Method for controlling the drive of a machine

Abstract

A method for controlling a drive of a machine having at least one first roll element, 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, that includes a speed correction method, which automatically compensates for deformation-related deviations of the peripheral velocity of the first roll element by way of an adaptation of the setpoint value for the velocity of the first roll element, and a retrospective method, which automatically compensates for deviations of the speed consistency of the first roll element 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, in a preceding revolution cycle or in multiple preceding revolution cycles, are applied in combination.

Claims

1. A method for controlling a drive of a machine having at least one first roll element, 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, comprising: a speed correction method, which automatically compensates for deformation related deviations of a peripheral velocity of the first roll element by way of an adaptation of a setpoint value for the velocity of the first roll element, and a retrospective method, which automatically compensates for deviations of a speed consistency of the first roll element within a revolution cycle by applying a correction signal determined from a curve of a control variable in a preceding revolution cycle or in multiple preceding revolution cycles, are applied in combination.

2. The method as claimed in claim 1, wherein, for the speed correction method, a time curve of a value characteristic of a drive torque of a roll arrangement is determined in at least one of the subintervals, a parameter for an increase of the value characteristic of the drive torque of the roll arrangement in the at least one of the subintervals is derived therefrom, and a guide variable of the peripheral velocity of the first roll element is adapted as a function of the derived parameter for the increase of the value characteristic of the drive torque of the roll arrangement in the at least one of the subintervals to minimize the increase.

3. The method as claimed in claim 2, wherein the parameter of the value characteristic of the drive torque is derived from the drive torque of the roll arrangement or from a variable physically proportional to the drive torque.

4. The method as claimed in claim 3, wherein the variable is physically proportional to the drive current or the drive power.

5. The method as claimed in claim 2, wherein the value characteristic of the drive torque of the roll arrangement is a force applied by the first roll element onto the counter surface.

6. The method as claimed in claim 2, wherein the derived parameter for the increase of the value characteristic of the drive torque of the roll arrangement in the at least one of the subintervals is a value averaged in the at least one subinterval.

7. The method as claimed in claim 2, wherein the derived parameter for the increase of the value characteristic of the drive torque of the roll arrangement in the at least one of the subintervals is a possibly smoothed slope of the drive torque of the roll arrangement in the at least one subinterval.

8. The method as claimed in claim 1, wherein the peripheral velocity of the first roll element is adapted by a change of a specified value for the velocity of the first roll element.

9. The method as claimed in claim 1, wherein the peripheral velocity of the first roll element is adapted by a change of a specified value for a diameter of the first roll element.

10. The method as claimed in claim 1, wherein the peripheral velocity of the first roll element is adapted by changing a specified value for a infeed of the first roll element.

11. The method as claimed in claim 1, wherein the retrospective method is a self-learning method for controlling cyclic sequences.

12. The method as claimed in claim 11, wherein the self-learning method for controlling cyclic sequences is a repetitive control method.

13. The method as claimed in claim 1, wherein the retrospective method uses a speed error which is scaled with a speed controller amplification as an input signal.

14. The method as claimed in claim 1, wherein the retrospective method passes through an initialization phase over at least one revolution cycle.

15. The method as claimed in claim 14, wherein the at least one revolution cycle is greater than at least two revolution cycles.

16. The method as claimed in claim 1, wherein the counter surface is formed by a surface of a second roll element, wherein the first roll element and the second roll element roll on one another, and wherein the second roll element includes a second drive that is controlled in a manner similar to that of the first roll element.

17. The method as claimed in claim 16, wherein the retrospective method uses a periodic drag error occurring in a contact point between the first roll element and the second roll element as the control variable.

18. The method as claimed in claim 1, wherein the machine is a printing press, wherein the first roll element is a form cylinder, wherein a counter pressure cylinder and an anilox roll on the form cylinder, and wherein an elastic printing form is applied to the form cylinder, which is in contact with at least one of the anilox roll or the counter pressure cylinder during at least one subinterval of a revolution of the form cylinder.

