WO2019112655A1 - Trajet d'outil en boucle fermée et génération de réglage de processus de fabrication additive - Google Patents

Trajet d'outil en boucle fermée et génération de réglage de processus de fabrication additive Download PDF

Info

Publication number
WO2019112655A1
WO2019112655A1 PCT/US2018/047871 US2018047871W WO2019112655A1 WO 2019112655 A1 WO2019112655 A1 WO 2019112655A1 US 2018047871 W US2018047871 W US 2018047871W WO 2019112655 A1 WO2019112655 A1 WO 2019112655A1
Authority
WO
WIPO (PCT)
Prior art keywords
toolpath
printing
initial
settings
printing device
Prior art date
Application number
PCT/US2018/047871
Other languages
English (en)
Inventor
Guannan REN
Zhen Song
Mark R. BURHOP
Original Assignee
Siemens Aktiengesellschaft
Siemens Corporation
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 Siemens Aktiengesellschaft, Siemens Corporation filed Critical Siemens Aktiengesellschaft
Publication of WO2019112655A1 publication Critical patent/WO2019112655A1/fr

Links

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
    • G05B19/4097—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by using design data to control NC machines, e.g. CAD/CAM
    • G05B19/4099—Surface or curve machining, making 3D objects, e.g. desktop manufacturing
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
    • G05B19/404—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia

Definitions

  • the present disclosure is directed, in general, systems and methods for machine- controlled additive manufacturing.
  • a method includes receiving an initial toolpath and the associated initial printing settings by an AM system.
  • the method includes manufacturing a portion of a part according to the initial toolpath and the initial settings.
  • the method includes inspecting the manufactured portion of the part by the AM system.
  • the method includes, based on the inspection, performing a feature detection process to identify at least one printing problem in the manufactured portion of the part.
  • the method includes, based on the at least one printing problem, performing a reinforcement learning process to produce an adjusted toolpath or optimized settings for a 3D printing device.
  • the method includes manufacturing a further portion of the part according to the adjusted toolpath or the optimized settings.
  • the inspecting, feature detection process, and reinforcement learning process are repeated as the part continues to be manufactured.
  • the initial toolpath is a static initial toolpath based on 3D geometry corresponding to the part.
  • the inspection is performed using at least one camera.
  • the inspection is performed using one or more of a position sensor, a speed sensor, an infrared camera, a camera, or a temperature sensor.
  • the feature detection process sends reward information to the reinforcement learning process.
  • the reinforcement learning process uses a reward table that implements a mapping between printing problems and actions.
  • the optimized settings for the 3D printing device include one or more of a tooltip end position, a tooltip moving speed, a tooltip temperature, a material feeding speed, an environment temperature or a material type.
  • the feature detection process uses a convolutional neural network or other neural network training process.
  • AM system comprising a 3D printing device and an AM control system connected to control the 3D printing device.
  • the AM control system is configured to perform processes as disclosed herein.
  • Other embodiments include a non-transitory computer-readable medium storing executable instructions that, when executed, cause AM control system to perform processes as described herein.
  • Figure 1 illustrates an architecture for a toolpath generation for additive manufacturing based on position feedback
  • Figure 2 illustrates an architecture for a toolpath generation for additive manufacturing in accordance with disclosed embodiments
  • Figure 3 illustrates an example of a robotic device for additive manufacturing in accordance with disclosed embodiments
  • Figure 4 illustrates aspects of a process in accordance with disclosed embodiments
  • Figure 5 illustrates a process in accordance with disclosed embodiments.
  • Figure 6 illustrates a block diagram of a data processing system in which an embodiment can be implemented.
  • 3D printing Traditional additive manufacturing (also referred to herein as 3D printing) methods have no closed-loop control on the product quality.
  • the standard 3D printing file formats e.g., STL, OBJ, or PLY, only specify geometry information, sometimes with speed of the tooltip.
  • Table 1 lists these common 3D printing problems, their causes, detection methods, toolpath adjustments, and setting adjustments that may provide solutions to each problem.
  • a typical 3D printer interface includes many possible settings, sometimes as many as forty individual selections, which can leave users overwhelmed. If settings are not configured properly, problems as shown in Table 1 may arise and the printing process may fail or produce unsatisfactory results.
  • Figure 1 illustrates an architecture 100 for a toolpath generation for additive manufacturing based on position feedback.
  • a static toolpath 102 is used to control actuators 104 to position the tool tip 106.
  • Position sensor 108 feeds back position information of actuators 104 (and therefore tool tip 106), and the feedback is used to adjust the actuators to correct or change the position of tool tip 106.
  • this static toolpath generation does not generate 3D printing settings, nor does it adjust the toolpath based on the inspection on printing quality.
  • FIG. 2 illustrates a system 200 for a toolpath generation for additive manufacturing, where the sensor feedbacks are used for online toolpath and setting adjustment, as described in detail below. Note that some elements of system 200 in particular can be implemented in an AM control system 250 that is connected to control the 3D printing device and other hardware described herein.
  • 3D geometry 202 is used to generate a static initial toolpath 204.
  • the 3D printing device information 206 includes default settings 208.
  • the static initial toolpath 204 and default settings 208 are feed to reinforcement learning 210, which produces an adjusted toolpath 212 and optimized printer setting 214.
  • Adjusted toolpath 212 and optimized printer setting 214 are sent to the 3D printing device 216, which manufactures printed parts 220 by moving and operating tooltip 218.
  • position and speed sensors 224 detect the position and speed of the tooltip 218 (or the actuators controlling it) and pass position and speed data to feature detection 226.
  • Additional sensors 222 detect additional sensor data regarding tooptip 218 and/or the printed part 220 being manufactured.
  • Additional sensors 220 can include, for example, infrared (IR) sensors, IR cameras, visible light or other cameras, position sensors, temperature sensors, or other sensors.
  • IR infrared
  • Feature detection 226 can also collect model data 230 from cloud data repository 228.
  • Model dada 230 can include, for example, a material model, an actuator model, a device model, and log data.
  • Feature detection 226 identifies printing problems as described herein provides the printing problems to reinforcement learning 210 as described below.
  • the feedback process continues as the printed part 220 is being manufactured, including in particular by refining adjusted toolpath 212 and optimized printer setting 214 according to the reinforcement learning 210.
  • a reinforcement learning (RL) process 210 is used for quality assured tooltip control. Based on the 3D geometry 202 and the information of 3D printing device, the initial static toolpath and default settings are sent to the RL process 210.
  • The“feature detection” block 226 can identify printing problems, including faults in the part that has been or is being printed. It can do so, for example, based on computer vision and sensor fusion algorithms such as deep learning or artificial neural networks. These issues can be determined from the additional sensor data (of additional sensors 222) and/or the position and speed data (of position and speed sensors 224), such as from the camera or the IR camera, and some issues may be detected by temperature sensors, etc.
