A friction stir welding real-time temperature field monitoring method, system and electronic equipment

Abstract

The application discloses a friction stir welding real-time temperature field monitoring method, system and electronic equipment, and relates to the technical field of welding. The method comprises the following steps: acquiring temperature field data of a friction stir welding area, calculating a heat distribution characteristic value according to the temperature field data, wherein the temperature field data is obtained through an infrared temperature measurement system; comparing and analyzing the heat distribution characteristic value with a preset temperature window to generate a temperature field deviation index; triggering a parameter adjustment instruction according to the temperature field deviation index, wherein the parameter adjustment instruction comprises a shoulder pressing amount correction value and a spindle speed compensation value; collecting the radiation temperature of a black mark area around a stir head in real time, dynamically matching the radiation temperature with a reference temperature model, and starting multi-parameter collaborative control when the temperature deviation exceeds a preset deviation threshold, wherein the reference temperature model is constructed based on acquired welding parameter information of the friction stir welding. The application improves the temperature monitoring and regulation effect of the friction stir welding under complex working conditions.

Description

Technical Field

[0001] This application relates to the field of welding technology, and in particular to a method, system and electronic equipment for real-time temperature field monitoring in friction stir welding. Background Technology

[0002] Friction stir welding, as a solid-state joining technology, relies heavily on the uniformity of heat input for its welding quality. Studies have shown that the axial temperature gradient of the welding temperature field needs to be controlled within ±15℃, and the circumferential temperature difference should not exceed 20℃; otherwise, it will lead to problems such as uneven microstructure and decreased mechanical properties.

[0003] Existing methods for monitoring welding temperature fields mostly rely on thermocouples or infrared thermometers for localized temperature acquisition, which cannot reflect the overall temperature distribution characteristics of the weld area. For example, although US6733010B2 introduces infrared temperature measurement technology, it only monitors the surface temperature of the shoulder and ignores the temperature gradient changes in the heat-affected zone. When welding dissimilar metals or composite materials, the existing welding strategies of fixed temperature windows (such as the fixed temperature window (450-520℃) for aluminum-magnesium alloys) and fixed temperature thresholds are prone to temperature overshoot. Moreover, when welding dissimilar materials such as copper and aluminum, temperature fluctuations can reach ±35℃. Under complex working conditions, it is difficult to ensure the stability of the temperature window. Therefore, there are shortcomings in the temperature monitoring and control of friction stir welding, and there is room for improvement. Summary of the Invention

[0004] To improve the temperature monitoring and control effect of friction stir welding under complex working conditions, this application provides a method, system and electronic equipment for real-time temperature field monitoring of friction stir welding.

[0005] Firstly, the objective of this invention is achieved through the following technical solution:

[0006] A method for real-time temperature field monitoring in friction stir welding includes:

[0007] Acquire temperature field data of the friction stir welding area, and calculate heat distribution characteristic values ​​based on the temperature field data, wherein the temperature field data is acquired by an infrared temperature measurement system;

[0008] The heat distribution characteristic value is compared and analyzed with a preset temperature window to generate a temperature field deviation index;

[0009] The parameter adjustment command is triggered according to the temperature field deviation index, wherein the parameter adjustment command includes the shoulder pressing amount correction value and the spindle speed compensation value;

[0010] The radiation temperature of the black marked area around the stirring head is collected in real time, and the radiation temperature is dynamically matched with the reference temperature model. When the temperature deviation exceeds the preset deviation threshold, multi-parameter collaborative control is initiated, wherein the reference temperature model is constructed based on the welding parameter information of friction stir welding.

[0011] By adopting the above technical solutions, the welding parameter information of friction stir welding includes welding speed, spindle speed, and welding pressure. The black marked area is formed by coating with a high-absorption coating to create an artificial heat radiation source, enhancing the temperature signal contrast. This application uses an infrared temperature measurement system to collect temperature field data in real time, breaking through the limitations of traditional single-point thermocouple monitoring and realizing global perception of the temperature distribution in the welding area. This application forms a closed-loop linkage mechanism with the temperature deviation index, the shoulder pressure, and the spindle speed to achieve multi-parameter coupled control, solving the process conflict problem that is easily caused by adjusting a single parameter, and realizing precise control of welding heat input. Through dynamic matching of the reference temperature model, differences in material properties and fluctuations in operating conditions can be automatically identified. By triggering a collaborative control strategy through a preset deviation threshold, the robustness of the process is significantly improved. Through multi-dimensional temperature feature analysis, this application can provide early warning of abnormal temperature areas, which helps to avoid defects such as coarse grains and flash caused by local overheating, improves the welding effect of friction stir welding, and thus helps to improve the temperature monitoring and control effect of friction stir welding under complex operating conditions.

[0012] In a preferred embodiment of this application, obtaining the temperature field data of the friction stir welding zone includes:

[0013] An infrared temperature camera mounted on the spindle support of the friction stir welding machine is used to perform a gridded temperature scan of the working area of ​​the stirring head to obtain an infrared thermal image.

[0014] The infrared thermal image is processed to grayscale to extract temperature gradient distribution features;

[0015] Based on the temperature gradient distribution characteristics and the blackbody radiation theory, a temperature-grayscale conversion model is established to convert the red thermal image into temperature field matrix data, thereby obtaining the temperature field data of the welding area.

