EP0290637B1 - Process for operating a calander machine and controlling device for carrying out the process - Google Patents

Abstract

The process for operating a calender machine, having a bending- compensating roll, fixes an operating pressure for each effective point to which pressure can be applied, so that a load parameter in the pressing nip has a predetermined set value. First of all, a pressure reaction matrix is formed, the elements of which indicate the change in the load parameter in all the zones in the case of a change in pressure at only one effective point in each case. Then, in the case of a change in set value, a change in pressure compensating fully or partially for the difference between actual value and set value of the load parameter is calculated using the pressure reaction matrix successively step by step in each case for the effective point of a zone. For all the other zones, the changed actual value of the load parameter resulting from this change in pressure is calculated. If the value drops below a tolerance value, the operating pressure for each zone can be corrected by the sum of all pressure changes. These iteration calculations are carried out in a control device by a programmed arithmetic device (16). <IMAGE>

Description

The invention relates to a method for operating a roller machine having at least two rollers for the treatment of web material in a press nip, in particular a calender or smoothing unit for paper, plastic or textile webs, in which the press nip has a number of zones, each one Actuating point which can be subjected to adjustable pressure - including individual bearing elements or groups of bearing elements which can be acted upon with the same pressure and which support the roller shell of a bending compensation roller on a non-rotatable carrier which penetrates the shell - in which method a working pressure is determined for each active point with the aid of a computing operation, which depends on the setpoint profile of a load parameter in the press nip, and on a control arrangement for a roll machine having at least two rolls for the treatment of web material in a press nip, in particular calender or smoothing unit for paper, plastics Off- or textile webs, in which the press nip has a number of zones, each of which has an effective point which can be subjected to adjustable pressure - including individual bearing elements or groups of bearing elements which are subjected to the same pressure and which support the roller jacket of a bending compensation roller on a non-rotatable support which penetrates the jacket - are assigned which control arrangement generates control signals for pressure control valves in the feed lines to the active points and has a computing device which is assigned input devices and memories for the target values of the load parameter assigned to the zones and outputs for the control signals for carrying out the method.

In the roll machines considered here, the web material is mainly influenced by the line load (force per unit length) or compressive stress (force per unit area) prevailing in the press nip. It is therefore of interest to specify a setpoint value for a load parameter which is equal to or depends on the aforementioned variables and to ensure in operation that this value is at least approximately maintained. However, this is encountering difficulties because it is not possible to measure the forces occurring in the press nip during operation.

A method of the type described at the outset and an associated control arrangement are known (GB-A-2 156 101) in which the desired gap pressure distribution can be controlled. For this purpose there is a computing device which can be supplied with setting values for the desired pressure profile and which emits the control signals for pressure control valves. These control signals are calculated as a function of the web width, the bending properties of the roll shell, the number and construction of the bearing elements, etc. The computing device can also approximate the pressure profile using a mathematical calculation model based on the possible settings of the roll. Details of the calculation model are not given.

A method is already known (DE-A-2 825 706), in which a simplified mechanical model of the roller machine is used to determine the force distribution in the press nip. For this purpose, the rollers are replaced by bars. Pressure measuring elements are arranged in zones distributed between two rollers which simulate the press nip. On the other side of the bar, they are each assigned a pressure element simulating a bearing element. A controller is provided for each zone, to which on the one hand an adjustable setpoint and on the other hand the actual value measured by the pressure measuring element of a load parameter prevailing in the zone in question are supplied. The controller specifies a working pressure for the zone in question, which is supplied to both the bearing element of the original machine and the pressure element of the mechanical model. If the setpoint is changed in one zone, this affects the neighboring zones because of the rigidity of the bars, so that the working pressure can also be readjusted there with the help of the controllers assigned to these zones.

Calenders, smoothing units and other roller machines are of considerable size. The rollers are several meters long. It is extremely difficult to build a mechanical model that can replicate the original machine in every detail. In addition, essential data of the original machine change, for example when rollers with an elastic cover are turned off, which changes weight and rigidity, or when overhanging weights are varied, for example when a guide roller arrangement is changed as a result of a different web guide. The mechanical model cannot take all of this into account.

A method is also known (DE-A-3 117 516) in which auxiliary correction signals are triggered by an external correction of the pressure control signal for a bearing element group, which act on the group control signals of adjacent bearing element groups in a compensating effect. The conditions in the press nip are completely ignored here. As in the prior art described above, a change in one zone leads to a compensating change in the neighboring zones. However, the auxiliary correction signals used for this do not guarantee that the conditions in the other zones will remain unchanged if the line load changes in one zone.

There is also a control arrangement (EP-A-0 140 776) in which the shape and position of the roll shell and thus the profile and the thickness of the sheet to be rolled are regulated. For this purpose, position sensors and pressure sensors are used in connection with a computing device.

The invention has for its object to provide a method of the type described above, with the help of which it is possible with relatively little effort and without a mechanical model to control the individual active points in such a way that in the event of a change in the setpoint of the Load parameters in one zone the actual value of the load parameter can be adjusted in this zone and retains its previous value practically unchanged in the other zones.

