Method for operating a device, computer program product, and device for manufacturing a product

Abstract

A method for operating an apparatus for manufacturing a product is described, the method comprising the following steps: - acquiring at least one quality data set, which includes measured values ​​of one or more quality parameters, each of which corresponds to a characteristic of the product; - acquiring at least one related machine data set, which includes actual values ​​of a plurality of adjustable machine parameters of the apparatus; - temporally assigning the quality data set to the machine data set and generating a first data set including temporally correlated measured values ​​and actual values; - repeating the aforementioned steps at least once for generating at least one second data set; - determining a correlation between one or more quality parameters and machine parameters by comparing the acquired data sets and creating a control model for the apparatus, and - based on the rated values ​​of one or more quality parameters, providing a corresponding rated value for at least one of the adjustable machine parameters according to the control model.

Description

Technical Field

[0001] The invention relates to a method for operating a device for producing a product, in particular a device for forming hollow bodies or injection-molded parts, a computer program product and a device for producing a product. Background Art

[0002] Different methods for operating a device for manufacturing a product are known from the prior art. Typically, when operating a device for manufacturing a product, a number of parameters are set on the device before the actual production process. During the production process, these parameters can change by manual adjustment / readjustment or by changing the properties of the device itself during the process. Thus, for example, the behavior of the device changes depending on the environment in which it is operated or which material is processed in the device. External parameters, such as room temperature or external temperature or air humidity, have an influence on the device. The properties of the device itself can also change. For example, the device can heat up during the production process, which, for example, causes certain parts of the device to change in their length, or the operating medium, such as hydraulic oil, etc., has a viscosity that changes during operation. All these changes have an impact on the product to be manufactured. Therefore, it is necessary to constantly monitor the properties of the product. In some processes, for example, it is sufficient to check once every 100 products, while in other processes, every product must be checked once.

[0003] This inspection process is time-consuming and expensive. In addition, not all products can be inspected non-destructively.

[0004] WO 2011 / 023155 A1 discloses a method for blow-molding containers, in which at least one property of the finished blow-molded container is calculated based on measured parameters characterizing the blow-molding process and compared with a setpoint value. Parameters influencing the blow-molding process are changed based on determined deviations. The relevant simulation is to be carried out with the aid of expert knowledge.

[0005] This known method is both expensive and imprecise. Providing expert knowledge requires a great deal of effort and is based on subjective judgements. Summary of the invention

[0006] The object of the present invention is to overcome one or more disadvantages of the prior art. In particular, a method and a device should be provided, which enable a device for manufacturing a product to be operated substantially automatically and to manufacture products of as consistent a quality as possible, and in particular to reduce the expenditure for testing the manufactured products.

[0007] This object is achieved by the device and method according to the invention. Further embodiments are described below.

[0008] A method according to the invention for operating a device for producing a product, in particular a device for forming a hollow body or an injection-molded part, comprises the following steps:

[0009] - collecting at least one quality data set, said quality data set comprising measured values ​​of one or more quality parameters, said quality parameters respectively corresponding to a characteristic of said product;

[0010] - recording at least one relevant machine data set, which comprises actual values ​​of a plurality, in particular all, of adjustable machine parameters of the device;

[0011] - assigning the quality data set to the machine data set in time and generating a first data set comprising temporally correlated measured values ​​and actual values;

[0012] - repeating the above steps at least once to generate at least one second data set;

[0013] - determining a correlation between the one or more quality parameters and the machine parameters by comparing the acquired data sets and creating a control model for the device, and

[0014] - Based on the desired value of the one or more quality parameters, a corresponding desired value is provided for at least one of the adjustable machine parameters by means of the control model.

[0015] Product quality parameters may be, for example, the material distribution in the finished product, the contour of the finished product, the wall thickness of the finished product, the weight of the finished product, the temperature distribution of directly demoulded products or also the color and surface properties of the finished product.

[0016] The product may also be an intermediate product. Correspondingly, the quality parameter corresponding to the product characteristic may also be a parameter of the intermediate product, such as the temperature of the plastic melt or the diameter of the hose preform during extrusion blow molding.

