WO2022159344A1 - Système et procédé d'analyse de qualité de fabrication additive - Google Patents

Système et procédé d'analyse de qualité de fabrication additive Download PDF

Info

Publication number
WO2022159344A1
WO2022159344A1 PCT/US2022/012540 US2022012540W WO2022159344A1 WO 2022159344 A1 WO2022159344 A1 WO 2022159344A1 US 2022012540 W US2022012540 W US 2022012540W WO 2022159344 A1 WO2022159344 A1 WO 2022159344A1
Authority
WO
WIPO (PCT)
Prior art keywords
slice
additive manufacturing
severity factor
defect
quality analysis
Prior art date
Application number
PCT/US2022/012540
Other languages
English (en)
Inventor
Lucas Stahl
Vijay DUA
Sourav Dutta
Srinivas Garimella
Onkar Pradeeprao BHISE
Mohit Singhal
Kelsey SNIVELY
Vinaya MANVATKAR
Jacob A. Kallivayalil
Karthikeyan Sundararajan CHOKAPPA
Original Assignee
Eaton Intelligent Power Limited
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Eaton Intelligent Power Limited filed Critical Eaton Intelligent Power Limited
Publication of WO2022159344A1 publication Critical patent/WO2022159344A1/fr

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B33—ADDITIVE MANUFACTURING TECHNOLOGY
    • B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
    • B33Y50/00—Data acquisition or data processing for additive manufacturing
    • B33Y50/02—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
    • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
    • B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
    • B29C64/30—Auxiliary operations or equipment
    • B29C64/386—Data acquisition or data processing for additive manufacturing
    • B29C64/393—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/40—Picture signal circuits
    • H04N1/409—Edge or detail enhancement; Noise or error suppression
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Definitions

