Asymmetric thermal management for laser-arc hybrid welding method and system

Abstract

The application discloses a laser-arc hybrid welding method and system of asymmetric heat management, and relates to the technical field of intelligent welding control. k A and k B , and calculates a thermal conductivity ratio R = k A / k B ; in the welding process, visual image data, arc plasma spectrum data and infrared temperature field distribution data of a molten pool region are synchronously acquired; based on the thermal conductivity ratio R and the infrared temperature field distribution data, a transverse offset amount of a laser beam is calculated and controlled d ; based on a penetration state extracted from the visual image data, energy matching of laser power and arc current is dynamically adjusted, and protective gas blowing flow rates of first and second base material sides are synchronously and independently controlled; the acquired data are input into a preset reinforcement learning decision model, a process parameter combination of a next moment is continuously output, and until a welding task is completed. The application realizes asymmetric heat input and asymmetric cooling.

Description

Technical Field

[0001] This invention relates to the field of intelligent welding control technology, and in particular to a laser-arc hybrid welding method and system with asymmetric thermal management. Background Technology

[0002] As high-end equipment manufacturing moves towards lightweight and high-performance, hybrid structures made of dissimilar materials, such as aluminum alloys and steel, or titanium alloys and stainless steel, are widely used. Laser-arc hybrid welding technology combines the high energy density of lasers with the excellent bridging ability of electric arcs, making it an ideal process for joining dissimilar materials. However, the significant differences in the thermophysical properties (such as thermal conductivity, melting point, and specific heat capacity) on both sides of the dissimilar materials place extremely high demands on the control of welding heat input.

[0003] Currently, the industry mainly relies on the following two traditional solutions for heat input control in laser-arc hybrid welding of heterogeneous materials, but both have obvious technical defects: The first category of existing technologies: welding schemes based on equal heat input or static fixed bias.

[0004] For example, some existing patent documents typically employ methods such as aligning the laser beam and the electric arc heat source with the center of the joint (equal heat input), or relying on human experience to set a fixed laser offset before welding (static offset).

[0005] The drawback is that this approach cannot adapt to the dynamic heat accumulation during the welding process. Because the material side, with its high thermal conductivity and low melting point, transfers heat extremely quickly and has a low melting threshold, heat rapidly spreads to a large area as welding progresses. Even with a fixed laser offset, this side will still collapse or overheat and vaporize rapidly due to the simultaneous softening of a large heat-affected zone. Conversely, the material side, with its low thermal conductivity, may not have fully melted due to the tendency for heat to accumulate locally and its high melting point. This static heat input method cannot dynamically correct itself based on real-time temperature field changes, easily leading to excessive growth of brittle intermetallic compounds at the interface, weakening the mechanical properties of the joint.

[0006] The second category of existing technologies: single adjustment of laser / arc power and symmetrical cooling scheme.

[0007] Other existing closed-loop control patents, although they introduce visual or voltage sensing to adjust the total power of the laser or electric arc in real time, still use traditional coaxial protective gas or symmetrical off-axis protective gas systems in the cooling and heat dissipation stages.

[0008] The drawback is that thermal imbalance at the interface of heterogeneous materials is not only caused by heating but also closely related to heat dissipation. Traditional symmetrical protective gas blowing gives the same convective heat transfer coefficient on both sides of the joint. When the high thermal conductivity side of the base material faces the risk of overheating and accelerated compound formation, the system can only cope by reducing the overall system power. This often directly leads to fatal defects such as incomplete fusion and incomplete penetration in the low thermal conductivity side of the base material, resulting in a process deadlock where one aspect is neglected while the other is addressed.

[0009] In summary, existing laser-arc hybrid welding technologies for heterogeneous materials lack the ability to dynamically and collaboratively control the global physical processes of heat input and heat dissipation at complex interfaces. How to break away from conventional symmetrical thermal management thinking and achieve dynamic asymmetrical heat input and asymmetrical cooling at the interface, thereby stably suppressing the growth of intermetallic compounds, is a pressing technical challenge in this field. Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides a laser-arc hybrid welding method and system with asymmetric thermal management, which can achieve asymmetric heat input and asymmetric cooling.

[0011] In order to achieve the objective of this invention, the following solution is proposed: A laser-arc hybrid welding method with asymmetric thermal management includes the following steps: S1. Parameter Acquisition: Obtain the thermal conductivity of the first and second base materials to be welded. k A and k B And calculate the thermal conductivity ratio. R = k A / k B .

