A method and apparatus for monitoring the quality of a build-up welding process

Abstract

The present application relates to the technical field of welding quality detection, and particularly relates to a surfacing process quality monitoring method and device, which comprises the following steps: acquiring a dynamic parameter set of a surfacing process in real time, wherein the dynamic parameter set comprises arc characteristic parameters, temperature field distribution parameters and mechanical state parameters; setting an edge computing node and inputting the dynamic parameter set into a welding quality prediction model to output a quality evaluation result comprising a defect probability and a penetration prediction value; generating a welding parameter adjustment instruction based on the quality evaluation result to adjust the operation state of a welding power source and a wire feeding mechanism in real time; and updating the dynamic parameter set according to the welding parameter adjustment instruction and transmitting the dynamic parameter set to an industrial computer based on the edge computing node. Through the present application, the problem of uncontrollable instantaneous quality fluctuation caused by strong coupling of multiple physical fields in the surfacing process is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of welding quality inspection technology, and in particular to a method and equipment for quality monitoring of the welding process. Background Technology

[0002] As an important additive manufacturing technology, surfacing is widely used in the field of surface strengthening and repair of mechanical parts. Its core process essentially involves melting a metal welding wire into droplets using an electric arc heat source, forming a metallurgically bonded surfacing layer on the substrate surface. This process involves the strong coupling of multiple physical fields, including arc physics, molten pool fluid dynamics, and solid-state phase transitions: arc energy fluctuations directly affect the molten pool temperature gradient, while the mechanical vibration of the wire feeding mechanism interferes with the droplet transition stability, ultimately resulting in the instantaneous generation of microscopic defects such as abrupt changes in weld depth, porosity, and cracks.

[0003] Current mainstream technologies in the industry suffer from monitoring lag. Traditional methods relying on post-weld X-ray or ultrasonic testing typically only yield results some time after the process is complete, failing to intercept defects arising on short timescales. Online monitoring only analyzes parameters such as arc characteristics, temperature field, and mechanical state separately. In reality, fluctuations in wire feed speed alter the molten pool flow field through droplet impact, thus affecting the arc plasma contraction effect. This multi-parameter chain reaction is not modeled. These shortcomings collectively point to the uncontrollable quality fluctuations under the transient coupling of multi-physics fields. This directly results in the persistently low pass rate of welding on core components of high-end equipment, coupled with high rework costs.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a method and equipment for quality monitoring during the welding process, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for quality monitoring during welding processes, the method comprising:

[0008] The dynamic parameter set of the welding process is acquired in real time. The dynamic parameter set includes arc characteristic parameters, temperature field distribution parameters and mechanical state parameters.

[0009] Set up edge computing nodes and connect them to the welding quality prediction model. Input a dynamic parameter set and output quality assessment results that include defect probability and weld penetration prediction values.

[0010] Welding parameter adjustment instructions are generated based on the quality assessment results, and the operating status of the welding power source and wire feeding mechanism is adjusted in real time.

[0011] The dynamic parameter set is updated according to the welding parameter adjustment instructions and transmitted to the industrial control unit based on the edge computing nodes.

[0012] Furthermore, the output includes quality assessment results that incorporate defect probability and melt penetration prediction values, including:

[0013] Obtain the quality impact characteristics of each parameter in the dynamic parameter set and establish a real-time interactive impact relationship based on the quality impact characteristics;

[0014] Cross-domain temporal correlation analysis is performed on the real-time interactive influence relationship, and the weight coefficients of each parameter in the dynamic parameter set on the influence of the weld overlay quality are dynamically allocated.

[0015] The defect probability and the predicted melt depth are generated synchronously based on the allocation results of the weighting coefficients.

[0016] Furthermore, when a high-frequency vibration characteristic indicating an abnormality in the wire feeding mechanism is detected in the mechanical state parameters, a device fault command is generated and the welding process is interrupted.

[0017] Furthermore, performing edge computing yields quality results, including:

[0018] The spatiotemporal alignment of the dynamic parameter set is completed locally on the edge computing node, and a real-time analysis window synchronized with the dynamic evolution of the molten pool is established.

