Temperature control method, system and storage medium for thermal processing

Abstract

The present application relates to the technical field of processing temperature control, and discloses a temperature control method, system, and storage medium for a hot working process. The method comprises: collecting temperature data of the entire hot working process through multiple sensors, calculating the temperature change rate, extreme points, and uniformity index, and obtaining key characteristic parameters; establishing a quantitative relationship matrix between process parameters and temperature field using orthogonal experiments; using a rolling time domain optimization algorithm based on the matrix to calculate the optimal process parameter combination to generate control instructions; predicting and correcting the execution timing and duration of the instructions to form precise target control instructions. The present application improves the temperature control accuracy and stability of the hot working process of large alloy steel components, thereby improving product quality, reducing energy consumption, and shortening the production cycle.

Description

Technical Field

[0001] The present application relates to the technical field of processing temperature control, and in particular to a temperature control method, system and storage medium for a thermal processing process. Background Art

[0002] In the manufacturing process of alloy steel components for heavy equipment, hot working is one of the key processes, including smelting, casting, forging and heat treatment. The hot working process is usually carried out in a high-temperature environment. It is a typical long-process manufacturing process with many process types and many process factors that affect manufacturing quality. The high-temperature working environment of hot working also makes it difficult to collect process data. Currently, hot working temperature control mainly relies on experience judgment and limited point measurement. Traditional temperature control methods are often based on discrete temperature measurement points, combined with the experience of process personnel for manual adjustment, and lack systematic real-time temperature field monitoring and precise control methods. This method can still cope with the processing of small or conventional components, but with the continuous increase in the size and weight of equipment components and the continuous improvement of performance requirements, traditional control methods can no longer meet the needs.

[0003] The main problems with current hot working temperature control are incomplete temperature field monitoring and imprecise control. Due to a limited number of temperature measurement points, it's impossible to obtain the complete temperature distribution inside and on the surface of a component, resulting in insufficient temperature field information. Temperature control relies primarily on empirical judgment, lacking a quantitative process parameter-temperature field relationship model, resulting in low control accuracy. Hot working equipment exhibits lag in response, with the actual execution of control instructions deviating from expectations, and a lack of effective prediction and correction mechanisms. These issues lead to unstable product quality, large temperature differences between the inside and outside of large components, poor structural uniformity, and the susceptibility to defects. This is particularly true when large components continuously push the limits of size, weight, and performance. Summary of the Invention

[0004] The present application provides a temperature control method, system and storage medium for a thermal processing process, which are used to improve the temperature control accuracy and stability of the thermal processing process of large alloy steel components, thereby improving product quality, reducing energy consumption and shortening the production cycle.

[0005] In the first aspect, the present application provides a temperature control method for a hot working process, and the temperature control method for a hot working process includes: collecting temperature data of each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the entire hot working process; calculating the temperature change rate, temperature extreme points and temperature uniformity index in the hot working process based on the spatiotemporal temperature distribution information to obtain key temperature characteristic parameters; performing orthogonal experimental analysis on process parameters based on the key temperature characteristic parameters to establish a quantitative relationship matrix of the influence of process parameters on the temperature field; according to the quantitative relationship matrix, using a rolling time domain optimization algorithm to calculate the optimal combination of heating power, cooling rate and holding time to generate a temperature control instruction; predicting the execution timing and execution duration of the temperature control instruction to obtain an execution prediction result, and correcting the control instruction based on the execution prediction result to obtain a target control instruction.

[0006] In a second aspect, the present application provides a temperature control system for a thermal processing process, the temperature control system for a thermal processing process comprising:

[0007] The acquisition module is used to collect temperature data from each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the entire thermal processing process;

[0008] a calculation module for calculating the temperature change rate, temperature extreme points and temperature uniformity index during the thermal processing process based on the spatiotemporal temperature distribution information to obtain key temperature characteristic parameters;

[0009] An analysis module is used to perform orthogonal test analysis on process parameters based on the key temperature characteristic parameters, and establish a quantitative relationship matrix of the influence of process parameters on temperature field;

[0010] A generation module is used to calculate the optimal combination of heating power, cooling rate and holding time according to the quantitative relationship matrix using a rolling horizon optimization algorithm to generate a temperature control instruction;

[0011] The prediction module is used to predict the execution timing and execution duration of the temperature control instruction to obtain an execution prediction result, and to modify the control instruction according to the execution prediction result to obtain a target control instruction.

[0012] In a third aspect, a temperature control device for a thermal processing process is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the temperature control device for the thermal processing process executes the above-mentioned temperature control method for the thermal processing process.

[0013] In a fourth aspect, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, which, when executed on a computer, enables the computer to execute the above-mentioned temperature control method for a thermal processing process.

[0014] In the technical solution provided by the present application, the temperature data of each process of smelting, casting, forging and heat treatment are collected through a multi-sensor network, thereby realizing the comprehensive acquisition of spatiotemporal temperature distribution information of the whole process of thermal processing, overcoming the problem of insufficient temperature field information caused by limited temperature measurement points in traditional methods; based on the acquired spatiotemporal temperature distribution information, key temperature characteristic parameters such as temperature change rate, temperature extreme point and temperature uniformity index are calculated, providing quantitative indicators for temperature field analysis, and enhancing the accuracy and comprehensiveness of temperature field analysis; through orthogonal experimental analysis, a quantitative relationship matrix of the influence of process parameters on temperature field is established, which converts the traditional temperature field analysis that relies on experience into a quantitative relationship matrix. The system transforms precision control into precise control based on data and models, realizing bidirectional mapping from process parameters to temperature field and from temperature field to process parameters; adopts rolling time domain optimization algorithm to calculate the optimal combination of heating power, cooling rate and holding time, generates temperature control instructions, and upgrades the temperature control strategy from static optimization to dynamic optimization, which can respond to changes in working conditions in real time and improve the adaptability and precision of control; by predicting the execution timing and execution duration of temperature control instructions and correcting the control instructions according to the prediction results, the problems of lag and deviation in the execution of control instructions are solved, and the execution precision of control instructions is significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 This is a schematic diagram of an embodiment of a temperature control method for a thermal processing process in an embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of an embodiment of a temperature control system used in a thermal processing process in an embodiment of the present application;

[0018] Figure 3 It is a schematic block diagram of the structure of a temperature control device used in a thermal processing process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] Embodiments of the present application provide a temperature control method, system, and storage medium for a thermal processing process. The terms "first," "second," "third," "fourth," and the like (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the data used in this way are interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or apparatus.

[0020] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of the temperature control method for a thermal processing process in the embodiment of the present application includes:

[0021] Step S101: collecting temperature data of each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the whole process of heat treatment;

[0022] Step S102: Calculate the temperature change rate, temperature extreme point, and temperature uniformity index during the thermal processing process based on the spatiotemporal temperature distribution information to obtain key temperature characteristic parameters;

[0023] Step S103: performing orthogonal test analysis on process parameters based on key temperature characteristic parameters to establish a quantitative relationship matrix of the influence of process parameters on temperature field;

[0024] Step S104: Calculate the optimal combination of heating power, cooling rate, and holding time using a rolling time domain optimization algorithm based on the quantitative relationship matrix to generate a temperature control instruction;

[0025] Step S105: predict the execution timing and execution duration of the temperature control instruction to obtain an execution prediction result, and modify the control instruction according to the execution prediction result to obtain a target control instruction.

[0026] It is understandable that the execution subject of the present application can be a temperature control system used in a thermal processing process, or a terminal or a server, which is not limited here. The embodiment of the present application is described by taking a server as the execution subject as an example.

[0027] Specifically, during the smelting process, infrared thermometers are primarily installed above the electric arc furnace, with a measurement range of 1000°C to 1800°C and an accuracy of ±0.5%, for monitoring the surface temperature of the molten metal. Thermocouples are embedded in the furnace wall, with a measurement range of 20°C to 1600°C and an accuracy of ±0.25%, for monitoring the temperature inside the furnace. Thermal imagers, with a spatial resolution of 320×240 pixels and a temperature resolution of 0.05°C, are installed at the edge of the smelting area to monitor overall temperature distribution. During the casting process, infrared thermometers are installed in the pouring area to measure pouring temperatures. A thermocouple array is embedded in the mold wall in a grid pattern with a spacing of 20 cm, enabling three-dimensional monitoring of the mold temperature. The thermal imager uses an automatic tracking system to record temperature changes throughout the entire ingot casting process, from pouring to cooling. During the forging process, infrared thermometers and thermal imagers are mounted around the 7000T and 18500T hydraulic presses, respectively, to measure the forging surface temperature in real time. Multi-point thermocouple arrays are arranged at different heights and locations within the heating furnace to form a temperature monitoring network. During the heat treatment process, multi-point thermocouple arrays are located on the top, bottom, and walls of the furnace to monitor the temperature distribution within the furnace. Infrared thermometers and thermal imagers are installed in the furnace exit area to record the temperature profile of the workpiece during cooling. Data collected by these sensors is transmitted to the data processing system via Industrial Ethernet and 5G wireless networks with an acquisition frequency of at least 100ms. A Kalman filter algorithm is used to eliminate measurement noise and errors. This algorithm processes sensor data in two stages: prediction and correction. First, the current state is predicted based on the previous state. Then, the predicted value is corrected based on actual measurement results to obtain a more accurate temperature estimate. The filtered temperature data is then used to reconstruct the continuous temperature field distribution using a three-dimensional interpolation algorithm, ultimately yielding complete spatiotemporal temperature distribution information, including temperature values, three-dimensional spatial coordinates, and timestamps.

