Back free-form welding method for high-strength steel telescopic boom tail end

Abstract

This invention relates to the field of welding technology, specifically to a back-side free-forming welding method for the tail end of a high-strength steel telescopic boom. The method includes: during the welding process, acquiring temperature data of the root region of the bevel and the base materials on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data; determining the molten pool flow stability; determining the root metal sinking and penetration consistency index; determining a compensation factor; and adjusting the welding arc voltage according to the compensation factor. This method, by constructing a two-dimensional evaluation system of molten pool flow stability and root metal sinking and penetration consistency index, and determining a compensation factor based on the coupling relationship between the two to dynamically adjust the welding arc voltage, achieves coordinated control of molten pool flow stability and root metal forming behavior. This effectively solves the coupling effect problem of uneven lateral temperature gradient and molten pool dynamic flow instability caused by abrupt stiffness changes and uneven heat dissipation in the thick plate box-shaped structure at the tail end of a high-strength steel telescopic boom.

Description

Technical Field

[0001] This invention relates to the field of welding technology, and specifically to a back free-forming welding method for the tail end of a high-strength steel telescopic boom. Background Technology

[0002] With the rapid development of the construction machinery industry, cranes, aerial work platforms, fire trucks, and large material handling equipment are placing increasingly higher demands on the load-bearing capacity, weight ratio, and structural reliability of telescopic boom structures. To improve structural strength and reduce overall weight, the main structure of telescopic booms is increasingly being manufactured using high-strength steel with a yield strength exceeding 700 MPa. High-strength steel offers advantages such as high strength, good toughness, and lightweight construction. However, during the welding of thick-plate box-shaped structures, the tail end region often experiences sudden changes in local stiffness and non-uniform heat dissipation. This leads to a coupling effect between uneven transverse temperature gradients and dynamic flow instability of the molten pool at the root of the bevel, placing higher demands on the welding process.

[0003] During the back-side free-forming welding process of the tail end structure of a high-strength steel telescopic boom, the tail end area typically adopts a thick plate box-shaped structure with local abrupt changes in stiffness and non-uniform heat dissipation conditions. The welding heat input in the root region of the groove is prone to a coupling effect of uneven transverse temperature gradient and dynamic flow instability of the molten pool. Under this coupling effect, the balance between the surface tension, gravity and arc pressure of the molten metal at the root gap fluctuates periodically, making it easy for the weld pool to experience transient collapse or local stagnation when penetrating the root of the groove. This results in a reduction in the uniformity of weld penetration at the root and makes it difficult to maintain stable back-side free-forming welding quality. Summary of the Invention

[0004] This invention provides a back-free forming welding method for the tail end of a high-strength steel telescopic boom to solve existing problems.

[0005] The present invention provides a back-side free-forming welding method for the tail end of a high-strength steel telescopic boom, which adopts the following technical solution:

[0006] One embodiment of the present invention provides a back-side free-forming welding method for the tail end of a high-strength steel telescopic boom. The method includes: during the welding process, acquiring temperature data of the root region of the bevel and the base materials on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data; determining the molten pool flow stability based on the temperature data, the welding heat input parameters, and the molten pool tail metal reflow state data; wherein the molten pool flow stability is used to quantify the stability of the molten pool tail metal reflow; determining a root metal sinking and penetration consistency index based on the root gap filling state data; wherein the root metal sinking and penetration consistency index is used to quantify the root gap filling consistency; determining a compensation factor based on the molten pool flow stability and the root metal sinking and penetration consistency index; and adjusting the welding arc voltage according to the compensation factor.

[0007] Further, determining the flow stability of the molten pool based on the temperature data, the welding heat input parameters, and the molten pool tail metal reflow state data includes: determining a lateral temperature gradient offset index based on the temperature data; wherein the lateral temperature gradient offset index is used to quantify the unevenness of heat distribution in the bevel section; determining the lateral imbalance of heat input based on the lateral temperature gradient offset index and the welding heat input parameters; wherein the lateral imbalance of heat input is used to quantify the imbalance of welding heat input in lateral distribution; determining the lateral expansion offset of the molten pool front based on the lateral imbalance of heat input and the asymmetry of the lateral expansion distance of the molten pool front; wherein the lateral expansion offset of the molten pool front is used to quantify the actual impact of heat input imbalance on the molten pool morphology; and determining the tail metal reflow stability based on the lateral expansion offset of the molten pool front and the fluctuation characteristics of the molten pool tail metal reflow velocity.

[0008] Furthermore, the method for determining the degree of asymmetry in the lateral expansion distance of the molten pool front edge includes: determining the position of the weld centerline, the position of the left edge boundary of the molten pool, and the position of the right edge boundary of the molten pool; wherein, the position of the weld centerline is determined based on the welding torch travel direction; calculating a first lateral expansion distance between the left edge boundary of the molten pool and the weld centerline, and a second lateral expansion distance between the right edge boundary of the molten pool and the weld centerline; determining the degree of asymmetry in the lateral expansion distance of the molten pool front edge based on the absolute difference between the first lateral expansion distance and the second lateral expansion distance; wherein, the larger the absolute difference, the higher the degree of asymmetry.

[0009] Further, determining the root metal sinking penetration consistency index based on the root gap filling state data includes: obtaining a filling height sequence and an occupancy ratio sequence of molten metal within the root gap; wherein the occupancy ratio is the ratio of the instantaneously occupied area of ​​the molten metal to the theoretically filled area of ​​the root region; determining the sinking fluctuation degree based on the degree of fluctuation in the height difference between adjacent moments in the filling height sequence; determining the filling sufficiency degree based on the average level of the occupancy ratio sequence; and determining the root metal sinking penetration consistency index based on the sinking fluctuation degree and the filling sufficiency degree; wherein the smaller the sinking fluctuation degree and the higher the filling sufficiency degree, the higher the consistency index.

[0010] Further, determining the compensation factor based on the molten pool flow stability and the root metal sinking penetration consistency index includes: calculating the product of the molten pool flow stability and the root metal sinking penetration consistency index; determining the compensation factor based on the product; wherein the compensation factor is negatively correlated with the product, and the smaller the product, the larger the compensation factor.

[0011] Further, adjusting the welding arc voltage according to the compensation factor includes: obtaining a set voltage value under normal and stable welding conditions and the maximum voltage compensation range allowed by the system; determining a voltage compensation amount based on the compensation factor and the maximum voltage compensation range, and superimposing the set voltage value and the voltage compensation amount to determine the adjusted welding arc voltage.

[0012] Further, the acquisition of temperature data of the root region of the bevel and the base materials on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data includes: acquiring welding heat input parameters; wherein, the welding heat input parameters include welding current, welding voltage, and welding speed; acquiring an image sequence of the tail region of the molten pool, and determining the molten pool tail metal reflow velocity based on the image sequence; acquiring root gap filling state data; wherein, the root gap filling state data includes the filling height and occupancy ratio of molten metal in the root gap; the occupancy ratio is the ratio of the instantaneously occupied area of ​​the molten metal to the theoretically filled area of ​​the root region; and acquiring temperature data of the root region of the bevel and the base materials on both sides.

[0013] Furthermore, the step of acquiring an image sequence of the tail region of the molten pool and determining the molten pool tail metal reflow velocity based on the image sequence includes: continuously imaging the tail region of the molten pool to acquire the image sequence; using an image recognition algorithm to track the flow trajectory of the tail molten metal in the image sequence; and calculating the molten pool tail metal reflow velocity based on the change in the molten metal flow position in adjacent image frames.

[0014] Furthermore, the method also includes: dividing the effective monitoring period of the welding process into several observation time windows, and repeatedly executing the data acquisition step, the molten pool stability determination step, the root metal sinking penetration consistency index determination step, the compensation factor determination step, and the welding arc voltage adjustment step within each observation time window, so as to dynamically update the welding arc voltage along the weld direction.

[0015] Furthermore, the method also includes: determining the effective monitoring time period of the welding process based on the starting conditions for the stable formation of the welding arc and the termination conditions for the completion of root penetration; wherein, the starting conditions include the welding current and voltage reaching a set stable value, and the termination conditions include the tail shape change of the molten pool tending to be stable or the arc electrical signal returning to a stable value.

[0016] The beneficial effects of the technical solution of the present invention are:

[0017] In this embodiment of the invention, during the welding process, temperature data of the root region of the groove and the base materials on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data are acquired. Based on the temperature data, welding heat input parameters, and molten pool tail metal reflow state data, the molten pool flow stability is determined; wherein, the molten pool flow stability is used to quantify the stability of the molten pool tail metal reflow. Based on the root gap filling state data, a root metal sinking and penetration consistency index is determined; wherein, the root metal sinking and penetration consistency index is used to quantify the root gap filling consistency. Based on the molten pool flow stability and the root metal sinking and penetration consistency index, a compensation factor is determined; and the welding arc voltage is adjusted according to the compensation factor.

