Octree data structure in additive manufacturing

Abstract

A computer-implemented method of assisting, operating, monitoring and / or controlling an additive manufacturing process, the method comprising the steps of obtaining operational data (OD) captured during an additive manufacturing process, assigning the operational data (OD) to nodes (63) in a first layer (L3) of a first octree (60) on a first computing device, the method includes aggregating operational data (OD) of nodes (63) in a first layer into aggregated data (AD), and allocating the aggregated data (AD) to nodes (62) of a second layer (L2) of a first octree (60), where the second layer (62) is above the first layer (63) in the first octree (60), transmitting and / or loading, for example, a selected layer in the first octree into a second octree on a second computing device, wherein the second octree comprises fewer layers than the first octree.

Description

Technical Field

[0001] The present disclosure relates to additive manufacturing. Background Art

[0002] In additive manufacturing, such as laser powder bed fusion (Lb-PBF), wire arc additive manufacturing (WAAM) or binder jetting (BJ), different types of defects may occur, such as porosity, slag / foreign inclusions, lack of fusion / delamination, incomplete penetration, cracks, undercuts, spattering, burn-through / keyholes / heat build-up, geometric / form deviations, deformation, tool wear and balling. These defects reduce the quality of the object and, as a result, the object may be scrapped. Production scrap leads to high costs, especially when defects are discovered after the part is finished, which may take several days. In addition, post-processing quality assurance is expensive and cannot be applied to large objects. Summary of the invention

[0003] If a defect has been detected during the additive manufacturing process, the process can be stopped and / or a compensation strategy can be applied. Therefore, it is obvious that in-situ monitoring is needed to reduce production costs, improve object quality and / or reduce production waste. Therefore, a suitable process monitoring system is necessary. This faces several challenges. For example, an additive manufacturing process performed by an additive manufacturing system, for example, including multiple sensors, may have a long processing time, which may result in a large amount of operating data being accumulated and / or an excessive size of one or more data files including operating data. Therefore, collecting and / or archiving operating data is very cumbersome and / or requires a lot of effort for data filtering, data aggregation and data storage. In addition, it is necessary to remotely access the operating data, for example, by an operator, such as a quality engineer and / or a machine operator. This requires real-time and / or online monitoring of operating data, especially quality data. Furthermore, the operating data needs to be consistent with respect to time and / or position in the object.

[0004] It is therefore an object of the present invention to address these challenges and to improve the provision of operational data in additive manufacturing processes.

[0005] This object is achieved by the independent claim. Advantageous embodiments are provided in the dependent claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 An additive manufacturing system for producing an object is shown.

[0007] Figure 2 An additive manufacturing system for assisting, operating, monitoring and / or controlling an additive manufacturing process is shown.

[0008] Figure 3A scheme for processing operational data at the back end of an additive manufacturing system and visualizing the operational data at the front end is shown.

[0009] Figure 4 Operational data acquired during the additive manufacturing process is shown.

[0010] Figure 5 The recursive subdivision of a cubic space into cubic subspaces is shown.

[0011] Figure 6 A hierarchical tree structure representing a cube space and its subspaces in the form of an octree is shown.

[0012] Figure 7 An octree and operational data associated with the octree nodes are shown.

[0013] Figure 8 and Fig. 9 Operational data processing schemes for different additive manufacturing processes are shown.

[0014] Fig.10 A visualization showing aggregate data for nodes of the first layer of the octree.

[0015] Fig.11 A visualization showing aggregate data for nodes of the second layer of the octree.

[0016] Fig.12 A visualization of the aggregated data of an octree node and an overlaid visualization of the data model of an object are shown.

[0017] Fig.13 A visualization sequence of aggregated data is shown.

[0018] Fig.13a and Fig.13b The first octree and the second octree are shown respectively.

[0019] Fig.14 Defects in an object during the additive manufacturing process are shown.

[0020] Fig.15 A visualization of the defect is shown.

[0021] Fig.15a and Fig.15b The creation of quality indicators based on measurement data from different sensors is shown.

[0022] Figures 16 to 31 Exemplary method steps for an additive manufacturing process are shown. DETAILED DESCRIPTION

[0023] Figure 1An additive manufacturing system is shown. Examples of additive manufacturing processes include fused deposition modeling, fused wire manufacturing, robotic casting, electron beam freeform manufacturing, direct metal laser sintering, electron beam melting, selective laser melting, selective thermal sintering, selective laser sintering, and stereolithography. Many of these processes involve depositing and melting / softening / bonding materials layer by layer at selective locations to build the desired object. A non-exhaustive list of exemplary materials that can be used for additive manufacturing includes metals, thermoplastics, and ceramics.

[0024] Additive manufacturing processes typically employ additive manufacturing machines configured to perform their respective processes, for example, including tools or deposition heads that follow tool paths. However, it should be understood that some additive manufacturing machines are also capable of performing machining / subtractive manufacturing processes and correspond to hybrid additive / subtractive manufacturing machines.

[0025] The additive manufacturing system 100 may include at least one processor 102 that is operably configured to generate instructions 104 that can be used by an additive manufacturing machine 106 to control the operation of the additive manufacturing machine to build an object via at least additive manufacturing. In an exemplary embodiment, one or more data processing systems 108 may include at least one processor 102. For example, the external data processing system may correspond to a workstation having various software components (e.g., programs, modules, applications) 110. The software components 110 are operably configured to cause the at least one processor 102 to perform functions and operations that implement the instructions 104. In an exemplary embodiment, the instructions 104 may be in a G-code format or other numerical control (NC) programming language format. Examples of G-code formats include formats that comply with standards such as RS-274-D, ISO 6983, and DIN 66025.

[0026] The additive manufacturing machine 106 may include a deposition head 112 and a build plate 114. The deposition head 112 may include an integrated heat source 116, such as a laser (or electrode), which is operable to melt / soften a material 118, such as a powdered metal (or metal wire) provided by the deposition head. The additive manufacturing machine 106 is operable to build an object 120 from the build plate 114 via depositing a layer on top of a layer 122 of material 118 in a build direction 130. In this example, the deposition head 112 is operable to simultaneously output and melt / soften a continuous stream of material bonded to the build plate and / or previously applied layers constituting the object. In the examples described herein, the material may correspond to a metal (in powder or wire form). However, it should be understood that in alternative embodiments, the additive manufacturing machine may be operable to deposit other types of materials, such as thermoplastics.

