A method and system for dynamic alignment in magnetically assisted welding

Abstract

This invention belongs to the field of welding-related technology, specifically providing a magnetically controlled assisted welding dynamic alignment method and system. The system includes: a magnetic control device, an infrared temperature measuring device, a motor, and a control system. The magnetic control device is mounted on the welding torch, and the infrared temperature measuring device is mounted on the magnetic control device. The infrared temperature measuring device includes a reciprocating infrared temperature sensor for real-time acquisition of the temperature on both sides of the molten pool. The motor drives the magnetic control device to rotate and adjust its position along the welding torch axis. The control system processes the temperature data in real time, calculates the deflection angle, and outputs a motor control signal. This invention can ensure the relative position of the magnetic field and the welding trajectory during complex path welding, thereby guaranteeing welding performance and quality under the assistance of the magnetic control device.

Description

Technical Field

[0001] This invention mainly relates to the field of welding-related technologies, specifically a magnetically assisted welding dynamic alignment method and system. Background Technology

[0002] Gas metal arc welding (GMAW) has become the main welding process for complex structures in marine engineering (especially robotic welding) due to its high efficiency, high quality, ease of automation, and wide adaptability. By precisely controlling the movement of the welding torch along a complex trajectory and maintaining stable protection and droplet transfer, it effectively solves the space, quality, and efficiency challenges in the manufacturing of marine structures and aerospace components. However, in industrial practice, increasing welding speed can lead to a series of quality problems such as insufficient penetration, lack of fusion, undercut, porosity, poor weld bead formation (hump), and increased spatter, and significantly increases the difficulty of operation and the requirements for equipment.

[0003] Optimizing the gas metal arc welding (GMAW) process by introducing magnetic control devices (such as external magnetic fields or electromagnetic stirring devices) is an effective technical approach to improve welding speed and production efficiency. Its core lies in using magnetic force to actively intervene in arc behavior and molten pool flow, thereby overcoming the process limitations of traditional GMAW at high speeds. However, existing methods and devices are only suitable for single-pass straight welding. When facing complex welding paths, the magnetic field and welding trajectory will rotate relative to each other, causing changes in the behavior of the molten metal droplets and the welding arc, leading to alterations in the liquid metal state of the molten pool and making it difficult to optimize weld performance. Summary of the Invention

[0004] To address the shortcomings of current technologies, this invention, based on existing technologies and practical applications, provides a magnetically controlled assisted welding dynamic alignment method and system. This method ensures the relative position of the magnetic field and the welding trajectory during complex path welding, thereby guaranteeing welding performance and quality assisted by the magnetically controlled device.

[0005] The technical solution of the present invention is as follows:

[0006] According to one aspect of the present invention, a magnetically assisted welding dynamic alignment method is provided, comprising the following steps:

[0007] S1. Collect temperature data in the detection area and locate the molten pool;

[0008] S2. Real-time acquisition of temperature data on both sides of the molten pool, noise reduction processing of the original temperature data, calculation of the temperature difference on both sides of the molten pool after filtering, obtaining the deflection angle between the molten pool symmetry axis and the magnetic control device axis, as a quantitative indicator of the alignment state of the magnetic control device.

[0009] S3. Based on the temperature difference and its trend on both sides of the molten pool, control commands are generated using an incremental PID algorithm, and dead zone processing is used to limit the maximum amplitude of the output control quantity.

[0010] S4. The control commands generated by the incremental PID algorithm are converted into target angle commands for the motor, and the position is adjusted by the motor-driven magnetic control device to achieve real-time matching between the magnetic field direction and the arc motion.

[0011] Furthermore, step S1 specifically includes:

[0012] S11. Collect the maximum value of the temperature at the current location, and determine whether the maximum value of the temperature is greater than or equal to the threshold for determining the presence of the molten pool. If yes, it is determined that the molten pool is within the detection range, and proceed to step S2; otherwise, it is determined that the molten pool is not within the detection range, and proceed to step S12.

[0013] S12. Search other locations along the circumference at a constant angular velocity and execute step S11. If no molten pool is found after scanning one full circle, it is considered abnormal, the search is stopped and an alarm is triggered.

[0014] Furthermore, in step S2, the method for denoising the original temperature data is as follows:

[0015] The raw temperature data is processed using a recursive filtering algorithm. The formula for processing the temperature data on both sides of the molten pool is as follows:

[0016]

[0017] The initial conditions are:

[0018]

[0019] in, This indicates the real-time temperature at the i-th position on the left side of the molten pool; The value represents the real-time temperature at the i-th position on the right side of the molten pool; α∈(0,1] represents the smoothing factor, which determines the filtering strength. The larger α is, the greater the weight of the new sampled value, the smaller the filtering lag but the weaker the noise reduction effect. The smaller α is, the greater the weight of the historical data, the better the noise reduction effect but the lag increases. This indicates that the temperature was measured at the i-th position on the left side of the molten pool. This indicates that the temperature was measured at the i-th position on the right side of the molten pool.

