WO2013027074A2

WO2013027074A2 - Magnetometer, method for its operation and non-destructive material testing apparatus

Info

Publication number: WO2013027074A2
Application number: PCT/HU2012/000077
Authority: WO
Inventors: Antal Gasparics; Tibor Farkas; János SZÖLLÖSY; Róbert György VÁMOS
Original assignee: ARACONSYS Kft.
Priority date: 2011-08-23
Filing date: 2012-08-23
Publication date: 2013-02-28
Also published as: WO2013027074A3

Abstract

The invention is a magnetometer comprising a sensor element (20), a drive means (31) operating the sensor element (20) and a processing means (34) processing the response signal of the sensor element (20), the sensor element (20) comprising a driving coil, and an iron core suitable to be driven into magnetic saturation by the driving coil. In the inventive magnetometer the drive means (31) is adapted to supply a constant voltage of varying polarity as a drive voltage to the driving coil, and the processing means (34) is adapted to detect surges of a current of the driving coil appearing at reaching iron core saturations, and to provide information characterising a detected magnetic field strength on the basis of positions in time of said surges. Furthermore, the invention is a method for operating the magnetometer and a non-destructive material testing apparatus using the magnetometer.

Description

MAGNETOMETER, METHOD FOR ITS OPERATION

AND NON-DESTRUCTIVE MATERIAL TESTING APPARATUS

TECHNICAL FIELD The invention relates to a single coil fluxgate magnetometer, a method for the operation thereof, and a material testing apparatus realized with the single coil fluxgate magnetometer.

BACKGROUND ART

In the special technical field of detecting and measuring magnetic fields, an element having coils and responding to an external magnetic field with a change in voltage/current is called a sensor or a sensor element. A combination of the sensor element, a control/drive unit required for its operation and a unit processing the response signal of the sensor element is called a magnetometer or a sensor module.

Magnetometers suitable for measuring relatively weak magnetic fields, and for detecting small changes in the magnetic field strength are known in the prior art. One group of such magnetometers is called 'fluxgate' magnetometers.

The basic operating principle of conventional fluxgate magnetometers enables the detection of a weak H field exhibiting a much lower magnetic field strength than required for the saturation of prior art iron cores (i.e. much smaller than the H_s field in Fig. 3) and (because the sensor body cannot pass through the medium boundary, instead of the H field, for the sake of simplicity, the induction density actually existing in the air, i.e.) the detection of the associated B field by using a driving field which is able to create magnetic saturation in the applied iron core. The iron core used in the fluxgate magnetometer is magnetised simultaneously by the driving field and the external field to be measured. Since the driving field is symmetric from the aspect of the origin of the iron core B-H characteristics, the added external magnetic field of given orientation disturbs the symmetry of periodic magnetisation, and this enables the detection of the symmetry destroying field.

In the case of a conventional fluxgate magnetometer, the driving field is a single frequency (i.e. sinusoidal) time function with such an amplitude that the iron core of the sensor stays within the linear section of B-H characteristics even when the field strength is the highest. Therefore, if there is no external field, a linear system is obtained, and the magnetisation in the iron core, i.e. the B field, exhibits a sinusoidal character inducing in the sensor coil of the iron core a sinusoidal voltage of an identical frequency with that of the driving field. When there is an external field, the iron core leaves the linear section during the magnetisation cycle, and the system becomes non-linear, as a result of which even harmonics appear in the sensor coil. If the external field to be measured is re-compensated by a given compensating current to the sinusoidal current of the driving coil, the linear operation can be restored, i.e. the upper harmonic signal appearing in the sensor coil can be eliminated and the field corresponding to the current necessary for compensation will be identical with the external field to be measured.

So-called 'pulse-position' fluxgate magnetometers are also known in prior art. The operation of a pulse-position fluxgate magnetometer can be deduced from the operation of a conventional fluxgate magnetometer. When a triangular function is applied instead of a single frequency, i.e. a sinusoidal driving field, the operating principle of these magnetometers becomes evident. A possible variant of a sensor element 10 of such a magnetometer is shown in longitudinal section in Fig. 1. An iron core 12 is arranged in a coil body 11. In the prior art variant of the so-called double coil design, a driving coil 13 and a sensor coil 14 are arranged on the coil body 11. In the depicted example, the iron core 12 is made of oblong cross section metal glass fibre. As shown in Fig. 2, according to the sectional view along the plane A-A of Fig. 1 , for the purpose of reducing the clearance between the driving coil 13 and the iron core 12, the driving coil 3 is mounted on the oblong shaped coil body to match the cross section of the iron core 2.

Assuming that the iron core 12 has an idealised so-called Z magnetising characteristics as shown in Fig. 3, in the case of a triangular time function l_E drive current, as shown in Fig. 4, the voltage signal U| induced in the sensor coil 14 in every half period will be the resultant of two square pulses: in the coils without an iron core 12 the sum of the reference signal proportional to the first time derivative of the field (the lower pulse in the figure) and the pulse generated between two saturations as a result of the magnetising of the iron core 12 (the higher pulse in the figure).

