Light-filament-powder three-beam coaxial self-adaptive laser melting forming method and device

Abstract

The invention discloses a light-wire-powder three-beam coaxial self-adaptive laser melting forming method and device, and relates to the technical field of metal additive manufacturing and laser welding. A high-precision real-time control system is started, and multi-source signals of a molten pool area are collected in real time through a monitoring module; transmitting a multi-source signal acquired by the monitoring module to a central processing unit, and establishing a mapping relation model of monitoring characteristics and process quality; based on the output of the mapping relation model, the laser power, the wire feeding speed and the powder feeding rate are fed back and controlled in real time; and repeating the self-adaptive regulation and control step to realize the whole-process self-adaptive closed-loop control of the whole forming path. The light-filament-powder three-beam coaxial self-adaptive laser melting forming method and device have the advantages of being good in collaboration and multiple in regulation and control dimension, online real-time closed-loop control in the manufacturing process can be achieved, and the problems of poor forming and component control in non-uniform gap repair and functionally graded material preparation are solved.

Description

Technical Field

[0001] This invention relates to the fields of metal welding and additive manufacturing technology, and in particular to a method and apparatus for coaxial adaptive laser melting forming of three beams: light, filament, and powder. Background Technology

[0002] In the fields of laser-directed energy deposition and large-gap weld repair, using single wire or powder as filler has inherent limitations. Wire has high utilization but limited bridging ability and is sensitive to assembly gaps; powder has strong adaptability but low utilization and easily generates fumes, affecting process stability. Existing wire-powder composite processes mostly use off-axis feeding, resulting in inaccurate energy-material coupling and a narrow process window. More importantly, existing technologies lack real-time, coaxial, and multi-dimensional sensing capabilities of the interaction mechanism between "light-wire-powder" in the molten pool, making it impossible to quickly and accurately control based on the actual state of the molten pool. This leads to significant challenges in forming quality and composition control when repairing uneven gaps or manufacturing functionally graded materials. Therefore, developing a novel laser melting forming technology capable of achieving coaxial coordination of three beams and possessing intelligent closed-loop feedback capabilities is urgently needed. Summary of the Invention

[0003] The purpose of this invention is to provide a three-beam coaxial adaptive laser melting and forming method and apparatus for light-filament-powder, which solves the problems mentioned in the background art. Through innovative coaxial optical design and multi-signal fusion technology, the forming process is made visible, measurable and controllable, and ultimately achieves high-quality and highly adaptive intelligent manufacturing.

[0004] To achieve the above objectives, the present invention provides a three-beam coaxial adaptive laser melting and forming method for optical-filament-powder, comprising the following steps: S1. Activate the protective gas system to create a high-purity inert atmosphere environment inside the inert gas protection chamber; S2. Start the high-precision real-time control system, and collect multi-source signals in the molten pool area in real time through the coaxial integrated monitoring module; S3. The multi-source signals collected by the monitoring module are transmitted to the central processing unit for data fusion and feature extraction. A mapping relationship model between monitoring features and process quality is established through intelligent algorithms. The monitoring features include the molten pool area, temperature gradient and spectral intensity, and the process quality includes porosity and composition deviation. S4, based on the mapping relationship model output of S3, provides real-time feedback control of laser power, wire feeding speed and powder feeding rate, and then uses a high-precision real-time control system to adjust the laser processing system and the wire and powder synchronous feeding mechanism in real time. S5. Repeat S2 to S4 to achieve adaptive closed-loop control of the entire forming path. S3 includes: S31. The intelligent algorithm first concatenates all features to form a comprehensive feature vector. ,in, This represents the transpose of a vector, converting a row vector into a column vector. Indicates the length of the molten pool. Indicates the width of the molten pool. Indicates the area of ​​the molten pool. Indicates the highest temperature of the molten pool. Indicates the average temperature of the molten pool. This indicates the maximum cooling rate behind the molten pool. This represents the intensity ratio of the characteristic spectral lines of Ti and Al. S32. Use pre-trained machine learning regression and classification models to analyze feature vectors. The process is executed to output predicted values ​​for process quality. S33. Compare the predicted value in S31 with the preset safety threshold and generate control instructions; The machine learning regression and classification model in S32 includes a logistic regression classifier and a linear regression model, and the processing of the feature vector includes: S321, Logistic Regression Classifier Based on Feature Vectors Calculate the probability of porosity defects occurring Its decision function is: ; in, The weight vector of the classifier. For bias terms; S322, Linear regression model based on eigenvectors Predicting compositional bias of key elements Its prediction function is: ; in, The weight vector of the regressor. For bias terms; In S4, when the high-precision real-time control system adjusts the laser processing system and the wire-powder synchronous feeding mechanism in real time, the selection and application are selected according to the working conditions. When the gap between the forming areas is less than 0.1mm, single laser manufacturing is used; when the gap is between 0.1 and 0.3mm, laser powder feeding manufacturing is used; when the gap is between 0.3 and 0.5mm, laser wire feeding manufacturing is used; and when the gap is greater than 0.5mm, laser-wire-powder synchronous manufacturing is used.