[0016] By adopting the above technical solutions, the spatial resolution is improved. The gridded scanning improves the spatial sampling density of the temperature field space compared with the traditional single-point infrared temperature measurement, and can capture local hot spots in the welding area (such as overheating at the edge of the stirring head). The temperature-grayscale conversion model based on the blackbody radiation theory, combined with grayscale nonlinear correction, helps to eliminate the deviation of the infrared camera response curve and reduce temperature measurement error. This application overcomes the problems of slow response and susceptibility to environmental interference of traditional temperature measurement methods, and improves the accuracy and stability of temperature field data acquisition.

[0017] In a preferred embodiment of this application, the step of generating the temperature field deviation index includes:

[0018] The similarity between the real-time temperature field matrix and the standard temperature field in the material process database is calculated to obtain the similarity difference.

[0019] Calculate the axial temperature gradient offset and the circumferential temperature fluctuation coefficient;

[0020] By combining the temperature gradient offset, the circumferential temperature fluctuation coefficient, and the real-time temperature extreme value, a temperature field deviation index containing the spatial distribution dimension is generated.

[0021] By adopting the above technical solution, and using the combined criteria of axial temperature gradient offset and circumferential temperature fluctuation coefficient, material overheating or cold welding defects can be detected 10 to 15 seconds earlier than traditional single-point temperature threshold alarms. The temperature field similarity difference (e.g., SSIM < 0.85) is directly related to the degree of mismatch between welding speed and pressure, guiding the direction of parameter adjustment to improve the accuracy of process parameter traceability. This application combines axial temperature gradient offset and circumferential temperature fluctuation coefficient to comprehensively generate a temperature field deviation index in the spatial dimension, realizing a comprehensive assessment of the temperature distribution state during welding. Compared with the single temperature threshold judgment method, this method can identify potential welding abnormal trends earlier.

[0022] In a preferred embodiment of this application: the welding parameter information includes welding speed, spindle speed, and welding pressure; the reference temperature model construction method includes:

[0023] Collect temperature field data under different combinations of welding parameters, and establish welding speed v and rotation speed. Multiple regression model of pressure F and temperature field characteristic parameters: Where a, b, c, and d are weighting coefficients. This is the error term;

[0024] Genetic algorithms are used to optimize model parameters a, b, c, and d, so that the root mean square error between the predicted temperature field and the measured temperature field is less than a preset temperature threshold.

[0025] The optimized reference temperature model is embedded into the dynamic temperature compensation unit, and noise is removed using the Kalman filter algorithm.

[0026] When a sudden change in welding parameters during friction stir welding is detected, the adaptive learning module is activated, and an LSTM neural network is used to predict the temperature field change trend within a specified time period in the future.

[0027] By adopting the above technical solution, a multivariate regression benchmark temperature model is constructed based on welding parameters (such as welding speed, rotation speed, and pressure). The model parameters are optimized by using a genetic algorithm. Combined with Kalman filtering for noise reduction and LSTM neural network prediction, the accuracy and dynamic adaptability of temperature field modeling are significantly improved. This application not only improves the response capability to sudden changes in the welding process, but also enhances the robustness and intelligence level of the system under different process conditions.

[0028] In a preferred embodiment of this application, the instruction to adjust parameters based on the temperature field deviation index specifically includes:

[0029] Establish a sensitivity matrix for welding parameters based on the rotation speed sensitivity coefficient, the pressure sensitivity coefficient, and the welding speed sensitivity coefficient;

[0030] A parameter adjustment vector is generated based on the welding parameter sensitivity matrix and the temperature field deviation index. The parameter adjustment vector includes a rotation speed compensation value, a downward pressure correction value, and a welding speed correction value.

[0031] Based on the parameter adjustment vector, parameter adjustment constraints are set, and a hierarchical adjustment instruction set is generated according to the parameter adjustment constraints.

[0032] By adopting the above technical solution, establishing a welding parameter sensitivity matrix, generating a parameter adjustment vector by combining the temperature field deviation index, and introducing a graded adjustment instruction set, refined and scientific control of welding parameter adjustment is achieved.

[0033] In a preferred embodiment of this application: the step of setting parameter adjustment constraints based on the parameter adjustment vector, and generating a hierarchical adjustment instruction set according to the parameter adjustment constraints, includes:

[0034] The parameter adjustment constraints include:

[0035] when Perform rapid adjustment when ∈ [-10rpm, +10rpm];

[0036] when Precision adjustment is performed when the range is ∈ [-0.05mm, +0.05mm].

[0037] Segmented regulation is performed when Δv ∈ [-5m / min, +5m / min];

[0038] Generate a hierarchical adjustment instruction set { , , },in Main spindle speed adjustment amount; Δv is the correction value for shoulder pressing; Δv is the correction value for welding speed. This is a single adjustment command. For gradual adjustment instructions, This is a compound adjustment command.

[0039] By adopting the above technical solution and setting clear parameter adjustment constraints, the variation range of spindle speed, shoulder pressing amount and welding speed is matched with different types of adjustment commands (rapid adjustment, precision adjustment and segmented adjustment), thereby constructing a graded adjustment command set. This application can select the most suitable control strategy according to the degree of temperature deviation under different working conditions, avoiding the occurrence of over-adjustment or under-adjustment.

