·Home ·Table of Contents ·Reliability and Validation

Eddy Current Probes Characterization By Means of a ' G. WACHE Company ALSTOM / ISS / CND / DT PH. NOURRISSON Company ALSTOM / ISS / CND / PSC TH. GARET Company ASCOT Contact

SCOPE

It has been demonstrated in the past that a « reference circuit » was able to produce eddy currents signatures similar to whose obtained with the notches used for calibrating the equipement [ 1 ], [ 2 ].

If the operator disposes of only one reference tube, with artificial defects, to characterize the probes used in association with the equipement, he can be surprised by sensitivity differences observed on special events during the test, from one probe to the other.

Among other reasons depending from the probe configuration or manufacturing process , the two main listed below can explain those differences :

• the test is generally performed by comparison between the signal providing from the unknown defect and the one providing from the reference notch used : the latter is mainly affected by the nature of the material and the accuracy of the machining process of the artificial defect.
• the coupling of the probe with the EC equipement is also a source of discrepancy : the nature of the electronic bridge used and the wire insuring the connection between the probe and the generator can affect the signal by introducing a filtering effect more or less significant.
Those injurious effects are attenuated by balancing the probe with the equipement on the reference tube but in fact they are still there, specially in absolute mode, and two probes providing from different suppliers and qualified for the process can lead to different responses.

By using a « reference circuit », it is possible to reduce significantly the first effect and to evaluate the bandwidth of the probe associated to the equipement independantly of the reference tube : this paper shows examples of characterization results obtained for condensors and steam generators axial probes.

CHARACTERIZATION OF THE PROBES USED FOR CONDENSORS

The purpose of the method proposed for axial condensor probes is to insure the suitable frequency bandwidth with the test frequencies used for the application : the method, based on a -6 dB cut off in amplitude response, is carried out with the convenient tubes in which the reference circuit is placed. The Appendix N° 1 describes by specific figures the general principle of the method.

The probe to be tested is moved at a constant speed inside the non-conductive reference tube containing the reference circuit as shown in the Figure No 1.

Fig 1: General description of the equipement used for the tests of characterization

Only one reference circuit is necessary for the bandwidth evaluation, but the tubes available on the market dispose of three reference circuits, as shown in the Figure No 2. We dispose of three circuits in order to compare more accurately amplitude and phase responses of the probe to be characterized, at the responses obtained with the qualified reference probe, for specific frequency values.

Fig 2: Cross section of an axial probe to be tested inside the non-conductive reference tube containing the three different « reference coils »

The shape of the amplitude response versus frequency, corresponding to the theoretical model, is observed in the Figure No 3 showing the upper and lower limits defined for the usefull frequency bandwidth of the probe.

Fig 3: Proposal of a -6 dB criteria for usefull frequency bandwidth of the probe

The Appendix No 2 shows the examples of experimental results obtained for different probes with different reference tubes and for two different frequency settings of the EC equipement : the Figure No 4 presents the results obtained with setting 1 for one batch of probes and the Figure No 5 presents the results obtained for another batch, in which we find the same probe reference f14,6/97-86 presented in the first batch.

Fig 4: Amplitude response of the reference coil B1 versus Frequency for four different axial probes and tubes associated with generator setting 1

Fig 5: Amplitude response of the reference coil B1 versus Frequency for four different axial probes and tubes associated with generator setting 2

The Figure No 6 gives for information, the theoretical curves obtained for three different probes with the same outside diameter but with three different coils configurations.

Fig 6: Theoretical Amplitude response of the reference coil B1 versus Frequency for three different axial probes ofconstant outside diameter

Two main observations can be done on this Appendix No 2 :

• the general shape of the experimental curves displayed in Figures No 4 and 5 are similar to the theoretical ones displayed in Figure No 6, and each probe has its specific bandwidth, which confirms the selectivity of the method proposed,
• the frequency setting 1 presents irregularities for specific frequency values, compared to the frequency setting 2, which confirms the dependancy of the EC equipement used in association with the probe (see reference f14,6/97-86).