[0028] The spatiotemporal temperature distribution information is segmented according to the process stage, and the ratio of the temperature difference between adjacent time points and the time interval is calculated for each measuring point to obtain the temperature change rate curve. For each point in the curve, its first-order derivative value is calculated. When the derivative changes from positive to negative or from negative to positive, it is marked as an inflection point. These inflection points represent the key temperature transition moments in the thermal processing process. For example, in the heat treatment stage, when the workpiece switches from heating to insulation, the temperature change rate drops sharply from a positive value to near zero, and the inflection point at this time is marked as the heating completion point; when the workpiece switches from insulation to cooling, the rate changes from near zero to negative, and the inflection point at this time is marked as the cooling start point. A heat map is constructed based on the spatiotemporal temperature distribution information. The heat map indicates the temperature by the depth of color, with red representing the high temperature area and blue representing the low temperature area. A peak search algorithm is applied to the thermal map. This algorithm compares the temperature of each point with its neighbors. When a point's temperature is higher than all its neighbors, it is marked as a local maximum; when its temperature is lower than all its neighbors, it is marked as a local minimum. This results in a distribution map of temperature extremes. The temperature gradient vectors between the extremes are calculated. The direction of the gradient vector points to the direction of the fastest temperature rise, and its magnitude indicates the severity of the temperature change. This provides information on the direction and intensity of the temperature change. Statistical parameters of the spatial temperature field are calculated for each time point, including the standard deviation (reflecting the dispersion of the temperature distribution), skewness (reflecting the asymmetry of the distribution), and kurtosis (reflecting the peakedness of the distribution). These parameters collectively constitute the temperature uniformity index. Finally, the temperature change rate, extreme point distribution, gradient vectors, and statistical parameters are integrated into a multidimensional feature vector to form the key temperature characteristic parameters of the thermal processing process.

[0029] For the smelting process, parameters such as arc power, vacuum level, bottom-blowing gas flow rate, and holding time were selected; for the casting process, parameters such as pouring temperature, pouring speed, mold preheating temperature, and riser size were selected; for the forging process, parameters such as heating temperature, forging deformation, forging speed, and mold temperature were selected; and for the heat treatment process, parameters such as heating rate, holding temperature, holding time, and cooling rate were selected. An orthogonal experimental design approach was used to rationally arrange the experimental plan, significantly reducing the number of experiments. For example, for the heat treatment process, if each of the four parameters had three levels, a full factorial experiment would require 81 experiments, while an orthogonal array only required nine experiments to analyze the main effects. In each set of experiments, temperature data was collected via a multi-sensor network, and key temperature characteristic parameters were extracted to establish a process parameter-temperature characteristic response dataset. This dataset was subjected to an analysis of variance to calculate the contribution rate and significance level of each process parameter to the temperature characteristic parameter. The contribution rate indicates the degree to which the parameter explains the variation in the results, and the significance level indicates whether the parameter's influence is statistically significant. Based on the results of variance analysis, influence weight coefficients were constructed, and polynomial regression was used to fit the nonlinear relationship between process parameters and temperature characteristic parameters, resulting in a process-temperature influence function. Partial derivatives of this function were calculated to obtain a Jacobian matrix, in which each element represents the degree to which a small change in a process parameter affects a specific characteristic of the temperature field. The Jacobian matrix was normalized to obtain a quantitative relationship matrix, which clearly describes the corresponding relationship between process parameter adjustments and temperature field changes.

[0030] Product quality indicators are converted into temperature control targets. For example, in the forging process, the internal and external temperature difference of the forging must not exceed 50°C, and in the heat treatment process, the heating rate must not exceed 150°C / hour and the temperature fluctuation must be controlled within ±5°C. A temperature field state prediction formula is constructed based on a quantitative relationship matrix. This formula can predict the temperature field evolution sequence over a period of time under given process parameters. The core idea of the rolling horizon optimization algorithm is to execute only the control variables for the first time step in each control cycle, then perform re-optimization within a sliding prediction window. Specifically, the algorithm first sets a prediction horizon (e.g., 30 minutes) and a control horizon (e.g., 5 minutes). At the beginning of each control cycle, based on the current temperature state and the quantitative relationship matrix, the temperature field evolution trajectory for different process parameter combinations within the future prediction horizon is predicted and the weighted error relative to the target temperature curve is calculated. A nonlinear programming method is used to solve the optimal process parameter combination that minimizes the prediction error. The parameter values for the first control cycle of the optimal process parameter combination are then executed. After the cycle ends, the model state is updated based on actual temperature feedback, and the prediction window is moved to re-optimize the calculation. Through this rolling optimization approach, the algorithm adapts to changing operating conditions and model errors, achieving precise control of the temperature field. Ultimately, the optimized sequence of process parameters, such as heating power, cooling rate, and holding time, is converted through piecewise linear interpolation into a continuous parameter control curve that meets the equipment's execution accuracy. Standard temperature control instructions are then generated according to the equipment's control interface requirements.

[0031] Temperature control instructions are matched and analyzed with historical execution data to extract instruction execution delay characteristics, temperature response characteristics, and device status characteristics. The execution delay characteristic is calculated by recording the historical instruction issuance time, execution start time, and execution end time, and then calculating the execution delay and actual duration. The temperature response characteristic is extracted by time-aligning historical temperature data with the control instructions, and then extracting the temperature change rate, temperature response time, and temperature stability parameters before and after instruction execution. Device status characteristics are extracted from device operation logs, including device start / stop status, workload, maintenance cycle, and fault history. Based on these characteristics, a long short-term memory (LSTM) neural network is constructed. The network consists of an input layer, two LSTM hidden layers, a fully connected layer, and an output layer. The input layer receives the instruction feature vector. The first hidden layer contains 128 neurons for extracting timing features, the second hidden layer contains 64 neurons for feature fusion, and the fully connected layer contains 32 neurons. The output layer generates execution timing offset and execution duration predictions. The temperature control instruction is converted into a feature vector and input into the neural network. The execution prediction result is calculated through forward propagation. Based on the prediction result, the instruction sequence is reordered and parameters are adjusted to eliminate execution conflicts and optimize the execution order. Finally, the temperature target achievement degree of the corrected instruction is evaluated, the temperature evolution trajectory is calculated using the heat conduction model, and compensation adjustments are made for the part that deviates from the target to obtain the final target control instruction.

[0032] In the embodiment of the present application, the temperature data of each process of smelting, casting, forging and heat treatment is collected through a multi-sensor network, which realizes the comprehensive acquisition of the spatiotemporal temperature distribution information of the whole process of thermal processing, overcoming the problem of insufficient temperature field information caused by the limited temperature measurement points in the traditional method; based on the acquired spatiotemporal temperature distribution information, key temperature characteristic parameters such as temperature change rate, temperature extreme point and temperature uniformity index are calculated, which provides quantitative indicators for temperature field analysis and enhances the accuracy and comprehensiveness of temperature field analysis; through orthogonal experimental analysis, a quantitative relationship matrix of the influence of process parameters on temperature field is established, which converts the traditional temperature control method that relies on experience into a quantitative relationship matrix. The system is transformed into precise control based on data and models, realizing bidirectional mapping from process parameters to temperature field and from temperature field to process parameters; the rolling time domain optimization algorithm is used to calculate the optimal combination of heating power, cooling rate and holding time, and generate temperature control instructions, so that the temperature control strategy is upgraded from static optimization to dynamic optimization, which can respond to changes in working conditions in real time and improve the adaptability and accuracy of control; by predicting the execution timing and execution duration of temperature control instructions and correcting the control instructions according to the prediction results, the problems of lag and deviation in the execution of control instructions are solved, and the execution accuracy of control instructions is significantly improved.

[0033] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0034] Infrared thermometers, thermocouples and thermal imagers are installed in the smelting process to measure the surface temperature of the molten metal, the temperature in the furnace and the overall temperature distribution to obtain the temperature data of the smelting process;

[0035] In the casting process, an infrared thermometer is set to measure the pouring temperature, a thermocouple array is used to embed the casting mold to monitor the casting temperature, and a thermal imager is used to track the cooling process of the ingot to obtain the casting process temperature data;

[0036] Infrared thermometers and thermal imagers are deployed in the forging process to measure the surface temperature of forgings, and multi-point thermocouple arrays are installed to monitor the temperature field of the heating furnace to obtain the temperature data of the forging process;

[0037] Arrange a multi-point thermocouple array for the heat treatment process to monitor the temperature distribution of the heat treatment furnace, use an infrared thermometer and a thermal imager to record the cooling curve of the workpiece after leaving the furnace, and obtain the temperature data of the heat treatment process;

[0038] The temperature data of the smelting process, casting process, forging process and heat treatment process are transmitted to the data processing system through the industrial control network, and the Kalman filter algorithm is used to eliminate the measurement error to obtain the fused temperature data;

[0039] The fused temperature data is interpolated and reconstructed in three dimensions to construct the continuous spatiotemporal temperature distribution information of the entire thermal processing process. The spatiotemporal temperature distribution information includes three dimensions: temperature value, spatial coordinates, and timestamp.