[0040] In a preferred embodiment of this application, the method further includes:

[0041] When the temperature gradient offset G of the infrared thermogram is compared with the axial strain acquired by the strain sensor Satisfy G× When the temperature exceeds 1.5℃ / %, the circuit breaker mechanism is triggered, and a safety command is forcibly generated. ;

[0042] Based on the noise covariance matrix Q of the Kalman filter output, the sensitivity matrix is ​​corrected. , The weighting coefficient of Δv: When the speed noise component of Q is greater than the preset speed noise threshold, The weighting coefficient is increased by 30%; when the downsampling noise component of Q is greater than the preset downsampling noise threshold... The weighting coefficient of Δv is reduced by 20%; when the welding speed noise component of Q is greater than the preset speed noise threshold, the weighting coefficient of Δv is increased by 5%.

[0043] or,

[0044] The execution logic of the hierarchical adjustment instruction set includes:

[0045] When the rapid adjustment command is met simultaneously and precision adjustment commands When the trigger condition is met, it will be executed first. And generate compensation instructions through a PID fuzzy controller. This results in a correction value for the shoulder undercut. satisfy: ,in To quickly adjust instructions Shoulder undercut correction value; For precision adjustment commands Shoulder undercut correction value; ∈[0.3, 0.7] represents the weighting factor dynamically assigned by the fuzzy rule.

[0046] By adopting the above technical solutions, a fusing mechanism and a dynamic weight correction mechanism are introduced on the basis of hierarchical control, which further improves the system's safety and adaptability. When the product of the temperature gradient offset and the axial strain exceeds the set threshold, a forced safety command is triggered, which can effectively prevent welding defects caused by abnormal temperature rise. At the same time, the sensitivity coefficient weight is dynamically adjusted based on the noise covariance matrix output by the Kalman filter, which enhances the robustness of the control system to measurement noise.

[0047] Secondly, the objective of this invention is achieved through the following technical solution:

[0048] A real-time temperature field monitoring system for friction stir welding is provided for executing the real-time temperature field monitoring method for friction stir welding as described above. The system includes:

[0049] Infrared temperature measurement module is used to collect temperature field data of the welding area during the welding process;

[0050] The data processing module is used to calculate the heat distribution characteristic value based on the temperature field data, and compare and analyze the heat distribution characteristic value with a preset temperature window to generate a temperature field deviation index.

[0051] The control module is used to generate parameter adjustment instructions based on the temperature field deviation index. The parameter adjustment instructions include a shoulder pressing amount correction value and a spindle speed compensation value.

[0052] The radiation temperature acquisition unit is used to collect the radiation temperature of the black marked area around the stirring head in real time.

[0053] A dynamic matching module is used to dynamically match the radiation temperature with a reference temperature model, which is constructed based on the welding parameter information of friction stir welding.

[0054] A multi-parameter collaborative controller is used to initiate parameter collaborative control when the temperature deviation exceeds a preset deviation threshold, so as to adjust the process parameters during the welding process.

[0055] Thirdly, the objective of this invention is achieved through the following technical solution:

[0056] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for real-time temperature field monitoring in friction stir welding.

[0057] Fourthly, the objective of this invention is achieved through the following technical solution:

[0058] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for real-time temperature field monitoring in friction stir welding.

[0059] In summary, this application includes at least one of the following beneficial technical effects:

[0060] 1. This application proposes a complete real-time temperature field monitoring system architecture for friction stir welding, which covers infrared thermometry, data processing, deviation analysis, parameter adjustment, radiation temperature acquisition, dynamic model matching, and multi-parameter collaborative control. It realizes high-precision sensing and closed-loop control of the temperature field during the welding process, and has good real-time performance, stability and intelligence. It is suitable for various welding materials and process parameter scenarios.

[0061] 2. By setting an infrared temperature camera on the spindle support of the friction stir welding machine to perform gridded scanning of the welding area, and by combining grayscale processing and blackbody radiation theory to establish a temperature-grayscale conversion model, high-precision, real-time acquisition of the temperature field of the welding area can be achieved. Attached Figure Description

[0062] Figure 1 This is a flowchart of a real-time temperature field monitoring method for friction stir welding according to an embodiment of this application;

[0063] Figure 2 This is a schematic diagram of a device according to one embodiment of this application. Detailed Implementation

[0064] The present application will be further described in detail below with reference to the accompanying drawings.

[0065] In one embodiment, such as Figure 1 As shown, this application discloses a method for real-time temperature field monitoring in friction stir welding, which specifically includes the following steps:

[0066] S1: Obtain temperature field data of the friction stir welding area and calculate the heat distribution characteristic value based on the temperature field data, wherein the temperature field data is acquired through an infrared temperature measurement system.

[0067] The infrared temperature measurement system is deployed as follows:

[0068] An infrared thermal imager (such as FLIR A655sc, temperature range 300-1200℃, frame rate ≥200Hz) is mounted on the spindle support of the friction stir welding machine. The lens axis is kept at an angle of ±5° with the rotation axis of the stirring head. The camera is controlled by a three-dimensional motion platform to perform grid-like scanning along the welding direction (X-axis) and depth direction (Y-axis), with a scanning interval of 1mm×1mm, covering the working area of ​​the stirring head (diameter ≤200mm), and welding current and voltage signals are acquired simultaneously (sampling rate ≥1kHz).

[0069] Specifically, non-uniformity correction is performed on the infrared thermal image, i.e., a piecewise linear correction method is used to eliminate non-linear errors in the detector response curve; dynamic background subtraction is implemented: that is, environmental thermal radiation interference is filtered out using a moving average method (window length of 5 frames), while retaining the characteristic temperature signal of the stirring head area. Thermal distribution feature extraction includes gradient calculation and hotspot localization. Gradient calculation refers to using the Sobel operator to calculate the axial (X-direction) and circumferential (θ-direction) gradients of the temperature field matrix, generating a gradient magnitude map G(x, y) = ,in This represents the temperature gradient offset along the X-axis of the temperature field matrix. This represents the temperature gradient offset along the Y-axis of the temperature field matrix. The hot spot in the contact area between the stirring head and the workpiece is extracted using the Otsu adaptive threshold segmentation method, and its centroid coordinates are calculated. , and equivalent diameter .