[0017] All these quality parameters are usually acquired by means of suitable sensors and provided as measured values, in particular automatically. However, it is also possible that quality parameters, such as color or surface properties, which can be indicators of surface quality, are evaluated by the machine operator and their magnitude is transmitted to the system as measured values ​​by manual input.

[0018] The machine parameters of the device can be, for example, the temperature of the tool, the temperature of the corresponding operating medium such as hydraulic oil or compressed air or cooling water, or in the case of a blow molding machine, for example, the temperature of the molding die, the volume of the blowing airflow or its temperature, the parameters of the stretching curve such as the stretching speed, the temperature of the preform, the corresponding mass pressure in the corresponding cavity, the speed of the screw conveyor, the volume, type or shape and temperature of the conveyed raw materials such as plastic particles. Similarly, for example, the electrostatic charge of the machine or the supplied plastic particles can be collected as a machine parameter. It goes without saying that not all of the machine parameters mentioned here can be directly adjusted. Therefore, for example, the shape of the particles can be changed only by replacing the supplied particles. Therefore, it can be provided that, in addition to the adjustable machine parameters, non-adjustable machine parameters are also collected.

[0019] All of these machine parameters can be detected or read out in particular by measurement technology and provided as actual values. Preferably, a plurality of machine parameters are detected simultaneously and provided in one machine data set.

[0020] In the context of the present invention, the relevant machine data set is the data set of the device used to manufacture the product from which the quality parameter is derived.

[0021] By collecting the quality data set, a plurality of measured values ​​of quality parameters of the product are also included. Preferably, a plurality of measured values ​​of a plurality of quality parameters are collected simultaneously and provided in the quality data set.

[0022] By acquiring a machine data set, a plurality of actual values ​​of the corresponding machine parameters are included, which can describe a time curve of the corresponding actual value.

[0023] The machine data set is temporally assigned to the quality data set, in other words linked to the quality data set, and a data set is generated which contains both measured values ​​and actual values. Preferably, the quality data set is hereby assigned to a machine data set which is relevant to the production of the product or the primary product and thus to the associated quality parameters.

[0024] In principle, it is not always necessary to record all quality parameters at a specific point in time; thus, for example, the weight of the finished product can always be recorded, while, for example, the surface quality can only be recorded for one tenth of the products. By comparing a number of such data sets, correlations between one or more quality parameters and one or more machine parameters can be determined. These correlations can be causal or indirect correlations, for example, linked to one another via a third variable.

[0025] Providing a setpoint value for at least one machine parameter according to the determined correlation based on the setpoint value of one or more quality parameters enables a precise control of the device for manufacturing the product. In particular, an inspection of the product after its manufacture may become superfluous, since by previously identifying the relationship between the quality parameter and the machine parameter, in other words determining the correlation between the respective machine parameter and the quality parameter, and by correspondingly adjusting the machine parameter, the result, i.e. the setpoint value of the quality parameter to be achieved, can be predicted.

[0026] Preferably, a plurality of desired values ​​for machine parameters are provided based on the desired value for a quality parameter, such as the wall thickness of the blow molded body. Preferably, all desired values ​​for machine parameters are provided which can influence the desired value in open-loop and / or closed-loop control technology.

[0027] It may also be provided that correlations between actual values ​​of a plurality of machine parameters are determined, since it is conceivable that a plurality of machine parameters also influence one another.

[0028] It is thereby possible to predict to what extent a change in one machine parameter influences another machine parameter and, subsequently, to what extent a quality parameter is also likely to change.

[0029] In addition, at least one environmental data set can be acquired. The environmental data set comprises measured values ​​of one or in particular a plurality of environmental parameters. The environmental data set is temporally assigned to the machine data set and thus forms a component of the corresponding acquired data set.

[0030] By acquiring the environmental data set, it is possible to additionally acquire the influence of the environmental conditions on the corresponding quality parameter and include it in the correlation and thus expand the control model in particular with one or more boundary conditions.