[0040] Specifically, during the smelting process, an infrared thermometer is fixed approximately 2 meters above the electric arc furnace, using a non-contact measurement method to accurately capture the surface temperature of the molten metal. The measurement range is 1000°C-1800°C, and the sampling frequency is 10Hz. High-temperature-resistant K-type thermocouples are embedded in different locations on the furnace wall to form a measurement network for monitoring the temperature distribution within the furnace. The measurement range is 20°C-1600°C, and the sampling frequency is 1Hz. A thermal imager is set up in a safe position at the edge of the smelting area, equipped with a cooling system and dustproof devices. It scans at a frequency of 50Hz and has a resolution of 320×240 pixels to obtain overall temperature distribution information. The three sensors work together to form complementary measurements: the infrared thermometer focuses on measuring surface temperature, the thermocouple measures internal temperature, and the thermal imager provides global temperature distribution, forming a complete smelting temperature monitoring network. During the casting process, an infrared thermometer is fixed above the pouring area, focused on the pouring flow, and measures the pouring temperature in real time. The sampling frequency is increased to 20Hz to ensure that temperature fluctuations are captured; the thermocouple array is embedded in the mold wall according to a computer-optimized layout. The number depends on the size of the mold. Generally, 50-100 measuring points are arranged in a large mold to form a three-dimensional grid structure with a spacing of 10-30cm. It penetrates into different depths inside the mold to comprehensively monitor temperature changes in various areas of the mold; the thermal imager uses an automatic tracking system to continuously scan the entire process from the beginning of ingot pouring to complete cooling. The scanning range covers the entire ingot surface, records the temperature change curve and temperature distribution changes during the cooling process, and provides data support for solidification process analysis.

[0041] During the forging process, infrared thermometers are installed at key locations within the forging operation area, such as around the press, to measure the surface temperature of the forgings at a sampling frequency of 5Hz. Thermal imagers are mounted above the forging equipment, providing temperature distribution information across the entire surface of the forgings at a scanning frequency of 30Hz. Multi-point thermocouple arrays are installed on the top, bottom, and walls of the heating furnace, forming a three-dimensional temperature measurement grid to monitor the temperature field distribution within the heating furnace. Thermocouples are spaced 20-50cm apart, and the number of thermocouples is determined by the size of the furnace, typically ranging from 30 to 80. The coordinated use of these three sensors ensures comprehensive temperature data collection during the forging process, monitoring both the heating process and temperature changes during the forming process.

[0042] During the heat treatment process, a multi-point thermocouple array is installed in an optimized layout at various locations within the furnace, covering all areas. The number of thermocouples typically ranges from 20 to 60, determined by the furnace dimensions and arranged in a three-dimensional grid structure, to monitor the temperature distribution within the furnace. Infrared thermometers and thermal imagers are stationed in the furnace exit area. The infrared thermometer is aimed at specific points on the workpiece exiting the furnace, while the thermal imager provides an image of the surface temperature distribution across the entire workpiece. Together, they record the temperature profile of the workpiece from exiting the furnace to completion of cooling, with sampling frequencies of 2Hz and 20Hz, respectively. Temperature data collected from each process is transmitted to a central data processing system via an industrial control network. The industrial control network utilizes a two-tier architecture: the field layer uses Industrial Ethernet with a transmission rate of 100Mbps, covering the fixed equipment area; the control layer utilizes a 5G wireless network with a transmission rate of 1Gbps, covering the mobile device area. The data collection frequency is dynamically adjusted according to the characteristics of the process stage. It is increased to 100ms / time at critical moments (such as the start of pouring and the beginning of forging), and reduced to 1s / time during the stable stage, ensuring that key temperature changes are not missed while the data volume is controllable.

[0043] After data is transmitted to the processing system, it undergoes preprocessing, including outlier removal, noise filtering, and data normalization. Outlier removal uses the 3σ criterion: data deviating from the mean by more than three standard deviations is considered an outlier and removed. Noise filtering utilizes the Kalman filter algorithm, which recursively combines measured and predicted values to minimize the covariance of the estimation error, thereby obtaining an optimal estimate. When processing temperature data, the Kalman filter algorithm first establishes the system state equation and the measurement equation. The state equation describes the evolution of temperature over time, while the measurement equation describes the relationship between the measured value and the true value. A prediction step then predicts the current state based on the previous state. A correction step then uses the current measurement result to correct the prediction to obtain an optimal estimate of the current state. This iterative prediction-correction process effectively eliminates random and systematic errors in sensor measurements, improving the accuracy of temperature data.

[0044] The temperature data processed by the Kalman filter is interpolated and reconstructed in three dimensions to construct continuous spatiotemporal temperature distribution information. This interpolation utilizes the Kriging interpolation method, which is based on regionalized variable theory and considers the spatial relationship between sample points and the spatial autocorrelation of the data to provide the best linear unbiased estimate of the values at unsampled points. The three-dimensional reconstruction, based on octree spatial partitioning and the Mach cube algorithm, transforms discrete temperature data points into a continuous three-dimensional temperature field. The resulting spatiotemporal temperature distribution information consists of three dimensions: temperature value (i.e., the temperature value of each spatial point at a specific moment), spatial coordinates (i.e., the three-dimensional position of the temperature point, referenced to the device coordinate system), and timestamp (i.e., the time the temperature data was collected, accurate to the millisecond level).

[0045] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0046] The spatiotemporal temperature distribution information is processed into time series segments, the ratio of the temperature difference between adjacent time points to the time interval is calculated, and the temperature change rate curve of each measuring point is obtained;

[0047] Apply the derivative extreme value detection algorithm to the temperature change rate curve, mark the inflection point where the slope of the rate curve changes significantly, and obtain the key temperature transition moment during the hot working process;

[0048] Based on the spatiotemporal temperature distribution information, a thermal map is constructed, and the peak search algorithm is used to identify the coordinates of the local highest and lowest points in the temperature field to obtain the temperature extreme point distribution map;

[0049] The temperature gradient between the extreme points of the temperature extreme point distribution diagram is calculated, and the temperature gradient vector field is constructed to obtain the direction and intensity of temperature change during the hot working process;

[0050] According to the spatiotemporal temperature distribution information, the standard deviation, skewness and kurtosis of the spatial temperature field are calculated at each time point to obtain statistical parameters that characterize the uniformity of temperature distribution;

[0051] The temperature change rate curve, key temperature transition moment, temperature extreme point distribution diagram, temperature gradient vector field and statistical parameters are fused to construct a multidimensional feature vector and obtain the key temperature characteristic parameters of the thermal processing process.

[0052] Specifically, the acquired spatiotemporal temperature distribution information is segmented into time series, dividing the temperature data into different time periods based on the process stages. For smelting processes, this is divided into stages such as material preparation, roughing, refining, and vacuum treatment; for casting processes, into stages such as pouring, solidification, and cooling; for forging processes, into stages such as heating, forging, and cooling; and for heat treatment processes, into stages such as heating, holding, quenching, and tempering. Within each time period, the ratio of the temperature difference between adjacent time points to the time interval is calculated to obtain the temperature change rate. Specifically, for each measurement point, the temperature and time values of two adjacent time points are taken, and the difference between them is calculated. The temperature difference divided by the time difference represents the temperature change rate. This calculation is repeated for all time points to generate a temperature change rate curve for each measurement point. This curve intuitively reflects the speed of temperature change during hot working and is an important basis for analyzing process transitions. A derivative extreme value detection algorithm is applied to the temperature change rate curve to identify inflection points where the slope changes significantly. These inflection points typically correspond to critical temperature transitions during hot working. The core of the derivative extreme value detection algorithm is to calculate the rate of change of the rate curve, that is, the rate trend. In practice, the derivative value is approximated by calculating the difference between each point on the curve and the previous and next points. When the derivative value changes from positive to negative or from negative to positive, and the magnitude of the change exceeds a preset threshold, the point is marked as an inflection point. These inflection points correspond to key moments in process transitions, such as the completion of heating, the start of holding, and the beginning of cooling, providing a time reference for subsequent temperature control strategy development.

[0053] A heatmap is constructed based on spatiotemporal temperature distribution information. A heatmap is a visual representation of temperature using color depth, typically using red to indicate high temperatures and blue to indicate low temperatures. During construction, the three-dimensional space is first divided into grid cells, each assigned a temperature value obtained by actual measurement or interpolation. These temperature values are then mapped to a predefined color space to generate a color heatmap. Based on the heatmap, a peak search algorithm is applied to identify local maxima and minima in the temperature field. The peak search algorithm compares the temperature of each point with its surroundings to determine whether it is an extreme point. If a point's temperature is higher than all of its neighboring points, it is marked as a local maximum; if it is lower than all of its neighboring points, it is marked as a local minimum. To avoid noise interference, a minimum temperature difference threshold is typically set; only when the temperature difference exceeds the threshold is it marked as an extreme point. The resulting temperature extreme point distribution map visually illustrates key areas where heat is concentrated or dissipated during the hot working process.