If there is an external magnetic field to be measured (H_M) in an identical orientation with the axis of the driving coil 13, the symmetry is destroyed. In one half period of excitation the external field accelerates and in the other half period delays the magnetisation, and therefore the second pulses induced as a result of the magnetisation of the iron core are shifted in time in the direction of arrows shown in Fig. 4, compared to the phase of the triangular driving field. The shifting of positive and negative pulses takes place in an opposite orientation. The extent of this pulse shift (if the triangular time function of the driving field consists of linear sections) is proportional with the external field to be measured.

Hence, in the case of pulse-position fluxgate magnetometers, it is not the actual shape of the induced voltage signal in the sensor coil 13, but only the phase position of the voltage pulses generated by the iron core 12 which carries the information concerning the external field. The phase modulation is much more favourable from the aspect of noise suppression than a measurement based on the conventional fluxgate amplitude modulation (the magnitude of the second upper harmonic).

Such a pulse-position magnetometer is described by way of example in US 7 378 843 B2. It is a disadvantage of this known solution that separate driving coil and sensor coil are needed for implementing the measurement. This precondition increases the inaccuracy of measurement and the uncertainty stemming from production tolerances on the one hand, and on the other limits the size reduction of sensor elements in such magnetometers.

In US 3 875 502, a magnetometer applied for material testing is described, which comprises a magnetic field generated by an AC electromagnet, a work piece moved across this field perpendicularly to the magnetic lines of force, and a coil arrangement which senses in the work piece the eddy currents so generated. By this prior art solution, conclusions are drawn about the presence of material defects in the work piece, on the basis of detecting the eddy currents. This known solution is also characterised by the disadvantages described above. DESCRIPTION OF THE INVENTION

An object of the invention is to provide a magnetometer which is exempt from the disadvantages of prior art approaches or substantially mitigates these drawbacks. It is an object of the invention to provide a magnetometer which only comprises a single coil on which both the driving and the detection are realizable. A further object of the invention is to provide a magnetometer to be produced at a low cost and by simple means, which by the simple design of its sensor element offers high accuracy and measuring safety. An object of the invention is furthermore to provide a method of operating the magnetometer, which ensures simplicity and a high measuring accuracy thanks to the single coil design. A further object of the invention is to provide a material testing apparatus, which is able to detect surface or internal material defects by non-destructive methods with a high accuracy and reliability.

The objects of the invention have been achieved by the magnetometer defined in claim 1 , the method defined in claim 6, and the material testing apparatus of claim 9. Preferred embodiments of the invention are defined in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the invention are described below by way of example with reference to the following drawings, where

Fig. 1 is a longitudinal section of a prior art sensor element,

Fig. 2 is a sectional view taken along plane A-A of the sensor element shown in Fig. 1 ,

Fig. 3 is a chart of ideal magnetic characteristics of an iron core,

Fig. 4 is a time diagram showing operating signals of the sensor element depicted in Fig. 1 ,

Fig. 5 is a time diagram showing the operation of a single coil sensor element according to the invention,

Fig. 6 is a perspective view of the sensor element according to the invention,

Fig. 7 is a block diagram of a preferred magnetometer according to the invention, Fig. 8 is a view of a sensor surface of a probe based on the magnetometers according to the invention,

Fig. 9 is a magnified view of section B of Fig. 8, Fig. 10 is a block diagram of a central signal processor in a material testing apparatus according to the invention,

Fig. 11 is a schematic view of a rolling process implementation, one of the preferred areas of non-destructive material testing,

Fig. 12 is a perspective view of an examination of a plate produced by rolling in accordance with Fig. 11 ,

Fig. 13 is a side view of the arrangement shown in Fig. 12,

Fig. 14 is a perspective view of the arrangement shown in Fig. 12, with the covering of a probe removed,

Fig. 15 is a side view of the arrangement shown in Fig. 14,

Fig. 16 is a view of the rolling unit providing a constant distance from the plate to be examined, in the arrangement shown in Fig. 12,

Fig. 17 is a perspective view of the unit creating the magnetic field in accordance with Figs. 12 to 16,

Fig. 18 is a view of an inclusion used by way of example in the examined plate, and Fig. 19 is a time diagram of the measured values in association with the detection of the inclusion shown in Fig. 8.

MODES FOR CARRYING OUT THE INVENTION

One of the ideas of the invention is that instead of driving the driving coil with a predetermined current function and instead of detecting the phase of voltage pulses appearing on the sensor coil, a single coil is applied, i.e. the driving coil also serves as a sensor coil. To this end, the driving coil is driven with a predetermined voltage function, and the shape and phase of the current function appearing on the same driving coil are detected. In such a way, the structure of the pulse-position fluxgate magnetometer can be simplified, which results not only in reduced size but also in other technological advantages stemming from the simplification of production.

