A glass mold cooling rate closed-loop control method and system

Abstract

The application provides a glass mold cooling rate closed-loop control method and system. In the application, infrared thermal image data of the surface of the glass mold is collected, and internal temperature conduction data is obtained. Then, combined with the cooling rate graph and the temperature data, the heat conduction compensation area is divided and the heat compensation coefficient is determined according to the three-dimensional geometric characteristics of the mold. Then, all data are fused to perform space-time three-dimensional processing to generate a dynamic thermal field model, identify and quantify hot spot areas of uneven cooling. At the same time, the phase transition critical point of the glass material is detected in real time to generate a phase transition correction factor. Finally, the cooling rate deviation trend is predicted by combining the hot spot quantification index and the phase transition factor, a predicted deviation value is generated, and the corresponding fan speed is adjusted in stages, thereby forming a regional closed-loop control. The application realizes accurate regional closed-loop control of the glass mold cooling rate, can effectively eliminate the problem of uneven hot spots and optimize the cooling process.

Description

Technical Field

[0001] This application relates to the field of glass manufacturing automation and thermodynamic process control technology, and in particular to a closed-loop control method and system for the cooling rate of glass molds. Background Technology

[0002] In high-end glass product pressing and molding processes, the uniformity and precise control of the mold cooling rate are core requirements for ensuring stress-free distortion and surface smoothness of the product. Because glass molds typically have complex curved surfaces and non-uniform wall thicknesses, their heat capacity and heat dissipation efficiency vary significantly across different regions. Therefore, precise control in two dimensions is crucial: in the spatial dimension, dynamically compensating for the hysteresis effect of heat conduction in areas of abrupt curvature changes and thickness transitions; and in the temporal dimension, precisely matching the critical cooling rate near the phase transition point of the glass material. Any uneven cooling or timing deviations will lead to excessive residual stress within the product or surface ripple defects.

[0003] Current mainstream solutions employ a control method that combines infrared thermal imaging with offline finite element simulation. This approach involves deploying an infrared thermal imager on the mold surface to collect two-dimensional temperature distribution data, combining this data with a pre-built finite element model of the mold to perform heat conduction simulation, and then pre-setting the fan power parameters for each cooling zone based on the simulation results to form an open-loop control strategy based on a static thermodynamic model.

[0004] However, this approach has significant limitations: First, the offline finite element model cannot respond in real time to dynamic disturbances during the production process, such as fluctuations in ambient temperature and changes in the initial temperature of the molten glass, leading to a mismatch between preset parameters and actual requirements. Second, relying solely on surface temperature data ignores changes in the three-dimensional heat flow state inside the mold, especially the delay in heat conduction in the thickness direction, resulting in misjudgments of the true thermal state in the curvature / wall thickness variation area. Third, the identification of the critical point of phase transformation of the glass material relies on offline calibration data and lacks online feedback on the phase transformation state, resulting in a lag in the speed regulation response of the fan, which can easily cause local overcooling or insufficient cooling rate defects in the phase transformation sensitive temperature zone. Summary of the Invention

[0005] This application provides a closed-loop control method and system for the cooling rate of glass molds, which solves the common problems of uneven cooling caused by dynamic thermal field response mismatch, lack of three-dimensional heat conduction state monitoring and phase change critical control hysteresis in the prior art.

[0006] In a first aspect, this application provides a closed-loop control method for the cooling rate of a glass mold, comprising:

[0007] Infrared thermal image data of the glass mold surface is collected, and the infrared thermal image data is analyzed by pixel-level temperature change. The real-time cooling rate map is calculated by combining the time change. At the same time, point temperature measurement arrays are deployed in the curvature change area and thickness transition area of ​​the glass mold surface to collect temperature conduction data inside the glass mold simultaneously.

[0008] Based on the cooling rate diagram and the temperature conduction data, the heat conduction compensation region is divided according to the three-dimensional geometric features of the glass mold, and the heat compensation coefficient corresponding to the compensation region is determined.

[0009] By integrating the cooling rate map, the temperature conduction data, and the thermal compensation coefficient, a dynamic thermal field model is generated through spatial and temporal three-dimensional fusion processing to identify and quantify hot spots with uneven cooling.

[0010] The phase transition critical point of the glass material is detected in real time in the dynamic thermal field model, and a phase transition correction factor is generated when the temperature reaches the preset phase transition threshold.

[0011] By combining the hot spot region quantification index of the dynamic thermal field model with the phase change correction factor, the cooling rate offset trend is predicted and a prediction deviation value is generated. Based on the prediction deviation value, the fan speed parameters of the corresponding cooling region are adjusted in stages to form a regional closed-loop control of the cooling rate of the glass mold.

[0012] Optionally, by combining the hot spot region quantification index of the dynamic thermal field model with the phase change correction factor, the cooling rate shift trend is predicted and a prediction deviation value is generated, including:

[0013] Read the quantitative index value of the hot spot area in the dynamic thermal field model. The quantitative index value represents the degree of difference between the temperature change rate at the current location and the preset uniform cooling standard.

[0014] The phase transition correction factor is obtained by acquiring the proximity coefficient of the current location point temperature to the phase transition temperature threshold in the same hot spot area, and the expected temperature change is calculated by combining the current temperature change rate of the hot spot area with the phase transition correction factor within a fixed time interval.

[0015] Calculate the difference between the expected temperature change and the preset standard temperature change, analyze the magnitude component of the difference with the fixed time interval length, and output a predicted deviation value representing the expected offset of the temperature change rate per unit time.

[0016] Optionally, the fan speed parameters of the corresponding cooling zones are adjusted in stages according to the predicted deviation value to form a regionalized closed-loop control of the glass mold cooling rate, including:

[0017] The glass mold cooling area is divided into multiple independent control blocks, and each control block is bound to a set of fan equipment;

[0018] Based on the prediction deviation value, a corresponding absolute size range mapping table is set, and the absolute size of the prediction deviation value is divided into low level, medium level and high level, with each level corresponding to the fixed speed increment of the wind turbine equipment.

[0019] For each hotspot area, identify the prediction deviation value and corresponding direction of the control block to which it belongs;

[0020] The adjustment level and speed increment are obtained by matching the predicted deviation value with the mapping table. When the deviation direction is positive, the speed is increased; when the deviation direction is negative, the speed is decreased. The fan speed parameter adjustment value is output to the speed regulation device of the fan equipment to realize closed-loop control.

[0021] Optionally, the phase transition critical point of the glass material is detected in real time in the dynamic thermal field model, and a phase transition correction factor is generated when the temperature reaches a preset phase transition threshold, including:

[0022] In the spatiotemporal data structure of the dynamic thermal field model, a continuous scanning operation is performed on all location points. When the current temperature value of a location point is found to be lower than the preset first temperature boundary, the location point is marked as a candidate monitoring point.

[0023] The absolute difference between the current temperature value of the candidate monitoring point and the preset phase transition temperature threshold for the solid-liquid phase transition of the glass material is calculated in real time.

[0024] A judgment operation is performed on the absolute difference of the candidate monitoring points. When the absolute difference is less than the preset second temperature boundary, the candidate monitoring point is judged as the phase transition critical point.

[0025] Based on the absolute difference between the current temperature value and the phase transition temperature threshold, a reverse numerical conversion operation is performed. When the absolute difference is equal to zero, a correction value of 1 is output, and when the absolute difference is equal to the second temperature boundary, a correction value of 0 is output. Within the range of the absolute difference, a correction coefficient is continuously calculated as a phase transition correction factor according to a linear rule.

