METHOD FOR CONTROLLING THE PRODUCTION OF A SHEET METAL COMPONENT AND METHOD FOR PRODUCING A SHEET METAL COMPONENT OR SEVERAL DIFFERENT SHEET METAL COMPONENTS

Description

Background of the invention

[0001] The invention relates to a method for controlling the production of a sheet metal component, comprising the steps Reading in CAD data from an initial design of the sheet metal component, parameterizing the CAD data so that parameter values ​​of the CAD data are obtained.

[0002] The invention further relates to a computer program comprising program instructions that, when executed by a computer, cause the computer to perform such a method. Finally, the invention relates to methods for producing a sheet metal component or several different sheet metal components.

[0003] When designing sheet metal components, it is often unclear in practice whether a design has potential for optimization. In addition to the original designer, additional experts are typically only consulted for optimization if the sheet metal component is to be manufactured in very large quantities.

[0004] From US 10,061,300 B1 a method is known in which a manufacturing request comprises a digital model of a physical object and is obtained from a computer, a set of features of the physical object is determined from the digital model, a predictive value for the manufacturing request is generated using a machine learning regression model and the set of features, and a non-deterministic answer is determined on the basis of the predictive value and a multi-criteria optimization model, wherein the non-deterministic answer contains a set of attributes of a manufacturing process of the physical object, and wherein the set of attributes fulfills a multi-criteria condition of the multi-criteria optimization model.

[0005] US 2020 / 0159886 A1 describes a method in which an optimal component design is generated based on an analysis of a plurality of similar component designs according to objectives and weightings specified by a user. Object of the invention

[0006] It is an object of the invention to enable the efficient production of sheet metal components. Description of the invention

[0007] This object is achieved according to the invention by a method for producing a plurality of different sheet metal components according to claim 1, and by a computer program according to claim 11.

[0008] According to the invention, a computer-implemented method for controlling the production of a sheet metal component is provided. The control method comprises the following steps: A) Reading in CAD data of an initial design of the sheet metal component, C) Parameterizing the CAD data so that parameter values ​​of the CAD data are obtained, D) Evaluating the optimizability of the initial design of the sheet metal component using artificial intelligence based on the parameter values ​​of the CAD data and outputting an evaluation result.

[0009] The procedural steps are generally carried out in the specified order.

[0010] A sheet metal component is understood, in particular, to be a component that is obtained or obtainable from at least one sheet of metal. Sheet metal refers to a semi-finished product whose thickness is significantly smaller (typically by at least a factor of 10) than its dimensions in spatial directions orthogonal to the thickness direction. The sheet metal component can be manufactured, in particular, by bending and welding from the at least one sheet of metal. It is understood that further processing can be performed on the sheet metal or sheet metal component.

[0011] In step A), CAD data from an initial design of the sheet metal component is read into a computer memory. The initial design was typically created by a designer. Alternatively, the initial design may have been created automatically.

[0012] Preferably, the CAD data is imported as 3D CAD data in step A). ​​In a subsequent step B), this 3D CAD data is then converted into 2D CAD data. This conversion allows a data format suitable for subsequent analysis to be obtained.

[0013] In step C), parameter values ​​of the CAD data are determined. For this purpose, the CAD data can be analyzed for specific properties and characteristic values ​​for these properties can be determined; these characteristic values ​​constitute at least some of the parameter values.

[0014] The parameter values ​​can be determined for workpiece parameters, preferably for one or more of the following parameters: Width, length, area, thickness, material, raw material, external waste, internal waste, total length of external contours, number of external contour elements, type of external contour elements, number of bending edges, number of contours, total length of internal contours, total length of bending lines, bend x bend angle, bend x type, bend x bend length, bend x orientation, range of bend orientation, number of welds, weld length, number of non-linear contours.

[0015] Alternatively or additionally, the parameter values ​​for source parameters can be determined, preferably for a name, especially of the commissioning company, the designer, and / or the component. This allows the artificial intelligence to consider the skills and experience of a designer or group of designers, as well as the purpose of a component (if the file name or component name provides clues to the purpose).

[0016] Alternatively or additionally, the parameter values ​​for metadata can be determined, preferably for one or more of the following parameters: File size, design software used, and / or design software version. This metadata can provide an indication of the presence or absence of certain typical defect patterns. The metadata parameters therefore indirectly relate to workpiece characteristics.