[0005] Preferably, S2 includes: S21. Acquire dynamic images of the molten pool morphology, wire droplet transition behavior, and powder flow field using a high-speed vision sensor, and analyze physical dimensional characteristics, including: molten pool length. molten pool width molten pool area ; S22. Obtain the two-dimensional temperature field distribution of the molten pool and its heat-affected zone using a thermal imager, and analyze to obtain the highest temperature of the molten pool. Average temperature of the molten pool Maximum cooling rate behind the molten pool ; S23. Obtain the characteristic spectrum of the molten pool radiation using a spectrometer, and extract the intensity ratio of the characteristic spectral lines of Ti and Al. It is used to retrieve information on the temperature and composition of the molten pool.

[0006] Preferably, in step S33, the predicted value is compared with a preset safety threshold, and the determination is based on: like Then a command to reduce laser power is generated; like Then, an instruction to increase the powder feeding rate is generated; If the area of ​​the molten pool This generates commands to decrease the scanning speed and increase the wire feed speed; in, To preset a safe threshold for the probability of pores, To preset the safety threshold for component deviation, This is a preset safe threshold for the molten pool area under stable welding conditions.

[0007] A three-beam coaxial adaptive laser melting and forming device for light, filament, and powder includes a three-beam coaxial integrated head, a laser processing system, a filament and powder synchronous feeding mechanism, a protective gas system, and a high-precision real-time control system.

[0008] Preferably, the interior of the three-beam coaxial integrated head is sequentially provided with a first collimating lens group, a reflector group, a second collimating lens group, and a hollow reflector group; One end of the three-beam coaxial integrated head is provided with a ring-shaped guide tube and a ring-shaped powder nozzle, and the ring-shaped powder nozzle is coaxially wrapped around the outside of the ring-shaped guide tube; the other end of the three-beam coaxial integrated head is connected to a powder feeding channel, which passes through the inside of the three-beam coaxial integrated head and is connected to the ring-shaped powder nozzle.

[0009] Preferably, the laser processing system includes a laser generator and a transmission optical fiber, wherein the laser generator generates a high-energy laser beam, which is guided to the three-beam coaxial integrated head through the transmission optical fiber; The synchronous feeding mechanism for filament and powder includes a filament feeder and a powder feeder. The filament feeder is connected to the annular shaft guide tube through a flexible filament guide tube. The powder feeder is connected to the powder storage tank and to the powder feeding channel through a powder feeding hose. The protective gas system includes a protective gas cylinder, an inert gas protective chamber, and a gas supply pipe. The inert gas protective chamber is installed on the welding work platform. The high-precision real-time control system includes a monitoring module, a motion control card, a six-axis robot, and a central controller. The signal input terminal of the central controller is connected to the monitoring module, which includes a high-speed vision sensor, a thermal imager, and a spectrometer. The control terminal of the motion control card is connected to the control terminals of the six-axis robot, the laser generator, the wire feeder, and the powder feeder, and is used to receive control commands issued by the central processor and drive the actuators to move.

[0010] Preferably, the laser generator, the wire feeder, and the powder feeder are coaxially and concentrically distributed in space. The wire fed by the wire feeder is transported along the central axis through the annular guide tube. The annular laser beam emitted by the laser generator is transported around the periphery of the wire feeding shaft. The powder flow transported by the powder feeder forms a conical flow field around the annular laser beam through the annular powder nozzle.

