·Home ·Table of Contents ·Materials Characterization and testing

Characterisation of steel grades by Magnetoinductive method D. Stegemann, W. Reimche, B. Heutling, A. Krys, Institute of Nuclear Engineering and Non Destructive Testing (IKPH) University of Hannover, Elbestrasse 38A, 30419 Hannover, Germany J. Kroos, M. Stolzenberg, G. Westkämper Salzgitter Stahl AG, Salzgitter, Germany R. Angerer Voest Alpine Stahl, Linz, Austria Contact

Abstract

For the characterization of different steel types in particular tensile strength and yield strength as well as hardness are being used. These quantities are normally determined by destructive tests, which are time consuming and costly. This paper deals with a non destructive method which is able to characterize mechanical and technological properties, like tensile and yield strength, by the magnetoinductive analysis of higher harmonics. The working principle of the technique is the relation between physical properties (electric conductivity, magnetic permeability) and mechanical properties of ferromagnetic materials. The characteristic properties of materials depend, amongst others, on alloy composition, grain size, texture and dislocation density. As a consequence of this the electrical and magnetic properties are formed. For the Non Linear Harmonic Analysis (NLHA) is of particular interest the shape of the magnetic Hysteresis Loop, which reflects all the magnetic properties, as well as its variation due to changes in material properties. Due to the fact that the physical properties are strongly related to the mechanical properties (yield strength, tensile strength, consolidation, anisotropy, hardness etc.) these mechanical properties can be determined by NLHA. The measurement of the complete hysteresis loop, or parts of it, is not a very efficient way to determine mechanical properties. NLHA, instead, uses a sinusoidal signal from a coil for excitation of the hysteresis loop within the material and analyses the output signal of a receiving coil modulated by the hysteresis loop. The higher harmonics determined from the analysis contain the complete information about the shape of the hysteresis loop. These magnetic data (amplitude and phase of the harmonics) are then correlated with the mechanical data, mentioned before, which are determined from destructive test in order to calibrate the technique.

Results have been obtained from various types of steels, ranging from low strength to high strength steels, which will be presented in the paper. The accuracy for the mechanical properties is in the order of 3%. For mechanical properties more difficult to determine, like anisotropy etc., in some cases less than 3%. Regarding hardness, core hardness and surface hardness can be determined, adjusting the measuring frequency in accordance with the Skin-Effect. By variation of the measuring frequency also the hardness penetration depth can be found.

Due to its electromagnetic character the NLHA-method is very fast. Measuring times, including evaluation, are in the order of seconds. Furthermore it works without contact with the material. Distances between the sensor and the material to be tested of up to several mm can be used. In conclusion the NLHA is a reliable and fast test technique for the determination of mechanical properties of ferritic steel types. It is also suited for in-line application to control the steel production process.

Introduction.

To maintain and improve steel qualities a thorough understanding of materials properties is essential. In particular mechanical-technological properties, like tensile strength, yield strength and anisotropy are important quantities to characterize steel qualities. A strong effort is therefore devoted to the task to measure this values in an economic and fast way. Due to the fact that destructive tests are time consuming and costly a non destructive testing technique will be described here which is suited for in-line and on-line application in steel mills to measure the characteristic quantities mentioned before.
First the working principle will be outlined and then some typical results are being presented from a great number of applications from the laboratory and from the production line in steel factories. The non destructive measuring technique is called "Non Linear Harmonics Analysis (NLHA)".

How we measure.

The working principle of the NLHA is based on the magneto-mechanical analogy, schematically shown in Figure 1. . It describes the correlation between the mechanical and the magnetic moments of the elementary atomic magnets. In the macro-scopic range of materials the results of the analogy are reflected in the correlation of macro-mechanical properties of ferro-magnetic materials, like yield strength, tensile strength and anisotropy on the one hand side and magnetic properties , characterized by the shape of the magnetic hysteresis curve and the quantities describing the shape, like permeability, coercivity, remanence and saturation magnetization on the other hand. For the method a sinusoidal input signal is used to generate the magnetic field strength as shown in Figure 2. The signal is modulated by the hysteresis curve and the output signal is measured by a receiving coil. In terms of electrical engineering the hysteresis curves acts as a transfer-function of the material. The shape of the hysteresis curve is completely determinable by the higher harmonics of the receiving signal. The NLHA-technique has to be calibrated by correlating the higher harmonics to mechanical values determined by destructive tests. Typical results of this calibration procedure show, for example, that the first and fifth harmonic decrease linearly with increasing tensile strength, whereas with increasing yield strength the third and fifth harmonic decreases linearly . Analysed are all uneven harmonics up to the eleventh. Thus, having amplitudes and phases of the six harmonics already twelve parameters are available for evaluation in principle. By means of varying also the measuring frequency and magnetic field strength even more parameters are available for evaluation. If there are several measured variables which correlate linearly with a material property the latter can be evaluated by using multidimensional linear regression analysis. If the measured quantities are linearly independent from each other the results are even improved. Regarding the time needed for data acquisition and evaluation at the moment less than two seconds have been achieved per measuring point. Therefore the method is well suited for on- and in-line applications.

