Method for optimizing additive manufacturing

Abstract

The invention relates to a computer-implemented method for optimizing additive manufacturing of a component, comprising the following steps: a) creating a virtual component model (1), b) analyzing the properties of the virtual component model (1) in order to obtain at least one property value comprising a plurality of localized individual values, c) subdividing the virtual component model (1) into at least two segments (2) on the basis of the localized individual values of the at least one property value, d) allocating a manufacturing parameter set from a plurality of manufacturing parameter sets to each segment (2) of the virtual component model (1) in order to obtain a manufacturing model, e) converting the manufacturing model into a machine file for manufacturing the component, and f) manufacturing the component by means of a device for additive component manufacturing using the machine file.

Description

[0001] Methods for optimizing additive manufacturing

[0002] The present invention relates to a computer-implemented method for optimizing additive manufacturing of a component.

[0003] Various additive manufacturing processes, also commonly referred to as 3D printing processes, are known from the state of the art. Such processes include selective laser melting, selective electron beam melting, and laser cladding. They enable the near-net-shape production of complex components that were previously only possible through complex combinations of existing processes, or even impossible to produce at all.

[0004] State-of-the-art additive manufacturing processes can use single metallic materials, metallic alloys (e.g., titanium-, nickel-, iron-, or aluminum-based materials), conventional polymers (e.g., thermoplastics), and fiber-reinforced or filled polymers as starting materials. The starting material is usually applied layer by layer using a powder or wire and thermal energy input, and the component is built up additively layer by layer.

[0005] In state-of-the-art additive manufacturing processes, there is a close relationship between the processing—i.e., between the process parameters, the manufacturing strategy, and the material formation of the manufactured component—on the one hand, and the resulting material properties of the manufactured component, on the other. In the selective laser melting (SLM) process, for example, the relevant process parameters are laser power, laser beam quality, scan speed, and / or line spacing, while in the manufacturing strategy, the relevant process parameters are the scan direction or the orientation of the component within the build space. Furthermore, the cooling rate of the starting material, after it has been heated and melted by the laser radiation, significantly influences the resulting material properties of the manufactured component during component formation.

[0006] Additive manufacturing processes that process metallic materials, such as selective laser melting, are characterized by high cooling rates, which significantly influence the metallurgy and thus the material properties of the processed material. The cooling rates have a detailed influence on the microstructure, microstructures, phases, anisotropy, and defects such as pores and cracks in the processed material. Repeated energy inputs from laser radiation and / or the melting of zones of adjacent component layers result in local, complex thermal profiles with repeated heating and cooling curves in the component. These complex thermal profiles lead to different material properties in different areas of a manufactured component – ​​despite identical process parameters.The microstructures and textures altered by thermal treatment with laser radiation significantly determine the mechanical, electrical, and thermal material properties and thus the material quality of the manufactured component. There is also a direct correlation between the thermal profiles resulting from the process parameters and the formation of thermal residual stresses, which influence undesired deformation of the manufactured component and possibly crack formation.

[0007] Furthermore, the process parameters determine not only the material or component quality but also the build rate of the additive manufacturing process. The build rate defines the ratio of additively manufactured volume per unit of time. For example, the build rate for selective laser melting can be in the range of 100 cm 3 / h. Process parameter sets that produce high material quality often lead to a low build rate. Conversely, process parameter sets that have a high build rate often lead to losses in the material quality of the manufactured component.

[0008] In state-of-the-art additive manufacturing processes, the component is typically manufactured starting from a virtual component model. In a pre-processing step, machine commands for the manufacturing device, such as the 3D printer, are first assigned to the virtual component model. Such machine commands can contain information about the energy input or the production speed of the 3D printer. Furthermore, the virtual component model is virtually divided into a multitude of layers using the linked machine commands, and process-specific layer information for generating the component contour is assigned to the individual layers. In this way, a first layer can be manufactured at a higher production speed or with a higher energy input than a second layer.According to the state of the art, the software that creates the machine instructions with regard to the component to be manufactured and the layers only specifies different parameters based on the geometry of the component to be manufactured. In the state of the art, however, it is not known to specify different parameters taking into account different acting loads, different functions and / or different requirements for the component to be manufactured.

