CA2079774A1

CA2079774A1 - Method and apparatus for magnetic inspection

Info

Publication number: CA2079774A1
Application number: CA002079774A
Authority: CA
Inventors: Seigo Ando; Yasuhiro Matsufuji
Original assignee: Individual
Current assignee: JFE Engineering Corp
Priority date: 1991-02-04
Filing date: 1992-01-31
Publication date: 1992-08-05
Also published as: US5502382A; DE69222989D1; EP0523249B1; AU1221092A; WO1992014145A1; DE69222989T2; US5357198A; EP0523249A1; EP0523249A4

Abstract

A B S T R A C T
In a magnetic inspection method and apparatus of the present invention, a magnetizer (4) is provided such that a pair of magnetic poles (4a, 4b) face an object (10) to be inspected. A magnetic field is generated by the magnetizer (4) in the object (10). A magnetic sen-sor (7a) is situated at substantially middle point of a line connecting the magnetic poles (4a, 4b) or a line parallel to this connecting line. Shield bodies (22a, 22b, 23a, 23b) with a low coercive force are arranged on those sides of the magnetic sensor (10) which face the magnetic poles. The magnetic sensor (7a) detects a leakage magnetic flux due to an internal or surface defect of the object (10).
In addition, according to the magnetic inspection apparatus of this invention, the distance (2A) between the shield bodies (22a, 22b) is 2.2 times or more, and 2.8 times or less, the distance (L) between the magnetic sensor (7a) and the object (10).
Furthermore, each shield body (23a, 23b) has an L-cross section composed of a vertical portion (24a) and a horizontal portion (24b), and the width (2A) of the horizontal portion (24b) in the direction of arrangement of the magnetic poles (4a, 4b) is 0.4 time or more, and 0.6 time or less, the distance (B) between the magnetic poles.

Description

~ ) 7 .) 7 ~ ~

S P E C I F I C A T I O N

"METHOD AND APPARAT~S FOR MAGNETIC INSPECTION"

Technical Field The present invention relates to a magnetic inspection method and a magnetic inspection apparatus wherein a magnetic field is generated by a magnetizer in a to-be-inspected object formed of a magnetic mate-rial such as steel plate and a leakage magnetic flux due to a defect on the object is detected by a magnetic sensor.
Back~round Art A magnetic inspection apparatus detects, by utiliz-ing magnetism, defects such as internal and surface flaws and inclusion in a thin steel strip or a to-be-inspected object. It was reported that a magnetic inspection apparatus, in which a magnetic sensor array comprising linearly arranged magnetic sensors for detecting magnetic fluxes is built, is capable of successively detecting defects on a running thin steel strip over the entire width thereof (Published Unexamined Japanese ~tility Model Application (PUJUMA
No. 63-107849).
Figs. 39 and 40 are schematic cross-sectional views showing, in different directions, the above-mentioned magnetic inspection apparatus for successively detecting the defects on the running thin steel strip. Fig. 41 is a side view showing the state in which the magnetic inspection apparatus is built in a support apparatus.
Referring to Fig. 41, a horizontal arm 12 is sup-ported by a pair of spring members 13a and 13b within a frame 11 set on the floor of a room. Accordingly, the arm 12 is vertically movable. A stationary shaft 2 of the magnetic inspection apparatus is fixed at the center of the arm 12. A pair of guide rolls 14a and 14b for guiding a thin steel strip 10 on the outer peripheral surface of a hollow roll 1 are arranged on both sides of the frame 11.
In Figs. 39 and 40, one end portion of the station-ary shaft 2 penetrates the hollow roll 1 of a non-magnetic material along the center axis of the roll 1.
The other end portion of the shaft 2 is fixed on the horizontal arm 12. The stationary shaft 2 is supported on the inner peripheral surfaces of both end portions of the hollow roll 1 by a pair of rolling bearings 3a and 3b such that the shaft 2 is situated along the center axis of the hollow roll 1. Accordingly, the hollow roll 1 is freely rotatable about the stationary shaft 2.
A magnetizing core 4c having a substantially U-cross section is fixed to the stationary shaft 2 by means of a support member 5 within the hollow roll 1, such that magnetic poles 4a and 4b of the core 4c are situated close to the inner peripheral surface of the hollow roll 1. A magnetizing coil 6 is wound around oJ 7 ~

the magnetizing core 4c. Thus, the magnetizing core 4c and magnetizing coil 6 constitute a magnetizer 4.
A magnetic sensor array 7 consistlng of magnetic sensors 7a arranged linearly along the axis of the hollow roll 1 is fixed to the stationary shaft 2 between the magnetic poles 4a and 4b of the magnetizing core 4c.
A power cable 8 for supplying an excitation current to the magnetizing coil 6 and a signal cable 9 for tak-ing out output signals from the magnetic sensors 7a of the magnetic sensor array 7 are led to the outside through the inside passage of the stationary shaft 2.
Accordingly, the positions of the magnetizer 4 and magnetic sensor array 7 are fixed, and the hollow roll 1 rotates around the magnetizer 4 and magnetic sensor array 7 with a small gap.
When the outer peripheral surface of the hollow roll 1 of the magnetic inspection apparatus with the above structure is pressed on one side surface of the thin steel strip 10 under a predetermined pressure which runs, for example, in a direction a, the hollow roll 1 rotates in a direction b since the stationary shaft 2 is fixed on the horizontal arm 12.
In the above magnetic inspection apparatus, when an excitation current is supplied to the magnetizing coil 6, a closed magnet'c path is formed by the magnetic poles 4a and 4b of the magnetizing core 4c and the run-ning thin steel strip 10. If there is an internal or r~J 7 1 surface defect of the thin steel strip, the magnetic path in the thin steel strip is disturbed and a leakage magnetic flux occurs. The leakage magnetic flux is detected by the magnetic sensor 7a which constitute a part of the magnetic sensor array 7 and faces the location of the defect. A signal corresponding to the defect is output from this magnetic sensor 7a.
The level of this output detection signal corre-sponds to the magnitude of the internal or surface defect of the thin steel strip 10. Thus, by measuring the level of the output signal, the width-directional position and magnitude of the internal or surface defect of the steel strip 10.
However, regarding the above-described magnetic inspection apparatus, there are the following problems to be solved.
When a small defect of the object such as the thin steel strip 10 is detected, the S/N does not basically increase unless the magnetic force is adequate.
In order to solve such a problem, there is an idea that the excitation current to the magnetlzing coil 6 of the magnetizer 4 is increased to intensify the magnetic field forming in the thin steel strip 10. The greater the magnetic flux in the steel strip 10, the higher the value of the leakage magnetic flux due to the defect.
In general, as shown in Fig. 42, when the magnetic 2~377 ~

poles 4a and 4b are situated near the thin steel strip 10, magnetic force lines extend from the pole 4a to the pole 4b through a magnetic gap, the thin steel plate 10 and another magnetic gap. Since the thin steel strip 10 is formed of a ferromagnetic material, magnetic fluxes do not leak out of the thin steel strip 10 while passing through the strip 10 if no defect is present in the strip 10.
However, as stated above, if the magnetic field applied to the thin steel strip 10 is increased so as to obtain a leakage magnetic flux of a sufficiently high signal level when a defect is present, the strip 10 is magnetically saturated, as shown in Fig. 42. As a result, a large floating magnetic flux 1`5 occurs even in a defect-free portion. The actual value of the floating flux 15 is extremely high, e.g. several Gauss to several-ten Gauss.
In addition, it is experimentally confirmed that the variation in a vertical component of the floating flux 15 depends greatly on the speed of the thin steel strip 10. Fig. 43 shows the relationship between the output voltage and the speed of the thin steel strip 10, in the case where the vertical floating magnetic flux in the defect-free portion was measured with the sensitiv-ity of the magnetic sensors 7a lowered intentionally.~s shown in Fig. 43, the output voltage rises as the speed of the thin steel strip 10 increases.

