A die changing device of a drawing die with intelligent positioning function

Abstract

The application discloses a die changing device of a drawing die with an intelligent positioning function, and relates to the technical field of metal wire processing. The die changing device comprises an intelligent positioning module, a quick die changing module and a central control system. The intelligent positioning module comprises a light source and an optical positioning sensor. The quick die changing module comprises a mechanical arm and a die locking mechanism. The mechanical arm is used for grabbing the drawing die. The die locking mechanism can lock or loosen the drawing die. The light source is arranged towards the die locking mechanism. The optical positioning sensor is arranged on the die locking mechanism and is used for monitoring the position of the drawing die on the die locking mechanism. The mechanical arm, the optical positioning sensor and the die locking mechanism are electrically connected with the central control system. The application can ensure that the axis of the drawing die coincides with the drawing axis and can quickly change the die.

Description

Technical Field

[0001] This invention relates to the field of metal wire processing technology, and in particular to a die changing device for a drawing die with intelligent positioning function. Background Technology

[0002] Copper wire drawing is a metal plastic processing technology that transforms copper rods into fine copper wires by gradually reducing their cross-sectional dimensions through multiple processes such as stretching, heating, and cooling using specialized equipment. This process is widely used in fields such as wire and cable, electronic devices, and precision manufacturing, and its processing procedures and production equipment constitute an important part of the modern copper deep processing system. Among these processes, the drawing die device, as a key component of copper wire drawing equipment, directly affects the dimensional accuracy, surface quality, and production efficiency of the finished copper wires.

[0003] Existing copper wire drawing equipment mostly adjusts the wire diameter by changing drawing dies of different specifications. However, the installation and positioning of drawing dies usually rely on manual operation, and it is difficult to maintain a stable and consistent center line alignment accuracy. In particular, errors tend to accumulate after multiple die changes. Summary of the Invention

[0004] The purpose of this invention is to provide a die changing device for a drawing die with intelligent positioning function, so as to solve the problems existing in the prior art, and to ensure that the axis of the drawing die coincides with the drawing axis, enabling rapid die changing.

[0005] To achieve the above objectives, the present invention provides the following solution: This invention provides a die-changing device for a drawing die with intelligent positioning function, comprising: an intelligent positioning module, a rapid die-changing module, and a central control system. The intelligent positioning module includes a light source and an optical positioning sensor. The rapid die-changing module includes a robotic arm and a die-locking mechanism. The robotic arm is used to grasp the drawing die, and the die-locking mechanism can lock or release the drawing die. The light source is positioned towards the die-locking mechanism. The optical positioning sensor is disposed on the die-locking mechanism and is used to monitor the position of the drawing die on the die-locking mechanism. The robotic arm, the optical positioning sensor, and the die-locking mechanism are all electrically connected to the central control system.

[0006] In some specific embodiments, the mold clamping mechanism includes a mold clamping body and several mold clamping claws. The optical positioning sensor is disposed on the mold clamping body, and the several mold clamping claws are arranged circumferentially along the mold clamping body. The mold clamping claws are made of shape memory alloy. When the mold clamping claws are energized, they can lock the drawing die. When the mold clamping claws are de-energized, they can release the drawing die.

[0007] In some specific embodiments, the manufacturing process of the clamping claw includes: First, the raw material for the mold clamping claw is heat-treated, and then the heat-treated raw material is tested to obtain the martensitic transformation initiation temperature M. s Martensitic normal phase transformation end temperature M f Martensitic reverse phase transformation initiation temperature A s and the end temperature of the reverse phase transformation of martensite A f The raw materials were trained using the stress-induced martensite cycle training method. During training, the raw materials are heated to temperature A, the end temperature of the reverse martensitic transformation. f The above process involves applying a bending load to the raw material until the applied load induces a stress-induced martensitic transformation and produces a predetermined deformation that matches the function of the clamping claw. The material is then held at the same bending load and temperature, unloaded, and cooled to the martensitic transformation end temperature M. f Now, complete one training session; repeat the training process described above.

[0008] In some specific solutions, the center of the illumination area formed by the light source coincides with the center of the mold-locking mechanism.

[0009] In some specific embodiments, the optical positioning sensor includes a substrate layer and a conductive ring, a conductive layer, a photoelectric region, four electrodes, and four array resistors disposed on the substrate layer. The photoelectric region is rectangular, and the four electrodes are located on the outer sides of the four corners of the photoelectric region. An array resistor is disposed between adjacent electrodes. The conductive layer is located between the photoelectric region and the array resistors, and the conductive ring is located on the outer sides of the electrodes and the array resistors.

[0010] In some specific solutions, the quick mold change module also includes a mold storage bin and a waste mold collection box. The mold storage bin is used to hold drawing dies to be used. The interior of the mold storage bin is divided into several areas, and different areas are used to place drawing dies of different specifications. The waste mold collection box is used to hold drawing dies after use. The waste mold collection box is equipped with a cushioning pad inside, and the side of the waste mold collection box is equipped with a removable door panel.

[0011] In some specific solutions, a safety monitoring module is also included. The safety monitoring module includes a force sensor and a displacement sensor. Both the force sensor and the displacement sensor are electrically connected to the central control system. The force sensor is mounted on the mold clamping mechanism and is used to monitor the force on the drawing die during the drawing process. The displacement sensor is positioned toward the end of the drawing die on the mold clamping mechanism and is used to monitor the displacement generated by the drawing die during the drawing process.