[0012] S2. Multi-source sensor real-time monitoring: During the welding process, visual image data of the molten pool area, arc plasma spectral data, and infrared temperature field distribution data are acquired simultaneously.

[0013] S3, Asymmetric Heat Distribution: Based on Thermal Conductivity Ratio R In addition to infrared temperature field distribution data, the lateral offset of the laser beam is calculated and controlled. d This causes the center of the laser beam spot to shift towards the substrate material on the side with lower thermal conductivity.

[0014] S4, Dynamic Coordinated Heat Control: S401. Based on the melting state extracted from visual image data, the energy ratio of laser power and arc current is dynamically adjusted, and the protective gas blowing flow rate on the first and second base material sides is simultaneously and independently controlled to achieve asymmetric thermal management of the interface. S402. Perform low-level phase binding between the laser pulse signal generator and the servo motor of the wire feeder; when the laser pulse is in the peak power range, control the wire feeder to accelerate forward feeding; at the instant the laser pulse switches to the base power range, control the wire feeder to perform mechanical retraction, and force the molten droplet to fall off through mechanical pulling force.

[0015] S5. Closed-loop optimization: Input the data obtained from steps S1 to S4 into the preset reinforcement learning decision model, and continuously output the process parameter combination for the next moment until the welding task is completed; the process parameters include laser power, arc current, wire feed / retraction speed, lateral offset and shielding gas blowing flow rate, etc.

[0016] Furthermore, step S2 also includes: acquiring acoustic signal characteristics during the welding process in real time through acoustic emission sensors; when specific acoustic signal characteristics representing the generation of microcracks are identified, an emergency process intervention command is triggered to reduce the overall heat input. The means of reducing the overall heat input include reducing laser power, arc current, and shielding gas blowing flow rate.

[0017] Furthermore, in step S3, the lateral offset d The calculation formula is: d=f ( R )+△ d ( T ) in, f ( R (Based on thermal conductivity ratio) R The initial offset reference value, △ d ( T This is a dynamic compensation amount based on the actual temperature difference between the parent material and the target temperature difference in the infrared temperature field distribution data. When the temperature of the parent material on the high thermal conductivity side approaches the collapse threshold, the offset to the parent material on the low thermal conductivity side is increased.

[0018] Furthermore, in step S4, the method for regulating the protective gas blowing flow rate is as follows: a first protective gas nozzle and a second protective gas nozzle are respectively set at the two base materials on both sides; when the critical high temperature zone width of one side of the base material in the infrared temperature field distribution data exceeds the safety threshold, the blowing flow rate of the corresponding protective gas nozzle of the base material on that side is automatically increased, thereby accelerating the cooling rate of that side by enhancing convective heat transfer and suppressing the excessive growth of intermetallic compounds.

[0019] Furthermore, in step S4, when the penetration state is incomplete, the laser power is increased while maintaining or decreasing the arc current; when the penetration state is over-penetrated, the laser power is decreased while increasing the arc current; when the penetration state is achieved, the laser power and arc current remain unchanged.

[0020] Furthermore, the arc plasma spectral data obtained in step S2 is used to monitor the vaporization state of the metal base material. Traditional welding systems rely solely on infrared surface thermometry, which suffers from a time lag in heat conduction. This invention uses arc plasma spectral data and infrared temperature field distribution data for cross-validation, which can significantly reduce thermal damage at heterogeneous interfaces. Once the fiber optic spectrometer detects an abnormal surge in metal vapor concentration on the high thermal conductivity base material side, it indicates that this side is on the verge of severe vaporization. At this point, even if the infrared thermometry has not yet reached the collapse threshold, the welding system of this invention will immediately intervene, treating it as a high-priority danger signal and forcibly reducing the laser peak power.

[0021] A laser-arc hybrid welding system with asymmetric thermal management is used to implement the welding method described above, including a parameter acquisition module, a multi-source sensor real-time monitoring module, a dynamic collaborative heat control module, and a collaborative optimization decision engine.