[0018] This invention, by constructing a two-dimensional evaluation system of molten pool flow stability and root metal penetration consistency index, and determining a compensation factor based on the coupling relationship between the two to dynamically adjust the welding arc voltage, achieves coordinated control of molten pool flow stability and root metal forming behavior. This effectively solves the coupling effect problem of uneven lateral temperature gradient and molten pool dynamic flow instability caused by abrupt stiffness changes and uneven heat dissipation in the thick-plate box-shaped structure at the tail end of a high-strength steel telescopic boom. Furthermore, by establishing a multi-level quantitative characterization model of lateral temperature gradient shift index, lateral imbalance of heat input, and molten pool leading edge expansion shift, the welding heat input distribution, molten pool geometry evolution, and tail metal return behavior are transformed into calculable stability parameters. Qualitative indicators enable quantitative analysis and precise identification of the thermo-mechanical coupling state during the welding process. Furthermore, an iterative observation time window mechanism is employed, real-time acquisition of multi-source monitoring data and dynamic updates of compensation factors and welding arc voltage within each window, forming a continuous closed-loop control along the weld direction. This overcomes the limitations of traditional empirical parameter control in adapting to the time-varying characteristics of the welding process. Additionally, through adaptive adjustment of the welding arc voltage by the compensation factor, heat input is enhanced to improve molten pool fluidity when tail metal backflow is unstable or root metal sinking is discontinuous. This suppresses transient collapse and local retention of molten metal in the root gap, improving root penetration consistency and back-side free-form quality. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of a high-strength steel telescopic boom structure provided in an embodiment of this application;

[0021] Figure 2 A schematic flowchart illustrating the back free-forming welding method for the tail end of a high-strength steel telescopic boom provided in an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of the molten pool front provided for an embodiment of this application. Detailed Implementation

[0023] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, provides a specific implementation method, structure, features, and effects of the invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] The following description, in conjunction with the accompanying drawings, details a specific solution provided by the present invention.

[0026] like Figure 1 As shown, during the free-forming welding process of the back of the high-strength steel telescopic boom tail structure, the tail area typically employs a thick-plate box-shaped structure (see...). Figure 1 (The dotted line in the image) Accompanied by abrupt changes in local stiffness and non-uniform heat dissipation, the welding heat input in the root region of the bevel easily creates a coupling effect between uneven transverse temperature gradients and dynamic flow instability of the molten pool. Under this coupling effect, the balance between surface tension, gravity, and arc pressure of the molten metal at the root gap will fluctuate periodically, making the weld pool prone to transient collapse or local stagnation when penetrating the root of the bevel, thereby reducing the uniformity of weld root penetration. Existing free-forming welding processes usually rely mainly on empirical parameters to control heat input, lacking a process mechanism for the coordinated control of molten pool flow stability and root metal forming behavior. Therefore, it is difficult to achieve stable and consistent back free-forming welding quality in the tail structure of high-strength steel telescopic booms.

[0027] In view of this, this application provides a back free-forming welding method for the tail end of a high-strength steel telescopic boom. This method constructs a two-dimensional evaluation system of the stability of the molten pool flow and the consistency index of the root metal sinking and penetration, and determines a compensation factor based on the coupling relationship between the two to dynamically adjust the welding arc voltage. This achieves the coordinated control of the molten pool flow stability and the root metal forming behavior, effectively solving the coupling effect problem of uneven lateral temperature gradient and dynamic flow instability of the molten pool caused by abrupt stiffness change and uneven heat dissipation in the thick plate box structure at the tail end of the high-strength steel telescopic boom.

[0028] like Figure 2 As shown in the embodiment of this application, a method for free-form welding of the back end of a high-strength steel telescopic boom is provided, including:

[0029] Step S110: During the welding process, acquire temperature data of the root region of the groove and the base material on both sides, welding heat input parameters, molten pool tail metal reflow status data and root gap filling status data.

[0030] The following sections describe the data obtained in step S110:

[0031] (1) Temperature data of the root area of ​​the bevel and the parent material on both sides;

[0032] This mainly includes the average temperature of the root region at the bottom of the bevel, as well as the average temperature of the base material on the left and right sides of the bevel, all collected in real time during the welding process. This set of data constitutes the temperature field distribution information on the transverse section of the weld joint, covering the heat-affected zone from the weld center to the edge of the base material, and is used to characterize the transfer and distribution of heat in the transverse direction during the welding process.

[0033] Temperature data from the root region of the bevel and the base metal on both sides can reflect the uniformity of the welding heat input distribution across the bevel cross-section. Specifically, this is manifested as the transverse temperature gradient caused by the temperature difference between the root region of the bevel and the base metal on both sides. The magnitude of the transverse temperature gradient directly characterizes the unevenness of heat accumulation and dissipation across the bevel cross-section, providing a thermodynamic basis for judging the melting state and flow conditions of the molten pool in the root region of the bevel, and thus identifying areas of heat concentration or dissipation that may occur during the welding process.

[0034] Temperature data of the root region of the bevel and the base material on both sides can be obtained by arranging a network of thermocouple temperature sensors in the welding area. A certain number of thermocouples are arranged at the bottom of the bevel to form a root temperature monitoring network, covering the root gap and its adjacent area; a certain number of thermocouples are arranged on each side of the left and right edges of the bevel, distributed along the thickness direction of the base material to cover the heat-affected zone of the base material. Each thermocouple collects temperature signals in real time at a set sampling frequency. After signal conditioning and data synchronization, the average temperature value of each region is calculated to form the average temperature of the root region and the average temperature of the base material on both sides.

[0035] After acquiring temperature data of the root region of the bevel and the parent material on both sides, the temperature data can be preprocessed. The preprocessing methods include: time synchronization and alignment of temperature signals from multiple measuring points based on a unified time reference to eliminate acquisition time lag errors; using filtering algorithms to remove high-frequency electromagnetic interference and random noise to improve the signal-to-noise ratio; eliminating outliers caused by sensor malfunctions or instantaneous disturbances through threshold discrimination or statistical methods; establishing a continuous temperature field distribution of the bevel cross section based on spatial interpolation algorithms to complete the temperature information between measuring points; calculating the average temperature value within the observation time window to characterize the regional thermal state; and performing normalization processing to eliminate dimensional differences and provide a standardized data basis for subsequent lateral temperature gradient calculations.

[0036] (2) Welding heat input parameters;

[0037] Welding heat input parameters, including welding current, welding voltage, and welding speed, constitute a set of key process parameters characterizing the energy obtained per unit length of weld during the welding process. Among them, the product of welding current and welding voltage characterizes the instantaneous heat power output by the arc, while welding speed reflects the heat input distribution density along the longitudinal direction of the weld. Together, these three parameters determine the heat input intensity and thermal cycle characteristics of the weld pool.

[0038] Welding heat input parameters reflect the stability of the welding arc and the dynamic changes in the heat input of the molten pool, specifically characterized by the energy intensity level of the molten pool within a short observation window. Changes in welding heat input parameters directly affect the temperature field distribution of the molten pool, the fluidity of the molten metal, and solidification behavior. They provide an energetic basis for assessing the risk of root retention, transient collapse tendency, and microstructural evolution of the heat-affected zone, serving as a fundamental data source for welding process quality control and defect prevention.

[0039] Welding thermal input parameters can be acquired in real time through a multi-source sensing system: an arc signal acquisition module is installed at the power output port to record the instantaneous values ​​of welding current and welding voltage in real time at a set sampling frequency; the displacement rate of the welding torch relative to the workpiece is monitored in real time by the encoder or speed sensor built into the welding equipment to obtain welding speed data. Each parameter acquisition channel is synchronized with the temperature monitoring and visual monitoring systems to ensure the correspondence of multi-source data under a unified time reference.

[0040] After acquiring the welding heat input parameters, they can be preprocessed. The preprocessing methods include: achieving strict time synchronization of current, voltage, and speed data based on timestamp information to establish a data set corresponding to the welding position; using digital filtering algorithms to remove grid harmonic interference and high-frequency noise to extract the true characteristics of the arc electrical signal; calculating the average current, average voltage, and average welding speed within the observation time window using the sliding window averaging method to characterize the heat input level within the time period; and removing outliers or smoothing abnormal jump data to eliminate interference from the transient process of arc ignition or extinguishing.

[0041] (3) Data on the reflux status of metal at the tail end of the molten pool;

[0042] The molten pool tail metal reflow state data refers to the dynamic behavioral parameters characterizing the reflow of molten metal from the front edge to the tail edge and its sinking into the root gap during back freeform welding. It mainly includes the molten pool tail metal reflow velocity and its fluctuation characteristic sequence. The molten pool tail metal reflow state data can be obtained by analyzing continuous image sequences of the molten pool tail region. It reflects the flow trajectory and transient velocity changes of the molten metal under the combined effects of surface tension, gravity, and arc pressure, constituting a key monitoring quantity for assessing the hydrodynamic stability inside the molten pool.

[0043] Data on the molten pool tail reflow status reflects the continuity and stability of the molten metal reflow within the molten pool, specifically characterized by the variation in reflow velocity between adjacent moments. In a stable freeform welding process, the molten metal reflow velocity towards the tail remains essentially continuous, with minimal variation between adjacent moments; however, when the molten pool flow becomes unstable, the reflow velocity exhibits drastic fluctuations. By quantifying the degree of fluctuation in the tail reflow velocity, dynamic flow instability in the molten pool can be identified, the continuity of molten metal transport within the molten pool can be determined, and thus the molten pool's ability to fill the root gap and the stability of the back freeform quality can be assessed.

[0044] Optionally, the above-mentioned acquisition of an image sequence of the tail region of the molten pool and determination of the molten pool tail metal reflow velocity based on the image sequence includes: continuously imaging the tail region of the molten pool to acquire an image sequence; using an image recognition algorithm to track the flow trajectory of the tail molten metal in the image sequence; and calculating the molten pool tail metal reflow velocity based on the change in the molten metal flow position in adjacent image frames.

