·Home ·Table of Contents ·General

Barkhausen noise at deep drawing steel after various degrees of cold plastic deformation J. Grum, B. Pecnik, Faculty of Mechanica Engineering, Aškerceva 6, 1000 Ljubljana, Slovenia D. Fefer Faculty of Electrical Engineering, Trzaška 25, 1000 Ljubljana, Slovenia

Abstract:

The magnetic Barkhausen effect is a phenomenon occurring during magnetisation of ferromagnetic materials and related to movements of magnetic domains under the influence of an alternating magnetic field. The Barkhausen noise is a collection of all information on a material; therefore, voltage signals should be captured, processed an compared to equal signals from the materials having specified properties. In the investigation, a deep-drawing steel Fe 360-B according to ISO standard in the soft condition, in the form of plate and strips having various initial thicknesses, which were subsequently cold rolled to a uniform thickness, was used. The study is based on an analysis of the voltage signal of the magnetic Barkhausen noise by means of its maximum amplitude, an analysis of its power spectra, and a comparison of the voltage-signal power. Direct determination of the degree of cold deformation can be performed by means of the calibration curve or by appropriate software support which transforms the signal obtained directly into the material property required.

1.0 Introduction

The existence of magnetic domains in a ferromagnetic material was first theoretically predicted by Weiss in 1907 so that this theory was later named after him, i.e., Weiss theory [1, 2]. The first experimental confirmation of the existence of magnetic domains was provided in 1919 by Professor Barkhausen. In magnetisation of ferromagnetic materials, changes in the direction of the Weiss domains occur in shocks, which produces, with the increase in field intensity, the material a step-like increase in magnetic flux density. The increase in magnetic flux by shocks induces voltage shocks in the measuring coil. Barkhausen detected induced voltage shocks by a measuring coil, amplified them and lead them to a loudspeaker. The internal magnetisation by shocks of the ferromagnetic material was registered by the loudspeaker as rustling or crepitating. Later researchers referred to this noise as Barkhausen noise (BN), the term which is found in the technical literature. It is characteristic of the present state of measuring and processing techniques that the captured voltage signals may be registered by a measuring (detecting) coil and then subjected to on-line processing; the magnetic method based on the Barkhausen noise (BN) is, therefore, very rapid, sufficiently reliable and as such suitable for control of materials in automated systems. The characteristic Barkhausen noise (BN) will confirm that a process of material magnetisation includes a number of small steps with an increase in magnetic flux density, which are a result of changes in magnetic domains of the order of magnitude of 10^-8 cm³ [2]. The magnetic method based on the Barkhausen noise (BN) permits, among other things, non-destructive testing of various properties of materials such as influences of grain size, microstructure, material hardness, degree of cold deformation, and density of dislocations. The method is suitable also for determination of size and variation of load tensions in various microstructures as well as of residual stresses in a thin surface layer of the material after machining and/or heat treatment [3-5].

2.0 Experimental procedure

2.1 Preparation of a specimen material
Specimens were cold rolled from different initial thicknesses of a common structural steel Fe360-B in accordance with ISO standard to three equal final thicknesses. The initial dimensions of the specimens were 300 mm x 40 mm with different thickness which was obtained by means of preliminary machining. The rolled pieces were rolled in the longitudinal direction. By cold rolling they were thinned to uniform final thicknesses, i.e., of 2.0 mm, 3.5 mm, and 5.0 mm respectively, and thus attained different degrees of cold deformation, i.e., of 20 %, 40 %, and 60 % respectively. The initial material in a soft state shows fine-grained pearlitic-ferritic normalised microstructure.

3.0 Experimental results

3.1 Hardness analysis of the specimen material
Fig. 1 shows measured Vickers hardness of the specimen material as a function of the final specimen thickness and the degree of cold deformation of the specimen material. It shows hardness variations for the final sheet thicknesses, i.e., 2.0 mm, 3.5 mm, and 5.0 mm, with 0 %, 20 %, 40 %, and 60 % degrees of cold deformation. The values of the average hardness of the individual specimen material are in the range expected with regard to the chemical composition of the steel and the given microstructure with the individual degrees of cold deformation of the common structural steel Fe360-B.

Fig 1: Variations of HV1.0 hardness as a function of the degree of cold deformation of steel with different final material thicknesses.