[0054] The temperature gradient between the extreme points of the temperature extreme point distribution diagram is calculated to construct a temperature gradient vector field. The temperature gradient reflects the spatial rate of change and direction of temperature and is a direct indicator of the intensity and direction of heat conduction. For any two extreme points in space, the temperature difference between them is calculated and divided by the spatial distance to obtain the magnitude of the temperature gradient, with the direction pointing from the low temperature point to the high temperature point. The specific operation is to calculate the three-dimensional spatial distance between the two points (considering the distance on the three coordinate axes of x, y, and z), and then divide the temperature difference between the two points by this distance to obtain the temperature gradient value, while recording the direction vector from the low temperature point to the high temperature point. The gradient vectors of all extreme point pairs are combined to form a temperature gradient vector field. This vector field clearly shows the direction and intensity of heat flow during thermal processing, providing a decision-making basis for temperature field uniformity control.

[0055] Based on the spatiotemporal temperature distribution information, the statistical parameters of the spatial temperature field are calculated for each time point, including standard deviation, skewness, and kurtosis. The standard deviation calculation is to average the square root of the sum of the differences between the temperature values of all spatial measurement points and their average temperature at the same time point, reflecting the degree of dispersion of the temperature distribution. The skewness calculation involves the cube of the temperature deviation from the average value, reflecting the asymmetry of the temperature distribution. Positive skewness indicates that the high temperature area is concentrated, and negative skewness indicates that the low temperature area is concentrated. The kurtosis calculation involves the fourth power of the temperature deviation from the average value, reflecting the sharpness of the temperature distribution. High kurtosis indicates that the temperature distribution is sharp, and low kurtosis indicates that the temperature distribution is flat. These three statistical parameters together constitute the temperature uniformity index, which comprehensively characterizes the uniformity of the temperature distribution and provides a quantitative indicator for the control of temperature field uniformity during thermal processing.

[0056] The temperature change rate curve, key temperature transition moments, temperature extreme point distribution diagram, temperature gradient vector field, and statistical parameters are fused to construct a multidimensional feature vector, which yields the key temperature characteristic parameters of the hot working process. Feature fusion first normalizes all features, converting features of different measurement units to the same range. Different weights are then assigned to different features based on process requirements to highlight the impact of important features. Finally, the weighted features are combined into a multidimensional vector to form the key temperature characteristic parameters of the hot working process. These characteristic parameters comprehensively reflect the dynamic changes and spatial distribution characteristics of the temperature field during hot working, providing a data foundation for subsequent process parameter optimization and temperature control strategy formulation.

[0057] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0058] For the smelting, casting, forging and heat treatment processes, key process parameter groups were selected, orthogonal test tables were designed, and process parameter combination experimental schemes were obtained;

[0059] Conduct hot working tests according to the process parameter combination experimental plan, collect the spatiotemporal temperature distribution information of each set of experiments, extract key temperature characteristic parameters, and obtain the process parameter-temperature characteristic response data set;

[0060] Perform variance analysis on the process parameter-temperature characteristic response data set, calculate the contribution rate and significance level of each process parameter to the temperature characteristic parameter, and obtain the main effect analysis results;

[0061] Based on the main effect analysis results, the influence weight coefficients of process parameters on temperature field are constructed, and the nonlinear relationship between process parameters and temperature characteristic parameters is fitted using polynomial regression to obtain the process-temperature influence function.

[0062] The partial derivatives of the process-temperature influence function on the process parameters are calculated to generate the Jacobian matrix and obtain the sensitivity distribution of the process parameter changes to the temperature field response;

[0063] The sensitivity distribution is normalized, and the forward mapping relationship and inverse solution relationship between the process parameters and the temperature field are constructed to generate a quantitative relationship matrix, which contains the correspondence between the process parameter adjustment amount and the temperature field change amount.

[0064] Specifically, key process parameter groups were selected for the smelting, casting, forging, and heat treatment processes. For the smelting process, key process parameters included arc power, vacuum level, bottom-blowing gas flow rate, and holding time; for the casting process, key parameters included pouring temperature, pouring speed, mold preheating temperature, and riser size; for the forging process, key parameters included heating temperature, forging deformation, forging speed, and mold temperature; and for the heat treatment process, key parameters included heating rate, holding temperature, holding time, and cooling rate. For these parameters, an orthogonal experimental design (OD) was used to arrange the experiments. Orthogonal experimental design is an efficient experimental method that uses carefully designed experimental combinations to obtain the most information with the fewest number of experiments. The specific operation involves first determining the value levels for each parameter, typically 3-5 levels, and then selecting an appropriate orthogonal array based on the number of parameters and levels. For example, for the four parameters of the heat treatment process, if each parameter has three levels, an L9 orthogonal array requires only 9 experiments, while a full factorial design would require 81 experiments. Each row of an orthogonal table represents a set of experimental solutions, each column represents a parameter, and the number in the table represents the level of that parameter in the experiment. The design of an orthogonal table ensures comprehensive testing of different parameter combinations while maintaining an economy of test times.

[0065] Hot working experiments were conducted according to a process parameter combination experimental plan, with a focus on collecting spatiotemporal temperature distribution information for each set of experiments. During the experiments, process parameters were set according to the parameter combinations specified by an orthogonal table. A multi-sensor temperature monitoring network was then used to record temperature data throughout the entire process. For smelting experiments, the surface temperature of the molten metal, the furnace temperature, and the overall temperature distribution were measured; for casting experiments, the pouring temperature, mold temperature, and the ingot cooling process temperature were measured; for forging experiments, the forging surface temperature and the heating furnace temperature field were measured; and for heat treatment experiments, the heat treatment furnace temperature distribution and the workpiece cooling curve were measured. After pre-processing, the collected temperature data was extracted to determine key temperature characteristic parameters, including the temperature change rate, temperature extreme point distribution, temperature gradient distribution, and temperature uniformity index. After each set of experiments, the set process parameter values were paired with the extracted temperature characteristic parameter values to form a process parameter-temperature characteristic response dataset. This dataset is a structured data table containing columns for process parameters and temperature characteristic parameters. Each row represents a set of experimental results, recording the temperature characteristic parameter values obtained for a specific process parameter combination.

[0066] A variance analysis was performed on the process parameter-temperature characteristic response dataset to calculate the contribution rate and significance level of each process parameter to the temperature characteristic parameter. A variance analysis is a statistical method used to determine the degree of influence of different factors on the variation in the results. Specifically, the mean value of the temperature characteristic parameter at each parameter level was first calculated to generate a parameter-response relationship table. The sum of squares due to each parameter (i.e., the sum of the squares of the differences between the mean values of each level and the overall mean) was then calculated. The total sum of squares (i.e., the sum of the squares of the differences between all experimental data and the overall mean) was then calculated. The contribution rate is equal to the parameter sum of squares divided by the total sum of squares, indicating the degree to which the parameter explains the total variation. The significance level was determined using an F-test: the F value (the ratio of the parameter mean square to the error mean square) was calculated and the P value was obtained from a table lookup. When the P value was less than 0.05, the parameter was considered to have a significant effect on the temperature characteristic parameter. The variance analysis yielded the results of the main effect analysis, clarifying the order and degree of influence of each process parameter on the temperature characteristic, providing a basis for subsequent modeling.

[0067] Based on the results of the main effect analysis, the influence weight coefficients of the process parameters on the temperature field are constructed. The influence weight coefficients reflect the relative importance of the different process parameters on the temperature field. The calculation method is to normalize the contribution rate of the parameters so that the sum of all weights is 1. After obtaining the weight coefficients, the nonlinear relationship between the process parameters and the temperature characteristic parameters is fitted using the polynomial regression method. Polynomial regression is a regression method that can describe nonlinear relationships. The basic form is that the temperature characteristic parameters are expressed as polynomial functions of the process parameters, including linear terms, quadratic terms, and interaction terms. During the fitting process, the order of the polynomial is first set, usually second or third order; then the least squares method is used to calculate the polynomial coefficients to minimize the mean square error between the predicted value and the actual value; finally, the fitting effect is evaluated by determining the coefficient and residual analysis. The polynomial equation obtained by fitting is the process-temperature influence function, which can predict the output temperature characteristic parameter value based on the input process parameter value, providing model support for temperature control.