  • the present disclosure relates to an additive manufacturing quality analysis system and method and finds particular, although not exclusive, utility in a system and method capable of quantifying a quality of an object manufactured with an additive manufacturing system.
  • Additive manufacturing otherwise known as 3D printing, is a manufacturing method which allows for a diverse range of objects or components to be produced from generic feedstock materials.
  • Additive manufacturing systems typically include a manufacturing bed, upon which an object is produced, and some form of material deposition system for depositing material on the manufacturing bed to form the object.
  • a computer based model of the object is produced and is numerically sliced into a number of discrete layers or slices. The layers are then produced and stacked in order on the manufacturing bed by the material deposition system to form the object.
  • an additive manufacturing system is a filament printing system.
  • a printer head is provided with feedstock material in the form of a filament, such as a thermoplastic filament.
  • the printer head has a heating element capable of melting the filament.
  • the printer head is controlled to move around the manufacturing bed and deposit molten filament according to a predetermined printing pattern based on the object slice, and the molten filament then cools and hardens to form the object slice.
  • the printer head then prints the next slice on top of the previous slice until the object has been produced.
  • Another example of an additive manufacturing system is a powder bed system.
  • a hopper containing a feedstock material, such as a metal, in a powder form deposits powder on the manufacturing bed.
  • a heat source such as a laser then scintillates the powder according to a predetermined printing pattern to bond powder particles to form the layer or slice.
  • the hopper may then deposit another layer of powder on the manufacturing bed, on top of the first slice, and the laser may then scintillate the next layer, bonding the two layers in the process.
  • a wiper blade or some other system may be used to ensure an even and constant thickness layer of powder is provided by the hopper.
  • a further example of an additive manufacturing system is a blown powder system, otherwise known as direct energy deposition.
  • a feedstock material such as a metal
  • Powder may be blown with an inert gas from a nozzle of a moveable printer head into the path of a laser, which heats the powder.
  • the heated powder which may be molten or sufficiently malleable, may then impact the manufacturing bed and form a layer of material.
  • the printer head may continue to deposit material in layers until the object is produced.
  • An object produced may differ from the computer model from which the object is produced. It is typically difficult or time consuming to take measurements of the manufactured object or otherwise verify the accuracy of each manufacturing process.
  • US 2018/0169948 A1 describes systems and methods by which parts produced by additive manufacturing can be assessed for conformity to a known master model. These systems and methods involve recording a thermal history of the manufacturing process of each part. The thermal history of a part is then compared to the thermal history of the master model. Significant deviations in thermal history may indicate irregularities in the build, and the part quality may then be assessed.
  • EP 3 388 907 A1 describes a method for providing a dataset for additive manufacturing.
  • the method includes collecting data related to the additive build-up of the additive manufacturing layers.
  • the structural quality of the layer is evaluated and the data is modified to delete portions of the data representing insufficient structural quality of the layer.
  • US 2017/0144378 A1 describes a system for adapting a product based on classification of product data using a trained machine learning classifier.
  • a learning system trains a production classifier to label or classify previously collected production data. The production data may then be used to classify product data in real time during production.
  • US 2015/0024233 A1 describes methods and apparatus to fabricate additive manufactured parts with in-process monitoring. As parts are formed layer-by-layer, a 3D measurement of each layer or layer group may be acquired. Dimensional data may then be acquired and accumulated until the part is fully formed.
  • US 2015/0037601 A1 describes an additive layer manufacture machine comprising an energy source for selectively melting a layer of metal powder and an electron beam for determining the shape of the melted metal layer once it has solidified. The shape is determined by detecting backscattered electrons from the interaction between the electron beam and the metal layer.
  • US 2015/0045928 A1 describes a system including an additive manufacturing device and a local network computer for controlling the additive manufacturing device.
  • a camera is provided with a view of the manufacturing volume of the device, and the computer is programmed to stop the manufacturing process when the object is identified as being defective.
  • EP 3 459 715 A1 describes a method and apparatus including a trained machine learning model that may be used to identify pre-warning indicators of defects in additive manufactured layers and predict defects in a manufactured layer.
  • an additive manufacturing quality analysis system for determining a quality of an object manufactured with an additive manufacturing system of the type configured to manufacture the object by producing and assembling a plurality of substantially planar slices of material on a manufacturing bed
  • the additive manufacturing quality analysis system comprising: an imaging system configured to capture images of the manufacturing bed; and a processor configured and arranged to: receive a computer based model of the object being manufactured, wherein the computer based model includes slice data corresponding to the plurality of substantially planar slices of material to be produced and assembled on the manufacturing bed; receive, from the imaging system, an image of each slice of material produced; identify, based on the image of each slice, a shape of each slice of material produced; determine a similarity between the shape of each slice of material and a corresponding slice of the computer based model; determine a part quality based on the similarity between the modified shape of each slice of material and a corresponding slice of the computer based model; and provide a final output including the part quality.
  • a key advantage of the present disclosure is that a quality of the object related to the similarity, both external and internal, of the computer model from which the object is manufactured may be provided. Accordingly, the accuracy and quality of the production may be known.
  • the object may be deemed to be of a relatively high quality if the object is relatively similar to the computer model, and may be deemed to be of a relatively low quality if the object is relatively dissimilar to the computer model.
  • a further key advantage of the present disclosure is that a quality of the object may be provided in real time during the manufacture of the object. Therefore, a user may stop or end the build of an object if the user decides, based on the quality of the object, that the final object will not be of a sufficient quality, thereby saving time and other resources. Additionally, image data may be extracted and analysed to provide a useful qualitative output related to the manufactured object.
  • Additive manufacturing may be 3D printing.
