Methods for optimized parameterization in the additive manufacturing of three-dimensional components

Abstract

The invention relates to a method for the additive manufacturing of a three-dimensional component by layer-by-layer application of a build-up material and locally selective solidification of the build-up material by means of at least one beam impacting the build-up material, wherein the guidance of the beam is carried out on the basis of a plurality of process parameter values.

Description

[0001] The invention relates to a method for the additive manufacturing of three-dimensional components and to a system for the additive manufacturing of three-dimensional components. Furthermore, the present invention relates to a computer-readable storage medium.

[0002] Manufacturing devices and corresponding methods for the additive manufacturing of three-dimensional components by layer-by-layer application and locally selective solidification of a build material are generally known from the prior art. For the locally selective solidification of the build material, at least one corresponding irradiation unit (e.g., comprising at least one laser) is usually provided, which emits a beam onto the build material, which typically consists of a powder, in order to melt and thereby solidify the powder at the point where the beam strikes.

[0003] After an initial layer has solidified on the original powder bed in this way, another layer of build material is typically applied, and the solidification process by blasting or sintering is repeated. This process is repeated until a finished component has been produced, from which unsolidified build material can easily be removed.

[0004] Often, several elements that are to be spatially separated in a later application can be manufactured as components of a single component, which is later cut at the relevant points to increase production efficiency.

[0005] Many of the physically relevant properties of the build-up material, as well as the solidified material, are not constant during a manufacturing process, but change with time and / or temperature, as well as other external influences. For example, the successive heating of layers leads to a decrease in the external dimensions of components with increasing layer thickness. Furthermore, it is comparatively difficult to estimate the extent to which different layer temperatures and geometries will affect the crystalline structures of the solidified material.

[0006] This leads to a situation where, even with constant parameter settings or parameter values ​​according to which the solidification jet is guided, and constant application of new build-up material, a distortion of the component occurs during manufacturing that is difficult to estimate, in particular shrinkage, which means that individual elements or components cannot be manufactured with consistent manufacturing tolerances.

[0007] The underlying phenomena that cause this delay are comparatively complex, making a purely theoretical prediction, and therefore compensation, impossible. For example, it is not possible to solve the heat conduction equation for all temperature, material, and geometric properties, and thus to determine a compensation for thermal expansion in the ongoing production process.

[0008] Furthermore, a linear adjustment of the process parameter values ​​for compensation purposes is often insufficient, as the underlying processes (for example, the solutions of the heat conduction equation) are nonlinear and therefore difficult to predict numerically. Due to these physically inherent imponderables, it is therefore difficult to maintain consistently high component quality throughout the entire manufacturing process.

[0009] If these factors, which often depend indirectly or directly on the layer height, i.e., the number of layers already applied, are not taken into account, unwanted inhomogeneities and deviations in the properties of the component can occur.

[0010] Therefore, there is a need to provide a process that ensures consistently high quality within a component.

[0011] This problem is solved in particular by a method according to claim 1, and by a device according to claim 14.

[0012] Specifically, this task is solved by a method for the additive manufacturing of a three-dimensional component by layer-by-layer application of a build material and locally selective solidification of the build material by at least one beam impacting the build material, whereby the guidance of the beam is carried out according to a multitude of process parameters, comprising the following steps: a) Creating a model comprising at least one target variable that at least partially represents the properties of the component, depending on at least one process parameter or parameter value; b) Specifying, based on test data comprising a multitude of process parameter values ​​and a multitude of target variables, a set of target variables comprising at least one target variable, and determining a set of construction parameters comprising at least one process parameter value, wherein the set of construction parameters realizes the at least one target variable, based on the model determined in step a); c) Control of a manufacturing process, based on the specified set of construction parameters, for the production of the component.

[0013] Process parameters, or process parameter values, encompass all adjustable parameters that may be present during the manufacturing of the component. These include path parameters, which describe the guidance of the solidifying beam, as well as other parameters describing the beam's intensity, frequency, dimensions, and / or exposure time (etc.), and parameters describing the composition, density, and / or thickness (etc.) of the build-up material.

[0014] Test data, in this context, refers specifically to all data collected through experiments or simulations prior to the manufacturing of an actual component, in order to investigate the relationship between process parameters and the properties of a manufactured component. This also includes the manufacturing and testing of test components.

