Method and apparatus for separating excess material from an additively manufactured part

Abstract

For separating excess material (WS) from an additively manufactured component (BT), spatially resolved structural data (CAD) of the component (BT) are received. With the structural data (CAD), a cavity (H) of the component (BT) is divided into grid cells (GZ) by a virtual spatial grid (G), wherein the alignment of the grid (G) with respect to the component (BT) in space is specified by an alignment specification (RA). From the alignment specification (RA), a discrete pattern (R) of orientations in space of the component (BT) is determined. Furthermore, the grid cells (GZ) of the cavity (H) are virtually filled with a virtual material (VWP). Then, with the structural data (CAD), a process of pouring the virtual material (VWP) from the component (BT) is simulated for orientations in space limited to the discrete pattern (R), wherein a time sequence of orientations (AP(1),..., AP(N)) limited to the discrete pattern (R) is determined. From the simulated pouring process, the component (BT) is subsequently rotated in succession to the orientations (AP(1),..., AP(N)) in the time sequence.

Description

Background Technology

[0001] Additive manufacturing is becoming increasingly important in modern production processes. It allows for the production of products with almost arbitrary contours and topologies with relatively low overhead. Compared to classical manufacturing methods, additive manufacturing requires fewer additional construction constraints.

[0002] Known additive manufacturing techniques include the so-called powder bed method, which is particularly used for manufacturing metal parts. In this case, the material for the part to be manufactured is provided in layers as a powder bed, fluidized bed, or bed of materials. To manufacture the part, the individual particles of the material are then bonded together layer by layer. In this case, the material can be a metal, polymer powder, inorganic material, or other powdery or liquid material. The bonding of material particles can be achieved through physical or chemical processes, such as sintering, bonding, melting, solidification, or other bonding methods. Known methods include laser melting and laser sintering, as well as using ultraviolet radiation to harden liquid materials.

[0003] Unbonded material, or excess material, serves as a passive support structure during manufacturing, thus enabling, for example, overhanging geometries. The same applies to fluid or fluidizable materials, such as monomers that can be cured by ultraviolet light. After manufacturing, the unbonded material typically completely or at least partially fills the component and must be removed from the cavity of the final part after it has been removed from the material bed.

[0004] To date, the removal of excess material has been done manually or by mechanically rotating or shaking the component. However, such methods can be very time-consuming, especially when the component has cavities with complex geometries.

[0005] A method is known from document EP 3527304 A1, in which excess workpieces are simulated and an optimized emptying process is performed accordingly. However, sufficiently accurate simulation of the emptying process requires very high computational costs. Summary of the Invention

[0006] The object of this invention is to create a method and apparatus that enable the efficient separation of excess material from additively manufactured parts.

[0007] This objective is achieved by a method having features according to the invention, by an apparatus having features according to the invention, by a computer program product having features according to the invention, and by a computer-readable storage medium having features according to the invention.

[0008] To separate excess material from an additively manufactured part, spatially resolved structural data of the part is received. Using this structural data, the cavity of the part is divided into grid cells using a virtual spatial mesh, the spatial alignment of which is specified by an alignment specification. A discrete pattern of the part's spatial orientation is determined based on this alignment specification. Further, virtual material is simulatedly filled into the grid cells of the cavity. Then, using the structural data, a process of pouring virtual material out of the part is simulated for the spatial orientation limited to the discrete pattern, wherein a time series of the orientation limited to this discrete pattern is determined. Based on the simulated pouring process, the part is subsequently rotated sequentially to the orientation in the time series.

[0009] To perform the method according to the invention, an apparatus for separating material from additively manufactured parts, a computer program product, and a computer-readable, preferably non-volatile, storage medium are provided.

[0010] The methods and apparatus according to the invention can be executed or implemented, for example, by one or more computers, processors, application-specific integrated circuits (ASICs), digital signal processors (DSPs), and / or so-called "field-programmable gate arrays" (FPGAs).

[0011] The advantages of this invention are particularly evident in that by restricting spatial orientation to a discrete pattern, the computational cost required for simulation can typically be significantly reduced. In most cases, these simulation results are sufficiently accurate, provided that the actual movement of the component is also restricted to this discrete pattern. Furthermore, by simulating the pouring process and, based on this, moving the component to a favorable pouring posture, excess material can typically be separated from the component very effectively.

[0012] Advantageous implementations and extensions of the invention are described below.

