A laser deposition additive manufacturing online monitoring system and method

Abstract

The application provides a kind of laser deposition additive manufacturing online monitoring system and method, comprising: laser deposition additive manufacturing module, robot system module and computer module;Firstly, in laser deposition module, laser is used to combine digital model path to deposit metal material, at the same time, additive manufacturing process data is collected in real time using robot system module, and all data is transmitted to computer module.Combined with algorithm model module, the surface of deposition layer, molten pool state can be monitored in real time, and the information such as defects or molten pool abnormal state monitored is fed back to laser deposition module to achieve real-time control, ensure the forming quality of workpiece surface, and improve product qualification rate.

Description

Technical Field

[0001] This application relates to the field of online monitoring technology in laser deposition additive manufacturing, and particularly to an online monitoring system and method for laser deposition additive manufacturing. Background Technology

[0002] Additive manufacturing, also known as 3D printing, primarily involves shaping materials layer by layer using a 3D digital model. Laser deposition additive manufacturing is a metal additive manufacturing process that uses focused thermal energy to melt raw metal to build workpieces, offering greater efficiency and flexibility. It can be used for rapid prototyping, manufacturing functionally graded materials, and repairing high-value-added components such as turbine blades. Over the past decade, this technology has been increasingly applied in the aerospace, defense, automotive, and biomedical industries.

[0003] In laser deposition additive manufacturing, the surface finish of workpieces can be poor due to the influence of various process parameters, environmental factors, and equipment. Defect generation and dimensional accuracy have consistently hampered the further popularization and application of this technology. Therefore, research on online monitoring technology for the additive manufacturing process is a necessary means to control the overall additive manufacturing process and ensure quality.

[0004] Currently, online monitoring technologies applied to address this issue include industrial CT (Computed Tomography), ultrasonic testing, and X-ray inspection. These methods primarily perform overall surface scanning and flaw detection on the workpieces produced after additive manufacturing. However, these methods lack the ability to monitor and control the additive manufacturing process itself, such as the detection of molten pool behavior, temperature, and surface morphology. Inspections based on these non-destructive testing methods are costly and cannot enable real-time repair. Summary of the Invention

[0005] To address the aforementioned issues of the inability to monitor and control the additive manufacturing process, such as the lack of detection of molten pool behavior, temperature, and surface morphology, and the inability to perform real-time repairs, this application proposes an online monitoring system for laser deposition additive manufacturing, comprising: a laser deposition additive manufacturing module, a robot system module, and a computer module.

[0006] The robot system module communicates with the computer module; the computer module communicates with the laser deposition additive manufacturing module.

[0007] The laser deposition additive manufacturing module is configured to melt and deposit powder or filament materials in conjunction with a digital model to form the target workpiece.

[0008] The robot system module is configured to collect data during the laser deposition additive manufacturing process and transmit the data to the computer module. The data includes: molten pool image data, molten pool temperature data, and 3D point cloud data.

[0009] The computer module includes an algorithm model module, which is configured to: train an algorithm model based on previous data, process the data in real time using the algorithm model, and save the processing results to the computer module.

[0010] The computer module is configured to process data through the algorithm model module, monitor the surface of the deposited layer and the state of the molten pool in real time, and feed back information such as detected defects or abnormal states of the molten pool to the laser deposition module.

[0011] In one feasible implementation, the robot system module includes a data acquisition device module and a signal transmission line module; the data acquisition device module is communicatively connected to the signal transmission line module; and the signal transmission line module is communicatively connected to the computer module.

[0012] The data acquisition device module is configured to acquire data during the laser deposition additive manufacturing process.

[0013] The signal transmission line module is configured to transmit data to the algorithm model module via a communication protocol or a high-speed transmission line.

[0014] In one feasible implementation, the data acquisition device module includes a high-speed camera, an infrared camera, and a laser 3D scanner, all of which are communicatively connected to the signal transmission line module.

[0015] The high-speed camera is configured to acquire image data of the molten pool.

[0016] The infrared camera was configured to collect molten pool temperature data.

[0017] The laser 3D scanner is configured to acquire 3D point cloud data.

[0018] In one feasible implementation, the computer module further includes a host and a display; the host is communicatively connected to the display, and the host is also communicatively connected to the algorithm model module.