[0012] In some specific solutions, the central control system controls the light source to turn on, and the optical positioning sensor is used to collect photoelectric signals from the four corners of the drawing die on the clamping mechanism and the contour data of the drawing die; the contour template of the drawing die is square. The central control system is used for: preprocessing the photoelectric signals of the four corner points of the drawing die to obtain the position coordinates of the four corner points; fitting the actual center coordinates based on the position coordinates of the four corner points of the drawing die; calculating the deviation between the actual center coordinates and the drawing axis reference coordinates; determining the coordinate system transformation matrix based on the deviation between the actual center coordinates and the drawing axis reference coordinates; denoising the contour data to obtain denoised coordinates; using the coordinate system transformation matrix to convert the denoised coordinates into drawing axis coordinates in the drawing axis coordinate system; extracting the midpoint coordinates of the four sides of the drawing die based on the drawing axis coordinates; fitting the actual centerline based on the midpoint coordinates of the four sides of the drawing die; calculating the deviation between the actual centerline and the drawing axis reference coordinates; and controlling the robotic arm to correct the position of the drawing die based on the deviation between the actual centerline and the drawing axis reference coordinates.

[0013] In some specific implementations, the central control system is used to: acquire basic historical data and real-time monitoring data of the drawing die for the target time period; input the basic historical data and real-time monitoring data of the drawing die for the target time period into a trained life prediction model to obtain the remaining life prediction result of the drawing die; and issue a replacement prompt based on the remaining life prediction result of the drawing die.

[0014] In some specific implementations, the central control system further includes a model training module, which is used for: Obtain the dataset; the dataset includes the die operation data for each historical moment in the historical time period of the drawing die and the remaining life label for each historical moment; the die operation data includes basic historical data and real-time monitoring data; The dataset is divided into a first time period dataset, a second time period dataset, and a third time period dataset according to time; the first time period, the second time period, and the third time period are consecutive time periods; the first time period dataset, the second time period dataset, and the third time period dataset are respectively designated as the training set, the validation set, and the test set; The first model is obtained by fitting the training set. Using the Bayesian optimization method, the model parameters of the first model are updated based on the validation set, and the model parameters with the smallest loss function value are selected as the optimized hyperparameters. The optimized hyperparameters are used as the initial model parameters for the GBRT model to determine the initial model; The samples in the validation set and the samples in the training set are concatenated in chronological order to obtain a concatenated dataset; the concatenated dataset includes concatenated samples and the remaining lifetime label of each concatenated sample. The initial model is trained using the spliced ​​dataset to obtain a trained lifespan prediction model; The trained lifetime prediction model is tested using the test set to obtain the remaining lifetime prediction results for each mold operation data in the test set.

[0015] The present invention achieves the following technical effects compared to the prior art: This invention uses a light source from an intelligent positioning module to illuminate the mold-locking mechanism. An optical positioning sensor collects the position data of the drawing die, which is then transmitted to a central control system. A robotic arm adjusts the position of the drawing die to ensure its axis coincides with the drawing axis. The mold-locking mechanism enables the locking and unlocking of the drawing die, facilitating rapid die changes. The intelligent positioning module also reduces errors caused by manual operation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of a die-changing device for a drawing die with intelligent positioning function in some embodiments of the present invention. Figure 2 This is a top view of the unlocked state of the locking mechanism in some embodiments of the present invention; Figure 3 This is a top view of an optical positioning sensor in some embodiments of the present invention; Figure 4 This is a control flowchart for some embodiments of the present invention; Figure 5 This is a flowchart of the copper wire drawing process. In the diagram: 100-base, 210-optical positioning sensor, 211-substrate layer, 212-electrode, 213-conductive ring, 214-photoelectric area, 215-array resistor, 216-conductive layer, 220-light source, 310-mold storage bin, 320-waste mold collection box, 330-robotic arm, 340-mold locking mechanism, 3401-mold locking body, 3402-mold locking claw, 410-force sensor, 420-displacement sensor, 430-gripper, 510-display screen, 520-PLC, 600-drawing die. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The purpose of this invention is to provide a die changing device for a drawing die with intelligent positioning function, so as to solve the problems existing in the prior art, and to ensure that the axis of the drawing die coincides with the drawing axis, enabling rapid die changing.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] like Figures 1 to 5As shown, this embodiment provides a die-changing device for a drawing die 600 with intelligent positioning function, including: a base 100, an intelligent positioning module, a quick die-changing module, and a central control system. The base 100 is integrally cast from a high-rigidity material, and its top has an installation groove adapted to the die-locking mechanism 340. The intelligent positioning module includes a light source 220 and an optical positioning sensor 210. The quick die-changing module includes a robotic arm 330 and a die-locking mechanism 340. The robotic arm 330 is used to grasp the drawing die 600. The die-locking mechanism 340 is mounted on the base 100 and can lock or release the drawing die 600. The light source 220 is positioned towards the die-locking mechanism 340. The optical positioning sensor 210 is mounted on the die-locking mechanism 340 and is used to monitor the position of the drawing die 600 on the die-locking mechanism 340. The robotic arm 330, the optical positioning sensor 210, and the die-locking mechanism 340 are all electrically connected to the central control system. In this embodiment, the light source 220 of the intelligent positioning module illuminates the mold-locking mechanism 340, and the optical positioning sensor 210 collects the position data of the drawing die 600, which is then transmitted to the central control system. The robotic arm 330 adjusts the position of the drawing die 600 so that its axis coincides with the drawing axis. The mold-locking mechanism 340 enables the locking or unlocking of the drawing die 600, achieving rapid die change. This embodiment, through the intelligent positioning module and the rapid die change module, eliminates errors from manual installation, improves equipment utilization and production efficiency, and achieves high-precision, high-efficiency, high-safety, and reliable drawing operations.

[0022] In some specific embodiments, the mold-locking mechanism 340 includes a mold-locking body 3401 and a plurality of mold-locking claws 3402. An optical positioning sensor 210 is disposed on the mold-locking body 3401, and the plurality of mold-locking claws 3402 are arranged circumferentially along the mold-locking body 3401. The mold-locking claws 3402 are made of shape memory alloy, preferably nickel-titanium alloy, with a density of approximately 6.5 g / cm³. 3 The shape memory recovery force is approximately 400 MPa, and the phase transformation temperature is approximately 95°C. The shape memory alloy, after training, can reversibly recover the shape of its high-temperature phase (austenitic phase) and low-temperature phase (martensitic phase) as temperature changes occur. When the clamping claw 3402 is energized, it can lock the drawing die 600; when the clamping claw 3402 is de-energized, it can release the drawing die 600. When the drawing die 600 is placed on the optical positioning sensor 210 and the axis of the drawing die 600 is completely aligned with the drawing axis, the clamping claw 3402 is energized and wraps around the drawing die 600, tightly locking it.