19. The method as claimed in claim 18, wherein the method is automatically executed during a printing procedure.

20. The method as claimed in claim 18, wherein the method is regularly executed during a printing procedure.

21. The method as claimed in claim 18, wherein an identical linear velocity is set by the adaptation of the peripheral velocity (velocities) to a machine having multiple printing mechanisms.

22. The method as claimed in claim 1, wherein the correction signal is determined from the curve of an actual velocity or actual position of the first roll element.

Description

(1) The present invention will be explained in greater detail hereafter with reference to FIGS. 1 to 8, which show exemplary, schematic, and nonrestrictive advantageous embodiments of the invention. In the figures:

(2) FIG. 1 shows a schematic illustration of the printing procedure of a flexographic printing press;

(3) FIG. 2 shows a schematic illustration of the drive torque to be expected during a revolution of the form cylinder;

(4) FIG. 3 shows a diagram of the dynamic behavior of a printing press in an experimental arrangement;

(5) FIG. 4 shows an enlarged illustration of a subregion of the drive torque of FIG. 3;

(6) FIG. 5 shows a diagram of the dynamic behavior of the printing press in the experimental arrangement after a first adaptation of the specified value for the diameter of the form cylinder;

(7) FIG. 6 shows a diagram of the dynamic behavior of the printing press in the experimental arrangement upon a use of a repetitive control method;

(8) FIG. 7 shows a diagram of the dynamic behavior of the printing press in the experimental arrangement after a second adaptation of the specified value for the diameter of the form cylinder, wherein a repetitive control method was additionally applied;

(9) FIG. 8 shows a diagram which illustrates the iterative functionality of the shooting method by successive interpolations and extrapolations;

(10) FIG. 9 shows a diagram of an exemplary inventive control loop for two roll elements rolling on one another;

(11) FIG. 10 shows a cross section of an idealized, non-deformed roll pair; and

(12) FIG. 11 shows a cross section of the roll pair from FIG. 10, in which an elastic deformation occurs due to the action of a contact force.

(13) As two roll elements roll on one another, deformations occur, which are described in general hereafter with reference to FIGS. 10 and 11.

(14) FIG. 10 shows an idealized roll pair composed of a first roll element 1′ and a second roll element 2′, which roll on one another at a contact point A (in relation to the illustrated cross section). The (nondeformed) normal radii R.sub.1.0 of the first roll element 1′ and R.sub.2.0 of the second roll element 2′ define the normal distance d.sub.0 of the roll axles. The illustration in FIG. 10 corresponds to the situation in which no contact force F acts between the roll elements (F=0) and no elastic deformation of the roll elements occurs.

(15) FIG. 11 schematically shows the deformation which occurs on the roll pair when the two roll elements 1′, 2′ are pressed against one another with a contact force F>0 (the deformations are shown greatly exaggerated in FIG. 11 for reasons of recognizability). The two roll elements no longer are in contact in a line (i.e., in a point in a cross section), but rather in a contact surface (which is shown as a line in the cross-sectional illustration in FIG. 11). The radii of the roll elements are also no longer constant, wherein the minimal radii R.sub.1, R.sub.2 are located in the middle of the contact surfaces. The distance of the roll axles d in the deformed state is therefore less than the normal distance d.sub.0. The peripheral velocity at the contact surface therefore also no longer corresponds to the value computed on the basis of the idealized representation. Similar considerations also apply if one roll element rolls on a flat counter surface with elastic deformation.

(16) Such deformations of roll elements rolling on one another are not always accurately predictable in practice and determining the accurate 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 immediate effects on the product quality, the inventive method is directed to minimizing the quality flaws which arise from these deformations. The invention will be described hereafter on the basis of an exemplary application in printing technology,

(17) FIG. 1 shows the roll arrangement of a flexographic printing press, consisting of an anilox roll 1, a form cylinder 2, and a counter pressure cylinder 3, at five different points in time t=t.sub.0 to t.sub.4, which are each related to one revolution of the form cylinder 2.

