A traceless welding method and system for liquid-cooled CDU thin-wall stainless steel pipeline fusion depth adaptive control

Abstract

The present application belongs to the technical field of liquid cooling equipment pipeline welding, and relates to a traceless welding method and system for fusion depth self-adaptive control of liquid cooling CDU thin-wall stainless steel pipeline. The method comprises the following steps: obtaining pipeline parameters to be welded, joint type and adjacent valve or manifold block heat capacity information, and generating a welding target window; coaxially clamping the thin-wall stainless steel pipe and the joint, and setting a retractable shaped back protection core in the pipe; performing non-contact pre-scanning and segmentation on the girth seam; collecting molten pool images, thermal radiation, reflected light intensity, oxygen content in the pipe and back protection pressure during welding, calculating fusion depth estimation value and outer surface trace index; adjusting welding energy, walking speed, swing amplitude and back protection gas flow according to combined deviation linkage, and performing energy step attenuation in the arc collection section. The present application is suitable for high consistency welding of liquid cooling CDU thin-wall stainless steel pipeline.

Description

Technical Field

[0001] This invention belongs to the field of liquid cooling equipment pipeline welding technology, specifically relating to a non-marking welding method and system for adaptive control of penetration depth in liquid-cooled CDU thin-walled stainless steel pipelines. Background Technology

[0002] As the coolant distribution and heat exchange unit in a data center liquid cooling system, the liquid-cooled CDU typically contains components such as supply lines, return lines, valves, sensors, filters, manifold blocks, and quick connectors. To balance corrosion resistance, cleanliness, and compact assembly requirements, CDU piping often uses thin-walled 304, 304L, 316, or 316L stainless steel tubing, connected to joints, valve shorts, or manifold blocks via circumferential welding. Due to the thin wall thickness, short circumferential weld length, and significant differences in heat capacity between adjacent components, problems such as insufficient penetration, overheating, excessive surface reinforcement, excessive inner wall protrusion, or backside oxidation can easily occur during welding, thus affecting the reliability of the piping seal, the stability of the coolant flow, and the cleanliness of the inner wall.

[0003] Among existing welding technologies, there are laser welding methods for controlling weld penetration. For example, Chinese patent CN105643104B discloses a laser welding method for precisely controlling weld penetration. This method determines the target penetration depth and selects corresponding welding process parameters, and also involves misalignment control, gap control, and energy attenuation settings at the weld tail end. Additionally, there is online monitoring technology for the dynamic process of the weld pool, which collects image information of the weld pool, welding process information, sound signals, and temperature distribution, and uses a control system to control parameters such as welding speed, current, and voltage.

[0004] However, existing technologies are mostly geared towards general weld penetration control or online monitoring of the weld pool, typically focusing on controlling welding heat input, weld pool morphology, or process parameters. They do not adequately consider the sudden changes in circumferential heat capacity caused by adjacent metal components such as valves and manifold blocks in liquid-cooled CDU thin-walled stainless steel pipelines, nor do they incorporate the control of oxygen content inside the pipe, back protection pressure, inner wall protrusion limits, and outer surface marks into the same closed-loop regulation process. For liquid-cooled CDU pipelines, relying solely on fixed parameters or single weld pool monitoring can easily lead to weld quality differences at different circumferential positions, making it difficult to simultaneously meet the requirements of penetration, sealing, inner wall cleanliness, and outer surface consistency.

[0005] Therefore, there is a need for a welding method and system with adaptive penetration control for thin-walled stainless steel pipes in liquid-cooled CDUs, in order to improve weld formation consistency and pipe assembly quality. Summary of the Invention

[0006] The purpose of this invention is to provide a non-marking welding method and system for adaptive control of the penetration depth of liquid-cooled CDU thin-walled stainless steel pipelines, so as to simultaneously ensure stable penetration depth, controlled inner wall formation, low back oxidation and low marks on the outer surface during the circumferential welding process of thin-walled pipelines, thereby improving the sealing reliability and welding consistency of CDU pipelines.

[0007] To achieve the above-mentioned technical objectives, the present invention provides the following technical solution.

[0008] In a first aspect, the present invention provides a non-marking welding method for adaptively controlling the penetration depth of thin-walled stainless steel tubing in a liquid-cooled CDU, comprising the following steps:

[0009] S1. Obtain the pipe diameter, wall thickness, joint type, circumferential position of the weld, and heat capacity information of adjacent valves or manifold blocks of the pipeline to be welded, and generate a welding target window that includes the target penetration ratio, outer surface reinforcement limit, inner wall protrusion limit, inner wall oxidation limit, and back protection pressure range.

[0010] S2. The thin-walled stainless steel tube and the joint are coaxially clamped. A retractable forming back protection core is set inside the tube so that the circumferential micro-hole gas outlet of the forming back protection core corresponds to the circumferential seam to be welded. The back of the weld is protected by a circumferential inert gas thin layer. The molten metal protrusion on the inner wall is limited by the forming limiting surface set at intervals with the inner wall of the tube. At the same time, the oxygen content and back protection pressure inside the tube are detected.

[0011] S3. Perform non-contact pre-scanning along the weld to obtain the weld gap, misalignment, local wall thickness and the location of sudden changes in joint heat capacity, and divide the circumferential weld into the arc initiation segment, steady state segment, heat capacity sudden change segment and arc termination and overlap segment.

[0012] S4. Generate initial welding parameters according to each segment and start welding. Real-time acquisition of molten pool image, welding heat radiation, reflected light intensity, oxygen content in pipe and back protection pressure. Calculate the current estimated molten depth and the outer surface trace index.

[0013] S5. Based on the combined deviation of the current estimated penetration depth, the outer surface trace index, the oxygen content inside the pipe, and the back shielding pressure relative to the welding target window, adjust the welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate in a coordinated manner.

[0014] S6. Perform energy step decay and short-stroke overlap welding in the arc-end overlap section, maintain back protection and follow-up cooling after welding, and output weld parameter records and sealing detection marks.

[0015] Furthermore, the thin-walled stainless steel pipe is 304, 304L, 316 or 316L stainless steel pipe, with a wall thickness of 0.4mm to 2.0mm, and the joint type includes one or more of the following: straight pipe butt joint, bent pipe butt joint, tee branch pipe joint, manifold block plug joint, valve short pipe joint or quick-connect transition joint.

[0016] Further, in step S1, the welding target window includes: a target penetration depth ratio of 70% to 95% of the pipe wall thickness, an outer surface excess height limit of no more than 0.15 mm, an inner wall protrusion limit of no more than 15% of the pipe wall thickness, an oxygen content in the pipe of no more than 50 ppm, and a back protection pressure of 0.5 kPa to 2.0 kPa; for the main supply and return liquid pipes of liquid-cooled CDUs with higher pressure ratings, the target penetration depth ratio is set to 80% to 100% of the pipe wall thickness.

