Method and apparatus for assessing the condition of a pressure punch

Abstract

To provide a method for assessing the state of a pressing punch of a rotary press.SOLUTION: There is provided a pressing punch which includes a punch shaft, a punch head arranged at one end of the punch shaft and configured to interact with a pressure apparatus of a rotary press during operation and a punch tip arranged at another end of the punch shaft and configured to press a material in a cavity of a die plate of the rotary press to form pellets. A method for assessing the state of the pressing punch includes the steps of: recording an image of at least one portion of the pressing punch using a camera; transmitting the recorded image to an evaluation apparatus; performing a state analysis of the at least one portion of the pressing punch by the evaluation apparatus using an image processing algorithm and assessing a state of the pressing punch based on the state analysis; and outputting a warning from the evaluation apparatus when the state of the pressing punch assessed by the evaluation apparatus is outside of a tolerance range. There is also provided a device which executes the method.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The present invention relates to a method for assessing the condition of a pressure punch of a rotary press, wherein the pressure punch has a punch shaft, at one end of which is arranged a punch head that cooperates with a pressure device of the rotary press during operation, and at the other end of which is arranged a punch tip that forces material into a cavity in a die plate of the rotary press to form pellets during operation. The present invention also relates to an apparatus for carrying out this method.

[0002] A rotary press typically has multiple upper and lower punches, each paired and assigned to one cavity of a die plate. During operation, the upper and lower punches rotate together with the die plate, their axial movement controlled by a control cam and guided by upper and lower punch guides. During rotation, the die plate passes through various rotary press devices, including a filling device and a pressure device. In the filling device, the material to be pressed, particularly powder material, is loaded into the die plate cavities. In the pressure device, the upper and lower punches are forced into the cavities, typically by upper and lower pressure rollers, to compress the material into pellets, e.g., tablets. After the pressure device, the upper punch is guided upward from the cavity, and the pellets produced in the cavity are pressed against the upper surface of the die plate by the lower punch. The pellets are then scraped from the die plate, e.g., by a scraper, to the output of the rotary press, where they are delivered for further processing.

[0003] Rotary presses operate at high rotational speeds and correspondingly high production rates. During numerous pressing processes, the pressing punches become worn. In addition to wear on the punch heads due to their interaction with the pressure rollers, the punch tips that press the material into the cavities also wear, for example due to abrasives or other abrasives. Wear and tear on the punch tips can lead to the produced pellets not meeting the acceptable specifications.

[0004] It is known to analyze and evaluate the wear of punch heads using the pressure curve. This method only allows difficult evaluation of the wear of punch tips, or it is not possible to do so at all. This is because the influence of the wear of punch tips, such as the radius of the punch tip or the possibility of a flat edge on the punch tip, on the pressure curve is small. However, such changes in the punch tip have a significant impact on the pellet manufacturing process and, therefore, on its quality.

[0005] It is common to determine the condition of pressure punches by visual inspection. However, this evaluation method is inaccurate and dependent on the operator performing the evaluation. It is difficult to define appropriate objective criteria for evaluating the condition of pressure punches and, if applicable, replacing them. Manual inspection methods are time-consuming. In practice, the insufficient condition of pressure punches, such as punch tips, is often only recognized when production problems occur or the quality of the produced pellets is insufficient. Furthermore, this method does not allow for real-time monitoring of pressure punches, especially punch tips.

[0006] A method for inspecting pressure punches is known from US Pat. No. 10,598,605 B2, in which the pressure punch is arranged in a punch holder. While the pressure punch is held in the holder, it is scanned with an LED micrometer on the one hand and a laser sensor on the other hand. Various geometric parameters of the scanned pressure punch are detected by the LED micrometer and the laser scanner and compared with tolerance ranges. Based on this, a pass or fail status is assigned to the parameters. However, scanning each point on the inspected surface of the pressure punch involves a great deal of effort. The sensors specified in US Pat. No. 10,598,605 B2 are also expensive.

[0007] In view of the prior art described, it is therefore an object of the present invention to provide a method and an apparatus of the type mentioned at the outset, which allows the state of a pressure punch to be assessed in a reliable, simple and cost-effective manner.