The operating principle of the magnetometer according to the invention is shown in Fig. 5. Assuming that the iron core has ideal Z shaped magnetising characteristics mentioned above, the driving coil and the sensor coil can be combined in a way that a constant drive voltage 1½ of varying polarity is fed to the driving coil. When constant voltage is switched to the unit, an excited current li increasing with a constant rate appears in the driving coil, until the magnetisation of the iron core reaches the level of saturation. At that point the impedance of the driving coil drops, which results in a sudden surge (increase in the rate of rise) of the current. The saturation of the iron core can be detected in this way, for example by monitoring the current of the coil by means of a comparator. By reversing the polarity of the constant voltage, the current of the driving coil will not exceed a dangerous rate.

Without an external field, the surge in the coil current occurs symmetrically, after reversals of the voltage polarity. In the case of an existing magnetic field strength H , an external field component parallel with the axis of the driving coil, similarly to the double coil arrangement, accelerates or delays the driving of the iron core into saturation and hence a surge in the coil current as shown by the arrows in Fig. 5. This occurs with an opposite orientation in the opposite polarity phases, and this emerging phase asymmetry is also proportional with the external field to be measured in the case of an ideal Z iron core characteristic. In the case of a non- ideal, i.e. real iron core, the detection can be linearized by feedback and by compensating the external field in the iron core.

Therefore, according to the invention, the surges, occurring when reaching iron core saturations, in the current of the driving coil 23 are detected, and on the basis of the time-related positions of these surges, information characterising the detected magnetic field strength HM is obtained. The single coil magnetometer according to the invention can be operated by two preferred basic methods, in a self-oscillating and in a controlled frequency mode.

In the case of the self-oscillating version, the polarity reversal of the voltage supplied to the driving coil takes place automatically when reaching a value of the driving coil current, which is associated with the magnetic saturation of the iron core. In this case, the frequency of the drive voltage is determined by the resultant impedance of the coil, thereby being a determined value. The time-related ratio of voltage polarity is proportional with the external field to be measured. In this case, therefore, the magnetometer 30 has a processing means 34 which, when a surge occurs, instructs the drive means 31 to reverse the polarity of the drive voltage UE and calculates the information concerning the magnetic field strength HM based on the proportion of periods spent in each polarity. The polarity reversal may also take place with a fixed and controlled frequency. In this case the period must be sufficiently long for allowing the field of the current increasing in the iron core to magnetise the iron core up to saturation. Once saturation is reached, the current can be limited to a permissible range by other electronic means. In this case the frequency and filling factor of the drive voltage are fixed. The response signal of the magnetometer is the output signal of the comparator which detects a surge in the driving coil current. This can even be used for triggering a flip-flop, in which case the filling factor of the flip-flop output is proportional with the external field to be measured. In this case, therefore, the magnetometer 30 has a drive means 31 which reverses the polarity of the drive voltage U_E with a constant period which is of sufficient length to drive the iron core into magnetic saturation.

Our experiments have shown that in the case of several sensor elements located closely side by side, where a magnetic coupling may occur among the sensors, the self-oscillating version cannot be applied due to the modulations arising among the various oscillation frequencies. For this reason, also in the material testing apparatus to be described later, the single coil version running synchronously and operating on a controlled frequency has been chosen.

The manufacturing and winding of the sensor element 20 shown in Fig. 6 may be carried out by way of example as follows. An elongated coil body 21 is made for example from a 0.6 mm printed circuit board. Many coil bodies can be manufactured simultaneously by cutting from a printed circuit board. The thickness of the coil body 21 is preferably between 0.2 and 3 mm, and the length may vary between 3 and 40 mm. An elongated iron core 22 sits on the coil body 21. The characteristic sizes of an iron core used by way of example are as follows: width 0.1 to 2 mm, length 1 to 50 mm, thickness 0.001 to 0.05 mm. The iron core can be produced by chilling, evaporation or atomization and also by chemical deposition. The material of the iron core can be for example Co66Fe3Cr₉Si8B ₄. The raw material of the iron core 22 can be produced by way of example as follows. The raw material of the sensor can be, for example, an amorphously structured ferromagnetic strip. In the course of production, a strip of approx. 100 m length is manufactured, from which a piece is cut off for making the sensor. A large quantity of amorphous alloy can be produced from the melt by chilling. The free-jet melt- spinning process is suitable for producing continuous strips. The apparatuses for manufacturing the products operate on the basis of the single sided cooling principle. The induction melted alloy melt is sprayed through a round cross section nozzle by means of a suitably selected overpressure on the outer surface of a rapidly turning disk which is cooled by water for example. When hitting the disk, the ejected melt is turned into a pool. The cooling surface of the disk is exposed to this pool, and therefore the moving surface takes with it a thin layer of melt (in the given case already partly solidified), and when this is solidified into a continuous strip, it flies off the disk partly as a result of the centrifugal force and partly due to cooling induced shrinkage. By means of this procedure, strips and wires of 10 to 100 μιη thickness and 1 to 10 mm width can be produced in arbitrary lengths. The cooling rate varies along the cross section of the strip and it is highest on the surface exposed to the disk and when approaching the free surface, it decreases according to the inverse of the square of thickness. In the case of 20 micron thickness this change is generally negligible.