[0027] The additive manufacturing machine is operable to move the deposition head horizontally (in the XY direction) and vertically (in the Z direction). The additive manufacturing machine is also operable to move the build plate (such as by rotating the build plate relative to one or more different axes). The additive manufacturing machine is operable to move the print head and / or the build plate relative to each other, depositing beads of material in a pattern, which build an object or a portion of an object in layers. The additive manufacturing machine may include a controller 124, which is operably configured to actuate hardware components (e.g., motors, circuits, and other components) of the additive manufacturing machine to selectively move the deposition head and / or the build plate to deposit materials in different patterns. Such a controller 124 may include at least one processor, which is operative to control hardware components (e.g., deposition heads and heat sources) of the additive manufacturing machine in response to software and / or firmware stored in the additive manufacturing machine. Such a controller is operable to directly control the hardware of the additive manufacturing machine by reading and interpreting the generated instructions 104. In an exemplary embodiment, such instructions may be provided to the controller or obtained by the controller via a network connection. In such an example, the controller 124 may include a wired or wireless network interface component operable to receive instructions. Such instructions 104 may come directly from the data processing system 108 over the network. However, in other examples, the instructions 104 may be saved by the data processing system on an intermediate storage location (such as a file server) that the additive manufacturing machine can access.

[0028] The software 110 is operable to receive a (3D) model 126 of an object and generate instructions 104 based on the (3D) model 126 of the object. In an example, the software may include a CAM software component that facilitates generating instructions 104 from the (3D) model. For example, such a (3D) model may correspond to a CAD file such as a STEP or IGES format. In an exemplary embodiment, the software component 110 may include a CAD / CAM / CAE software suite such as the NX application available from Siemens.

[0029] In addition to generating G-code for the (3D) model, the exemplary CAM software component may also be configured to cause the data processing system to output a visual representation of the object 120 based on the (3D) model on a display operatively connected to the processor. In addition, the CAM component may be configured to cause the data processing system to provide a graphical user interface for use with input from an input device providing parameters that can be used to generate instructions 104 for building the object. Such user-provided parameters may include a build direction to be associated with the object (or portions of the object), the thickness and width of each bead of deposited material, the speed at which the material is deposited, the mode in which the head travels relative to the build plate to deposit material to the object, and any other parameters that define characteristics for the operation of the additive manufacturing machine.

[0030] Defects may occur during the additive manufacturing process. To detect these defects, the machine operator is responsible for process monitoring of the machine throughout the process. The operator monitors operational data such as process data and / or parameters such as laser power, current or voltage. In additive manufacturing machines, the video stream from the camera can be used to visually monitor the process. However, this monitoring is done manually based on operator experience. If the operator detects a defect, the machine can be stopped. However, many high-frequency patterns in the data are not detectable by humans, and the machine user's attention tends to fade after a while. In addition, extraordinary experience is required to detect defects. Therefore, many defects cannot be detected in situ and a quality assurance step after the manufacturing process must be added. Due to the large amount of data acquired, current systems usually display offline data after the process is finished. Other solutions may provide online systems with some degree of data aggregation, but this is highly specialized for each processing technology and sensor class. To date, a complete monitoring system for in-situ part quality monitoring (i.e., during the additive manufacturing process) has not been realized.

[0031] Now go to Figure 2 , Figure 2 An additive manufacturing system 100 for assisting, operating, monitoring and / or controlling an additive manufacturing process is shown. The system includes an additive manufacturing machine 106 and a first edge device E1 and a second edge device E2. The edge devices E1, E2 may be devices for data processing, and to this end, may include one or more processors that may execute instructions according to one or more software programs stored on the edge devices E1, E2. The system may further include sensors M1, ..., M6 for monitoring the additive manufacturing process. One or more sensors M1, ..., M6 may capture one or more physical properties associated with the additive manufacturing process of an object. One or more sensors may provide measurement data associated with the captured one or more physical properties. Various measurement techniques may be used, such as vision, ultrasound, liquid penetrant, magnetic particles, eddy currents, radiography and / or other applicable measurement techniques. As Figure 2As shown, current sensors, acoustic sensors, cameras, pyrometers, spectral emission sensors and / or acoustic sensors may be used to monitor the additive manufacturing process. The sensors may be communicatively coupled to the first edge device E1. Each sensor may, for example, provide its measurement data to the first edge device E1 via a network (such as a data bus). The first edge device E1 may also be communicatively coupled to the additive manufacturing machine 106, for example, a controller of the additive manufacturing machine. In addition, the additive manufacturing machine 106 and / or the first edge device E1 may be communicatively coupled to the second edge device E2. The first edge device E1 may be used to process measurement data obtained from one or more sensors and / or to process timestamps and / or spatial positions obtained from the additive manufacturing machine 106. The second edge device E2 may include and / or be communicatively coupled to a display D for displaying measurement data and / or timestamps and / or spatial positions on the display D. As described in the present invention, the display D may also be used to display visualizations of operational data and / or aggregated data. The second edge device E2 may also be communicatively coupled to a remote computing platform O, for example, a cloud computing platform. The second display D2 communicatively connected to the remote computing platform can also be used to display (visualization of) operational data and / or aggregated data located in the cloud. The operational data may include measurement data and / or may additionally include one or more timestamps and / or spatial locations. Therefore, the measurement data can be enhanced with time and / or spatial data. Based on time and / or spatial data, the measurement data can then be indexed as described in the present invention and can be organized into a data structure in the form of a hierarchical tree (such as an octree). The operational data may also include one or more quality indicators and / or anomaly scores associated with the measurement data.

[0032] Figure 3 A scheme for processing operational data in a backend B of an additive manufacturing system and visualizing the operational data in a frontend F is shown. The backend B may include a first edge device that obtains measurement data and distributes the operational data to (leaf) nodes of a hierarchical tree structure. The first edge device may also aggregate operational data and distribute the aggregated operational data to nodes of a second layer of the hierarchical tree structure. To this end, the first edge device may index the measurement data according to a timestamp and / or spatial position obtained from the additive manufacturing machine. To this end, a Morton code of the measurement data may be used. For example, a vector comprising measurement data of one or more sensors may be formed. The components of the vector may represent measurement data of one of the sensors at a specific time and / or a specific space.