[0008] The present invention is defined in independent claim 1 and 14 This object is achieved by: Advantageous embodiments are disclosed in the dependent claims, the description and the drawings.

[0009] For a method of the type mentioned at the beginning, the invention achieves this object by the following steps: recording an image of at least a portion of the pressure punch with a camera; sending the recorded image to an evaluation device; an evaluation device performing a condition analysis of at least a portion of the pressure punch using an image processing algorithm, and evaluating the condition of the pressure punch based on the condition analysis; A step in which the evaluation device outputs a warning if the state of the pressure punch evaluated by the evaluation device is outside the tolerance range.

[0010] The invention also achieves this object by an apparatus for carrying out the method according to the invention, which apparatus comprises a camera for recording images of at least a portion of the pressure punch and an evaluation device, the evaluation device being designed to perform a condition analysis of at least a portion of the pressure punch using at least one image processing algorithm and to evaluate the condition of the pressure punch based on the condition analysis, and the evaluation device being designed to output a warning if the condition of the pressure punch evaluated by the evaluation device is outside an acceptable range.

[0011] The basic configuration of a rotary press in which the pressure punch evaluated according to the present invention is used is described above. Thus, a rotary press generally comprises a plurality of upper and lower pressure punches, each of which presses a material, particularly a powder, into a cavity in a die plate to form pellets, particularly tablets. The pressure punch has a plurality of parts, particularly a punch shaft, at one end of which a punch head is arranged, which cooperates with the pressure device of the rotary press, particularly a pressure roller, during operation, and at the other end of which a punch tip is arranged, which presses the material into the cavity in the die plate during operation. The design of the punch tip depends on the shape and size of the pellets to be produced. Different rotary presses, i.e., different pellets to be produced, use different pressure punches, particularly those with punch tips of different shapes and / or sizes.

[0012] The end face of the punch tip of the pressing punch, which is particularly important in terms of wear, may have an annular flat in its outer region, particularly in a plane perpendicular to the longitudinal axis of the punch shaft. The annular flat surrounds the concave contour that forms the pellets. The annular flat reinforces the outer edge of the end face of the punch tip and protects the contour, particularly in its edge region, from wear and damage, ensuring that the desired pellet shape is always produced. The width of the annular flat is typically less than 1 mm, particularly less than 0.5 mm. Such narrow annular flats are subject to particular loads during operation of the rotary press and are therefore particularly susceptible to wear and damage.

[0013] The present invention is based on the idea of ​​first recording an image or photograph of at least a portion of a pressure punch, such as the end face of the punch tip, the punch head, and / or the punch shaft, using a camera, particularly a digital camera. The recording can be performed, in particular, in a plan view, i.e., in the axial direction of the punch shaft, of the punch tip and / or punch head. The camera can operate in the visible wavelength range, i.e., in the infrared range. The camera can, for example, have a two-dimensional CMOS or CCD sensor. The recorded digital image is sent to an evaluation device, which performs a condition analysis of the recorded portion, and thus the pressure punch. To do so, the evaluation device processes the recorded image with an image processing algorithm or, respectively, applies an image processing algorithm to the recorded image. Corresponding software can be stored in the evaluation device. The tolerance ranges of certain parameters evaluated as part of the condition analysis can be defined, for example, by the operator or by the evaluation device. If this tolerance is exceeded, the evaluation device outputs a warning, for example, in the form of a visual and / or audible signal and / or a message to the operator, for example, via a display or app on a PC, laptop, tablet, or smartphone. The condition evaluated by the evaluation device can, in particular, be the wear state of at least a portion of the or each pressing punch. As explained above, wear of the punch tip occurs during operation of the rotary press, for example, due to the application of abrasive material. As part of the condition analysis according to the present invention, damage to the surface of the punch tip, in particular scratches, bumps, or other deformations on the end face, can also be detected and evaluated. This also applies to the punch head and punch shaft. Real-time monitoring of the pressing punch is also conceivable. Furthermore, automatic evaluation of the pressing punch by an evaluation device coupled to a camera is also possible. The condition evaluation is performed, in particular, independently of individual inspection by an operator. At the same time, the method according to the present invention or the device according to the present invention, respectively, can be implemented more cost-effectively than using laser scanners as proposed in the prior art.The use of image processing algorithms ensures a reliable assessment of the state of the pressing punches, thus allowing the quality characteristics of the produced pellets to be continuously optimized, the process stability to be increased and unplanned downtimes during operation of the rotary press to be minimized in a simple and reliable manner.

