WO1988006268A1

WO1988006268A1 - Apparatus for measuring or testing dimension or contour through measuring distance

Info

Publication number: WO1988006268A1
Application number: PCT/SE1988/000050
Authority: WO
Inventors: Bengt Hialmar TÖRNBLOM
Original assignee: Individual
Current assignee: Individual
Priority date: 1987-02-18
Filing date: 1988-02-11
Publication date: 1988-08-25
Anticipated expiration: 1989-08-18
Also published as: SE456606B; EP0302099A1; SE8700659L; SE8700659D0

Abstract

Device for testing and/or measuring test objects such as rolled wire (1) with respect to quantities, e.g. dimension, form, etc. The invention illustrates how quantity can be calculated via a reverse procedure of distance measurement. By measuring the distance between the surface of the test object, or a part therefor, and reference points (RP), often fictive, pertaining/reference to the eddy current transducer, a quantity such as diameter (D) can be calculated/measured since the distance (A) between e.g. diametrically located reference points or the like is known. By combining the invention with a crack-detecting device, the same transducer arrangement, for instance, can be used to detect both cracks and dimension.

Description

Apparatus for measuring or testing dimension or contour through measuring distance.

The present invention shall be deemed a device and relates to the field of measuring and/or testing, more particularly in the field of eddy-current technique.

The invention shows how test objects can be measured and/or tested, e.g. detected, with respect to a quantity where said quantity may, for instance, be the dimension of a hot-rolled wire during the rolling processing.

When measuring the dimension of wire, rods or pipes during production, it is often necessary to measure the test object without coming into contact with it. The object may be moving quickly, e.g.20 m/s and its temperature is about 850ºC The known measuring methods most frequently used are generally based on various optical principles such as shadow measurement, etc. However, it has been found that a considerable amount of maintenance is necessary with these methods, due to the dirty environment and so on.

The present invention, based on the use of currents with associated magnetic fields induced in the test object, shows how the problems described here, and other associated problems, can be solved.

The following description shall be considered as one of many feasible embodiments enabling realization of the invention.

It is well known that the lift-off problem is a limiting and disturbing factor in the detection of cracks with the aid of the eddy-current technique. In the present invention the lift-signal is instead used, or better lift-off vector, to measure the distance between transducer and test object. Swedish patent applications 8302738-3 and 8400698-0 described, with the aid of impedance diagrams, how the impedance of the transducer is affected by the lift-off distance (LO). In so-called dynamic transformation according to 8400698-0, the LO-signal is used to control transformation and amplification of the crack signal. In the present invention several surface transducers are placed e.g. round a rolled wire as shown in Figure 1, and simultaneously measure the distance to the wire (1). Each transducer (2) may be said to correspond to one reference point (RP), as indicated in Figure 2.

Figure 1 of the drawings shows transducers (2) with surface transducer coils (4). The transducers are, for instance, arranged symmetrically around the test object (1), the cross section of the test object being shown. The transducers are connected to an electronics device (3). The distance between the transducers, or better between the coils and the surface of the test object is designated LO. If the transducers are compensated and/or balanced without any wire being present, i.e. LO = ∞, the signal (V) which is obtained either directly or indirectly from the transducer as a function of LO will have approximately the appearance shown in Figure 3.

Figure 3 thus shows the signal (V) which is obtained indirectly from the transducer as a function of LO.

If, as in this case, the transducer is compensated or balanced at infinite LO, i.e. with no wire present, the LO function curve will asymtotically approach zero with increasing LO. One might say that the signal V in Figure 3, deriving from the transducer, indirectly corresponds to the distance LO in Figure 1. By converting the function in Figure 3, e.g. via function conversion, a linear function according to Figure 4 can be obtained, where LO can correspond to the actual LO distance in Figure 2. This linear function is of course easier to utilize than the non-linear exponential function in Figure 3. Note, however, that the function according to Figure 4 need not necessary show the distance LO between transducer and test object but, if preferred, may show the distance LO to a reference position pertaining to the transducer, designated reference point (RP). The reference points of the transducers are marked in Figure 2, as well as their distance LO to the test object (1). The diameter of the test object is designated D. The distance between two diametrically located reference points is designated A. Observe that the curve in Figure 4 starts from V=O in reference point RP. The reference point RP need not be a point in the transducer but may be a fictive, imagined point at a distance LO from the surface of the test object.