CHARACTERIZATION OF THE PROBES USED FOR STEAM GENERATORS

The non destructive test of steam generators tubing by means of axial probes requires that each acquisition must be validated through a quality verification, based on amplitude and phase responses of the artificial events of the calibration tube.

This verification costs a lot of time, in the shop incoming process, applied for the probes ordered outside the company and we have proceeded to experiments in order to compare requirements imposed on each event of the calibration tube to the reference circuit response, for the same setting of the EC equipement, and the same scanning of the probe, as presented in the Figure No 1 of Appendix 1.

The experiment was done with two batches of qualified probes providing from two different suppliers : the Figure No 7 shows the amplitude response of the reference coil versus frequency, for two examples of those probes providing from suppliers A and B. One can observe on this figure significant differences especially for the two different frequency values reference F2 = 280 kHz and F3 = 120 kHz.

Fig 7: Amplitude response of the reference coil B1 versus frequency for two different suppliers of process

The Figure No 8 presents the correlation observed between the amplitude response of the 7 events of the calibration tube and the amplitude response of the reference coil, for a batch of 17 probes providing from the same supplier A, and for the same gain and frequency settings of the EC equipement. One can observe for each event a linear response with a dispersion around the straight regression line generally low, excepted for the event reference VD. We can notice also on this figure the principle of the threshold defined for acceptance of the probes, tested according to this method, during the shop incoming process.

Fig 8: Amplitude response of the reference coil B2 versus amplitude response of the 7 events of the calibiration tube : selection principle

The Appendix No4 presents the results obtained for the specific event reference VD with the 2 batches of probes tested providing from the 2 different suppliers (17 probes from supplier A and 10 from supplier B). The Figures No 9, 10 and 11 present respectively the results obtained for the 3 frequency values F1, F2 and F3. One can observe on those figures two different behaviours related to the two different suppliers, which confirm the ability of the method to separate the batches and explain some significant differences in sensitivity.

Fig 9: Amplitude response of the reference coil B1 versus amplitude response of the event VD for the frequency mode F1D

Fig 10: Amplitude response of the reference coil B1 versus amplitude response of the event VD for the frequency mode F2D

In the same way, the Appendix No 5 presents the results obtained for 3 different events with the same 2 batches of probes as previously, and the same frequency setting F2. The figures Nows 12, 13 and 14 present respectively the results obtained for the 3 different event references GE 10%, 4 PHI 1 and GE LAR. We can observe that the value of RČ, reflecting the quality of the regression law, is closer to 1 for supplier B than supplier A, which confirms the ability of the method to quantify and classify the origin of the probes.

Fig 11: Amplitude response of the reference coil B1 versus amplitude response of the event VD for the frequency mode F3D

Fig 12: Amplitude response of the reference coil B1 versus amplitude response of the event GE 10% for the frequency mode F2D

Fig 13: Amplitude response of the reference coil B1 versus amplitude response of the event 4PHI 1 for the frequency mode F3D

Fig 14: Amplitude response of the reference coil B1 versus amplitude response of the event GE LAR for the frequency mode F2D

CONCLUSION

For the condensor tubing, a method based on the evaluation of the bandwith at - 6dB cut off is proposed, in order to verify that the probe meets the frequency values required for the test : the results obtained for several probes have confirmed that some probes were not suitable for the specific application required.

For the steam generator tubing, the correlation have confirmed that the amplitude response of the reference circuit was significant of the amplitude response of all the events of the calibration tube presently used and that it was possible to quantify and explain the differences observed on the two batches providing from two different suppliers of probes.

REFERENCES

1. SURECA- Système Unique de Référence et d'Evaluation des Capteurs utilisés en Courants de Foucault. G. WACHE, Revue Pratique de Contrôle Industriel No 176 bis (Octobre 92) p. 52-59
2. Caractérisation des capteurs utilisés en contrôle non destructif par Courants de Foucault : aspects théoriques et expérimentaux, G. WACHE - J. JARDET - R. LINK, 6^ème Conférence Européenne sur les Contrôles Non Destructifs. NICE 94

© AIPnD , created by NDT.net

|Home|    |Top|