[0017] Further, in step S2, the formed back protective core includes a flexible heat-resistant sealing ring, a circumferential microporous venting band, a forming limiting surface, and a retraction traction component. A forming gap of 0.1 mm to 0.5 mm is formed between the forming limiting surface and the inner wall of the pipe. The circumferential microporous venting band is used to form a circumferential inert gas thin layer on the back of the weld. The retraction traction component is used to remove the formed back protective core from the end of the pipe or the reserved interface after welding.

[0018] Further, in step S3, the non-contact pre-scanning uses a line laser profile sensor, a coaxial vision sensor, or a structured light sensor to obtain the weld centerline, circumferential gap sequence, misalignment sequence, and local curvature sequence; when the misalignment of any segment is greater than 0.08 mm or the gap is greater than 0.10 mm, the segment is marked as a compensation segment, and in step S4, the initial walking speed is reduced or the swing amplitude is increased.

[0019] Furthermore, in step S4, the current weld depth estimate is calculated based on the weld pool width, weld pool tail length, peak thermal radiation, attenuation of reflected light intensity, and welding energy per unit length; the external surface trace index is calculated based on the weld reinforcement height, weld width change rate, surface discoloration level, and number of spatter spots.

[0020] Furthermore, the current estimated melting depth satisfy:

[0021]

[0022] In the formula, This is the current estimated melting depth. The width of the molten pool; The length of the molten pool tail; This represents the peak value of thermal radiation. This represents the attenuation of reflected light intensity. The welding energy per unit length; This is a heat capacity correction factor determined by the joint type, local wall thickness, and heat capacity of adjacent metals; function The depth of penetration was determined by calibrating the cross-sectional penetration data of the test weld sample.

[0023] Furthermore, in step S5, the coordinated adjustment of welding energy, travel speed, oscillation amplitude, and back shielding gas flow includes:

[0024] When the current estimated penetration depth is lower than the lower limit of the welding target window and the external surface trace index does not exceed the limit, increase the welding energy or decrease the walking speed.

[0025] When the current estimated penetration depth is higher than the upper limit of the welding target window, or the outer surface trace index exceeds the limit, reduce the welding energy, increase the walking speed, or increase the oscillation amplitude.

[0026] When the oxygen content inside the pipe increases or the back shielding pressure fluctuates beyond the preset range, increase the back shielding gas flow rate and reduce the adjustment range of welding energy.

[0027] When the heat capacity abrupt change segment enters the welding area, the welding energy and travel speed are adjusted in advance according to the heat capacity correction coefficient K_h.

[0028] In a second aspect, the present invention also provides a non-marking welding system for adaptively controlling the penetration depth of liquid-cooled CDU thin-walled stainless steel tubing, used in the non-marking welding method for adaptively controlling the penetration depth of liquid-cooled CDU thin-walled stainless steel tubing described in the first aspect, the system comprising:

[0029] The pipe clamping and rotation module is used to coaxially clamp thin-walled stainless steel pipes and fittings, and drive the pipes or welding heads to move relative to each other along the circumferential seam.

[0030] The molded back protection module includes a retractable molded back protection core, a back protection air supply unit, an oxygen content sensor, and a back protection pressure sensor.

[0031] The pre-welding scanning module is used to obtain weld gap, misalignment, local wall thickness, local curvature, and locations of sudden changes in joint heat capacity;

[0032] The welding execution module is used to perform one of the following: laser welding, micro-beam plasma welding, automatic TIG welding, or laser-arc hybrid welding.

[0033] The online sensing module is used to collect images of the molten pool, welding thermal radiation, reflected light intensity, oxygen content inside the tube, and back protection pressure.

[0034] The adaptive control module is used to generate the welding target window, calculate the current penetration depth estimate and the outer surface mark index, and adjust the welding energy, travel speed, oscillation amplitude and back shielding gas flow rate.

[0035] The follow-up cooling module is used to cool the area adjacent to the weld after welding;

[0036] The quality recording module is used to record welding parameters, segmented adjustment process, back protection status, and sealing inspection marks.

[0037] Furthermore, the adaptive control module includes a segmented modeling unit, a melt depth estimation unit, a flawless evaluation unit, and a closed-loop adjustment unit;

[0038] The segmented modeling unit is used to divide the circumferential weld into an arc initiation segment, a steady-state segment, a heat capacity mutation segment, and an arc termination and overlap segment based on the pre-welding scanning results.

[0039] The penetration depth estimation unit is used to calculate the current penetration depth estimate based on the molten pool image, welding thermal radiation, reflected light intensity, and welding energy per unit length.

[0040] The tracelessness evaluation unit is used to calculate the trace index on the outer surface;

[0041] The closed-loop adjustment unit is used to adjust the welding energy, travel speed, oscillation amplitude and back shielding gas flow rate in conjunction with the combined deviation of the current estimated penetration depth, external surface trace index, tube oxygen content and back shielding pressure.

[0042] This invention, by acquiring information on pipe wall thickness, joint type, circumferential weld position, and heat capacity of adjacent valves or manifold blocks, can identify heat dissipation differences in different sections of the circumferential weld before welding. Based on this, it can divide the weld into arc-starting, steady-state, heat capacity abrupt change, and arc-ending overlap sections, allowing for targeted welding energy and travel speed for different sections. By setting a retractable forming back protective core inside the pipe, the circumferential microporous venting zone 5 corresponds to the circumferential weld 2 to be welded, forming a thin circumferential inert gas layer on the back of the weld. Simultaneously, the forming limiting surface limits the molten metal protrusions on the inner wall, which helps reduce the risk of inner wall oxidation and excessive inner wall protrusions. Furthermore, by acquiring molten pool images online... The system measures welding heat radiation, reflected light intensity, oxygen content inside the pipe, and back shielding pressure, and calculates the current estimated weld penetration and external surface trace index. It can adjust welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate in conjunction with the combined deviations of weld penetration, appearance, back shielding status, and oxygen content to keep weld penetration, external surface reinforcement, internal wall protrusion, and internal oxidation state within the target range. Through energy step decay in the arc-ending and folding section, short-stroke folding, and post-weld back shielding and follow-up cooling, it can reduce the probability of incomplete fusion, overheating, oxidation, and surface discoloration at the arc-ending point. It is suitable for batch and traceable welding of liquid-cooled CDU thin-walled stainless steel pipelines. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the non-marking welding system of the present invention.

[0044] Figure 2This is a partial cross-sectional schematic diagram of the shaped back protective core and the thin-walled stainless steel tube circumferentially joined in the present invention.

[0045] Figure 3 This is a schematic diagram of the traceless welding method of the present invention.