- 6 - ~7~ 17 :1 Accordingly, the varia-tion in vertical component of the floating magnetic flux 15 rises in accordance with the increase in speed.
Since the floating magnetic flux 15 is always generated, the leakage magnetic flux due to a defect is superimposed on the floating magnetic flux, when the defect is present on the thin steel strip 10. In addition, the floating magnetic flux is greater than the leakage magnetic flux. Each magnetic sensor 7a detects the floating flux and leakage magnetic flux simultaneously, as shown in Fig. 42.
The same phenomenon occurs in the case where the magnetic sensor 7a is situated on the magnetic pole side of the thin steel strip 10, as indicated by a solid line in Fig. 42, and in the case where the magnetic sensor 7a is situated on that side of the strip 10 opposite to the magnetic poles 4a and 4b, as indicated by a broken line.
~ n the other hand, in order to detect the defect of the thin steel strip 10 with high precision, it is necessary to increase the sensitivity of the magnetic sensors 7a. However, as stated above, the variation component of the leakage magnetic flux due to the defect is superimposed on the high-level floating flux in the defect-free portion. Thus, if the high-sensitivity mag-netic sensor array 7 is used, the magnetic sensors aresaturated by the floating flux because of their narrow dynamic range, and the leakage magnetic flux due to the ~7~7~7-1 defect cannot be detected with high precision.
Disclosure of the Invention A first object of the invention is to provide a magnetic inspection method and apparatus capable of remarkably decreasing a floating magnetic flux inter-secting a magnetic sensor for detecting a leakage mag-netic flux due to a defect, and capable of preventing saturation of output of the magnetic sensor, exactly detecting a small defect, and enhancing inspection accuracy greatly.
A second object of the invention is to provide an optimal shield shape for a magnetic sensor, thus attaining a high S/N in an output signal from the magnetic sensor.
A third object of the invention is to provide a magnetic inspection apparatus capable of eliminating a noise component contained in an output signal from a magnetic sensor, which is due to a local variation in magnetic permeability of an object to be inspected, thereby remarkably enhancing the inspection accuracy of the defect.
In order to achieve the first object, in the mag-netic inspection method and apparatus of this invention, a magnetic sensor for detecting a leakage magnetic flux due to an internal or surface defect of the to-be-inspected object is provided at a substantially middle point of a line connecting a pair of magnetic poles of ~ ~ 7 ~ rJ 7 :~

a magnetizer for generating a magnetic field in the object or a line parallel to this connecting line. A
shield body with a low coercive force ls provided on each of sides of the magnetic sensor, which face the s magnetic poles of the magnetizer.
Consideration will now be given to a magnetic path and magnetic force lines produced by the magnetizer having the pair of magnetic poles and the object situ-ated to face the magnetic poles.
Referring to Fig. 5, magnetic poles 41a and 41b of a magnetizer faces a to-be-inspected object 42 formed of, e.g. a thin steel strip. A magnetic sensor 71a is provided at the center of a line parallel to a line connecting the magnetic poles 41a and 41b. A first shield body 43a is situated on one side of the magnetic sensor 71a, which faces the magnetic pole 41a. A second shield body 43b is situated on the other side of the magnetic sensor 71a, which faces the other magnetic pole 41b.
In this state, when a magnetic field generated by the magnetizer is small, magnetic force lines coming out of the magnetic pole 41a pass through a magnetic gap, enter the object 42, pass through the object 42 and an opposite magnetic gap, and enter the opposite magnetic pole 41b. When the magnetic field generated by the magnetizer is increased, the object 42 is magnetically saturated and the magnetic resistance increases.

~ 'J ~ ~) 7 J 1 _ g .

Thus, as shown in the figure, a great deal of ~loating magnetic flux occurs.
However, since the shield bodies with low coercive force are provided on both sides of the magnetic sensor 71a, the floating magnetic flux crossing the magnetic sensor 71a is remarkably reduced. The magnetic sensor 71a faces the object 42. Thus, if the object 42 has a defect at a location facing the magnetic sensor 71a, the sensor 71a detects a leakage magnetic flux corresponding to the defect. In this case, since the ambient floating magnetic flux 15a is small, the magnetic sensor is not saturated and only the leakage magnetic flux can effi-ciently be detected.
The detection of the vertical component of the leakage magnetic flux by means of the magnetic sensor 71a will now be described with reference to Fig. 6.
If a defect 44 is present in the object 42, a magnetic field 45 generated around the defect 44 exhib-its a magnetic flux distribution characteristic of a very small magnetic pole, as shown in the figure. When the object 42 runs in the direction a and the defect 42 is moved to a location just below the magnetic sensor 7~a, the magnetic sensor 71a detects a magnetic flux distributed vertically, as indicated by a solid line 46.
The width Wh of the magnetic field 45 due to the defect 44 is only several mm. Even if the shield bodies ~ ~1 rJ ~ r~

43a and 43b are provided, the magnetic field 45 can be put between the shield bodies 43a and 43b. Thus, the magnetic sensor 71a can detect the leakage magnetic flux of the magnetic field 45 due to the defect 44, without being influenced by the shield bodies 43a and 43b. That is, the defect 44 can be detected with a high S/N.
In order to achieve the second object, in the magnetic inspection apparatus of the present invention, the distance between the shield bodies is set to be 2.2 times or more, and 2.8 times or less, the distance between the magnetic sensor and the object.
In addition, each shield body has an L-cross sec-tion composed of a vertical portion and a horizontal portion, and the width of the horizontal portion in the direction of arrangement of the magnetic poles is 0.4 time or more, and 0.6 time or less, the distance between the magnetic poles.
When the distance A between the magnetic sensor and the shield body is too large, the ratio of a component of the floating magnetic flux in a defect-free portion, which is not shielded by the shield body and reaches the magnetic sensor, increases. If the distance A between the magnetic sensor and the shi.eld body is decreased, the ratio of the component of the floating magnetic flux, which reaches the magnetic sensor, decreases.
However, if the distance A becomes too small, the leakage magnetic flux due to the defect does not easily 2~7~77`:~

reach the magnetic sensor. On the other hand, the ratio of the component of the magnetic flux, which reaches the magnetic sensor, depends largely on the lift-off L
represented by the distance between the magnetic sensor and the object.
Accordingly, in this invention, the relationship between the distance 2A between the shield bodies and the lift-off L is defined by 2.2 < 2A/L ~ 2.8 ... (1) Thereby, the ratio of the leakage magnetic flux to the floating magnetic flux in the magnetic flux crossing the magnetic sensor can be increased, and the leakage magnetic flux can efficiently be detected. The value of formula (1) was calculated on the basis of the result of a computer simulation conducted on the magnetic field at the position of the magnetic sensor, by making and using a test model.
The relationship between the width W of the hori-zontal portion of each L-cross sectional shield body and the inter-magnetic pole distance B of the magnetizer will now be described.
Specifically, the floating magnetic flux and leak-age magnetic flux reaching the magnetic sensor are attenuated by the presence of the shield bodies. If the ratio (W/B) of the width W of each shield body to the inter-magnetic pole distance B is varied, the atte-nuation amount varies. The attenuation amount differs 7 - ~
-- l2 --between the floati.ng magnetic flux and the leakage magnetic flux. Accordingly, a range in which the att-enuation amount of the floating magnetic flux is large and the attenuation amount of the leakage magnetic flux is small is found. In this invention, the ratio (W/B) is defined by 0.4 < W/B ~ 0.6 ...(2) Thereby, the ratio of the leakage magnetic flux to the floating magnetic flux in the magnetic flux crossing lo the magnetic sensor can be increased, and the leakage magnetic flux can efficiently be detected.
In order to achieve the third object, in the mag-netic inspection apparatus of the present invention, a plurality of magnetic sensors are arranged at regular intervals in the width direction of the running object.
A difference signal between output signals from those ones of said magnetic sensors which are separated by a predetermined distance is calculated by a corresponding subtraction circuit. An absolute value of each differ-ence signal output from each subtraction circuit is cal-culated by a corresponding absolute value circuit. An arithmetic operation circuit evaluates the defect of the object on the basis of the output signal from each absolute value circuit.
For example, in the case of the object of a thin steel strip, etc., local non-uniformity in magnetic permeability occurs in the to-be-inspected body due to 2 ~ rf internal stress, non-uniformity in material quality, a variation in thickness of the object, etc. caused at the time of processing the steel strip. Accordingly, a variation component of leakage magnetic flux due to non-uniform magnetic permeability is included as noise in the detection signal of the magnetic sensor, even if the defect does not present.
In general, an area of non-uniformity of magnetic permeability is much greater than an area of a defect.
Thus, the variation component of the leakage magnetic flux due to non-uniform magnetic permeability is detected simultaneously by a number of adjacent magnetic sensors. On the other hand, the leakage magnetic flux due to the defect is detected by a smalI number of magnetic sensors, e.g. on or two sensors. Thus, if a difference signal representing a difference between the output signals from the,magnetic sensors separated by a predetermined distance is obtained, noise component due to non-uniform magnetic permeability can be removed from the difference signal. Therefore, the S/N of the output signal of the magnetic sensor ls improved.
Brief Description of the Drawings Fig. 1 is a cross-sectional view showing a magnetic inspection apparatus according to an embodiment of the present invention, taken along a plane parallel to the direction in which a thin steel strip runs;
Fig. 2 is a cross-sectional view showing the r;J ~

