Method and apparatus for production monitoring and in particular real-time control in a system for producing, filling and / or packaging containers for beverages or similar products

Abstract

The invention describes a method and an apparatus for production monitoring and in particular real-time control in a system for producing and / or filling and / or packaging containers for beverages or similar products. For this purpose, the containers are continuously transported through a stationary imaging region of at least one sensor that provides two-dimensional imaging, wherein position data relating to the containers are determined from recordings of the containers, which are generated by the sensor, by means of electronic evaluation. Furthermore, times for subsequent location-based machine manipulation and / or inspection of individual containers or all containers are predicted on the basis of the position data. This provides a simplified and more reliable sensor system for triggering container manipulations and inspections.

Description

[0001] Method and device for production monitoring and in particular real-time control in a plant for the production, product filling and / or packaging of containers for beverages or similar products

[0002] As is well known, containers for beverages or similar filling products in bottling plants or similar production facilities are monitored at specific points during transport using stationary individual sensors, such as light barriers, dynamic pressure switches, or flap switches, with regard to their absolute position, orientation, relative arrangement to one another, or similar transport-relevant parameters. This allows for any necessary transport-related manipulation of the containers, such as the rejection of faulty containers, the distribution of containers to transport lanes, or the creation of a specific positional overview of the containers relative to one another before packaging. Such sensors typically provide time-dependent status information, for example, whether a container is present in the respective detection area.Combined with drive data from the respective transport means, for example, transport speeds determined by rotary encoders, it is possible to calculate when a specific container enters the range of a container manipulator, allowing any necessary machine actions to be carried out, such as the removal of a faulty container. For this purpose, the time when the respective container arrives at the manipulator should be calculated as precisely as possible, given the typical machine performance in filling plants.

[0003] The problem with this approach is that, due to the selective condition recording with such individual sensors, a large number of them must usually be combined at different locations and possibly of different types in order to obtain the necessary information for subsequent container manipulation. Disadvantages include, for example, the effort required for mounting and adjusting the sensors for type-dependent sensor changes, the wear and tear of mechanical sensors, the contamination and corresponding cleaning requirements of optical sensors, the sensitivity of the sensors to mechanical damage and various environmental influences, the usually short measuring distances to the containers, the generally necessary plausibility checks of the condition data determined by the sensors, and the limited suitability for determining container conditions, such as for identifying overturned or damaged containers.

[0004] There is therefore a need for improvement.

[0005] The stated object is achieved with a method and a device according to the independent claims. Preferred embodiments are specified in the subclaims. The inventive use of at least one area-imaging sensor enables real-time container monitoring within a fixed imaging range of the sensor during the container's transport through the imaging range. The area-imaging sensor thus provides more information than point-based condition sensors such as light barriers, which can only detect the presence of an object at a specific location at a specific time. In other words, the area-imaging sensor enables transport tracking of the container within the fixed imaging range.

[0006] For example, a transport speed and / or position of the container can be determined from at least one image captured by the area-imaging sensor, in particular from several such images. Thus, in the simplest case, only a single area-imaging sensor is required to determine the container's position data. On this basis, a prediction can be made as to when the container thus represented (imaged) will arrive at a container manipulator, in order to control the container accordingly at a suitable time and thus perform location-specific (at the location of the manipulator) container manipulation.

[0007] The area-imaging sensor (in the sense of at least one detector and the associated readout electronics) represents the moving containers by definition in images of the stationary imaging area and is, for example, a camera. Instead or in addition, a surface scan of the imaging area is also possible, for example, using a laser beam and associated time-of-flight measurement. The area-imaging sensor could be designed as at least one 2D camera, a 3D camera, a combination of line sensors, 2D and / or 3D cameras, a radar sensor, or the like. For the sake of simplicity, the area-imaging sensor will hereinafter also be referred to simply as an area sensor or sensor.

[0008] In addition to absolute positions and / or relative transport positions of the containers, the position data determined with the area sensor can contain further position information, such as: position of the object in terms of lying or upright; containers crowded together; container shape; container size; container volume; and / or container rotational position.

[0009] At least one electronic evaluation unit and control unit is available for the real-time evaluation of the raw image data supplied by the area sensor, so that monitored containers can be manipulated and / or inspected in the ongoing product flow (and thus in real time), for example as described below.

