A method, system, product and medium for correcting track deviation of ultra-long cantilever friction stir welding

Abstract

A method, system, product and medium for correcting the trajectory offset of ultra-long cantilever stir friction welding. The method comprises: during the welding process, acquiring the acoustic signal of the welding head in real time and extracting the acoustic feature data, comparing it with the acoustic basic feature library constructed by no-load operation, and calculating the acoustic difference. When the acoustic difference exceeds a preset threshold, it is identified as an acoustic anomaly and the abnormal frequency characteristics are recorded. According to the abnormal frequency characteristics, the welding head speed is adjusted according to the fine-tuning amplitude and fine-tuning time proportional to the acoustic difference, and the speed is restored after the fine-tuning is completed. At the same time, the friction force, relative speed and acoustic signal after fine-tuning are obtained, and the friction energy per unit time is calculated. When the friction energy fluctuation coefficient exceeds the threshold and the instantaneous value is lower than the preset value, or the acoustic anomaly persists and the welding head deviates from the preset trajectory, the welding head is adjusted to the device trajectory. The implementation of the technical solution provided in the present application improves the trajectory accuracy of ultra-long cantilever stir friction welding.

Description

Technical Field

[0001] The present application relates to the field of welding, and in particular to a method, system, product and medium for correcting the offset of an ultra-long cantilever friction stir welding trajectory. Background Art

[0002] With the rapid development of advanced manufacturing industries such as aerospace, rail transit, and new energy, the processing and assembly of ultra-long cantilever structures has gained widespread application. As an efficient and environmentally friendly solid-state joining process, friction stir welding (FSW) holds significant application value in these fields. The quality of FSW connections is directly related to the strength, durability, and overall performance of the component.

[0003] Currently, to ensure the trajectory accuracy of friction stir welding, related technologies typically use visual sensors, laser tracking systems, or displacement sensors for real-time monitoring of the welding path. These methods directly detect the relative position of the welding head and the welding path, identify trajectory deviations, and adjust the position or posture of the welding head through a control system to achieve trajectory correction and ensure weld quality.

[0004] However, in the related art, the detection and compensation methods that directly rely on geometric offsets have problems with response lag and insufficient adaptability. When ultra-long cantilever components induce local micro-resonance or vibration due to welding heat input, periodic forging pressure or structural characteristics, these high-frequency dynamic changes are usually difficult to be accurately captured by visual or geometric sensors. If the correction system is too sensitive to these micro-offset changes, new disturbances may be introduced due to frequent high-frequency corrections, destroying the plastic flow state of the material during stir friction welding, resulting in welding defects (such as tunnel defects, surface ripples) or reduced process stability; if the correction system ignores the cumulative effect of these dynamic changes, it may cause the weld to gradually deviate from the ideal center line, ultimately affecting the welding accuracy and component strength. Therefore, how to effectively identify and compensate for the trajectory offset caused by the dynamic characteristics of the component during the welding process, while avoiding the damage to the welding quality and process stability caused by the correction action itself, is still a problem that is difficult to solve in the related art. Summary of the Invention

[0005] The present application provides a method, system, product and medium for correcting the trajectory offset of ultra-long cantilever friction stir welding, which are used to improve the trajectory accuracy of ultra-long cantilever friction stir welding.

[0006] In a first aspect of the present application, a method for correcting an offset of an ultra-long cantilever friction stir welding track is provided, the method comprising:

[0007] During the welding process, the acoustic signal of the welding head is obtained in real time, and the real-time acoustic feature data of the acoustic signal is extracted; the real-time acoustic feature data is compared with the acoustic basic feature library, and the acoustic difference is calculated; when the acoustic difference exceeds the preset acoustic difference threshold, it is identified as an acoustic anomaly and the abnormal frequency characteristics of the acoustic anomaly are recorded; according to the abnormal frequency characteristics, the rotation speed of the welding head is adjusted according to the fine-tuning amplitude and fine-tuning time, and the welding head is controlled to return to the rotation speed before fine-tuning after the fine-tuning time ends; the friction force, relative speed and real-time acoustic signal after the fine-tuning time are obtained in real time, and the friction energy within a preset unit time length is calculated; when the fluctuation coefficient of the friction energy exceeds the preset fluctuation coefficient threshold and the instantaneous value of the friction energy is lower than the preset friction energy threshold, or when there is still acoustic anomaly in the real-time acoustic signal, and the shortest distance between the real-time position of the welding head and the preset device trajectory exceeds the preset offset threshold, the position of the welding head is adjusted to the device trajectory.

[0008] In the above embodiment, by acquiring the acoustic signal of the welding head in real time, extracting the acoustic feature data and comparing it with the acoustic basic feature library, high-frequency dynamic anomalies caused by periodic forging pressure, small resonance or vibration during the welding process are identified, overcoming the defects of the traditional detection method that relies on geometric offset response lag and high-frequency dynamic changes that are difficult to capture. At the same time, by recording the frequency characteristics of the acoustic anomaly and implementing fine-tuning of the amplitude and speed of fine-tuning time, the authenticity and severity of the anomaly are verified, avoiding the disturbance damage caused by frequent high-frequency correction. In addition, combined with the multi-dimensional analysis of friction force, relative speed and friction energy fluctuations, when the anomaly persists or the offset exceeds the threshold, the welding head position is dynamically adjusted to the preset trajectory, thereby suppressing the cumulative effect of dynamic changes and avoiding the weld from deviating from the center line. This solution improves the real-time and adaptability of trajectory compensation, avoids the generation of welding defects such as defects and surface ripples, while ensuring the welding path accuracy and process stability, and improving the trajectory accuracy of ultra-long cantilever stir friction welding.

[0009] In conjunction with some embodiments of the first aspect, in some embodiments, adjusting the rotational speed of the welding head according to the fine-tuning amplitude and fine-tuning time based on the abnormal frequency characteristics, and controlling the welding head to return to the rotational speed before fine-tuning after the fine-tuning time ends, specifically includes:

[0010] The heat flux density in the area of ​​the preset heat flux monitoring area size below the welding head is imaged in real time, and the rotation speed of the welding head is adjusted according to the fine-tuning amplitude and fine-tuning time. The adjustment transient response characteristics of the heat flux density at the beginning of the fine-tuning time are recorded, and the welding head is controlled to return to the rotation speed before fine-tuning after the fine-tuning time ends; the speed adjustment data and the adjustment transient response characteristics are input into the preset ideal heat generation and heat transfer dynamic mathematical model of stir friction welding to obtain the actual heat generation efficiency and the actual heat flux density distribution; the speed adjustment data is the difference between the speed values ​​before and after the adjustment; when the heat generation efficiency deviation value exceeds the preset heat generation efficiency deviation value threshold or the heat flux density distribution deviation value exceeds the preset heat flux density distribution deviation threshold, an interface friction state instability warning information is issued, the axial force of the welding head is adjusted to the preset axial force optimization range, and the tool inclination angle of the welding head is adjusted according to the adjustment transient response characteristics.

[0011] In the above-described embodiment, by capturing real-time heat flux density in a pre-set heat flux monitoring area below the welding head and periodically adjusting the welding head rotational speed by fine-tuning the amplitude and time, the transient response characteristics of the heat flux density adjustment can be recorded and input into the ideal dynamic mathematical model of heat generation and heat transfer in friction stir welding to calculate the actual heat generation efficiency and heat flux density distribution. When the heat generation efficiency deviation or heat flux density distribution deviation exceeds a preset threshold, the axial force of the welding head is adjusted to the optimized range, and the tool inclination angle is dynamically adjusted to restore the stability of the interface friction state. This method uses real-time heat flux monitoring to keenly detect friction instability caused by uneven heat input or abnormal welding head movement. By adjusting the axial force and tool inclination angle in conjunction, it optimizes the heat input distribution and material flow behavior, avoiding the drawbacks of delayed response to dynamic changes in local heat diffusion and a single adjustment method. Ultimately, the heat input uniformity and process stability of the welding process are improved, ensuring the consistency and reliability of weld quality under complex working conditions. In particular, the accuracy of the welding trajectory is significantly improved in friction stir welding with ultra-long cantilevers.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, after inputting the speed adjustment data and the adjustment transient response characteristics into a preset ideal heat generation and heat transfer dynamic mathematical model of stir friction welding, the method further includes:

[0013] If it is found that the adjustment transient response characteristics are related to the rotation period of the welding head and match the preset acoustic-thermal coupling fingerprint of the specific defect of the welding head, the heat flux fluctuation frequency, heat flux fluctuation phase and heat flux fluctuation amplitude of the periodic heat flux fluctuation are obtained by analyzing the adjustment transient response characteristics; based on the correlation between the heat flux fluctuation amplitude, heat flux fluctuation phase and the abnormal spectrum characteristics corresponding to the acoustic anomaly, the preliminary defect type and preliminary defect location of the welding head are determined by comparing with the preset welding head defect fingerprint library; based on the ratio of the heat flux fluctuation amplitude to the preset fluctuation amplitude deviation threshold, the speed adjustment amplitude of the welding head is calculated, and based on the heat flux fluctuation phase, the adjustment phase point of the speed adjustment is calculated. At the adjustment phase point, the speed of the welding head is adjusted for a second time according to the speed adjustment amplitude; the second heat flux fluctuation amplitude and the second heat flux fluctuation phase distribution characteristics before and after the second adjustment are obtained, and the real-time offset of the welding head on the welding trajectory is calculated; based on the real-time offset, the offset compensation path of the welding head is generated using the multi-point interpolation method; and the motion trajectory of the welding head is adjusted again according to the offset compensation path.