[0011] Therefore, the present invention employs the above-mentioned three-beam coaxial adaptive laser melting and forming method and device of light-filament-powder, which realizes the precise coordinated transport of energy and materials and the comprehensive perception of process status. Through multi-source information fusion and intelligent closed-loop feedback, it achieves a leap from "passive execution" to "active control", significantly improving the forming accuracy, microstructure and composition control capabilities in the repair of uneven gaps, manufacturing of complex structural parts and preparation of functional graded materials. It has the advantages of high integration and strong adaptability.

[0012] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating an embodiment of the optical-filament-powder three-beam coaxial adaptive laser melting and forming method and apparatus of the present invention; Figure 2 This is a schematic diagram of the structure of the optical-filament-powder three-beam coaxial adaptive laser melting and forming method and apparatus of the present invention; Figure 3 This is a schematic diagram of the internal structure of the three-beam coaxial integrated head of the optical-filament-powder three-beam coaxial adaptive laser melting forming method and device of the present invention; Reference numerals: 1. Protective gas cylinder; 2. Inert gas protective chamber; 3. Powder feeder; 4. Powder storage tank; 5. Laser generator; 6. Transmission fiber optic cable; 7. Wire feeder; 8. Welding work platform; 9. High-speed vision sensor; 10. Thermal imager; 11. Three-beam coaxial integrated head; 111. First collimating lens group; 112. Reflector group; 113. Second collimating lens group; 114. Hollow reflector group; 115. Circular shaft wire guide tube; 116. Powder feeding channel; 117. Circular shaft powder nozzle; 118. Gas delivery pipe; 12. Spectrometer; 13. Central processing unit; 14. Six-axis robot; 15. Motion control card. Detailed Implementation

[0014] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0015] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0016] Example Please see Figures 1-3 This invention provides a three-beam coaxial adaptive laser melting and forming method for light, filament, and powder. This method employs a monitoring module to observe and detect the interaction between light, filament, and powder and the geometric and thermophysical characteristics of the molten pool in real time during laser melting. Through multi-source information fusion algorithm compensation calculations, the calculation results are fed back, and process parameters such as laser power, filament feeding speed, and powder feeding rate are adaptively adjusted based on the results, ultimately achieving shape and property coordinated control of the laser melting and forming process. Specifically, it includes the following steps: S1. Activate the protective gas system to create a high-purity inert atmosphere environment inside the inert gas protection chamber 2.

[0017] S2. Activate the high-precision real-time control system, which uses a coaxial integrated monitoring module to acquire multi-source signals from the molten pool region in real time. This includes: S21. Acquire dynamic images of the molten pool morphology, wire droplet transition behavior, and powder flow field using a high-speed vision sensor 9, and analyze physical dimensional characteristics, including: molten pool length. molten pool width molten pool area ; S22. The two-dimensional temperature field distribution of the molten pool and its heat-affected zone is obtained by thermal imager 10, and the highest temperature of the molten pool is obtained by analysis. Average temperature of the molten pool Maximum cooling rate behind the molten pool ; S23. Obtain the characteristic spectrum of the molten pool radiation using spectrometer 12, and extract the intensity ratio of the characteristic spectral lines of Ti and Al. It is used to retrieve information on the temperature and composition of the molten pool.

[0018] S3. The multi-source signals collected by the monitoring module are transmitted to the central processing unit 13 for data fusion and feature extraction. An intelligent algorithm is used to establish a mapping model between monitoring features (molten pool area, temperature gradient, and spectral intensity) and process quality (porosity and compositional deviation). This includes: S31. The intelligent algorithm first concatenates all features to form a comprehensive feature vector. ,in, This represents the transpose of a vector, converting the original row vector into a column vector. S32. Use pre-trained machine learning regression and classification models to analyze feature vectors. The process involves processing the data to output predicted values ​​for process quality. Machine learning regression and classification models, including logistic regression classifiers and linear regression models, process the feature vectors as follows: S321, Logistic Regression Classifier Based on Feature Vectors Calculate the probability of porosity defects occurring Its decision function is: ; in, The weight vector of the classifier. For bias terms; S322, Linear regression model based on eigenvectors Predicting compositional bias of key elements Its prediction function is: ; in, The weight vector of the regressor. This is a bias term.