[0045] At the hardware level, a high-speed industrial camera positioned to the side or rear of the welding torch can be used to continuously image the tail region of the molten pool. The frame rate can be set according to the welding speed to capture dynamic details of the molten pool surface. Alternatively, a binocular vision system or a structured light 3D measurement device can be used to obtain the 3D morphology and surface velocity field distribution of the molten pool tail. In specific application scenarios, infrared thermal imaging technology can also be combined to monitor the temperature field distribution of the molten pool tail to indirectly reflect the metal flow state. The imaging system maintains a relatively fixed geometric relationship with the welding arc to ensure a one-to-one correspondence between the image sequence and the welding position.

[0046] At the image processing level, the determination of the molten pool tail metal reflow velocity can be implemented using various image recognition and motion estimation algorithms. Based on traditional computer vision methods, optical flow can be used to calculate pixel-level motion vector fields in the image sequence to characterize the apparent flow velocity of the molten metal, or particle image velocimetry can be used to project tracer particles onto the molten pool surface to track the centroid displacement. Based on feature extraction methods, edge detection or texture analysis algorithms can be used to identify the spatial position changes of specific morphological features at the molten pool tail (such as solidified grains, surface ripples, or oxide traces), and the instantaneous reflow velocity can be calculated by the ratio of the displacement difference of feature points in adjacent image frames to the time interval. Based on deep learning methods, convolutional neural network or recurrent neural network models can be constructed, and the nonlinear mapping relationship between the molten pool image and the flow velocity can be learned through training to directly regress and predict the reflow velocity value. These methods can be used independently or in combination to construct a reflow velocity sequence within the observation time window to characterize the degree of fluctuation.

[0047] In the above-described scheme, during the back-side freeform welding process, the molten metal recirculation at the tail end of the molten pool is a crucial step in transporting molten metal from the heated area at the front edge of the molten pool to the cooled area at the tail end and ultimately sinking into the root gap. Its flow stability directly determines the continuity of the root gap filling and the quality of the back-side forming. By capturing continuous image sequences of the molten pool tail end using an industrial camera and employing image recognition algorithms to track the motion trajectory of the molten metal surface features, the invisible internal metal flow is transformed into observable and quantifiable velocity field data, enabling real-time perception of the dynamic flow state of the molten pool. The recirculation velocity and its fluctuation characteristics, calculated based on the changes in the metal flow position between adjacent image frames, can effectively characterize whether there are local stagnation, recirculation interruptions, or flow instability phenomena within the molten pool. This provides a data foundation for calculating the stability of the tail-side metal recirculation by combining the degree of expansion offset at the front edge of the molten pool, and further enables closed-loop control of the molten pool flow stability through compensation factors and voltage adjustments.

[0048] (4) Data on the filling status of the root gap;

[0049] The aforementioned root gap filling status data refers to a set of quantitative parameters characterizing the geometric filling characteristics and dynamic deposition behavior of molten metal within the root gap during back-side freeform welding. These parameters primarily include the filling height sequence and occupancy ratio sequence of the molten metal within the root gap. The filling height sequence reflects the instantaneous deposition position change of the molten metal along the depth direction of the root gap under the combined effects of gravity, arc pressure, and surface tension. The occupancy ratio sequence reflects the ratio of the instantaneous filling area of ​​the molten metal on the cross-section of the root gap to the theoretically filled area of ​​the root region. The root gap filling status data constitutes a core monitoring indicator for evaluating the root metal sinking and penetration process and the final forming quality.

[0050] Root gap filling status data reflects the continuity and sufficiency of molten metal sinking in the root gap. Specifically, it is characterized by the fluctuation range of the height difference between adjacent moments during the molten metal sinking process and the average level of molten metal filling the root gap. The smaller the fluctuation in filling height, the smoother the molten metal sinking process and the better the penetration continuity; the closer the occupancy ratio is to the theoretical full state, the more sufficiency of root gap filling and the higher the back-side forming integrity. By coupling the sinking fluctuation and filling sufficiency, transient collapse, local retention, or lack of fusion defects of molten metal in the root gap can be identified, providing direct forming status basis for judging the weld root penetration consistency and back-side free forming quality.

[0051] Root gap filling status data can be obtained through various non-contact measurement methods. In a visual monitoring approach, an industrial camera or high-speed camera continuously images the root gap area from the back of the weld. Image recognition algorithms extract the molten metal surface contour, and the instantaneous filling height of the molten metal within the root gap is calculated. Combined with two-dimensional cross-sectional analysis or three-dimensional reconstruction techniques, the instantaneous area occupied by the molten metal within the root gap is calculated, thus obtaining an occupancy ratio sequence. In a laser measurement approach, a laser rangefinder or laser scanning device measures the distance to the molten metal surface within the root gap, directly acquiring the filling height data. Laser contour scanning obtains the geometric morphology of the root gap cross-section, and the ratio of the molten metal-occupied area to the theoretically filled area is calculated. These measurement methods can be used independently or in combination, synchronously recording the filling height and occupancy ratio within an observation time window at a set sampling frequency, forming a root gap filling status data set.

[0052] Step S120: Based on temperature data, welding heat input parameters, and molten pool tail metal reflow status data, determine the molten pool flow stability; wherein, the molten pool flow stability is used to quantify the stability of molten pool tail metal reflow.

[0053] The aforementioned melt pool flow stability is a comprehensive evaluation index used to quantify the continuity and stability of the molten metal recirculation behavior in the tail region of the melt pool during back-side freeform welding. It characterizes the dynamic flow state of the molten metal as it is transported from the front to the tail and further sinks into the root gap, reflecting the coupling relationship between heat input distribution, melt pool morphology evolution, and hydrodynamic behavior during welding. A higher melt pool flow stability value indicates a smoother and less volatile recirculation process at the tail of the melt pool, and better continuity of molten metal transport within the melt pool. Conversely, a lower value indicates recirculation interruption, local stagnation, or flow instability, directly affecting the filling capacity of the root gap and the consistency of the back-side freeform welding quality.

[0054] Optionally, step S120 may include: determining a transverse temperature gradient offset index based on temperature data; wherein the transverse temperature gradient offset index is used to quantify the unevenness of heat distribution in the groove section; determining the degree of transverse imbalance of heat input based on the transverse temperature gradient offset index and welding heat input parameters; wherein the degree of transverse imbalance of heat input is used to quantify the degree of imbalance in the transverse distribution of welding heat input; determining the degree of transverse expansion offset of the molten pool front based on the degree of transverse imbalance of heat input and the asymmetry of the transverse expansion distance of the molten pool front; wherein the degree of transverse expansion offset of the molten pool front is used to quantify the actual impact of heat input imbalance on the molten pool morphology; and determining the stability of the tail metal reflow based on the degree of transverse expansion offset of the molten pool front and the fluctuation characteristics of the tail metal reflow velocity.

[0055] The transverse temperature gradient offset index is a dimensionless indicator characterizing the uniformity of heat distribution across the bevel cross-section during welding. It quantifies the temperature deviation between the root region of the bevel and the base material on both sides, identifying uneven transverse heat accumulation caused by abrupt changes in local stiffness and differences in heat dissipation conditions. A higher index value indicates a more uneven heat distribution across the bevel cross-section, a significant temperature gradient between the molten pool root region and the base material, and a higher risk of molten metal stagnation or transient collapse at the root gap. The calculation method can be:

[0056] ;

[0057] The above formula reflects the degree of heat accumulation or loss at the root of the molten pool relative to the base material by calculating the absolute deviation between the temperature of the root region and the average temperature of the base material on both sides. In the formula, and These represent the average monitored temperatures of the parent material on the left and right sides of the bevel, respectively. This indicates the average temperature in the root region at the bottom of the bevel. This represents a sufficiently small non-zero constant (e.g., 0.0001) to ensure that the calculation result is not zero and to avoid numerical anomalies.

[0058] The degree of lateral heat input imbalance is a comprehensive indicator characterizing the coupling effect between the intensity of welding heat input and the degree of uneven lateral heat distribution within a short observation window. This indicator not only reflects the energy input level of the molten pool but also considers the imbalance state of heat distribution laterally, thus distinguishing between two working conditions: simply high heat input and heat input imbalance with potential instability risk. This provides an energy and thermodynamic coupling basis for assessing the risk of molten pool root retention or collapse. The aforementioned degree of lateral heat input imbalance... The calculation method can be:

[0059] ;

[0060] The above formula couples the energy input intensity per unit length of weld (characterizing the heat power density of the molten pool) with the degree of unevenness in the transverse temperature gradient; when the temperature gradient offset exponent is large (uneven heat distribution) and the heat input per unit length is high, An increase in the value indicates that the molten pool is in a dangerous state of high heat input and lateral imbalance, with a higher risk of molten pool root stagnation or transient collapse. In the formula, This represents the average welding voltage within the current observation window; This represents the average welding current within the current observation window; Indicates the average welding speed; This represents the hyperbolic tangent function, used to measure the exponential shift of the lateral temperature gradient. Normalization is performed. Characterizes the energy intensity of the molten pool within a short time window.