apparatus, taken along a plane perpendicular to the direction in which the thin steel strip runs;
Fig. 3 is a side view showing the apparatus as built in a support apparatus;
Fig. 4 is a partly enlarged view showing an important portion of the apparatus;
Fig. 5 is a schematic diagram for illustrating the operational principle of the present invention;
Fig. 6 shows the positional relationship between a leakage flux and shield bodies, for illustrating the operational principle of the invention;
Fig. 7 shows a simulation result of a magnetic flux intensity near a magnetic sensor when the shield body of the apparatus of the embodiment is not provided;
Fig. 8 shows a simulation result of a magnetic flux intensity near the magnetic sensor when the shield body of the apparatus of the embodiment is provided;
Fig. 9 shows a simulation result of the magnetic flux intensity near the magnetic sensor when the shield body of the apparatus, which has an L-cross section, is provided;
Fig. lOA shows a waveform at the time the apparatus is not provided with the shield plate;
Fig. lOB shows a waveform at the time the apparatus is provided with the shield plate;
Fig. 11 shows the relationship between the floating magnetic flux and the presence/absence of the shield ~7~.3 ~ 7 1 plates in the apparatus;
Fig. 12 shows the relationship between the magnetic sensor output and the presence/absence of the shield plates in the apparatus;
Fig. 13 shows the positional relationship between the distance between the magnetic poles of the magnet-izer and the distance between the shield bodies;
Fig. 14 shows the relationship between the distance between the magnetic poles of the magnetizer and the shape of the shield body;
Fig. 15 shows the relationship between the shield distance of the shield shape shown in Fig. 13 and the S/N of the output signal from the magnetic sensor;
Fig. 16 shows the relationship between the shield distance of the shield shape shown in Fig. 13 and the S/N of the output signal from the magnetic sensor;
Fig. 17 shows -the relationship between the shield distance of the shield shape shown in Fig. 13 and the S/N of the output signal from the magnetic sensor;
Fig. 113 shows the relationship between the inter-magnetic pole distance of the shield shape shown in Fig. 13 and the optimal shield distance;
Fig. 19 shows the relationship between the inter-magnetic pole distance of the shield shape shown in Fig. 14 and the S/N of the output signal from the magnetic sensor in a different shield mode;
Fig. 20 shows a cross-sectional view showing 2 i~ r~ ~ ~ rJ r a schematic structure of a magnetic inspection apparatus accordirlg to another embodiment of the invention;
Fig. 21 is a front view of an important portion of the apparatus;
Fig. 22A is a cross-sectional view showing a schematic structure of a magnetic inspection apparatus according to still another embodiment of the invention;
Fig. 22B is a perspective view of the apparatus;
Fig. 23 is a cross-sectional view showing a magnetic inspection apparatus according to still another embodiment of the invention;
Fig. 24 is a cross-sectional view showing a magnetic inspection apparatus according to still another embodiment of the invention;
Fig. 25 is a partly enlarged view showing an important portion of the apparatus;
Fig. 26 shows a test model for use in a simulation for obtaining numerical values in the invention;
Fig. 27 shows an arrangement of shield plates for use in the simulation;
Fig. 28 shows an arrangement of other shield plates for use in the simulation;
Fig. 29A shows a magnetic field characteristic indicating the simulation result associated with the shield shape shown in Fig. 27;
Fig. 29B shows a magnetic field charac-teristic indicating the simulation result associated with the J 7 7 ~
"

shield shape shown in Fig. 27;
Fig. 30 shows another magnetic field characteristic indicating the simulation result;
Fig. 31A shows a floating magnetic field character-istic indicating the simulation result associated withthe shield shape shown in Fig. 28;
Fig. 31s shows a defect magnetic field characteris-tic indicating the simulation result associated with the shield shape shown in Fig. 28;
Fig. 32 shows a relative floating magnetic field characteristic indicating the simulation result;
Fig. 33 shows an optimal characteristic indicating the simulation result;
Fig. 34 shows an arrangement of magnetic sensors in a magnetic inspection apparatus according to still another embodiment of the invention;
Fig. 35 is a block diagram showing an electric structure of the apparatus;
Fig. 36 shows the relationship between the position of the magnetic sensors and the output in the apparatus;
Fig. 37 shows the relationship between the position of the magnetic sensors and the output in a conventional apparatus;
Fig. 38 shows the relationship between the interval of the magnetic sensors and the output in the apparatus of the embodiment;
Fig. 39 is a cross-sectional view showing 2 ~ 7 7 1 - l8 --a conventional magnetic inspection apparatus, taken along a plane parallel to the direction in which a thin steel strip runs;
Fig. 40 is a cross-sectional view showing the conventional apparatus, taken along a plane perpendicu-lar to the direction in which the thin steel strip runs;
Fig. 41 is a side view showing the apparatus as built in a support apparatus;
Fig. 42 is a diagram for illustrating the problem in the conventional apparatus; and Fig. 43 shows the relationship between a vertical component of a floating magnetic flux and the speed at which the thin steel plate runs in the conventional apparatus.
Best Mode of Carryinq Out the Invention An embodiment of the present invention will now be described with reference to the accompanying drawings.
Figs. 1, 2 and 3 are cross-sectional views showing a magnetic inspection apparatus according to the embodi-ment which is incorporated in an inspection line in a factory. The same parts as in the conventional magnetic inspection apparatus shown in Figs. 39, 40 and 41 are denoted by like reference numerals. Thus, detailed descriptions of the common parts are omitted.
In this embodiment, a thin steel strip 10 or 7~t an object to be inspected is interposed between upper and lower hollow rolls 1 and la. In Fig. 3, two hori-zontal arms 12 and 12a are supported within a frame 11 by means of spring members 13a, 13b, 13c and 13d. Thus, the horizontal arms 12 and 12a are vertically movable.
Stationary shafts 2 and 2a of the magnetic inspection apparatus are fixed at middle parts of the horizontal arms 12 and 12a. A pair of guide rolls 14a and 14b for guiding the thin steel strip 10 between the hollow rolls 1 and la of the magnetic inspection apparatus are pro-vided on both sides of the frame 11.
In Figs. 1 and 2, an end portion of the stationary shaft 2 penetrates a center shaft of the lower hollow roll 1 made of a non-magnetic material. The stationary shaft 2 is rotatably supported by a pair of rolling bearings 3a and 3b such that the shaft 2 is situated along the center axis of the hollow roll 1. Thus, the hollow roll 1 is freely rotatable about the stationary shaft 2.
Within the hollow roll 1, a magnetizing core 4c around which a magnetizing coil 6 of a magnetizer 4 is wound is fixed to the stationary shaft 2 via a support member 5, such that magnetic poles 4a and 4b are situ-ated close to the inner peripheral surface of the hollow roll 1.
On the other hand, the upper hollow roll la, which is provide above the lower hollow roll 1 with the ~ ~, 7 ~ ~ r! ~

thin steel strip 10 interposed, is rotatable about a stationary shaft 2a. When the thin steel strip 10 runs in the direction a, the roll la rotates in the direction c. A magnetic sensor array 7 is fixed to the stationary shaft 2a of the hollow roll la via a support rod 21a so as to face the magnetic pol.es 4a and 4b of the magnetizer 4 housed in the lower hol].ow roll 1. The magnetic sensor array 7 cornprises magnetic sensors 7a arranged linearly in the width direction of the thin steel strip 10. Signal cables of the magnetic sensors 7a are led out through the stationary shaft 2a.
One shield plate 22a is situated on one side of the magnetic sensor array 7 so as to face the magnetic pole 4a. Similarly, the other shield plate 22b is situated on the other side of the sensor array 7 so as to face the magnetic pole 4b. The shield plates 22a and 22b are made of a material having high magnetic permeability and low coercive force. In this embodiment, the shield plates 22a and 22b are made of Permalloy. The shield plates 22a and 22b are fixed to the center shaft 2a by means of support members (not shown).
Fig. 4 is an enlarged view of an important portion.
Each magnetic sensor 7a of the magnetic sensor array 7 is a saturable-type magnetic sensor formed by winding a detection coil around a rod-shaped core of ferromagnetic material. The height H of each shield plate 22a, 22b is greater than the length D of the rod-shaped core of each 2~9~

