System, method for operating a system, and use

Abstract

The invention relates to a system comprising a packing machine, a control device and a monitoring device, wherein the monitoring device comprises at least one sensor device and a computer device and wherein the packing machine comprises a filling machine with a filling nozzle. The computer device is designed and configured to check the data recorded by the sensor device for predetermined events on the basis of an evaluation based on an automated application. The automated application comprises an AI application and / or a computer vision application. In the method for operating a system of this type, a certain event is indicated on the basis of the evaluation of the data from the sensor device based on the automated application and a corresponding predetermined action is initiated. The invention also relates to the use of an automated application comprising an AI application and / or a computer vision application in the monitoring of events in a packing machine.

Description

[0001] System, method for operating a system and use

[0002] Description

[0003] The present invention relates to a system comprising at least one packaging machine, in particular for filling bulk material, at least one control device, and at least one monitoring device, wherein the monitoring device comprises at least one sensor device and at least one computing device. The packaging machine comprises at least one filling module with at least one filling nozzle. The present invention also relates to a method for operating such a system and to a use thereof.

[0004] Packaging machines, for example, for filling bulk goods into bags, are known in a wide variety of designs in the state of the art. For example, there are packaging machines with a single filling module with a filling spout, to which, depending on the design, an open-mouth bag, a valve bag, and / or a large bag (big bag) can be attached. Depending on the design, several filling modules can also be arranged in series. There are also rotating packaging machines in which several filling modules are distributed around the circumference of the packaging machine, with the bags attached to the filling spout(s) being filled while the packaging machine rotates.

[0005] Such packaging machines have many automated processes that monitor the filling process and general operation of the packaging machine. For example, the filling pressure can be used to check whether a bag is correctly attached to the filling nozzle and / or whether the bag is leaking or bursts or tears during the filling process. If a bag is leaking, more filling pressure than normal must be applied to maintain the desired filling pressure. Depending on the design, a sudden drop in pressure can also be detected, e.g. in the case of a bursting bag, and / or a slower increase in filling pressure in a bag that was already leaking. Depending on the design, a leak in a bag can also be detected by an unexpected increase in weight.

[0006] A disadvantage of this monitoring method, however, is that detecting a leak based on the filling pressure can sometimes take several seconds, depending on the design, e.g., 2 to 4 seconds. During this time interval from the onset of the leak to its detection, for example, when filling cement on a high-performance machine, several kilograms of cement can be pumped into the environment rather than into the bag, which can lead to significant contamination and enormous dust generation. This needs to be improved.

[0007] It is therefore the object of the present invention to improve the monitoring of processes on a packaging machine, in particular the detection of malfunctions, and to accelerate the reaction time.

[0008] This object is achieved by a system having the features of claim 1, by a method having the features of claim 6 and by a use having the features of claim 11. Preferred developments of the invention are the subject of the dependent claims. Further advantages and features of the present invention emerge from the general description and the description of the exemplary embodiment. The system according to the invention comprises at least one packaging machine, at least one control device and at least one monitoring device, wherein the monitoring device comprises at least one sensor device and at least one computing device. The packaging machine comprises at least one filling module with at least one filling nozzle.The computing device is suitable and designed to check the data acquired by the sensor device for predetermined events using at least one evaluation based on at least one automated application.

[0009] The packing machine is in particular a packing machine for filling bulk goods and / or generally dusty, flowable and / or pourable products.

[0010] Such a packing machine can, in particular, comprise at least one filling module with at least one filling spout. Multiple filling modules can also be provided in series. Depending on the design, the packing machine can also be designed as a so-called rotating packing machine with multiple filling spouts distributed around the circumference, so that, for example, a bag to be filled is suspended from a filling spout at one position, filled during rotation, and then ejected.

[0011] An automated application is in particular at least one processor-based application or a computer-based application which processes, further processes or analyses the data acquired by the sensor device.

[0012] According to the invention, the automated application is at least one AI application and / or at least one computer vision application and / or comprises at least one of these applications. Computer vision refers in particular to the analysis of images and videos using computer programs. The aim of these analyses is, in particular, to extract information contained in the image. Depending on the design, probabilistic methods, methods of image processing, projective geometry, deep learning, and computer graphics are preferably used.

[0013] In the area of ​​AI application, any application in the field of artificial intelligence can be used which can be advantageously used for monitoring in relation to packaging machines.

[0014] The use of machine learning (ML) is particularly advantageous here, with such machine learning being a subfield of artificial intelligence (AI). Central components of machine learning are algorithms and methods that enable programs to learn from data without being explicitly programmed. This allows these programs to make decisions, preferably independently. A distinction is made between supervised, unsupervised, and semi-supervised learning. Supervised learning methods include convolutional neural networks (also known as space-invariant artificial neural networks) and further developments of convolutional neural networks, such as region-based convolutional neural networks.These modifications of conventional neural networks are particularly suitable for image processing applications, since, depending on the design, one or more preprocessing steps are carried out before the neural network, including, for example, convolution operations depending on the application.

