A flexible intelligent heating method based on heat source discretization

Abstract

The application discloses a flexible intelligent heating method based on heat source discretization, comprising the following steps: obtaining the contour, posture and target temperature information of each performance partition of a plate to be heated; controlling the opening and closing of corresponding discrete heat sources in a heating furnace and setting heating parameters according to the above information; feeding the plate into the heating furnace for partition heating; detecting the actual temperature of each partition after heating; comparing the actual temperature with the target temperature to obtain deviation information; and correcting and optimizing the heating parameters based on the deviation. The method effectively solves the problems of low efficiency, difficulty in realizing conformal and gradient heating of the traditional heating process by means of the layout of discrete, independently controllable heat sources, intelligent identification and feedback control, realizes efficient, accurate and flexible partition temperature control, and meets the production rhythm requirements of high-performance automobile component hot forming and the like.

Description

Technical Field

[0001] This invention relates to a material thermal processing method, and more particularly to a flexible intelligent heating method based on heat source discretization. Background Technology

[0002] Lightweighting of automobiles is a key path to achieving energy conservation and emission reduction goals, and hot stamping is a core technology for producing ultra-high strength steel (UHSS) automotive components. It obtains a martensitic structure with extremely high strength by forming at austenitic temperatures and then rapidly quenching. However, to meet collision safety requirements, certain body structural components (such as B-pillars and door anti-collision beams) need to possess gradient properties, meaning different areas have different strengths and toughness—essentially "custom strength" or "performance gradient" components. For example, the upper part of the B-pillar requires high strength to resist intrusion, while the lower part needs higher toughness to absorb collision energy.

[0003] Currently, the mainstream processes for preparing such performance gradient components mainly include laser welded plate (TWB) technology, thermal partitioning mold technology, and direct hot stamping technology after partitioning heating. Laser-welded plate (TWB) technology involves pre-welding steel plates of different strengths or thicknesses into a single unit, followed by hot stamping. This method suffers from drawbacks such as weakened performance in the welded area, limited design freedom, the need for additional welding steps, and higher costs.

[0004] Thermal zoning mold technology involves setting different cooling rates in different areas of the mold to achieve performance zoning. This method is complex in mold design, costly, difficult to control thermal deformation, and has a high scrap rate.

[0005] The technology of direct hot stamping after zone heating is exemplified by the TemperBox process from the Swedish company. This process involves partitioning and shielding the sheet metal within a heating furnace to create a gradient temperature field in different areas of the billet, followed by integrated stamping and quenching. This method currently offers superior overall performance, high design flexibility, and high repeatability.

[0006] However, existing zoned heating technologies, especially the TemperBox process, have a significant bottleneck: low heating efficiency. To achieve a temperature gradient, certain areas need to be shaded, which severely hinders heat conduction and convection, causing a significant decrease in the heating rate of unshaded areas (e.g., from 3.0°C / s to approximately 1.8°C / s). This is severely out of sync with the approximately 20-second production cycle of mainstream hot stamping production lines, limiting production efficiency. Another approach is zoned cooling after heating, where the entire sheet is first heated to a high temperature, and then localized areas are forcibly cooled to achieve the gradient. While this method has high heating efficiency, it consumes a huge amount of energy, and excessively high heating temperatures can lead to coarsening of the microstructure and deterioration of performance in some areas, as well as insufficient temperature control precision.

[0007] In summary, existing technologies struggle to achieve graded heating of complex-shaped sheet materials while maintaining high efficiency, high precision, low energy consumption, and high design freedom. Therefore, an innovative heating method is urgently needed to resolve the conflict between heating efficiency and forming cycle time, and to achieve precise, flexible, and intelligent zoned temperature control. Summary of the Invention

[0008] To address the shortcomings of the aforementioned technologies, this invention provides a flexible intelligent heating method based on heat source discretization.

