阅读说明：本技术 提供用于增材制造的数据集的方法及对应的质量控制方法 (Method of providing a data set for additive manufacturing and corresponding quality control method ) 是由 P·阿弗多维克 J·福南德 A·格莱琴 于 2018-04-12 设计创作，主要内容包括：提出了一种提供用于增材制造的数据集(DS、DT)的方法,该方法包括在待制造构件(10)的至少一层(L<Sub>N</Sub>)的增材构造期间,收集(S1)用于数据集的第一类型数据(I1),借助于第一类型数据来评估(S2)层的结构质量,修改(S3)第一类型数据(I1),其中从第一类型数据(I1)中删除数据中的表示层(L<Sub>N</Sub>)的结构质量不足的部分数据,以及将第二类型数据(I2)叠加(S4)至第一类型数据(I1),其中第二类型数据(I2)适用于支持对所制造构件(10)的结构质量地验证。此外,还提出了一种对应的质量控制方法。(it is proposed a method of providing a data set (DS, DT) for additive manufacturing, the method comprising at least one layer (L) of a component (10) to be manufactured N ) During the additive building, collecting (S1) first type data (I1) for the data set, evaluating (S2) the structural quality of the layer by means of the first type data, modifying (S3) the first type data (I1), whereinRemoving a presentation layer (L) in data from a first type of data (I1) N ) And superimposing (S4) the second type of data (I2) onto the first type of data (I1), wherein the second type of data (I2) is adapted to support a structural quality verification of the manufactured component (10). In addition, a corresponding quality control method is also provided.)

1. a method of providing one data set (DS, DT) for additive manufacturing, comprising the steps of:

-at least one layer (L) of a component (10) to be manufactured_N) During the additive building, collecting (S1) a first type of data (I1) for the Data Set (DS),

-evaluating (S2) the layer (L) by means of the first type of data_N) The structural quality of (a) is high,

-modifying (S3) said first type of data, wherein said layer (L) is represented in data deleted from said first type of data_N) Wherein an algorithm acquires data at each layer (L) for the component (10)_N) A plurality of regions in which the powder in one powder bed is not solidified or is solidified erroneously, and

-superimposing (S4) a second type of data (I2) to the first type of data (I1), wherein the second type of data is adapted to support a verification of the structural quality of the manufactured component (10).

2. Method according to claim 1, wherein said first type of data (I1) comprises said layer (L1)_N) Or for said layer (L)_N) Optical or image data, microscopic data, CAD data, such as geometric information, CAM data and/or numerical control data.

3. Method according to claim 1 or 2, wherein the second type of data (I2) and/or the further type of data (I3, IN) comprises data from the layer (L)_N) Or for said layer (L)_N) Temperature information, pressure information or gas flow information, or information about beam properties, or information about the melt bath or powder bed, collected or read by one manufacturing device or another sensor device.

4. The method as claimed in any of the preceding claims, wherein the second type of data (I2) and third type of data (I3) are collected and superimposed with the first type of data (I1), and wherein the second type of data (I2) and/or the third type of data (I3) can be switched on and off in a data processing device (S5) in order to verify a manufactured layer (L) of the component (10)_N) The structural quality of (a).

5. The method according to any one of the preceding claims, wherein the Data Set (DS) is a visualization 3D data set and/or a tomography 3D data set of the manufactured component (10).

6. The method according to any one of the preceding claims, wherein a color, texture or otherwise graphically enhanced 3D representation (S6) of the Data Set (DS) is generated.

7. The method according to any one of the preceding claims, wherein correlation algorithms (S7) are implemented on different and/or superimposed types of data, for example by means of mathematical models, and wherein these correlation information are in turn input into machine learning algorithms (S8).

8. A quality control method for additive manufacturing of a component (10), comprising providing a Data Set (DS) according to the method of any of the preceding claims.

Technical Field

The present invention relates to a method of providing a data set for additive manufacturing of a component and a corresponding quality control method. Furthermore, a corresponding computer program, AM method and correspondingly manufactured component are described.

Preferably, the component means a component applied to a turbomachine, for example, a component in a flow path hardware of a gas turbine. The component is therefore preferably made of a superalloy or a nickel-based alloy, in particular a precipitation-hardened alloy, a solution-hardened alloy or an age-hardened alloy.

In the context of manufacturing, the term "additive" shall denote, inter alia, a layer-by-layer, generative and/or bottom-up manufacturing process. Additive manufacturing as described herein may also involve rapid prototyping.