[0013] According to an advantageous embodiment of the invention, the simulated corresponding motion steps can be limited to discrete motion of the virtual material from the corresponding grid cell to its adjacent grid cells. The discrete motion of the virtual material can be simulated here by moving virtual material particles from the corresponding grid cell to their respective adjacent grid cells. In particular, motion within the grid cell and / or continuous motion of the virtual material can be excluded.

[0014] Accordingly, the direction of motion of virtual materials in the simulation can be restricted to the direction from the corresponding grid cell to its adjacent grid cell.

[0015] In particular, a grid-based three-dimensional cellular automata can be used to perform the simulation.

[0016] Restrictions on discrete motion from grid cell to grid cell, restrictions on the direction of discrete motion, and / or the use of cellular automata often greatly simplify the simulation and typically lead to a significant reduction in the required computational cost. In fact, it has been shown that in many cases, computational cost can be reduced by at least an order of magnitude.

[0017] According to another advantageous embodiment of the invention, the direction guiding from a mesh cell to an adjacent mesh cell can be incorporated into a discrete pattern of spatial orientation. For these directions, in many cases, material motion can already be simulated sufficiently accurately and with minimal computational cost using simple simulation models.

[0018] Discrete patterns can preferably be formed by spatial orientations that are discretized in increments of approximately 45°.

[0019] Furthermore, it is possible to identify mesh cells that are filled with more virtual material than their neighboring mesh cells. The directions guiding from the identified mesh cells to their neighboring mesh cells can therefore be incorporated into the time series of these orientations.

[0020] According to an advantageous extension of the invention, material movement can be detected separately by sensors for each orientation of the determined time series. As a result of detecting a decrease in material movement, a trigger signal can then be generated, causing the component to rotate to the next orientation in the time series. In particular, the trigger signal is considered to be preferential over the simulated pouring process. By preferentially triggering rotation to the subsequent orientation, delays caused by an overestimation of the pouring duration due to simulation can be avoided. Attached Figure Description

[0021] Embodiments of the present invention will now be explained in more detail with reference to the accompanying drawings. Here, they are illustrated schematically:

[0022] Figure 1 The apparatus according to the invention is explained.

[0023] Figure 2 The simulation of the pouring process is explained.

[0024] Figure 3 The discrete patterns of the directions guided to adjacent grid cells are elucidated, and

[0025] Figure 4 A flowchart illustrating the method according to the present invention is provided.

[0026] If the same or corresponding reference numerals are used in the accompanying drawings, then the same or corresponding entities are indicated therefrom, which are preferably implemented or realized as described in conjunction with the relevant figures. Detailed Implementation

[0027] Figure 1 The apparatus A according to the invention is illustrated schematically for separating excess material WS from an additively manufactured part BT, i.e., manufactured by an additive manufacturing method. The part BT is preferably manufactured by a 3D printer in a powder bed method, wherein, as described at the outset, the individual particles of the powdered or fluid material are interconnected layer by layer. The excess material WS, which is not connected during the additive manufacturing process, should be removed accordingly. Figure 1 The part BT is shown after the layered manufacturing process is completed but before the removal of excess unconnected material WS.

[0028] The device A has a vibration device SV, a positioning device PV, a controller CTL, and a sensor system S.

[0029] Component BT is mechanically coupled to vibration device SV, which in turn is mechanically coupled to positioning device PV. Positioning device PV, preferably designed as a robotic arm, is used to position component BT in different orientations. Orientation indicates the spatial state of the component and includes its position and spatial orientation or alignment. The component can therefore be moved or rotated to different positions and spatial orientations via positioning device PV. According to the invention, the spatial orientation of component T directed by positioning device PV is limited to discrete patterns of directions. These discrete directional patterns are advantageously formed by spatial orientation, with angular values ​​discretized substantially in 45° increments.

[0030] A vibration device SV is used to subject a component BT to mechanical vibration. The amplitude, frequency, and / or direction of the mechanical vibration are preferably variable in this case. A positioning device PV allows the component BT to rotate and translate together with the vibration device SV about one or more axes of rotation. The component BT can preferably be removed from a 3D manufacturing environment, such as a 3D printer, and fixed to the vibration device SV by means of the positioning device PV, which is designed as a robotic arm. The positioning device PV and the vibration device SV are part of a motion device BV for moving the component BT.

[0031] The sensor system S, the positioning device PV, the vibration device SV, or the motion device BV are coupled to the controller CTL. The controller CTL is used to control the motion device BV, that is, to control the positioning and alignment of the component BT by the positioning device PV, and to control the vibration to be caused by the vibration device SV.