[0022] The parameter acquisition module is used to input and retrieve the thermal conductivity of the first and second base materials. k A and k B ; The multi-source sensing real-time monitoring module includes an industrial camera, a fiber optic spectrometer, and an infrared thermal imager mounted coaxially or off-axis, which are used to acquire visual image data of the molten pool area, arc plasma spectral data, and infrared temperature field distribution data, respectively; it also includes an acoustic emission sensor for acquiring acoustic signal characteristics during the welding process. The dynamic collaborative heat control module includes an optical galvanometer system or a precision micro-motion slide at the end of the laser beam to control the lateral offset and positioning, and a dual-channel gas proportional valve to control the blowing flow of protective gas on both sides respectively. The collaborative optimization decision engine, built into the industrial control computer, is based on a pre-trained reinforcement learning decision model. It receives data from the parameter acquisition module and the multi-source sensor real-time monitoring module, and issues real-time control commands to the dynamic collaborative heat control module, laser, arc welding machine and wire feeder.

[0023] The beneficial effects of this invention are as follows: 1. By combining laser spatial deflection with independent control of dual-path protective gas, the thermodynamic state of the heterogeneous interface is balanced, effectively suppressing the formation of brittle compounds and improving tensile strength. 2. It can automatically correct welding parameters based on real-time infrared temperature and visual status, which greatly reduces the cost of changing and debugging heavy non-standard parts. 3. Introduce acoustic signal characteristics as an auditory early warning mechanism for microcracks. Once a defect is detected, it can trigger the welding system to cool down or reduce power, minimizing the defect rate. 4. The physical pulling force of the mechanical retraction of the welding wire is used to forcibly cut off the molten droplets, making the droplets extremely small and uniform, and the additional heat input brought into the molten pool is strictly controlled. 5. Using arc plasma spectral data and infrared temperature field distribution data for cross-validation can greatly reduce thermal damage at heterogeneous interfaces. Attached Figure Description

[0024] Figure 1 A flowchart of the welding method is shown; Figure 2 The overall architecture and data flow diagram of the welding system are shown. Figure 3 A schematic diagram of the asymmetric welded end is shown; Figure 4 A schematic diagram showing the isotherm profile and the width of the critical high-temperature zone is shown. Detailed Implementation

[0025] Example 1 like Figure 1 As shown, this embodiment provides a laser-arc hybrid welding method with asymmetric thermal management, including the following steps: S1. Parameter Acquisition: Obtain the thermal conductivity of the first base material to be welded. k A Thermal conductivity of the second base material k B And calculate the ratio of their thermal conductivity. R = k A / k B This ratio directly reflects the inherent difference in the heat absorption and dissipation capabilities of the two parent materials.

[0026] S2. Multi-source sensor real-time monitoring: During the welding process, the central controller of the welding system synchronously sends high-frequency hardware trigger signals (such as TTL level pulses) to the industrial camera, fiber optic spectrometer and infrared thermal imager to simultaneously acquire visual image data of the molten pool area, arc plasma spectral data and infrared temperature field distribution data.

[0027] Preferably, step S2 also incorporates an acoustic emission sensor for auditory defect monitoring. The acoustic emission sensor is attached to the surface of the base material and collects acoustic signal characteristics during the welding process in real time. Since brittle intermetallic compounds release specific frequency acoustic signals when they crack or microcracks initiate, the welding system will trigger an emergency process intervention command when it detects these specific frequency signals, reducing the overall heat input. Methods for reducing overall heat input include reducing laser power, arc current, and shielding gas flow rate.

[0028] S3, Asymmetric Heat Distribution: Based on Thermal Conductivity Ratio RIn addition to infrared temperature field distribution data, the lateral offset of the laser beam is calculated and controlled. d This causes the center of the laser beam spot to shift towards the side of the substrate with lower thermal conductivity.

[0029] Lateral offset d The calculation formula is: d=f ( R )+△ d ( T ) f ( R )= K ·( k B - k A ) / ( k B + k A ) in, f ( R (Based on thermal conductivity ratio) R The initial offset reference value is determined before welding based on the inherent difference in thermal conductivity between the base materials on both sides. K It is an empirical proportionality constant (dimensions in mm), typically taken as 0.5 to 1.5 times the laser spot radius, and needs to be determined through pre-welding calibration experiments. If the base materials on both sides are the same material, f ( R The value is 0. If the thermal conductivity of the second base material is... k B Thermal conductivity much greater than that of the first substrate k A ,but( k B - k A ) / ( k B + k A This part of the value will be close to 1, so the laser will be deflected towards the first substrate (with lower thermal conductivity). K The distance.

