METHOD FOR TESTING A MATERIAL

Abstract

The present invention relates to a method for testing a material for manufacturing a component for a turbomachine, in particular an aircraft engine, comprising the steps: i) providing the material in powder form; ii) sequentially measuring each portion of the powdered material with a hyperspectral camera system; iii) determining the suitability of the powdered material for manufacturing the component based on measurement results obtained in step ii).

Description

Technical area

[0001] The present invention relates to a method for testing a material intended for the production of a component for a turbomachine. State of the art

[0002] The components of a turbomachine, especially an aircraft engine, can be exposed to high temperatures and mechanical stress during operation. This can affect components located in the gas duct, for example, but also components outside of it, such as rotor disks. Regardless of the specific application, the high demands placed on the components require the use of special materials, such as special alloys, which in turn can result in increased manufacturing requirements. The quality of the material itself can influence that of the component, and can therefore be particularly important in light of the potential consequences of component failure. Description of the invention

[0003] The present invention is based on the technical problem of providing an advantageous method for testing a material.

[0004] This is achieved with the features of claim 1. The material provided in powder form (step i) is measured with a hyperspectral camera system (step ii). The measurement is carried out sequentially on one part of the powdered material at a time; in other words, a plurality of images are taken one after the other (each of a different part of it). The measurement results are then used to determine whether the powdered material is suitable for manufacturing a component of a predefined quality, i.e., whether it meets a component-specific required degree of purity.

[0005] By determining suitability based on hyperspectral measurement results, the operational reliability of the component in the turbomachine can be increased, for example, and / or potential rejects during subsequent component production can be reduced. Since component production itself is time- and cost-intensive, overall optimization can be achieved despite the additional step, i.e., the upstream material testing. The hyperspectral camera system generally expands the data space obtained; for example, a foreign particle can be detected, regardless of any size difference, due to altered heat radiation (IR range) or a different excitation / absorption behavior.

[0006] Deviations in the powdered material that could impair component quality can arise, for example, from impurities such as foreign particles or non-metallic inclusions, but also from particles with gas pores, etc. However, the suitability of the material can also be determined, for example, by excessive divergence in particle shape, form, and size, whereby the various properties can be recorded more reliably in the overall view using hyperspectral measurement than, for example, with images in the visible spectral range alone.

[0007] Preferred embodiments can be found throughout the disclosure and in particular in the dependent claims, whereby the presentation of the features does not always distinguish in detail between process, use, or device aspects; in any case, the disclosure is to be read implicitly with regard to all claim categories, in particular also to a corresponding testing and manufacturing system. Thus, if, for example, specific process steps or process steps performed with specific means are described, this is to be read simultaneously as a disclosure of a corresponding system that is configured to perform the process steps, in particular, has corresponding means.

[0008] The hyperspectral camera system can have sensitivity in the UV and / or IR range, in addition to the visible spectral range. Overall, the detected spectral range can extend, for example, at least over 500 nm - 1500 nm, and further and particularly preferably 400 nm - 2000 nm or 350 nm - 2500 nm (but also beyond that). By measuring the powdered material sequentially in parts, it can be captured with a sufficiently good resolution (not too many particles at once), which can increase the probability of detecting deviations. Furthermore, the testing method with sequential measurement can then also be easily integrated into a production process, for example.

[0009] According to a preferred embodiment, the powdered material is moved during the sequential measurement using a conveyor unit to which the hyperspectral camera system is assigned. Viewed in a stationary coordinate system, this can be stationary, so that the sequence results solely from the movement of the material by means of the conveyor unit. The conveyor unit can be, in particular, a conveyor belt, but alternatively, for example, also a slightly inclined vibrating plate. The hyperspectral camera system can be arranged above the conveyor belt or vibrating plate, so that its detection range is directed from top to bottom onto the conveyor belt or vibrating plate.

[0010] The conveying unit can also convey the powdered material as a particle stream through the detection zone of the hyperspectral camera system. For example, it can be designed in the form of a funnel, from whose outlet the material trickles down and passes through the detection zone. In this case, the camera system can be arranged offset laterally, for example, with a detection zone directed essentially horizontally toward the particle stream.