  • Additive manufacturing may be a manufacturing method in which an object or component is produced from generic feedstock materials such as thermoplastic filament or metal powder.
  • additive manufacturing such as filament printing, powder bed and blown powder manufacturing.
  • the present disclosure may be relevant to any manufacturing method or system in which an object or component is produced in layers, slices or other sub-pieces based on a computer model of the part or component.
  • the manufacturing system typically includes a manufacturing bed upon which the layers, slices or other sub-pieces are created and assembled.
  • the imaging system may comprise a camera and/or a sensor capable of capturing images or other light based data.
  • the imaging system may comprise one or more of: a visual light camera, an infrared camera, an ultraviolet camera and a photodiode. Other types of camera and sensor are envisaged.
  • the or each camera or sensor may be configured to capture data at a sample rate of 10 kHz, 15 kHz, 20 kHz, 30 kHz, 40 kHz, or any other suitable rate, which may be dependent on parameters of the additive manufacturing system. If a plurality of cameras and/or sensors are provided, the cameras and/or sensors may each have different sampling rates.
  • the imaging system may comprise a plurality of cameras and/or sensors.
  • the plurality of cameras and/or sensors may be spaced to provide different views of the manufacturing bed.
  • the imaging system may further comprise a light source.
  • the light source may be complimentary to the camera and/or sensor.
  • an infrared light source may be provided along with an infrared camera.
  • the light source may be configured to illuminate the manufacturing bed.
  • the processor may be networked with another processor, such as a processor of a second additive manufacturing quality analysis system according to the first aspect of the present disclosure. In this way, data related to a greater number of object builds or prints may be captured, stored and analysed for patterns or anomalies.
  • the processor may comprise an edge computing device.
  • the processor may receive the computer based model of the object being manufactured from a user and/or from a separate computing system.
  • the computer based model may be received by the processor with the slice data fully defined.
  • the processor may be configured to receive the computer based model and determine a suitable division of the model into slices, thereby generating the slice data.
  • the processor may be configured to receive a single image of each slice from the imaging system. For example, after each slice of material has been produced, an image may be captured and provided to the processor. Alternatively, a plurality of images of each slice may be provided to the processor.
  • the processor may be continuously provided with images at a predefined frequency. For example, images may be provided to the processor every second, or at some other suitable time interval. Alternatively, the processor may be continuously provided with images at an irregular frequency. For example, when manufacturing a relatively simple section of a part or object, the frequency may be reduced and when manufacturing a relatively complex section of a part or object, the frequency may be increased.
  • the relative complexity of part or object sections may be predetermined and/or determined by the processor.
  • the processor may be configured to determine a shape of each slice of material produced by identifying an outline of the slice and/or by identifying a boundary between materials.
  • the material slices may be a different colour and/or texture to the manufacturing bed and the processor may use colour and/or texture analysis to identify a boundary and outline of the slice.
  • the processor may use depth analysis to identify the material slice as the material slice may be at a relatively different depth to the manufacturing bed.
  • the processor may be configured to determine a property of each slice of material produced.
  • the property may be the shape, a colour, a texture, a geometry, a temperature, a heat conductivity, an electrical conductivity, a reflectivity, an alloy, a material mixture or any combination thereof.
  • the similarity between the shape of each slice of material and the corresponding slice of the computer based model may be defined numerically, such as with a relative value and/or a percentage similarity value.
  • the part quality may be defined numerically.
  • the part quality may be defined relatively, for example ‘good’, ‘average’ or ‘bad’.
  • the part quality may be defined relative to the computer based model and/or other manufactured parts.
  • the final output may be provided to a user.
  • the final output may be provided following completion of an object, a part, a section of an object or part, an article or a series of articles.
  • a user interface may be provided in this regard.
  • the user interface may comprise a display or some other feedback means suitable for presenting the final output to the user.
  • the user interface and the processor may form a single computing system.
  • the processor may be further configured and arranged to modify the identified shape of each slice of material based on a predetermined viewing distance and angle of the imaging system and the manufacturing bed to remove perspective and scale distortion.
  • Some spatial distortion, perspective and/or parallax error may be present in an image due to the viewing position of the imaging system relative to the slice being imaged.
  • the relative position of the imaging system may be predetermined. Accordingly, the necessary modification of the identified shape to remove errors due to the relative position of the imaging system and the slice being imaged may be predetermined.
  • the imaging system may be placed, or moved to, a position in which no viewing errors are present, such as directly above the slice being imaged.
  • some scale modification may still be necessary to match the scale of the computerised model. The modification may instead be performed on the computerised model such that the computerised model fits the scale and perspective of the images.
  • the processor may be further configured and arranged to identify, based on the determined similarity between the shape of each slice of material and the corresponding slice of the computer based model, a defect in a slice of material.
  • the defect may include dimensional variation, a hole, a bump and/or any other undesirable feature. In this way, a qualitative and/or quantitative property, such as dimensional variability, may be determined.
  • a defect may result from super-elevation, incomplete spreading, re-coater hopping, re-coater streaking, debris on the manufacturing bed and/or any other known defect cause.
  • the processor may be further configured and arranged to determine a build failure probability based on an identified defect, wherein the output includes the build failure probability.
  • the processor may be configured and arranged to determine if the layer or slice including the defect is suitable to act as a base support for subsequent layers or slices.
  • a relatively large hole in a layer or slice may mean that the layer or slice is unsuitable for supporting material above the hole and a relatively high build failure probability may be determined.
  • a relatively small hole in a layer or slice may mean that the layer or slice is still suitable for supporting material above the hole because the layer may be able to bridge or fill the hole, and a relatively low build failure probability may be determined.
  • the processor may use a trained build failure machine learning model to determine a build failure probability.