Background

Additive manufacturing techniques (AM) include, for example, powder bed methods such as Selective Laser Melting (SLM), or Selective Laser Sintering (SLS), or Electron Beam Melting (EBM). These methods are relatively well known methods for making, prototyping or manufacturing parts or components, for example from powder materials. Conventional devices or apparatuses for such methods typically include a fabrication or build platform on which the component is built layer by layer, after a layer of substrate is fed, which may then be melted and subsequently solidified, for example by the energy of a laser beam. The layer thickness is determined by a doctor blade, which is automatically moved over the powder bed and removes excess material, for example. Typical layer thicknesses amount to 20 μm or 40 μm. During manufacture, the laser beam is scanned over the surface and melts the powder on selected areas, which may be predetermined by the CAD file according to the geometry of the component to be manufactured.

For example, a selective laser melting method is described in EP2601006B 1.

EP2666612a1 describes an imaging method of a three-dimensional component manufactured by a generative manufacturing method. The method comprises the following steps: at least two layers of images of the component are acquired during manufacturing via a detection unit configured to spatially resolved acquire measurements characterized as energy input into the component. The method further comprises generating, in the processing unit, a 3D image of the component from the acquired layer image, and finally imaging the 3D component with the display unit.

The term "layer" as used herein denotes in particular a physical layer or a slice of a component to be manufactured, as well as a virtual layer or data describing a virtual layer, for example in a CAD or image file.

One particular problem in the AM field is the reproducibility of the process and the predictability and verification of the structural properties of the manufactured materials/components.

At present, the true complex correlation or correlations between any kind of process parameters, e.g. physical entities, and the structure to be manufactured or a model thereof cannot be easily visualized at all. Inferring the cause of, for example, a 3D printing or AM process failure depends on the experience of the AM technician or extremely time-consuming digital data comparisons (if there is a large amount of sensor data) that are performed manually.

A related field of action is so-called "optical tomography" using layered photographs or other single sensor data.

During Layered Additive Manufacturing (LAM), such as Selective Laser Melting (SLM), complex physical processes can affect the outcome of 3D printing or manufacturing. At present it is not clear in most cases what physical influence over time and geometry determines the quality of the process. Examples of these effects may be SLM-machine characteristics, gas flow pattern over the powder bed, powder mass, temperature of the powder bed, actual layer thickness when exposed to beam radiation, laser or beam mass, radiation from the powder bed surface, etc.

Disclosure of Invention

The present invention addresses the challenge of how to visualize multiple correlations over time, and more importantly, the present invention directly records or provides corresponding information into a 3D representation of a printed part.

the above object is achieved/solved by the solutions of the independent claims. Preferred embodiments are the subject matter of the dependent claims.

One aspect of the invention relates to a method of providing a data set for additive manufacturing, comprising the steps of:

-collecting a first type of data for a data set during additive building of at least one layer of a component to be manufactured,

-evaluating and/or modifying the structural quality of the layer by means of the first type of data,

-modifying the first type of data, wherein parts of the data representing a layer of insufficient structural quality are removed from the first type of data.

The above-mentioned structural deficiency of the layer may be related to the powder or particle properties in the layer, for example in the region of the powder layer not intended to be solidified, depending on the intended geometry of the component. Additionally or alternatively, poor structural quality (e.g., solidification and/or process failure) of the powder bed region that has been exposed to the energy beam may be addressed.

The method further comprises superimposing a second type of data, e.g. acquired or predefined, to the first type of data, wherein the second type of data is adapted to support verification of structural quality of the manufactured component.

The term "adapted to" in a given context means in particular that observation and/or variation of a given type of data (second type) causes a structural change in the manufactured part.

It should also be understood that the first type of data (e.g. optical image information) is of course also suitable for supporting or allowing verification of the quality of the structure.

One advantage of the proposed concept is that the multi-sensor data and the composite sensor data can be accessed graphically as a three-dimensional overlay for the purposes of in-depth process understanding, on-line process monitoring, autonomous real-time process control, and/or machine learning. By the present invention, in-situ as well as ex-situ monitoring of a specific additive manufacturing process may be facilitated. Optical tomography provides the user with a good understanding or profound understanding of the 3D construction process or process history. Through an in-depth process understanding that enables reproducible and reliable prediction of the structural, mechanical and/or thermomechanical properties of the manufactured part, expensive non-destructive testing methods, such as ultrasonic or X-ray testing, can be omitted after the process. For example, creating a graphical representation of potential defects by spatially or hierarchically resolved optical microscopy is only a good step into the correct orientation, but stays "only" in the fact that there are defects or flaws. Even if other parameters, such as the parameters of the energy beam or heat introduction into the powder bed, are related to the optical or image data in a layered manner, the in-depth process understanding contemplated by the present invention is not achieved.