This object is achieved according to the invention in that a pressure reaction matrix is formed, the elements of which indicate the change in the load parameter in all zones when there is a change in pressure at only one active site, in order to adapt the actual value of the load parameter to the target value using the pressure reaction matrix for the active site Zone a pressure change which completely or partially compensates for the difference between the actual value and the setpoint value and for all other zones a changed actual value resulting from this pressure change is calculated so that this calculation is repeated, the compensating pressure change being provided successively in each case at a different active point until an error function dependent on the differences falls below a tolerance value, and that the working pressure for each active point is changed by the sum of all pressure changes calculated for this active point.

In this procedure, a machematic tool is created by forming the pressure reaction matrix, which describes the roller machine to be controlled very precisely. Changes to the machine (turning off elastic rollers; replacing rollers; converting overhanging weights etc.) can be taken into account very simply by changing the die or individual die members.

With the pressure reaction matrix specified in this way, an iteration calculation process is carried out during operation, in which the effect of each pressure change on all zones is calculated and in which the errors occurring in the individual zones are eliminated by pressure changes until a tolerance value is undershot. The correct control signal for each zone can then be derived from all pressure changes, which leads to the desired setpoint profile of the load parameter in the press nip. Because of the presence of the pressure reaction matrix, this calculation method requires relatively little effort, so that one can get by with small memories and computers. The computing time is so short, even if 20 to 100 iteration steps are carried out, that this can be interrupted.

• a) für jede Zone wird ermittelt, um welchen Betrag sich der Lastparameter ändert, wenn der Druck in einer Wirkstelle um einen Betrag geändert wird, in allen anderen Wirkstellen aber gleich bleibt,
• b) diese Ermittlung wird für eine Druckänderung in allen Wirkstellen wiederholt,
• c) es wird eine Druckreaktionsmatrix gebildet, deren Glieder Quotienten aus Lastparameteränderung und Druckänderung sind, wobei die Zeilen jeweils einer zone und die Spalten jeweils einer Wirkstelle zugeordnet sind.

The following steps can be carried out before starting operation to form the pressure reaction matrix:

• a) it is determined for each zone by what amount the load parameter changes if the pressure in one active point is changed by an amount but remains the same in all other active points,
• b) this determination is repeated for a pressure change in all active points,
• c) a pressure reaction matrix is formed, the elements of which are quotients of the change in load parameters and the change in pressure, the rows each being assigned to a zone and the columns each being assigned to an active site.

All essential data of the original machine are systematically obtained insofar as they play a role in the calculation. The rows can run both horizontally and vertically, the reverse applies to the columns.

The elements of the pressure reaction matrix can be obtained in various ways. For example, they can be determined by measurements on the machine using material which reacts in a pressure-dependent manner and is to be introduced into the press nip. This includes NCR paper, which is then evaluated with a whiteness measuring device (e.g. from Elrepho).

Another preferred possibility is that the terms of the pressure response matrix are determined by calculations using a mathematical model of the machine. Such a model includes all the essential properties of the machine, such as the stiffness of the roll or of the support and of the roll casing, elasticity modules of the hard and covered rolls, overhanging weights and the like.

The calculation using the finite element method is particularly recommended, as is used in practice for numerous cases. However, there are also other types of calculation, for example using the transfer matrix method.

It has proven to be particularly favorable that when determining the terms of the pressure reaction matrix, a setpoint value of the load parameter which is constant over the length of the press nip and which is changed zone by zone is assumed. Comparable conditions then prevail for all members of the matrix.

• d) aus dem der Zone größter Differenz und der zugeordneten Wirkstelle zugehörigen Glied der Reaktionsmatrix wird eine Druckänderung berechnet, die eine der Differenz zwischen Istwert und Sollwert entsprechende Lastparameteränderung bewirkt,
• e) aus dieser Druckänderung wird mit Hilfe der in der gleichen Spalte der Druckreaktionsmatrix stehenden Glieder eine Lastparameteränderung in den übrigen Zonen berechnet,
• f) für jede Zone wird aus der Summe des bisherigen Istwerts des Lastparameters und seiner Änderung ein neuer Istwert gebildet,
• g) für eine zweite Zone wird aus dem dieser Zone und der zugeordneten Wirkstelle zugehörigen Glied der Druckreaktionsmatrix eine Druckänderung berechnet, die eine der Differenz zwischen neuem Istwert und Sollwert entsprechende Lastparameteränderung bewirkt,
• h) aus der letztgenannten Druckänderung wird mit Hilfe der in der gleichen Spalte der Druckreaktionsmatrix stehenden Glieder eine Lastparameteränderung in den übrigen Zonen berechnet,
• i) für jede Zone wird aus der Summe des zuletzt gültigen Istwerts des Lastparameters und seiner Änderung ein neuer Istwert gebildet,
• j) die Schritte g) bis i) werden für weitere Zonen wiederholt, bis eine die Differenz in den einzelnen Zonen berücksichtigende Fehlerfunktion unter einen Toleranzwert sinkt,
• k) für jede Wirkstelle wird aus der Summe des dort vorherrschenden Arbeitsdrucks und aller zugehörigen Druckänderungen ein neuer Arbeitsdruck gebildet, und es werden entsprechende Steuersignale an die Maschine gegeben.

