EP0522301A1 - Method for monitoring and controlling the printing process, especially in offset printing machines - Google Patents

Abstract

A method for monitoring and controlling the printing process, especially in offset printing machines is proposed, in which increases in tonal value, for example, are not to be evened out solely by varying the inking. The method according to the invention is divided into a learning phase and an ability phase. The quality characteristics which can be recorded by measurement on the printing product form a characteristics space in which association vectors from process-related evaluations of trial prints are assigned to certain categories and/or causes of interferences (learning phase). In the ability phase, that is to say the control and monitoring of the printing process, on the basis of the association with the category and/or the cause of the interference determined, said category is displayed and/or interventions are made in the process via the cause determined in this manner with the aim of eliminating said cause. <IMAGE>

Description

The invention relates to a method for monitoring and regulating the printing process, in particular on offset printing machines, according to the preamble of claim 1.

The production of printed products on an offset printing machine, for example sheet-fed offset printing press, is a complex process because of the large number of influencing variables which are dependent on one another. With offset printing, the quality of the printed product, in particular the sizes of ink guide, dampening solution guide, type of printing material, type of ink and rubber blanket etc. have an effect. Quality is understood here to mean compliance with certain minimum requirements for the printed product. These are, for example, color rendering, sharpness and contrast etc. Because of the complexity of the offset printing process in particular, complete control is theoretically only feasible with great effort in terms of process analysis. However, a concrete form of such a multi-size regulation has so far not become known.

Instead of a multivariable control, so-called partial control loops are formed today, such as a control loop for ink control, one for dampening solution control, one for register control, etc. In particular, one control loop is declared the so-called main control loop, the other sub-loops then only serve the other control variables to hold setpoints found all at once (DE 3 717 904 A1). The mutual dependency of the influencing variables is thus eliminated, so any quality disturbance that occurs on the printed product is always over tries to correct the main control loop.

In practice, it has long been known to make color control the main control loop, i.e. to detect the ink application on the printed product (inking) by sensor and to influence the ink flow by means of a controller and corresponding actuators in order to achieve a desired inking. The type of sensory detection of the printing ink application is generally photoelectric, i.e. either via the optical color density (densitometric) or via the detection of a color location (colorimetric) or using video technology and image processing. Devices and methods for determining the coloration and regulating the color flow are described in the publications DE 2 728 738 A1, EP 228 347 B1, DE 3 440 706 A1, DD 274 786 A1 and DD 274 787 A1. The setpoints, according to which the control of the ink flow takes place, can be taken, for example, by measurement from a template, found out by trying it out, or can also be taken from instructions for standardizing the printing process (BVD-Fogra). An example of a sub-control circuit for dampening solution guidance is shown in DE 3 636 507 A1.

A disadvantage of the main regulation of the printing process via the color guide described above, however, is that an attempt is made to correct any disturbances in the quality of the printed product via the color guide. For example, the increase in tonal value in offset printing is a parameter that decisively determines the reproduction quality of the print. If the optical area coverage achieved in printing is significantly larger or significantly smaller in several tonal value levels, this disturbs the contrast range in the case of single-color prints and the color reproduction in the case of multicolor prints. Certain strategies for regulating the color guidance with a focus on area coverage have become known (DE 3 543 444 A1, DE 3 440 706 A1), but here too any change in the area coverage achieved in printing is attempted to be achieved by changing the color guidance. For example, it is well known that an increased ink supply optically effective area coverage is also able to increase (increasing halftone dots) such as a reduction in the dampening solution supply or an increase in the pressure provided between the plate and blanket cylinders. It would certainly be nonsensical to reduce the area coverage by reducing the ink supply if, for example, the color density of a full-tone measuring field is almost in the range of a predetermined target value. In this case, the dampening solution guide or the cylinder pressure described above would rather be changed. The type and quality of the printing material, the type of printing blanket, the properties of the printing ink, the type and quantities of so-called printing aids, the machine temperature and ultimately the subject also play a decisive role. If the influence of the dot gain in a printed product can be achieved particularly well by changing the cylinder pressure, for example, the dot gain in another printed product (different subject, paper, etc.) can best be brought about by changing the dampening solution. So far, the decision as to what is the best corrective action has been left to the experienced printer. Based on his knowledge and the resulting ability to judge the present printing process, he selected the influencing variable by changing which he was able to limit the described dot gain most effectively.

The object of the present invention is therefore to provide a method for monitoring and regulating the printing process, according to the preamble of claim 1, by means of which a deviation in the quality of a printed product can be corrected via the most suitable influencing variable, in particular the actual cause, in order to achieve the desired quality.

This object is achieved by the characterizing features of claim 1. Further developments of the invention result from the subclaims and the description and the drawings.