[0020] The formula for calculating the temperature difference between the two sides of the molten pool after filtering is as follows:

[0021]

[0022] Where ΔT(i) represents the real-time temperature difference at position i, and its positive and negative values ​​and amplitudes represent the direction and degree of deviation, respectively. When the axis of the magnetic control device is coplanar with the axis of symmetry of the molten pool, the temperature distribution is symmetrical. When the magnetic control device deflects, the two sides are in different temperature zones.

[0023] Furthermore, in step S3, the incremental PID algorithm dynamically quantifies the system asymmetry and combines historical state memory and temperature-related corrections, as shown in the following expression:

[0024]

[0025]

[0026] in, (This represents the increment of the output control quantity, used to drive the motor to adjust the angle). , , The temperature difference between the current time, the previous time, and the time before that is used as the tracking error term for the current region in the PID control algorithm; K p The proportional term is used to instantly respond to changes in the deviation trend; the larger the deviation, the stronger the compensation. K i K represents the integral term, used to accumulate historical deviations and eliminate static errors; d This represents the differential term, used to predict future changes in deviation and suppress overshoot. This indicates an asymmetric compensation term, used to compensate for differences between the two sides of the molten pool.

[0027] Furthermore, in step S3, the asymmetric compensation term The expression is as follows:

[0028]

[0029] in, This represents the temperature difference direction factor, a sign function that determines the compensation polarity based on the current temperature difference direction. If ΔT>0, it is +1; if ΔT<0, it is -1. This represents the average temperature of the left and right filters, indicating the current absolute temperature level. The asymmetry correction coefficient represents the strength of the correction to the inherent asymmetry of the system and is calibrated experimentally. This represents the historical average temperature range.

[0030] Furthermore, in step S3, the use of dead-time processing to limit the maximum amplitude of the output control quantity includes:

[0031] Set a dead zone threshold ϵ. When the temperature difference is less than the dead zone threshold ϵ, increase the output control quantity. To return to zero, the expression is as follows:

[0032]

[0033] Among them, the dead zone threshold ϵ is a positive constant set according to the noise level of temperature acquisition, and the dead zone threshold ϵ is greater than the noise amplitude.

[0034] Furthermore, step S4 specifically includes:

[0035] The control commands generated by the incremental PID algorithm are converted into target angle commands for the motor, as shown in the following expression:

[0036]

[0037] in, This represents the increment of the control quantity output by the incremental PID algorithm; Kscale represents the proportional factor, used to calibrate the mapping relationship between the control quantity and the physical angle. Indicates the current angle of the magnetic control device; Indicates the target angle of the magnetic control device;

[0038] Target angle The linear mapping is to the PWM duty cycle, which drives the motor to rotate to a specified position. The expression is as follows:

[0039]

[0040] in, This indicates the duty cycle of the PWM signal that needs to be output to the motor driver. , These represent the maximum and minimum duty cycles, respectively. , These represent the maximum and minimum allowable rotation angles of the magnetic control device, respectively. .

[0041] Furthermore, in step S4, an adaptive speed regulation mechanism is introduced, making the motor's reference speed proportional to the rate of change of the target angle, as expressed below:

[0042]

[0043] In the formula, Indicates the reference speed of the motor. Indicates the speed response gain; Indicates the rate of change of the target angle; This represents the minimum sustaining speed required to maintain the static friction of the system.

[0044] When the error between the current position and the target position is large, an additional acceleration is introduced to enable the motor to start with the maximum allowable acceleration, as shown in the following expression:

[0045]

[0046] in, Indicates additional acceleration; Indicates the acceleration factor; Indicates the error threshold; When the error is less than this value, this compensation term is zero to avoid system jitter in the small error range;

[0047] The final commanded speed of the motor is determined by the reference speed and the integral of the acceleration compensation, as shown in the following expression:

[0048]

[0049] in, This indicates the final speed of the motor.

[0050] According to another aspect of the present invention, a magnetically controlled assisted welding dynamic alignment system is provided for implementing the above-described method. The system includes: a magnetic control device, an infrared temperature measuring device, a motor, and a control system.

[0051] The magnetic control device is mounted on the welding torch, and the infrared temperature measuring device is mounted on the magnetic control device. The infrared temperature measuring device includes a reciprocating infrared temperature sensor for real-time acquisition of the temperature on both sides of the molten pool. The motor is used to drive the magnetic control device to rotate and adjust its position along the welding torch axis. The control system is used to process the temperature data in real time, calculate the deflection angle, and output the motor control signal.

[0052] Furthermore, the control system includes a coarse positioning unit and a precise positioning unit;

[0053] The coarse positioning unit includes:

[0054] The status judgment module is configured to determine whether the molten pool is within the detection range of the infrared temperature sensor based on the collected temperature data.

[0055] The search module is configured to control the motor to drive the infrared temperature sensor to perform a 360° rotation scan when the status judgment module cannot detect the molten pool.

[0056] The precise positioning unit includes:

[0057] The acquisition module is configured to acquire temperature data from both sides of the molten pool in real time, process the temperature data, and calculate the temperature difference after filtering.

[0058] The control module is configured to generate motor control commands from the temperature difference values ​​acquired by the acquisition module using an incremental PID algorithm.