[0010] Preferably, the stationary imaging area is so large that several containers present on the means of transport can be imaged in individual images of the area sensor in order to determine the position data of these containers at a specific imaging time, if necessary jointly from one image.

[0011] Assigning the position data to time series enables systematic transport tracking of the containers and, if necessary, determining the associated transport speeds in the area immediately before a container manipulation to be controlled, possibly together with other transport sections that the container passes through in the production facility. In this way, the controller can also access position data collected upstream of the described device and / or status and / or location information of the respective container, which can also be components of the time series and thus be assigned to the respective position data.

[0012] Such time series preferably consist of several points in time and associated position data as well as, if applicable, associated status data of the containers and / or drive data of associated means of transport at the corresponding points in time.

[0013] The time intervals in the respective time series are usually fixed and / or known as a result of the monitoring process, but can also be estimated or interpolated if necessary. Individual time intervals in the time series can differ from one another. The time intervals can also be derived from movement intervals and / or machine cycle intervals. Movement intervals can be recorded, for example, by position sensors, cycle intervals by rotary encoders on the respective transport device, or by transport pitches in a forced container guide.

[0014] The predicted time for an assigned machine manipulation and / or inspection is calculated, for example, by extrapolating the time series.

[0015] The described area sensor enables the monitoring of different container formats without mechanical adjustment of the sensor or the like.

[0016] The described area-imaging sensor can, for example, comprise one or more cameras. Such cameras can capture the imaging area seamlessly or with suitable gaps, forming multiple partial imaging areas. The partial imaging areas then have a fixed and known arrangement relative to one another.

[0017] The at least one area sensor can, for example, be arranged above and / or next to the respective means of transport and, for example, be aligned in its transport direction, opposite to its transport direction, or transversely thereto. An artificial intelligence can, for example, be trained using the time series described. For example, such time series can be examined for anomalies, as these are particularly meaningful for AI training runs. Determined time series can be stored centrally and / or across systems, for example in a data cloud. Time series stored in this way are suitable for the training, validation, and / or test runs of newly installed neural networks or the like. For example, a fundamentally identical neural network can be used in different production systems orFilling lines can be used for corresponding control tasks, but possibly with different line layouts or machine layouts, and time series data can be exchanged across plants.

[0018] A container manipulator in the sense of the present invention can, for example, be a pusher, printer or label dispenser assigned to the containers.

[0019] Container manipulation within the meaning of the present invention is understood to mean an action that specifically influences an individual container, for example its targeted treatment, processing, handling and / or transport guidance.

[0020] An inspection is understood here as a check of the properties and / or conditions of individual containers that does not directly influence and / or change the containers.

[0021] The described calculation of times for triggering a container manipulator, including the associated output of trigger signals or the like, can be applied equally to times for triggering an inspection unit or at least one inspection sensor contained therein, such as an inspection camera (and vice versa). This means that not only container manipulations, but also the recording of container properties and / or states by associated inspection sensors, such as cameras, can be triggered using the described time series and the times determined from them.

[0022] In principle, any sequence of container manipulators and / or inspection units is conceivable for this purpose. An action such as a measurement could be initiated or triggered by the inspection sensor. Continuously recorded inspection sensor signals could be sampled. Examples of inspection sensors include: load cells; inspection cameras, for example, for cap inspection; code readers; initiators; and heat sensors such as pyrometers or thermal imaging cameras.

[0023] A preferred embodiment of the invention is illustrated in the drawings.

[0024] Figure 1 shows a device according to the invention; and Figure 2 shows a schematic representation of a time series.

[0025] As can be seen in a schematic representation in Figure 1, the device 1 for production monitoring and in particular real-time control in a system (not shown) for the manufacture and / or product filling and / or packaging of containers 2, such as bottles or cans, for beverages or similar products (not shown), comprises: at least one transport means 3, designed for example as a conveyor belt, for the continuous (in the sense of non-clocked) transport of the containers 2; at least one sensor (area sensor) 5 which produces a planar image in a stationary imaging region 4 and which is used to generate images 6 of the containers 2 during their transport on the transport means 3;and an electronic evaluation unit 7 for determining position data PD1 to PD5 (see Figure 2) of the containers 2 from the recordings 6 and for predicting times TS based thereon for subsequent location-related mechanical control of a container manipulator 8, such as a pusher for diverting faulty containers 2.;

[0026] The planar imaging sensor / area sensor 5 can, for example, consist of one or more cameras. The stationary imaging area 4 then consists, for example, either of the image area of ​​a single camera or, if appropriate, of the spatially fixed arrangement of image areas of multiple cameras. Such image areas can be seamlessly adjacent to one another, overlap, or even spaced apart from one another.

