WO1988000101A1

WO1988000101A1 - Method for measuring and recording roll gap for continuous casters

Info

Publication number: WO1988000101A1
Application number: PCT/SE1987/000278
Authority: WO
Inventors: Örjan AXELSSON; Lars Ove Larsson
Original assignee: Axelsson Oerjan; Lars Ove Larsson
Priority date: 1986-06-27
Filing date: 1987-06-12
Publication date: 1988-01-14
Also published as: SE464465B; SE8602859D0; AU7644787A; SE8602859L; EP0274500A1

Abstract

Method for measuring and registration of the width and geometry of a roll gap between rolls of two parallel roll paths in a continuous caster, forming e.g. a vertical zone, a curved zone and a horizontal zone, the procedure being that one or more measuring yokes (11) are brought in contact with two supporting rolls (14) in the first roll path causing them to be located along the tangent to the supporting rolls (14) and causing a measuring device (15.1) to measure the distance between a reference plane, e.g. a tangent plane to the supporting rolls (14) and the other rolls (17) within the length of the measuring yoke, the corresponding distance being measured to the rolls in the second roll path and the measurement values registered and compared with the corresponding values setting out from the geometry of the roll paths in an ideal roll gap.

Description

METHOD FOR MΞAS HIXG AX RECORDING ROLL GA? ?0R CO"T^π^τOϋS

FIELD OF TECHNOLOGY

The present invention relates to a method and apparatus for measuring and registration of the width and geometric relationship between the roll faces of two parallel roll paths in a continuous caster characterized by e.g. a vertical zone, a curved zone and a horizontal zone.

BACKGROUND TECHNOLOGY

A device for measuring of the width of the roll gap in continuous caste is known from the Swedish patent description no 7709637-8. The device comprises a fixing frame adjustable in the width direction of the roll gap which frame can be moved to the intended point by means of a chain and a conveying device. This device is equipped with centering rolls cradled in pairs on both sides of the device, of which rolls at least one on each side is reversibly turnable. Centrally between the rolls and at right angl to their axes the device is furthermore equipped with measurement sensing means which are via a cable connected with a measurement value indicator. The conveying device consists of two motor-driven conveyor belts attached to the one alignment roll side of the fixing frame, which belts extend alo at least three rolls on the roll conveyor to be checked on.

A similar measuring device is known from the Swedish patent description no 7709638-6 which has, in addition to the measurement sensing means refer to in the aforementioned patent description, measurement sensing means att to both sides of the first-mentioned measurement sensing means in the late direction of the roll path. The purpose of these additional measurement sensing means is to achieve a measurement of the curved path represented by the conveyor rolls.

An inconvenience entailed by the existing devices is that they do not permit continuous measurement while the device is being moved along the roll path. Another drawback is that these devices do not permit measuremen of the width of the roll gap except when using non-excentric, i.e. pair-wise and opposite each other located rolls positioned at the roll gap along a line perpendicular to the longitudinal direction of the roll path. DESCRIPTION OF THE INVENTION

The purpose of the present invention is to achieve a method and an apparat for measuring and registration of the width and geometric relationship betw the roll faces of the roll gap between rolls in two parallel roll paths in a continuous caster, characterized by e.g. a vertical zone, a curved zone and a horizontal zone, which method and apparatus according to the invention do not entail the inconveniences pertaining to existing devices.

The purpose of the invention has been adiieved by means of a method according to the present invention characterizedby one ormore measuring yoke being brought in contact with two supporting rolls^" on the first roll path and being positioned along the tangent of the supporting rolls and by a measurement device being applied to measure the distance between a referenc plane, e.g. a tangent plan to the supporting rolls, and the other rolls within the lenght of the measurement yoke, and by the measurement values being registered and compared with the corresponding values, setting out from the geometric relationship between the roll paths of an ideal roll gap. The method is furthermore characterized by the measurement yoke being caused to move continually along the first roll path during measurement in progress. Additionally, the method is characterized by the measurement values being transformed into a coordinate system common to all the rolls.