[0030] Among them, △ d ( T The value represents the dynamic compensation amount Δ, calculated based on the actual temperature difference between the base material and the target temperature difference in the infrared temperature field distribution data. In dissimilar material welding, the high thermal conductivity base material is prone to collapse due to heat accumulation. The welding system calculates the temperature difference between the target and actual temperatures of the high thermal conductivity base material in real time. When the actual temperature of the high thermal conductivity base material approaches the collapse threshold, the welding system immediately calculates the dynamic compensation amount Δ. d ( TThis further increases the deflection of the laser beam towards the low thermal conductivity side of the substrate. △ d ( T This is obtained through a PI (proportional-integral) control model, as shown in the following formula: in, e ( t The actual temperature value and the target temperature value at time ( ) t The deviation, i.e., the actual temperature minus the target temperature, occurs when the actual temperature does not exceed the target temperature. e ( t The value is 0. K p For proportional gain, K i This is the integral gain. The proportional (P) output is proportional to the current deviation, i.e. K p · e ( t The output of the integral (I) part is proportional to the integral of the deviation over time, i.e. The PI (proportional-integral) control model is a linear feedback controller, which is existing technology and will not be elaborated on further.

[0031] S4, Dynamic Coordinated Heat Control: S401, Energy-Gas Ratio Coordination: Based on visual image data, the feature contour of the keyhole is extracted and its pixel area is calculated using an image grayscale threshold segmentation algorithm (a mature existing technology). Simultaneously, the maximum width of the molten pool is extracted, and the ratio of the keyhole area to the molten pool width is used as a feature index characterizing the current penetration state. Then, the energy ratio of laser power and arc current is dynamically adjusted according to the penetration state. When the keyhole is extremely small or not open, the current penetration state is determined to be incomplete, and the laser power is increased while maintaining or decreasing the arc current. When the keyhole is excessively expanded, the current penetration state is determined to be over-penetrated, and the laser power is decreased while increasing the arc current. When the keyhole expansion is within a certain range, the current penetration state is determined to be adequate, and the laser power and arc current remain unchanged.

[0032] In addition, the protective gas blowing flow rates on the first and second base material sides are simultaneously and independently controlled to achieve asymmetric thermal management of the interface. The method for controlling the protective gas blowing flow rate is as follows: Figure 3 As shown, a first protective gas nozzle and a second protective gas nozzle are respectively installed on the two sides of the base material. When the critical high temperature zone width of one side of the base material in the infrared temperature field distribution data exceeds the safety threshold, the blowing flow rate of the corresponding protective gas nozzle on that side of the base material is automatically increased. By strengthening convective heat transfer, the cooling rate of that side is accelerated, and the excessive growth of intermetallic compounds is suppressed.

[0033] The method for obtaining the critical high-temperature zone width is as follows: A critical temperature threshold for the softening risk of the base material is preset; the acquired infrared temperature field distribution data is segmented by a threshold, and pixels with temperatures greater than or equal to the critical temperature threshold are extracted to fit and form an isothermal contour; subsequently, the maximum pixel distance from the isothermal contour to the weld center is calculated, which is the critical high-temperature zone width, as illustrated in the diagram. Figure 4 As shown. It should be noted that the method for obtaining the width of the critical high-temperature zone is existing technology and will not be elaborated here.

[0034] S402, Temporal-Domain Electromechanical Coordination: To prevent large drops of liquid metal from falling into the molten pool and disrupting the thermal balance under low heat input conditions, the pulse signal generator of the laser and the servo motor of the wire feeder were phase-bound at the underlying electrical signal level.

[0035] The operating logic is as follows: When the laser pulse is in its peak power range (intense heating period), the wire feeder accelerates forward, and the tip of the welding wire rapidly melts under the intense light to form a droplet; at the microsecond instant when the laser pulse switches to the base power range (weak cooling period), the wire feeder servo motor rapidly reverses, performing a mechanical retraction with a displacement of approximately 0.5mm-1.5mm. The backward physical pulling force forcibly severs the droplet neck, causing the droplet to detach. This mechanism achieves highly stable single-pulse, single-droplet transition, effectively suppressing additional heat input into the molten pool due to excessively large droplet volume.