The fuzzy method known per se is used Classification, ie from the characteristics that reflect the quality status of an existing printed product, is determined by using an unsharp classifier which quality deviations are present. On the other hand, it is determined which influencing variable can be used to correct this quality deviation. Condition monitoring using the unsharp classification is known, for example, from DD 251 534 A1. Compared to the regulation of malfunctions in the printing process exclusively via the ink guide, the method according to the invention has the advantage that the controller, in conjunction with the fuzzy classification, tries to remedy the malfunction via the most probable cause. In a way, this is comparable to human behavior, since even the experienced printer corrects a disturbance in the printing process about the cause that seems most likely to him.

Furthermore, an embodiment of the invention is based on FIGS. 1 and 2, each of which shows part of a feature space.

A printed sheet is printed on a sheetfed offset printing machine and then subjected to a quality diagnosis using measurement technology. A pressure control strip (its measuring fields) printed on the printed sheet is measured densitometrically (test sheet). The print control strip contains measuring fields for the solid color density DV for each printing ink, as well as halftone fields for the halftone color density (80%) DR 80 and halftone fields for the halftone color density (40%) DR 40, as well as quality diagnostic fields such as gray balance or possibly so-called sliding and duplicating fields . An example of such a print control strip is the 6-color CCI print control strip from MAN Roland Druckmaschinen AG. A densitometric scanning of this print control strip thus results in the characteristics spot color density DV, screen tone color density (40%) DR 40, screen tone color density (80%) DR 80 for each color. A feature space can thus be formed, the components of which are determined by the spot color density DV, the screen tone color density (40 %) DR 40, and the screen color density (80%) DR 80 are formed. A densitometric measurement of the print control strip of the The present printed product thus yields a point for a color in this feature space in accordance with the measured values. If there are several printed colors, there are therefore several dots corresponding to the number of colors.

On the basis of experience, an expert defines classes (quality classes) that are of interest in the printing process, for example a class K1 "going deep", a class K2 "tonal value increase compared to standard too high", a class K3 "tonal value increase compared to standard too low",
a class K4 "characteristic curve in the standard". According to the number of classes declared (here K1 to K4), a so-called membership N-tuple is assigned to each point in the feature space. With 4 classes according to the example above, 4 points would have to be assigned to each point. These numerical values can be standardized so that they are between 0 and 1. The membership N-tuple (for N classes) can also be referred to as a so-called membership vector, which should also be the case. The numerical size of a component of the membership vector now indicates how "strongly" a point in the feature space belongs to a certain class. With 4 classes, the membership vector thus has four components, the first component shows the degree of belonging of the one point to class K1, the second component the degree of membership of the point to class K2, etc. The components of the membership vectors of the individual points In the feature space, a large number of pressure tests are determined in a so-called learning phase through process engineering evaluation. In the above-described standardization of the components of the membership vector as numerical values between 0 and 1, for example a numerical value in a component close to 0 stands for very low membership in the corresponding class and a numerical value close to 1 accordingly corresponds to a high membership for the respective class . If the points of the feature space are occupied in this way with membership vectors, it can now be determined after a densitometric scan of the print control strip to which class the present print product belongs (optional phase).

1 shows the level which is formed by the components solid color density DV, halftone color density (40%) DR 40, furthermore the classes K1 to K4 are entered in this level according to the definition given above. The lines of constant membership in a class are shown (membership = 0.5). Within the lines of the classes shown, the affiliation is therefore greater than 0.5. What is striking and procedural is that these areas of the classes formed in this way can overlap each other. Also, the lines of belonging = 0.5 do not have to enclose closed areas.

According to the invention it is now further provided that the points of the feature space are further assigned membership vectors which indicate the membership of each point to a cause which has a negative influence on the print quality. FIG. 2 shows a plane of the feature space, which is formed by the coordinates solid color density DV, screen color density (40%) DR 40. The lines of affiliation = 0.5 are again drawn in for the causes U1 to U3, where, for example, U1 means "wrong ink supply", U2 "wrong print delivery", U3 "wrong fountain solution control". Here, too, the assignment of the feature space to the membership vectors - in this example, the membership vector has 3 components, since three causes U1 to U3 are specified - is carried out by process-related evaluations of pressure tests. The components of the membership vector can also be normalized between 0 and 1 here.