  • the build failure machine learning model may be trained with a labelled dataset including data related to builds including a characterised defect and a build failure or success. In this way, trained machine learning algorithms may be used for image processing which has been found to provide a desirable result.
  • the processor may be further configured and arranged to: determine a severity factor for the defect based on a predetermined set of severity parameters; compare the severity factor to a predetermined acceptable severity factor threshold range; and determine the severity factor for the defect is within or outside of the acceptable severity factor threshold range. In this way, the severity of a defect may be quantified and the build may be stopped if the severity of the defect is too high.
  • the predetermined set of severity parameters may include, for example, a width, length, diameter, area, position and proximity to an edge or other feature.
  • the predetermined set of severity parameters may be dependent on the type of defect identified. Accordingly, other severity parameters are envisaged.
  • the processor may be further configured and arranged to determine and apply a correction procedure to modify the slice of material such that the severity factor for defect is within the acceptable severity factor threshold range.
  • a defect may be corrected or remedied such that the build is not stopped and a desirable build quality is produced.
  • the correction procedure may include providing a new slice to be manufactured immediately after the slice of material including the defect. For example, if a hole is identified, the processor may cause the additive manufacturing system to fill the hole before moving on to the next layer or slice.
  • the correction procedure may include modifying a source code for a slice of material that is to be manufactured following the slice of material including the defect.
  • the slice of material that is to be manufactured following the slice of material including the defect may be modified to minimise the effect of and/or correct the defect.
  • Further slices may be modified in a similar way, if necessary. For example, a relatively large hole may be bridged with several subsequent layers.
  • the processor may be further configured and arranged to modify a shape of a subsequent material layer such that the severity factor for the defect is reduced. In this way, the effect of the defect may be reduced.
  • the processor may be further configured and arranged to produce a computer model of an object manufactured with the additive manufacturing system from the identified shape of each slice of material produced.
  • the processor may be configured to convert the identified shape of each slice of material into a corresponding computer based model of the respective slice and then assemble the slices into a computer based model of the manufactured object.
  • a slice thickness may be predetermined or derivable from the parameters of the additive manufacturing system. In this way, a user may be provided with a computer based model of the as-produced part, including any defects and/or anomalies.
  • the processor may be further configured and arranged to continually provide an intermediate output, wherein the intermediate output includes a part progress.
  • the part progress may comprise a percentage or other numerical representation of the completeness of the build.
  • the intermediate output may include a remaining build time. Accordingly, a user may be notified of a remaining build time.
  • the intermediate output may comprise a list and characterisation of any defect identified. In this way, a user may be notified of any defects before the build is complete.
  • the processor may be further configured and arranged to predict, with a trained build quality prediction machine learning model, a final build quality based on one or more slices of material produced.
  • the build quality prediction machine learning model may be trained with a labelled dataset including data related to build quality including a determined similarity of at least a partially produced product and a final build quality.
  • the processor may be further configured and arranged to predict, with a trained slice quality prediction machine learning model, a subsequent slice quality based on one or more preceding slices of material.
  • the slice quality prediction machine learning model may be trained with a labelled dataset including data related to build quality including a quality of a first slice and a quality of a subsequent second slice. In this way, trained machine learning algorithms may be used for image processing which has been found to provide a desirable result.
  • the additive manufacturing quality analysis system may further comprise a user input device configured to receive an input from a user.
  • the user input device may comprise a graphical user interface (GUI).
  • GUI graphical user interface
  • the user input device may be used to provide the output, both intermediate and final as appropriate, to the user.
  • the user input device may be used to provide details of any defect identified.
  • the user input device may be used to allow a user to modify or adjust the operation of the additive manufacturing system, particularly in response to the intermediate output.
  • the processor may be further configured and arranged to modify operation of the additive manufacturing system in response to a user input.
  • the determination of a similarity between the shape of each slice and the corresponding slice of the computer based model may include image moment analysis and/or a structural similarity index measure analysis. Accordingly, the similarity between the captured image and an image derived from the computer based model may be compared and a similarity determined.
  • the processor may be further configured and arranged to store, in an associated memory, a part quality.
  • the output may include an average part quality of a plurality of objects manufactured with the additive manufacturing system. Accordingly, an average quality of parts manufactured with the additive manufacturing system may be determined and provided to a user. Furthermore, a part quality relative to the average part quality may be determined and provided such that a user may differentiate parts based on quality.
  • the processor may be further configured and arranged to determine and provide a material property relating to a property of the material of the manufactured part.
  • Properties of the feedstock material may be predetermined, identified and/or provided to the processor.
  • the processor may be configured and arranged to determine, based on the properties of the feedstock material and the manufacturing system parameters, a material property of the manufactured part or object. For example, a feedstock including nickel, aluminium and titanium powder may form a nickel based superalloy when formed into a part or object.
  • the processor may be able to make this determination based on the composition of the feedstock material and the operating or bonding temperature of the additive manufacturing system.
  • the additive manufacturing quality analysis system may include an infrared camera and a photodiode configured to measure meltpool intensity and area.
  • the camera may have, for example, a sampling rate of 15 kHz and the photodiode may have, for example, a sampling rate of 50 kHz.
  • a scanner with a scanner mirror may also be provided.
  • Mirrors may be provided to direct laser and emission light to the appropriate locations.
  • a mirror configured to reflect laser wavelength light and configured to be transparent to emission wavelength light may be used to direct the laser light onto the meltpool whilst allowing emission light to pass through.