[0033] The front end F may include a second edge device that may be communicatively coupled to the first edge device using one or more application programming interface APIs, such as a RESTful API. Therefore, the second edge device may be used to select operational data corresponding to a spatial region, a temporal region and / or a sensor from a plurality of sensors. Therefore, as described in the present invention, upon selection and / or transmission of an indication, the second edge device may receive data, such as measurement data, operational data and / or aggregated data, from the first edge device. The operational data and / or aggregated data obtained by the second edge device may then be used as a basis for visualization. Depending on the circumstances, the second edge device and / or a display coupled to the second edge device may have specific storage capabilities and / or specific display capabilities. Therefore, the amount of data that may be stored on the second edge device and / or visualized via the second edge device may depend on the capabilities of the second edge device and / or the display.

[0034] Therefore, to achieve process monitoring and / or anomaly detection in the additive manufacturing process (in real time), an edge computing approach is adopted. The system can include two types of edge devices, Figure 2 The first edge device and the second edge device in are exemplarily represented. Any number of edge devices of the first type are used to process and / or monitor received measurement data with the aid of machine learning techniques in a computing cluster and may include a GPU for accelerating anomaly detection operations. The second type may be used for visualization and / or for connecting the additive manufacturing system with a cloud computing platform for process monitoring and / or visualization.

[0035] Figure 4 Measurement data 40 acquired during the additive manufacturing process is shown. The measurement data 40 may be raw data obtained from each sensor. The measurement data may be stored as a time series in a database. The database may be part of a first edge device. During the additive manufacturing process, a large amount of data is generated, which therefore requires a memory of corresponding capacity to store the data. In order to retrieve the data and / or visualize the data, efficient data management is required.

[0036] Figure 5 shows the recursive subdivision of the cubic space 50 into cubic subspaces 51, and Figure 6 A hierarchical tree structure 60 in the form of an octree representing the cubic space 50 and its subspaces 51 is shown. Figure 5 The cubic space and its subspaces in are equivalent to Figure 6 The cube space represents measurements, operational data, and / or aggregate data, as appropriate.

[0037] Data aggregation can be performed by means of a hierarchical tree structure 60, for example in the form of an octree. A stack of octrees or quadtrees can be used for data aggregation, i.e. for combining measurement data captured, for example, in the form of a time series with their position in the object. An octree is a hierarchical tree structure 6, in which each internal node has (exactly) eight child nodes. Thus, an octree can be used to aggregate the three-dimensional space by recursively subdividing it into Figure 5 The eight octants shown are used to divide the three-dimensional space. An octree is a three-dimensional analogue of a quadtree. Using an octree to organize and / or store measurement data (e.g., for visualization) provides the possibility of reducing the data displayed to the user without the need for additional aggregation, filtering, or deletion of measurement data. The interface between the first edge device and the second edge device can be used to transmit data required for visualization. If necessary, the octree can be used for data aggregation, for example, for quality reporting.

[0038] The root node 61 or root octant spanning the entire cubic space or domain is located at level 0 of the tree. Each child node 62 or child octant of the cubic space is one level higher in the tree structure than its parent node. The smallest octant or leaf node 63 may correspond to measurement data or raw data from a sensor. The layer L2 between the root node L2 and the smallest node L3 or leaf may correspond to aggregated data. Thus, different degrees of data aggregation are obtained. Different degrees of data aggregation can be obtained by different levels L1, L2, L3 of the hierarchical tree structure, in this case an octree.

[0039] Thus, using the hierarchical tree structure 60 enables storage and / or visualization of high frequency data, such as measurement data from one or more sensors, especially in the case of large amounts of measurement data such as is present in LB-PBF, and especially data with frequencies above 100 kHz.

[0040] Interface (such as Figure 3 The API shown in ) can be used to display operational data remotely with low latency without losing detailed information of the original data. If necessary, the spatial and / or temporal area of ​​interest can be evaluated by zooming in or filtering the visualization. Thus, spatial and / or temporal areas can be viewed. An area can be defined as one or more tool paths of a tool of an additive manufacturing machine, such as a deposition head. The tool path may include information from external or internal data sources. Internal data sources may include monitoring data (processed sensor data) that show, for example, abnormal behavior. External data sources may include simulation data (in which critical ranges in an object have been identified) or CAM / CAD data (through which objects in the same additive manufacturing process are separated, or through which specific requirements are defined, such as the stability and / or surface roughness of an object).

[0041] As described in the present invention, a software application running on a first edge device, such as type 1, performs the task of collecting, storing, and analyzing the received additive manufacturing process information, for example, with low latency. The first edge device can be connected to an additive manufacturing machine and one or more sensors, for example, via Ethernet. Data from the additive manufacturing machine and data from the sensor can be received by the first edge device, for example, via an IP protocol.

[0042] Thus, the proposed additive manufacturing system enables the implementation of a scalable, and preferably portable and / or containerized microservice that executes on any number of type 1 edge devices, thereby continuously receiving a data stream of measurement data and process parameters from the described sensors and additive manufacturing machines. Software applications can be implemented as microservices, which can be encapsulated with the help of container virtualization on top of an operating system to achieve scalability, portability and / or hardware abstraction. Inter-service communication can be achieved with the help of a central data bus. The received measurement data and additive manufacturing information can be published via the data bus and / or received from other subscribing microservices on other edge devices. In addition, microservices for receiving measurement data, analyzing measurement data, monitoring the additive manufacturing process, and storing measurement data can be utilized.

[0043] To reduce memory usage, for example on edge devices, the octree may additionally be connected to a database (DB), for example a time series database such as InfluxDB, where the measurement data is saved. Figure 7 Such an embodiment is shown. Figure 7 An octree 70 and operational data (OD), as well as aggregated data AD associated with a node 71 of the octree are shown. The octree position can be mapped to a timestamp via a position code, such as a Morton code, which can preferably be simultaneously saved in a database. For fast access, the most relevant sensor signals can be saved as a cache in a memory of the edge device, such as a RAM, as a hash map. If the user selects a specific sensor or sensor combination, for example including sensor values ​​and an aggregation method, such as tool path energy given by voltage, current and welding speed, the data can be loaded from the database or cache data. Therefore, different levels of detail are available for the area of ​​interest, where the level of detail can be obtained by querying the layers of the octree. Therefore, the solution provides a general and scalable method for in-process monitoring of additive manufacturing processes.

[0044] Thus, an additive manufacturing system including sensors and edge devices can be used in additive manufacturing (processes) for in-situ assisting, operating, monitoring and / or controlling the additive manufacturing process. The octree-based architecture and its capabilities allow for remote visualization of a large number of data points on a display. Additional benefits include scalability and flexibility in a multivariate sensor framework.