The apparatus according to the invention is characterized by one or more measuring yokes, intended to be brought in contact with two supporting roll on the first roll path which are positioned e.g. along a tangent plane of the support rolls, and by one or more initial measurement devices by aid of which the distance between a reference plane, e.g. the tangent plane and the other rolls of the first roll path is measured, within the longitudinal extention of the measuring yoke,and by one or more other measurement organs by aid of which the distance between the reference plane and the other rolls of the other roll path are measured.

The apparatus is primarily characterized by the measuring yokes having two ends resting against one outer support roll each, and by the first measurement organ resting against an intermediate roll . The other measuremen device can have two or more measurement points, e.g. measurement rolls or similar, positioned in the longitudinal direction of the measuring yoke. The measurement^'devics are in connection with one transducer each, which each one emits signals to a multiplexer controlled by a a computer in a way to sequentially switching-in one transducer at a time to an analogue digital converter also controlled by the computer, and in which analoαue voltage signals are transformed into digital information. The transducer contains e.g. an optical sensor, the signal of which is tranformed in a converter into an electric voltage or current signal the magnitude whereof is a function to the measurement distance.

By means of the invention a method and apparatus has been achieved for measuring and registration of width and geometry of a roll gap, superior to methodsand apparatuses as yet known. In particular, the invention permi continuous and rapid measurement of the roll paths in a roll gap, where the rolls may have an entirely independent position in the first path in relation to the other of the two paths along the roll gap.

DESCRIPTION OF FIGURES

The invention is described below as a construction example related to attached illustrations.

Figure 1 schematically shows a roll path of a continuous caster at a vertical longitudinal section.

Figure 2 displays apart of the roll path according to Fig.l with measurement values inserted.

Figure 3 illustrates the principle of converting measurement values int a joint coordinate system.

Figure 4 shows a coordinate system applied to the loose side of the roll path.

Fi^'gure 5 shows a coordinate system applied to the fixed side as well as to the loose side of the roll path.

Figure 6 shows the straight part of the roll path adapted to a straight line.

Figure 7 shows the curved zone adapted to a circle.

Figure 8 is a graphic presentation of the measurement results.

Figures 9 and 10 are flow-charts of the measuring system according to the invention.

Figure 11 shows a measuring yoke of a measurement cart which is moved along the roll path.

Figure 12 shows a calibration device for the measuring yoke according to Fiα.ll. The roll path is defined as the envelope of the roll casing surfaces on the inside of the roll gap. The upper part of the roll path is referred to as its loose side whereas the lower part is called the fixed side. With regard to the parts of the roll gap the envelope constitutes mathematically a straight line, a circular arc with a fixed radius, or a continuous transition where the curve changes from a circular arc to a straight line. Figure 1 displays the roll path with envelope. When the strand passes throu the roll gap (from point A to point B in the Figure) the strand will adopt the shape of the inside of the roll gap. When there are rolls in the roll path which deviate from the ideal position the casting crust is forced to deform which may entail cracking in the crust. The setting of the rolls is consequently of decisive importance to the quality of the casting strand The roll gap is designated C in the Figures.

In order to measure and check the width of the roll gap and the mutual positions of the rolls, i.e. the geometry of the roll paths, a measuring yoke is placed across three adjacent rolls. The yokes are supported on the outermost rolls on the so-called fixed side of the continuous caster, i.e. the lower roll path, and are moved along the roll gap and the measurem distance from a straight line between the supporting points of the measurin yoke to the intermediate roll. The distance is also measured to the rolls at the so-calTed loose side, i.e. the upper roll path. Figure 2 illustrates the measuring principle applying to a randomly chosen group of rolls. The measurement values are called rough coordinates. They constitute coordinate in local coordinate systems represented by the tangent to the outer rolls (the abscissa) and the normal to the latter through the centre (ordinate) of the one of the outer rolls. The rough coordinates are designated Hf (j) with regard to the fixed side and HI (j) with regard to the loose side. Both Hf and HI are measured as being the distance from the straight line through the contact points of the measuring yoke to the rolls designated (j) and (j+2) to the roll (j+1). Should the total number of rolls be N the will be N-2 rough coordinates of the above described type in the strand.

