Plasma arc welding method for hardware product machining

Abstract

The invention relates to the technical field of plasma arc welding, in particular to a plasma arc welding method for hardware product machining, which comprises the following steps: acquiring the normal and curvature change of a stainless steel thin-walled pipe joint, judging the structural state of a workpiece, acquiring the posture of a welding gun for an aluminum alloy section corner piece based on the state, and judging the focus offset and the incident angle. And the carbon steel reinforcing plate molten pool continuity is monitored, the jet flow stability of the special-shaped stamping part is judged, and a high-strength fastener welding track correction scheme is generated. According to the method, the normal, the curvature and the concave-convex state of the stainless steel thin-walled pipe joint are recognized, the posture change of the aluminum alloy corner piece is matched, the focus position and the incident angle are adjusted, the carbon steel reinforcing plate molten pool bubble and metal flowing trend is combined, the forming continuity risk is judged, and the jet flow pressure and the reverse pressure of the special-shaped stamping piece are monitored; the jet flow direction and speed are adjusted, the welding path is corrected according to the posture of the high-strength fastener and the state of a molten pool, and track continuity and forming consistency are improved.

Description

Technical Field

[0001] The invention relates to the technical field of plasma arc welding, in particular to a plasma arc welding method for processing hardware products. Background Art

[0002] The field of plasma arc welding technology includes processing methods that use plasma arc as a heat source to weld metal materials, and is an important branch of arc welding technology. The core content of this technical field is to quickly melt metal workpieces and achieve connections through the high-temperature concentrated heat source generated by a high-energy plasma arc column. Plasma arc welding technology has the characteristics of concentrated heat source, fast welding speed, and large welding depth. It is widely used in aerospace, automobile manufacturing, pressure vessels, medical equipment, and high-precision metal processing. The overall technical field includes keyhole plasma welding, non-transfer plasma welding, transfer plasma welding, and plasma powder surfacing. Among them, transfer plasma welding is the mainstream process, usually combined with shielding gases such as argon for high-quality welding operations.

[0003] Among them, the plasma arc welding method for hardware processing refers to a manufacturing method that uses a plasma arc welding process to fix and connect metal hardware accessories. The technical matters targeted by this patent subject cover the welding treatment of different types of hardware materials such as stainless steel, aluminum alloy, carbon steel, etc. Specifically, the transferred plasma arc is used as a heat source to heat and melt the metal surface in an inert gas environment, and then the welding connection of the metal workpiece is achieved by controlling parameters such as arc current, electrode spacing and welding speed. The technical basis involved is mainly the B23K9 / 30 classification in plasma arc welding, which is characterized by the use of a pinhole effect to generate a concentrated heat source for weld formation. The welding process usually uses a CNC platform to achieve workpiece positioning and welding trajectory control, and cooperates with a high-frequency arc ignition device to ensure the stability of the welding arc.

[0004] Existing welding techniques generally rely on preset trajectory paths and static parameter control during welding, lacking real-time awareness of complex boundary changes within hardware products. This results in spatial offsets of the arc focus when welding concave, convex, or hyperbolic surfaces, preventing it from consistently maintaining optimal alignment with the workpiece surface normal and causing uneven energy distribution. Traditional processes rely solely on surface observation or static heat input adjustment to address changes in the molten pool state during welding. These processes are unable to capture the dynamics of bubbles within the molten pool and metal flow trends in real time, easily leading to molten pool collapse and discontinuous formation. This deficiency is particularly pronounced when welding thin-walled pipe fittings or small corner pieces. Jet flow pressure regulation also suffers from response lag and lacks real-time control of the pressure balance between the jet and liquid metal, making the pinhole effect susceptible to instability. Trajectory control primarily relies on a preset path and fails to adjust to posture changes, boundary dynamics, or molten pool state. Welding path offsets frequently occur, particularly in the spatial welding of irregularly shaped workpieces and high-strength fasteners. This can easily lead to weld misalignment, deformation, or cold welds, directly impacting weld quality and structural strength. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a plasma arc welding method for processing hardware products.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a plasma arc welding method for processing hardware products, comprising the following steps:

[0007] S1: Obtain the normal direction, curvature change, and concave-convex shape of the surface of the stainless steel thin-walled pipe joint. Based on the normal continuity and curvature change, determine whether the workpiece is concave, convex, hyperbolic, or flat, and obtain the workpiece structure matching status;

[0008] S2: Based on the workpiece structure matching state, the posture direction of the aluminum alloy profile corner piece welding gun is obtained. According to the change in the angle between the normal direction and the posture direction, combined with the relationship between the curvature change and the incident direction, the focus normal or tangential offset is determined, and the incident angle is determined to be tilted forward or reversed, thereby obtaining a focus posture adjustment instruction;

[0009] S3: Based on the focus attitude adjustment instruction, the bubble rising, cracking and metal flow state in the carbon steel structure reinforcement plate molten pool are monitored, and according to the matching between the bubble rising trend and the metal flow trend, whether there is a forming discontinuity in the molten pool is determined, and a molten pool forming continuity state determination value is obtained;

[0010] S4: According to the molten pool forming continuity state judgment value, the changes in the jet axial dynamic pressure and the molten pool reverse pressure during the welding process of special-shaped stamping hardware are obtained, and based on the balance relationship, it is determined whether to adjust the jet angle or airflow speed. Combined with the nozzle outlet state and the airflow fluctuation trend, the jet stability control parameters are obtained.

[0011] As a further solution of the present invention, the workpiece structure matching state includes normal direction, curvature change, and concave-convex shape; the focus posture adjustment instruction includes normal offset, tangential offset, positive tilt of the incident angle, and reverse adjustment of the incident angle; the molten pool forming continuity state judgment value includes bubble floating trend, metal flow trend, and forming discontinuity state; the jet stability control parameters include jet angle adjustment, airflow speed increase, nozzle outlet state, and airflow fluctuation trend.

[0012] As a further solution of the present invention, the specific steps of S1 are:

[0013] S101: Acquire three-dimensional point cloud data of the surface of a stainless steel thin-walled pipe joint, calculate the normal vector direction of each sampling point based on the spatial coordinates of the surface sampling points and the spatial position difference of adjacent points, and generate normal vector distribution data based on the change in the angle between the normal vectors;

[0014] S102: calling the normal vector distribution data, calculating the local principal curvature and Gaussian curvature around each sampling point based on the continuous change of the angle between adjacent normal vectors, and establishing the surface curvature change characteristics by comparing the sign relationship between the principal curvature value and the Gaussian curvature;

[0015] S103: Based on the surface curvature variation characteristics, for the numerical range of Gaussian curvature and principal curvature, determine whether the surface morphology corresponding to the sampling point meets the concave, convex, hyperbolic or flat classification criteria, form a distribution structure of the classified morphology in space, and output the structural matching status.

