EP3504062B1 - Method for controlling the drive of a machine - Google Patents

Description

The present invention relates to a method for controlling the drive of a machine with at least one first rolling element, which rolls on a surface at least in partial intervals of a rotation cycle under the action of a contact force with elastic deformation on a counter surface.

Due to the effect of the contact force, a body, which is referred to in the following in a non-restrictive manner as a roller element, is elastically deformed when it rolls on a counter surface. This elastic deformation changes the effective radius of the roller element. When two roller elements roll against each other, both radii change and thus, together with the roller speed, the line speed of a product that is transported through the pair of rollers.

In the context of the present disclosure, “roller elements” generally refer to machine elements that rotate about a fixed or moving axis. The roller elements can be essentially cylindrical, or can be designed as profile rollers.

In general, the speed of the roll surface, or the relative speed of roll surfaces rolling against one another, is set by specifying a target speed for the rolls. Since the effective radii under the action of a contact force are not known a priori, there is a torque exchange between the rollers due to the elastic deformation. This means that one roller exerts an accelerating torque on the other roller, while the other roller exerts a braking torque on the former.

In addition, periodic movement disturbances (disturbing moments) occur in practice. Examples of such disorders are e.g. a printing plate which is mounted on one of the rollers and does not enclose the entire circumference of the roller, as well as the printing motif which is stamped on this printing plate.

Processes of the type mentioned at the outset are used, for example, in rotary printing processes with elastic printing forms, such as the flexographic printing process. Numerous problems can arise in everyday printing technology, the handling and solution of which require a good training and extensive experience from the operator of the printing press. The appearance of horizontal stripes in the printed image is such an undesirable phenomenon in flexographic printing and these horizontal stripes are among the well-known problems in everyday printing life.

If the printing form does not completely cover the circumference of the forme cylinder, it is known that the gap in the area of the surface of the form cylinder not covered by the printing form can cause vibrations in the printing block, which have a negative effect on the printed image. These vibrations occur when the printing form comes into contact with the anilox roller or the impression cylinder or releases this contact again.

It is also known that the geometric shape of the printing form surface determined by the printing motif can cause motif-induced vibrations, which likewise have a negative effect on the printed image.

Even if such vibrations play a role, the exact causes of horizontal stripes in the printed image are often not easy to determine, so that measures to prevent them are usually found out by trial and error. In printing technology practice, attempts are made to overcome this undesirable phenomenon by various measures such as Change in the print delivery, change in the print image length, distribution of a printing ink over several printing units, manual change of roller diameters on the operating device of the machine control, use of special sleeves and adapters, suitable choice of adhesive tape, etc. to suppress. All of these measures require manual interventions by experienced operators of the printing press that are specifically tailored to the respective print job, or they have to be taken into account during the production of the printing form.

DE 10 2012 013532 A1 discloses a method in which full contact is made in the printing nip between printing form and printing material. Due to the extremely low relief depth on the entire surface, very small tolerances must be observed, which can prove difficult in practice.

DE 69400403 T2 discloses a printing method in which an instantaneous intensity or a motor torque of a drive motor is measured and the rotational speeds of printing cylinders are adjusted so that when a rubber element is passed between a first and a second cylinder on the one hand and between a first and a third cylinder on the other hand, a minimal change in the instantaneous intensity or the motor element is effected.

EP 1 759 839 A2 discloses a printing press with a forme cylinder, a rubber cylinder rolling on the forme cylinder and a dampening roller rolling on the forme cylinder. The synchronism of the forme cylinder and blanket cylinder is ensured by a control device, the position or angular position of the forme cylinder and blanket cylinder being monitored with appropriate sensors. In a so-called "Delta operation" the surfaces of the dampening roller and the forme cylinder roll on each other at different surface speeds. As a result, a torque builds up with each revolution, which "breaks down" when the dampening roller rolls over the clamping channel of the forme cylinder, which leads to a disturbed print image due to the asynchrony of the forme cylinder and blanket cylinder. These deviations are compensated for by an adaptive control device which uses a retrospective method.

DE 10107135 A1 discloses a vibration damping method in a web printing press. Deviations in position of the forme cylinder are determined with sensors and compensated for with a non-linear control, in particular a sliding mode control.

There is also a need for methods and devices for improving the processing quality and in particular for avoiding horizontal stripes in a printed image, which should function as independently as possible from the experience of the operator. If possible, the process should not require any additional effort in production preparation or conversion, for example in printing form manufacture, and should be simple to implement.

This and other objectives are achieved according to the invention by a method according to independent claim 1. A speed correction method, which automatically compensates for deformation-related deviations in the circumferential speed of the first roller element by adapting the setpoint for the speed of the first roller element, and a retrospective method, which deviations in the speed constancy of the first roller element within a revolution cycle by applying a control variable , in particular an actual speed or an actual position of the roller element, automatically compensates for the correction signal determined in one or more previous rotation cycle or cycles, combined. In experiments, the inventor found that neither the use of a speed correction method alone nor the use of a retrospective method to increase the speed constancy alone could solve the quality problems. Only through the combined use of these two processes was it possible to achieve an increase in quality that was not to be expected due to the disappointing results that were achieved with the individual processes alone. The method according to the invention has the advantage that it can also be easily implemented in existing machines.