[0014] In the above-described embodiment, by analyzing and adjusting the correlation between transient response characteristics and the weld head rotation period in real time and comparing them with a preset library of acoustic-thermal coupling fingerprints for specific weld head defects, the frequency, phase, and amplitude of periodic heat flux fluctuations are identified. This allows for the detection of potential anomalies caused by periodic defects or dynamic changes, and a preliminary determination of the weld head defect type and location. Based on the ratio of the heat flux fluctuation amplitude to a preset deviation threshold, the weld head speed adjustment range is calculated. The optimal adjustment phase point is determined by combining the heat flux fluctuation phase information, and a second adjustment of the weld head speed is performed at the optimal timing for dynamic response. This periodic feature-based control approach not only verifies the authenticity of the defect but also optimizes the dynamic behavior of the weld head, avoiding new disturbances caused by delayed or over-adjusted adjustments. After the speed adjustment is completed, the offset of the weld head along the welding trajectory is calculated in real time by analyzing the heat flux fluctuation amplitude and phase distribution characteristics before and after the second adjustment. A smooth offset compensation path is generated using multi-point interpolation, allowing for dynamic adjustment of the weld head trajectory to ensure welding stability and trajectory accuracy. Ultimately, by combining multi-dimensional acoustic and heat flow characteristic analysis with dynamic adjustment, the problems of delayed response to periodic defects and insufficient path compensation accuracy of traditional methods were overcome, and the accuracy of welding trajectories and weld quality were improved, especially in ultra-long cantilever friction stir welding, ensuring the stability and reliability of the welding process under complex working conditions.

[0015] In conjunction with some embodiments of the first aspect, in some embodiments, after comparing the real-time acoustic feature data with the acoustic basic feature library and calculating the acoustic difference, the method further includes:

[0016] When it is detected that the welding path is a long-distance straight line area, the vibration characteristics of the welding head are monitored in real time, and the vibration accumulation and vibration change trend are calculated, and the heat input gradient distribution in the welding area is monitored at the same time; when the vibration accumulation, vibration change trend or the change rate of the heat input gradient distribution exceeds the corresponding preset threshold and there is no acoustic anomaly in the acoustic difference, the rotation speed of the welding head is adjusted according to the preset fine-tuning parameters.

[0017] In the above-described embodiment, when the weld path is detected as a long, straight region, the vibration characteristics of the weld head are monitored in real time, the vibration accumulation is calculated, and the heat input gradient distribution in the weld region is monitored simultaneously. When the vibration accumulation, vibration trend, or rate of change of the heat input gradient distribution exceeds corresponding preset thresholds, and no abnormality is detected by acoustic variability, the weld head rotation speed is adjusted by a preset fine-tuning amplitude to balance the vibration state and optimize the heat input distribution. This method, through the coordinated monitoring and dynamic adjustment of vibration and heat input, can identify and suppress early abnormal trends caused by vibration accumulation or uneven heat input, avoiding the limitation of relying solely on acoustic detection, which may overlook non-acoustic anomalies. Compared with traditional methods, this solution overcomes the shortcomings of delayed response to heat input gradient changes and vibration accumulation and insufficient adjustment accuracy in long-distance, straight-line welding through real-time analysis and control of multi-dimensional parameters. This improves weld uniformity and the accuracy of the straight-line weld path, while ensuring the stability of the welding process. This further enhances the accuracy of the weld trajectory and process adaptability, particularly in ultra-long cantilever friction stir welding.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, when it is detected that the welding path is a long straight line area, the vibration characteristics of the welding head are monitored in real time, and the vibration accumulation amount and vibration change trend are calculated, and the heat input gradient distribution in the welding area is monitored at the same time, further comprising:

[0019] When the vibration change trend in the vibration characteristics exceeds the preset change trend threshold, the fine-tuning parameters are adjusted according to the vibration frequency and vibration amplitude in the vibration characteristics; according to the heat input gradient distribution in the welding path direction, the welding head speed corresponding to the area with excessive heat input is reduced, and the axial pressure of the welding head is increased in the area with excessive heat input.

[0020] In the above-described embodiment, when the vibration trend in the vibration characteristics exceeds a preset trend threshold, the fine-tuning amplitude of the welding head is dynamically adjusted based on the vibration frequency and amplitude. Simultaneously, based on the heat input gradient distribution along the welding path, the welding head speed corresponding to the area with excessive heat input is reduced, and the welding head axial pressure in this area is increased to achieve heat input balance and optimize material flow. This method, by capturing anomalies in vibration trend in real time and combining them with heat input characteristics for coordinated adjustment, can quickly respond to anomalies caused by the cumulative effect of vibration or localized excessive heat input, avoiding weld quality degradation and process instability. Compared with traditional methods that rely solely on speed adjustment or have a delayed response to vibration anomalies, this solution overcomes the lack of adaptability to dynamic changes through multi-dimensional coordinated control of vibration characteristics and heat input distribution, thereby improving welding path accuracy and weld consistency. Ultimately, this solution improves welding trajectory accuracy and process stability in ultra-long cantilever friction stir welding, meeting the requirements of high-precision welding under complex working conditions.

[0021] In conjunction with some embodiments of the first aspect, in some embodiments, before obtaining the friction force, relative speed, and real-time acoustic signal after the fine-tuning time in real time and calculating the friction energy within a preset unit time length, the method further includes:

[0022] When, by analyzing historical welding data and the CAD model of the current workpiece, it is predicted that an inherent low-frequency modal vibration mode is activated in the unwelded path section, the corresponding predicted anti-instability parameter combination is selected from the preset anti-instability welding parameter library according to the predicted vibration frequency and predicted vibration amplitude; when the welding head reaches the high-risk section corresponding to the activation of the inherent low-frequency modal vibration mode, the predicted anti-instability parameter combination is applied with a preset trial length, and the probe acoustic signal and the trial dynamic stability of the welding head are monitored in real time; if the trial dynamic stability is within the preset stability threshold range and the trial acoustic signal is within the preset acoustic signal threshold range, the predicted anti-instability parameter combination is applied to complete the welding of the high-risk section; if the trial dynamic stability is not within the preset stability threshold range or the trial acoustic signal is not within the preset acoustic signal threshold range, the preset small disturbance signal is superimposed on the control instruction of the welding head.

[0023] In the above embodiment, by analyzing historical welding data and the CAD model of the current workpiece, the inherent low-frequency modal vibration mode that may be activated in the unwelded path section is predicted, and based on the predicted vibration frequency and vibration amplitude, the corresponding predicted anti-instability parameter combination is selected from the preset anti-instability welding parameter library. In the high-risk section where the inherent low-frequency modal vibration mode is activated, the predicted anti-instability parameter combination is applied with a preset trial length, and the probe acoustic signal and the trial dynamic stability of the welding head are monitored in real time to judge the stability and reliability of the welding process. If the trial dynamic stability and the trial acoustic signal are both within the preset threshold range, the predicted anti-instability parameter combination is directly applied to complete the welding of the high-risk section; if the threshold requirements are not met, a small disturbance signal is superimposed on the control instruction of the welding head to dynamically optimize the vibration state and suppress the adverse effects of the low-frequency modal vibration mode. By combining the predictive capabilities of historical welding data and CAD model analysis, as well as tentative parameter verification and dynamic adjustment of small disturbance signals, it is possible to identify and respond in advance to welding instability problems caused by low-frequency modal vibration modes in high-risk sections. In ultra-long cantilever stir friction welding, prediction and real-time control are used to improve the accuracy of welding trajectories and weld quality in high-risk sections. At the same time, the stability and adaptability of the welding process are significantly improved, meeting the requirements for high-precision welding under complex working conditions.

[0024] In conjunction with some embodiments of the first aspect, in some embodiments, after comparing the real-time acoustic feature data with the acoustic basic feature library and calculating the acoustic difference, the method further includes:

[0025] When the acoustic difference is continuously high and there is no acoustic anomaly in the acoustic difference, the high baseline, high reproducibility characteristics of the acoustic difference and the corresponding workpiece type and welding section are identified; the preset digital twin model of the workpiece type is retrieved, and the equivalent thermal-mechanical load applied to the corresponding welding section during the stir friction welding process is simulated in the digital twin model; the inherent structural acoustic response characteristics generated by the welding section are analyzed through simulation calculations; if the inherent structural acoustic response characteristics are consistent with the real-time acoustic feature data, the inherent structural acoustic response characteristics are replaced with the acoustic basic feature library.

[0026] In the above embodiment, when the acoustic difference is persistently high and no acoustic anomalies are present, the high baseline, high reproducibility characteristics of the acoustic difference and their corresponding workpiece type and weld section are identified, and a preset digital twin model of the workpiece type is retrieved. The equivalent thermal-mechanical load of the weld section during the friction stir welding process is simulated in the model. The inherent structural acoustic response characteristics of the weld section are analyzed through simulation calculations. If the simulation results match the real-time acoustic feature data, the inherent structural acoustic response characteristics are replaced with the acoustic basic feature library to update and optimize the applicability of the feature library. By combining acoustic anomalies with the digital twin model, the acoustic high baseline phenomenon caused by the inherent structural characteristics of the workpiece can be identified and its rationality verified through simulation, avoiding the risk of misjudging the inherent acoustic characteristics of the workpiece as welding anomalies. By optimizing the basic acoustic feature library, the monitoring accuracy and reliability are improved. By accurately identifying and dynamically updating the basic acoustic feature library, the adaptability and accuracy of acoustic monitoring during stir friction welding are improved. In ultra-long cantilever welding, unnecessary adjustments or misoperations caused by misjudgment of inherent acoustic characteristics are avoided, further improving the accuracy of the welding trajectory and the stability of the process, meeting the needs of high-precision welding under complex working conditions.