[0027] The area sensor 5 is preferably arranged above and / or next to the transport means 3 in such a way that the imaging area 4 is larger than the largest dimensions of the containers 2 transported upright and / or horizontally. This enables overall views of at least one container 2, in particular of several containers 2, in individual images 6. Thus, the respective position, transport orientation, and possibly a state of the containers 2 can be monitored using only a single sensor 5.

[0028] In the example shown, the area sensor 5 is oriented transversely to the transport direction 3a of the transport means 3. An orientation in or against the transport direction 3a can be advantageous in order to monitor a longer section of the transport means 3 with, for example, only a single camera as described. A sensor arrangement / orientation on curved sections or the like can also be advantageous, such that as much of the respective transport device as possible is detected with one sensor 5. For example, one sensor 5 could thus detect a transport section with a U-formation.

[0029] In the example shown, the device 1 is a component of an inspection device 11 for inspecting the containers 2, for example, for empty bottle inspection. The inspection device 11 comprises at least one inspection unit 12 with an inspection sensor 13, here, for example, for imaging contour control of the containers 2. The container manipulator 8 is then, in particular, a pusher for rejecting containers 2 that have been identified by the inspection unit 12 as being of inadequate quality.

[0030] In principle, corresponding rejection criteria could also be output directly by the evaluation unit 7 as a result of an image evaluation of the recordings 6, for example in the case of containers 2 lying flat or grouped too closely together.

[0031] During ongoing operation, the containers 2 arriving on the transport means 3 are recorded several times by the surface imaging sensor 5, preferably at suitable time intervals.

[0032] From this, a time series 21 is created, shown schematically in simplified form in Figure 2, here, for example, with five imaging times TA1 to TA5, at each of which an image 6 of the imaging area 4 with at least one container 2 transported through it is created using the area sensor 5. From the respective positions of the at least one container 2 in the images 6, position data PD1 to PD5 of the container 2 are determined by electronic image analysis and assigned to the imaging times TA1 to TA5 as a time series 21.

[0033] The respective time series 21 can be newly created on the basis of the mapping times TA1 to TA5 or, if it has already been generated previously, for example in connection with a production process upstream, it can be supplemented accordingly.

[0034] At the same time as the recordings 6, for example, rotary encoder cycles or similar information of the means of transport 3 can be recorded and included in the time series 21 as drive data AD1 to AD5.

[0035] On the basis of the time series 21 with the mapping times TA1 to TA5 and associated position data PD1 to PD5, it can be predicted at which time TS the respective container 2 arrives at the container manipulator 8 to which position data PD6 is assigned.

[0036] Likewise, corresponding times TS for the respective arrival of the containers 2 at the inspection sensor 13 could be determined. In this way, i.e., using a specific time series 21, times TS for different locations for the respective manipulation or inspection of the containers 2 can also be determined.

[0037] Based on the time series 21, it can be calculated automatically at which control time TS the container manipulator 8 is triggered and / or the inspection unit 12 is triggered, for example by a trigger signal output by the evaluation unit 7, in order to carry out an inspection and, if necessary, the desired manipulation of the container 2, in this case the discharge of a container 2 that is recognized as being of inadequate quality.

[0038] Thus, several location-dependent times TS (trigger times) can be calculated on the basis of the respective image evaluation, in the example for the image recording (container inspection) with the inspection sensor 13 of the inspection unit 12 and for the drive of the container manipulator 8.

[0039] When transporting containers 2 at a constant or changing transport speed, the acquisition and assignment of drive data AD1 to AD5 is not absolutely necessary for predicting control times TS; in principle, a time series 21 with the position data PD1 to PD5 is sufficient for this purpose.

[0040] On the other hand, the drive data AD1 to AD5 for drive control are usually available anyway, so that they are preferably taken into account in the time series 21 in order to trigger / control the container manipulator 8 and / or the inspection sensor 13 as precisely as possible.