[0026] Optionally, by fusing the cooling rate map, the temperature conduction data, and the thermal compensation coefficient, a three-dimensional spatial and temporal fusion processing is performed to generate a dynamic thermal field model, identifying and quantifying hotspot regions with uneven cooling, including:

[0027] Create a spatiotemporal data structure that includes three-dimensional spatial location (x, y, z) and continuous time points;

[0028] The temperature change rate data at the surface location in the cooling rate diagram is mapped to the surface location point in three-dimensional space, and the temperature conduction data of the point temperature measurement array is mapped to the internal location point in three-dimensional space.

[0029] For unmapped location points, a weighted average is used to assign values ​​based on neighboring location point data.

[0030] In the spatiotemporal data structure, a correction temperature value is calculated based on the temperature value of each location point and the thermal compensation coefficient of the corresponding compensation area, and the spatiotemporal data structure is corrected based on the correction temperature value.

[0031] Based on the corrected spatiotemporal data structure, the standard deviation of temperature values ​​at all locations is calculated at each time point. Regions where the standard deviation exceeds a set threshold are identified as uneven cooling regions. The difference between the standard deviation and the set threshold is used as a quantification index value to identify and quantify the hotspot regions of the uneven cooling regions.

[0032] Optionally, based on the cooling rate diagram and the temperature conduction data, a heat conduction compensation region is divided according to the three-dimensional geometric features of the glass mold, and a heat compensation coefficient corresponding to the compensation region is determined, including:

[0033] Input three-dimensional geometric model data of the glass mold containing surface shape, thickness and curvature information, and based on the three-dimensional geometric model data, identify areas where the curvature change exceeds a set threshold and areas where the thickness difference exceeds a set threshold;

[0034] The region is divided into multiple independent blocks to form a compensation region. Within each compensation region, the average temperature change rate at the surface position in the cooling rate map and the average temperature value of the temperature conduction data from the point temperature measurement array are extracted. The initial value of the thermal compensation coefficient is calculated according to a preset rule.

[0035] The initial value of the thermal compensation coefficient is adjusted according to the relative positional relationship between the compensation area and the overall three-dimensional geometric model to obtain the final thermal compensation coefficient.

[0036] Optionally, the infrared thermal image data is analyzed by pixel-level temperature changes, and a real-time cooling rate map is calculated by combining it with time changes, including:

[0037] Continuous thermal image data is acquired from an infrared imaging device on the surface of a glass mold. The thermal image data consists of multiple uniformly distributed measurement points, each containing a temperature value.

[0038] A recording sequence with a fixed time interval is set for each measurement point, and the temperature value at the previous moment and the temperature value at the current moment are stored in the recording sequence;

[0039] For each measurement point, the difference between the temperature value at the previous moment and the temperature value at the current moment is calculated, and the rate of temperature change at each measurement point is calculated by combining the difference with a fixed time interval.

[0040] Arrange the temperature change rates of all measurement points according to their positions in the original thermal image data, and output a cooling rate map.

[0041] Secondly, this application provides a closed-loop control system for the cooling rate of a glass mold, comprising:

[0042] The acquisition module is used to acquire infrared thermal image data of the glass mold surface, analyze the infrared thermal image data through pixel-level temperature changes, and calculate the real-time cooling rate map by combining it with time changes. At the same time, a point-like temperature measurement array is deployed in the curvature change area and thickness transition area of ​​the glass mold surface to synchronously acquire the temperature conduction data inside the glass mold.

[0043] The partitioning module is used to partition the heat conduction compensation region based on the cooling rate diagram and the temperature conduction data, according to the three-dimensional geometric features of the glass mold, and determine the heat compensation coefficient corresponding to the compensation region.

[0044] The generation module is used to fuse the cooling rate map, the temperature conduction data and the thermal compensation coefficient, perform spatial and temporal three-dimensional fusion processing to generate a dynamic thermal field model, and identify and quantify hot spots with uneven cooling.

[0045] The generation module is also used to detect the phase transition critical point of the glass material in real time in the dynamic thermal field model, and generate a phase transition correction factor when the temperature reaches the preset phase transition threshold.

[0046] The prediction module is used to combine the hot spot area quantification index of the dynamic thermal field model with the phase change correction factor to predict the cooling rate offset trend and generate a prediction deviation value. Based on the prediction deviation value, the fan speed parameters of the corresponding cooling area are adjusted in stages to form a regional closed-loop control of the cooling rate of the glass mold.

[0047] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are to be invoked and executed by the processing component to implement a closed-loop control method for cooling rate of a glass mold as described in the first aspect above.

[0048] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a closed-loop control method for the cooling rate of a glass mold as described in the first aspect.

[0049] This application achieves real-time surface cooling rate and internal three-dimensional heat conduction state by performing pixel-level temperature change analysis of infrared thermography on the mold surface and simultaneously acquiring point array temperature data of the curvature / thickness transition zone. Based on this data, combined with the three-dimensional geometric features of the mold, a heat compensation region is divided and a heat compensation coefficient is determined, which can correct the heat conduction distortion error caused by structural differences. Furthermore, by fusing multi-source data and performing spatiotemporal three-dimensional processing to generate a dynamic thermal field model, it can accurately identify and quantify the spatial distribution parameters of uneven cooling hotspots. By detecting the glass phase transition critical point in real time in the model and generating a phase transition correction factor, it can dynamically respond to changes in the thermal properties of the material's phase transition sensitive region. Finally, by combining the hotspot quantification index and the phase transition factor to predict the cooling rate offset trend, generating a prediction deviation value and adjusting the fan speed in stages, it can achieve regional closed-loop control of the cooling rate for spatial heterogeneity.

[0050] Furthermore, by reading the temperature change rate difference values ​​(i.e., quantitative indicators) of hot spot regions in the dynamic thermal field model and combining them with the proximity coefficient of the phase change threshold in the same region (i.e., phase change correction factors), the expected temperature change is calculated within a fixed time interval. The time derivative of the difference between the expected value and the standard value is output as the prediction deviation value, which characterizes the offset intensity of the cooling rate per unit time. By establishing a coupling calculation mechanism between the quantitative indicators and the phase change factors, the predicted temperature change trend is transformed into a control variable per unit time. This transforms the abstract thermal state difference into a deviation control signal that can directly drive the actuator (fan), significantly improving the temperature control response speed and accuracy in phase change sensitive areas.

[0051] These or other aspects of this application will become more apparent from the description of the following embodiments. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart of a closed-loop control method for cooling rate of a glass mold provided in this application is shown;

[0054] Figure 2 A schematic diagram of a scenario for a closed-loop control method for the cooling rate of a glass mold provided in this application is shown.

[0055] Figure 3 This paper shows a schematic diagram of the structure of a closed-loop control system for the cooling rate of a glass mold provided in this application;

[0056] Figure 4 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0057] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0058] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0059] In the production of precision glass products, the uniformity of mold cooling directly affects the product yield. Current mainstream solutions suffer from three major drawbacks: First, they rely on pre-set, fixed mathematical models for cooling regulation, which cannot dynamically adjust strategies when faced with changes in ambient temperature or material fluctuations—like using an old map to navigate a new route. Second, infrared monitoring can only acquire the surface temperature of the mold, failing to detect the internal heat dissipation of complex curved surfaces and areas with varying thicknesses, creating monitoring blind spots. Third, there is a lag in response to the critical solidification point where glass transitions from liquid to solid, leading to excessive or insufficient heat dissipation in localized areas, causing product deformation or microcracks. These defects stem from the disconnect between static control models and the dynamic production environment, a lack of three-dimensional understanding of heat dissipation, and sluggish response at critical points.