[0027] In a second aspect, an embodiment of the present application provides a trajectory offset correction system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the trajectory offset correction system to perform the method described in the first aspect and any possible implementation of the first aspect.

[0028] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions. When the computer program product is run on a trajectory deviation correction system, the trajectory deviation correction system executes the method described in the first aspect and any possible implementation of the first aspect.

[0029] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium comprising instructions. When the instructions are executed on a trajectory deviation correction system, the trajectory deviation correction system executes the method described in the first aspect and any possible implementation of the first aspect.

[0030] It is understood that the trajectory deviation correction system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to implement the ultra-long cantilever friction stir welding trajectory deviation correction method provided in the embodiments of this application. Therefore, the beneficial effects achievable by these methods can be referenced to the beneficial effects of the corresponding methods and will not be further elaborated here.

[0031] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0032] 1. This application obtains the acoustic signal of the welding head in real time, extracts the acoustic feature data and compares it with the acoustic basic feature library to identify high-frequency dynamic anomalies caused by periodic forging pressure, small resonance or vibration during the welding process, thereby overcoming the defects of the traditional detection method that relies on geometric offset response lag and high-frequency dynamic changes that are difficult to capture. By recording the frequency characteristics of the acoustic anomaly and implementing fine-tuning of the amplitude and speed of fine-tuning time, the authenticity and severity of the anomaly are verified, while avoiding the disturbance damage caused by frequent high-frequency correction. Combined with the multi-dimensional analysis of friction, relative speed and friction energy fluctuations, when the anomaly persists or the offset exceeds the threshold, the welding head position is dynamically adjusted to the preset trajectory, thereby suppressing the cumulative effect of dynamic changes and preventing the weld from deviating from the center line. This solution improves the real-time and adaptability of trajectory compensation, avoids the generation of welding defects such as defects and surface ripples, while ensuring the welding path accuracy and process stability, and improving the trajectory accuracy of ultra-long cantilever stir friction welding.

[0033] 2. This application uses real-time imaging of the heat flux density in a preset heat flux monitoring area below the welding head, and combines the fine-tuning amplitude and fine-tuning time to adjust the welding head speed in stages. It can record the transient response characteristics of the heat flux density adjustment and input it into the ideal heat generation and heat transfer dynamic mathematical model of stir friction welding to calculate the actual heat generation efficiency and heat flux density distribution. When the heat generation efficiency deviation value or the heat flux density distribution deviation value exceeds the preset threshold, the axial force of the welding head is adjusted to the optimized range, and the tool inclination angle is dynamically adjusted to restore the stability of the interface friction state. This method uses real-time heat flux monitoring to keenly capture friction instability problems caused by uneven heat input or abnormal welding head movement, and optimizes the heat input distribution and material flow behavior by adjusting the axial force and tool inclination angle in a linked manner, avoiding the defects of delayed response to dynamic changes in local heat diffusion and a single adjustment method. Ultimately, the heat input uniformity and process stability of the welding process are improved, ensuring the consistency and reliability of weld quality under complex working conditions, especially in ultra-long cantilever stir friction welding, significantly improving the accuracy of the welding trajectory.

[0034] 3. This application analyzes and adjusts the correlation between transient response characteristics and the weld head rotation period in real time, and compares this with a preset library of acoustic-thermal coupling fingerprints for specific weld head defects. This identifies the frequency, phase, and amplitude of periodic heat flux fluctuations, thereby capturing potential anomalies caused by periodic defects or dynamic changes and preliminarily determining the type and location of the weld head defect. Based on the ratio of the heat flux fluctuation amplitude to a preset deviation threshold, the weld head speed adjustment amplitude is calculated. The optimal adjustment phase point is determined by combining the heat flux fluctuation phase information, and the weld head speed is adjusted a second time at the optimal timing for dynamic response. This periodic feature-based control method not only verifies the authenticity of the defect but also optimizes the dynamic behavior of the weld head, avoiding new disturbances caused by delayed or over-adjusted adjustments. After the speed adjustment is completed, the offset of the weld head on the welding trajectory is calculated in real time by analyzing the heat flux fluctuation amplitude and phase distribution characteristics before and after the second adjustment. A smooth offset compensation path is generated using multi-point interpolation, dynamically adjusting the weld head's motion trajectory to ensure the stability of the welding process and the accuracy of the trajectory. Ultimately, by combining multi-dimensional acoustic and heat flow characteristic analysis with dynamic adjustment, the problems of delayed response to periodic defects and insufficient path compensation accuracy of traditional methods were overcome, and the accuracy of welding trajectories and weld quality were improved, especially in ultra-long cantilever friction stir welding, ensuring the stability and reliability of the welding process under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of a method for correcting the offset of an ultra-long cantilever friction stir welding trajectory according to an embodiment of the present application;

[0036] Figure 2 This is another flow chart of the method for correcting the offset of the ultra-long cantilever friction stir welding trajectory in an embodiment of the present application;

[0037] Figure 3 It is a schematic diagram of an exemplary hardware structure of the trajectory deviation correction system in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The terms used in the following examples of the present application are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular expressions "a," "an," "said," "above," "the," and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to and encompasses any or all possible combinations of one or more of the listed items.

[0039] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of the technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, "plurality" means two or more.

[0040] To ensure the trajectory accuracy of friction stir welding (FSW), the welding path is typically monitored in real time using methods such as visual sensors, laser tracking systems, or displacement sensors. These methods directly detect the relative position of the welding head and the welding path, identify trajectory deviations, and adjust the position or motion of the welding head through a control system to correct the welding trajectory and ensure weld quality. However, these methods have significant limitations when dealing with high-frequency dynamic changes caused by welding heat input, cyclic forging pressure, or workpiece structural characteristics. Visual or geometric sensors struggle to accurately capture the subtle resonances or chatter that occur during the welding of extremely long cantilever components. These dynamic changes can cause the weld to gradually deviate from the ideal trajectory. Furthermore, correction methods that directly rely on geometric deviations suffer from response lags and are unable to quickly adapt to high-frequency changes. If the correction system is overly sensitive to subtle deviations, frequent high-frequency corrections may disrupt the material's plastic flow state, further leading to weld defects (such as tunneling defects and weld ripples) or reduced process stability. Therefore, accurately identifying dynamic anomalies during welding and appropriately compensating for them to avoid the dual effects of trajectory deviations and correction disturbances remain difficult problems.

[0041] In an embodiment of the present application, by acquiring the acoustic signal of the welding head in real time, extracting the acoustic feature data and comparing it with the acoustic basic feature library, the acoustic difference is calculated, and the acoustic anomalies caused by periodic forging pressure, small resonance or vibration during the welding process are identified. The acoustic basic feature library is constructed through no-load acoustic signals to ensure the reliability of the difference calculation. When the acoustic difference exceeds the preset threshold, the welding head speed is adjusted according to the fine-tuning amplitude and fine-tuning time based on the abnormal frequency characteristics, and the speed is restored after the fine-tuning time ends, thereby verifying the authenticity of the anomaly and avoiding frequent corrections to cause new disturbances. At the same time, by acquiring friction, relative speed and friction energy in real time, the stability of the welding interface can be dynamically quantified, and when the friction energy fluctuation coefficient or the real-time acoustic signal is abnormal, the position of the welding head can be adjusted to the preset trajectory. By combining multi-dimensional dynamic analysis of acoustic signals, friction energy, and welding head position, this solution overcomes the response lag problem of traditional geometric offset detection, avoids the damage to welding quality and stability caused by frequent corrections, improves the real-time and adaptability of trajectory compensation, ensures weld quality and welding path accuracy, and further improves process stability and welding trajectory accuracy, especially in ultra-long cantilever stir friction welding.

[0042] Figure 1 The present invention is a flowchart of a method for correcting the offset of an ultra-long cantilever friction stir welding track using an embodiment of the present invention, comprising the following steps:

[0043] S101. Acquire an acoustic signal of a welding head in real time during a welding process, and extract real-time acoustic feature data of the acoustic signal.

[0044] Specifically, real-time acquisition of the weld head's acoustic signals and extraction of acoustic signature data during the welding process is accomplished through highly sensitive acoustic sensors (such as piezoelectric sensors or MEMS microphones). These sensors are installed near the weld head or workpiece and can capture the vibration and sound signals generated at the interface between the weld head and the workpiece. These signals are collected in analog form and converted into digital signals through analog-to-digital conversion (ADC).

[0045] The acquired acoustic signal contains time domain data and frequency domain data. The signal is spectrally analyzed through fast Fourier transform (FFT) to extract its frequency characteristics (such as main frequency, harmonic frequency, amplitude) and energy characteristics (such as total energy and energy distribution).

[0046] S102: Compare the real-time acoustic feature data with the acoustic basic feature library and calculate the acoustic difference.

[0047] Specifically, the construction of the acoustic basic feature library is based on the acoustic signals collected when the welding head is in a no-load state. In the no-load state, the welding head has no contact with the workpiece, and its acoustic signal is mainly composed of the normal vibration of the equipment itself and the operating noise. Through a high-sensitivity acoustic sensor, the acoustic signal is continuously collected during the no-load operation, recorded in the form of time domain, and then converted into frequency domain data through signal processing methods. Fast Fourier transform (FFT) is usually used to decompose the original signal into different frequency components, and characteristic values ​​such as main frequency, harmonic frequency, frequency amplitude and energy distribution are extracted. After these characteristic values ​​are sampled multiple times, statistical methods (such as calculating the average value, standard deviation, etc.) are used to form a set of parameters that can represent the acoustic characteristics of the equipment in the no-load state, and stored as the acoustic basic feature library.