[0015] In particular, sensors (e.g. optical sensors such as a color or black / white camera, thermal image, etc.) are used to record data on specific conditions (e.g. bag breakage). These images and their conditions are then put into context by assigning the corresponding label to each condition (an image = a matrix of information). This can occur directly through interaction between the packing machine and the inspection device, e.g. if the packing machine detects a bag leak using conventional inspection methods that the inspection device itself did not recognize. A link is then created between the detected bag leak and the corresponding image, which further trains the AI ​​and makes detection more reliable. A bag breakage or pure leak detected by the inspection device can also be used as an additional data basis for training. Feedback learning and not feedback learning.

[0016] This declared information is then transferred / trained into a CNN (convolutional neural network), preferably using machine learning or AI methods or AI applications.

[0017] If, for example, a sensor observes the filling nozzle during a filling process, the data is continuously passed through the CNN, which then evaluates the condition directly and the status is further processed by the packaging machine or the control device (dust detected = packaging machine stops the filling process).

[0018] A sensor device can, for example, be and / or comprise a camera device, or, depending on the design, can also comprise or be designed as an acoustic sensor, a temperature sensor and / or at least one other suitable sensor.

[0019] The system according to the invention offers many advantages. A significant advantage is that, depending on the design, predetermined events can be detected much more quickly by the automated application based on the data acquired by the sensor device than by conventional monitoring functions. This results from the fact that, based on the data acquired by the sensor device, a predetermined event can be detected extremely quickly by the automated application, independent of other operating processes and parameters.

[0020] Preferably, the evaluation of the automated application is transferable to the control device. This makes it possible for the control device to directly control or regulate the filling process of the packaging machine, for example, upon detection of a predetermined event.

[0021] In expedient developments, the sensor device comprises at least one camera device. Depending on its design, such a camera device or camera can meet certain requirements. Depending on the design, for example, a conventional webcam may be sufficient. In this case, the image data from the camera device is preferably used for the evaluation.

[0022] In expedient embodiments, the sensor device is assigned to or in front of at least the filling module, whereby this is to be understood in particular that the detection range of the sensor device, for example a camera, can detect at least the filling nozzle.

[0023] Preferably, by means of such a configuration, it can be monitored, checked or detected, for example, whether a bag to be filled is correctly attached to a filling nozzle and / or whether a bag leak exists and / or develops during the filling process. However, type detection can also preferably be achieved based on the type of bag and it can be detected, for example, whether the valve is correctly opened when valve bags are used. Depending on the configuration, it can also be detected whether a filled bag has been correctly closed. In particular, the correct closure of a valve bag and / or an open-mouth bag can be checked. In particular, the correct closure of a valve bag and / or an open-mouth bag can be detected via the evaluation of conventional image data, preferably from a camera device such as a conventional webcam. This list of application cases is only an example.

[0024] The method according to the invention is suitable for operating a system as described above. Based on the evaluation of the data from the sensor device, a specific event is inferred using the automated application, and a predetermined action is subsequently initiated when a predetermined event is detected. The automated application is at least one AI application and / or at least one computer vision application and / or comprises at least one of these applications.

[0025] The same definitions of artificial intelligence and computer vision apply here as previously described for the system.

[0026] The method according to the invention also offers the advantages already described above for the system according to the invention.

[0027] The automated application preferably checks the data online. For example, a calculation process can be executed decentrally on multiple computing centers or on an online computing application based on the data to determine a predetermined event.

[0028] Particularly preferably, the automated application is trained and / or usable offline on the computing device. Depending on the design, the automated application can initially be trained online in order to have the broadest and greatest possible computing power available for training. Depending on the design, any additional functionality and further learning can also take place locally.

[0029] The automated application is further trained in appropriate training courses. This can be done online, offline, or using other methods and procedures.

[0030] Particularly preferably, the leakage of a bag is detected during the filling process. Based on the detection of such an event, the control device can, for example, stop the filling process. By using an automated application and in particular at least one AI application and / or at least one computer vision application, for example, the time span from the bag break until the packaging machine or the filling module or the filling process at the corresponding filling nozzle is switched off can be significantly shortened. With conventional methods, in which, for example, a bag break is detected by the drop in filling pressure of the packaging machine, the switch-off time can be shortened from sometimes several seconds to less than one second, preferably less than half a second and in particular to less, for example 1 / 10 of a second.