  • the data used for the feature detection algorithm can be collected and stored by a storage in the local printing device 216, the AM control system 250, or the cloud data repository 228.
  • users of the printers can decide to share their training data on the cloud without releasing their proprietary design information.
  • Other users may download parameters of trained neural networks, but not the raw data directly.
  • the outputs of the reinforcement learning process 210 can include both adjusted toolpath 212 and optimized printer settings 214, which are fed to the 3D printing device 216.
  • the 3D printing device 316 can be, for example, a 3D printer or a robotic device with a tooltip on its arm.
  • Figure 3 illustrates an example of a robotic device for additive manufacturing.
  • Position and speed sensors 224 and/or additional sensors 222 can be mounted, for example, on either the tooltip or in the vicinity of the printed parts.
  • a process as disclosed optimizes one or more of the following aspect (to which it is not limited), to improve the operation of the 3D printing device:
  • Tooltip end position The Tooltip moving speed.
  • the RL process 210 can operate as described below, in some embodiments.
  • a P represents a set of the actions to adjust the path, such as faster, slower, more or less spacing.
  • a s represents a set of actions to adjust the 3D printing parameters, such as increasing temperature, reducing extruding speed, etc. is an action within the set of all possible actions.
  • S is the set of states, including the ID of different defects/problems and their severity levels.
  • s k is a specific state at the time instance k, where s k 6 S.
  • Q(s, a) is a reward function, which maps the current state (e.g., detected problems or printing parameters) to a product quality metric as a reward value (e.g., adjusted toolpath or settings).
  • Q: S ® Pr(d a
  • Q(s, a) can be implemented, for example, as a reward table with every possible s in a row (alternately, column) and every possible a a column (alternately, row).
  • the reward table can be stored in a storage or memory of the local printing device 216, the AM control system 250, or the cloud data repository 228.
  • reward table 232 is stored on the AM control system 250.
  • Q(s, a) or the reward table 232 can be stored in a Deep Q-learning Network (DQN), which is a type of neural network.
  • DQN Deep Q-learning Network
  • r is the current reward. If the current part has no defects, the r is the largest, such as 1. If there are many defects, r is reduced to a small number, such as 0.
  • g is a fixed parameter as the learning gain.
  • the RL process 210 updates Q(s, a) (for example, as reward table 232) using the following equation:
  • the system can also perform a feature identification process using a feature identification algorithm /(m).
  • m is the sensor measurement data, where 6 M .
  • M is the set of all possible sensor measurements, which may have unlimited m.
  • M c is the collected measurement, such as a set of photos of different defects.
  • M c c M .
  • M c is always a countable set, i.e., with limited number of elements.
  • M ® S is the ideal mapping from measurement to state, i.e., the defect type number.
  • f k (m) ⁇ M c ® S is the f(m ) function in the k-th iteration.
  • /(m) is a mapping function from limited measurement to the faulty state.
  • f(m ) can be implemented by a convolutional neural network (CNN), but it can be any linear or nonlinear functions.
  • CNN convolutional neural network
  • the update of f k (m ) can use any neural network training process.
  • a ® M is the simulation model of the 3D printing process. Given actions as the input, the model generates sensor measurements. This simulation is optional. If there are no simulation tools available, the system can use real sensor data. If simulation model is available, the system can generate a large amount of training data for the feature identification process. The simulation can be based on physics model or based on neural network. Given initial training, in some embodiments, the Generative Adversarial Network (GAN) can be used to generate training data.
  • GAN Generative Adversarial Network
  • the static toolpath generator 102 in the AM process illustrated in Fig. 1 controls only the tooltip end position 106. Other important manufacturing factors are not controlled by the AM system illustrated in Fig. 1.
  • Disclosed embodiments improve on other AM systems by being able to identify faults in the printed parts and“tune” the system and printing process accordingly.
  • the system can use both simulation models and historical data for parameter optimization and toolpath adjustment.
  • the system can use the data collected from one 3D printing device to improve the tuning of another 3D printing device.
  • Various embodiments can use GAN to generate faulty printing by simulation, then learn to identify the faults.
  • Disclosed embodiments can be implemented as a closed-loop system.
  • the system adjusts the path and settings based on observation of the quality of the printing results.
  • the parameters are updated continuously for the best results.
  • Disclosed embodiments can use reinforcement learning in combination with feature detection processes. These distinct processes can be connected for optimal results.
  • Disclosed embodiments can use different sources to provide the training data for the feature detection process.
  • the training data can be any one or combination of real data, simulation data, or generative data (that are generated by neural networks).
  • the system can create a detailed lookup table and ensure uniform conditions. For example, with the exact same material and exact same environmental temperature, the quality of the final product is similar. However, when the condition is changed, such as shifting material vendors, the AM process parameters must be retuned. During the tuning process, manual experiments and measurements may be required, which increases costs and delays to the manufacturing process.
  • Figure 4 illustrates aspects of a process in accordance with disclosed embodiments.
  • the feature detection process 408 receives printing process data that can be one or more of real data 402, simulation data 404, or generative data 406.
  • the feature detection process 408 can be receiving actual feedback of the ongoing printing process.
  • 3D printing device 424 is printing part 414, and has completed finished area 416
  • one or more sensors 420 as described herein can inspect the finished area 416.
  • feature detection process 408 can use a feature detection algorithm as described above. Based on all the data received, feature detection process 408 can use the reward table described above to send reward information to RL process 410. RL process 410 already“knows” the current or initial toolpath 418 and uses it to control 3D printing device 414.