(18) Direct printing methods, for example, flexographic printing, have been generally routine and known for some time in the prior art, and therefore each individual component of the printing press will not be discussed herein. The illustration of some components was also omitted in FIG. 1 for the sake of comprehensibility, since they are well known to a person skilled in the art.

(19) The form cylinder 2 bears a printing form 4 made of a flexible material, on which raised points define the regions to be printed according to the known flexographic printing method. The anilox roll 1 applies the printing ink to the raised points of the printing form 4. The printing ink is then applied to the printed material between the form cylinder 2 and the counter pressure cylinder 3.

(20) Since the length of the printing form 4 can be shorter than the circumference of the form cylinder 2, there can in general be a region not covered by the printing form 4 on the form cylinder 2, which is also referred to herein as a printing gap 5. During a revolution of the form cylinder 2, i.e., during a printing cycle, it therefore, for example, passes through the following points in time t=t.sub.0 to t.sub.4: t.sub.0: The printing form 4 comes into contact with the counter pressure cylinder 3 at the contact point A between the form cylinder 2 and the counter pressure cylinder 3 (beginning of a printing cycle), while the anilox roll 1 is still in contact with the printing form 4; t.sub.1: The contact between anilox roll 1 and the printing form 4 applied to the form cylinder 2 ends at contact point B, while the counter pressure cylinder 3 is still in contact with the printing form 4; t.sub.2: After the printing gap 5, the anilox roll 1 moves back into contact with the printing form 4, while the counter pressure cylinder 3 is still in contact with the printing form 4; t.sub.3: The contact between counter pressure cylinder 3 and printing form 4 ends at contact point A, while the anilox roll 1 is still in contact with the printing form 4;

(21) t.sub.4: The counter pressure cylinder 3 (and/or the printed material carried thereon) moves back into contact with the printing form 4, while the anilox roll 1 is still in contact with the printing form 4. The location corresponds to the point in time t.sub.0, wherein the printing cycle ends and a new printing cycle begins.

(22) The arrangement illustrated in FIG. 1, which results in the described sequence of contact changes, is solely by way of example and is not restrictive. As is clear to a person skilled in the art, the printing form 4 can be shorter or longer and the relative arrangement of the roll elements in relation to one another can also differ. Such changes can also result in a different sequence of contact changes. For example, in the event of a shorter printing form 4 and a corresponding roll arrangement, a time window could occur in which neither the counter pressure cylinder 3, nor the anilox roll 1 is in contact with the printing form 4. On the other hand, it is also possible that the printing form 4 encloses the entire circumference of the form cylinder 2, so that no contact changes occur. The invention may also be advantageously applied to such cases.

(23) In position-controlled or speed-controlled roll elements, the rotational velocities of the individual roll elements are adapted to one another on the basis of the respective diameter, so that no relative velocities exist between the roll elements in the contact points in the theoretical model. In practice, however, it has been shown that such relative velocities can occur in the contact point because of the elastic deformation of the roll elements. The drive torque is higher if both the anilox roll 1 and also the counter pressure cylinder 3 are simultaneously engaged with the form cylinder, and it is lower if no contact point of one or more roll elements is presently located in the region of the printing gap 5.

(24) FIG. 2 shows a schematic illustration of the drive torque to be expected during a revolution of the form cylinder, in relation to the points in time t.sub.0 to t.sub.4, as shown in FIG. 1. This theoretical scheme can be used to analyze actual measurement results. It is to be noted that different roll arrangements and/or different lengths of the printing form 4 would result in different curves of the drive torque, wherein a person skilled in the art can readily transfer the teachings of the present application to such cases.

(25) After these theoretical considerations, the invention will be explained on the basis of an experimental series carried out by the applicant and with reference to FIGS. 3 to 7 by way of example and in a nonrestrictive manner.