[0046] Reference numerals: 1. Inner wall of the pipe; 2. Circumferential seam to be welded; 3. Formed back protective core; 4. Formed limiting surface; 5. Circumferential microporous venting zone; 6. Inert gas; 7. Flexible temperature-resistant sealing ring; 8. Retraction component. Detailed Implementation

[0047] The embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the following embodiments are for illustrating the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Without departing from the technical principles of the present invention, those skilled in the art can make adaptive adjustments to the specific parameters according to the specifications of the liquid-cooled CDU piping, the type of connector, the type of welding equipment, and the on-site assembly conditions.

[0048] The liquid-cooled CDU referred to in this invention is a coolant distribution and heat exchange unit used in liquid cooling systems for data centers, server racks, or high heat flux density electronic equipment. Its internal components typically include supply lines, return lines, pump sets, heat exchangers, filters, valves, sensors, manifolds, and quick-connect fittings. Liquid-cooled CDUs typically use thin-walled stainless steel tubing made of 304, 304L, 316, or 316L stainless steel. The tubing wall thickness is usually small, and the tubing is arranged adjacent to valves, manifolds, quick-connect fittings, and other components. Due to the thin tubing wall, short weld circumferential length, and significant differences in local heat dissipation conditions, if welding is performed using only fixed process parameters, insufficient weld penetration may occur on the side near the valves or manifolds, while excessive surface reinforcement, excessive inner wall protrusion, or inner wall oxidation and discoloration may occur on the side farther from high heat capacity components. These issues can affect the sealing reliability of the tubing, the cleanliness of the inner wall, and the flow stability of the liquid coolant.

[0049] The term "seamless welding" in this invention does not mean that there are no visible weld seams after welding. Rather, it means that, under the premise of meeting the requirements for pipeline sealing and penetration depth, the weld seam reinforcement, weld seam width variation, surface discoloration, and spatter marks are controlled within a preset range, while limiting inner wall protrusion and inner wall oxidation, so that the welded pipeline can meet the requirements of liquid-cooled CDU for appearance consistency, cleanliness, and pressure sealing.

[0050] I. General Description of Implementation Methods

[0051] 1.1, Seamless Welding System

[0052] like Figure 1As shown, the non-marking welding system for liquid-cooled CDU thin-walled stainless steel pipes with adaptive penetration control provided in this embodiment includes a pipe clamping and rotation module, a forming back protection module, a pre-welding scanning module, a welding execution module, an online sensing module, an adaptive control module, a follow-up cooling module, and a quality recording module.

[0053] The pipe clamping and rotation module is used to coaxially clamp the thin-walled stainless steel pipe and the fitting, forming a weldable circumferential seam 2 between the end of the thin-walled stainless steel pipe and the fitting. The pipe clamping and rotation module can employ a three-jaw centering clamp, a double-end support clamp, a fine-adjustable V-shaped support, or a dedicated clamp with angle positioning. When the dimensions of the pipe and fitting to be welded are small, the welding execution module can be fixed and the thin-walled stainless steel pipe and fitting can be driven to rotate together. When the pipe to be welded is connected to valves, manifold blocks, or other components that are inconvenient to rotate, the pipe can be fixed and the welding head can move circumferentially along the circumferential seam. Preferably, the pipe clamping and rotation module is equipped with a fine-adjustment mechanism to adjust the coaxiality between the thin-walled stainless steel pipe and the fitting, reducing the pre-welding misalignment to a preset range.

[0054] The forming back protection module is used to provide inert gas protection to the back of the weld during the welding process and to limit the protrusion of molten metal on the inner wall. The forming back protection module includes a retractable forming back protection core 3, a back protection gas supply unit, an oxygen content sensor, and a back protection pressure sensor. The forming back protection core 3 is located inside the thin-walled stainless steel tube, and its position corresponds to the circumferential weld 2 to be welded.

[0055] 1.2 Molded back protective core

[0056] like Figure 2 As shown, the formed back protective core 3 includes a flexible high-temperature resistant sealing ring 7, a circumferential microporous venting band 5, a forming limiting surface 4, and a retraction traction component 8. The flexible high-temperature resistant sealing ring 7 can be made of high-temperature resistant silicone rubber, fluororubber, flexible graphite composite ring, or heat-resistant polyimide composite ring, and is used to form a local closed space on both sides of the formed back protective core 3 to prevent the protective gas from directly and rapidly leaking out along the cavity. The circumferential microporous venting band 5 is arranged circumferentially along the formed back protective core 3. The circumferential microporous venting band 5 is an annular array of micropores arranged circumferentially along the outer periphery of the formed back protective core 3. The annular micropore array corresponds axially to the weld seam 2. The venting direction of the micropores is towards the inner wall 1 of the tube or the back of the weld seam 2, and is inclined at 0° to 45° relative to the radial direction of the formed back protective core 3. The micropore diameter can be 0.05mm to 0.30mm, and the center distance between adjacent micropores can be 0.5mm to 2.0mm. The back protection gas supply unit introduces argon or an argon-helium mixture into the formed back protection core 3, causing the inert gas 6 to flow out uniformly from the circumferential microporous outlet zone 5, forming a thin circumferential inert gas layer on the back side of the weld. This thin circumferential inert gas layer covers the back side area of ​​the circumferential weld to reduce the risk of inner wall oxidation.

[0057] 1.3 Forming and limiting surface

[0058] The forming limiting surface 4 is located on the outer periphery or adjacent to the circumferential microporous venting zone 5, and a forming gap is maintained between the forming limiting surface 4 and the inner wall 1 of the pipe. The forming gap can be 0.1 mm to 0.5 mm, preferably 0.15 mm to 0.30 mm. During welding, if the molten metal on the back of the weld protrudes into the pipe, the forming limiting surface 4 can limit the protrusion height, keeping the inner wall protrusion within a preset range. The forming limiting surface 4 can be a polished cermet surface, a silicon nitride ceramic surface, or a stainless steel surface coated with a heat-resistant and non-stick coating to reduce the risk of molten metal adhesion. The retraction traction component 8 can be a durable and flexible cable, a thin rod, or a traction rod with positioning scales, used to remove the formed back protective core 3 from the pipe end or reserved interface after welding is completed and cooled.

[0059] 1.4 Oxygen content sensor

[0060] An oxygen content sensor is used to detect the oxygen content inside the pipe, and a back protection pressure sensor is used to detect the back protection pressure within the local protective space where the formed back protection core 3 is located. Both the oxygen content inside the pipe and the back protection pressure are input into the adaptive control module for subsequent combined deviation judgment. The back protection gas supply unit may include an inert gas source, a pressure regulating valve, a mass flow controller, a filter, and a solenoid valve. The back protection gas flow rate can be adjusted according to the welding section, pipe diameter, oxygen content inside the pipe, and back protection pressure.