Another way of describing the reference point RP is to say that RP is the point or the like at which the part-function describing the distance to the test object of its surface is equal to zero. This function can be represented by either signals or digital measured values, etc. The reference point always refers to the relevant transducer in so much as it is a function, direct or indirect, of the position of the transducer. One might also say that the reference point is an imagined aid to facilitate calculating a quantity.

Figure 5 is an example of what is meant by reference point since D has been calculated from the distance between the reference points RP₁ and RP₂.

Since it is assumed that Figure 4 is applicable to all transducers, the function curves for two diametrically placed transducers can be combined as in Figure 5, so that e.g. RP₁ and RP₂ are at a distance A from each other along the LO-axis. If signals V₁ and V₂, respectively, are obtained from the transducers, the corresponding LO values i.e. LO₁ and LO₂, can easily be determined.

The diameter, D, of the test object can thereafter be calculated as follows: D = A - (LO₁ + LO₂).

This calculation, like the other function conversions, linearizations and so on, can of course be performed using either hardware or software, e.g. via programmable electronics such as computers or the like.

The function curve according to Figure 3 is already known. However, as far as we know, the principle according to Figure 5 is not known. Note also that if linearization has not occurred and the functions are thus non-linear as in Figure 3, the D-value will vary if the position of the test object between RP₁ and RP₂ is varied, which is of course entirely unacceptable.

The linear relationship between LO and signal and/or measured value according to Figure 4 thus greatly simplifies calculation of the dimensions of the test object.

Placing several surface transducers around the test object, e.g. as described in Swedish patent application 8700472-7, enables the diameter D to be measured at exactly the same time in a plurality of directions and dimensions, as well as allowing a picture of the cross-sectional profile, contour, etc., of the wire. It is also possible, if suitable, to combine the transducer coils, thus obtaining several simulated reference points and thereby several opportunities for measurement.

As mentioned earlier, in Figure 3 the transducer has been compensated at LO = ∞. However, within the scope of the invention the compensation may be performed at some other LO if suitable.

"Compensation" refers for instance to the type of compensation, including balancing, described in Swedish patent applications 7507857-6, 7613708-2, 7813344-4, 8302738-3 and 8400698-0 and the terminology used is substantially covered by the present invention.

The electronics, part (3) in Figure 1 may be conventionally constructed with respect to the eddy-current part. A good example of how the eddy-current part can be designed is revealed in Swedish patent application 8400698-0 in which Figure 3 exactly describes one of many feasible examples of what item 3 in Figure 1 may include. The output signal from the eddy-current part can then be further signal-processed, for instance, via associated analog-digital convertors and computer including the calculation part (4).

This example is also satisfactory since at least two carrier frequencies are usually required to enable effective suppression of undesired effects via vector transformation. The frequencies should be selected so high that the current penetration depth will be limited. This is also important so that the measurement will not be disturbed too much due to the speed at which the wire moves through the transducer arrangement. At low frequencies the response from the test object to the transducer will be delayed due to the inductive connection of the currents and/or magnetic field deep in the test object. This will give rise to a disturbing lag effect which is particularly noticeable at high speeds.

To ensure high measurement precision the distance (A) between the reference points, RP, must be as exact and stable as possible. To achieve this, the transducers must often be compensated immediately before the leading end of the wire passes in through the transducer arrangement. All previous functions such as temperature-dependent winding resistors and the like are then eliminated. It is thus advisable to provide a photocell to sense the wire immediately before the transducers and to give this photocell as compensation pulse during which the transducers are rapidly compensated. Thereafter, e.g. after 10 ms, the wire may entire the new transducer arrangement. The performance of the device is thus always at its highest before measurement.