[0059] The power drive module is configured to convert the commands generated by the control module into precise motor movements, drive the magnetic control device to adjust its position, and achieve real-time matching between the magnetic field direction and the welding trajectory.

[0060] The beneficial effects of this invention are:

[0061] 1. Dynamic Magnetic Field Real-Time Alignment System: This invention develops a closed-loop control method and system that can drive the magnetic control device to rotate and align with the molten pool in real time. It continuously scans the temperature on both sides of the molten pool through a unique mechanical oscillating infrared thermometer, intelligently judges the angular deviation between the axis of the magnetic control device and the axis of symmetry of the molten pool (i.e., the center line of the welding heat source), and instructs the external rotor motor to compensate for this angle in real time. This allows the magnetic field direction of the magnetic control device to always automatically maintain the best match with the welding direction on complex curves or three-dimensional welding paths, breaking through the limitation of existing technologies that can only be used for straight-line welding.

[0062] 2. Hierarchical Intelligent Control Strategy: The system employs an innovative dual-layer control logic of "coarse positioning and precise positioning." Upon startup, the system automatically scans and locates the molten pool, ensuring rapid target positioning. After finding the target, the system enters a precision control mode: first, it uses real-time acquired temperature data from both sides of the molten pool and advanced signal processing algorithms (such as dynamic noise reduction and asymmetric compensation) to accurately calculate the deviation angle; then, it uses an improved intelligent control algorithm (integrating historical error correction and a specialized compensation mechanism for the system's inherent asymmetry) to generate precise motor drive commands. This hierarchical strategy significantly improves the system's response speed and final control accuracy.

[0063] 3. Integrated Hardware Collaborative Innovation: The entire device employs an innovative mechanical design, integrating the magnetic control unit, a specially designed swingable infrared temperature probe, and an external rotor motor into a compact unit, directly mounted on the welding torch. The magnetic control unit and temperature probe are rigidly connected and rotate as a whole, eliminating relative motion errors and ensuring strict alignment between the temperature measurement reference and the magnetic field direction. This integrated design is the physical basis for the effective operation of the algorithm, jointly achieving high-precision dynamic alignment. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0065] Figure 2 This is a schematic diagram of the system structure of the present invention.

[0066] Figure 3 This is a three-dimensional structural diagram of the welding torch of the present invention.

[0067] Figure 4 This is a schematic diagram of the welding principle without using a dynamic alignment structure.

[0068] Figure 5 This is a schematic diagram of the welding principle when using the dynamic alignment structure of the present invention. Detailed Implementation

[0069] The present invention will be further described in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined in this application.

[0070] This embodiment provides a magnetically controlled welding dynamic alignment system. The system principle is as follows: Figure 2 , Figure 3 As shown, this system mainly includes hardware devices and control algorithms. It mainly monitors the deflection angle θ between the symmetry axis of the molten pool and the symmetry axis of the magnetic control device in real time, and drives the motor to rotate to compensate for this angle in order to achieve dynamic alignment of the magnetic field and the molten pool, and ensure the welding effect under complex paths.

[0071] In this embodiment, the hardware device mainly includes a magnetic control device 1, an infrared temperature measuring device 2, a motor 3, and a control system 4. The magnetic control device 1 is fixed to the welding torch 5 and can rotate around the axis of the welding torch 5. The infrared temperature measuring device 2 is fixed to the magnetic control device 1 and includes a reciprocating infrared temperature sensor 6. In this structure, the oscillation of the infrared temperature sensor 6 can be controlled independently, achieving reciprocating oscillation to monitor the temperature on both sides of the molten pool. During the rotation of the magnetic control device 1, the infrared temperature sensor 6 rotates synchronously to ensure their relative positions. The motor 3 is coaxially connected to the magnetic control device 1 and is used to drive the magnetic control device 1 and the infrared temperature measuring device 2 to rotate. The control system 4 is mainly used to process temperature data in real time, calculate the deflection angle, and output motor control signals.

[0072] In this embodiment, the control algorithm is mainly integrated into the control system 4, which is divided into two modules: a coarse positioning module and a precise positioning module. The coarse positioning module is simple but important, equivalent to the system's "power-on self-test". When the system starts up, it first confirms the existence of the molten pool (state judgment module). If it cannot find it, it controls the motor to perform a 360-degree rotation search (search module). The precise positioning module processes in three layers: the acquisition module is like a "sensory organ", using dual-channel infrared temperature measurement and filtering algorithm to obtain the temperature difference; the control module is the "brain", using an improved PID algorithm to calculate the motor adjustment amount; and the power module is the "hands and feet", translating the instructions into the actions of the motor 3, driving the motor 3 to rotate, and thus adjusting the position of the magnetic control device 1.

[0073] like Figure 1 As shown in the figure, this embodiment also provides a dynamic alignment method for magnetic welding, which is mainly implemented through a control algorithm, as detailed below.