During operation, it is advisable to carry out the following steps to adapt the actual value of the load parameter to the setpoint:

• d) a pressure change is calculated from the zone of the greatest difference and the associated active point of the reaction matrix, which causes a change in load parameters corresponding to the difference between the actual value and the setpoint,
• e) a change in the load parameters in the other zones is calculated from this change in pressure using the elements in the same column of the pressure reaction matrix,
• f) a new actual value is formed for each zone from the sum of the previous actual value of the load parameter and its change,
• g) for a second zone, a pressure change is calculated from the element of the pressure reaction matrix associated with this zone and the associated active site, which causes a change in load parameters corresponding to the difference between the new actual value and the setpoint,
• h) a change in the load parameters in the other zones is calculated from the last-mentioned pressure change using the elements in the same column of the pressure reaction matrix,
• i) a new actual value is formed for each zone from the sum of the last actual value of the load parameter and its change,
• j) steps g) to i) are repeated for further zones until an error function taking into account the difference in the individual zones drops below a tolerance value,
• k) a new working pressure is formed for each active point from the sum of the working pressure prevailing there and all associated pressure changes, and corresponding control signals are given to the machine.

In a further embodiment of the invention, a plurality of two-dimensional pressure reaction matrices are formed for different operating states of the machine and optionally used for the calculation depending on the operating state. This takes into account the fact that the conditions within the machine do not change linearly, so that optimum accuracy can only be obtained if different matrices are used for the calculation in different operating states. The matrices can be selected automatically or by the machine operator.

For example, pressure reaction matrices can be provided for at least two different target value ranges of the load parameter, for at least two different diameters of at least one roller, or for several mean temperatures of the roller surfaces. Different matrices can also be provided for different roll weights when changing rolls, for different overhanging weights, for different roll hardnesses, bed numbers or web properties.

• I) für jede Zone wird ermittelt, um welche Beträge sich der Lastparameter ändert, wenn die Temperatur in dieser Zone sich um mehrere vorbestimmte Werte ändert,
• m) die temperaturabhängige Lastparameteränderung wird jeweils als Korrekturglied in der Differenz zwischen Istwert und Sollwert des Lastparameters berücksichtigt.

The following additional steps are provided in a further embodiment:

• I) for each zone it is determined by what amounts the load parameter changes when the temperature in this zone changes by several predetermined values,
• m) the temperature-dependent load parameter change is taken into account as a correction element in the difference between the actual value and the setpoint of the load parameter.

In this way, a different temperature influence and the associated change in diameter of the rollers are taken into account. If the temperature rises in a zone, the pressure supplied to the associated active site can usually be reduced.

In this context, it is expedient if the temperature is measured over the length of the roller and the corresponding pressure reaction matrix or the temperature-dependent correction element is selected automatically as a function thereof.

If there are at least two bending compensation rollers, a reaction matrix with links for all zones and effective points of all bending compensation rollers should be formed. This takes into account the fact that when the pressure at the point of action of a roller changes, not only the other zones of this roller but also all zones of each additional roller experience a change in the load parameter.

If the bending compensation roller has external hydraulic cylinders as additional active points, it is advisable to assign an edge zone to each of them to determine the change in load parameters. In this way, the pressure for these outer hydraulic cylinders can also be calculated in the sense of an adaptation to the desired target value of the load parameter in the press nip.

A particularly quick calculation results if the pressure change is carried out for the active point of the zone in which the greatest difference between the actual value and the setpoint of the load parameter exists. This gives the smallest number of iteration steps required.

The calculation steps should be repeated at least as often as there are zones. As a rule, however, at least twice the number of iteration steps is run through before the tolerance value is undershot.

It is important in many cases that the calculation steps are repeated at least once for the zone with which the calculation was started. It has been found that the pressure changes which have been carried out in the other zones for the purpose of rectifying faults, in turn, have repercussions on the first zone which can only be compensated for by a correction of the pressure there.

The square root of the sum of the error squares for all zones has proven to be particularly suitable for the error function. This function ensures that the deviation of the calculated new actual value of the load parameter from the associated setpoint is particularly small in all zones.

The method described so far can also be integrated into a higher-level control loop. In particular, the setpoint profile can be changeable as a function of a web data control loop.