[0061] Figure 3 The diagram illustrates the top-view geometry of the molten pool during back-side free-form welding. The molten pool front refers to the foremost boundary of the molten metal region (molten pool) formed below the welding arc in the direction of the welding torch's travel. This boundary constitutes the dynamic interface between the molten metal and the unmelted base material. Figure 3 As shown, the leading edge is located at the very front of the weld centerline (i.e., the trajectory of the welding torch), and extends laterally to the base material on both sides of the bevel to form the width boundary of the molten pool. The lateral extension distance on both sides (i.e., the first distance between the left leading edge boundary of the molten pool and the weld centerline, and the second distance between the right leading edge boundary of the molten pool and the weld centerline) together determine the overall geometric shape of the molten pool leading edge. When the heat input is evenly distributed laterally, it exhibits an approximately symmetrical extension state, while when the heat input is unbalanced, it exhibits an asymmetrical shape, which directly affects the path and stability of the molten metal flowing back to the tail.

[0062] The degree of expansion and offset of the molten pool front is a quantitative indicator characterizing the actual impact of lateral heat input imbalance on the geometry of the molten pool. In stable free-form welding, the expansion range of the molten pool front to the left and right is usually symmetrical. When the lateral distribution of heat input is unbalanced, changes in the temperature field and surface tension distribution inside the molten pool lead to changes in the flow direction of the molten metal, causing the molten pool front to exhibit an asymmetrical expansion state. This indicator quantifies this asymmetry to reflect the response of the molten pool morphology evolution to heat input imbalance.

[0063] The asymmetry of the lateral expansion distance of the molten pool front can be determined in at least one of the following ways:

[0064] The first method: Determined based on geometric features;

[0065] Optionally, the method for determining the degree of asymmetry in the lateral expansion distance of the molten pool front includes: determining the position of the weld centerline, the position of the left edge boundary of the molten pool, and the position of the right edge boundary of the molten pool; wherein, the position of the weld centerline is determined based on the direction of the welding torch; calculating the first lateral expansion distance between the left edge boundary of the molten pool and the weld centerline, and the second lateral expansion distance between the right edge boundary of the molten pool and the weld centerline; determining the degree of asymmetry in the lateral expansion distance of the molten pool front based on the absolute difference between the first lateral expansion distance and the second lateral expansion distance; wherein, the larger the absolute difference, the higher the degree of asymmetry.

[0066] The above-mentioned degree of expansion and offset of the molten pool front The calculation method is as follows:

[0067] ;

[0068] The above calculation formula couples and quantifies the degree of lateral imbalance in heat input with the asymmetry of the molten pool geometry, characterizing the actual impact of heat input imbalance on the molten pool morphology through the deviation in left and right expansion distances; when the lateral imbalance in heat input is severe and the left and right expansion of the molten pool is significantly asymmetrical, A significant increase in the value indicates that the molten pool is in a state of morphological instability. In the formula, It indicates the degree of lateral imbalance in heat input, reflecting the degree of imbalance in the lateral distribution of welding heat input; This indicates the first lateral extension distance between the left leading edge boundary of the molten pool and the weld centerline; This indicates the second lateral extension distance between the right leading edge boundary of the molten pool and the weld centerline; This represents a sufficiently small non-zero constant (e.g., 0.0001) to ensure the calculation result is not zero and to avoid numerical anomalies; absolute value term. It directly quantifies the degree of asymmetry in the left-right expansion of the molten pool front.

[0069] The determination of the weld centerline position is based on the welding torch travel direction and can be implemented in several ways: In one embodiment, the weld centerline position is directly obtained through pre-set welding path planning data, and this position remains constant in a fixed coordinate system; In another embodiment, the weld centerline is dynamically determined by tracking the displacement trajectory of the welding torch relative to the workpiece in real time, combined with real-time position information fed back by an encoder or vision positioning system; In yet another embodiment, the welding torch center position is inferred as the weld centerline by analyzing the axis of symmetry of the molten pool image or the direction of the highest temperature gradient in the temperature field.

[0070] The embodiments of this application can determine the position of the left and right leading edge boundaries of the molten pool in at least one of the following ways: (1) Determination based on image grayscale gradient: Use an industrial camera to acquire a grayscale image of the molten pool area, extract the contour line of the molten pool by edge detection algorithm (such as Canny operator, Sobel operator or Laplacian operator), and identify the foremost point of the molten pool leading edge contour on the left and right sides as the position of the left and right leading edge boundaries; In the image processing process, threshold segmentation can be used to separate the bright area of ​​the molten pool from the background of the base material, and noise interference can be eliminated by morphological operations (such as corrosion, expansion), and then the coordinates of the leading edge boundaries can be determined by skeleton extraction or contour tracking. (2) Determination based on temperature field abrupt change: Use an infrared thermal imaging camera to acquire an image of the temperature field distribution on the surface of the molten pool, identify the position where the maximum temperature gradient is located as the solid-liquid interface line, that is, the leading edge boundary of the molten pool; By scanning the temperature gradient peak point along the weld seam laterally, locate the abrupt change positions of the temperature gradient on the left and right sides respectively, and map them as the left and right leading edge boundaries of the molten pool. (3) Determining the contour based on laser scanning: The surface of the molten pool is scanned using a line laser scanning sensor or a structured light three-dimensional measurement system to obtain the three-dimensional point cloud data or two-dimensional cross-sectional contour of the molten pool; the boundary of the molten pool front is identified by analyzing the abrupt change in surface height or curvature change, that is, the junction between the molten metal surface and the base material surface; the position of the left and right front boundaries of the molten pool is determined by extracting the coordinates of the highest point or curvature extreme point on the left and right sides.

[0071] The above scheme takes the weld centerline as the theoretical axis of symmetry of the welding heat source, and the expansion distance of the molten pool front on both sides should theoretically remain equal. However, when the lateral heat input distribution becomes unbalanced due to sudden changes in local stiffness or differences in heat dissipation conditions during actual welding, the temperature field inside the molten pool exhibits an asymmetrical distribution. The surface tension gradient drives the molten metal to flow towards the low-temperature side, resulting in an increase in the expansion distance of the molten pool front on the low-temperature side and a decrease in the expansion distance on the high-temperature side. By accurately measuring the expansion distance of the frontal boundaries on both sides relative to the weld centerline and calculating their difference, the morphological deviation caused by this thermo-mechanical coupling effect can be quantified into a calculable physical quantity, thus providing a direct geometric criterion for evaluating the degree of response of the molten pool morphology to the imbalance of heat input.

[0072] The second method: determined based on the centroid offset of the molten pool;

[0073] In this method, the overall contour of the molten pool can be extracted using an image processing algorithm, and its geometric centroid position can be calculated to determine the lateral offset of the centroid relative to the weld centerline. The weighted distance integrals of the left and right contours of the molten pool relative to the centroid are calculated to obtain the left weighted extension distance and the right weighted extension distance. The degree of asymmetry in the lateral extension distance of the molten pool front edge is determined based on the ratio or difference between the left weighted extension distance and the right weighted extension distance. The greater the deviation of the ratio from 1 or the greater the difference, the higher the degree of asymmetry.

[0074] The third method: Determined based on fitting the axis of symmetry;

[0075] In this method, the least squares method or principal component analysis method can be used to fit the symmetry axis of the molten pool front profile point set to obtain the actual symmetry axis direction of the molten pool front; the included angle or lateral offset distance between the actual symmetry axis and the weld centerline is calculated; the degree of asymmetry of the lateral expansion distance of the molten pool front is determined according to the size of the included angle or the lateral offset distance, where the larger the included angle or the larger the lateral offset distance, the higher the degree of asymmetry.

[0076] The fourth method: Determined based on temperature field gradient analysis;

[0077] In this method, infrared thermal imaging can be used to obtain the temperature field distribution on the surface of the molten pool, and the location of the maximum temperature gradient can be identified as the boundary of the molten pool front. The lateral distances of the temperature gradient peak positions on the left and right sides of the molten pool front relative to the weld centerline are extracted to obtain the first temperature gradient distance and the second temperature gradient distance. The degree of asymmetry of the lateral expansion distance of the molten pool front is determined based on the absolute difference between the first temperature gradient distance and the second temperature gradient distance. The larger the absolute difference, the higher the degree of asymmetry.

[0078] The aforementioned tail-end metal reflow stability (i.e., molten pool flow stability) is a quantitative indicator characterizing the continuity and stability of molten metal reflow behavior in the tail region of the molten pool. This indicator comprehensively considers the coupling effect of molten pool leading-edge morphological shift on the tail-end reflow path, as well as the fluctuation characteristics of the reflow velocity itself. A higher value indicates a smoother molten metal reflow process towards the tail, smaller velocity changes between adjacent moments, a more stable internal fluid dynamic state of the molten pool, and a more ideal back-side free-form quality. The calculation formula is:

[0079] ;

[0080] in, This represents the number of sampling points within the observation window. and They represent the first The and the first Monitoring values ​​of tail metal reflux velocity at each sampling point; For sufficiently small nonzero constants; denominator terms The fluctuation intensity of the tail metal reflow velocity is characterized. To determine the fluctuation characteristics of the tail metal reflow velocity, motion estimation techniques based on optical flow can be used to calculate the motion vectors of particles on the molten metal surface in the image sequence, or particle image velocimetry can be used to obtain the velocity field distribution through tracer particle tracking. Alternatively, deep learning models can be used to directly regress and predict the reflow velocity. The degree of velocity fluctuation can be calculated using statistical methods such as standard deviation, variance, or the average of the absolute values ​​of the differences between adjacent frames to quantify the stability of the reflow process.

[0081] Step S130: Based on the root gap filling status data, determine the root metal sinking penetration consistency index; wherein, the root metal sinking penetration consistency index is used to quantify the root gap filling consistency.