A further development of the melt-spinning technology is the planar flow casting method. In this case a tip of a quartz tube touches the melt pool, stabilising it in space and time by reducing the hydrodynamic fluctuations. In this way it becomes possible to make a homogenous length broad strip (up to 200 mm width in our case) disregarding the initial and final transients of approx. half a meter length. The thickness of the strip mostly depends on the quantity of material flowing through the opening of the quartz tube in a given time. Accordingly, the most important technological parameters are as follows:

- The temperature of the melt, which makes an impact on the material flow mainly through the viscosity. It is generally 50 to 100 K higher than the melting point of the given alloy.

- The distance between the nozzle and the disk, which is typically 0.1 to 0.3 mm.

- In the case of iron-, cobalt- and nickel-based alloys, generally a Cu(Zr) disk is used, but in order to improve adherence, the application of a disk of a different material composition may also be justified. For high Si content Fe-based alloys, for example, a disk made of a stainless material has been proven to be more favourable.

An amorphous structure is practically free of the grain boundaries and precipitations characteristic of polycrystalline materials, therefore it has isotropic soft magnetic characteristics, where the crystal anisotropy is averaged to zero. However, during the production, due to the heterogeneous cooling conditions, residual stress freezes into the material, which introduces anisotropy into the sample as a result of magnetostriction. This heterogeneous magnetostrictive anisotropy deteriorates the theoretically expectable soft magnetic characteristics: the coercive space expands and the permeability decreases. Since the frozen residual stresses may only be reduced, but not fully eliminated by heat treatment, ideal soft magnetic characteristics may be approached, if the composition is adjusted in a way that the magnetostriction constant is close to zero. As currently described in the literature, this is only possible by using a high Co content amorphous alloy. The Co content makes the raw material more expensive, but in the case of sensors this is not a significant point, because a limited material quantity is needed.

On the basis of comparative measurements, the magnetostriction of the material Co73Fei ,5Bi₆Si5,2Moo,i5Ni₀,35Mn3,8 proved to be the lowest.

Next, preferably with a magnetic field heat treatment induced anisotropy is generated in the iron core 22 in the direction of the magnetic field, by which the shape of the hysteresis loop can be substantially influenced. By means of a longitudinally induced anisotropy, the magnetising curve is squared, and by a transversal heat treatment it is flattened. It is important to note that the coercitive space is reduced under the effect of magnetic field heat treatment, regardless of the direction of the magnetic field.

The method according to the invention deducts from time-measurement the magnetic flux asymmetry caused by the field to be measured in a magnetic coil, which is excited by a symmetrically changing magnetic field. A compensating current eliminating the time shift between the positive and negative voltage pulses will be proportional with the field to be measured. To make the measurement accurate, the voltage pulses are to be as much distortion-free as possible. The slightly tilted (associated with a small remanence), but steep magnetising curve facilitates this measuring accuracy.

A driving coil 23 is wound on a unit consisting of the coil body 21 and the iron core 22 fitted thereon. By way of example, the driving coil 23 can be made of an enamelled copper wire having a 0.05 mm diameter. For the experiments, a driving coil 23 consisting of a total of 030 windings in 5 layers was used.

The installation of the sensor elements 20 at the ends of the elongated coil body 21 is assisted by orienting holes in the printed circuit board. When the sensor elements 20 are placed between adjusting pins, the driving coil 23 can be appropriately positioned, and the two poles of the driving coil 23 can be soldered to the appropriate soldering points.

Fig. 7 shows a block diagram of a preferred magnetometer 30 as an example. The magnetometer 30 comprises a sensor element 20 according to the above description, a drive means 31 operating the sensor element 20, and a processing means 34 which processes the response signal of the sensor element 20. The drive means 31 preferably comprises a control unit 32 and a drive unit 33 for driving the sensor element 20.