[0014] All upper and / or lower pressing punches of the rotary press may be evaluated using the method according to the invention. The evaluation device may display the results of the condition analysis to an operator. The evaluation device may also include an input device with which the operator can set parameters for the condition analysis, such as tolerances for certain evaluated parameters or the type of punch to be evaluated. The results of the condition analysis and the operator's input can be displayed on the display of the rotary press, or on an input device of the evaluation device, or on an input device of a separate device, such as a PC, laptop, tablet, or smartphone. In particular, the display and input can be performed via a corresponding application (app).

[0015] According to one embodiment, the condition analysis using an image processing algorithm may include comparing the recorded image with at least one reference image. The at least one reference image may represent a target condition of at least a portion, such as the punch tip or its end face. The image processing algorithm may be used to compare the recorded image with at least one or more reference images. Certain parameters, such as the outer diameter of the end face of the punch tip, if present, the inner diameter of the annular outer flat area of ​​the end face or the flatness of the end face, and in particular the occurrence of irregular areas in the recorded image compared with the at least one reference image, may be compared. This also applies to punch heads, for example. A database of reference images for each punch type, e.g., each punch tip shape and / or size, may be created so that the recorded image can be compared with the reference images stored in the database. Therefore, by directly comparing the recorded image with the reference image, edge damage, for example in the area of ​​the punch tip, may be quickly and reliably detected.

[0016] According to another embodiment, the at least one reference image may be at least one reference image of at least a portion of an unused pressure punch and / or at least one reference image of at least a portion of the pressure punch to be evaluated before a final production process of the pressure punch in the rotary press. Thus, a comparison can be made with a new state of the pressure punch or with a previous state of the pressure punch, in particular a state just before a final production process of the pressure punch in the rotary press or a state just before multiple production processes of the pressure punch in the rotary press.

[0017] Generally, the recorded image can be compared to a reference image of the exact pressure punch being evaluated, so that the individual characteristics of each pressure punch can be evaluated and taken into account, although it is also contemplated that the reference image may be a standard image of the type of punch being evaluated.

[0018] The recorded image and the at least one reference image can be processed by the evaluation device using at least one image processing algorithm before being compared. Various image processing methods can be used that simplify further processing, such as smoothing, grayscale conversion, etc.

[0019] According to another embodiment, the condition analysis using an image processing algorithm may include extracting a region of interest from the recorded image, preferably the end face of the punch tip. By extracting what is known as a region of interest (ROI) from the recorded image, for example, the edge area of ​​the punch tip can be analyzed particularly reliably. The image to be evaluated is limited to the end face of the punch tip. The area visible behind the pressure punch, in particular the punch shaft and, if applicable, the punch head, is not taken into account. This makes the evaluation simple and fast.

[0020] The recorded image may be a color image (RGB). Then, according to another embodiment, the condition analysis using an image processing algorithm may include grayscaling the color image. Furthermore, the condition analysis using an image processing algorithm may include converting the recorded image into a binary image. The binarization may be performed using an adaptive threshold. The grayscale conversion and binarization make it easy to find optical characteristics of the recorded image related to the wear state. The binary image may also be composed of floating-point numbers between 0 and 1.

[0021] As already mentioned, at least a portion of the pressure punch recorded by the camera may be the end face of the punch tip of the pressure punch. Then, according to another embodiment, the condition analysis using an image processing algorithm may include determining the center point of the recorded end face of the punch tip. In this context, the center point refers specifically to the geometric center point or geometric center, since the end face is not necessarily circular. Furthermore, the condition analysis using an image processing algorithm may include converting the coordinates of the recorded image into a polar coordinate system. It may be useful to convert the recorded image into polar coordinates and thus first determine the center point of the end face of the punch tip, particularly for evaluating the flat annular region outside the punch tip.