[0068] The partial derivatives of the process-temperature influence function with respect to the process parameters are calculated to generate the Jacobian matrix. The Jacobian matrix is a matrix that describes the sensitivity of a multivariate function to its respective variables. Each element in the matrix represents the partial derivative of the dependent variable with respect to a certain independent variable. In this method, the rows of the Jacobian matrix correspond to different temperature characteristic parameters, and the columns correspond to different process parameters. The matrix elements represent the sensitivity of a specific temperature characteristic parameter to changes in a specific process parameter. The calculation method is to calculate the partial derivative of the fitted process-temperature influence function for each process parameter. The larger the partial derivative value, the more sensitive the temperature characteristic parameter is to changes in the process parameter. For example, if the partial derivative of the heat treatment holding temperature with respect to the temperature difference between the inside and outside of the workpiece is large, it means that a small change in the holding temperature will lead to a significant change in the temperature difference between the inside and outside of the workpiece, and precise control is required. By calculating the Jacobian matrix, the sensitivity distribution of the process parameter changes to the temperature field response is obtained, which provides an accurate quantitative basis for the formulation of temperature control strategies.

[0069] The sensitivity distribution is normalized to construct a forward mapping relationship and an inverse solution relationship between the process parameters and the temperature field. Normalization converts the elements in the Jacobian matrix to the same scale to facilitate comparison and calculation. A commonly used normalization method is maximum-minimum normalization, which maps element values to the range of 0-1. The calculation formula is to subtract the minimum value from the element value and divide it by the difference between the maximum and minimum values. The normalized Jacobian matrix clearly shows the relative influence of different process parameters on different temperature characteristics. Based on the normalized matrix, a forward mapping relationship is constructed between the process parameters and the temperature field. That is, given a given process parameter combination, the temperature field characteristics are predicted. At the same time, an inverse solution relationship is constructed. That is, given a target temperature field characteristic, the required process parameter combination is inferred. The inverse solution usually uses an optimization algorithm, setting the objective function as the difference between the predicted temperature field and the target temperature field. The objective function is minimized by adjusting the process parameters. The resulting quantitative relationship matrix contains the corresponding relationship between the process parameter adjustments and the temperature field changes.

[0070] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0071] Parameterize the thermal processing requirements, convert product quality indicators into temperature control targets, and obtain the target values and constraints for optimization.

[0072] Based on the quantitative relationship matrix, a temperature field state prediction formula is constructed. The temperature characteristic parameters and candidate process parameters at the current moment are input to calculate the temperature field evolution sequence in the prediction time domain to obtain the temperature response prediction result.

[0073] The error between the temperature response prediction result and the target temperature curve is calculated, and a weighted quadratic performance index evaluation method is constructed to obtain the evaluation index for process parameter optimization;

[0074] Based on the evaluation indicators, the process parameters are optimized in a rolling manner using a nonlinear programming solution. Each time, only the control amount of the first time step is executed, and then the sliding prediction window is used for re-optimization to obtain a dynamically updated process parameter sequence.

[0075] According to the process parameter sequence, the heating power, cooling rate and holding time are calculated by piecewise linear interpolation to obtain a continuous parameter control curve that meets the equipment execution accuracy;

[0076] The continuous parameter control curve is converted into a protocol and arranged in a timing sequence according to the requirements of the equipment control interface, and a standard temperature control instruction including the execution time, execution object and execution value is generated.

[0077] Specifically, parametric expression involves converting product quality indicators into quantifiable temperature control objectives, including temperature control accuracy, temperature field uniformity, heating and cooling rates, and control ranges for key temperature points. To implement this, quality requirements from the product technical standards, such as forging grain size, microstructure uniformity, and mechanical properties, are first collected. Then, drawing on materials science theory and practical experience, these quality requirements are converted into temperature control parameters. For example, for large alloy steel forgings, a grain size requirement of grades 4-6 can be translated into a heat treatment temperature control range of 880±10°C; a requirement for high microstructure uniformity can be translated into a workpiece internal and external temperature difference of no more than 50°C; and a requirement for no heat treatment cracks can be translated into a heating rate of no more than 150°C / hour and a cooling rate of no more than 80°C / hour. These specific temperature control parameters constitute the target values and constraints for the optimization solution. The target value is typically temperature control accuracy (e.g., a furnace temperature fluctuation range of ±5°C), while the constraints include upper and lower temperature limits, maximum temperature difference, and maximum heating and cooling rates. Based on the previously established quantitative relationship matrix, a temperature field state prediction formula is constructed to predict temperature field changes over future time periods. The quantitative relationship matrix represents the correspondence between process parameter changes and temperature field changes. This matrix allows calculation of the impact of specific process parameter adjustments on the temperature field. The process for constructing the temperature field state prediction formula is as follows: first, the prediction time step and prediction horizon are determined. The time step is typically 5-10 minutes, and the prediction horizon is 30-60 minutes. Then, using the current temperature characteristic parameters as the initial state, the candidate process parameter changes are substituted into the quantitative relationship matrix to calculate the temperature field change for the next time step. The temperature field change is then superimposed on the current temperature field to determine the temperature field state for the next time step. This process continues, step by step, to obtain a temperature field evolution sequence for the entire prediction horizon. The prediction calculation considers the physical characteristics of the thermal processing process, such as thermal inertia and heat transfer lag, and uses correction coefficients to adjust the prediction accuracy. The resulting temperature response prediction is a time series representing the temperature field trend over a future period under given process parameters.

[0078] The error between the temperature response prediction result and the target temperature curve is calculated, and a performance index is constructed to evaluate the quality of the process parameters. The error calculation first compares the values of the predicted temperature curve and the target temperature curve at the same time point to calculate the temperature difference; then a weighted quadratic performance index evaluation method is used, that is, the square of the temperature difference is multiplied by the weight coefficient of the corresponding time point, and the total error is obtained by summing. The weighted quadratic evaluation method emphasizes the influence of points with large deviations, and at the same time, the importance of different time points or different areas is reflected through the weight coefficient. The principle of setting the weight coefficient is as follows: high weight for key process transition points, such as the end point of heating and the start point of cooling; high weight for key parts, such as thick parts of the workpiece and hot spots; high weight for time periods with high target control accuracy requirements. Through this weighted evaluation method, a comprehensive process parameter optimization evaluation index is obtained. The smaller the index value, the closer the process parameter combination is to the optimal one.

[0079] Based on the evaluation indicators, a nonlinear programming approach is used to perform rolling optimization of process parameters. Nonlinear programming is a mathematical optimization method used to find the optimal solution to an objective function under nonlinear constraints. In this method, the objective function is the weighted quadratic evaluation indicator described above, and the constraints are the reasonable adjustment range of the process parameters. The core concept of rolling optimization is to execute only the control variables for the first time step in each control cycle, and then perform re-optimization within a sliding prediction window. Specifically, a prediction horizon (e.g., 30 minutes) and a control horizon (e.g., 5 minutes) are set. At the beginning of each control cycle, a nonlinear programming algorithm is used to calculate the optimal process parameter sequence for the entire prediction horizon. Only the parameter values for the first control horizon (5 minutes) are then executed. After this execution, the prediction window is moved, using the new temperature field state as the initial state, to recalculate the optimal parameter sequence. This rolling optimization approach can promptly respond to actual temperature changes, adapt to operating condition fluctuations and model errors, and achieve precise control of the temperature field. Through continuous rolling optimization, a dynamically updated process parameter sequence containing the optimal parameter values for each control cycle is obtained.

[0080] Based on the optimized process parameter sequence, piecewise linear interpolation is performed on key control parameters such as heating power, cooling rate, and holding time to obtain a continuous parameter control curve. Piecewise linear interpolation means that between two adjacent control points, the parameter values are assumed to change linearly and the parameter values at intermediate time points are obtained through interpolation. During implementation, the time interval between the interpolation points must meet the equipment control accuracy requirements, typically in the order of seconds or minutes. For example, for heating power, if the times of two adjacent control points are t1 and t2, and the power values are P1 and P2, then the power value P at any time t between t1 and t2 can be calculated using linear interpolation: P = P1 + (P2-P1)*(t-t1) / (t2-t1). In this way, the discrete process parameter sequence is converted into a continuous control curve, ensuring a smooth transition in temperature control and avoiding temperature fluctuations caused by sudden changes in parameters. The continuous parameter control curve is converted to a protocol and timed according to the requirements of the equipment control interface to generate standard temperature control instructions. Protocol conversion involves converting the data format of a control curve into a communication protocol format recognizable by a specific device, such as Modbus, OPC UA, or TCP / IP. Scheduling involves determining the execution order and timing of each control instruction to ensure that the instructions are executed according to the predetermined plan. Standard temperature control instructions contain three core elements: execution time (when the instruction should be executed), execution target (which device or control unit the instruction targets), and execution value (the specific set value of the control parameter). For example, a complete control instruction might be: At 10:30:00, set the power of heating zone 1 to 65kW; at 10:35:00, set the flow rate of the cooling system to 120L / min. These standardized control instructions are distributed to the control systems of each executing device via the industrial control network, enabling automated temperature control.

[0081] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0082] Match and analyze the temperature control instructions with historical execution data, extract instruction execution delay characteristics, temperature response characteristics, and device status characteristics, and obtain the instruction execution feature set;

[0083] A long short-term memory (LSTM) neural network is constructed based on the instruction execution feature set. The LSTM neural network consists of an input layer, two LSTM hidden layers, a fully connected layer, and an output layer. The input layer receives the instruction feature vector. The first hidden layer contains 128 neurons for extracting timing features, the second hidden layer contains 64 neurons for feature fusion, and the fully connected layer contains 32 neurons. The output layer generates execution timing and duration predictions.