[0082] The aforementioned root metal penetration consistency index is a quantitative parameter used to comprehensively evaluate the dynamic deposition behavior and final geometric filling state of molten metal in the root gap of the groove during back free-forming welding. The root metal penetration consistency index characterizes the stability of molten metal penetrating the root of the groove under the combined effects of gravity, arc pressure, and surface tension by coupling the continuity of molten metal penetration along the depth direction of the root gap with the sufficiency of filling along the cross-sectional direction. A higher value indicates a smoother molten metal penetration process, more sufficient root gap filling, and a more uniform and complete back weld formation.

[0083] The purpose of determining the root metal sinking and penetration consistency index in the above scheme is to establish a quantitative correlation between the flow state at the tail of the molten pool and the final forming quality of the root. By monitoring the changes in the filling height and occupancy ratio of the molten metal in the root gap in real time, the actual motion state information of the molten metal after flowing back from the tail of the molten pool and depositing downwards and penetrating the root can be directly obtained. This index transforms the invisible metal transport process inside the molten pool into a quantifiable forming state parameter, providing direct data support for judging whether the molten metal forms a continuous and uniform back forming structure. The determination of the root metal sinking and penetration consistency index realizes the closed-loop control basis for the final forming quality of the welding process. Through coupled calculation with the molten pool flow stability, it is possible to distinguish complex working conditions such as stable molten pool flow but poor root filling, or unstable molten pool flow but acceptable root filling. Based on the feedback adjustment of this index, when the molten metal sinking is discontinuous or the filling is insufficient, voltage compensation can be used to enhance the heat input, improve the molten pool fluidity, and promote the full filling of the root gap, thereby ensuring the formation of a consistent free-forming back weld along the weld direction.

[0084] Optionally, step S130 may include: obtaining the filling height sequence and occupancy ratio sequence of molten metal in the root gap; wherein, the occupancy ratio is the ratio of the instantaneously occupied area of ​​molten metal to the theoretically filled area of ​​the root region; determining the sinking fluctuation degree based on the degree of change in the height difference between adjacent moments in the filling height sequence; determining the filling sufficiency degree based on the average level of the occupancy ratio sequence; and determining the root metal sinking penetration consistency index based on the sinking fluctuation degree and the filling sufficiency degree; wherein, the smaller the sinking fluctuation degree and the higher the filling sufficiency degree, the higher the consistency index.

[0085] The above-mentioned root metal sinking penetration consistency index The calculation formula is:

[0086] ;

[0087] The above formula evaluates both the continuity of molten metal sinking and the integrity of root filling by multiplying the reciprocal of the height variation fluctuation by the average occupancy ratio. When the molten metal sinking process is stable (small height difference between adjacent moments) and the root gap is fully filled (high occupancy ratio), The significantly increased value indicates high consistency in the downward penetration of the metal at the root and good back-side forming quality. In the formula, and They represent the first The and the first Normalized value of the molten metal filling height in the root gap at each sampling time; Represent a sufficiently small nonzero constant; Characterizes the average fluctuation amplitude of height changes between adjacent time points in the filled height sequence; This indicates the proportion of the root gap that is instantly filled by molten metal (i.e., the occupancy ratio). The closer it is to 1, the more fully it is filled. This represents the average level of the occupancy sequence, reflecting the adequacy of the filling of the root gap.

[0088] The embodiments of this application can obtain the filling height sequence and occupancy ratio sequence of molten metal in the root gap through at least one of the following methods: (1) Acquisition based on back visual monitoring: The root gap area is continuously imaged using an industrial camera or high-speed camera set on the back of the weld. The outline of the molten metal surface is extracted by image preprocessing (including filtering, contrast enhancement and background subtraction). The boundary line between the molten metal and the background of the base material in the root gap is identified based on the edge detection algorithm. The vertical distance of the liquid surface relative to the bottom of the root gap is calculated as the filling height. The pixel area occupied by the molten metal is calculated by two-dimensional image segmentation. The theoretical filling area is calculated by combining the known geometric dimensions of the root gap, and then the occupancy ratio is obtained. Images are continuously acquired and processed at a set sampling frequency to construct the filling height sequence and the occupancy ratio sequence. (2) Based on laser triangulation: The distance of the molten metal surface in the root gap is measured from the back of the weld using a laser range sensor or a line laser scanning device to directly obtain the molten metal liquid level height data; the geometric topography point cloud data of the root gap cross section is obtained by laser contour scanning, and the ratio of the actual cross-sectional area occupied by the molten metal to the theoretical cross-sectional area of ​​the root gap is calculated as the occupancy ratio; continuous scanning is performed at a set sampling frequency to construct the filling height sequence and the occupancy ratio sequence; this method can obtain high-precision three-dimensional geometric information and is suitable for scenarios with strict requirements on the root forming size. (3) Based on infrared thermal imaging: The temperature field distribution of the root gap area is monitored from the back of the weld using an infrared thermal imaging camera, and the filling boundary of the molten metal is identified based on the temperature difference between the molten metal and the solid base material; the filling height of the molten metal (based on the location of the temperature gradient change) and the filling area (based on the high temperature region range) are estimated by the temperature field reconstruction algorithm, and then the occupancy ratio is calculated; this method can indirectly obtain the root filling state under conditions of strong arc light interference or difficulty in visible light imaging. (4) Acquisition based on laser-vision fusion: Structured light stripes are projected into the root gap using a line laser projector, and deformed light stripe images are captured by an industrial camera. The three-dimensional morphology of the molten metal surface is calculated based on the structured light three-dimensional reconstruction algorithm. Height information is extracted from the three-dimensional morphology data to construct a filling height sequence. The ratio of the area occupied by the molten metal to the theoretically filled area is calculated by cross-sectional analysis to construct an occupancy ratio sequence. This method combines the precision of laser measurement with the intuitiveness of visual monitoring and is suitable for high-precision monitoring under complex working conditions.

[0089] The above scheme is based on a coupled evaluation mechanism of the deposition kinetics and geometric filling quality of molten metal in the root gap: after the molten metal flows back from the tail of the molten pool, it settles downwards along the root gap under the action of gravity. The continuity of this settling process directly determines whether the root gap can be continuously filled without producing defects such as incomplete fusion or collapse. By monitoring the degree of fluctuation in the filling height, it is possible to identify whether there are transient stagnation, rebound, or violent fluctuations in the molten metal during the settling process, thereby evaluating the continuity of penetration. At the same time, the filling ratio of molten metal on the cross-section of the root gap directly determines the forming width and protrusion height of the back weld. By monitoring the average level of the occupancy ratio, the sufficiency of filling can be evaluated. Multiplying the reciprocal of the degree of sinking fluctuation (the smaller the fluctuation, the larger the value) with the average occupancy ratio (the closer to full, the larger the value) achieves the synergistic quantification of the two dimensions of continuous sinking and sufficient filling. Only when the molten metal sinks smoothly and fills sufficiently can the filling be achieved. The value increases significantly, thus providing a precise basis for calculating the compensation factor in conjunction with the stability of the molten pool flow and for achieving coordinated control of molten pool flow and root forming.

[0090] Step S140: Determine the compensation factor based on the stability of the molten pool flow and the consistency index of the root metal sinking and penetration.

[0091] The aforementioned compensation factor is a dimensionless control parameter used to quantify the intensity of voltage compensation requirements during back-side free-forming welding. Through the coupling relationship between the molten pool flow stability and the root metal penetration consistency index, it maps the dual evaluation results of molten pool dynamic behavior and root forming quality into the adjustment range of welding heat input. A larger compensation factor value indicates a greater deviation of the current welding state from the ideal stable condition, requiring increased welding heat input to improve molten pool fluidity and promote sufficient filling of the root gap.

[0092] The purpose of setting the compensation factor in the above scheme is to establish a quantitative feedback bridge between the internal fluid dynamics of the molten pool and external welding process parameters. By coupling the stability of the tail metal backflow (reflecting the continuity of metal transport within the molten pool) and the consistency of root metal sinking and penetration (reflecting the final forming quality) in calculation, critical conditions of unstable tail backflow or discontinuous root sinking can be accurately identified. Under such conditions, fixed welding parameters set solely based on experience are difficult to adapt to the time-varying characteristics of the welding process, while the compensation factor can dynamically quantify the required voltage adjustment range, achieving a closed-loop response from quality monitoring to process control. By setting the compensation factor, the above scheme achieves adaptive control of the special constraints of the thick plate box-shaped structure at the tail end of the high-strength steel telescopic boom. Due to the sudden changes in local stiffness and non-uniform heat dissipation conditions in the tail region, the welding heat input in the root region of the groove is prone to a coupling effect of uneven transverse temperature gradient and dynamic instability of the molten pool. The compensation factor evaluates the synergistic state of molten pool flow and root formation in real time. When the risk of molten pool instability or poor root filling is detected, the voltage compensation mechanism is automatically triggered. By enhancing the arc heat input, the molten pool temperature is increased and the fluidity of the molten metal is improved, making the tail metal reflow more stable. At the same time, it promotes the continuous sinking and full penetration of the root metal, thereby suppressing the occurrence of transient collapse or local retention, and ensuring a stable and consistent back free-forming welding quality along the weld direction.

[0093] The compensation factor can be determined in at least one of the following ways according to the embodiments of this application:

[0094] The first approach: Determining the negative correlation mapping strategy based on product coupling;

[0095] Optionally, step S140 may include: calculating the product of the molten pool flow stability and the root metal sinking penetration consistency index; determining a compensation factor based on the product; wherein the compensation factor is negatively correlated with the product, and the smaller the product, the larger the compensation factor.