[0045] Octrees provide a mechanism for aggregating similar spatial data regions, which can save space and time. Many basic operations (such as neighborhood search) are implemented through tree operations such as tree traversal. Octrees and their octants can be represented as sorted sequences. Therefore, octrees can aggregate a large number of small volumes of space into a single larger representative volume. The octree provides a discrete point cloud spatial representation with semantically enhanced data.

[0046] Figure 8 and Fig. 9 OD processing schemes for different additive manufacturing processes are shown. For each object to be built, an octree 80 can be initialized based on the machine context. Therefore, each new object to be built can be associated with a new octree. Figure 8 and Fig. 9 As shown, data processing can be organized in different ways. For example, Figure 8 In the case of the WAAM additive manufacturing process / system / machine shown, the octree can be filled progressively. The machine context can be provided by the additive manufacturing machine. Binary values ​​indicating the start and stop of the entire manufacturing program R20, the start and stop commands R21 of the WAAM process, and the actual start and stop R22 of the WAAM process can be used. Based on the machine context, the octree 80 can be instantiated, preferably on the edge device E1, and the measurement data can be incrementally entered in the octree.

[0047] Similarly, in the case of SLM additive manufacturing processes, such as Fig. 9 As shown, the process start may serve as a trigger to instantiate the octree 80 , for example, on an edge device.

[0048] For example, a data item including a spatial position may be obtained from an additive manufacturing machine via an edge device. The spatial position x, y, z may be a coordinate of a coordinate system (such as a Cartesian coordinate system). The data item may also include payload data. The payload data may include measurement data from the additive manufacturing machine and / or from additional sensors.

[0049] For a WAAM additive manufacturing process, all coordinate data may be used to populate the octree, whereas in the case of an SLM additive manufacturing process, for example, the octree may be populated by first determining a quadtree 81 and stacking the quadtrees 81 to obtain an octree 82.

[0050] Fig.10 A visualization V1 of operational data of nodes of a first layer of an octree is shown. The first layer of the octree may include measurement data obtained by an edge device from one or more sensors. Quality indicators or anomaly scores may be determined and assigned to the measurement data and / or nodes of the octree. Then, as Fig.10As shown, the quality indicators and / or anomaly scores can be visualized. Color-based representations of the quality indicators and / or anomaly scores can be used for visualization. Fig.10 The visualization V1 shown in may correspond to the lowest level of the octree, ie comprising the raw measurement data or raw data.

[0051] Fig.11 A visualization V2 of aggregated data for nodes of a second layer of an octree is shown. As shown, (the visualization of) the second layer includes a smaller number of nodes than the first layer, and is layered (e.g., directly) above the first layer. The nodes of the first layer and their corresponding operational data have been combined (i.e., aggregated) to form new values ​​that are assigned to corresponding nodes of the second layer. To this end, one or more statistical properties of the operational data of the first layer have been determined, and the resulting data (i.e., aggregated data) has been assigned to the nodes of the second layer.

[0052] Fig.12 A visualization of the aggregated data of the octree nodes and an overlay of the data model of the object is shown V3. Thus, different degrees of data aggregation (via the octree) can be obtained, and a visual comparison with the original CAD file (i.e., the (3D) model) of the object can be obtained.

[0053] Fig.13 A sequence of visualizations of aggregated data is shown a), b), c), d). The visualization may depend on one or more characteristics of the display. The visualization may depend on the degree of detail necessary to monitor the additive manufacturing process in order to help a user detect defects in the additive manufacturing process. Thus, the resolution of the visualization may vary between different views, such as Fig.13a )to Fig.13 d) is shown as a series of views. Indicators can be used to retrieve the necessary aggregate data. Thus, the user can zoom in or out on the area of ​​interest. Views can be based on different levels of the octree. Fig.13a ) view, get the high level of the octree and visualize the aggregate data of that level, where Fig.13b ), Fig.13 c) and Fig.13 d), the resolution is increased by retrieving aggregated data from lower levels of the octree.

[0054] Fig.13a and Fig.13b The first octree 60a and the second octree 60b are shown respectively. Fig.13aAs shown, the first octree can be created on a device for data processing, that is, a first edge device or a first type of edge device. Therefore, the first octree can be located on the first device for data processing. The first edge device may include multiple layers L1, ..., L5. Among them, the lowest layer may be associated with the original sensor data. The middle layer may include aggregated data, where the degree of aggregation or detail depends on the position of the layer in the octree. The lowest layer with the lowest degree of aggregation and the highest layer with the highest degree of aggregation and the middle layer with successively higher / lower degrees of aggregation. As shown in the figure, the resolution of data aggregation between layers may depend on the number of layers of the octree 60a. More layers can provide finer resolution (fewer layers provide coarser resolution), such as finer resolution of spatial and / or temporal position. The number of layers of the first octree can be determined based on the required resolution and / or based on the processing power and / or storage capacity of the first processing device.

[0055] Now go to Fig.13b , a second octree 60b may be created on a second edge device (or typically a second device for data processing) or an operating device. Thus, the second octree may be located on the second edge device or the operating device. Now, the second octree may include a smaller number of layers than the first octree. The number of layers of the second octree may also be based on the processing power and / or storage power and / or display power of the second edge device and / or the operating device, respectively. Because the processing, storage and / or display capabilities of the first edge device on the one hand and the second edge device and / or the operating device on the other hand are different, it may be necessary to provide octrees with different numbers of layers on different devices. For example, layer L1' of the second octree 60b may correspond to layer L1 of the first octree 60a, layer L2' of the second octree 60b may correspond to layer L3 of the first octree 60a, and layer L3' of the second octree may correspond to layer L5 of the first octree 60a. In any case, other mappings between layers may be selected based on the required resolution and / or capabilities of the first edge device and / or the second edge device and the operating device, respectively. For example, only a predetermined number of (e.g., equidistant) layers of the first octree may be used for (creating) the second octree. Therefore, a first octree including a first number of layers on a first edge device and a second octree including a second number of layers on a second edge device and / or an operating device may be created, wherein the second octree may include fewer layers than the first octree. In addition, one or more layers may therefore be selected and transmitted from the first edge device to the second edge device and / or the operating device. Therefore, the second edge device and / or the operating device may load, for example, selected layers of the first octree into the second octree. As the case may be, it should be understood that the layers and the nodes of the layers may include allocated operating data and / or aggregated data.