[0048] Next, when comparing the real-time acoustic feature data with the basic acoustic feature library, the corresponding frequency features in the real-time signal and the basic feature library are compared one by one. The difference between the main frequency of the real-time signal and the main frequency in the basic feature library, as well as the amplitude difference of the harmonic frequencies, are compared to quantify the degree of deviation between the real-time signal and the idle signal. To accurately assess this deviation, a common method is to use the Euclidean distance formula, averaging the squared difference between the real-time and basic features. Other algorithms (such as cosine similarity) can also be combined to quantify the overall similarity of the signals.

[0049] Ultimately, the acoustic difference calculation results in a quantitative value that describes the degree to which the real-time signal deviates from the basic feature library. This difference value can be used to determine the magnitude of changes in the acoustic signal during the welding process.

[0050] In some embodiments, when it is detected that the welding path is a long straight line area, the vibration characteristics and heat input gradient distribution of the welding head can be monitored in real time, and the welding parameters can be dynamically adjusted by combining data such as vibration accumulation, vibration change trend and acoustic signal to further optimize the welding stability and weld quality.

[0051] First, after detecting a long, straight welding path, the vibration characteristics of the welding head are monitored in real time, including vibration amplitude, vibration frequency, and vibration accumulation. The vibration accumulation is calculated by integrating the vibration amplitude per unit time and is used to quantify the cumulative effect of vibration during the welding process. When the vibration trend exceeds a preset trend threshold, the welding head's fine-tuning amplitude is dynamically adjusted based on the vibration characteristics (such as vibration frequency and vibration amplitude). The fine-tuning amplitude is proportional to the vibration trend; that is, the more severe the vibration change, the larger the fine-tuning amplitude. This operation can quickly respond to vibration anomalies and suppress weld interface instability caused by vibration accumulation.

[0052] Secondly, the heat input gradient distribution in the weld area is monitored in real time. Using thermal sensors or infrared imaging technology, heat input data along the weld path is acquired, and its mean and gradient distribution are calculated. When the heat input gradient distribution in a particular area exceeds the mean, that area is identified as having excessive heat input. In this case, the weld head speed is reduced to reduce heat input, while the axial pressure of the weld head is increased to improve the plastic flow of the material. By dynamically controlling the heat input and pressure distribution, weld defects caused by overheating, such as material burnout or internal porosity, can be prevented.

[0053] Next, if the cumulative vibration, vibration trend, or rate of change in the heat input gradient distribution exceeds a preset threshold, and if the acoustic variability is normal, the welding head speed is adjusted by a preset fine-tuning increment. This operation combines the dynamic monitoring results of vibration characteristics and heat input, and further considers the stability of acoustic characteristics, to ensure the comprehensive balance of various welding parameters during the process, helping to maintain process stability in a dynamically changing environment.

[0054] Finally, to ensure the effectiveness of adjustments, the welding head's vibration characteristics, heat input distribution, and acoustic signal changes are dynamically monitored throughout the entire process, and the response of the adjusted parameters is evaluated in real time. Through a closed-loop control system, the welding head's parameter adjustments are continuously optimized to prevent secondary disturbances from interfering with the weld interface, thereby ensuring the stability and accuracy of the welding path over long distances.

[0055] In summary, real-time monitoring of vibration characteristics and heat input gradient distribution, combined with comprehensive analysis of vibration accumulation, trends, and acoustic signals, can address the issues of abnormal vibration and uneven heat input in linear welding paths. Adjusting the vibration characteristics can suppress the cumulative vibration effect and prevent weld interface instability. Controlling the heat input gradient balances heat distribution, reducing the incidence of weld defects, thereby improving welding quality and efficiency and enhancing the adaptability and reliability of the welding system.

[0056] S103: When the acoustic difference exceeds a preset acoustic difference threshold, it is identified as an acoustic anomaly and the abnormal frequency characteristics of the acoustic anomaly are recorded.

[0057] Specifically, when the real-time calculated acoustic difference exceeds a preset acoustic difference threshold, an acoustic anomaly is identified. At this point, the acoustic signal is further analyzed to extract the anomaly's frequency characteristics. Specifically, the frequency characteristics of the acoustic anomaly are extracted by performing spectral analysis (such as fast Fourier transform (FFT)) on the real-time acoustic signal. Spectral analysis can identify key information in the anomaly signal, such as the main frequency, harmonic frequencies, and their amplitudes. These characteristic values ​​are recorded to characterize the acoustic anomaly.

[0058] During the processing, the dynamic changes of abnormal frequency characteristics must also be considered. If the amplitude or frequency distribution of the abnormal frequency changes over time, the signal will be continuously sampled and the abnormal frequency characteristics at a series of time points will be recorded to form a dynamic characteristic description of the abnormal signal.

[0059] Furthermore, noise filtering is performed to avoid misjudgments. For example, if background noise is present in the environment, which may cause short-term fluctuations in acoustic differences, filtering algorithms (such as bandpass filtering or wavelet transforms) can be used to remove the effects of this noise, ensuring that the recorded abnormal frequency characteristics accurately reflect the dynamic state of the welding process.

[0060] S104. According to the abnormal frequency characteristics, the rotation speed of the welding head is adjusted according to the fine-tuning amplitude and fine-tuning time, and after the fine-tuning time ends, the welding head is controlled to return to the rotation speed before the fine-tuning.

[0061] Specifically, the calculation of the fine-tuning amplitude and time is proportional to the acoustic difference. Acoustic difference is the quantitative deviation between the real-time acoustic signal and the basic acoustic feature library, reflecting the degree of dynamic change during the welding process. The fine-tuning amplitude is calculated using the formula: Fine-tuning amplitude = k1 × acoustic difference, where k1 is the adjustment sensitivity coefficient, which controls the speed adjustment range. The fine-tuning time is calculated using the formula: Fine-tuning time = k2 × acoustic difference, where k2 is the time adjustment coefficient, ensuring that the fine-tuning time matches the intensity and duration of the acoustic anomaly.

[0062] Based on the calculated results, the welding head speed is adjusted to the sum of the original speed and the fine-tuning amplitude (the direction of adjustment depends on the type of anomaly; resonance typically reduces the speed) and remains in this adjusted state for the fine-tuning time. After the fine-tuning time expires, the welding head speed is gradually restored to its original speed before fine-tuning using a smooth transition to avoid new dynamic instabilities caused by sudden speed changes.

[0063] In the above steps, the rotation speed of the welding head is dynamically adjusted according to the abnormal frequency characteristics of the acoustic anomaly, thereby improving the response capability and adaptability to high-frequency dynamic changes (such as resonance or chatter) during the welding process. Unlike the existing technology that relies on geometric offset correction and has the risk of response lag and disturbance, this step performs targeted rotation speed adjustment by calculating the fine-tuning amplitude and time proportional to the acoustic difference, and smoothly restores the original rotation speed after the fine-tuning time, avoiding damage to the welding interface caused by frequent compensation. The cumulative effect of dynamic changes is suppressed, preventing the weld from deviating from the ideal trajectory, while protecting the plastic flow state of the material during the welding process and reducing the risk of welding defects (such as tunnel defects and surface ripples). Ultimately, the stability of the welding process and the quality of the weld are improved, and the damage to the welding quality and process stability caused by the frequent correction action itself is avoided, providing a more reliable technical guarantee for improving welding accuracy.

[0064] S105 , acquiring the friction force, relative speed, and real-time acoustic signal after the fine-tuning time in real time, and calculating the friction energy within a preset unit time length.

[0065] Specifically, sensors are used to collect the friction force at the contact interface between the welding head and the workpiece in real time. The friction force is usually detected by a force sensor installed on the welding device, and its tangential component is the friction force. The relative speed is calculated by combining the encoder of the welding head drive device and the workpiece displacement sensor, representing the actual movement speed of the contact interface between the welding head and the workpiece. In addition, acoustic signals are captured in real time by acoustic sensors to assist in determining changes in the friction state. After these real-time parameters are collected, they are calculated using the friction energy formula: friction energy = ∫ (friction force × relative speed) dt, and discretized into friction energy = Σ (friction force × relative speed × unit time) within a unit time length. The energy generated by friction at the friction interface is quantified through real-time integral calculation.

[0066] In some embodiments, due to the high flexibility and complex dynamic characteristics of structural parts, the contact interface during welding may be affected by resonance or thermal expansion, resulting in a sudden increase in friction, manifested as abnormal contact pressure. This abnormality will cause a significant increase in friction energy, accompanied by the appearance of high-frequency acoustic noise (such as a high-amplitude signal at an abnormal frequency). If not handled in a timely manner, such dynamic changes may destroy the stability of the contact interface between the welding head and the workpiece, causing the plastic flow state of the interface material to be disturbed, thereby causing welding defects (such as uneven welds, tunnel defects) or reduced welding trajectory accuracy. To address the above problems, the rotation speed of the welding head is dynamically adjusted according to the amplitude of the change in friction energy and the characteristic frequency of the acoustic signal to reduce the friction peak caused by resonance or thermal expansion; at the same time, the pressure between the welding head and the workpiece is adjusted according to the duration of the abnormal signal to restore the stability of the contact interface. Throughout the process, the changes in friction energy and acoustic signals are continuously monitored through closed-loop control.

[0067] Through the above technical steps, the pressure anomaly caused by resonance or thermal expansion can be alleviated, the stability of the welding interface can be restored, the further development of welding defects can be prevented, the dynamic adaptability of the welding process can be improved, and the process requirements under complex working conditions can be met.