magnetic sensor 7a. The lower ends of the shield plates are substantially on a level with the lower end of each magnetic sensor 7a. It suffices if the height H is at least 1/2 of the length D of the rod-shaped core.
In this embodiment, the length D of each magnetic sensor 7a is 5 mm. The height H of each shield plate 22a, 22b is 16 mm, and the thickness of each shield plate is 0.2 mm. Each shield plate 22a, 22b is sepa-rated from the center axis of each magnetic sensor 7a by 4 mm (A = 4 mm). A lift-off or a distance between each magnetic sensor 7a and each shield plate 22a, 22b, on the one hand, and the thin steel strip 10, on the other, is 3.6 mm. The length of each shield plate 22a, 22b in the width direction of the thin steel strip 10 is greater than the width (length) of the magnetic sensor array 7.
In the magnetic inspection apparatus having the above structure, when the thin steel strip 10 is trav-eled in the direction a while it is clamped between the hollow rolls 1 and la under a predetermined pressure, the hollow rolls 1 and la rotate in the directions b and c.
In this state, an excitation current is supplied to the magnetizing coil 6 and a closed magnetic path is formed by the running thin steel strip 10 and the mag-netic poles 4a and 4b of the magnetizing core 4c stored in the lower hollow roll 1. When the thin steel strip - 22 - '~ 977 i 10 has an internal or surface defect, a leakage magnetic flux occurs. The leakage magnetic flux is detected by the magnetic sensor 7a of the sensor array 7 housed in the upper hollow ~oll la, which magnetic sensor 7a cor-responds to the position of the defect. This magneticsensor 7a outputs a detection signal.
In the above structure, the gravitational force of the thin steel strip 10 is not directly exerted on the upper hollow roll la. Thus, the thickness of the upper hollow roll la is made less than that of the lower hol-low roll 1. Accordingly, the distance between the mag-netic sensor array 7 and the thin steel strip 10 can be decreased, and the detection sensitivity of the sensor array 7 can be enhanced.
Figs. 7 and ~ are graphs showing computer simula-tion results, indicating floating magnetic flux distri-butions obtained before and after the shield plates 22a and 22b as shown in Fig. 4 are provided. In each graph, the original point P is an intersection between the upper surface of the thin steel strip 10 (in Fig. 4) and the center line between the magnetic poles 4a and 4b.
The horizontal axis indicates the horizontal distance X
(unit = mm), and the right end point Q corresponds to the tip point of the magnetic pole 4b. The vertical axis indicates the vertical distance (unit = mm). Solid lines indicate magnetic force lines produced between the magnetic poles 4a and 4b.

- 23 ~ 7 ~ ~1 In Figs. 7 and 8, a, b ... _ indicate vertical locations near the magnetic sensor 7a at the distance X = 5 mm. The values a to f stated on the right side of each graph indicate vertical components of magnetic s field (floating magnetic field) at the corresponding locations. Accordingly, the (-) signs indicate downward magnetic field.
The vertical magnetic field at the distance X = o is symmetric and becomes 0 in the simulation result.
Thus, the respective values are obtained at the distance X = 1 mm.
Fig. 7 shows the case where the shield plate 22b is not provided, and Fig. 8 shows the case where the shield plate 22b is provided. As shown in the figures, when the shield plate 22b is provided, it is understood that the vertical component of the floating magnetic flux intersecting at the respective locations of the magnetic sensor 7a is remarkably reduced.
Fig. 9 shows the case where a shield plate 23 hav-ing an L-cross section was used as a shield body. As shown in Fig. 9, the shield plate 23 comprises a verti-cal portion 24a having the same shape as the shield plate 22b shown in Fig. 4, and a horizontal portion 24b extending outwardly from the lower end of the vertical portion 24a. The horizontal width W of the horizontal portion 24b is 2 mm.
It is suggested that by attaching the horizontal portion 24b to the lower end of the shield plate 23, the floating magnetic flux extending to the inside of the shield plate 23 from the lower end region of the plate 23 can be remarkably reduced.
Based on the above simulation results, two types of shield plates 22a, 22b and 23 were manufactured and arranged on both sides of the magnetic sensor array 7.
Standard defective samples having artificial defects (through holes) of 0.2 mm~, 0.3 mm~, 0.6 mm~ and 0.9 mm~
were substituted for the thin steel strip 10, and the defects of the samples were inspected. Thus, the inspection results were obtained.
Fig. lOA shows an output waveform of each magnetic sensor 7a before the shield plates are not provided, and Fig. lOB shows an output waveform of each magnetic sen-sor 7a after the shield plates 22a and 22b shown in Fig. 4 are provided. The sensitivity of the magnetic sensor 7a at the time the waveform of Fig. lOB was observed is set to be higher that the sensitivity at the time the waveform of Fig. lOA was observed.
Regarding the waveform of Fig. lOA obtained in the case where the shield plates are not provided, when the magnetizing force is increased, the magnetic sensor is saturated due to the increase in floating magnetic flux.
However, if the magnetizing current is decreased, the output values (amplitudes) corresponding to the magni-tude of defects of 0.2 to 0.9 mm~ are not obtained. In 1 i) r~

addition, the S/N is low. By contrast, in the case of the waveform of Fig. lOB, the shield plates 22a and 22b are attached, and the output values (amplitudes) corre-sponding the magnitude of defects are obtained and the S/N is remarkably enhanced.
Fig. 11 shows actual measurement data on the intensity of the floating magnetic flux detected by each magnetic sensor 7a in the case where the defect-free thin steel strip 10 was inspected by the apparatus of the embodiment. The measurement was conducted under the conditions that no shield plate was provided, the shield plates 22a and 22b shown in Fig. 4 were provided, and the shield plate 23 having the L-cross section was provided.
lS As shown in the figure, by providing the shield plates, it is understood that the output voltage due to floating magnetic flux detected by each magnetic sensor 7a is remarkably decreased. As stated above, the leak-age magnetic flux due to the defect is hardly influenced by the presence of the shield plate. As a result, the ratio of the leakage magnetic flux due to the defect to the floating magnetic flux increases, and the defect detection S/N increases.
Fig. 12 is a graph showing the output ratio (Vs/Vi) between the output voltage Vi due to the floating mag-netic flux of the defect-free portion and the output voltage vs obtained by compounding the leakage voltage 2~3~!7 ~