[0061] 1.5 Pre-welding scanning

[0062] The pre-weld scanning module is used for non-contact pre-scanning along the circumferential weld. This module can employ a line laser profile sensor, a coaxial vision sensor, or a structured light sensor. Before welding, the pre-weld scanning module acquires information on the weld centerline, circumferential gap sequence, misalignment sequence, local curvature sequence, and local wall thickness. Based on the geometric relationships of the thin-walled stainless steel pipe, joint, adjacent valve, or manifold block, it determines the location of abrupt changes in joint heat capacity. For weld areas adjacent to manifold blocks, valves, or quick-connect fittings, the large volume of adjacent metal and high heat conduction rate can easily lead to differences in actual weld depth at different circumferential locations within the same circumferential weld. Therefore, the pre-weld scanning module is not only used to align the weld position but also to provide fundamental data for segmented control of the circumferential weld.

[0063] 1.6 Welding Execution

[0064] The welding execution module is used to perform one of the following: laser welding, micro-plasma welding, automatic TIG welding, or laser-arc hybrid welding. For liquid-cooled CDU thin-walled stainless steel pipes with a wall thickness of 0.4 mm to 1.2 mm, low-power continuous fiber laser welding or pulsed laser welding is preferred; for pipes with a wall thickness of 1.0 mm to 2.0 mm, micro-plasma welding, automatic TIG welding, or laser-arc hybrid welding can be used. The welding execution module can adjust the welding energy, travel speed, and oscillation amplitude. Welding energy can be expressed by laser power, arc current, pulse duty cycle, or combined heat source power; travel speed refers to the speed at which the welding heat source moves relative to the circumferential weld; oscillation amplitude refers to the lateral oscillation amplitude of the welding heat source relative to the weld centerline.

[0065] 1.7 Online Sensing

[0066] The online sensing module is used to acquire images of the molten pool, welding thermal radiation, reflected light intensity, oxygen content inside the pipe, and back shielding pressure in real time during the welding process. The online sensing module can include a molten pool image acquisition unit, a thermal radiation acquisition unit, and a reflected light intensity acquisition unit. The molten pool image acquisition unit can use an industrial camera with a narrowband filter to acquire the molten pool width, molten pool tail length, molten pool boundary, and changes in molten pool morphology. The thermal radiation acquisition unit can use an infrared temperature probe or an infrared thermal imaging unit to acquire the peak value of welding thermal radiation and the temperature rise in the vicinity of the weld. The reflected light intensity acquisition unit can use a coaxial photodetector to acquire changes in reflected light intensity during the welding process. The oxygen content inside the pipe is acquired by an oxygen content sensor, and the back shielding pressure is acquired by a back shielding pressure sensor, and these are recorded synchronously with the molten pool image, welding thermal radiation, and reflected light intensity.

[0067] 1.8 Adaptive Control

[0068] The adaptive control module generates the welding target window, calculates the current penetration depth estimate and the external surface mark index, and adjusts the welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate in conjunction with the combined deviation. The adaptive control module may include a segmented modeling unit, a penetration depth estimation unit, a mark-free evaluation unit, and a closed-loop adjustment unit.

[0069] 1.9 Segmented Modeling

[0070] The segmented modeling unit divides the circumferential weld into four segments based on the weld gap, misalignment, local wall thickness, location of abrupt changes in joint heat capacity, and circumferential weld position obtained from the pre-weld scanning module: the arc-initiating segment, the steady-state segment, the heat capacity abrupt change segment, and the arc-ending overlap segment. The arc-initiating segment refers to the section from the start of the welding heat source until the molten pool stabilizes; the steady-state segment refers to the section where the weld gap, misalignment, and adjacent heat capacity changes are relatively small; the heat capacity abrupt change segment refers to the section near valves, manifold blocks, quick-connect fittings, or thick-walled joints; and the arc-ending overlap segment refers to the section where the circumferential weld end connects to the arc-initiating segment and undergoes a short-stroke overlap. This segmentation method allows for the application of different initial welding parameters and subsequent adjustment strategies for different circumferential positions.

[0071] 1.10. Melting Depth Estimation

[0072] The melt depth estimation unit is used to calculate the current melt depth estimate. Current melt depth estimate. Calculated based on weld pool width, weld pool tail length, peak thermal radiation, reflected light intensity attenuation, welding energy per unit length, and heat capacity correction factor. Current estimated weld depth. satisfy:

[0073]

[0074] In the formula, This is the current estimated melting depth. The width of the molten pool; The length of the molten pool tail; This represents the peak value of thermal radiation. This represents the attenuation of reflected light intensity. The welding energy per unit length; This is a heat capacity correction factor determined by the joint type, local wall thickness, and heat capacity of adjacent metals; function The depth of penetration was determined by calibrating the cross-sectional penetration data of the test weld sample.

[0075] In this embodiment, the function The results can be obtained using polynomial fitting, piecewise linear fitting, or lookup table interpolation, without the need for a neural network model. Taking piecewise linear fitting as an example, test weld samples can be fabricated for different pipe diameters, wall thicknesses, and joint types. Test welds can be performed using different welding energies, travel speeds, and back protection conditions. After welding, cross-sections are cut at different positions along the circumferential weld to measure the actual weld penetration, outer surface reinforcement, and inner wall protrusion. The actual cross-sectional weld penetration is then compared with the corresponding... , , , , and Establish the correspondence to obtain the calibration table. During actual welding, interpolate the values ​​from the calibration table based on the online acquired data to obtain the current estimated penetration depth. For the same liquid-cooled CDU product series, calibration tables can be stored separately by pipe diameter, wall thickness, and connector type.

[0076] Welding energy per unit length It can be calculated using the following formula:

[0077]

[0078] In the formula, The welding energy per unit length; Output power for welding heat source; This refers to walking speed. When an electric arc heat source is used, It can be calculated from welding current, arc voltage, and thermal efficiency; when laser-arc hybrid welding is used, It can be the sum of laser power and effective arc power.

[0079] Heat capacity correction factor The determination can be made based on the joint type, local wall thickness, and heat capacity of adjacent metals. For ordinary straight pipe butt joints, A value of 1.00 can be used; for areas near valve short pipe fittings, A value of 1.05 to 1.20 can be used; for areas near the manifold block connector, You can take values ​​between 1.15 and 1.35. The specific value can be corrected using the test weld section data. A larger value indicates that the section has a strong heat dissipation capacity, and the actual penetration depth may be smaller under the same welding energy. It is necessary to increase the welding energy or reduce the walking speed in advance.

