Copper / stainless steel pipeline adaptive intelligent welding control method and system based on dynamic electrical parameter feedback

Abstract

The application provides a copper / stainless steel pipeline adaptive intelligent welding control system and method based on dynamic electrical parameter feedback, which comprises the following steps: continuously collecting welding electrical parameter data output by a power supply in a welding process to construct a welding state characteristic quantity; calculating a stable margin index and determining a maximum parameter change amplitude of parameter adjustment for a next welding cycle to generate a recoverability constraint condition; constructing a plurality of candidate parameter vectors and a candidate welding parameter set based on reference welding parameters of a current welding cycle; identifying a candidate parameter vector exceeding the risk index as unusable to construct an unusable welding parameter set; eliminating the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; selecting a target welding parameter in the executable candidate set and issuing the target welding parameter to a welding execution mechanism to complete welding of the next welding cycle. The application improves the first-time pass rate of heterogeneous pipeline welding and stabilizes the production rhythm.

Description

Technical Field

[0001] This invention belongs to the field of adaptive intelligent welding of copper / stainless steel pipelines, and particularly relates to an adaptive intelligent welding control method and system for copper / stainless steel pipelines based on dynamic electrical parameter feedback. Background Technology

[0002] In the air conditioning and refrigeration equipment manufacturing industry, dissimilar material welding of copper and stainless steel pipes is widely used in high-pressure pipelines, heat exchange circuits, and complex connection parts. Due to the significant differences in thermal conductivity and melting behavior between the two materials, the available heat input window for the welding process is narrow, and the weld quality is highly sensitive to changes in process parameters. Existing production lines generally adopt the preset parameter process of inverter welding power supply. Before welding, parameters such as pulse waveform, wire feed, and welding speed are determined based on experience or experiments, and the welding process is executed in a fixed or simple segmented manner. This method can meet the basic requirements when the assembly accuracy is stable and the gap consistency is good. However, in actual pipeline production, the welding process exhibits obvious non-steady-state characteristics due to factors such as batch differences in pipe materials, fluctuations in assembly gaps and misalignments, deviations in pipe end roundness, and changes in on-site heat dissipation conditions. Fixed parameters are difficult to maintain stable formation under different operating conditions, and defects such as insufficient wetting, lack of fusion, molten pool fluctuations, or local overheating are prone to occur. Some existing technologies introduce real-time monitoring of arc voltage or welding current and correct parameters after detecting deviations. However, most solutions still use instantaneous electrical parameter deviations as adjustment triggers, lacking systematic utilization of the evolution characteristics of electrical parameters over time segments. They also lack a constraint mechanism for the risk of "parameter change amplitude and linkage direction in the next cycle". This makes it easy for adjustment lag, overcompensation, or multiple parameters to be superimposed and amplified when the working conditions change suddenly, thus pushing the welding process into an unstable state that is difficult to recover.

[0003] Especially under the narrow window conditions of welding dissimilar copper / stainless steel pipes, the combined changes in process parameters are more likely to induce irreversible defects than single parameter deviations. Existing solutions lack operable parameter exclusion boundaries for such linkage risks, making it difficult to simultaneously ensure first-pass yield and production line cycle stability. Summary of the Invention

[0004] The purpose of this invention is to propose an adaptive intelligent welding control method and system for copper / stainless steel pipelines based on dynamic electrical parameter feedback, thereby solving the above-mentioned problems.

[0005] To achieve the above objectives, a first aspect of the present invention provides an adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback, the method comprising the following steps: S1. Continuously collect welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and construct welding state characterization quantities by combining the corresponding time segments. S2. Based on the welding state characterization quantity of the current time segment, calculate the stability margin index, which is used to measure the stability of the welding process; according to the stability margin index, determine the maximum parameter change range for parameter adjustment in the next welding cycle, and generate recoverability constraints; wherein, the recoverability constraints include the upper limit of parameter change range and adjustment direction information; S3. Based on the baseline welding parameters of the current welding cycle, and combined with the preset discrete template, construct multiple candidate parameter vectors and a set of candidate welding parameters; based on the recoverability constraint and the coupling sensitive direction vector, calculate the corresponding risk index for each candidate parameter vector, and identify the candidate parameter vectors whose risk index exceeds the limit as unusable, and construct an unusable welding parameter set. S4. Remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, select the candidate parameter vector with the smallest change range from the current reference welding parameter as the target welding parameter, and send it to the welding execution mechanism to complete the welding of the next welding cycle.