In connection with the present disclosure, the term “being in contact” means a roller element with a counter surface or roller pairs with one another not only in direct contact but also in contact with an intermediate one Product, in particular a printing material, which is passed between a pair of rollers during printing, for example. Furthermore, the term “being in contact” does not necessarily mean that the roller element touches the counter surface during the entire rotation time.

The term “roll on an opposing surface with elastic deformation” is understood in the context of the present invention in such a way that the surfaces of the elements rolling against each other lie essentially without sliding, at least in the contact area.

The term “contact area” in connection with the present invention denotes the area in which a pair of rollers (or a roller element and a counter surface) touch, possibly with the interposition of a product or printing material. In the case of an idealized, rigid pair of rollers, the contact area can be represented in a cross section as a "contact point". Surface or relative speeds, for example, are generally related to a calculated point of contact, where it is clear to the person skilled in the art that real, elastically deformable rollers are contact surfaces. The terms "area of contact" and "point of contact" can therefore generally be used synonymously.

In the context of the present invention, the term "revolution cycle" refers to the period of time in which the characteristic peaks in the controlled variable are typically repeated cyclically, the revolution cycle in particular being able to correspond to the period of the first roller element or another roller element.

For example, the speed (or a value derived from the speed, such as the peripheral speed) of the roller element can be used as the control variable. The speed is generally derived from an angle of rotation signal generated by a rotary encoder on the drive motor or on the roller element. The retrospective method for increasing the speed constancy compensates for peaks in the deviation of the controlled variable that occur regularly at certain points in the rotation cycle by the correction signal. Instead of the controlled variable per se, the correction signal can also be derived from a variable influenced by the controlled variable.

The method according to the invention is described in connection with the present disclosure based on the speed. However, the person skilled in the art is readily able to implement the invention on the basis of the angle of rotation (or of rotational positions). The setpoints are then not presented as constant speed specifications, but as linearly increasing position or angle specifications. Such designs are to be regarded as analog embodiments.

In order to compensate for the dead time of the control loop (i.e. the time shift between the controller output and the actual cross current on the drive motor), the correction signal must be "reset" by this dead time in order to synchronize the correction with the value peaks to be corrected.

The invention is based on the knowledge that differences in the effective circumferential speeds - due to the elastic deformation which leads to changes in the effective roller diameters which are extremely difficult to predict beforehand - of roller pairs represent a cause of quality errors, for example of transverse stripes in a printed image. The same quality defects can also occur if a roller element generally rolls on a counter surface. By eliminating these speed differences, these quality errors can largely be prevented.

Advantageously, for the speed correction method, the time profile of a value characteristic of a drive torque of the roller arrangement can be determined in at least one of the partial intervals, a parameter for an increase in this value in the partial interval can be derived therefrom and the reference variable of the peripheral speed of the first roller element as a function be adjusted by this parameter to minimize the increase.

To detect a speed difference between the surface of the first roller element and the counter surface, the course of the drive torque for the roller element or the course of a variable proportional to the drive torque, such as, for example and without limitation, the drive current or the drive power in the selected subinterval can be evaluated.

According to a further advantageous embodiment, the value characteristic of the drive torque can be a force applied by the at least one roller element to the counter surface.

The characteristic value for a drive torque of the roller arrangement can be a parameter representative of a speed error of the roller element and / or a drag error between the roller element and the counter surface, such as, for example and without limitation, an average or effective torque, an average slope of a torque, an average force effect (torques evaluated with the roller radius) on the counter surface or on another roller element or a sum of force effects of roller pairs on each other. Based on this, the peripheral speed is adjusted so that the representative parameter is minimized in this subinterval.

The derived parameter can furthermore be a value averaged in a partial interval or it can be an optionally smoothed increase in the drive torque in the partial interval.

The subinterval can include one or more complete revolutions of one of the roller elements. Advantageously, a peripheral speed of the first roller element can be adjusted by changing the default value for the speed of this roller element.

In a further advantageous embodiment, a peripheral speed of the first roller element can be adjusted by changing the default value for the diameter of this roller element. The deviation (e.g. in percent) of the default value from the actual diameter can be evaluated as a characteristic value, for example to avoid quality problems, that announce itself by a change in this value, to recognize early.

In a further advantageous embodiment, a peripheral speed of the first roller element can be adjusted by changing the default value for the feed of this roller element. As a result, the value for the diameter, which is relevant for the relative speed between two roller elements, also changes due to elastic deformation. In printing machines, for example, the ink absorption from an anilox roller to a forme cylinder and the pressure from the forme cylinder on the printing material can be advantageously influenced at the same time.