A control arrangement for a roller machine having at least two rollers for the treatment of web material in a press nip, in particular a calender or smoothing unit for paper, plastic or textile webs, in which the press nip has a number of zones, each of which has an active point which can be subjected to an adjustable pressure. Among them, individual bearing elements or groups of bearing elements subjected to the same pressure, which support the roll shell of a bending compensation roller on a non-rotatable carrier passing through the shell, are assigned, which control arrangement generates control signals for pressure control valves in the feed lines to the active points and has a computing device which is assigned input devices and memories for the target values of the load parameter assigned to the zones, as well as outputs for the control signals, for carrying out the method according to the invention is characterized in that the computing device has memories for the members associated with at least one pressure reaction matrix which change the Specify the load parameters in all zones when there is a pressure change at only one effective point, and that the computing device for obtaining the control signals for carrying out calculation steps in which the actual reaction value of the load parameter is adjusted to the target value using the pressure reaction matrix for the effective point of a zone Difference between the actual value and the target value, completely or partially compensating for the pressure change, and for all other zones a changed actual value resulting from this pressure change is calculated, and on the repetition of the calculation is programmed, the compensating pressure change being provided step by step at a different active point.

A control device is expediently connected between the computing device and the pressure control valves, which converts sudden changes in the control signals output by the computing device into a ramp function. The ramp function ensures a gradual change in the actual load parameter value in the press nip. This ensures that no undesirable vibrations or the like. occur.

Furthermore, it is favorable that a temperature measuring device is provided which can measure the roller temperature in the individual zones, and that the computing device has an input for the temperature measured values. This measuring device can have an individual measuring point for each zone or a measuring sensor which is moved back and forth along the roller.

In a further embodiment, a web data measuring device is recommended, which is capable of measuring actual values of web data at least in several places across the web width, and a converter connected upstream of the zone setpoint input devices, which determines the zone setpoints based on the web data. In this way, the computing device can be integrated into a higher-level control loop or a control system.

• Fig. 1 schematisch eine Biegeausgleichswalze mit zugehöriger Steueranordnung,
• Fig. 2 einen Kalander mit einer solchen Biegeausgleichswalze,
• Fig. 3 einen Kalander mit zwei Biegeausgleichswalzen,
• Fig. 4 einen Superkalander mit zwölf Walzen, davon zwei Biegeausgleichswalzen,
• Fig. 5 zweidimensionales Modell für den Superkalander der Fig. 5 zwecks Berechnung nach der finite Elementmethode und
• Fig. 6 die Darstellung der Fig. 5 bei Druckbelastung der einzelnen Wirkstellen.

The invention is explained in more detail below with reference to preferred exemplary embodiments illustrated in the drawing. Show it:

• 1 schematically shows a bending compensation roll with associated control arrangement,
• 2 shows a calender with such a bending compensation roller,
• 3 shows a calender with two bending compensation rollers,
• 4 shows a supercalender with twelve rollers, two of which are bending compensation rollers,
• FIG. 5 two-dimensional model for the supercalender of FIG. 5 for the calculation according to the finite element method and
• Fig. 6 shows the representation of Fig. 5 with pressure loading of the individual active points.

In the roller machine 1 according to FIGS. 1 and 2, an upper roller 2 and a lower roller 3 work together, which form a press nip 4 between them. The upper roller 2 is fixed in place in the frame 5. The lower roller 3 has a jacket 6, which is supported with the interposition of the press nip 4 facing primary bearing elements 7 and secondary bearing elements 8 arranged on the opposite side, and via roller bearings 9 and 10 located at the ends on a carrier 11 penetrating the jacket. Both the top roller 2 and the bottom roller 3 can be provided with an elastic cover. The carrier is held in a rotationally fixed manner at its free ends in spherical bearings 12 and 13 which can be pressed upwards in the active plane by means of hydraulic cylinders 14 and 15, respectively.

Hydraulic cylinders 14 and 15 are supplied with hydraulic fluid via pressure control valves V _L and V _R. The primary bearing elements 7 are combined in pairs into groups which are supplied with pressure fluid via pressure control valves V1 to V6. Similar valves can also be provided for the pairs of secondary bearing elements 8. The hydraulic cylinders 14 and 15 mentioned and the groups of primary bearing elements 7 are referred to below as "active points" which can be subjected to adjustable pressure. Each working point is assigned a specific zone in the press nip, namely the hydraulic cylinder 14 the one edge zone Z _L and the other hydraulic cylinder 15 the edge zone Z _R. The zones Z to Z ₆ located in between each correspond to the groups of primary bearing elements 7 shown below. The secondary bearing elements 8 only serve to clamp the roll shell and are supplied with constant pressure. Only if they should be loaded with changeable pressure during operation are they to be regarded as "active sites" in the aforementioned sense and would then be assigned to zones Z and Z _s .

In order to determine the control signals which are supplied to the pressure control valves mentioned in order to determine the pressure to be emitted by them, a programmable computing device 16 is provided which uses input points 17 with setpoints q _{sa "} for a load parameter prevailing in the press nip 4, in particular the line load or compressive stress, The computing device 16 outputs control signals pson, which correspond to the pressure to be supplied to the individual active points, via a data line 18. These control signals are fed to a programmable logic controller 19, which compares these control signals with the actual pressure values p _is , via the lines 20 are supplied, and then emits corresponding actuation signals y to the valves via lines 21. In addition, the controller 19 ensures that, in the event of sudden changes in the pressure setpoint p _, the actuation signals output via lines 21 should run according to a ramp function, a so only a gradual change occurs.