(26) FIG. 3 shows a diagram of the dynamic behavior of the experimental arrangement, wherein the uppermost curve shows the drag error (difference between setpoint position and actual position in relation to the surface of the form cylinder), the middle curve shows the velocity curve, and the lowermost curve shows the drive torque of the form cylinder, wherein the points in time t.sub.0 to t.sub.4 during a revolution according to the illustrations in FIGS. 1 and 2 are shown in the diagram,

(27) The setpoint position corresponds in this case to the setpoint value of the position controller, the actual position was measured using an encoder. A non-constant curve of the drag error indicates that the peripheral velocities of the roll elements which are in contact do not match with one another.

(28) The velocity curve has clearly pronounced and broad value peaks at the points in time t.sub.0, t.sub.1, t.sub.2, and t.sub.4 (whenever a roll pair engages or disengages).

(29) The curve of the drive torque shown in FIG. 3 is shown once again enlarged in FIG. 4. It can be clearly recognized therein that the drive torque increases approximately linearly in each of the intervals t.sub.1 to t.sub.2 and t.sub.3 to t.sub.4. It was possible to show in the scope of the experiments that these increases of the drive torque are to be attributed to different peripheral velocities between the roll pairs, wherein the increase in the interval t.sub.1 to t.sub.2 is to be attributed to a velocity difference between counter pressure cylinder and form cylinder and the increase in the interval t.sub.3 to t.sub.4 is to be attributed to a velocity difference between anilox roll and form cylinder.

(30) A strongly pronounced horizontal stripe was recognizable in the printed image at t.sub.1, where the printing form cylinder loses the contact to the anilox cylinder, and a less pronounced, but still clearly visible horizontal stripe is recognizable at t.sub.2, where printing form cylinder and anilox cylinder come back into contact with one another.

(31) It was possible to show that the horizontal stripes arise in particular because of distorted pixels on the printed material, which are perceived as stripes in the event of a macroscopic observation of the printed image by the human eye.

(32) An elastic deformation of the two roll elements (of the entire printing mechanism construction) occurs due to the contact force between the respective roll pairs (caused by the pressure provision). This elastic deformation in turn results in changed effective diameters of the respective roll elements which deviate from the diameters set by the machine operator. These effects can cause a peripheral velocity difference of the respective pairs, even if the rotational velocities are supposedly set correctly, and a direct measurement of the exact effective diameter values is not possible with running printing press.

(33) This velocity difference has the result that a torque exchange between the roll elements takes place due to the contact of the roll elements, which is characterized in that the roll element rotating faster drives the roll element rotating slower and vice versa (the roll element rotating slower brakes the roll element rotating faster).

(34) This results, in the contact phase of the two roll elements (when the printing form is engaged), in the case that the roll pairs are operated in a position-controlled manner, in the torque increasing over time on the roll element rotating faster, on the one hand, and in a torque decreasing over time on the roll element rotating slower, on the other hand. This is recognizable in the curve of the diagrams of FIGS. 3 and 4.

(35) The effect of the torques building up and acting against one another over the time of the engagement is caused by the drag error of the associated drive control loops increasing over time.

(36) At the end of the contact phase, the drag errors accumulated up to this point dissipate again and result in a compensation behavior determined by the interference behavior of the drive control loops (aperiodic decay or decay with damped oscillation). Depending on the compensation of a viewer (determined by the interference dynamic response of the closed drive control loop), stripes are caused in the printed image.

(37) One consideration on which the invention is based is to prevent the occurrence of horizontal stripes in the printed image in that different peripheral velocities are recognized in roll pairs and the peripheral velocities of the roll pairs are automatically adapted to one another. For this purpose, the torque curves of the associated roll drives are analyzed in the context phase and the roll speeds are adapted until the peripheral velocities of the roll pairs correspond and essentially constant torque curves result in the middle of the contact phase.