[0031] By means of the method according to the invention, it is preferably and by way of example also possible to monitor or check whether a bag to be filled is correctly attached to a filling nozzle and whether the bag is leaking. However, it is also preferably possible to achieve type recognition based on the type of bag (e.g. size, print image, material) and, for example, to detect whether the valve is correctly opened when valve bags are used. Depending on the design, it is also possible to detect whether a filled bag has been correctly closed. In particular, the correct closure of a valve bag and / or an open-mouth bag can be checked. The correct closure of a valve bag and / or an open-mouth bag can be detected in particular via the evaluation of conventional image data, preferably from a camera device such as a conventional webcam. This list of application cases is only exemplary.These monitoring operations can also be carried out very quickly using the method according to the invention.

[0032] The use according to the invention relates to the use of an automated application, in particular an application comprising at least one AI application and / or at least one computer vision application, in monitoring events on a packaging machine.

[0033] The use according to the invention also offers the advantages already described above for the system according to the invention and the method according to the invention.

[0034] The events are preferably monitored based on image data from at least one camera device and at least one sensor device.

[0035] Preferably, a system as described above is used. In particular, all features of the previously described system and / or method can also be integrated individually and / or in combination into the use according to the invention.

[0036] Particularly preferably, the event detected is the leakage of a bag (200) during filling and / or whether a bag (200) to be filled is correctly attached to the filling nozzle and / or whether a filled bag (200) has been correctly closed and / or the type of bag used for filling and / or whether the bag is correctly opened for attachment. Further advantages and features of the present invention emerge from the exemplary embodiment, which is explained below with reference to the accompanying figures.

[0037] The figures show:

[0038] Fig. 1 is a purely schematic representation of an embodiment of a system according to the invention in a perspective view;

[0039] Fig. 2 shows the embodiment according to Figure 1 in a supplemented perspective view;

[0040] Fig. 3 shows the embodiment according to Figure 1 in a further supplemented perspective view;

[0041] Fig. 4 shows the embodiment according to Figure 1 in another supplemented perspective view;

[0042] Fig. 5 is a purely schematic representation of a further embodiment of a system according to the invention in a front view;

[0043] Fig. 6 is a purely schematic representation of a next embodiment of a system according to the invention in a side view; and

[0044] Fig. 7 is a purely schematic representation of a further embodiment of a system according to the invention in a side view.

[0045] Figure 1 shows a purely schematic embodiment of a system 1 according to the invention, wherein the system 1 according to the invention in the embodiment shown here comprises a packaging machine 10, a control device 30 and a monitoring device 50.

[0046] In the exemplary embodiment shown here, the control device 30 is included in the packaging machine 10, or in the exemplary embodiment shown, the control device 30 controls the packaging machine 10 or the filling process.

[0047] In the embodiment shown here, the packaging machine 10 is designed as a so-called single-spout packaging machine, which only comprises a filling module 11 with a single filling spout 12.

[0048] In the embodiment shown, empty bags or bags 200 or, in this case, valve bags can be attached to this filling nozzle 12 and then filled with a free-flowing and / or pourable bulk material.

[0049] In the exemplary embodiment shown, the control device 50 comprises a sensor device 51 and a computing device 52. In the exemplary embodiment shown, the computing device 52 comprises a computer 53, wherein the computing device 52 or the computer 53 is connected to the Internet, i.e., is online, as indicated here in the exemplary embodiment shown.

[0050] In the embodiment shown, the sensor device 51 comprises a camera device 54, which is provided here by a conventional webcam 55.

[0051] However, in other embodiments not shown, the packaging machine 10 can also comprise a plurality of filling modules 11, which, depending on the configuration, can also be assigned to a rotating packaging machine. Even in such a configuration, only one control device 50 with only one sensor device 51 and only one computing device 52 can be provided. However, multiple control devices 50, sensor devices 51, and / or computing devices 52 can also be provided.

[0052] Figure 2 shows a purely schematic example of an application of the system according to the invention shown in Figure 1. In the embodiment shown, the control device

[0053] 50 to check whether a bag 200 is leaking when it is filled, which is detected by a cloud of dust or a development of dust in a certain area.

[0054] For this purpose, the detection area 56 of the sensor device 51 or, in this case, the camera device 54 is selected such that only the relevant image section is recorded and / or viewed or used. The detection area 56 can actually be a detection area 56 of the camera or a software-selected image section of the sensor device.

[0055] 51 or the camera device 54.

[0056] Figure 3 shows a purely schematic illustration of a leak in the bag 200 during filling, or rather, that the bag is generally leaky. This creates a cloud of dust in the upper area of ​​the detection zone 56.

