Table of Contents ECNDT '98 Session: Materials characterization

Nondestructive Analysis of Stress States in Components using Micromagnetic and Ultrasonic Techniques An Overview E. Schneider Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP) Saarbrücken, Germany Email: schneider@izfp.fhg.de

Summary

Both, the electromagnetic, magneto-elastic and the ultrasonic techniques are, compared with the x-ray- or bore-hole technique, new techniques to evaluate mechanical stress states. Their application allows a fast and easy to perform locus and/or time continuous analysis of stress states in order to detect stress inhomogeneities and to evaluate stress states of large parts of a component. The fundamentals of the techniques have mainly been developed in the 70th and early 80th.
The application of electromagnetic, magneto-elastic techniques is recommended if stresses or stress gradients in surface near layers are of interest. The use of one measuring quantity allows the evaluation of stress states in parts with a controlled microstructure. The combined use of different electro-magnetic and magneto-elastic quantities enables the evaluation of stress states of parts with an inhomogeneous microstructure and the characterization of microstructural states in terms of hardness or strength values. Consequently, the main application is to characterize stress and microstructure in mechanically or thermally treated surfaces of technical parts. The set-ups available on the market mainly differ in the measuring quantities used. In all cases, the quantities have to be calibrated using samples with a known microstructural state.
The major area of ultrasonic applications covers the analysis of stress states in surface layers of about 5 mm of thickness and of stress states in the bulk of large components. The quantitative stress analysis presumes the knowledge of the material dependent elastic properties. Set-ups optimized for the evaluation of stress states of bolts and screws, of railroad wheels and rails are on the market. Different laboratory prototypes are in use to evaluate stress states of specific components.

Fundamentals

Neither the electromagnetic, magneto-elastic nor the ultrasonic technique is a just-take-it-from-the-shelf-and-apply-technique. In both cases, some preparations have to be made. The most re-cent publications offering a comprehensive overview of the electromagnetic, magneto-elastic techniques are given by Tiitto [1] and by Theiner [3]. Overview articles describing the back-ground and applications of ultrasonic techniques are written by Thompson, Lu and Clark [2] and by Schneider [4].

Since all electromagnetic and magneto-elastic quantities of a ferromagnetic material are influenced by stress and microstructure and their changes, the measuring quantities need to be calibrated. Using representative samples in a tensile test or bending experiment is the most widely used procedure. It is this procedure which is the main difference between the different available systems. The one dimensional calibration is traditionally used for the magneto-elastic Barkhausen noise measurement. The amplitude of the magnetic Barkhausen noise is measured as function of tension and compression applied to the specimen. This calibration seems to be sufficient as long as the stress in the second principal direction is less than about 25% of the elastic limit of the material. Otherwise biaxial calibrations have to be made. Fig .1 shows a three dimensional graph of a biaxial calibration surface for an isotropic material [1].

Fig. 1: Three dimensional graph of a biaxial calibration surface for an isotropic material [1]

It is clear that the application of this calibration is limited to components with the same microstructural state. A more general way of calibration includes the measurements of different quantities like Barkhausen noise amplitude, the coercivity derived from the maximum of the Barkhausen noise signal, the distortion factor of time signal of the tangential magnetic field strength, the coercivity derived from the mentioned field strength. The stress dependences of these quantities are measured using samples with different microstructural states. The cali-bration function is calculated by a multiple regression analysis. The advantage of this procedure is that the microstructural influences are taken into account. Hence the stress state can be eva-luated even if the microstructure changes along the measuring traces. Fig.2 displays a calibra-tion line for stress analysis in the hardened layer and in the base material of the same steel.

Fig. 2: Calibration function based on four different micromagnetic quantities for stress analysis in hardened layers and in the bulk of plates [5]

