A method for scheduling production resources for container flooring

Abstract

The application discloses a kind of container floor production resource scheduling methods, it is related to container manufacturing technical field, including S1, through sensor array acquisition floor processing process heat conductivity coefficient data and local temperature gradient distribution data, using fourier transform method processing data extraction low frequency domain characteristics, obtain the trend of space-time uneven change of heat conductivity coefficient;S2, according to the trend of space-time uneven change, judge whether the current processing state deviates from the preset heat conduction performance threshold, if deviates then through constraint condition setting and objective function definition start linear programming algorithm, optimize material addition proportion and temperature control parameter, determine adjusted parameter combination;The container floor production resource scheduling method, significantly improve the stability of heat conduction performance, reduce resource consumption, realize the efficient and accurate control of microscale processing.

Description

Technical Field

[0001] This invention relates to the field of container manufacturing technology, and more specifically to a method for scheduling production resources for container flooring. Background Technology

[0002] In modern container manufacturing, especially for special scenarios such as cold chain transportation and high-temperature chemical transportation, the thermal conductivity of the container floor is crucial to the container's thermal management efficiency and cargo quality assurance. Existing manufacturing methods mostly employ fixed material formulations and standardized temperature control processes, lacking the ability to respond to and adjust to dynamic changes in thermal conductivity over time and space, making it difficult to meet personalized transportation needs. Particularly in complex processing environments, the thermal conductivity may exhibit uneven changes in space and time due to factors such as uneven material distribution and fluctuations in temperature control parameters, forming localized thermal bridging effects. This disrupts the consistency of the overall heat conduction path, affecting the stability of the finished container floor's thermal conductivity and its thermal response speed.

[0003] Thermal bridging typically originates from weak thermal conductivity zones or interfacial discontinuities within the microstructure, interfering with the container floor's ability to regulate local temperature differences. This is particularly prevalent during large-area installation and multi-point curing. Furthermore, traditional manufacturing systems lack the capability for real-time monitoring and feedback control of thermal conductivity. When thermal conductivity deviates from preset thresholds, it cannot be identified and corrected promptly, leading to localized performance degradation, overall thermal management failure, and increased rework probability and resource waste. In addition, under continuous production models, floor processing stability becomes a major bottleneck restricting production line efficiency and product quality. Temperature control and material flow behavior during curing are highly coupled; without a closed-loop adjustment mechanism for key parameters, batch-to-batch product performance fluctuations can easily occur, making it difficult to maintain consistent thermal conductivity. Existing production systems generally lack the ability to monitor and analyze the cumulative effects of parameter deviations, making it impossible to establish a macroscopic stability assessment system or achieve precise microscopic scheduling of processing details. Therefore, the current technical challenges are mainly reflected in two aspects: first, the lack of a thermal bridge identification and suppression mechanism based on dynamic changes in thermal conductivity; and second, the lack of processing stability control methods that combine real-time detection and intelligent scheduling. Summary of the Invention

[0004] The purpose of this invention is to provide a method for scheduling production resources for container flooring, thereby solving the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a production resource scheduling method for container floor plates, comprising: S1, collecting thermal conductivity data and local temperature gradient distribution data during the floor plate processing using a sensor array, processing the data using Fourier transform to extract low-frequency domain features, and obtaining the spatiotemporal non-uniform variation trend of thermal conductivity; S2, determining whether the current processing state deviates from a preset thermal conductivity performance threshold based on the spatiotemporal non-uniform variation trend, and if so, initiating a linear programming algorithm by setting constraints and defining an objective function to optimize the material addition ratio and temperature control parameters, and determining the adjusted parameter combination; S3, obtaining the allocation requirements for heating and cooling resources from the adjusted parameter combination, limiting the variable range and performing an iterative solution process for the allocation requirements to obtain a resource allocation sequence; S4, obtaining real-time equipment status data for the resource allocation sequence and performing optimal solution verification and parameter sensitivity analysis to determine whether the sequence matches the equipment capacity, and if not, using error information... S5. Based on the final scheduling instruction, the operating parameters of the processing equipment are updated in the feedback control loop using an integral term accumulation and differential term prediction mechanism to determine the real-time correction effect of the base plate's thermal conductivity. S6. Through the real-time correction effect, the cumulative deviation value of the performance index is obtained, and system response monitoring and stability analysis are performed to determine whether the cumulative deviation value is lower than the preset threshold, thus obtaining the stability confirmation of the processing process. S7. Optimization log data is extracted from the stability confirmation, and historical adjustment records are grouped using a resource consumption minimization and multi-objective trade-off method to determine the reference parameter template for future processing. S8. For the reference parameter template, state variable feedback and closed-loop gain optimization data are obtained to determine whether the template is suitable for processing the new batch of base plates. If it is not suitable, the linear programming algorithm is started to re-optimize the material addition ratio and temperature control parameters through constraint setting and objective function definition to obtain the updated spatiotemporal change trend.