[0060] To address the aforementioned issues, this application proposes an intelligent mold cooling dynamic control system. This system forms a three-dimensional monitoring network by deploying an infrared thermal imager on the mold surface (constructing an overall temperature field) and embedding miniature temperature sensors in curved areas and at thickness boundaries prone to uneven heat dissipation (acquiring actual internal heat dissipation data). The system dynamically integrates real-time multi-dimensional data with the mold's three-dimensional structure, automatically identifying areas of abnormal heat dissipation and quantifying the degree of risk. Particularly during the critical stage nearing glass solidification, the system proactively predicts changes in heat dissipation trends and adjusts the cooling fan speeds in each area according to the degree of deviation. This method replaces static presets with dynamic models, overcomes monitoring blind spots with three-dimensional perspective, and eliminates response lag with predictive control. It completely solves the hidden dangers of uneven heat dissipation caused by "model rigidity, lack of perception, and passive control" in traditional methods, significantly improving product yield and equipment response accuracy.

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] Figure 1 A flowchart of a closed-loop control method for the cooling rate of a glass mold is provided in this application embodiment, as follows: Figure 1 As shown, the method includes:

[0063] 101. Collect infrared thermal image data of the glass mold surface, analyze the infrared thermal image data through pixel-level temperature change, and calculate the real-time cooling rate map by combining it with time change. At the same time, deploy point temperature measurement arrays in the curvature change area and thickness transition area of ​​the glass mold surface to synchronously collect temperature conduction data inside the glass mold.

[0064] Optionally, step 101 may specifically include the following steps:

[0065] 1011. Continuous thermal image data is acquired from an infrared imaging device on the surface of a glass mold, wherein the thermal image data consists of multiple uniformly distributed measurement points, each measurement point containing a temperature value;

[0066] 1012. Set a recording sequence with a fixed time interval for each measurement point, and save the temperature value at the previous moment and the temperature value at the current moment in the recording sequence;

[0067] 1013. For each measurement point, calculate the difference between the temperature value at the previous moment and the temperature value at the current moment, and combine the difference with a fixed time interval to calculate the rate of temperature change at each measurement point.

[0068] 1014. Arrange the temperature change rates of all measurement points according to the positions of the original thermal image data, and output a cooling rate map.

[0069] In the above scheme, infrared thermal image data specifically refers to the temperature distribution image of the mold surface captured by an infrared thermal imager, with each pixel corresponding to the temperature value of a tiny area on the mold surface. The point-like temperature measurement array refers to a group of miniature temperature sensors installed in areas of abrupt changes in mold shape, such as the curved section of the bottle mouth, and in areas where thickness changes, such as the transition from the bottle body to the bottom. Temperature conduction data refers to the rate of heat transfer inside the mold collected in real time by these sensors. The cooling rate graph is a dynamic chart generated by analyzing temperature changes at continuous time points, visually displaying the rate of temperature decrease per second in different areas of the mold surface using different colored blocks.

[0070] In this embodiment, firstly, in step 1011, an infrared thermal imager is used to acquire thermal image data of the temperature distribution on the surface of the glass mold at a rate of 10 frames per second. Each frame contains several rows and columns of measurement points, and the temperature value at each measurement point is recorded as the current temperature value. For example, the temperature value recorded at measurement points B77 to B89 at time t1 = 14:00:00.000 is 560.0 degrees Celsius. Secondly, in step 1012, a time series record is established for each measurement point. A fixed time interval Δt = 0.5 seconds is set, and the same point is measured again at the next time point t2 = 14:00:00.500, recording a new temperature value of 555.0 degrees Celsius. At this time, the temperature record sequence for the corresponding point is updated to include the data at times t1 and t2. Next, step 1013 is used to calculate the rate of temperature change: For any measurement point, two adjacent temperature data points are extracted from the recording sequence, the difference between the temperature value at the previous moment and the temperature value at the current moment is calculated, and the rate of temperature change at each measurement point is calculated using a formula based on the difference and a fixed time interval.

[0071]

[0072] The parameters are defined as follows: This indicates the temperature value at the time of the previous record was 560.0 degrees Celsius. This indicates the current recorded temperature value is 555.0 degrees Celsius, and Δt is a fixed time interval of 0.5 seconds. The specific calculation process consists of two steps: first, the absolute value of the temperature change is calculated as |560.0 - 555.0| = 5.0 degrees Celsius, then divided by the time interval to obtain v = 5.0 ÷ 0.5 = 10.0 degrees Celsius per second. This value quantifies the rate of temperature decrease per second in this area. Finally, step 1014 remaps the calculation results of all measurement points according to their original spatial locations. For example, the area from point B77 to B89 is marked as having a temperature decrease of 10.0 degrees Celsius per second, and the adjacent area from point C90 to B105 is calculated to have a temperature decrease of 6.5 degrees Celsius per second. The final output is a cooling rate distribution map encoded with color levels.

[0073] In practical applications, on the lampshade mold production line of glass product factory C, the system uses an infrared thermal imager (numbered A) to photograph the mold surface, generating a temperature matrix of 1024 x 1024 measurement points. Ten thermocouple sensors are installed in area X (the curved area at the bottle mouth) and another ten in area Y (the transition area between the thick and thin sections at the bottom). Taking internal sensor Y3 as an example: at 2 PM, the temperature is recorded as 420.0 degrees Celsius; at 2 PM 00:00.5 seconds, the temperature is recorded as 416.5 degrees Celsius. The temperature change is calculated as 420.0 degrees Celsius minus 416.5 degrees Celsius, yielding 3.5 degrees Celsius. Dividing this by the time interval of 0.5 seconds gives a temperature change rate of 7.0 degrees Celsius per second. Simultaneously, infrared measurement points B77 to B89 recorded 560.0 degrees at 2 PM and 555.0 degrees at 2 PM:00:00.5. The temperature change of 560.0 minus 555.0 equals 5.0 degrees, which, when divided by the 0.5-second time interval, calculates a temperature drop of 10.0 degrees per second. The system categorizes and encodes the calculation results from over 100,000 points across the entire map, ultimately generating a dynamic cooling rate map where red represents rapid heat dissipation zones (greater than 10 degrees per second) and blue represents slow heat dissipation zones (less than 3 degrees per second).

[0074] This solution achieves, for the first time, joint monitoring of the three-dimensional heat dissipation status of a mold through simultaneous acquisition of surface infrared scanning and an internal sensor network. A cooling rate map is generated based on continuous temperature difference calculations over time, allowing operators to directly observe the precise heat dissipation amplitude per second in each area. For complex curved surfaces and thickness-to-thickness interfaces of the mold, direct measurement of internal temperature data corrects errors caused by angular obstruction in surface infrared monitoring, eliminates blind spots in traditional methods, and provides complete and reliable data support for subsequent precise adjustment of cooling intensity.

[0075] 102. Based on the cooling rate diagram and the temperature conduction data, divide the heat conduction compensation region according to the three-dimensional geometric features of the glass mold, and determine the heat compensation coefficient corresponding to the compensation region;

[0076] Optionally, step 102 may specifically include the following steps:

[0077] 1021. Input three-dimensional geometric model data of the glass mold containing surface shape, thickness and curvature information, and based on the three-dimensional geometric model data, identify areas where the curvature change exceeds a set threshold and areas where the thickness difference exceeds a set threshold;

[0078] 1022. Divide the region into multiple independent blocks to form a compensation region. In each compensation region, extract the average temperature change rate of the surface position in the cooling rate diagram and the average temperature value of the temperature conduction data of the point temperature measurement array, and calculate the initial value of the thermal compensation coefficient according to the preset rules.

[0079] 1023. Adjust the initial value of the thermal compensation coefficient according to the relative positional relationship between the compensation area and the overall three-dimensional geometric model to obtain the final thermal compensation coefficient.