Ultrasonic techniques use the acousto-elastic effect which describes the influence of stress or strain states on the propagation velocities of ultrasonic waves. The effect is analytically des-cribed in terms of elastic constants of the material. Besides the well known Young´s and shear moduli, also the third order elastic constants are needed. These constants are experimentally evaluated using a representative sample.
The ultrasonic stress analysis is prepared in such a way that the elastic constants are evaluated. Using these constants, the influences of any one-, two-, or three axial stress state on the ultra-sonic velocities or times-of-flight are calculated.
In this way, wave modes which are strongly influenced by the stress state to be analyzed are identified as well as modes which are not much influenced by the stress. The last mentioned modes are used to compensate influences of the change of ultrasonic path length or microstructure. A-priori information concerning the stress states of the component of interest (e.g. one principal stress is negligible) or concerning the texture (e.g. texture is homogeneous along the measured trace) or the stress equilibrium conditions is used for simplifications of the technique and for minimization of the measuring effort.
The ultrasonic application for evaluating stress states is not limited to metals. The acousto-elastic effect, as formulated by Hughes and Kelly [6] is valid for originally isotropic materials with a cubic crystal symmetry only. But the same equations can be used to evaluate stress states in materials for which linear dependences of the relative change of ultrasonic velocities with the elastic strain or stress are measured. In fundamental investigations using monolithic ZrO₂ and Al₂O₃ ceramic samples, it was found that the sound velocities change linearly with the elastic strain. This effect is of the same magnitude in ZrO₂ as it is in ferritic steels. The stress or strain influence on the sound velocities was found to be much smaller in Al₂O₃ ceramics [4].

Disturbing Influences

The influence of the second principal stress in the surface layer on the electromagnetic and magneto-elastic quantities has to be calibrated if the stress is about 25% of the elastic limit [1]. The influence of texture on the quantities was found to be not negligible [7], hence it has to be calibrated. The temperature influence on the quantities are not yet systematically investigated.
Using the elastic constants of the material, the influence of the three principal stresses on the velocity or time-of-flight of ultrasonic waves can be taken into account. The elastic constants and hence the ultrasonic velocities are temperature dependent. Velocities decrease linearely with increasing environmental temperatures. The temperature dependences for different materi-als are published. Whereas the elastic moduli are not or neglectably influenced by the change of microstructure, the third order elastic constants show a dependence on the microstructural state. These influences are evaluated and the elastic constants of different steels and Al-alloys are published e.g. by Schneider [4]. Texture also influences the elastic constants. There are different techniques described in literature [e.g. 2, 4] to evaluate stress states of textured com-ponents, but there is not one technique which can generally be applied. Depending on the stress state to be evaluated, on the strength of texture and on the geometry of the component, different procedures are in use.

Set-Ups and Sensors

The STRESS SCAN set-ups, developed by AST [8] are the most widely used ones. The amplitude of the magnetic Barkhausen noise is measured and surface stress states are evaluated using one or two dimensional calibrations. Also the INTROMAT and the TOMOSCOP set-ups, developed at the Academy of Science of Belo Russia [9], are based on the magnetic Barkhausen noise measurement. RAILSCAN and RAILTEST, developments of the Hungarian Railroads MAV [10], and STRESSTEST, developed by METALELEKTRO at Budapest [11], use Barkhausen noise measurements in order to evaluate the stress states of rails. MUTTON and LANGMAN [12] developed a set-up to evaluate the stress states of railroad wheels by measuring the direction dependence of the magnetic properties. The MMA system, developed at the KFKI Atomic Energy Research Institute [13] at Budapest, offers the possibility to meas-ure the magnetic Barkhausen noise and the magneto-acoustic emission. Different magnetic quantities are derived from the hysteresis curve and/or from the change of the permeability with the magnetic field. The Special Power Module of the MAXWELL Equipment, developed by IXTREM in France [14], also allows the measurement of the Barkhausen noise and the incremental permeability. MAPS is a set-up developed by AEA at Harwell [15], but the author has no more information yet. The 3MA system of the IZFP [16] has been widely used to evaluate stress states in parts and components and an optional software program of the QUALIMAX set-up is developed for stress analysis [17]. In all cases a large number of sen-sors, optimized with respect to the particular measuring quantity and to the geometry of the part under test are available.