[0068] In other embodiments, when analyzing historical welding data and the CAD model of the current workpiece to predict that inherent low-frequency modal vibration modes may be activated in the unwelded path section, an appropriate anti-instability welding parameter combination can be selected and trial welding can be performed to dynamically evaluate the welding stability and optimize the welding parameters, thereby suppressing welding defects caused by low-frequency modal vibrations and ensuring welding quality and process stability in high-risk sections.

[0069] First, by analyzing historical welding data and the CAD model of the current workpiece, it is possible to predict whether the inherent low-frequency modal vibration mode of the unwelded path section is at risk of activation. Specifically, the dynamic characteristic parameters of the unwelded path section, such as the geometric characteristics, material characteristics, and structural support conditions, are extracted from the CAD model of the current workpiece, and a similarity comparison is performed in conjunction with the historical welding data. When the dynamic characteristic parameters of the path are similar to those in the historical data, the corresponding historical vibration frequency and historical vibration amplitude in the historical welding data are used to predict the vibration frequency and vibration amplitude of the similar unwelded path section. If the deviation between the predicted vibration frequency and the inherent modal vibration frequency is less than a preset activation frequency threshold, and the vibration amplitude exceeds a preset activation multiple of the historical average vibration amplitude, it is determined that the inherent low-frequency modal vibration mode is likely to be activated. Based on the predicted vibration frequency and vibration amplitude, an anti-destabilization parameter combination that matches the dynamic characteristics is selected from the anti-destabilization welding parameter library to suppress the occurrence of modal activation during the welding process.

[0070] Secondly, once a high-risk section where natural low-frequency modal vibrations are activated is identified, the selected predicted anti-instability parameter combination is applied over a preset trial length, while the probe acoustic signal and the probe dynamic stability of the weld joint are monitored in real time. The trial length is set to conduct a short-term dynamic verification within the high-risk section to determine whether the selected parameter combination can suppress low-frequency modal vibrations. The real-time monitoring of the probe acoustic signal, including spectral characteristics, signal intensity, and abnormal frequency changes, is used to assess the acoustic stability of the weld interface; the probe dynamic stability is determined by monitoring the operating status of the weld joint (such as vibration amplitude and motion trajectory).

[0071] Next, if the dynamic stability during the trial is within the preset stability threshold and the trial acoustic signal is within the preset acoustic signal threshold, the selected predicted anti-instability parameter combination is confirmed to meet the requirements and is applied to complete the welding of the corresponding high-risk section. This parameter application strategy ensures the reliability of the welding process based on verification and avoids the degradation of welding quality caused by the direct application of unverified parameters.

[0072] Finally, if the dynamic stability during the trial is outside the preset stability threshold or the trial acoustic signal exceeds the acoustic signal threshold, a preset small perturbation signal is superimposed on the control instructions of the welding head to further optimize the welding parameter combination. The role of the small perturbation signal is to break the possible vibration resonance state and help the system explore a new stable welding state. In this way, welding parameters can be dynamically adjusted under complex working conditions, gradually improving welding stability and ensuring welding quality.

[0073] In summary, by analyzing historical welding data and the current workpiece CAD model, predicting and dynamically verifying welding parameters in high-risk areas can mitigate the risk of low-frequency modal activation. The application of trial welding and small perturbation signals ensures the reliability and adaptability of parameter adjustments, thereby suppressing vibration instability, improving welding process stability and weld quality consistency, and providing a key guarantee for high-precision welding in complex working conditions.

[0074] S106. When the fluctuation coefficient of the friction energy exceeds the preset fluctuation coefficient threshold and the instantaneous value of the friction energy is lower than the preset friction energy threshold, or there is still acoustic anomaly in the real-time acoustic signal, and the shortest distance between the real-time position of the welding head and the preset device trajectory exceeds the preset offset threshold, adjust the position of the welding head to the device trajectory.

[0075] Specifically, during the welding process, when the following conditions are met, the welding head position adjustment will be triggered to realign it with the preset device trajectory. First, the dynamic abnormality condition must be met at the same time, that is, the fluctuation coefficient of the friction energy exceeds the preset fluctuation coefficient threshold and the instantaneous value of the friction energy is lower than the preset friction energy threshold, or there is still an acoustic abnormality in the real-time acoustic signal, that is, the acoustic difference exceeds the preset acoustic difference threshold. This indicates that the welding interface may have an unstable friction state, insufficient pressure or resonance problem. Secondly, the trajectory offset condition must also be met, that is, the shortest distance between the real-time position of the welding head and the preset device trajectory exceeds the preset offset threshold, indicating that the welding head has deviated from the ideal path, which may lead to a decrease in weld quality or an increase in welding defects. Position adjustment is triggered only when both the dynamic abnormality and trajectory offset conditions are met at the same time.

[0076] When the position adjustment is triggered, the welding head is repositioned to the preset trajectory through a position adjustment mechanism (such as a servo system or a multi-axis motion control system). The adjustment process includes path planning and closed-loop control. First, the position error between the real-time position of the welding head and the target trajectory point is collected, and the optimal path and target position for adjustment are calculated using a motion control algorithm (such as shortest path planning or smooth curve fitting). When performing the adjustment, the movement of the welding head is controlled in real time by the servo motor, while the friction energy and acoustic signals are continuously monitored to ensure that no new disturbances are generated during the adjustment process. The closed-loop control mechanism compares the real-time position of the welding head with the target position, continuously reducing the position error and ensuring the adjustment accuracy. In addition, in order to avoid secondary disturbances to the welding interface caused by the adjustment action, a flexible transition mechanism is adopted to gradually decelerate when approaching the target position, so that the welding head can smoothly transition to the trajectory.

[0077] In the above embodiment, by acquiring the acoustic signal of the welding head in real time, extracting the acoustic feature data and comparing it with the acoustic basic feature library, high-frequency dynamic anomalies caused by periodic forging pressure, small resonance or vibration during the welding process are identified, overcoming the defects of the traditional detection method that relies on geometric offset response lag and high-frequency dynamic changes that are difficult to capture. By recording the frequency characteristics of the acoustic anomaly and implementing speed adjustment of fine-tuning amplitude and fine-tuning time, the authenticity and severity of the anomaly are verified, while avoiding the disturbance damage caused by frequent high-frequency correction. Combined with the multi-dimensional analysis of friction force, relative speed and friction energy fluctuations, when the anomaly persists or the offset exceeds the threshold, the welding head position is dynamically adjusted to the preset trajectory, thereby suppressing the cumulative effect of dynamic changes and preventing the weld from deviating from the center line. This solution improves the real-time and adaptability of trajectory compensation, avoids the generation of welding defects such as defects and surface ripples, while ensuring the welding path accuracy and process stability, and improving the trajectory accuracy of ultra-long cantilever stir friction welding.

[0078] In other embodiments of the present application, when the acoustic difference during welding remains high but no abnormalities are present, the inherent structural characteristics of the workpiece may cause an offset in the acoustic characteristics, leading to misjudgment. Using the ultra-long cantilever friction stir welding trajectory offset correction method provided in this application, a digital twin model of the workpiece can be retrieved to simulate equivalent thermal-mechanical loads and analyze the inherent structural acoustic response characteristics. This can then be compared with real-time acoustic data to identify acoustic offsets caused by the inherent characteristics of the workpiece, dynamically update the acoustic basic feature library, and improve monitoring accuracy.

[0079] like Figure 2 FIG. 1 is another flow chart of a method for correcting an offset of an ultra-long cantilever friction stir welding track provided in an embodiment of the present application, comprising the following steps:

[0080] S201. Acquire an acoustic signal of a welding head in real time during a welding process, and extract real-time acoustic feature data of the acoustic signal.

[0081] S202: Compare the real-time acoustic feature data with the acoustic basic feature library, and calculate the acoustic difference.

[0082] S203. When the acoustic difference continues to be high and there is no acoustic anomaly in the acoustic difference, identify the high baseline and high reproducibility features of the acoustic difference and the corresponding workpiece type and welding section.

[0083] Specifically, by monitoring the changing curve of acoustic difference over a long period of time, its high baseline characteristics are identified. The acoustic difference data is segmented according to a fixed time window, and its mean and fluctuation range are calculated within each time segment. If the fluctuation range of the mean acoustic difference within multiple time segments is small and the central value remains within a specific high range, it is determined to have a high baseline characteristic. In addition, time series analysis methods (such as weighted moving average or exponential smoothing) are used to further confirm whether the trend of acoustic difference throughout the welding cycle is stable, eliminating the influence of short-term fluctuations. The high baseline characteristic refers to the continuous and stable state of acoustic difference within a high range, which is manifested by a small fluctuation range and a central value close to a fixed range. By identifying the high baseline characteristic, it is possible to determine whether the high acoustic difference is a normal state, avoiding misjudgment and unnecessary adjustments.

[0084] At the same time, acoustic variability records under the same or similar working conditions from historical welding tasks are extracted and compared with their variations across different tasks. The similarity between tasks is quantified by calculating correlations (e.g., the Pearson correlation coefficient) between the acoustic variability means. If the correlation coefficient exceeds a set threshold, it indicates that the acoustic variability is highly consistent under the same conditions. Furthermore, a reproducibility model is established using machine learning algorithms (e.g., support vector machines or cluster analysis) to correlate high reproducibility features with factors such as workpiece type and weld section. This approach identifies inherent patterns in high acoustic variability and provides a basis for optimizing welding parameters. The reproducibility of the acoustic features is calculated based on acoustic data from multiple repeated welding processes. High reproducibility indicates that the high acoustic variability is highly consistent under the same or similar welding conditions. This consistency may be due to the acoustic response characteristics of the specific workpiece material or the unique geometry of the weld area. By combining the workpiece material type, local characteristics of the weld path, and historical welding data, the cause of the high acoustic variability can be identified, such as the heat input characteristics of the weld section or the acoustic wave propagation pattern within the material.