[0032] The motion device BV is specifically controlled in such a way that components BT rotate successively to different orientations in a discrete space, where excess material WS is emptied as quickly and / or as completely as possible. Ideally, the time required to empty excess material WS, i.e., to clear the accumulated material from the components, should be minimized.

[0033] This minimization is achieved by simulating one or more emptying processes. Here, the time requirements for emptying are determined through simulation for different emptying postures, specifically the discretized spatial orientation of component BT. The accumulated time requirements are then minimized using optimization methods. Here, the time series of emptying postures is determined in a targeted manner, resulting in the shortest possible emptying time. Numerous standard optimization methods are available for this type of optimization.

[0034] The simulation is performed using the three-dimensional cellular automata ZA of the controller CTL, based on the volumetric CAD model of the component BT. The volumetric CAD model is represented here by the spatially resolved structural data of the component BT, which is transmitted to the controller CTL. Based on this simulation, the controller CTL determines appropriate motion data BD for manipulating the positioning device PV and the vibrating device SV or the motion device BV. The orientation and position in space to be occupied by the component BT, as well as the frequency, amplitude, and / or direction of vibration, are quantified using the motion data BD. Preferably, the corresponding outward posture of the component BT can be specified by a positional description quantifying the position of the component BT, along with a discretized angular description quantifying the spatial orientation of the component BT. In particular, Euler angles can be used as the angular description.

[0035] The motion device BV receives motion data BD from the controller CTL and is controlled by the motion data BD. In this way, the positioning device PV moves the component BT to a dispensing posture to dispose of excess material WS as quickly and efficiently as possible. Furthermore, the vibration device SV causes the component BT to vibrate mechanically, thereby accelerating the dispensing process as much as possible. The material WS is dispensed from the component BT to a certain extent.

[0036] The sensor system S detects the material WS poured out during the pouring process or the material WS falling from the component BT. The sensor system S is specifically designed to detect the movement of the material WS caused by the pouring process, particularly the reduction of this movement. For this purpose, the sensor system S can have different sensors. Specifically, a scale W can be provided to continuously weigh the material WS poured out from the component BT. Alternatively or attached to the scale W, another volume sensor can be provided to detect the poured material WS. Furthermore, an optical sensor OS, such as a camera or grating, can be used to detect the poured material WS, preferably in conjunction with a laser L used to illuminate the poured material WS. Additionally, an acoustic sensor AS can be used to detect the noise of the material WS falling from the component BT or the noise of the material WS moving within the component BT. Furthermore, a motion sensor BS, such as a laser interferometer or a so-called MEMS (microelectromechanical systems), can be used to detect the movement of the material WS within the component BT.

[0037] Once the sensor system S determines that there is no longer any perceptible material movement, it can generally be assumed that the movement of the material WS inside component BT has also decreased. Once this state is reached, the emptying of the component and any possible tipping process inside component BT ceases for the current state. Since many simulations overestimate the duration of the tipping process, it has proven highly advantageous to guide the component to the next tipping position immediately after detecting a decrease in material movement, without waiting for the time period determined by the simulation to elapse.

[0038] Therefore, as a result of detecting a decrease in material movement, the sensor system S generates a trigger signal TR, which is transmitted to the controller CTL. Specifically, the trigger signal TR is generated once the weight of the poured material WS, measured by the scale W, remains at least approximately constant, i.e., no longer changes or no longer changes significantly. In the case of other sensors OS, AS, and BS, the trigger signal TR is generated once the detection result falls below a predetermined threshold.

[0039] As a result of receiving the trigger signal TR, the controller CTL causes the motion device BV to move the component BT to the next outgoing posture.

[0040] Figure 2 The simulation of the process of pouring excess material WS from component BT is explained.

[0041] To simulate the pouring process, method steps S1 to S7 are preferably executed by the controller CTL.

[0042] In method step S1, the controller CTL reads in the volume model CAD of component BT in the form of structural data obtained from spatial analysis. The volume model CAD can specifically exist as a so-called CAD model (CAD: Computer-Aided Design). Based on the volume model CAD, the cavity H of component BT is determined. After additive manufacturing of component BT, cavity H is filled with excess material WS and should be emptied through opening E of component BT.

[0043] To simulate the unloading process, component BT, and especially its cavity H, are divided into multiple mesh elements GZ based on structural data (CAD) using a virtual, three-dimensional mesh G. For clarity, in... Figure 2 Only a few of these grid cells in GZ are given attached figure labels.