[0022] According to another embodiment, in which the end face of the punch tip has an annular flat in its outer region as described above, the condition analysis using an image processing algorithm may include identifying the outer and inner boundaries of the annular flat. The lines defining the annular flat may, for example, be circular in the recorded image, but ideally appear as parallel lines after conversion to polar coordinates. This significantly simplifies the evaluation of the annular flat's boundary using the image processing algorithm. Therefore, the condition analysis using an image processing algorithm may include determining the distance around the circumference of the annular flat between the outer and inner boundaries. This distance can be determined quite simply from the perspective of the circumferential evaluation technique, especially after conversion to polar coordinates, by determining pixels on the processed recorded image along an axis that intersects the parallel lines representing the boundary. The principal maxima occurring in the processed recorded image represent the inner and outer edges of the annular flat on the end face of the punch tip. Therefore, by determining the maxima in the image processed as part of the condition analysis, the outer and inner boundaries of the annular flat can be identified. The distance evaluation can be easily performed from the difference between the principal maxima occurring before and after conversion from polar coordinates to Cartesian coordinates. However, it is of course also possible to determine the outer and inner boundaries of the annular plateau in Cartesian coordinates without determining the center point and converting to polar coordinates, in particular by determining the pixels that form the boundaries on an array sensor, for example. The distance values ​​of the outer and inner boundaries of the annular plateau determined around the circumference of the end face of the punch tip can be compared, for example, with a target distance value. If individual or multiple distance values ​​determined around the circumference deviate unreliably from the target value, the evaluation device can infer that wear or damage to the annular plateau or its outer edge has increased. It is also conceivable to measure the variation range of the distance values ​​determined around the circumference, for example, the minimum and maximum distance values. If the variation range exceeds an acceptable range, increased wear or damage to the annular plateau and thus the punch tip can be inferred.

[0023] Reference to an annulus in this context includes not only circular annulus but also other annular geometric shapes such as elliptical or oval annulus, or non-circular annular shapes such as angular or polygonal annulus, for example.

[0024] According to another embodiment, the condition analysis using the image processing algorithm may include applying a Canny algorithm to the recorded image, which can effectively remove image information that is not relevant to the condition assessment, such as light reflections from a light source provided to illuminate at least a portion of the image, from the recorded image, thereby simplifying the subsequent condition analysis.

[0025] For example, using an image processing algorithm based on the Canny algorithm, condition analysis may involve detecting anomalies on at least a portion of the punch, such as the end face of the punch tip. Anomalies are deviations from an expected state, such as surface defects, resulting in the surface partially deviating from an expected shape, such as flat or concave. Typical anomalies include scratches, cracks, and dents. Image processing algorithms for cluster analysis can also be used to detect anomalies. Using this, anomalies are assessed as particularly critical if they occur frequently in one location or area. For example, in addition to the edge width of the annular flat of the punch tip, the entire punch tip surface, especially the recessed surface that is usually concave and forms the pellets, can also be inspected by the image processing algorithm. To this end, anomalies can be detected from the processed image, for example, after extracting regions of interest, grayscaling if applicable, and, for example, smoothing the recorded image. For example, after smoothing the image, it is possible to use the DBSCAN (density-based spatial clustering of applications with noise) algorithm to evaluate anomalies formed, for example, by scratches, cracks, or dents in the punch tip. In this way, particularly serious defects in the punch tip that lead to defects in the corresponding pellets produced can be detected early and reliably, allowing the affected pressing punch to be identified and, for example, replaced.

[0026] According to another embodiment, the condition analysis can be performed by a machine learning algorithm. In particular, a self-learning algorithm continuously optimizes the evaluation criteria for the condition of the pressing punch, particularly the punch tip, based on training data from previous evaluation processes and / or empirical data. The limits for the deviation of parameters determined as part of the image processing according to the present invention from target parameters can be optimized during operation by this type of self-learning software based on training data and / or empirical data. The machine learning algorithm can, for example, consist of a neural network.