[0057] Through the automated application or through the use of artificial intelligence or computer vision, such a bag leak can be detected extremely quickly, whereby the control device 30 can stop the filling process based on the event detected by the monitoring device 50, e.g., a bag break or a bag leak. This allows the filling process to be stopped significantly more quickly than with conventional methods. Figure 4 shows that a bag leak can occur and be detected at various locations. This, too, can be detected depending on the training of the automated application, the AI ​​application, or the computer vision application.

[0058] Figure 5 shows a similar embodiment to Figure 1, showing a similar application. In the embodiment shown here, the detection area 56 of the sensor device 51 or the camera device 54 is focused on the upper region of the bag 200 to be filled.

[0059] Figure 6 shows a further application example for monitoring an area of ​​a packaging machine 10 by means of an automated application or by means of a Kl application or by means of a computer vision application, which are executed on a control device 50 or on the computing device 52 of a control device 50.

[0060] The area of ​​the filling nozzle 11 is monitored here, whereby the correct position of the bag on the filling nozzle can be monitored depending on the training of the automated application or the AI ​​application or the computer vision application.

[0061] In Figure 7, the embodiment according to Figure 6 is shown again in a purely schematic manner, with a further application being shown here.

[0062] Here, the detection area 56 is selected such that an automatic bag opening plug 100 positioned adjacent to the packing machine 10 is detected. The bag opening plug 100 is then preferably also part of the system 1.

[0063] In the application shown here, for example, it can be checked whether the valve of a bag 200 in the firing channel 101 is correctly opened and / or whether the correct type of bag is provided by the bag opening plug 100.

[0064] The other applications mentioned in the description and many more can also be monitored using an automated application or an AI application or a computer vision application.

[0065] List of reference symbols

[0066] I System

[0067] 10 packing machine

[0068] II Filling module

[0069] 12 filling nozzles

[0070] 30 Control device

[0071] 50 Control device

[0072] 51 Sensor device

[0073] 52 computing device

[0074] 53 computers

[0075] 54 Camera setup

[0076] 55 webcams

[0077] 56 Detection area

[0078] 100 Sackauf plugs

[0079] 101 shot channel

[0080] 200 bags

Claims

Claims:

1. System (1) comprising at least one packaging machine (10), at least one control device (30) and at least one monitoring device (50), wherein the monitoring device (50) comprises at least one sensor device (51) and at least one computing device (52), and wherein the packaging machine (10) comprises at least one filling module (11) with at least one filling nozzle (12), characterized in that the sensor device (51) comprises at least one camera device (52) and that the computing device (52) is suitable and designed to check the image data captured by the camera device (52) for predetermined events using an evaluation based on at least one automated application, wherein the automated application comprises at least one AI application and / or at least one computer vision application.

2. System according to claim 1, wherein the evaluation of the automated application is transferable to the control device (30).

3. System according to one of the preceding claims, wherein the sensor device (51) is associated with the filling module (11).

4. System according to one of the preceding claims, wherein the control device (50) is suitable and configured to detect a leak in a bag (200) during filling and / or to detect whether a bag (200) to be filled is correctly attached to the filling nozzle and / or to detect whether a filled bag (200) has been correctly closed and / or to detect which type of bag is required for filling is used and / or determine whether the bag is correctly opened for hanging.

5. Method for operating a system (1) according to one of the preceding claims, characterized in that the control device (51) uses the automated application to infer a specific event based on the evaluation of the image data of the camera device (52) of the sensor device (51) and to initiate a correspondingly predetermined action, wherein the automated application comprises at least one Kl application and / or at least one computer vision application.

6. Method according to the preceding claim, wherein the automated application checks the data online.

7. Method according to one of the two preceding claims, wherein the automated application is trained and / or operates offline on the computing device.

8. Method according to one of the preceding claims 5 to 7, wherein the automated application is retrained.

9. Method according to one of the preceding claims 5 to 8, wherein the leakage of a bag (200) is detected during filling and / or wherein it is detected whether a bag (200) to be filled is correctly attached to the filling nozzle and / or wherein it is detected whether a filled bag (200) has been correctly closed and / or wherein it is detected which type of bag is used for filling and / or whether the bag is correctly opened for attachment.

10. Use of an automated application comprising at least one AI application and / or at least one computer vision application in the monitoring of Events on a packaging machine (10) based on image data from at least one camera device (52) and at least one sensor device (51).

11. Use according to the preceding claim, wherein a System according to one of claims 1 to 4 is used.

12. Use according to one of the two preceding claims, wherein the event detected is the leakage of a bag (200) during filling and / or wherein it is detected whether a bag (200) to be filled is correctly attached to the filling nozzle and / or wherein it is detected whether a filled bag (200) has been correctly closed and / or wherein it is detected which type of bag is used for filling and / or whether the bag is correctly opened for attachment.

Images