due to a defect in the case where the defect exists and the output voltage Vi due to the floating magnetic flux.
The output ratio ~vs/vi) was experimentally obtained under the aforementioned measurement conditions.
As can be seen from this graph, by providing the shield plate, the ratio of the output voltage due to the leakage magnetic flux contained in the entire output voltage (compound voltage vs) of the magnetic sensor 7a increases.
It was thus proved by the measurement results shown in Figs. lOA to 12 that the simulation results shown in Figs. 7 to 9 are correct.
Accordingly, even if the generated magnetic field generated by the magnetizer is increased to increase the leakage magnetic flux due to the defect, most of the floating magnetic flux of the defect-free portion resulting from the increase in the magnetic field is shield by the shield plates 22a, 22b, 23 and does not reach the magnetic sensor array 7. Accordingly, by increasing the generated magnetic field, only the leakage magnetic flux can be increased and the S/N of the output voltage due to the defect detected by each magnetic sensor 7a is remarkably increased.
Furthermore, since the output voltage is not saturated, the output voltage corresponding to the mag-nitude of the defect is obtained. Thus, a small defect in the thin steel strip 10 can be detected with high 2 ~ 7 ~ ~ 7 -1 precision.
Besides, when the speed at which the thin steel strip 10 is increased, the vertical component of the floating magnetic flux increases, as shown in Fig. 43.
However, the floating magnetic flux itself input to each magnetic sensor 7a decreases. Thus, even if the speed of the thin steel strip 10 is increased, the S/N does not decrease in particular. Therefore, by increasing the speed of the strip 10, the inspection efficiency can be enhanced.
However, if the distance 2A between the shield plates 22a and 22b is set to be too small, not only the floating magnetic flux but also the leakage magnetic flux due to the defect in the thin steel strip 10 is shielded. As a result, the S/N of the output signal of the magnetic sensor 7a is lowered, to the contrary.
Inversely, if the distance 2A between the shield plates 22a and 22b is too large, the floating magnetic flux is shielded. Thus, an optimal value should be present with respect to the distance 2A.
Under the circumstances, the inventors experimen-tally produced a test model of the magnetizer 34 having the shape as shown in Fig. 13. The distance B between magnetic poles 34a and 34b is variable between 10 mm, 20 mm and 30 mm. The vertical component of the magnetic density detected by the magnetic sensor 7a situated at the center between the magnetic poles was calculated by 3 ~ ~ 1 - 2~ -the aforementioned computer simulation in the case where the distance 2A between the shields was variously changed.
It is supposed that the defect is present at the center of the thin steel strip 10. As stated above, since the vertical magnetic field on the center line between the magnetic poles cannot be calculated, the vertical component of the magnetic flux density at a point P displaced towards the magnetic pole from the center by 1 mm (x = l mm) is calculated. And the S/N of the vertical component is calculated.
Specifically, S indicates the differential magnetic flux density between the leakage magnetic flux due to the defect and the floating magnetic flu`x of the defect-free portion of the thin steel strip lO, and N indicatesthe variation in floating magnetic flux density of the entire thin steel strip in the case where the magnetic sensor 7a oscillates in the X-direction by + 0.2 mm with respect to the point P as a central point. That is, N
is noise due to the leakage magnetic flux which occurs when the magnetic sensor 7a oscillates owing to the thin steel strip lO running on the rolls and by the rolls.
Fig. 15 is a characteristic graph showing the vari-ation in the S/N of the vertical component at point P inthe case where the distance 2A between the shield plates is varied in the state in which the inter-magnetic pole ~ ~ 7 .il ~' 7 .~

distance B of the magnetizer 3'L is set at 10 mlll. The horizontal axis is set at the value (2A/B) obtained by normalizing the inter-shield plate distance 2A by the inter-magnetic pole distance B.
As can be seen from Fig. 15, in the case where the inter-magnetic pole distance B is 10 mm, the S/N takes a maximal value when the inter-shield plate distance 2A
normalized ~y the distance B is about 0.5.
Fig. 16 is a characteristic graph showing the variation in the S/N in the case where the distance B
is 20 mm. In the case where the inter-magnetic pole distance B is 20 mm, the S/N takes a maximal value when the inter-shield plate distance (2A/B) is about 0.4.
Similarly, Fig. 17 is a characteristic graph show-ing the variation in the S/N in the case where the inter-magnetic pole distance B is 30 mm. In the case where the distance B is 30 mm, the S/N takes a maximal value when the inter-shield plate distance (2A/B) is about 0.3.
Referring to Figs. 15, 16 and 17, it is assumed that there is a optimal inter-shield plate distance [2A]b. Fig. l.8 is a graph showing the relationship between the inter-magnetic pole distance B and the inter-shield plate distance 2A at which the maximal S/N
is obtained when the distance B is 10 mm, 20 mm and 30 mm, i.e. the optimal inter-shield plate distance [2A]b. The optimal inter-shield plate distance [2A]b 7 7 ~

and the inter-magnetic pole distance B meet the relationship of equation (3):
[2A]b = -O.OlB ~ 0.6 . (3) wherein the unit of 2A and B is mm.
Accordingly, if the inter-shield plate distance [2A] is set to meet equation (3) in relation to the inter-magnetic pole distance B, the optimal S/N is obtained. In the actual apparatus, if the distance [2A]
meets formula (4), a practically sufficiently high S/N
is obtained:
-O.OlB + 0.55 ~ 2A < -O.OlB + 0.65 ...(4) The shape of the shield body provided on that side of the magnetic sensor 7a which faces the magnetic pole 4a, 4b will now be considered.
Possible shapes of the shield body are a pair of shield plates 22a, 22b shown in Fig. 13, shield plate 23 having the L-cross section shown in Fig. 9, and a shield body 25 having a U-cross section as shown in Fig. 14.
Specifically, the shield body 25 can magnetically shield not only the sides of the magnetic sensor 7a but also that face of the sensor 7a which is not opposed to the thin steel strip 10. Accordingly, a floating magnetic flux parallel to the thin steel strip 10 enters one of the vertical portions of the magnetic shield body 25, passing through an intermediate horizontal portion and goes out of the other vertical portion. As a result, it becomes difficult for the floating magnetic flux - 31 - 2~3771 to enter the inside of the shield body 25 having the - U-cross section. Thus, the S/N of the output signal of the magnetic sensor 7a is further enhanced.
Suppose that the distance between the vertical portions of the shield body 25 is 2A, like the distance between the shield plates 22a and 22b shown in Fig. 13.
Using the shield body 25 with the U-cross section, the relationship between the optimal distance [2A]b and the inter-magnetic pole distance B for obtaining the optimal SJN was examined by the same method as with the shield plates 22a and 22b shown in Fig. 13. Like the [S/N]22 of the optimal shield distance [2A]b of the shield plates 22a and 22b shown in Fig. 13, [S/N]25 of the optimal distance [2A]b of the shield body 25 shown in Fig. 14 was obtained with respect to each inter-magnetic pole distance B.
For each inter-magnetic pole distance B, the ratio ([S/N]25/[S/N]22) of the optimal [S/N]25 of the shield body 25 to the optimal [S/N] of the shield plates 22a and 22b was calculated. Fig. 19 is a graph showing the calculation result. As can be seen from the character-istic graph of Fig. 19, when the inter-magnetic pole distance B is less than a limit distance of 29.3 mm, a higher S/N is obtained with the shield body 25 having the U-cross section shown in Fig. 14. Inversely, when the distance B exceeds 29.3 mm, a higher S/N is obtained with the shield plates 22a and 22b arranged in parallel 2~7~7'~

as shown in Fig. 13.
The reason why the inversion phenomenon of the S/N
characteristic occurs at a specific value of the inter-magnetic pole distance B is as follows.
When the distance B is less than 29.3 mm, the floating magnetic flux parallel to the thin steel strip 10 is greatly attenuated while passing through the hori-zontal portion of the shield body 25 with the U-cross section. However, when the distance B exceeds 29.3 mm, the magnetic flux extending normally towards the thin steel plate 10 is biased to the shield body 25.
Consequently, the floating magnetic flux of the sensor unit increases, and the S/N decreases.
Accordingly, the limit distance 29.3 mm is defi-nitely determined when the width 2A of the shield plates 22a and 22b, or shield body 25, in the direction of arrangement of the magnetic poles and the inter-magnetic pole distance B meet the condition of equation (3).
Fig. 20 is a cross-sectional side view showing an important portion of a magnetic inspection apparatus according to another embodiment of the present invention, and Fiy. 21 is a front view of the important portion of this apparatus. The same parts as in the embodiment of E~ig. 1 are denoted by like reference numerals. Thus, detailed descriptions of the common parts are omitted.
The magnetic sensor array 7 arranged along the 2 ~ 7 ~ r! 7 i.