[0044] In this embodiment, the mesh G is crossed by orthogonal basis vectors, such that the mesh element GZ is cuboid or cubic in shape. Alternatively or additionally, a tetrahedral or hexahedral mesh G may also be provided. The alignment of the mesh G defined by the basis vectors relative to the component G is specified by the alignment specification RA, preferably relative to the ground-fixed coordinate system of the component BT. The alignment specification RA may in particular include one or more of the basis vectors.

[0045] The discrete pattern R of the spatial orientation of component BT is determined based on the alignment specification RA. The discrete pattern R is derived from... Figure 3 To clarify, in this embodiment, the discretization mode R follows a preferred simulation direction pre-defined by the grid G. Given a pre-defined grid G ​​for discretizing component BT, it is certainly possible to simulate, particularly well and / or easily, those material movements in component BT that originate from grid cell GZ and proceed to adjacent grid cells BGZ. Conversely, movements in other directions must be approximated across cells or handled through a more refined simulation of material movements within the grid cells. However, this significantly increases the computational cost required.

[0046] Therefore, the discrete pattern R is precisely formed by, or constrained by, the directions from one mesh cell GZ to the adjacent mesh cell BGZ. In the case of a two-dimensional mesh, the corresponding (internal) mesh cell GZ is surrounded by eight adjacent mesh cells BGZ. In this case, the pattern R will consist of those discrete directions that are tilted with a 45° step relative to the basis vectors of the two-dimensional mesh. In the true three-dimensional case, the corresponding (internal) mesh cell GZ is surrounded by 26 adjacent mesh cells BGZ. As long as one direction in the three-dimensional mesh G can be specified by two angle values, the discrete pattern R in the three-dimensional case can be formed by the following 26 directions, whose angle values ​​are tilted with a 45° step relative to the mesh planes spanned by the two basis vectors of the mesh G.

[0047] exist Figure 2 In the further method step S2 shown, the path length of the path extending from the cavity H to the opening E is determined based on the volume model CAD for the corresponding mesh element GZ. The corresponding path length is... Figure 2 The upper right portion is shown in shaded areas, with darker areas positioned closer to the opening E than brighter areas. Advantageously, the path length can be determined using a so-called fast-travel method, which identifies the shortest path from each grid cell GZ to the opening E.

[0048] Furthermore, for the corresponding grid cell GZ, a corresponding local tilting direction is selected or determined from the discrete mode R. This direction leads from the grid cell GZ to the grid cell BGZ among these adjacent grid cells BGZ that has the shortest local path length to the opening E. The determined tilting direction and the determined shortest local path length are assigned to the corresponding grid cell GZ.

[0049] In method step S3, based on the volumetric CAD model, the cavity H is completely or partially filled with virtual material particles (VWP) in a simulated manner. In this case, the virtual material particles (VWP) can represent many real material particles in the simulation. When using other simulations, virtual material filling can be achieved in other ways accordingly.

[0050] Furthermore, the counter K used for the time series of the pour-out posture AP(K) to be determined is initialized to K=1. As described above, such a pour-out posture AP(K) specifies the position of component BT and its orientation in discretized space, respectively.

[0051] In method step S4, a favorable pouring posture AP(K) and appropriate vibration of the component BT are determined for the current filling state of the component BT with virtual material particles VWP. The selection of the current pouring posture AP(K) determines the emptying process, where gravity is a driving force and is supported by the shaking of the component BT. The selection of the pouring posture AP(K) is based on the analysis of the current filling level of the component BT and the determined path length and local pouring direction. In this case, the pouring posture AP(K) specifically includes the orientation and position of the component BT.

[0052] To determine a favorable pouring posture AP(K), a mesh cell GZ is sought that currently contains more virtual material particles (VWP) than its neighboring mesh cell BGZ and is assigned a short path length to the opening E, particularly as short as possible. If necessary, a pouring parameter can be determined for the corresponding mesh cell GZ, where the number of virtual material particles (VWP) currently contained in that mesh cell is calculated together with the assigned path length, for example, in the form of a weighted sum. Such a pouring parameter is preferably increased by a shorter path length and a larger number of currently contained virtual material particles (VWP), and otherwise decreased. In this case, the mesh cell with the highest pouring parameter can be selected from the mesh cells GZ.

[0053] Then, the local tilting direction assigned to the found or selected mesh cell is chosen as the global tilting direction. Based on the chosen tilting direction, a tilting posture AP(K) is defined, by which the component BT is oriented in such a way that the global tilting direction points downwards in the direction of gravity. This also applies to the orientation specified by the tilting posture AP(K), provided that the local tilting direction is limited to the discrete mode R.