[0019] The host also includes an image processing unit, a computing processing unit, and a data storage unit.

[0020] The image processing unit is configured to process molten pool image data.

[0021] The computational processing unit is configured to process molten pool temperature data and 3D point cloud data.

[0022] The data storage unit is configured to store both the stored data and the processed data.

[0023] The display is configured to show real-time data and enable human-computer interaction.

[0024] In one feasible implementation, the algorithm model module includes: an LSTM time-series prediction method module, a deep learning object detection algorithm module, and a neural network classification module; wherein the LSTM time-series prediction method module and the neural network classification module are both communicatively connected to the computing processing unit, and the deep learning object detection algorithm module is communicatively connected to the image processing unit.

[0025] The LSTM timing prediction method module is configured to: obtain predicted molten pool temperature data by combining timing methods with process parameters, and compare the predicted molten pool temperature data with the real-time acquired molten pool temperature data to obtain the comparison results.

[0026] The deep learning object detection algorithm module is configured to: input the molten pool image data after proportional scaling, train it, output fixed-standard molten pool image data for molten pool state detection, and obtain the detection result.

[0027] The neural network classification module is configured to classify 3D point cloud data by constructing a three-layer backpropagation neural network to obtain 3D defect information.

[0028] In one feasible implementation, the process parameters include time series, laser power, powder / wire feeding speed, and scanning speed.

[0029] In one feasible implementation, the proportionally scaled molten pool image data size is 224. Multiples of 224 pixels, with a fixed standard that the image dimensions range from 20 to 60 pixels.

[0030] In one feasible implementation, the laser deposition additive manufacturing module includes: a laser deposition head, a rotary processing platform, a high-power laser, a powder / filament feeding device, and a protective gas device.

[0031] The rotary processing platform is used to place the substrate and perform rotary processing at a certain angular velocity; the laser deposition head is located above the rotary processing platform and is used to deposit materials using a laser with a specific spot size.

[0032] A high-power laser is connected to the laser deposition head via an optical fiber, which is used to transmit lasers of different powers to the laser deposition head through the optical fiber.

[0033] The powder / wire feeding device is connected to the output end of the laser deposition head; the powder / wire feeding device is used to directionally output the material and works together with the laser deposition head to deposit the material.

[0034] The protective gas device is also connected to the output end of the laser deposition head. The protective gas device is used to output protective gas to prevent high-temperature oxidation during the material deposition process, which would affect the performance of the target workpiece.

[0035] It also includes a controller, which communicates with the computer module and with the high-power laser and the powder / wire feeding device, and is used to adjust process parameters based on information fed back from the computer module.

[0036] This application also provides an online monitoring method for laser deposition additive manufacturing, applicable to any of the above-mentioned online monitoring systems for laser deposition additive manufacturing. The monitoring method includes:

[0037] Laser deposition additive manufacturing of the target workpiece is performed using a laser deposition additive manufacturing module.

[0038] In the laser deposition additive manufacturing process, the robot system module acquires data of the target workpiece in real time and transmits the acquired data to the computer module for storage.

[0039] The computer module processes the data in real time and stores the results; it also feeds back information such as detected defects or abnormal melt pool conditions to the laser deposition module.

[0040] In one feasible implementation, the acquired data includes: molten pool image data, molten pool temperature data, and 3D point cloud data.

[0041] As described in the above technical solution, the laser deposition additive manufacturing online monitoring system of this application includes: a laser deposition additive manufacturing module, a robot system module, and a computer module. In the laser deposition module, metal materials are first deposited using a laser combined with a digital model path. Simultaneously, the robot system module collects data on the additive manufacturing process in real time, transmitting all data to the computer module. The acquired data is then combined with an algorithm model module to monitor the surface of the deposited layer and the state of the molten pool in real time. Information such as detected defects or abnormal molten pool states is fed back to the laser deposition module for real-time control, ensuring the surface forming quality of the workpiece and improving the product qualification rate. The laser deposition additive manufacturing online monitoring system of this application is mainly achieved by each module performing different functions and the modules interacting with each other. The content of each module can be modified or further developed according to specific requirements. This invention features high integration, high efficiency, ease of operation and modification, and significantly improves the performance of laser deposition additive manufacturing equipment. Attached Figure Description