[0019] S33. Compare the predicted value in S31 with the preset safety threshold and generate a control command. The comparison between the predicted value and the preset safety threshold is based on the following criteria: like Then a command to reduce laser power is generated; like Then, an instruction to increase the powder feeding rate is generated; If the area of ​​the molten pool This generates commands to decrease the scanning speed and increase the wire feed speed; in, To preset a safe threshold for the probability of pores, To preset the safety threshold for component deviation, This is a preset safe threshold for the molten pool area under stable welding conditions.

[0020] S4, based on the mapping model output of S3, provides real-time feedback control of laser power, wire feeding speed, and powder feeding rate. This, in turn, enables real-time regulation of the laser processing system and the wire-powder synchronous feeding mechanism through a high-precision real-time control system. During real-time regulation, the application is selected based on the working conditions: single-laser manufacturing is used when the gap between the forming areas is less than 0.1mm; laser powder feeding is used when the gap is between 0.1 and 0.3mm; laser wire feeding is used when the gap is between 0.3 and 0.5mm; and simultaneous laser-wire-powder manufacturing is used when the gap is greater than 0.5mm.

[0021] S5. Repeat S2 to S4 to achieve adaptive closed-loop control of the entire forming path.

[0022] A three-beam coaxial adaptive laser melting and forming device for light, filament, and powder includes a three-beam coaxial integrated head 11, a laser processing system, a filament and powder synchronous feeding mechanism, a protective gas system, and a high-precision real-time control system.

[0023] The interior of the three-beam coaxial integrated head 11 is sequentially equipped with a first collimating lens group 111, a reflector group 112, a second collimating lens group 113, and a hollow reflector group 114. The first and second collimating lens groups 111 and 113 collimate the incident laser beam, while the reflector group 112 alters the optical path of the collimated laser beam. The hollow reflector group 114 achieves a 90° bend in the optical path and has a hollow diameter of 50-100mm, thus altering the optical path and acting on the workpiece surface to form a molten pool. One end of the three-beam coaxial integrated head 11 is equipped with a ring-shaped guide tube 115 and a ring-shaped powder nozzle 117, with the ring-shaped powder nozzle 117 coaxially surrounding the outside of the ring-shaped guide tube 115. The other end of the three-beam coaxial integrated head 11 is connected to a powder feeding channel 116, which passes through the interior of the three-beam coaxial integrated head 11 and connects to the ring-shaped powder nozzle 117. The powder is connected to the annular powder nozzle 117 by the powder feeding channel 116, which is coaxially wrapped around the outside of the annular wire guide tube 115 to form a conical powder flow and coaxially intersect with the laser beam and wire on the workpiece surface.

[0024] The laser processing system includes a laser generator 5 and a transmission fiber 6. The laser generator 5 generates a high-energy laser beam, which is guided to a three-beam coaxial integrated head 11 through the transmission fiber 6.

[0025] The synchronous feeding mechanism for filament and powder includes a filament feeder 7 and a powder feeder 3. The filament feeder 7 is connected to the annular shaft guide tube 115 via a flexible filament guide tube. The powder feeder 3 is connected to the powder storage tank 4 and to the powder feeding channel 116 via a powder feeding hose.

[0026] The protective gas system includes a protective gas cylinder 1, an inert gas protective chamber 2, and a gas supply pipe 118. The inert gas protective chamber 2 is installed on the welding work platform 8, so that the workpiece to be manufactured is always in an inert gas atmosphere.

[0027] The high-precision real-time control system includes a monitoring module, a motion control card 15, a six-axis robot 14, and a central controller. The central controller receives and processes multi-source monitoring data, runs a built-in intelligent algorithm model, and generates process parameter control commands. The signal input terminal of the central controller is connected to the monitoring module, which includes a high-speed vision sensor 9, a thermal imager 10, and a spectrometer 12. The control terminal of the motion control card 15 is connected to the control terminals of the six-axis robot 14, the laser generator 5, the wire feeder 7, and the powder feeder 3, and is used to receive control commands issued by the central processor 13 and drive the actuators to move.