[0019] The weight coefficients are dynamically allocated based on the real-time analysis window, while the historical weight allocation sequence of at least three consecutive windows is cached.

[0020] When the communication delay of the industrial control unit exceeds a preset threshold, the historical weight allocation sequence is activated to generate the quality assessment result, which is then transmitted in parallel to the industrial control unit and the welding power supply for adjustment.

[0021] Further, generating the quality assessment result using the historical weight allocation sequence includes:

[0022] Extract the fluctuation characteristics from the historical weight allocation sequence to identify the weight offset pattern between the arc energy parameter and the mechanical vibration parameter;

[0023] The current trend of coupling strength change in the real-time analysis window is matched with the weight offset pattern using pattern similarity matching.

[0024] When the matching degree is lower than the preset tolerance range, a compensation factor based on the solidification of the molten pool is generated;

[0025] The compensation factor is used to correct the mean calculation result of the historical weight allocation sequence to generate the quality assessment result.

[0026] Further, based on the quality assessment results, welding parameter adjustment instructions are generated, including:

[0027] The coupling relationship between the defect probability and the predicted melt depth is analyzed, and when the defect probability exceeds a preset value, an arc energy adjustment command is generated.

[0028] Based on the degree to which the predicted melt depth deviates from the process reference, the wire feeding speed compensation is dynamically calculated and a mechanical state correction command is generated.

[0029] The arc energy adjustment command and the mechanical state correction command are fused together to generate a set of coordinated control commands that simultaneously apply to the pulse waveform parameters of the welding power source and the servo motor of the wire feeding mechanism.

[0030] Furthermore, the operating status of the welding power source and wire feeding mechanism is adjusted in real time, including:

[0031] The collaborative control instruction set is decomposed into a welding power supply execution subset and a wire feeding mechanism execution subset;

[0032] Based on the hardware abstraction layer interface, the welding power supply is used to generate a subset of pulse waveforms, and the current rise rate and pulse frequency are dynamically adjusted to match the requirements for molten pool oscillation suppression.

[0033] The torque loop controller of the servo motor driven by the wire feeding mechanism performs gradient compensation of the wire feeding speed within a preset time window.

[0034] Synchronously collect and execute molten pool status feedback data to verify the synergistic effectiveness of the arc energy adjustment command and the mechanical state correction command.

[0035] Furthermore, based on the updated dynamic parameter set transmitted by multiple edge computing nodes, cross-regional welding quality statistical analysis is performed to generate a global quality process optimization strategy and distribute it to the corresponding edge computing nodes.

[0036] A welding process quality monitoring device is used to implement the aforementioned welding process quality monitoring method.

[0037] The technical solution of this invention can achieve the following technical effects:

[0038] It effectively solves the problem of uncontrollable instantaneous quality fluctuations caused by strong coupling of multiple physical fields during the welding process.

[0039] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0040] 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 recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating a method for quality monitoring during the welding process.

[0042] Figure 2 A flowchart illustrating the process for obtaining quality assessment results;

[0043] Figure 3 A flowchart illustrating the process of generating welding parameter adjustment instructions. Detailed Implementation

[0044] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0045] 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. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0046] Example 1;

[0047] like Figure 1 As shown, this application provides a method for quality monitoring during the welding process, the method comprising:

[0048] S10: Real-time acquisition of dynamic parameter set of the welding process, which includes arc characteristic parameters, temperature field distribution parameters and mechanical state parameters;

[0049] S20: Set up edge computing nodes and connect them to the welding quality prediction model. Input the dynamic parameter set and output the quality assessment results containing the defect probability and the predicted penetration value.

[0050] S30: Generates welding parameter adjustment instructions based on quality assessment results, and adjusts the operating status of welding power source and wire feeding mechanism in real time;

[0051] S40: Update the dynamic parameter set according to the welding parameter adjustment command and transmit it to the industrial control unit based on the edge computing node.