[0009] With regard to process parameters, a distinction is generally made between those parameters that describe the production of a volume area inside the component and those that describe the production of a contour or edge area. The parameter sets used generally aim to achieve optimal material and surface properties throughout the entire component and take into account the geometry of the component to be manufactured, while neglecting the fact that optimal material and surface properties are only required in areas of the component subject to particular stress. This results in a component being manufactured that, while possessing optimal material and surface properties throughout its entire volume, is not manufactured in a time- and resource-efficient manner.In summary, state-of-the-art additive manufacturing processes produce components that, on the one hand, have areas where material and surface properties do not necessarily prevail and, on the other hand, have a low build rate, which makes the production of a component time- and cost-intensive.

[0010] Based on the described disadvantages of the additive manufacturing processes known from the prior art, it is therefore the object of the present invention to provide a computer-implemented method for optimizing the additive manufacturing of a component.

[0011] The computer-implemented method according to the invention is defined by the features of claim 1.

[0012] The computer-implemented method according to the invention for optimizing additive manufacturing of a component comprises the following steps: a) creating a virtual component model, b) analyzing the properties of the virtual component model to obtain at least one property value comprising a plurality of localized individual values, c) dividing the virtual component model into at least two segments based on the localized individual values ​​of the at least one property value, d) assigning a manufacturing parameter set of a plurality of manufacturing parameter sets to each segment of the virtual component model to obtain a manufacturing model, e) converting the manufacturing model into a machine file for manufacturing the component, and f) manufacturing the component with a device for additive component manufacturing using the machine file.

[0013] The computer-implemented method according to the invention advantageously allows additively manufactured components to be produced at an optimized build rate, i.e., in a shorter production time. At the same time, less starting material is used, which additionally saves resources and costs. A particularly advantageous feature of the method according to the invention is that the virtual component model is subdivided into at least two segments based on the localized individual values ​​of at least one property value.

[0014] Within the scope of the method according to the invention, the creation of a virtual component model according to step a) can be carried out, for example, using software, such as commercially available CAD software. A virtual component model created in this way consists of a geometric model that can be available in various formats. The virtual component model can be created, for example, as a geometric model in the form of a mesh of triangles or polygons, as a vector-based surface model, or as a file in a native software format from a manufacturer of the corresponding software. The virtual component model can already be available in the form of a mesh geometry.

[0015] According to step b) of the method according to the invention, the properties of the virtual component model are analyzed to obtain at least one property value comprising a plurality of localized individual values. The properties can be, for example, mechanical, thermal, or aerodynamic properties of the component model, which can be analyzed using appropriate software. The analysis of the properties generates at least one property value, which can, for example, comprise a mechanical stress or a stress state of the component model.

[0016] The property value comprises a plurality of localized individual values, which in turn contain information about the property value at a specific location in the virtual component model. This means that, for example, a specific location in the virtual component model can be assigned multiple pieces of stress information, such as a normal stress, which act in different spatial directions at this specific location. Of course, multiple properties of the virtual component model can also be analyzed simultaneously or sequentially, so that at least one localized individual value can be assigned in a different way to a specific location in the virtual component model. Since information about the component model is obtained from the analysis of the properties of the virtual component model according to step b), a geometrically adapted virtual component model can be generated after the properties have been analyzed.With this component model, the analysis reveals, for example, which mechanical load—for example, in the form of normal stress—acts at a specific location on the component. With the geometrically adjusted component model, areas subject to more intense mechanical stress can have greater wall thickness, a denser material, or otherwise optimized material properties. At the same time, areas of the component subject to less intense mechanical stress can have non-optimized material properties or be removed entirely from the geometric model, thus reducing production time.

[0017] According to step c) of the method according to the invention, the virtual component model is divided into at least two segments based on the localized individual values ​​of the at least one property value. In this way, the virtual component model is divided into several segments, for example into a first segment in which a high load state is to be expected and a second segment in which a low load state is to be expected. The respective load state is described by the localized individual values. Thus, after the subdivision, the virtual component model is differentiated with regard to the properties. If, according to step b), initially only the properties of the entire virtual component model were analyzed, according to step c), the virtual component model is divided into several segments.

[0018] According to step d), each segment of the virtual component model is assigned a manufacturing parameter set of a plurality of different manufacturing parameter sets to obtain a manufacturing model.