[0084] The temperature control instructions are converted into feature vectors and input into the long short-term memory neural network. Through forward propagation calculation, the execution timing offset and execution duration prediction value of each instruction are obtained;

[0085] Perform equipment constraint verification on the execution timing offset and execution duration prediction values, and modify the prediction results based on the response characteristics and physical limitations of the hot processing equipment to obtain executable execution prediction results.

[0086] Based on the execution prediction results, the temperature control instruction sequence is re-arranged and the parameters are adjusted to eliminate instruction execution conflicts and optimize the instruction execution order to obtain preliminary revised control instructions;

[0087] The temperature target achievement degree of the initially revised control instructions is evaluated, and the corrected temperature evolution trajectory is calculated using the heat conduction model. Compensation adjustments are made for the parts that deviate from the target to obtain the final target control instructions.

[0088] Specifically, it is necessary to match and analyze temperature control instructions with historical execution data to extract three key features. The instruction execution delay feature refers to the time difference between the issuance of a control instruction and its actual execution. By time-stamping the historical temperature control instruction database, the issuance time, execution start time, and execution end time of each instruction are recorded, and the execution delay and actual duration are calculated. The time-stamping method assigns timestamps to each instruction in the database, including the instruction generation timestamp, the instruction execution start timestamp, and the instruction execution completion timestamp. The execution delay and actual duration are calculated by taking the difference between these timestamps. The temperature response feature refers to the temperature change after the execution of a control instruction. By time-aligning the historical temperature data with the control instruction, the temperature change rate before and after instruction execution, the temperature response time, and the temperature stability parameters are extracted. The time alignment method synchronizes the time series of the temperature data with the instruction execution time series to determine which instruction execution stage the temperature value at each time point corresponds to. The device status feature refers to the impact of the device operating status on instruction execution. Based on the device operation log, the device start / stop status, workload, maintenance cycle, and fault record information are extracted to construct a device status feature vector. These three types of features are mapped into a comprehensive feature association table, and then through principal component analysis and feature importance sorting, the feature subset with the highest correlation is selected to generate the instruction execution feature set.

[0089] Based on the extracted instruction execution feature set, a long short-term memory (LSTM) neural network is constructed to predict instruction execution timing and duration. LSTM neural networks are a special type of recurrent neural network that excels at processing time series data and capturing long-term dependencies. The network architecture consists of an input layer, two LSTM hidden layers, a fully connected layer, and an output layer. The input layer receives instruction feature vectors, which contain information such as instruction type, parameter size, and device status. These feature vectors are normalized and then fed into the network. The first hidden layer, consisting of 128 neurons, is primarily responsible for extracting temporal features and capturing the dependencies between instruction executions at different time points. The LSTM units control information flow through three gating structures (input gate, forget gate, and output gate), effectively memorizing long-term temporal patterns. The second hidden layer, consisting of 64 neurons, is primarily responsible for feature fusion, integrating extracted time series features with device status features. The fully connected layer, consisting of 32 neurons, performs dimensionality reduction and nonlinear transformation on the features. Finally, the output layer generates predicted execution timing offsets and execution durations. The network training adopts historical data sets, uses mean square error as the loss function, and optimizes the network parameters through the back propagation algorithm until the loss function converges.

[0090] Temperature control instructions are converted into feature vectors and fed into a long short-term memory (LSTM) neural network. A forward propagation calculation yields the execution timing offset and execution duration prediction for each instruction. Feature vector conversion encodes various instruction attributes into numerical form, including instruction type (e.g., heating power adjustment, cooling rate adjustment), parameter size, device type, and current device status. This encoding method combines one-hot encoding with numerical normalization, converting categorical variables into one-hot vectors and normalizing continuous variables to a specific range. After the feature vectors are fed into the neural network, a series of forward calculations ultimately yield predictions. The forward propagation calculation process proceeds as follows: the input layer receives the feature vectors and passes them to the first LSTM layer via a weight matrix. The first LSTM layer processes the timing information and passes its output to the second LSTM layer. The second LSTM layer performs feature fusion and passes its output to a fully connected layer. The fully connected layer performs dimensionality reduction and nonlinear transformations, ultimately generating two predictions at the output layer: the execution timing offset (the difference between the actual and planned execution times of the instruction) and the execution duration (the actual execution time of the instruction).

[0091] The execution timing offset and execution duration prediction values are checked for equipment constraints to ensure that the prediction results are consistent with the actual operating characteristics of the equipment. Equipment constraint verification is to check whether the prediction results meet the physical limitations and response characteristics of the hot processing equipment to prevent the generation of unrealistic control plans. Specific verification contents include: whether the instruction execution timing exceeds the equipment response time range, such as the power adjustment response time of the smelting furnace is not less than 30 seconds; whether the instruction execution duration is consistent with the sustainable working time of the equipment, such as continuous maximum power heating does not exceed 60 minutes; whether the instruction execution parameters are within the equipment capacity, such as the heating power does not exceed the rated power of the equipment. After verification, the prediction results that do not meet the constraints are corrected, and the execution timing offset and execution duration are adjusted to make them consistent with the equipment characteristics, so as to obtain executable execution prediction results.

[0092] Based on the execution prediction results, the temperature control instruction sequence is re-timing and parameter adjustment is performed to optimize the instruction execution order. Re-timing is to adjust the order of instruction issuance based on the predicted execution timing offset to ensure that each instruction can be executed in the expected time sequence. The specific method is to construct an instruction dependency graph, analyze the order of dependencies between instructions, and then re-order instructions without order dependencies based on the predicted execution timing to eliminate instruction execution conflicts. Parameter adjustment is to fine-tune the control parameters based on the predicted execution duration to ensure that the control effect is not affected by changes in execution time. For example, if the predicted execution delay of the heating power adjustment instruction is 5 minutes, the heat loss caused by the delay is compensated by appropriately increasing the heating power. Through re-timing and parameter adjustment, a preliminary revised control instruction sequence is obtained.

[0093] The temperature target achievement degree of the initially revised control instructions is evaluated to ensure that the revised instruction sequence can achieve the expected temperature control effect. The evaluation method is to use the heat conduction model to calculate the temperature evolution trajectory after the execution of the revised instruction sequence. The heat conduction model is based on the thermophysical parameters of the material and the finite element method, and can predict the temperature field distribution and changes during the thermal processing process. By comparing the calculated temperature trajectory with the target temperature curve, the part that deviates from the target is identified, and targeted compensation adjustments are made. The compensation adjustment adopts a feedforward correction method. According to the size and nature of the deviation, the parameter value or execution timing of the relevant instructions is adjusted to make the temperature trajectory closer to the target curve. After evaluation and compensation adjustment, the final target control instruction is formed for actual temperature control execution.

[0094] In a specific embodiment, the process of performing matching analysis on the temperature control instruction and the historical execution data may specifically include the following steps:

[0095] Perform time-stamping on the historical temperature control instruction database, record the issuance time, start execution time, and end execution time of each instruction, calculate the execution delay time and actual duration, and obtain the instruction execution timing characteristics;

[0096] Based on the instruction execution timing characteristics, a hierarchical cluster analysis is performed on the instruction type, process stage and equipment load status to obtain the distribution pattern of instruction execution delay under different conditions;

[0097] Time-align the historical temperature data with the control instructions, extract the temperature change rate, temperature response time, and temperature stability parameters before and after the instruction execution, and obtain the temperature response characteristics;

[0098] Based on the equipment operation log, the equipment start and stop status, workload, maintenance cycle and fault record information are extracted to construct the equipment status feature vector and obtain the equipment operation status characteristics;

[0099] Correlate and map the instruction execution timing characteristics, instruction execution delay distribution law, temperature response characteristics and equipment operation status characteristics, establish a multi-dimensional feature correlation table, and obtain comprehensive feature correlation results;

[0100] Perform principal component analysis and feature importance ranking on the comprehensive feature correlation results, select the feature subset with the highest correlation, and generate the instruction execution feature set.