[0009] To solve the above technical problems, the technical solution adopted by the present invention is: a flexible intelligent heating method based on heat source discretization, comprising the following steps: S1. Obtain the contour information, posture information, and target heating temperature information of each performance zone of the plate to be heated; S2. Based on contour information and attitude information, control the opening of the heat source at the corresponding position in the heating furnace, and set the heating parameters for the corresponding heat source based on the heating temperature information of each target. S3. The plate to be heated is transported into the heating furnace and heated in sections by the heat source set in step S2. S4. Temperature detection is performed on the heated board to obtain the actual temperature information of each performance zone; S5. Compare the actual temperature information with the target heating temperature information to obtain the temperature deviation information; S6. Based on temperature deviation information, correct and optimize the heating parameters.

[0010] Furthermore, in step S1, obtaining the contour and posture information of the plate to be heated is achieved through a visual recognition system deployed at the front end of the heating furnace, specifically including: obtaining the planar contour information of the plate through 2D visual imaging; and obtaining the spatial posture and surface information of the plate through 3D contour reconstruction.

[0011] Furthermore, in step S2, the heat dissipation source is a short-wave infrared lamp; the heating parameters include at least one of the following: heating power, on / off sequence, and distance between the heat source and the surface of the plate.

[0012] Furthermore, in step S4, temperature detection employs a combination of an infrared thermal imager and a contact thermocouple; the infrared thermal imager is used to acquire the temperature distribution across the entire board surface, while the contact thermocouple is used for point-to-point calibration at key locations.

[0013] Furthermore, in step S5, the comparison includes: Calculate the actual temperature uniformity within each performance zone and compare it with the target uniformity range; Calculate the actual temperature gradient value at the boundary of adjacent performance zones and compare it with the target temperature gradient value.

[0014] Furthermore, in step S6, the heating parameters are corrected and optimized based on the temperature deviation information, specifically as follows: Store the temperature deviation information and corresponding heating parameters of this heating process in the system database; Based on historical databases, a mapping model based on board information, heating parameters, and heating effect is established through intelligent self-learning algorithms to optimize the heating parameters of subsequent similar boards.

[0015] Furthermore, step S3 also includes: The heated sheet metal is quickly transferred into a mold for hot stamping and pressure holding quenching. Subsequently, strength tests were conducted on different performance zones of the formed component to verify whether the heating effect met the final performance requirements.

[0016] Furthermore, the heat sources are arranged in a matrix within the heating furnace, with a center-to-center distance of 20mm between adjacent heat sources, and each heat source is independently controllable, in order to achieve precise micro-area control of the temperature field on the surface of the board.

[0017] Furthermore, the intelligent self-learning algorithm specifically includes the following steps: Temperature deviation and deviation change rate are input into fuzzy rule base, and optimized PID parameters are obtained through fuzzy inference; Based on real-time temperature feedback and PID parameter calculation, the power regulation amount is dynamically controlled by PWM signal to regulate the power output of each heat dissipation source. When the temperature deviation exceeds 5%, a power compensation of 10%-20% is added, and fine-tuning is performed when the temperature approaches the target temperature to avoid overshoot.

[0018] Furthermore, the system also includes a process database and a self-learning optimization module, used for: Store the characteristics of the plate material, heating parameters, measured temperature field and deviation information for each heating cycle; A machine learning model is trained based on historical data to establish a mapping relationship between board characteristics, heating parameters, and heating effect; For boards with new shapes or materials, the system automatically recommends initial heating parameters and continuously optimizes the control strategy during the production process.

[0019] This invention discloses a flexible intelligent heating method based on heat source discretization, which is an innovative heating method that solves the contradiction between heating efficiency and forming cycle time, and achieves precise, flexible and intelligent zoned temperature control. Attached Figure Description

[0020] Figure 1This is a schematic diagram showing the arrangement of heat dissipation sources.

[0021] Figure 2 This is a schematic diagram of performance gradient heating.

[0022] Figure 3 This is a schematic diagram for verifying the uniformity of strength.