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with implementations of the invention and are configured together with the description to explain the principles of embodiments of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0043] Figure 1 This is a schematic diagram of the online monitoring system for laser deposition additive manufacturing according to this application;

[0044] Figure 2 An online monitoring system for laser deposition additive manufacturing, which is an exemplary embodiment of this application;

[0045] Figure 3 This is a schematic flowchart of the online monitoring method for laser deposition additive manufacturing according to this application;

[0046] Figure 4 This is a schematic diagram of the LSTM time series prediction method module of this application;

[0047] Figure 5 This is a schematic diagram of the deep learning object detection algorithm module of this application;

[0048] Figure 6 This is a schematic diagram of the neural network classification module of this application. Attached image description:

[0050] 1-Laser deposition additive manufacturing module; 2-Robot system module; 3-Computer module;

[0051] 11-Laser deposition head; 12-Rotary processing platform; 13-High-power laser; 14-Powder / wire feeding device; 15-Protective gas device; 16-Controller; 21-Data acquisition equipment module; 22-Signal transmission line module; 31-Algorithm model module; 32-Main unit; 33-Display;

[0052] 211-High-speed camera; 212-Infrared camera; 213-Laser 3D scanner; 321-Image processing unit; 322-Computational processing unit; 323-Data storage unit;

[0053] 3111 - Function module; 3112 - Forget gate; 3113 - Update gate; 3114 - Output gate; 3121 - Melt pool image dataset; 3122 - Pre-feature layer; 3133 - Feature layer data; 3134 - First feature layer; 3135 - Second feature layer; 3134a - First output feature vector layer; 3135a - Second output feature vector layer. Detailed Implementation

[0054] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the embodiments of the invention will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of how embodiments of the invention are carried out.

[0055] In laser deposition additive manufacturing, the surface accuracy of the workpiece is often poor due to the influence of different process parameters, environment, and equipment. The generation of defects and dimensional inaccuracies have consistently hindered the further popularization and application of this technology. Therefore, research on online monitoring technology for the additive manufacturing process is a necessary means to control the overall additive manufacturing process and ensure quality. Currently, online monitoring technologies applied to address this issue include industrial CT inspection, ultrasonic inspection, and X-ray inspection, which mainly perform overall surface scanning and flaw detection on the workpiece after additive manufacturing. However, these methods cannot effectively monitor and control the additive manufacturing process, lacking features such as detection of molten pool behavior, temperature monitoring, and surface morphology.

[0056] This application addresses the aforementioned problems of the inability to monitor and control the additive manufacturing process, such as the lack of detection of molten pool behavior, temperature, and surface morphology, and the inability to perform real-time repairs. It proposes an online monitoring system for laser deposition additive manufacturing, referring to... Figure 1 As shown, Figure 1 This is a schematic diagram of the online monitoring system for laser deposition additive manufacturing according to this application, including: laser deposition additive manufacturing module 1, robot system module 2, and computer module 3.

[0057] Robot system module 2 is communicatively connected to computer module 3; computer module 3 is communicatively connected to laser deposition additive manufacturing module 1. The integrated equipment in robot system module 2 monitors the laser deposition additive manufacturing process of laser deposition additive manufacturing module 1. The equipment on robot system module 2 can monitor laser deposition additive manufacturing module 1 through methods such as photography, scanning, and infrared sensing, without requiring a direct communication connection with laser deposition additive manufacturing module 1, thus improving the flexibility of equipment use.

[0058] The laser deposition additive manufacturing module 1 is configured to melt and deposit powder or filament materials in conjunction with a digital model to form the target workpiece. Any existing laser deposition additive manufacturing equipment can be used in this application. The monitoring of the laser deposition additive manufacturing module 1 in this application is universal and is not affected by the equipment structure. Laser deposition additive manufacturing equipment can be selected according to actual needs, and no restrictions are imposed in this application.

[0059] Robot system module 2 is configured to collect data during the laser deposition additive manufacturing process and transmit the data to computer module 3. The data includes molten pool image data, molten pool temperature data, and 3D point cloud data. Molten pool image data, molten pool temperature data, and 3D point cloud data are crucial data determining workpiece quality during laser deposition additive manufacturing. These three data points cover all problems and defects encountered during workpiece fabrication, ensuring no omissions or unresolved defects. This application achieves workpiece quality control through monitoring these three data points.