A shooting method can advantageously be used to determine the default value for the speed of rotation of a roller element, the default value for the diameter of the roller element or the default value for the infeed of the roller element. The shooting process is an iterative process that can run automatically and in which, based on a starting value, e.g. for the specified value of the speed of the roller element, the associated target parameter is determined from the resulting torque curve in the sub-interval mentioned. The starting value is then slightly changed and the resulting deviation is determined. The optimal operating point is set automatically by successive linear inter- and extrapolations. The iterations are stopped when the desired operating point has been reached with sufficient accuracy and therefore convergence has been achieved.

The retrospective method can advantageously be a self-learning method for controlling cyclical processes, in particular a repetitive control method. Such methods are well suited for increasing the speed constant. In the context of the present invention, self-learning methods for controlling cyclical processes are generally methods in which faults (for example setpoint deviations or errors) are determined and stored in at least one first cycle, measures for suppressing these faults being determined on the basis of the faults determined , and wherein these measures are used in at least one further cycle in order to suppress corresponding disturbances in this further cycle. The permanently repetitive application of this methodology suppresses self-learning in the best possible way.

The repetitive control process is a well-known process which is described, for example, in the technical article " Repetitive control for systems with uncertain period-time ", Maarten Steinbuch, Automatica 38 (2002) 2103-2109 , is described. The repetitive control method can be used to determine the occurrence of (then also periodic) To minimize disturbances. The process is self-learning due to the permanent repeated application per se and is therefore very easy to use. In the application at hand, it can advantageously be used to increase the speed constant, which is achieved, for example, by a self-learning additive current application.

Advantageously, the retrospective method can use a speed error, possibly scaled with a speed controller gain, and / or a periodic following error occurring at the point of contact between the work roll and the counter-pressure cylinder, whereby a switchover between different variants of determining a feedback signal can be provided, with several alternatives Provide operating modes. For example, the feedback can be a signal determined by a motor encoder, representative of the controlled variable, or a signal determined by a load encoder representative of the controlled variable. The feedback signal can also be generated by a virtual load transmitter which, for example, determines an estimated value for the controlled variable on the basis of the drive current and the motor torque.

The retrospective method can advantageously go through an initiation phase over at least one rotation cycle, preferably over at least two rotation cycles. A revolution cycle can in particular correspond to one revolution of the first roller element (or also another roller element if this defines the revolution cycle), so that in many application areas it is not necessary to determine the cycle from a signal. After switching on the self-learning method, an internal memory is initiated over the first revolution cycle. In the second rotation cycle (and possibly in further cycles), the process is activated continuously or step-by-step by means of a loop-in control, so that the interference suppression is fully active after two cycles (or two rotations).

In a further advantageous embodiment of the invention, the counter surface can be formed by the surface of a second roller element, the first roller element and the second roller element rolling against one another and the drive of the second roller element being regulated analogously to the first roller element.

The retrospective method for increasing the speed constancy can advantageously use a periodic following error occurring at the point of contact between the first roller element and the second roller element as a controlled variable.

In an advantageous embodiment, in which the machine is a printing press, the first roller element can be a forme cylinder, an impression cylinder and a Roll the anilox roller on the forme cylinder, and an elastic printing form is applied to the forme cylinder, which is in contact with the anilox roller and / or the impression cylinder during at least a partial interval of the rotation of the forme cylinder. This enables horizontal stripes to be prevented reliably in the printed image.

The present disclosure thus also relates to a method for controlling a printing press with a plurality of roller elements, namely at least one forme cylinder, an anilox roller and an impression cylinder, an elastic printing forme being applied to the forme cylinder, which during at least a partial interval of the rotation of the forme cylinder with the anilox roller and / or is in contact with the impression cylinder, the course of a value characteristic of the drive torque of at least one of the roller elements being determined in the partial interval, a parameter being derived therefrom and the peripheral speed of at least one of the roller elements being adapted as a function of this parameter.

The subinterval evaluated for the calculation is selected in connection with the present invention on the basis of the respective machine. In the case of printing presses, the size of the forme cylinder, the size, shape and position of the printing form and the arrangement of the further roller elements are taken into account. The partial interval can be determined on the basis of an evaluation of the course of characteristic values, such as the drive torque, the measured roller speed, the speed error, the tracking error or other suitable characteristic values, during a test run or when starting up the printing press and can also include one or more complete revolutions .

The selected subinterval should be selected in such a way that interference is minimized. In a further advantageous embodiment of the invention, the partial interval is therefore selected such that the points of contact between the printing form and the anilox roller and between the printing form and the impression cylinder are free of changes of contact in the partial interval. This enables a stable evaluation of the parameters with minimal interference. A "change of contact" is understood to mean coming into contact and releasing the contact between the printing form and another roller element.

In one embodiment of the method according to the invention, the method can be carried out automatically and optionally also regularly during the printing process. This allows an automatic elimination of differences in peripheral speed between contacting roller pairs and the associated print image errors. This enables an automatic process to be created which is controlled independently by the machine software. No manual interaction by the operating personnel is necessary, no additional prepress work is required and no additional printing units are required. The feature of the automatic and optionally also regular execution of the method can also be applied according to the invention to machines that are not printing machines.

In an advantageous embodiment, an identical line speed can be set by adapting the peripheral speed on a machine with several printing units. This feature can also be applied analogously to other machines.