[0023] In some specific embodiments, the manufacturing process of the mold clamping claw 3402 includes: First, the raw material of the clamping claw 3402 is heat-treated to eliminate work hardening and improve its mechanical properties. During heat treatment, the raw material is held at 400℃ for 30 minutes. Then, the heat-treated raw material is tested to obtain the phase transformation temperature after heat treatment. During the test, a 5mg sample is taken from the raw material and tested using a differential scanning calorimeter (DSC). The heating / cooling rate is 15℃ / min. The test results can be used to obtain the martensitic positive phase transformation initiation temperature M using the tangent method. s Martensitic normal phase transformation end temperature M f Martensitic reverse phase transformation initiation temperature A s and the end temperature of the reverse phase transformation of martensite A f The raw material is trained using a stress-induced martensite (SIM) cyclic training method, which is a thermo-mechanical treatment process that imparts a stable two-way shape memory effect to the clamping claw 3402. During training, the raw materials are heated to temperature A, the end temperature of the reverse martensitic transformation. f At temperatures above 30–50°C, ensuring the material is fully in the austenitic state, apply a bending load to the component until the applied load induces a stress-induced martensitic transformation in the raw material, producing a predetermined deformation that matches the function of the clamping claw 3402. Hold the component at this temperature for 1–3 minutes while maintaining the bending load and temperature to allow the martensitic variant to fully orient and stabilize. Then, slowly unload the load and cool the component to the martensitic transformation end temperature M. f The temperature is kept between 30 and 50°C to allow for the full formation of low-temperature martensite, completing one training cycle. This training process is repeated 15 to 30 times to establish a stable and repeatable internal stress field and preferred martensite orientation within the raw material. After standardized SIM cycle training, the clamping jaw 3402 can achieve autonomous bidirectional shape memory action: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] f Upon reaching the above position, it automatically changes to the locked shape and cools naturally to M after power is cut off. f The device automatically restores its original unlocked shape upon activation, with sufficient deformation, reliable operation, and good repeatability. It can stably perform core functions such as locking, positioning, tightening, and releasing. Furthermore, it exhibits stable temperature response and excellent fatigue performance after training, enabling repeated opening and closing with precise movements.

[0024] In some specific embodiments, the light source 220 includes multiple LEDs arranged circumferentially around the top of the clamping mechanism 340 to provide a uniform and stable detection light source for the optical positioning sensor 210. The center of the illumination area formed by the light source 220 coincides with the center of the clamping mechanism 340, avoiding the shadow cast by the light source 220 on the drawing die 600 from affecting the optical positioning sensor 210. This ensures that the optical positioning sensor 210 can accurately identify the reference surface of the drawing die 600 (i.e., the side of the drawing die 600 that contacts the clamping body 3401) in complex production environments.

[0025] In some specific embodiments, the optical positioning sensor 210 includes a substrate 211 and a conductive ring 213, a photoelectric sensing area, four electrodes 212, and four array resistors 215 disposed on the substrate 211. The photoelectric sensing area is rectangular, and the area of ​​the photoelectric sensing area is larger than the area of ​​the reference surface of the drawing die 600. The four electrodes 212 are rectangularly distributed on the outer sides of the four corners of the photoelectric sensing area. An array resistor 215 is disposed between adjacent electrodes 212. The four array resistors 215 are rectangularly distributed. The photoelectric sensing area includes a plurality of photoelectric areas 214. A conductive layer 216 is disposed on the edge of each photoelectric area 214. The conductive layer 216 is used to conduct the current of the photoelectric area 214 to the array resistor 215. The conductive ring 213 surrounds the outer side of the electrodes 212 and the array resistors 215. The optical positioning sensor 210 collects the reference surface position data of the drawing die 600, which is transmitted in real time to the PLC (Programmable Logic Controller) of the central control system. The built-in algorithm calculates the deviation between the die centerline and the drawing axis and feeds it back to the quick die change module for fine adjustment to ensure that the two are completely aligned and eliminate positioning errors. When the robotic arm 330 places the drawing die 600 on the clamping mechanism 340, the reference surface of the drawing die 600 covers the photoelectric sensing area. The light generated by the light source 220 shines on the uncovered part of the photoelectric sensing area. The uncovered part of the photoelectric sensing area generates a current after being photosensitive. The conductive layer 216 on the outside of the photoelectric area 214 will make the current take the shortest path to reach the array resistor 215 corresponding to the photoelectric sensing area. The four electrodes 212 correspond to the current magnitude at the four edges of the photoelectric sensing area, and the conductive ring 213 can prevent the current signal inside the optical positioning sensor 210 from leaking outward.

[0026] The optical positioning sensor 210 is electrically connected to the PLC520 of the central control system via wires. The position data of the mold reference surface collected by the optical positioning sensor 210 is transmitted to the PLC520 in real time. The PLC520 collects multi-point array data of the mold reference surface, and after preprocessing (filtering, feature extraction), the actual center line of the mold is obtained by fitting using the least squares method. This is then compared with the calibrated drawing axis (drawing axis reference coordinates) to obtain the radial and axial deviations (i.e., the vertical and horizontal coordinate deviations between the actual center and the drawing axis reference coordinates). The central control system drives the servo system of the robotic arm 330 to perform closed-loop position correction based on the deviation value until the deviation meets the accuracy requirements, achieving precise alignment between the mold center line and the drawing axis. The central control system can automatically identify the reference surface of the newly installed mold through the intelligent positioning module, calculate the deviation value between the actual center line of the mold and the drawing axis reference coordinates using the built-in positioning algorithm, and feed it back to the quick mold change module for fine adjustment to ensure that the mold center line and the drawing axis are completely coincident, eliminating positioning errors.