[0041] In addition, the inspection device 11 and thus also the transport means 3 usually operate at different transport speeds during production, which are determined, for example, by a filling machine (not shown) downstream depending on the production situation there. This requires acceleration and braking processes on the transport means 3, which is why the containers 2 can change their transport positions relative to the transport means 3. This is also due to the low static friction of the containers 2 on the transport means 3, which is usually intended for other reasons.

[0042] However, relative position changes of the containers 2 with respect to the transport means 3 can be compensated for by the control system as a result of container monitoring using the area sensor 5. The position data PD1 to PD5 are evaluated as a time series 21 and, if necessary, the drive data AD1 to AD5 are included. However, the transport-related resynchronization that was previously necessary when using individual sensors for point-based monitoring can be avoided.

[0043] The integration of drive data AD1 to AD5 into the described procedure is optional.

[0044] The area-imaging sensor 5 enables not only the acquisition of absolute position data PD1 to PD5 of the container 2, but also the detection of the position of the container 2, for example, whether a container 2' has fallen over and is lying on the transport means 3. Defective containers 2 can also be detected, for example, if the body of the container 2 is missing and only a base cup 2" instead of a container 2 moves through the stationary imaging area 4. In this case, too, the container manipulator 8 can be controlled accordingly, for example, to divert the base cup 2" and avoid transport problems further downstream.

[0045] In principle, the time series 21 can be updated downstream of the device 1. For example, it can be entered into the time series 21 that a specific container 2 has been discharged and is missing downstream. An otherwise mandatory exit inspection immediately after or at the exit of the inspection device 11 could then be omitted, for example.

[0046] The integration of position data PD1 to PD5 and / or, if applicable, drive data AD1 to AD5 into time series 21 enables particularly error-resilient operation, for example, even if individual position data PD1 to PD5 are missing due to incorrect measurements. In this case, data gaps can be filled or otherwise compensated by analyzing time series 21, for example, through interpolation.

[0047] The generation of the images 6 and the resulting calculation of the position data PD1 to PD5 requires detection time and computing time, which cause a latency 22 for the described area monitoring as a whole. As schematically indicated in Figure 2, a time interval 23 between at least one imaging time, here TA5, and the time TS for controlling / triggering the container manipulator 8 is preferably greater than the latency 22.

[0048] For example, the ratio of the time interval 23 and the latency 22 is between 1.1 and 5. The starting point for the time interval 23 is preferably, but not necessarily, the last (available) imaging time TA5 of the time series 21. The latency 22 is device-dependent and can be measured and / or calculated in a known manner.

[0049] The time series 21 is preferably continuously supplemented. This means, for example, that a further (not shown) imaging time could lie between the imaging time TA5 and the triggering time TS, depending on the design and / or application of the computing unit used, possibly even after this, possibly even within the period of latency 22. This results from the detection time and the computing time for individual recordings 6 (images). The computing time for individual recordings 6 can overlap with the detection time for the immediately following recording 6. The described method and the described device 1 can optionally be used in a particularly advantageous manner if the time series 21 are evaluated using artificial intelligence.The input data for such calculations then serve, for example, as the position data PD1 to PD5 of the area-imaging sensor 5 and the evaluation unit 7 as well as the drive data AD1 to AD5 of the transport means 3. Corresponding output signals then include, for example, the control times TS (trigger times) for the container manipulator 8 and, if applicable, corresponding control times for the inspection unit 12 or its inspection sensor 13.

[0050] The time series 21 determined during operation can be stored across all systems at a central location, for example, in a data cloud. From there, the time series 21 can be used for machine training, validation, and / or quality testing both in the data-generating device 1 or the associated production system, as well as across other production systems, for example, during new installations or the like. Anomalies can also be analyzed using centrally stored time series 21 in order to train artificial intelligence systems particularly efficiently.

Claims

Claims 1. Method for production monitoring and in particular real-time control in a plant for the manufacture and / or product filling and / or packaging of containers (2) for beverages or similar products, wherein the containers are continuously transported through a stationary imaging area (4) of at least one planar imaging sensor (5), wherein position data (PD1 - PD5) of the containers are determined from images (6) of the containers generated thereby by means of electronic evaluation, and wherein times (TS) for a subsequent location-related mechanical manipulation and / or inspection of individual or all containers are predicted on the basis of the position data.