[0027] According to another embodiment, the press punch to be evaluated can be attached to the rotary press during the method, or can be held in a holder outside the rotary press, in particular in a cleaning device for the press punch. The camera can also be held on a support arm that can be moved into the press space of the rotary press or into the cleaning device for the press punch. To evaluate the press punches of the rotary press, the rotor can be rotated, for example, sequentially, so that the camera can sequentially record at least a portion of the press punches of the rotary press, for example, the end faces of the punch tips. In this way, all press punches can be evaluated. To this end, the support arm can be rotated, for example, 360 degrees, so that the upper and lower press punches of the rotary press can be sequentially recorded. According to the above-described embodiment, the condition analysis of all press punches can be largely automated, both when installed in the rotary press and when removed from the rotary press. It is also conceivable to integrate the camera used according to the present invention into an additional tactile measurement system for detecting the geometric parameters of the press punches. The support arm can be part of a robot, for example, so that automatic control of the support arm can be performed to perform the condition analysis. This can also be controlled by the evaluation device.

[0028] The method can also be performed in a separate measuring device, where punches, individually or in groups, are guided into a support sheet of a measuring device provided for this purpose. The support sheet is positioned so that images can be recorded with a camera, for example, in a light tent. When measuring a group of punches, each punch is held in the support sheet, for example, with a connector for each punch. This configuration allows images of each punch to be recorded without removing the punch from the support sheet. To achieve this, the support sheet is fixed, for example, in the aforementioned light tent, using a holder or the like, while a camera is moved to each punch and images are recorded.

[0029] Further advantages can be achieved by incorporating the method of the present invention into a cleaning device for the pressure punch of a rotary press. Meanwhile, regularly storing the pressure punch in a holder within the cleaning device and fixing its position (GMP environment) facilitates image recording of the end face of the punch tip. Furthermore, recording images within the cleaning device can avoid problems that may arise in the pressure space of the rotary press due to existing production dust, etc. If the cleaned punch is polished before the condition assessment according to the present invention, additional parameters that may affect the condition assessment can be reliably excluded. For example, when incorporated into the cleaning device, at least a portion of the camera images can be taken before and / or after each cleaning cycle. By incorporating the method of the present invention into a rotary press, for example, camera images of the pressure punch can be taken before the start of each production process, and a corresponding condition assessment can be performed.

[0030] The present invention also relates to a cleaning device for pressure punches of a rotary press, comprising a holder for holding the upper and / or lower pressure punches in a cleaning chamber, the cleaning device further comprising a cleaner for cleaning the upper and / or lower pressure punches, and the cleaning device may also comprise a polishing device for polishing the pressure punches.

[0031] The present invention also relates to a rotary press. The rotary press comprises a rotor rotatable by means of a rotary drive, the rotor comprising an upper punch guide for an upper pressure punch, a lower punch guide for a lower pressure punch, and a die plate disposed between the punch guides. The pressure punch cooperates with a cavity in the die plate. The rotary press further comprises a filling device, by means of which the powder material to be pressed is filled into the cavity in the die plate. The rotary press comprises a pressure device having upper and lower pressure rollers that cooperate with the upper and lower pressure punches during operation to press the powder material to be pressed into the cavity in the die plate, and the rotary press comprises the device according to the present invention.

[0032] With the method according to the invention and the device according to the invention, it is also possible to record additional components that are placed on the pressure punch, such as bellows or dust caps, with a camera and to evaluate the recorded images according to the method according to the invention, so that the state of the other components can also be evaluated according to the invention. [Brief explanation of the drawings]

[0033] Exemplary embodiments of the invention are explained in more detail below with reference to the figures.