center axis between the magnetic poles 4a and 4b housed within the hollow roll 1 and the shield plates 22a and 22b provided on both sides of the magnetic sensor array 7 are attached to a support frame 51 situated along the axis of the hollow roll 1. The support frame 51 is sup-ported on the upper surface of the thin steel strip 10 by means of a number of rollers 52 and 53 arranged in two lines along the axis of the hollow roll 1 via sup-port arms 54 and 55.
Thus, the lift-off L ( or the distance) between the magnetic sensors 7a of the magnetic sensor array 7 and the surface of the thin steel strip 10 is kept at a constant value determined by the outside diameter of each roller 52, 53 and the shape of the support frame 51. The horizontal movement of the support frame 51 is restricted by a support mechanism (not shown) so as to keep a predetermined relationship between the magnetic sensors 7a and the magnetic poles 4a and 4b.
According to this magnetic inspection apparatus with this structure, the outside diameter of each roller 52, 53 is much smaller than that of the hollow roll la shown in Fig. 1. Thus, the moment of inertia of each roller 52, 53 is much smaller than that of the hollow roll la shown in Fig. 1. Therefore, it is not necessary to increase the strength of the support frame 51 and rollers 52 and 53 in particular.
In addition, the maintenance of the magnetic sensor 2~)7r3 ~7'1 - 3~ -array 7 is easier than in the case where the array 7 is housed within the hollow roll la.
Fig. 22A is a cross-sectional front view showing an important portion of a magnetic inspection apparatus according to still another embodiment of the present invention, and Fig. 22s shows an external appearance of this apparatus. The same parts as in the embodiment of Fig. l are denoted by like reference numerals, and detailed descriptions of the common parts are omitted.
According to this embodiment, a pair of rollers 59 and 60 of, e.g. rubber are attached via shafts 57 and 58 on both sides of a support frame 56 storing the magnetic sensor array 7 facing the thin steel strip 10 and the shield plates provided on both sides of the array 7.
The rubber rollers 59 and 60 are put in contact with outer peripheral edge portions of the hollow roll l.
In the magnetic inspection apparatus with this structure, the lift-off L (or the distance) between the magnetic sensor array 7 and the surface of the thin steel strip 10 is kept at a constant value determined by the outside diameter of each roller 59, 60 and the shape of the support frame 51. Accordingly, the mag-netic sensor array 7 is not influenced by vibration due to running of the thin steel strip lO.
Like the embodiment of Figs. 20 and 21, the hori-zontal movement of the support frame 56 is restricted by a support mechanism (not shown) so as to keep -- 35 - 2~7~77:~

a predetermined relationship between the magnetic sensor array 7 and the magnetic poles 4a and 4b.
Fig. 23 is a cross-sectional side view showing an important portion of a magnetic inspection apparatus according to still another embodiment of the present invention. The same parts as in the embodiment of Fig. 1 are denoted by like reference numerals, and detailed descriptions of the common parts are omitted.
In this embodiment, the thin steel strip 10 is wound around the hollow roll 1 storing the magnetizer 4 over about 9oo with a predetermined tensile force applied. A support frame 61 storing the magnetic sensor array 7 and shield plates 22a and 22b is situated outside the hollow roll 1 so as to face`the magnetic poles 4a and 4b of the magnetizer 4.
According to the magnetic inspection apparatus having the above structure, the angle of contact of the thin steel strip 10 with the hollow roll 1 is very large, i.e. 90. Thus, vibration of the thin steel strip 10 is remarkably damped while the steel strip 10 is being in contact with the hollow roll 1. Therefore, the llft-off L between the magnetic sensor array 7 and steel plate 10 can be decreased, and the magnetic inspection sensitivity and S/N can be enhanced.
Fig. 24 is a cross-sectional view schematically showing the structure of the magnetic inspection appara-tus according to still another embodiment of the - 36 - ~ ~ 7 ~J ~ 7 !1 invention. The same parts as in the embodiment of Fig. 1 are denoted by like reference numerals, and detailed description of the common parts is omitted.
In this embodiment, shield plates 23a and 23b each having an L-cross section are supported by support mem-bers (not shown) on both sides of the magnetic sensor 7a in the upper hollow roll la, which are opposed to the magnetic poles 4a and 4b. Each of the shield plates 23a and 23b has the same shape as the shield plate 23 shown in Fig~ 9. As shown in Fig. 2s, each of the shield plates 23a and 23b has a vertical portion 24a with a height H in a direction perpendicular to the thin steel strip 10, and a horizontal portion 24b with a width W in a direction parallel to the steel strip`10. Each shield plate 23a, 23b has a thickness t. Each shield plate 23a, 23b is situated at a distance A from the magnetic sensor array 7. The lower ends of the shield plates 23a and 23b are substantially on a level with the lower end of each magnetic sensor 7a.
In this apparatus, the length D of each magnetic sensor 7a is 5 mm, the height H of each shield body 23a, 23b is 16 mm, and the thickness of each shield body is 0.2 mm. The width w of the horizontal portion 24b is 6 mm. The distance A between each shield body 23a, 23b and the center axis of each magnetic sensor 7a is 4 mm.
The lift-off L represented by the distance between the thin steel strip 10, on one hand, and each magnetic ~ ~ 7 rl 7 1 sensor 7a and each shield body 23a, 23b, on the other hand, is 3.6 mm. The length of each shield plate 23a, 23b in the axial direction of the hollow roll l is greater than that of the magnetic sensor array 7.
Further, the inter-magnetic pole distance B of the mag-netizer 4 is 15 mm.
The relationship between the inter-shield plate distance 2A and width W of horizontal portion 24b of each shield plate 23a, 23b with the L-cross section and lift-off L and inter-magnetic pole distance B will now be described. Specifically, the vertical component of the magnetic field obtained by the magnetic sensor 7a when the dimensions and position of the shield plates 23a and 23b are varied were computer-simulated.
Suppose a test model of the magnetizer 34 with the shape as shown in Fig. 26. Like the magnetizer shown in Figs. 13 and 14, the distance B between the magnetic poles 34a and 34b is variable between lO mm, 20 mm and 30 mm. The thickness of each steel plate forming a mag-netizing core 34c is 0.4 mm. The current density of a current flowing in a magnetizing coil 36 is 1.25 x 105 A/m2 is 3 mm. A center point o of the thin steel strip 10 or the ob;ect to be inspected is the original point of (X, Y) coordinates.
Suppose that the shield plates 22a and 22b having only vertical portions without horizontal portions, as shown in Fig. 27, and the shield plates 23a and 23b 7 7 :~
- 3~ --having both vertical portions 24a and horizontal portions 24b, as shown in Fig. 28, are located in the (x~ Y) coordinates. The length D of the magnetic sensor 7a provided between the shield plates 22a and 22b (23a and 23b) is 5 mm.
Suppose two samples of the object to be inspected (i.e. the thin steel strip lO), one being free of defects and the other having a standard defect of 0.4 mm in diameter.
Using the test models of the magnetizer and shield plates, the vertical component of the magnetic field was calculated, which is obtainable by the magnetic sensor 7a while varying the parameters: the distance A between the shield plates 22a to 23b and the center of the mag-netic sensor 7a, the lift-off L, the inter-magnetic pole distance B, the presence/absence of shield plates 22a to 23b, and the presence/absence of a defect in the to-be-inspected object. Since the magnetic sensor 7a has the length D (= 5 mm)~ the vertical component at every l mm in the vertical direction is integrated from 0 mm to 5 mm, thereby obtaining the vertical component of the magnetic field. The vertical component of the magnetic field at distance X = o is replaced with the values at X = 1 mm. Further, when the sample with the standard defect is used, it assumed that the defect is located at the original point O.
Fig. 29A shows a ratio ~ between a magnetic field 2~7~377 1 (vertical component) ~1 in the case where the shield plates 23a and 23b are provided and a magnetic field (vertical component) ~2 in the case where the shield plates 22a and 22b are not provided, under the condition that the sample with the standard defect is used. ~he horizontal axis of Fig. 29A indicates (A/L). That is, Fig. 29A shows the relative ratio ~ of the defect mag-netic field, depending on the presence/absence of the shield.
On the other hand, Fig. 29B shows a ratio a between a magnetic field (vertical component) al in the case where the shield plates 22a and 22b are provided and a magnetic field (vertical component) a2 in the case where the shield plates 22a and 22b are not provided, under the condition that the defect-free sample is used. The horizontal axis of Fig. 29B indicates (A/L). That is, Fig. 29B shows the relative ratio a of the defect mag-netic field, depending on the presence/absence of the shield.
A relative ratio Rl (= ~/a) between the defect mag-netic field indicated by relative ratio a of Fig. 29s and the floating magnetic field indicated by relatlve ratio ~ of Fig. 29A is calculated. Fig. 30 shows the relative ratio R1 (= ~/a), with the horizontal axis thereof indicating (A/L). The relative ratio Rl in the state in which the shield plates 22a and 22b are not provided is set at 1.