[0017] In step D), the optimizability of the initial design of the sheet metal component is evaluated using artificial intelligence based on the parameter values ​​of the CAD data. In other words, the optimization potential of the initial design is determined. For this purpose, the parameter values ​​are analyzed using artificial intelligence. Each individual parameter extracted in step C), as well as the combination of all parameters, can be analyzed. An evaluation result is output. The evaluation result can be displayed, for example, on a screen. Alternatively or additionally, the evaluation result can be written to a file and / or transmitted to a preferably autonomous design system. The evaluation result typically indicates a measure of optimizability. Evaluating optimizability can be performed with relatively little computational effort.In contrast, implementing the optimization described in the aforementioned prior art requires considerable effort. The method according to the invention makes it possible to use available resources for optimization only when the optimization promises to achieve a sufficiently significant advantage.

[0018] Optimization potential can be expressed as a numerical value, particularly a percentage, which can indicate the extent to which optimization potential exists with regard to a specific criterion. For example, it can be stated that the weld seam length can be reduced by an estimated 50%.

[0019] Advantageously, the evaluation of optimizability includes classification into a predetermined number of categories, preferably three. The categories can be, for example, "highly optimizable," "optimizable," and "not optimizable." The category into which the optimizability was classified is output as the evaluation result. By specifying fewer categories, a certain - unavoidable - uncertainty in the prediction of optimizability can be concealed. This categorization makes it easier for the user to select sheet metal components for optimization.

[0020] The artificial intelligence can be based on the following models: kNN, AdaBoost, Random Forest, preferably in the Extra Trees variant, logistic regression, MLP / neural network, SVC, and / or Gaussian Processes. The artificial intelligence can use a machine learning algorithm. The machine learning algorithm can be trained with labeled data containing expert know-how. For this purpose, CAD files can be manually evaluated by experts and divided, for example, into the three categories "highly optimizable," "optimizable," and "not optimizable." A semi-supervised learning method can be used to train the algorithm. Using a large amount of data, the algorithm learns to what extent a sheet metal component can be optimized under which parameter constellations.

[0021] Advantageously, a constellation of parameter values ​​is evaluated for evaluation in step D). In this respect, the constellation refers in particular to the relationships between two or more parameter values ​​and the presence of elements of a certain type, such as weld seams, bends, or internal contours. Evaluating the constellation of parameter values ​​can make it possible to identify sub-aspects of the sheet metal component that can be optimized, for example, optimization potential with regard to the weld seam length or the number and length of bends.

[0022] The scope of the present invention further includes a computer program comprising program instructions which, when the computer program is executed by a computer, cause the computer to carry out the method according to the invention described above.

[0023] Furthermore, a computer program product on which such a computer program is stored falls within the scope of the present invention. A computer program product is understood to mean a computer-readable storage medium containing the computer program in computer-readable form.

[0024] Furthermore, the present invention relates to a method for producing a sheet metal component. The manufacturing method comprises the following steps: A) Reading in CAD data of an initial design of the sheet metal component, C) Parameterising the CAD data so that parameter values ​​of the CAD data are obtained, D) Assessing the optimisability of the initial design of the sheet metal component using artificial intelligence based on the parameter values ​​of the CAD data and outputting an evaluation result, F) Designing an optimised design of the sheet metal component if the evaluation result reaches or exceeds a predetermined level of optimisability, G) Determining a finished design of the sheet metal component which corresponds to the optimised design if it was designed according to step F) and which otherwise corresponds to the initial design, H) Manufacturing the sheet metal component according to the finished design.

[0025] Steps A) to D) correspond to the control method according to the invention described above. For advantages and advantageous developments of these steps, reference is made to the above description.

[0026] If the initial design can be optimized to a sufficient extent, an optimized design of the sheet metal component is created in step F). The optimized design can be created manually or automatically.

[0027] If an optimized design has been developed based on the previously determined degree of optimizability, this is defined as the final design of the sheet metal component in step G). Otherwise, i.e., if the optimizability of the initial design does not reach the predetermined level according to the evaluation in step D), the initial design is defined as the final design. Thus, a complex optimization of the sheet metal component's design is only performed if it is expected that sufficiently significant advantages can be achieved.

[0028] The optimized design can be checked for optimizability by repeating steps A) to D), with the optimized design being considered the starting design for the next iteration. If necessary, an optimized design can be designed multiple times, especially until its optimizability falls below the predetermined level.