[0006] Furthermore, S1 specifically includes: Continuously collect welding electrical parameter data output by the power supply during the welding process; The continuously acquired welding electrical parameter data is divided into several continuous time segments according to a fixed time length. Each time segment covers multiple electrical parameter sampling points. The time length of the corresponding time segment is set in association with the control cycle or pulse cycle of the welding power source, so that each time segment contains complete welding behavior characteristics. For each time segment, the arc voltage sequence and welding current sequence are jointly processed to extract combined features reflecting the continuity of energy input and arc stability. The combined features are scaled and range-mapped to maintain a consistent expression scale under different pipe diameters, material batches, or process conditions. Then, the feature combination within the segment is output as a welding state characterization quantity.

[0007] Furthermore, the combined features are features used to reflect the smoothness of the changes in arc voltage and welding current over time, the persistence of abrupt changes, and the correspondence between the changes in the two.

[0008] Furthermore, the calculation of the stability margin index based on the welding state characterization quantity of the current time segment specifically includes: The importance of the welding state characterization quantities is weighted by a weight vector to generate a weighted summation result; Based on the weighted summation result, a smooth nonlinear mapping function is introduced to compress it, generating a compressed result; Based on the compression results, a stability margin index is calculated by combining an additional term based on the overall amplitude of the welding state characterization quantity; the additional term based on the overall amplitude of the state is used to accelerate the constraint tightening speed when the state deviation increases.

[0009] Furthermore, the step of determining the maximum parameter change range for parameter adjustment in the next welding cycle based on the stability margin index, and generating recoverability constraints, specifically involves: Based on the stability margin index, the maximum allowable variation range is scaled using a power function to generate the maximum allowable parameter variation range within the central region of the welding process window. The recoverability constraints are generated by combining the maximum parameter change magnitude and the adjustment direction information; wherein, the adjustment direction information is determined by the weighted summation result.

[0010] Furthermore, the construction of multiple candidate parameter vectors and a candidate welding parameter set based on the baseline welding parameters of the current welding cycle and a preset discrete template is specifically as follows: During the debugging phase, the controller writes several sets of neighborhood adjustment templates, where each set of templates is composed of a finite number of relative change levels; The receiver saves the reference welding parameter vector of the previous cycle or the current cycle, and adds the relative change given by the neighborhood adjustment template to the reference welding parameter vector to generate a candidate parameter vector, while obtaining the candidate change.

[0011] Furthermore, the risk index of the candidate parameter is calculated and generated based on the candidate change amount, the maximum parameter change amplitude, and the coupling sensitive direction vector; the coupling sensitive direction vector is pre-configured according to the process test results. Candidate parameters with values ​​greater than 1 are recorded as unavailable and included in the set of unavailable welding parameters.

[0012] Furthermore, the magnitude of the change is measured using Euclidean distance to prioritize the most stable parameter adjustment scheme while satisfying the constraints.

[0013] Furthermore, in step S4, the target welding parameters include at least the normalized adjustment amount of the corresponding welding power pulse channel, the normalized adjustment amount of the wire feed drive channel, and the normalized adjustment amount of the welding motion channel.

[0014] A second aspect of the invention provides an adaptive intelligent welding control system for copper / stainless steel pipelines based on dynamic electrical parameter feedback, the system comprising: The electrical parameter acquisition module is used to continuously acquire welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and combine them with the corresponding time segments to construct welding state characterization quantities. The constraint generation module is used to calculate a stability margin index based on the welding state characterization quantity of the current time segment. The stability margin index is used to measure the stability of the welding process. Based on the stability margin index, the module determines the maximum parameter change range for parameter adjustment in the next welding cycle and generates recoverable constraints. The recoverable constraints include the upper limit of the parameter change range and adjustment direction information. The parameter filtering module is used to construct multiple candidate parameter vectors and a set of candidate welding parameters based on the baseline welding parameters of the current welding cycle and a preset discrete template; based on the recoverability constraint and the coupling sensitive direction vector, it calculates the corresponding risk index for each candidate parameter vector, and identifies the candidate parameter vectors whose risk index exceeds the limit as unusable, thus constructing a set of unusable welding parameters. The execution control module is used to remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, the candidate parameter vector with the smallest change from the current reference welding parameter is selected as the target welding parameter and sent to the welding execution mechanism to complete the welding of the next welding cycle.