A memory 22 is connected to the computing device 16, which on the one hand holds the target values of the load parameter in the individual zones and on the other hand a plurality of pressure reaction matrices, as will be explained in detail later. The latter are introduced via the input point 23.

Furthermore, the computing device 16 is connected to a temperature sensor 24 which, in a known manner, measures the surface temperature T of one roller, in particular of the related roller 2, at various points along its length, as is known, for example, from DE-PS 31 31 799.

The setpoint qs.11 of the load parameter can be set manually at the input points 17, as illustrated on the left in FIG. 1. However, the setpoint can also come from an upstream converter 25, to which a measuring device 26 - also known from DE-PS-31 31 799 - supplies web data w measured over the width of the web, such as web thickness, gloss, smoothness or the like. As is known, this train data can be influenced by changing the line load in corresponding zones.

While only one press nip 4 is present in the exemplary embodiment described so far, FIG. 3 shows a roller machine 101 in which a central roller 102 is fixedly mounted in the frame 105. A lower roller 103 can be pressed upward in a manner similar to that of FIGS. 1 and 2, while an upper roller 127 can be pressed against the central roller 102 in the manner of a spatula. Thus two press nips 104 and 128 are available.

In the embodiment according to FIG. 4, a supercalender 201 is provided, in which six related rollers 229 to 234 and four hard rollers 235 to 238 are arranged between a lower bending compensation roller 203 and an upper bending compensation roller 227. The lower roller 203 corresponds to the roller 3 in Fig. 1 with the difference that the bearings 12 and 13 for the carrier 11 are held fixed to the frame during operation. The roller 227 corresponds to an upside down roller 3 of FIG. 1 with the difference that the coupling of the roller shell 6 to the carrier 11 by the roller bearings 9 and 10 is omitted and the shell 6 as a whole thus move radially relative to the carrier 11 can.

• a) Festlegung einer Druckreaktionsmatrix für die betreffende Walzenmaschine und
• b) Berechnung der erforderlichen Steuersignale unter Verwendung dieser Matrix.

In all of the roller machines described above, efforts are made to keep the actual value of the load parameter, such as plug load or compressive stress, equal to a desired setpoint profile in the press nip and to track it zone by zone when setpoint changes occur due to web monitoring or measurement. Since such roller systems not only react immediately to a zone correction where an adjustment has been made, a control is necessary which adjusts the effective point pressures in such a way that the desired effects really occur where they are desired. According to the invention, two measures are provided for this, namely

• a) definition of a pressure reaction matrix for the roller machine concerned and
• b) Calculation of the required control signals using this matrix.

a) Definition of a pressure reaction matrix

To create such a pressure reaction matrix, as explained in connection with FIGS. 5 and 6, a finite element model of the roller machine is created. The finite element method is a numerical calculation method with which complex problems are broken down into small individual problems (elements) that are accessible to a solution. Depending on the desired accuracy of the calculation, a roller system can be broken down into three-dimensional elements or into two-dimensional elements. A three-dimensional description reproduces the structure more precisely, but leads to a more complex calculation. A two-dimensional calculation model for the supercalender of FIG. 4 is shown in FIGS. 5 and 6.

The horizontal lines correspond from top to bottom of the roller shell 6 of the top roller 227, the covered roller 229, the hard roller 235, the covered roller 230, the hard roller 236, the covered roller 231, the hard roller 237, the covered roller 232, the covered roller 233, the hard roller 238, the related roller 234 and the roller shell 6 of the lower roller 203. The latter is supported by its roller bearings 9 and 10 at the specified points. The horizontal lines a thus correspond to the rolls or roll shells. The vertical connections b are contact elements that simulate the elastic behavior of the roll covers - or when the web material becomes smooth. The influence of the bearing elements 7 and 8 and the hydraulic cylinders 14 and 15 is represented by forces at the corresponding points of attack. The subdivision into individual fields is such that a finite element is present at least for each of the zones, so that a zone-by-zone assignment is exactly possible for the load application. Each roller is included in the calculation with regard to its rigidity and its weight, whereby outer diameter, inner diameter, modulus of elasticity, transverse number and density can be entered. The compression behavior of the elastic coverings is also entered depending on the material and diameter pairing for the contact elements b. The overhanging weights caused by bearings, guide rollers, protective brackets etc. are applied as forces at the roller bearing points.

The two-dimensional model of FIG. 5 changes under load, as is indicated in FIG. 6 in a greatly enlarged deformation. It can be seen that the compression elements b in particular have been greatly reduced. Considerable compression can be found in the area of the two adjacent rollers 232 and 233.