[0023] Figure 4 This is a schematic diagram of intelligent flexible heating for trapezoidal blanks with heat dissipation sources.

[0024] Figure 5 This is a schematic diagram of intelligent flexible heating for a circular blank corresponding to a heat source.

[0025] Figure 6 This is a schematic diagram of intelligent flexible heating for a circular blank corresponding to a heat source.

[0026] Figure 7 This is a timing diagram showing the start-up, shutdown, and power changes of each heat source during the heating process.

[0027] Figure 8 This is a diagram showing the vertical layout.

[0028] Figure 9 This is a temperature difference diagram arranged vertically.

[0029] Figure 10 This is a diagram showing the horizontal layout.

[0030] Figure 11 This is a horizontally arranged temperature difference diagram. Detailed Implementation

[0031] This invention discloses a flexible intelligent heating method based on heat source discretization, comprising the following steps: A flexible intelligent heating method based on heat source discretization includes the following steps: S1. Obtain the contour information, posture information, and target heating temperature information of each performance zone of the material to be heated. The acquisition of the contour and posture information of the material is achieved through a vision recognition system deployed at the front end of the heating furnace. Specifically, this includes: acquiring the planar contour information of the material through 2D visual imaging; and acquiring the spatial posture and surface information of the material through 3D contour reconstruction. During the hardware deployment phase, the 2D vision system uses an area scan camera or a line scan camera, paired with a telecentric lens to eliminate perspective distortion. It also employs a combination of coaxial light sources and ring or strip light source arrays, combined with adaptive light source control technology to suppress steel plate reflection and ambient light interference. The camera and light source are fixed to a gantry frame above the production line and equipped with a temperature-resistant and dustproof protective cover. The camera's intrinsic parameters, distortion coefficients, and pixel equivalent calibration are completed using the Zhang Zhengyou calibration method to establish the correspondence between pixels and physical dimensions. The 3D contour reconstruction system, due to the characteristics of the selected automotive door ring material, uses laser triangulation. Laser triangulation requires the deployment of a line laser emitter and an industrial camera, and hand-eye calibration must be completed to achieve coordinate system consistency with the heating furnace control system.

[0032] In the software processing stage, 2D images are first subjected to median filtering, Gaussian filtering, or bilateral filtering to remove noise. Brightness and contrast are optimized through histogram equalization and grayscale stretching. Then, adaptive threshold binarization or Otsu's method is used to complete background segmentation. After morphological operations are used to fill holes and eliminate artifacts, precise edges are extracted using operators such as Canny. After 3D data acquisition, point cloud noise is removed through statistical filtering and radius filtering. The segmented scanning data of long pieces need to be registered and stitched using the ICP algorithm. Then, the data volume is reduced by downsampling. Methods such as Poisson reconstruction are used to fill holes and convert them into triangular mesh models. Finally, the 6DoF attitude parameters of the piece are calculated.

[0033] The combined implementation of the two methods not only obtains high-precision planar contours through 2D vision, but also makes up for the lack of spatial posture and surface information through 3D reconstruction, providing comprehensive and accurate material feature data for the heating furnace control system, and supporting the precise control of flexible intelligent zone heating.

[0034] S2. Based on contour and attitude information, control the activation of the heat source at the corresponding position within the heating furnace, and set heating parameters for the corresponding heat source based on the target heating temperature information; the heat source is a short-wave infrared lamp; the heating parameters include at least one of the following: heating power, on / off sequence, and distance between the heat source and the surface of the sheet material. Specifically, the heating power is adjustable within 200W / cm²; the on / off sequence is achieved by an infrared recognition system at the front of the furnace, which automatically turns on the heating switch after the system recognizes the sheet material entering the furnace; the furnace has a real-time point temperature measuring device that automatically shuts off the heating system after reaching the desired temperature; the distance between the heat source and the surface of the sheet material is 40mm.