The invention has been developed further because it is more helpful to explain why a defect or process failure has occurred. In contrast to prior art imaging methods, the proposed concept is implemented by providing a data set, providing (almost) true digital replicas of a single component made by AM.

Furthermore, when the present invention is fully implemented, an enhanced AM manufacturing process (actually an on-line manufacturing process, i.e., a manufacturing process when the process is performed (real-time control)) will be provided, and eventually the occurrence of similar process deviations will be avoided before they occur.

The approach presented addresses the key requirements of today's manufacturing technology, namely better AM process reproducibility and structural understanding. Due to the increasing importance of digitization and/or the increasing impact of industrial 4.0 or internet of things, additive manufacturing processes are now undergoing tremendous, even subversive changes.

The need for better process understanding becomes more apparent when it comes to additive manufacturing of geometrically complex components and/or structures, e.g. comprising complex (alloy) compositions. Thermally stressed components of turbines made from superalloys, for example, need to provide excellent mechanical strength, resistance to thermal creep deformation, good surface stability, and corrosion or oxidation resistance. Therefore, the development of superalloy components has been heavily dependent on physical, chemical, and particularly process innovations.

Another aspect of the invention relates to a quality control method using or comprising a data set as described above.

In one embodiment, the first type of data includes optical or image data or information, microscopic data, CAD data (e.g., geometric information), CAM data, and numerical control data of or for the layer.

In one embodiment, the second type of data and/or the further type of data (e.g. the third type of data) comprises information such as temperature, pressure, gas flow, beam property information or information about the melt pool or powder bed collected or read from the layer or a manufacturing facility or another sensor facility for the layer.

In one embodiment, the second type of data is collected and overlaid with the first type of data, as well as other types of data (third type of data). This embodiment further addresses the problem of the present invention.

In one embodiment, the second type of data and/or the third type of data can be switched on and off in the data processing device in order to verify the structural quality of the component or of the manufactured layer for the component.

in one embodiment, the algorithm obtains areas where the powder in the powder bed in each layer for the component is not solidified or is solidified incorrectly. With this embodiment, it is further possible to easily modify the first type of data as described above.

In one embodiment, the dataset is a visualized 3D dataset and/or a tomographic 3D dataset of the fabricated component.

In one embodiment, a color, texture, or otherwise graphically enhanced 3D representation of the data set is generated. This allows, for example, an operator of the manufacturing apparatus to conveniently manage or study the data set and thus the digital copy of the component or the manufacturing process.

In one embodiment, the correlation algorithm is implemented on a plurality of types of data that are different and/or superimposed, for example by means of mathematical methods, and wherein these correlation information are in turn input into a machine learning algorithm.

Another aspect of the invention relates to a quality control method comprising a data set as described above and/or a providing method of the data set.

Another aspect of the invention relates to a computer program which, when being executed, is adapted to instruct a data processing device to carry out at least some of the method steps.

Another aspect of the invention relates to a method of additive manufacturing of a component, wherein the above quality control method and/or the providing method is applied.

Another aspect of the invention relates to a component that is additively or additively manufactured by the additive manufacturing method as described above.

The advantages related to the described providing method, the computer program and/or the described quality control method may also be related to the additive manufacturing method and/or the component, and vice versa.

Drawings

Other features, conveniences and advantageous improvements will become apparent from the following description of exemplary embodiments taken in conjunction with the accompanying drawings.

Fig. 1 shows a simplified flow chart illustrating the method steps of the present method.

Fig. 2 shows a schematic pictogram of the concept of the method of the invention.

Fig. 3 schematically shows an additive manufacturing apparatus.

Similar elements, homogeneous elements and identically acting elements may be referred to by the same reference numerals in the figures.

Detailed Description

Fig. 1 shows in particular a flow chart of the following steps of providing a data set (see DS below):

Step 1 (S1): each layer of the layered additive manufacturing process (LAM) is recorded by means of photography (first type of data, see I1 below), most likely by a camera located at the top of the chamber (i.e. e.g. in the upper part of the build space BS as shown in fig. 3), which camera has a resolution of either high or low size. This provides the possibility of creating a three-dimensional (data) model if the spacing in the z-direction is the same as the layer thickness of the CAM file sent to the 3D printer or AM manufacturing device. As described above, step S1 may involve the step of collecting the first type of data.