[0070] Furthermore, temperature field data of the friction stir welding zone is obtained, including:

[0071] S11: The working area of ​​the stirring head is scanned in a grid pattern using an infrared temperature camera mounted on the spindle support of the friction stir welding machine to obtain an infrared thermal image.

[0072] In this embodiment, gridded scanning refers to dividing the welding area into a regularly arranged discrete array of points (e.g., 1mm × 1mm) and collecting temperature data point by point. Infrared thermal images are two-dimensional grayscale images used to reflect the temperature distribution of the welding area; the grayscale value is positively correlated with temperature. The thermal imager's trigger mode is set to "external trigger," synchronized with the welding power supply signal to avoid motion blur. Each frame of the infrared thermal image is accompanied by a timestamp and spatial coordinate labels for easy subsequent data fusion.

[0073] S12: Perform grayscale processing on the infrared thermal image to extract temperature gradient distribution features.

[0074] In this embodiment, grayscale processing refers to mapping the original pixel values ​​(such as 14-bit digital signals) of the infrared thermal image to 0-255 grayscale levels to enhance contrast; temperature gradient refers to the rate of change of the temperature field in space, reflecting the direction and rate of heat transfer.

[0075] Specifically, the cv2-based image processing toolkit performs detection on the grayscale processed infrared thermal image: the Sobel algorithm is used to detect the changes in brightness in the left and right directions (i.e., the horizontal edges) to obtain the horizontal gradient; similarly, the changes in brightness in the up and down directions (i.e., the vertical edges) are detected to obtain the vertical gradient. Based on the image change gradients in the horizontal and vertical directions, the square root is used to calculate the total "edge intensity" at each point, and the total edge intensity of each pixel is synthesized to form a gradient map.

[0076] S13: Based on the temperature gradient distribution characteristics and blackbody radiation theory, a temperature-grayscale conversion model is established to convert the red thermal image into temperature field matrix data, thereby obtaining the temperature field data of the welding area.

[0077] In this embodiment, blackbody radiation theory refers to the quantitative relationship between the radiant energy of an ideal object and its temperature (Planck's law); the temperature-grayscale conversion model refers to a mathematical model that maps the grayscale values ​​of an infrared thermogram to real temperature values.

[0078] Specifically, a constant-temperature blackbody furnace (temperature range 300-1200℃, accuracy ±1℃) was used as the standard source to acquire grayscale images at different temperatures.

[0079] The model construction of the temperature-grayscale conversion model includes:

[0080] Derive the relationship between radiance L and temperature T based on Planck's law:

[0081] ,in, Emissivity of the material (0.3-0.4 for aluminum alloys); Let G1 be the Stefan-Boltzmann constant; and then, by fitting a linear relationship between the gray value G1 and the radiance L using the least squares method: G1 = a1 × L + b1, where a1 and b1 are weighting coefficients. After combining this with emissivity correction, the temperature calculation formula is obtained. Insert a standard temperature block (such as a copper block with a known melting point of 1085℃) every 10 minutes of welding, and dynamically adjust parameters a1 and b1.

[0082] S2: Compare and analyze the heat distribution characteristic values ​​with the preset temperature window to generate a temperature field deviation index.

[0083] In this embodiment, a standard temperature field library based on typical materials and fixed process parameters is first constructed:

[0084] The method for constructing a standard temperature field library includes: selecting typical materials (such as 2219 aluminum alloy and 316L stainless steel) to prepare standardized specimens; collecting temperature field data under fixed process parameters (v=300mm / min, ω=1500rpm, F=5kN); simulating the temperature field distribution through finite element simulation (ANSYS Workbench); calibrating the data with measured data; and establishing a library that includes the material's thermal conductivity λ and specific heat capacity. A three-dimensional standard temperature field database.

[0085] Specifically, the structural similarity index (SSIM) is used to quantify the real-time temperature field. (x, y) and standard temperature field The spatial matching degree of (x, y), and the multi-scale similarity calculation are as follows:

[0086] Where N is the number of scale layers (usually 3-5 layers). The mean; The real-time temperature field spatial average temperature; The average temperature of the standard temperature field. This represents the covariance between the real-time temperature field and the standard temperature field. For variance; For covariance; , This is a constant term, typically ranging from 0.01 to 0.03; Multi-scale SSIM (MS-SSIM) integrates similarity weights from different spatial frequencies, and outputs the similarity difference. ,like If the value is greater than 0.2, the temperature field is considered abnormal.

[0087] Further, step S2 includes:

[0088] S21: Calculate the similarity between the real-time temperature field matrix and the standard temperature field in the material process database to obtain the similarity difference.

[0089] S22: Calculate the axial temperature gradient offset and the circumferential temperature fluctuation coefficient.

[0090] S23: Combines temperature gradient offset, circumferential temperature fluctuation coefficient and real-time temperature extreme value to generate a temperature field deviation index that includes spatial distribution dimension.