[0096] The above compensation factors The calculation method is as follows:

[0097] ;

[0098] The above formula utilizes the monotonically decreasing characteristic of the exponential function to achieve a negative correlation between the compensation factor and the product of the molten pool flow stability and the root metal sinking penetration consistency index; when the molten pool tail recirculation is stable and the root metal sinking penetration consistency is good ( (product is large), compensation factor Approaching zero indicates that no additional compensation is needed; when the molten pool flow is unstable or the root filling is poor ( (small product), compensation factor A value approaching 1 indicates the need for maximum voltage compensation. In the formula, This represents the compensation factor, which is a dimensionless parameter. It represents an exponential function with the natural constant as the base, which acts as an inverse proportional normalization function, ensuring that the compensation factor increases exponentially as the product decreases.

[0099] It is understandable that the calculation of the compensation factor based on the product can be implemented using various mathematical mapping forms. One implementation method uses an exponential function form, that is, directly using the above formula. One approach is to calculate a smooth transition, providing a gradually increasing compensation amount as the welding state approaches critical stability, avoiding abrupt changes in the compensation value. Another approach is to use a reciprocal function, which provides a larger compensation factor when the product is small, but the curve slope is steeper. Yet another approach is to use a linear negative correlation, which is easy to calculate. A third approach is to use a piecewise function, setting a product threshold to divide the welding state into a stable region, a transition region, and an unstable region, and using different compensation factor calculation strategies in different regions to adapt to specific process window requirements.

[0100] The second approach: Determining the evaluation strategy based on weighted synthesis;

[0101] Considering that the influence weights of molten pool flow stability and root formation consistency on the final welding quality may differ under different welding conditions, corresponding weight coefficients are assigned to the molten pool flow stability and root metal sinking and penetration consistency indicators, respectively, and a comprehensive stability index is calculated by weighted summation; then, a compensation factor is determined based on this comprehensive stability index, wherein the lower the comprehensive stability index, the larger the compensation factor, so as to achieve differentiated compensation control for different process sensitivities.

[0102] The third approach: Determining the minimum value based on bottleneck identification;

[0103] The smaller value between the stability of the molten pool flow and the consistency of root metal penetration is taken as the control bottleneck parameter of the current welding process. Based on this minimum value, a compensation factor is determined. When either index decreases significantly (the minimum value decreases), the compensation factor is increased accordingly to ensure that sufficient voltage compensation can be triggered when instability occurs in either the molten pool flow or the root forming, so as to avoid the overall welding quality deterioration caused by a single defect.

[0104] The fourth approach: Determining the rule-based reasoning strategy based on fuzzy logic;

[0105] A fuzzy rule base is established, which includes rules such as "the compensation factor is small if the molten pool flow is stable and the root penetration is consistent" and "the compensation factor is large if the molten pool flow is unstable or the root penetration is poor". The stability of the molten pool flow and the consistency index of the root metal sinking and penetration are used as input variables for fuzzification processing. The fuzzy output value of the compensation factor is calculated by fuzzy inference engine.

[0106] Step S150: Adjust the welding arc voltage according to the compensation factor.

[0107] The aforementioned welding arc voltage refers to the potential difference applied between the end of the welding wire (or tungsten electrode, etc.) and the workpiece during the welding process. Its physical essence is the product of the electric field strength required to maintain arc discharge and the arc length, characterizing the energy state and stability of the arc plasma column. In arc welding, this voltage, along with the welding current, determines the arc's output power (i.e., heat input intensity), directly affecting the temperature field distribution of the molten pool, the fluidity of the molten metal, the penetration depth, and the weld formation quality. In this embodiment, the welding arc voltage, as a core controlled process parameter, determines the arc's heat input level and the dynamic behavior of the molten pool. Through adaptive adjustment based on compensation factors, the voltage can be increased in real time to improve heat input when molten pool flow instability or poor root filling is detected. This improves molten pool fluidity, promotes uniform distribution and stable penetration of molten metal into the root gap, and ensures the quality and consistency of the back-side free-form welding.

[0108] Optionally, step S150 may include: obtaining the set voltage value and the maximum voltage compensation range allowed by the system under normal and stable welding conditions; determining the voltage compensation amount based on the compensation factor and the maximum voltage compensation range; and superimposing the set voltage value and the voltage compensation amount to determine the adjusted welding arc voltage.

[0109] The aforementioned set voltage value refers to the reference arc voltage value pre-calibrated for specific welding materials, plate thicknesses, and joint types under normal and stable welding conditions. It characterizes the energy level required to maintain stable arc combustion, ensure good molten pool fluidity, and achieve sufficient root penetration under ideal working conditions. The aforementioned maximum voltage compensation range refers to the upper limit of safe adjustment allowed by the welding power supply system or process specifications. It is used to limit the maximum value of voltage increment during dynamic compensation, preventing quality defects such as welding spatter, burn-through, undercut, or overheating of weld metal caused by excessive heat input. The voltage compensation amount refers to the dynamic adjustment range that needs to be superimposed on the set reference value, calculated based on the degree to which the current welding process deviates from the ideal stable state (i.e., the compensation factor). The adjusted welding arc voltage is the real-time voltage command value actually used to drive the welding arc, which is linearly superimposed from the set voltage value and the voltage compensation amount.

[0110] The above-mentioned adjusted welding arc voltage The calculation method is as follows:

[0111] ;

[0112] The above formula maps the dimensionless compensation factor to the actual voltage adjustment and performs limited incremental compensation based on the reference setting, ensuring that the welding heat input is adaptively enhanced within a safe and controllable range. Where, This represents the adjusted welding arc voltage, which is the actual control parameter output to the welding power source to drive the arc. This represents the set voltage value under normal and stable welding conditions, serving as a reference value for regulation. This represents the maximum permissible voltage compensation range in the welding system, serving as a safety constraint to prevent overcompensation; product term This refers to the voltage compensation amount, which represents the voltage increase based on the actual needs determined according to real-time operating conditions.

[0113] Optionally, the above-mentioned back free-forming welding method for the tail end of a high-strength steel telescopic boom may further include: dividing the effective monitoring time period of the welding process into several observation time windows, and repeatedly executing the data acquisition step, the molten pool stability determination step, the root metal sinking and penetration consistency index determination step, the compensation factor determination step, and the welding arc voltage adjustment step within each observation time window, so as to dynamically update the welding arc voltage along the weld direction.

[0114] In the implementation of the above scheme, the effective monitoring period of the welding process (i.e., the stage from the stable formation of the welding arc to the completion of root penetration) is divided into several equal-length or adaptive observation time windows along the longitudinal direction of the weld. Each window corresponds to a specific welding position interval. Within each observation time window, multi-source monitoring data of that position interval are collected synchronously, and the steps of calculating the stability of the molten pool flow, evaluating the consistency index of root metal sinking and penetration, determining the compensation factor, and adjusting the welding arc voltage are executed in sequence to complete the identification of the welding state and the correction of process parameters for that position interval. Then, the next observation time window is entered, and the above data collection and control process is repeated to achieve continuous segment-by-segment optimization along the weld direction. The above scheme constructs a real-time feedback control loop through iterative iteration to ensure the dynamic stability of welding quality: the welding arc voltage adjusted in the previous window serves as the initial process parameter for the subsequent window, and the control results in each window are accumulated and superimposed along the longitudinal direction of the weld to form a continuous voltage waveform that adapts to the time-varying characteristics of the welding process; since the control of each window is based on the monitoring data acquired in real time within that interval rather than preset fixed parameters, the system can respond in a timely manner to deviations in the welding state caused by sudden changes in local stiffness, changes in heat dissipation conditions, or fluctuations in material properties, and suppress the generation and expansion of transient defects through adaptive voltage adjustment.

[0115] It is understandable that due to the strong time-varying and spatial non-uniformity of the welding process of the thick plate box-shaped structure at the tail end of the high-strength steel telescopic boom, there are local abrupt changes in stiffness and differences in heat dissipation conditions along the weld direction. This causes the heat input distribution, molten pool behavior, and root forming state to continuously change with the welding position. A single fixed welding parameter cannot adapt to this spatial heterogeneity. By discretizing the continuous welding process into observation time windows and implementing segmented control, the process parameters are precisely adapted to local working conditions. This avoids excessive heat input in areas with good heat dissipation or insufficient heat input in areas with poor heat dissipation, ensuring a consistent penetration depth and back forming quality along the entire weld length.

[0116] Optionally, the above-mentioned back free-forming welding method for the tail end of a high-strength steel telescopic boom may further include: determining the effective monitoring time period of the welding process based on the starting conditions for the stable formation of the welding arc and the termination conditions for the completion of root penetration; wherein, the starting conditions include the welding current and voltage reaching a set stable value, and the termination conditions include the tail shape change of the molten pool tending to be stable or the arc electrical signal returning to a stable value.

[0117] The aforementioned activation conditions are determined based on the physical state of stable welding arc formation, specifically through real-time monitoring of welding current and welding voltage. When the arc signal acquisition module detects that both welding current and welding voltage have reached preset stable values ​​and remain within a stable fluctuation range for a set duration, it determines that the welding arc has entered a stable combustion state, triggering the start of the effective monitoring period. This mechanism effectively avoids abnormal data generated during the arc ignition stage due to insufficient arc establishment and drastic fluctuations in current and voltage, ensuring that all data entering the observation time window originates from a welding process with stable heat input, providing a reliable data foundation for calculating parameters such as the transverse temperature gradient offset index.