[0019] A manufacturing parameter set can, for example, contain information about the process parameters of the additive manufacturing process, with such information defining, in particular, the manufacturing speed and manufacturing quality. If the method according to the invention is used, for example, in selective laser melting, the manufacturing parameter set can contain information regarding the laser power and the scanning speed and / or define whether a segment is manufactured with a hatching pattern or with a fill pattern. Thus, each segment can be manufactured with a different manufacturing parameter set assigned based on the property analysis. Depending on the manufacturing parameter sets assigned to the segments, the manufactured areas of a real component have individual material properties that result from manufacturing with the manufacturing parameter sets.

[0020] By assigning a manufacturing parameter set to each segment, a manufacturing model of the virtual component model is created, which thus contains information regarding the properties of the individual segments and the manufacturing parameter sets of the virtual component model.

[0021] Thus, the build rate of each segment is adjusted according to the analysis performed in step b) using an individually selected manufacturing parameter set from a plurality of different manufacturing parameter sets, and the build rate of the entire component, which consists of multiple segments, is optimized so that the actual component can be manufactured more quickly and efficiently using fewer resources. The method according to the invention therefore selects a single, most suitable manufacturing parameter set from a repertoire of a plurality of different manufacturing parameter sets based on the previously determined local property values ​​and assigns it to the respective segment.

[0022] According to step e), the manufacturing model is converted into a machine file for manufacturing a component. Conversion refers to the programming language translation of the information from the manufacturing model into a machine file that can be read and processed by a device for additive component manufacturing.

[0023] In step f) of the method according to the invention, the component is manufactured using a device for additive component manufacturing using the machine file. Such a device can be used, for example, but not exclusively, in a 3D printing, electron beam melting, laser beam melting, laser sintering, or stereolithography process.

[0024] The computer-implemented method according to the invention advantageously enables a component to be additively manufactured. By analyzing the properties and subsequently subdividing the virtual component model into several segments, the component is manufactured in an optimized manner not only with regard to its geometry, but also with regard to the required physical and, in particular, mechanical properties. Furthermore, by assigning different sets of manufacturing parameters to the respective segments, the manufacturing time is advantageously reduced and the build rate optimized, since the analysis of the properties assigns a suitable set of manufacturing parameters to each segment.Thus, for example, a segment that is subject to less mechanical stress and therefore requires non-optimal material and surface properties can be manufactured in a shorter time than a segment that is subject to greater mechanical stress and therefore requires optimal material and surface properties. Thus, the process according to the invention optimizes the build rate of an additively manufactured component.

[0025] It is preferably provided that in step c) the step of dividing into at least two segments takes place based on a comparison of at least one limit value with the localized individual value. Such a limit value can, for example, be a previously defined value, for example a voltage value. This limit value is then compared with the localized individual value of the previously determined at least one property value. By dividing the virtual component model based on a comparison of at least one limit value with the localized individual value, it is defined to which segment a localized individual value is to be assigned. Thus, for example, a first localized individual value that describes a first voltage and lies above the limit value can be assigned to a first segment and a second localized individual value that describes a second voltage value and lies below the limit value can be assigned to a second segment.Consequently, it is possible to define in a particularly advantageous manner which localized individual values ​​are assigned to a segment.

[0026] In a preferred embodiment of the method according to the invention, in step a), after creation, the virtual component model is divided into a plurality of network elements, each network element having at least one node. Such network elements with nodes can be created if the virtual component model has been created using software that analyzes the properties of a virtual component model. By dividing the virtual component model into a plurality of network elements, the component model can be virtually structured in a particularly advantageous manner and processed using commercially available software.