[0091] When calculating the gradient offset, the axial gradient is calculated as the temperature gradient magnitude along the welding direction (X-axis): The circumferential gradient is calculated along the direction of rotation of the stirring head ( (Axis) Calculate the temperature gradient magnitude: The formula for quantifying the offset is: axial temperature gradient offset. in, This is the axial gradient magnitude matrix of the real-time temperature field; The standard temperature field axial gradient magnitude matrix; circumferential temperature fluctuation coefficient. ,in This is the circumferential gradient magnitude matrix of the real-time temperature field; denoted as the circumferential gradient magnitude matrix of the standard temperature field; A represents the area of ​​the region of interest.

[0092] Calculation of spatial standard deviation based on spatial average temperature: Here, M and N1 should represent the number of rows and columns of the temperature field matrix, which is the image resolution. For example, if the temperature field is divided into an M x N1 grid, then the position of each pixel can be represented by... Let x range from 1 to M, and y range from 1 to N1; The space average temperature.

[0093] Calculate the energy percentage in the 10-50Hz frequency band: in, The results are obtained from the Fast Fourier Transform of the temperature field data. The temperature field deviation index is obtained by weighted fusion calculation based on the weight coefficients (specific values ​​can be defined by the user) corresponding to the axial temperature gradient offset, circumferential temperature fluctuation coefficient, spatial standard deviation and energy proportion of the 10-50Hz frequency band.

[0094] S3: Triggers parameter adjustment commands based on temperature field deviation index, including shoulder pressure correction value and spindle speed compensation value.

[0095] In this embodiment, step S3 includes:

[0096] S31: Establish a sensitivity matrix for welding parameters based on the rotation speed sensitivity coefficient, the pressure sensitivity coefficient, and the welding speed sensitivity coefficient.

[0097] In this embodiment, the speed sensitivity coefficient refers to the change in speed per unit rotational speed ( The change in temperature field deviation caused by ); the pressure sensitivity coefficient refers to the change in pressure per unit amount of pressure ( The temperature field deviation index change caused by a unit change in welding speed (Δv) is the amount of change in temperature field deviation index.

[0098] Specifically, an orthogonal experiment (such as L9(3^4)) was designed, with a fixed pressure F=5kN, combined variables v=[200, 400, 600]mm / min, ω=[1000, 1500, 2000]rpm, and the temperature field deviation index BI was recorded.

[0099] A sensitivity model was established using partial least squares regression (PLSR):

[0100] ,in, , These are the weighting coefficients; For intersecting terms; Error term; extract principal component contribution rate and determine weighting coefficients; rotational speed sensitivity: =0.38, Downward pressure sensitivity: =0.45, Welding speed sensitivity: =0.22.

[0101] Next, construct a 3×3 diagonal sensitivity matrix S:

[0102] S= .

[0103] S32: Generate parameter adjustment vectors based on welding parameter sensitivity matrix and temperature field deviation index. The parameter adjustment vectors include rotation speed compensation value, pressure correction value and welding speed correction value.

[0104] Specifically, the normalized index BI is normalized to obtain... The adjustment amount is calculated using matrix multiplication:

[0105] Where Δv is the adjustment amount of welding speed, in m / min (meters per minute); The adjustment amount for the shoulder pressing down, in mm (millimeters). The amount of adjustment for the spindle speed, in rpm (revolutions per minute).

[0106] S33: Set parameter adjustment constraints based on parameter adjustment vectors, and generate a hierarchical adjustment instruction set according to the parameter adjustment constraints.

[0107] Specifically, the tiered adjustment instruction set generates tiered single-time adjustment instructions based on process constraints and the magnitude of the adjustment. ), gradual adjustment command ( ), composite adjustment command ( The constraint definitions for tiered adjustments include: The rapid adjustment range is ±10 rpm, the gradual adjustment range is ±3 rpm, and the combined adjustment range is ±5 rpm + ±3 m / min. The rapid adjustment range is ±0.1 mm, the gradual adjustment range is ±0.2 mm, and the combined adjustment range is ±0.05 mm ± ±2 m / min. Δv has a rapid adjustment range of ±5 m / min, a gradual adjustment range of ±1 m / min, and a combined adjustment range of ±3 m / min ± ±2 rpm.

[0108] Furthermore, parameter adjustment constraints are set based on the parameter adjustment vector, and a hierarchical adjustment instruction set is generated according to the parameter adjustment constraints, including:

[0109] S331: Parameter adjustment constraints include:

[0110] when Perform rapid adjustment when ∈ [-10rpm, +10rpm];

[0111] when Precision adjustment is performed when the range is ∈ [-0.05mm, +0.05mm].

[0112] Segmented regulation is performed when Δv ∈ [-5m / min, +5m / min];

[0113] Generate a hierarchical adjustment instruction set { , , },in This is a single adjustment command. For gradual adjustment instructions, This is a compound adjustment command.

[0114] Furthermore, the execution logic of the hierarchical adjustment instruction set includes:

[0115] When the rapid adjustment command is met simultaneously and precision adjustment commands When the trigger condition is met, it will be executed first. And generate compensation instructions through a PID fuzzy controller. This results in a correction value for the shoulder undercut. satisfy: ,in To quickly adjust instructions Shoulder undercut correction value; For precision adjustment commands Shoulder undercut correction value; ∈[0.3, 0.7] represents the weighting factor dynamically assigned by the fuzzy rule.

[0116] S4: Real-time acquisition of the radiation temperature of the black marked area around the stirring head, dynamic matching of the radiation temperature with the reference temperature model, and activation of multi-parameter collaborative control when the temperature deviation exceeds the preset deviation threshold. The reference temperature model is constructed based on the acquired welding parameter information of friction stir welding.