[0080] In the above scheme, the cooling rate map refers to the heat dissipation rate distribution image of each area of ​​the mold surface generated in step 101. Temperature conduction data refers to the internal heat dissipation rate measured by miniature sensors installed in key parts of the mold. Three-dimensional geometric features refer to the three-dimensional spatial structure data of the mold surface, such as its undulating shape, thickness variations, and degree of curvature. The compensation area specifically refers to local blocks where uneven heat dissipation is caused by abrupt changes in shape or thickness, such as the arc at the bottle mouth or the thickness boundary area at the bottom of the bottle. The thermal compensation coefficient refers to the heat dissipation intensity correction value calculated individually for each compensation area, used to adjust subsequent control parameters.

[0081] In this embodiment, firstly, a three-dimensional structural model file of the glass mold is input in step 1021. This model file includes data on the surface irregularities, thickness variations, and curvature of the mold. The system automatically scans the model to identify high-risk areas for heat dissipation: when a curvature angle change exceeds a set threshold, it is marked as a curvature abrupt change area, such as the area with a curvature angle of 30 to 70 degrees at the bottle neck arc; when the thickness difference between adjacent areas exceeds a set threshold, it is marked as a thickness transition area, such as the gradient area from 3 mm to 8 mm at the bottle body. Secondly, in step 1022, the marked areas are divided into several independent blocks: for example, the bottle neck arc area is divided into three compensation blocks numbered Q1 to Q3. Dual index extraction is performed on each block: the average heat dissipation rate at the surface position of the cooling rate map within the block is calculated and recorded as... The average heat dissipation rate measured by the sensors inside the block is recorded as follows: Calculate the initial value of the thermal compensation coefficient using the formula:

[0082]

[0083] in This is the standard heat dissipation rate constant. For example, the heat dissipation rate constant for block Q1. =8.5℃ / second, =7.2℃ / second, When the heat dissipation rate is 8.0℃ / second: (0.6×8.5 + 0.4×7.2) ÷ 8.0 = 1.02. Finally, in step 1023, the initial value of the heat compensation coefficient is corrected according to the relative position of the block and the core area of ​​the mold to obtain the final heat compensation coefficient: for example, if block Q1 is located at the top edge of the mold, the system determines that the heat dissipation efficiency is reduced by 8%, and the initial coefficient 1.02 is multiplied by 1.08 to obtain the final heat compensation coefficient 1.10.

[0084] In practical applications, during the lampshade mold production process at glass product factory C, the system inputs a three-dimensional geometric model containing surface shape data, thickness distribution, and curvature parameters. Based on this model data, it automatically identifies region A (located in the middle of the bottle mouth arc surface with a curvature change exceeding a preset threshold of 10 degrees / mm) and transition region B (located at the bottom edge of the bottle) with a thickness difference exceeding 5 mm. Region A is divided into compensation blocks A1-A3, and region B is divided into compensation blocks B1-B2. For region A1, the average temperature change rate data corresponding to the surface position in the cooling rate diagram is calculated to be 8.3 degrees Celsius per second. Simultaneously, the average temperature conduction data from three sensors in the point-like temperature measurement array within this block is calculated to be 7.1 degrees Celsius per second. According to preset rules, a weighting coefficient of 0.6 is assigned to the surface data and a weighting coefficient of 0.4 to the internal data to calculate the initial thermal compensation coefficient: 0.6 × 8.3 = 4.98 + 0.4 × 7.1 = 2.84, totaling 7.82. Dividing this by the standard heat dissipation rate of 8.0 degrees Celsius per second yields an initial coefficient of 0.9775. Based on the relative position of block A1 to the core area of ​​the mold, it was detected that it is in a high thermal conduction delay zone, which requires an increase of 3% according to the rules. The initial coefficient of 0.9775 was multiplied by the correction coefficient of 1.03, resulting in a thermal compensation coefficient of 1.006 for this compensation area. The same process was used to process block A2, with an initial coefficient of 0.99, which was corrected to 1.02 after position correction. The initial coefficient of block B1 was 1.05, which was corrected to 1.08. Thus, the thermal compensation coefficients for all compensation blocks were determined.

[0085] This solution, by integrating three-dimensional structural features and real-time temperature data, achieves for the first time intelligent identification and quantitative compensation of areas with uneven heat dissipation. Based on geometric features, high-interest areas are automatically delineated, and compensation coefficients are generated by combining surface and internal heat dissipation rates, effectively eliminating misjudgments of heat dissipation caused by complex mold shapes. Furthermore, by correcting with positional correlation coefficients, the impact of heat conduction delays in corner areas is further reduced. Finally, precise compensation parameters for each key block are output, providing a spatial difference adjustment benchmark for subsequent closed-loop control.

[0086] 103. By integrating the cooling rate diagram, the temperature conduction data, and the thermal compensation coefficient, a three-dimensional spatial and temporal fusion processing is performed to generate a dynamic thermal field model, and hot spots with uneven cooling are identified and quantified.

[0087] Optionally, step 103 may specifically include the following steps:

[0088] 1031. Create a spatiotemporal data structure that includes three-dimensional spatial location (x, y, z) and continuous time points;

[0089] 1032. Map the temperature change rate data of the surface position in the cooling rate diagram to the surface position point in three-dimensional space, and map the temperature conduction data of the point temperature measurement array to the internal position point in three-dimensional space;

[0090] 1033. For unmapped location points, assign a weighted average value using data from neighboring location points;

[0091] 1034. In the spatiotemporal data structure, a correction temperature value is calculated based on the temperature value of each location point and the thermal compensation coefficient of the corresponding compensation area, and the spatiotemporal data structure is corrected based on the correction temperature value.

[0092] 1035. Based on the corrected spatiotemporal data structure, calculate the standard deviation of temperature values ​​at all locations at each time point. When the standard deviation exceeds a set threshold, the region is identified as an uneven cooling region. The difference between the standard deviation and the set threshold is used as a quantitative index value to identify and quantify the hot spot region of the uneven cooling region.

[0093] In the above scheme, the cooling rate map refers to the heat dissipation rate distribution map on the mold surface, the temperature conduction data refers to the heat dissipation rate data of key internal points, and the thermal compensation coefficient refers to the correction weight value calculated for poorly heat-dissipated areas. Spatial and temporal three-dimensional fusion processing refers to simultaneously labeling the mold temperature data at three-dimensional spatial locations and continuous time points. The dynamic thermal field model refers to a three-dimensional animation of the mold temperature field changing over time, created by a computer. Hotspot areas specifically refer to dangerous blocks where the heat dissipation rate is much slower than the surrounding area, leading to excessively high temperatures; the quantitative index value indicates the severity of uneven heat dissipation in this area.