Prior to cold deformation, i.e., in the soft state, individual specimens showed hardness ranging from 152 HV_1.0 to 164 HV_1.0, which, with an increase in the degree of cold deformation, increased to 277 HV_1.0. Fig. 1 indicates that the highest hardness values were obtained with the final specimen thickness of 2.0 mm, a bit lower values with the final thickness of 3.5 mm, and the lowest values with the material thickness of 5.0 mm. This is to say that homogeneity of cold deformation decreases with an increase in the specimen thickness. With smaller material thicknesses the material becomes more cold hardened at the upper and lower specimen surfaces than in the central part.

3.2 Experimental set-up with a sensor unit
The experimental set-up used to capture the induced voltage consists of four basic units:

• a power unit for magnetisation consisting of a function dynamic generator, a power amplifier, a slide resistance and a magnetising yoke with turns;
• a unit for detecting Barkhausen noise (BN) voltage signals consisting of a search coil, i.e. detecting coil, a signal amplifier and a band-pass filter;
• a unit for display of the captured voltage signals consisting of a personal computer including an oscilloscope card SCLITE 1.5 of National Instruments;
• a unit for processing, evaluation and analysis of voltage signals of the Barkhausen noise by means of LabView software.

The laboratory sensor unit consists of a magnetising yoke with 1250 turns used to apply an alternating magnetic field to the ferromagnetic specimen material. Good contact between the magnetising yoke and the specimen surface may be achieved by an appropriate preliminary preparation of the specimens by grinding and calibration rolling. Fig. 2 shows the sensor unit consisting of three major parts, i.e., the magnetic yoke with turns, the detecting coil and components for connecting a power amplifier and BNC terminal for transmission of the captured BN voltage signal to the amplifier, the band-pass filter and finally the unit for display of the captured voltage signal.

Fig 2: Sensor unit with yoke for the specimen magnetisation and capturing of voltage signals

The sensor unit consists of the magnetising yoke with turns and with the detecting coil designed for capturing the BN voltage signal. The detecting (measuring) coil is the most important element of the experimental set-up since it is on the detecting coil that the quality of the captured voltage signal depends, which may have a decisive influence on preliminary processing and processing of the voltage signals. The detecting coil consists of 40 turns with a thickness of 0.08 mm arranged in one winding as a spiral coil.

4.0 Processing of voltage signals

For the processing and analysis of the BN voltage signal, the slice uniformity, i.e., a unit of time record of the signal regardless of the type of sensor unit, the selected excitation and sampling frequencies are the most important. Thus BN voltage signals occurring with a negative slope of a magnetisation curve in the vicinity of the transition of the time x-axis were selected for the analysis (Fig. 3). The highest amplitude of the BN voltage signal was attained with the strongest changes in magnetic flux density. Efficient evaluation of the state, i.e. specimen-material properties, may be expected if the following are selected:
• optimum magnetising parameters,
• optimum size and shape of the detecting coil,
• optimum sampling frequency.

Fig 3: Size of the slice of BN voltage signal on time axis.

For an efficient analysis of BN voltage signals, the most suitable magnetising-current intensity and frequency for external magnetisation should be ensured; therefore, very extensive preliminary studies were performed. The results indicated that the most distinctive BN voltage signals are obtained with a current intensity of 0.5 A and an excitation frequency of 10 Hz. A sampling frequency of 50 kHz was used. Analyses of the captured voltage signals were made in terms of four characteristics, i.e.:

• determination of the maximum amplitude of the voltage signal value,
• calculation of the variance of the voltage signal,
• calculated voltage-signal power,
• calculated value of V_rms of the voltage signal.

4.1 Comparison of measurements of the BN voltage signal with a constant magnetising current and a constant magnetic flux
Fig. 4 shows calibration curves during the phase of cold deformation of the material as a function of time delay of the voltage signal, the calculated signal power, the signal variance, and the value of V_rms of the BN voltage signal in the case of magnetisation of the specimen material with the constant magnetising current and the constant magnetising flux.

Fig 4: Calibration curves for degrees of cold deformation with constant magnetising current and constant magnetising flow for a sheet with a final thickness of 3.5 mm.