[0028] The laser generator 5, wire feeder 7, and powder feeder 3 are spatially coaxial and concentrically distributed. The wire fed by the wire feeder 7 is conveyed along the central axis through the annular guide tube 115. The annular laser beam emitted by the laser generator 5 is conveyed around the periphery of the wire feeder shaft. The powder flow conveyed by the powder feeder 3 is formed by the annular powder nozzle 117 surrounding the annular laser beam to create a conical flow field. The annular laser beam, wire, and powder coaxially converge on the workpiece surface, forming a spatial coupling structure with "wire in the center, laser ring in the middle, and powder ring outside," achieving synchronous, point-to-point, and unidirectional transport of energy and materials, ensuring thermal-mechanical-mass triple coupling of light, wire, and powder within the molten pool.

[0029] The present invention is further illustrated below through a specific embodiment. This embodiment addresses the surface defect repair and functional graded layer fabrication of 20mm thick titanium alloy TC4 plates, solving the problems of weak wire bridging ability, low powder utilization, and difficulty in controlling forming quality in traditional repair processes. It achieves high-precision, high-quality forming of 20mm thick titanium alloy TC4 components. Specifically, it includes the following steps: S1. Assemble the 20mm thick TC4 titanium alloy test plate to be welded onto the tooling fixture, with the reserved butt joint gap increasing linearly from 0.2mm to 1.5mm along the welding direction. Clean the area to be welded and the surface of the filler with acetone to remove oil stains. Close the door of the inert gas protection chamber 2, turn on the protective gas system, and introduce high-purity argon gas to replace the air in the chamber for at least 5 minutes to ensure that the oxygen content is always below 100ppm during the welding process.

[0030] S2. Turn on the power to the central processing unit 13, start the real-time monitoring software, and initialize the high-speed camera, thermal imager 10, and spectrometer 12 respectively: (1) The high-speed camera acquires dynamic images of the molten pool according to the set parameters (resolution 2592×1944, frame rate 1000fps, exposure time 50μs, gain 10dB), stores one frame every 10ms, and analyzes the length of the molten pool using an image recognition algorithm. molten pool width molten pool area Initial monitoring showed =3.5mm =2.0mm =5.2mm 2 .

[0031] (2) Thermal imager 10 acquires molten pool temperature field images according to the set parameters (resolution 640×512, temperature range 500-1800℃, frame rate 60fps), outputs temperature field data every 20ms, and extracts the highest temperature of the molten pool through temperature field analysis algorithm. Average temperature of the molten pool Maximum cooling rate behind the molten pool Initial monitoring showed =1650℃ =1200℃ =80℃ / ms.

[0032] (3) The spectrometer 12 collects the radiation spectrum of the molten pool according to the set parameters (wavelength range 200-1100nm, resolution 0.02nm, integration time 50ms), and outputs the spectral data once every 50ms. The intensity ratio of the characteristic spectral lines of Ti (498.17nm) and Al (394.40nm) is extracted by the spectral analysis algorithm. Initial monitoring showed =8.5.

[0033] Start the motion control card 15 and the six-axis robot 14, set the robot scanning path to a zigzag shape, the scanning speed to 5mm / s, the scanning interval to 1.5mm, and the number of forming layers to 10. The robot drives the three-beam coaxial integration head 11 to move along the set path, and maintains a distance of 15mm between the integration head and the substrate surface during the movement.

[0034] S3, the central processing unit 13 receives multi-source monitoring data transmitted from the high-speed camera, thermal imager 10, and spectrometer 12, aligns the data according to the timestamp, and concatenates the feature values ​​into a comprehensive feature vector: ; in, The transpose operator represents the transpose of a vector, converting a single-row vector into a column vector.

[0035] The built-in pre-trained machine learning model is invoked (the training data comes from 100 sets of 20mm titanium alloy TC4 forming experiments, covering monitoring data and process quality inspection results under different laser power, wire feeding speed, and powder feeding rate): (1) Logistic Regression Classifier: ; in, The weight vector of the classifier. As the bias term, the calculation yields (Preset porosity probability safety threshold) =0.3, < .

[0036] (2) Linear regression model: in, The weight vector of the regressor. As the bias term, the calculation yields (Preset component deviation safety threshold) =0.15%, (The ingredients are qualified).

[0037] S4, the central processing unit 13 outputs the process quality prediction value from the machine learning model ( =0.12%, =0.08%) was compared with the preset safety threshold, and the current forming state was determined to be stable, but the molten pool area was... =5.2mm 2 Slightly below the optimal molten pool area range (5.5-6.0 mm) 2 (This requires fine-tuning of process parameters to improve molding quality.)