  • a further mirror configured to be semi-transparent to emission light may be used to split the emission light into two separate beams such that a beam may be provided to the camera and the photodiode.
  • a meltpool machine learning algorithm may be trained with a dataset including measured meltpool characteristics, corresponding defects and associated root causes.
  • the development of the algorithm may include the extraction and correlation of data with a process to define data filtering criteria.
  • the data filtering criteria may be used to clean data.
  • Meltpool area and intensity maps may then be developed from cleaned data.
  • the algorithm may be tested and verified with a test geometry including known and intentional defects or anomalies.
  • meltpool level data may be captured during the production of a part or object.
  • Useful data may then be extracted and processed with the meltpool machine learning algorithm.
  • the meltpool machine learning algorithm may then provide an output including one or more of: layer wise images including identified defect locations mapped on the layer geometry, a distribution map of defects, root causes for defects, and corrective measures for defects.
  • an additive manufacturing quality analysis method including the steps: receiving a computer based model of an object to be manufactured with an additive manufacturing system, wherein the computer based model includes slice data corresponding to a plurality of substantially planar slices of material to be produced and assembled on a manufacturing bed of the additive manufacturing system; receiving, from an imaging system, an image of each slice of material produced; identifying, based on the image of each slice, a shape of each slice of material produced; determining a similarity between the shape of each slice of material and a corresponding slice of the computer based model; determining a part quality based on the similarity between the modified shape of each slice of material and a corresponding slice of the computer based model; and providing a final output including the part quality.
  • the additive manufacturing quality analysis method may further include the step: identifying, based on the determined similarity between the shape of each slice of material and the corresponding slice of the computer based model, a defect in a slice of material.
  • a defect may be, for example, a hole, a protrusion, an incorrect shape and/or any other known defect.
  • the images may be separated into a plurality of corresponding sub-images and the similarity between the sub-images identified.
  • a defect may be identified based on, for example, a relatively low similarity in a single or a relatively small number of sub-images.
  • the additive manufacturing quality analysis method may further include the steps: determining a severity factor for the defect based on a predetermined set of severity parameters; comparing the severity factor to a predetermined acceptable severity factor threshold range; determining the severity factor for the defect is outside of the acceptable severity factor threshold range; determining the defect cannot be corrected; and controlling the additive manufacturing system to stop production of the object. Accordingly, a defect that may result in failure of the build may be identified build stopped immediately. In this way, time and other resources may not be spent completing a build after a fatal defect has been produced.
  • the additive manufacturing quality analysis method may further include the steps: determining a severity factor for the defect based on a predetermined set of severity parameters; comparing the severity factor to a predetermined acceptable severity factor threshold range; determining the severity factor for the defect is inside of the acceptable severity factor threshold range, or determining the severity factor for the defect is outside of the acceptable severity factor threshold range and applying a correction such that the severity factor is within the severity factor threshold range; and controlling the additive manufacturing system to complete production of the object. In this way, the severity and influence of a defect on the overall build may be reduced and a more desirable part or object may be produced.
  • the second aspect of the present disclosure may include any or each of the features described herein in relation to the first aspect of the present invention.
  • the method of the second aspect of the present disclosure may include a step that is equivalent to an operational step of the processor of the first aspect of the present disclosure.
  • Figure 1 is a schematic diagram of an additive manufacturing system and an associated quality analysis system
  • Figure 2 is a method diagram showing the steps of an additive manufacturing quality analysis method
  • Figure 3 is a method diagram showing the steps of an additive manufacturing method including quality analysis and assurance procedures.
  • FIG. 4 is a schematic diagram of a meltpool analysis system. Detailed Description
  • Figure 1 is a schematic diagram of an additive manufacturing system 100 and an associated quality analysis system 200.
  • the additive manufacturing system 100 shown is a powder bed type additive manufacturing system. Other types of additive manufacturing system are envisaged.
  • the additive manufacturing system 100 includes a manufacturing bed 110 upon which material is deposited in layers to form a part.
  • a moveable laser 120 is also provided above the manufacturing bed 110 and is arranged to, in use, direct laser light 130 toward the manufacturing bed 110.
  • a computer system (not shown) is provided to control movement and operation of the laser.
  • feedstock material such as a metal
  • a wiper or other system may be provided to spread the powder evenly on the manufacturing bed 110.
  • the laser 120 is then controlled to illuminate, or scintillate, specified areas of the powder to at least partially melt the powder particles such that they bond and form a layer of a part.
  • a part 140 is produced on the manufacturing bed 110.
  • the part 140 is partially buried in unbonded feedstock powder 150 which may be removed and reused.
  • the quality analysis system 200 includes an imaging system 210, such as a camera, and a processor 220.
  • the imaging system 210 is trained on the manufacturing bed 110 and is controlled by the processor 220 to capture images of the manufacturing bed 110 and a part 140 being manufactured thereon.
  • the processor 220 may communicate with the imaging system 210 wirelessly or via a wired connection.
  • the imaging system 210 is shown outside of the footprint of the additive manufacturing system 100, the imaging system 210 may alternatively be positioned inside the footprint of the additive manufacturing system 100, such as alongside the laser 120.
  • the quality analysis system 200 is configured and arranged to carry out a quality analysis method such as that described in relation to Figure 2.
  • Figure 2 is a method diagram showing the steps of an additive manufacturing quality analysis method 300.
  • the method 300 is carried out on a processor, such as the processor 220 of the quality analysis system 200 shown in Figure 1.
  • a first step of the method 300 is to receive 310 a computer based model of an object.
  • the computer based model may be a computer-aided design (CAD) file.
  • the CAD file may include data related to slices of material that are to be produced by an additive manufacturing system to produce the object.
  • the processor of the quality analysis system or a processor associated with the additive manufacturing system may be configured and arranged to determine a number of slices of material to be produced, based on inherent manufacturing parameters of the additive manufacturing system.