Each measurement value, i.e. rough coordinate, is specified in a local coordinate system ( , ) belonging to the measuring yoke. The direction of the coordinate system is determined by the direction assumed by the yok when resting on two of the rolls. The direction of the coordinate system may consequently vary from one measurement value to another, for instance in the curve zone, which is shown in Figure 2 where the two local coordina systems ( , ) and ( , ) are reproduced. It is consequently not possibl to make any direct comparison between the different measurement values. In order to compare the positions of several individual rolls it is necessary to retransform the rough coordinates into a coordinate system common to the individual rolls. The X axis of the joint coordinate syste may coincide with the tangent into two random rolls. See Figure 3. For practical reasons the direction of the measuring yoke, as assumed by it w it rests on the first and the third roll, has been chosen as the direct of the coordinate system. If the relationship of the direction to the ear coordinates (horizontal and vertical direction) is known, the geometry of the rolls can be expressed in terms of these coordinates by a simple coordinate transformation. This is not necessary, however, if the task is to only determine the mutual positions of the rolls and to calculate the dimension of the gap. As an example a group of M subsequent rolls may be chosen. The first roll begins in a random roll which we have chosen to si designate as 0, the second roll 1, etc, all the way to the last roll whic is designated M-l. The X axis of the joint coordinate system extends thro the contact points of the measuring yoke to rolls 0 and 2. The Y axis ext through the centre to roll 0.

Mathematically it is possible to derive an interconnection between the H coordinates and the XY coordinates with regard to the contact points of the rolls of the fixed side of the roll path. This interconnection is difficult to express by analytical formulas but it can be recursively calculated in a computer. By this calculation the XY coordinates are exac specified provided that the variables Hf, HI and Lf, LI according to, Fig 2 are exact.

The rolls of the loose side and the fixed side, respectively, are meas with the measuring yoke supported on the same rolls on the fixed side. Ordinarily, a measurement value is obtained for each supporting position assumed by the measuring yoke. There may, however, be border cases where no measurement values or two measurement values are obtained in the same measuring position. This is because the rolls of the fixed respectively loose side are not always located straight across each other along a line at right angles to the roll gap. Figure 4 shows the measurement values in relation to the yoke and the local coordinate system ( , ).

Presuming a line parallel with the -axis of the local coordinate sy the coordinates of the loose side constitute the tangential point of the roll to the line. The measured roll distance HI constitutes the -valu and the distance (LI) of the roll from the first roll of the fixed side its -value.

The distance can e.g. be determined on the basis of a drawing or be measur When the XY coordinates of a group of rolls of the fixed side have been calculated the direction of each one of the local coordinate systems is known in relation to the coordinate system passing through the first and the,-third rolls on the fixed side of the group. This direction is designat and^*is determined in the same way as earlier described. It is consequentl possible to retransform the coordinates of the loose side into the same XY system as that of the fixed side and thereby determine the relationship of the rolls of the fixed respectively loose side to each other. This is illustrated in Figure 5.

Assuming that the roll distances in the longitudinal direction of the roTT path is on each and every measurement occasion in accord with the drawings,- a certain simplification may be allowed in measuring the rough coordinates. It is then not necessary to measure the roll distances, whic may be difficult unless the traction speed of the measuring yoke is const Such arr. assumption naturally introduces a certain measurement error to th roll distances specified which will influence the accuracy of the calcula XY coordinates. However, it is difficult to give an analytically exact me to the error. This depends on a number of parameters, i.a. on the number of rolls used when analysing the deviations. In the curved zone the error appears in both the X and the Y directions. In the horizontal straight se it appears mainly in the Y direction.

RoTT gap width as well as deviations between the rolls of the fixed respectively loose side can be determined by means of a series of XY coordinates determined as earlier described. The values measured are next to be compared with the original model in order to enable gap width and roll deviations to be established.

Mathematical adaption of curves is the most frequently employed method to adapt measured coordinates to a model. This requires, however, that th model can be expressed as a mathematical function. The method is called linear regression. A mathematical function is in this case adapted across M rolls. The first roll may be chosen randomly in the strand since the pu is to only try to find deviations from the assumed curve.