[0031] Environmental parameters may include parameters such as air temperature or air humidity, but may also include air pressure, time of day (in particular day or night), the geographical location of the machine and / or of the factory in which the machine is operated (and correspondingly the climatic conditions associated therewith), or parameters about the state of the factory floor in which the machine is located, such as open or closed doors or windows, which may, for example, indicate some information about air movements, such as air currents.

[0032] Thus, for example, during a blow-molding process, the ambient air temperature may have an effect on the cooling rate of the product, which in turn may have an effect on the corresponding wall thickness of the product, for example when the material solidifies more slowly at certain points.

[0033] It goes without saying that a plurality of mutually different combinations of desired values ​​for the machine parameters may be suitable for achieving a specific desired value for the quality parameter.

[0034] Depending on boundary conditions, such as the ambient temperature, one or more machine parameters must be set differently. These relationships are acquired by determining the correlation between measured values ​​and actual values ​​and determining the measured values ​​of the ambient parameters with the ambient parameters as boundary conditions. These relationships are then provided in the control model.

[0035] The data set can be determined at least once based on test results of a device for producing a product and provided for creating a control model for the device.

[0036] This enables the data sets to be provided in a specific form. It is thus conceivable that individual machine parameters remain constant for a certain period of time and that other machine parameters vary within this period of time. Over the length of this period of time, one or more quality parameters can be measured on the product and, correspondingly, the measured values ​​of the quality parameters and the corresponding actual values ​​of the machine parameters can be transferred to the corresponding quality data set or machine data set.

[0037] Additionally or alternatively, it can be provided that a plurality of measured values ​​and actual values ​​are collected from a production facility having a device for manufacturing a product and provided, for example, in a database. A plurality of quality data sets and machine data sets can be created from these measured values ​​and actual values, and a plurality of data sets can be generated from these quality data sets and machine data sets. By comparing these plurality of data sets, the determination of correlations is simplified and the corresponding correlations are more precise. This enables the creation of more precise control models.

[0038] Thus, a data set can be determined by collecting a plurality of measured values ​​and actual values ​​from a production facility and provided for use in creating a control model for the device.

[0039] The method steps according to the invention substantially correspond to a learning phase.

[0040] Through this learning phase, a very close connection is created between machine parameters, quality parameters and environmental parameters, which in turn leads to a very accurate relationship model and thus to a very accurate control model.

[0041] The provision of a setpoint value for at least one machine parameter described in the method can follow this learning phase.

[0042] The provision of the above-mentioned setpoint value or multiple setpoint values ​​for machine parameters can be carried out, for example, automatically or manually in the form of transferring these values ​​to the controller of the corresponding device for manufacturing the product. The controller can be part of a computer and can have one or more computer program products. After the transfer, the device for manufacturing the product can operate according to these regulations and perform the corresponding production process or manufacturing process.

[0043] It can be provided that actual values ​​of a plurality of the machine parameters and of at least one of the environmental parameters are acquired and, based on these actual values, one or more values ​​of the quality parameter corresponding to these actual values ​​are output according to a control model. The one or more values ​​are compared with the setpoint values ​​of the one or more quality parameters and deviations are determined. Based on this deviation, the device is then open-loop controlled, in particular closed-loop controlled, according to the control model, in particular taking into account the actual value of the at least one environmental parameter.

[0044] Alternatively or additionally, it can be provided that the actual value of the machine parameter is compared with the corresponding setpoint value. The device can be open-loop controlled, in particular closed-loop controlled, as a function of the deviation of the actual value of the machine parameter from the corresponding setpoint value.

[0045] It can be provided that only actual values ​​of machine parameters and / or environmental parameters are recorded when the device is adjusted. That is, no quality parameters are recorded. If the actual values ​​of the machine parameters correspond to the setpoint values ​​of the machine parameters, these quality parameters are automatically and correctly determined.

[0046] It can be provided that the corresponding value of a machine parameter, for example the speed of opening or closing the mold, is directly influenced. Other machine parameters cannot be adjusted directly. This is, for example, the viscosity of the operating medium. However, by means of a suitable cooling device, for example, the temperature can be influenced, which in turn has an influence on the viscosity. This influence can in turn be derived from one of the predetermined correlations.