[0015] The beneficial technical effects of the present invention are at least as follows: This invention proposes a dynamic electrical parameter-driven control scheme for copper / stainless steel pipeline welding. Its core lies in organizing the continuously acquireable electrical parameter time series from the welding power source into welding state characterization quantities formed by time segments. Based on these state characterization quantities, recoverable constraints for parameter adjustment in the next welding cycle are generated. These constraints are further implemented as availability judgment and elimination rules for candidate welding parameter combinations, thereby identifying and eliminating unsuitable parameter linkage schemes before execution. On this basis, target welding parameters are determined from the remaining executable candidates and sent to the welding power source, wire feeding mechanism, and welding motion mechanism to complete the next welding cycle, ensuring that parameter updates follow recoverable constraints and maintain smooth execution. Through this mechanism, this invention elevates welding process control from relying on instantaneous deviation correction to a periodic adjustment process centered on state evolution, recoverable constraints, and parameter combination elimination. This can suppress the instability risk caused by unfavorable linkage adjustments under typical operating conditions such as assembly gap fluctuations and changes in heat dissipation conditions, improving the first-pass yield of dissimilar pipeline welding and stabilizing production cycle time. Attached Figure Description

[0016] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0017] Figure 1This is a flowchart of the adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback, as described in this invention. Detailed Implementation

[0018] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0019] like Figure 1 As shown in the embodiment of the present invention, an adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback is provided. The method includes: S1. Continuously collect welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and construct welding state characterization quantities by combining the corresponding time segments.

[0020] Specifically, during the welding of copper / stainless steel pipelines, to provide a unified data foundation for subsequent constraint determination and parameter selection, this step transforms the stable electrical parameter changes obtainable from the welding power source side into structured welding process status inputs. The input for this step is the welding electrical parameter data acquired in real-time during the welding process. The welding electrical parameters data The electrical parameter acquisition unit originates from within the welding power source and is integrated into the output circuit. It acquires the arc voltage over time via a voltage sampling circuit and the welding current over time via a current sampling circuit. These two acquisitions are synchronized to form a time series of electrical parameters reflecting the dynamic behavior of the welding process. Due to the differences in thermal conductivity and melting characteristics between copper and stainless steel, the arc state and molten pool behavior fluctuate during welding depending on the nozzle assembly gap, misalignment, and local heat dissipation conditions. The arc voltage and welding current exhibit mixed characteristics, including both trend changes and transient disturbances, over a short timescale. Therefore, this step constructs the state input based on the overall performance within a short time interval to improve the stability and consistency of the process state description.

[0021] In practice, during the welding process, the continuously collected welding electrical parameter data will be... The welding process is divided into several continuous time segments of fixed length, each covering multiple electrical parameter sampling points. The duration of each segment can be set in relation to the control cycle or pulse cycle of the welding power source, ensuring each segment contains relatively complete welding behavior characteristics. For example, under pulsed welding conditions, a time segment can cover one or more pulse cycles, simultaneously reflecting the electrical parameter changes during the peak and baseline phases. For each time segment, the arc voltage sequence and welding current sequence are jointly processed to extract combined features reflecting the continuity of energy input and arc stability. Emphasis is placed on the smoothness of voltage and current changes over time, the persistence of abrupt changes, and the correspondence between these changes. During feature construction, the obtained features are scaled and mapped to maintain a consistent expression scale across different pipe diameters, material batches, or process conditions. Finally, the combined features within each segment are output as a welding state characterization quantity. The relationship is expressed as: ; in, This represents the welding electrical parameter data, which is obtained synchronously by the voltage sampling circuit and the current sampling circuit of the welding power supply output circuit. It represents the welding state characterization quantity, used to describe the overall state of the welding process within a certain time segment and as input for subsequent steps; The state construction process specifically includes time segment division, joint feature extraction of voltage and current, and scaling and interval mapping of the feature results, followed by combined output. For ease of engineering implementation, scaling and interval mapping can be a calculation process driven by a set of preset configuration parameters: for example, setting the configuration parameters as voltage and current reference ranges within a segment, and mapping the observed values ​​within the segment to a standard interval to form a feature representation of a uniform scale; in one embodiment, if the change in arc voltage within a segment falls in the middle to high range of the preset voltage reference range, and the change in welding current falls in the middle to low range of the preset current reference range, after scaling, corresponding standardized feature values ​​can be formed respectively, which are then combined with features reflecting the "consistency of voltage and current changes" to form the feature representation of that segment. This allows subsequent steps to use the same set of constraint logic under different operating conditions. Process it.