[0085] In some cases, due to the presence of various curvatures, sharp angles, depressions, or irregular boundaries on the material surface, sound waves may reflect, overlap, or interfere in local areas during welding, resulting in a short-term increase in acoustic variability. This increase may be due to changes in the sound wave propagation path caused by specific geometric shapes, or it may be caused by a mismatch between the welding parameters and the local structural characteristics of the workpiece. If this phenomenon is not identified and addressed, it may be misjudged as an abnormality, leading to unnecessary parameter adjustments, affecting the stability of the welding process and the quality of the weld.

[0086] To address this issue, first, the changing trend of acoustic variability is monitored in real time. Baseline analysis is used to determine whether the elevated state is a temporary disturbance. The specific method involves calculating the mean and fluctuation amplitude of the acoustic variability within a time window and combining it with time series analysis to determine its stability. Secondly, the reproducibility characteristics of the acoustic variability are analyzed in conjunction with historical data. Multi-task comparison and correlation calculation are used to determine whether the elevated state is related to the workpiece geometry. If the elevated state is determined to be related to geometric characteristics and has high reproducibility, the welding path will be optimized or welding parameters (such as welding head speed, heat input, or pressure distribution) will be adjusted based on the analysis results to reduce the acoustic wave reflection effect and improve the dynamic stability of the weld.

[0087] The above technical steps enable the distinction between short-term perturbations and stable characteristics of acoustic variability, as well as the identification of their source and their correlation with the workpiece geometry. This avoids misjudging normal-to-high-sounding conditions. Furthermore, by optimizing the welding path or parameters, the adverse effects of acoustic reflections on the welding process are reduced, ensuring process stability and improving the consistency and reliability of weld quality.

[0088] S204 , calling a preset digital twin model of the workpiece type, and simulating the equivalent thermal-mechanical load applied to the corresponding welding section during the friction stir welding process in the digital twin model.

[0089] Specifically, during the welding process, the preset digital twin model corresponding to the workpiece type is called up, and the equivalent thermal-mechanical load applied to the welding section during the friction stir welding process is simulated in the model, which can provide predictions and guidance for welding parameter optimization and process control.

[0090] A digital twin model is a high-precision virtual model constructed based on workpiece CAD data, material properties, and welding process parameters. It dynamically maps the thermal and mechanical state of the actual welding process. By inputting the geometric characteristics, material properties, and process parameters of the weld zone into the digital twin model, finite element analysis (FEA) or other physical field simulation methods can be used to calculate the heat input distribution, stress and strain distribution, and material flow behavior generated during friction stir welding, and predict the impact of equivalent thermal and mechanical loads on the weld area.

[0091] According to the workpiece type, the digital twin model of the workpiece is retrieved. The model contains the workpiece's geometry, material properties (such as thermal conductivity, thermal expansion coefficient, elastic modulus, etc.), and welding path information. Then, the welding process parameters (such as welding head rotation speed, axial pressure, welding speed) are input into the model, and an equivalent thermal-mechanical load is applied in the simulation environment. The thermal load is simulated by the friction heat generated by the welding head and the workpiece surface during the welding process, and the force load is calculated by the axial pressure of the welding head and the plastic flow of the material caused by the welding movement. Through numerical calculation, the temperature distribution, stress distribution and material deformation of the welding section are obtained, and the location and type of possible defects (such as overheating, uneven welds or residual stress concentration) are predicted.

[0092] S205. Analyze the inherent structural acoustic response characteristics of the welding section through simulation calculation.

[0093] Specifically, the geometric features and material properties (such as density, elastic modulus, Poisson's ratio, etc.) of the weld section are extracted using a digital twin model or CAD model of the workpiece, and a local mesh is divided according to the welding path. On this basis, the inherent modal parameters of the weld section are calculated through modal analysis, including the natural frequency, modal vibration shape, and modal participation factor. The core of modal analysis is to establish the dynamic equations of the workpiece and obtain its inherent modal characteristics by solving the eigenvalue problem. Next, an equivalent thermal-mechanical load is applied to the weld section, and the stress distribution and deformation behavior of the material during the welding process are calculated in combination with thermal-mechanical coupled field simulation, thereby correcting the dynamic response characteristics of the weld section.

[0094] Frequency response analysis simulates the acoustic behavior of the weld zone under specific excitations. The excitation source can be friction generated at the interface between the weld head and the workpiece, or acoustic waves induced by local material vibrations during the welding process. By solving the dynamic equations, the response amplitude and phase distribution of the weld zone under different excitation frequencies can be determined, allowing the acoustic wave propagation path and acoustic pressure distribution characteristics to be calculated.

[0095] In some embodiments, the workpiece surface may contain features such as multiple curvatures, depressions, sharp corners, or irregular boundaries. These geometric complexities can significantly affect the propagation behavior of acoustic waves. Sound waves may reflect, scatter, and interfere in these areas, resulting in abnormally high local sound pressure or non-uniform distribution of acoustic response. This abnormal acoustic behavior can trigger dynamic instability in the weld zone, affecting the operation of the weld head and potentially leading to weld quality degradation or weld defects such as overheating, incomplete fusion, or an uneven weld interface. To address this issue, first, the workpiece geometric features are extracted through high-precision meshing, and the propagation path and reflection characteristics of the acoustic waves in different regions are defined based on boundary conditions. Second, acoustic finite element analysis or boundary element method is used to calculate the sound pressure distribution characteristics of the sound waves in complex areas, identifying areas of high sound pressure caused by local sound wave superposition or interference. Finally, by combining welding process parameters and material properties, the sound wave propagation conditions are optimized by adjusting parameters such as the weld head speed, welding path, or heat input, thereby reducing the impact of abnormal acoustic response on the welding process.

[0096] The above technical steps can identify problems caused by abnormal acoustic wave propagation in complex geometric areas and avoid acoustic instability by optimizing welding parameters. This not only improves the accuracy of simulation results but also enhances the stability of the welding process, ensuring consistent weld quality and process reliability.

[0097] S206: If the inherent structural acoustic response feature is consistent with the real-time acoustic feature data, the inherent structural acoustic response feature is replaced into the acoustic basic feature library.

[0098] Specifically, the inherent structural acoustic response characteristics of the welding section are obtained through simulation calculation, including information such as natural frequency, modal vibration shape, and sound pressure distribution. At the same time, the acoustic feature data recorded by the acoustic sensor during the welding process, such as frequency spectrum, sound pressure amplitude, phase change, etc., are collected in real time, and its key feature parameters are extracted through data processing. Feature matching algorithms (such as Pearson correlation coefficient, dynamic time warping algorithm, etc.) are used to perform similarity analysis on the simulation features and real-time features. If the similarity index is higher than the set threshold, it is determined that the two sets of feature data are consistent. The acoustic features calculated by simulation are used as new reference features and replaced or supplemented to the acoustic basic feature library. The update process of the feature library is usually carried out in an incremental manner, that is, new features are added on the basis of retaining historical data to ensure the comprehensiveness and dynamic adaptability of the feature library.

[0099] S207: When the acoustic difference exceeds a preset acoustic difference threshold, identify it as an acoustic anomaly and record the abnormal frequency characteristics of the acoustic anomaly.

[0100] S208. Perform real-time imaging of the heat flux density in an area of ​​the size of a preset heat flux monitoring area below the welding head, and adjust the rotation speed of the welding head according to the fine-tuning amplitude and fine-tuning time, record the adjustment transient response characteristics of the heat flux density at the beginning of the fine-tuning time, and control the welding head to return to the rotation speed before fine-tuning after the fine-tuning time ends.

[0101] Specifically, a pre-set heat flux monitoring area is set below the critical section of the weld path, and the heat flux density in this area is imaged in real time using an infrared thermal imager or thermopile sensor. Heat flux density imaging captures the thermal radiation generated during the welding process and, in conjunction with a heat conduction model, calculates the heat flux distribution at each point in time. The real-time imaging results are presented as a high-resolution heat flux map, which is continuously updated to reflect the dynamic heat input status of the weld section.

[0102] The welding head's rotational speed is adjusted to implement a short-term perturbation using fine-tuning amplitude and fine-tuning time as parameters. Speed ​​fine-tuning is performed with small changes, maintained within the preset fine-tuning time range. Speed ​​fine-tuning directly affects the rate of frictional heat generation between the welding head and the workpiece, thereby causing changes in the heat flux density. Through this perturbation process, the transient response characteristics of the heat flux density adjustment are observed and recorded, including the heat flux density change amplitude, response time, and recovery process curve.

[0103] After the fine-tuning operation is completed, the welding head speed is controlled to return to the set value before fine-tuning to ensure the overall stability and continuity of the welding process. The smooth transition of speed recovery is achieved by the PID control algorithm to avoid the secondary impact of large parameter jumps on welding quality.

[0104] S209: Input the speed adjustment data and the adjustment transient response characteristics into a preset ideal heat generation and heat transfer dynamic mathematical model of stir friction welding to obtain actual heat generation efficiency and actual heat flux density distribution.

[0105] Specifically, the welding head speed values ​​before and after adjustment are recorded, and the speed adjustment data (i.e., the speed difference) is calculated as the input variable. Simultaneously, real-time thermal flow imaging is used to capture the transient response characteristics of the adjustment, including the change in heat flux density, response time, and distribution. These transient characteristic data reflect the transfer and diffusion of heat input in the welding section.