[0060] Computer module 3 includes algorithm model module 31, which is configured to: train an algorithm model based on previous data; process the data in real time using the algorithm model; and save the processing results to computer module 3. Algorithm model module 31 can establish an algorithm model based on previous equipment process parameters. The algorithm model can be matched with three types of data to be processed: molten pool image data, molten pool temperature data, and 3D point cloud data, and processed separately. This ensures the accuracy and efficiency of the processing. The algorithm model can be arbitrarily built according to needs and is continuously updated by collecting data during monitoring, making subsequent data processing increasingly accurate. It offers high flexibility and high work efficiency.

[0061] Computer module 3 is configured to process data through algorithm model module 31, monitor the surface of the deposited layer and the state of the molten pool in real time, and feed back information such as detected defects or abnormal states of the molten pool to the laser deposition module. Computer module 3 processes data through algorithm model module 31, and can save and feed back data. Real-time communication between computer module 3 and laser deposition additive manufacturing module 1 creates a closed-loop information feedback system, enabling real-time monitoring. Laser deposition additive manufacturing module 1 responds to the feedback information from computer module 3, promptly modifying or remedying defects, greatly improving the quality of the finished product and increasing the pass rate during mass production.

[0062] Furthermore, this application also provides an online monitoring method for laser deposition additive manufacturing, applicable to any of the aforementioned online monitoring systems for laser deposition additive manufacturing, with reference to... Figure 3 As shown, Figure 3 This is a schematic flowchart of the online monitoring method for laser deposition additive manufacturing according to this application. The monitoring method includes:

[0063] S100: Laser deposition additive manufacturing of the target workpiece is performed through laser deposition additive manufacturing module 1;

[0064] S200: In the laser deposition additive manufacturing process, the robot system acquires data of the target workpiece in real time and transmits the acquired data to computer module 3 for storage; further, the acquired data includes: molten pool image data, molten pool temperature data and three-dimensional point cloud data.

[0065] S300: Data is processed in real time through computer module 3 and the processing results are stored; information such as detected defects or abnormal states of the molten pool is fed back to the laser deposition module.

[0066] The information regarding defects or abnormal molten pool conditions includes: defect data, abnormal molten pool condition data, and abnormal molten pool temperature data. Specifically, defect data, such as porosity and cracks, requires transformation of the returned defect coordinate data into the machining platform coordinate system. Abnormal molten pool condition data, such as molten pool splashing and molten pool offset, requires timely adjustment of process parameters and machining path repair based on the currently set process parameters. Abnormal molten pool temperature data, such as excessively high or low temperatures, requires timely adjustment of process parameters to restore the molten pool temperature to the normal range.

[0067] As can be seen from the above technical solution, in this application, the laser deposition module first uses a laser combined with a digital model path to deposit metallic materials. Simultaneously, the robot system module 2 collects data from the additive manufacturing process in real time, transmitting all data to the computer module 3. The acquired data is then combined with the algorithm model module 31 to monitor the surface of the deposited layer and the state of the molten pool in real time. Information such as detected defects or abnormal molten pool states is fed back to the laser deposition module for real-time control, ensuring the surface forming quality of the workpiece and improving the product qualification rate.

[0068] In one feasible implementation, refer to Figure 2 As shown, Figure 2This application discloses an exemplary embodiment of an online monitoring system for laser deposition additive manufacturing. The robot system module 2 includes a data acquisition device module 21 and a signal transmission line module 22. The robot system module 2 integrates a vision sensing system and a scanning path tracking system. The integration method involves using a specific fixture to fix the robot control system and data acquisition device, and performing hand-eye calibration to determine the coordinate transformation and precise positioning of each data point. The data acquisition device module 21 is communicatively connected to the signal transmission line module 22, and the signal transmission line module 22 is communicatively connected to the computer module 3.

