·Home ·Table of Contents ·Materials Characterization and testing

Assessment of Steels ageing by Rayleigh wave velocity measurements C. Pecorari, X. Gros, B. Acosta, L. Debarberis, M. Beers and G. Manna Commission of the European Communities, Joint Research Centre Institute for Advanced Materials PO BOX 2, 1775 ZG Petten, The Netherlands Tel.: (+31) 224 565435, Fax: (+31) 224 561432 Contact

ABSTRACT

The phase velocity of Rayleigh waves propagating of the stress-released surface of thermally aged steel samples is measured by means of an interferometric technique. Results are reported which show that the Rayleigh wave velocity is sensitive to changes of the material structures brought about by the thermal treatment. Two mechanisms are offered as plausible explanations for the phase velocity change. The acoustic measurements are correlated with results obtained by other techniques.

INTRODUCTION

Prior to cracking, and possibly consequent catastrophic failure, materials usually undergo transformations that result in the degradation of their physical properties. The generation of defects such as vacancies and dislocations, or the migration of metals with low melting temperature towards grain boundaries are examples of possible mechanisms eventually leading to loss of structural integrity of engineering materials and components. Understandably, issues of this nature are of great relevance to the operation of nuclear power plants insofar as their safety depends on the assessment the ageing process occurring to the plant's structural components.
The AMES (Ageing of Materials Evaluation and Studies) program supported by the DG XII of the European Commission provides the framework to investigate, among other issues on the above-mentioned subject, the suitability of several nondestructive techniques to assess material ageing induced by both thermal and irradiation treatment [1]. In this work, measurements of Rayleigh wave phase velocity have been carried out to assess the sensitivity of this technique to the aging process induced by thermal treatment in JRQ steel samples. Two additional samples, one untreated (fresh) and treated (heat-treated) that had been especially introduced in the project to provide reference results were also tested. An attempt to provide a physical explanation of behavior of the Rayleigh phase velocity is offered. Finally, the measured values of the Rayleigh velocity are correlated to measurements of other physical properties of the same materials obtained by other techniques.

EXPERIMENTAL SET-UP

A schematic diagram shown in Fig. 1 illustrates the experimental set-up utilized in this investigation. Two surface acoustic wave transducers, operating at nominal frequencies of 2.25 MHz and 10 MHz, respectively, were used to excite Rayleigh waves in the samples of interest. The acoustic signals consisted of a train of sinusoidal oscillations containing about 15 cycles. The temporal length of the train was determined by the need to produce a wave that was as monochromatic as possible in order to avoid undesired

Fig 1: Experimental set-up, and time-domain signals.

dispersion effects. The normal particle displacement was detected by a laser interferometer, the output of which was displayed by a digital oscilloscope operating at a sampling frequency of 500 MHz. The acoustic signal was detected at two locations (P(z₁) and P(z₂) in Figure 1) along the travel path of the wave. The distance between such points was chosen to be of the order of 20 mm, and was measured with an absolute error of 1 mm. The delay of the second signal with respect to the first one was measured with a relative accuracy better than 10^-3. The central peak of the finite pulse was chosen as a reference to measure the delay. All considered, the relative error associated to the measurement system was estimated to be not greater than 10^-3.

EXPERIMENTAL RESULTS

The material used in this works is JRQ steel cut in small bars with dimension of 10 x 10 x 55 mm³. Prior to annealing for 1 hour at temperatures ranging between 400 ^oC and 750 ^oC, a series of samples was heat treated at 900 ^oC and subsequently water quenched. Of the two samples used as reference, the aged one was heat treated at 700 ^oC and then water quenched.
Figure 2 reports the values of Rayleigh phase velocity on the JRQ steel samples measured according to the procedure described in the preceding section. The results show an increase of the phase velocity with annealing temperature for the whole sequence with the exception of the last sample that was annealed at a temperature of 750^o C. Similarly, Figure 3 illustrating the results obtained on the samples named fresh and thermally treated, shows a clear difference in the Rayleigh phase velocity measured in the two samples. In both figures the vertical bars indicate the maximum variation of the measured values.


Fig 2: Rayleigh phase velocity versus annealing temperature for the JRQ samples.