[0027] The central control system controls the light source 220 to turn on, and the optical positioning sensor 210 is used to collect photoelectric signals from the four corners of the drawing die 600 on the mold clamping mechanism 340 as well as the contour data of the drawing die 600; the contour template of the drawing die 600 is square.

[0028] The central control system is used for: preprocessing the photoelectric signals of the four corner points of the drawing die 600 to obtain the position coordinates of the four corner points; fitting the actual center coordinates based on the position coordinates of the four corner points of the drawing die 600; calculating the deviation between the actual center coordinates and the reference coordinates of the drawing axis; determining the coordinate system transformation matrix based on the deviation between the actual center coordinates and the reference coordinates of the drawing axis; denoising the contour data to obtain the denoised coordinates; converting the denoised coordinates to the drawing axis coordinates in the drawing axis coordinate system using the coordinate system transformation matrix; extracting the midpoint coordinates of the four sides of the drawing die 600 based on the drawing axis coordinates; fitting the actual centerline based on the midpoint coordinates of the four sides of the drawing die 600; calculating the deviation between the actual centerline and the reference coordinates of the drawing axis; and controlling the robotic arm 330 to correct the position of the drawing die 600 based on the deviation between the actual centerline and the reference coordinates of the drawing axis.

[0029] The localization algorithm specifically includes the following steps 1 to 4.

[0030] Step 1, reference calibration (executed after initial startup / sensor replacement): Place the standard calibration drawing die 600 on the base 100 and lock it. The optical positioning sensor 210 collects the coordinates of the four corner points of the die. The least squares method is used to fit the center coordinates. The deviation from the reference coordinates (0, 0) of the drawing axis is calculated, and the coordinate system transformation matrix is ​​generated and stored. Specifically, it includes the following steps 1.1 to 1.7.

[0031] Step 1.1, Load standard calibration mold parameters: PLC520 retrieves preset parameters, standard mold side length, theoretical center coordinates (0, 0), and the contour template is square with four sides parallel to the coordinate axes.

[0032] Step 1.2, robotic arm 330 places drawing die 600: PLC520 sends a command to robotic arm 330 to grab drawing die 600 and place it on base 100, and locking mechanism 340 locks it (below the upper limit of 20000N drawing force to avoid deformation of drawing die 600). Step 1.3, start the light source 220 and optical positioning sensor 210: PLC520 controls the light source 220 to turn on, and the optical positioning sensor 210 enters the data acquisition mode.

[0033] Step 1.4, Acquire the corner coordinates of the drawing die 600: The four electrodes 212 respectively acquire the photoelectric signals of the four corners of the drawing die 600. The photoelectric signals are preprocessed. The preprocessing process includes: the photoelectric signals are filtered by RC low-pass filter (cutoff frequency 1kHz) to remove electromagnetic interference, and then converted into digital coordinate signals by 16-bit ADC analog-to-digital conversion. The raw corner coordinate set raw_corners=[(x1, y1), (x2, y2), (x3, y3), (x4, y4)] is output, which includes the position coordinates of the four corners of the drawing die 600.

[0034] Step 1.5, Fit the center and calculate the deviation: Use the least squares method to fit the coordinates of the above corner points, the formula is center actual =( , ), calculate the deviation between the actual center and the reference coordinate (0, 0) of the drawing axis: x-axis deviation offset x =center x -0, ordinate offset y =center y -0, center x With center y Let x1 = 10.5, y1 = -3.2, x3 = 14.7, y3 = 5.8, and offset be the x and y coordinates of the actual center. x =12.6, offset y =1.3.

[0035] Step 1.6, Generate coordinate system transformation matrix: Construct a 2D rigid body transformation matrix, i.e., a coordinate system transformation matrix (only compensating for translation errors). The data is stored in the PLC520 register. Its working principle is as follows: Assume there is a target point (x5, y5), where x5 and y5 are the x and y coordinates of the target point, respectively. In homogeneous coordinates, this is represented as a column vector (x5, y5, 1). Multiplying this column vector (x5, y5, 1) by matrix M on the left yields the result. The new coordinates are: (x5 + offset) x y5+offset y ).

[0036] Step 1.7, Verify calibration accuracy: Fine-tune the standard mold of the robotic arm 330 to the theoretical center, re-collect the corner coordinates, and verify that the deviation is ≤0.001mm. If it is not qualified, repeat steps 1.2-1.6.

[0037] Step 2, Data Acquisition and Preprocessing (executed after each mold change): The optical positioning sensor 210 acquires mold contour data at a frequency of 100Hz, which is then processed by RC low-pass filtering, moving average filtering, and 3... After denoising according to the criteria, the coordinates are calibrated to the drawing axis coordinate system through the coordinate system transformation matrix; including the following steps 2.1 to 2.5.

[0038] Step 2.1, collect contour data: After the robotic arm 330 grasps the drawing mold 600 to be positioned and places it on the base 100, the optical positioning sensor 210 collects the photoelectric signal of the mold reference surface at a frequency of 100Hz, and continuously collects 10 frames of data.

[0039] Step 2.2: Perform RC low-pass filtering on the contour data obtained in step 201 to obtain RC low-pass filtered contour data.

[0040] Step 2.3, Time Domain Filtering: Perform moving average filtering on the contour data after RC low-pass filtering, using the following formula: This reduces instantaneous fluctuations caused by vibration; The coordinates are after moving average filtering; The coordinates are after RC low-pass filtering.

[0041] Step 2.4, Spatial Domain Denoising: Outliers are removed using the 3σ criterion. The specific logic is as follows: calculate the coordinate mean. Standard deviation , Original coordinates , Indicates counting, The number of original coordinates; remove those that meet the criteria. For outliers, retain valid contour data and use it as coordinates after denoising.