[0027] Preferably, the analysis of the properties in step b) is carried out node-related to obtain the localized individual values. In the case of mechanical, thermal or aerodynamic analysis using appropriate software, for example, the geometry together with the local load state is also available in a mesh-based manner. Through a node-related analysis of the properties, the information from the analysis is assigned to the virtual component model locally in a node-related manner. Thus, in addition to the geometric information, the virtual component model now also has node-related information regarding the analyzed properties, which is present in the form of localized individual values. In other words: each node of the virtual component model can now be assigned at least one localized individual value, so that, for example, after a mechanical load analysis, information regarding a stress state is available at each node of the virtual component model.Geometric information can thus also be assigned to the nodes of the network consisting of network elements, for example information regarding the distances of a node to the edge regions of the component model, which enables the segments to be subdivided into different geometric zones. Such a geometric zone can, for example, contain struts, thick-walled or thin-walled regions, and strongly or slightly curved regions. Through the node-related analysis of the properties to obtain the localized individual values, information regarding the properties is advantageously locally assigned to the virtual component model. Such a node-related analysis can, for example, be carried out in the form of a stress analysis using appropriate software. Preferably, in step c) the nodes are assigned to a segment based on a comparison of the localized individual value with at least one threshold value.By comparing the localized individual value with at least one threshold value, the nodes are assigned to a segment in a defined manner. In this way, a segment only has nodes that have a localized individual value that lies above or below at least one threshold value. For example, a segment can have nodes that are only assigned stress values ​​above a predetermined threshold value. Such a threshold value can, for example, represent a limit stress, so that a segment only contains nodes with localized individual stress values ​​that lie above the respective threshold value. Based on this threshold value, a manufacturing parameter set can then be selected in the subsequent process steps, with which the segment is manufactured according to the stress state.Thus, the respective segments, which represent an individual geometric portion of the component, can be individually grouped according to the information of the localized individual values, or the local individual values ​​can be assigned. Consequently, the size and local assignment of the individual segments can be advantageously determined using a threshold value depending on the information of the localized individual values ​​assigned to the nodes.

[0028] The subdivision according to step c) based on the preferential comparison with a threshold value differs from the subdivision based on a preferential comparison with a limit value by including the information of the localized individual values ​​that are available in a node-related manner. The comparison with a limit value can be carried out without the presence of network elements and nodes, whereas the comparison with a threshold value requires network elements and nodes. In a particularly preferred embodiment of the method according to the invention, the network elements arranged at a boundary region of segments are divided into at least two sub-network elements, which are assigned to the respective adjacent segments or the network elements are assigned to one of the adjacent segments. For geometric reasons, the boundary between two segments runs exactly through the nodes of the network elements.

[0029] If two segments border each other, the adjacent network elements can share nodes that have localized individual values ​​above a threshold, as well as nodes that have localized individual values ​​below a threshold. To enable a particularly advantageous assignment of the network elements to the segments, the corresponding network element can be divided into at least two subnetwork elements, with one subnetwork element being assigned to one segment and the other subnetwork element being assigned to the other segment.

[0030] By dividing the network elements that are arranged at the boundary area of ​​segments into at least two sub-network elements, the sub-network elements can be assigned to the adjacent segments, thus ensuring a clear assignment of the network elements to the segments.

[0031] Alternatively, the network elements are assigned directly to the adjacent segments without any division. By assigning the network elements to the respective segments in this way, the component geometry advantageously has no gaps in the boundary area, so that each network element is advantageously assigned to a segment.

[0032] It can also be provided that segments are separated from each other by gaps. In this case, the network elements in the edge area are not assigned to either segment. By not assigning network elements to one of the adjacent segments in this way, segments can be spatially separated from each other.

[0033] Additionally, it is possible to create an overlap of the segments by adding additional thresholds. Even more complex overlap concepts, such as checkerboard or zigzag arrangements of different segments, can be created using additional thresholds.

[0034] It should also be noted that in a component model with multiple segments, the individual segments do not necessarily have to be connected, i.e., directly adjacent to one another. However, to counteract excessive fragmentation, i.e., the division of the component into too many non-directly connected segments, the segments consisting of a few network elements can also be assigned to other segments. In other words, many small segments are integrated into another adjacent segment to simplify the process. This advantageously prevents excessive fragmentation of the component. The component model then has a reduced number of segments, which advantageously optimizes the build rate.

[0035] Preferably, the analysis of the properties of the virtual component model according to step a) is carried out by simulation and / or by data on a load and / or by measurements on a real component. Numerical simulation analysis can be performed, for example, using finite element simulation (FE simulation).

[0036] The analysis of properties using data on a load condition can be performed, for example, using data from a strain gauge or another sensor or measuring device. Such measuring devices can be arranged directly on a real component model and thus generate data, for example, on a load condition of the real component model. Measurements on a real component model can also be performed, for example, in the form of flow measurements in a wind tunnel. By analyzing the properties through simulation and / or using data on a load and / or through measurements on a real component model, at least one property value can advantageously be assigned to the virtual component model in various ways.The method according to the invention can thus be advantageously applied to already existing simulations and component models, and / or use already existing data and measurements of a real component model.