[0035] S3. The plate to be heated is transported into the heating furnace and heated in sections by the heat source set in step S2. The heated plate is then quickly transferred to the mold for hot stamping and pressure holding quenching. Subsequently, the strength of different performance sections of the formed component is tested to verify whether the heating effect meets the final performance requirements.

[0036] S4. Temperature detection is performed on the heated board to obtain the actual temperature information of each performance zone. Temperature detection is performed using a combination of infrared thermal imager and contact thermocouple. The infrared thermal imager is used to obtain the temperature distribution of the entire board, and the contact thermocouple is used for point calibration at key locations.

[0037] S5. Compare the actual temperature information with the target heating temperature information to obtain temperature deviation information; the comparison includes calculating the actual temperature uniformity within each performance zone and comparing it with the target uniformity range; calculating the actual temperature gradient value at the boundary of adjacent performance zones and comparing it with the target temperature gradient value. Based on the temperature deviation information, the heating parameters are corrected and optimized, specifically as follows: Store the temperature deviation information and corresponding heating parameters of this heating process in the system database; Based on historical databases, an intelligent self-learning algorithm is used to establish a mapping model based on board information, heating parameters, and heating effect to optimize the heating parameters of subsequent similar boards. The intelligent self-learning algorithm adopts a "PID + fuzzy control" composite process. First, a fuzzy rule base is established with temperature deviation and deviation change rate as inputs and PID parameter correction as outputs. After fuzzification and inference, the optimized PID parameters are obtained. Then, the power adjustment is calculated by combining real-time temperature feedback. The heat source power is dynamically adjusted through PWM or analog output. When the deviation exceeds 5%, 10%-20% power compensation is added. When the temperature approaches the target temperature, fine-tuning is performed to avoid overshoot. The process is iterative to maintain temperature stability.

[0038] The heat source groups are divided according to the area, shape complexity, and temperature control accuracy of the performance zones: small, regular zones with high precision requirements adopt a "one-to-one" principle (one zone corresponds to one independent heat source group), while large, irregular zones adopt a "many-to-one" principle (multiple heat sources work together to cover one zone). The coordinate calibration and precise binding of the material zone and the heat source group are achieved through a visual recognition system. Each heat source group is equipped with an independent sensing and control channel. Combined with pulse switch control synchronized with the production line cycle, the performance gradient heating of different areas of automotive components is finally achieved, taking into account heating accuracy, production efficiency, and energy consumption optimization.

[0039] like Figure 7 It can be seen that the required heating temperature can be achieved by controlling the power of the heat source and the start-stop time during the heating process. Figure 7The central shape represents a specific board material. The upper section requires a heating temperature of 950℃, while the lower section requires 750℃. A red line divides the area into zones. Therefore, the board material needs to be heated in a gradient manner. This patent offers two heating methods. The left-hand diagram shows the controlled power variation for gradient heating. To achieve different gradients in the two areas, the power density of the upper lamp tube is controlled at 100W / cm², and the power density of the lower lamp tube is controlled at 75W / cm². After 15 seconds of operation, both zones reach the required heating temperature, thus achieving zoned heating. The right-hand diagram shows the controlled start-stop time to achieve the desired temperature. Both zones use a lamp tube power density of 100W / cm². The upper zone is heated first, and after 3 seconds, the lower lamp tube is turned on for heating. The desired zoned heating is completed after 15 seconds of operation.

[0040] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0041] Example: Heating of ultra-high strength steel plate for automotive door ring components

[0042] First, information is acquired. The automotive door ring blank to be heated is fed to the preparatory station in front of the heating furnace by the feeding mechanism. Here, a vision recognition system is deployed. The 2D vision module uses a high-resolution industrial area array camera equipped with a telecentric lens to eliminate perspective distortion, and is combined with an adjustable brightness coaxial light source and a ring light source to suppress reflections on the steel plate surface and ensure clear imaging. The 3D contour reconstruction module uses a 3D scanner based on the principle of laser triangulation. A line laser beam is projected onto the surface of the sheet material, and the deformation of the laser line is captured by an industrial camera on one side. The three-dimensional point cloud data of the sheet material surface is obtained through calculation. This system is particularly suitable for detecting spatial posture defects such as warping and wrinkles in sheet materials.