[0015] The target parameter here is understood to be, in particular, a property of the component after completion and / or (immediately) after the solidification step. This can be an extrinsic property, such as dimensions, surface roughness and / or diameter (and / or similar), as well as an intrinsic property, such as crystal structure, electrical or thermal conductivity and / or flexibility (and / or similar).

[0016] A model encompasses, in particular, a set of known relationships that connect the process parameter values ​​with the target variables. These relationships can be either deterministic or statistical in nature. A deterministic model would, for example, specify a target variable as a (continuous) function of a process parameter, at least over a certain interval (e.g., the density of the finished component as a function of the radiation intensity). However, the model can also include statistical relationships, such as correlations between process parameters and target variables (e.g., a correlation coefficient between the exposure time and the external dimensions of the component). Similarly, the model can include tables that list the realized target variables observed in experiments with a (varying) set of process parameters.The model therefore reflects, in particular, the amount of theoretical and empirical information available about the effects of process parameter values ​​on the target variables.

[0017] In principle, it is also possible to deduce process parameter values ​​from a target variable using the model (for example, through at least a local inverse function). This allows the best target variables to be determined in advance and the corresponding process parameter values ​​to be preselected and / or chosen as design parameters.

[0018] Construction parameters are understood in particular to be the selection of process parameter values ​​which are used for the actual manufacturing of a component, i.e., for example, the actual path parameters which are to be controlled for the manufacturing of a component.

[0019] By creating a model according to the invention, based on which a set of manufacturing parameters is selected, it is possible to compensate for even sources of distortion that are not precisely traceable, thus ensuring consistent component quality. In particular, this is achieved because the empirical relationships between the target variables and the process parameters facilitate the selection of the path parameters required to produce a desired result.

[0020] In particular, this also makes it possible to record the effect of a change in certain process parameter values ​​on specific target variables through special test series, and to take this into account in the future production of components.

[0021] The model can be used, for example, to determine an advantageous combination of target parameters or a combination that enables the achievement of specific (optimized) component properties. This determination (optimization) can also be layer-dependent.

[0022] In a preferred embodiment, the model is created based on at least one correlation determined from the test data between at least one target variable and at least one process parameter value.

[0023] Since, as mentioned above, it is often not possible to establish clear relationships between process parameter values ​​and target variables, correlation is well-suited to incorporating information about a relationship, at least from a statistical perspective. For example, when determining the set of construction parameters, it can be considered whether there is a positive, negative, or no correlation between a process parameter value and a target variable in order to determine an optimized set of construction parameters as quickly as possible. Furthermore, considering correlations helps to minimize the number of failed attempts, as a general trend is known in advance, even if the exact quantitative effects cannot be predicted with certainty.

[0024] In a further preferred embodiment, the model comprises at least one target variable as a function of one or more process parameter values.

[0025] Given a known functional relationship between a target variable and a process parameter value (preferably referring to continuous functions), the change in the target variable when a process parameter value is adjusted can be predicted (essentially) accurately. This is particularly useful because it allows for the precise selection of design parameters to generate a desired component property. The functional relationship can be used in addition to or as a supplement to known correlations to accelerate the selection of a path parameter set, especially when a functional relationship to the target variables is known for only some of the process parameter values.

[0026] Method according to one of the preceding claims, in particular according to claim 3, wherein the function is determined by interpolation or extrapolation from the test data.

[0027] Since, as mentioned above, the underlying equations whose solutions would establish a functional relationship between process parameter values ​​and one or more target variables are often not analytically solvable, it is particularly advantageous to approximate a target variable as a function of the process parameter values. As the use of functions offers advantages in terms of speed in selecting path parameters, as well as predictive power and accuracy, a suitably shaped function can be determined from test data. If the relationship between the target variable and the process parameter value is determined through experiments, interpolating the experimental data to create such a function is a particularly efficient way to incorporate the acquired knowledge into the model. In particular, this makes it possible to minimize the number of required experiments, thereby saving costs.This allows the correlation that has already been demonstrated in certain parameter ranges to be easily interpolated or extrapolated to further ranges without having to conduct further experiments.

[0028] In another preferred embodiment, the function is determined by fitting from the test data.

[0029] Fitting functions to existing test data also minimizes the number of required trials and the associated costs. Furthermore, fitted functions are often easier to use and determine, especially when certain analytical properties can be anticipated.

[0030] According to a still preferred embodiment, the model is created, at least partially, on the basis of a numerical simulation.