[0037] In a particularly preferred embodiment of the method according to the invention, it is provided that in step e) when converting the manufacturing model into the machine file, the manufacturing model is divided into layers.

[0038] Alternatively or in addition to layers, the manufacturing model can also be virtually divided into paths, vectors, or alternative elements. While the division into layers, also known as "slicing," is preferably used in conventional 3D printing processes, the division into paths and vectors is preferred in deposition welding processes and other additive manufacturing processes. By dividing the manufacturing model into layers, paths, and / or vectors, the manufacturing model can be further processed in a particularly advantageous manner.

[0039] Preferably, the component model is formed by a three-dimensional component model. A real three-dimensional component model can be recorded, for example, by a camera or scanner device, so that a virtual component model is generated from the recorded data of the three-dimensional component model. Thus, the method according to the invention can advantageously also be used with the inclusion of three-dimensional component models.

[0040] In a particularly preferred embodiment of the method according to the invention, the three-dimensional component model is a CAD / CAM model or a 3D scan of the actual component. Thus, the method according to the invention can advantageously be applied to existing CAD / CAM models, and commercially available CAD / CAM software can interact with the method according to the invention. 3D scans of any actual components can also be used, so that the method according to the invention can advantageously interact with commercially available software for creating 3D scans.

[0041] Preferably, the component model includes at least one material property of the component. Such a material property can be, for example, thermal conductivity, mechanical properties, or other material parameters. Including at least one material property in the component model has proven particularly advantageous for the method according to the invention.

[0042] Preferably, the method according to the invention comprises, after step f), the further method step g): checking the manufactured component using a validation process. By checking the manufactured component, defects in the component can be used to identify non-optimal process sequences, so that the process, in particular the selection of the assigned manufacturing parameter sets, can be advantageously optimized.

[0043] In a particularly preferred embodiment of the method according to the invention, the manufactured component is inspected using resonance analysis, whereby the manufactured structure is mechanically tested. Specific acoustic parameters are calculated for the respective component from the resonance frequencies used in the resonance analysis, to which quality characteristics such as "OK," "cracked," or "material and structural defects" are assigned. The assignment is made by comparing the measured parameters with the values ​​of comparison objects with known properties. The parameters depend on the geometry, material, and internal structure of the manufactured component. Resonance analysis has proven particularly advantageous for inspecting the manufactured component when using the method according to the invention.However, according to the invention, other inspection methods can also be used, such as computer tomography or X-ray methods to check the manufactured component for defects or deviating material densities.

[0044] Preferably, the machine file generated according to method step e) contains a G-code. G-code, also known as RS-274 code, is the most widely used CNC programming language. G-code is primarily used in computer-aided manufacturing to control automated machine tools and production machines. By using a G-code in the machine file, the method according to the invention can be advantageously used by a variety of existing additive manufacturing devices.

[0045] Preferably, the assignment of a manufacturing parameter set to the segment in step d) is carried out on the basis of at least one localized individual value of at least one property value. Thus, the assigned manufacturing parameter set can be adapted to the localized individual value of the at least one property value. The property value can, for example, be a degree of porosity of the material, so that a manufacturing parameter set having a manufacturing speed is assigned accordingly to the local porosity corresponding to the local individual value. Thus, in the method according to the invention, the manufacturing parameter set can always be assigned in such a way that it is advantageously adapted to the respective localized individual value.Of course, the at least one localized individual value of the at least one property value can also be compared with other localized individual values ​​within the segment or within the entire component, and on the basis of this comparison, a manufacturing parameter set can then be assigned to the segment.

[0046] For example, if a segment has several localized individual values ​​regarding the porosity of the material, these porosities can be compared with each other. A manufacturing parameter set is then assigned to the segment, which, for example, generates the lowest porosity of all local individual values ​​in the segment. This makes it particularly advantageous to assign a customized manufacturing parameter set to a segment, and optimize the build rate according to the desired manufacturing result.

[0047] The invention is explained in more detail below with reference to the following figures.