2. Method according to claim 1, wherein the sensor (5) images the containers (2) at least twice, in particular multiple times, offset in time in the stationary imaging area (4), and wherein from the images (6) of the containers obtained thereby, time series (21) are determined or supplemented by means of electronic evaluation at least with the position data (PD1 - PD5) of the respective container.

3. Method according to claim 1 or 2, wherein time intervals (23) between recording times (TA1 - TA5) of at least one, in particular the respective last recording (6) of the containers (2) and a control time (TS) triggering the manipulation and / or inspection are in particular by a factor of 1.1 to 5 greater than a latency (22) caused for imaging the containers and for determining the associated position data (PD1 - PD5).

4. Method according to at least one of the preceding claims, wherein the sensor (5) records overall views of the containers (2), in particular of at least two containers in each case.

5. Method according to at least one of the preceding claims, wherein the sensor (5) further detects status data of the containers (2) and the status data are assigned to the position data (PD1 - PD5), in particular in the form of time series (21).

6. Method according to one of the preceding claims, wherein drive data (AD1 - AD5) of a transport means (3) transporting the containers (2) between the sensor (5) and manipulation and / or inspection are further determined and assigned to the position data (PD1 - PD5) of the containers, in particular in the form of common time series (21).

7. Method according to at least one of the preceding claims, wherein the manipulation and / or inspection is carried out on the basis of the predicted control times (TS) and a discharge of individual containers (2) from a product flow, a distribution of the containers on partial product streams, and / or a correction of the location and / or relative transport position of individual containers.

8. Device (1) for production monitoring and in particular real-time control in a plant for the manufacture and / or product filling and / or packaging of containers (2) for beverages or similar products, comprising: a transport means (3) for the continuous transport of the containers; at least one sensor (5) which produces a planar image in a stationary imaging area (4) and for generating images (6) of the containers on the transport means; an electronic evaluation unit (7) for determining position data (PD1 - PD5) of the containers from the images and for predicting control times (TS) based thereon for subsequent location-related mechanical control of a container manipulator (8) and / or an inspection unit (12) for inspecting the containers.

9. Device according to claim 8, wherein the sensor (5) is designed to image the containers (2) at least twice, in particular multiple times, offset in time in the stationary imaging area (4), and wherein the evaluation unit (7) is designed to determine and / or supplement time series (21) with at least the position data (PD1 - PD5) of the respective container from the images (6) of the containers obtained thereby by means of electronic image evaluation.

10. Device according to claim 8 or 9, wherein the sensor (5) is arranged upstream of the container manipulator (8) and / or the inspection unit (12) and the evaluation unit (7) is designed such that time intervals (23) between recording times (TA1 - TA5) of at least one, in particular the last, recording (6) of the containers (2) and a triggering of the container manipulator and / or the inspection unit are greater than a latency (22) caused by individual images of the containers and the determination of the associated position data (PD1 - PD5), in particular by a factor of 1.1 to 5.

11. Device according to at least one of claims 8 to 10, wherein the sensor (5) is designed and arranged with respect to the transport means (3) to record overall views of the containers (2), in particular of at least two containers in each case.

12. Device according to at least one of claims 8 to 11, wherein the sensor (5) and the evaluation unit (7) are further designed to detect status data of the containers (2) and to assign the status data to the position data (PD1 - PD5), in particular in the form of time series (21).

13. Device according to at least one of claims 8 to 12, which is further configured to receive drive data (AD1 - AD5) of the sensor (5) extending at least from the imaging area (4) to determine the position data (PD1 - PD5) of the transport means (3) extending to the container manipulator (8) and to assign them to the position data (PD1 - PD5) of the containers (2), in particular in the form of common time series (21).

14. Device according to at least one of claims 8 to 13, wherein the container manipulator (8) is a pusher for diverting individual containers (2) from a product stream, a switch for distributing the containers to partial product streams or a guide / stop element for correcting the position and / or relative transport position of individual containers.

15. Inspection device (11) for inspecting containers (2) for beverages or similar products in a plant for the production and / or product filling and / or packaging of the containers, with the device (1) according to at least one of claims 8 to 14 and with an inspection unit (12) arranged between the imaging area (4) of the at least one planar imaging sensor (5) and the container manipulator (8), in particular for imaging quality control of the containers, wherein the container manipulator is in particular a pusher for diverting containers of inadequate quality.

Images