[0074] The status determination module acts as the system's start / stop controller, determining whether the molten pool is within the sensor's detection range. This is a prerequisite for all control actions, ensuring that the system only initiates the positioning process when a valid target exists. The molten pool determination follows this expression:

[0075]

[0076] in, This indicates the maximum temperature value collected by the infrared temperature sensor. It represents the maximum temperature value collected by the sensor within a single scan cycle; this is a real-time measurement and represents the temperature of the hottest spot within the sensor's field of view. This indicates the presence of a threshold for judging the molten pool. This is a preset constant whose value depends on the type of material being welded (e.g., the melting temperature of aluminum alloys is approximately 660°C, and that of steel is approximately 1500°C). This threshold is set to distinguish the high-temperature molten pool from the workpiece body and the surrounding environment.

[0077] The above molten pool determination uses simple logical judgment conditions. The system continuously monitors the maximum temperature. Only when the highest detected temperature exceeds the material-related melt pool temperature threshold. Only when the melt pool is detected is the system considered to exist, and the system will enter or remain in the precise positioning state. Otherwise, the melt pool is considered to be outside the detection range, triggering the search module.

[0078] In the search module, when the status judgment module cannot detect the molten pool, the motor is controlled to perform a 360° rotation scan. This is the initial positioning phase of the system, aimed at quickly locating the molten pool in an unknown environment. The rotation formula is as follows:

[0079]

[0080] in, This represents the real-time search angle, specifically the angle of the motor (i.e., the magnetic control device) at time t. This is a quantity that changes over time. This indicates the initial rotation angle, the initial position angle of the magnetic control device when the system starts searching. The constant search angular velocity is a preset constant, measured in degrees per second or radians per second, which determines the speed at which the motor rotates during the search; t represents the current time since the start of the search.

[0081] When the status judgment module determines that the molten pool does not exist, the control system commands the motor to move at a constant angular velocity. Start rotating. The process of the motor's rotation angle increasing linearly with time is described, i.e., performing a 360° scan. The termination condition of the scan is as follows:

[0082]

[0083] The constraints are as follows:

[0084]

[0085] The above constraints ensure that the search range does not exceed a full circle (360°). When the search angle is an integer multiple of 360°, it means that a full circle scan has been completed without finding a molten pool. The system can determine this as an anomaly (such as welding not starting) and stop the search or issue an alarm.

[0086] The data acquisition module plays a crucial role in the complex trajectory welding system with dynamic magnetic field coupling. By measuring the temperature on both sides of the molten pool, a reliable instantaneous temperature difference ΔT(t) is calculated. The temperature values ​​on the left and right sides of the molten pool are monitored in real time by an oscillating infrared temperature sensor. An exponential smoothing filtering algorithm is used to denoise the original temperature signal, and the difference between the filtered temperatures on both sides is calculated as a quantitative indicator of the alignment state of the magnetic control device. This provides an accurate and reliable basis for subsequent control decisions. The specific process is as follows.

[0087] Original signal acquisition: The temperature of the molten pool on both sides of the molten pool is monitored in real time using an infrared temperature sensor. , The expression is as follows:

[0088]

[0089] in: , These represent the measured temperatures at the i-th positions on the left and right sides of the molten pool, respectively. These temperatures are the raw temperature data and include the effects of noise and other factors. , ε represents the actual temperature at the i-th position on the left and right sides of the molten pool, respectively; EMI Indicates electromagnetic interference noise; ε spatter This indicates that the noise is blocked by splashing.

[0090] The formula above describes the composition of the actual measured signal, namely, the "true temperature" superimposed with two main, unavoidable process noises. This illustrates that the original signal cannot be used directly and must undergo subsequent signal denoising processing.

[0091] Signal denoising: During welding, sudden interferences such as electromagnetic noise may occur, causing anomalies in the temperature data collected by the sensors. If these anomalies directly enter the filtering algorithm, they will contaminate the filtering results and affect subsequent control decisions. Therefore, anomalies need to be processed before filtering to avoid adverse effects on the system. The original temperature signal contains high-frequency noise (arc interference, spatter obstruction), and exponential smoothing filtering is required to track data faster and respond promptly. The independent filtering calculation formulas for the temperature data on the left and right sides of the molten pool are as follows:

[0092]

[0093] The initial conditions are:

[0094]

[0095] in, This indicates the real-time temperature at the i-th position on the left side of the molten pool; The value represents the real-time temperature at the i-th position on the right side of the molten pool; α∈(0,1] represents the smoothing factor, which determines the filtering strength. The larger α is, the greater the weight of the new sampled value, the smaller the filtering lag but the weaker the noise reduction effect. The smaller α is, the greater the weight of historical data, the better the noise reduction effect but the lag increases.

[0096] The above employs a recursive filtering algorithm. The current filter output is a weighted average of the current measurement value and the filter output from the previous time step. It effectively filters out ε. EMI ε spatter High-frequency random noise was used to obtain a stable reflection of the heat distribution trend of the molten pool. The value of α needs to be adjusted by compromise based on the welding process (noise level) and the required response speed.