[0056] Thus, the storage space (memory) is saved on the second edge device and / or the operating device, respectively. Nevertheless, the (aggregated) data associated with the layer and the node are still available, for example, for further processing or visualization by the second edge device and / or the operating device, respectively.

[0057] The resolution and / or number of layers of the second octree 60b can be determined based on the number of pixels of the display of the operating device. Based on the number of pixels, the number of layers can be determined for the second octree. For example, the maximum resolution of the octree can correspond to the number of pixels available on the display of the operating device. Additionally and / or alternatively, the number of layers of the second octree 60b can be determined based on the storage capacity of the second edge device and / or the operating device. Therefore, a series of visualizations (with different resolution levels) of operating data depending on the capabilities of the second edge device and / or the operating device can be provided based on the second octree.

[0058] Fig.14 Defects in an object during an additive manufacturing process are shown. The spatial correlation of defect correlations can be defined based on the dimensions of the weld bead during the welding process. These dimensions x, y, z can depend on the welding speed along the tool path, the substrate temperature, the heat input per unit length, and the material. A cross-sectional view of the double ellipsoid with the current working point and valid data points is shown in Fig.14 Shown in.

[0059] Fig.15 A visualization of the defect is shown. Fig.15 In the figure, anomaly scores are plotted in logarithmic colors along the tool path. Regions where defects occur have much higher anomaly scores than non-anomalous regions. In the region above the defect, the anomaly score increases significantly as the effect of the defect propagates.

[0060] Defects can be detected by setting a threshold for the anomaly score. Fig.15 In the visualization, the detected defects are visualized according to the selected threshold. For example, a user-set threshold can be used. Spatially correlated detected defects are colored in the same way. The expansion of each detected defect cluster can be monitored along the additive manufacturing process.

[0061] Fig.15a and 15b In general, the creation of quality indicators based on measurement data or operational data from different sensors is shown. As shown, measurement data from different sensors M1, ..., M4 may be available, for example, stored in a database. Now, as described in the present invention, the measurement data may be aggregated for each sensor separately. However, depending on the situation, the measurement data of different sensors may also be aggregated and / or combined, for example, in order to determine a quality indicator and / or anomaly score. For example, Fig.15aAs shown in , the measurement data of sensors M1, M2, M3 and M4 are combined to form a first quality indicator Q1. On the other hand, the measurement data of sensors M3 and M4 can be combined to form a second quality indicator Q2. Thus, a quality indicator can be determined based on the measurement data of different sensors. Furthermore, a plurality of quality indicators can be combined to form a (superior) quality indicator Q3. One or more quality indicators can be visualized on a display of the operating device. Additionally or alternatively, a notification can be issued to inform a user or operator about the quality indicator and / or a change in the quality indicator.

[0062] exist Fig.15b In the example, the quality indicator is shown by a corresponding signal light. Therefore, the quality indicator can take discrete values. Fig.15b As shown in , each sensor M1, M2, M3, M4 provides measurement data, wherein for each sensor, a quality indicator is determined (individually) based on the corresponding measurement data. The quality indicators Q1, Q2, Q3, Q4, Q5 can be combined to form a (superordinate) quality indicator Q5. One or more of the quality indicators Q1, ..., Q5 can be visualized. Based on the quality indicator Q5, a notification can be issued in order to inform a user or operator about the quality indicator Q5 and / or a change in the quality indicator Q5.

[0063] Figures 16 to 31 Exemplary method steps for an additive manufacturing process are shown.

[0064] Go to Fig.16 In a first step S1, operational data may be obtained. The operational data may include measurement data from one or more sensors. The operational data may also include spatial information, for example from an additive manufacturing machine. The spatial information may be used to construct an octree of the object. In addition to the machine context, the spatial and / or temporal context may affect the current deposition process, resulting in a spatial-temporal relationship of the data, since some defects may propagate from one layer of the object to subsequent layers. Thus, the operational data may include, for example, measurement data that has been enhanced or annotated with temporal and / or spatial information.

[0065] For each path increment of the additive manufacturing process, an item with a spatial index based on the Cartesian position of the data point can be added to a hierarchical tree structure, such as an octree. As a payload, measurement data, anomaly scores, defect types, defect identifiers, locations of previous data points and / or timestamps can be stored. Therefore, in step S2, operational data can be assigned to the nodes of the first layer of the hierarchical tree structure. The operational data can be used as the basis for one or more visualizations. In step S3, the operational data is aggregated. For example, the operational data of the nodes of the first layer are aggregated into aggregated data. In step S4, the aggregated data is assigned to the nodes of the second layer of the hierarchical tree structure. Among them, the second layer can be located above the first layer in the hierarchical tree structure.

[0066] Go to Fig.17 In step S5, a first spatial volume of the object is represented by nodes in the first layer, and in step S6, a second spatial volume of the object is represented by nodes in the second layer. The second spatial volume may be larger than the first spatial volume.

[0067] Go to Fig.18 In step S7, the aggregated data is assigned to nodes of the second level of the hierarchical tree structure, and in step S8, the aggregated data of the second level of the hierarchical tree structure is visualized. The visualization may be performed on a display of an operating device for controlling and / or monitoring the additive manufacturing process. Thus, the visualization may be presented to a user or operator of the additive manufacturing system / machine.

[0068] Go to Fig.19 In step S9, the operation data is aggregated. In step S10, the statistical characteristics of the operation data of the nodes of the first layer can be determined. In step S11, the statistical characteristics are used as aggregated data.

[0069] Steering Fig. 20 In steps S12 to S14, the operation data is aggregated, statistical characteristics of the operation data of the nodes of the first layer are determined, and the statistical characteristics are used as the aggregated data. According to step S15, the average value, median value, variance, minimum value, maximum value, sum and / or count of the operation data can be used as the statistical characteristics.

[0070] Go to Fig.21 , aggregated data can be obtained. Aggregated data can be assigned to nodes of the (first) middle layer of the hierarchical tree structure. In step S17, the aggregated data is further aggregated. Further aggregation of the aggregated data can also be based on statistical properties of the aggregated data of the (first) middle layer, such as mean, median, variance, minimum, maximum, sum and / or count. In step S18, the further aggregated data is assigned to nodes of the latter layer (i.e., the second (middle) layer) of the hierarchical tree structure.

[0071] Go to Fig. 22 , the operating data may be obtained, for example, by the edge device in step S19. In step S20, the operating data may be aggregated and thus the aggregated data may be obtained in step S21. As described above, further aggregation may be performed on all intermediate layers of the hierarchical tree structure. In step S22, an indication of a storage capacity and / or a display capacity of the operating device may be obtained, for example, by the edge device.