[0038] (1) The central processing unit 13 generates control instructions: "reduce the scanning speed to 4.8 mm / s and increase the wire feeding speed to 5.2 mm / s", and sends the instructions to the motion control card 15 via Ethernet.

[0039] (2) After receiving the instruction, the motion control card 15 drives the six-axis robot 14 to reduce the scanning speed from 5 mm / s to 4.8 mm / s and drives the wire feeder 7 to increase the wire feeding speed from 5 mm / s to 5.2 mm / s, while keeping the laser power (initially set to 2500W), powder feeding rate (initially set to 10g / min), and protective gas flow rate (10L / min) stable.

[0040] S5. Repeat S2 to S4, adjust the process parameters, and then use a high-speed camera to acquire real-time images of the molten pool and analyze the molten pool area. Increased to 5.6mm 2 (Entering the optimal range); thermal imager 10 monitored... =1680℃ =1220℃ =75℃ / ms (more stable temperature field); monitored by spectrometer 12 =8.3 (more uniform composition); the central processing unit 13 processes the multi-source signal again and calculates... =0.08% < , The optimal forming state was determined, and no further adjustment of process parameters was required. This ultimately solved the problems of weak wire bridging ability, low powder utilization rate and difficulty in controlling forming quality in traditional repair processes, and achieved high-precision and high-quality forming of 20mm titanium alloy TC4 components.

[0041] Therefore, the present invention employs the above-mentioned three-beam coaxial adaptive laser melting and forming method and device of light-filament-powder, which realizes the precise coordinated transport of energy and materials and the comprehensive perception of process status. Through multi-source information fusion and intelligent closed-loop feedback, it achieves a leap from "passive execution" to "active control", significantly improving the forming accuracy, microstructure and composition control capabilities in the repair of uneven gaps, manufacturing of complex structural parts and preparation of functional graded materials. It has the advantages of high integration and strong adaptability.

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A light-filament-powder three-beam coaxial self-adapting laser melting additive manufacturing method, characterized in that, The method comprises the following steps: S1, turn on the protective gas system to form a high-purity inert atmosphere environment in the cabin; S2, start the high-precision real-time control system, and collect multi-source signals of the molten pool area in real time through the coaxial integrated monitoring module; S3, transmit the multi-source signals collected by the monitoring module to the central processor for data fusion and feature extraction, and establish a mapping relationship model of monitoring features and process quality through intelligent algorithms, wherein the monitoring features include molten pool area, temperature gradient and spectral intensity, and the process quality includes porosity and composition deviation; S4, based on the output of the mapping relationship model of S3, real-time feedback control laser power, wire feeding speed and powder feeding rate, and then real-time regulation and control of the laser processing system and the wire-powder synchronous feeding mechanism through the high-precision real-time control system; S5, repeat S2 to S4 to realize whole-process adaptive closed-loop control of the whole forming path; The S3 comprises: S31, the intelligent algorithm first splices all the features to form a comprehensive feature vector wherein, represents the transposition of the vector, and the original row vector is transposed into a column vector, represents the length of the molten pool, represents the width of the molten pool, represents the area of the molten pool, represents the highest temperature of the molten pool, represents the average temperature of the molten pool, represents the maximum cooling rate behind the molten pool, represents the characteristic spectral line intensity ratio of Ti and Al elements; S32, using a pre-trained machine learning regression and classification model on the feature vector processing, output a predicted value of the process quality; S33, compare the predicted value in S31 with the preset safety threshold, and generate a control instruction; The machine learning regression and classification model in S32 comprises a logistic regression classifier and a linear regression model, and processing the feature vector comprises: S321、the logic regression classifier determines the probability of the gas hole defect occurring according to the feature vector calculating the probability of the gas hole defect occurring and the decision function thereof is ； wherein, is a weight vector of the classifier, is a bias term; S322、the linear regression model according to the feature vector predicting compositional bias of key elements with a prediction function of: ； wherein, is a weight vector of the regressor, is a bias term; When the laser processing system and the wire-powder synchronous feeding mechanism are real-time regulated and controlled through the high-precision real-time control system in S4, the application is selected according to the working conditions, single laser manufacturing is adopted when the gap between the forming areas is less than 0.1mm, laser powder manufacturing is adopted when the gap is between 0.1mm and 0.3mm, laser wire manufacturing is adopted when the gap is between 0.3mm and 0.5mm, and laser wire-powder synchronous manufacturing is adopted when the gap is greater than 0.5mm.