[0094] In this embodiment, firstly, step 1031 creates a container containing three-dimensional coordinates and time: the length, width, and height of the glass mold space are represented by x, y, and z respectively, and a recording point is generated every 0.5 seconds along the time axis. For example, a data structure with 20 time layers numbered t1 to t20 is created. Secondly, step 1032 maps the temperature change rate data of the cooling rate map surface position to spatial points on the glass mold surface: for example, surface point x50y60z0 is marked with a heat dissipation rate of 9℃ / second at time t1; simultaneously, internal sensor data is mapped to internal spatial points: for example, point x55y65z-20 is marked with a heat dissipation rate of 7.2℃ / second at time t1. Next, step 1033 assigns values ​​to points without data: for each unmapped point, the nearest neighbor data within a sphere with a radius of 3 mm is collected and assigned according to the distance d using an inverse weighting:

[0095]

[0096] For example, at location x52y62z-10, which has no sensor, the data from the four nearest points are weighted and averaged, with closer points having higher weights. Taking this location as an example: the weights of the four nearest points are 0.3, 0.3, 0.2, and 0.2, and their heat dissipation rates are 8.0, 8.2, 7.9, and 8.1℃ / second, respectively. Therefore, (0.3×8.0+0.3×8.2+0.2×7.9+0.2×8.1) = 8.06℃ / second. Subsequently, in step 1034, within the spatiotemporal data structure, based on the temperature value of each location point and the corresponding thermal compensation coefficient of the compensation area, thermal compensation coefficient correction is applied: calculations are performed for each point within the compensation area. For example, at point A1 in the bottle neck area (x50y60z0), the thermal compensation coefficient is 1.006. The original temperature of 560℃ at time t1 is corrected to 560 × 1.006 = 563.36℃. Finally, hotspot areas are identified through step 1035: at each time point, the standard deviation σ of the dispersion of the overall mold temperature values ​​is calculated. At time point t_k, the corrected temperature values ​​at N locations within the compensation area S are calculated. calculate:

[0097]

[0098] in The average temperature of the region. A hot zone warning is triggered when σ > 2℃, and the quantification index value Q = σ - 2. When the standard deviation of the region continuously exceeds the threshold of 2℃, it is marked as a region of uneven cooling. The difference between the standard deviation and the set threshold is used as the quantification index value to identify and quantify the hot spots of the uneven cooling region. For example, at time t1, the temperature dispersion of block B1 at the bottom of the bottle reaches 3.5℃, exceeding the threshold of 1.5℃, and the quantification index value is 1.5.

[0099] In practical applications, in the lampshade mold scenario of glass factory C: the system creates a three-dimensional spatial grid containing 100×100×20 location points, and records 10 time layers every 0.5 seconds on the time axis. Step 1032 maps the infrared data: the cooling rate data of the bottle mouth point x30y40z0 at time t1 is 8.3℃ / second; the internal sensor point x32y42z-15 is mapped to temperature conduction data of 7.1℃ / second. Step 1033: Process the data point x31y41z-5 (no data): Take the four neighboring points x30y40z0 (weight 0.4, data 8.3℃), x32y42z-15 (weight 0.3, data 7.1℃), x30y42z0 (weight 0.2, data 8.0℃), and x32y40z-10 (weight 0.1, data 7.3℃). Calculate 0.4×8.3+0.3×7.1+0.2×8.0+0.1×7.3=7.84℃ / second. Step 1034: Apply compensation coefficients: For point A1 at the bottle mouth (x30y40z0), the thermal compensation coefficient is 1.006, correcting the original temperature of 542℃ to 542×1.006=545.25℃; for point B1 at the bottle bottom (x80y90z0), the coefficient is 1.08, correcting the original temperature of 588℃ to 588×1.08=635.04℃. Step 1035: Calculate the standard deviation of the entire mold temperature at time t1: The temperature range of 10 points within block B1 is 635.04±3.5℃, with a standard deviation of 2.8℃. Exceeding the set threshold of 2.0℃ triggers an alarm, with a quantified index value of 0.8; simultaneously, the standard deviation of block A1 is 1.5℃, not exceeding the threshold. The system output block B1 is a hotspot area and requires priority control.

[0100] This solution achieves three-dimensional dynamic visualization of the mold temperature field for the first time, accurately reproducing the heat distribution evolution process through spatiotemporal data fusion. Geometric errors are eliminated using a thermal compensation coefficient, making the temperature field data closer to the actual heat dissipation state. A statistical standard deviation-based hot zone identification mechanism overcomes the limitations of traditional experience-based judgment, providing a scientific basis for subsequent precise control with quantitative indicators, and effectively capturing hidden heat dissipation defect areas invisible to the naked eye.

[0101] 104. The phase transition critical point of the glass material is detected in real time in the dynamic thermal field model, and a phase transition correction factor is generated when the temperature reaches the preset phase transition threshold.

[0102] Optionally, step 104 may specifically include the following steps:

[0103] 1041. In the spatiotemporal data structure of the dynamic thermal field model, a continuous scanning operation is performed on all location points. When the current temperature value of a location point is found to be lower than the preset first temperature boundary, the location point is marked as a candidate monitoring point.

[0104] 1042. Calculate in real time the absolute difference between the current temperature value of the candidate monitoring point and the preset phase transition temperature threshold for the solid-liquid phase transition of the glass material;

[0105] 1043. Perform a judgment operation on the absolute difference of the candidate monitoring points. When the absolute difference is less than the preset second temperature boundary, the candidate monitoring point is judged as a phase transition critical point.

[0106] 1044. Based on the absolute difference between the current temperature value and the phase transition temperature threshold, perform a reverse numerical conversion operation. When the absolute difference is equal to zero, output a correction value of 1. When the absolute difference is equal to the second temperature boundary, output a correction value of 0. Within the range of the absolute difference, continuously calculate the correction coefficient as the phase transition correction factor according to a linear rule.

[0107] In the above scheme, the dynamic thermal field model refers to the comprehensive data volume generated in step 103, which includes the three-dimensional spatial temperature field and its temporal changes. The phase transition critical point refers to the key temperature position where the glass material changes from a liquid to a solid state. The phase transition temperature threshold specifically refers to the critical temperature value at which the glass undergoes a solid-liquid transition. The phase transition correction factor refers to the control weight coefficient calculated for the phase transition critical point, used to eliminate the temperature control hysteresis problem in the phase transition process.

[0108] In this embodiment, firstly, step 1041 involves continuously scanning the temperature data of all locations in the spatiotemporal data structure of the dynamic thermal field model. When a location's current temperature is detected to be lower than a preset first temperature boundary value T_b1, that point is marked as a candidate monitoring point. For example, the bottom of the bottle, with coordinates x80y90z0, is marked at time t3 when its temperature is 577 degrees Celsius, which is lower than the set value of 580 degrees Celsius. Secondly, step 1042 calculates the current temperature value of the candidate monitoring point in real time. Phase transition temperature threshold of glass materials The absolute difference d is calculated using the following formula: For example, point =577℃ When the temperature is 580℃, d = |577-580| = 3℃. Next, step 1043 determines the difference d: when d is less than the preset second temperature boundary value T_b2, the corresponding point is determined as the phase transition critical point. For example, T_b2 = 5℃ and d = 3℃ meets the condition and is determined. Finally, step 1044 performs a reverse numerical transformation to generate the phase transition correction factor k: the mapping rule between the difference d and the correction coefficient k is established as follows:

[0109] When d=0℃, the output k=1.0.

[0110] When d= When k=0.0,

[0111] When 0 <d< The coefficients are calculated according to the linear relationship:

[0112]

[0113] The parameter is defined as follows: d represents the absolute difference in degrees Celsius. The second temperature boundary value is in degrees Celsius. Taking d=3℃ as an example, substituting into the formula: 1.0 minus the fraction 1.0 in parentheses, divided by 5 and multiplied by 3 equals 1.0 minus 0.2 in parentheses, multiplied by 3 equals 1.0 minus 0.6, resulting in k=0.4. This is the phase transition correction factor output for this location.