[0016] As a further solution of the present invention, the specific steps of S2 are:

[0017] S201: Based on the workpiece structure matching state, detecting the welding gun posture direction of the aluminum alloy profile corner piece at the current workstation, calling the space vector of the workpiece surface normal direction and the welding gun posture direction, extracting the angle change trend between the two, and obtaining the angle change trend value;

[0018] S202: Based on the angle change trend, the surface curvature change data of the workpiece corner piece is retrieved, combined with the change trend of the welding gun incident direction, and based on the matching relationship between the curvature change and the incident direction, the focus offset state along the normal or tangential direction is determined to obtain the offset dynamic prediction analysis result;

[0019] S203: Based on the dynamic prediction analysis result of the offset, the angle change trend between the normal direction of the workpiece and the incident direction is called, and according to the corresponding relationship between the angle trend and the offset state, whether the incident angle is in a forward tilt or reverse adjustment state is judged, and a focus posture adjustment instruction parameter set is established.

[0020] As a further solution of the present invention, the specific steps of S3 are:

[0021] S301: Based on the focus attitude adjustment instruction parameter set, an imaging signal of the carbon steel structure reinforcement plate molten pool is obtained, the bubble position, size and floating speed are detected, the speed change and direction distribution of the metal flow are monitored, and a bubble floating trend value is generated based on the corresponding relationship between the bubble floating speed and the flow speed;

[0022] S302: calling the bubble upward trend value, extracting the speed change and direction sequence of the metal flow, and obtaining the metal flow trend coefficient based on the consistency of the speed change and the upward trend;

[0023] S303: Extracting the offset interval based on the matching relationship between the metal flow trend coefficient and the bubble floating trend value, determining whether it exceeds the forming continuity offset threshold, and obtaining the molten pool forming continuity state judgment value.

[0024] As a further solution of the present invention, the specific calculation formula of the bubble upward tendency value is:

[0025]

[0026] Among them, H represents the comprehensive displacement trend dimension of the bubble in the molten pool affected by the metal flow and structural characteristics, λ represents the dynamic resistance coefficient of the molten pool, and v b represents the instantaneous rising velocity of the bubble, v m represents the metal flow velocity vector modulus, θ represents the angle between the flow direction and the vertical axis, σ represents the surface roughness factor of the carbon steel structure, ρ b represents the bubble density, ρ m represents the density of molten metal, Δt represents the time sampling interval, represents the Frobenius norm of the metal flow velocity gradient tensor, and n represents the number of sampling points in the effective observation time window.

[0027] As a further solution of the present invention, the specific steps of S4 are:

[0028] S401: Based on the molten pool formation continuity state determination value, the axial dynamic pressure of the plasma jet and the reverse pressure of the liquid metal in the molten pool during the welding process are collected. For the numerical sequence of the two in the same time period, the average change characteristics and fluctuation trends are extracted, the dynamic trend comprehensive index is calculated, and the dynamic pressure and reverse pressure change curve coefficient is generated;

[0029] S402: Based on the dynamic pressure and reverse pressure variation curve coefficient, extract the time period in which the jet dynamic pressure and reverse pressure are in equilibrium. Combine the jet angle offset and the airflow velocity fluctuation amplitude within the time period, and determine whether there is jet offset or airflow instability based on the jet angle offset reference value and the airflow stability fluctuation threshold, thereby obtaining a dynamic prediction analysis result of the offset.

[0030] S403: According to the dynamic prediction analysis results of the offset, combined with the nozzle outlet cross-sectional morphological parameters, outlet flow fluctuation sequence and airflow velocity distribution, adjust the nozzle flow channel structure and airflow output pressure, match the airflow velocity fluctuation changes before and after the adjustment with the outlet flow fluctuation trend, determine whether it enters the airflow stability fluctuation threshold, and generate the jet stability control parameters.

[0031] As a further solution of the present invention, the specific calculation formula for calculating the dynamic trend comprehensive index is:

[0032]

[0033] Among them, W represents the comprehensive index of dynamic trend, μ1 represents the arithmetic mean of the axial dynamic pressure sequence of the plasma jet, and μ2 represents the arithmetic mean of the reverse pressure sequence of the liquid metal in the molten pool. represents the variance of the plasma jet axial dynamic pressure series, represents the variance of the reverse pressure sequence of the liquid metal in the molten pool, α represents the correction coefficient of the temperature gradient of the molten pool, β represents the dynamic weight factor of the jet velocity, and p k represents the instantaneous value of the axial dynamic pressure at the kth sampling point, q k Represents the instantaneous value of the reverse pressure at the kth sampling point, T k represents the measured value of the melt pool temperature at the kth sampling point, v k represents the plasma jet velocity measurement value of the kth sampling point, and s represents the total number of sampling points in the same time period.

[0034] As a further embodiment of the present invention, the method further comprises:

[0035] S5: Based on the jet stability control parameters, the trajectory deviation trend caused by boundary changes, posture adjustment and molten pool state during the welding process of high-strength fastening hardware is obtained. According to the welding gun posture direction, trajectory path changes and workpiece edge changes, it is determined whether the welding trajectory needs deviation correction or path smoothing, and a welding trajectory correction plan is obtained;

[0036] The welding trajectory correction scheme includes trajectory offset correction, path smoothing adjustment, welding gun posture direction change, trajectory path change trend, and workpiece edge change.

[0037] As a further solution of the present invention, the specific steps of S5 are:

[0038] S501: Obtain the jet stabilization control parameters, workpiece boundary change parameters, posture adjustment parameters, and molten pool state parameters, and establish a mapping relationship between posture direction change and boundary change based on the correlation between the welding posture direction and the workpiece boundary change and the morphological change caused by the molten pool state to obtain key morphological features;

[0039] S502: Based on the key morphological features, extracting trajectory path change trends, molten pool disturbances, and posture change parameters, analyzing the corresponding relationship between trajectory deviation trends and posture direction changes, establishing a corresponding mapping between trajectory path deviation trends and posture direction changes, and obtaining deviation dynamic prediction analysis results;

[0040] S503: Call the dynamic prediction analysis results of the offset, combine the jet stability control parameters, posture adjustment parameters and workpiece boundary change parameters, and determine the trajectory correction and path smoothing adjustment requirements based on the matching conditions of the posture direction change and the trajectory offset trend to obtain a welding trajectory correction plan.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are:

[0042] In the present invention, by identifying the normal, curvature and concave-convex state of the stainless steel thin-walled pipe joint, matching the posture changes of the aluminum alloy corner parts, adjusting the focus position and the incident angle, combining the bubbles and metal flow trends of the carbon steel reinforcement plate molten pool, determining the forming continuity risk, monitoring the jet flow pressure and reverse pressure of the special-shaped stamping parts, adjusting the jet direction and speed, and correcting the welding path according to the posture of the high-strength fastener and the molten pool state, thereby improving the trajectory continuity and forming consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0045] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0046] See also Figure 1 , a plasma arc welding method for hardware product processing, comprising the following steps:

[0047] S1: Obtain the normal direction, curvature change, and concave-convex shape of the surface of the stainless steel thin-walled pipe joint. Based on the normal continuity and curvature change, determine whether the workpiece is concave, convex, hyperbolic, or flat, and obtain the workpiece structure matching status;

[0048] S2: Based on the workpiece structure matching status, the welding gun posture direction of the current workstation of the aluminum alloy profile corner piece is obtained. According to the change trend of the angle between the workpiece normal direction and the posture direction, combined with the matching relationship between the curvature change and the incident direction, it is determined whether the focus is offset along the normal or tangential direction, and whether the incident angle is tilted forward or adjusted backward, and the focus posture adjustment instruction is obtained;

[0049] S3: Based on the focus attitude adjustment instruction, the bubble rising, cracking and metal flow status in the carbon steel structure reinforcement plate molten pool are monitored. According to the matching between the bubble rising trend and the metal flow trend, it is judged whether there is a forming discontinuity in the molten pool, and the molten pool forming continuity state judgment value is obtained;

[0050] S4: Based on the continuity of the molten pool, the relationship between the axial dynamic pressure of the plasma jet and the reverse pressure of the liquid metal in the molten pool during the welding of special-shaped stamping hardware is obtained. Based on the equilibrium state of the jet dynamic pressure and reverse pressure, it is determined whether the jet angle needs to be adjusted or the airflow velocity needs to be increased. At the same time, the nozzle outlet state and the airflow fluctuation trend are combined to obtain the jet stability control parameters.

[0051] S5: Based on the jet stability control parameters, the trajectory deviation trend caused by boundary changes, posture adjustment and molten pool state during the welding process of high-strength fastening hardware is obtained. According to the welding gun posture direction, trajectory path change trend and workpiece edge changes, it is determined whether the welding trajectory needs to be corrected for deviation or smoothed, and a corresponding operation plan is formed to obtain a welding trajectory correction plan;

[0052] The workpiece structure matching status includes normal direction, curvature change, and concave and convex shape. The focus posture adjustment instructions include normal offset, tangential offset, positive tilt of the incident angle, and reverse adjustment of the incident angle. The molten pool forming continuity status judgment value includes bubble floating trend, metal flow trend, and discontinuous forming state. The jet stability control parameters include jet angle adjustment, airflow speed increase, nozzle outlet status, and airflow fluctuation trend. The welding trajectory correction scheme includes trajectory offset correction, path smoothing adjustment, welding gun posture direction change, trajectory path change trend, and workpiece edge change.

[0053] The specific steps of S1 are:

[0054] S101: Acquire three-dimensional point cloud data of the surface of a stainless steel thin-walled pipe joint, calculate the normal vector direction of each sampling point based on the spatial coordinates of the surface sampling points and the spatial position difference of adjacent points, and generate normal vector distribution data based on the change in the angle between the normal vectors;

[0055] To obtain three-dimensional point cloud data of the surface of a stainless steel thin-walled pipe joint, a laser scanner or structured light scanner is used, with the scanning accuracy set to within 0.02 mm. A high-density, omnidirectional scan of the pipe joint surface is performed. The resulting point cloud data contains the X, Y, and Z spatial coordinates of each sampling point. First, the original point cloud is denoised. By setting the standard deviation multiple to 1.0, isolated points that deviate significantly from surrounding points are removed. For each sampling point, a neighborhood search is performed to extract 20 neighboring points. The spatial positions of these neighboring points are counted, and the average coordinate value of the neighborhood center is calculated. The covariance matrix is further derived based on the spatial differences of these neighboring points relative to the average value. This matrix is then decomposed to obtain the vector corresponding to the minimum eigenvalue, which is used as the normal vector direction for the point. For example, the average coordinates of the 20 points surrounding a point are 10.02 in the X direction, 5.14 in the Y direction, and 3. 86, where the coordinates of a certain neighborhood point are 10.05 in the X direction, 5.18 in the Y direction, and 3.89 in the Z direction. Based on these spatial differences, the normal vector direction of the point is determined to be 0.12 in the X direction, -0.98 in the Y direction, and 0.14 in the Z direction. The normal vectors of all sampling points on the entire surface are extracted according to the same rule. Then, based on the normal vector directions between adjacent points, the angles between any two adjacent points are calculated. For example, the normal vector of a certain point is 0.12 in the X direction, -0.98 in the Y direction, and 0.14 in the Z direction. The normal vectors of the adjacent points are 0.14 in the X direction, -0.96 in the Y direction, and 0.18 in the Z direction. The angle between the two is about 4.5 degrees. By batch calculating the angle changes between the normal vectors of all adjacent points in the full surface point cloud data, a normal vector distribution matrix is formed. Each item in the matrix represents the angle value between the normal vectors of the corresponding adjacent points, and finally the surface normal vector distribution data is constructed.

[0056] S102: calling normal vector distribution data, calculating the local principal curvature and Gaussian curvature around each sampling point based on the continuous change of the angle between adjacent normal vectors, and establishing the surface curvature change characteristics by comparing the sign relationship between the principal curvature value and the Gaussian curvature;

[0057] First, with each point as the center, the neighborhood radius is set to 1.5mm, and all adjacent points within the radius are extracted. By counting the angles between the normal vector of the center point and the normal vector of each adjacent point, the average change value of these angles is calculated to reflect the gradient change level of the normal vector. Then, according to the spatial distribution trend of the normal vector in the neighborhood, the principal component analysis method is used to extract the main direction. The point and its neighboring points are fitted with a local quadratic surface to derive the values of the principal curvature and secondary curvature. For example, the principal curvature obtained by fitting a certain sampling point is 0.23, and the secondary curvature is -0.1. The Gaussian curvature obtained by multiplying the two is -0.023. By observing the sign relationship between the principal curvature and the Gaussian curvature , establish curvature change characteristics. For example, when the Gaussian curvature is less than zero, and the absolute values of the main curvature and the secondary curvature are both greater than 0.15, the area can be judged as a saddle-shaped hyperboloid. When the Gaussian curvature is greater than zero, and the main curvature and the secondary curvature are similar, with a relative error within 5%, it can be judged as a spherical shape. If the Gaussian curvature is greater than zero and the main curvature is greater than 0.05, it can be judged as convex. If the main curvature is less than -0.05, it is judged as concave. If the absolute value of the Gaussian curvature is less than 0.005 and the absolute value of the main curvature is less than 0.01, it can be judged as a plane area. According to the above rules, continuous calculation and judgment are performed on all sampling points to complete the description of the curvature change characteristics of the entire surface.