[0097] Temperature difference calculation: The temperature difference between the two sides after filtering is calculated as a quantitative indicator of the alignment state of the magnetic control device. The sign and amplitude of ΔT represent the direction and degree of deviation, respectively. The expression is as follows:

[0098]

[0099] Wherein, ΔT(i) represents the real-time temperature difference at position i (positive and negative indicate direction, and amplitude indicates the magnitude of deviation). (1) When the axis of the magnetron device is coplanar with the axis of symmetry of the molten pool, the temperature distribution is symmetrical, i.e., ΔT≈0. (2) When the magnetron device rotates (angle θ), the sensor deviates from the axis of symmetry, and the two sides are in different temperature zones, i.e., |ΔT|∝|θ|.

[0100] The control module is the core decision-making unit of the magnetic welding device. Its main function is to convert the temperature difference (ΔT) acquired by the acquisition module into angle adjustment commands for the magnetic control device, thereby achieving closed-loop feedback control. It receives ΔT from the acquisition module, analyzes its changing trend (increasing, decreasing, or stabilizing) in real time, and determines whether to immediately adjust the angle of the magnetic control device or maintain the current state. Through an incremental PID algorithm, it generates precise motor action commands and uses dead-zone processing to limit the maximum amplitude of the output control quantity to prevent overload or runaway. This drives the magnetic control device to remain relatively stationary with respect to the welding trajectory, thus achieving real-time matching between the magnetic field direction and the arc movement. The specific method is as follows.

[0101] An improved PID control algorithm, employing dynamic quantification of system asymmetry and incorporating historical state memory and temperature-related corrections, effectively suppresses measurement deviations caused by thermocouple effects and structural asymmetry compared to traditional incremental PID algorithms, significantly improving the control accuracy of the molten pool symmetry. This improved PID algorithm plays a core role in the control module of the magnetic control device. Its function is to minimize computational load and avoid integral saturation, achieving an efficient and stable real-time angle correction formula (improved PID) based on temperature difference and real-time response.

[0102]

[0103]

[0104] in, This indicates the increment of the output control quantity (used to drive the motor to adjust the angle). , , The temperature difference between the current time, the previous time, and the time before that is used as the tracking error term for the current region in the PID control algorithm; K p The proportional term represents the immediate response to changes in deviation; the larger the deviation, the stronger the compensation. K i This represents the integral term, which accumulates historical deviations and eliminates static errors (such as residual deviations caused by mechanical hysteresis); K d It represents the differential term, predicts future deviation changes, and suppresses overshoot (sensitive to high-frequency noise, requiring low-pass filtering in conjunction with it). This represents the asymmetric compensation term.

[0105] When there are differences in the installation location, material properties, or aging degree of a mechanically oscillating infrared sensor, different thermoelectric potentials will be generated on both sides of the molten pool, even when exposed to the same thermal radiation. This difference will cause a systematic deviation in the left and right readings of the molten pool, and this deviation is usually related to the absolute value of the temperature (the higher the temperature, the greater the deviation). Compensation formula:

[0106]

[0107] in, This represents the temperature difference direction factor, a sign function that determines the compensation polarity based on the current temperature difference direction. If ΔT>0, it is +1; if ΔT<0, it is -1. This represents the average temperature of the left and right filters, indicating the current absolute temperature level, because the sensor's asymmetry error is often related to the absolute temperature value. The asymmetry correction coefficient represents the strength of the correction to the inherent asymmetry of the system. It is a constant obtained through experimental calibration and represents the degree of inherent asymmetry of the system caused by differences in sensor installation and performance. The historical average temperature difference, ΔT, is the average value obtained over a long period of time under stable system conditions and known good alignment (such as during the calibration phase), representing the zero-position offset of the system.

[0108] Incremental PID algorithm outputs the change in control quantity. Instead of absolute position values, this is more suitable for driving motors. The algorithm comprehensively considers the current state (proportional term), history (integral term), and future trend (differential term) of the error. In particular, it incorporates... This feature enables the controller to proactively compensate for known system measurement deviations during decision-making, thereby significantly improving the control accuracy for the true symmetry of the molten pool.

[0109] Inaccurate deviations may occur due to sensor noise or minute fluctuations in the molten pool temperature. Dead zones can shield against such interference, preventing frequent actuator jitter. The dead zone threshold ϵ needs to be slightly larger than the sensor noise amplitude (e.g., if the experimentally measured noise peak is 0.5°C, then set...). After the control input is returned to zero, the actuator stops operating, reducing mechanical wear and energy consumption. The expression is as follows:

[0110]

[0111] in, The dead zone threshold is a positive constant set based on the sensor noise level.

[0112] The power drive module is the core of the magnetic control device. Its function is to convert the instructions generated by the control module into the precise movement of the motor, drive the magnetic head and temperature sensor to adjust their positions, and achieve real-time matching between the magnetic field direction and the welding trajectory. The specific process is as follows.

[0113] Angle control: Receives output from the control module. This is interpreted as a target angle command for the servo motor, which drives the magnetic control device to rotate.

[0114]

[0115] Where Kscale represents the scaling factor, which is used to calibrate the mapping relationship between the control quantity and the physical angle; Indicates the current angle of the magnetic control device; This represents the target angle of the magnetic control device. The above formula will control the increment of the algorithm output. The position superimposed on the motor at the previous moment. Up, obtain the new target location. . It is a calibration parameter that ensures the commanded angle matches the actual mechanical angle.