In the straight parts of the roll gap the adaption is preferably to a straight line. See Figure 6.

^Yl ^{= q}l ^Xl ^{+ q}o (loose side) Y- = p. ,X_f + p_{Q (fiχed side)} e a ap on s e ec e e.g. y us ng e m n mum-square me o , .e. in such a way that the square total of the deviation between the right and the left parts of the equations is minimal. When calculating gap widths and roll deviation in the straight part, one should split the straight pa into a suitable number of groups, e.g. 5-10 rolls to each group. There ma be a deviation, i.e. an error, also on these points. The adapted straight line will still pass through these points. In order to determine the erro at the contact points, the coordinate system must be turned in a way to make it coincide with the adapted line. Curve adaption should consequentl take place stepwise as follows:

1. Determining of constants pO and pi for the regression line to the coordinate points of the original coordinate system XY (which extends thr the first and third roll of a random group of rolls).

2. Transforming of the original coordinates with regard to both the fixe and the loose side in order to make them coincide with the regression lin obtained in Step 1.

3. Calculation of gap width and roll deviations.

The mathematical solution of the curve adaption to a straight line constitutes a standard solution in statistics. The solution is for this reason not described here. The coordinate system XY constitutes the local coordinate system of the group of rolls analysed. X'Y¹ constitutes a coordinate system, the X axis of which coincides with the regression line p X+p . Curve adaption in respect to the loose side can then be effected by X'Y' coordinates, in which case the distance from the adapted line to the axis constitutes the mean gap width.

The curve adaption in the transmission zone can be made to a polynomial of a suitable number of degrees. The derivator of the polynomial at the transitions between the straight respectively circular section is to be in accord with the derivators of the straight line respectively the circul arc. Also the curvature is to coincide at these points. When the number of degrees of the polynomial has been determined the curve adaption is eff in the same manner as in the case of the straight line. The difference is that the number of constants to be determined by the minimum conditions are several. Also the polynomial adaption is a mathematical routine and need not be described. One should keep in mind, however, that possible def in the supporting roll will displace the coordinate system causing the erro to be zero there. For this reason the obtained curve must be retransformed in the same way as described with regard to the three steps above, before the deviation can be determined. In the curved zone the XY coordinates are adapted to a circle. From a given random roll on the circular arc M subsequent rolls are chosen to be included in the curve adaption. The first roll in the sequence is designated 1. The last roll is designated M. The radius of the circle is R. Its origin of coordinates in relation to the local coordinate system through the first and the third roll is (a,b). The equation of the circle can tie expressed in local coordinates as follows:

The task will now be to determine a and b in such a way that the error iϊi the adapted circle will be minimum in the sense of the minimum square method. In the ordinary case, this results in a very complicated non-linea equation system, but by assuming that a and b are cathetuses to a triangle with the hypotenuse R, the problem can be simplified. This actually results in a. linear equation system from which a and b can be solved iteratively. The coordinate system will then contact the circle closer to the origin of coordinates of the new coordinate system. When the coordinate trans¬ formation has been completed, a and b are calculated again (iteratively) as earlier mentioned. The iteration is repeated until a=o which entails b=r. Ordinarily, this ensues after 2-3 iterations. Finally, the deviation e, from the obtained circle is calculated. See Figure 7.

e_fc - - X_k ^{2 +} < - R)²

When adapting the coordinate points to a circle, it is favourable from the mathematical viewpoint to have many points involved. This yields a safer value of the circle obtained. For practical reasons, however, one should not have too many, since the roll distance error will in such case be notica le. Approx. 5-10 rolls should be a suitable number.

The roll positions are shown graphically in Figure 8. The graphical pict constitutes the mean lines of the fixed respectively loose sides of the strand. The mean lines have, for the sake of simplicity, been straightened out into a straight line and the mutual positions of the rolls in relation to the mean line (the adapted curve) havebeen indicated. Should the measurement take place at several axial planes each plane can be similarl drawn.