[0052] Specifically, firstly, a dynamic parameter set needs to be acquired in real time during the welding process. This dynamic parameter set includes arc characteristic parameters, temperature field distribution parameters, and mechanical state parameters. Arc characteristic parameters can be captured by arc sensors to detect voltage and current fluctuations in the arc. Temperature field distribution parameters are obtained through thermal imagers to capture changes in the temperature field. Mechanical state parameters are obtained through mechanical sensors to capture the operating status of the welding equipment. This data can be processed by edge computing nodes, and a welding quality prediction model is used to analyze the data and generate quality assessment results. The quality assessment results are not limited to detecting existing welding defects but also include predicting the weld penetration depth, thereby enabling early intervention in the welding process. In a preferred embodiment, the edge computing nodes employ an intelligent prediction model. This model can learn and update by combining historical data from the welding process, thereby performing more accurate analysis of the real-time acquired data. Then, based on the quality assessment results, real-time welding parameter adjustment instructions are generated, including adjusting the weld... By adjusting welding parameters such as current, voltage, and wire feed speed, the welding parameter adjustment commands are directly fed back to the welding power source and wire feed mechanism, enabling them to adaptively adjust based on real-time data to ensure the stability and consistency of welding quality. For example, when the weld penetration depth is detected to be lower than the preset standard, the edge computing node can automatically generate commands to increase the welding current or slow down the wire feed speed to ensure that the penetration depth reaches the required standard. At the same time, all parameter updates during the adjustment process are further integrated into the dynamic parameter set to ensure the real-time nature and integrity of the data. These optimization steps not only improve welding quality but also reduce operational complexity by minimizing human intervention and enhance adaptive capabilities. Finally, all processed data is transmitted to the industrial control unit through the edge computing node for global data analysis and storage, providing a basis for subsequent welding process optimization. Through this real-time monitoring and adjustment mechanism, the quality control capability of the welding process is improved, resulting in enhanced stability and precision of the entire welding process.

[0053] The technical solution of this invention effectively solves the problem of uncontrollable instantaneous quality fluctuations caused by strong coupling of multiple physical fields during the welding process.

[0054] Furthermore, such as Figure 2 As shown, the output includes quality assessment results containing defect probability and melt depth prediction values, including:

[0055] Obtain the quality impact characteristics of each parameter in the dynamic parameter set and establish real-time interactive impact relationships based on the quality impact characteristics;

[0056] Cross-domain temporal correlation analysis is performed on the real-time interactive influence relationship, and the weight coefficients of each parameter in the dynamic parameter set on the influence of the weld overlay quality are dynamically allocated.

[0057] Based on the allocation results of the weighting coefficients, the defect probability and the predicted melting depth are generated simultaneously.

[0058] As a preferred embodiment of the above, firstly, in practice, real-time acquisition of the dynamic parameter set utilizes a high-precision sensor array to capture accurate data on arc characteristics, including arc length, stability, and current and voltage fluctuations. Equally important is the use of infrared thermometers to monitor the temperature field distribution, ensuring accurate data entry in high-heat areas. Mechanical state parameters are primarily obtained using vibration monitors and displacement sensors to collect dynamic information about the welding station and mechanism, assisting in the analysis of weld quality. Secondly, the setting of edge computing nodes requires a certain level of computing power. Typically, embedded systems with high processing capabilities are used to run welding quality prediction models. These models, based on machine learning algorithms, analyze the input dynamic parameter set and output a probability of defects. To improve the accuracy and real-time performance of the analysis, the model needs to continuously update the training data based on the quality assessment results of the weld penetration prediction, in order to cope with changes in different materials and welding environments. In the step of generating welding parameter adjustment instructions based on the quality assessment results, control logic is typically used to adjust the welding power supply and wire feeding mechanism in real time. Specifically, the welding current and voltage can be fine-tuned, or the wire feeding speed can be adjusted to adapt to the dynamically changing welding environment. This real-time adjustment ensures the stability of welding quality, thereby achieving effective control of welding defects. Finally, the dynamic parameter set is updated according to the adjustment instructions, while ensuring the real-time performance and integrity of the data. The updated dynamic parameter set is transmitted to the industrial control unit in real time via edge computing nodes, enabling centralized data management and subsequent analysis and optimization.