• Fig.1 eine schematische Darstellung des Druckvorgangs einer Flexodruckmaschine;
• Fig. 2 eine schematische Darstellung des zu erwartenden Antriebsmoments während einer Umdrehung des Formzylinders;
• Fig. 3 ein Diagramm des dynamischen Verhaltens einer Druckmaschine in einer Versuchsanordnung;
• Fig. 4 eine vergrößerte Darstellung eines Teilbereichs des Antriebsmoments der Fig. 3;
• Fig. 5 ein Diagramm des dynamischen Verhaltens der Druckmaschine in der Versuchsanordnung nach einer ersten Anpassung des Vorgabewerts für den Durchmesser des Formzylinders;
• Fig. 6 ein Diagramm des dynamischen Verhaltens der Druckmaschine in der Versuchsanordnung bei einer Verwendung eines Repetitive-Control-Verfahrens;
• Fig. 7 ein Diagramm des dynamischen Verhaltens der Druckmaschine in der Versuchsanordnung nach einer zweiten Anpassung des Vorgabewerts für den Durchmesser des Formzylinders, wobei zusätzlich ein Repetitive-Control-Verfahren angewendet wurde;
• Fig. 8 ein Diagramm, das die iterative Funktionsweise des Schießverfahrens durch sukzessive Inter- und Extrapolationen veranschaulicht;
• Fig. 9 ein Diagramm eines beispielhaften Regelkreises für zwei aneinander abrollende Walzenelemente;
• Fig. 10 einen Querschnitt eines idealisieren, unverformten Walzenpaars und
• Fig. 11 einen Querschnitt des Walzenpaars der Fig. 10, bei dem es durch die Einwirkung einer Kontaktkraft zu einer elastischen Verformung kommt.

The subject invention is described below with reference to the Figures 1 to 8 explained in more detail, which show exemplary, schematic and non-limiting advantageous embodiments of the invention. It shows

• Fig. 1 a schematic representation of the printing process of a flexographic printing machine;
• Fig. 2 is a schematic representation of the expected drive torque during one revolution of the forme cylinder;
• Fig. 3 a diagram of the dynamic behavior of a printing press in a test arrangement;
• Fig. 4 an enlarged view of a portion of the drive torque of the Fig. 3 ;
• Fig. 5 a diagram of the dynamic behavior of the printing press in the test arrangement after a first adjustment of the default value for the diameter of the forme cylinder;
• Fig. 6 a diagram of the dynamic behavior of the printing press in the test arrangement when using a repetitive control method;
• Fig. 7 a diagram of the dynamic behavior of the printing press in the test arrangement after a second adjustment of the default value for the diameter of the forme cylinder, a repetitive control method additionally being used;
• Fig. 8 a diagram that illustrates the iterative mode of operation of the shooting method by successive inter- and extrapolations;
• Fig. 9 a diagram of an exemplary control loop for two rolling elements rolling against each other;
• Fig. 10 a cross section of an idealized, undeformed roller pair and
• Fig. 11 a cross section of the pair of rollers Fig. 10 , in which there is an elastic deformation due to the action of a contact force.

When two roller elements roll against each other, deformations occur, which are described below with reference to 10 and 11 are generally described.

Fig. 10 shows an idealized pair of rollers from a first roller element 1 'and a second roller element 2' which roll against each other at a contact point A (based on the cross section shown). The (undeformed) normal radii R _{1.0 of} the first roller element 1 'and R _{2.0 of} the second roller element 2' define the standard distance d _{0 of} the roller axes. The representation in Fig. 10 corresponds to the situation in which there is no contact force F between the roller elements (F = 0) and no elastic deformation of the roller elements occurs.

Fig. 11 shows schematically the deformation that occurs on the pair of rollers when the two roller elements 1 ', 2' are pressed together with a contact force F> 0 (the deformations are in Fig. 11 shown greatly exaggerated for reasons of recognizability). The two roller elements no longer touch each other in a line (ie in a cross section at a point), but in a contact surface (which is shown in the cross section in Fig. 11 is shown as a line). The radii of the roller elements are also no longer constant, the minimum radii R ₁ , R _{2 being} in the middle of the contact surfaces. The distance between the roller axes d in the deformed state is therefore smaller than the standard distance d ₀ . Therefore, the peripheral speed at the contact surface no longer corresponds to the value calculated on the basis of the idealized representation. Analogous considerations also apply if a roller element rolls on a flat counter surface under elastic deformation.

Such deformations of roll elements rolling against each other are not always precisely predictable in practice, and determining the exact extent of the deformation by means of measuring methods is very complex and often cannot be carried out in practice. However, since the deformation often has direct effects on the product quality, the method according to the invention aims to minimize the quality errors which result from these deformations. The invention is described below on the basis of an exemplary application in printing technology.

Fig. 1 shows the roller arrangement of a flexographic printing machine, consisting of an anilox roller 1, a forme cylinder 2 and an impression cylinder 3, at five different times t = t ₀ to t ₄ , which are each related to one revolution of the forme cylinder 2.