(38) The peripheral velocities of the roll elements can be adapted, for example, by a change of the specified value for the roll diameter, in a further test run, the specified value for the diameter of the form cylinder was therefore increased in a first step by 0.6%, to achieve a correspondingly lower peripheral velocity of the form cylinder. FIG. 5 shows a diagram of the dynamic behavior of the printing press in the above-described experimental arrangement after this adaptation of the specified value for the diameter of the form cylinder by +0.6%. It can be seen clearly that the drive torque has significantly reduced value peaks (approximately 6 Nm in FIG. 5 in relation to approximately 13 Nm in FIG. 4) and a reduced average value. Furthermore, the drive torque has a consistent curve on average in the subintervals of the contact phases (intervals t.sub.1 to t.sub.2 and t.sub.3 to t.sub.4). Horizontal stripes could no longer be seen in the printed image.

(39) The conclusion was drawn from the above considerations that an adaptation of the peripheral velocities as a function of the curve of the drive torque can prevent printing errors, and in particular the formation of horizontal stripes. This adaptation can take place automatically, wherein, for example, the slope of the drive torque in the “constant” regions (i.e., the regions in which no change occurs at the contact points A and B, cf. FIG. 1) is analyzed, and the peripheral velocity (velocities) of the roll(s) is (are) adapted accordingly, for example, by way of a change of specified values for the roll diameter or diameters.

(40) The equalization of the peripheral velocities of the roll pairs can be achieved not only by way of a change of the speed of one of the participating roll elements, but rather a change of the infeed between the roll pairs can also be performed to influence the desired length of the printed motif on the substrate and thus compensate for distortions which can result in the printed image.

(41) In a further approach for improving the printed image, an attempt was made to implement a dynamic adaptation of the drive control as a function of the curve of the drive torque. For this purpose, the consistency of the roll velocity was enhanced with the aid of a repetitive control method.

(42) A high quality of the printed image is first inventively achieved by an enhanced consistency of the roll velocities with the aid of the repetitive control method by an additive current activation. The repetitive control method can be executed in a self-learning manner in principle and is therefore very easily applicable.

(43) FIG. 6 shows a diagram of the dynamic behavior of the experimental arrangement if the RC method is applied. The specified value for the roll diameter was not changed in relation to the starting value (FIGS. 3 and 4). The curves of the following values are indicated in FIG. 6, from top to bottom: drag error speed error speed drive torque status of the RC controller (0: not active, 2 and 3: initialization phases, 4: active) starting value of RC controller (additive current activation)

(44) After the activation of the RC controller (status=4) it can be seen that the actual speed does not have relevant value peaks and can therefore be considered to be constant. The drag error was also massively reduced. Nonetheless, the curve of the drive torque shows that it is always positive and a continuous increase occurred in the contact phases. In spite of the substantial improvements, stripes were noticeable in the printed image, although to a lesser extent than before the use of the RC controller.

(45) To nonetheless be able to use the recognizable advantages of the RC controller, further experiments were carried out, wherein the adaptation of the specified values for the roll velocity and the RC, controller were combined. The measurement result of this experiment is shown in FIG. 7. FIG. 7 shows an approximately constant velocity curve having a very small drag error and low drive torque. No horizontal stripes were recognizable in the printed image.

(46) FIG. 8 illustrates the iterative functionality of a shooting method, using which an optimum specified value for the diameter of a roll element can be ascertained. Proceeding from a starting value D.sub.0 for the diameter, a corresponding value k.sub.0 is determined for the slope of the torque (this corresponds, for example, to the slope shown in FIG. 4 between the points in time t.sub.1 and t.sub.2). Thereafter, the starting value D.sub.0 is changed slightly to the value D.sub.1 and the corresponding value k.sub.1 of the slope of the torque is determined. The next specified value D2 for the diameter is then determined as the intersection of the line through the points (D.sub.0, k.sub.0) and (D.sub.1, k.sub.1) with the abscissa axis. The method is continued iteratively until a specified value D.sub.x is found, for which the slope of the torque k.sub.x is sufficiently small. In FIG. 8, this is the case at the value D.sub.4, which only still has a very small slope k.sub.4.

(47) The shooting method can be applied to various specified values, wherein it can run automatically at the beginning of each printing procedure.