[0116] PWM duty cycle-angle mapping: mapping the target angle Linear mapping to PWM duty cycle drives the motor to rotate to a specified position:

[0117]

[0118] in, This indicates the duty cycle of the PWM signal that needs to be output to the motor driver. , This indicates the maximum and minimum duty cycle values, and sets the duty cycle boundary (D). min / D max This prevents the motor from stalling or overvoltage. , This indicates the allowable rotation angle range of the magnetic control device. .

[0119] The above is a linear mapping function. It maps the target angle... (exist[ , Linear conversion to PWM duty cycle within the specified range (exist[ , (Within the specified range). The control system precisely controls the motor speed and angle by outputting a PWM wave with this duty cycle.

[0120] Dynamic response and adaptive speed regulation: To adapt to the rapid changes in trajectory during high-speed welding, this embodiment introduces an adaptive speed regulation mechanism to ensure that the system can respond quickly to large deviations and make smooth and precise fine adjustments.

[0121] The motor's base speed is proportional to the rate of change of the target angle. When the welding path direction changes drastically, the control module will output a large value. This causes the target angle to change rapidly, which in turn drives the motor to run at a higher speed, thus enabling a model that quickly follows angle changes.

[0122]

[0123] in, Indicates the motor's reference speed; Indicates the speed response gain; Indicates the rate of change of the target angle; This represents the minimum sustaining speed required to maintain static friction in the system.

[0124] In the above formula, the motor's base speed is proportional to the rate of change of the target angle, achieving self-adaptation. When welding along straight lines or small bends, Small, with low-speed fine-tuning of the motor; suitable for sharp turns during welding. Large motor automatically accelerates and quickly aligns.

[0125] To further improve response speed, when the error between the current position and the target position is large, an additional acceleration is introduced, causing the motor to start at its maximum permissible acceleration, quickly eliminating the large error.

[0126]

[0127] Where 'a' represents the additional acceleration; Indicates the acceleration factor; Indicates the error threshold; When the error is less than this value, this compensation term is zero to avoid system jitter in the small error range.

[0128] When the system experiences a large angular error due to disturbances or other reasons, this module provides an additional acceleration. This allows the motor to start quickly with maximum permissible acceleration, rapidly eliminating large errors and shortening adjustment time.

[0129] Real-time synthesized speed: The final commanded speed of the motor is determined by both the reference speed and the integral of the acceleration compensation, thus achieving a smooth and rapid dynamic response.

[0130]

[0131] The final speed command sent to the motor Adaptive reference speed The integral value of the acceleration compensation is combined with the integral value of the acceleration compensation. This ensures that the system can respond smoothly to normal trajectory changes and react quickly to sudden large deviations, achieving fast, stable and accurate dynamic tracking performance overall.

[0132] Figure 4 The diagram shown illustrates the welding principle without the use of a dynamic alignment structure. Figure 5 This is a schematic diagram illustrating the welding principle when using the dynamic alignment structure of this invention. Figure 4As can be seen from the diagram, this illustrates the problems that arise when welding along complex curved paths if the direction of the magnetic field (usually indicated by the magnetic pole axis NS) of the magnetron control device (coil shown in the diagram) remains fixed. As illustrated, the welding direction (along the tangent to the curved path) is constantly changing, while the magnetic field direction always points to the initially set direction. This results in a continuous deflection angle θ between the magnetic field direction and the instantaneous welding forward direction. The existence of this deflection angle causes the direction of the electromagnetic force applied to the arc and molten pool to deviate from the ideal state, making it impossible to achieve optimal control over the molten pool flow, arc morphology, and droplet transfer. The consequence is that in the curved sections of the welding path, the weld formation may be asymmetrical, easily leading to defects such as undercut, humps, or lack of fusion, and welding quality and stability cannot be guaranteed. This is precisely the limitation mentioned in the background art: existing magnetron control technology is only applicable to straight-line welding.

[0133] from Figure 5 As can be seen from the diagram, this schematic intuitively illustrates the core working mechanism of the system of this invention. As shown, the system integrated on the welding torch (including a magnetic control device, an oscillating infrared temperature sensor, and a drive motor) can sense and dynamically adjust in real time. Its workflow is as follows: The infrared temperature sensor (probe shown in the diagram) rapidly oscillates back and forth, scanning both sides of the molten pool (points A and B), acquiring temperature distribution data in real time. The control system calculates the deflection angle θ between the current molten pool symmetry axis and the magnetic control device axis based on the temperature difference ΔT between the two sides. The control system immediately drives the motor to rotate, causing the entire magnetic control device (along with the temperature sensor) to rotate by a corresponding compensation angle. Figure 5 As shown, after dynamic adjustment, the magnetic field axis (NS direction) of the magnetron control device is always aligned with the current welding direction (curve tangent) or at a preset optimal relative angle. At this time, the deflection angle θ is eliminated or controlled within a very small range.