  • RL process 410 Based on the received data, RL process 410 produces an adjusted toolpath 422 and optimized printer settings 412. 3D printing device 424 then continues to manufacture part 414 using an adjusted toolpath 422 and optimized printer settings 412. As it does so, the finished area 416 increases and the feedback loop continues as described above with new inspections using sensors 420.
  • Figure 5 illustrates a process in accordance with disclosed embodiments that can be performed in system 200 as described above, referred to generically as the“system” below.
  • the process illustrated in Fig. 5 can be combined with or implemented in conjunction with any other processes or devices as described above.
  • the system receives an initial toolpath and initial settings of a 3D printing device (502). This can be a static toolpath as described above, or any toolpath that is then adjusted as described below.
  • the initial settings can be default settings, settings from a previous manufacturing process, manufacturer settings, or other settings.
  • the system manufactures a portion of a part according to the initial toolpath and the initial settings (504).
  • the system inspects the manufactured portion of the part (506).
  • the system Based on the inspection, the system performs a feature detection process to identify printing problem(s) in the manufactured portion of the part (508).
  • the system performs a reinforcement learning process to produce an adjusted toolpath and/or optimized settings for the 3D printing device (510). That is, in any given iteration of the feedback process, the system may determine that an adjusted toolpath is necessary to ensure or improve print quality, that new optimized settings for the 3D printing device are necessary to ensure or improve print quality, or both.
  • the reinforcement learning process can be used to tune at least one print setting or part of the toolpath in manufacturing the rest of the part. In some cases, no adjustment may be needed, but the process can continue to repeat as below until an adjustment is needed.
  • the system then manufactures a further portion of the part according to the adjusted toolpath and/or optimized settings (512).
  • FIG. 6 illustrates a block diagram of a data processing system in which an embodiment can be implemented, for example as part of a system 200 for a toolpath generation for additive manufacturing as described herein, or as an AM control system 250 as described herein, particularly configured by software or otherwise to perform the processes as described herein, and in particular as each one of a plurality of interconnected and communicating systems as described herein.
  • the data processing system depicted includes a processor 602 connected to a level two cache/bridge 604, which is connected in turn to a local system bus 606.
  • Local system bus 606 may be, for example, a peripheral component interconnect (PCI) architecture bus.
  • PCI peripheral component interconnect
  • main memory 608 Also connected to local system bus in the depicted example are a main memory 608 and a graphics adapter 610.
  • the graphics adapter 610 may be connected to display 611.
  • Peripherals such as local area network (LAN) / Wide Area Network / Wireless (e.g . WiFi) adapter 612, may also be connected to local system bus 606.
  • Expansion bus interface 614 connects local system bus 606 to input/output (I/O) bus 616.
  • I/O bus 616 is connected to keyboard/mouse adapter 618, disk controller 620, and I/O adapter 622.
  • Disk controller 620 can be connected to a storage 626, which can be any suitable machine usable or machine readable storage medium, including but not limited to nonvolatile, hard-coded type mediums such as read only memories (ROMs) or erasable, electrically programmable read only memories (EEPROMs), magnetic tape storage, and user-recordable type mediums such as floppy disks, hard disk drives and compact disk read only memories (CD-ROMs) or digital versatile disks (DVDs), and other known optical, electrical, or magnetic storage devices.
  • ROMs read only memories
  • EEPROMs electrically programmable read only memories
  • CD-ROMs compact disk read only memories
  • DVDs digital versatile disks
  • I/O bus 616 Also connected to I/O bus 616 in the example shown is audio adapter 624, to which speakers (not shown) may be connected for playing sounds.
  • Keyboard/mouse adapter 618 provides a connection for a pointing device (not shown), such as a mouse, trackball, trackpointer, touchscreen, etc.
  • I/O adapter 622 can be connected to communicate with or control printing hardware 628, which can include any of the 3D printing devices, sensors, imagers, systems, or other devices or hardware described herein.
  • a data processing system in accordance with an embodiment of the present disclosure includes an operating system employing a graphical user interface.
  • the operating system permits multiple display windows to be presented in the graphical user interface simultaneously, with each display window providing an interface to a different application or to a different instance of the same application.
  • a cursor in the graphical user interface may be manipulated by a user through the pointing device. The position of the cursor may be changed and/or an event, such as clicking a mouse button, generated to actuate a desired response.
  • One of various commercial operating systems such as a version of Microsoft WindowsTM, a product of Microsoft Corporation located in Redmond, Wash may be employed if suitably modified.
  • the operating system is modified or created in accordance with the present disclosure as described.
  • LAN/ WAN/Wireless adapter 612 can be connected to a network 630 (not a part of data processing system 600), which can be any public or private data processing system network or combination of networks, as known to those of skill in the art, including the Internet.
  • Data processing system 600 can communicate over network 630 with server system 640 (such as cloud systems as described herein), which is also not part of data processing system 600, but can be implemented, for example, as a separate data processing system 600.
  • machine usable/readable or computer usable/readable mediums include: nonvolatile, hard-coded type mediums such as read only memories (ROMs) or erasable, electrically programmable read only memories (EEPROMs), and user-recordable type mediums such as floppy disks, hard disk drives and compact disk read only memories (CD-ROMs) or digital versatile disks (DVDs).
  • ROMs read only memories
  • EEPROMs electrically programmable read only memories
  • user-recordable type mediums such as floppy disks, hard disk drives and compact disk read only memories (CD-ROMs) or digital versatile disks (DVDs).