[0054] Furthermore, motion data quantified for the determined pouring posture AP(K) were determined. Additionally, other motion data was determined that could cause the component BT to vibrate during the accelerated pouring process. This other motion data could specifically specify the amplitude, frequency, and direction of these vibrations. The determined motion data was summarized in a dataset BD(K), which quantified the specific motion for component BT.

[0055] In method step S5, the physical simulation of the process of virtual material particles VWP being poured out of the opening E of component BT caused by this specific motion is performed using a three-dimensional cellular automaton ZA based on the volume model CAD. For each motion step in the simulation, the cellular automaton only considers or simulates the discrete motion of the material particle VWP from the corresponding mesh cell GZ to its adjacent mesh cell BGZ. This simplification can typically significantly reduce the computational cost required. Accordingly, for each motion step in the simulation, the direction of motion of the material particle VWP is limited to the direction from the corresponding mesh cell GZ to the adjacent mesh cell BGZ, i.e., the discrete pattern R.

[0056] Within the scope of the simulation, component BT is virtually brought into the pouring posture AP(K) determined in method step S4 and subjected to vibration. In particular, the distribution of virtual material particles VWP in component BT caused by motion is simulated. As described above, the pouring process is driven by gravity and supported by vibration.

[0057] As a result of the simulation of a specific motion of component BT, namely a specific pouring process in the pouring posture AP(K), the temporal variation of the distribution of virtual material particles VWP in component BT derived therefrom is determined. The current pouring posture AP(K) and the resulting variation of the distribution of virtual material particles VWP are assigned to the corresponding motion of component BT.

[0058] In step S6, it is checked whether cavity H has been virtually emptied. For this purpose, a target value for emptying can be pre-defined, such as the remaining fill level; if this remaining fill level is below, component BT is considered emptied or ready for use. If cavity H has not yet been emptied, counter K is incremented and the process jumps back to method step S4, where the next pouring posture AP(K+1) is determined. Otherwise, method step S7 is executed.

[0059] In method step S7, the simulation results are recorded and a control file is generated based on them. The control file contains the time series of the determined outward postures AP(1),...,AP(N), the associated motion data BD(1),...,BD(N) used to manipulate the motion device BV, and other simulation results. In the current embodiment, the control file for the Kth simulated motion of component BT includes: the associated outward posture AP(K), the motion data BD(K) for the specified motion, and the change in the distribution of virtual material particles VWP in component BT derived from this motion.

[0060] Figure 4 A flowchart of the method according to the present invention for separating excess material WS from component BT is shown.

[0061] In the initial method step S10, the volumetric CAD model is simulated through method steps S1 to S7, as combined with... Figure 2 The described pouring process. As a result of the simulation, a control file is generated as described above, which contains the time series of the determined pouring postures AP(1), ..., AP(N), the associated motion data BD(1), ..., BD(N) for manipulating the motion device BV, and the temporal variation of the distribution of virtual material particles VWP in the component BT for the corresponding pouring posture AP(K). The simulation can preferably be performed in advance or offline.

[0062] Before executing the next method step S11, the counter K for the pouring posture is initialized to K=1.

[0063] In method step S11, based on motion data BD(K), the motion device BV is caused to move component BT to an outward posture AP(K) that is discretized in terms of its orientation and cause it to vibrate. The motion of component BT, or its temporal variation, follows the temporal variation of the simulated material distribution.

[0064] In a further method step S12, the sensor system S continuously detects the material WS being poured out from the component BT in the pouring posture AP(K)—as in combination with… Figure 1 As explained. In this case, in particular, the movement of the poured material WS is detected.

[0065] In the relevant method step S13, it is checked whether the detected material motion has weakened, i.e., significantly weakened or stopped. Once it is determined that there is no longer any perceptible material motion, a trigger signal TR is generated—as in combination with... Figure 1The described process is transmitted to the controller CTL. The trigger signal TR causes the controller CTL to perform a clear check S14 and, if necessary, jump back to method step S11. On the other hand, if no reduction in material movement is detected, the process returns to method step S12.

[0066] Within the scope of the emptying check S14, it is checked whether component BT has been emptied. This check can be performed, in particular, by comparing the weight of the poured-out material WS with the weight of the material WS initially present in component BT. For example, the latter can be determined based on the solid model CAD. If the check concludes that component BT has been emptied, then the separation according to the invention is successfully completed and the target state ST is reached.