[0031] Numerical simulation offers the advantage of eliminating the need for physical experiments and their associated material costs. Furthermore, simulations can be easily parallelized, allowing for the rapid generation of large amounts of test data, thereby increasing the model's reliability and quality. The use of numerical simulation can be particularly beneficial for generating test data, as well as for identifying correlations or functional relationships within the model itself.

[0032] Method according to one of the preceding claims, wherein the set of target variables also includes ranges, preferably intervals, for one or more target variables.

[0033] Often, it is neither possible nor necessary to set a target parameter to a precise value. Since different process parameter values ​​can sometimes have conflicting effects on one or more target parameters, selecting design parameters that reproduce exact values ​​for all desired target parameters is not always possible or desirable. Therefore, a component of the present invention can be to specify intervals or ranges for certain target parameters that are still acceptable as a manufacturing result. For example, an outer diameter can have a tolerance range between a minimum and a maximum dimension within which the resulting component can still be considered sufficiently good. This simplifies the selection of a design parameter set and allows for greater flexibility in component production.

[0034] Combinations are also possible. For example, a value for the volume and an interval for the electrical conductivity could be specified. The specified volume can be realized for any diameter-z-height combination (assuming a fixed diameter and a fixed z-height). The cross-sectional area (A) of the grid elements, and thus the electrical conductivity (σ, σ~1 / A), changes with the z-height. Therefore, not all diameter-z-height combinations are possible; rather, the z-height must be consistent with a cross-sectional area that results in an electrical conductivity value within the specified interval. In this case, the model specifically includes the relationship (in the form of a geometric formula) between diameter, z-height, cross-sectional area, and volume, as well as the relationship between cross-sectional area and electrical conductivity.

[0035] Method according to one of the preceding claims, wherein one or more target variables comprising the set of target variables are already selected based on the model.

[0036] The model makes it possible to calculate a set of target variables from the process parameter values. If there are limitations with certain process parameters, this can result in certain target variables being difficult or expensive to implement. Using the model allows such relationships to be considered when defining the desired target variables. This makes it possible to eliminate incompatible specifications in advance when defining target variables.

[0037] In another embodiment, the set of target variables comprises one or more conditions on one or more target variables, and the set of construction parameters is determined such that the one or more conditions are met.

[0038] It is often not necessary for a certain target parameter to assume a specific value, but merely not to exceed or fall below a certain value. Such an approach to determining construction parameters simplifies and accelerates the production of the component without neglecting critical properties. For example, in the manufacture of insulating elements, their specific conductivity is not critical, as long as it does not exceed a certain threshold.

[0039] It is also possible to distinguish between hard and soft conditions. Hard conditions are those that must be absolutely implemented in the finished component; otherwise, the result is unacceptable. Soft conditions, on the other hand, are those that are not strictly necessary but are nevertheless desirable.

[0040] Specifically, it is therefore possible to select process parameter values ​​in such a way that all hard conditions are met, and then to further modify the process parameter values ​​in order to meet as many of the soft conditions as possible without affecting the fulfillment of the hard conditions, in order to finally arrive at the construction parameters.

[0041] In a further embodiment, the model includes further information about the dependence between two or more target variables, preferably a correlation, and more preferably a function dependence.

[0042] By considering how target parameters are interrelated, a more efficient selection of construction parameters can be ensured. For example, the cross-sectional area and the height of a component are directly related to its volume, which is simply the product of the two.

[0043] In a further preferred embodiment, a dependency between two or more target variables is taken into account when determining the set of construction parameters.

[0044] By taking into account how different target variables are related to each other, a more efficient choice of construction parameters can be made, as redundant adjustments and / or calculations can be avoided or at least reduced.

[0045] It is still preferred that the set of construction parameters includes layer-specific process parameter values.

[0046] By allowing the construction parameters to be individually defined for each layer (or at least for some layers), it is possible to achieve an even more precise adaptation to the local conditions and an even more precise realization of the model.

[0047] In a further preferred embodiment, one or more target variables are optimized when determining a set of construction parameters.