[0022] The output of this step is a welding condition characterization value. The welding state characterization quantity This will serve as the input for generating recoverable constraints in the next step. By organizing the continuously acquired electrical parameter data from the welding power source into a structured welding process state input, the subsequent control process can generate constraints and select parameters based on the overall performance of the welding process, thereby supporting stable control of copper / stainless steel pipeline welding under fluctuating operating conditions.

[0023] S2. Based on the welding state characterization quantity of the current time segment, calculate the stability margin index, which is used to measure the stability of the welding process; according to the stability margin index, determine the maximum parameter change range for parameter adjustment in the next welding cycle, and generate recoverability constraints; wherein, the recoverability constraints include the upper limit of parameter change range and adjustment direction information.

[0024] Specifically, the welding state characterization quantity has been obtained in S1. Building upon this foundation, the core task of this step is to transform the state variable into a recoverable constraint that directly constrains the parameter adjustment for the next welding cycle. A common engineering practice in welding control is to map the continuous state to a scalar index representing "safety margin" or "adjustment intensity," and then limit the range of control variable changes based on this index. This step follows this approach, its mathematical foundation derived from the classical control theory method of weighted summation of state variables and obtaining constraint factors through nonlinear mapping. Furthermore, it incorporates engineering experience regarding "state deviation" in welding scenarios, and specifically modifies the original form.

[0025] Specifically, welding condition characterization quantity The stability margin index is constructed from the joint characteristics of arc voltage and welding current within a time segment, as described in step one. Each component reflects the stability and energy continuity of the welding process within that time segment. To compress the multidimensional state into a single index directly usable for constraint generation, a basic form widely used in mathematics and control engineering—linear weighted summation—is first introduced. This involves weighting the importance of each state component using a weight vector. Subsequently, to prevent the weighted result from growing unbounded numerically and causing abrupt constraint changes, a smooth nonlinear mapping function is introduced for compression. This mapping function is derived from the logistic function form in classical mathematics, characterized by gradual changes at small inputs and gradual saturation at larger inputs, making it suitable for describing the "from loose to tight" adjustment pattern. Based on this, and considering the exceptional sensitivity to state deviation in copper / stainless steel dissimilar material welding, an additional term based on the overall amplitude of the welding state characterization is introduced to accelerate constraint tightening as state deviation increases. This yields the stability margin index. The calculation relationship is as follows: ; in, Indicates the characteristic quantity of welding condition The weighted summation result, which is derived from the vector dot product in linear algebra, is used to comprehensively reflect the overall trend of each state component; This is the state weight vector, whose values ​​are determined during the system debugging phase based on typical stable welding states. Configured to reflect the relative impact of different state components on weld recoverability; This is a bias term used to adjust the center position of the weighted result in the nonlinear mapping; Exponential operations form the basis of logical functions; This represents the magnitude index obtained by taking the absolute values ​​of each component of the state characterization quantity and summing them, and is used to describe the overall deviation of the current state from the stable interval. The weighting coefficient for the amplitude addition term is used to adjust the proportion of the influence of the deviation in the stability margin; This is a smoothing factor used to ensure that the amplitude term changes gradually when the deviation is small, and gradually saturates as the deviation increases. Through the above combination, As the welding state changes from stable to unstable, it increases monotonically, and its value range is naturally limited to a finite range, which facilitates subsequent constraint calculations.

[0026] Furthermore, in obtaining the stability margin index This step further transforms this into a direct constraint on the parameter adjustment range for the next welding cycle. This transformation stems from the inverse relationship between "margin" and "gain" in classical control engineering: when the system state approaches the instability boundary, the control action amplitude should be reduced; when the system state is far from the boundary, the control action can be relatively relaxed. Considering that welding parameter adjustments are typically applied in the form of relative changes, this step uses a power function to scale the maximum allowable variation amplitude, thereby obtaining the upper limit of the parameter variation amplitude. The calculation relationship is as follows: ; in, This indicates the maximum allowable parameter variation within the central region of the welding process window; its value is determined during the commissioning phase based on the equipment's response capability and process test results. The convergence exponent is used to adjust the tightening speed of the constraint as the stability margin changes. When it is large, the constraint is It tightens noticeably in the early stages of enlargement, when When the value is small, the constraint tightening process is relatively gradual. This formula shows that when the welding state is stable, When smaller, Approaching 1, the parameter variation range is close to When the welding condition deviates from the stable range, When it increases, The rapid decrease causes the allowable range of parameter variation to shrink accordingly.