  • the additive manufacturing system can then begin producing the slices of material.
  • the second step of the method 300 is to receive 320, from an imaging system such as the camera 210 shown in Figure 1 , an image of each slice produced. The image may then be analysed with any suitable analysis method, such as those described herein, and a shape of each slice of material identified 330. After the shape of a slice of material has been identified 330, a similarity between the shape of the slice of material and a corresponding slice of the computer based model may be determined 340.
  • the determination 340 of a similarity between the shape of each slice and the corresponding slice of the computer based model may include image moment analysis and/or a structural similarity index measure analysis.
  • the similarity between the shape of the slice of material and the corresponding slice of the computer based model may be defined numerically, such as a percentage.
  • the steps receiving 320 images, identifying 330 shapes and determining 340 a similarity may be carried out immediately after each slice of material is produced.
  • the next step is to determine 350 a part quality based on the similarity.
  • a relatively high similarity may be indicative of a relatively high quality part, whereas a relatively low similarity may be indicative of a relatively low quality part.
  • An output including the part quality may then be provided 360, such as to a user via a user interface (not shown).
  • Figure 3 is a method diagram showing the steps of an additive manufacturing method 400 including quality analysis and assurance procedures.
  • the additive manufacturing system such as that shown in Figure 1 , may begin to construct the part to be manufactured for example by depositing and sintering a metal powder.
  • Powder bed and melt pool level real-time build process data 410 may be collected via appropriate sensors, such as infrared cameras and/or other sensors described herein, and associated computing hardware and software.
  • the powder bed and melt pool level real-time build process data can be used, along with part CAD data 415, to diagnose or predict anomalies at the powder bed level 420. Anomalies may be, for example, porosity or lack of fusion.
  • the CAD part data 415 may include layer wise inputs, such as the shape and thickness of each layer, along with quality control (QC) requirements which may, for example, include a temperature range within which adequate fusion of the metal powder is obtained.
  • QC quality control
  • the acceptable limits for the anomalies may be included in part CAD data 415 as QC requirements, or otherwise. If, following the determination, one or more anomalies are not within acceptable limits, a determination is made as to whether or not the anomaly can be correct 430 in such a way that the anomaly, post correction, is within the acceptable limits.
  • the type of correction used may be predetermined and selected from a database of corrective actions. For example, if a pore is detected, the database may include instructions to insert a new layer to fill only the pore, before continuing with the build process. Alternatively or additionally, a trained machine learning model, as described herein, may be used. If it determined that the anomaly cannot be corrected, the build process may be stopped 435.
  • corrective action may be, for example, dispensing more powder and sintering for a longer time in a region containing a pore in a previous layer.
  • the determination as to whether or not corrective action is need in upcoming layers 440 may be predetermined and selected from a database of corrective actions. Alternatively or additionally, a trained machine learning model, as described herein, may be used.
  • a corrective action may be suggested 445 to a user. The suggested action takes account of process parameters 450 to ensure that any corrective actions do not conflict with or contravene process parameters. A user may then decide to implement the suggested corrective action.
  • the system may be pre-authorised to take corrective actions, if necessary.
  • the build process may be altered or amended to take account of the corrective actions.
  • the powder bed level quality report 465 may include details or information related to the quality of the manufactured part based, at least in part, on the powder bed and melt pool level real-time build process data 410 collected during the build process.
  • FIG 4 is a schematic diagram of a meltpool analysis system 500.
  • the system 500 includes a laser 505 configured to provide laser light 510 at a sufficient power and intensity to melt a feedstock material used in an additive manufacturing process.
  • the laser light 510 is directed onto a first mirror 515 at an angle of approximately 45° and is wholly reflected into a scanner 520.
  • the scanner 520 includes a second mirror 525, also arranged at an angle of approximately 45° relative to the laser light 510 such that the second mirror 525 wholly reflects the laser light 510 toward a lens 530.
  • the lens 530 is optional.
  • the lens 530 focusses the laser light 510 onto the feedstock material to melt the feedstock material and create a meltpool 535.
  • Emissions 540 such as infrared light, are emitted from the meltpool 535.
  • the emissions 540 pass back through the scanner 520 and are reflected by the second mirror 525 towards the first mirror 515.
  • the first mirror 515 is configured to be transparent to emission 540 wavelengths so that the emissions 540 can pass through the first mirror 515.
  • a third mirror 545 is positioned behind the first mirror 515.
  • the third mirror is arranged at a 45° angle relative to the emission 540 path and is configured to be semi-transparent to emission 540 wavelengths. A first portion of the emissions 540 may therefore be reflected by the third mirror 545 to an infrared camera 550.
  • the portion of the emissions 540 that is not reflected by the third mirror 545 will pass through the third mirror 545 and enter a photodiode 555.
  • Other cameras and sensors are envisaged. Only a single, or more that two, cameras or sensors may be provided.
  • the meltpool analysis system 500 may be used to capture meltpool level data.
  • a powder bed type additive manufacturing system 100 is shown in Figure 1
  • other types of additive manufacturing system are envisaged such as a filament printing or a blown powder system.
  • the quality analysis system 200 and method 300 described herein may be used alongside any manufacturing system in which material is deposited in layers or slices to produce an object.
  • the system 200 may include a plurality of cameras 210 or any other known imaging device. If a plurality of imaging devices are provided, each imaging device may be positioned and arranged to capture an image from a different angle and/or viewpoint.
  • the quality analysis system 200 may include a user interface, such as a graphical user interface (GUI).
  • GUI graphical user interface
  • the user may interact with the GUI to modify operation of the additive manufacturing system 100, in response to receiving the output via the GUI or otherwise.
  • the method 300 shown in Figure 2 may include further steps, as will be apparent from the present disclosure. For example, a defect may be identified in a produced layer and corrective actions taken in subsequent layers.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Materials Engineering (AREA)
  • Manufacturing & Machinery (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Mechanical Engineering (AREA)
  • Optics & Photonics (AREA)