[0011] In a preferred embodiment, the powdered material is kept under clean room and / or inert gas conditions during measurement. The measurement can be carried out, for example, in a suitable room or chamber in whose internal volume a certain maximum number of particles per unit volume is permitted (e.g., no more than 10,000, 1,000, or 100 particles > 0.5 µm per cubic foot). Alternatively or additionally, the volume can be filled with an inert gas, e.g., argon. Regardless of the specific implementation, the clean room / inert gas atmosphere can, for example, prevent any changes to the material during or after measurement. Preferably, the material is then kept under atmospheric conditions at least until downstream production or transported to the downstream production unit.

[0012] According to a preferred embodiment, the material has specific fractionated particle sizes. A particle size distribution of at least 44 µm (325 mesh) and at most 150 µm (100 mesh) is particularly preferred. Alternatively, the material may preferably have a particle size distribution of less than 63 µm (230 mesh). The particle size distributions can be adjusted, for example, by appropriately sieving the material.

[0013] According to a preferred embodiment, the hyperspectral camera system has at least two hyperspectral cameras with which the powdered material is captured from different angles. The cameras can, for example, be oriented at different angles to the direction of movement of the material in the conveyor unit, i.e., capture the passing powdered material obliquely from the front and / or substantially perpendicularly and / or obliquely from the rear. Regardless of the specific orientation, at least three hyperspectral cameras are generally preferred, e.g., exactly three cameras. Preferably, the cameras of the system are tilted relative to one another but directed at the same capture area, meaning that the same part is captured from different directions at a given time.

[0014] In a preferred embodiment, the measurement result generally includes not only spatial and spectral information, but also angular information. The latter is preferably determined by simultaneous measurements with multiple hyperspectral cameras (see the previous paragraph), but could generally also be obtained by measurements with just one camera from different angles. The spatial information can be, in particular, two-dimensional, i.e., represent the position within a respective detection area, i.e., correspond to the pixel coordinates of the image.

[0015] Combined with the spectral and angular information, this can result in, for example, at least a four-dimensional tensor, which can then be evaluated using AI (see below for details). In general, the dimensionality of the tensor or data structure can be adjusted as needed if additional properties are measured or determined. The spectral information can, for example, include information on intensity and wavelength, especially different intensities across a wavelength range; see the initial remarks.

[0016] According to a preferred embodiment, the measurement result also includes time information, which can be used, for example, for ongoing real-time monitoring. Thus, a time stamp can be assigned to each part or image of the part; a time entry can be added as a further dimension to the aforementioned tensor. Figuratively speaking, the time information can make a conspicuous / unsuitable particle trackable, so that it can then be removed from the conveyor unit or discharged at the end of the conveyor unit, for example. In general, the temporal information can be captured, for example, using a hyperspectral camera system with the appropriate temporal resolution.

[0017] According to a preferred embodiment, the measurement results are used to derive material properties. This can be done, for example, in the course of hyperspectral image analysis and can yield properties such as particle size and / or shape, but also, for example, powder purity. Furthermore, a surface texture, in particular surface roughness, can also be determined depending on the resolution of the powder or individual particles. Specific limit values ​​can then be set for the various properties, for example, depending on the quality level required for the component to be manufactured, and suitability can then be determined by comparing them with these limit values.

[0018] According to a preferred embodiment, the powdered material is additionally frequency-excited, i.e., it is irradiated with electromagnetic radiation of a corresponding wavelength to excite eigenmodes. The energy of the eigenmodes and thus the wavelength can depend specifically on the material and the particle size; excitation can be carried out selectively in bands or broadband. As a result of the vibrational excitation, some heat is generated in the particle, which can be detected, for example, in the IR range and thus with the camera(s) of the hyperspectral camera system. Since the eigenmode excitation can be particularly dependent on the particle size, the heat radiation can be used to infer the particle size (the better the particle size matches the wavelength of the excitation, the greater the heating). This can, for example, also support classification according to specific particle sizes.

[0019] According to a preferred embodiment, the powdered material is fractionated prior to measurement, i.e., it is divided into fractions with different particle sizes. This can preferably be done by sieving, so that the different fractions result from the sieve overflow and passage, depending on the opening width. By fractionating beforehand, the particle size distribution that occurs during fractionation can be monitored during measurement. On the other hand, contamination can also occur, for example from fragments or abraded pieces that become detached from the sieve, and measurement can prevent these from entering the component. Furthermore, it is also conceivable that the hyperspectral camera system can also be used to inspect the sieves themselves; for this purpose, the actual powder conveying can be temporarily interrupted.