[0117] In this embodiment, the welding parameter information includes welding speed, spindle speed, and welding pressure.

[0118] Specifically, step S4 includes:

[0119] S41: Collect temperature field data under different welding parameter combinations and establish welding speed v and rotation speed. Multiple regression model of pressure F and temperature field characteristic parameters: Where a, b, c, and d are weighting coefficients. This is the error term.

[0120] In this embodiment, welding speed refers to the speed at which the weld moves per unit time (unit: mm / min); rotational speed refers to the angular velocity of the stirring head (unit: rpm); welding pressure refers to the axial force corresponding to the shoulder pressing amount (unit: kN); temperature field characteristic parameters include spatial average temperature, axial gradient amplitude, circumferential fluctuation coefficient, and error term. Follows a normal distribution N(0, ).

[0121] S42: Use a genetic algorithm to optimize the model parameters a, b, c, and d so that the root mean square error between the predicted temperature field and the measured temperature field is less than a preset temperature threshold.

[0122] In this embodiment, the genetic algorithm optimization includes crossover and mutation operations based on the fitness function, where the fitness function is the reciprocal of the root mean square error (RMSE), i.e.:

[0123] ,in, To predict the temperature field; The measured temperature field is represented by RMSE (Root Mean Square Error), which measures the difference between the predicted and measured temperature fields. The reciprocal of RMSE is used when converting it to a fitness function; the smaller the RMSE (the more accurate the model prediction), the higher the fitness value. The mutation operation simulates binary crossover (SBX) with a crossover probability Pc = 0.9. The mutation operation uses polynomial mutation with a mutation probability Pm = 0.1.

[0124] S43: The optimized reference temperature model is embedded into the dynamic temperature compensation unit, and noise is removed using the Kalman filter algorithm.

[0125] Specifically, the dynamic temperature compensation unit adjusts the output of the reference model based on real-time welding parameters. The state vector of the Carl filtering algorithm... Includes temperature field predictions and control inputs. Including welding speed v, rotation speed Pressure F, observed value The data is based on measured temperature field data, and the recursive state estimation algorithm formula for Kalman filtering is as follows: , ,in = Indicates the state at step K; For the previous step ( k The state estimate of -1); H represents system noise; H is the observation matrix (mapping the state vector to the observations); the state transition matrix A = This represents the evolution of the system state over time. The temperature term decays (retaining 90% from the previous step, plus the influence of the current input), while other parameters remain unchanged. Control input gain matrix: B = , indicating control input The degree of influence on each state variable.

[0126] S44: When a sudden change in welding parameters of friction stir welding is detected, the adaptive learning module is activated to use an LSTM neural network to predict the temperature field change trend within a specified time period in the future.

[0127] In this embodiment, the STM network is a long short-term memory neural network, suitable for time series prediction; the training data consists of historical welding parameters and temperature field data, sampled at a time step of Δt = 1 second.

[0128] Specifically, the input layer of the STM network consists of 3 nodes (v, ω, F), the LSTM layer has 64 hidden units, the forgetting factor γ = 0.8, and the output layer has 1 node (predicting the temperature field). The loss function used is mean squared error (MSE), and the optimizer used is the Adam optimizer (learning rate 0.001).

[0129] In one embodiment, a method for real-time temperature field monitoring in friction stir welding further includes:

[0130] S10: When the temperature gradient offset G of the infrared thermal image is different from the axial strain acquired by the strain sensor... Satisfy G× When the temperature exceeds 1.5℃ / %, the circuit breaker mechanism is triggered, and a safety command is forcibly generated. .

[0131] In this embodiment, the temperature gradient offset G refers to the absolute difference between the real-time temperature field gradient amplitude and the standard temperature field gradient amplitude, reflecting abnormal heat distribution patterns; axial strain This refers to the amount of plastic deformation in the axial direction (welding direction) of the stirring head, which is collected by a strain sensor (such as HBM C16A).

[0132] Specifically, the temperature gradient offset G is related to the axial strain acquired by the strain sensor. Satisfy G× When the temperature exceeds 1.5℃ / %, a risk of thermo-coupling failure is identified (such as stuck stirring head or excessive softening of material), and a safety command is forcibly triggered. .

[0133] Specifically, security instructions The execution mechanism includes: immediately interrupting the welding current (cutting off the heat source); controlling the shoulder to rise at a speed of 60mm / s for 200ms to avoid material adhesion; triggering an audible and visual alarm and recording the process parameters at the time of the fault.

[0134] S20: Based on the noise covariance matrix Q output by the Kalman filter, correct the sensitivity matrix in... , The weighting coefficient of Δv: When the speed noise component of Q is greater than the preset speed noise threshold, The weighting coefficient is increased by 30%; when the downsampling noise component of Q is greater than the preset downsampling noise threshold... The weighting coefficient is reduced by 20%.

[0135] In this embodiment, the noise covariance matrix ,in , , These are the noise components for welding speed, rotation speed, and pressure, respectively. To account for temperature observation noise, the noise covariance matrix Q is output in real time through a Kalman filter state estimator.

[0136] For example, the basis value of the weighting coefficients is: the initial sensitivity matrix. (corresponding to Δv, , T noise weight).

[0137] Example of dynamic adjustment:

[0138] like =0.0006>0.0005, The weight is adjusted to 0.4 × 1.3 = 0.52.

[0139] The constraints include: the adjusted total weights must be normalized. This application combines temperature gradient offset (thermal field) with axial strain (force field), overcoming the limitations of single-parameter threshold alarms.