[0058] S103: Based on the surface curvature variation characteristics, for the numerical range of Gaussian curvature and principal curvature, determine whether the surface morphology corresponding to the sampling point meets the classification criteria of concave, convex, hyperbolic or flat, form a distribution structure of the classified morphology in space, and output the structure matching status;

[0059] First, set the classification standard and divide the Gaussian curvature and the principal curvature into different numerical intervals. For example, the area with Gaussian curvature between -0.005 and 0.005 and the principal curvature between -0.01 and 0.01 is considered a plane. When the Gaussian curvature is greater than 0.02 and the principal curvature is greater than 0.05, it is considered a convex area. When the Gaussian curvature is greater than 0.02 and the principal curvature is less than -0.05, it is considered a concave area. When the Gaussian curvature is less than zero and the absolute value of the principal curvature is greater than 0.05, it is considered a hyperbolic area. Based on this classification standard, each sampling point is judged one by one. For example, if the Gaussian curvature of a point is 0.03 and the principal curvature is 0.07, it is judged to be convex. Another point is judged to be concave. The Gaussian curvature is -0.025 and the principal curvature is 0.12, which is judged to be a hyperbolic surface. Through three-dimensional coordinate mapping, the classification label of each sampling point is embedded in the point cloud data to form the spatial morphological distribution of the point cloud. The classification label is identified by a numerical value, for example, 0 for a plane, 1 for a convex shape, -1 for a concave shape, and 2 for a hyperbolic surface. Then, based on the spatial label structure, the similarity score of the classification matching is calculated. The scoring method is the number of matched points divided by the total number of points. For example, if the detection object contains 500 points and 450 points are successfully matched, the similarity score is 90%. The scoring threshold is set to 85%. When the score is greater than or equal to 85%, the structural matching status is determined to be qualified, otherwise it is determined to be unqualified.

[0060] The specific steps of S2 are:

[0061] S201: Based on the workpiece structure matching state, the welding gun posture direction of the current workstation of the aluminum alloy profile corner piece is detected, the space vector of the workpiece surface normal direction and the welding gun posture direction is called, the angle change trend between the two is extracted, and the angle change trend value is obtained;

[0062] First, the spatial three-dimensional coordinate information of the corner pieces is extracted through the CAD model of the aluminum alloy profile. Combined with the visual inspection system on the workstation, the actual spatial posture of the current corner piece in the workstation is collected in real time, the theoretical posture is matched with the actual posture, and the position deviation is corrected through the six-degree-of-freedom spatial coordinate transformation matrix. Then, the welding gun posture detection module is used to obtain the current posture direction vector through the inertial navigation sensor on the welding gun. At the same time, the spatial position of the gun head is obtained by combining the position sensor, and the angle between the current welding gun posture direction and the normal of the workpiece surface is calculated. The angle calculation is based on the vector dot product. The change trend is tracked through data from multiple consecutive time points. For example, at time point t1, the angle is 20 degrees, rises to 22 degrees at t2, and increases to 25 degrees at t3, reflecting that the posture direction is gradually deviating from the normal of the workpiece surface. By calculating the angle difference between adjacent time points, the change trend is extracted. If the continuous change trend is positive, it means that the deviation from the normal is increasing, and if the trend is negative, it is approaching the normal. Finally, the angle change trend is obtained as the basis for subsequent analysis.

[0063] S202: Based on the angle change trend, the surface curvature change data of the workpiece corner is retrieved. Combined with the change trend of the welding gun incident direction, the focus offset along the normal or tangential direction is determined based on the matching relationship between the curvature change and the incident direction, and the offset dynamic prediction analysis result is obtained.

[0064] First, the curvature information in the workpiece CAD model is called, and the surface of the corner parts is divided into discrete grids. The main curvature and minor curvature of the surface are calculated for each node, and the node curvature value is obtained by taking the square root of the sum of the two. For example, if the main curvature of a node is 0.02 and the minor curvature is 0.03, the total curvature of the node is about 0.036. Then, combined with the change of the incident direction of the welding gun in adjacent time periods, the trend of the incident direction vector change is calculated to analyze whether it is offset toward the normal or along the tangential direction. When the curvature value is greater than 0.05, it is determined that there is a significant curvature change on the surface. Combined with the change in the incident direction, the welding The trend of the gun's incident angle change, for example, if the current incident angle increases from 28 degrees to 31 degrees, it means that the incident direction is deviating from the normal. If it decreases from 28 degrees to 25 degrees, it means that the incident direction is adjusted toward the normal. The change in the incident angle is associated with the change in surface curvature. By judging the positive and negative values of the angle trend combined with the curvature change, we can further analyze whether the focus is shifted along the normal or tangential direction. For example, if the incident angle increases by 3 degrees and the corresponding curvature node value is 0.06, it is judged to be a normal shift. If the incident angle decreases by 2 degrees and the curvature node is 0.045, it is judged to be a tangential shift, thus forming a complete dynamic prediction analysis of the shift.

[0065] S203: Based on the results of the dynamic prediction analysis of the offset, the angle change trend between the workpiece normal direction and the incident direction is called, and according to the corresponding relationship between the angle trend and the offset state, whether the incident angle is in a forward tilt or reverse adjustment state is determined, and a focus posture adjustment instruction parameter set is established;

[0066] Based on the results of the dynamic prediction analysis of the offset, the trend of the angle change between the workpiece normal and the incident direction is continued to be called, and the change state of the incident angle is judged by the time series. For example, the angle is 18 degrees at time point t1, decreases to 15 degrees at t2, and continues to decrease to 12 degrees at t3. The trend is continuously decreasing, and it is determined to be a reverse adjustment. If the angle continues to increase, it is determined to be a positive tilt state. Subsequently, the attitude adjustment parameter set is established in combination with the angle change trend and the offset state. The adjustment gain coefficient range is set between 0.1 and 0.5. When the angle change is greater than 5 degrees and the predicted offset exceeds 3 degrees, the gain coefficient is set to 0.5 for large adjustments. When the angle change is less than 2 degrees and the offset is less than 1 degree, the gain coefficient is set to 0.1 for fine-tuning. The attitude adjustment instruction is described in the form of quaternion parameters. For example, the current attitude parameter is a set of quaternions, and the adjustment target forms a new attitude quaternion. The attitude update is achieved by multiplying the two sets of quaternions. Finally, the new attitude instruction is output for welding gun focus attitude adjustment, thereby forming a dynamically adjusted instruction parameter set.