[0027] To facilitate understanding of this calculation process, a set of configuration parameters can be used as an example. For instance, in a pipeline welding scenario, the settings during the commissioning phase... To ensure stable welding Close to 0, configuration The second term causes the state amplitude to increase. It has a significant improvement effect, setting This represents the maximum relative adjustment range allowed by the process, and is selected as follows: To match the device's response speed. When the data is obtained within a certain time segment during operation... The welding condition shows a slight deviation. Substituting this into the above calculation yields the following result. At a moderate level, corresponding for Part of it; when the state deviates further, Increase rise, This further reduces the value, thus automatically adopting a more conservative change range when adjusting parameters in the next cycle.

[0028] Through the above calculations, this step ultimately forms the recoverability constraints. Its core content consists of the upper limit of parameter variation. and by The symbols and distributions imply the adjustment direction information. This constraint is directly used to construct and screen candidate welding parameters in the next step, ensuring that parameter adjustments always revolve around the recoverable range of the welding process, thereby achieving stable and controllable parameter adjustments in the next cycle under the narrow process window conditions of copper / stainless steel pipeline welding.

[0029] S3. Based on the baseline welding parameters of the current welding cycle, and combined with the preset discrete template, construct multiple candidate parameter vectors and a set of candidate welding parameters; based on the recoverability constraint and the coupling sensitive direction vector, calculate the corresponding risk index for each candidate parameter vector, and identify the candidate parameter vectors whose risk index exceeds the limit as unusable, and construct an unusable welding parameter set.

[0030] Specifically, this step uses recoverability constraints. As input, the upper limit of amplitude is... The information on adjustment direction is organized as the basis for the calculation of "parameter availability determination", and the set of unusable welding parameters is identified within the discrete candidate parameter set. In copper / stainless steel pipe welding, pulse waveform correlation adjustment, wire feed rhythm adjustment, and welding progress adjustment often work together to affect local heat input and molten pool behavior. When the operating conditions are close to the edge of the process window, the unidirectional superposition of parameters is more likely to trigger forming fluctuations or incomplete fusion. This step focuses on this scenario characteristic, merging the "overall change intensity" and "superposition intensity along the heat input coupling sensitive direction" of candidate parameters into a single risk index, and classifying candidate parameters accordingly. This allows the next step to directly determine the executable parameters within the exclusion boundary.

[0031] Furthermore, the construction of the candidate welding parameter set is accomplished using a production line-executable discrete template method: during the debugging phase, the controller writes several sets of neighborhood adjustment templates, each set consisting of a finite number of relative change levels. These levels correspond to the increase or decrease of pulse waveform related parameter sets, wire feed related parameter sets, and welding travel related parameter sets, respectively. During welding operation, the controller saves the reference welding parameter vector from the previous cycle or the current cycle. And add the relative change given by the template to Generate candidate parameter vectors At the same time, the candidate change amount is obtained. Here , and All are expressed in a relative form (e.g., using normalized control values ​​for each execution channel or relative scale), so that the risk indicators calculated subsequently maintain the same numerical scale and are easy to correlate with... Direct comparison. The aforementioned "variance norm" uses the Euclidean norm from linear algebra as a measure of overall variation intensity, derived from classical geometric distance metrics. To reflect the coupling sensitivity of parameter linkages in copper / stainless steel welding, a squared penalty term for vector projection is introduced. This penalty term's form originates from the quadratic penalty in classical regularization, and its target is selected as the candidate variation. The projection along the coupling-sensitive direction is used to highlight the characteristic that "the stronger the superposition in the same direction, the faster the risk increases." The coupling-sensitive direction is represented by a vector. This indicates that the configuration is determined by trial welding during the commissioning phase and written into the controller configuration area: commissioning personnel select several sets of typical parameter linkage modes, record the sensitive directions of weld formation and process stability, and normalize the corresponding relative change directions as... ,make This indicates the degree of superposition of candidate changes along this direction. Based on the above sources and modifications, the relationship between the risk indicators and unavailability determination in this step is as follows: ; in, Risk indicators for candidate parameters; The candidate change vector is determined by the candidate parameters. With reference parameters The result is obtained by direct subtraction within the controller. The Euclidean norm measures the intensity of overall change. For recoverability constraints The upper limit of the amplitude is given in step two; This is a coupling-sensitive direction vector, written into the configuration area during the debugging phase and called during runtime; This is the secondary penalty weighting coefficient, which is also set during the commissioning phase based on the process window and equipment response characteristics. The first term in the above formula... The overall change intensity is normalized according to the allowable range to make it a directly comparable scale; the second term The squared penalty of the projected amount amplifies the superposition along the sensitive direction in a quadratic manner; the sum of the two terms yields... As a comprehensive risk indicator, when it exceeds a threshold, the corresponding candidate parameters will be... Record as unavailable and add to the set .because Expressed in the form of relative change For normalized direction vectors, As the upper limit of relative amplitude, both terms in the above formula are quantization results at the same numerical scale, which facilitates direct superposition to form the upper limit. .