If a densitometric measurement of the printed product is carried out, the location of the feature point of this printed product in the feature space is then known. Due to the fact that the points in the feature space were assigned with membership vectors (assignment) in a learning phase, the position of the point in the feature space determined by measurement means that both the print product belongs to a certain class and the print product belongs to a specific cause, which is a quality problem has caused can be determined. This results from the maximum, ie the maximum value of a component of the respective membership vector. The following cases can occur, for example, if they belong to a class K1 to K4:

The membership in the class, for example K3, is clear, since the membership vector in this feature point has a high value in the third component and a very small value in the other components,
the membership vector supplies a broad maximum, ie it shows membership values of approximately the same size in two or more components, which means that the printed sheet or its quality cannot be clearly assigned to a class or cause. In the first-mentioned case, ie the assignment to the class or cause is unambiguous, the appropriate measure to remedy this defect can be initiated directly. If, for example, the characteristic of the printed product clearly assigns the cause of "fountain solution wrong", for example a measurement of the dampening solution supply can take place, whereupon the fountain solution guidance is set to a new value, by means of which the quality defect (for example, penetration in the depths) is eliminated. In the second case, in which the affiliation to a class or cause is not clear, several options are available. Sometimes it would be possible to display the equal opportunities and let the printer decide, it would also be possible to make a random choice about a cause and then correct the defect accordingly. Furthermore, it is also conceivable to additionally subclassify or subdivide into sub-causes when assigning the points of the feature space with membership vectors for the individual classes or causes. It would thus be possible to subdivide the cause U1 "wrong color management" into several sub-causes, each sub-cause being associated with the changes required to correct the wrong color management (eg necessary changes in color density). Accordingly, these measures were still in the learning phase respectively.

In summary, the method according to the invention can thus be represented in its application as follows:
Learning phase - the points of a feature space are assigned membership vectors, which indicate the membership to certain classes of print quality or causes of printing disturbances, whereby printing tests evaluated in terms of process engineering are evaluated and
Optional phase - the metrological results, for example a densitometric measurement of the measuring field, result in a point in the feature space for the printed product, so that from the corresponding affiliation vectors (maximum component) an assignment of the printed product (actual quality) to a defined class or cause of a printing disturbance (quality defect) ) he follows. The pressure disturbance is then remedied via the cause determined in this way.

So far, the method according to the invention has been described with reference to a feature space, the axes of which were formed exclusively by measurement (densitometric) quantities. As is well known, the type of substrate has a major influence on the print quality that can be achieved. It is therefore also possible in the learning phase to assign the corresponding membership vectors for different types of printing material, which result in the same properties in terms of printing technology, in the feature spaces (formed by measuring technology variables). Thus, a feature space for highly absorbent types of printing material will have a different assignment of membership vectors for certain classes and causes than, for example, a feature space for coated papers or very smooth types of paper. The same would also be conceivable for different types of blankets, for example. In general, the type of printing material and the type of blanket are selected once before the start of the print job, so that in the optional phase of the method according to the invention, the corresponding feature space along with the associated assignment is also included Affiliation vectors is established and used.

A video-technical evaluation, for example of specially designed quality diagnostic fields (sliding / doubling fields) can also be used in the method according to the invention. Measured values obtained in this way and processed with the methods of image processing (dot gain, geometry of the grid points, line broadening) can form one or more axes of the feature space, in which case the membership vectors would then have to be assigned to the points of the feature space according to the defined classes or causes via pressure tests. A complete recording of the printed product is also conceivable. The classes to be determined by the experts can then be defined by specific contrast histograms in the RGB color space of the camera. Depending on the class to which the printed product turns out to be the most belonging, certain target values (e.g. color density values) can be specified for the color guide.

The method according to the invention is advantageously implemented in a computer system which is connected to a color density measuring system, a colorimetric or video measuring system. After a measurement on the printed product in the optional phase of the method, the causes in question can then be displayed via an assigned screen device, in particular in conjunction with a rule recommendation for the elimination of the pressure disturbance. It can also be provided that the computer on which the method according to the invention runs is connected directly to corresponding actuating means on the printing press via special interfaces and bus systems. In accordance with the determined cause of a printing difficulty, the provision of printing can then be changed automatically, the dampening solution guide corrected, or the actuators for the ink guide activated.

Claims (3)

Process for monitoring and regulating the printing process, in particular on offset printing machines, in which features reflecting the quality of the printed product are recorded and, in conjunction with other variables influencing the printing process, the printing process is intervened in order to bring the printed product to a desired quality.
characterized,
that from pressure tests evaluated in terms of process engineering, the points in a characteristic space formed by the measuring-technical characteristics are assigned to certain previously defined classes and / or causes of printing difficulties (quality defects) in the manner of a fuzzy classification (learning phase) and
that the monitoring of the printing process from the characteristics obtained on the printed product by measurement technology and via the corresponding membership vector of this point in the feature space determines the membership of the printed product to a specific class or cause, whereupon the membership to this class is displayed and / or to the printing process the cause determined in this way is intervened (optional phase). Method according to claim 1,
characterized,
that the coordinates spanning the feature space are formed by the solid color density (DV) and at least one halftone color density (DR) determined halftone value. The method of claim 1 or 2,
characterized,
that different types of printing material, each with a similar technical behavior, are assigned a characteristic space with a corresponding assignment of the membership vectors.

Images