[0067] The specific steps of S3 are:

[0068] S301: Based on the focus attitude adjustment instruction parameter set, an imaging signal of the carbon steel structure reinforcement plate molten pool is obtained, the bubble position, size and floating speed are detected, the speed change and direction distribution of the metal flow are monitored, and a bubble floating trend value is generated based on the corresponding relationship between the bubble floating speed and the flow speed;

[0069] The specific calculation formula for establishing the bubble floating trend value is:

[0070]

[0071] Among them, H represents the comprehensive displacement trend dimension of the bubble in the molten pool affected by the metal flow and structural characteristics, λ represents the dynamic resistance coefficient of the molten pool (0<λ≤1), and v b represents the instantaneous rising velocity of the bubble (mm / ms), v m represents the metal flow velocity vector modulus (mm / ms), θ represents the angle between the flow direction and the vertical axis (0≤θ≤π / 2), σ represents the surface roughness factor of the carbon steel structure (σ≥0), ρ b Represents bubble density (g / mm 3 ), ρ m Represents the density of molten metal (g / mm 3 ), Δt represents the time sampling interval (ms), Frobenius norm representing the metal flow velocity gradient tensor (mm / ms 2 ), n represents the number of sampling points in the effective observation time window;

[0072] λ = 0.5 was measured by a high-temperature molten pool rheological properties tester. This value was obtained by normalizing the ratio of the metal viscosity coefficient (23.6 Pa·s) to the temperature gradient (85°C / mm). The fluctuation range is positively correlated with the oscillation frequency of the molten pool.

[0073] v b = 3.2 mm / ms The bubble trajectory is captured by a 20,000 fps high-speed camera system, and the displacement data within a 10 ms time window is extracted and calculated as the mean using image analysis software;

[0074] v m =5.8mm / ms A laser Doppler velocimeter was used to continuously collect 100 sets of velocity vector data in the middle area of the molten pool and take the modulus average;

[0075] θ = 0.4887 rad (28°) The velocity vector field is obtained by particle image velocimetry and the mean angle with the vertical axis is calculated;

[0076] σ=0.45 Use a three-dimensional surface profiler to measure the surface roughness of the carbon steel structural plate, taking 5×5mm 2 RMS height value in the area;

[0077] ρ b =0.0012g / mm 3 The bubble composition (78% H2, 15% CO, 7% N2) was analyzed by gas chromatography and then calculated according to the ideal gas state equation;

[0078] ρ m =7.8g / mm 3 Measured by Archimedes method in the molten state at 1600℃;

[0079] Δt = 0.1ms is set according to the maximum sampling frequency of the high-speed camera system;

[0080] After obtaining the velocity field through the particle image velocimetry system, the Frobenius norm of the spatial derivative matrix is calculated;

[0081] n=100 is obtained by dividing the 10 ms effective observation time window by Δt.

[0082] Calculation process:

[0083] Calculate the denominator:

[0084]

[0085] Calculate the speed difference:

[0086]

[0087] Calculate the drag product term:

[0088] 0.5·1.354=0.677;

[0089] Calculate the density ratio term:

[0090]

[0091] Compute the gradient time term:

[0092]

[0093] Single sampling value accumulation:

[0094] 100·(0.0001538+0.0728)=100·0.07295=7.295;

[0095] Final calculation results:

[0096] H = 0.677·7.295 = 4.94;

[0097] The results show that the bubble floating tendency value is positively correlated with the metal flow velocity gradient. When the value increases by 10%, the H value increases by 6.2%. The calculated result 4.94 is used as the direct output of the bubble floating tendency value, and the numerical dimension is mm / ms·s. -1 , reflecting the net displacement trend of the bubble affected by fluid dynamics per unit time.

[0098] S302: calling the bubble upward trend value, extracting the speed change and direction sequence of the metal flow, and obtaining the metal flow trend coefficient based on the consistency of the speed change and the upward trend;

[0099] Based on the bubble rising trend value, the change sequence of metal flow velocity is extracted. For example, in the past 10 seconds, the metal flow velocity changed to 50, 55, 60, 65, 70, 75, 80, 78, 76, and 74 mm / s, respectively. The change amount of this velocity sequence in adjacent time periods is calculated, and the change amount sequence is 5, 5, 5, 5, 5, 5, -2, -2, and -2 mm / s, respectively. At the same time, the corresponding changes in metal flow direction are recorded, for example, the direction gradually changes from 0 degrees to 90 degrees and then slowly falls back to 60 degrees. Subsequently, a correlation analysis is performed on the change synchronization between the velocity change sequence and the bubble rising trend value. The correlation index is close to 92%, reflecting a significant synchronization relationship between the two. The metal flow trend function is further constructed through regression analysis. If the regression coefficient is 1.2 and the intercept is 5, when the bubble rising trend value is 4 mm / s, the metal flow trend is approximately 9.8 mm / s. Finally, the fluctuation coefficient of the metal flow trend is calculated based on the velocity change sequence, and its value is approximately 10.24. This fluctuation coefficient reflects the stability and fluctuation amplitude of the metal flow velocity within this time period.

[0100] S303: extracting the offset interval based on the matching relationship between the metal flow trend coefficient and the bubble floating trend value, determining whether it exceeds the forming continuity offset threshold, and obtaining the molten pool forming continuity state judgment value;

[0101] According to the matching relationship between the metal flow trend coefficient and the bubble floating trend value, a two-dimensional feature space is constructed, with the metal flow trend coefficient as the horizontal axis and the bubble floating trend value as the vertical axis. The current metal flow trend coefficient is 10.24, and the bubble floating trend value is 4. Referring to historical data, the mean of the metal flow trend is set to 8, the standard deviation is set to 2, the mean of the bubble floating trend value is set to 3.5, and the standard deviation is 1. The offset is calculated based on the position offset in the feature space, and the offset is approximately 2.31. The forming continuity offset threshold is set to 2.5. When the offset is less than the threshold, the current molten pool forming state is judged to be continuous. If the offset is equal to or greater than the threshold, it is judged that the molten pool forming is unstable. In this example, the offset does not exceed the threshold, so the forming continuity state of the molten pool is continuous.

[0102] The specific steps of S4 are:

[0103] S401: Based on the molten pool formation continuity state judgment value, the axial dynamic pressure of the plasma jet and the reverse pressure of the liquid metal in the molten pool during the welding process are collected. For the numerical sequence of the two in the same time period, the average change characteristics and fluctuation trends are extracted, the dynamic trend comprehensive index is calculated, and the dynamic pressure and reverse pressure change curve coefficient is generated;

[0104] The specific calculation formula for calculating the dynamic trend comprehensive index is:

[0105]

[0106] Among them, W represents the comprehensive index of dynamic trend, μ1 represents the arithmetic mean of the axial dynamic pressure sequence of the plasma jet, and μ2 represents the arithmetic mean of the reverse pressure sequence of the liquid metal in the molten pool. represents the variance of the plasma jet axial dynamic pressure series, represents the variance of the reverse pressure sequence of the liquid metal in the molten pool, α represents the correction coefficient of the temperature gradient of the molten pool, β represents the dynamic weight factor of the jet velocity, and p k represents the instantaneous value of the axial dynamic pressure at the kth sampling point, q k Represents the instantaneous value of the reverse pressure at the kth sampling point, T k represents the measured value of the melt pool temperature at the kth sampling point, v k represents the plasma jet velocity measurement value at the kth sampling point, and s represents the total number of sampling points in the same time period;

[0107] The arithmetic mean μ1 of the plasma jet axial dynamic pressure sequence is calculated using 100 consecutive sampling point data collected by the pressure sensor. The specific value is

[0108] The arithmetic mean μ2 of the reverse pressure sequence of the molten pool liquid metal is calculated by the reaction force data measured by the laser displacement sensor. The specific value is

[0109] Variance of the plasma jet axial dynamic pressure series Calculated from pressure sensor data,

[0110] Variance of the reverse pressure series of the liquid metal in the molten pool Calculated from the reaction force data,

[0111] The molten pool temperature gradient correction coefficient α is determined based on the molten pool temperature field data measured by the infrared thermal imager. When the maximum temperature gradient is 15K / mm, α = 0.85. This coefficient decreases linearly with the increase of the temperature gradient.