[0101] Specifically, a database of historical temperature control instructions is time-stamped. This process involves adding a clear timestamp to each control instruction. In practice, each control instruction record is extracted from the database, including information such as the instruction ID, instruction type, parameter values, and target device. Three key time points are then added: the instruction issuance time (the moment the control instruction is generated and sent), the execution start time (the moment the device actually begins responding to the instruction), and the execution end time (the moment the device completes execution). The execution delay is calculated by calculating the difference between the execution start time and the issuance time, while the actual duration is calculated by calculating the difference between the execution end time and the execution start time. These time differences constitute the timing characteristics of instruction execution, reflecting the response characteristics of the hot processing equipment to different types of control instructions. Based on the extracted instruction execution timing characteristics, a hierarchical cluster analysis is performed to identify the distribution patterns of instruction execution delays under different conditions. Hierarchical cluster analysis is a bottom-up clustering method that first treats each sample as an independent category and then gradually merges similar categories until a termination condition is met. In this method, cluster analysis is based on three key dimensions: instruction type (such as heating power adjustment, cooling rate adjustment, and holding time setting), process stage (such as the roughing and refining stages of smelting; the pouring and solidification stages of casting; the heating and forging stages of forging; and the heating, holding, and cooling stages of heat treatment), and equipment load status (such as light, medium, and heavy load). The clustering process uses a hierarchical aggregation method. First, a distance matrix is calculated between samples (using Euclidean distance to measure similarity between samples). Then, based on the distance matrix, the closest clusters are gradually merged to form a tree-like cluster structure. The final clustering results are obtained by setting an appropriate threshold or specifying the number of clusters. The clustering results reveal the distribution of instruction execution delays under different combinations of conditions. For example, during the heating phase of the forging process, when the equipment is under heavy load, the execution delay of the heating power increase instruction is significantly higher than when the equipment is under light load.

[0102] Historical temperature data and control instructions are time-aligned to extract temperature response features. Time alignment involves matching the time series of temperature data with the time series of control instruction execution to determine which instruction execution phase the temperature value at each time point corresponds to. In specific implementation, the temperature data and control instructions are first arranged in chronological order. The temperature data is then segmented into corresponding segments based on the instruction execution time period (from the start to the end of execution). For each segment, three key features are extracted: temperature change rate (the rate of temperature change over time during instruction execution, calculated by calculating the ratio of the temperature difference between adjacent time points to the time interval), temperature response time (the time interval from the start of instruction execution to the significant temperature change, determined by detecting the inflection point of the temperature curve), and temperature stability parameter (the degree of temperature fluctuation after instruction execution, represented by calculating the standard deviation or coefficient of variation of the temperature values). These features together constitute the temperature response feature, reflecting the impact of different control instructions on temperature changes.

[0103] Extract equipment operating status features based on equipment operation logs. Equipment operation logs are data records that record changes in the equipment's operating status, including equipment start and stop times, operating parameters, maintenance records, fault alarms, and other information. Four key types of information are extracted from the logs: equipment start and stop status (whether the equipment is in operation, and the duration of the start or stop), workload (the equipment's current workload level, usually expressed as power load rate or capacity utilization), maintenance cycle (the time since the last maintenance, and the time of the next planned maintenance), and fault records (the type, frequency, and handling of recent faults). This information is structured to form an equipment status feature vector, which describes the equipment's operating status at different time points and provides a device-level basis for predicting the effectiveness of instruction execution.

[0104] The previously extracted instruction execution timing features, instruction execution delay distribution patterns, temperature response features, and equipment operation status features are associated and mapped to establish a multi-dimensional feature association table. Association mapping is to establish connections between features from different sources according to time and object relationships to form a comprehensive data structure. In specific implementation, the primary key of the association is first determined, and the instruction ID or timestamp is usually selected as the basis for the association; then, various features are matched and merged according to the primary key to ensure that each record contains complete feature information; for matching the time dimension, the time window method is used to associate the equipment status features within a specific time window with the instruction execution features. The multi-dimensional feature association table is a structured data table, where each row represents an instruction execution instance and each column represents a feature dimension, including information on multiple dimensions such as instruction type, parameter value, process stage, equipment status, execution delay, and temperature response.

[0105] Principal component analysis and feature importance ranking are performed on the comprehensive feature correlation results to generate the final instruction execution feature set. Principal component analysis is a dimensionality reduction technique that transforms the original features into a set of mutually orthogonal principal components through linear transformation, preserving the key information in the data while reducing the number of features. The specific steps are: first, the data is normalized to make features of different dimensions comparable; then, the feature covariance matrix is calculated to analyze the correlation between features; then, the eigenvalues and eigenvectors of the covariance matrix are calculated. The eigenvalues indicate the importance of the principal components. Eigenvalues are ranked based on their magnitude, and the top principal components whose cumulative contribution reaches a preset threshold (typically 85%-95%) are selected as the new feature space. Feature importance ranking uses machine learning algorithms to assess the impact of each feature on the target variable. Common methods include feature importance scoring using random forests, coefficient size using linear models, and scoring based on information gain. After ranking, the most important feature subset is selected to form the final instruction execution feature set. This feature set contains the key information required to predict the timing and duration of instruction execution, and is of moderate dimensionality and rich in information.

[0106] The above describes the temperature control method for the thermal processing process in the embodiment of the present application. The following describes the temperature control system for the thermal processing process in the embodiment of the present application. Figure 2 In one embodiment of the present application, a temperature control system for a thermal processing process includes:

[0107] The acquisition module 201 is used to collect temperature data of each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the entire thermal processing process;

[0108] A calculation module 202 is configured to calculate the temperature change rate, temperature extreme points, and temperature uniformity index during the thermal processing process based on the spatiotemporal temperature distribution information to obtain key temperature characteristic parameters;

[0109] An analysis module 203 is configured to perform an orthogonal test analysis on the process parameters based on the key temperature characteristic parameters, and establish a quantitative relationship matrix of the effects of the process parameters on the temperature field;

[0110] A generation module 204 is configured to calculate the optimal combination of heating power, cooling rate, and holding time based on the quantitative relationship matrix using a rolling horizon optimization algorithm to generate a temperature control instruction;

[0111] The prediction module 205 is used to predict the execution timing and execution duration of the temperature control instruction to obtain an execution prediction result, and to modify the control instruction according to the execution prediction result to obtain a target control instruction.

[0112] Through the collaborative cooperation of the above components, the temperature data of each process of smelting, casting, forging and heat treatment are collected through a multi-sensor network, which realizes the comprehensive acquisition of spatiotemporal temperature distribution information of the whole process of hot processing, overcoming the problem of insufficient temperature field information caused by limited temperature measurement points in traditional methods; based on the acquired spatiotemporal temperature distribution information, key temperature characteristic parameters such as temperature change rate, temperature extreme point and temperature uniformity index are calculated, which provides quantitative indicators for temperature field analysis and enhances the accuracy and comprehensiveness of temperature field analysis; through orthogonal experimental analysis, a quantitative relationship matrix of the influence of process parameters on temperature field is established, which replaces the traditional reliance on experience The temperature control is transformed into precise control based on data and models, realizing bidirectional mapping from process parameters to temperature field and from temperature field to process parameters; the rolling time domain optimization algorithm is used to calculate the optimal combination of heating power, cooling rate and holding time, and generate temperature control instructions, so that the temperature control strategy is upgraded from static optimization to dynamic optimization, which can respond to changes in working conditions in real time and improve the adaptability and accuracy of control; by predicting the execution timing and execution duration of temperature control instructions, and correcting the control instructions according to the prediction results, the problems of lag and deviation in the execution of control instructions are solved, and the execution accuracy of control instructions is significantly improved.

[0113] above Figure 2 The temperature control system for the thermal processing process in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The temperature control device for the thermal processing process in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0114] Figure 3 The diagram is a schematic structural diagram of a temperature control device for a thermal process, provided in an embodiment of the present invention. The temperature control device 300 for a thermal process may vary significantly depending on its configuration or performance. The device may include one or more central processing units (CPUs) 310 (e.g., one or more processors), memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage media 330 may be either transient or persistent storage. The program stored in the storage medium 330 may include one or more modules (not shown), each of which may include a series of instructions for operating the temperature control device 300 for a thermal process. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, executing the series of instructions stored in the storage medium 330 on the temperature control device 300 for a thermal process, thereby implementing the steps of the aforementioned method for controlling a thermal process.

[0115] The temperature control device 300 for thermal processing may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 3 The structure of the temperature control device for thermal processing shown does not constitute a limitation on the temperature control device for thermal processing provided by the present invention, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0116] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, cause the computer to execute the steps of the temperature control method for a thermal processing process.