First, the effective point pressures are calculated so that there is a constant basic line load in the lower press nip. This can be done for different stress levels. With the characteristic field obtained in this way, uniform distance loads can be set in the calender.

In order to be able to control the calender zone by zone, you need information on how the roller system reacts to a change in a zone. For this purpose, the pressure of each individual active point is changed by a certain amount based on the constant setpoint of the load parameter. The change in the load parameter is determined at certain reference points, in particular in the middle of the zones Z to Z ₆ and at the edge of the zones Z _L and Z _R. If these changes are summarized in a matrix, the so-called pressure reaction matrix R _{ij of} the calender is obtained, as shown in the appendix to the formula (1). Ap means the change in pressure, Aq the change in the load parameter, the numbers 1, 2 ... i, j ... n mean the numbering of the zones or active points. The rows each correspond to a zone, the columns each to an effective point.

In the supercalender of FIG. 4 and in a compact calender according to FIG. 3, where two bending compensation rollers work against each other, the pressure reaction matrix R, _{j has} a number of rows and columns corresponding to twice the number of zones, because every change in the pressure in an effective point of the one bending compensation roller not only has an influence on the other zone of this roller, but also on all zones of the other bending compensation roller. For example, if you change the working pressure of an effective point in the upper bending compensation roller, the line load in the gap of the lower bending compensation roller also changes.

If hydraulic cylinders also play a role, additional zones must be taken into account in the matrix R _ij ^LR (Tm), as is illustrated in (2).

It has already been mentioned that different matrices can be set up for different loading conditions. (2) shows that different matrices can also be determined for different temperature mean values T _m . Changes must also be made when the machine is tampered with, for example by turning the rollers or changing the overhanging weights.

b) calculation of the control signals

It is assumed that the actual value of the load parameter in the individual zones is equal to the specified target value q _soll if there are corresponding working pressures p _io , _Pjo . Now comes the command to change the setpoint in zone i by the value Δq. This change in the setpoint corresponds to a change in pressure △ _pi at the associated active point according to the formula (3), the running number n = 1 here. When adjusting in zone i, however, deviations arise, for example in zone j, k etc., as indicated by the formulas (4). Now a new actual value of the load parameter can be calculated in each zone according to the formulas (5). In the zone in which the actual value has the greatest deviation from the target value, the difference is compensated for by another pressure change. This step-by-step calculation is repeated until the error F "according to the function (6) is less than a certain tolerance value.

The pressures p, p _j for the individual active points, which are _{to be sent} to the machine as control signal p _, are calculated according to the formulas (7) from the original working pressure and the sum of all pressure changes calculated in the iteration steps. The error function F "corresponds to the square root of the sum of the error squares of the load parameters in the individual zones.

The iteration approximation can also be used when the calender is to be put into operation. Then the actual value of the load parameter in the columns of the reaction matrix is set equal to the basic line load. The computing device 16 checks in which zone the greatest deviation between the target value and the actual value is present. This zone is fully regulated in one step, whereupon the calculation scheme proceeds as described.

In some cases, it is advisable not to correct the difference completely, but only by 80%, for example, if this means that the tolerance value can be fallen below more quickly.

As already mentioned, the setpoint can be specified by path data w with the aid of the converter 25, so that the process described is carried out by the path or even integrated into a higher-level control loop.

The computing device can also automatically select the correct reaction matrix for the respective calculation process. The average load that comes closest to one of the matrices can be derived from the setpoint profile. In the same way, the temperature reaction pressure matrix can also be selected with the aid of the temperature sensor 24.

When the roll temperature changes, its diameter changes and, in the case of plastic-related rolls, the hardness (modulus of elasticity) of the roll surface also changes. This can lead to a change in the distributed load distribution. If the overall temperature level changes, this can be taken into account by means of a different pressure reaction matrix. However, if the temperature changes in the longitudinal direction of the roller, undesirable changes in the load parameter result. If, for example, the line load in one zone is increased compared to the other zones, the roller cover in this zone heats up due to the increased flexing work, which results in an increase in diameter. As a result, the line load continues to increase until the desired setpoint of the load parameter can no longer be maintained. Taking into account the measurement of the roller temperature T, the control can make such a correction that the desired setpoint remains set despite the heating of the cover.

For this purpose, temperature reaction matrices D _ij (T _m ) are created for different mean temperatures, each taking into account the change Aq of the load parameter in a zone for different temperature changes △ T ₁ , AT ₂ ..., as shown in (8) . Here corresponds to Numbering of parameter changes and temperature changes of the zone numbering.

This control works as follows: From the temperature measurements, the mean value is calculated, which stands for the relevant temperature level. With the mean roller temperature, the temperature deviation in each zone is now determined, as indicated in (9). With these temperature differences, the parameter changes in the press nip according to formula (10) can now be enriched using the temperature reaction matrix D _ij (T _m ). The actual value of the load parameter in each zone therefore results from the current pressure setting in the active points and from the temperature distribution, as indicated by (11). This proportion of the load parameter, which is dependent on the temperature, must be taken into account when comparing the actual value of the load parameter with the target value, for example in the context of the formulas (12) or (13). The internal iteration steps for enriching the pressure setting can then be carried out with the setpoint specified in this way.