[0112] The dynamic weight factor β of the jet velocity is calculated using the plasma jet velocity data measured by high-speed video. When the standard deviation of the jet velocity is 0.3 m / s, β corresponds to 1.2. This factor increases with the increase of the velocity fluctuation amplitude.

[0113] The instantaneous value of the axial dynamic pressure at the kth sampling point p k The pressure is obtained by the pressure sensor at a sampling frequency of 1kHz, the typical value is p 50 =12.7MPa;

[0114] Instantaneous value of reverse pressure q k Synchronously measured by reaction force sensor, typical value q 50 =8.1MPa;

[0115] Melt pool temperature measurement value T k The infrared thermometer is used to collect data at a frequency of 200 Hz, with a typical value of T 50 =1850K;

[0116] Plasma jet velocity measurement value v k Obtained by particle image velocimetry, typical value v 50 =2.1m / s;

[0117] The total number of sampling points s = 100 corresponds to a 0.1 second time window;

[0118] Substitute specific numerical values for calculation:

[0119]

[0120] First calculation:

[0121]

[0122] The second calculation selects the typical sampling point k=50:

[0123]

[0124] The accumulated value of 100 sampling points is approximately 0.118.

[0125] Final calculation results:

[0126] W=0.85×(10.35+0.548)+1.2×0.118=9.264+0.142=9.406;

[0127] This result indicates that the dynamic trend composite index reflects the mean difference and fluctuation characteristics of the axial dynamic pressure and the reverse pressure. A larger value indicates lower system stability. The calculated result of 9.406 is used as an input parameter in the curve fitting algorithm, which generates the dynamic pressure and reverse pressure variation curve coefficients through cubic spline interpolation.

[0128] S402: Based on the dynamic pressure and back pressure variation curve coefficient, the time period in which the jet dynamic pressure and back pressure are in equilibrium is extracted. The jet angle offset and the airflow velocity fluctuation amplitude within the time period are combined. Based on the jet angle offset reference value and the airflow stability fluctuation threshold, it is determined whether there is jet offset or airflow instability, and the offset dynamic prediction analysis result is obtained.

[0129] The time period in which the jet flow pressure and the reverse pressure are in equilibrium is screened based on the variation curve coefficient. The equilibrium state is determined by a rate of change of no more than 5% and a variation curve coefficient between 2.0 and 3.0. The jet angle offset and the airflow velocity fluctuation amplitude within this time period are then extracted. The angle offset is determined by image recognition of the angle between the jet centerline and the theoretical vertical direction. For example, if the theoretical angle is 90 degrees and the actual measurement is 92.5 degrees, the offset angle is 2.5 degrees. The airflow velocity fluctuation amplitude is obtained by using a velocity sensor to obtain a velocity sequence and calculate the standard deviation. When the offset angle exceeds the baseline value by 3 degrees or the airflow velocity fluctuation amplitude exceeds 5 m / s, it is determined that the jet is offset or the airflow is unstable. If the offset angle is 4.1 degrees and the velocity fluctuation is 6.3 m / s, it is identified as an offset and unstable state. If the offset angle is 1.8 degrees and the velocity fluctuation is 4.2 m / s, it is determined to be normal. The prediction analysis results are used for subsequent nozzle flow channel structure adjustment and airflow parameter correction.

[0130] S403: Based on the results of the dynamic prediction analysis of the offset, combined with the nozzle outlet cross-sectional morphological parameters, the outlet flow rate fluctuation sequence, and the airflow velocity distribution, the nozzle flow channel structure and the airflow output pressure are adjusted. The airflow velocity fluctuation changes before and after the adjustment are matched with the outlet flow rate fluctuation trend. The determination is made as to whether the airflow falls within the stable fluctuation threshold, and the jet stability control parameters are generated.

[0131] First, the current nozzle outlet cross-sectional morphological parameters are called, including the outlet diameter of 3.5mm and the ellipticity of 0.05. At the same time, the outlet flow fluctuation sequence and the airflow velocity distribution are extracted. The outlet flow is collected in real time by the mass flow meter with a sampling frequency of 100Hz and a collection time of 5 seconds to obtain 500 sets of flow data. The ratio of the difference between the maximum flow and the minimum flow to the average flow is calculated. When the ratio is greater than 10%, it is determined that the flow fluctuation is too large. For example, the maximum flow is 28L / min, the minimum flow is 24L / min, and the average flow is 26L / min. The fluctuation ratio is 15.4%, which is abnormal. Combined with the airflow velocity distribution, the velocity skewness is calculated to determine whether the airflow is skewed. Away from the center, when the skewness coefficient is greater than 0.8 or less than -0.8, the velocity distribution is considered abnormal. The nozzle structure parameters are adjusted according to the skewness results. For example, the outlet diameter is adjusted from 3.5mm to 3.2mm or 3.8mm, the internal contraction angle is adjusted from 20 degrees to 15 degrees, and the airflow output pressure is increased from 0.6MPa to 0.7MPa. Through adjustment, the airflow velocity fluctuation amplitude is reduced to within 5m / s, and the flow fluctuation ratio is reduced to within 8%. When the airflow velocity fluctuation does not exceed 5m / s and the flow fluctuation does not exceed 10%, it is determined that the jet is within the stable fluctuation threshold, forming the jet stability control parameters including outlet diameter, outlet pressure, and contraction angle.

[0132] The specific steps of S5 are:

[0133] S501: Obtain jet stabilization control parameters, workpiece boundary change parameters, posture adjustment parameters, and molten pool state parameters. Based on the correlation between welding posture direction and workpiece boundary change, combined with the morphological change caused by the molten pool state, a mapping relationship between posture direction change and boundary change is established to obtain key morphological features.