[0069] Continue to refer to Figure 2 As shown, the data acquisition device module 21 is configured to acquire data during the laser deposition additive manufacturing process. The data includes molten pool image data, molten pool temperature data, and 3D point cloud data. Molten pool image data can be acquired using devices such as cameras, and molten pool temperature data can be acquired using temperature sensors, infrared sensors, etc. This application provides a data acquisition device module 21 comprising: a high-speed camera 211, an infrared camera 212, and a laser 3D scanner 213; all three are communicatively connected to the signal transmission line module 22. Specifically, the high-speed camera 211 is configured to acquire molten pool image data; the infrared camera 212 is configured to acquire molten pool temperature data; and the laser 3D scanner 213 is configured to acquire 3D point cloud data. The data acquisition device module 21 is positioned close to the laser deposition additive manufacturing module 1. Without affecting the manufacturing process of the laser deposition additive manufacturing module 1, it acquires data in real time. The high-speed camera 211, infrared camera 212, and laser 3D scanner 213 are positioned side-by-side, easily acquiring three types of workpiece data at the same coordinate system. The data coordinates are highly consistent, allowing for synchronous monitoring and processing, and enabling synchronous repair after feedback. Other devices with similar functions can also be used in the data acquisition device module 21 to acquire the aforementioned three types of data; this application does not impose any restrictions.

[0070] The signal transmission line module 22 is configured to transmit data to the algorithm model module 31 via a communication protocol or a high-speed transmission line such as a USB 3.0 interface. The signal transmission line module 22 is connected to all devices in the data acquisition equipment via data communication. It can internally implement a communication protocol or use a high-speed transmission line for communication, and transmit data to the computer module 3. Configuring a dedicated signal transmission line module 22 can improve communication stability and efficiency.

[0071] In one feasible implementation, continue to refer to Figure 2As shown, computer module 3 also includes a host 32 and a display 33; the host 32 is communicatively connected to the display 33, and the host 32 is also communicatively connected to the algorithm model module 31. The host 32 also includes an image processing unit 321, a computational processing unit 322, and a data storage unit 323. The image processing unit 321 is configured to process molten pool image data. The computational processing unit 322 is configured to process molten pool temperature data and 3D point cloud data. The data storage unit 323 is configured to store the data and the processed data. The host 32 is configured with two processing units to process different types of data, or three processing units can be used to process three types of data. The configuration can be adjusted according to actual needs. When the data volume is large and the processing efficiency is low, three processing units can be used to process the three types of data. When the data volume is small and the processing speed is fast, only two processing units are needed, which can save some costs.

[0072] The data storage unit 323 can save past data and record the data being monitored and generated in real time, which can be used for later model updates and analysis of the current status of the equipment.

[0073] The display 33 is configured to display real-time data and enable human-machine interaction. The display 33 can present real-time data to the operator, who can manually intervene in the laser deposition module as needed. For example, if the data shows a major defect, the operator can determine that the workpiece cannot be repaired successfully and can pause the laser deposition module to minimize losses. The display 33 also has human-machine interaction capabilities, allowing personnel to control data processing or analyze past data.

[0074] In one feasible implementation, continue to refer to Figure 2 As shown, the algorithm model module 31 includes: an LSTM time-series prediction method module, a deep learning object detection algorithm module, and a neural network classification module; wherein, the LSTM time-series prediction method module and the neural network classification module are both communicatively connected to the computing processing unit 322, and the deep learning object detection algorithm module is communicatively connected to the image processing unit 321.

[0075] The LSTM time-series prediction method module is configured to: obtain predicted molten pool temperature data by combining time-series methods with process parameters, and compare the predicted molten pool temperature data with real-time acquired molten pool temperature data to obtain the comparison result. Process parameters include time series, laser power, powder / wire feeding speed, and scanning speed. The configuration data of the LSTM time-series prediction method module can be modified according to changes in process parameters. LSTM (Long Short-Term Memory) is a type of temporal recurrent neural network (RNN), primarily designed to address the gradient vanishing and gradient exploding problems during long sequence training. Simply put, compared to ordinary RNNs, LSTM performs better with longer sequences. In this application, it can maintain data comparison functionality during real-time monitoring.