As a complement to the graphical picture a list can be obtained of the displacements of the rolls in relation to each other or in relation to the expected gap width. The displacement can be listed for the total of rolls or for only certain of them, e.g. rolls which extend toofar into the casting strand or section of too large or too little gap width.

A complete measuring system consists of one or more transducers 1 and a measuring instrument 2, and devices 3 for result presentation may be in the form of a character display, a terminal, a printer or a central proces unit. See Figure 9.

The process is the following: The measurement transducer 1 yields durin the measurement sequence continuosly information on the measurement distan to the measuring instrument 2. In the latter the information is processed and the measurement data reduced to a suitable number. By analyses of measurement data also the number of rolls can be determined and a relative distance value be obtained for each roll. Depending on the capacity of the software in the measurement instrument additional computations can be made Thus a curve adapation can be effected and the deviation of each roll from the ideal position be shown in tables. Also limit values can be included. The measurement signals from the different transducers 1 are connected to multiplexer 4. This is controlled by a computer 5 in a way to sequentially switch in one transducer 1 at a time to an analogue/digital converter 6 which is also controlled by the computer 5. The converter 6 converts the analogue voltage signals into digital information.

In the course of the measuring procedure approx. 300 measurement values per second are produced. Most of these values lack interest and are to be sorted out. Only the values representing the least distance between the individual roll and the measurement device are to be saved. For this reason all measurement data are continually processed by the computer, whereupon they are entered for computation of minimum distance. When this has been determined the measurement value for each roll is stored, partly in the electronic memory (RAM) in the computer and, partly, on a diskette of a diskette unit 7. This diskette unit has two functions:

1. Reading from the diskette of instructions to the instruments in order to determine its function.

2. Storing of measurement data as an extra safety measure should the electronic memory be erased e.g. due to power failure or faulty operation. For the operation of the instrument (inform when a measurement operation begins and ends, reading in of programs, etc) there is a display 8 (a character screen) as well as a keyboard 9. The display shows i.a. instructi to the operator and receipts of commandos given. The instrument also contai a date and time generator 10 (clock) marking mathematically each measuremen vn^'th the correct date and time. The measurement instrument is battery-opera and is charged between the measurement periods by means of an external char device connected between the mains and the instrument. The instrument is also provided with two outlets for connection of peripherals, e.g. a centra computer which is to collect and supply information to the instrument. In this case the instrument will be connected to the existing data network and some simple commands will be entered on the keyboard of the instrument whereupon the transfer will be controlled by the program in the CPU. An alternative is to connect a printer to the instrument. This is preferabl done via a parallel outlet. Printout can be made of the measurement values for each individual roll. Only deviations larger than a given limit value or than what the instrument has been programmed to compute, will be printe out. The print-out is ordered via the keyboard.

The measuring device according to Figure 11 consists of a measurement cart comprising one or more measuring yokes 11 in the form of an extended chassis 12, at each end of which a supporting plate 13 is arranged in such a way that the individual supporting plates 13 are at ,one and the same plan and which are intended to each one rest against one conveyor roll 14. The length of the supporting plates 11 in the longitudinal direction of the ^• roll path is such that the plates permit a varying distance between the conveyor rolls 14 and 17 in the first roll path. A first measuring bar 15 is pivotally fitted on the one side of the chassis 12 around an axis parall to the axes of the supporting rolls 14. The first measuring rod 15.^τ su^pport at its free end a measuring roll 16 which is located between the two supporting plates 13 and which is in contact with a conveyor roll 17.1 positioned between the two supporting rolls 14 in the first roll path, whic are in connection with the supporting plates 13. A second measuring bar 15.2 is pivotally fitted on the other side on the chassis 12 around an axis parallel with the axes of the rolls 14. The second measuring bar 15.2 suppo at its free end a measuring roll 16 which is in contact with a conveyor roll 17.2 positioned in the second roll path. As a result of the supporting plate 13 having been given a certain minimum length in the longitudinal direction of the roll path, the measuring yoke 11 can be displaced in the longitudinal direction of the roll path with unchanged reference plane permitting the measuring roll 16 on the other measuring bar 15.2 to be caus to pass the conveyor roll 17.2 in contact with it, particularly if this conveyor roll 17.2 is positioned excentrically, i.e. at the side of the center point normal from the tangent between supporting rolls 14, i.e. to the connection line between the contact points of the measuring yoke 11 on the supporting rolls 14.