[0006] Preferably, step S1 includes: acquiring microscopic thermal conductivity and local temperature gradient data during the base plate processing in real time using a sensor array, storing the data as a time series dataset to obtain raw heat conduction data; performing frequency domain transformation on the raw heat conduction data using Fourier transform to extract low-frequency domain features to obtain a frequency domain feature dataset, wherein frequency components with frequencies below a preset threshold in the frequency domain data obtained after Fourier transform are used as low-frequency domain features; if the amplitude of the low-frequency components in the frequency domain feature dataset exceeds the preset threshold, the low-frequency components are weighted to generate a weighted low-frequency feature set; performing an inverse Fourier transform on the weighted low-frequency feature set to reconstruct the spatiotemporal variation trend of thermal conductivity to obtain a spatiotemporal variation distribution; using a k-means clustering algorithm to divide the thermal conductivity into regions based on the spatiotemporal variation distribution to obtain thermal conductivity characteristic partitions in different regions during the processing; if there are abnormal regions in the thermal conductivity characteristic partitions, the thermal conductivity stability of the abnormal regions is determined by verifying the thermal conductivity stability of the abnormal regions; and using an interpolation algorithm to generate a continuous distribution of the spatiotemporal variation trend of thermal conductivity based on the thermal conductivity stability of the abnormal regions to obtain the final heat conduction trend.

[0007] Preferably, step S2 includes: acquiring real-time heat flux data during processing based on the adjusted parameter combination; generating a continuous heat flux distribution using data smoothing to obtain a smoothed heat flux dataset; extracting heat flux anomalies from the smoothed heat flux dataset; if the heat flux anomalies exceed a preset fluctuation threshold, generating a corrected heat flux distribution using an interpolation algorithm to obtain a corrected heat flux dataset; calculating the heat flux uniformity index of the processing area using the corrected heat flux dataset to obtain a heat flux uniformity distribution; acquiring heat conduction deviation data of the processing area based on the heat flux uniformity distribution; optimizing the power allocation of the processing equipment using a gradient descent algorithm to obtain optimized power parameters based on the heat conduction deviation data exceeding a preset deviation threshold; generating a control instruction set for the processing equipment from the optimized power parameters; predicting the heat flux change trend using time series analysis to obtain a predicted heat flux distribution; adjusting the operating parameters of the processing equipment based on the predicted heat flux distribution to generate a real-time control signal to obtain a stable processing state; collecting new thermal conductivity data from the stable processing state; processing the data using a mean filtering algorithm to obtain an updated thermal conductivity dataset.

[0008] Preferably, step S3 includes extracting scheduling instructions for heating and cooling resources from the resource allocation sequence, dividing the scheduling periods using a time series segmentation method to obtain a segmented scheduling dataset; calculating the resource utilization rate index for each period based on the segmented scheduling dataset; if the resource utilization rate index is lower than a preset threshold, adjusting the allocation of scheduling periods using a greedy algorithm to obtain an optimized scheduling period; obtaining resource consumption data for each period from the optimized scheduling period, processing the consumption data using a mean filtering method to obtain a smoothed resource consumption dataset; analyzing the resource fluctuation characteristics for each period based on the smoothed resource consumption dataset; if the fluctuation characteristics exceed a preset fluctuation threshold, generating a corrected resource consumption distribution using an interpolation method to obtain a corrected consumption dataset; extracting resource scheduling deviation data from the corrected consumption dataset, calculating the statistical characteristics of the deviation data to obtain a resource scheduling deviation distribution; adjusting the allocation ratio of heating and cooling resources based on the resource scheduling deviation distribution, generating a real-time resource control signal to obtain a stable resource scheduling state; collecting new resource allocation data from the stable resource scheduling state, processing it using a data standardization method to obtain an updated resource allocation dataset.