According to the above description, the drive unit 33 switches a DC voltage of varying polarity, i.e. a driving voltage U_E to the driving coil 23 of the sensor element 20. As described below, during the material testing, several sensor elements are arranged side by side and therefore several magnetometers are required for covering a certain testing width. For the synchronised operation thereof the control unit 32 of the drive means 31 has an input which receives a CLOCK signal of a central control unit. Furthermore, the control unit 32 has an input receiving a compensating OFFSET voltage signal, the magnitude of which depends on the movement velocity of the material to be tested. In the experiments it has been recognised that even in the case of the most accurate production and positioning, each sensor element 20 needs a different offset signal depending on the velocity. In the case of the preferred sensor element 20 also depicted in the figures, the velocity dependence of the compensating offset signal is linear, and as a result of the production/positioning tolerances, it is a function determined by individual constants for each sensor element 20. According to the preferred embodiment shown in Fig. 7, this velocity dependent offset signal also comes from the central control unit (not shown) to the control unit 32 of the drive means 31. According to the invention, the current l| detected on the driving coil 23 of the sensor elements 20 is measured by a current meter 35 and the surge is processed by a signal processor 37, using a comparator 36 for detection. The current meter 35, the comparator 36 and the signal processor 37 represent parts of the processing means 34. After an A/D conversion, the measurement results provided by the signal processor 37 are supplied to the central control unit through the control unit 32.

Consequently, the magnetometer 30 shown in Fig. 7 comprises the sensor element 20, the drive means 31 operating the sensor element 20 and the processing means 34 which processes the response signal of the sensor element 20. The sensor element 20 comprises the driving coil 23, and the iron core 22 which can be driven into magnetic saturation by the driving coil 23. According to the invention, the magnetometer 30 has a drive means 31 which feeds the driving voltage U_E of varying polarity to the driving coil 23, and it also has the processing means 34 which senses the surges at reaching iron core saturations of the current of the driving coil 23, and which provides information characteristic of the detected magnetic field strength H according to the positions in time of the surges.

Fig. 8 shows by way of example a preferred arrangement consisting of the sensor elements 20. Fig. 9 shows the magnified view of section B in Fig. 8. In the arrangement, there are driving coils 23 which are uniaxial and are arranged with equal spacings. In the arrangement, the driving coils 23 of the sensor elements 20 are arranged in two adjacent rows, and the driving coils 23 in one row are shifted by half a period compared to the driving coils 23 in the other row.

The measurement results converted to digital format and coming from each magnetometer 30 are supplied to the central control unit mentioned above. The central control unit comprises a central signal processor 40 which evaluates the measurement results. The block diagram thereof is shown in Fig. 10. The measurement results arrive through the appropriate number of channels corresponding to the number of magnetometers 30 at a filtering means 41 , i.e. the pre-processing unit of the central signal processor 40.

After filtering, a signal energy calculating means 42 which amplifies the information provided by the measurement results preferably calculates the energy of the measuring signals as a further signal processor.

The preferred operation of the signal energy calculating means 42 is by way of example as follows. Instead of the signal energy, (discrete) signal power density (SPD) is calculated, which is simply expressed as follows:

SP

SPD =— , where n is the number of samples. (1 ) n

The SP filter implements the following calculation:

n

Where the average is

* = - ·∑*/ (³) n

SP can be expressed in the following form:

n n

but, because of (3):

(5) n

and therefore

SP = T x_; ² - 2 · n■ ( )² + n■ (xf =∑x,² - n■ ( )² (6) n n

that is

&P ==∑x, ² - ,7 . (x)² (7) n

For this calculation, only two sums are to be calculated: the sum of samples and the sum of the square of samples, and the other elements are constant.

It follows from this that in the case of a moving filter, the calculation can be substantially accelerated, because only these two sum values are to be 'updated' by deducting the drop-out element or by adding the new element, and when the two sums are used in (7), the value of the new SP is arrived at by one calculation.

The new sample sum can be calculated as follows: x ₌ - ^L ₊ ^i±L _{= +} "÷' ^~ ' (8) n+\ n _n Yi n _n

and

∑*,² =∑*,² - *.² + *.² ₊i (9) n n

n+1 n

After this, by means of an evaluation means 43, threshold and correlation evaluations are carried out on the processed signals.

Preferably rules based on calculations and experiences are recorded in the evaluation means 43. A material defect is indicated only if the detected signal energy exceeds a certain level on one channel. Because the effects generated by the material defects can be observed spread in space, even the smallest material defect provides a response signal on more than one sensor element 20 according to the experiences. By means of a correlation between the various channels, measuring defects when only one channel signals a material defect can be screened.

From the evaluation means 43, on the basis of the calculated data, a coloured real time display is presented on a display unit 44. Statistics can be drawn up by means of a statistic unit 45 about the material defects detected.

The non-destructive material testing apparatus according to the invention can be preferably applied in rolling mills. The rolling operation shown in Fig. 11 depicts rolls 50 and 51 , which roll a plate 52 from the incoming material.