[0076] Specifically, such as Figure 4 As shown in the figure, The data represents the molten pool temperature as a time series distribution. This serves as the forward propagation activation function module 3111, and also as a storage unit for recording the current state and passing it to the next unit; symbol and They represent dot product and dot addition, respectively; tanh and Let these represent the hyperbolic tangent function and the sigmoid function, respectively; where This refers to the unit state that will be passed on to the next unit structure. Indicates the state of the previous unit. Based on the current input and The candidate cell states of the previous activation function at that location. exist and The space between them is carefully adjusted, and there are two door structures, namely... and , Forgotten Gate 3112, To update gate 3113. In an LSTM network, the gating unit controls the information entering and exiting the cell structure. Cell state. via output gate 3114 A new activation function is generated. The output of the activation function module 3111 here also represents the current input. The predicted value is obtained. As can be seen from the above, the predicted value of the molten pool temperature data can be obtained through the LSTM time series prediction method module. By comparing the predicted value with the actual value, it can be determined whether the molten pool temperature is correct, and further, whether a defect has occurred.

[0077] The deep learning object detection algorithm module is configured to: input the molten pool image data after proportional scaling, train it, and output fixed-standard molten pool image data for molten pool state detection to obtain the detection result. Specifically, in this embodiment, the molten pool image dataset 3121 is used as input. The molten pool image dataset is connected to four CBS modules, where each CBS module includes a convolutional layer, a batch normalization layer, and a SILU (Sigmoid Weighted Liner) module. The system uses a Unitsilu activation function to output the initial feature layer 3122, which is then connected to an ELAH module (this module inputs the initial feature layer 3122 into two channels, with the upper channel consisting of four CBS modules and the lower channel consisting of one CBS module, and finally concatenates the output information of the two channels into a matrix as the output of the overall model), outputting feature layer data 3133. The feature layer data 3133 is then imported into the MP modules of the two channels (this module inputs the feature layer data into two channels, with the upper channel consisting of one max-pooling layer and one CBS module connected, and the lower channel connected to a CBS module, and finally concatenates the output information of the two channels into a matrix as the overall output), outputting two feature layers in the upper and lower parts respectively, including a first feature layer 3134 and a second feature layer 3135. The first feature layer 3134 needs to be connected to two convolutional layers to output a first output feature vector layer 3134a; the second feature layer 3135 is connected to one convolutional layer to output a second output feature vector layer 3135a.

[0078] In this invention, a dataset of molten pool state images is used as input. The image size is configured according to the parameters in the deep learning object detection algorithm module and computer module 3, and the output is determined based on the actual situation. Specifically, as shown... Figure 5 As shown, with an input size of 640 Taking a 640p 3-channel RGB image as an example, the size of the output initial feature layer is 320. 320 64, the output feature layer data size is 160. 160 256; The size of the first output feature vector layer is 20. 20 (80+5) 3; The size of the second output feature vector layer is 40. 40 (80+5) 3; Based on the module's processing capabilities, this application suggests that the size of the scaled molten pool image data should be 224. Multiples of 224 pixels, with a fixed standard that the image dimensions range from 20 to 60 pixels.

[0079] The neural network classification module is configured to classify 3D point cloud data by constructing a three-layer backpropagation neural network to obtain 3D defect information. For example... Figure 6 As shown, the 3D point cloud data in this application, after undergoing corresponding data preprocessing operations, is input into a fully connected layer with 64 neural units, and then connected to two fully connected layers with sizes of 32 neural units and 1 neural unit respectively. Finally, the detection result is output. The fully connected layers need to add a ReLU (Rectified Linear Unit) activation function, and the output layer needs to add a Linear activation function. The 3D defect information can be obtained from the detection result.

[0080] The LSTM time series prediction method module, deep learning object detection algorithm module, and neural network classification module mentioned above are models that have been trained based on the target dataset and have excellent performance. In addition, the models can be designed, developed, or modified to suit different target needs according to different requirements.

[0081] In one feasible implementation, continue to refer to Figure 2 As shown, the laser deposition additive manufacturing module 1 includes: a laser deposition head 11, a rotary processing platform 12, a high-power laser 13, a powder / wire feeding device 14, and a protective gas device 15. In this embodiment, the laser deposition additive manufacturing module 1 mainly uses the laser deposition head 11 to perform a laser deposition additive manufacturing process on a substrate placed on the rotary processing platform 12. This process is mainly carried out under inert gas protection, adjusting different process parameters such as laser power, laser scanning speed, and powder / wire feeding speed to perform laser melting and deposition of powder or wire.