[0134] First, the jet stability control parameters are collected, including the wire feeding speed set to 12m / min, the gas flow rate to 20L / min, the current to 180A, and the voltage to 25V. These parameters are obtained in real time through the data interface of the welding equipment. At the same time, the molten pool temperature is monitored by infrared thermal imaging equipment. The current molten pool temperature is 1600℃, and the molten pool contour morphology changes are extracted in real time with the help of the machine vision system. The workpiece boundary change parameters are measured by the laser displacement sensor. The current boundary fluctuation is within the range of plus or minus 1.5mm. The posture adjustment parameters come from the position feedback of the six-axis robot arm. The current posture changes are 5 degrees on the X axis, 3 degrees on the Y axis, and 2 degrees on the Z axis. Then the posture changes are compared with the workpiece boundary changes. Through calculation, it is found that the two have a highly linear correlation. For example, when the X-axis posture change increases by 5 degrees, the boundary offset of the workpiece increases synchronously by about 1.5mm. In terms of the change of the molten pool state, the current detected aspect ratio of the molten pool is 1.3 to 1, and the molten pool contour width fluctuates significantly, with a fluctuation amplitude of 40%. Combining the boundary change trend with the molten pool morphology change, a mapping model is established between the posture direction change and the boundary change. Through fitting, it is found that when the X-axis offset is about 5 degrees, the Y-axis offset is 3 degrees, and the Z-axis offset is 2 degrees, the corresponding boundary offset is approximately in the range of plus or minus 1.5mm, and the molten pool remains in a state of aspect ratio of 1.3 and fluctuation rate of 40%. Finally, a complete set of key morphological features is obtained.

[0135] S502: Based on the key morphological features, the trajectory path change trend, the molten pool disturbance and the posture change parameters are extracted, the corresponding relationship between the trajectory deviation trend and the posture direction change is analyzed, and the corresponding mapping between the trajectory path deviation trend and the posture direction change is established to obtain the deviation dynamic prediction analysis result;

[0136] Based on the extraction of key morphological features, the trajectory path change trend is acquired in real time. Monitoring by a laser tracking system shows that the current trajectory is continuously deflecting at a rate of 0.2 mm per second along the Y-axis. Molten pool disturbances are detected by a high-speed camera system. The current melt pool surface fluctuation frequency is 10 Hz, and the amplitude reaches 0.8 mm. Simultaneously, the posture change trend shows a continuous deviation of 4 degrees along the X-axis, 2.5 degrees along the Y-axis, and 1.8 degrees along the Z-axis. Subsequently, data fitting analysis of the trajectory deviation trend and posture direction changes revealed that for every 1-degree increase in posture change, the trajectory deviation increases by approximately 0.6 mm. Further deviation levels are classified: deviations less than or equal to 0.5 mm are considered normal, deviations between 0.5 mm and 1.5 mm are considered minor, and deviations exceeding 1.5 mm are defined as severe. The currently observed deviation is 0.9 mm, which is classified as minor. Based on the current trajectory change rate and posture change trend, if the X-axis posture continues to increase to more than 5 degrees, the deviation is expected to reach 1.5 mm within the next 5 seconds, approaching the severe deviation threshold. This provides the results of the dynamic deviation prediction analysis.

[0137] S503: Calling the offset dynamic prediction analysis results, combining the jet stability control parameters, posture adjustment parameters and workpiece boundary change parameters, and according to the matching conditions between the posture direction change and the trajectory offset trend, determining the trajectory correction and path smoothing adjustment requirements, and obtaining the welding trajectory correction plan;

[0138] Based on the results of dynamic prediction analysis of offset, combined with the jet stability control parameters (wire feeding speed 12m / min, gas flow rate 20L / min), posture adjustment parameters (X-axis 5 degrees, Y-axis 3 degrees, Z-axis 2 degrees) and workpiece boundary changes (plus or minus 1.5mm), according to the relationship between posture change and trajectory offset fitted in the early stage, when the X-axis posture change increases to 5.5 degrees, the predicted trajectory offset reaches 1.9mm, which exceeds the slight offset standard. According to the judgment rule, when the predicted offset is greater than or equal to 1.5mm, trajectory correction needs to be triggered. After this condition is met, the trajectory correction scheme is started. Through gain adjustment, the current error of 1.9mm is calculated, and the corresponding correction amount is 1.33mm. Compensation adjustment is performed along the negative direction of the Y axis, and posture smoothing adjustment is performed at the same time. The X axis is adjusted from the current 5.5 degrees to 5.3 degrees, the Y axis remains at 3 degrees, and the Z axis remains at 2 degrees. Finally, a complete welding trajectory correction scheme is formed, including compensation of 1.33mm along the negative direction of the Y axis and smooth fine-tuning of the posture direction.

[0139] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A plasma arc welding method for hardware processing, characterized in that: The following steps are involved: S1: Obtain the normal direction, curvature change, and concave-convex shape of the surface of the stainless steel thin-walled pipe joint. Based on the normal continuity and curvature change, determine whether the workpiece is concave, convex, hyperbolic, or flat, and obtain the workpiece structure matching status; S2: Based on the workpiece structure matching state, the posture direction of the aluminum alloy profile corner piece welding gun is obtained. According to the change in the angle between the normal direction and the posture direction, combined with the relationship between the curvature change and the incident direction, the focus normal or tangential offset is determined, and the incident angle is determined to be tilted forward or reversed, thereby obtaining a focus posture adjustment instruction; S3: Based on the focus attitude adjustment instruction, the bubble rising, cracking and metal flow state in the carbon steel structure reinforcement plate molten pool are monitored, and according to the matching between the bubble rising trend and the metal flow trend, whether there is a forming discontinuity in the molten pool is determined, and a molten pool forming continuity state determination value is obtained; S4: According to the molten pool forming continuity state judgment value, the changes in the jet axial dynamic pressure and the molten pool reverse pressure during the welding process of special-shaped stamping hardware are obtained, and based on the balance relationship, it is determined whether to adjust the jet angle or airflow speed. Combined with the nozzle outlet state and the airflow fluctuation trend, the jet stability control parameters are obtained.

2. The plasma arc welding method for hardware product processing according to claim 1, characterized in that: The workpiece structure matching state includes normal direction, curvature change, and concave-convex shape; the focus posture adjustment instruction includes normal offset, tangential offset, positive tilt of the incident angle, and reverse adjustment of the incident angle; the molten pool forming continuity state judgment value includes bubble floating trend, metal flow trend, and forming discontinuity state; the jet stability control parameters include jet angle adjustment, airflow speed increase, nozzle outlet state, and airflow fluctuation trend.

3. The plasma arc welding method for hardware product processing according to claim 1, characterized in that: The specific steps of S1 are: S101: Acquire three-dimensional point cloud data of the surface of a stainless steel thin-walled pipe joint, calculate the normal vector direction of each sampling point based on the spatial coordinates of the surface sampling points and the spatial position difference of adjacent points, and generate normal vector distribution data based on the change in the angle between the normal vectors; S102: calling the normal vector distribution data, calculating the local principal curvature and Gaussian curvature around each sampling point based on the continuous change of the angle between adjacent normal vectors, and establishing the surface curvature change characteristics by comparing the sign relationship between the principal curvature value and the Gaussian curvature; S103: Based on the surface curvature variation characteristics, for the numerical range of Gaussian curvature and principal curvature, determine whether the surface morphology corresponding to the sampling point meets the concave, convex, hyperbolic or flat classification criteria, form a distribution structure of the classified morphology in space, and output the structural matching status.