When aluminium plates are manufactured, various thicknesses and widths of plate and strip products are made from thick aluminium plate in the rolling plant, in several machining phases. According to the experiences, already at the time of casting, surface and inner defects may develop in the material, and during rolling the already existing material defects may come to the surface and be aggravated. A typical defect can be an air or gas inclusion, which represents a discontinuity defect inside the material. The characteristic of such an inclusion is that during further machining, the rolled plate may be punctured, or the defect may be come to the surface on one side. An inclusion does not necessarily represent one large defect, it may also mean that in one area the strip exhibits lots of small defects, i.e. it is porous (has lots of pinholes).

Further characteristic defects are the surface damages of the strip, arising during the handling of the strip (transport, conveyance or the introducing of a contaminant or chip into the system). These damages can be scratches, indentations or punches. It may also happen that part of a chip is pressed into the surface of the strip by the rolling mill. This invisible defect may represent a problem in rolling thin strips, because during the process of gradual thinning, at one time the plate may suddenly be split into two.

It is a general characteristic of the defects that their depths and locations within the thickness change during the rolling process, and their sizes increase in the direction of rolling, i.e. they are 'spreading'. An early recognition of defects is important because the more machining phases the material is subjected to, the defect is less detectable, and its contour is less defined (the only exceptions are the air or gas inclusions, which appear when the sheet is so thin that it is punctured on one side). With the current methods these defects can only be noticed after the originally defective aluminium plate has been subjected to machining for several hours. Consequently, a reliable measuring method is required which enables the screening of the defective plate before further processing. When a section of the material is detected to be defective, it is to be removed from the manufacturing process, so as to avoid unnecessary machine and work time spent on inferior raw material. One must be able to pinpoint each of these defects at any point of the strip width and also on the top and bottom rolled surfaces.

Aluminium strips machined by rolling arrive at the rolling mill in the form of reels, in several thicknesses and widths. The reels to be machined are taken by forklift trucks to the appropriate machine of the plant. The typical dimensions of the strips are as follows: 150 to 2000 mm width, 2.4 to 12 mm thickness. The strip is introduced on one side of the rolling mill by the machine operators. They adjust the so-called pass (pass is the size by which the plate thickness is reduced during the rolling process, and the same term applies when the plate is rolled in one direction (one pass)), and while the rolling mill moves slowly, the strip is guided across under the machining roll, and is introduced to the unwinding roll. Next, the speed of rolling is increased, and the rolling mill winds the complete strip onto the unwinding roll, while the thickness is reduced. The typical rolling data are as follows: pass 1 to 3 mm, strip velocity 50 to 120 m/min, strip temperature 30 to 40 °C, and other environmental characteristics are the presence of oil mist, oil and vapour. Figs. 2 to 16 show an arrangement, an assembly in several views, which consists of a probe 60 comprising the magnetometers 30 according the invention, and a unit 61 which preferably produces a constant magnetic field. The function of the latter is to create a magnetic field which is as homogeneous as possible, and has an advantageous direction and distribution from the aspect of the probe 60. The plate 52 produced by rolling crosses the arrangement shown in Fig. 12. As shown in Fig. 13, the material to be tested, i.e. the plate 52 moves practically in a perpendicular direction both with respect to the magnetic lines of force generated by the unit 61 and to the axes of the driving coils 23 arranged on the bottom surface of the probe 60, in the testing region 62 or space section. Since the material defects can be detected mostly at the beginning of the rolling process, the probe 60 preferably collects data in the case of each strip at the time of the first pass. In the reset position, the probe 60 is in an enclosed protective housing, from which it is moved out for the measurement. Next, the probe 60 and the magnetic unit 61 'closes' onto the plate 52. For example, the movement of the units can be carried out by linear actuators and pneumatic work cylinders. Once the measurement is completed, the probe 60 moves back into the protective housing.

The measurement is based on the principle that the sensor elements 20 applied for the measurement are able to detect the changes of the magnetic field located in their vicinity. The magnetic field can be created by an electromagnet or permanent magnet. In the measuring arrangement used by way of example, a permanent magnet excitation has been chosen. For operating the equipment according to the invention, eddy currents are to be induced in the tested material by means of relative displacement. If the tested material is displaced compared to the magnetic field, and the displacement has a component which is normal to the direction of the lines of force, and the lines of force intersect the tested material, then voltage is induced in the latter. As this voltage finds an enclosed path, it is able to drive a current along this path in the conductor. This current is the eddy current, the direction of which can be determined by Lenz's law and which is converted into heat.

The criteria of the excitation field are as follows:

a magnetic field of appropriate magnitude is to be established for exciting the eddy current of necessary extent;

the lines of force of the excitation magnetic field should be normal (perpendicular) to the longitudinal axis of the group of sensors;

the excitation field must be as homogeneous as possible in the whole length of the group of sensors;

the excitation should be able to penetrate the whole cross section of the tested material.

In the arrangement shown by the figure, the permanent magnet unit 61 can be seen at the bottom. The rolled aluminium strip is moved between this permanent magnet unit 61 and the probe 60. The magnetic sensors in the probe detect the constant magnetic field, and measure the changes thereof. From these change signals, the measuring system can draw conclusions about material defects and other irregularities.