[0118] The termination conditions mentioned above are based on the process node of root penetration completion, indicating that the molten metal has completely filled the root gap and the welding process has entered the final stage. Specific criteria include: the tail-end morphology of the weld pool stabilizing, i.e., the molten metal flow at the tail end is stagnant, the solidification front is stable, or the weld pool geometry no longer changes significantly, as identified by the weld pool visual monitoring device; or the arc electrical signal recovering to a stable value, i.e., the welding current and welding voltage returning to a background level similar to that before arc initiation or a preset termination threshold. Once any of the above conditions are met, the effective monitoring period is considered over, and data acquisition and voltage adjustment for that welding cycle are stopped. This mechanism effectively eliminates invalid information generated during the post-weld cooling stage due to weld pool solidification and changes in heat conduction patterns without molten metal flow, avoiding unnecessary voltage compensation for solidified welds.

[0119] The determination of the aforementioned effective monitoring time period is a dynamic process based on real-time signal monitoring and logical judgment. The arc signal acquisition module continuously monitors the instantaneous values ​​of welding current and voltage and compares them in real-time with the start-up threshold. Simultaneously, a visual monitoring device analyzes the geometric feature changes of the molten pool tail image sequence or monitors the attenuation trend of the arc signal. This judgment logic based on multi-source signal fusion ensures accurate identification of the time period boundaries, preventing missed effective data due to arc ignition delays and avoiding omissions of the final stage of root forming due to premature termination. By strictly defining the effective monitoring time period, precise focus on monitoring resources and control actions is achieved, preventing misjudgments caused by arc instability during the arc ignition stage and interference from the solidification process during the cooling stage on the assessment of molten pool flow stability. This ensures that the calculation of the compensation factor and voltage adjustment only apply to the active molten pool stage that has a substantial impact on the quality of the back free forming, significantly improving data quality, calculation reliability, and the accuracy of dynamic control.

[0120] To facilitate understanding of the working principle of the above-described free-form welding method for the tail end of a high-strength steel telescopic boom, this application embodiment also provides a specific application example of this method in a certain application scenario. In this application scenario, the free-form welding method for the tail end of a high-strength steel telescopic boom mainly includes:

[0121] Step 1: Acquire multi-source monitoring data;

[0122] Before performing freeform welding on the back side, basic process parameters such as welding current, voltage, welding speed, and shielding gas flow rate are set. An arc signal acquisition module, a molten pool visual monitoring device, and a temperature monitoring sensor are arranged near the welding torch to construct a data acquisition system for monitoring the welding process. During welding, welding current and voltage signals are acquired in real time through an arc signal acquisition module, and arc voltage fluctuations, current changes, and instantaneous power changes are recorded at a set sampling frequency. The welding equipment's built-in encoder or speed sensor monitors the displacement rate of the welding torch relative to the workpiece in real time, acquiring welding speed data. A certain number of thermocouples are arranged at the bottom of the bevel to form a root temperature monitoring network, and a certain number of thermocouples are arranged on each side of the left and right edges of the bevel to cover the thickness direction of the base material, acquiring temperature data of the root region of the bevel and the base material on both sides in real time. An industrial camera positioned to the side or rear of the welding torch continuously images the weld pool area, acquiring image data of the weld pool's geometry, length, width, and tail metal flow state in real time. A back-side visual monitoring device or laser rangefinder is arranged on the back of the weld at the tail end of the telescopic arm to acquire real-time data on the filling height and area of ​​molten metal in the root gap during welding. The above arc signal data, weld pool visual data, temperature field data, and root filling data are synchronously recorded according to a unified time reference, forming a multi-source monitoring data set for the corresponding welding position.

[0123] Step 2: Determine the stability of the molten pool flow based on temperature data, welding heat input parameters, and molten pool tail metal reflow status data;

[0124] It is understandable that, since the tail end structure of high-strength steel telescopic boom is usually a thick plate box structure, the local stiffness difference and heat dissipation condition difference can easily lead to uneven accumulation of welding heat in the transverse direction. By analyzing the temperature gradient change, it is possible to identify the heat concentration or loss area that may occur during the welding process in advance, and provide a reliable basis for judging the melting state and flow conditions of the molten pool in the root area of ​​the bevel.

[0125] During welding, a certain number of thermocouples (adjustable according to the bevel width and thickness) are placed at the bottom of the bevel to form a root temperature monitoring network; a certain number of thermocouples are placed on each side of the left and right edges of the bevel, covering the thickness direction of the base material; thus, the lateral temperature gradient shift index can be determined based on the temperature difference between the base material on both sides of the bevel and the root region. The lateral temperature gradient shift index within the current observation time window is calculated as follows:

[0126] ;

[0127] The aforementioned transverse temperature gradient offset index is used to quantify the unevenness of heat distribution in the bevel section. The larger the value, the greater the deviation between the root temperature and the temperature of the base material on both sides, and the easier it is for the molten pool root to stagnate or collapse.

[0128] Subsequently, based on the lateral temperature gradient offset index and welding heat input parameters, the degree of lateral heat input imbalance was determined:

[0129] ;

[0130] The degree of lateral imbalance of heat input is used to quantify the degree of imbalance in the lateral distribution of welding heat input. The larger the value, the higher the heat input and the uneven lateral distribution, and the higher the risk of molten pool root retention or transient collapse.

[0131] Furthermore, based on the degree of lateral imbalance in heat input and the asymmetry of the lateral expansion distance of the molten pool front, the degree of molten pool front expansion offset is determined:

[0132] ;

[0133] The degree of expansion and offset of the molten pool front is used to quantify the actual impact of heat input imbalance on the molten pool morphology. The larger the value, the greater the lateral imbalance of heat input and the more asymmetrical the expansion of the molten pool front.

[0134] Finally, based on the degree of expansion and offset of the molten pool front and the fluctuation characteristics of the metal reflow velocity at the tail of the molten pool, the stability of the metal reflow at the tail (i.e., the stability of the molten pool flow) is determined:

[0135] ;

[0136] The stability of the tail metal reflow is used to quantify the stability of the tail metal reflow in the molten pool. The larger the value, the higher the stability of the tail metal reflow in the molten pool and the more ideal the forming quality.

[0137] Step 3: Based on the root gap filling status data, determine the root metal sinking penetration consistency index;

[0138] First, the filling height sequence and occupancy ratio sequence of molten metal in the root gap are obtained, where the occupancy ratio is the ratio of the instantaneously occupied area of ​​molten metal to the theoretically filled area of ​​the root region. Based on the degree of fluctuation in the height difference between adjacent time points in the filling height sequence, the degree of sinking fluctuation is determined; based on the average level of the occupancy ratio sequence, the degree of filling adequacy is determined.

[0139] Then, the root metal penetration consistency index is calculated:

[0140] ;

[0141] The root metal sinking penetration consistency index is used to quantify the consistency of root gap filling. The larger the value, the more stable and sufficient the molten metal sinks, that is, the higher the root metal sinking penetration consistency.

[0142] Step 4: Determine the compensation factor based on the molten pool flow stability and the root metal sinking penetration consistency index. Calculate the molten pool flow stability (i.e., the tail metal backflow stability). Consistency index with root metal sinking and penetration The product of these factors is used to determine the compensation factor.

[0143] ;

[0144] The compensating factor's stability with tail metal reflux Consistency index with root metal sinking penetration The product decreases and increases; when the tail recirculation is unstable or the root sinking is discontinuous, the compensation factor increases to indicate the need for increased heat input.

[0145] Step 5: Adjust the welding arc voltage according to the compensation factor;

[0146] Obtain the set voltage value under normal and stable welding conditions and the maximum allowable voltage compensation amplitude of the system The voltage compensation amount is determined based on the compensation factor and the maximum voltage compensation amplitude. The set voltage value is then superimposed with the voltage compensation amount to determine the adjusted welding arc voltage. This adjustment increases the welding arc voltage as the compensation factor increases, thereby enhancing the welding heat input, improving the fluidity of the molten pool, stabilizing the tail metal reflow, and promoting root metal penetration.

[0147] Step 6: Implement dynamic closed-loop control;

[0148] The effective monitoring period for the welding process is determined based on the starting conditions for the stable formation of the welding arc and the termination conditions for the completion of root penetration. The starting conditions include the welding current and voltage reaching set stable values, and the termination conditions include the stabilization of the molten pool tail morphology or the recovery of the arc electrical signal to a stable value, thus eliminating invalid information during the post-weld cooling phase. This effective monitoring period is divided into several observation time windows. Within each observation time window, steps one through five are repeated to dynamically update the welding arc voltage along the weld direction. This iterative process ensures the continuity and consistency of the entire weld along the weld direction.

[0149] This invention is now complete.

[0150] In summary, in this embodiment of the invention, during the welding process, temperature data of the root region of the groove and the base materials on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data are acquired. Based on the temperature data, welding heat input parameters, and molten pool tail metal reflow state data, the molten pool flow stability is determined; wherein, the molten pool flow stability is used to quantify the stability of the molten pool tail metal reflow. Based on the root gap filling state data, a root metal sinking and penetration consistency index is determined; wherein, the root metal sinking and penetration consistency index is used to quantify the root gap filling consistency. Based on the molten pool flow stability and the root metal sinking and penetration consistency index, a compensation factor is determined; and the welding arc voltage is adjusted according to the compensation factor.