[0042] Step 2.5, Coordinate System Transformation: Use the coordinate system transformation matrix M from Step 1.6 to transform the denoised coordinates to the drawing axis coordinate system. The formula is: , M ij The element in the i-th row and j-th column of the transformation matrix M ; and These are the x and y coordinates of the denoised coordinates, respectively; and These are the abscissa and ordinate in the drawing axis coordinate system, respectively.

[0043] Step 3, Deviation Calculation (executed after each mold change): Step 3.1, Extract the midpoints of the four sides of the mold: Left midpoint: mid left =(x min , ), x min This represents the minimum value in the abscissa of the pull-out axis. These represent the minimum and maximum ordinates of all points on the left side of the drawing die 600; right midpoint: mid right =(x max , ), x max This represents the maximum value in the abscissa of the pull-out axis. These represent the minimum and maximum ordinates of all points on the right side of the drawing die 600; Midpoint: top =( y max ), These are the minimum and maximum ordinates of all points on the upper side of the drawing die 600, respectively. max The maximum value in the vertical coordinate of the pull-out axis; lower midpoint: mid bottom =( y min ), These represent the minimum and maximum ordinates of all points on the lower side of the drawing die 600, respectively. min The minimum value in the ordinate of the drawing axis; Step 3.2 Fitting the actual centerline: Calculate the geometric center of the midpoint set, i.e., the actual centerline formula is: , [0] represents the actual centerline coordinates, and [0] represents the x-coordinate component. and Let x be the left midpoint and right midpoint respectively, and [1] be the y-coordinate component. and These are the y-coordinates of the upper and lower midpoints, respectively. Calculate the deviation between the actual centerline and the reference coordinates of the drawing axis, including: x-coordinate deviation. ordinate deviation .

[0044] Step 4, Step Correction (executed when deviation exceeds threshold), includes: Step 4.1, Threshold Judgment: If the absolute value of the deviation between the actual centerline and the reference coordinates of the drawing axis is greater than the set deviation value (the set deviation value can be 0.01mm), that is, if... mm or mm, triggering the correction process; Step 4.2, PLC520 generates the robot arm 330 fine-tuning instruction: calculate the reverse compensation amount adjust. x =-Δx, adjust y =-Δy, adjust x and adjust y These are the x-axis and y-axis components of the reverse compensation amount, respectively, and the fine-tuning amount is decomposed in 0.005mm steps. The formula is as follows: , This is the fine-tuning compensation amount in 0.005mm steps. Step 4.3, the robotic arm performs fine-tuning: PLC520 sends a single-step command to robotic arm 330, pausing for 50ms after each step to wait for the mold to stabilize. Step 4.4, re-verification: After fine-tuning, repeat steps 2-3 to recalculate the deviation until it is ≤0.01mm; if three consecutive fine-tuning attempts fail, a sensor fault alarm is triggered.

[0045] In some specific embodiments, the drawing die 600 is hexahedral in shape, and the diameter of the wire drawn by the die is within 5.5mm (e.g., 5.5mm, 4.6mm, 3.9mm, 3.3mm, 2.8mm, 2.4mm, 2.1mm, 1.9mm, 1.6mm), and the corresponding drawing force of the metal wire should be less than 20000N.

[0046] In some specific embodiments, the quick mold change module further includes a mold storage bin 310 and a waste mold collection box 320. The mold storage bin 310 is used to hold drawing dies 600 to be used. The top of the mold storage bin 310 is open. The interior of the mold storage bin 310 is divided into several areas, and different areas are used to place drawing dies 600 of different specifications, which facilitates precise calling by the central control system. The waste mold collection box 320 is located next to the mold storage bin 310. The waste mold collection box 320 is used to hold drawing dies 600 after use. The top of the waste mold collection box 320 is open. The interior of the waste mold collection box 320 is equipped with a cushioning pad to prevent damage when waste molds fall. The side of the waste mold collection box 320 is equipped with a removable door panel, which facilitates the staff to clean the waste molds regularly. In use, the robotic arm 330 can grab the drawing die 600 on the base 100 according to production instructions and accurately transport it to the waste die collection box 320, or grab the drawing die 600 of a specified specification from the die storage bin 310 and accurately transport it to the die locking mechanism 340 on the base 100. The components of the quick die changing module work together to realize the automatic grabbing, replacement, locking of the drawing die 600 and the recycling of waste dies.

[0047] In some specific embodiments, the robotic arm 330 adopts a six-axis robotic arm structure. Its fixed end is mounted on the ground on one side of the base 100 via a bracket, and its free end is equipped with a pneumatic gripper. The inner side of the pneumatic gripper is provided with an anti-slip rubber pad, which can adaptively adjust the gripping force according to the specifications of the drawing die 600 to avoid damage to the drawing die 600 during the gripping process. The range of motion of the robotic arm 330 covers the entire working area of ​​the mounting groove of the base 100, the die storage bin 310, and the waste die collection box 320.

[0048] In some specific embodiments, a safety monitoring module is also included. The safety monitoring module includes a force sensor 410 and a displacement sensor 420. The force sensor 410 is a piezoelectric force sensor and is mounted on the clamping claw 3402 of the clamping mechanism 340. When the clamping claw 3402 locks the drawing die 600, the force sensor 410 is used to monitor the axial impact force on the drawing die 600 during the drawing process. The displacement sensor 420 is a spectral confocal displacement sensor and is mounted on the clamp 430 on the base 100. The displacement sensor 420 is positioned facing the end of the drawing die 600 on the clamping mechanism 340 and is used to monitor the displacement generated by the drawing die 600 during the drawing process. Both the force sensor 410 and the displacement sensor 420 are electrically connected to the central control system and transmit the monitoring data to the PLC 520 in real time. The PLC 520 compares the data with a preset safety threshold. When the monitored value exceeds the threshold, an alarm signal is immediately triggered, and the equipment is stopped to prevent safety accidents.