Direct printing processes, such as flexographic printing, have long been common in the art and are well known, and therefore not every single component of the printing press is discussed here. Also on the representation of some components in Fig. 1 omitted for the sake of clarity, since they are sufficiently known to the person skilled in the art.

The forme cylinder 2 carries a printing form 4 made of a flexible material, on which raised areas according to the known flexographic printing process define the areas to be printed. The anilox roller 1 applies the printing ink to the raised areas of the printing form 4. The printing ink is then applied to the printing material between the forme cylinder 2 and the impression cylinder 3.

t₀:

Die Druckform 4 gelangt beim Berührungspunkt A zwischen dem Formzylinder 2 und dem Gegendruckzylinder 3 in Kontakt mit dem Gegendruckzylinder 3 (Beginn eines Druckzyklus), während die Rasterwalze 1 weiterhin in Kontakt mit der Druckform 4 ist;

t₁:

Die Berührung zwischen Rasterwalze 1 und der auf dem Formzylinder 2 aufgebrachten Druckform 4 bei Berührungspunkt B endet, während der Gegendruckzylinder 3 weiterhin in Kontakt mit der Druckform 4 ist;

t₂:

Nach der Drucklücke 5 gelangt die Rasterwalze 1 wieder in Kontakt mit der Druckform 4, während der Gegendruckzylinder 3 weiterhin in Kontakt mit der Druckform 4 ist;

t₃:

Die Berührung zwischen Gegendruckzylinder 3 und Druckform 4 im Berührungspunkt A endet, während die Rasterwalze 1 weiterhin in Kontakt mit der Druckform 4 ist;

t₄:

Der Gegendruckzylinder 3 (bzw. der auf diesem mitgeführte Bedruckstoff) gelangt wieder in Kontakt mit der Druckform 4, während die Rasterwalze 1 weiterhin in Kontakt mit der Druckform 4 ist. Die Lage entspricht dem Zeitpunkt t₀, wobei der Druckzyklus endet und ein neuer Druckzyklus beginnt.

Since the length of the printing form 4 can be shorter than the circumference of the forme cylinder 2, there can generally be an area on the form cylinder 2 which is not covered by the printing form 4 and which is also referred to herein as the printing gap 5. With one revolution of the forme cylinder 2, ie during a printing cycle, the following times t = t ₀ to t _{4 are} run through, for example:

t ₀ :

The printing form 4 comes into contact with the impression cylinder 3 at the point of contact A between the form cylinder 2 and the impression cylinder 3 (beginning of a printing cycle), while the anilox roller 1 is still in contact with the printing form 4;

t ₁ :

The contact between the anilox roller 1 and the printing form 4 applied to the form cylinder 2 ends at contact point B, while the impression cylinder 3 is still in contact with the printing form 4;

t ₂ :

After the printing gap 5, the anilox roller 1 again comes into contact with the printing form 4, while the impression cylinder 3 is still in contact with the printing form 4;

t ₃ :

The contact between the impression cylinder 3 and the printing form 4 ends at the contact point A, while the anilox roller 1 is still in contact with the printing form 4;

t ₄ :

The impression cylinder 3 (or the printing material carried on it) comes into contact with the printing form 4 again, while the anilox roller 1 is still in contact with the printing form 4. The position corresponds to time t ₀ , the printing cycle ending and a new printing cycle beginning.

In the Fig. 1 The arrangement shown, which leads to the sequence of touch changes described, is purely exemplary and is not restrictive. As is clear to the person skilled in the art, the printing form 4 can be shorter or longer and the relative arrangement of the roller elements to one another can also differ. Such changes can also lead to a different sequence of touch changes. For example, with a shorter printing form 4 and a corresponding roller arrangement, a time period could occur in which neither the impression cylinder 3 nor the anilox roller 1 is in contact with the printing form 4. On the other hand, it is also possible for the printing form 4 to encompass the entire circumference of the forme cylinder 2, so that no changes in contact occur. The invention can also be used advantageously in such cases.

In the case of position or speed-controlled roller elements, the rotational speeds of the individual roller elements are coordinated with one another based on the respective diameter, so that there are no relative speeds between the roller elements at the points of contact in the theoretical modeling. In practice, however, it has been found that due to the elastic deformation of the roller elements, such relative speeds can occur at the point of contact. The drive torque is higher if both the anilox roller 1 and the impression cylinder 3 are simultaneously in engagement with the forme cylinder, and it is less if the point of contact of one or more roller elements is precisely in the area of the pressure gap 5.

Fig. 2 shows a schematic representation of the drive torque to be expected during one revolution of the forme cylinder, based on the times t ₀ to t ₄ , as shown in FIG Fig. 1 are shown. This theoretical scheme can be used to evaluate actual measurement results. It should be noted that different roller arrangements or different lengths of the printing form 4 would lead to different courses of the drive torque, whereby the person skilled in the art can readily apply the teachings of the present application to such cases.

After these theoretical considerations, the invention will now be based on a series of tests carried out by the applicant and with reference to the 3 to 7 exemplified and explained in a non-limiting manner.