[0059] Furthermore, when a high-frequency vibration characteristic indicating an abnormality in the wire feeding mechanism is detected in the mechanical state parameters, an equipment fault command is generated and the welding process is interrupted.

[0060] As a preferred embodiment of the above, to ensure the stability and normal operation of the wire feeding mechanism, a vibration sensor is used to monitor the mechanical state of the mechanism at high frequency during operation. The vibration sensor collects the vibration data of the wire feeding mechanism in real time, forming a detailed record of vibration characteristics. This vibration characteristic information is a subtle manifestation of the machine's state. When the wire feeding mechanism is operating normally, the vibration characteristics remain within a stable frequency range. However, when abnormal high-frequency vibration occurs, it often indicates a mechanical failure of the wire feeding mechanism, such as bearing failure or component wear. Therefore, a vibration characteristic recognition algorithm is set up. This algorithm can compare the real-time vibration data with the pre-set normal vibration characteristics in real time. Once a high-frequency vibration exceeding the normal range is detected, the algorithm will identify the abnormal vibration characteristics. Based on dynamic characteristics, the algorithm instantly identifies and classifies abnormal signals, and then quickly generates equipment fault commands. These fault commands can be transmitted to the welding control system through the industrial control system to interrupt the current welding process and prevent potential equipment damage or welding quality problems. In a preferred embodiment, to ensure timely and accurate anomaly judgment, the vibration recognition algorithm has a self-learning function, which can continuously update and optimize based on historical operating data to improve the accuracy and sensitivity of detection. For example, if high-frequency vibration characteristics under a specific state are detected multiple times but do not cause damage to the equipment, the threshold of this characteristic value can be adjusted to improve the adaptability and stability of the system. In addition, after triggering the equipment fault command, a fault log is automatically recorded to facilitate subsequent maintenance and repair operations.

[0061] Furthermore, performing edge computing to obtain quality results includes:

[0062] The spatiotemporal alignment of the dynamic parameter set is completed locally on the edge computing node, and a real-time analysis window is established in sync with the dynamic evolution of the molten pool.

[0063] Dynamic allocation of weight coefficients is performed based on a real-time analysis window, while caching historical weight allocation sequences of at least three consecutive windows;

[0064] When the communication delay of the industrial control unit exceeds the preset threshold, the historical weight allocation sequence is activated to generate quality assessment results, which are then transmitted in parallel to the industrial control unit and the welding power supply for adjustment.

[0065] As a preferred embodiment of the above, firstly, the edge computing node locally performs spatiotemporal alignment processing on the dynamic parameter set. Through high-precision timestamps and spatial identification of sensor data, a real-time analysis window synchronized with the dynamic evolution of the molten pool is constructed. This real-time analysis window ensures the capture and analysis of instantaneous changes during the welding process, providing timely feedback on the current welding state. After the real-time analysis window is established, weight coefficients are assigned to each dynamic parameter, and dynamic allocation of these weight coefficients is performed. This dynamic allocation mechanism adjusts the importance of each parameter according to real-time data changes during the welding process, ensuring the accuracy of welding quality assessment. To effectively utilize historical data, the edge node caches at least three consecutive windows of historical weight allocation sequences. This caching mechanism not only improves the robustness of data processing but also enhances network performance. Alternative solutions are provided in case of delays or anomalies; a preferred embodiment includes enabling the generation of quality assessment results by using a pre-cached historical weight allocation sequence when the communication delay between the control system and the industrial control unit exceeds a preset threshold. This mechanism ensures that even in the case of network congestion, the edge computing node can still provide stable quality assessment results. The generated assessment results are then simultaneously transmitted to the industrial control unit and the welding power source to realize the instant adjustment of welding parameters and the execution of real-time adjustment commands to maintain the stability and consistency of the welding process. This implementation improves the accuracy and efficiency of quality monitoring of the welding process by using a real-time analysis window, dynamic weight allocation, and historical caching mechanism, especially in terms of stability under high communication delay conditions, ensuring efficient operation of the welding process under different working conditions.