[0048] Because a set of target variables cannot always be ideally determined, as individual target variables may "contradict" each other, it can be advantageous to optimize only one target variable when defining the set of construction parameters, rather than specifying a rigid range for it. For example, a certain degree of wall stability may be desired, as well as the lightest possible component. While a set of construction parameters with comparatively thick walls would maintain high wall stability, it would also result in a very high weight, which is suboptimal. Instead of indirectly specifying this through flexible conditions, it is also advantageous to choose the construction parameters (e.g., while adhering to all strict conditions) in such a way that they optimize one target variable. In the example above, a wall thickness could be specified, and the remaining construction parameters chosen to achieve the lowest possible weight.Preferably, a constant wall thickness and density of the produced wall is sought, since this can represent a constant electrical resistance.

[0049] Another aspect of the present invention comprises a manufacturing plant, in particular a laser sintering or laser melting plant, according to the above.

[0050] A manufacturing plant according to the invention preferably comprises a control unit designed to control the manufacturing plant according to the method described above. In particular, the manufacturing plant may also include a processor, especially a microprocessor, particularly to instruct the system to implement the method as described above. The manufacturing plant may include a memory, preferably containing instructions that the processor can access to instruct the control unit to execute the method described above, and / or at least one input and / or output device for transferring data.

[0051] Furthermore, the invention comprises a computer-readable storage medium containing instructions that cause at least one processor to implement a method according to the preceding descriptions when the instructions are executed by the at least one processor.

[0052] The invention is described below with regard to further details, features, and advantages, which are explained in more detail with reference to the figures. The described features and combinations of features, as shown below in the figures and described with reference to the drawing, are applicable not only in the combinations specified, but also in other combinations or individually, without departing from the scope of the invention.

[0053] This shows: Fig. 1: a representation of testing or simulation to establish a relationship between process parameter values ​​and target variables; Fig. 2: the summarization of process parameter values ​​and target variables, as well as relationships between them, into a model; Fig. 3: the manufacturing of a component based on a set of target variables using the model; and Fig. 4: the selection of construction parameters within a given interval of a target variable.

[0054] In Fig. Figure 1 shows a step for creating a model 100 according to the present disclosure. In a first step, a set of process parameter values ​​1 is specified, and subsequently, by means of a test or simulation 3, these are related to a set of target variables 2. The test can, for example, consist of manufacturing a component using the specified set of process parameter values ​​1, and then measuring the target variables 2 of the resulting component to determine the relationship between the selected process parameter values ​​1 and the target variables 2.

[0055] Here and in the following, curly brackets around process parameter values ​​1 or target variables 2 denote a set of such process parameter values ​​1 or target variables 2, where the index indicates the element of the respective set. The cardinality of the sets can differ; for example, there can be more or fewer process parameter values ​​1 than target variables 2, or vice versa. Fundamentally, Ω denotes the set of possible indices of the process parameter values ​​1, and Ψ the set of possible indices of the target variables 2. Ξ additionally denotes the set of possible indices of the design parameters 4. The elements of the sets can (as described above) also include conditions or intervals of the respective variables. To avoid unnecessary confusion, the notation is retained, so that, for example, Z1 can refer directly to the target variable Z1 or to an interval of possible values ​​of Z1.

[0056] In Fig. Figure 2 shows the integration of information obtained from a test or simulation 3 into a model 100. Model 100 comprises the set of available information about the relationships between process parameter values ​​1 and target variables 2. For example, it is conceivable that a target variable 2 is known as a function 5 of a set of process parameter values ​​1 (for example, the volume of a cylindrical component is known as a function of a curve radius and a height, or number of layers).

[0057] Furthermore, the model 100 can also include correlations 6 between different quantities, for example between process parameter values ​​1 and target variables 2, as well as between process parameter values ​​1 or target variables 2 themselves.

[0058] Furthermore, the model 100 can also include relations 7 of a more general nature. For example, it might be known that a certain process parameter value 1 is always smaller than another process parameter value 1, or that certain process parameter values ​​1 imply or exclude certain target variables 2. These "qualitative" relations 7 can also be part of a model 100 according to the invention.

[0059] Fig. Figure 3 shows the manufacturing process of a component 200 using the method according to the invention. After creating a model 100, a set of desired target parameters 2 for a component 200 is specified, and a corresponding set of construction parameters 4 is selected from the set of process parameter values ​​1 based on the model 100 (it is understood that Ξ is a (not necessarily real) subset of Ω). The selection of the construction parameters 4 according to the model 100 is carried out such that the specified target parameters 2 are realized in the actual component 200. As discussed above, it is also possible to specify only an interval for certain and / or all target parameters 2.