[0047] Furthermore, there may be parameters that cannot be influenced in practice, such as environmental parameters that are naturally hardly changeable. In order to compensate for their deviations from the desired values, the relevant machine parameters can be correspondingly open-loop controlled or closed-loop controlled.

[0048] The deviation of the actual value from the corresponding desired value of the respective machine parameter means that the corresponding desired value of the quality parameter used to determine the desired value of the machine parameter also deviates. Therefore, an open-loop or closed-loop control of the corresponding machine parameter or, if appropriate, one or more machine parameters having a corresponding influence on and / or being dependent on the machine parameter enables the corresponding desired value of the quality parameter to be achieved.

[0049] The respective actual values ​​of the machine parameter and of the at least one environmental parameter can be detected continuously during operation of the device.

[0050] This enables continuous monitoring of quality parameters.

[0051] Alternatively, it can be provided that the actual values ​​of the respective machine parameters and of the at least one environmental parameter are recorded at predeterminable time intervals (cyclically) during operation of the device.

[0052] This is particularly advantageous in the case of parameters that change only slowly or have only a very small influence on the quality parameter. The amount of data and the effort for recording the corresponding machine parameters can thus be reduced.

[0053] In a preferred form of the method, the dependencies, i.e. the control model, are only transferred to the device once. Due to the very close connection of the machine parameters, the quality parameters and the environmental parameters, the quality parameters can be specified with fixed machine parameters without having to measure the quality parameters again. In other words, the device can be adjusted to the quality parameters to be achieved and the device selects the corresponding machine parameters based on the control model. As soon as the measured machine parameters, i.e. the actual values, agree with the selected machine parameters, i.e. the setpoint values, it can be assumed that the actual values ​​of the quality parameters correspond to the selected setpoint values ​​of the quality parameters.

[0054] The invention therefore also relates to the operation of an apparatus, in particular a method for operating an apparatus for producing products, in particular an apparatus for forming hollow bodies or injection-molded parts, wherein the correlations obtained during a learning phase using the method according to the invention are stored statically, in particular as a static control model, in the apparatus. In this case, the corresponding setpoint values ​​of the machine parameters are selected and adjusted based on the desired setpoint values ​​of the quality parameters.

[0055] However, it can also be provided that the nominal parameters are entered in an external database and fed into the device from the external database. This can be achieved, for example, manually by means of input by a machine operator. Automatic transmission, for example via an electronic interface, is also possible.

[0056] It may be provided that an actual value of at least one of the quality parameters is cyclically compared with a desired value of the quality parameter.

[0057] These comparisons are particularly advantageous when the device for manufacturing the product is in test operation, in other words during a learning phase. Through these comparisons, it can be determined whether the determined correlation is consistent with practice and / or further measured values ​​can be obtained.

[0058] Preferably, the control model can be determined by machine learning based on one or more correlations between one or more quality parameters and machine parameters taking into account at least one quality parameter. In particular, one or more correlations can be determined by means of an artificial intelligence implementation, in particular using a neural network.

[0059] This allows relationships, i.e. correlations, to be determined and identified, independently of whether there is a direct or causal relationship between the individual measured values ​​and the actual values. By using machine learning, far-reaching and / or higher-level patterns of multiple individual measured values ​​and actual values ​​can be identified and / or compared to each other and / or correlated with each other.

[0060] The machine data set may include a plurality of machine parameters, in particular actual values ​​of a plurality of machine parameters. In this case, each machine parameter may be assigned a weight of its influence on each quality parameter according to its correlation with one or more quality parameters of the quality data set. Thus, the control model has weights for the individual machine parameters.

[0061] That is, the greater or better the correlation between a machine parameter and a quality parameter, the greater the influence of a change in the corresponding machine parameter on the corresponding quality parameter. Accordingly, it can be provided that a correspondingly high weight is assigned to such a machine parameter, in other words a corresponding correlation between a corresponding quality parameter and a corresponding machine parameter.