[0140] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0141] In one embodiment, a real-time temperature field monitoring system for friction stir welding is provided, which corresponds to a real-time temperature field monitoring method for friction stir welding described in the above embodiment.

[0142] A real-time temperature field monitoring system for friction stir welding includes an infrared temperature measurement module, a data processing module, a control module, a radiation temperature acquisition unit, a dynamic matching module, and a multi-parameter collaborative controller. Detailed descriptions of each functional module are as follows:

[0143] Infrared temperature measurement module is used to collect temperature field data of the welding area during the welding process;

[0144] The data processing module is used to calculate the heat distribution characteristic value based on the temperature field data, compare and analyze the heat distribution characteristic value with the preset temperature window, and generate the temperature field deviation index.

[0145] The control module is used to generate parameter adjustment commands based on the temperature field deviation index. The parameter adjustment commands include the shoulder pressing amount correction value and the spindle speed compensation value.

[0146] The radiation temperature acquisition unit is used to collect the radiation temperature of the black marked area around the stirring head in real time.

[0147] The dynamic matching module is used to dynamically match the radiation temperature with the reference temperature model, which is constructed based on the welding parameter information of friction stir welding.

[0148] A multi-parameter collaborative controller is used to initiate parameter collaborative control when the temperature deviation exceeds a preset deviation threshold, so as to adjust the process parameters during the welding process.

[0149] For specific limitations regarding a real-time temperature field monitoring system for friction stir welding, please refer to the limitations of a real-time temperature field monitoring method for friction stir welding mentioned above, which will not be repeated here. Each module in the aforementioned real-time temperature field monitoring system for friction stir welding can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or it can be stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0150] In one embodiment, an electronic device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 2 As shown, this electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores temperature field data, radiation temperature of the black marked area, etc. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a real-time temperature field monitoring method for friction stir welding.

[0151] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0152] S1: Obtain temperature field data of the friction stir welding area and calculate the heat distribution characteristic value based on the temperature field data. The temperature field data is acquired by an infrared temperature measurement system.

[0153] S2: Compare and analyze the heat distribution characteristic values ​​with the preset temperature window to generate a temperature field deviation index;

[0154] S3: Trigger parameter adjustment command based on temperature field deviation index, where the parameter adjustment command includes shoulder pressure correction value and spindle speed compensation value;

[0155] S4: Real-time acquisition of the radiation temperature of the black marked area around the stirring head, dynamic matching of the radiation temperature with the reference temperature model, and activation of multi-parameter collaborative control when the temperature deviation exceeds the preset deviation threshold. The reference temperature model is constructed based on the acquired welding parameter information of friction stir welding.

[0156] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0157] S1: Obtain temperature field data of the friction stir welding area and calculate the heat distribution characteristic value based on the temperature field data. The temperature field data is acquired by an infrared temperature measurement system.

[0158] S2: Compare and analyze the heat distribution characteristic values ​​with the preset temperature window to generate a temperature field deviation index;

[0159] S3: Trigger parameter adjustment command based on temperature field deviation index, where the parameter adjustment command includes shoulder pressure correction value and spindle speed compensation value;

[0160] S4: Real-time acquisition of the radiation temperature of the black marked area around the stirring head, dynamic matching of the radiation temperature with the reference temperature model, and activation of multi-parameter collaborative control when the temperature deviation exceeds the preset deviation threshold. The reference temperature model is constructed based on the acquired welding parameter information of friction stir welding.

[0161] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0162] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0163] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for monitoring a temperature field in real time in friction stir welding, characterized by, The method comprises the following steps: acquiring temperature field data of a friction stir welding area, and calculating a heat distribution characteristic value according to the temperature field data, wherein the temperature field data is obtained by an infrared temperature measurement system; comparing and analyzing the heat distribution characteristic value with a preset temperature window to generate a temperature field deviation index; triggering a parameter adjustment instruction according to the temperature field deviation index, wherein the parameter adjustment instruction comprises a shoulder pressing amount correction value and a spindle speed adjustment amount; real-time acquisition of the radiation temperature of a black mark area around the stir head, dynamic matching of the radiation temperature with a reference temperature model, and starting of multi-parameter collaborative control when the temperature deviation exceeds a preset deviation threshold, wherein the reference temperature model is constructed based on acquired welding parameter information of the friction stir welding; the black mark area is an artificial heat radiation source formed by coating a high-absorptivity coating to enhance the temperature signal contrast; the step of acquiring the temperature field data of the friction stir welding area comprises: carrying out grid temperature scanning on the working area of the stir head by an infrared temperature camera installed on the spindle support of the friction stir welding to obtain an infrared thermal image; carrying out gray scale processing on the infrared thermal image to extract temperature gradient distribution characteristics; establishing a temperature-gray scale conversion model based on the temperature gradient distribution characteristics and blackbody radiation theory to convert the red thermal image into temperature field matrix data and obtain the welding area temperature field data; the step of generating the temperature field deviation index comprises: carrying out similarity calculation of the real-time temperature field matrix with the standard temperature field in the material process database to obtain a similarity difference value; calculating an axial temperature gradient offset and a circumferential temperature fluctuation coefficient; generating a temperature field deviation index containing a spatial distribution dimension by comprehensively considering the axial temperature gradient offset, the circumferential temperature fluctuation coefficient, and the real-time temperature extreme value; the welding parameter information comprises welding speed, spindle speed, and welding pressure; and the reference temperature model construction method comprises: acquiring temperature field data under different welding parameter combinations, and establishing a multivariate regression model of welding speed v, speed ω, pressure F, and temperature field characteristic parameters: T model (x, y) = a x v 2 +b x omega + c x F + d x v x omega x F + epsilon, wherein a, b, c, d are weight coefficients, and epsilon is an error term; genetic algorithm is used to optimize model parameters a, b, c, d, so that the root mean square error between the predicted temperature field and the measured temperature field is less than a preset temperature threshold; embedding the optimized reference temperature model into a dynamic temperature compensation unit, and carrying out denoising through a Kalman filtering algorithm; when a welding parameter mutation of the friction stir welding is identified, activating an adaptive learning module to predict the temperature field change trend in a specified future period of time by using an LSTM neural network.