[0009] Preferably, step S4 includes extracting real-time device status data from the resource allocation sequence, processing the status data using a data standardization method to obtain a standardized status dataset; calculating the matching degree between the device status and the device capacity constraint based on the standardized status dataset; generating an initial error signal if the matching degree is lower than a preset threshold; optimizing the initial error signal using the least squares method to obtain an optimized error signal; calculating the adjustment gain of the resource allocation ratio using a proportional controller based on the optimized error signal to obtain adjustment gain data; updating the resource allocation sequence using the adjustment gain data to generate a temporary scheduling sequence; verifying the matching degree between the device status data and the device capacity constraint based on the temporary scheduling sequence; iteratively adjusting the gain data and updating the temporary scheduling sequence if the matching degree does not reach a preset threshold; and generating a final scheduling instruction using the finally matched temporary scheduling sequence.

[0010] Preferably, step S5 includes: parsing the processing equipment operating parameters contained in the final scheduling instruction; processing the parameter data using a data standardization method to obtain a standardized operating parameter set; calculating the current heat flux density of the processing equipment based on the standardized operating parameter set; analyzing the thermal conductivity of the base plate using a Fourier heat conduction model to obtain initial heat flux distribution data; if the initial heat flux distribution data does not match the preset heat flux uniformity threshold, generating a heat flux deviation signal; calculating the control cycle gain using a proportional-integral-derivative controller to obtain gain adjustment data; updating the processing equipment operating parameters using the gain adjustment data to generate a temporary operating parameter set; verifying the correction effect of the temporary operating parameter set on the thermal conductivity of the base plate using finite element analysis to obtain temporary heat flux distribution data; if the temporary heat flux distribution data still does not reach the preset heat flux uniformity threshold, iteratively optimizing the control cycle gain; updating the temporary operating parameter set to obtain an optimized operating parameter set; recalculating the thermal conductivity of the base plate using heat flux density analysis based on the optimized operating parameter set; verifying the thermal bridge elimination effect to obtain final heat flux distribution data; and generating the final operating parameter instruction for the processing equipment using the final heat flux distribution data to determine the real-time correction result of the thermal conductivity of the base plate.

[0011] Preferably, step S6 includes acquiring continuous monitoring data of the equipment's operating status through sensors on the processing equipment, generating performance index data using data standardization processing; calculating cumulative deviation values ​​using time series analysis based on the performance index data to obtain a deviation value sequence; if any value in the deviation value sequence exceeds a preset threshold, optimizing the performance index data using a Kalman filter algorithm to generate an optimized deviation value sequence; analyzing the system response characteristics using Fourier transform through the optimized deviation value sequence to obtain response characteristic parameters; if the response characteristic parameters do not reach a preset stability threshold, adjusting the equipment operating parameters using an adaptive control algorithm to generate a temporary operating parameter set; verifying the stability of the processing process using heat flux density analysis based on the temporary operating parameter set to obtain stability state data; and generating final operating parameter instructions for the processing equipment using the stability state data to determine the stability state of the processing process.

[0012] Preferably, step S7 includes extracting optimized log data from the stability confirmation data, removing outliers using a data cleaning method to obtain a cleaned log dataset; grouping historical adjustment records using a K-means clustering algorithm based on the cleaned log dataset to generate a categorized adjustment parameter set; calculating a trade-off between resource consumption and heat conduction uniformity using a multi-objective optimization algorithm based on the categorized adjustment parameter set to obtain an optimized pre-adjustment parameter set; smoothing parameters by using a linear interpolation method if any parameter in the optimized pre-adjustment parameter set exceeds a preset threshold to obtain a smoothed pre-adjustment parameter set; analyzing the distribution characteristics of microscale heat conduction non-uniformity using thermal conduction simulation based on the smoothed pre-adjustment parameter set to obtain heat conduction distribution data; generating a reference parameter template for the processing equipment using a data mapping method based on the heat conduction distribution data to determine the pre-adjustment scheme; and comparing the reference parameter template with historical processing data using a parameter verification method to determine the applicability of the pre-adjustment scheme.

[0013] Preferably, step S8 includes extracting feature vectors from spatiotemporal trend data, performing dimensionality reduction on the feature vectors using principal component analysis to obtain a dimensionality-reduced feature dataset; establishing a mapping model between processing parameters and spatiotemporal trends using regression analysis based on the dimensionality-reduced feature dataset to obtain a parameter mapping model; adjusting the model weights using gradient descent algorithm if the prediction error of the parameter mapping model exceeds a preset threshold to obtain an optimized parameter mapping model; and generating a processing parameter configuration set for a new batch of base plates using data interpolation based on the optimized parameter mapping model.