[0062] However, if the machine parameter has an influence, which can also be negative, on another quality parameter of the product, for example, it can be provided that a slightly lower weight is assigned to the machine parameter. That is to say, the machine parameter which has the greatest influence on the respective quality parameter is not necessarily the machine parameter with the highest weight. The highest weight can, for example, be a machine parameter which, although has a correspondingly large influence on a certain quality parameter of the product, at the same time has a very small influence on other quality parameters and / or on other machine parameters.

[0063] It may also be provided that a plurality of weightings are determined, for example with respect to rapid changes and with respect to the smallest possible influence on other machine parameters or quality parameters.

[0064] In order to achieve a desired value for one or more of the quality parameters, one or more of the corresponding machine parameters can be open-loop controlled or closed-loop controlled according to a control model with corresponding weightings with respect to the one or more quality parameters.

[0065] Correspondingly, precise intervention and precise open-loop control or closed-loop control can be achieved.

[0066] In addition, it may be provided that a check is performed to determine whether one or more actual values ​​of the machine parameters or environmental parameters detected during the control of the device are within a value range of the machine parameters and / or environmental parameters detected during the creation of the control model.

[0067] If, for example, the ambient temperature is recorded as one of the ambient parameters in order to create a control model, then typically there is a maximum recorded value and a minimum recorded value. The minimum value represents the lower limit of the value range, the maximum value represents the upper limit. Thus, the value range in which the ambient parameter is recorded can extend, for example, from 15° to 30°. For actual values ​​outside this value range, there is no longer a value based on the measured data in the control model.

[0068] It can be set that, in the case where one or more of the actual values ​​of the machine or environmental parameters collected when the control device is outside the value range, it is checked whether the deviation of the corresponding actual value occurs once or repeatedly, in particular within a specific time period, and a fault signal can be output if the corresponding deviation occurs repeatedly.

[0069] This enables the machine operator or a higher-level controller to be informed that a process parameter is outside the range determined by the measurement. However, the fault message can likewise be evaluated as an indication that the measuring sensor or measuring transducer is defective.

[0070] Additionally or alternatively, if one or more of the actual values ​​of the machine or environmental parameters detected during the control of the device are outside the value range, it can be checked whether the deviation from the value range is significant. If the deviation is significant, a fault signal can be output.

[0071] This prevents the output of correspondingly unnecessary fault signals even in the event of deviations from a value range which has little or no effect on the quality parameter.

[0072] A significant deviation may be present when the actual value of the deviation relates to a machine parameter which is set with a high weighting for the setpoint value of one or more quality parameters, or when the deviation has exceeded an in particular adjustable threshold value.

[0073] It can be provided that the control model switches off the device or at least requires manual control intervention if the check shows that the deviation is significant or that the deviation occurs repeatedly.

[0074] This prevents overproduction of products with different quality parameters and / or at least requires corresponding checks and instructions by an operator.

[0075] Another aspect of the present invention relates to a computer program product. The computer program product comprises instructions which, when executed on a computer, cause the computer to execute the steps of the method of the present invention.

[0076] This makes it possible to implement the method according to the invention in an open-loop control and / or in a closed-loop control of a device for producing a product.

[0077] Another aspect of the present invention relates to a device for manufacturing a product, in particular a device for forming a hollow body, comprising a computer program product according to the present invention.

[0078] This makes it possible to provide a device for manufacturing products on which not only a learning process / learning phase but also a production process / production phase can be carried out. In other words, a device designed according to the invention makes it possible to determine all dependencies between the device itself and the product to be manufactured and subsequently makes it possible to dispense with a comprehensive check of the respectively manufactured product in the subsequent production process.

[0079] The device according to the invention can be a component of a device for injection molding or blow molding. For example, the device is an extruder, a dryer for raw materials for manufacturing products, an extrusion blow molding device or a stretch blow molding device. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The invention is further explained below with reference to schematic diagrams which only illustrate exemplary embodiments.