[0080] 1.11 Seamless Evaluation

[0081] The tracelessness evaluation unit is used to calculate the external surface trace index. The external surface trace index can be calculated based on weld reinforcement, weld width variation rate, surface discoloration level, and the number of spatter spots. For example, the external surface trace index... satisfy:

[0082]

[0083] In the formula, The outer surface trace index; For weld reinforcement; This is the limit value for the excess height of the outer surface; This represents the change in weld width. This refers to the limit value for weld width variation; The surface discoloration level; To allow for color-changing levels; The number of splashing bright spots per unit length; To allow the number of splash highlights; , , , These are the weighting coefficients, and .when When the value is not greater than 1, it indicates that the marks on the outer surface are within the set range; when A value greater than 1 indicates that it is necessary to reduce heat input, increase walking speed, or adjust the swing amplitude.

[0084] 1.12 Closed-loop regulation

[0085] The closed-loop control unit is used to adjust the current melt depth estimate. Surface Trace Index The combined deviation of the oxygen content inside the pipe and the back shielding pressure relative to the welding target window is used to adjust the welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate. This combined deviation is not a single parameter deviation, but rather a state deviation that simultaneously considers the penetration depth, appearance, inner wall oxidation, and back shielding stability. For example, when the current penetration depth estimate is... Below the lower limit of the welding target window, and the outer surface trace index When the limit is not exceeded, the closed-loop control unit increases the welding energy or decreases the travel speed; when the current penetration depth estimate is... Exceeding the upper limit of the welding target window, or the outer surface trace index When the limit is exceeded, the closed-loop control unit reduces the welding energy, increases the travel speed, or increases the oscillation amplitude; when the oxygen content inside the pipe increases or the back shielding pressure fluctuates beyond the preset range, the closed-loop control unit increases the back shielding gas flow rate and reduces the adjustment amplitude of the welding energy to avoid inner wall oxidation due to unstable back shielding; when the heat capacity abrupt change section enters the welding zone, the closed-loop control unit adjusts according to the heat capacity correction coefficient. Adjust welding energy and walking speed in advance.

[0086] The follow-up cooling module is used to cool the area adjacent to the weld after welding. The follow-up cooling module can employ controlled micro-volume inert gas cooling, indirect compressed air cooling, contact heat conduction with a cooling copper block, or heat conduction through a water-cooled fixture. For liquid-cooled CDU thin-walled stainless steel piping, inert gas cooling or water-cooled fixture heat conduction, which do not directly contaminate the weld, is preferred. The follow-up cooling module is typically positioned at a certain angle behind the welding heat source, for example, 15° to 60°, to avoid excessively rapid cooling affecting weld formation.

[0087] 1.13 Quality Records

[0088] The quality recording module records welding parameters, segmented adjustment processes, back shielding status, and sealing inspection marks. It can bind and store the corresponding pipe number, joint type, welding target window, welding energy curve, travel speed curve, oscillation amplitude curve, back shielding gas flow rate curve, pipe oxygen content curve, back shielding pressure curve, current penetration depth estimate curve, and external surface trace index curve for each weld. After post-weld airtightness testing or water pressure maintenance testing, the test results are written to the sealing inspection mark. This record can be used for quality traceability of liquid-cooled CDU piping welding.

[0089] II. Specific Implementation Steps of Welding Method

[0090] like Figure 3 As shown, the welding method of this embodiment includes the following steps.

[0091] S1. Obtain the pipe diameter, wall thickness, joint type, circumferential position of the weld seam, and heat capacity information of adjacent valves or manifold blocks of the pipeline to be welded, and generate the welding target window.

[0092] Specifically, operators or the control system read the product process documents to determine the material grade, outer diameter, wall thickness, joint type, and circumferential position of the weld seam for the thin-walled stainless steel pipe. Different welding target windows can be set for the main supply and return pipes, branch pipes, valve short pipes, and manifold plug pipes of the liquid-cooled CDU. The welding target window includes at least the target penetration depth ratio, the outer surface reinforcement limit, the inner wall protrusion limit, and the inner wall oxidation limit. The inner wall oxidation limit can be represented by the oxygen content limit inside the pipe and / or the post-weld inner wall discoloration level limit. The target penetration depth ratio can be 70% to 95% of the pipe wall thickness; for the main supply and return pipes of the liquid-cooled CDU with higher pressure ratings, the target penetration depth ratio can be set to 80% to 100% of the pipe wall thickness. The outer surface reinforcement limit can be no greater than 0.15 mm, the inner wall protrusion limit can be no greater than 15% of the pipe wall thickness, and the oxygen content inside the pipe can be no greater than 50 ppm.

[0093] When generating the welding target window, the segmented modeling unit also reads the heat capacity information of adjacent valves or manifold blocks. The heat capacity of adjacent metals can be calculated based on the material, volume, wall thickness, distance from the weld, and contact area of ​​adjacent components, or it can be pre-defined according to the process documents. For the weld area near the manifold block, the control system marks this location as a heat capacity abrupt change segment, providing a basis for subsequent segmented adjustments.

[0094] S2. The thin-walled stainless steel tube and the connector are coaxially clamped together, and a retractable shaped back protection core is installed inside the tube.

[0095] Specifically, a thin-walled stainless steel tube is inserted or connected to the joint, and positioned using a pipe clamping and rotation module to form an annular seam between the thin-walled stainless steel tube and the joint. After positioning, the formed back protective core 3 is inserted into the tube from the end of the pipe or the reserved interface, and the circumferential microporous venting band 5 is aligned with the annular seam 2 to be welded. The position of the formed back protective core 3 is finely adjusted by the retraction traction component 8 to maintain a forming gap between the forming limiting surface 4 and the back area of ​​the annular seam.

[0096] Subsequently, inert gas 6 is introduced into the back protection gas supply unit, initially at a relatively high flow rate to displace the oxygen content in the pipe, reducing it to below the preset value. Once the oxygen content meets the requirements, the flow rate is switched to welding hold flow. The oxygen content sensor continuously monitors the oxygen content in the pipe, and the back protection pressure sensor continuously monitors the back protection pressure. If the oxygen content in the pipe does not reach the target value before welding, the system will not start the welding execution module; if the back protection pressure is below the lower limit, the system will prompt to check whether the flexible high-temperature resistant sealing ring 7 is properly sealed or whether there is a leak in the protective gas passage.

[0097] S3. Perform non-contact pre-scanning along the circumferential seam to obtain the weld gap, misalignment, local wall thickness and the location of sudden changes in joint heat capacity, and divide the circumferential seam into the arc initiation segment, steady state segment, heat capacity sudden change segment and arc termination overlap segment.

[0098] Specifically, the pre-welding scanning module scans around the circumferential seam to obtain the weld centerline, circumferential gap sequence, misalignment sequence, and local curvature sequence. For straight pipe butt joints, if the circumferential gap and misalignment change little, most of the circumferential seam is classified as a steady-state segment; for manifold block plug joints, the portion near the thick-walled area of ​​the manifold block is classified as a heat capacity abrupt change segment; for valve short pipe joints, the portion near the valve body is classified as a heat capacity abrupt change segment.