For example, an IBM 7535 device from IBM or a DEC 11/53 device from Digital Equipment Corporation can be used as computing device 16. A commercially available memory of 500 kB is sufficient as the memory 22. Devices S ⁵ , for example, come as programmable logic controller 19 _­ 150 U from Siemens or the device A 500 from AEG. Formula attachment

Claims (24)

1. Method for operating a roll machine, having at least two rolls, for the treatment of web material in a pressing nip, in particular a calender or glazing rolls for paper, plastic or textile webs, in which the pressing nip (4) has a number of zones (Z1 to Z6, ZL, ZR), each of which is assigned to one point of action subjectable to adjustable pressure (p) - including individual bearing elements or groups of bearing elements (7, 8) subjectable to the same pressure, which support the roll shell (6) of a bending-compensation roll (3) on a support (11) which passes through the shell (6) and is fixed in terms of rotation, in which method a working pressure (p) is determined for each point of action with the aid of a computing operation, said working pressure depending on the setpoint value profile of a load parameter (q), characterized in that a pressure reaction matrix (R) is formed, the elements of which indicate the change of the load parameter (q) in all zones (Z1 to Z6, ZL, ZR) in the case of a pressure change (Ap) at in each case only one point of action, in that, to adapt the actual value (qactua_l) of the load parameter to the setpoint value (qset) using the pressure reaction matrix (R), a pressure change (Ap) completely or partially compensating the difference between actual value and setpoint value is calculated for the point of action of one zone (e.g. Z1) and a changed actual value (q_actual) resulting from this pressure change is calculated for all other zones (e.g. Z2 to Z6, ZL, ZR), in that this calculation is repeated, the compensating pressure change being provided stepwise in succession at another point of action in each case, until an error function (F") dependent on the differences undershoots a tolerance value, and in that, for each point of action, the working pressure (p) is changed by the sum of all pressure changes (Ap) calculated for this point of action.

2. Method according to Claim 1, characterized in that, before starting operations, the following steps are carried out:

a) it is determined for each zone (Z1 to Z6, ZL, ZR) by what amount the load parameter (q) changes when the pressure (p) in one point of action is changed by an amount that remains the same in all other points of action,

b) this determination is repeated for a pressure change (Ap) in all points of action,

c) a pressure reaction matrix (R) is formed, the elements of which are quotients of load parameter change (Aq) and pressure change (Ap), the lines in each case being allocated to a zone and the columns in each case being allocated to a point of action.

3. Method according to Claim 1 or 2, characterized in that the elements of the pressure reaction matrix (R) are determined by measurements at the machine using material which is to be introduced into the pressing nip (4) and reacts as a function of the pressure.

4. Method according to Claim 1 or 2, characterized in that the elements of the pressure reaction matrix (R) are determined by calculations using a mathematical model of the machine.

5. Method according to Claim 3, characterized in that the calculation is performed according to the finite element method.

6. Method according to one of Claims 1 to 5, characterized in that, in the determination of the elements of the pressure reaction matrix (R), a setpoint value (q_set), constant over the length of the pressing nip, of the load parameter, which is altered zonewise, is taken as a starting point.

7. Method according to one of Claims 1 to 6, characterized in that, in operation, the following steps are carried out in order to adapt the actual value (q_act_ual) of the load parameter to the setpoint value (qset):

d) from that element of the reaction matrix which is associated with the zone (e.g. Z1) of greatest difference and with the allocated point of action a pressure change (Ap) is calculated which effects a load parameter change (Aq) corresponding to the difference between actual value and setpoint value,

e) with the aid of the elements in the same column of the pressure reaction matrix (R), a load parameter change (Aq) in the remaining zones (e.g. Z2 to Z6, ZL, ZR) is calculated from this pressure change (Ap),

f) a new actual value is formed for each zone from the sum of the previous actual value of the load parameter (q) and its change (Aq),

g) a pressure change (Ap) is calculated for a second zone (e.g. Z2) from that element of the pressure reaction matrix (R) which is associated with this zone and with the allocated point of action, which pressure change effects a load parameter change (Aq) corresponding to the difference between new actual value and setpoint value,

h) with the aid of the elements in the same column of the pressure reaction matrix (R), a load parameter change (Aq) is calculated in the remaining zones (e.g. Z1, Z3 to Z6, ZL, ZR) from the last-mentioned pressure change,

i) a new actual value is formed for each zone from the sum of the last valid actual value of the load parameter (q) and its change (Aq),

j) steps g) to i) are repeated for further zones (e.g. Z3) until an error function taking into account the difference in the individual zones falls below a tolerance value,

k) a new working pressure is formed for each point of action from the sum of the working pressure (p) prevailing there and all associated pressure changes (Ap) and corresponding control signals are sent to the machine.