[0029] In the final step H), the sheet metal component is manufactured according to the final design. The manufacturing method according to the invention provides a balance between a simple design process without iterations (when the initial design offers little or no optimization potential) and improved manufacturability of the sheet metal component according to the optimized design, if it has been determined that the initial design has sufficient optimization potential. The use of artificial intelligence to evaluate the optimizability in step D) enables a reliable, fast, and resource-efficient process.

[0030] Finally, a method for producing a variety of different sheet metal components falls within the scope of the present invention. The manufacturing method for the many different sheet metal components comprises the following steps: A) Reading in CAD data of an initial design of the respective sheet metal component, C) Parameterizing the CAD data so that parameter values ​​of the CAD data are obtained, D) Evaluating the optimizability of the initial design of the respective sheet metal component using artificial intelligence based on the parameter values ​​of the CAD data and outputting an evaluation result.

[0031] Steps A) to D) correspond to the control method according to the invention described above, which is carried out here for each of the different sheet metal components. For advantages and advantageous further developments of these steps, please refer to the above description.

[0032] The manufacturing process for the many different sheet metal components also includes the following steps: e) Selecting those sheet metal components for which the evaluation result reaches or exceeds a predetermined level of optimizability in order to optimise their respective design, f) Designing an optimised design of the sheet metal components selected in step e), g) Defining a finished design for each of the different sheet metal components, whereby the finished design corresponds to the optimised design for the sheet metal components selected in step e) and to the respective initial design for the sheet metal components not selected in step e), h) Manufacturing at least one example of each of the different sheet metal components.

[0033] In step e), those sheet metal components are selected for optimization for which sufficiently significant benefits can be expected from optimizing the design. This can be considered based on how many copies of the respective sheet metal component are to be manufactured. In particular, with larger production runs, smaller optimization benefits may justify the optimization effort. If the evaluation of optimizability involves classification into a predetermined number of categories, it can be provided that the boundaries of the categories are defined depending on the number of units of a particular sheet metal component to be manufactured.

[0034] In step f), an optimized design is created for each of the previously selected sheet metal components. The optimized designs can be created manually or automatically.

[0035] If an optimized design has been developed for one of the sheet metal components based on the previously determined degree of optimizability, this design is defined as the final design of the respective sheet metal component in step g). Otherwise, i.e., if the optimizability of the initial design does not reach the predetermined level according to the evaluation in step d), the initial design is defined as the final design of the respective sheet metal component. Thus, a complex design optimization is only performed for those sheet metal components for which it is expected that sufficiently significant advantages can be achieved.

[0036] In the final step h), at least one copy of each sheet metal component is manufactured according to the respective finished design. The manufacturing method according to the invention provides a balance between a simple design process without iterations (for those sheet metal components whose initial designs offer little optimization potential) and improved manufacturability of the sheet metal components according to the optimized designs, once it has been determined that the initial designs have sufficient optimization potential. The use of artificial intelligence to evaluate the optimizability in step D) enables a reliable, fast, and resource-efficient process.

[0037] It can be provided that after step f) has been carried out for the selected sheet metal components, steps A) to f) are repeated, whereby the optimized design is used as the starting design for each repeat execution of these steps. In this way, the designs of the sheet metal components previously selected for optimization can be further improved. Preferably, steps A) to f) are repeated for the respectively selected sheet metal components until no more sheet metal components are selected when step e) is carried out. In this way, it can be achieved that an at least almost optimal design is available for each of the different sheet metal components. In this variant, increased resources are used for the design of designs and their verification in order to improve manufacturability.

[0038] Further features and advantages of the invention will become apparent from the description and the drawings. According to the invention, the features mentioned above and those further described can be used individually or in combination in any convenient way. The embodiments shown and described are not intended to be exhaustive, but rather are exemplary in nature for describing the invention. Detailed description of the invention and drawing

[0039] The invention is illustrated in the drawing and explained in more detail using exemplary embodiments. In the drawings: Fig. 1 shows a schematic flow diagram of a method according to the invention for controlling the production of a sheet metal component, in which the optimizability of an initial design of the sheet metal component is determined using artificial intelligence on the basis of parameter values ​​extracted from CAD data of the initial design; Fig. 2 shows a schematic flow diagram of a method according to the invention for producing a sheet metal component, wherein, before the sheet metal component is manufactured, the optimizability of its initial design is assessed by artificial intelligence and, if the optimizability is sufficiently high, an optimized design is designed; Fig.3 a schematic flow diagram of a method according to the invention for producing a plurality of different sheet metal components, wherein, before the sheet metal components are manufactured, the optimizability of their respective initial design is assessed by artificial intelligence and, if the optimizability is sufficiently high, an optimized design is designed for the corresponding sheet metal components.