[0066] Furthermore, the quality assessment results are generated using historical weighted sequence assignments, including:

[0067] Extract fluctuation characteristics from historical weight allocation sequences to identify the weight shift patterns of arc energy parameters and mechanical vibration parameters;

[0068] The current real-time analysis window's coupling strength change trend and weight offset pattern are matched using pattern similarity.

[0069] When the matching degree is lower than the preset tolerance range, a compensation factor based on the solidification of the molten pool is generated;

[0070] The mean calculation result of the historical weight allocation sequence is corrected by using a compensation factor to generate the quality assessment result.

[0071] As a preferred embodiment of the above, in the historical sequence analysis, by performing distribution detection and curve fitting on the weight data of different welding stages, the optimal weight shift pattern between arc energy and mechanical vibration parameters is identified. The weight shift pattern can reveal the intrinsic relationship of parameter changes under specific conditions and guide the subsequent analysis process. Next, the trend of coupling strength change in the current real-time analysis window is detected, and by quantifying the correlation between data, pattern similarity matching is performed with the identified weight shift pattern. This similarity matching determines whether the current state meets the expected quality standard by comparing the real-time window data with typical patterns in the historical weights. If the matching degree is lower than the preset tolerance range, it means that the observed data does not match the historical optimal result. If there is a significant deviation between the results, a compensation factor based on the solidification of the molten pool is generated. The generation of the compensation factor takes into account the uncertainties that may occur during the welding process, such as uneven heat distribution or differences in material properties. The application of the compensation factor in the compensation effect is specifically manifested in the detailed correction of the mean calculation results of the historical weight allocation sequence. By adjusting the compensation factor, the weight allocation results can be adjusted so that the final evaluation results are more suitable for the current welding environment and conditions. For example, when the weld quality is lower than expected due to excessive mechanical vibration during a welding process, this anomaly is identified and the weight sequence is adjusted through the compensation factor so that the quality evaluation results can reflect the current welding status after compensation, thereby further optimizing and adjusting the welding parameters.

[0072] Furthermore, such as Figure 3 As shown, welding parameter adjustment instructions are generated based on the quality assessment results, including:

[0073] The coupling relationship between defect probability and melt depth prediction is analyzed, and when the defect probability exceeds the preset value, an arc energy adjustment command is generated.

[0074] Based on the degree to which the predicted melt depth deviates from the process baseline, the wire feeding speed compensation is dynamically calculated and a mechanical state correction command is generated.

[0075] The arc energy adjustment command and the mechanical state correction command are integrated to generate a set of coordinated control commands that simultaneously apply to the pulse waveform parameters of the welding power source and the servo motor of the wire feeding mechanism.

[0076] As a preferred embodiment of the above, firstly, after receiving the quality assessment results, the welding parameter adjustment strategy is determined by analyzing the coupling relationship between the defect probability and the predicted penetration depth. When the defect probability exceeds a preset threshold, it means that certain unstable factors in the welding process may lead to a decrease in welding quality. Therefore, an arc energy adjustment command is generated to achieve real-time optimization control of the arc energy. The increase or decrease of arc energy can directly affect the stability of welding, and thus affect the welding quality. At the same time, for situations where the predicted penetration depth deviates from the process reference, the compensation amount of the wire feed speed is calculated. This calculation is performed dynamically and can comprehensively consider the degree of deviation between the current welding parameters and the process reference to ensure the compensation amount. The effectiveness and adaptability of the system are improved by precisely adjusting the wire feed speed to achieve the ideal penetration depth. The generated mechanical condition correction commands not only adjust the wire feed speed but may also involve fine control of the wire feed tension and angle to achieve optimal state matching during welding. Furthermore, the arc energy adjustment commands and mechanical condition correction commands are integrated. This is based on a global optimization consideration of the entire welding process, generating a set of coordinated control commands that simultaneously act on the welding power source and the wire feed mechanism. For the control of the welding power source, the command set includes adjusting pulse waveform parameters, such as changes in frequency and amplitude. For the wire feed mechanism, the wire feed speed is flexibly adjusted through precise control of the servo motor.