[0081] Figure 1 : Schematic diagram of an apparatus for manufacturing a product;

[0082] Figure 2 : Flowchart of the method according to the invention. DETAILED DESCRIPTION

[0083] Figure 1 A simplified schematic diagram of a device for manufacturing a product is shown, which in the present invention is configured, for example, as a blow molding machine 100. The blow molding machine 100 has a feed device 110, a blow molding area 120 and a discharge device 130. The preforms are fed into the blow molding machine in a known manner by means of the feed device 120. In the blow molding area 120, they are inflated with compressed air and stretched by means of a stretch rod. The inflated containers are collected and / or removed in the discharge device 130. This manufacturing process is known per se and is therefore not explained in detail here.

[0084] Figure 1 A controller 200 is also shown. The controller 200 is connected to sensors arranged on the blow molding machine 100 via a connection 102, which is indicated here by a dashed line. The controller 200 can be implemented as a separate unit, but is usually an integrated component of the device. The sensor can be a temperature sensor, a metronome, a position sensor, etc. The sensor can be used to acquire the actual value of the machine parameter. The sensor is only shown schematically here. The temperature sensor 101 is shown as a placeholder for a plurality of sensors.

[0085] In the embodiment shown, the connection 102 between the controller 200 and the blow molding machine is wired via a cable. However, these connections can also be realized wirelessly or via optical waveguides.

[0086] In addition, Figure 1, a human-machine interface (HMI unit), namely an operating unit 103, is shown. Via this operating unit 103, a machine operator can monitor the device and, for example, predetermine desired values ​​for quality parameters.

[0087] Figure 2 A schematic flow chart of a method for operating a device for manufacturing a product is shown, for example, Figure 1 A method of blow molding a machine 100 is provided.

[0088] In a first step, a quality data set 10 is acquired. The quality data set 10 comprises, for example, a plurality of measured values ​​of the wall thickness of the container.

[0089] In a second step, a machine data set 20 is acquired. The machine data set 20 comprises the actual value of the temperature sensor 101 (see Figure 1 ).

[0090] The quality data set 10 and the machine data set 20 are acquired simultaneously. Accordingly, the quality data set 10 can be temporally assigned to the machine data set 20 and can form a first data set 30. In the first data set 30, measured values ​​and actual values ​​are temporally correlated.

[0091] This time dependency is further explained with reference to the following simplified example. At time t1, for example, the wall thickness D1 of the first container B1 in the discharge device 130 is measured (see Figure 1 ). The wall thickness D1 corresponds to the mass parameter. At the same time, in the blow molding area 120 (see Figure 1 ) is measured in the cavity of the second container B2 which is inflated at the same time t1. The temperature K1 corresponds to the machine parameter. The feed device 110 (see Figure 1 ) The temperature V1 of the preforms in the third container B3 in the embodiment of the present invention is as follows. The temperature V1 can be treated as a machine parameter or as a quality parameter. In this example, the temperature V1 is treated as a machine parameter.

[0092] In the next step, all containers B1, B2 and B3 continue to move one station. That is, the preform of the third container B3 is further moved from the feed device 110 to the blow molding area 120, and the inflated second container B2 is further moved from the blow molding area 120 to the discharge area 130, while the first container B1 is taken out of the discharge area 130. A new preform of the fourth container B4 is provided in the feed device 110.

[0093] At time t2, the wall thickness D2 of container B2 is now measured. The preform of container B3 is located in the cavity for filling container B3, where the temperature K2 is measured. At the same time, the temperature V2 of the new preform of container B4 is measured. These processes are now repeated for the other containers B2, B3 to Bn.

[0094] That is, the temperature K1 of the container B2 is measured at the time t1, but the effect of this temperature K1 on the wall thickness D2 of the container B2 is not measured until the time t2. In other words, for the measured value of the quality parameter, in this example D2, there are a plurality of actual values ​​which are located at different distances in time before the acquisition time of the measured value of the quality parameter. These time correlations are created in the first data set 30.

[0095] After the creation of the first data set 30 , the second data set 30 ′ is created. This includes the acquisition of the second machine data set 20 ′ and the acquisition of the second quality data set 10 ′ and the corresponding generation of the second data set 30 ′ as described above.