[0034] [Figure 1] FIG. 1 is a side view of a pressure punch of a rotary press. [Figure 2] Figure 1 shows an enlarged view of the punch tip of the pressure punch. [Figure 3] Figure 2 shows a top view of the punch tip. [Figure 4] An apparatus according to the invention for assessing the condition of the pressure punch shown in Figures 1 to 3. [Figure 5] 10A and 10B illustrate a condition analysis according to the present invention using an image processing algorithm. [Figure 6] 4A and 4B are further diagrams illustrating the condition analysis according to the invention using image processing algorithms; [Figure 7] 4A and 4B are further diagrams illustrating the condition analysis according to the invention using image processing algorithms; [Figure 8] 4A and 4B are further diagrams illustrating the condition analysis according to the invention using image processing algorithms; [Figure 9] 4A and 4B are further diagrams illustrating the condition analysis according to the invention using image processing algorithms; [Figure 10] 4A and 4B are further diagrams illustrating the condition analysis according to the invention using image processing algorithms;

[0035] Unless otherwise specified, the same reference symbols refer to the same objects in the figures.

[0036] FIG. 1 shows a schematic representation of a pressure punch for a rotary press. The pressure punch comprises, for example, a substantially cylindrical punch shaft 10, one end of which is provided with a punch head 12 that cooperates with the pressure device of the rotary press, particularly the pressure roller, during operation. The other end of the punch shaft 10 is provided with a punch tip 14 that is inserted into a cavity in the die plate of the rotary press during operation and, in cooperation with the punch tip of the opposing pressure punch, presses the powder material filled in the cavity to form pellets, particularly tablets. FIG. 2 shows an enlarged view of a portion of the punch tip 14. It can be seen that the punch tip 14 has an annular flattened portion 18 (edge ​​width portion) in the outer region of its end face 16 (pressure surface), which lies in a plane perpendicular to the longitudinal axis of the punch shaft 10, which runs vertically in FIG. 2. FIG. 3 shows a plan view of the punch tip 14 as seen from above in FIG. 2. The annular flattened portion 18 may be, for example, circular or elliptical. This defines a central recess 20 (pressing surface without edge width) which forms the outline of the pellets produced during operation of the rotary press. Further pressing punches, which are paired with and assigned to the opposite side of the pressing punch shown in Figures 1 to 3, can be designed identically, so that the annular flats 18 of the opposing pressing punches abut each other during operation when pressing the powder material.

[0037] During operation, for example, abrasive powder material or other reasons can cause wear and tear on the end face 16 of the punch tip 14, particularly the annular flat 18, which is relatively sensitive due to its narrow width.

[0038] FIG. 4 shows an apparatus according to the present invention for evaluating the condition of a pressure punch, in particular the punch tip 14. Again, for clarity, only a portion of the punch tip 14 is shown enlarged. The apparatus includes a camera 22, in particular a digital camera 22, e.g., with a two-dimensional CMOS or CCD array sensor, with which an image of the end face 16 of the punch tip 14 is recorded in a plan view, as indicated by the field of view 23, which is diagrammatically depicted in FIG. 4. The recorded image is transmitted via a wired or wireless connection 24 to an evaluation device 26. The evaluation device 26 performs a condition analysis of the punch tip 14 of the pressure punch using at least one image processing algorithm and evaluates the condition of the pressure punch based on the condition analysis. To this end, at least one image processing algorithm is applied to the recorded image by the camera 22. To this end, corresponding software is stored in the evaluation device 26. The evaluation device 26 may also include a display and / or operating devices, e.g., for displaying a warning if the condition of the pressure punch evaluated by the evaluation device 26 is outside of an acceptable range and / or for an operator to input evaluation parameters. The evaluation device 26 can also communicate with external devices, such as PCs, laptops, tablets, and smartphones, that can store apps for displaying and inputting data from the evaluation device 26.

[0039] The evaluation of the images recorded by the camera 22 using at least one image processing algorithm will be explained in more detail with the aid of Figures 5 to 10. On the left side of Figure 5, an exemplary image of the end face 16 of the punch tip 14 recorded by the camera 22 is shown. In addition to the end face 16 with the annular flat 18 and the recess 20, the punch shaft 10 can also be seen outside the focus of the camera 22. In a first evaluation step, the evaluation device 26 uses an image processing algorithm to extract the end face 16 of the punch tip 14 as a region of interest from the image on the left side of Figure 5, as shown on the right side of Figure 5.

[0040] The left side of Figure 6 shows the processed image of the right side of Figure 5. In a further evaluation step, the evaluation unit 26 uses an image processing algorithm to grayscale and convert the image of the left side of Figure 6 into a binary image, as shown on the right side of Figure 6.