[0077] Furthermore, real-time adjustment of the operating status of the welding power source and wire feeding mechanism includes:

[0078] The collaborative control instruction set is broken down into a welding power source execution subset and a wire feeding mechanism execution subset;

[0079] Based on the hardware abstraction layer interface, a subset of the welding power supply is executed to generate pulse waveforms, and the current rise rate and pulse frequency are dynamically adjusted to match the requirements for suppressing molten pool oscillations.

[0080] The torque loop controller based on the subset of the wire feeding mechanism drives the servo motor to complete the gradient compensation of the wire feeding speed within a preset time window.

[0081] Synchronously collect and execute molten pool status feedback data to verify the synergistic effectiveness of arc energy adjustment commands and mechanical status correction commands.

[0082] As a preferred embodiment of the above, firstly, the cooperative control instruction set is decomposed into two execution subsets: a welding power supply execution subset and a wire feeding mechanism execution subset. These two subsets involve specific control parameters for the power supply and wire feeding, respectively, and are operated through a hardware abstraction layer interface. For the welding power supply execution subset, the current rise rate and pulse frequency are dynamically adjusted by meticulously generating pulse waveforms. The purpose of this is to match the requirements for molten pool oscillation suppression, ensure the stability of the molten pool during the welding process, and improve welding quality. The precise control of the current and the oscillation suppression function make the regulation of arc energy more effective. Secondly, the wire feeding mechanism execution subset is controlled by a torque loop controller that drives the servo motor. The torque loop controller can complete the operation within a preset time window. Gradient compensation of wire feeding speed, along with the flexibility and high-precision control of the servo motor, ensures stable speed changes during wire feeding, guaranteeing synchronization between material supply and arc energy. This design ensures the stability of the wire feeding mechanism under high loads and allows for rapid response to changes in welding power supply requirements. In the optimized embodiment, feedback data on the molten pool state after execution is collected. This feedback mechanism verifies the synergistic effect of arc energy adjustment commands and mechanical state correction commands. The verification process includes detailed comparisons of molten pool temperature distribution, cooling rate, and solidification morphology to ensure that the adjusted welding parameters achieve the expected results. The feedback data is not only used for current process verification but also serves as source data for future optimization and deep learning, providing a reference for subsequent welding processes.

[0083] Furthermore, based on the updated dynamic parameter set transmitted from multiple edge computing nodes, cross-regional welding quality statistical analysis is performed to generate a global quality process optimization strategy and distribute it to the corresponding edge computing nodes.

[0084] As a preferred embodiment of the above, firstly, multiple edge computing nodes distributed in different regions continuously transmit updated dynamic parameter sets. These parameter sets contain real-time welding process data for each region. This real-time transmission ensures the immediacy and comprehensiveness of the data, enabling the collection of diverse data samples across different welding environments and process settings. After receiving the dynamic parameter sets from each node, a global welding quality statistical analysis is performed. By combining data mining and machine learning techniques, common characteristics and differences between different nodes are identified, and the welding quality of each region is evaluated. In particular, the statistical analysis covers the comprehensive performance of key parameters such as arc characteristics, temperature field distribution, and mechanical condition, providing a basis for further optimization. Quantitative basis: By comparing and integrating parameter data transmitted from different regions, a global quality trend is derived. By evaluating these trend changes and outliers, a global quality process optimization strategy is generated. The strategy includes adjusting welding current, voltage, and wire feed speed to adapt to welding challenges in different regions and improve overall quality. Then, the generated global quality process optimization strategy is distributed to the corresponding edge computing nodes. Each node executes an appropriate optimization strategy based on the specific conditions of its corresponding region to improve welding results in real time. For example, for regions with slight differences in steel thickness, the strategy may focus on fine adjustments to the current, while for regions with significant mechanical vibration, the focus is on stabilizing the operation of the wire feed mechanism.

[0085] Example 2;

[0086] Based on the same inventive concept as the welding process quality monitoring method in the foregoing embodiments, the present invention also provides a welding process quality monitoring device for implementing a welding process quality monitoring method.