2. The method of claim 1, wherein, the step of triggering the parameter adjustment instruction according to the temperature field deviation index specifically comprises: establishing a welding parameter sensitivity matrix based on a speed sensitivity coefficient, a pressing amount sensitivity coefficient, and a welding speed sensitivity coefficient; generating a parameter adjustment vector based on the welding parameter sensitivity matrix and the temperature field deviation index, wherein the parameter adjustment vector comprises a spindle speed adjustment amount, a shoulder pressing amount correction value, and a welding speed correction value; setting a parameter adjustment constraint condition based on the parameter adjustment vector, and generating a hierarchical adjustment instruction set according to the parameter adjustment constraint condition.

3. The method of claim 2, wherein the method further comprises: the step of setting a parameter adjustment constraint condition based on the parameter adjustment vector, and generating a hierarchical adjustment instruction set according to the parameter adjustment constraint condition comprises: the parameter adjustment constraint condition comprises: The fast adjustment is performed when Δω ∈ [-10 rpm, +10 rpm]; The fine adjustment is performed when Δh ∈ [-0.05 mm, +0.05 mm]; The subsection adjustment is performed when Δv ∈ [-5 m / min, +5 m / min]; The hierarchical adjustment instruction set {C1, C2, C3} is generated, wherein Δω is a spindle speed adjustment amount, Δh is a shoulder depression amount correction value, Δv is a welding speed correction value, C1 is a single adjustment instruction, C2 is a gradual adjustment instruction, and C3 is a compound adjustment instruction.

4. The method of claim 3, wherein the method further comprises: The method further comprises: When the temperature gradient offset G of the infrared thermal image and the axial strain ε collected by the strain sensor meet the condition of G x ε > 1.5℃ / %, a fuse mechanism is triggered, and a safety instruction I4 is forced to be generated. z When the temperature gradient offset G of the infrared thermal image and the axial strain ε collected by the strain sensor meet the condition of G x ε > 1.5℃ / %, a fuse mechanism is triggered, and a safety instruction I4 is forced to be generated. z When the temperature gradient offset G of the infrared thermal image and the axial strain ε collected by the strain sensor meet the condition of G x ε According to the noise covariance matrix Q output by the Kalman filter, the weight coefficients of Δω, Δh, and Δv in the sensitivity matrix are corrected: when the rotational speed noise component of Q is greater than a preset rotational speed noise threshold, the weight coefficient of Δω is increased by 30%; when the shoulder depression amount noise component of Q is greater than a preset shoulder depression amount noise threshold, the weight coefficient of Δh is reduced by 20%; and when the welding speed noise component of Q is greater than a preset speed noise threshold, the weight coefficient of Δv is increased by 5%; Or, The execution logic of the hierarchical adjustment instruction set comprises: When the trigger conditions of the fast adjustment instruction I1 and the fine adjustment instruction I2 are both met, I2 is preferentially executed, and a compensation instruction I'2 is generated by a PID fuzzy controller, so that the correction value Δh of the shoulder depression amount satisfies: wherein is the shoulder depression amount correction value of the fast adjustment command I1 ; is the shoulder depression amount correction value of the fine adjustment command I2 ; and α ∈ [0.3, 0.7] is a weight factor dynamically assigned by a fuzzy rule.

5. A friction stir welding real-time temperature field monitoring system, characterized by, The system is used for executing the real-time temperature field monitoring method of friction stir welding according to any one of claims 1-4, and the system comprises: An infrared temperature measurement module is configured to collect temperature field data of a welding area during a welding process. A data processing module is configured to calculate a heat distribution characteristic value based on the temperature field data, and compare and analyze the heat distribution characteristic value with a preset temperature window to generate a temperature field deviation index. A control module is configured to generate parameter adjustment instructions based on the temperature field deviation index, wherein the parameter adjustment instructions comprise a shoulder depression amount correction value and a spindle speed adjustment amount. A radiation temperature acquisition unit is configured to acquire a radiation temperature of a black mark area around a friction stir welding head in real time. A dynamic matching module is configured to dynamically match the radiation temperature with a reference temperature model, wherein the reference temperature model is constructed based on welding parameter information of the friction stir welding. A multi-parameter collaborative controller is configured to start parameter collaborative control to adjust process parameters during a welding process when a temperature deviation exceeds a preset deviation threshold.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor implements the steps of the real-time temperature field monitoring method of friction stir welding according to any one of claims 1-4 when executing the computer program.

7. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: receiving a request for a resource from a client; determining whether the client is authorized to access the resource; and if the client is authorized to access the resource, providing the resource to the client. The computer program implements the steps of the real-time temperature field monitoring method of friction stir welding according to any one of claims 1-4 when executed by the processor.

Images