[0043] All vision hardware is installed in a high-temperature resistant and dustproof protective cover, and high-precision system calibration is completed in advance to unify the camera coordinate system, laser coordinate system and the world coordinate system of the heating furnace and conveyor line.

[0044] In the software processing flow, the 2D image undergoes filtering and denoising, contrast enhancement, threshold segmentation, and morphological processing to extract the precise outer contour and inner hole contour of the plate. The 3D point cloud data, after filtering and denoising, registration, and stitching, reconstructs the 3D surface model of the plate and calculates its spatial attitude (including translation and rotation) relative to the heating furnace plane. By fusing the 2D contour information with the 3D attitude information, the precise spatial position and shape representation of the plate within the heating furnace are obtained.

[0045] Simultaneously, based on the design requirements of the door ring component (e.g., high strength required for area A, high toughness required for area B), the sheet metal is divided into multiple performance zones for graded heating, and a target heating temperature is set for each zone (e.g., ...). Figure 2 As shown, the yellow arrow in area A represents areas requiring heating to 920°C, and the blue arrow in area B represents areas requiring heating to 850°C. This zone information and target temperature information are used as process parameters and are pre-inputted or sent from the host computer to the intelligent control system of the heating furnace.

[0046] Next, parameter settings and heat source preparation are performed. The intelligent control system receives the board outline, posture information, and target temperature of each zone from the vision system. Based on the precise outline of the board, it calculates which areas are within the board's projection range from the heat source. Figure 4 and Figure 5 As shown, only the corresponding heat sources are turned on, while all heat sources outside the plate area are turned off, thus achieving true conformal heating. This avoids energy waste on ineffectively heating air and the furnace body. The system logically groups the activated heat sources according to pre-defined performance zones, with each group responsible for heating a specific performance zone. For zones with high target temperatures (e.g., zone A), a higher power percentage (e.g., 90%) is set for their corresponding heat source groups; for zones with low target temperatures (e.g., zone B), a lower power percentage (e.g., 70%) is set. Power is steplessly adjusted via PWM (Pulse Width Modulation) signals. Currently, the heat source mainly uses short-wave infrared heat sources with a wavelength range of 0.76–1.4μm; the controllable linear power density is within 200W / cm²; and the response time is within 100ms.

[0047] When arranging the lamps, the filament distance between adjacent lamps is 20mm in both the length and width directions. This ensures good uniformity of heating the underlying material, with the error controlled within 10℃. For example... Figure 8-11 As shown, the temperature error in the longitudinal arrangement is generally kept within 10℃, indicating good temperature uniformity. In the transverse arrangement, the central area shows relatively uniform temperature, with the temperature difference in the central part controlled within 10℃, minimizing its impact on the board. Furthermore, the area with a temperature difference controlled within 5℃ accounts for approximately 30%, demonstrating good uniformity. Specifically, Figure 8 as well as Figure 10 Both simulations depict the heating of a small number of lamps arranged horizontally and vertically, with the central temperature change observable at the thicker line. Figure 9 and Figure 11 Temperature change line, from Figure 8 and Figure 9It can be observed that when the lamps are arranged longitudinally, the temperature in the middle section gradually and steadily increases over time, stabilizing around 70 seconds. Following this stabilization period, it can be observed that the temperature change in the middle section is relatively small along the length, with the temperature difference between 150mm and 450mm controlled within 10℃, indicating good uniformity. Furthermore, due to... Figure 10 and Figure 11 It can be observed that when the lamps are arranged horizontally, the temperature rises in the middle and stabilizes after about 70 seconds, and the temperature uniformity is improved to a certain extent. The temperature difference between 150mm and 400mm is generally within 10℃. It can also be found that the temperature difference at the very center can be controlled within 5℃, and the range is about 30%, indicating good uniformity.