Fig. 3 shows a diagram of the dynamic behavior of the test arrangement, the top curve showing the following error (difference between the target and actual position in relation to the surface of the forme cylinder), the middle curve showing the speed curve and the bottom curve the drive torque of the forme cylinder, with the times t ₀ to t ₄ for one revolution as shown in Fig. 1 and 2nd are plotted in the diagram.

The nominal position corresponds to the nominal value of the position controller, the actual position was measured with an encoder. A non-constant course of the following error indicates that the peripheral speeds of the roller elements that are in contact do not match.

The speed curve at points in time t ₀ , t ₁ , t ₂ and t ₄ (whenever a pair of rollers comes into or out of engagement) has clearly pronounced and broad value peaks.

The in Fig. 3 shown curve of the drive torque is in Fig. 4 shown again enlarged. This clearly shows that the drive torque increases approximately linearly in the intervals t ₁ to t ₂ and t ₃ to t ₄ . In the course of the tests, it was possible to show that these increases in drive torque can be attributed to different circumferential speeds between the pairs of rollers, the increase in the interval t ₁ to t _{2 being} due to a speed difference between the impression cylinder and the forme cylinder and the increase in the interval t ₃ to t ₄ a speed difference between anilox roller and forme cylinder can be attributed.

In the printed image, a pronounced horizontal stripe at t ₁ , where the printing form cylinder loses contact with the anilox cylinder, and a less pronounced but still clearly visible transverse stripe at t ₂ , where the printing form cylinder and anilox cylinder come into contact with one another, were discernible.

It could be shown that the horizontal stripes result in particular from distorted image points on the printing material, which are perceived as stripes by the human eye when the printed image is viewed macroscopically.

The contact force between the respective roller pairs (caused by the pressure provided) leads to an elastic deformation of the two roller elements (of the entire printing block structure). This elastic deformation in turn leads to changed effective diameters of the respective roller elements, which differ from the diameters set by the machine operator. These effects can cause a difference in the circumferential speed of the respective pairings, even if the rotational speeds are supposedly set correctly and a direct measurement of the exact effective diameter values is not possible while the printing press is running.

This difference in speed means that the contact of the roller elements results in an exchange of torque between the roller elements, which is characterized in that the faster rotating roller element drives the slower rotating roller element and vice versa (the slower rotating roller element brakes the faster rotating roller element).

This results in the contact phase of the two roller elements (when the printing form is in engagement) in the event that the roller pairs are operated in a position-controlled manner, in the torque which increases over time on the faster rotating roller element on the one hand, and in a torque which decreases over time slower rotating roller element on the other hand. This is in the course of the diagrams Fig. 3 and 4th recognizable.

The effect of the torques which build up and act against one another over the time of the intervention is caused by the following errors of the assigned drive control loops increasing over time.

At the end of the contact phase, the lag errors accumulated up to that point are reduced again and lead to a compensation behavior determined by the disturbance behavior of the drive control loops (aperiodic decay or decay with damped oscillation). Depending on the compensation behavior (determined by the disturbance dynamics of the closed drive control loop), stripes are caused in the printed image.

One consideration on which the invention is based is to prevent the occurrence of transverse stripes in the printed image by recognizing different peripheral speeds in roller pairs and automatically adapting the peripheral speeds of the roller pairs to one another. For this purpose, the torque profiles of the assigned roller drives are evaluated in the contact phase and the roller speeds are adjusted until the peripheral speeds of the roller pairs match and, on average, essentially constant torque profiles result in the contact phase.

The peripheral speeds of the roller elements can be adapted, for example, by changing the default value for the roller diameter. In a further test run, the default value for the diameter of the forme cylinder was therefore increased by 0.6% in a first step in order to achieve a correspondingly lower peripheral speed of the forme cylinder. Fig. 5 shows a diagram of the dynamic behavior of the printing press in the test arrangement described above after this adjustment of the default value for the diameter of the forme cylinder by + 0.6%. It can be clearly seen that the drive torque has significantly reduced value peaks (approx. 6 Nm in Fig. 5 compared to approx. 13 Nm in Fig. 4 ) and had a reduced average. Furthermore, the drive torque in the subintervals of the contact phases (intervals t ₁ to t ₂ and t ₃ to t ₄ ) showed an average constant course. There were no longer any horizontal stripes in the printed image.

From the above investigations it was concluded that an adaptation of the peripheral speeds depending on the course of the drive torque pressure errors, and especially prevent the formation of horizontal stripes. This adjustment can be automated, for example the increase in the drive torque in the "constant" areas (ie the areas in which there is no change at the contact points A and B, cf. Fig. 1 ) is evaluated, and the peripheral speed (s) of the roller (s) is (are) adjusted accordingly, for example by changing the default values for the roller diameter (s).

The adjustment of the circumferential speeds of the roller pairs can not only be achieved by changing the speed of one of the roller elements involved, but also changing the infeed between the roller pairs in order to influence the desired length of the print motif on the substrate and thereby distortions which can arise in the printed image.