[0134] Example:

[0135] Example 1: Using a 300×100×5mm 6061 aluminum alloy plate, an existing GMAW welding power supply was used, with the welding current adjusted to 217A and the welding voltage to 27.5V. The DC pulse adjustable power supply was adjusted to a dual-channel pulse waveform with a magnitude of 8A, a frequency of 80Hz, a duty cycle of 50%, and a phase difference of 180°. The welding speed was 1.4m / min.

[0136] Example 2: Using a 300×100×5mm 6061 aluminum alloy plate, an existing GMAW welding power supply was used. The welding parameters were exactly the same as in Example 1. The difference was that there was magnetron-assisted welding. An external magnetic field was applied during the welding process, but the magnetic field was not dynamically aligned, that is, the direction of the magnetic field remained unchanged.

[0137] Example 3: Using a 300×100×5mm 6061 aluminum alloy plate, and an existing GMAW welding power source, the same welding parameters as above are used. The difference is that the dynamic alignment system provided by this invention is used. The angle of the magnetic control device is adjusted by detecting the temperature of the molten pool, and the adjusted external magnetic field is applied to the welding process.

[0138] Verification and analysis revealed that in Example 1, at a relatively high welding speed (1.4 m / min), the GMAW weld without any magnetic field assistance exhibited typical "high-speed welding defects." The weld surface showed discontinuous hump-shaped formations, a clear indication of molten pool instability; simultaneously, undercut was observed at the weld edges, with uneven overall width, suggesting potentially insufficient penetration. This confirms that under high-speed welding conditions prioritizing efficiency, traditional GMAW processes face severe quality challenges.

[0139] In Example 2, the weld quality was improved to some extent after applying a magnetic field with the same parameters, indicating that the magnetic field played a role in stirring and stabilizing the molten pool. However, since the direction of the magnetic field was fixed during welding, it could not be adjusted in real time according to changes in the welding trajectory direction (i.e., Figure 4 As shown in the figure, the optimization effect is local and incomplete. The weld may be of good quality in some sections (such as when the magnetic field direction accidentally matches the welding direction), but slight asymmetry or inconsistent forming may still occur at the point where the direction changes. It fails to completely eliminate the inherent defects of high-speed welding, and the improvement in process stability is limited.

[0140] In Example 3, the introduction of the dynamic alignment system of this invention resulted in a qualitative leap in weld quality. The weld formation was continuous, uniform, and smooth, completely eliminating humps and undercut defects. The transition between the weld and the base material was smooth on both sides, with consistent width, exhibiting excellent geometric symmetry. This directly proves that by using real-time infrared temperature measurement feedback and dynamically rotating the magnetic control device, the direction of the applied magnetic field can always maintain optimal matching with the instantaneous welding direction (i.e., Figure 5 (As shown in the diagram). This dynamic and precise magnetic field coupling enables continuous, stable, and optimized control of the molten pool flow and arc behavior throughout the entire welding path, thereby achieving both high forming quality and high process stability at high speeds.

Claims

1. A magnetically assisted welding dynamic alignment method, characterized in that, Includes the following steps: S1. Collect temperature data in the detection area and locate the molten pool; S2. Real-time acquisition of temperature data on both sides of the molten pool, noise reduction processing of the original temperature data, calculation of the temperature difference on both sides of the molten pool after filtering, obtaining the deflection angle between the molten pool symmetry axis and the magnetic control device axis, as a quantitative indicator of the alignment state of the magnetic control device. S3. Based on the temperature difference and its trend on both sides of the molten pool, control commands are generated using an incremental PID algorithm, and dead zone processing is used to limit the maximum amplitude of the output control quantity. S4. The control commands generated by the incremental PID algorithm are converted into the target angle commands of the motor, and the position is adjusted by the motor-driven magnetic control device to achieve real-time matching between the magnetic field direction and the arc motion. In step S2, the method for denoising the original temperature data is as follows: The raw temperature data is processed using a recursive filtering algorithm. The formula for processing the temperature data on both sides of the molten pool is as follows: ， The initial conditions are: ， in, This indicates the real-time temperature at the i-th position on the left side of the molten pool; The value represents the real-time temperature at the i-th position on the right side of the molten pool; α∈(0,1] represents the smoothing factor, which determines the filtering strength. The larger α is, the greater the weight of the new sampled value, the smaller the filtering lag but the weaker the noise reduction effect. The smaller α is, the greater the weight of the historical data, the better the noise reduction effect but the lag increases. This indicates that the temperature was measured at the i-th position on the left side of the molten pool. This indicates that the temperature was measured at the i-th position on the right side of the molten pool. The formula for calculating the temperature difference between the two sides of the molten pool after filtering is as follows: ， Wherein, ΔT(i) represents the real-time temperature difference at position i, and its positive and negative values ​​and amplitudes represent the direction and degree of deviation, respectively. When the axis of the magnetic control device is coplanar with the axis of symmetry of the molten pool, the temperature distribution is symmetrical. When the magnetic control device deflects, the two sides are in different temperature zones. In step S3, the incremental PID algorithm dynamically quantifies the system asymmetry and combines historical state memory and temperature-related corrections, as shown in the following expression: ， in, (This represents the increment of the output control quantity, used to drive the motor to adjust the angle). , , The temperature difference between the current time, the previous time, and the time before that is used as the tracking error term for the current region in the PID control algorithm; K p The proportional term is used to instantly respond to changes in the deviation trend; the larger the deviation, the stronger the compensation. K i K represents the integral term, used to accumulate historical deviations and eliminate static errors; d This represents the differential term, used to predict future changes in deviation and suppress overshoot. This indicates an asymmetric compensation term, used to compensate for differences on both sides of the molten pool; In step S3, the asymmetric compensation term The expression is as follows: ， in, This represents the temperature difference direction factor, a sign function that determines the compensation polarity based on the current temperature difference direction. If ΔT>0, it is +1; if ΔT<0, it is -1. This represents the average temperature of the left and right filters, indicating the current absolute temperature level. The asymmetry correction coefficient represents the strength of the correction to the inherent asymmetry of the system and is calibrated experimentally. Indicates the historical average temperature range; In step S3, the maximum amplitude of the output control quantity is limited by dead-time processing, including: Set dead zone threshold When the temperature difference is less than the dead zone threshold Increase the output control quantity To return to zero, the expression is as follows: ， Among them, the dead zone threshold It is a positive constant set based on the noise level of temperature acquisition, and the dead zone threshold. Greater than the noise amplitude; Step S4 specifically includes: The control commands generated by the incremental PID algorithm are converted into target angle commands for the motor, as shown in the following expression: ， in, This represents the increment of the control quantity output by the incremental PID algorithm; Kscale represents the proportional factor, used to calibrate the mapping relationship between the control quantity and the physical angle. Indicates the current angle of the magnetic control device; Indicates the target angle of the magnetic control device; Target angle The linear mapping is to the PWM duty cycle, which drives the motor to rotate to a specified position. The expression is as follows: ， in, This indicates the duty cycle of the PWM signal that needs to be output to the motor driver. , These represent the maximum and minimum duty cycles, respectively. , These represent the maximum and minimum allowable rotation angles of the magnetic control device, respectively. .