[0032] To demonstrate the operability of this judgment process, a calculation example with a set of parameters is provided. Assume that within a certain welding cycle, the output of step two... Configuration during the debugging phase and take (This example vector represents the concatenation of three execution channels in the same direction, considered a sensitive direction.) For candidate parameters Change Calculate the Euclidean norm The first item is The projected amount is The square of the projection is The second item is ,get This candidate parameter will not be included. If the change in another candidate is ,but The first item is approximately The projected amount is The square of the projection is The second item is ,get This candidate parameter also does not enter the selection process. When candidate changes exhibit stronger synergistic superposition across all three channels, Increase and rapidly rise under squared penalty This allows such candidates to be identified as unavailable and included. During production line operation, the controller calculates the results for each candidate in the set. And update the set ,Will This serves as a basis for exclusion when determining the target welding parameters in the next step.

[0033] The output of this step is a set of unusable welding parameters. Its recording method can use candidate indexes or candidate changes. This allows the next step to directly exclude unusable items from the candidate set and determine executable target welding parameters. By applying the recoverability constraints of step two to the candidate parameters in the form of "amplitude normalization + coupling direction secondary penalty", this step translates the scenario sensitivity of dissimilar pipeline welding into calculable, screenable, and executable exclusion rules, thereby establishing clear parameter boundaries for stable welding execution in the next cycle.

[0034] S4. Remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, select the candidate parameter vector with the smallest change range from the current reference welding parameter as the target welding parameter, and send it to the welding execution mechanism to complete the welding of the next welding cycle.

[0035] Specifically, this step involves the set of unusable welding parameters obtained in the previous stage. Based on this, the unusable portions of the candidate welding parameter set are removed, and the target welding parameters are determined from the remaining executable candidates. This step is used for the actual execution of the next welding cycle. It corresponds to the transition from "parameter availability screening" to "equipment action issuance." The controlled objects include the pulse output channel of the welding power source, the wire feed drive channel, and the welding motion channel, ensuring that welding in the next cycle can proceed continuously under recoverability constraints and the exclusion of unavailable parameters. To ensure a uniform scale in the calculation process for engineering applications, candidate welding parameter vectors are used. and the current reference welding parameter vector All variables are expressed using relative change scales, with each component corresponding to a normalized control quantity for an adjustable execution channel. Candidate change quantities are obtained by subtracting the components from each other in memory by the controller. This data is then directly used for subsequent distance calculations and instruction generation.

[0036] Furthermore, in the fields of control theory and engineering optimization, to maintain the stationarity of the controlled system during discrete periodic updates, the "minimum change" principle is often used to select the control variable. This means selecting the candidate solution with the smallest distance from the current control variable within the feasible set that satisfies the constraints. This principle can be seen as a direct application of the classical least squares approach to discrete candidate sets: using Euclidean distance as a measure of change, the "minimum change" problem is transformed into a minimum L2 norm problem, thereby suppressing control variable jumps without introducing complex predictions. Following this classical form, this step first performs set filtering operations within the controller, removing unusable sets from the candidate welding parameter set. Eliminating elements yields a set of executable candidate sets. ; then in The target welding parameters are selected using the minimum L2 distance criterion. This selection relationship is expressed as: ; in, The target welding parameter vector to be actually executed in the next welding cycle; To remove unusable welding parameters from the candidate welding parameter set The resulting set of executable candidates is built in real time by the controller and stored in the cache during the current cycle; The reference welding parameter vector for the current welding cycle is read from the execution register by the controller at the beginning of the cycle and remains unchanged during the cycle; The Euclidean norm distance is used to measure the overall change of candidate parameters relative to the baseline parameters. Because... and All are similar vectors with relative scales. This distance is a metric result under the same numerical scale and can be directly used for comparison and minimization.