Under the impact of the vicinity of a magnetisable material, the excited or permanent magnetic field is greatly modified, i.e. distorted even if this material does not move. Because in this case the aluminium to be tested is not a magnetisable material, we can only rely on the changes triggered by the eddy current field generated by the movement of the strip during measurement. While the strip moves on, if an irregularity or a material defect having a modified electric conductivity or magnetic characteristic passes in front of the magnetic field, this will change the generated eddy currents, therefore the spatial distribution of the resultant magnetic field and with the magnetic sensors showing this deviation. ln the perspective view of Fig. 14, the probe 60 is shown with the cover removed. The vertically oriented printed circuit boards in the probe 60, as shown in the figure, carry those parts of the magnetometers 30 which are beyond the sensor elements 20. It can also be seen in the side view of Fig. 15 that these printed circuit boards are connected to a common bus, which ensures communication with the central control unit.

Fig. 16 shows a roller unit 63 fixed firmly to the probe 60, and which is adapted to ensure a constant distance between the probe 60 and the plate 52.

Fig. 17 depicts by way of example a preferred structure of the unit 61 creating the magnetic field. The unit 61 generating the constant magnetic field comprises a socket, in which neodymium permanent magnets 64 are arranged in appropriate nests, preferably in a special Halbach array.

This arrangement provides a favourable distribution of the lines of force, thereby enhancing the reliability of the measurement. From the aspect of measurement the most important thing is that the field strength of the permanent magnet changes as little as possible depending on the location. This is because the sensors are very sensitive to a change in the magnetic field. If the sensor body and the magnet are displaced minimally, this displacement appears in the measurement as a disturbance signal. By means of the above arrangement, the field strength of magnet assembly has value of 82 to 86 mT, being approx. constant within the width of the pole piece. In designing the magnet assembly, it is to be kept in mind that the reliability of measurement does not depend on the power of the permanent magnet, but on the uniformity of the magnetic field. Too high field strength generates undesirably high values in the measurement. Installing the magnets in a Halbach array requires careful attention, because the magnets must be fixed very close to each other and they are to be installed in repelling positions. This has been assisted by an appropriately designed magnetic socket. The polarity of the magnets has been measured by a Tesla meter, marking the north pole. Next, the magnets were pressed one by one using pincers into the nests. Finally, they were pushed to the bottom of the nests one by one using a rubber hammer and a mandrel. A magnetic field can be implemented also by an electromagnet and by the combination of two types of magnet. This is because the fixed field of a permanent magnet can be modified by using an electromagnet and by appropriate adjustment.

Fig. 18 is a view of an example of an inclusion in plate 52. Such a typical inclusion has a diameter of approx. 2 mm and it results in the measuring diagram depicted in Fig. 19. Because the deviation caused by the inclusion emerges both in positive and negative directions, a signal characteristic of the material defect can be obtained in the course of signal processing when the signal energy is calculated preferably with the square of the deviation from the mean level.

During the measurement according to the invention, the axes of the driving coils 23 are normal to the lines of force of the magnetic field generated by permanent magnets or an AC electromagnet, and the tested material is moved in a direction normal to the said axes. Of course, the measurement may also be done, if the directions are only close to being perpendicular, i.e. when they are practically normal. According to the invention, such a transversal arrangement is understood to be 'essentially perpendicular' when the functionalities of the invention can still be performed.

In the measuring arrangement shown in the figures, the probe 60 and the unit 61 creating the permanent magnetic field are arranged on the opposite sides of the tested material. According to the invention, the probe 60 and unit 61 creating the permanent magnetic field may also be arranged on the same side of the tested material, because the benefits of the invention prevail also in this arrangement.

Regarding the driving coils 23 of each sensor element 20, the geomagnetism, the production uncertainties and orientation errors result on the one hand in a static offset. This can be compensated, for example, by feeding to the system a DC voltage of a determined value. During the experiments the appearance of velocity dependent, so-called dynamic offset has also been recognised. This offset may also be traced back to geometrical defects. In the case of rolling mills, by way of example, the velocity signal by which this velocity dependent offset can be supplied to each magnetometer 30 through the central control unit can be obtained from the rolling mill. Regarding the magnetometers located in the probe, the factors of both the static and dynamic offsets may have different values for each magnetometer.

The evaluation of characteristic values measured by the probe 60 is carried out by a software running on the computerized central control unit. On the basis of the predetermined signal processing method and error limits, the software must give at the end of the rolling the number of defects found in the strip and their actually locations within the strip length.

According to the observations of the experiments, it is advantageous if the permanent magnet unit 61 and the probe 60 are steadily positioned with reference to each other. If they are displaced in relation to each other, this will be measured by the probe 60 as a distortion of the magnetic field, and this change will generate a disturbance signal making evaluation difficult.