[0072] In step S23, the aggregated data may be transmitted to an operating device, for example including a memory for storing the aggregated data and / or a display for displaying the aggregated data. Based on the indication, the edge device may determine a corresponding layer of the hierarchical tree structure. The layer includes an amount of data or refers to an amount of aggregated data that may be stored in the memory and / or displayed on the display of the operating device.

[0073] Go to Fig.23 In step S24, an indication of the storage capability and / or display capability of the operating device may be transmitted, for example, (from the operating device) to the edge device. In step S25, the aggregated data may then be obtained by the operating device. In step S26, the aggregated data may then be visualized in a first view on the display. It should be understood that the aggregation may correspond to aggregated data of any intermediate layer of the hierarchical tree structure and / or aggregated data assigned to any intermediate layer.

[0074] Go to Fig.24 In step S24, the aggregate data is obtained by the operating device (e.g., from an edge device). The operating device may also include an edge device that includes and / or is operatively coupled to a display. In step S28, the aggregate data may be visualized on the display (in a first view).

[0075] Go to Fig.25 In step S29, aggregate data of nodes of the first layer are obtained by the operating device. In step S30, the aggregate data is visualized in a first view, such as a display. In step S31, aggregate data of nodes of the second layer are obtained by the operating device and visualized in a second view, such as on a display. Therein, a first indicator can be obtained by the edge device, which indicates the first layer / nodes of the first layer. Subsequently, a second indicator can be obtained by the edge device, which indicates the second layer / nodes of the second layer. The visualization of the first view can then be replaced by the visualization of the second view on the display.

[0076] Go to Fig.26 In step S33, a code value (e.g., a Morton code value) representing the volume of the space may be assigned to the operation data. In step S34, the code values ​​may be sorted according to the hierarchical tree structure. In step S35, the operation data may be assigned to (leaf) nodes in the hierarchical tree structure based on the code values.

[0077] Go to Fig. 27, a simulation of the additive manufacturing process can be performed. Thus, simulated operating data is obtained in step S36. In step S37, a coding value (e.g., a Morton coding value) can be assigned to the simulated operating data. In step S38, the coding values ​​are sorted according to the hierarchical tree structure, and in step S39, the simulated operating data is compared with the (actual) operating data.

[0078] Go to Fig.28 In step S40, a quality indicator may be determined based on the operational data. In step S41, one or more quality indicators may be assigned to one or more coded values ​​and / or nodes in the hierarchical tree structure.

[0079] Go to Fig.29 In step S42, one or more anomalies in the operational data may be determined. To this end, an anomaly score may be determined. In step 43, the one or more anomalies may be assigned to one or more coded values ​​and / or nodes in the hierarchical tree structure.

[0080] Go to Fig.30 In step S44, for example, a first software application on a first computing device (such as a first edge device) may determine based on the operational data, the quality indicator, and / or the one or more anomalies. In step S45, the quality indicator and / or the one or more anomalies may be preferably visualized on a display of a second computing device (such as a second edge device).

[0081] Go to Fig.31 , in step S46 operational data may be obtained and in step S47 a timestamp and / or a spatial position may be assigned to each measurement value of a sensor monitoring the additive manufacturing process.

[0082] The proposed aspects enable real-time process monitoring and anomaly detection for any number of sensors with low latency for multiple types of additive manufacturing systems. The proposed aspects provide the possibility of monitoring the quality of an object during operation (i.e., in situ and in near real time). Timestamps and spatial information are managed, for example, by a controller and / or edge device. As a result, the lag between each sensor time step is minimized. This enables the measurement data to accurately match the actual processing timestamps it measures. Different aspects make it possible to scale the architecture and allow additional sensors to be added. All data is highly synchronized, so defects can be accurately localized. The connector between the edge device and each sensor (type) enables a fast plug-and-play solution.

[0083] Therefore, a distributed system for visualization, for example on a tablet, PC, smartphone or other operating device, for example, based on the web is proposed, which allows data filtering and / or aggregation for data visualization, as well as low latency without discarding the original raw data. Furthermore, temporary and / or spatial region of interest data aggregation is proposed without the need for subsequent alignment. Furthermore, the use of octrees or stacked quadtrees as hierarchical tree structures allows the use of direct energy deposition techniques (WAAM, LMD, FDM, etc.) as well as bed-based (LB-PBF, EB-PBF, SLS, BJ, etc.) techniques. Furthermore, a combination of a time series database with an octree via a Morton code or other position code is proposed in order to handle large amounts of data.

[0084] Other implementations are described below:

[0085] A first embodiment includes a preferably computer-implemented method for assisting, operating, monitoring and / or controlling an additive manufacturing process, the method comprising the steps of obtaining operation data (OD) captured during the additive manufacturing process, assigning the operation data (OD) to nodes (63) in a first layer (L3) of a hierarchical tree structure (60), aggregating the operation data (OD) of the nodes (63) in the first layer into aggregated data (AD), and assigning the aggregated data (AD) to nodes (62) in a second layer (L2) of the hierarchical tree structure (60), wherein the second layer (62) is located above the first layer (63) in the hierarchical tree structure (60).

[0086] A second embodiment comprises a method according to the preceding embodiment, wherein the nodes (63) in the first layer (L3) represent a first spatial volume of an object, for example a first spatial volume of an object to be manufactured, and the nodes (62) in the second layer (L2) represent a second spatial volume of the object, wherein the second spatial volume is larger than the first spatial volume.

[0087] A third embodiment comprises a method according to any of the preceding embodiments, visualizing the aggregate data (AD) of the second layer (L2) of the hierarchical tree (60) structure on a display (D1, D2) of an operating device for controlling and / or monitoring an additive manufacturing process, and / or superimposing a visualization of a data model of the object on a visualization of the aggregate data (AD) on the display (D1, D2), or vice versa.

[0088] A fourth embodiment comprises a method according to any one of the preceding embodiments, wherein the step of aggregating the operational data (OD) comprises determining statistical characteristics of the operational data (OD) of the nodes (63) of the first layer (L3), and using the statistical characteristics as aggregated data (AD), i.e., the aggregated data (AD) is the statistical characteristics of the operational data (OD) of the nodes (63) of the first layer (L3), and in particular, using the mean, median, variance, minimum, maximum, sum and / or count of the operational data (OD) associated with the nodes in the first layer (L3) as the statistical characteristics.