Ih an alternative version of the invention the second measuring bar 15.2 is provided with three measuring rolls 16 arranged in the longitudinal direction of the measuring yoke 11, of which rolls two are shown in dotte outline in Figure 11, and by aid of which one can measure the distance fr the reference plane to a roll 17.2 in the second roll path, pointedly excentrically located in relation to a corresponding roll 17.1 in the fir roll path.

With the first-mentioned version one achieves that a certain excentric placing of the rolls in the second roll path can be tolerated as a result of the displacement of the measuring yoke 11 on the supporting rolls 14. The alternative version in combination with the first-mentioned one permi an additionally excentricity to be tolerated, i.e. a random location of the rolls in the second roll path in relation to the rolls in the first roll path.

A transducer 18 is fitted on the chassis 12 at the respective measurin bar 15 on which a corresponding measuring plate 19 is fitted. The transdu 18 consists of a measuring body (sensor) and of adaption electronics (converter). The output signal from the transducer consists of a voltage of current the strength of which is a function of the measurement distanc The output signals from several individual transducers can be connected to the multiplexer 4 (channel selector) to be sequentially connected to an analogue/digital converter 6 which produces a numerical value for furt processing.

Measuring of the distance between two conveyor rolls 17 is effected by th measuring cart being connected to a so-called starting chain to a continu caster not shown in the figures and displaced by means of this chain alon the roll path of the machine. Measurement values are recorded when the measuring rolls 16 pass each individual roll in the roll path and are converted as described above. The measuring system is regularly calibrated by means of a calibration device according to Figure 12. The calibration device consists of a base plate 20 which is fixed with screws to the supporting plates 13 of the measuring yoke 11. A calibrated measurement transducer 21 (electronic micrometer) is attached straight in front of the one measuring roll 16 permitting this to be moved within its measuring area by the micrometer being turned. The output signal from the micrometer 21 is connected to the instrument via a 7th channel. According as the micrometer 21 is turned, existing as well as standard values are generated which are read into the instrument. The reference distance of the measurement transducer 21 is designated D in Figure 12. A measuring roll 16 having in this way been caused to move from min. to max. position one or more times, a command is grrven which automatically sets the longitudinal coefficient in the instrume tσ this very transducer. Micrometer 21 is thereupon moved to the other measuring roll 16 and the procedure is repeated. The calibration device is next moved to the second respectively third measuring yoke 11. The measuring system is now fully calibrated.

Claims

PATENT CLAIMS

1. Method for measuring and registration of the width and geometry of a roll gap between rolls of two parallel roll paths in a continuous caster, forming e.g. a vertical zone, a curved zone and a horizontal zone, the procedure being that one or more measuring yokes (11) are brought in cont with two supporting rolls (14) in the first roll path causing them to be located along the tangent to the supporting rolls (14) and causing a meas device (15.1) to measure the distance between a reference plane, e.g. a tangent plane to the supporting rolls (14) and the other rolls (17) withi the length of the measuring yoke, the corresponding distance being measur to the rolls in the second roll path and the measurement values registere characterized by the measurement values being transformed into a coordina system (x,y) common to all the rolls, and by a random function being adapt to the position of the rolls in the joint coordinate system, and by the result being compared with the corresponding values setting out from the geometry of the roll paths in an ideal roll gap.

2. Method according to Claim no 1, characterized by the X axis of the x,y coordinate system being caused to coincide with tangent to two random roll (14).

3. Method according to Claim no 2, characterized by the positions of the rolls in the two roll pathsbeing transformed into the x,y coordinate syste

4. Method according to Claim no 3, characterized by the coordinates of the rolls in an earth coordinate system being computed by means of a coordinat transformation.

5. Method according to Claim no 4, characterized by the x,y coordinate being adapted to a straight line in the straight parts .of the roll path.

6. Method according to Claim no 4, characterized by the x,y coordinate being adapted to a circle in the curved zone of the roll paths.