[0099] When the misalignment of any section exceeds 0.08 mm or the gap exceeds 0.10 mm, that section is marked as a compensation section. For compensation sections, the initial travel speed can be reduced, the oscillation amplitude increased, or the welding heat source position appropriately adjusted to ensure that the molten pool covers the weld gap. If the misalignment or gap exceeds the equipment's allowable range, the system can output an assembly adjustment prompt, and welding can be performed after repositioning.

[0100] S4. Generate initial welding parameters for each segment and start welding. Real-time acquisition of molten pool images, welding heat radiation, reflected light intensity, oxygen content inside the pipe, and back shielding pressure; calculate the current estimated weld depth. and external surface trace index .

[0101] Specifically, the adaptive control module generates initial welding parameters based on the welding target window and the circumferential weld segmentation results. The initial arc-starting phase uses a lower welding energy that is gradually increased to ensure a smooth weld pool establishment; the steady-state phase uses a reference welding energy and a reference travel speed; and the heat capacity abrupt change phase uses a heat capacity correction coefficient. Increase welding energy or reduce walking speed in advance; use energy step decay and short stroke back-overlap welding in the arc-end overlap section.

[0102] During the welding process, the molten pool image acquisition unit extracts the molten pool width from the molten pool image. and molten pool tail length The thermal radiation acquisition unit extracts the peak thermal radiation value. The reflected light intensity acquisition unit extracts the amount of reflected light intensity attenuation. The adaptive control module calculates the welding energy per unit length based on the welding heat source power and the travel speed. And combined with the heat capacity correction factor Obtain the current melting depth estimate Simultaneously, the tracelessness evaluation unit calculates the trace index on the outer surface. .

[0103] S5. Based on the current estimated melting depth Surface Trace Index The combined deviation of the oxygen content inside the pipe and the back shielding pressure relative to the welding target window is used to adjust the welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate.

[0104] Specifically, when the current melting depth estimate is... Below the lower limit of the welding target window, and at the same time, the external surface trace index If the limit is not exceeded, it indicates that the weld penetration is insufficient but there is still adjustment margin on the outer surface. In this case, the closed-loop adjustment unit can increase the welding energy or decrease the travel speed. Increasing the welding energy can be achieved by increasing the laser power, increasing the arc current, or increasing the pulse duty cycle; decreasing the travel speed can increase the welding energy per unit length.

[0105] When the current melting depth is estimated Exceeding the upper limit of the welding target window, or the outer surface trace index When the limit is exceeded, it indicates that the welding heat input is too high or the weld pool is too wide. In this case, the closed-loop control unit reduces the welding energy, increases the travel speed, or increases the oscillation amplitude. Increasing the oscillation amplitude can disperse the peak heat input within a certain range, reducing the surface reinforcement and local discoloration. The oscillation amplitude should not be too large to avoid excessive weld width or unstable fusion boundary.

[0106] When the oxygen content inside the tube increases or the back protection pressure fluctuates beyond the preset range, it indicates that the back protection state is unstable. At this time, the closed-loop adjustment unit increases the back protection gas flow rate and reduces the adjustment range of welding energy to avoid significant inner wall oxidation under insufficient back protection. When the back protection pressure remains below the lower limit, the system can pause welding or enter a low-energy holding state and prompt to check the sealing status between the formed back protection core 3 and the inner wall 1 of the tube.

[0107] When the heat capacity abrupt change segment enters the welding zone, the closed-loop control unit adjusts according to the heat capacity correction coefficient. Adjust welding energy and travel speed in advance. For example, in sections near the manifold block, where the manifold block has a strong heat absorption capacity, the welding energy can be increased by 5% to 15% or the travel speed reduced by 3% to 10% before the welding heat source reaches this section. After the welding heat source leaves the section with a sudden change in heat capacity, the parameters can be gradually restored to the steady-state parameters. By making these adjustments in advance, the risk of insufficient local penetration due to control lag can be reduced.

[0108] S6. Perform energy step decay and short-stroke overlap welding in the arc-end overlap section, maintain back protection and follow-up cooling after welding, and output weld parameter records and sealing detection marks.

[0109] Specifically, when the welding heat source enters the arc-end overlap section, the closed-loop control unit gradually reduces the welding energy according to a preset step. For example, it first reduces the energy to 90% of the steady-state energy, then to 75%, and finally to 60%, with each step held for 0.05s to 0.30s. The short-stroke overlap length can be 2% to 8% of the circumference of the circumference, or 1mm to 5mm. Through energy step decay and short-stroke overlap, the risks of arc-end craters, incomplete fusion joints, and localized overheating can be reduced.

[0110] After welding, the back protection gas supply unit continues to provide back protection for a period of time, such as 5 to 30 seconds, until the temperature on the back of the weld decreases to a temperature range where oxidation is less likely. The follow-up cooling module cools the area adjacent to the weld, controlling the discoloration of the weld surface and the extent of the heat-affected zone. After welding, the formed back protection core 3 is removed by the pull-back traction component 8. The quality recording module outputs the weld parameters for this weld and generates a sealing test mark after the airtightness test or water pressure maintenance test is completed.

[0111] III. Example 1

[0112] This embodiment focuses on the 316L thin-walled stainless steel pipe and manifold block plug-in connector in the supply and return liquid branch of a liquid-cooled CDU. The thin-walled stainless steel pipe has an outer diameter of 12.7 mm and a wall thickness of 0.8 mm. The connector is a 316L stainless steel manifold block plug-in connector with a plug-in depth of 2.5 mm. Continuous fiber laser welding is used, with a laser rated power of 500 W and a shielding gas of 99.999% argon. The welding target window is set as follows: target penetration depth is 80% to 95% of the wall thickness, i.e., target penetration depth is 0.64 mm to 0.76 mm; the outer surface excess height limit is no greater than 0.12 mm; the inner wall protrusion limit is no greater than 0.10 mm; the oxygen content inside the pipe is no greater than 50 ppm; and the back protection pressure is controlled between 0.8 kPa and 1.5 kPa.

[0113] Before welding, the formed back protective core 3 is inserted from the end of the pipe, aligning the circumferential microporous venting band 5 with the circumferential seam 2 to be welded. The forming gap between the forming limiting surface 4 and the inner wall 1 of the pipe is 0.20 mm. The micropore diameter of the circumferential microporous venting band 5 is 0.12 mm, and the center distance between adjacent micropores is 1.0 mm. Before welding, the gas in the pipe is first replaced with argon gas at a flow rate of 1.8 L / min. After the oxygen content in the pipe drops below 30 ppm, the back protective gas flow rate is adjusted to 0.8 L / min to 1.2 L / min, and the back protective pressure is maintained at 0.8 kPa to 1.5 kPa.