[0114] In practical applications, during the lampshade mold production process at glass factory C, the system scans the dynamic thermal field model at time t3. The coordinates of the bottle area location (x50y60z0) show a current temperature of 577 degrees Celsius, which is lower than the preset first temperature boundary value of 580 degrees Celsius. This triggers step 1041, marking it as a candidate monitoring point. Step 1042 is then executed to calculate the absolute difference d between this point and the glass phase transition temperature threshold of 580 degrees Celsius: according to the formula: The difference result is 3 degrees Celsius. Step 1043 determines that since the preset second temperature boundary value of 5 degrees Celsius is greater than the current difference of 3 degrees Celsius, this point is determined to be the phase transition critical point. Step 1044 is executed to perform reverse numerical conversion to generate a phase transition correction factor: taking the current difference d = 3℃, the linear conversion formula is applied:

[0115] Calculate the fractional term:

[0116] Calculate the product term: 0.2 × 3 = 0.6, the final output k = 1.0 − 0.6 = 0.4. Simultaneously detect the coordinates of the bottle mouth area point x30y40z0: current temperature 581℃, lower than 580℃? No, do not mark (step 1041). New detection point coordinates x55y65z0, temperature 579℃: lower than 580℃, marked as a candidate point, and calculate d = |579 − 580| = 1℃, substitute into the formula:

[0117]

[0118] The system outputs two phase transition critical point correction factors: 0.4 for the body of the bottle and 0.8 for the mouth of the bottle, forming a spatial distribution map of the phase transition state.

[0119] This scheme uses a triple boundary determination mechanism to accurately pinpoint the spatial location and time point of phase transition; it generates a dynamic correction factor based on the linear transformation of the difference interval, giving higher control priority to the region about to undergo phase transition, effectively mitigating the problem of temperature response hysteresis, and realizing intelligent synchronous control of the glass solidification process, laying the foundation for state-sensitive decision-making for the final closed-loop regulation.

[0120] 105. Combining the hot spot region quantification index of the dynamic thermal field model with the phase change correction factor, predict the cooling rate offset trend and generate a prediction deviation value. Then, adjust the fan speed parameters of the corresponding cooling region according to the prediction deviation value to form a regional closed-loop control of the cooling rate of the glass mold.

[0121] Optionally, step 105 may specifically include the following steps:

[0122] 1051. Read the quantitative index value of the hot spot area of ​​the dynamic thermal field model. The quantitative index value represents the degree of difference between the temperature change rate at the current location and the preset uniform cooling standard.

[0123] 1052. Obtain the phase transition correction factor, which represents the proximity coefficient of the current location point temperature to the phase transition temperature threshold in the same hot spot area, and calculate the expected temperature change by combining the current temperature change rate of the hot spot area with the phase transition correction factor within a fixed time interval.

[0124] 1053. Calculate the difference between the expected temperature change and the preset standard temperature change, analyze the amplitude component of the difference with the fixed time interval length, and output the prediction deviation value representing the expected offset of the temperature change rate per unit time.

[0125] 1054. Divide the glass mold cooling area into multiple independent control blocks, and bind each control block to a set of fan equipment;

[0126] 1055. Based on the prediction deviation value, set up a corresponding absolute size range mapping table, and divide the absolute size of the prediction deviation value into low level, medium level and high level, with each level corresponding to the fixed speed increment of the wind turbine equipment;

[0127] 1056. For each hotspot area, identify the prediction deviation value and corresponding direction of the control block to which it belongs;

[0128] 1057. Based on the predicted deviation value matching mapping table, the adjustment level and speed increment are obtained. When the deviation direction is positive, the speed is increased; when the deviation direction is negative, the speed is decreased. The fan speed parameter adjustment value is output to the speed regulation device of the fan equipment to realize closed-loop control.

[0129] In the above scheme, the hotspot area quantification index value refers to the temperature fluctuation degree of the uneven heat dissipation area detected in step 103, in °C. The phase change correction factor refers to the dimensionless weighting coefficient for glass solidification process regulation generated in step 104. The prediction deviation value refers to the predicted amount of cooling rate deviation per unit time calculated by combining the above parameters, in °C / s. The fan speed parameter refers to the fan speed per minute that controls the cooling intensity. The regional closed-loop control refers to the feedback mechanism that dynamically adjusts the fan speed according to spatial differences.

[0130] In this embodiment, firstly, step 1051 reads the quantitative index value Q of the hot spot region in the dynamic thermal field model. For example, a Q value of 0.8℃ for the bottom block B1 indicates that the temperature fluctuation in this region is 0.8℃ greater than the standard. Secondly, step 1052 obtains the phase change correction factor k for the same region, such as 0.4, and combines it with the current temperature change rate V_now, such as 8℃ per second. Within a fixed time interval Δt = 5 seconds, the expected temperature change ΔT_predict is calculated, which equals V_now multiplied by k multiplied by Δt, i.e., 8 multiplied by 0.4 multiplied by 5, equaling 16℃. Next, step 1053 calculates the difference δT between the expected temperature change and the standard change ΔT_standard. For example, if the standard value is 18℃, then δT equals 16 minus 18, which equals -2℃. This difference and the time interval are calculated using the formula:

[0131] The offset is calculated as -2 divided by 5, which equals -0.4°C per second. Then, step 1054 divides the mold surface into independent control blocks and binds them to fan groups, such as block B1 bound to fans F1 to F3. Step 1055 classifies according to preset rules: when the absolute value of P is less than 0.3, it's a low-level increment of 50 revolutions; 0.3 to 0.6 is a medium-level increment of 100 revolutions; and greater than 0.6 is a high-level increment of 150 revolutions. Step 1056 identifies the direction of P; a negative value indicates insufficient cooling. Step 1057 executes the adjustment: since P equals -0.4, it's at the medium-level adjustment, so an increment of 100 revolutions is selected. Since the direction is negative, the speed needs to be increased, raising the speeds of F1 to F3 by 100 revolutions per minute to achieve closed-loop control.

[0132] In practical applications, during the lampshade mold production process at glass factory C, the system performs closed-loop control on the hot spot area B1 at the bottom of the bottle: First, it reads the quantitative index value of this area as 0.8 degrees Celsius, and simultaneously obtains the phase change correction factor of 0.4 and the current cooling rate of 9 degrees Celsius per second; then, it executes step 1052 to calculate the expected temperature change: taking a fixed time window of 5 seconds, it calculates...

[0133]

[0134] Execute step 1053 to compare with the standard value: preset standard change amount 20 degrees Celsius, calculate the difference: / s, convert the difference into a prediction deviation value. / s, synchronously process bottle neck block A1: quantification index 0.6℃, correction factor 0.8, current rate 7℃ / s; expected change calculation: Standard comparison calculation: / s, Perform graded adjustment (steps 1055-1057):

[0135] Block B1 (bound to fans F01-F03): P absolute value 0.4, intermediate adjustment range (preset: 0.3-0.6 range), base value for speed increment: 100 rpm; a negative P indicates insufficient cooling, instructing the speed to increase by 100 rpm. Implementation calculation: if the current speed is 1500 rpm, then the new setting is 1600 rpm. Block A1 (bound to fan F11): P absolute value 0.4, same level intermediate adjustment, base value for speed increment: 100 rpm; a positive P indicates overcooling risk, instructing the speed to decrease by 100 rpm. Implementation calculation: if the current speed is 1450 rpm, then the new setting is 1350 rpm, ultimately achieving: enhanced heat dissipation at the bottom of the bottle at 1600 rpm, and reduced heat dissipation at the top of the bottle at 1350 rpm, a regional closed-loop feedback control is achieved. The calculation process is fully embedded in the implementation flow, forming a technical closed loop.

[0136] This solution predicts cooling offset trends in advance by integrating quantitative indicators and phase change factors; it transforms complex thermodynamic deviations into concise speed commands using a hierarchical mapping table; it achieves precise point-to-point control of uneven heat dissipation areas using partition binding technology; and finally, it constructs a fully closed-loop control system that monitors, predicts, and regulates, completely solving the problem of alternating oscillations between overcooling and undercooling caused by response lag in traditional methods.