Landscapes

  • Engineering & Computer Science (AREA)
  • Manufacturing & Machinery (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)

Abstract

L'invention concerne des procédés de fabrication additive et des systèmes ainsi que des supports lisibles par ordinateur correspondants. Un procédé consiste à recevoir (502) un trajet d'outil initial (204) et des réglages initiaux (208) d'un dispositif d'impression 3D (216) par un système AM (200). Le procédé consiste à fabriquer (504) une partie (416) d'une pièce (414) conformément au trajet d'outil initial (204) et aux réglages initiaux (208). Le procédé consiste en outre à inspecter (506) la partie fabriquée (416) de la pièce (414) par le système AM (200). Le procédé consiste également, sur la base de l'inspection, à mettre en œuvre (508) un processus de détection de caractéristiques (408) permettant d'identifier au moins un problème d'impression dans la partie fabriquée (416) de la partie (414). Le procédé consiste par ailleurs, sur la base dudit problème d'impression, à exécuter (510) un processus d'apprentissage de renforcement (410) permettant de produire un trajet d'outil ajusté (422) ou des réglages optimisés (412) du dispositif d'impression 3D (216). Le procédé consiste enfin à fabriquer (512) une autre partie de la pièce (414) conformément au trajet d'outil ajusté (422) ou aux réglages optimisés (412).
PCT/US2018/047871 2017-12-06 2018-08-24 Trajet d'outil en boucle fermée et génération de réglage de processus de fabrication additive WO2019112655A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201762595095P 2017-12-06 2017-12-06
US62/595,095 2017-12-06

Publications (1)

Publication Number Publication Date
WO2019112655A1 true WO2019112655A1 (fr) 2019-06-13

Family

ID=63524381

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2018/047871 WO2019112655A1 (fr) 2017-12-06 2018-08-24 Trajet d'outil en boucle fermée et génération de réglage de processus de fabrication additive

Country Status (1)

Country Link
WO (1) WO2019112655A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021262413A1 (fr) * 2020-06-22 2021-12-30 Autodesk, Inc. Génération de trajectoire d'outil par apprentissage par renforcement pour la fabrication assistée par ordinateur
WO2022233206A1 (fr) * 2021-05-07 2022-11-10 苏州奇流信息科技有限公司 Système de base de données d'échantillons, procédé d'entraînement et de vérification de paramètre d'impression et ordinateur
WO2023059627A1 (fr) * 2021-10-05 2023-04-13 Foshey Michael J Apprentissage de politiques de contrôle en boucle fermée pour la fabrication