Fig 1: Magneto-mechanical analogy

Fig 2: Working principle of Non Linear Harmonics (NLHA)

Measuring System.

Figure 3 shows the schematic view of a system which was designed, fabricated and installed for on-line materials characterization in a production line to determine mechanical-technological quantities of steel sheets before pressing.

Fig 3: On-line measuring system for Non Linear Harmonics Analysis

What we measure.

For the characterization of steel and its quality, different mechanical-technological quantities are important. Typical examples are shown in Figure 4, which presents the stress-strain diagram of two different steel types. Given is the tensile strength (R_m), the upper yield point (R_eh), the lower yield point (R_el) and the 0.2% yield strength (R_p0.2). These quantities can be determined by destruc-tive test which are time consuming and costly. Therefore a non destructive technique has been developed to determine these quantities in a fast and reliable manner.

Fig 4: Mechanical-technological quantities of steel

Due to the production procedure of steel, in particular by rolling, the mechanical properties depend on the rolling direction, resulting in an anisotropic behavior. For many steel forming processes this is not of advantage and therefore it is also important to determine values which define the anisotropy. Examples are given in Figure 5.

Fig 5: Quantities to characterize anisotropy

Results

In this section results are given from the application of NLHA in steel works, using the above described types of system.
To establish a good calibration curve between the values measured non destructively by NLHA and the mechanical values determined from destructive tensile tests a sufficient number of specimens has to be used for a good statistical accuracy. Furthermore the samples should have a sufficient range of values in order to cover a larger area for the calibration curve. In addition it has to be considered that the specifications of the different steel types have larger or smaller tolerances according to the quality of the specific type. Also the destructive tests used for the calibration have error margins. For the results to be shown steel sheets of a main production line were chosen, where steel types of DC01 to DC05 (new European Norm) are fabricated with thickness in the range from 0.5 to 3.0 mm. The steel sheets are divided in seven groups: group 1 with thickness of less than 0.7 mm up to group 7 with thickness of more than 1.6 mm. The procedure of evaluation is the following: The measured parameters of the NLHA (amplitudes and phases of the harmonics, analysing frequency, magnetic field strength) are correlated with the mechanical value under consideration, using multidimensional regression analysis. As soon as the calibration curve is established no further destructive tests are necessary. The calibration curve, off course, is only valid for the steel type under consideration. For other steel types equivalent calibration curves have to be made.
Furthermore an automated Harmonic Analysis measuring system for use in the testing laboratory of the SZAG (Salzgitter Steel Works)was installed providing the means to turn the test specimen and thus allowing direction dependent measurements. Furthermore, the measuring system is integrated into the course of production so that the test specimen are integrally measured right before they are cut into specimen for tensile tests. Thus, it is possible to make sure that the Harmonic measuring values are gained right where the destructive tests are performed and only thereby allowing the direct correlation of destructive and non destructive measuring values.

As part of the non destructive tests with this Harmonic measuring system 1512 steel sheets of the qualities DC01, DC03, DC04, and DC05 with sheet thicknesses varying from 0.6 mm to 3.0 mm are measured and analysed. Analysis showed that the sheet thickness influences the Harmonic measuring values, especially those of the lower frequencies. To minimize this effect the measured harmonic values are divided into groups consisting of steel sheets with roughly equal sheet thickness.

Regression ana-lysis regarding the tensile strength R_m for each of those groups already showed excellent results for five dimensions. The joint presen-tation of the group results of the 5-dimensional regres-sions shows that 85% of the values deviate less than +/- 3% and 98% deviate less than +/- 6% from their corresponding de-structive characteristic value. Since the reference values are relatively equally distributed despite a certain accumulation in the lower third the reliability of the regression results are good (Fig. 6).

Fig 6: Calibration curve for tensile strength

Fig 7: Calibration curve for yield strength

Additionally, regression analysis were conducted regarding the yield strength R_p0,2 for the different groups of sheet thickness and their results are jointly shown (Fig. 7). In this case an absolute scaling is used since a relative scaling would result in a distortion of the visual results because of the lower interval bounds. Thus, it is shown that the results for R_p0,2 are nearly as good as those for R_m.

Measurements of anisotropic and isotropic steel sheets along, diagonally, and crosswise to the direction of rolling show that the difference of the harmonic measuring values of the DC04 steel sheets is readily pronounced whereas the harmonic measuring values of the isotropic steel sheets differ only marginally.