[0096] The quality data sets 10 and 10' can include further quality parameters, which are preferably all acquired simultaneously. Thus, in addition to the wall thickness, a further quality parameter can be, for example, the opacity of the container wall or the concentricity of the closure relative to the container bottom.

[0097] The machine data sets 20 and 20 ′ can also contain measured values ​​of further machine parameters.

[0098] In a next step, the first data set 30 and the second data set 30 ′ are combined and a correlation 40 between the measured values ​​and the actual values ​​is determined.

[0099] Based on these dependencies, a desired value 50 of the acquired machine parameter can be determined for each desired value of the quality parameter or, in the case of a plurality of machine parameters in the corresponding machine data set 20, 20', a desired value 50 can be determined for each acquired machine parameter. As described above, there can be a plurality of combinations of desired values ​​50 of machine parameters which lead to the same desired value of the quality parameter. For the case where, for example, a machine parameter cannot be changed (or where necessary an unchangeable environmental parameter is included in the method), the desired value must be determined based on two fixed values ​​(the desired value of the quality parameter and the unchangeable value of the device / environment), which reduces the possible combinations. These relationships are stored in a control model for the device or a control model for the device is created based on these relationships.

[0100] Since a plurality of first data sets 30, second data sets 30' (shown as data set 30") are suitably generated, machine learning is used to determine correlations 40. This enables multiple values ​​to be compared with each other and similarities or patterns etc. to be determined and associated with each other even if the relationship is no longer apparent.

[0101] In a subsequent step of the method, a desired value for the quality parameter is selected and the corresponding desired value 50 for the machine parameter is transmitted to the controller 200 (see Figure 1), in order to operate the blow molding machine 100 accordingly (see Figure 1 ). Here, the control model can be transferred to the controller, or only the corresponding values ​​of the machine parameters determined according to the control model can be transferred to the controller. Once the blow molding machine 100 (see Figure 1 ) deviate from the setpoint values ​​50, the controller can readjust these actual values. If, for example, a value that cannot be influenced by the controller (e.g. the external temperature) changes, the setpoint value 50 can be adjusted according to the presetting of the corresponding uninfluenceable value. For this purpose, the setpoint value 50 is anchored in a data matrix, in particular in a control model, which is transmitted to the controller 200 as part of the setpoint value 50 of the quality parameter (see Figure 1 ).

Claims

1. A method for operating a device for manufacturing a product, the method comprising a learning phase and a production phase, so that a comprehensive inspection of the manufactured product can be omitted in the production phase, wherein the learning phase comprises at least the following steps: - collecting at least one quality data set (10, 10'), said quality data set comprising measured values ​​of one or more quality parameters, said quality parameters respectively corresponding to a characteristic of said product; - recording at least one relevant machine data set (20, 20'), which comprises actual values ​​of a plurality of adjustable machine parameters of the device, wherein the measured values ​​and the actual values ​​are recorded simultaneously; - temporally assigning the quality data set (10, 10') to the machine data set (20, 20') and generating a first data set (30) comprising temporally correlated measured values ​​and actual values; - repeating the above steps at least once to generate at least one second data set (30'); - determining a correlation (40) between the one or more quality parameters and the machine parameters by comparing the acquired first data set (30) and the second data set (30'), and creating a control model for the device, wherein the production phase comprises the following steps: - providing a corresponding desired value (50) for each of the adjustable machine parameters by means of the control model based on the desired value of the one or more quality parameters, It is characterized in that actual values ​​of a plurality of the machine parameters are acquired, the actual values ​​of the machine parameters are compared with corresponding setpoint values ​​(50), and the device is open-loop controlled or closed-loop controlled according to a control model depending on the deviation of the actual values ​​of the machine parameters from the corresponding setpoint values ​​(50) of the machine parameters.

2. The method according to claim 1, characterized in that The device is a device for forming hollow bodies or injection-molded parts, and the machine data record comprises the actual values ​​of all adjustable machine parameters of the device.