[0117] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0118] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a temperature control device used in a thermal process (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0119] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A temperature control method for a thermal processing process, characterized in that: include: Collect temperature data from each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the entire hot working process; The temperature change rate, temperature extreme points and temperature uniformity index during the hot working process are calculated based on the spatiotemporal temperature distribution information to obtain the key temperature characteristic parameters, specifically including: performing time series segmentation processing on the spatiotemporal temperature distribution information, calculating the ratio of the temperature difference between adjacent time points to the time interval, and obtaining the temperature change rate curve of each measuring point; applying the derivative extreme value detection algorithm to the temperature change rate curve, marking the inflection point where the slope of the rate curve changes significantly, and obtaining the key temperature transition moment during the hot working process; constructing a thermal map based on the spatiotemporal temperature distribution information, using the peak search algorithm to identify the coordinates of the local highest and lowest points in the temperature field, and obtaining a temperature extreme point distribution map; calculating the temperature gradient between the extreme points on the temperature extreme point distribution map, constructing a temperature gradient vector field, and obtaining the direction and intensity of temperature change during the hot working process; calculating the standard deviation, skewness and kurtosis of the spatial temperature field at each time point based on the spatiotemporal temperature distribution information, and obtaining statistical parameters that characterize the uniformity of the temperature distribution; performing feature fusion on the temperature change rate curve, key temperature transition moment, temperature extreme point distribution map, temperature gradient vector field and statistical parameters to construct a multidimensional feature vector, and obtaining the key temperature characteristic parameters of the hot working process; Based on key temperature characteristic parameters, orthogonal experimental analysis is conducted on process parameters to establish a quantitative relationship matrix of the impact of process parameters on the temperature field. Based on the quantitative relationship matrix, a rolling time domain optimization algorithm is used to calculate the optimal combination of heating power, cooling rate, and holding time to generate temperature control instructions. The execution timing and execution duration of the temperature control instruction are predicted to obtain the execution prediction result, and the control instruction is corrected according to the execution prediction result to obtain the target control instruction, which specifically includes: matching and analyzing the temperature control instruction with the historical execution data, extracting the instruction execution delay characteristics, temperature response characteristics and equipment status characteristics, and obtaining the instruction execution feature set; constructing a long short-term memory neural network based on the instruction execution feature set, the long short-term memory neural network includes an input layer, two long short-term memory hidden layers, a fully connected layer and an output layer, the input layer receives the instruction feature vector, the first hidden layer includes 128 neurons for extracting timing features, the second hidden layer includes 64 neurons for feature fusion, the fully connected layer includes 32 neurons, and the output layer generates the execution timing and duration prediction value; converting the temperature control instruction into a feature vector, inputting the long short-term memory neural network, and performing the multi-layered ... The long-term memory neural network obtains the execution timing offset and execution duration prediction value of each instruction through forward propagation calculation; the execution timing offset and execution duration prediction value are checked for equipment constraints, and the verification content includes: whether the instruction execution timing exceeds the equipment response time range, whether the instruction execution duration conforms to the sustainable working time of the equipment, and the prediction results are corrected according to the response characteristics and physical limitations of the thermal processing equipment to obtain executable execution prediction results; based on the execution prediction results, the temperature control instruction sequence is re-arranged and the parameters are adjusted to eliminate instruction execution conflicts and optimize the instruction execution order to obtain preliminary corrected control instructions; the temperature target achievement of the preliminary corrected control instructions is evaluated, and the corrected temperature evolution trajectory is calculated using the heat conduction model. Compensation and adjustment are made for the part that deviates from the target to obtain the final target control instructions.

2. The temperature control method for a thermal processing process according to claim 1, characterized in that: The temperature data of each process of smelting, casting, forging and heat treatment are collected to obtain the spatiotemporal temperature distribution information of the whole process of hot processing, including: setting infrared thermometers, thermocouples and thermal imagers for the smelting process to measure the surface temperature of the molten metal, the temperature in the furnace and the overall temperature distribution to obtain the temperature data of the smelting process; setting infrared thermometers for the casting process to measure the pouring temperature, using thermocouple arrays embedded in the casting mold to monitor the casting temperature, using thermal imagers to track the cooling process of the ingot to obtain the temperature data of the casting process; deploying infrared thermometers and thermal imagers for the forging process to measure the surface temperature of the forging, installing multi-point thermocouple arrays to monitor the temperature field of the heating furnace, and obtaining the temperature data of the forging process. According to the data; a multi-point thermocouple array is arranged for the heat treatment process to monitor the temperature distribution of the heat treatment furnace, and an infrared thermometer and a thermal imager are used to record the cooling curve of the workpiece after leaving the furnace to obtain the temperature data of the heat treatment process; the temperature data of the smelting process, the casting process, the forging process and the heat treatment process are transmitted to the data processing system through the industrial control network, and the Kalman filter algorithm is used to eliminate the measurement error to obtain the fused temperature data; the fused temperature data is interpolated and 3D reconstructed to construct the continuous spatiotemporal temperature distribution information of the entire hot processing process. The spatiotemporal temperature distribution information includes three dimensions: temperature value, spatial coordinate and timestamp.

3. The temperature control method for a thermal processing process according to claim 1, characterized in that: Based on the key temperature characteristic parameters, an orthogonal test analysis is conducted on the process parameters to establish a quantitative relationship matrix of the influence of the process parameters on the temperature field, including: selecting key process parameter groups for smelting, casting, forging and heat treatment processes respectively, designing orthogonal test tables, and obtaining a process parameter combination experimental plan; conducting hot working experiments according to the process parameter combination experimental plan, collecting the spatiotemporal temperature distribution information of each group of experiments, extracting key temperature characteristic parameters, and obtaining a process parameter-temperature characteristic response data set; performing variance analysis on the process parameter-temperature characteristic response data set, and calculating the contribution rate and significance level of each process parameter to the temperature characteristic parameter. Obtain the main effect analysis results; based on the main effect analysis results, construct the influence weight coefficients of the process parameters on the temperature field, use polynomial regression to fit the nonlinear relationship between the process parameters and the temperature characteristic parameters, and obtain the process-temperature influence function; calculate the partial derivatives of the process-temperature influence function on the process parameters, generate the Jacobian matrix, and obtain the sensitivity distribution of the process parameter changes to the temperature field response; normalize the sensitivity distribution, construct the forward mapping relationship and inverse solution relationship between the process parameters and the temperature field, and generate a quantitative relationship matrix, which contains the correspondence between the process parameter adjustment amount and the temperature field change amount.

4. The temperature control method for a thermal processing process according to claim 1, characterized in that: Based on the quantitative relationship matrix, a rolling time domain optimization algorithm is used to calculate the optimal combination of heating power, cooling rate, and holding time to generate temperature control instructions. This includes: parameterizing the thermal processing requirements, converting product quality indicators into temperature control targets, and obtaining the target value and constraints for the optimization solution. A temperature field state prediction formula is constructed based on the quantitative relationship matrix. The current temperature characteristic parameters and candidate process parameters are input to calculate the temperature field evolution sequence within the prediction time domain to obtain the temperature response prediction result. The error between the temperature response prediction result and the target temperature curve is calculated, and a weighted quadratic performance index evaluation method is constructed to obtain the evaluation index for process parameter optimization. Based on the evaluation index, a nonlinear programming solution is used to perform rolling optimization of the process parameters, executing only the control variables of the first time step each time, and then re-optimizing within the sliding prediction window to obtain a dynamically updated process parameter sequence. Based on the process parameter sequence, piecewise linear interpolation is performed on the heating power, cooling rate, and holding time to obtain a continuous parameter control curve that meets the equipment execution accuracy. The continuous parameter control curve is converted to a protocol and time-series-scheduled according to the equipment control interface requirements to generate a standard temperature control instruction containing the execution time, execution object, and execution value.

5. The temperature control method for a thermal processing process according to claim 1, characterized in that: The temperature control instructions are matched and analyzed with historical execution data, and the instruction execution delay characteristics, temperature response characteristics and equipment status characteristics are extracted to obtain the instruction execution feature set, including: time-stamping the historical temperature control instruction database, recording the issuance time, start execution time and end execution time of each instruction, calculating the execution delay time and actual duration, and obtaining the instruction execution timing characteristics; based on the instruction execution timing characteristics, hierarchical clustering analysis is performed on the instruction type, process stage and equipment load status to obtain the instruction execution delay distribution law under different conditions; historical temperature data and control instructions are time-aligned to extract the temperature change rate, temperature response time and temperature stability parameters before and after instruction execution to obtain the temperature response characteristics; based on the equipment operation log, the equipment start and stop status, workload, maintenance cycle and fault record information are extracted to construct the equipment status feature vector to obtain the equipment operation status characteristics; the instruction execution timing characteristics, instruction execution delay distribution law, temperature response characteristics and equipment operation status characteristics are correlated and mapped, and a multidimensional feature correlation table is established to obtain the comprehensive feature correlation result; principal component analysis and feature importance ranking are performed on the comprehensive feature correlation result, and the feature subset with the highest correlation is selected to generate the instruction execution feature set.

6. A temperature control system for a thermal processing process, characterized in that: For implementing the temperature control method for a thermal processing process according to any one of claims 1 to 5, the temperature control system for the thermal processing process comprises: The acquisition module is used to collect temperature data from each process of smelting, casting, forging and heat treatment to obtain the spatiotemporal temperature distribution information of the entire thermal processing process; The calculation module is used to calculate the temperature change rate, temperature extreme point and temperature uniformity index during the thermal processing process based on the spatiotemporal temperature distribution information to obtain key temperature characteristic parameters; Analysis module, used to conduct orthogonal experimental analysis on process parameters based on key temperature characteristic parameters and establish a quantitative relationship matrix of the influence of process parameters on temperature field; A generation module is used to calculate the optimal combination of heating power, cooling rate and holding time based on the quantitative relationship matrix using a rolling time domain optimization algorithm to generate temperature control instructions; The prediction module is used to predict the execution timing and execution duration of the temperature control instruction, obtain the execution prediction result, and modify the control instruction according to the execution prediction result to obtain the target control instruction.

7. A temperature control device for a thermal processing process, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the temperature control method for a thermal processing process according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the processor is caused to execute the temperature control method for a thermal processing process according to any one of claims 1 to 5 .

Images