[0060] Fig. Figure 4 illustrates the selection procedure under certain conditions. A portion of model 100 is shown, which defines the target variables Z1 and Z2 as a function of a process parameter value P. k shows (i.e., Z1(P) k) (thick dashed curve) and Z2(P k (solid curve). A strict condition is given that Z1 must not exceed a certain value (shown as a thick, dashed horizontal line). Simultaneously, it is specified that the value of Z2 must lie within a first interval 8. It can therefore be derived from the functions 5 contained in model 100 that the process parameter value P k assigned construction parameters B k A second interval 9 must be selected to produce component 200 (not shown) within the desired specifications.

[0061] Fig. 5 shows a similar situation to Fig.4, however with different function curves, and the condition that Z1 must be greater than a predefined value (horizontal dashed line). In this case, it is not possible to fulfill both conditions simultaneously. If both conditions are defined as hard conditions, no solution can be found, and based on model 100, an error message can preferably be issued before the construction parameter 4 is selected. If, on the other hand, the condition on Z1 is formulated only as a soft condition, a construction parameter 4 can still be selected from the third interval 10 to ensure that Z2 lies in the first interval. In order to still get as close as possible to fulfilling the soft condition, the invention would then use the lowest possible value for P. k selected as construction parameter 4.

[0062] It should be noted here that all the parts described above, considered individually and in any combination, especially the details shown in the drawings, are claimed as essential to the invention. Modifications to this are familiar to those skilled in the art.

[0063] Furthermore, it is noted that the broadest possible scope of protection is sought. Therefore, the disclosure contained in the claims can also be specified by features that are described by further features (even if these further features are not necessarily included). It is explicitly pointed out that parentheses and the term "in particular" are intended to highlight the optionality of features in the respective context (which does not imply that a feature is to be considered mandatory in the corresponding context without such indication).

Claims

[1] Method for the additive manufacturing of a three-dimensional component by layer-by-layer application of a build material and locally selective solidification of the build material by at least one beam impacting the build material, wherein the guidance of the beam is carried out on the basis of a plurality of process parameter values, comprising the following steps: a) Creating a model comprising at least one target variable that at least partially represents the properties of the component, depending on at least one process parameter value; b) Specifying, based on test data comprising a multitude of process parameter values ​​and a multitude of target variables, a set of target variables comprising at least one target variable, and determining a set of construction parameters comprising at least one process parameter value, wherein the set of construction parameters implements the at least one target variable, based on the model determined in step a); c) Control of a manufacturing process, based on the specified set of construction parameters, for the production of the component. [2] Method according to claim 1, wherein the model is created on the basis of at least one correlation determined from the test data between at least one target variable and at least one process parameter value. [3] Method according to any of the preceding claims, wherein the model comprises at least one target variable as a function of one or more process parameter values. [4] Method according to any of the preceding claims, in particular according to claim 3, wherein the function is determined by interpolation or extrapolation from the test data. [5] Method according to any of the preceding claims, in particular according to claim 3, wherein the function is determined by fitting from the test data. [6] Method according to any of the preceding claims, wherein the model is created, at least partially, on the basis of a numerical simulation. [7] Method according to one of the preceding claims, wherein the set of target variables comprises ranges, preferably intervals, for one or more target variables. [8] Method according to one of the preceding claims, wherein one or more target variables comprising the set of target variables are selected based on the model. [9] Method according to one of the preceding claims, wherein the set of target variables comprises one or more conditions on one or more target variables, and the set of construction parameters is determined such that the one or more conditions are met. [10] Method according to any of the preceding claims, wherein the model further comprises information about the dependence between two or more target variables, preferably a correlation, more preferably a function dependence. [11] Method according to any of the preceding claims, in particular according to claim 10, wherein a dependency between two or more target variables is taken into account when determining the set of construction parameters. [12] Method according to any of the preceding claims, wherein the construction parameter set includes layer-specific process parameter values. [13] Method according to one of the preceding claims, wherein in determining a set of construction parameters, one or more target variables are optimized. [14] Manufacturing plant for the additive manufacturing of objects, in particular laser sintering or laser melting plant, comprising a control unit which is configured to control an adjustment or setting unit according to the method according to any one of claims 1 to 13. [15] Computer-readable storage medium containing instructions that cause at least one processor to implement a method according to any one of the preceding claims 1 to 13 when the instructions are executed by the at least one processor.

Images