~ ~ r~ r~ r/ ~ ~
- ~o -Specifically, Fig. 30 shows the ratio Rl between the magnetic field (~)~ in which the leakage magnetic flux due to the defect and the floating magnetic flux are added, and the magnetic field (a) of only the float-ing magnetic flux. Thus, the higher the relative ratioRl, the higher the ratio of the leakage magnetic field due to the defect included in the magnetic field detected by the magnetic sensor 7a.
Accordingly, it is understood that the optimal range of the ratio (A/L) of the half distance A of the distance 2A between the shield plates 22a and 22b to the lift-off L is 1.1 to 1.4 expressed by formula (l).
Fig. 31A shows a relative ratio R2 ( = yl/y2 ) of a floating magnetic flux yl in the magnetic sensor 7a, in the case where the L-cross sectional shield plates 23a and 23b with horizontal portions 24b are used, to a floating magnetic flux y in the case where the shield plates 23a and 23b are not provided under the same condition, when the defect-free sample is used. The horizontal axis of Fig. 31A indicates (W/B). That is, Fig. 31A shows a relative ratio R2 of the floating magnetic flux, depending on the presence/absence of the shield.
The direction of the vertical component of the magnetic field intersecting the magnetic sensor 7a may changed to an upward direction or a downward direction, depending on the presence/absence of the shield plates 2 ~ J ~ ;l 23a and 23b. As a result, when the relative ratio R2 ~= yl/y2) is calculated, the calculated value may be a ~-) value. Thus, the region of the (-) value is evalu-ated by an absolute value.
Accordingly, in Fig. 31A, in the region where the relative ratio R2 (= yl/y2) is low, the ratio of the floating magnetic field in the magnetic field detected by the magnetic sensor 7a is very low due to the pres-ence of the shield plates 23a and 23b. Thus, the region where the relative value R2 (= yl/y2) is low is set to be an optimal range.
Fig. 31B shows a relative ratio R3 (= ~1/62) of a magnetic field 61 in the magnetic sensor 7a, in the case where the shield plates 23a and 23b are provided, to a magnetic field ~2 in the case where the shield plates 23a and 23b are not provided, under the condition that the sample with the standard defect is used. The hori-zontal axis of Fig. 31B indicates (W/B). Accordingly, Fig. 31B shows the relative value R3 of the magnetic field including the floating magnetic flux and the leak age magnetic flux due to the defect, deyendlng on the presence/absence of the shield.
Specifically, Fig. 31B shows that the ratio of the leakage magnetic flux in the det~cted magnetic field becomes higher in accordance with the increase in rela-tive ratio R3 (= ~ 2). It is thus understood that a lower (W/B) is advantageous.

- ~2 - 2~7 ~rJ'~

Comparing the characteristic of Fig. 31A and that of Fig. 31, the optimal relationship between the width W
of the horizontal portion 24b of each shield plate 23a, 23b and the inter-magnetic pole distance B is expressed by 0.4 < W/B -< 0.6 of formula (2).
Fig. 32 shows the simulation result of the vertical component of the magnetic field detected by the rnagnetic sensor 7a in the case where the defect-free sample was used and the ratio (H/D) of the height H of the magnetic shield 22a, 22b to the length D of the magnetic sensor 7a was varied. The lift-off L and inter-magnetic pole distance B were used as parameters.
According to Fig. 32, when the height H of the shield plate 22a, 22b is increased, the floating magnetic field detected by the magnetic sensor 7a decreases. For example, when a region having a magnetic field level lower than a reference level by 6 dB or more is set to be an allowable range of floating magnetic field, values of the ratio (H/D) at locations where the characteristic lines intersect the -6 dB line are Yl, Y2 and Y3. Y1, Y2 and Y3 can be approximated by the characteristic o~ Fig. 33, when the horizontal a~ls indicates the ratio (B/L). This characteristic is line-arly approximated by formula (5):
(H/D) = -0.108 (B/L) + 1.27 ........ (5) The allowable range is above the linear character-istic line of formula (5). It is thus desirable that ~ 7 (~ r~ 7 il - ~3 -the relationship ( H/D ) between the height H of the shield plate 22a, 22b and the height D of the magnetic sensor 7a be set at a value determined by formula (6) with use of the ratio (s/L) of the inter-magnetic pole distance B and lift-off L.
(H/D) > -0.1 (B/L) + 1.2 ...(6) As has been described above, the shield plates 22a to 23b are provided on both sides of the magnetic sensor 7a and the relationship between the width of each shield plate, height H of each shield plate, distance 2A
between the shield plates, inter-magnetic pole distance B of the magnetizer, lift-off L, and height D of the magnetic sensor 7a is determined by formulae (l), (2~
and (3). Thereby, the floating magnetic flux component in the magnetic field detected by the magnetic sensor 7a can be decreased, and only the leakage magnetic flux component due to the defect can be remarkably increased.
Thus, saturation of the output of the magnetic sensor 7a can be prevented, the defect detection sensitivity increased, and the detection precision enhanced.
The present invention is not limited to the above embodiments. In the apparatuses of the embodiments, the magnetic sensor array 7 is provided within the upper hollow roll la. However, the shield plates 22a and 22b, for example, may be provided on both sides of the mag-netic sensor array 7 within the hollow roll 1 of the conventional apparatus shown in Fig. 39.

rJ t`) ~

In the embodiments, the shield plates 22a and 22b or L-cross sectional shield plates 23a and 23b are provided on both sides of the magnetic sensor array 7.
However, for example, a cylindrical shield body sur-rounding the entire magnetic sensor array 7 may beprovided.
Fig. 34 shows schematically an important portion of a magnetic inspection apparatus according to another embodiment of the present invention. This apparatus has the same structure as shown in Figs. 1, 2 and 3. A num-ber of magnetic sensors 7a are arranged within the upper hollow roll 1, as shown in Fig. 34, at regular intervals S in the width direction of the thin steel strip 10.
Fig. 35 is a block diagram showing an electric structure of the apparatus of this embodiment.
Outputs from the magnetic sensors 7a are input to subtraction circuits 62, Each subtraction circuit 62 receives a pair of output signals from a pair of mag-netic sensors with another magnetic sensor interposed therebetween, arranged in the width direction of the thin steel strip 10. The subtraction circuit 62 sub-tracts one output signal from the other output signal and outputs a difference signal. Difference signals from the subtraction circuits 62 are input to absolute value circuits 63. The absolute value circuits 63 cal-culate absolute values of the input difference signals and output absolute value signals. The absolute value 7 ~ ) 7 ~ ~
~5 .

signals output from the absolute value circuits 63 are input to summing circuits 64.
Each summing circuit 64 sums absolute value signals from adjacent two absolute value circuits 63 and deliv-ers a sum signal to an arithmetic operation circuit 65.
The arithmetic operation circuit 65 calculates the defect position in the width direction of the thin steel strip 10 and defect magnitude on the basis of the signal level of each sum signal input from the summing circuits 64. The calculated defect position and defect magnitude are output to an output device (not shown) such as a CRT
display device.
In the magnetic inspection apparatus with the above structure, each subtraction circuit 62 outputs a differ-ence signal representing a difference between output signals from two magnetic sensors 7a arranged in the width direction at a distance E (=2S) from each other.
The absolute value of the difference signal is calcu-lated by the absolute value circuit 63.
As stated above, local non-uniformity in magnetic permeability occurs in the to-be-inspec-ted body due to internal stress, non-uniformity in material quality, a variation in thickness of the object, etc. caused at the time of processing the steel strip. Accordingly, a var-iation component of leakage magnetic flux due to non-uniform magnetic permeability is included as noise in the detection signal of the magnetic sensor.