[0137] Figure 2 This application provides a scenario diagram illustrating a closed-loop control method for the cooling rate of a glass mold, as shown in the embodiments below. Figure 2 As shown, a complete embodiment of steps 101-105 includes:

[0138] On the lampshade mold production line of glass factory C, the system executes a closed-loop control process from t0 to t5: Step 101 uses an infrared thermal imager to collect mold surface temperature images every second, such as t0: 560℃ / t1: 555℃ at coordinates (80, 90) on the bottom of the glass bottle. At the same time, 6 thermocouples are deployed in the curvature change zone of the bottle mouth, such as sensor S1 recording t0: 535℃ / t1: 532℃, and 4 thermocouples are deployed in the thickness transition zone of the bottle bottom, such as S7 recording t0: 520℃ / t1: 516℃. The surface cooling rate is calculated to be (560-555) / 0.5=10℃ / s, and the internal conduction rate is (520-516) / 0.5=8℃ / s, and a cooling rate map is generated. Step 102: Based on the 3D model, identify region A (where the curvature of the bottle mouth changes by more than 40 degrees / mm) and region B (where the thickness difference at the bottom of the bottle exceeds 5mm); divide the compensation blocks A1-A2 / B1-B2, extract the surface average velocity of 8.3℃ / s and the internal average velocity of 7.1℃ / s for region A1, calculate the initial coefficient according to (0.6×8.3+0.4×7.1) / 8.5≈0.977, and obtain the thermal compensation coefficient of 1.006 after position correction. Step 103: Create a spatiotemporal data container, mapping the surface rate to spatial points, such as (80,90): 10℃ / s, and mapping internal data, such as S7 coordinates (82,92): 8℃ / s; for the point (81,91) without data, take the weighted average of the four nearest points, 8.1℃ / s; apply a compensation coefficient for correction, such as correcting the original temperature of (80,90) from 555℃ to 555×1.006=558.3℃; calculate the standard deviation of block B1, 3.5℃ exceeding the threshold of 2.0℃, and mark it as a hotspot region with a quantization index value of 1.5. Step 104: Scan the dynamic model at time t3, detecting the point (80,90) with a temperature of 578℃, which is lower than 580℃, making it a candidate point; calculate the phase transition difference |578-580|=2℃; since it is less than the boundary value of 5℃, it is determined to be a critical point for phase transition; generate a correction factor according to k=1.0-0.2×2=0.6. Step 105 reads the quantized value of 1.5℃ for hotspot area B1, combines it with the phase change factor of 0.6 and the current rate of 8.5℃ / s; calculates the expected temperature change in 5 seconds as 8.5×0.6×5=25.5℃; the difference from the standard value of 26℃ is -0.5℃; the prediction deviation P=-0.5 / 5=-0.1℃ / s; binds the fan F1-F3 graded adjustment: P=-0.1 is a low-level +50 rpm command, and the speed is increased to 1550 rpm to complete the closed loop due to the negative value.

[0139] This solution achieves precise control of mold cooling through multi-level technological collaboration: In the three-dimensional data acquisition stage, surface infrared thermal images and internal key point temperature conduction monitoring are integrated to simultaneously capture the real heat dissipation dynamics of complex geometric areas, solving the spatial coverage blind spot problem of traditional monitoring methods; in the adaptive thermal compensation stage, high distortion risk areas are identified based on the mold's three-dimensional features, and compensation parameters are dynamically generated based on the surface and internal heat dissipation differences to effectively balance the conduction imbalance caused by the structure; in the thermal field visualization stage, a time-space air conditioning fusion model is constructed to identify and quantify the core areas of uneven heat dissipation through abnormal temperature fluctuations, achieving precise location of potential hazard areas; in the phase change pre-response stage, a temperature difference conversion mechanism is established to dynamically generate control weight factors near the solidification critical point, significantly improving the control sensitivity of the phase change sensitive area; in the predictive closed-loop stage, heat dissipation anomaly indicators and solidification risk coefficients are integrated to convert thermal state differences into graded adjustment commands, ultimately achieving differentiated feedback control of cooling intensity through zoned fan binding. The entire process forms an adaptive closed loop from monitoring, modeling, early warning to execution, ensuring the heat dissipation uniformity of complex structure molds during the critical phase change period, achieving overall optimization of cooling process stability and product molding quality.

[0140] This plan Figure 3 This application provides a schematic diagram of the structure of a closed-loop control system for the cooling rate of a glass mold, as shown in the embodiment of the present application. Figure 2 As shown, the system includes:

[0141] The acquisition module 31 is used to acquire infrared thermal image data of the glass mold surface, analyze the infrared thermal image data through pixel-level temperature changes, and calculate the real-time cooling rate map by combining it with time changes. At the same time, a point-like temperature measuring array is deployed in the curvature change area and thickness transition area of ​​the glass mold surface to synchronously acquire the temperature conduction data inside the glass mold.

[0142] The partitioning module 32 is used to partition the heat conduction compensation region based on the cooling rate diagram and the temperature conduction data, according to the three-dimensional geometric features of the glass mold, and determine the heat compensation coefficient corresponding to the compensation region.

[0143] The generation module 33 is used to fuse the cooling rate map, the temperature conduction data and the thermal compensation coefficient, perform spatial and temporal three-dimensional fusion processing to generate a dynamic thermal field model, and identify and quantify hot spots with uneven cooling.

[0144] The generation module 33 is also used to detect the phase transition critical point of the glass material in real time in the dynamic thermal field model, and generate a phase transition correction factor when the temperature reaches the preset phase transition threshold.

[0145] The prediction module 34 is used to combine the hot spot area quantification index of the dynamic thermal field model with the phase change correction factor to predict the cooling rate offset trend and generate a prediction deviation value, and to adjust the fan speed parameters of the corresponding cooling area in stages according to the prediction deviation value, so as to form a regional closed-loop control of the cooling rate of the glass mold.

[0146] Figure 3 The aforementioned closed-loop control system for the cooling rate of a glass mold can perform... Figure 1 The implementation principle and technical effects of the closed-loop control method for the cooling rate of a glass mold described in the above embodiment will not be repeated here. The specific methods by which each module and unit of the closed-loop control system for the cooling rate of a glass mold performs its operations have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0147] In one possible design, Figure 3 The glass mold cooling rate closed-loop control system of the embodiment shown can be implemented as a computing device, such as... Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42;

[0148] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 42.

[0149] The processing component 42 is used for the above Figure 1 The embodiment describes a closed-loop control method for the cooling rate of a glass mold.

[0150] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0151] Storage component 41 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0152] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0153] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0154] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0155] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0156] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown illustrates a closed-loop control method for the cooling rate of a glass mold.

[0157] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0158] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0159] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A closed-loop control method for the cooling rate of a glass mold, characterized in that, include: Infrared thermal image data of the glass mold surface is collected, and the infrared thermal image data is analyzed by pixel-level temperature change. The real-time cooling rate map is calculated by combining the time change. At the same time, point temperature measurement arrays are deployed in the curvature change area and thickness transition area of ​​the glass mold surface to collect temperature conduction data inside the glass mold simultaneously. Based on the cooling rate diagram and the temperature conduction data, a heat conduction compensation region is divided according to the three-dimensional geometric features of the glass mold, and a heat compensation coefficient corresponding to the compensation region is determined, including: Input three-dimensional geometric model data of the glass mold containing surface shape, thickness and curvature information, and based on the three-dimensional geometric model data, identify areas where the curvature change exceeds a set threshold and areas where the thickness difference exceeds a set threshold; The region is divided into multiple independent blocks to form a compensation region. Within each compensation region, the average temperature change rate at the surface position in the cooling rate map and the average temperature value of the temperature conduction data from the point temperature measurement array are extracted. The initial value of the thermal compensation coefficient is calculated according to a preset rule. The initial value of the thermal compensation coefficient is adjusted according to the relative positional relationship between the compensation area and the overall three-dimensional geometric model to obtain the final thermal compensation coefficient. By integrating the cooling rate map, the temperature conduction data, and the thermal compensation coefficient, a dynamic thermal field model is generated through spatial and temporal three-dimensional fusion processing to identify and quantify hot spots with uneven cooling. The phase transition critical point of the glass material is detected in real time in the dynamic thermal field model, and a phase transition correction factor is generated when the temperature reaches the preset phase transition threshold. By combining the hot spot region quantification index of the dynamic thermal field model with the phase change correction factor, the cooling rate offset trend is predicted and a prediction deviation value is generated. Based on the prediction deviation value, the fan speed parameters of the corresponding cooling region are adjusted in stages to form a regional closed-loop control of the cooling rate of the glass mold.