[0049] In some specific embodiments, the central control system includes a display screen 510 and a PLC 520, which coordinates the intelligent positioning module and the rapid die-changing module. Its built-in control algorithm is used to: receive production orders and parse the specifications of the required drawing die 600; direct the rapid die-changing module to execute a fully automatic die-changing sequence; process data from the optical positioning sensor 210; monitor the signals from the force sensor 410 and the displacement sensor 420; calculate the usage time of the drawing die 600, enabling predictive maintenance of the drawing die 600 and prompting for replacement before the drawing die 600 reaches its critical lifespan.

[0050] The central control system is located on the equipment's operation panel, including a display screen 510 and a PLC 520. The display screen 510 uses a touchscreen structure to display production order information, equipment operating status, mold positioning data, safety monitoring data, and fault alarm information. It also supports manual input of commands and modification of parameters by operators. The PLC 520 uses a Siemens S7-1500 series programmable logic controller as the control core of the entire device. It incorporates a dedicated control algorithm based on a finite state machine architecture to achieve switching of working states and timing linkage between modules. Combining a least-squares mold centerline fitting algorithm with a PID closed-loop control algorithm, it completes real-time compensation and calibration of mold positioning accuracy (see steps 1-4 above). It also embeds threshold judgment and anomaly response algorithms to monitor safety status in real time and trigger the intelligent positioning module and rapid mold change module for linkage protection when deviations occur. Ultimately, it achieves coordinated control of rapid mold change, accurate positioning, and safe operation, improving the automation level and operational reliability of mold processing. It is responsible for coordinating the work of the intelligent positioning module, rapid mold change module, and safety monitoring module. The PLC520 is electrically connected to the sensors and actuators of each module via a bus, enabling bidirectional data transmission and precise command issuance. It also features a built-in mold life calculation algorithm.

[0051] In some specific embodiments, the central control system further includes a model training module, which is used to: train a trained life prediction model, specifically including the following steps S1 to S3.

[0052] S1, Obtain the dataset; the dataset includes the die operation data for each historical moment in the 600-year historical time period of the drawing die and the remaining life label for each historical moment; the die operation data includes basic historical data and real-time monitoring data.

[0053] First, data preprocessing is performed: the input mold operation data is preprocessed, specifically including the following steps S1.1~S1.2.

[0054] S1.1 Data Definition: The remaining life of the mold is the target variable y, and a set of basic historical data and real-time monitoring data are the covariate features. The characteristics of each covariate are as follows: , ... The historical target variable is y. D-1 y D-2 ...; Historical covariate characteristics are , ...; The target variable to be predicted is y D .

[0055] S1.2 Data mapping processing: using interpolation functions Map data from different acquisition frequencies to the same length; Basic historical data X T2 This refers to 288 time-point data points (T2=288) at 5-minute intervals for the mold management system and production records. This includes mold material, mold bore / size, theoretical lifespan at factory, theoretical maximum drawing time, historical drawing time, historical process conditions, historical failure causes (wear, chipping, scratches, normal failure), and historical maintenance, grinding, and repair records. Real-time monitoring data X T1 The data published for sensors and industrial control systems includes 48 time points (T1=48) at 30-minute intervals, including drawing speed, real-time drawing force, inlet wire diameter, outlet wire diameter, lubricant temperature, copper rod material, cumulative drawing time, micro-displacement, die temperature, surface roughness of copper wire at die exit, and scratches on copper wire at die exit.

[0056] Using interpolation function X the basic historical data T2 Mapped to real-time monitoring data X T1 Data with the same dimensions, specifically: mapped basic historical data. , , The collection point number for the basic historical data This represents the integer value i / 6 rounded up, for example, i=5. The value is 1, and the 5th 288th time point (corresponding to the 25th minute of basic historical data collection) is mapped to the 1st time point of the 48th time point ( (corresponding to 0-30 minutes), the process parameter values ​​at this time point are processed by linear interpolation to ensure that the two types of data have consistent dimensions, and the mapped basic historical data. With real-time monitoring data X T1 The dimensions are the same, all being 48 time points with 30-minute intervals.

[0057] S2, Hyperparameter Tuning and Optimization: The input data and hyperparameters are tuned and optimized, and the results are concatenated as the optimized hyperparameters. The hyperparameters are tuned and optimized using the XGBoost model combined with Bayesian optimization. Specifically, the dataset is divided into three time periods: a first time period dataset, a second time period dataset, and a third time period dataset. These three time periods are consecutive. The first time period dataset, the second time period dataset, and the third time period dataset are designated as the training set, the validation set, and the test set, respectively. A first model is fitted based on the training set. Using the Bayesian optimization method, the model parameters of the first model are updated based on the validation set, and the model parameters with the smallest loss function value are selected as the optimized hyperparameters. The hyperparameter tuning and optimization in S2 includes the following steps S2.1~S2.3.

[0058] S2.1 Input data adjustment and optimization: The current day is used as the third time period, and the covariate feature x for that day... D-1 Input X for the test set test Take n days back as the second time period, and in the second time period x D-2 x D-3 , ...x D-n Input X to the validation set val Let's take m days back as the first time period, and x in the first time period... D-n-1 x D-n-2 , ...x D-n-m Input X to the training set train The target variable y corresponds to the covariate characteristics in the second time period. D-1 y D-2 , ...y D-n+1 Output Y for the validation set val The target variable y corresponding to the covariate features in the first time period D-n y D-n-1 , ...y D-n-m+1 Output Y for the training set train ; where n and m can be optimized within the ranges of (20, 167) and (60, 667) respectively.

[0059] S2.2 Hyperparameter Tuning and Optimization: Based on the XGBoost model, the hyperparameters are optimized using the Bayesian optimization method within a given adjustment range.