Numerous manufacturers offered ultrasonic systems to evaluate the strain or stress state of screws or bolts. The measuring quantity is the increase of the ultrasonic time-of-flight after the screw is fastened and its length is increased. Hence the principle is the measurement of the elongation rather than the use of the acousto-elastic effect to evaluate the applied or residual stress of the screw.
The ultrasonic set-ups to evaluate applied or residual stress states which are in industrial use are the DEBRO 30 [18], the DEBBIE [19], developed by Polish Academy of Sciences, the AUSTRA [20], the UER [21], the UER-T [22], the UES [23], developed by the IZFP and a system developed by SNCF [24]. AUSTRA enables the evaluation of surface and bulk stress states; all the other systems are mainly used to evaluate stress states in railroad wheels or rails. Using other sensors, all mentioned systems can also be applied for the analysis of stress states of other components. A RITEC - system [25] using the ultrasonic resonance technique is opti-mized to evaluate stress states in thin sheets. The ultrasonic sensors are usually not in stock, they are optimized for the particular application.

With respect to the numerous systems available on the market, bench mark tests to compare the accuracy, the reliability and the applicability of the systems have been suggested a few times already. As to the knowledge of the author, there is only one test of this sort made: The DEBRO 30, the UER and the set-up of the SNCF were used to evaluate the stress states in the rims of about 100 railroad wheels. The results are found to be in good agreement, the major difference between the systems was the time needed for the stress analysis and the needed handling effort.

Applications

Since both, the electromagnetic, magneto-elastic and the ultrasonic techniques presume a calibration or the evaluation of the material dependent elastic constants, the application of the techniques pays off only if numerous parts with the same microstructural state or the local or time continuous stress analysis on the same component is of interest. In order to evaluate the stress at just one position in a particular component, the application of the established x-ray- or drilling hole- or ring core technique is recommended.
As to published experiences, set-ups using the Barkhausen noise can be successfully applied to evaluate stress states in parts with a reasonably well controlled microstructural state. Typical applications cover the surface and subsurface stress states of machined, grinded and shot peened parts [1, 3].
If the microstructural state varies along the measuring trace, the measurement of different electromagnetic and magneto-elastic quantities is needed in order to separate or to discriminate the microstructural influence. Based on a comprehensive calibration, the change of the residual stress parallel to two welds is evaluated along a trace perpendicular to the welds. The result, shown in Fig. 3 is achieved using 15 quantities derived from the incremental permeability, the time signal of the tangential magnetic field strength and from the frequency dependence of the eddy current impedance. Using the same quantities, also the change of the Vickers hardness is evaluated and shown in Fig.4. The agreement with the stress values evaluated by x-ray diffraction and with the mechanically determined hardness values is very satisfying.

Fig. 3: Stress parallel to the welds versus a trace perpendicular to the two weld seams. Nondestructive evaluation using magnetic (ND) and X-ray techniques [5]

Fig. 4: Vickers Hardness along a trace perpendicular to two weld seams, nondestructively evaluated (ND) using magnetic technique and mechanically evaluated hardness [5]

BEMI [26], a laboratory prototype system, allows the stress analysis with a high lateral resolu-tion. The magnetic Barkhausen noise or the eddy current impedance can be used as measuring quantity. The change of the Barkhausen noise amplitude versus a trace across two narrow areas with high residual stresses is shown in Fig. 5 together with the result of the x-ray stress analysis [3].

Fig. 5: Residual stress profile, evaluated with BEMI using the Barkhausen noise signal (Mmax) and the results of the x-ray diffraction technique [26]

Fig. 6: SLongitudinal stress in the tread of a rail overrolled in a test stand

The application of an ultrasonic SH-wave propagating the tread of a rail is successfully used to evaluate the stress in new and used rails. In order to discriminate influences of microstructural changes along the length of the rail, a second SH-wave is simultaneously applied in the outer side of the rail head. The influence of the train traffic on the stress state was simulated in a rolling test stand. The decrease of the original tensile stress in the overrolled part is clearely to be seen in Fig. 6. The ultrasonic results are found to be in good agreement with the results of the ring core technique (RCT).

Fig. 7: Circumferential stress along the length of a hollow gear shaft before (A) and after (B;C;D) stepwise mechanical treatment to optimize the stress state

The immediate availability of the results render the electromagnetic , magneto-elastic and the ultrasonic techniques suitable for optimizations of the stress states in components. Fig. 7 shows the cicumferential stress in a hollow gear shaft before and after a stepwise mechanical treat-ment to reduce the tensile stress. An ultrasonic technique was applied for the stress analysis.