[0082] Specifically, the rotary processing platform 12 is used to place the substrate and perform rotary processing at a certain angular velocity; the laser deposition head 11 is located above the rotary processing platform 12 and is used to deposit materials using lasers with specific spot sizes; the high-power laser 13 is connected to the laser deposition head 11 via optical fiber and is used to transmit lasers of different powers to the laser deposition head 11 through the optical fiber; the powder / wire feeding device 14 is connected to the output end of the laser deposition head 11; the powder / wire feeding device 14 is used to directionally output the material and works together with the laser deposition head 11 to deposit the material. The protective gas device 15 is also connected to the output end of the laser deposition head 11 and is used to output protective gas to prevent high-temperature oxidation during material deposition, which could affect the performance of the target workpiece.

[0083] The laser deposition additive manufacturing module 1 of this application also includes a controller 16, which is communicatively connected to a computer module 3 and to a high-power laser 13 and a powder / wire feeding device 14. The controller 16 adjusts process parameters based on information fed back from the computer module 3. The controller 16 receives defect feedback from the computer module 3 and adjusts the process parameters accordingly. The high-power laser 13 and the powder / wire feeding device 14 determine the forming structure of the workpiece. By modifying the process parameters of the high-power laser 13 and the powder / wire feeding device 14, the purpose of repairing defects can be achieved.

[0084] As described above, the laser deposition additive manufacturing online monitoring system of this application collects data (including molten pool image data, molten pool temperature data, and 3D point cloud data) using data acquisition equipment integrated into the robot system during the laser deposition additive manufacturing process. This data is then transmitted to the algorithm model of the computer module for decision-making, classification, and detection. The final information is fed back to the laser deposition additive manufacturing module through the computer module to adjust process parameters and correct errors caused by defects and precision issues during processing. Through these operations, the control accuracy and feedback speed of the laser deposition additive manufacturing system are improved. Furthermore, the model can be updated and modified in real time based on actual needs and the data collected, making system development more convenient and operation easier.

[0085] Following the disclosure herein, other embodiments of this disclosure will readily occur. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0086] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. An online monitoring system for laser deposition additive manufacturing, characterized in that, include: Laser deposition additive manufacturing module (1), robot system module (2) and computer module (3); The robot system module (2) is communicatively connected to the computer module (3); the computer module (3) is communicatively connected to the laser deposition additive manufacturing module (1); The laser deposition additive manufacturing module (1) is configured to: melt and deposit powder or filament materials in combination with digital models to complete the forming of the target workpiece; The robot system module (2) is configured to: collect data during the laser deposition additive manufacturing process and transmit the data to the computer module (3), the data including: molten pool image data, molten pool temperature data and three-dimensional point cloud data; The computer module (3) includes an algorithm model module (31), which is configured to: train an algorithm model based on previous data, process the data in real time through the algorithm model, and save the processing results to the computer module (3); The computer module (3) is configured to process the data through the algorithm model module (31), monitor the surface of the deposited layer and the state of the molten pool in real time, and feed back the monitored defect data information, abnormal data information of the molten pool state and abnormal data information of the molten pool temperature to the laser deposition additive manufacturing module (1). The algorithm model module (31) includes: an LSTM time-series prediction method module, a deep learning target detection algorithm module, and a neural network classification module; wherein, the LSTM time-series prediction method module and the neural network classification module are both connected to the computing processing unit (322), and the deep learning target detection algorithm module is connected to the image processing unit (321); the LSTM time-series prediction method module is configured to: obtain predicted molten pool temperature data by combining time-series methods with process parameters, and compare the predicted molten pool temperature data with the real-time collected molten pool temperature data to obtain a comparison result; the deep learning target detection algorithm module is configured to: input the molten pool image data after proportional scaling, train it, and output the fixed standard molten pool image data for molten pool state detection to obtain a detection result; the neural network classification module is configured to: classify the three-dimensional point cloud data by constructing a three-layer backpropagation neural network to obtain three-dimensional defect information; the laser deposition additive manufacturing module (1) responds to the feedback information of the computer module (3).