[0020] According to a preferred embodiment, the measurement results are evaluated using AI, i.e., the data acquired with the hyperspectral camera system is evaluated using AI algorithms. This can, for example, allow for automation and make a larger data space analyzable compared to visual evaluation (compare the multidimensional tensor discussed above). For evaluation using AI (artificial intelligence), so-called unsupervised learning algorithms can be used. This allows the data to be examined for patterns without previously known target values, so that, for example, anomalies (outliers) can be distinguished from normal variance (background noise). The data can, for example, be clustered (sorted according to commonalities) according to their characteristics, so-called "features."

[0021] According to an alternatively preferred embodiment, the AI-based evaluation is carried out using a supervised learning algorithm, which is, for example, trained and / or monitored. For this purpose, the measurement results of the hyperspectral camera system can be compared, for example, with measurement data obtained from samples taken from the powdered material, e.g., using metallographic or mechanical measurement methods. The algorithm can be trained based on the measurements taken from the samples, which can be done, for example, by labeling the measurement results of the hyperspectral camera system. The trained algorithm can then use regression models to calculate the target properties in question from the measurement results of the hyperspectral camera system.

[0022] Of course, the two approaches can also be combined, meaning the AI-based evaluation can be evaluated using both a supervised learning algorithm and an unsupervised learning algorithm. Unsupervised learning can be used particularly for those quantities for which separate experimental validation (mechanical or metallurgical) would be either impossible or too complex. Regardless of the specific implementation, the powdered material can then be tested in situ using the appropriate algorithm, for example, during ongoing production or batch filling (see below), thus integrating the suitability determination according to step iii) into the ongoing process (real-time monitoring).

[0023] The invention also relates to a method for producing a component for a turbomachine from a powdered material, wherein the material is measured with the hyperspectral camera system prior to component manufacture and tested for suitability. Subsequent component manufacture can then take place, for example, using a powder-based process, such as an additive process, e.g., selective laser melting (SLM) or electron beam melting (EBM), in which case the powdered material is processed directly. However, the powdered material can also be further processed, for example, by metal injection molding (MIM), i.e., it can be mixed with a binder system to form the appropriate shape. In all of these processes, the quality and suitability of the powdered material plays a key role, particularly for components designed for high temperatures, but also in general.

[0024] As an alternative to additive manufacturing, this testing method can be used particularly advantageously in powder metallurgical component production. This can, on the one hand, involve special alloys that can no longer be molded into the component shape using conventional casting or forging processes, for example, because the number or size of defects would become too large (e.g., segregation or demixing). Therefore, corresponding components are manufactured from the powdered material (each particle contains the alloy in the desired form), with the required batch of powder being placed in a collection container.

[0025] This is then evacuated and compressed; in what is known as hot isostatic pressing, a solid is created from the powder, a process also known as compaction. This solid can then be further processed into the desired shape, for example, by extrusion and forging. In this context, the present testing method can be particularly advantageous in that powder metallurgical production entails particular complexity and no corrections are possible during production. On the other hand, those components for which the special alloy is intended can be subjected to particular stresses during operation, so quality can be of particular importance. Hyperspectral measurement, for example, can ensure the high requirements for purity, uniformity, shape and surface quality of the powder.

[0026] Regardless of the specific manufacturing process, the powdered material can be an alloy, particularly a nickel-based alloy or nickel-based superalloy for applications involving high temperatures. Examples include ME16, ME3, N18, N19, or RR1000. Regardless of the specific alloy, the manufactured component can then be, for example, an airfoil, a blade, or even a rotor disk.

[0027] The invention also relates to a system for testing a powdered material, which comprises a conveying unit (see above), a hyperspectral camera system associated with the conveying unit (see above), and an evaluation unit. This is a computer configured to determine the suitability of the powder for component production based on the measurement results obtained with the hyperspectral camera system, preferably using appropriate AI algorithms (see above). For further details, including possible configurations of the conveying unit and hyperspectral camera system, reference is made to the above disclosure.