[0067] On the other hand, if the check finds that component BT has not been cleared, the counter K is incremented and the process returns to step S11, where component BT is moved to the next discharge position AP(K+1). If necessary, a predetermined time period can be waited before starting the movement to the next discharge position AP(K) to ensure or increase the probability that any remaining movement of the material WS inside component BT will also reach a stop.

[0068] As described above, the trigger signal TR causes a jump back to method step S11. In this case, the trigger signal TR is taken into account over time the simulated change in material distribution. Specifically, instead of waiting for the duration determined by simulation until the material motion decays in the pouring posture AP(K) to pass, the motion is based on the actual measured weakening of material motion. That is, unlike the simulation-driven manipulation of the motion device BV by the controller CTL, once the sensor system S determines that no more perceptible material motion is occurring, the motion is prompted to the next pouring posture AP(K+1). The pouring process can often be significantly accelerated, as the decay time is often overestimated in simulations.

[0069] Due to the limitation of discrete patterns of several spatial orientations of the component BT according to the present invention, the computational cost required for simulation can generally be significantly reduced. This is particularly applicable because the pattern is specifically coordinated with the grid G ​​on which the simulation is based. It has been shown that the simulation is particularly accurate if the actual pouring direction or orientation of the component lies in the pattern induced by this grid. Therefore, high simulation accuracy can also be guaranteed by limiting the pouring posture to be taken for emptying the component BT to the same pattern.

Claims

1. A method for separating excess material (WS) from an additively manufactured part (BT), wherein: a) Receive the spatially resolved structural data (CAD) of the component (BT). b) Using the structural data (CAD), the cavity (H) of the component (BT) is divided into grid cells (GZ) using a virtual spatial grid (G), wherein the spatial alignment of the grid (G) relative to the component (BT) is specified by an alignment specification (RA). c) Determine the discrete pattern (R) of the spatial orientation of the component (BT) according to the alignment specification (RA). d) Simulatedly fill the mesh cells (GZ) of the cavity (H) with virtual material (VWP). e) Using the structural data (CAD) for spatial orientation limited to the discrete pattern (R), simulate the process of pouring the virtual material (VWP) from the component (BT), wherein a time series of orientations (AP(1), ..., AP(N)) limited to the discrete pattern (R) is determined, and f) According to the simulated pour-out process, the component (BT) is rotated sequentially to the orientation (AP(1), ..., AP(N)) in the time series.

2. The method according to claim 1, characterized in that, During the pouring process, the component (BT) is rotated to the orientation (AP(1), ..., AP(N)) in the time sequence by means of a motion device (BV) and / or the component (BT) is subjected to mechanical vibration.

3. The method according to any one of the preceding claims, characterized in that, The corresponding motion steps in the simulation are restricted to discrete motions of the virtual material (VWP) from its corresponding mesh cell (GZ) to its neighboring mesh cell (BGZ).

4. The method according to claim 1 or 2, characterized in that, In the simulation, the direction of motion of the virtual material (VWP) is restricted to the direction from the corresponding grid cell (GZ) to its adjacent grid cell (BGZ).

5. The method according to claim 1 or 2, characterized in that, The simulation was performed using a grid-based three-dimensional cellular automaton (ZA).

6. The method according to claim 1 or 2, characterized in that, The direction from a grid cell (GZ) to an adjacent grid cell (BGZ) is incorporated into the discrete pattern of orientation in the space.

7. The method according to claim 1 or 2, characterized in that, The discrete pattern is formed by spatial orientations discretized by angle values ​​in increments of approximately 45°.

8. The method according to claim 1 or 2, characterized in that, Identify mesh cells (GZs) that are filled with more virtual material (VWP) than their neighboring mesh cells (BGZs). Furthermore, the direction from the determined grid cell (GZ) to the adjacent grid cell (BGZ) is incorporated into the time series of the orientation (AP(1),...,AP(N)).

9. The method according to claim 1 or 2, characterized in that, For the corresponding orientation (AP(K)) of the determined time series: - Detect material movement using sensors -A trigger signal (TR) is generated as a result of detecting a decrease in material motion, and - The trigger signal (TR) causes the component (BT) to rotate to the next orientation (AP(K+1)) in the time series.

10. An apparatus (A) for separating excess material (WS) from an additively manufactured part (BT), the apparatus being configured to perform the method according to any one of the preceding claims.

11. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 9.

12. A computer-readable storage medium comprising the computer program product according to claim 11.

Images