[0060] The specific content is as follows: During model training, the training objective is to find a set of hyperparameters that, under these hyperparameters, fit the input X of the training set. train And label Y train Obtain the first model And the covariate features X of the validation set val Make predictions so that the remaining life prediction value With the remaining lifetime label Y of the validation set val The error between them is the smallest; since the model output is the remaining life of the mold, it belongs to a continuous numerical regression prediction problem; the mean squared error can intuitively reflect the degree of deviation between the predicted value and the true value, and a higher penalty is applied to larger prediction deviations to ensure that the model does not deviate significantly from the true life. The formula is: ; Where, f(X) val ) represents the predicted remaining lifetime; L is the loss function, using mean squared error loss, expressed by the following formula: ; N val The number of samples in the validation set; For the verification set The remaining lifetime label for each sample; For the verification set The remaining lifetime prediction value of each sample. Bayesian optimization is used to optimize the optimizer, specifically as follows: Within the hyperparameter range, a set of hyperparameter points is randomly selected as the initial point P0. The validation error L0 of P0 is calculated, and P0 is added to the known point set IP. Based on the assumption that other unknown points in the IP hypothesis space follow a Gaussian distribution, the hyperparameter point P1 with the smallest expected validation error on the Gaussian distribution is selected. The actual validation loss error L1 of P1 is calculated, and P1 is added to IP. The Gaussian distribution of other points in the hypothesis space is updated based on IP. This process of selecting new hyperparameter points is repeated until the number of known points in IP reaches a set threshold (preferably 20-30). The point with the smallest validation loss error L1 within IP is then selected. min The corresponding hyperparameter point P min The learning rate is used as the final optimized hyperparameter. The adjustment ranges for tree depth, number of trees M, and subsampling ratio are (0.01, 0.3), (3, 8), (100, 300), and (0.5, 1.0), respectively.

[0061] S2.3, Concatenating and Optimizing Hyperparameters: The input data parameters (n, m) and model hyperparameters are concatenated together to form the optimization hyperparameters.

[0062] S3, Model Prediction: The GBRT model is configured with optimized hyperparameters. The validation set samples and training set samples are concatenated in chronological order as input to the model for fitting, resulting in a trained lifetime prediction model. This trained lifetime prediction model is then used to predict the remaining lifetime of the molds on the test set, outputting the mold's remaining lifetime prediction results. Specifically: The optimized hyperparameters are used as the initial model parameters for the GBRT model to determine the initial model; samples from the validation set and training set are concatenated in chronological order to obtain a concatenated dataset; the concatenated dataset includes the concatenated samples and the remaining lifetime label for each concatenated sample; the initial model is trained using the concatenated dataset to obtain a trained lifetime prediction model; the trained lifetime prediction model is tested using the test set to obtain the remaining lifetime prediction results for each mold's operating data in the test set.

[0063] The specific content of the model prediction is as follows: The GBRT model is reconfigured with optimized hyperparameters, and the validation set samples (X) are used... val Y val ) and training set samples (x train Y train The samples are concatenated in chronological order to obtain the concatenated dataset. ( (For time-dimension splicing), splicing sample input X fit =[X train ;X val ], the remaining lifetime label Y of the spliced ​​sample fit =[Y train ;Y val The final GBRT prediction model is obtained through training, and this final GBRT prediction model is used as the trained lifetime prediction model. The trained lifetime prediction model is then used to test the input covariate features X on the test set. test To make predictions, the GBRT model consists of multiple regression trees, and the iterative formula for the GBRT model is: ; where f k (x) represents the predicted output of the k-th regression tree in the GBRT model; f k-1 (x) represents the predicted output of the (k-1)th regression tree; h k (x) represents the residuals of the first k-1 trees fitted to the k-th regression tree, which yields the predicted target variable y. 1,D y 2,D , ..., y 288,D The average value is taken as the final prediction result, and a replacement prompt is given before the drawing die 600 reaches its lifespan critical point. This enables early prediction and intelligent control of the replacement timing of the drawing die 600, and a replacement prompt is issued through the display screen 510 before the lifespan critical point is reached.

[0064] After the above training process, a trained life prediction model is obtained. The central control system is used to: acquire basic historical data and real-time monitoring data of the drawing die 600 for the target time period; input the basic historical data and real-time monitoring data of the drawing die 600 for the target time period into the trained life prediction model to obtain the remaining life prediction result of the drawing die 600; and issue a replacement prompt based on the remaining life prediction result of the drawing die 600.

[0065] This embodiment fundamentally eliminates manual installation errors through an intelligent positioning module, ensuring the dimensional accuracy and consistency of drawn products. The rapid mold change module reduces mold change time from tens of minutes to minutes, significantly improving equipment utilization and production efficiency. The dual protection provided by the mold locking mechanism 340 and the safety monitoring module completely eliminates the risk of safety accidents such as mold displacement and mold flying. This embodiment enables fully automated operation from order placement to mold replacement, calibration, production, and monitoring.

[0066] In the description of this invention, it should be understood that the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0067] In the description of this invention, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0068] If this invention discloses or relates to components or structural parts that are fixedly connected to each other, then, unless otherwise stated, a fixed connection can be understood as: a detachable fixed connection (e.g., using bolts or screws) or a non-detachable fixed connection (e.g., riveting, welding). Of course, a fixed connection can also be replaced by an integral structure (e.g., manufactured in one piece using a casting process) (except where it is obviously impossible to use an integral molding process).

[0069] In addition, unless otherwise stated, the terms used in any of the technical solutions disclosed in this invention to indicate positional relationships or shapes include states or shapes that are similar to, close to, or approximate with those states or shapes.

[0070] Any component provided by this invention can be assembled from multiple individual components or can be a single component manufactured by a one-piece molding process.

[0071] It should be noted that the structures, proportions, sizes, etc., depicted in the accompanying drawings of this specification are only used to complement the content disclosed in the specification, so as to enable those skilled in the art to understand and read them, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0072] It should also be noted that in the embodiments of this application, the same reference numerals are used to denote the same component or the same part.

[0073] Any adaptive changes made according to actual needs are within the scope of protection of this invention.

