulation, asbestos, fire proofing, concrete or coating • INCOTEST is aimed at the detection of larger area (not localised) corrosion • Pulsed eddy current technology V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 3

Transmitter coil

Pulsed Eddy Currents

Receiver coil I l d

 

object

V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 4

100

U

Pulsed Eddy Currents

(

t

) 

A



 ,  ,

l



  1 .

5

Θ

10 1 inverse power law

t -1.5

0.1

0.01

-3 dB point exponential decay 0.001

0.01

V1.0; May 27, 2004 0.1

1/3 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal

x

1 5

RTD-INCOTEST

© V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 6

RTD-INCOTEST

© Typical applications insulated vessels, storage tanks, piplines riser/caisson (splash zone) inspection rough corroded or coated surfaces V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 7

RTD-INCOTEST

© Advantages: measurement at a distance no cleaning needed fast Disadvantages: reference needed large footprint V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 8

U

Principe of new method

(

t

) 

F



 ,  ,

l

,

d

,

t



  1 .

5

Θ

 

  = permeability = conductivity

l

= lift-off

d

= thickness object

t

= time

F

= depends on geometry of probe V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 9

U F

Principe of new method



(

t

 ) ,   ,

F l

,





d

, ,

t





, 

l

, 1

d A



,

t

   ,



 

t

,  1 .

5

probe l



Θ

 



probe

   (

l

,

d

) 2   

d

2 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 10

U

(

t

)

Pulsed Eddy Currents

   1 .

5

Θ

  

 ,  ,

l

,

d

,

t



100 10

F = 1

1 0.1

0.01

0.001

0.01

F < 1 0.1

x

1 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 11

Dimensions of probe

r t2 r t1

Transmitting coil

n t

turns Receiving coil

n r

turns

l r1 l r2 l t1 l t2 r r1 r r2 d

  7.6 mm Conducting plate

Transmitter Receiver l 1 / mm l 2 / mm r 1 / mm r 2 / mm n

2.15

7.05

5.25

7.70

400 2.15

7.05

7.70

10.25

1280 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 12

Measurement example

-5 1E-5 -6 1E-6 0.001

V1.0; May 27, 2004 0.01

16th WCNDT, Aug 30 - Sep 3, 2004 Montreal

Time (s)

0.1

13

object  ms/mm 2 D2 A3 A4 D36 V1.0; May 27, 2004

Results

d mm d fit mm error %  2 0.78

18.87

12.97

6.87

19.8

13.5

7.5

6.4

3.8

21.0

0.8

0.7

0.2

2.01

1.32

1.09

19.12

13.23

7.09

18.77

13.15

6.99

18.53

12.76

6.85

21.1

14.1

9.6

18.9

15.0

10.4

20.3

13.0

7.5

9.9

4.6

4.9

47.0

16.0

9.8

4.0

4.1

8.5

16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 0.7

0.8

1.1

2.4

1.6

0.3

0.2

0.9

1.4

 % 5.0

4.4

9.3

10.5

6.2

36.0

0.9

14.1

48.4

9.6

2.2

9.5

 mm 0.9

0.6

0.6

2.0

0.8

2.6

0.2

1.9

3.4

1.8

0.3

0.7

14

Discussion

• Only for the smallest thickness a deviation occurs larger than 2 mm. With the old model and a (not) favorable reference measurement the relative deviation would be at least (60) 30 %. • As  2 varies around 1 the fit error is not determined by systematic differences but by the accuracy of the measurements.

V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 15

Discussion

• 1% of the signal is generated by eddy currents at a distance larger than 7 cm from the center of the coils. The plates dimensions were 16.5 by 16.5 cm, hence the footprint was a little larger than the object. V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 16

Discussion

• The smaller the wall thickness of the object, the less accurate the fitted wall thickness. This is due to the special form of function

F

and is a limitation of the method.

V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 17

Conclusions

1 Simultaneous fit of two time constants from the signal measured by RTD-INCOTEST © pulsed eddy current method facilitates the ( determination of both material properties  ) and absolute thickness of an object. 18 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal

Conclusions

2 Measurements show that accurate geometry is very important. The object must be larger than the (effective) footprint and the distance between probe and object must be accurately known or measured.

19 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal

Conclusions

3 Accuracy of 10% is possible and can be reduced by enhancement of probe design and measurement accuracy. Further research is needed to increase the accuracy of the method.

20 V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal

SIMULTANEOUS MEASUREMENT OF MATERIAL PROPERTIES AND THICKNESS OF CARBON STEEL PLATES USING PULSED EDDY CURRENTS

Thank you for your attention

V1.0; May 27, 2004 16th WCNDT, Aug 30 - Sep 3, 2004 Montreal 21