[0048] If calculations show that power adjustment alone cannot bring different zones to their target temperatures simultaneously, timing control is introduced. For example, the heat source array above the zone requiring higher temperatures can be turned on earlier, or intermittent heating can be used. A plate lifting device can be integrated into the heating furnace. By adjusting the distance between the plate and the heat source array, the radiant flux density can be changed, serving as an auxiliary means of fine-tuning the temperature. Dynamic distance adjustment can be performed for areas requiring more uniform heating or special gradient shapes. In this embodiment, the heat dissipation source is a short-wave infrared lamp with a wavelength range of 0.76-1.5μm, which has advantages such as strong penetration, controllable heating depth, and extremely fast response speed (millisecond level). Figure 1 As shown, they are densely arranged in a matrix at the top and / or bottom of the heating furnace.

[0049] like Figure 6 As shown, after the parameters are set, the conveying mechanism uniformly transports the initial material sheet into the heating chamber of the furnace. At this time, each heat source group begins to work according to the preset power and timing. Because the intelligent heating system of the heat source is discrete and independently controlled, each zone begins to heat up under the irradiation of its dedicated heat source group. Thanks to the efficient radiative heat transfer of short-wave infrared, the material heats up rapidly. The entire heating process is carried out under the monitoring of the intelligent control system, which may make minor power adjustments based on feedback from a few fixed temperature measuring points inside the furnace.

[0050] After heating, the sheet material is rapidly conveyed out of the heating furnace for high-temperature testing and enters the intelligent temperature measurement zone. A high-precision infrared thermal imager is suspended above the temperature measurement zone to perform infrared temperature detection, instantly capturing an image (thermal image) of the temperature distribution across the entire sheet surface. Its measurement range covers room temperature to over 1000°C, and its accuracy is periodically calibrated using a standard blackbody source. Array-type contact thermocouples are arranged at specific locations on the temperature measurement platform (corresponding to the center and boundaries of key zones in a typical sheet material). As the sheet material is conveyed into position, these thermocouples, propelled by a pneumatic device, contact the lower surface of the sheet material for targeted temperature measurement.

[0051] The temperature measurement area has a semi-enclosed structure to reduce interference from ambient airflow and stray light on infrared temperature measurement.

[0052] The infrared thermal image is processed by software, and through a pre-calibrated coordinate mapping relationship, the pixel temperature values ​​in the image are mapped to the physical location of the board, thereby extracting the temperature field data of each performance zone, including average temperature, maximum / minimum temperature, temperature distribution standard deviation, etc.

[0053] The fixed-point temperature data measured by thermocouples is used to calibrate and verify infrared temperature measurement data at the same location, correcting errors caused by changes in the emissivity of the plate surface or environmental interference, and ensuring that the final obtained actual temperature information is accurate and reliable.

[0054] The control system compares the actual temperature information of each zone obtained in step four with the target heating temperature information input in step one. It checks whether the average temperature of each zone falls within the allowable error band of the target temperature (e.g., ±10°C). It calculates the standard deviation of the temperature at all temperature measurement points within each zone. If the standard deviation exceeds a set threshold (e.g., ±5°C), the zone is considered to have uneven heating, and the power distribution of the corresponding heat source group needs to be optimized. Temperature data is extracted from a certain width band near the boundary of adjacent zones, and its temperature change curve is analyzed to calculate the actual gradient value. This is compared with the designed theoretical gradient curve to determine whether the shape and steepness of the gradient transition zone meet the requirements.

[0055] The comparison results were quantified into specific temperature deviation information, including the average temperature deviation of each zone, uniformity deviation index, gradient fit index, etc.