The closer the magnet unit 61 and the probe 60 are to the material, the more favourable the signal to noise ratio of the measurement will be. The probe 60 comprises sensitive components and electronic elements, and therefore it should be located at a safe distance from the plate to be tested. In the arrangement of the implemented example, this represents a distance of approx. 1.5 mm. The permanent magnet unit is much less prone to damage, and therefore it is sufficient even to leave a clearance of 0.5 mm. Therefore, the distance between the two elements varies between 4.4 and 14 mm subject to the thickness (2.4 to 12 mm) of the tested plate.

Of course, the invention is not limited to the preferred embodiments described above, but further modifications and changes are possible within the scope determined by the claims.

Claims

1. A magnetometer comprising a sensor element (20), a drive means (31) operating the sensor element (20) and a processing means (34) processing the response signal of the sensor element (20), the sensor element (20) comprising a driving coil (23), and an iron core (22) suitable to be driven into magnetic saturation by the driving coil (23), characterised in that

- the drive means (31) is adapted to supply a constant voltage of varying polarity as a drive voltage (UE) to the driving coil (23), and

- the processing means (34) is adapted to detect surges of a current ( ) of the driving coil (23) appearing at reaching iron core saturations, and to provide information characterising a detected magnetic field strength (H_M) on the basis of positions in time of said surges.

2. The magnetometer according to claim 1 , characterised in that the processing means (34) is adapted to instruct the drive means (31) to reverse the polarity of the drive voltage (UE) when a surge occurs, and to calculate the information characterising the magnetic field strength (HM) on the basis of the ratio of periods spent in each polarity.

3. The magnetometer according to claim 1 , characterised in that the drive means (31) is adapted to reverse the polarity of the drive voltage (UE) with a constant periodicity having a sufficient period to achieve magnetic saturations of the iron core.

4. The magnetometer according to any of claims 1 to 3, characterised in that the sensor element (20) has an elongated coil body (21) made of a printed circuit board, on which the iron core (22) also having an elongated shape is arranged, and the driving coil (23) is wound around the unit consisting of the coil body (21) and the iron core (22).

5. The magnetometer according to any of claims 1 to 4, characterised in that the iron core (22) is made of a ferromagnetic material having an amorphous structure and being magnetically heat treated.

6. A method of operating a magnetometer, comprising the steps of switching a voltage to a driving coil (23) of a sensor element (20) of the magnetometer (30), the sensor element (20) comprising the driving coil (23) and an iron core (22) suitable to be driven into magnetic saturation by the driving coil (23), and processing a response signal of the sensor element (20), characterised by

- supplying a constant voltage of varying polarity as a drive voltage (UE) to the driving coil (23), and

- detecting surges of a current ( ) of the driving coil (23) appearing at reaching iron core saturations, and providing information characterising a detected magnetic field strength (HM) on the basis of positions in time of said surges.

7. The method according to claim 6, characterised by reversing the polarity of the drive voltage (UE) when a surge occurs, and calculating the information characterising the magnetic field strength (H_M) on the basis of a ratio of periods spent in each polarity.

8. The magnetometer according to claim 6, characterised by reversing the polarity of the drive voltage (U_E) with a constant periodicity having a sufficient period to achieve magnetic saturations of the iron core.

9. A non-destructive material testing apparatus comprising

- a unit (61) creating a magnetic field,

- magnetometers (30) comprising driving coils (23) with axes being essentially perpendicular to the lines of force of the created magnetic field, and

- a testing region (62) enabling a movement of the tested material in a direction being essentially perpendicular to both the lines of force and the axes,

characterised by comprising magnetometers (30) according to any of claims 1 to 5.

10. The apparatus according to claim 9, characterised by comprising uniaxial driving coils (23) arranged with equal spacings.

11. The apparatus according to claim 10, characterised in that the driving coils (23) are arranged in two adjacent rows, and the driving coils (23) in one row are shifted by half a period compared to the driving coils (23) in the other row.

12. The apparatus according to any of claims 9 to 11 , characterised in that the magnetometers (30) are assigned to a common central control unit and each drive means (31) comprises an input receiving a clock signal of the central control unit, and an input receiving a compensating voltage signal depending on a velocity of the tested material.

13. The apparatus according to claim 12, characterised in that the processing means (34) comprise an A/D converter for converting the measurement results into a digital form, the central control unit is adapted to digitally read the measurement results, and the central control unit comprises a central signal processor (40) evaluating the measurement results.

14. The apparatus according to claim 13, characterised in that the central signal processor (40) comprises a filtering means (41) for filtering consecutive measurement results of each magnetometer, a signal energy calculating means (42) for amplifying information in the measurement results, and an evaluating means (43) performing threshold and correlation evaluations of the so obtained values.

15. The apparatus according to claim 9, characterised by comprising a unit (61) creating a permanent magnetic field.