4. The plasma arc welding method for hardware processing according to claim 3, characterized in that: The specific steps of S2 are: S201: Based on the workpiece structure matching state, detecting the welding gun posture direction of the aluminum alloy profile corner piece at the current workstation, calling the space vector of the workpiece surface normal direction and the welding gun posture direction, extracting the angle change trend between the two, and obtaining the angle change trend value; S202: Based on the angle change trend, the surface curvature change data of the workpiece corner piece is retrieved, combined with the change trend of the welding gun incident direction, and based on the matching relationship between the curvature change and the incident direction, the focus offset state along the normal or tangential direction is determined to obtain the offset dynamic prediction analysis result; S203: Based on the dynamic prediction analysis result of the offset, the angle change trend between the normal direction of the workpiece and the incident direction is called, and according to the corresponding relationship between the angle trend and the offset state, whether the incident angle is in a forward tilt or reverse adjustment state is judged, and a focus posture adjustment instruction parameter set is established.

5. The plasma arc welding method for hardware processing according to claim 4, characterized in that: The specific steps of S3 are: S301: Based on the focus attitude adjustment instruction parameter set, an imaging signal of the carbon steel structure reinforcement plate molten pool is obtained, the bubble position, size and floating speed are detected, the speed change and direction distribution of the metal flow are monitored, and a bubble floating trend value is generated based on the corresponding relationship between the bubble floating speed and the flow speed; S302: calling the bubble upward trend value, extracting the speed change and direction sequence of the metal flow, and obtaining the metal flow trend coefficient based on the consistency of the speed change and the upward trend; S303: Extracting the offset interval based on the matching relationship between the metal flow trend coefficient and the bubble floating trend value, determining whether it exceeds the forming continuity offset threshold, and obtaining the molten pool forming continuity state judgment value.

6. The plasma arc welding method for hardware processing according to claim 5, characterized in that: The specific calculation formula of the bubble floating tendency value is: Among them, H represents the comprehensive displacement trend dimension of the bubble in the molten pool affected by the metal flow and structural characteristics, λ represents the dynamic resistance coefficient of the molten pool, and v b represents the instantaneous rising velocity of the bubble, v m represents the metal flow velocity vector modulus, θ represents the angle between the flow direction and the vertical axis, σ represents the surface roughness factor of the carbon steel structure, ρ b represents the bubble density, ρ m represents the density of molten metal, Δt represents the time sampling interval, represents the Frobenius norm of the metal flow velocity gradient tensor, and n represents the number of sampling points in the effective observation time window.

7. The plasma arc welding method for hardware processing according to claim 5, characterized in that: The specific steps of S4 are: S401: Based on the molten pool formation continuity state determination value, the axial dynamic pressure of the plasma jet and the reverse pressure of the liquid metal in the molten pool during the welding process are collected. For the numerical sequence of the two in the same time period, the average change characteristics and fluctuation trends are extracted, the dynamic trend comprehensive index is calculated, and the dynamic pressure and reverse pressure change curve coefficient is generated; S402: Based on the dynamic pressure and reverse pressure variation curve coefficient, extract the time period in which the jet dynamic pressure and reverse pressure are in equilibrium. Combine the jet angle offset and the airflow velocity fluctuation amplitude within the time period, and determine whether there is jet offset or airflow instability based on the jet angle offset reference value and the airflow stability fluctuation threshold, thereby obtaining a dynamic prediction analysis result of the offset. S403: According to the dynamic prediction analysis results of the offset, combined with the nozzle outlet cross-sectional morphological parameters, outlet flow fluctuation sequence and airflow velocity distribution, adjust the nozzle flow channel structure and airflow output pressure, match the airflow velocity fluctuation changes before and after the adjustment with the outlet flow fluctuation trend, determine whether it enters the airflow stability fluctuation threshold, and generate the jet stability control parameters.

8. The plasma arc welding method for hardware processing according to claim 7, characterized in that: The specific calculation formula for calculating the dynamic trend comprehensive index is: Among them, W represents the comprehensive index of dynamic trend, μ1 represents the arithmetic mean of the axial dynamic pressure sequence of the plasma jet, and μ2 represents the arithmetic mean of the reverse pressure sequence of the liquid metal in the molten pool. represents the variance of the plasma jet axial dynamic pressure series, represents the variance of the reverse pressure sequence of the liquid metal in the molten pool, α represents the correction coefficient of the temperature gradient of the molten pool, β represents the dynamic weight factor of the jet velocity, and p k represents the instantaneous value of the axial dynamic pressure at the kth sampling point, q k Represents the instantaneous value of the reverse pressure at the kth sampling point, T k represents the measured value of the melt pool temperature at the kth sampling point, v k represents the plasma jet velocity measurement value of the kth sampling point, and s represents the total number of sampling points in the same time period.

9. The plasma arc welding method for hardware processing according to claim 1, characterized in that: The method further comprises: S5: Based on the jet stability control parameters, the trajectory deviation trend caused by boundary changes, posture adjustment and molten pool state during the welding process of high-strength fastening hardware is obtained. According to the welding gun posture direction, trajectory path changes and workpiece edge changes, it is determined whether the welding trajectory needs deviation correction or path smoothing, and a welding trajectory correction plan is obtained; The welding trajectory correction scheme includes trajectory offset correction, path smoothing adjustment, welding gun posture direction change, trajectory path change trend, and workpiece edge change.

10. The plasma arc welding method for hardware processing according to claim 9, characterized in that: The specific steps of S5 are: S501: Obtain the jet stabilization control parameters, workpiece boundary change parameters, posture adjustment parameters, and molten pool state parameters, and establish a mapping relationship between posture direction change and boundary change based on the correlation between the welding posture direction and the workpiece boundary change and the morphological change caused by the molten pool state to obtain key morphological features; S502: Based on the key morphological features, extracting trajectory path change trends, molten pool disturbances, and posture change parameters, analyzing the corresponding relationship between trajectory deviation trends and posture direction changes, establishing a corresponding mapping between trajectory path deviation trends and posture direction changes, and obtaining deviation dynamic prediction analysis results; S503: Call the dynamic prediction analysis results of the offset, combine the jet stability control parameters, posture adjustment parameters and workpiece boundary change parameters, and determine the trajectory correction and path smoothing adjustment requirements based on the matching conditions of the posture direction change and the trajectory offset trend to obtain a welding trajectory correction plan.

Images