[0089] The fifth embodiment includes a method according to any of the preceding embodiments, further aggregating the aggregated data (AD) of the nodes in a specific layer into further aggregated data, and distributing the further aggregated data to the nodes of a subsequent layer of the hierarchical tree structure (60), wherein the subsequent layer is above the specific layer in the hierarchical tree structure (60).

[0090] A sixth embodiment comprises a method according to any of the preceding embodiments, transmitting aggregate data (AD) associated with nodes of a layer of a hierarchical tree structure (60) to an operating device based on an indication indicating storage capabilities and / or display capabilities of the operating device.

[0091] A seventh embodiment includes a method according to any of the preceding embodiments, visualizing aggregate data (AD) of nodes of a first layer in a first view on a display based on an indication received from an operating device, and / or visualizing aggregate data (AD) of nodes of a second layer in a second view on a display based on further indications received from the operating device.

[0092] An eighth embodiment includes a method according to any of the preceding embodiments, creating a first octree including a first number of layers on a first edge device, and creating a second octree including a second number of layers on a second edge device and / or an operating device, the second octree including fewer layers than the first octree.

[0093] A ninth embodiment includes transferring and / or loading, for example, selected layers in the first octree into the second octree according to the method of any one of the preceding embodiments.

[0094] The tenth embodiment includes a method according to any of the preceding embodiments, assigning a coded value representing a spatial volume, such as a Morton coded value, to operation data, wherein the operation data includes measurement data (40) and / or parameter data related to an additive manufacturing process, and based on the coded value, assigning the operation data (OD) to a node in a hierarchical tree structure (60), and / or sorting the coded values ​​according to the hierarchical tree structure (60) to associate the operation data (OD) to at least one node in a first instance of the hierarchical tree structure (60).

[0095] An eleventh embodiment includes obtaining simulated operation data according to a method according to any of the preceding embodiments, wherein the simulated operation data includes simulated measurement values ​​and / or simulated parameter values ​​related to the additive manufacturing process, assigning a coded value representing a spatial volume, such as a Morton coded value, to the simulated operation data, storing the simulated operation data in a hierarchical tree structure based on the coded value, sorting the coded value according to the hierarchical tree structure to associate the simulated operation data to at least one node in a second instance of the hierarchical tree structure, and comparing the operation data with the simulated operation data by visualizing the same layer of the first instance and the second instance of the hierarchical tree structure on a display of an operating device.

[0096] A twelfth embodiment comprises the method according to any of the preceding embodiments, determining quality indicators based on operational data, and assigning one or more quality indicators to one or more coded values ​​and / or nodes in a hierarchical tree structure.

[0097] The thirteenth embodiment comprises a method according to any of the preceding embodiments, for example, determining one or more anomalies in operational data based on an ellipsoidal influence space surrounding a spatial volume, and assigning the one or more anomalies to one or more coded values ​​and / or nodes in a hierarchical tree structure.

[0098] A fourteenth embodiment comprises the method of any of the preceding embodiments, wherein a first software application determines, on a first computing device, quality indicators and / or one or more anomalies based on operational data.

[0099] A fifteenth embodiment comprises the method according to any of the preceding embodiments, obtaining operational data by assigning a time stamp and / or a spatial position to each measurement value of a sensor monitoring the additive manufacturing process.

[0100] A sixteenth embodiment includes a system (100), for example, an additive manufacturing system including one or more edge devices (E1, E2) and / or operating devices, operable to perform the method steps of any of the preceding embodiments.

[0101] The seventeenth embodiment includes a computer program stored, for example, on a non-volatile medium, the computer program including program code, which, when executed, performs the method steps of any one of the above-mentioned embodiments 1 to 15.

[0102] Thus, as described in the present invention, an additive manufacturing process may be performed and / or may include a device for operating, monitoring and / or controlling a first and second computing device, for example, the first and / or second edge device and / or operating device. One or more of those computing devices may include a display for displaying operating data and / or aggregated data of the first and / or second octree. For example, the first computing device may obtain operating data from one or more sensors. These one or more sensors may be used to monitor the additive manufacturing process. Thus, one or more sensors may generate operating data by monitoring the additive manufacturing process. The operating data may be stored in the first computing device. There may be limitations on the processing power and / or storage capacity of the first computing device and / or the second computing device. Therefore, a second computing device for (further) processing, storing and / or visualizing the operating data and / or aggregated data stored in the first octree on the first computing device is proposed. Therefore, a second octree may be provided on the second computing device. Data of one or more layers of the first octree from the first computing device may be transmitted to the second computing device. Thus, the data of the first octree may be at least partially or only partially stored in the second octree on the second computing device. To this end, the (operated and / or aggregated) data associated with the nodes of one or more layers of the first octree may be transmitted to and / or loaded into the second octree on the second computing device. That is, the data associated with one or more nodes or all nodes of the first octree layer may be transmitted to and / or loaded into one or more nodes of the corresponding layer of the second octree. Thus, one or more layers of the first octree may be transmitted and / or loaded into the second octree. However, the second octree may include or have fewer layers than the first octree. For example, one or more layers of the (bottom or lower) layer of the first octree including raw data or operation data may be missing in, for example, the second octree. That is, not all layers, but only the layers of the first octree (i.e., the data associated with the nodes of that layer) are selected to be transmitted and / or loaded into the second octree. For example, only the n upper layers of the first octree may be transmitted and / or loaded into the second octree, while the m lower layers of the first octree are not transmitted and / or loaded into the second octree. Wherein m+n may correspond to the total number of layers of the first octree. Therefore, using the first octree for operational data and using the second octree for visualization, for example, provides the possibility of reducing the amount of data without repeatedly aggregating, filtering and / or deleting operational data. Thus, the second octree may be loaded with operational data and / or aggregated data. Preferably, the second octree does not include operational data or raw data on the lowest layer (i.e., the leaf nodes of the first octree). Preferably, the second octree only includes aggregated data, i.e., one or more intermediate layers of the first octree. Therefore, multiple consecutive layers of the first octree may be transferred to and / or loaded into the second octree.