[0036] S5. Closed-loop optimization: The core is a reinforcement learning decision model (such as the Proximal Policy Optimization (PPO) algorithm model). In the actual welding process, the data obtained from steps S1 to S4 above are input into this model. The model will continuously output the optimal combination of process parameters that can balance heat input and heat dissipation at the next moment with a millisecond refresh rate, thus completing dynamic closed-loop optimization until the welding task is completed.

[0037] The process of constructing and training a reinforcement learning decision model includes: Define the state space: construct a state vector from the thermal conductivity of the base material, real-time monitoring data from multi-source sensors, and current process parameters; Define the motion space: construct motion vectors from laser power, arc current, wire feed / retraction speed, lateral offset, and the flow rate of protective gas on both sides; Define a reward function that includes physical constraints: the reward function takes the achievement of the penetration state as a positive incentive and the high thermal conductivity side base material temperature exceeding the collapse threshold or generating microcracks as a negative penalty; Model training: The near-end policy optimization algorithm is used to continuously iterate the interaction between state and action in a virtual welding environment or historical experimental data. By maximizing the cumulative reward function, the reinforcement learning decision model is converged.

[0038] Another point to note is that in heterogeneous material composite welding, the high thermal conductivity base material often undergoes severe vaporization before macroscopic molten pool collapse when facing thermal overload. As a crucial cross-validation mechanism for the asymmetric thermal management of this invention, the arc plasma spectral data acquired in step S2 is used to monitor the vaporization state of the metal base material. A fiber optic spectrometer extracts the intensity of characteristic element spectral lines corresponding to the high thermal conductivity base material from the arc plasma spectral data; the relative concentration of the characteristic metal vapor in the arc is calculated based on the intensity of the characteristic element spectral lines; when a sudden change in the relative concentration of the characteristic metal vapor is detected and exceeds a preset vaporization burn-off threshold, it is determined that excessive vaporization has occurred in the high thermal conductivity base material, and this excessive vaporization state is input as a high-priority constraint into the reinforcement learning decision model to forcibly reduce the laser peak power. This invention, through dual sensing of optical composition analysis and thermodynamic surface temperature measurement, can significantly reduce thermal damage at heterogeneous interfaces.

[0039] Example 2 like Figure 2 As shown, this embodiment provides a laser-arc hybrid welding system with asymmetric thermal management to implement the welding method described in Embodiment 1, including a parameter acquisition module, a multi-source sensor real-time monitoring module, a dynamic collaborative heat control module, and a collaborative optimization decision engine.

[0040] The parameter acquisition module includes a human-computer interaction interface and a materials physics database, used to input and analyze the thermal conductivity of the first parent material. k A Thermal conductivity of the second base material k B .

[0041] The multi-source sensing real-time monitoring module includes an industrial camera, a fiber optic spectrometer, and an infrared thermal imager, which are mounted coaxially or off-axis, respectively, to acquire visual image data of the molten pool area, arc plasma spectral data, and infrared temperature field distribution data; the multi-source sensing real-time monitoring module also includes an acoustic emission sensor, which is used to collect acoustic signal characteristics during the welding process for microcrack early warning.

[0042] The dynamic collaborative heat control module includes an optical galvanometer system or a precision micro-motion slide at the end (driven by a high-frequency servo motor) for controlling the lateral offset and positioning of the laser beam, and dual-channel gas proportional valves that are independently connected to the protective gas nozzles on both sides to control the blowing flow rate.

[0043] The collaborative optimization decision engine, built into the industrial control computer, is based on a pre-trained reinforcement learning decision model. As the brain of the welding system, it receives thermophysical parameters from the parameter acquisition module, multi-dimensional sensing data from the multi-source real-time monitoring module, and the current operating process parameters of the welding system. This engine not only sends spatial offset and gas blowing commands to the dynamic collaborative heat control module, but also synchronously sends real-time control commands to the external laser control motherboard, arc welding machine power supply, and wire feeder servo driver via the industrial real-time Ethernet bus, thereby achieving closed-loop optimization across the entire heat input, spatial allocation, and heat dissipation chain.

[0044] The above embodiments are only used to illustrate the technical ideas and features of the present invention, and are not intended to be unique or to limit the present invention. Those skilled in the art should understand that various changes or equivalent substitutions made to the present invention without departing from its scope are all within the scope of protection of the present invention.