Arranging the transducers in Figure 1 in two halves, for instance, which can move in relation to each other, the throughput channel of the transducer can be varied automatically, for instance. This may be valuable as it allows the transducer opening to be made temporarily large enough to prevent ribs, etc., at the leading end of the wire from becoming wedged in the transducer. Transit transducers may be used instead of surface transducers, or a combination of surface and transit transducers.

The surface transducers in Figure 1 may also be provided with automatic distance setters, e.g. ball-bearing screws and motors, which can set the transducers at predetermined LO-distances to suit the dimension of the wire currently being tested, for instance.

A special calculating function may then be provided for each such distance setter. The computer memory/program is then suitably provided with a library of suitable functions, including correction programs for relevant wire dimensions and so on.

It is often advantageous to be able to measure the distance (LO) over a relatively large working area, e.g.1-10 mm. It is therefore suitable to provide the surface transducers with ferrite cores, for instance, which amplify the magnetic flow between transducer and test object.

An extremely troublesome complication in the type of transducer arrangement shown in Figure 1 is that in practice the wire may vibrate vigorously, as is indicated in the Figure by another position marked in broken lines, i.e. the centre P1 has been displaced to P2. To come to terms with this problem it may be advantageous for the transducers to be sensed instantaneously, i.e. at the same instant, and for signals and/or measured values to be stored in a computer memory, for instance, and for D, etc., to be calculated later on, for instance. Another problem of great significance to the measurement is that the curve according to Figure 3 is only valid when the test object, i.e. the wire, is immediately opposite the transducer. This is because the inductive connection between wire and surface transducer coil varies when the wire is moved a distance S, as in Figure 6, but when the distance X is constant. The position shown in broken lines corresponds to the curve/function shown in broken lines in Figure 3. (These curves are not drawn to scale and should therefore be considered as illustrating the principle.) The result of this is that the measurement precision is noticeably reduced if the centre of the wire is moved from P1 to P2, for instance, as shown in Figure 1.

This problem can also be overcome by correcting the measurement with the aid of the deviation of the wire from a normal position such as P1 in Figure 1. The dimension of the wire is thus also calculated as a function of the position of the test object, e.g. its centering in relation to the transducers, and thus taking this into account. An excellent method of determining the position of the wire is to utilize the transducers as described in Swedish patent 8101284-1. The transducers can then be used to measure both the dimension and the position of the wire on the basis of two different devices/principles. Furthermore, there is nothing to prevent the same transducer also being used for crack detection, for instance, in which case the application will be more complete and sophisticated.

The scope of the present invention of course includes all the examples described herein, as well as combinations thereof.

Another complication arises from the technical measuring point of view if the test object, in this case the wire, has surface cracks or other defects. A surface crack (5) has been drawn in by way of example in Figure 2. This surface crack disturbs the spread of the eddy currents induced on the surface of the test object, which in turn is understood by the surface transducer coil as a change in impedance. This change in impedance can be described as a vector (refer to Patent application 8302738-3 for instance), the direction of which usually substantially coincides with the distance direction, LO, in the impedance plane of the transducer. This may cause the transducer to believe that the surface of the test object is further away than it really is and the dimension calculation is disturbed by the crack, with the result that it is generally incorrect. To deal with this problem also, vector transformation technique based on the use of several carrier frequencies can be used. Swedish patent application 8700359-6 describes how a transformation device excellent for this purpose can be designed and used.

This vector transformation thus enables those variables/parameters which disturb the dimension measurement, for instance, to be separated out, thus avoiding presentation of incorrect dimension values.