Step 2(S2) is optional (see dashed outline): for example, the algorithm identifies or evaluates which portions of each picture in each picture are not affected (not cured) in the powder bed (or fluid in the case of stereolithography as an AM process). In this way, algorithms that may be known to the skilled person also evaluate the structural quality of the respective layer. This identification of the uncured powder bed section is preferably carried out directly after taking the individual pictures (online monitoring). If step S2 is omitted, the focus of the analysis is not only on the part to be printed, but can also be used to include powder bed related phenomena in the analysis work.

Step 3 (S3): the powder bed segment of each picture was subtracted or deleted from the pixel representation of each picture. The uncured powder bed sections were removed from each picture directly after taking the individual pictures. Step S3 may be a real-time calculation and further involve the step of modifying the first type of data as described above.

Step 4 (S4): on each of the pictures prepared from step 1 to step 3, one or more additional sensor data types (second data type and/or third type data, see I2, I3 below) are collected and/or superimposed in graphical form. As one example, temperature readings from an additional infrared camera may be envisaged. By doing so, the additional sensor data can be visualized not only with the correct temporal correlation ("at what seconds of the printing process"), but also directly in the correct spatial order: for example, in layer 2333 of the structure of the member, the sensor data X1 (see below) changed, which shows that the approved tolerance band of X1 ("good treatment") was abandoned-this is why the structure collapsed at layer 2334.

Step 5(S5) is optional: switching one or several or all but one of the additional sensor data sets on and off gives the premise of more insight to the AM technician or operator.

Step 6(S6) is optional: the combined color, texture or otherwise graphically enhanced 3D representation of the layered visual evidence of the structure, plus one or several additional sensor datasets in graphical form on the right layer picture, can then be used for virtual "roaming" (fly-through) of datasets, research or evaluation. This generation has been shown to be effective for profound material or structural insights in a complex correlation approach.

Step 7(S7) is optional: the correlations identified from step 5 and step 6 can then be described in mathematical models and algorithms.

Step 8(S8) is optional: the algorithms and mathematical models described above may then be used for machine learning exercises.

The current data set may be in the ". jz" format and provide all the information relevant to reproducibly and comprehensively describing the AM manufacturing process. The data set may also be or include CAM data.

The described method steps may also be applied to the proposed quality control method and/or the method of additive manufacturing a component (not explicitly indicated in the figures).

The pictogram of fig. 2 preferably shows a representation or dataset identifying the characteristics of the AM printing process, i.e. the dynamic parameters describing the transfer functions of the subsystems presented by the printing process.

F₀(t) (see below) may represent or include the following parameters: powder properties as a function of time, inert gas flow, laser properties, etc.

On the other hand, Δ d (t) may describe the deviation of the component, which relates to the defects involved in the manufacture, which thus relates to the geometric integrity check.

The data sets or corresponding tools, or processing devices, provided herein, may also enable an operator to use the tool or device without specific physical or chemical knowledge (see numeral 100 in fig. 3).

The operator should also be able to influence the manufacturing process (adaptive process) on-line, for example to speed up and slow down the construction according to complexity, and to analyze the component or specific layers thereof in terms of geometry, dimensions, surface integrity, and to report data or find specific interests.

The parameters shown in the pictogram of fig. 2 may be characterized as follows:

x (t): relative displacement between the laser and the CAD model.

F (t): instantaneous value, and F₀(t) -nominal value.

Δ d (t): total deviation of the relative displacement x (t).

p (t) and pd (t) are perturbations due to thermal expansion, changes in the rate of parameter build-up, thickness changes, etc.

The parameter af may represent a correlation factor, as shown in fig. 2, such as build speed, argon or inert gas feed/speed, information on a CAD model, information on the melt pool or its dynamics, information on the powder bed, information on optical and/or microscopic pictures or images of the manufactured layer, layer thickness, recoater feed/speed and/or any information, for example, on the Manufacturing Execution System (MES).

The correlation factors described above may also represent or describe information about post-fabrication (structural and/or chemical) analysis, such as computer tomography, electron microscopy, or other means.

The parameter Δ d (t) may represent the difference of the manufactured component or its structure compared to the corresponding CAD model.