Fig 3: Rayleigh wave velocity in the fresh and heat-treated reference JRQ samples.

Two phenomena, possibly concurrent, may intervene to cause the behavior illustrated by the results of Figs. 2 and 3. The first one is grain-boundary segregations, generated during quenching and diffusing more efficiently within the bulk of the grains with increasing annealing temperature. The presence segregations at the grain boundary may hinder the transfer of acoustic energy from grain to grain, causing an increase of scattering-induced attenuation and, therefore, a decrease of the phase velocity. The sudden decrease of the Rayleigh velocity value in the sample annealed at 750^o C may be explained in terms of phase transformation occurring in the bulk of the sample at temperatures above (about) 725 ^oC. The second mechanism that may be called upon to explain the increase of the phase velocity of the Rayleigh wave with increasing annealing temperature is the reduction of the average grain size. Direct observation of the microstructure of the samples annealed at 400 ^oC, 575 ^oC, and at 675 ^oC yielded estimates of the average grain size of about 62 mm, 40 mm, and 46 mm, respectively. Materials with small grains attenuate and, therefore, disperse the propagating ultrasonic wave to a lesser degree than materials with large grains. Finally, the systematically lower values of the phase velocity obtained using a transducer working at a nominal frequency of 10 MHz compared to those at 2.25 MHz, are easily understood as the consequence of the higher attenuation and dispersion suffered by the wave with higher frequency.
The results of measurements of Rayleigh wave phase velocity were correlated with those yielded by two other techniques: STEAM and the hardness measurements.
STEAM (Seebeck and Thomson Effects on Aged Material) is a nondestructive technique that utilizes the generation of a steady-state thermoelectric potential, DS, between two points of a metal between which a temperature difference, DT, is applied. The increase of thermoelectric potential due to a variation of one degree Celsius, i.e., the ratio DV/DT, that is proportional to the thermoelectric power, is a function, among other system parameters, of the temperature of the metallic sample, of its chemical composition and its atomic arrangement. In particular, changes of the crystalline lattice may modify its thermoelectric power, to the extent that thermoelectric measurements can be used to monitor changes of the material mechanical properties [1]. At the JRC-IAM a prototype system to evaluate the ageing of ferritic steels by using the Seebeck-Thomson effect has been developed [2]. The results obtained by means of this instrument on the samples investigated in this work are shown in Fig. 4. The measurements were performed at room temperature. The increase of the Rayleigh velocity correlates with a decrease of the variation of the electric potential DV produced by temperature difference of one degree Celsius, the behavior of the latter quantity being explained by the disappearance of the material defects.


Fig 4: Correlation between Rayleigh velocity values and STEAM measurements.

Fig 5: Correlation between Rayleigh velocity values and sample's hardness.

Finally, Fig. 5 illustrates the comparison between Rayleigh phase velocity values and hardness values measured on the same samples. As in the previous comparison, the increase of the Rayleigh velocity correlates with the decrease of hardness that shows the recovery of the material's original ductility.

CONCLUDING REMARKS

Measurements of Rayleigh wave phase velocity have been carried out on thermally treated JRQ steel samples. The values obtained in this work indicate that the phase velocity is sensitive to changes of the physical properties caused by the treatment. The most likely physical mechanisms responsible for the increase of the wave velocity with annealing temperature are the segregations at the grain boundaries and the decrease of the average grain size. The velocity measurements presented here correlate well with those obtained by means of other techniques on the same samples.

ACKNOWLEDGMENTS

This work has been carried out as part of the AMES-NDT project sponsored by the DGXII (Science, Research and Development) of the European Commission.

REFERENCES

1. AMES-NDT Final Report, AGE-AMES-NDT(00)-D06, February 2000.
2. R. D. Barnard, "Thermoelectricity in Metals and Alloys", University of Salford, Taylor and Francis LTD, 1972
3. B. Acosta, L. Debarberis, M. Beers and C. Mc Girl, "STEAM, A Non-destructive Evaluation for Assessing Materials Ageing", EC-JRC-IAM Petten, Proceedings 25th Annual Meeting of Spanish Nuclear Society, November 1999

© AIPnD , created by NDT.net

|Home|    |Top|