Abstract

La fabrication additive, autrement appelée impression 3D, est un procédé de fabrication qui permet de fabriquer une gamme variée d'objets ou d'éléments à partir de matières premières génériques. Cependant, il est typiquement onéreux ou difficile d'identifier des défauts ou des anomalies dans des objets issus de fabrication additive, en particulier si le défaut est interne. La présente divulgation concerne un système d'analyse de la qualité de la fabrication additive (200) comprenant un système d'imagerie (210) et un processeur (220). Le système d'analyse de qualité (200) est conçu et agencé pour capturer une image de chaque tranche d'une pièce issue de la fabrication additive et pour comparer une forme de chaque tranche à une forme de tranche correspondante d'un modèle informatique de la pièce à fabriquer. La qualité de la pièce est déterminée (350) sur la base de la similarité, une sortie, comprenant la qualité de la pièce, est ensuite fournie (360).
PCT/US2022/012540 2021-01-20 2022-01-14 Système et procédé d'analyse de qualité de fabrication additive WO2022159344A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IN202111002785 2021-01-20
IN202111002785 2021-01-20

Publications (1)

Publication Number Publication Date
WO2022159344A1 true WO2022159344A1 (fr) 2022-07-28

Family

ID=80445960

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2022/012540 WO2022159344A1 (fr) 2021-01-20 2022-01-14 Système et procédé d'analyse de qualité de fabrication additive

Country Status (1)

Country Link
WO (1) WO2022159344A1 (fr)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150024233A1 (en) 2013-07-19 2015-01-22 The Boeing Company Quality control of additive manufactured parts
US20150037601A1 (en) 2013-08-02 2015-02-05 Rolls-Royce Plc Method of manufacturing a component
US20150045928A1 (en) 2013-08-07 2015-02-12 Massachusetts Institute Of Technology Automatic Process Control of Additive Manufacturing Device
US20160236414A1 (en) * 2015-02-12 2016-08-18 Arevo Inc. Method to monitor additive manufacturing process for detection and in-situ correction of defects
US20170144378A1 (en) 2015-11-25 2017-05-25 Lawrence Livermore National Security, Llc Rapid closed-loop control based on machine learning
US20180169948A1 (en) 2015-06-12 2018-06-21 Materialise N.V. System and method for ensuring consistency in additive manufacturing using thermal imaging
EP3388907A1 (fr) 2017-04-13 2018-10-17 Siemens Aktiengesellschaft Procédé de fourniture d'un ensemble de données pour la fabrication additive et procédé de contrôle de qualité correspondant
EP3459715A1 (fr) 2017-09-26 2019-03-27 Siemens Aktiengesellschaft Procédé et appareil pour prédire l'apparition et le type de défauts dans un processus de fabrication additive
EP3459714A1 (fr) * 2017-09-26 2019-03-27 Siemens Aktiengesellschaft Procédé et appareil pour surveiller une qualité d'un objet d'une série de travaux d'impression en 3d d'objets identiques
US20200242496A1 (en) * 2019-01-25 2020-07-30 General Electric Company System and methods for determining a quality score for a part manufactured by an additive manufacturing machine