The proposed method of providing a data set may be a method of visualizing, controlling or monitoring properties of an AM manufacturing component. The proposed process may be an adaptive process and/or an interactive process by means of which the manufacturing configuration of the component may be supervised and/or controlled online. For example, when a certain threshold or tolerance for a defect is exceeded, more particularly when a break occurs in the component structure, the process parameters may be adjusted online during construction of the component.

The proposed method allows the use of the method or tool (when implemented in a data processing device) without specific knowledge about the structure or material of the manufactured component itself.

Further, all necessary information may be displayed on a screen such as a touch screen.

The components in the model may also be rotated and touched.

Furthermore, the process parameters can be read or edited both inside and outside the structure to be manufactured.

Fig. 3 schematically shows an additive manufacturing apparatus 100. The apparatus comprises or provides a three-dimensional construction space BS (see coordinate system with translation dimensions x, y and z on the left). Within the construction space BS, the component 10 is preferably arranged in a state in which it is actually manufactured by an additive process (e.g. selective laser melting outside the powder bed PB). Thus, the member 10 passes through the plurality of layers L shown_NAnd (4) manufacturing. The number of layers can easily reach a thousand or even ten thousand or more.

In the upper right corner of the device 100 a display 20 is shown, which display 20 comprises inter alia a touch screen TS, which display 20 shows digital (image) data of the component 10.

The information or data of the presented data set can be edited, modified, scrolled through, altered or changed (e.g. in terms of its value and orientation) by means of the touch screen TS to obtain knowledge about the built structure of the component 10. The image of the component 10 may be part of a data set DS or a data structure. The digital replica DT for the component 10 or the component 10 may be referred to synonymously with the data set DS. In other words, the data set DS described above is preferably a visualization, graphical and/or tomographic 3D data set of the component 10.

The component on the right side of fig. 3 (or its digital model) is subdivided into a plurality of layers L, as shown to the left of the physical component 10 in the construction space BS_N。

According to the invention, each layer comprises a set of sub data or information I1, I2, I3 up to IN. The data or information may relate to optical or image data, microscopic data, CAD data such as geometric information, CAM data and/or numerical control data and/or temperature information, pressure information, gas flow information or information about the layer or the melt bath or powder bed for the layer. I1, I2, I3 through IN may also relate to any information described herein IN connection with fig. 2.

For example, I1 may represent optical image data for the fabricated layer. Alternatively, micrographs or other microscopic image information may also be included in I1. Preferably, each layer L is during the additive construction of the component 10_NOne image is recorded so that for the entire building block, N images are collected and stored (e.g., in a data processing facility) for only child data I1. The reference I2 may for example represent temperature data or pressure data or any other data as described above.

According to the invention, at least the data I1 and I2 are then preferably superimposed, as indicated by the bubble in the graphical display 20, so that the advantages of the invention can be exploited.

The mentioned sub-data I1, I2, IN may further implement a correlation algorithm by means of mathematical methods. The relevant information or further collected data can thus in turn be processed by machine learning algorithms, allowing the most convenient and realistic investigation of process and material properties by means of the "digital copy" of the data set DS provided by the present invention.

thus, the number of parameters describing or (comprehensively) characterizing the layers of the structurally complex component 10 can easily exceed 100. To give just other examples of the mentioned values I1, I2 to IN, the above information or pieces of information may relate to: layer thickness, bath geometry, thermal shock per unit volume or area, laser wavelength, shadow line distance (i.e., spacing of adjacent scan lines), beam velocity, beam spot geometry, beam angle, type of purge gas, flow rate of possible vent gases, state of gas valves, set ambient pressure before or during build operation, state (i.e., mass) of the substrate, and the like.

By means of parameters, values or information which can be assigned to each of the above-mentioned layers, it is obvious that the overall quality control means of the component 10 can also be provided by a data set by means of which the data set DS is scanned or scrolled through after the component 10 has been manufacturedDifferent layer L_NWhen used, the "quality" can also be controlled retrospectively, for example, as to mechanical structure and chemical composition. IN the data set DS, the proposed method or data set also allows for an improved analysis process, since a single piece of information (see e.g. I3, IN) can be deactivated or switched on and off.

The scope of protection of the invention is not limited to the examples given above. The invention is embodied in each novel feature and each combination of features, including especially each combination of any features set forth in the claims, even if that feature or combination of features is not explicitly stated in the claims or examples.