[0041] This binary image, shown again on the left in Figure 7, is further processed by the evaluation unit 26 using image processing algorithms so that the center point 28 of the annular flat 18, and therefore the end face 16 of the punch tip 14, and the rectangle 30 that defines the annular flat 18 are determined.

[0042] The processed image shown on the right side of FIG. 7 is then converted by the evaluation device 26 using an image processing algorithm into polar coordinates as shown on the left side of FIG. 8. The inner and outer boundaries of the annular flat portion 18, which is approximately circular in FIG. 7, appear here as straight lines that are parallel in the ideal case. As shown on the right side of FIG. 8, the distance between two intensity maxima 32, 34 corresponds to the distance between the inner and outer boundaries, and therefore corresponds to the width of the annular flat portion 18, and is therefore in the x direction, i.e., from left to right in FIG. 8. Using an image processing algorithm, the evaluation device 26 alternately determines the maxima 32, 34 shown on the right side of FIG. 8 on the y axis, i.e., from bottom to top in FIG. 8, so that the distance between the inner and outer boundaries of the annular flat portion 18 can be measured over the entire circumference of the annular flat portion 18.

[0043] As explained with reference to Figures 5 to 8, condition analysis using image processing algorithms makes it possible to reliably and simply determine the condition of the annular plateau 18, and in particular defects at the edge of this portion 18. The determined parameters, for example the distance between the inner and outer boundaries of the annular plateau 18, can then be compared with tolerances and, in the event of deviations from the tolerances, a warning signal can be output as explained.

[0044] Figures 9 and 10 are used to explain additional or alternative possible condition analyses using at least one image processing algorithm by the evaluation device 26, by means of which the surface condition of the end face 16 of the punch tip 14 can also be evaluated.

[0045] Figure 9 shows four diagrams. The upper left shows the image captured by the camera 22 and reduced to the region of interest, as also shown on the right side of Figure 5. This image is then converted to a grayscale image by the evaluation unit 26 using an image processing algorithm, as shown in the upper right of Figure 9. Furthermore, as shown in the lower left of Figure 9, the evaluation unit 26 can smooth the grayscale image using an image processing algorithm. The processed image can then be subjected to, for example, a Canny edge image processing algorithm. The result is shown in the lower right of Figure 9. The surface of the end face 16 of the punch tip 14 exhibits anomalies that may be caused by scratches or similar surface defects. Therefore, the evaluation unit 26 can detect these anomalies on the end face 16 of the punch tip 14 and output a warning that the punch tip 14 is becoming increasingly worn.

[0046] The upper left of Figure 10 shows a sequence of binary images generated by the evaluation device 26 using an image processing algorithm, for example, from the grayscale image shown in the upper right of Figure 9. In this binary image, the evaluation device 26 can again use an image processing algorithm to determine the inner and outer boundaries of the annular flat 18. Furthermore, as shown on the right side of Figure 10, the evaluation device 26 can apply an image processing algorithm for cluster analysis to detect anomalies on the surface of the end face 16 of the punch tip 14. For example, a density-based spatial clustering of applications with noise (DBSCAN) algorithm can be used. The evaluation device 26 can detect accumulations of detected anomalies, i.e., clusters, as shown in Figure 10, with four clusters 36, 38, 40, and 42. These clusters 36, 38, 40, and 42 are evaluated by the evaluation device 26 as surface defects, and a warning can again be output.

[0047] The condition analysis can be performed by the evaluation device 26 using machine learning algorithms. To this end, training data of image morphologies of the end face 16 of the punch tip 14 that are within the acceptable range and training data of image morphologies of the end face 16 that are outside the acceptable range can be used.