[0028] The invention further relates to a system for producing a component for a turbomachine, which system comprises, on the one hand, a production unit with which the component can be produced from the powdered material. This can be additive, in particular powder-based production (see above), but in particular also powder metallurgical production. In addition to the production unit, the system has a testing system as described in the previous paragraph, which is arranged upstream of the production unit. In the case of the powder metallurgical route, the two can, for example, be connected to one another in such a way that the powdered material passes from the conveying unit, preferably under a clean room / inert atmosphere, directly into the collecting container, where it can then be compressed as described above. Short description of the drawings

[0029] In the following, the invention is explained in more detail using an exemplary embodiment, whereby the individual features within the scope of the independent claims can also be essential to the invention in other combinations and no distinction is made in detail between the different claim categories.

[0030] In detail, Fig. 1 a system for testing a material in a schematic representation; Fig. 2 one to Fig. 1 alternative structure of a system; Fig. 3 an overview of some process steps in a flow chart; Fig. 4 a turbofan engine in a schematic section to illustrate an application environment. Preferred embodiment of the invention

[0031] Fig. 1 shows a system 1 for testing a material 10 provided in powder form 11. The system 1 has a hyperspectral camera system 5, which comprises several hyperspectral cameras 5.1-5.3. These are assigned to a conveyor unit 6, in which the system 1 according to Fig. 1 a conveyor belt 7. The powdered material 10 passes from an output container 12 onto the conveyor belt 7, from which it is conveyed in a direction of movement 17 to a collection container 22, which then also serves for component production.

[0032] The powdered material 10 passes through a detection zone 15 of the hyperspectral camera system 5. The part 10.1 of the material 10 located there is measured with the hyperspectral camera system 5, specifically with the different hyperspectral cameras 5.1-5.3 from different angles 25.1, 25.2. The hyperspectral cameras 5.1-5.3 each have a sensitivity in the visible, UV, and IR spectral ranges. Accordingly, in addition to the angular information, spectral information is also available, as well as location information regarding the position within the detection zone 15 and time information. This measurement result is fed to an evaluation unit 20, i.e., a measuring computer in which the data is evaluated using AI (see Fig. 3 and also the introduction to the description. This allows, for example, foreign particles or unsuitable powder particles (e.g., too large / small, etc.) to be identified and removed before reaching the collection container 22, which can help ensure the desired component quality (see the details above).

[0033] The entire assembly is provided in a clean room 30, in which clean room and / or protective gas conditions 31 can be set. This can prevent the introduction of contamination during the measurement, and the collection container 22 can then be fed directly to a manufacturing unit 40, shown only schematically here, for component production. The production itself can be carried out, in particular, using powder metallurgy; thus, the collection container 22 can be evacuated and compressed or extruded, and a solid body can be created from the powder by hot isostatic pressing. Overall, this system 100 can then be used to manufacture components of defined quality, depending on the application's requirements profile.

[0034] Fig. 2 shows a Fig. 1 alternative arrangement, namely a conveyor unit 6, designed as a funnel 8, which generates a falling particle stream 21. In this case, too, the powder enters the collection container 2 and is measured upstream using the hyperspectral camera system 5. This in turn comprises several hyperspectral cameras 5.1-5.3, and the measurement results are acquired via the evaluation unit 20. A sieve 9 is shown schematically, with which the material 10 is previously fractionated.

[0035] Fig. Figure 3 summarizes several process steps in a flowchart. The material is provided in powder form 51 and then sequentially measured in parts 52. The resulting measurement result 50 is evaluated 53 to determine the material's suitability for component production 54. For this purpose, the spatial information 50.1, spectral information 50.2, angular information 50.3, and temporal information 50.4 obtained from the measurement are used. The evaluation 53 is performed using an AI algorithm 70. Both supervised and unsupervised learning algorithms can be used.

[0036] Fig.4 shows an axial flow machine 60, specifically an aircraft engine 61. This is functionally divided into a compressor 62, a combustion chamber 63, and a turbine 64, with both the compressor 62 and the turbine 64 being constructed from multiple stages, each with a stator and a rotor (not referenced in detail). In the compressor 62, intake air is compressed, which is then combusted in the combustion chamber 63 with added kerosene. The resulting hot gas is expanded in the turbine 64, with the gained kinetic energy being used to drive the compressor 62 and generate thrust. Also in view of the high mechanical and / or thermal loads, various components 65, only a few blades referenced here as examples, can be provided from high-performance alloys and manufactured from a powder tested according to the present disclosure. LIST OF REFERENCE SYMBOLS 1 system 5 Hyperspectral camera system 5.1.–5.3 at least two hyperspectral cameras 6 conveyor unit 7 Conveyor belt 8 funnels 9 Sieve 10 Methods for testing a material 10.1 Sequentially measured part 11 Material in powder form 12 dispensing containers 15 Detection range 17 Direction of movement 20 Evaluation unit 21 Falling stream of particles 22 collection containers 25.1, 25.2 Different angles 30 cleanroom 31 Shielding gas condition 40 manufacturing units 50 measurement results 50.1 Location information 50.2 Spectral information 50.3 Angle information 50.4 Time information 51 Provision (of the material in powder form) 52 Sequential measuring 53 Evaluation 54 Determining suitability 60 Turbomachine 61 aircraft engine 62 compressors 63 Combustion chamber 64 turbines 65 component 70 AI algorithms 100 system