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150024233A1 (en) 2013-07-19 2015-01-22 The Boeing Company Quality control of additive manufactured parts
US20150037601A1 (en) 2013-08-02 2015-02-05 Rolls-Royce Plc Method of manufacturing a component
US20150045928A1 (en) 2013-08-07 2015-02-12 Massachusetts Institute Of Technology Automatic Process Control of Additive Manufacturing Device
US20160236414A1 (en) * 2015-02-12 2016-08-18 Arevo Inc. Method to monitor additive manufacturing process for detection and in-situ correction of defects
US20180169948A1 (en) 2015-06-12 2018-06-21 Materialise N.V. System and method for ensuring consistency in additive manufacturing using thermal imaging
US20170144378A1 (en) 2015-11-25 2017-05-25 Lawrence Livermore National Security, Llc Rapid closed-loop control based on machine learning
EP3388907A1 (fr) 2017-04-13 2018-10-17 Siemens Aktiengesellschaft Procédé de fourniture d'un ensemble de données pour la fabrication additive et procédé de contrôle de qualité correspondant
EP3459715A1 (fr) 2017-09-26 2019-03-27 Siemens Aktiengesellschaft Procédé et appareil pour prédire l'apparition et le type de défauts dans un processus de fabrication additive
EP3459714A1 (fr) * 2017-09-26 2019-03-27 Siemens Aktiengesellschaft Procédé et appareil pour surveiller une qualité d'un objet d'une série de travaux d'impression en 3d d'objets identiques
US20200242496A1 (en) * 2019-01-25 2020-07-30 General Electric Company System and methods for determining a quality score for a part manufactured by an additive manufacturing machine

Similar Documents

Publication Publication Date Title
US20210146447A1 (en) Additive manufacturing method and apparatus
US11607875B2 (en) Method and system for monitoring additive manufacturing processes
US20160098825A1 (en) Feature extraction method and system for additive manufacturing
Zenzinger et al. Process monitoring of additive manufacturing by using optical tomography
US9724876B2 (en) Operational performance assessment of additive manufacturing
CN113119470B (zh) 实时反馈控制3d打印过程的方法以及用于其的3d打印系统
US10183329B2 (en) Quality control of additive manufactured parts
US20150177158A1 (en) Operational performance assessment of additive manufacturing
US20160339519A1 (en) In-process monitoring of powder bed additive manufacturing
EP3747571B1 (fr) Surveillance de la fusion sur lit de poudre
US20210197275A1 (en) Three dimensional printing system and method capable of controlling size of molten pool formed during printing process
WO2020185169A1 (fr) Système de surveillance et procédé d'identification d'anomalies dans un processus d'impression 3d
EP3756859A1 (fr) Système de fabrication d'un objet de fabrication additive et méthode de fabrication d'un objet de fabrication additive
CN112996615B (zh) 识别增材制造中的异常事件的方法
EP3659727A1 (fr) Procédé d'identification automatique de déficiences de dépôt de matériaux au cours d'un procédé de fabrication additive et dispositif de fabrication
JP2021009126A (ja) 付加製造物の品質推定装置
US20230294173A1 (en) Method and Apparatus for Additive Manufacturing of a Workpiece using Statistical Learning Model
US20220197262A1 (en) Computer-Implemented Method for Determining Defects of an Object Produced Using an Additive Manufacturing Process
Dickins et al. Design of a multi-sensor in-situ inspection system for additive manufacturing
CN111319259B (zh) 用于监测增材制造系统中粉末铺展的系统和方法
CN115775249A (zh) 一种增材制造造件成形质量监测方法、系统及存储介质
Colosimo et al. On-machine measurement, monitoring and control
US20210197282A1 (en) Method and apparatus for estimating height of 3d printing object formed during 3d printing process, and 3d printing system having the same
JP7494448B2 (ja) 付加製造物の品質推定装置
US20220212396A1 (en) Additive manufacturing method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22704037

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22704037

Country of ref document: EP

Kind code of ref document: A1