[0040] Figure 1 shows a flowchart of a computer-implemented method for controlling the production of a sheet metal component. First, an initial design of the sheet metal component is created in one step 100 in the form of CAD data. The CAD data of the initial design are converted 102into a memory of a computer on which the method is running. Preferably, the CAD data are provided and read in as 3D CAD data, for example in .step format. The read 3D CAD data can be assigned to a step 104 converted into 2D CAD data, for example, into the .geo format. If the CAD data of the original design is already provided and imported as 2D CAD data, the step of converting to 2D CAD data can be omitted. Alternatively, the process can also be performed with 3D CAD data.

[0041] The imported and, if necessary, converted CAD data of the initial design are processed in one step 106 parameterized so that parameter values ​​of the CAD data are obtained. The parameter values ​​can indicate one or more of the following parameters of the original design: Name, width, length, area, thickness, material, raw material, external waste, internal waste, file size, total length of external contours, number of external contour elements, type of external contour elements, number of bending edges, number of contours, total length of internal contours, total length of bending lines, bend x bend angle, bend x type, bend x bend length, bend x orientation, range of bend orientation, number of welds, weld length, number of non-linear contours.

[0042] Then, in one step 108The optimizability of the initial design is evaluated by artificial intelligence based on the previously determined parameter values. To evaluate the optimizability, the artificial intelligence can, in particular, evaluate a combination (constellation) of parameters. The artificial intelligence can be a neural network that has been trained in a semi-supervised learning process using example sheet metal components whose optimizability has been assessed by experts. The evaluation is preferably carried out by classifying the optimizability of the initial design into one of several predefined categories (classes) of optimizability. Preferably, three categories are considered. The initial design can, for example, be classified into the categories "highly optimizable," "optimizable," and "not optimizable."

[0043] The evaluation result from step 108 is processed in a step 110Output can be done by displaying the results on a screen. Alternatively or additionally, the evaluation result can be written to a file. In particular, a CAD file of the initial design can be assigned the degree of optimizability as an additional attribute.

[0044] Steps 102 to 110 are executed by a computer program running on a computer. The computer program includes program instructions that cause the computer to execute steps 102 to 110 when executing the computer program. The computer program with the program instructions can be implemented on a computer program product 10be stored. The computer program product 10 can, for example, be a memory card or a hard disk. The computer program can be downloadable via the Internet. Preferably, the execution of the computer program can be started via the Internet. In particular, it can be provided that the computer program runs on a server that reads in the CAD data of the initial design from a user PC. In this case, the evaluation result is typically transmitted back to the user PC in step 110 and can be further disseminated, in particular displayed, via the latter and / or its peripheral devices.

[0045] Figure 2 shows a flow chart of a method for manufacturing a sheet metal component using the control method of Figure 1 . In the manufacturing process, the steps 100 to 110 described above are first carried out. The evaluation result obtained is then 114checks whether the optimizability of the initial design reaches or exceeds a predetermined level of optimizability. If this is the case, in a step 116 An optimized design for the sheet metal component is designed. The optimized design can then be 118 as a finished design of the sheet metal component. If the evaluation of the initial design has shown that it cannot be optimized to a sufficient extent, the initial design is defined as the finished design of the sheet metal component in step 118. In a final step 120 the sheet metal component is manufactured according to the previously defined finished design.

[0046] Figure 3 shows a flow chart of a process for manufacturing several different sheet metal components. In Figure 3The method is illustrated by way of example for three different sheet metal components. The method is based on the at least partially temporally overlapping execution of the previously described steps 102 to 118 in process strands 12, 14 and 16 for each of the different sheet metal components.

[0047] After performing steps 100 to 110 for the different sheet metal components, in one step 112 Based on the respective queries 114, those sheet metal components are selected for design optimization for which a sufficiently high degree of optimizability was determined. When determining the degree of optimizability to be applied, the number of copies of the respective sheet metal component to be manufactured can be taken into account. Step 112 is preferably performed simultaneously for all different sheet metal components to be manufactured within the framework of the process.