Claims

1. A laser-arc hybrid welding method with asymmetric thermal management, characterized in that, include: S1. Obtain the thermal conductivity of the first and second base materials to be welded. k A and k B And calculate the thermal conductivity ratio. R = k A / k B ; S2. During the welding process, simultaneously acquire visual image data of the molten pool area, arc plasma spectral data, and infrared temperature field distribution data. S3, Asymmetric Heat Distribution: Based on Thermal Conductivity Ratio R In addition to infrared temperature field distribution data, the lateral offset of the laser beam is calculated and controlled. d This causes the center of the laser beam spot to shift towards the base material on the side with lower thermal conductivity; S4, Dynamic Coordinated Heat Control: S401. Based on the melting state extracted from visual image data, the energy ratio of laser power and arc current is dynamically adjusted, and the protective gas blowing flow rate on the first and second base material sides is simultaneously and independently controlled to achieve asymmetric thermal management of the interface. S5. Closed-loop optimization: Input the data obtained from steps S1 to S4 into the preset reinforcement learning decision model, and continuously output the process parameter combination for the next moment until the welding task is completed; the process parameters include lateral offset, laser power, arc current, and shielding gas blowing flow rate.

2. The laser-arc hybrid welding method with asymmetric thermal management according to claim 1, characterized in that, Step S2 also includes: real-time acquisition of acoustic signal characteristics during the welding process through acoustic emission sensors; when specific acoustic signal characteristics representing the generation of microcracks are identified, an emergency process intervention command is triggered to reduce the overall heat input, including reducing laser power, reducing arc current, and increasing shielding gas blowing flow rate.

3. The laser-arc hybrid welding method with asymmetric thermal management according to claim 1, characterized in that, In step S3, the lateral offset d The calculation formula is: d=f ( R )+△ d ( T ) in, f ( R (Based on thermal conductivity ratio) R The initial offset reference value, △ d ( T This is a dynamic compensation amount based on the actual temperature difference between the parent material and the target temperature difference in the infrared temperature field distribution data. When the temperature of the parent material on the high thermal conductivity side approaches the collapse threshold, the offset to the parent material on the low thermal conductivity side is increased.

4. The laser-arc hybrid welding method with asymmetric thermal management according to claim 1, characterized in that, In step S4, the method for controlling the flow rate of the protective gas is as follows: a first protective gas nozzle and a second protective gas nozzle are respectively set at the base material on both sides; when the critical high temperature zone width of one side of the base material in the infrared temperature field distribution data exceeds the safety threshold, the blowing flow rate of the protective gas nozzle corresponding to that side of the base material is automatically increased.

5. The laser-arc hybrid welding method with asymmetric thermal management according to claim 1, characterized in that, In step S4, when the penetration state is incomplete, the laser power is increased while maintaining or decreasing the arc current; when the penetration state is over-penetrated, the laser power is decreased while increasing the arc current.

6. The laser-arc hybrid welding method with asymmetric thermal management according to claim 1, characterized in that, Step S4 also includes: S402, performing low-level phase binding between the laser pulse signal generator and the servo motor of the wire feeder; controlling the wire feeder to accelerate forward feeding when the laser pulse is in the peak power range; controlling the wire feeder to perform mechanical retraction at the instant the laser pulse switches to the base power range, forcing the molten droplet to fall off through mechanical pulling force.

7. A laser-arc hybrid welding system with asymmetric thermal management, characterized in that, The welding method for implementing any one of claims 1 to 6 is characterized by comprising: The parameter acquisition module is used to input and retrieve the thermal conductivity of the first and second base materials. k A and k B ; The multi-source sensing real-time monitoring module includes an industrial camera, a fiber optic spectrometer, and an infrared thermal imager mounted coaxially or off-axis, which are used to acquire visual image data of the molten pool area, arc plasma spectral data, and infrared temperature field distribution data, respectively. The dynamic collaborative heat control module includes an optical galvanometer system or a precision micro-motion slide at the end of the laser beam to control the lateral offset and positioning, and a dual-channel gas proportional valve to control the blowing flow of protective gas on both sides respectively. The collaborative optimization decision engine, built into the industrial control computer, is based on a pre-trained reinforcement learning decision model. It receives data from the parameter acquisition module and the multi-source sensor real-time monitoring module, and issues real-time control commands to the dynamic collaborative heat control module, laser, arc welding machine and wire feeder.

Images