[0074] Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. Furthermore, those skilled in the art will recognize that, based on the ideas of this invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A die-changing device for a drawing die with intelligent positioning function, characterized in that: include: The system includes an intelligent positioning module, a rapid mold changing module, and a central control system. The intelligent positioning module comprises a light source and an optical positioning sensor. The rapid mold changing module comprises a robotic arm and a mold locking mechanism. The robotic arm is used to grasp the drawing die, and the mold locking mechanism can lock or release the drawing die. The light source is positioned towards the mold locking mechanism, and the optical positioning sensor is mounted on the mold locking mechanism to monitor the position of the drawing die on the mold locking mechanism. The robotic arm, the optical positioning sensor, and the mold locking mechanism are all electrically connected to the central control system. The mold clamping mechanism includes a mold clamping body and a plurality of mold clamping claws. The optical positioning sensor is disposed on the mold clamping body. The plurality of mold clamping claws are arranged along the circumference of the mold clamping body. The mold clamping claws are made of shape memory alloy. When the mold clamping claws are energized, they can lock the drawing die. When the power to the clamping claw is cut off, the clamping claw can release the drawing die. The central control system is used to: acquire basic historical data and real-time monitoring data of the drawing die for a target time period; input the basic historical data and real-time monitoring data of the drawing die for a target time period into a trained life prediction model to obtain the remaining life prediction result of the drawing die; and issue a replacement prompt based on the remaining life prediction result of the drawing die. The central control system further includes a model training module, which is used for: Obtain the dataset; the dataset includes the die operation data for each historical moment in the historical time period of the drawing die and the remaining life label for each historical moment; the die operation data includes basic historical data and real-time monitoring data; The dataset is divided into a first time period dataset, a second time period dataset, and a third time period dataset according to time; the first time period, the second time period, and the third time period are consecutive time periods; the first time period dataset, the second time period dataset, and the third time period dataset are respectively designated as the training set, the validation set, and the test set; The first model is obtained by fitting the training set. Using the Bayesian optimization method, the model parameters of the first model are updated based on the validation set, and the model parameters with the smallest loss function value are selected as the optimized hyperparameters. The optimized hyperparameters are used as the initial model parameters for the GBRT model to determine the initial model; The samples in the validation set and the samples in the training set are concatenated in chronological order to obtain a concatenated dataset; the concatenated dataset includes concatenated samples and the remaining lifetime label of each concatenated sample. The initial model is trained using the spliced ​​dataset to obtain a trained lifespan prediction model; The trained lifetime prediction model is tested using the test set to obtain the remaining lifetime prediction results for each mold operation data in the test set.

2. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: The manufacturing process of the clamping claw includes: First, the raw material for the mold clamping claw is heat-treated, and then the heat-treated raw material is tested to obtain the martensitic transformation initiation temperature M. s Martensitic normal phase transformation end temperature M f Martensitic reverse phase transformation initiation temperature A s and the end temperature of the reverse phase transformation of martensite A f The raw materials were trained using the stress-induced martensite cycle training method. During training, the raw materials are heated to temperature A, the end temperature of the reverse martensitic transformation. f The above process involves applying a bending load to the raw material until the applied load induces a stress-induced martensitic transformation and produces a predetermined deformation that matches the function of the clamping claw. The material is then held at the same bending load and temperature, unloaded, and cooled to the martensitic transformation end temperature M. f Now, complete one training session; repeat the training process described above.

3. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: The center of the illumination area formed by the light source coincides with the center of the mold-locking mechanism.

4. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: The optical positioning sensor includes a substrate layer and a conductive ring, a photoelectric sensing area, four electrodes, and four array resistors disposed on the substrate layer. The photoelectric sensing area is rectangular, and the four electrodes are located on the outer sides of the four corners of the photoelectric sensing area. An array resistor is disposed between adjacent electrodes. The conductive ring is located on the outer sides of the electrodes and the array resistors. The photoelectric sensing area includes several photoelectric areas, and a conductive layer is disposed on the edge of each photoelectric area.

5. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: The quick mold change module also includes a mold storage bin and a waste mold collection box. The mold storage bin is used to hold drawing dies to be used. The interior of the mold storage bin is divided into several areas, and different areas are used to place drawing dies of different specifications. The waste mold collection box is used to hold drawing dies after use. The waste mold collection box is equipped with a cushioning pad inside, and the side of the waste mold collection box is equipped with a removable door panel.

6. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: It also includes a safety monitoring module, which includes a force sensor and a displacement sensor. Both the force sensor and the displacement sensor are electrically connected to the central control system. The force sensor is mounted on the mold clamping mechanism and is used to monitor the force on the drawing die during the drawing process. The displacement sensor is mounted toward the end of the drawing die on the mold clamping mechanism and is used to monitor the displacement generated by the drawing die during the drawing process.

7. The die changing device for a drawing die with intelligent positioning function according to claim 1, characterized in that: The central control system controls the light source to turn on, and the optical positioning sensor is used to collect photoelectric signals from the four corners of the drawing die on the mold clamping mechanism and the contour data of the drawing die. The outline template of the drawing die is square; The central control system is used for: preprocessing the photoelectric signals of the four corner points of the drawing die to obtain the position coordinates of the four corner points; fitting the actual center coordinates based on the position coordinates of the four corner points of the drawing die; calculating the deviation between the actual center coordinates and the drawing axis reference coordinates; determining the coordinate system transformation matrix based on the deviation between the actual center coordinates and the drawing axis reference coordinates; denoising the contour data to obtain denoised coordinates; using the coordinate system transformation matrix to convert the denoised coordinates into drawing axis coordinates in the drawing axis coordinate system; extracting the midpoint coordinates of the four sides of the drawing die based on the drawing axis coordinates; fitting the actual centerline based on the midpoint coordinates of the four sides of the drawing die; calculating the deviation between the actual centerline and the drawing axis reference coordinates; and controlling the robotic arm to correct the position of the drawing die based on the deviation between the actual centerline and the drawing axis reference coordinates.

Images