[0106] Next, the above data was input into a dynamic mathematical model of ideal heat generation and heat transfer in friction stir welding. This model is based on the principles of thermodynamics and heat transfer, comprehensively considering frictional heat generation between the weld head and the workpiece surface, the thermal conductivity characteristics of the material, and dynamic heat diffusion. The model primarily consists of the following components: a frictional heat generation model, which calculates the thermal power generated at the interface based on weld head rotational speed, axial pressure, and the material friction coefficient; a heat transfer model, which uses the Fourier heat conduction equation to simulate the diffusion of heat flux within the workpiece, taking into account the thermal conductivity, density, and specific heat capacity of the material; and a dynamic response model, which couples the heat source model with the conduction model and describes the temporal and spatial variations of the heat flux density using partial differential equations. Numerically solving these equations predicts the transient heat flux distribution and heat generation efficiency within the weld zone, providing theoretical support for welding process optimization. Solving the model reveals the actual heat generation efficiency and heat flux density distribution within the weld zone.

[0107] In some embodiments, after the speed adjustment data and the adjusted transient response characteristics are input into the preset ideal heat generation and heat transfer dynamic mathematical model of stir friction welding, if it is found that the adjusted transient response characteristics are related to the rotation period of the welding head and match the preset welding head specific defect acoustic heat flow coupling fingerprint, the potential defect type and location of the welding head can be further judged by analyzing the heat flow fluctuation characteristics, and the stability of the welding process and the weld quality can be optimized by dynamically adjusting the speed and motion trajectory of the welding head.

[0108] By analyzing and adjusting the transient response characteristics, the heat flux fluctuation characteristics associated with the welding head rotation period are identified, and the heat flux fluctuation frequency, fluctuation phase, and fluctuation amplitude are locked. This process uses real-time heat flux monitoring data combined with the kinematic characteristics of the welding head to separate the periodic fluctuations from the complex heat flux density signal through frequency domain analysis or time domain signal processing methods. This can capture heat flux anomalies caused by the periodic movement of the welding head, identify possible mechanical defects or interface instability, and provide basic data for subsequent defect diagnosis.

[0109] Based on the correlation between the locked heat flux fluctuation amplitude and phase, and the abnormal spectral characteristics corresponding to the acoustic anomaly, these characteristics are compared with a preset weld joint defect fingerprint library to determine the type and location of the weld joint defect. This defect fingerprint library, established based on extensive experimental and theoretical analysis, covers the characteristic information of common defects such as wear, imbalance, and cracks. This comparative analysis can quickly locate defects and clarify their nature, providing targeted guidance for subsequent optimization of welding parameters and path adjustments.

[0110] The weld head speed adjustment amplitude is calculated based on the ratio of the heat flux fluctuation amplitude to a preset fluctuation amplitude deviation threshold. Simultaneously, the adjustment phase point is determined based on the heat flux fluctuation phase. This step aims to adjust the speed at the optimal moment in the weld head's cyclical motion, minimizing interference with the welding process while ensuring a dynamic balance of heat input. By controlling both timing and amplitude, the thermal stability and process controllability of the welding process can be improved.

[0111] At the determined adjustment phase point, the welding head speed is adjusted a second time according to the calculated speed adjustment range. The heat flux fluctuation amplitude and phase distribution characteristics before and after the adjustment are captured in real time. This adjustment aims to further verify and optimize the heat flux distribution by changing the welding head's motion state, while also providing high-precision input data for subsequent dynamic trajectory adjustments. This approach can reduce abnormal heat flux fluctuations during welding and improve the uniformity of welding heat input.

[0112] By analyzing the heat flux fluctuation amplitude and phase distribution characteristics after the second adjustment, the real-time offset of the welding head along the welding trajectory is calculated—that is, the deviation between the welding head center point and the target trajectory center point. Subsequently, based on the offset trend, a multi-point interpolation method is used to generate a compensation path for the welding head offset. This compensation path, calculated by fitting the real-time offset trend, corrects the welding head's trajectory, preventing offset accumulation and uneven welds during welding.

[0113] The welding head's trajectory is adjusted according to the calculated offset compensation path, and a smoothing algorithm is used to ensure the continuity and smoothness of the welding path. This operation avoids uneven welds or surface defects caused by trajectory adjustment, while ensuring the overall stability and weld quality of the welding process.

[0114] Through these steps, potential weld defects can be identified and located in real time, and the weld head rotation speed and trajectory can be dynamically optimized to ensure stable heat input and smooth weld seam during the welding process. This technical step reduces the risk of weld defects and improves the consistency of weld quality, providing a strong technical support for friction stir welding under complex working conditions.

[0115] S210. When the heat generation efficiency deviation value exceeds the preset heat generation efficiency deviation value threshold or the heat flux density distribution deviation value exceeds the preset heat flux density distribution deviation threshold, an interface friction state instability warning message is issued, the axial force of the welding head is adjusted to the preset axial force optimization range, and the tool inclination angle of the welding head is adjusted according to the transient response characteristics.

[0116] Specifically, the absolute value of the difference between the actual heat generation efficiency and the preset theoretical heat generation efficiency is calculated through real-time monitoring as the heat generation efficiency deviation value. At the same time, the real-time heat flux imaging technology captures the actual heat flux density distribution, compares it with the preset theoretical heat flux density distribution, and calculates the absolute value of the difference between the two as the heat flux density distribution deviation value. The preset thresholds are determined based on the welding material, workpiece geometry and process conditions, and there are heat generation efficiency deviation thresholds and heat flux density distribution deviation thresholds. If any deviation value exceeds the threshold range, it is determined that the interface friction state is unstable and a warning message is issued in real time.

[0117] Next, in order to restore the stable state, the axial force of the welding head is first adjusted to the preset axial force optimization range. The axial force is a key parameter that affects the frictional heat generation between the welding head and the workpiece surface, and the main basis for adjustment is the deviation amplitude of the actual heat generation efficiency. By increasing or decreasing the axial force, the frictional heat generation rate can be adjusted, thereby restoring the stability of the interface heat input. In addition, the tool inclination angle of the welding head is dynamically optimized based on the transient response characteristics of the heat flux density. The adjustment of the tool inclination angle will change the force distribution and material flow direction at the contact interface between the welding head and the workpiece, thereby affecting the spatial distribution of the heat flux density. By adjusting the inclination angle, the problem of uneven heat flux density distribution can be improved.

[0118] In some embodiments, due to the high thermal conductivity or large heat capacity of the material, heat diffuses or is absorbed rapidly inside the workpiece during welding, resulting in insufficient local heat input in the welding section and excessive deviation in the heat generation efficiency. In this case, the interface friction state may become unstable, thereby affecting the temperature field and material flow state of the welding section, which may eventually lead to weld quality problems, such as incomplete welding, insufficient fusion of materials, or insufficient weld strength. To solve the above problems, first, the axial force of the welding head is increased to a preset axial force optimization range to improve the interface friction heat generation efficiency and supplement the heat input of the welding section; at the same time, according to the transient response characteristics of the heat flux density, the tool inclination angle of the welding head is dynamically adjusted to optimize the material flow behavior and improve the spatial distribution of the heat flux density. By adjusting the linkage operation of the axial force and the tool inclination angle, it is possible to dynamically adapt to the complex heat diffusion environment of the workpiece to ensure that the welding section obtains uniform and sufficient heat input.

[0119] The above technical steps can restore heat input stability in the weld zone and optimize heat flux distribution, ensuring weld quality. This solution significantly reduces the risk of welding defects caused by insufficient heat input under special working conditions, while also improving the adaptability and efficiency of the welding process, providing reliable technical support for high-quality welding of complex materials and thick-walled workpieces.

[0120] S211 , obtaining in real time the friction force, relative speed, and real-time acoustic signal after the fine-tuning time, and calculating the friction energy within a preset unit time length.

[0121] S212. When the fluctuation coefficient of the friction energy exceeds the preset fluctuation coefficient threshold and the instantaneous value of the friction energy is lower than the preset friction energy threshold, or there is still acoustic anomaly in the real-time acoustic signal, and the shortest distance between the real-time position of the welding head and the preset device trajectory exceeds the preset offset threshold, adjust the position of the welding head to the device trajectory.

[0122] Steps S201-S202, S207, S211-S212 and Figure 1 In the illustrated embodiment, steps S101 - S103 and S105 - S106 are similar, and the descriptions of steps S101 - S103 and S105 - S106 may be referred to, and will not be repeated here.

[0123] In the above embodiment, when the acoustic difference is continuously high and there is no abnormality, its high baseline and high reproducibility characteristics are identified, and the digital twin model is retrieved in combination with the workpiece type and welding section to simulate the equivalent thermal-mechanical load applied to the welding section during the stir friction welding process, and the inherent structural acoustic response characteristics are analyzed through simulation calculations. If the simulation results are consistent with the real-time acoustic feature data, the inherent structural acoustic response characteristics are replaced with the acoustic basic feature library, thereby dynamically optimizing the adaptability of the basic feature library. This solution identifies welding anomalies and dynamically optimizes welding parameters through multi-dimensional linkage analysis of acoustic signals, heat flux density and friction energy, overcoming the problems of response lag and insufficient dynamic adaptability of traditional geometric detection methods, and ensuring the smoothness of the welding path and trajectory accuracy. Ultimately, this method significantly improves the welding quality and process stability of ultra-long cantilever stir friction welding, meeting the needs of high-precision welding under complex working conditions.