Further investigations are carried out on a rotating steel sheet. It was found that the measuring values are sinusoidal for a half turn showing that the amplitude decreases with an increasing measuring frequency (Fig. 8).

Fig 8: Measuring system for moving steel sheets

Aside from the mechanical-technological characteristic values it is important to know about the sheet steel's behaviour during forming, for example the anisotropy value r₂₀ describes the deep-drawing quality of steel sheets in the respective direction to the direction of rolling.

The implemented Harmonic Measuring system was used to investigated the de-pendencies of direction dependent mea-suring values and the respective anisotropy values of a range of steel sheets compo-sed of DC01- to DC05-qualities. The results show that a representation of the r₂₀-values by means of harmonic mea-suring values is easily possible and that the directional dependencies of the destructive characteristic values and their distinct differences in the three measuring directions are well distinguished .

Compensation of disturbance variables

In the eddy current technology the effects of lifting the coil off or tilting it against the test specimen's surface are the essential disturbance variables so that it is important to minimise them. Since the usage of air-core coils causes an exponential decrease of the higher harmonics when increasing the distance between coil and sheet surface, the higher harmonics are referred to the first harmonic of the measuring signal thus improving the formation of the higher harmonics and essentially ridding the harmonics of the lift-off effect (Figure 9).

Fig 9: Investigation of lift-off effect for NLHA

When using the Harmonic Measuring system in the inspection line more disturbance variables come into effect. It is to be expected that the most important effects will be the sheet>s velocity, steel strip fluttering as well as the tensile stress caused by the necessary front tension. Since these effects require elimination respectively compensation investigations were conducted to gather data for simulations to describe and then eliminate the said effects.

If the supporting rolls are placed rather widely in the inspection line it is possible that the front tension lessens inducing strip-fluttering during transient movements thus causing the system to become unstable. The fluttering causes stochastically changing lift-off effects between the coil and the steel sheet strip and the shown effects on the measuring values. To compensate for this disturbance variable and to enable the measuring system to integrate over a sufficient sheet volume the measuring system uses upper and lower coil-systems. This layout compensates for fluttering because a change of the distance between the strip and one coil causes an increased signal in one coil and a decreased signal in the other one so that by a suitable circuitry and little fluttering a compensation is possible (Figure 10).

Fig 10: Compensation of disturbance variables

For investigation of the disturbance variables sheet velocity and front tension suitable testing devices were designed and installed: to simulate the sheet velocity within the inspection line a steel disk is rotated in a way that it turns with a perimeter-velocity of 8m/s and by means of a sophisticated levering system it is possible to measure steel sheets under stress (see Figure 8). Since within the inspection line of SZAG into which the measuring system is to be integrated front tensions of up to 20 N/mm5 to occur, this interval was investigated. The results presented in the complex plane show that the influence of tensile stress on the harmonic values depends on the amount of stress and situation within the complex plane.

Integration of the measuring system into the inspection system

In preparation of the integration of the measuring system into the inspection system a suitable position within the inspection line was located which provides not only easy access for mounting and maintenance but also the advantage that additional rolls may be installed to minimise strip-fluttering. Additionally, a positioning device was designed in order to position the coil-systems exactly despite varying sheet thicknesses and thus minimising lift-off effects

Conclusion

The employment of the Harmonic Measuring system in the testing workshop of SZAG allowed the measurement and analysis of a large number of steel sheets in three directions of measurement. A major improvement of the correlation coefficients regarding tensile strength and yield strength as well as anisotropy was achieved by means of subdividing the data into groups of corresponding sheet thickness. Further investigations confirmed that the expected disturbance variables for the usage of the Harmonic Measuring system in the inspection line can be measured and compensated.

Acknowledgement

We acknowledge the partial support of the European Union, Commission Steel (EGKS).

References

1. Gruska, K.; Use of Harmonic analysis for inspection of ferromagnetic materials Soviet Journal NDT 19, 1983, p. 399
2. Sablik M J., Burkhardt G L., Kwun H., Jiles D C. A model for the effect of stress on the low-frequency harmonic content of the magnetic induction in ferromagnetic materials. J. Appl. Phys. 63(8), 1988
3. Stegemann D., Reimche W., Feiste K L., Heutling B. Determination of mechanical properties of sheet steel by electromagnetic techniques. Nondestructive Charaterization of Materials VIII, Plenum Press, 1998 p 269
4. Stegemann D., Reimche W.,.Krys, A., Heutling B. Feiste K. L. Kroos, J., Stolzenberg, M., Westkämper, G., Angerer, R. Characterisation of Mechanical-Technological Steel Properties by Non Linear Harmonics Analysis within the Production Line Nondestructive Charaterization of Materials IX, AIP Conf. Proc. 497, 1999, p. 196-201.