In a further approach to improve the print image, an attempt was made to implement a dynamic adaptation of the drive control as a function of the course of the drive torque. For this purpose, the constancy of the roller speed was increased using a repetitive control process.

A high quality of the printed image can only be achieved by increasing the roll speeds with the help of the repetitive control process by means of an additive current feed. The repetivite control process can basically be carried out in a self-learning manner and is therefore very easy to use.

• Schleppfehler
• Drehzahlfehler
• Drehzahl
• Antriebsmoment
• Status der RC-Steuerung (0: nicht aktiv, 2 und 3: Initialisierungsphasen, 4: aktiv)
• Ausgangswert RC-Steuerung (additive Stromaufschaltung)

Fig. 6 shows a diagram of the dynamic behavior of the test arrangement when the RC method is applied. The default value for the roll diameter was compared to the initial value ( Fig. 3 and 4th ) not changed. In Fig. 6 the courses of the following values are given, from top to bottom:

• Following error
• Speed error
• rotational speed
• Drive torque
• Status of the RC control (0: not active, 2 and 3: initialization phases, 4: active)
• Output value RC control (additive current application)

After activating the RC control (status = 4) it can be seen that the actual speed has no relevant value peaks and can therefore be regarded as constant can. The following error was also massively reduced. Nevertheless, the course of the drive torque shows that this is always positive and that there was a steady increase in the contact phases. In spite of the significant improvements, stripes were noticeable in the printed image, albeit to a lesser extent than before the RC control was used.

In order to nevertheless take advantage of the recognizable advantages of the RC control, further tests were carried out, combining the adjustment of the default values for the roller speed and the RC control. The measurement result of this experiment is in Fig. 7 shown. Fig. 7 shows an almost constant speed curve with a very low tracking error and low drive torque. No horizontal stripes were visible in the printed image.

Fig. 8 illustrates the iterative mode of operation of a shooting process with which an optimal default value for the diameter of a roller element can be determined. Starting from a starting value D ₀ for the diameter, a corresponding value k ₀ for the increase in the torque is determined (this corresponds, for example, to that in FIG Fig. 4 shown slope between the times t ₁ and t ₂ ). Thereafter, the starting value D _{0 is changed} slightly to the value D ₁ and the corresponding value k _{1 of} the increase in the torque is determined. The next default value D2 for the diameter is then determined as the intersection of the line through the points (D ₀ , k ₀ ) and (D ₁ , k ₁ ) with the axis of abscissa. The method is continued iteratively until a default value D _{x is} found for which the slope of the torque k _{x is} sufficiently small. In Fig. 8 this is the case for the value D ₄ , which only has a very slight slope k ₄ .

The shooting process can be applied to different default values, whereby it can run automatically at the beginning of each printing process.

Even if the examples set out above for the method according to the invention are each described using a control of the forme cylinder, it is clear to the person skilled in the art that the other roller elements involved in the printing process, such as, for example, the anilox roller, the impression cylinder, or further intermediate roller elements, are analogous Way can be optimized to improve the printed image.

Fig. 9 shows an exemplary control circuit for drive control of two roller elements 1 'and 2' which roll against each other and are each driven by a drive motor M _A , M _B. The two drive motors are each controlled by a control circuit, the function of the controller being described below with reference to the first roller element 1 '.

The target value w corresponds to the default value for the engine speed, this target value w being based on the dimensions of the undeformed roller elements. This setpoint is corrected on the one hand by an adaptation value a, which was determined in accordance with the speed correction method. The adjustment value is determined by the speed adjustment D, which is described in more detail below. The feedback y _M (t) is subtracted from the desired value w corrected with the adaptation value a in order to determine the control deviation e (t), which represents the input value into a speed controller R _A. The speed controller R _A outputs a control variable u (t).

The control variable u (t) is stored by a repetitive controller RC _A , which has an internal memory, in the internal memory in at least a first revolution cycle, the repetitive controller RC _{A being} based on the stored values in a subsequent revolution cycle Correction signal k (t) outputs. The correction signal k (t) is combined with the control variable u (t) to form a corrected control variable u _k (t). A current controller S _A creates a manipulated variable u _s (t) based on the corrected control variable u _k (t), which controls the drive motor M _A , for example in the form of a drive current.

The repetitive controller (RC _A ) thus uses the speed error scaled with the speed controller gain as an input variable and tries to regulate the speed error to zero during one revolution. This case is shown in the block diagram. Alternatively, the speed setpoint can be specified by a higher-level position controller whose actual value is the integrated value of the feedback y _M (t). In this case the use of the following error scaled with the position controller gain can be useful. The RC then tries to keep the following error constant at zero during one revolution.

The controlled variable y (t) is the speed of the roller element 1 '. To create the feedback y _M (t) from this controlled variable y (t), the control loop offers the Fig. 9 three options:
The speed can either be measured via a motor encoder MG _A provided on the drive motor, or via a load encoder LG _A arranged on the roller element 1 '. Alternatively, the feedback y _M (t) can be created by a virtual load generator VG _A. On the basis of the manipulated variable u _s (t), the virtual load generator creates an estimated value for the controlled variable y (t), which is based on the current or the motor torque (ie the manipulated variable u _s (t)), the motor speed and a model of the dynamic behavior between the motor and load sensor.