[0124] The following introduces an exemplary trajectory deviation correction system 300 provided in an embodiment of the present application. Figure 3 3 is a schematic diagram of an exemplary hardware structure of the trajectory deviation correction system 300 provided in an embodiment of the present application.

[0125] In some embodiments, the trajectory deviation correction system 300 is a computer device or the trajectory deviation correction system 300 includes a computer device. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface, and in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, the method in the embodiment of the present application is implemented.

[0126] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0127] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0128] As used in the above embodiments, the term “when…” may be interpreted as “if…” or “after…” or “in response to determining…” or “in response to detecting…”, depending on the context. Similarly, the phrases “upon determining…” or “if (stated condition or event) is detected” may be interpreted as “if determining…” or “in response to determining…” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0129] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product comprises one or more computer instructions. When loaded and executed on a computer, the computer program instructions fully or partially perform the processes or functions described in the embodiments of this application. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium accessible by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be magnetic media (e.g., floppy disk, hard disk, magnetic tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state drive).

[0130] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for correcting the offset of an ultra-long cantilever friction stir welding track, characterized in that: include: Acquiring an acoustic signal of the welding head in real time during the welding process, and extracting real-time acoustic feature data of the acoustic signal; Comparing the real-time acoustic feature data with an acoustic basic feature library and calculating an acoustic difference; the acoustic basic feature library is constructed by acquiring a no-load acoustic signal during no-load operation and extracting feature data of the no-load acoustic signal; When the acoustic difference exceeds a preset acoustic difference threshold, it is identified as an acoustic anomaly and the abnormal frequency characteristics of the acoustic anomaly are recorded; According to the abnormal frequency characteristics, the rotation speed of the welding head is adjusted according to the fine-tuning amplitude and fine-tuning time, and after the fine-tuning time ends, the welding head is controlled to return to the rotation speed before fine-tuning; the fine-tuning amplitude and the fine-tuning time are both proportional to the acoustic difference; Acquire the friction force, relative speed and real-time acoustic signal after the fine-tuning time in real time, and calculate the friction energy within a preset unit time length; The friction energy is the energy generated by friction at the contact interface between the welding head and the workpiece; When the fluctuation coefficient of the friction energy exceeds a preset fluctuation coefficient threshold and the instantaneous value of the friction energy is lower than a preset friction energy threshold, or the real-time acoustic signal still has the acoustic anomaly and the shortest distance between the real-time position of the welding head and the preset device trajectory exceeds a preset offset threshold, the position of the welding head is adjusted to the device trajectory; The fluctuation coefficient is a quantitative value of the stability of the friction energy in the unit time.

2. The method according to claim 1, characterized in that The adjusting the rotation speed of the welding head according to the fine-tuning amplitude and fine-tuning time based on the abnormal frequency characteristics, and controlling the welding head to return to the rotation speed before fine-tuning after the fine-tuning time ends, specifically includes: Real-time imaging of the heat flux density within an area of ​​a preset heat flux monitoring area below the welding head is performed, and the rotational speed of the welding head is adjusted according to the fine-tuning amplitude and fine-tuning time. The adjustment transient response characteristics of the heat flux density at the beginning of the fine-tuning time are recorded, and after the fine-tuning time ends, the welding head is controlled to return to the rotational speed before the fine-tuning; Inputting the speed adjustment data and the adjustment transient response characteristics into a preset ideal heat generation and heat transfer dynamic mathematical model of stir friction welding to obtain the actual heat generation efficiency and actual heat flux density distribution; the speed adjustment data is the difference between the speed values ​​before and after the adjustment; When the heat generation efficiency deviation value exceeds the preset heat generation efficiency deviation value threshold or the heat flux density distribution deviation value exceeds the preset heat flux density distribution deviation threshold, an interface friction state instability warning message is issued, the axial force of the welding head is adjusted to the preset axial force optimization range, and the tool inclination angle of the welding head is adjusted according to the adjustment transient response characteristics; the heat generation efficiency deviation value is the absolute value of the difference between the actual heat generation efficiency and the preset theoretical heat generation efficiency; the heat flux density distribution deviation value is the absolute value of the difference between the actual heat flux density distribution and the preset theoretical heat flux distribution.

3. The method according to claim 2, characterized in that After inputting the speed adjustment data and the adjustment transient response characteristics into a preset ideal heat generation and heat transfer dynamic mathematical model of friction stir welding, the method further includes: If it is found that the adjusted transient response characteristics are related to the rotation period of the welding head and match the preset welding head specific defect acoustic heat flow coupling fingerprint, the heat flow fluctuation frequency, heat flow fluctuation phase and heat flow fluctuation amplitude of the periodic heat flow fluctuation are obtained by analyzing the adjusted transient response characteristics; According to the correlation between the heat flux fluctuation amplitude, the heat flux fluctuation phase, and the abnormal spectrum characteristics corresponding to the acoustic anomaly, the preliminary defect type and preliminary defect location of the weld joint are determined by comparing with a preset weld joint defect fingerprint library; Calculating a rotation speed adjustment amplitude of the welding head based on a ratio of the heat flux fluctuation amplitude to a preset fluctuation amplitude deviation threshold, and calculating an adjustment phase point for the rotation speed adjustment based on the heat flux fluctuation phase, and performing a second adjustment on the rotation speed of the welding head at the adjustment phase point according to the rotation speed adjustment amplitude; Obtaining the second heat flux fluctuation amplitude and the second heat flux fluctuation phase distribution characteristics before and after the second adjustment, and calculating the real-time offset of the welding head on the welding trajectory; the real-time offset is the deviation value between the center point of the welding head and the center point of the target trajectory; Based on the real-time offset, a multi-point interpolation method is used to generate an offset compensation path for the welding head; the compensation path is calculated by fitting a change trend of the real-time offset; The motion trajectory of the welding head is readjusted according to the offset compensation path.

4. The method according to claim 1, wherein After comparing the real-time acoustic feature data with the acoustic basic feature library and calculating the acoustic difference, the method further includes: When it is detected that the welding path is a long straight line area, the vibration characteristics of the welding head are monitored in real time, and the vibration accumulation and vibration change trend are calculated, and the heat input gradient distribution in the welding area is monitored at the same time; When the vibration accumulation, the vibration change trend or the change rate of the heat input gradient distribution exceeds the corresponding preset threshold and the acoustic difference does not have the acoustic anomaly, the rotation speed of the welding head is adjusted according to the preset fine-tuning parameters.

5. The method according to claim 4, characterized in that After monitoring the vibration characteristics of the welding head in real time when the welding path is detected to be a long straight line area, calculating the vibration accumulation and vibration change trend, and monitoring the heat input gradient distribution in the welding area, the method further includes: When the vibration change trend in the vibration characteristic exceeds a preset change trend threshold, adjusting a fine-tuning parameter according to the vibration frequency and vibration amplitude in the vibration characteristic; the fine-tuning parameter is proportional to the vibration change trend; According to the heat input gradient distribution in the welding path direction, the welding head rotation speed corresponding to the area where the heat input is too high is reduced, and the axial pressure of the welding head is increased in the area where the heat input is too high; the area where the heat input is too high is the area where the heat input gradient distribution exceeds the mean.

6. The method according to claim 1, characterized in that Before obtaining the friction force, relative speed and real-time acoustic signal after the fine-tuning time in real time and calculating the friction energy within a preset unit time length, the method further includes: When, by analyzing historical welding data and a CAD model of the current workpiece, it is predicted that an inherent low-frequency modal vibration mode is activated in an unwelded path section, a corresponding predicted anti-instability parameter combination is selected from a preset anti-instability welding parameter library based on the predicted vibration frequency and predicted vibration amplitude; the inherent low-frequency modal vibration mode is activated when the deviation between the vibration frequency and the inherent modal vibration frequency is less than a preset activation frequency threshold and the vibration amplitude exceeds a preset activation multiple of the historical average vibration amplitude; When the welding head reaches a high-risk section corresponding to the activation of the natural low-frequency mode shape, the predicted anti-instability parameter combination is applied with a preset trial length, and the trial acoustic signal and the trial dynamic stability of the welding head are monitored in real time; If the tentative dynamic stability is within a preset stability threshold range and the tentative acoustic signal is within a preset acoustic signal threshold range, completing welding of the high-risk section by applying the predicted anti-instability parameter combination; If the probe dynamic stability is not within a preset stability threshold range or the probe acoustic signal is not within a preset acoustic signal threshold range, a preset small disturbance signal is superimposed on the control instruction of the welding head.

7. The method according to claim 1, characterized in that After comparing the real-time acoustic feature data with the acoustic basic feature library and calculating the acoustic difference, the method further includes: When the acoustic difference is continuously high and the acoustic anomaly does not exist in the acoustic difference, identifying a high baseline, high reproducibility feature of the acoustic difference and a corresponding workpiece type and welding section; Retrieving a preset digital twin model of the workpiece type, and simulating in the digital twin model an equivalent thermal-mechanical load applied to the corresponding welding section during the friction stir welding process; Analyzing the inherent structural acoustic response characteristics of the welding section through simulation calculation; If the inherent structural acoustic response feature is consistent with the real-time acoustic feature data, the inherent structural acoustic response feature is replaced into the acoustic basic feature library.

8. A trajectory deviation correction system, characterized in that: The trajectory deviation correction system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the trajectory deviation correction system to perform the method described in any one of claims 1 to 7.

9. A computer program product comprising instructions, characterized in that When the computer program product is run on a trajectory offset correction system, the trajectory offset correction system is caused to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a trajectory offset correction system, the trajectory offset correction system is caused to perform the method according to any one of claims 1 to 7.

Images