[0056] Finally, parameter correction and optimization are performed. The system doesn't simply adjust the next heating cycle proportionally based on the current deviation; instead, it establishes a continuously learning optimization engine. The complete case package for this heating cycle is stored in the process database. This package includes: the profile / orientation characteristics of the sheet material (digital representation), all preset heating parameters (power matrix, timing table, etc.), measured temperature field data for each zone, and calculated temperature deviation information. For the next identical sheet material, the system first attempts to make preliminary corrections to the heating parameters based on the current deviation, using classic PID or fuzzy control rules. For example, if the average temperature of a certain zone is too low, the power of its heat source group will be increased by 3% in the next cycle. The system utilizes the historical database to run machine learning algorithms (e.g., neural network-based regression models or reinforcement learning Q-learning algorithms). The goal of the algorithm is to learn a function F such that: optimal heating parameters ≈ F(sheet material characteristics, target temperature field, environmental conditions).

[0057] When encountering a new type of sheet material (such as a crash beam of a different shape), the system first searches the database for historical cases with the most similar features. Then, using the learned model F, it predicts the recommended heating parameters for this new sheet material as initial settings. This significantly reduces the trial-and-error time and material waste required for traditional process development. Production data from each sheet material serves as a new training sample to update and optimize model F, making the system's control strategy increasingly accurate and efficient over time.

[0058] To fully verify the effectiveness of this method, after the heating step, the austenitized sheet metal was rapidly transferred to a water-cooled mold within seconds for hot stamping and pressure quenching. After the formed component was removed from the mold, standard tensile or microhardness test specimens were cut from different performance zones (e.g., zone A, zone B) and their transition zones. Mechanical property testing verified that: Whether each zone has achieved the designed strength (hardness) and toughness targets.

[0059] Whether the intensity is uniform within the partition, such as... Figure 3 The strength uniformity is shown in the diagram. This sheet material is divided into two regions: 0mm-40mm and 40mm-90mm. The two regions require different performance characteristics, necessitating performance gradient zoning heating to achieve this. After completing the performance gradient zoning heating, the Vickers hardness of the entire sheet material is tested for verification. It can be seen that the performance uniformity is good across all regions, and the gradient transition at the junction between the two regions is excellent, perfectly achieving the required performance gradient.

[0060] Whether the intensity gradient between partitions is gentle and controllable.

[0061] These mechanical property results can, in turn, serve as the basis for evaluating and further optimizing the heating temperature field target, forming a large closed-loop quality control system of "heating-forming-performance".

[0062] The experimental procedure followed the relationship between heating temperature and final strength:

[0063] <850℃: Insufficient austenitization, insufficient final strength, tensile strength < 1200MPa.

[0064] 880–900℃: Complete austenitization and fine grains, resulting in the best overall performance, with tensile strength reaching 1500–1570 MPa.

[0065] 900–930℃: Austenite grains grow slightly, strength is stable but slightly reduced, tensile strength is 1400–1500MPa.

[0066] >950℃: Austenite grains coarsen, strength and toughness decrease, tensile strength < 1400MPa, toughness is significantly reduced and cracking is easy.

[0067] This patent is based on the relationship between heating temperature and final strength, and achieves performance gradient heating by separating the heat source from the heat source, so that each zone generates a different temperature after heating, resulting in different performance strengths after cooling.

[0068] During strength testing, sampling locations should be selected at the center of each zone and adjacent areas between two different zones. The sample shape should be dumbbell-shaped. Then, an axial tensile load is applied to the sample at room temperature until fracture, and the force-displacement curve is recorded. From the curve, two strengths of the plate can be determined: tensile strength (Rm) (the maximum load before fracture divided by the original cross-sectional area of ​​the sample) and yield strength (Rp0.2) (for high-strength steel without a significant yield point, the stress at which 0.2% plastic deformation occurs). After obtaining the data, error correction is performed to obtain the performance strength data of each zone of the plate after performance-zone heating.

[0069] In other implementations, the heat source is not limited to short-wave infrared lamps. For example, it could be an array of resistive miniature heating rods, an array of focusing halogen lamps, or even an array of miniature flame nozzles with independently controllable gas flow. Furthermore, in addition to power and timing, for some heat sources, their radiation angle or spectral characteristics can be adjusted to achieve more complex temperature field control.