Another related problem is temperature influence of various kinds on the measurement. When the temperature varies after compensation, the inner resistance of the transducer coils, as well as the resistance over-transformed from the metallic transducer casing. These resistance fluctuations cause various types of disturbances.

One method of avoiding these is to control the temperature of the coolant to the transducers, another is to correct the measured values on the basis of the temperature fluctuations measured.

Undesired measuring effects caused by varying temperatures in the test object, for instance, can also be suppressed in similar manner. A varying alloy composition in the test object can, for instance, produce similar effects to those obtained from a variation in temperature. This is because both influence the electrical conductivity of the test object. Since the alloy composition is generally known and the temperature of the wire can be measured, the signals required to correct the calculation functions for the quantity are available.

The invention is intended for use in measuring the dimension of hot rolled wire and the like where the test object has a temperature above its Curie point. The material is then non-magnetic. However, by choosing suitable carrier frequencies, the invention is also suitable for measuring/testing cold materials, both magnetic and non-magnetic. Similarly the invention can be used to measure quantities such as the level of a steel melt, etc. If the LO distance is calculated directly from the non-linear function in Figure 3, via a mathematicaly model/algorithm, for instance, this can be considered a linearization since a type of function conversion is used based on at least one real or fictive reference point. This therefore also falls within the scope of the invention.

To ensure correct calibration of a device according to the invention, it is advisable to provide it with automatic calibration aids. By manufacturing various reference objects or standards with varying quantities, the function conversion routines and so on used for calculating quantities can be updated within the scope of the invention. In other words, the reference objects replace the test object during the automatic calibration procedure. Computer programs are used to perform more advanced calibrations.

In practice the wire diameter, for instance, is changed when the final product is changed. Fine adjustment of the functions controlling, for instance by computer program, the correction/compensation of e.g. the influence of deviations from the centre, must therefore be performed. It may therefore be advisable to have a small series of "standards" of various diameters, these being placed one by one in the measuring position in the transductor for the computer to store the relevant measured values. The computer then suitably calculates the functions describing the influence of the different variables on the measured result, i.e. quantity, and can then also generate suitable and optimum correction functions. Calibration standards can of course also be automatically inserted into the transducer by mechanical means on suitable occasions. A calibration standard might consist, for instance, of a shaft which is turned in steps so that it has different diameters in axial direction. The shaft material is suitably non-magnetic steel with the same electrical properties as the test object.

The transducer arrangement according to Figure 1 can advantageously be secured in a guide tube immediately after the roller stand in a wire rolling mill. The transducers thus move relative to the rolled wire due to the movement of the wire. However, the invention also includes applications in which the transducers rotate or oscillate around the test object. The invention also includes all imaginable forms of test object. Particularly in rolling mills it may be of great importance to measure the dimension of the rolled wire after each reduction step so that the rolling process and its parameters can be controlled and corrected on the basis of information as to the diameter of the wire.

Another interesting application for the invention is measuring flatness, e.g. of steel strip and the like.

As is clear from the description so far, numerous parameters influence the final measured result of a quantity. Figure 7 is provided to give a clearer view of this and should be considered as a comprehensive figure shown in principle the variables normally included in the calculation of a quantity, such as dimension.

The test object (1) is surrounded by the surface transducer coils of the transducer arrangement (2) connected to the electronics system (3) which is in turn connection to the calculating part (4) in which the quantity (ST) is calculated. In principle the measuring and calculating parts may be included in the same electronic unit.

The following signals are received by the calculating part from the measuring electronics (3):

M1, N Lift-off function values, e.g. in accordance with Figure 3 M2, Centre deviation including the direction α e.g. R

M3, Separation signal e.g. crack or diameter M4, Transducer position e.g. LO position, in other words A Other signals transmitted to the calculating part (4) are, for instance:

P1, Material constants e.g. nominal diameter, alloy, resistivity, etc. P2, Form constants P3, Transducer temperature, e.g. temperature of coolant P4, Temperature of test object

K1, Process signal, photocell - leading end of wire K2, Process signal, rolling speed K3, Process signal, tolerance limits

The quantity is calculated with the aid of these input signals.