Claims

[1] Method for testing a material (10) for producing a component (65) for a turbomachine (60), in particular an aircraft engine (61), comprising the steps: i) providing (51) the material (10) in powder form (11); ii) sequentially measuring (52) one part (10.1) of the powdered material (11) with a hyperspectral camera system (5); iii) determining (54) a suitability of the powdered material (10) for the production of the component (65) on the basis of measurement results (50) determined in step ii). [2] Method according to claim 1, wherein the powdered material (10) is moved during the sequential measuring (52) in step ii) with a conveyor unit (6), in particular on a conveyor belt (7), wherein the hyperspectral camera system (5) is assigned to the conveyor unit (6). [3] Method according to claim 1 or 2, wherein the powdered material (10) in step ii) is kept under clean room and / or protective gas conditions (31). [4] Method according to one of the preceding claims, in which the powdered material (10) has a particle size distribution of at least 44 µm and at most 150 µm or has a particle size distribution of at most 63 µm. [5] Method according to one of the preceding claims, in which the hyperspectral camera system (5) has at least two hyperspectral cameras (5.1-5.3) with which the part (10.1) in step ii) is captured from different angles (25.1, 25.2). [6] Method according to one of the preceding claims, wherein the measurement results (50) in step iii) each comprise at least one item of location information (50.1), one item of spectral information (50.2) and one item of angle information (50.3). [7] Method according to one of the preceding claims, wherein the measurement results (50) in step iii) comprise time information (50.4) for ongoing real-time monitoring. [8] Method according to one of the preceding claims, in which at least one of the following properties of the powdered material (10) of the respective part (10.1) is determined from the measurement results (50): Purity, surface quality, especially surface roughness, Shape and / or size of the particles, wherein the size of the particles is preferably determined using frequency excitation and heat detection. [9] Method according to one of the preceding claims, in which the powdered material (10) is fractionated before step ii), whereby a particle size distribution is adjusted, preferably by sieving. [10] Method according to one of the preceding claims, in which the measurement results (50) in step iii) are evaluated on an AI-based basis in order to determine deviations, in particular impurities. [11] Method according to claim 10, wherein a supervised-learning algorithm is used for the AI-based evaluation, which algorithm has been trained and / or is monitored on the basis of tests carried out on the powdered material (10). [12] Method for producing a component (65) for a turbomachine (60) from a powdered material (10), in particular in a powder metallurgical process, wherein the powdered material (10) is tested in a method according to one of the preceding claims before the production of the component (65). [13] Method according to claim 12, wherein the sequentially measured parts (10.1) of the material (10) are collected in a collecting container (22), wherein - parts (10.1) for which non-suitability is determined in step iii) are rejected and / or - the measurement results (50) associated with the parts (10.1) of the powdered material (10) collected in the collecting container (22) then remain associated with the component (65) produced therefrom. [14] Installation (1) for testing a powdered material (10) intended for the production of a component (65) of a turbomachine (60), in particular in a method according to one of claims 1 to 11, with a conveyor unit (6) for moving the powdered material (10), a hyperspectral camera system (5) associated with the conveyor unit (6), and an evaluation unit (20), wherein the evaluation unit (20) is configured to determine a suitability of the powdery material (10) for the production of the component (65) from measurement results (50) determined on the powdery material (10) using the hyperspectral camera system (5). [15] Plant for producing a component (65) for a turbomachine (60) from a powdered material (10), in particular in a method according to claim 12 or 13, with a manufacturing unit (40) for producing the component (65) from the powdered material (60) and a system (1) according to claim 14 for testing the powdered material (10), which is arranged upstream of the production unit (40).

Images