~73~'7 1 - ~6 -In general, an area of non-uniformi-ty of magnetic permeability is much greater than an area of a defect.
Thus, the variation component of the leakage magnetic flux due to non-uniform magnetic permeability is S detected simultaneously by two magnetic sensors 7a. On the other hand, the leakage magnetic flux due to the defect has such a magnitude as can be detected by one magnetic sensor 7a. Thus, if a difference signal repre-senting a difference between output signals from every two magnetic sensors 7a arranged with another magnetic sensor interposed, as shown in the figures, is obtained, noise component due to non-uniform magnetic permeability can be removed from the difference signal. Therefore, the S/N of the output signal of the magnetic sensor 7a is improved.
In order to confirm the advantage of this embodiment, the inventors conducted actual inspection tests on two types of thin steel strips 10 made for tests, which have defects of 0.2 mm and 0.3 mm in diameter. The interval S of the magnetic sensors 7a is 10 mm. Thus, the distance E between the two magnetic sensors 7a whose output signals are supplied to the corresponding subtraction circuit 62 is 20 mm. Output signals delivered from the summing circuits 64 at width-directional positions X, when the position of the defectof the two types of thin steel strips for tests is dis-placed from the center of one magnetic sensor 7a in 2 ~ rJ ~ 7 1 -- ~17 --opposite directions along the width in units of 1 mm (6 mm in each direction, and 12 mm in total). Fig. 36 shows the relationship between each width-directional position of the defect and the relative value of the output signal.
Fig. 37 shows the relationship between each width-directional position of the defect and the relative value of the output signal when the subtraction circuits 62, absolute value circuits 63 and summing circuits 64 are removed and the detection signals from the magnetic sensors 7a are directly delivered to the arithmetic operation circuit 65. As shown in Figs. 36 and 37, by adopting the structure of this embodiment, the measure-ment error at the same defect position can remarkably reduced. As a result, the S/N of the output signals is greatly improved.
Regarding the arrangement of the magnetic sensors 7a shown in Fig. 34, if the distance E between the two magnetic sensors 7a for calculating the difference sig-nal is too small, both magnetic sensors 7a undesirablydetect the leakage magnetic flux due to the same defect.
Inversely, if the distance E between the magnetic sen-sors 7a is too large, it is difficult to detect a common leakage magnetic flux due to the non-uniform magnetic permeability. Accordingly, there is an optimal range of the distance E between the magnetic sensors.
Further, if the lift-off between the magnetic -- 48 ~ 3 7 7 ~7 sensors 7a and the thin steel strip 10 is too great, the S/N lowers. Thus, the optimal range of the distance E
between the magnetic sensors relates to the lift-off.
Fig. 38 shows the relationship between a value ( EL ) obtained by multiplying the lift-off L by the distance E
between magnetic sensors 7a, and a relative output of a sum signal input to the arithmetic operation circuit 65 at each value EL (mm) and a relative S/N of the sum signal.
As shown in Fig. 38, in the case where the lift-off L is constant, when the distance E increases, the rela-tive output increases up to a fixed saturation value.
However, if the distance E increases, the relative S/N
decreases to the contrary. Accordingly, if a practi-cally sufficient allowable range of the relative S/N and relative output value is 0.7 or more, an optimal range of the value (EL) is 17 to 78 expressed by formula (7):
17 ~ EL < 78 ...(7) For example, if the lift-off L is set at 3 mm, the optimal range of the distance E between the magnetic sensors is 6 to 16 mm.
In the embodiment shown ln Fig. 35, the subtraction circuits 62, absolute value circuits 63 and summing cir-cuits 64 are constituted by ordinary analog circuits.
However, these circuits may be constituted by, e.g.
digital circuits.
As described above, the distance E between the 207~77~

magnetic sensors for calculating the difference signal is set at a value determined by formula (7) in accor-dance with the lift-off L. Thereby, the noise component due to the non-uniform permeability in the thin steel strip 10 can be eliminated and a small defect in the steel strip 10 can be detected with a high S/N.

~, ., . . . . . ~ :
.. -, ...

.. . ~ .

... .. : ~.. :
: ., - ~ :
.

Claims

C L A I M S

1. A magnetic inspection method wherein a magnetizer is provided such that a pair of magnetic poles face an object to be inspected, a magnetic field is generated by the magnetizer in the object, and a leakage magnetic flux due to an internal or surface defect of the object is detected by a magnetic sensor, a shield body with a low coercive force being provided on each of the sides of the magnetic sensor which face the magnetic poles.

2. The method according to claim 1, wherein said mag-netic sensor detects a vertical component of the leakage magnetic flux due to the defect, which component inter-sects the surface of the object at right angles.

3. The method according to claim 2, wherein a shield body having a low coercive force and a .pi.-cross section surrounding said sides of the magnetic sensor which face the magnetic poles and a side of the magnetic sensor which does not face the object is used in the case where the distance between the magnetic poles is less than a limit distance definitely determined by the dimension of the shield body in the direction of arrangement of the magnetic poles, and a pair of shield plates having a low coercive force provided on the sides of the magnetic sensor which face the magnetic poles are used in the case where the distance between the magnetic poles exceeds the limit distance.

4. The method according to claim 3, wherein said limit distance is 29.3 mm.

5. A magnetic inspection apparatus comprising:
a magnetizer having a pair of magnetic poles pro-vided to face an object to be inspected, said magnetizer generating a magnetic field in the object;
a magnetic sensor for detecting a leakage magnetic flux due to an internal or surface defect of the object;
and a shield body having a low coercive force and pro-vided to surround a side of the magnetic sensor which does not face the object.

6. A magnetic inspection apparatus comprising:
a magnetizer having a pair of magnetic poles pro-vided to face an object to be inspected, said magnetizer generating a magnetic field in the object;
a magnetic sensor for detecting a leakage magnetic flux due to an internal or surface defect of the object;
and a pair of shield bodies having a low coercive force and provided on sides of the magnetic sensor which face the magnetic poles.

7. The apparatus according to claim 6, wherein said magnetic sensor is a vertical type magnetic sensor for detecting a vertical component of the leakage magnetic flux due to the defect, which component intersects the surface of the object at right angles.

8. The apparatus according to claim 6, wherein the distance between the shield bodies is 2.2 times or more, and 2.8 times or less, the distance between the magnetic sensor and the object.

9. The apparatus according to claim 6, wherein the relationship between a distance 2A (unit = mm) between the shield bodies and a distance B (unit = mm) between the magnetic poles is defined by -0.01B + 0.55 ? 2A ? -0.01B + 0.65.

10. A magnetic inspection apparatus comprising:
a magnetizer having a pair of magnetic poles pro-vided to face an object to be inspected, said magnetizer generating a magnetic field in the object;
a magnetic sensor for detecting a leakage magnetic flux due to an internal or surface defect of the object;
and a shield body having a low coercive force and a .pi.-cross section surrounding sides of the magnetic sensor which face the magnetic poles and a side of the magnetic sensor which does not face the object.

11. The apparatus according to claim 6, wherein each of said shield bodies has an L-cross section composed of a vertical portion perpendicular to the object and a hori-zontal portion attached to an end portion of the verti-cal portion, which faces the object, and extending in the direction of arrangement of the adjacent magnetic poles.

12. The apparatus according to claim 11, wherein the width of the horizontal portion of each of the shield bodies in the direction of arrangement of the magnetic poles is 0.4 time or more, and 0.6 time or less, the distance between the magnetic poles.

13. The apparatus according to claim 7, wherein the relationship between a height D of the magnetic sensor in a direction perpendicular to the object, a height H
of each shield body in the direction perpendicular to the object, a distance 2A between the shield bodies, a distance B between the magnetic poles, and a distance L
between the magnetic sensor and the object is defined by (H/D) ? -0.1 (B/L) + 1.2.

14. The apparatus according to claim 6, wherein said magnetizer is situated within a hollow roll which is rotatably supported on a stationary shaft extending at right angles with a running path of the object and which is rotated when the hollow roll is put in contact with the object running along the running path.

15. The apparatus according to claim 14, comprising a support frame with a .pi.-cross section for supporting the pair of shield bodies and the magnetic sensor provided between the shield bodies, and a plurality of rollers for supporting the support frame on the surface of the running object.

16. The apparatus according to claim 14, comprising a support frame with a .pi.-cross section for supporting the pair of shield bodies and the magnetic sensor provided between the shield bodies, and a plurality of rollers for supporting the support frame on the outer peripheral surface of the hollow roll which is rotating.

17. A magnetic inspection apparatus comprising:
a magnetizer having a pair of magnetic poles pro-vided to face a running object to be inspected;
a plurality of magnetic sensors, arranged in the width direction of the object, for detecting a leakage magnetic flux due to an internal or surface defect of the object;
a plurality of subtraction circuits each for calcu-lating a difference signal between output signals from those ones of said magnetic sensors which are separated by a predetermined distance;
a plurality of absolute value circuits each for calculating an absolute value of each difference signal output from each subtraction circuit; and an arithmetic operation circuit for evaluating the defect of the object on the basis of the absolute value signal output from each absolute value circuit.

18. The apparatus according to claim 17, wherein the relationship between a distance E (unit = mm) between the magnetic sensors associated with the output signals subjected to subtraction in each subtraction circuit and a distance L (unit = mm) between each magnetic sensor and the object is defined by 17 ? EL ? 78.