The calculating unit (4) may suitably include a microcomputer with a hardware-based calculating modules for quick calculation. Combining the invention with the devices described in Swedish patent application 8700472-7 offers the advantage that the electronics part (3) and calculating part (4) can be common for a number of transducers and/or transducer arrangements.

Some explanatory examples follow:

The test objects may be billets, rods, wire, pipes, sheet-metal, molten metal, etc.

The transducer may be eddy-current based surface transducers, Hall-elements, various types of sensors, etc., or combinations thereof.

The quantity may be dimension, form, position, profile, temperature, speed, etc., or combinations thereof. "Quantity" is thus a broad concept.

Carrier frequency may be the frequency of the relevant eddy current or frequency components included therein. LO may be what is normally termed lift-off. Examples of this are given in Swedish patent applications 8400689-0 and 8400861-4, for instance.

In other respects, reference is made to the terminology as defined in the patents and applications cited in this description.

The invention can be varied and is applicable in many ways within the scope of the following claims.

When the transducer moves, for instance a relative movement is intended.

Claims

1. A device for measuring and/or testing, for example detection of test objects (1), such as rolled wire, with respect to dimension and/or form, comprising at least one transducer (2) which moves on/over, for example along, the surface of the test object or a part thereof, and which via associated electronics (3) generates and/or calculates signals and/or measured values which directly or indirectly represent the distance (LO) between the transducer and the surface or part thereof of the test object, c h a r a c t e r i s e d in that these signals and/or measured values are fully or partially and directly or indirectly measured and/or calculated with the aid of currents such as eddy currents induced in/on the test object, and that dimension and/or form is/are fully or partially automatically measured and/or calculated as a function of at least two of these e.g. instantaneously measured and/or calculated signals and/or measured values.

2. A device as claimed in the preceding claim, c h a r a c t e r i s e d in that signals and/or measured values representing the distance (LO) between transducer and test object are measured and/or calculated on the basis of at least one reference distance and/or reference position, e.g. fixed reference positions/points which can be considered as pertaining to the transducer.

3. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that signals and/or measured values obtained directly or indirectly via transducers are processed, e.g. linearized, function-converted, etc., so that the information concerning distance, e.g. the measured value V is substantially proportional to the distance LO within the LO-working range of the transducer, e.g. so that V = LO × k, where k is a constant.

4. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that vector transformation, e.g. based on the use of at least two carrier frequencies, is utilized to separate and/or suppress the effect of disturbing variables/quantities such as surface cracks.

5. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that meansurement and/or calculation of a quantity also occurs as a function of the position of the test object, e.g. centering relative to the transducers/arrangement of the transducers on each measuring occasion.

6. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that magnitudes are measured and/or calculated with the aid of signals, fully or partially, deriving from at least two transducers located substantially diametrically opposite each other at the test object, e.g. round rolled wire.

7. A device as claimed in one or more of the preceding claims, cha r a c te r i s e d in that temperature compensation and/or maintaining a constant temperature occurs taking into consideration at least one of the following points/variables, - the temperature variation/dependence of the transducer winding - the temperature variation/dependence of the transducer casing- the temperature variation/dependence of the test object.

8. A device as claimed in one or more of the preceding claims, c ha r a c t e r i s e d in that the distance (LO) between the transducers and the surface of the test object can be altered automatically by photocell control, for instance, e.g. so that the thicker end of the rolled wired can pass through the transducer arrangement, after which the transducers immediately return to their normal position for dimension measurement.

9. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that calculation of quantities is effected with the aid of programmable electronics, e.g. a computer.

10. A device as claimed in one or more of the preceding claims, c h a r a c t e r i s e d in that the device is included in and/or used in combination with another measuring and/or control device.