[0037] To demonstrate the operability of this selection process, a calculation example with parameter substitution is given. Let the current baseline parameter be... The three components correspond to the normalized adjustment values ​​of the welding power pulse channel, the wire feed drive channel, and the welding motion channel, respectively. Step three identifies and eliminates the set of unusable parameters. Then, an executable candidate set is obtained. It contains three sets of candidate parameters: , , The controller calculates the distance separately: for The difference vector is , Distance is ;right The difference vector is The distance is ;right The difference vector is The distance is Therefore, the minimum distance correspondence is obtained. Therefore, this step is selected. This serves as the parameter for the next welding cycle. The calculation involves only candidate set traversal and Euclidean distance calculation, and can be completed within the welding power control cycle.

[0038] Sure Then, the controller breaks it down into execution instructions according to the channel and writes them into the corresponding interface register: welding power supply according to The pulse channel component updates the pulse output duty cycle or amplitude scale for the next cycle. The wire feed drive updates the drive setpoint based on the wire feed channel component, and the welding motion channel updates the motion setpoint based on the travel channel component, thus completing the actual welding execution within the next welding cycle. During execution, the welding power source continuously generates welding electrical parameter data and enters the next round of state construction, forming a continuous operation flow that connects with the previous steps. The output of this step is the determined and executed target welding parameters. And the welding process of the next welding cycle triggered by it. The electrical parameter data generated in the next cycle will be used as the input for step one in the subsequent cycle, completing the closed loop from parameter elimination to equipment execution.

[0039] This invention also provides an adaptive intelligent welding control system for copper / stainless steel pipelines based on dynamic electrical parameter feedback, the system comprising: The electrical parameter acquisition module is used to continuously acquire welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and combine them with the corresponding time segments to construct welding state characterization quantities. The constraint generation module is used to calculate a stability margin index based on the welding state characterization quantity of the current time segment. The stability margin index is used to measure the stability of the welding process. Based on the stability margin index, the module determines the maximum parameter change range for parameter adjustment in the next welding cycle and generates recoverable constraints. The recoverable constraints include the upper limit of the parameter change range and adjustment direction information. The parameter filtering module is used to construct multiple candidate parameter vectors and a set of candidate welding parameters based on the baseline welding parameters of the current welding cycle and a preset discrete template; based on the recoverability constraint and the coupling sensitive direction vector, it calculates the corresponding risk index for each candidate parameter vector, and identifies the candidate parameter vectors whose risk index exceeds the limit as unusable, thus constructing a set of unusable welding parameters. The execution control module is used to remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, the candidate parameter vector with the smallest change from the current reference welding parameter is selected as the target welding parameter and sent to the welding execution mechanism to complete the welding of the next welding cycle.

[0040] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0041] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0042] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0043] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. An adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback, characterized in that, The method includes: S1. Continuously collect welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and construct welding state characterization quantities by combining the corresponding time segments. S2. Based on the welding state characterization quantity of the current time segment, calculate the stability margin index, which is used to measure the stability of the welding process; according to the stability margin index, determine the maximum parameter change range for parameter adjustment in the next welding cycle, and generate recoverability constraints; wherein, the recoverability constraints include the upper limit of parameter change range and adjustment direction information; S3. Based on the baseline welding parameters of the current welding cycle, and combined with the preset discrete template, construct multiple candidate parameter vectors and a set of candidate welding parameters; based on the recoverability constraint and the coupling sensitive direction vector, calculate the corresponding risk index for each candidate parameter vector, and identify the candidate parameter vectors whose risk index exceeds the limit as unusable, and construct an unusable welding parameter set. S4. Remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, select the candidate parameter vector with the smallest change range from the current reference welding parameter as the target welding parameter, and send it to the welding execution mechanism to complete the welding of the next welding cycle.

2. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 1, characterized in that, S1 specifically includes: Continuously collect welding electrical parameter data output by the power supply during the welding process; The continuously acquired welding electrical parameter data is divided into several continuous time segments according to a fixed time length. Each time segment covers multiple electrical parameter sampling points. The time length of the corresponding time segment is set in association with the control cycle or pulse cycle of the welding power source, so that each time segment contains complete welding behavior characteristics. For each time segment, the arc voltage sequence and welding current sequence are jointly processed to extract combined features reflecting the continuity of energy input and arc stability. The combined features are scaled and range-mapped to maintain a consistent expression scale under different pipe diameters, material batches, or process conditions. Then, the feature combination within the segment is output as a welding state characterization quantity.

3. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 2, characterized in that, The combined features are used to reflect the smoothness of the changes in arc voltage and welding current over time, the persistence of abrupt changes, and the correspondence between the changes in the two.

4. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 1, characterized in that, The calculation of the stability margin index based on the welding state characterization quantity of the current time segment specifically includes: The importance of the welding state characterization quantities is weighted by a weight vector to generate a weighted summation result; Based on the weighted summation result, a smooth nonlinear mapping function is introduced to compress it, generating a compressed result; Based on the compression results, a stability margin index is calculated by combining an additional term based on the overall amplitude of the welding state characterization quantity; the additional term based on the overall amplitude of the state is used to accelerate the constraint tightening speed when the state deviation increases.

5. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 4, characterized in that, The step of determining the maximum parameter change range for parameter adjustment in the next welding cycle based on the stability margin index and generating recoverability constraints is as follows: Based on the stability margin index, the maximum allowable variation range is scaled using a power function to generate the maximum allowable parameter variation range within the central region of the welding process window. The recoverability constraints are generated by combining the maximum parameter change magnitude and the adjustment direction information; wherein, the adjustment direction information is determined by the weighted summation result.

6. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 1, characterized in that, The reference welding parameters based on the current welding cycle, combined with a preset discrete template, are used to construct multiple candidate parameter vectors and a candidate welding parameter set, specifically as follows: During the debugging phase, the controller writes several sets of neighborhood adjustment templates, where each set of templates is composed of a finite number of relative change levels; The receiver saves the reference welding parameter vector of the previous cycle or the current cycle, and adds the relative change given by the neighborhood adjustment template to the reference welding parameter vector to generate a candidate parameter vector, while obtaining the candidate change.

7. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 6, characterized in that, The risk index of the candidate parameter is calculated and generated based on the candidate change amount, the maximum parameter change amplitude, and the coupling sensitive direction vector; the coupling sensitive direction vector is pre-configured according to the process test results. Candidate parameters with values ​​greater than 1 are recorded as unavailable and included in the set of unavailable welding parameters.

8. The adaptive intelligent welding control method for copper / stainless steel pipelines based on dynamic electrical parameter feedback according to claim 1, characterized in that, The magnitude of the change is measured using Euclidean distance to prioritize the most stable parameter adjustment scheme while satisfying the constraints.

9. The method according to claim 1, characterized in that, In step S4, the target welding parameters include at least the normalized adjustment amount of the corresponding welding power pulse channel, the normalized adjustment amount of the wire feed drive channel, and the normalized adjustment amount of the welding motion channel.

10. An adaptive intelligent welding control system for copper / stainless steel pipelines based on dynamic electrical parameter feedback, characterized in that: The system includes: The electrical parameter acquisition module is used to continuously acquire welding electrical parameter data output by the power supply during the welding process, including arc voltage and welding current, and combine them with the corresponding time segments to construct welding state characterization quantities. The constraint generation module is used to calculate a stability margin index based on the welding state characterization quantity of the current time segment. The stability margin index is used to measure the stability of the welding process. Based on the stability margin index, the module determines the maximum parameter change range for parameter adjustment in the next welding cycle and generates recoverable constraints. The recoverable constraints include the upper limit of the parameter change range and adjustment direction information. The parameter filtering module is used to construct multiple candidate parameter vectors and a set of candidate welding parameters based on the baseline welding parameters of the current welding cycle and a preset discrete template; based on the recoverability constraint and the coupling sensitive direction vector, it calculates the corresponding risk index for each candidate parameter vector, and identifies the candidate parameter vectors whose risk index exceeds the limit as unusable, thus constructing a set of unusable welding parameters. The execution control module is used to remove the unusable welding parameter set from the candidate welding parameter set to obtain an executable candidate set; in the executable candidate set, the candidate parameter vector with the smallest change from the current reference welding parameter is selected as the target welding parameter and sent to the welding execution mechanism to complete the welding of the next welding cycle.

Images