2. The magnetically assisted welding dynamic alignment method according to claim 1, characterized in that, Step S1 specifically includes: S11. Collect the maximum value of the temperature at the current location, and determine whether the maximum value of the temperature is greater than or equal to the threshold for determining the presence of the molten pool. If yes, it is determined that the molten pool is within the detection range, and proceed to step S2; otherwise, it is determined that the molten pool is not within the detection range, and proceed to step S12. S12. Search other locations along the circumference at a constant angular velocity and execute step S11. If no molten pool is found after scanning one full circle, it is considered abnormal, the search is stopped and an alarm is triggered.

3. The magnetically assisted welding dynamic alignment method according to claim 1, characterized in that, In step S4, an adaptive speed regulation mechanism is introduced, making the motor's base speed proportional to the rate of change of the target angle, as shown in the following expression: ， In the formula, Indicates the reference speed of the motor. Indicates the speed response gain; Indicates the rate of change of the target angle; This represents the minimum sustaining speed required to maintain the static friction of the system. When the error between the current position and the target position is large, an additional acceleration is introduced to enable the motor to start with the maximum allowable acceleration, as shown in the following expression: ， in, Indicates additional acceleration; Indicates the acceleration factor; Indicates the error threshold; When the error is less than this value, this compensation term is zero to avoid system jitter in the small error range; The final commanded speed of the motor is determined by the reference speed and the integral of the acceleration compensation, as shown in the following expression: ， in, This indicates the final speed of the motor.

4. A magnetically controlled assisted welding dynamic alignment system, used to implement the method according to any one of claims 1-3, characterized in that, The system includes: a magnetic control device, an infrared temperature measurement device, a motor, and a control system; The magnetic control device is mounted on the welding torch, and the infrared temperature measuring device is mounted on the magnetic control device. The infrared temperature measuring device includes a reciprocating infrared temperature sensor for real-time acquisition of the temperature on both sides of the molten pool. The motor is used to drive the magnetic control device to rotate and adjust its position along the welding torch axis. The control system is used to process the temperature data in real time, calculate the deflection angle, and output the motor control signal.

5. The magnetically assisted welding dynamic alignment system according to claim 4, characterized in that, The control system includes a coarse positioning unit and a precise positioning unit; The coarse positioning unit includes: The status judgment module is configured to determine whether the molten pool is within the detection range of the infrared temperature sensor based on the collected temperature data. The search module is configured to control the motor to drive the infrared temperature sensor to perform a 360° rotation scan when the status judgment module cannot detect the molten pool. The precise positioning unit includes: The acquisition module is configured to acquire temperature data from both sides of the molten pool in real time, process the temperature data, and calculate the temperature difference after filtering. The control module is configured to generate motor control commands from the temperature difference values ​​acquired by the acquisition module using an incremental PID algorithm. The power drive module is configured to convert the commands generated by the control module into precise motor movements, drive the magnetic control device to adjust its position, and achieve real-time matching between the magnetic field direction and the welding trajectory.

Images