The calibration curves in Fig. 4 show the dependence of the degree of cold deformation from the time delay of the voltage signal, the calculated signal power, the signal variance and value of V_rms of the BN voltage signal captured at the final material thickness of 3.5 mm. The emphasis is laid on the plotted calibration curves as well as their comparison with the constant magnetising current or constant magnetising flux. With the curves of linear approximation of the calibration curves, also an equation is given which defines the relationship between the degree of cold deformation and, indirectly, the specimen material hardness. The equation of a calibration curve is determined directly as a result of the characteristic searched for from the captured BN voltage signal. A deficiency of the calibration curves determined in this way is that they are valid only for the given material and the given state of the same specimen material.

The experimental results shown in Fig. 4 permit a conclusion that the voltage signals captured with the constant magnetising current provide more useful data than those captured with the constant magnetising flux. That is to say that a strong enough change of the measured as well as calculated parameters of the voltage signal was sensed with the selected degrees of cold deformation of the specimen material. The experimental results indicate that the time delays give the highest value of correlation coefficient. Then follow the calculated value of V_rms, the variance, and the voltage-signal power. If the results of V_rms are compared with the results of variance and the voltage-signal power, then the inclination of the straight line and the value of correlation coefficient are to be considered in the selection of a suitable comparative curve. The data obtained from the voltage signal indicate that it is better to select the variance or the voltage-signal power as a feature due to a steeper inclination of the straight line in spite of a poorer correlation. It turned out anyway that magnetisation with the constant current is more reliable for the prediction of the degree of cold deformation. The calibration curves between the degree of cold deformation and the calculated feature of V_rms indicate that the state of the stressed structural steel Fe360-B is satisfactory.

The repeatability of measurements of the BN voltage signal was statistically verified by means of the single-factor analysis of the variance. Based on the single-factor analysis of the time delay variance, the variance, and the calculated voltage-signal power and the value of V_rms of the BN voltage signal, it can be confirmed that the differences are significant and that by means of the selected procedure and the selected characteristics of the voltage signal, the degree of cold deformation may efficiently be predicted or the microstructure, i.e. the material hardness attained, may be assessed.

5.0 Conclusions

The results obtained confirm the applicability of the micromagnetic NDT method for the assessment of the material state. The initial tests showed that it is very important to select an appropriate detecting (measuring) coil as well as appropriate magnetising parameters. For the analysis of the voltage signal, a uniform time slice of the voltage signal should be selected. With the selected features it requires suitable processing and final assessment of the results with regard to the calibration curves with the highest correlation coefficient with different degrees of material hardening. The BN voltage signals captured with the constant magnetising current are in the cases given more reliable than the voltage signals captured with the constant magnetising flux. After a comparison of values of individual features and their correlation coefficients with regard to the degree of cold material deformation it was found that the micromagnetic method including magnetisation with the constant magnetising current is more reliable and applicable since a steeper inclination of the straight lines between feature and the degree of cold deformation and a better correlation coefficient are obtained.

6.0 References

1. R. M. Bozorth: Ferromagnetism, IEEE Magnetics Society Liaisom to IEEE Press, Roger F. Hoyt, IBM Almaden Research Lentck, 1-13, 476-554, 811-837,1993.
2. Chikazumi S.: Physics of magnetism, Ed.: S. H. Charap, John Wiley Sons, 1964, 3-59, 111-127, 161-185
3. Tiitto K. : Use of Barkhausen Effect in Testing for Residual Stresses and Material Defects, Non-Destructive Testing, Australia, Vol. 26, No. 2, 1989, 36-41
4. Mitra A., Govindaraju M.R. and Jiles D.C.: Influence of Microstructure on Micromagnetic Barkhausen Emissions in AISI 4140 Steel, IEEE Transactions on Magnetics, Vol. 31, No. 6, 1995, 4053-4055
5. Tonshoff H.K., Regent C., Karpuschewski B.: Intelligent Sensor Systems to Avoid Thermal Demage in Grinding, 1 st International Machining and Grinding Conference, Michigan, September, 1995, 95-182
6. http://www.stresstech.fi/BarkNoisAnal.html#top
7. Swartzendruber L.J., Hicho G.E. : Effect of Sensor Configuration on Magnetic Barkhausen Observations, Res. Nondestructive Evaluation, New York, ZDA; 1993, 42-50

© AIPnD , created by NDT.net

|Home|    |Top|