2. The online monitoring system for laser deposition additive manufacturing according to claim 1, characterized in that, The robot system module (2) includes a data acquisition device module (21) and a signal transmission line module (22); the data acquisition device module (21) is communicatively connected to the signal transmission line module (22); and the signal transmission line module (22) is communicatively connected to the computer module (3); The data acquisition device module (21) is configured to: acquire data during the laser deposition additive manufacturing process; The signal transmission line module (22) is configured to transmit the data to the algorithm model module (31) via a communication protocol or a high-speed transmission line.

3. The online monitoring system for laser deposition additive manufacturing according to claim 2, characterized in that, The data acquisition device module (21) includes: a high-speed camera (211), an infrared camera (212), and a laser 3D scanner (213). The high-speed camera (211), the infrared camera (212), and the laser 3D scanner (213) are all communicatively connected to the signal transmission line module (22). The high-speed camera (211) is configured to acquire image data of the molten pool; The infrared camera (212) is configured to collect the molten pool temperature data; The laser 3D scanner (213) is configured to acquire the 3D point cloud data.

4. The online monitoring system for laser deposition additive manufacturing according to claim 2, characterized in that, The computer module (3) further includes a host (32) and a display (33); the host (32) is communicatively connected to the display (33), and the host (32) is also communicatively connected to the algorithm model module (31); The host (32) further includes: an image processing unit (321), a computing processing unit (322), and a data storage unit (323). The image processing unit (321) is configured to process the molten pool image data. The computing processing unit (322) is configured to process the molten pool temperature data and the three-dimensional point cloud data; The data storage unit (323) is configured to store the data and the processed data; The display (33) is configured to display real-time data and enable human-computer interaction.

5. The online monitoring system for laser deposition additive manufacturing according to claim 1, characterized in that, The process parameters include time series, laser power, powder / wire feeding speed, and scanning speed.

6. The online monitoring system for laser deposition additive manufacturing according to claim 1, characterized in that, The proportionally scaled molten pool image data has a size of 224. Multiples of 224 pixels, the fixed standard being that the image length and width range is 20 to 60 pixels.

7. The online monitoring system for laser deposition additive manufacturing according to claim 1, characterized in that, The laser deposition additive manufacturing module (1) includes: a laser deposition head (11), a rotary processing platform (12), a high-power laser (13), a powder / wire feeding device (14), and a protective gas device (15). The rotary processing platform (12) is used to place the substrate and perform rotary processing at a certain angular velocity; the laser deposition head (11) is located above the rotary processing platform (12) and is used to deposit materials using a laser with a specific spot size. The high-power laser (13) is connected to the laser deposition head (11) via an optical fiber, and is used to transmit lasers of different power to the laser deposition head (11) via the optical fiber; The powder / wire feeding device (14) is connected to the output end of the laser deposition head (11); the powder / wire feeding device (14) is used to directionally output the material and works together with the laser deposition head (11) to deposit the material; The protective gas device (15) is also connected to the output end of the laser deposition head (11); the protective gas device (15) is used to output protective gas to prevent high-temperature oxidation during material deposition, which would affect the performance of the target workpiece. It also includes a controller (16), which is communicatively connected to the computer module (3) and to the high-power laser (13) and the powder / wire feeding device (14), for adjusting process parameters based on the information fed back by the computer module (3).

8. A method for online monitoring in laser deposition additive manufacturing, characterized in that, The monitoring method, applied to any one of the laser deposition additive manufacturing online monitoring systems of claims 1-7, comprises: The laser deposition additive manufacturing of the target workpiece is performed using the laser deposition additive manufacturing module (1). During the laser deposition additive manufacturing process, the robot system module (2) acquires data of the target workpiece in real time and transmits the acquired data to the computer module (3) for storage. The computer module (3) performs real-time data processing and stores the processing results; and feeds back the monitored defect data, abnormal melt pool status data and abnormal melt pool temperature data to the laser deposition additive manufacturing module (1).

9. The online monitoring method for laser deposition additive manufacturing according to claim 8, characterized in that, The acquired data includes: molten pool image data, molten pool temperature data, and three-dimensional point cloud data.

Images