[0087] The device described above in this invention can effectively implement a method for quality monitoring of the welding process, and the technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0088] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for quality monitoring during the welding process, characterized in that, The method includes: The dynamic parameter set of the welding process is acquired in real time, and the dynamic parameter set includes arc characteristic parameters, temperature field distribution parameters and mechanical state parameters; Set up edge computing nodes and connect them to the welding quality prediction model. Input the dynamic parameter set and output the quality assessment results containing the defect probability and the predicted penetration value. Based on the quality assessment results, welding parameter adjustment instructions are generated to adjust the operating status of the welding power source and wire feeding mechanism in real time. The dynamic parameter set is updated according to the welding parameter adjustment command and transmitted to the industrial control unit based on the edge computing node; The output includes quality assessment results containing defect probability and melt depth prediction values, including: Obtain the quality impact characteristics of each parameter in the dynamic parameter set and establish a real-time interactive impact relationship based on the quality impact characteristics; Cross-domain temporal correlation analysis is performed on the real-time interactive influence relationship, and the weight coefficients of each parameter in the dynamic parameter set on the influence of the weld overlay quality are dynamically allocated. The defect probability and the predicted melt depth are generated synchronously based on the allocation results of the weighting coefficients. Performing edge computing based on the edge computing node includes: The spatiotemporal alignment of the dynamic parameter set is completed locally on the edge computing node, and a real-time analysis window synchronized with the dynamic evolution of the molten pool is established. The weight coefficients are dynamically allocated based on the real-time analysis window, while the historical weight allocation sequence of at least three consecutive windows is cached. When the communication delay of the industrial control unit exceeds a preset threshold, the historical weight allocation sequence is activated to generate the quality assessment result, which is then transmitted in parallel to the industrial control unit and the welding power supply for adjustment. The quality assessment result is generated by using the historical weight allocation sequence, including: Extract the fluctuation characteristics from the historical weight allocation sequence to identify the weight offset pattern between the arc characteristic parameters and the mechanical state parameters; The current trend of coupling strength change in the real-time analysis window is matched with the weight offset pattern using pattern similarity matching. When the matching degree is lower than the preset tolerance range, a compensation factor based on the solidification of the molten pool is generated; The compensation factor is used to correct the mean calculation result of the historical weight allocation sequence to generate the quality assessment result.

2. The method for quality monitoring of the welding process according to claim 1, characterized in that, When a high-frequency vibration characteristic indicating an abnormality in the wire feeding mechanism is detected in the mechanical state parameters, a device fault command is generated and the welding process is interrupted.

3. The method for quality monitoring of the welding process according to claim 1, characterized in that, Based on the quality assessment results, welding parameter adjustment instructions are generated, including: The coupling relationship between the defect probability and the predicted melt depth is analyzed, and when the defect probability exceeds a preset value, an arc energy adjustment command is generated. Based on the degree to which the predicted melt depth deviates from the process reference, the wire feeding speed compensation is dynamically calculated and a mechanical state correction command is generated. The arc energy adjustment command and the mechanical state correction command are fused together to generate a set of coordinated control commands that simultaneously apply to the pulse waveform parameters of the welding power source and the servo motor of the wire feeding mechanism.

4. The method for quality monitoring of the welding process according to claim 3, characterized in that, Real-time adjustment of the operating status of the welding power source and wire feeding mechanism, including: The collaborative control instruction set is decomposed into a welding power supply execution subset and a wire feeding mechanism execution subset; Based on the hardware abstraction layer interface, the welding power supply is used to generate a subset of pulse waveforms, and the current rise rate and pulse frequency are dynamically adjusted to match the requirements for molten pool oscillation suppression. The torque loop controller of the servo motor driven by the wire feeding mechanism performs gradient compensation of the wire feeding speed within a preset time window. Synchronously collect and execute molten pool status feedback data to verify the synergistic effectiveness of the arc energy adjustment command and the mechanical state correction command.

5. The method for quality monitoring of the welding process according to claim 1, characterized in that, Based on the updated dynamic parameter set transmitted from multiple edge computing nodes, cross-regional welding quality statistical analysis is performed to generate a global quality process optimization strategy and distribute it to the corresponding edge computing nodes.

6. A welding process quality monitoring device, used to implement the welding process quality monitoring method as described in any one of claims 1-5.

Images