(48) Although the above-described examples of the inventive method have each been described on the basis of a control of the form cylinder, it is clear to a person skilled in the art that the other roll elements participating in the printing procedure, such as the anilox roll, the counter pressure cylinder, or further interconnected roll elements, can be optimized in a similar manner to improve the printed image.

(49) FIG. 9 shows an exemplary control loop for the drive control of two roll elements 1′ and 2′ rolling on one another, which are each driven by one drive motor M.sub.A, M.sub.B. The two drive motors are each controlled via one control loop, wherein the function of the controller will be described hereafter with reference to the first roll element 1′.

(50) The setpoint value w corresponds to the specified value for the motor speed, wherein this setpoint value w is based on the dimensions of the nondeformed roll elements. This setpoint value is corrected, on the one hand, by an adaptation value a, which was determined according to the speed correction method. The determination of the adaptation value is performed by the speed equalization D, which is described in greater detail hereafter. The return y.sub.M(t) is subtracted from the setpoint value w corrected by the adaptation value a, in order to determine the control deviation e(t), which represents the input value in a speed controller R.sub.A. The speed controller R.sub.A outputs a control variable u(t).

(51) The control variable u(t) is stored in the internal memory by a repetitive controller RC.sub.A, which has an internal memory, in at least a first revolution cycle, wherein the repetitive controller RC.sub.A outputs a correction signal k(t) in a subsequent revolution cycle on the basis of the stored values.

(52) The correction signal k(t) is linked to the control variable u(t) to form a corrected control variable u.sub.k(t). A current controller S.sub.A produces a positioning variable u.sub.s(t), which activates the drive motor M.sub.A in the form of a drive current, for example, on the basis of the corrected control variable u.sub.k(t).

(53) The repetitive controller (RC.sub.A) thus uses the speed error scaled by the speed controller amplification as an input variable and attempts to control the speed error during a revolution down to zero. This case is shown in the block diagram. Alternatively, the speed setpoint value can be specified by a higher-order position controller, the actual value of which is the integrated value of the return y.sub.M(t). In this case, the use of the drag error scaled with the position controller amplification can be useful. The RC then attempts to control the drag error curve constantly to zero during a revolution.

(54) The control variable y(t) is the speed of the roll element 1′. To produce the return y.sub.M(t) from this control variable y(t), the control loop of FIG. 9 offers three options:

(55) The speed can be measured either via a motor encoder MG.sub.A provided on the drive motor, or via a load encoder LG.sub.A arranged on the roll element 1′. Alternatively thereto, the return y.sub.M(t) can be produced by a virtual load encoder VG.sub.A. The virtual load encoder produces an estimated value for the control variable y(t) on the basis of the positioning variable u.sub.s(t), which is based on the current or the motor torque (i.e., the positioning variable u.sub.s(t)), the motor-side speed, and a model of the dynamic behavior between motor encoder and load encoder.

(56) The type of the return can be selected via a selection switch SW.sub.A.

(57) The control loop of the second roll element 2′ has the same elements once again and functions in a similar manner as described above for the first roll element 1′. The elements of the control loop which are to be associated with the second roll element 2′ are identified in FIG. 9 by a subscript B, in contrast, the elements to be associated with the first roll element 1′ are identified by a subscript A. For the sake of comprehensibility, the variables or signals of the control loop w, e(t), u(t), u.sub.k(t), u.sub.s(t), y(t), y.sub.M(t), a are only indicated in FIG. 9 for the control loop of the first roll element 1′ The control loop of the second roll element uses similar signals.

(58) The two control loops are linked by the above-mentioned speed equalization D, which analyzes the values obtained on the basis of the positioning variable u.sub.s(t) (of both control loops) and/or the values obtained from the selected encoder signals according to the above-described speed correction method and produces the adaptation value a for both roll elements 1′, 2′.

(59) With a sufficiently fast current control loop (this condition is usually met), instead of the current actual value (i.e., the positioning variable 140), the current setpoint value (i.e., the corrected control variable u.sub.k(t)) can be used for the virtual load encoder and the speed equalization.