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150045928A1 (en) * 2013-08-07 2015-02-12 Massachusetts Institute Of Technology Automatic Process Control of Additive Manufacturing Device
US20150331402A1 (en) * 2014-05-13 2015-11-19 Autodesk, Inc. Intelligent 3d printing through optimization of 3d print parameters
US20150328839A1 (en) * 2014-05-13 2015-11-19 Autodesk, Inc. Dynamic real-time slice engine for 3d printing
US20160236414A1 (en) * 2015-02-12 2016-08-18 Arevo Inc. Method to monitor additive manufacturing process for detection and in-situ correction of defects
US20180136633A1 (en) * 2016-05-20 2018-05-17 Moog Inc. Outer space digital logistics system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150045928A1 (en) * 2013-08-07 2015-02-12 Massachusetts Institute Of Technology Automatic Process Control of Additive Manufacturing Device
US20150331402A1 (en) * 2014-05-13 2015-11-19 Autodesk, Inc. Intelligent 3d printing through optimization of 3d print parameters
US20150328839A1 (en) * 2014-05-13 2015-11-19 Autodesk, Inc. Dynamic real-time slice engine for 3d printing
US20160236414A1 (en) * 2015-02-12 2016-08-18 Arevo Inc. Method to monitor additive manufacturing process for detection and in-situ correction of defects
US20180136633A1 (en) * 2016-05-20 2018-05-17 Moog Inc. Outer space digital logistics system

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021262413A1 (fr) * 2020-06-22 2021-12-30 Autodesk, Inc. Génération de trajectoire d'outil par apprentissage par renforcement pour la fabrication assistée par ordinateur
US11782396B2 (en) 2020-06-22 2023-10-10 Autodesk, Inc. Toolpath generation by reinforcement learning for computer aided manufacturing
EP4325401A3 (fr) * 2020-06-22 2024-05-22 Autodesk, Inc. Génération de trajectoire d'outil par apprentissage par renforcement pour la fabrication assistée par ordinateur
WO2022233206A1 (fr) * 2021-05-07 2022-11-10 苏州奇流信息科技有限公司 Système de base de données d'échantillons, procédé d'entraînement et de vérification de paramètre d'impression et ordinateur
WO2023059627A1 (fr) * 2021-10-05 2023-04-13 Foshey Michael J Apprentissage de politiques de contrôle en boucle fermée pour la fabrication

Similar Documents

Publication Publication Date Title
US11520322B2 (en) Manufacturing optimization using a multi-tenant machine learning platform
US10753955B2 (en) Systems and method for advanced additive manufacturing
WO2019112655A1 (fr) Trajet d'outil en boucle fermée et génération de réglage de processus de fabrication additive
US10747202B2 (en) Systems and method for advanced additive manufacturing
US9138893B2 (en) Robot teaching system, robot teaching assistant device, robot teaching method, and computer-readable recording medium
US11027535B2 (en) Systems and method for advanced additive manufacturing
US10086568B2 (en) Seamless scanning and production devices and methods
US9346219B2 (en) Manufacturing process for 3D printed objects
CN109202074A (zh) 用于先进增材制造的系统和方法
US10520912B2 (en) Robot controller having function that simplifies learning, and robot control method
CA3055818C (fr) Procede et appareil de reduction robuste d'erreur de forme dans un processus de fabrication additive base sur le depot de poudre laser en raison d'une incertitude
JP6849643B2 (ja) 出力装置、制御装置、及び評価関数と機械学習結果の出力方法
US11537111B2 (en) Methods and apparatus for 2-D and 3-D scanning path visualization
JP2017030067A (ja) 加工時間測定機能とオンマシン測定機能を有する制御装置付き加工装置
KR20230142650A (ko) 적층 제조에서 인공 지능 공정 제어를 위한 시스템, 방법 및 매체
JP6354178B2 (ja) 画像処理装置、管理システムおよび管理方法
WO2019066767A1 (fr) Système et procédé permettant de fournir une variation de taille de bille pour améliorer la précision géométrique de couches déposées dans un procédé de fabrication additive
KR20180014364A (ko) Cnc 공구 마모 보정 장치 및 방법
CN110658785A (zh) 输出装置、控制装置、以及评价函数值的输出方法
TWI770910B (zh) 焊接條件調整裝置
US10549349B2 (en) Method and apparatus for controlling a 3D-printing device and 3D-printing device
JP2020086759A (ja) ３次元モデル作成装置、加工シミュレーション装置、工具経路自動生成装置
US20200001529A1 (en) Fabrication prediction system, fabrication system, information processing apparatus, and fabrication prediction method
JP7079592B2 (ja) 消費量調整装置、溶接システム、消費量調整方法及びコンピュータプログラム
US20220134567A1 (en) Robot system, robot control device, and robot control program

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: 18766093

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: 18766093

Country of ref document: EP

Kind code of ref document: A1