[0048] Although the recording and evaluation of images of the punch tip 14, and in particular its end face 16, have been described with reference to the figures, other parts of the pressure punch, such as the punch head 12 and / or punch shaft 10, can likewise be recorded by a camera and thus evaluated in the described manner. As mentioned above, further components connected to the pressure punch, such as bellows or dust caps, can also be recorded and evaluated in this manner. [Explanation of symbols]

[0049] 10...Punch shaft 12...Punch head 14...Punch tip 16...End face (pressure surface) 18...Annular flat portion (edge ​​width portion) 20...Concave portion (pressure surface without edge width portion) 22...Camera 23...Field of view 24...Connection 26...Evaluation device 28...Center point 30…Rectangular part 32...Maximum value 34...Maximum value 36...Cluster 38...Cluster 40...cluster 42...Cluster

Claims

1. 1. A method for evaluating the condition of a pressure punch of a rotary press, the pressure punch having a punch shaft (10) at one end of which is disposed a punch head (12) for cooperating with a pressure device of the rotary press during operation, and at the other end of which is disposed a punch tip (14) for forcing material into a cavity in a die plate of the rotary press to form a pellet during operation, the method comprising: - recording an image of at least a portion of said pressure punch with a camera (22); - sending said recorded image to an evaluation device (26); the evaluation device (26) performing a condition analysis of at least a portion of the pressure punch using an image processing algorithm and evaluating the condition of the pressure punch based on the condition analysis; - if the state of the pressing punch evaluated by the evaluation device (26) is outside a tolerance range, the evaluation device (26) outputs a warning, During the method, the press punch to be evaluated is held in a holder external to the rotary press; at least a portion of the pressure punch is an end face (16) of the punch tip (14) of the pressure punch; the condition analysis using an image processing algorithm includes determining a center point (28) of the recorded end face (16) of the punch tip (14); The condition analysis using an image processing algorithm includes converting the coordinates of the recorded image into a polar coordinate system; a method characterized in that the end face (16) of the punch tip (14) has an annular flat (18) in its outer region, and the condition analysis using an image processing algorithm includes identifying an outer boundary and an inner boundary of the annular flat (18).

2. 2. The method of claim 1, wherein the condition analysis using an image processing algorithm includes comparing the recorded image with at least one reference image.

3. 3. The method of claim 2, wherein the at least one reference image is at least one reference image of at least a portion of an unused pressure punch and / or the at least one reference image is at least one reference image of at least a portion of the pressure punch being evaluated before a final manufacturing process of the pressure punch being evaluated in the rotary press.

4. 2. The method of claim 1, wherein the condition analysis using an image processing algorithm comprises extracting an area of ​​interest from the recorded image, the area of ​​interest preferably being an end face (16) of the punch tip (14).

5. 2. The method of claim 1, wherein the recorded image is a color image, and wherein the condition analysis using an image processing algorithm includes grayscaling the color image.

6. 2. The method of claim 1, wherein said condition analysis using an image processing algorithm includes converting said recorded image into a binary image.

7. 2. The method of claim 1, wherein the condition analysis using an image processing algorithm also includes determining the distance around the circumference of the annular flat (18) between the outer boundary and the inner boundary.

8. 2. The method of claim 1, wherein the outer and inner boundaries of the annular plateau can be identified by determining maxima (32, 34) of images processed as part of the condition analysis.

9. 2. The method of claim 1, wherein the condition analysis using an image processing algorithm includes the use of a Canny algorithm.

10. 2. The method of claim 1, wherein the condition analysis using an image processing algorithm includes detecting abnormalities in at least a portion of the pressure punch.

11. 11. The method of claim 10, wherein an image processing algorithm for cluster analysis is used to detect anomalies in at least a portion of the pressure punch.

12. The method of claim 1 , wherein the condition analysis is performed by a machine learning algorithm.

13. 2. The method of claim 1, wherein the pressure punch to be evaluated is held in a pressure punch cleaning device of the rotary press while the method is being performed.

14. An apparatus for performing the method described in any one of claims 1 to 13, comprising a camera (22) for recording images of at least a portion of the pressure punch and an evaluation device (26), the evaluation device (26) being designed to perform a condition analysis of at least a portion of the pressure punch using at least one image processing algorithm and evaluate the condition of the pressure punch based on the condition analysis, and the evaluation device (26) being also designed to output a warning if the condition of the pressure punch evaluated by the evaluation device (26) is outside an acceptable range.

Images