[0048] They show:

[0049] Figure 1 is a schematic representation of a virtual component model according to method step a) of the method according to the invention and

[0050] Figure 2 shows a topology-optimised component model after the properties of the virtual component model have been analysed according to step b) and

[0051] Figure 3 shows a virtual component model with two segments in plan view, Figure 4 shows a virtual topology-optimised component model with a network structure with a plurality of network elements and nodes,

[0052] Figure 5 is a schematic representation of two segments with clearly assigned network elements,

[0053] Figure 6 is a schematic representation of two segments with network elements divided into subnetwork elements and

[0054] Figure 7 is a schematic representation of two segments that are spatially separated from each other by network elements represented as gaps.

[0055] Figure 1 schematically illustrates a virtual component model 1. The virtual component model 1 shown is a three-dimensional component model or volume model, created, for example, using a 3D scan of the actual component or with the aid of CAD software. Thus, the method according to the invention can advantageously be used with commercially available software.

[0056] Figure 2 shows the three-dimensional virtual component model according to Figure 1, which has a topology-optimized shape. Such a topology-optimized shape of the virtual component model 1 can be created after analyzing the properties of the virtual component model. In the topology-optimized virtual component model shown in Figure 2, the geometry has been adapted, for example, after carrying out an FE simulation with regard to equivalent stresses. By analyzing the equivalent stress, where the equivalent stress represents a property of the virtual component model, it is known in which areas the geometry of the later component must be reinforced, for example due to expected high local stresses, or in which areas of the geometry material can be omitted due to expected low local stresses. The component model 1 according to Figure 2 therefore has, in comparison to the component model according to Figure1 has several recessed areas. After analyzing the properties, a component model can advantageously be created that, thanks to the recesses, has a reduced volume while simultaneously strengthening the structure.

[0057] By analyzing the properties of the virtual component model 1, the geometry of the subsequent component has been advantageously optimized and production material saved. The topology-optimized component model 1 shown in Figure 2 is therefore manufactured in a shorter production time and with less material than the initial geometry shown in Figure 1. Of course, an analysis of the properties of the virtual component model can also be carried out with regard to other properties, such as thermal or aerodynamic properties. An analysis with regard to further properties can lead to a different form of the topology-optimized component model, since the component must be reinforced in other areas with regard to mechanical properties, for example, i.e. with regard to thermal properties.

[0058] Figure 3 shows a plan view of the virtual topology-optimized component model 1 according to Fig. 2, wherein the virtual component model 1 has been divided into two segments 2. The virtual component model 1 has been divided into a first segment 2a and a second segment 2b based on the localized individual values ​​of the at least one property value. The segments 2a, 2b are not necessarily connected to one another, as can be seen by way of example from Figure 3. The division of the virtual component model 1 can preferably be carried out based on a comparison of at least one limit value with a localized individual value. Such a limit value can, for example, be a defined wall thickness or component thickness, so that the respective localized individual value, i.e. the local wall thickness, is compared with the limit value, i.e. a defined wall thickness.If the localized individual value of the wall thickness is above the limit value, this area is assigned, for example, to the first segment 2a. If the localized individual value of the wall thickness is below the defined limit value, the local area is assigned to segment 2b. In a further process step, a manufacturing parameter set is then assigned to each segment 2, so that each segment 2 is manufactured with a different manufacturing parameter set. Since, in the additive manufacturing process according to the invention, the respective manufacturing parameter set defines the material properties of the respective segment 2, the segments 2 are manufactured with the desired quality by selecting the respective manufacturing parameter set.

[0059] Figure 4 shows a perspective oblique view of the topology-optimized component model 1, with a mesh structure comprising a plurality of mesh elements 3 and nodes 4. The virtual component model 1, which has a plurality of mesh elements 3 with multiple nodes 4, can, for example, already be created in the form of a mesh-based geometry during the creation of the CAD model. Another such mesh-based geometry of the virtual component model 1 can be generated, for example, through an FE simulation.

[0060] Figure 5 shows a schematic representation of two segments 2a, 2b with uniquely assigned network elements 3 and several nodes 4. The representation according to Fig. 5 shows segment 2a, which borders segment 2b. Each network element 3 has four nodes 4. For clarity, the dark-marked network elements 3 are uniquely assigned to segment 2a, and the light-marked network elements 3 are uniquely assigned to segment 2b. Such a unique assignment of the network elements is particularly advantageous if the localized individual values ​​assigned to the respective nodes are compared with a threshold value, and based on this comparison, the network element is assigned to a segment 2.Since each network element 3 has several nodes 4, a network element 3 can, for example, also have a node 4 with a localized property value that is above the threshold value and another node 4 with a localized property value that is below a threshold value, so that the respective network element cannot be unambiguously assigned to a segment 2 based on the comparison with a threshold value. In this case, for example, a unique assignment of the network elements 3 to a segment 2a, 2b is particularly advantageous.