[0114] The pre-welding scanning module scans the circumferential seam once, obtaining a weld gap of 0.03mm to 0.07mm and a maximum misalignment of 0.05mm. Due to the larger heat capacity on one side of the manifold block, the segmented modeling unit divides approximately 80° of the circumferential region near the thick wall of the manifold block into a heat capacity abrupt change segment, the initial 20° segment into an arc initiation segment, the final 30° segment into an arc termination and overlap segment, and the remaining portion into a steady-state segment. The heat capacity correction coefficient for the heat capacity abrupt change segment is used. Take 1.18, for the steady-state segment. Take 1.00.

[0115] The initial welding parameters were set as follows: in the steady state, the laser power was 260W, the travel speed was 11mm / s, and the oscillation amplitude was 0.10mm; in the arc initiation stage, the laser power was gradually increased from 180W to 260W; in the heat capacity abrupt change stage, the laser power was increased to 285W, and the travel speed was reduced to 10mm / s; in the arc termination and overlap stage, a three-level stepped attenuation of 260W, 230W, and 190W was used, and the overlap welding was 2.0mm.

[0116] During welding, the online sensing module acquires real-time images of the molten pool, welding heat radiation, and reflected light intensity. The weld depth estimation unit calculates the current weld depth estimate based on a calibration function. Before the heat source enters the heat capacity abrupt change zone, the adaptive control module increases the laser power and reduces the walking speed in advance. If a sudden change in heat capacity is detected during the heat capacity abrupt change zone... The thickness decreased to 0.63 mm, below the lower limit of the target window, and the outer surface trace index... If the value is 0.72, which does not exceed the limit, the system will increase the laser power to 292W and reduce the walking speed to 9.8mm / s. If the local surface trace index in the steady-state section is... If the value increases to 1.05, the system reduces the laser power to 250W and increases the travel speed to 11.5mm / s. During the welding process, the oxygen content inside the tube is maintained between 25ppm and 42ppm, and the back protection pressure is maintained between 0.9kPa and 1.3kPa.

[0117] After welding, the back protection core 3 is kept in place for 15 seconds, and the area adjacent to the weld is cooled by the follow-up cooling module. After the weld temperature drops, the formed back protection core 3 is removed by the pull-back traction component 8. Subsequently, the weld is subjected to visual inspection, cross-sectional penetration depth inspection, inner wall formation observation, and airtightness inspection.

[0118] To verify the results, Comparative Example 1 and Comparative Example 2 were set up. Comparative Example 1 used conventional fixed laser parameter welding, without heat capacity segmentation correction, without setting the forming back protection core 3, and only performed ordinary tube argon purging protection. Comparative Example 2 used closed-loop adjustment of laser power based on the molten pool image, but did not collect the oxygen content in the tube and the back protection pressure, and did not set the forming limiting surface 4.

[0119] The testing method is as follows: 20 samples were welded in each group. After welding, the appearance of each sample was measured, and the maximum external surface height was recorded. The weld cross-section of 10 samples from each group was randomly cut, and the minimum penetration depth and the maximum internal wall protrusion were measured. A helium leak test was performed on all 20 samples in each group, with a leak test pressure of 0.6 MPa and a holding time of 10 min. The oxidation of the internal wall of all 20 samples in each group was observed by endoscopy, and recorded as no obvious oxidation, slight oxidation, and obvious oxidation.

[0120] The welding quality test results of the examples and comparative examples are shown in Table 1.

[0121] Table 1 Welding quality test results

[0122]

[0123] As shown in Table 1, the minimum melt depth of Example 1 was 0.66 mm, which was within the target window, and the melt depth fluctuation range was narrow; the maximum outer surface clearance was 0.09 mm, which did not exceed the limit of 0.12 mm; the maximum inner wall protrusion was 0.07 mm, which did not exceed the limit of 0.10 mm; no obvious inner wall oxidation was observed in any of the 20 samples, and all helium leak tests were qualified. Comparative Example 1 did not undergo segmented heat capacity correction, resulting in insufficient melt depth on the side closer to the manifold block, while the side farther from the manifold block showed excessive outer surface clearance and excessive inner wall protrusion. Although Comparative Example 2 adjusted the laser power based on the molten pool image, it did not incorporate the oxygen content inside the tube, back protection pressure, and forming limit into the control, and some samples still showed excessive inner wall oxidation and excessive inner wall protrusion.

[0124] In summary, this embodiment determines the circumferential weld segmentation and the location of abrupt changes in heat capacity through pre-welding non-contact scanning. A circumferential inert gas layer is formed inside the pipe by the forming back protective core 3, limiting inner wall protrusion. The molten pool, thermal radiation, reflected light intensity, oxygen content inside the pipe, and back protective pressure are acquired by an online sensing module. The welding energy, travel speed, oscillation amplitude, and back protective gas flow rate are adjusted in conjunction with these parameters by an adaptive control module. This enables the liquid-cooled CDU thin-walled stainless steel pipe to maintain a relatively stable weld depth under different circumferential heat dissipation conditions, and controls the outer surface reinforcement, inner wall protrusion, and inner wall oxidation. The sensors, controllers, laser welding equipment, fixtures, and gas path components used in the above implementation process can all be implemented using commonly used industrial components. Those skilled in the art can implement the technical solution of this invention according to the description in this specification.

Claims

1. A method for seamless welding of thin-walled stainless steel pipes for liquid-cooled CDUs with adaptive penetration control, characterized in that, Includes the following steps: S1. Obtain the pipe diameter, wall thickness, joint type, circumferential position of the weld, and heat capacity information of adjacent valves or manifold blocks of the pipeline to be welded, and generate a welding target window that includes the target penetration ratio, outer surface reinforcement limit, inner wall protrusion limit, inner wall oxidation limit, and back protection pressure range. S2. The thin-walled stainless steel tube and the joint are coaxially clamped. A retractable forming back protection core is set inside the tube so that the circumferential micro-hole gas outlet of the forming back protection core corresponds to the circumferential seam to be welded. The back of the weld is protected by a circumferential inert gas thin layer. The molten metal protrusion on the inner wall is limited by the forming limiting surface set at intervals with the inner wall of the tube. At the same time, the oxygen content and back protection pressure inside the tube are detected. S3. Perform non-contact pre-scanning along the weld to obtain the weld gap, misalignment, local wall thickness and the location of sudden changes in joint heat capacity, and divide the circumferential weld into the arc initiation segment, steady state segment, heat capacity sudden change segment and arc termination and overlap segment. S4. Generate initial welding parameters according to each segment and start welding. Real-time acquisition of molten pool image, welding heat radiation, reflected light intensity, oxygen content in pipe and back protection pressure. Calculate the current estimated molten depth and the outer surface trace index. S5. Based on the combined deviation of the current estimated penetration depth, the outer surface trace index, the oxygen content inside the pipe, and the back shielding pressure relative to the welding target window, adjust the welding energy, travel speed, oscillation amplitude, and back shielding gas flow rate in a coordinated manner. S6. Perform energy step decay and short-stroke overlap welding in the arc-end overlap section, maintain back protection and follow-up cooling after welding, and output weld parameter records and sealing detection marks.