[0103] In addition, for example, based on an indication of the storage capacity and / or display capacity of a second computing device (such as an operating device and / or a visualization device), the aggregated data (AD) associated with the nodes of one or more layers of the first octree may be transmitted to the second computing device (e.g., an operating device). Therefore, only the data of the selected nodes of the first octree may be transmitted to and / or loaded into the second octree. As mentioned above, the second octree may include fewer layers than the first octree. For example, the second octree may include only half or less than half of the layers of the first octree. Now, for example, depending on the interests of the user, further aggregated data of the operating data on the leaf nodes of the first octree may be transmitted to and / or loaded into the second octree. To this end, additional nodes and / or layers may be created and / or added to the second octree. Thus, data from the first octree may be transmitted and / or loaded into these other nodes of the second octree. For example, based on user input (e.g., a user selects a spatial volume or a spatial volume determined based on user input), data from a first octree belonging to the spatial volume may be transferred to and / or loaded into a second octree. Again, nodes and / or layers belonging to the spatial volume (and data associated with those nodes and layers) are transferred to and / or loaded into the second octree. There, data is appended or added to the second octree by creating additional nodes and layers that also represent the spatial volume. Therefore, a dynamic scheme for transmitting, loading and / or visualizing data for an additive manufacturing process is proposed. The above steps can be repeated to transfer additional data from the first octree to the second octree upon user input, for example, until a leaf node of the first octree is reached. It should be understood that after determining one or more spatial volumes, the spatial resolution and / or data resolution increases, i.e., becomes more granular. Thus, by determining the nested spatial volumes, only the data associated with the nodes and / or layers of these nested spatial volumes (as well as the operational data and / or aggregated data associated therewith) are transferred to the first octree and / or loaded from the first octree to the second octree. Thus, for example, the visualization on the display of the second computing device (e.g., the said operating device or visualization device) can be updated and / or made more detailed, for example, on demand (e.g., by user input or if further processing of the data in the second computing device becomes necessary).

Claims

1. A computer-implemented method for assisting, operating, monitoring and / or controlling an additive manufacturing process, the method comprising the following steps: obtaining operational data (OD) captured during the additive manufacturing process, distributing the operation data (OD) to nodes (63) in a first layer (L3) of a first octree (60) on a first computing device, aggregating the operational data (OD) of the nodes (63) in the first layer into aggregated data (AD), and distributing the aggregated data (AD) to nodes (62) of a second layer (L2) of the first octree (60), wherein the second layer (62) is above the first layer (63) in the first octree (60), Transferring and / or loading, for example, selected layers in the first octree into a second octree on a second computing device, wherein the second octree includes fewer layers than the first octree.

2. The method according to the preceding claim, wherein: The nodes (63) in the first layer (L3) represent, for example, a first spatial volume of an object to be manufactured, and the nodes (62) in the second layer (L2) represent a second spatial volume of the object, wherein the second spatial volume is larger than the first spatial volume.

3. The method according to any one of the preceding claims, visualizing the aggregated data (AD) of the second layer (L2) of the first octree on a display (D1, D2) of the first computing device and / or a second computing device, such as an operating device for controlling and / or monitoring the additive manufacturing process, and / or On the display (D1, D2), a visualization of the data model of the object is superimposed on a visualization of the aggregated data (AD), or vice versa.

4. According to the method described in any one of the preceding claims, the step of aggregating the operation data (OD) includes determining the statistical characteristics of the operation data (OD) of the nodes (63) of the first layer (L3), and using the statistical characteristics as aggregated data (AD), that is, the aggregated data (AD) is the statistical characteristics of the operation data (OD) of the nodes (63) of the first layer (L3), in particular using the mean, median, variance, minimum, maximum, sum, and / or count of the operation data (OD) associated with the nodes in the first layer (L3) as statistical characteristics.

5. The method according to any one of the preceding claims, further aggregating the aggregated data (AD) of the nodes in a specific layer into further aggregated data, and distributing the further aggregated data to nodes of subsequent layers of the first octree (60), wherein: The subsequent layer is above the particular layer in the hierarchical tree structure (60).

6. The method according to any of the preceding claims, transmitting the aggregated data (AD) associated with the nodes of the layer of the first octree to the operating device based on an indication indicating the storage capacity and / or display capacity of the operating device.

7. The method according to any one of the preceding claims, Based on an instruction received from the operating device, visualizing the aggregated data (AD) of the nodes of the first layer in a first view on a display, and / or Based on a further instruction received from the operating device, the aggregated data (AD) of the nodes of the second layer are visualized in a second view on the display.

8. The method according to any one of the preceding claims, creating a first octree including a first number of levels on the first edge device, and A second octree including a second number of layers is created on a second edge device and / or an operating device, the second octree including fewer layers than the first octree.

9. The method according to any one of the preceding claims, A coded value representing a spatial volume, such as a Morton coded value, is assigned to the operation data, wherein: The operational data includes measurement data (40) and / or parameter data related to the additive manufacturing process, and Based on the coded value, assigning the operation data (OD) to nodes in the first octree, and / or The encoded values ​​are sorted according to the first octree, thereby associating the operation data (OD) to at least one node in a first instance of the first octree (60).

10. The method according to any one of the preceding claims, Simulated operating data is obtained, wherein: The simulated operational data comprises simulated measurement values ​​and / or simulated parameter values ​​associated with the additive manufacturing process, assigning a code value representing a spatial volume, such as a Morton code value, to the simulated operation data, and storing the simulated operation data in the first octree based on the code value, sorting the encoded values ​​according to the first octree to associate the simulated operational data with at least one node in a second instance of the first octree, and The operating data is compared to the simulated operating data by visualizing the same layer of the first instance and the second instance of the first octree on a display of an operating device.

11. The method according to any one of the preceding claims, determining a quality indicator based on the operational data, One or more quality indicators are assigned to one or more coded values ​​and / or nodes in the first octree.

12. The method according to any one of the preceding claims, determining one or more anomalies in the operational data, such as based on an ellipsoidal influence space surrounding the volume of space, and The one or more exceptions are assigned to one or more encoded values ​​and / or nodes in the first octree.

13. The method of any preceding claim, wherein a quality indicator and / or one or more anomalies are determined by a first software application on a first computing device based on the operational data.

14. The method according to any of the preceding claims, wherein the operating data are obtained by assigning a time stamp and / or a spatial position to each measurement value of a sensor monitoring the additive manufacturing process.

15. A system (100), such as an additive manufacturing system, comprising one or more computing devices (E1, E2), such as one or more edge devices and / or one or more operating devices, wherein the one or more computing devices are operable to perform the method steps according to any of the preceding claims.

16. A computer program comprising program code which, when executed, performs the method steps according to any of the preceding claims 1 to 15.

Images