[0061] Figure 6 shows a schematic representation of the segments 2a, 2b with clearly assigned network elements 3 and a network element 3 that is divided into two sub-network elements 3a, 3b, as well as with several nodes 4. As an alternative to the procedure described in Figure 5 regarding the treatment of network elements that are arranged at a border area of ​​several segments, a network element 3 that is arranged at a border area can also be virtually separated by manually adding further nodes 4a. By manually adding the additional nodes 4a, a network element 3 can be divided particularly advantageously, so that not an entire network element 3 has to be assigned to a segment 2, but only the corresponding sub-network element. The segments 2 can thus be delimited from one another with a particularly high resolution, and production of the individual segments with different production parameter sets can take place separately from one another with a high resolution.

[0062] Figure 7 shows a schematic representation of two segments 2a, 2b having network elements 3, which are spatially separated from one another and are connected to one another only by two network elements, each designed as a gap 5. Although the aim is generally to use a component model that has no gaps, it can be advantageous to separate individual segments 2a, 2b from one another only by network elements designed as a gap 5. In this way, individual segments 2a, 2b can be arranged as close to one another as possible, but without having a common boundary. The individual segments 2a, 2b, to which a set of manufacturing parameters is each assigned according to the invention, can thus be manufactured like an independent component. By separating or spacing the individual segments 2a, 2b in this way, the additive manufacturing process is thus carried out in an advantageous manner.

[0063] Reference symbol list component model segment a first segment b second segment network element a first subnetwork element b second subnetwork element node a further node gap

Claims

Patent claims 1. Computer-implemented method for optimizing additive manufacturing of a component, comprising the following steps: a) creating a virtual component model (1), b) analyzing the properties of the virtual component model (1) to obtain at least one property value comprising a plurality of localized individual values, c) dividing the virtual component model (1) into at least two segments (2) based on the localized individual values ​​of the at least one property value, d) assigning a manufacturing parameter set of a plurality of manufacturing parameter sets to each segment (2) of the virtual component model (1) to obtain a manufacturing model, e) converting the manufacturing model into a machine file for manufacturing the component, and f) manufacturing the component with a device for additive component manufacturing using the machine file.

2. The method according to claim 1, wherein in step c) the step of subdividing is carried out based on a comparison of at least one limit value with the localized individual value.

3. Method according to claim 1 or 2, wherein in step a) after creation, the virtual component model (1) is divided into a plurality of network elements (3), each network element (3) having at least one node (4).

4. The method according to claim 3, wherein in step b) the analysis of the properties to obtain the localized individual values ​​is carried out on a node-related basis.

5. The method according to claim 3 or 4, wherein in step c) the nodes (4) are assigned to a segment (2) based on a comparison of the localized individual value with at least one threshold value.

6. Method according to one of claims 3 to 5, wherein the network elements (3) arranged at a boundary region of segments (2) are divided into at least two sub-network elements (3a, 3b) which are assigned to the respective adjacent segments (2) or the network elements (3) are assigned to one of the adjacent segments (2).

7. Method according to one of claims 1 to 6, wherein the analysis of the properties is carried out by simulation and / or by data on a load and / or by measurements on a real component model.

8. Method according to one of claims 1 to 7, wherein in step e) when converting the manufacturing model into the machine file, the manufacturing model is divided into layers. 9 Method according to one of claims 1 to 8, wherein the component model (1) is formed by a three-dimensional component model.

10. The method according to claim 9, wherein the three-dimensional component model is a CAD / CAM model or a 3D scan of the real component.

11. The method according to any one of claims 1 to 10, wherein the component model comprises at least one material property of the component.

12. The method according to any one of claims 1 to 11, wherein the method after step f) comprises the further method step g): checking the manufactured component by a validation method.

13. The method according to claim 12, wherein the checking is carried out by means of a resonance analysis.

14. The method according to any one of claims 1 to 13, wherein the machine file contains a G-code.

15. The method according to any one of claims 1 to 14, wherein the assignment of a manufacturing parameter set to the segment in step d) is carried out on the basis of at least one localized individual value of at least one property value.

Images