References

1. Tiitto, S.: Magnetic Methods. Handbook of Mesurement of Residual Stresses; J.Lu (ed.) Society For Experimental Mechanics; The Fairmont Press Inc Lilburn (1996) 179-224.
2. Thompson, R.B.; Lu, W.-Y.; Clark, A.V.: Ultrasonic Methods. ibid, 149-178.
3. Theiner, W.A.: Micromagnetic Techniques. Structural and Residual Stress Analysis by Nondestructive Methods; V. Hauk(ed.) Elsevier Science B.V.Amsterdam (1997) 564-589.
4. Schneider,E.: Ultrasonic Techniques. ibid, 522-563.
5. Kern, R.: Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP),
6. Hughes, D.S.; Kelly, J.L.: Second Order Elastic Deformation of Solids. Physical Review 92 (1953) 5, 1145-1149.
7. Schneider, E.; Altpeter, I.; Deryke, P.: Zerstörungsfreie Bestimmung von Texturen in Walzprodukten mit Ultraschall- und magnetischen Verfahren. Abschlußbericht zum EKGS-Vorhaben P 2007, E 1.1/88; Vertragsnummer 7210.G3/111 (1991).
8. Product information. American Stress Technologies Inc, Pittsburg, PA, USA.
9. Vengrinovich, V.L., Minsk, Belo Russia, Product information.
10. Product information. MAV KFV KFT, Budapest, Hungaria.
11. Product information. Metal Elektro, Budapest, Hungaria.
12. Langman R.A.; Mutton,P.J.: Estimation of Residual Stresses in Railway Wheels by Means of Stress Induced Magnetic Anisotropy. NDTE INT. (1993) 26, 4, 195-205.
13. Product information; KFKI Atomic Energy Research Institute, Budapest, Bulgaria.
14. Crescenzo, E.: Advanced Multi-Electromagnetic Technique. DGZfP Bericht 55, DGZfP Berlin (1996) 93-100.
15. Carter, C.; AEA Technology, Harwell, UK; e-mailed information.
16. Product information. Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP).
17. Product information. Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP).
18. Product information. Polish Academy of Sciences, Warsaw, Poland.
19. Product information. DEBRO UMS,Warsaw, Poland.
20. Herzer, R.; Schneider, E.: Instrument for the Automated Ultrasonic Time-of-Flight Mea-surement. - A Tool for Materials Characterization. Nondestructive Characterization of Materials. Höller, P.; Hauk, V.; Dobmann, G.; Ruud, C.; Green, R. (eds.) Springer Verlag Berlin-Heidelberg (1989) 673-680.
21. Herzer, R.; Frotscher, H.; Schillo, K.; Bruche, D.; Schneider, E.: Ultrasonic Set-Up to Characterize Stress States in Rims of Railroad Wheels. Nondestructive Characterization of Materials VI; Green, R. E.; Kozaczek, K. J.; Ruud, C. (eds.) Plenum Press New York (1994) 699-706.
22. Schneider, E.; Herzer, R.: Ultrasonic Evaluation of Stresses in Rims of Railroad Wheels. 7th ECNDT, Copenhagen 1998.
23. Product information; Fraunhofer Institut Zerstörungsfreie Prüfverfahren (IZFP).
24. Limal, J.L.; Del Fabro, V.: Prevention of Thermal Damage in Railroad Wheels thanks to the Monitoring of Residual Stresses by Ultrasonic Examination. Proc. 10th International Symposium Assessment of Materials Aging and Damage Evolution by Non Destructive Evaluation Methods, TOME XVI, Cercle D'Etudes Des Metaux, St. Etienne (1995) Chapter 17.
25. Fukuoka, H.; Hirao, M.; Yamasaki, T.; Ogi, H.; Petersen, G.L.; Fortunko, C.M.: Ultra-sonic Resonance Method with EMAT for Stress Measurement in Thin Plates. Review of Progress in Quantitative Nondestructive Evaluation; D.O. Thompson, D.E. Chimenti (eds.) Plenum Press New York (1993) 12, 2129-2136.
26. Altpeter, I.; Bender, J.; Hoffmann, J.; Kopp, M.: Die Barkhausenrausch- und Wirbel-strommikroskopie, eine neue Rastersondentechnik zur Werkstoffcharakterisierung im Mikrometerbereich. Prakt. Metallogr. 35 (1998) 2 , will be published soon.