8. Method according to one of Claims 1 to 7, characterized in that a plurality of two-dimensional pressure reaction matrices (R) are formed for various operating states of the machine and used alternatively, as a function of the operating state, for the calculation.

9. Method according to Claim 8, characterized in that pressure reaction matrices (R) are provided for at least two different setpoint value ranges of the load parameter.

10. Method according to Claim 8 or 9, characterized in that pressure reaction matrices (R) are provided for at least two different diameters of at least one roll.

11. Method according to Claims 8 to 10, characterized in that pressure reaction matrices (R) are provided for a plurality of mean temperatures of the roll surfaces.

12. Method according to one of Claims 1 to 11, characterized by the following additional steps:

I) for each zone (Z1 to Z6, ZL, ZR) it is determined by what amounts the load parameter (q) changes when the temperature in this zone changes by a plurality of predetermined values,

m) the temperature-dependent load parameter change (Aq) is in each case allowed for as correction term in the difference between actual value (qactua_l) and setpoint value (qset) of the load parameter.

13. Method according to Claim 11 or 12, characterized in that the temperature over the length of the roll (3) is measured and the corresponding pressure reaction matrix (R) or the temperature-dependent correction term (D) automatically selected as a function of this.

14. Method for a roll machine having at least two bending-compensation rolls according to one of Claims 1 to 13, characterized in that a reaction matrix with elements for all zones and points of action of all bending-compensation rolls (103, 127; 203, 227) is formed.

15. Method according to one of Claims 1 to 14, characterized in that the bending-compensation roll (3) has outer hydraulic cylinders (14, 15) as additional points of action and one edge zone (ZL, ZR) is allocated to each of them for the determination of the load parameter change.

16. Method according to one of Claims 1 to 15, characterized in that the pressure change (Ap) is in each case carried out for the point of action of that zone (e.g. Z1) in which the greatest difference between actual value and setpoint value of the load parameter (q) exists.

17. Method according to one of Claims 1 to 15, characterized in that the calculation steps are repeated at least as often as there are zones.

18. Method according to one of Claims 1 to 17, characterized in that the calculation steps are repeated at least once for the zone (e.g. Z1) with which the calculation was begun.

19. Method according to one of Claims 1 to 18, characterized in that the error function is formed by the square root of the sum of the residual squares for all zones (Z1 to Z6, ZL, ZR).

20. Method according to one of Claims 1 to 19, characterized in that the setpoint value profile can be changed as a function of a web data control loop.

21. Control arrangement for a roll machine having at least two rolls, for the treatment of web material in a pressing nip, in particular a calender or glazing rolls for paper, plastic or textile webs, in which the pressing nip (4) has a number of zones (Z1 to Z6, ZL, ZR), each of which is allocated to one point of action subjectable to adjustable pressure (p) - including individual bearing elements or groups of bearing elements (7, 8) subjected to the same pressure, which support the roll shell (6) of a bending-compensation roll (3) on a support (11) which passes through the shell (6) and is fixed in terms of rotation, which control arrangement produces control signals (p_set) for pressure control valves (V1 to V6, VL, VR) in the supply lines to the points of action and has a computing device (16) to which are allocated input devices (17, 23) and memories (22) for the setpoint values (qset), allocated to the zones (Z1 to Z6, ZL, ZR), of the load parameter and outouts (18) for the control signals (pset), for carrying out the method according to one of Claims 1 to 20, characterized in that the computing device (16) is allocated memories (22) for the elements of at least one pressure reaction matrix (R), which indicates the change of the load parameter (q) in all zones (Z1 to Z6, ZL, ZR) in the case of a pressure change at in each case only one point of action, and in that, to obtain the control signals (pset) the computing device (16) is programmed to carry out calculation steps in which, to adapt the actual value (qactua_l) of the load parameter to the setpoint value (qset) using the pressure reaction matrix (R), a pressure change (Ap) completely or partially compensating the difference between actual value and setpoint value is calculated for the point of action of one zone (e.g. Z1) and a changed actual value (qactuai) resulting from this pressure change is calculated for all other zones (e.g. Z2 to Z6, ZL, ZR), and to repeat the calculation steps, the compensating pressure change being provided stepwise in succession at another point of action in each case.

22. Control arrangement according to Claim 21, characterized in that a control device (19) which converts sudden changes of the control signals (q_set) output by the computing device into a ramp function is connected between computing device (16) and pressure control valves (V).

23. Control arrangement according to Claim 21 or 22, characterized in that a temperature measuring device (24) is provided which is able to measure the roll temperature in the individual zones, and in that the computing device (16) has an input for the measured temperature values (T).

24. Control arrangement according to one of Claims 21 to 23, characterized by a web data measuring device (26) which is able to measure actual values of web data (w) at at least a plurality of points, transversely across the web width, and by a converter (25) which is connected upstream of the zone setpoint value input devices (17) and, on the basis of the web data, determines the zone setpoint values.

Images