2. The method for seamless welding of thin-walled stainless steel pipelines with adaptive penetration control for liquid-cooled CDUs according to claim 1, characterized in that, The thin-walled stainless steel pipe is 304, 304L, 316 or 316L stainless steel pipe with a wall thickness of 0.4mm to 2.0mm. The joint types include one or more of the following: straight pipe butt joint, bent pipe butt joint, tee branch pipe joint, manifold block plug joint, valve short pipe joint or quick-connect transition joint.

3. The seamless welding method for adaptive penetration control of thin-walled stainless steel pipelines in liquid-cooled CDUs according to claim 1, characterized in that, In step S1, the welding target window includes: a target penetration depth ratio of 70% to 95% of the pipe wall thickness, an outer surface excess height limit of no more than 0.15 mm, an inner wall protrusion limit of no more than 15% of the pipe wall thickness, an oxygen content in the pipe of no more than 50 ppm, and a back protection pressure of 0.5 kPa to 2.0 kPa; for the main supply and return liquid pipes of liquid-cooled CDUs with higher pressure ratings, the target penetration depth ratio is set to 80% to 100% of the pipe wall thickness.

4. The method for seamless welding of thin-walled stainless steel pipelines with adaptive penetration control for liquid-cooled CDUs according to claim 1, characterized in that, In step S2, the formed back protective core includes a flexible heat-resistant sealing ring, a circumferential microporous venting band, a forming limiting surface, and a retraction traction component. A forming gap of 0.1 mm to 0.5 mm is formed between the forming limiting surface and the inner wall of the pipe. The circumferential microporous venting band is used to form a circumferential inert gas thin layer on the back of the weld. The retraction traction component is used to remove the formed back protective core from the end of the pipe or the reserved interface after welding.

5. The non-marking welding method for adaptive penetration control of thin-walled stainless steel pipelines in liquid-cooled CDUs according to claim 1, characterized in that, In step S3, the non-contact pre-scanning uses a line laser profile sensor, a coaxial vision sensor, or a structured light sensor to obtain the weld centerline, circumferential gap sequence, misalignment sequence, and local curvature sequence. When the misalignment of any segment is greater than 0.08 mm or the gap is greater than 0.10 mm, the segment is marked as a compensation segment, and in step S4, the initial walking speed is reduced or the swing amplitude is increased.

6. The method for seamless welding of thin-walled stainless steel pipelines with adaptive penetration control for liquid-cooled CDUs according to claim 1, characterized in that, In step S4, the current weld depth estimate is calculated based on the weld pool width, weld pool tail length, peak thermal radiation, attenuation of reflected light intensity, and welding energy per unit length; the external surface trace index is calculated based on the weld reinforcement height, weld width change rate, surface discoloration level, and number of spatter spots.

7. The method for adaptive control of penetration depth in liquid-cooled CDU thin-walled stainless steel pipelines according to claim 6, characterized in that, Current estimated melting depth satisfy: ； In the formula, This is the current estimated melting depth. The width of the molten pool; The length of the molten pool tail; This represents the peak value of thermal radiation. This represents the attenuation of reflected light intensity. The welding energy per unit length; This is a heat capacity correction factor determined by the joint type, local wall thickness, and heat capacity of adjacent metals; function The depth of penetration was determined by calibrating the cross-sectional penetration data of the test weld sample.

8. The method for adaptive control of penetration depth in liquid-cooled CDU thin-walled stainless steel pipelines according to claim 1, characterized in that, In step S5, the coordinated adjustment of welding energy, travel speed, oscillation amplitude, and back shielding gas flow includes: When the current estimated penetration depth is lower than the lower limit of the welding target window and the external surface trace index does not exceed the limit, increase the welding energy or decrease the walking speed. When the current estimated penetration depth is higher than the upper limit of the welding target window, or the outer surface trace index exceeds the limit, reduce the welding energy, increase the walking speed, or increase the oscillation amplitude. When the oxygen content inside the pipe increases or the back shielding pressure fluctuates beyond the preset range, increase the back shielding gas flow rate and reduce the adjustment range of welding energy. When the heat capacity abrupt change segment enters the welding area, the welding energy and travel speed are adjusted in advance according to the heat capacity correction coefficient K_h.

9. A non-marking welding system for adaptive control of penetration depth in liquid-cooled CDU thin-walled stainless steel pipelines, characterized in that, A non-marking welding method for implementing adaptive penetration control of thin-walled stainless steel tubing in a liquid-cooled CDU according to any one of claims 1 to 8, the system comprising: The pipe clamping and rotation module is used to coaxially clamp thin-walled stainless steel pipes and fittings, and drive the pipes or welding heads to move relative to each other along the circumferential seam. The molded back protection module includes a retractable molded back protection core, a back protection air supply unit, an oxygen content sensor, and a back protection pressure sensor. The pre-welding scanning module is used to obtain weld gap, misalignment, local wall thickness, local curvature, and locations of sudden changes in joint heat capacity; The welding execution module is used to perform one of the following: laser welding, micro-beam plasma welding, automatic TIG welding, or laser-arc hybrid welding. The online sensing module is used to collect images of the molten pool, welding thermal radiation, reflected light intensity, oxygen content inside the tube, and back protection pressure. The adaptive control module is used to generate the welding target window, calculate the current penetration depth estimate and the outer surface mark index, and adjust the welding energy, travel speed, oscillation amplitude and back shielding gas flow rate. The follow-up cooling module is used to cool the area adjacent to the weld after welding; The quality recording module is used to record welding parameters, segmented adjustment process, back protection status, and sealing inspection marks.

10. The seamless welding system for adaptive penetration control of liquid-cooled CDU thin-walled stainless steel pipelines according to claim 9, characterized in that, The adaptive control module includes a segmented modeling unit, a melt depth estimation unit, a flawless evaluation unit, and a closed-loop adjustment unit; The segmented modeling unit is used to divide the circumferential weld into an arc initiation segment, a steady-state segment, a heat capacity mutation segment, and an arc termination and overlap segment based on the pre-welding scanning results. The penetration depth estimation unit is used to calculate the current penetration depth estimate based on the molten pool image, welding thermal radiation, reflected light intensity, and welding energy per unit length. The tracelessness evaluation unit is used to calculate the trace index on the outer surface; The closed-loop adjustment unit is used to adjust the welding energy, travel speed, oscillation amplitude and back shielding gas flow rate in conjunction with the combined deviation of the current estimated penetration depth, external surface trace index, tube oxygen content and back shielding pressure.

Images