The type of feedback can be selected using a SW _A selection switch.

The control loop of the second roller element 2 'has the same elements again and functions in an analogous manner to that described above for the first roller element 1'. The Elements of the control loop which are to be assigned to the second roller element 2 'are shown in Fig. 9 characterized by a subscript B, the elements assigned to the first roller element 1 ', on the other hand, with a subscript A. For the sake of clarity, the variables or signals of the control loop w, e (t), u (t), u _k (t), u _s (t), y (t), y _M (t), a in Fig. 9 specified only for the control loop of the first roller element 1 '. The control circuit of the second roller element uses analog signals.

The two control loops are linked by the above-mentioned speed adjustment D, which evaluates on the basis of the manipulated variable u _s (t) (both control loops) or the values obtained from the selected encoder signals in accordance with the previously described speed correction method and the adaptation value a for both roller elements 1 ' , 2 'created.

In the case of a sufficiently fast current control loop (this condition is usually fulfilled), the current setpoint (ie the corrected control variable u _k (t)) can be used instead of the actual current value (ie the manipulated variable us ( _t )) for the virtual load transmitter and the speed adjustment.

Claims (15)

1. A method for controlling a drive of a machine having at least one first roll element (1'), which rolls with a surface on a counter surface at least in subintervals of a revolution cycle with the action of a contact force under elastic deformation, characterized in that a speed correction method, which automatically compensates for deformation-related deviations of the peripheral velocity of the first roll element (1') by way of an adaptation of the setpoint value for the velocity of the first roll element (1'), and a retrospective method, which automatically compensates for deviations of the speed consistency of the first roll element (1') within a revolution cycle by applying a correction signal determined from the curve of a control variable, in particular an actual velocity or actual position of the first roll element (1'), in a preceding revolution cycle or in multiple preceding revolution cycles, are applied in combination.
2. The method as claimed in claim 1, characterized in that, for the speed correction method, the time curve of a value characteristic of a drive torque of the roll arrangement is determined in at least one of the subintervals, a parameter for an increase of this value in the subinterval is derived therefrom, and the guide variable of the peripheral velocity of the first roll element (1') is adapted as a function of this parameter to minimize the increase.
3. The method as claimed in claim 2, characterized in that the parameter is derived from the drive torque or from a variable physically proportional to the drive torque, for example, the drive current or the drive power.
4. The method as claimed in any one of claims 2 or 3, characterized in that the value characteristic of the drive torque is a force applied by the first roll element (1') onto the counter surface.
5. The method as claimed in any one of claims 2 to 4, characterized in that the derived parameter is a value averaged in the at least one subinterval.
6. The method as claimed in any one of claims 2 to 5, characterized in that the derived parameter is a smoothed slope of the drive torque in the at least one subinterval.
7. The method as claimed in any one of claims 1 to 6, characterized in that the peripheral velocity of the first roll element (1') is adapted by a change of the specified value for the speed of this roll element.
8. The method as claimed in any one of claims 1 to 6, characterized in that the peripheral velocity of the first roll element (1') is adapted by a change of the specified value for the diameter of this roll element.
9. The method as claimed in any one of claims 1 to 6, characterized in that the peripheral velocity of the first roll element (1') is adapted by changing the specified value for the infeed of this roll element.
10. The method as claimed in any one of claims 1 to 9, characterized in that the retrospective method is a self-learning method for controlling cyclic sequences, in particular a repetitive control method.
11. The method as claimed in any one of claims 1 to 10, characterized in that the retrospective method uses a speed error which is scaled with a speed controller amplification as the input signal.
12. The method as claimed in any one of claims 1 to 11, characterized in that the retrospective method passes through an initialization phase over at least one revolution cycle, preferably over at least two revolution cycles.
13. The method as claimed in any one of claims 1 to 12, characterized in that the counter surface is formed by the surface of a second roll element (2'), wherein the first roll element (1') and the second roll element (2') roll on one another, wherein the drive of the second roll element (2') is controlled similarly to the first roll element (1'), and wherein the retrospective method preferably uses a periodic drag error occurring in the contact point (A) between the first roll element (1') and the second roll element (2') as the control variable.
14. The method as claimed in any one of claims 1 to 13, characterized in that the machine is a printing press, wherein the first roll element (1') is a form cylinder (2), wherein a counter pressure cylinder (3) and an anilox roll (1) roll on the form cylinder (2), and wherein an elastic printing form (4) is applied to the form cylinder (2), which is in contact with the anilox roll (1) and/or the counter pressure cylinder (3) during at least one subinterval of the revolution of the form cylinder (2), and wherein the method is preferably automatically and/or regularly executed during the printing procedure.
15. The method as claimed in claim 14, characterized in that an identical linear velocity is set by the adaptation of the peripheral velocity on a machine having multiple printing mechanisms.

Images