2. The method according to claim 1, characterized in that, Combining the hotspot region quantification index of the dynamic thermal field model with the phase change correction factor, the cooling rate shift trend is predicted and a prediction deviation value is generated, including: Read the quantitative index value of the hot spot area in the dynamic thermal field model. The quantitative index value represents the degree of difference between the temperature change rate at the current location and the preset uniform cooling standard. Obtain the proximity coefficient of the same hot spot area indicating that the temperature at the current location point has reached the phase transition temperature threshold. The proximity coefficient is the phase transition correction factor. Then, within a fixed time interval, combine the current temperature change rate of the hot spot area with the phase transition correction factor to calculate the expected temperature change. Calculate the difference between the expected temperature change and the preset standard temperature change, analyze the magnitude component of the difference with the fixed time interval length, and output a predicted deviation value representing the expected offset of the temperature change rate per unit time.

3. The method according to claim 1, characterized in that, Based on the predicted deviation value, the fan speed parameters of the corresponding cooling zone are adjusted in stages to form a regionalized closed-loop control of the glass mold cooling rate, including: The glass mold cooling area is divided into multiple independent control blocks, and each control block is bound to a set of fan equipment; Based on the prediction deviation value, a corresponding absolute size range mapping table is set, and the absolute size of the prediction deviation value is divided into low level, medium level and high level, with each level corresponding to the fixed speed increment of the wind turbine equipment. For each hotspot area, identify the prediction deviation value and corresponding direction of the control block to which it belongs; The adjustment level and speed increment are obtained by matching the predicted deviation value with the mapping table. When the deviation direction is positive, the speed is increased; when the deviation direction is negative, the speed is decreased. The fan speed parameter adjustment value is output to the speed regulation device of the fan equipment to realize closed-loop control.

4. The method according to claim 1, characterized in that, The phase transition critical point of the glass material is detected in real time in the dynamic thermal field model. When the temperature reaches a preset phase transition threshold, a phase transition correction factor is generated, including: In the spatiotemporal data structure of the dynamic thermal field model, a continuous scanning operation is performed on all location points. When the current temperature value of a location point is found to be lower than the preset first temperature boundary, the location point is marked as a candidate monitoring point. The absolute difference between the current temperature value of the candidate monitoring point and the preset phase transition temperature threshold for the solid-liquid phase transition of the glass material is calculated in real time. A judgment operation is performed on the absolute difference of the candidate monitoring points. When the absolute difference is less than the preset second temperature boundary, the candidate monitoring point is judged as the phase transition critical point. Based on the absolute difference between the current temperature value and the phase transition temperature threshold, a reverse numerical conversion operation is performed. When the absolute difference is equal to zero, a correction value of 1 is output, and when the absolute difference is equal to the second temperature boundary, a correction value of 0 is output. Within the range of the absolute difference, a correction coefficient is continuously calculated as a phase transition correction factor according to a linear rule.

5. The method according to claim 1, characterized in that, By integrating the cooling rate map, the temperature conduction data, and the thermal compensation coefficient, a dynamic thermal field model is generated through spatial and temporal three-dimensional fusion processing. This model identifies and quantifies hotspot regions exhibiting uneven cooling, including: Create a spatiotemporal data structure that includes three-dimensional spatial location (x, y, z) and continuous time points; The temperature change rate data at the surface location in the cooling rate diagram is mapped to the surface location point in three-dimensional space, and the temperature conduction data of the point temperature measurement array is mapped to the internal location point in three-dimensional space. For unmapped location points, a weighted average is used to assign values ​​based on neighboring location point data. In the spatiotemporal data structure, a correction temperature value is calculated based on the temperature value of each location point and the thermal compensation coefficient of the corresponding compensation area, and the spatiotemporal data structure is corrected based on the correction temperature value. Based on the corrected spatiotemporal data structure, the standard deviation of temperature values ​​at all locations is calculated at each time point. Regions where the standard deviation exceeds a set threshold are identified as uneven cooling regions. The difference between the standard deviation and the set threshold is used as a quantification index value to identify and quantify the hotspot regions of the uneven cooling regions.

6. The method according to claim 1, characterized in that, The infrared thermal image data is analyzed at the pixel level for temperature changes, and a real-time cooling rate map is calculated by combining the results with time-varying parameters. This includes: Continuous thermal image data is acquired from an infrared imaging device on the surface of a glass mold. The thermal image data consists of multiple uniformly distributed measurement points, each containing a temperature value. A recording sequence with a fixed time interval is set for each measurement point, and the temperature value at the previous moment and the temperature value at the current moment are stored in the recording sequence; For each measurement point, the difference between the temperature value at the previous moment and the temperature value at the current moment is calculated, and the rate of temperature change at each measurement point is calculated by combining the difference with a fixed time interval. Arrange the temperature change rates of all measurement points according to their positions in the original thermal image data, and output a cooling rate map.

7. A closed-loop control system for the cooling rate of a glass mold, characterized in that, include: Infrared thermal image data of the glass mold surface is collected, and the infrared thermal image data is analyzed by pixel-level temperature change. The real-time cooling rate map is calculated by combining the time change. At the same time, point temperature measurement arrays are deployed in the curvature change area and thickness transition area of ​​the glass mold surface to collect temperature conduction data inside the glass mold simultaneously. Based on the cooling rate diagram and the temperature conduction data, a heat conduction compensation region is divided according to the three-dimensional geometric features of the glass mold, and a heat compensation coefficient corresponding to the compensation region is determined, including: Input three-dimensional geometric model data of the glass mold containing surface shape, thickness and curvature information, and based on the three-dimensional geometric model data, identify areas where the curvature change exceeds a set threshold and areas where the thickness difference exceeds a set threshold; The region is divided into multiple independent blocks to form a compensation region. Within each compensation region, the average temperature change rate at the surface position in the cooling rate map and the average temperature value of the temperature conduction data from the point temperature measurement array are extracted. The initial value of the thermal compensation coefficient is calculated according to a preset rule. The initial value of the thermal compensation coefficient is adjusted according to the relative positional relationship between the compensation area and the overall three-dimensional geometric model to obtain the final thermal compensation coefficient. By integrating the cooling rate map, the temperature conduction data, and the thermal compensation coefficient, a dynamic thermal field model is generated through spatial and temporal three-dimensional fusion processing to identify and quantify hot spots with uneven cooling. The phase transition critical point of the glass material is detected in real time in the dynamic thermal field model, and a phase transition correction factor is generated when the temperature reaches the preset phase transition threshold. By combining the hot spot region quantification index of the dynamic thermal field model with the phase change correction factor, the cooling rate offset trend is predicted and a prediction deviation value is generated. Based on the prediction deviation value, the fan speed parameters of the corresponding cooling region are adjusted in stages to form a regional closed-loop control of the cooling rate of the glass mold.

8. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a closed-loop control method for cooling rate of a glass mold as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements a closed-loop control method for the cooling rate of a glass mold as described in any one of claims 1 to 6.

Images