3. The method according to claim 1, characterized in that In addition, at least one environmental data set is acquired, which includes measured values ​​of one or more environmental parameters, and is temporally assigned to the machine data set (20, 20') in order to thereby form a component of the acquired first data set (30) and second data set (30').

4. The method according to any one of claims 1 to 3, characterized in that The first data set (30) and the second data set (30') are determined at least once based on test results of a device for manufacturing a product and are provided for creating a control model for the device.

5. The method according to claim 1, characterized in that The first data set (30) and the second data set (30') are determined by collecting a plurality of measured values ​​and actual values ​​from a production facility and are provided for creating a control model for the device.

6. The method according to claim 1, characterized in that Actual values ​​of a plurality of the machine parameters and at least one environmental parameter are collected, and based on these actual values, one or more values ​​of the one or more quality parameters corresponding to these actual values ​​are output according to the control model, and the one or more values ​​are compared with the rated values ​​of the one or more quality parameters, and deviations are determined, and based on the deviations, the device is open-loop controlled or closed-loop controlled according to the control model.

7. The method according to claim 6, characterized in that The actual value of the machine parameter is compared with a respectively associated setpoint value (50), and the device is closed-loop controlled as a function of the deviation of the actual value of the machine parameter from the setpoint value (50).

8. The method according to claim 1 or 6, characterized in that: The actual values ​​of the machine parameter and of the at least one environmental parameter are continuously detected during operation of the device.

9. The method according to claim 1 or 6, characterized in that: During operation of the device, actual values ​​of the machine parameter and of the at least one environmental parameter are recorded cyclically.

10. The method according to any one of claims 1 to 3, characterized in that The control model is determined by means of machine learning, taking into account the at least one quality parameter, based on a correlation between the one or more quality parameters and the machine parameter.

11. The method according to claim 1 or 6, characterized in that: The control model is determined by means of machine learning implemented by artificial intelligence, taking into account the at least one quality parameter, based on a correlation between the one or more quality parameters and the machine parameter.

12. The method according to claim 1 or 6, characterized in that: The control model is determined based on a correlation between the one or more quality parameters and the machine parameter by means of machine learning using a neural network, taking into account the at least one quality parameter.

13. The method according to claim 1 or 6, characterized in that: The machine data set (20, 20') comprises a plurality of machine parameters and, depending on its correlation (40) with one or more quality parameters of the quality data set, a weighting of its influence on each quality parameter is assigned to each machine parameter.

14. The method according to claim 13, characterized in that In order to achieve the desired value for one or more of the quality parameters, one or more of the corresponding machine parameters are closed-loop controlled according to the control model with corresponding weightings in relation to the one or more quality parameters.

15. The method according to claim 6, characterized in that In addition, it is checked whether one or more actual values ​​of the machine parameters or environmental parameters detected in the open-loop control are within a value range of the machine parameters or environmental parameters detected in the creation of the control model.

16. The method according to claim 15, characterized in that In the event that the actual value is outside the value range, it is checked whether a deviation of the actual value occurs repeatedly, and if a deviation of the actual value occurs repeatedly, a fault signal is output.

17. The method according to claim 15, characterized in that For the case where the actual value is outside the value range, it is checked whether the deviation from the value range is significant for the control model, when the deviation exceeds a threshold value, a significant deviation exists, and if the deviation is significant, a fault signal is output.

18. The method according to claim 17, characterized in that The threshold value is adjustable.

19. The method according to claim 15, characterized in that If the check reveals that the deviation is significant or recurring, the control model switches off the device or at least requires manual control intervention.

20. A computer program product comprising instructions which, when executed on a computer, cause the computer to perform the steps of the method according to any one of claims 1 to 19.

21. An apparatus for manufacturing a product, comprising a computer program product according to claim 20.

22. The device according to claim 21, characterized in that The device is a device for forming a hollow body.

23. The device according to claim 21 or 22, characterized in that The device is a component of a device for injection molding or blow molding.

24. The device according to claim 23, wherein the component is an extruder, a dryer for raw materials for producing a product, an extrusion blow molding device or a stretch blow molding device.

Images