[0048] For the sheet metal components not selected (shown here as an example in process sequence 16), the initial design is defined as the finished design in a respective step 118. For the selected sheet metal components (shown here as an example in process sequences 12 and 14), an optimized design is designed in a respective step 116. It can be provided that the optimized design is used as previously in Figure 2 described when the finished construction is determined, compare dashed arrows between steps 116 and 118 in Figure 3Alternatively, it can be provided that the optimized design is examined for its optimizability by repeating steps 102 to 110, wherein the optimized design is considered the initial design during the repeat run, compare the return arrows from steps 116 to steps 102. In this case, it can be provided that an optimized design is designed as often as necessary until none of the sheet metal components is selected for optimization in steps 112 and 114. Alternatively, the number of optimization loops can be limited to a predetermined value. The optimized design is finally defined in a respective step 118 as the finished design for the respective sheet metal component. It is understood that the iterative optimization of the design of the sheet metal component also applies to the method for manufacturing a (single) sheet metal component according to Figure 2 may be provided.

[0049] After determining the respective finished construction, at least one copy of each of the different sheet metal components is produced in a respective step 120. List of reference symbols

[0050] computer program product 10 Procedural strands 12, 14, 16 Provide 100 Reading CAD data from an initial design 102 Converting the CAD data 104 Parameterize the CAD data 106 Evaluate the CAD data 108 Output of optimizability 110 Select an assessment result 112 of sheet metal components query 114: Does the optimizability reach / exceed a predetermined level? 116 an optimized design 118 a prefabricated construction 120 of the sheet metal component

Claims

1. A method for producing a plurality of different sheet metal components, wherein the method comprises the following steps for each of the different sheet metal components: A) importing (102) CAD data of an initial design of the respective sheet metal component, C) parameterising (106) the CAD data so as to obtain parameter values for one or more of the workpiece parameters of the CAD data, D) evaluating (108) the optimisability of the initial design of the respective sheet metal component with regard to a weld length or a number and length of bends by means of artificial intelligence based on the parameter values of the CAD data and outputting (110) an evaluation result, and wherein the method further comprises the following steps: e) selecting (112) those sheet metal components for which the evaluation result has reached or exceeded a predetermined degree of optimisability, in order to optimise their respective design, f) drafting (116) an optimised design of the sheet metal components selected in step e), g) determining (118) a final design for each of the different sheet metal components, wherein the final design for the sheet metal components not selected in step e) corresponds to the respective initial design and wherein the final design for the sheet metal components selected in step e) corresponds to the optimised design, h) manufacturing (120) at least one copy of each of the different sheet metal components in accordance with the final design.

2. The method according to claim 1, characterised in that when selecting the sheet metal components in step e), the number of copies of the respective sheet metal component to be manufactured is taken into account.

3. The method according to claim 1 or 2, characterised in that after step f) has been carried out for the selected sheet metal components, steps A) to f) are repeated, wherein the optimised design is used as the initial design for the renewed execution of these steps.

4. The method according to claim 3, characterised in that steps A) to f) are repeated for the respective selected sheet metal components until no more sheet metal components are selected when step e) is carried out.

5. The method according to any one of claims 1 to 4, characterised in that the CAD data are imported as 3D CAD data in step A) and are converted (104) into 2D CAD data in step B).

6. The method according to any one of claims 1 to 5, characterised in that in step C) parameter values are determined for original parameters, preferably for a name, in particular of the commissioning company, the designer and / or the component.

7. The method according to any one of claims 1 to 6, characterised in that parameter values for metadata are determined in step C), preferably for one or more of the following parameters: - file size, - design software and / or design software version used.

8. The method according to any one of claims 1 to 7, characterised in that in step D) the evaluation of the optimisability comprises a classification into a predetermined number of, preferably three, categories.

9. The method according to any one of claims 1 to 8, characterised in that the artificial intelligence uses a machine learning algorithm.

10. The method according to any one of claims 1 to 9, characterised in that a constellation of the parameter values relative to one another is evaluated for the evaluation in step D).

11. A computer program comprising program commands that, when the computer program is executed by a computer, cause the computer to carry out the method according to one of claims 1 to 10.