[0151] This invention constructs a two-dimensional evaluation system for the stability of the molten pool flow and the consistency index of root metal sinking and penetration. Based on the coupling relationship between the two, a compensation factor is determined to dynamically adjust the welding arc voltage. This achieves the coordinated control of molten pool flow stability and root metal forming behavior, effectively solving the coupling effect problem of uneven lateral temperature gradient and dynamic flow instability of the molten pool caused by abrupt stiffness changes and uneven heat dissipation in the thick plate box-shaped structure at the tail end of a high-strength steel telescopic boom. On the other hand, by establishing a multi-level quantitative characterization model of lateral temperature gradient offset index, lateral imbalance of heat input, and molten pool leading edge expansion offset, the welding heat input distribution, molten pool geometry evolution, and tail metal return behavior are transformed into calculable stability factors. The system employs a cyclical iterative mechanism with observation time windows to collect multi-source monitoring data in real time within each window and dynamically update the compensation factor and welding arc voltage, forming a continuous closed-loop control along the weld direction. This overcomes the limitations of traditional empirical parameter control, which is unable to adapt to the time-varying characteristics of the welding process. Furthermore, by adaptively adjusting the welding arc voltage through the compensation factor, the system enhances heat input to improve the fluidity of the molten pool when the tail metal backflow is unstable or the root metal sinking is discontinuous. This suppresses the transient collapse and local retention of molten metal in the root gap, improving the consistency of root penetration and the quality of free forming on the back side of the weld.

[0152] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for free-form welding of the back end of a high-strength steel telescopic boom, characterized in that, The method includes: During the welding process, temperature data of the root region of the groove and the base material on both sides, welding heat input parameters, metal reflow status data of the tail of the molten pool, and root gap filling status data are obtained. Based on the temperature data, the welding heat input parameters, and the molten pool tail metal reflow status data, the molten pool flow stability is determined; wherein, the molten pool flow stability is used to quantify the stability of the molten pool tail metal reflow. The step of determining the flow stability of the molten pool based on the temperature data, the welding heat input parameters, and the molten pool tail metal reflow state data includes: Based on the temperature data, a lateral temperature gradient offset index is determined; wherein, the lateral temperature gradient offset index is used to quantify the unevenness of heat distribution in the bevel section. The formula for calculating the lateral temperature gradient offset index is: ； In the formula, Indicates the lateral temperature gradient offset index; and These represent the average monitored temperatures of the parent material on the left and right sides of the bevel, respectively. This indicates the average temperature in the root region at the bottom of the bevel. This represents a non-zero constant, with a value of 0.0001. Based on the lateral temperature gradient offset index and the welding heat input parameters, the degree of lateral imbalance of heat input is determined; wherein, the degree of lateral imbalance of heat input is used to quantify the degree of imbalance in the lateral distribution of welding heat input. The formula for calculating the degree of lateral imbalance in heat input is: ； In the formula, Indicates the degree of lateral imbalance in heat input; This represents the average welding voltage within the current observation window; This represents the average welding current within the current observation window; Indicates the average welding speed; Represents the hyperbolic tangent function; Based on the degree of lateral imbalance of heat input and the degree of asymmetry of the lateral expansion distance of the molten pool front, the degree of molten pool front expansion offset is determined; wherein, the degree of molten pool front expansion offset is used to quantify the actual impact of heat input imbalance on the molten pool morphology; The formula for calculating the degree of expansion and offset of the molten pool front is: ； In the formula, Indicates the degree of expansion and offset of the molten pool front; This indicates the first lateral extension distance between the left leading edge boundary of the molten pool and the weld centerline; This indicates the second lateral extension distance between the right leading edge boundary of the molten pool and the weld centerline; This represents a non-zero constant, with a value of 0.0001. Based on the degree of expansion and offset of the molten pool front and the fluctuation characteristics of the metal reflow velocity at the tail of the molten pool, the flow stability of the molten pool is determined; the formula for calculating the flow stability of the molten pool is: ； In the formula, Indicates the stability of the molten pool flow; This represents the number of sampling points within the observation window. and They represent the first The and the first Monitoring values ​​of tail metal reflux velocity at each sampling point; It is a non-zero constant, with a value of 0.0001; Based on the root gap filling status data, a root metal sinking and penetration consistency index is determined; wherein, the root metal sinking and penetration consistency index is used to quantify the root gap filling consistency. The step of determining the root metal sinking penetration consistency index based on the root gap filling state data includes: Obtain the filling height sequence and occupancy ratio sequence of the molten metal in the root gap; wherein, the occupancy ratio is the ratio of the instantaneously occupied area of ​​the molten metal to the theoretically filled area of ​​the root region; The degree of subsidence fluctuation is determined based on the degree of change in the height difference between adjacent times in the filling height sequence; The degree of filling adequacy is determined based on the average level of the occupancy ratio sequence; Based on the degree of subsidence fluctuation and the degree of filling adequacy, a root metal subsidence penetration consistency index is determined; wherein, the smaller the degree of subsidence fluctuation and the higher the degree of filling adequacy, the higher the consistency index; the calculation formula for the root metal subsidence penetration consistency index is: ； In the formula, This indicates the consistency index of metal penetration at the root; and They represent the first The and the first Normalized value of the molten metal filling height in the root gap at each sampling time; This represents a non-zero constant, with a value of 0.0001. This indicates the proportion of the root gap that is instantaneously filled by molten metal; This indicates the average level of the proportion sequence; Based on the stability of the molten pool flow and the consistency index of the root metal sinking and penetration, a compensation factor is determined, specifically including: Calculate the product of the molten pool flow stability and the root metal sinking penetration consistency index; Based on the product, a compensation factor is determined; wherein the compensation factor is negatively correlated with the product, and the smaller the product, the larger the compensation factor. The formula for calculating the compensation factor is as follows: ； In the formula, Indicates the compensation factor; Represents an exponential function with the natural constant as its base; Adjusting the welding arc voltage according to the compensation factor specifically includes: Obtain the set voltage value and the maximum voltage compensation range allowed by the system under normal and stable welding conditions; The voltage compensation amount is determined based on the compensation factor and the maximum voltage compensation amplitude, and the set voltage value is superimposed with the voltage compensation amount to determine the adjusted welding arc voltage. The formula for calculating the adjusted welding arc voltage is as follows: ； In the formula, This indicates the adjusted welding arc voltage; This indicates the set voltage value under normal and stable welding conditions; This indicates the maximum allowable voltage compensation range in the welding system.

2. The back free-forming welding method for the tail end of a high-strength steel telescopic boom according to claim 1, characterized in that, The method for determining the degree of asymmetry in the lateral expansion distance of the molten pool front includes: The positions of the weld centerline, the left leading edge boundary of the molten pool, and the right leading edge boundary of the molten pool are determined; wherein, the position of the weld centerline is determined based on the direction of the welding torch. Calculate the first lateral extension distance between the left leading edge boundary of the molten pool and the center line of the weld, and the second lateral extension distance between the right leading edge boundary of the molten pool and the center line of the weld; The degree of asymmetry in the lateral expansion distance of the molten pool front is determined based on the absolute difference between the first lateral expansion distance and the second lateral expansion distance; wherein, the larger the absolute difference, the higher the degree of asymmetry.

3. The back free-forming welding method for the tail end of a high-strength steel telescopic boom according to any one of claims 1 and 2, characterized in that, The acquisition of temperature data of the root region of the bevel and the base metal on both sides, welding heat input parameters, molten pool tail metal reflow state data, and root gap filling state data includes: Obtain welding thermal input parameters; wherein, the welding thermal input parameters include welding current, welding voltage, and welding speed; Acquire an image sequence of the tail region of the molten pool, and determine the metal reflow velocity at the tail of the molten pool based on the image sequence; Obtain root gap filling status data; wherein, the root gap filling status data includes the filling height and occupancy ratio of molten metal in the root gap; the occupancy ratio is the ratio of the instantaneously occupied area of ​​the molten metal to the theoretically filled area of ​​the root region; Obtain temperature data of the root region of the bevel and the parent material on both sides.

4. The back free-forming welding method for the tail end of a high-strength steel telescopic boom according to claim 3, characterized in that, The step of acquiring an image sequence of the tail region of the molten pool and determining the metal reflow velocity at the tail of the molten pool based on the image sequence includes: The tail region of the molten pool is continuously imaged to obtain the image sequence; The flow trajectory of the molten metal at the tail end in the image sequence is tracked using an image recognition algorithm; The reflux velocity of the molten metal at the tail of the molten pool is calculated based on the change in the position of the molten metal flow in adjacent image frames.

5. The back free-forming welding method for the tail end of a high-strength steel telescopic boom according to any one of claims 1 and 2, characterized in that, The method further includes: The effective monitoring period of the welding process is divided into several observation time windows. Within each observation time window, the data acquisition step, the molten pool stability determination step, the root metal sinking and penetration consistency index determination step, the compensation factor determination step, and the welding arc voltage adjustment step are repeatedly executed to dynamically update the welding arc voltage along the weld direction.

6. The back free-forming welding method for the tail end of a high-strength steel telescopic boom according to claim 5, characterized in that, The method further includes: The effective monitoring time period of the welding process is determined based on the starting conditions for the stable formation of the welding arc and the termination conditions for the completion of root penetration; wherein, the starting conditions include the welding current and voltage reaching a set stable value, and the termination conditions include the tail shape of the molten pool stabilizing or the arc electrical signal returning to a stable value.

Images