2. The light-filament-powder three-beam coaxial self-adapting laser melting forming method according to claim 1, characterized in that, The S2 comprises: S21, acquiring dynamic images of the molten pool shape, the wire droplet transfer behavior and the powder flow field by a high-speed visual sensor, and analyzing physical size characteristics, including: the length of the molten pool , the width of the molten pool , the area of the molten pool ; S22, obtain the two-dimensional temperature field distribution of the molten pool and its heat-affected zone by the thermal imager, and obtain the highest temperature of the molten pool by analysis , the average temperature of the molten pool , the maximum cooling rate behind the molten pool ; S23, obtain the characteristic spectrum of the radiation of the molten pool by the spectrometer, and extract the characteristic spectral line intensity ratio of Ti and Al elements for inverting the temperature and composition information of the molten pool.

3. The light-filament-powder three-beam coaxial self-adapting laser melting forming method according to claim 2, characterized in that: In S33, the predicted value is compared with the preset safety threshold, and the judgment is based on: If a reduced laser power command is generated; If then generate an increase powder feed rate command; If the molten pool area then generate a decrease the scan speed and increase the wire feed speed command; wherein, is a preset porosity probability safety threshold value, is a preset composition deviation safety threshold value, is a preset molten pool area safety threshold value under a stable welding process condition.

4. A light-fiber-powder coaxial adaptive laser melting forming device applied to the light-fiber-powder coaxial adaptive laser melting forming method of any one of claims 1-3, characterized in that: It comprises a three-beam coaxial integrated head, a laser processing system, a wire-powder synchronous feeding mechanism, a protective gas system and a high-precision real-time control system.

5. The light-filament-powder three-beam coaxial self-adapting laser melting forming device according to claim 4, characterized in that: The first collimating mirror group, the mirror group, the second collimating mirror group and the hollow mirror group are arranged in the three-beam coaxial integrated head in sequence; One end of the three-beam coaxial integrated head is provided with a ring shaft wire guide pipe and a ring shaft powder nozzle, the ring shaft powder nozzle is coaxially surrounded outside the ring shaft wire guide pipe; the other end of the three-beam coaxial integrated head is connected with a powder feeding channel, the powder feeding channel passes through the inside of the three-beam coaxial integrated head and is connected with the ring shaft powder nozzle.

6. The light-filament-powder three-beam coaxial self-adapting laser melting forming device according to claim 5, characterized in that: The laser processing system comprises a laser generator and a transmission optical fiber, the laser generator generates a high-energy laser beam which is guided to the three-beam coaxial integrated head through the transmission optical fiber; The wire-powder synchronous feeding mechanism comprises a wire feeder and a powder feeder, the wire feeder is connected with the ring shaft wire guide pipe through a flexible wire guide pipe; the powder feeder is connected with a powder storage tank and connected with the powder feeding channel through a powder feeding hose; The protective gas system comprises a protective gas cylinder, an inert gas protection cabin and a gas conveying pipe, the inert gas protection cabin is arranged on a welding work platform; The high-precision real-time control system comprises a monitoring module, a motion control card, a six-axis robot and a central controller, a signal input end of the central controller is connected with the monitoring module, the monitoring module comprises a high-speed visual sensor, a thermal imager and a spectrometer; a control end of the motion control card is connected with control ends of the six-axis robot, the laser generator, the wire feeder and the powder feeder, used for receiving a regulation and control instruction issued by the central processor and driving an actuator to act.

7. The light-filament-powder three-beam coaxial self-adapting laser melting forming device according to claim 6, characterized in that: The laser generator, the wire feeder and the powder feeder are coaxially and concentrically distributed in space, wire material conveyed by the wire feeder is conveyed along a central axis through the ring shaft wire guide pipe, the annular laser beam emitted by the laser generator is conveyed around the periphery of the wire shaft, and the powder flow conveyed by the powder feeder forms a conical flow field around the periphery of the annular laser beam through the ring shaft powder nozzle.

Images