[0070] It should be noted that the present invention is not only applicable to hot stamping of automotive steel sheets, but can also be widely used in aerospace, home appliance and other industries where complex partitioned heat treatment or forming of metal or non-metal sheets is required.

[0071] The above embodiments are not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the technical solution of the present invention are also within the protection scope of the present invention.

Claims

1. A flexible intelligent heating method based on heat source discretization, characterized in that, Includes the following steps: S1. Obtain the contour information, posture information, and target heating temperature information of each performance zone of the plate to be heated; the acquisition of the contour information and posture information of the plate to be heated is achieved by a vision recognition system deployed at the front end of the heating furnace, specifically including: acquiring the planar contour information of the plate through 2D vision imaging; and acquiring the spatial posture and surface information of the plate through 3D contour reconstruction. S2. Based on the contour information and posture information, control the opening of the heat dissipation source at the corresponding position in the heating furnace, and set the heating parameters for the corresponding heat dissipation source based on the target heating temperature information. S3. The plate to be heated is transported into the heating furnace and heated in sections by the heat source set in step S2. S4. Temperature detection is performed on the heated board to obtain the actual temperature information of each performance zone; S5. Compare the actual temperature information with the target heating temperature information to obtain temperature deviation information; S6. Based on the temperature deviation information, the heating parameters are corrected and optimized; In step S2, the heat dissipation source is a short-wave infrared lamp; the heating parameters include at least one of the following: heating power, on / off sequence, and distance between the heat source and the surface of the plate. In step S4, the temperature detection uses a combination of an infrared thermal imager and a contact thermocouple; the infrared thermal imager is used to acquire the temperature distribution across the entire plate, and the contact thermocouple is used for point calibration at key locations. In step S5, the comparison includes: Calculate the actual temperature uniformity within each performance zone and compare it with the target uniformity range; Calculate the actual temperature gradient value at the boundary of adjacent performance zones and compare it with the target temperature gradient value; In step S6, the heating parameters are corrected and optimized based on the temperature deviation information, specifically as follows: Store the temperature deviation information and corresponding heating parameters of this heating process in the system database; Based on historical databases, a mapping model based on board information, heating parameters, and heating effect is established through intelligent self-learning algorithms to optimize the heating parameters of subsequent similar boards.

2. The flexible intelligent heating method based on heat source discretization according to claim 1, characterized in that, Step S3 also includes: The heated sheet metal is quickly transferred into a mold for hot stamping and pressure holding quenching. Subsequently, strength tests were conducted on different performance zones of the formed component to verify whether the heating effect met the final performance requirements.

3. The flexible intelligent heating method based on heat source discretization according to claim 1, characterized in that, The heat sources are arranged in a matrix within the heating furnace, with a center-to-center distance of 20mm between adjacent heat sources, and each heat source is independently controllable, thereby achieving precise micro-area control of the temperature field on the surface of the plate.

4. The flexible intelligent heating method based on heat source discretization according to claim 1, characterized in that, The intelligent self-learning algorithm specifically includes the following steps: Temperature deviation and deviation change rate are input into fuzzy rule base, and optimized PID parameters are obtained through fuzzy inference; Based on real-time temperature feedback and PID parameter calculation, the power regulation amount is dynamically controlled by PWM signal to regulate the power output of each heat dissipation source. When the temperature deviation exceeds 5%, a power compensation of 10%-20% is added, and fine-tuning is performed when the temperature approaches the target temperature to avoid overshoot.

5. The flexible intelligent heating method based on heat source discretization according to claim 1, characterized in that, It also includes a process database and a self-learning optimization module, used for: Store the characteristics of the plate material, heating parameters, measured temperature field and deviation information for each heating cycle; A machine learning model is trained based on historical data to establish a mapping relationship between board characteristics, heating parameters, and heating effect; For boards with new shapes or materials, the system automatically recommends initial heating parameters and continuously optimizes the control strategy during the production process.

Images