[0014] Preferably, step S8 further includes extracting key control parameters from the processing parameter configuration set, verifying the stability of the parameter configuration during the processing using simulation analysis methods, and obtaining stability verification data; if the fluctuation range of the stability verification data exceeds a preset threshold, adjusting the key control parameters using a constraint optimization method to obtain an adjusted processing parameter set; and generating a parameter update template for the processing equipment using a data storage method based on the adjusted processing parameter set to determine the final processing parameter template.

[0015] As can be seen from the above technical solution, the present invention has the following beneficial effects:

[0016] This method for scheduling production resources for container flooring solves the operational challenges of thermal bridging and insufficient processing stability caused by spatiotemporal variations in thermal conductivity. The invention uses a sensor array to collect thermal conductivity and temperature gradient data in real time, employs Fourier transform to extract low-frequency features, and accurately identifies the spatiotemporal non-uniformity trend of thermal conductivity. When a deviation from a preset threshold is detected, the invention utilizes a linear programming algorithm to optimize the material addition ratio and temperature control parameters, generating a resource allocation sequence. Equipment status verification and parameter sensitivity analysis ensure that the sequence matches the actual capacity. Finally, scheduling commands drive a feedback control loop, combining integral and differential mechanisms to correct processing parameters in real time and eliminate microscopic thermal bridges. After continuously monitoring cumulative deviations and performing stability analysis, the invention extracts optimization logs and constructs reference parameter templates to adapt to the processing requirements of new batches. Its technical effects include significantly improving the stability of thermal conductivity, reducing resource consumption, and achieving efficient and precise control of microscale processing. Attached Figure Description

[0017] Figure 1 This is a flowchart of the production resource scheduling method for the container floor of the present invention. Detailed Implementation

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

[0019] like Figure 1As shown, this invention provides a technical solution: a method for scheduling production resources for container flooring, comprising: S1, collecting thermal conductivity data and local temperature gradient distribution data during flooring processing using a sensor array, processing the data using Fourier transform to extract low-frequency domain features, and obtaining the spatiotemporal non-uniform variation trend of thermal conductivity; S2, determining whether the current processing state deviates from a preset thermal conductivity performance threshold based on the spatiotemporal non-uniform variation trend, and if so, initiating a linear programming algorithm through constraint setting and objective function definition to optimize the material addition ratio and temperature control parameters, and determining the adjusted parameter combination; S3, obtaining the allocation requirements for heating and cooling resources from the adjusted parameter combination, performing variable range limitation and iterative solution process for the allocation requirements, and obtaining a resource allocation sequence; S4, obtaining real-time equipment status data for the resource allocation sequence and performing optimal solution verification and parameter sensitivity analysis, determining whether the sequence matches the equipment capacity, and if not, calculating through error signals... S5. Based on the final scheduling instruction, the operating parameters of the processing equipment are updated in the feedback control loop using an integral term accumulation and differential term prediction mechanism to determine the real-time correction effect of the base plate's thermal conductivity. S6. Through the real-time correction effect, the cumulative deviation value of the performance index is obtained, and system response monitoring and stability analysis are performed to determine whether the cumulative deviation value is lower than the preset threshold, thus obtaining the stability confirmation of the processing process. S7. Optimization log data is extracted from the stability confirmation, and historical adjustment records are grouped using a resource consumption minimization and multi-objective trade-off method to determine the reference parameter template for future processing. S8. For the reference parameter template, state variable feedback and closed-loop gain optimization data are obtained to determine whether the template is suitable for processing the new batch of base plates. If it is not suitable, the linear programming algorithm is started to re-optimize the material addition ratio and temperature control parameters through constraint setting and objective function definition to obtain the updated spatiotemporal change trend.

[0020] This method, based on dynamic modeling and control optimization theory of thermodynamic thermal conduction behavior, combines multi-source sensing data with Fourier transform to analyze the heat conduction trend during base plate processing. By quantifying the spatiotemporal non-uniformity of thermal conductivity, early identification of abnormal processing states is achieved. Within a linear programming framework, a constrained optimization model is constructed using material proportions and temperature control parameters as controllable variables, allowing for timely adjustment of process parameters when deviations from preset thresholds occur. The extracted resource allocation requirements are used to generate a resource allocation sequence through a limited variable range and iterative optimization process, and dynamically adapted in conjunction with equipment status data to ensure resource allocation matches equipment capacity. A feedback control mechanism incorporates integral accumulation and differential prediction to enhance the real-time performance and stability of the adjustment. Finally, the system confirms the effectiveness of the correction based on cumulative deviation and response stability, and constructs parameter templates through historical data mining to form preset strategies suitable for future batches, achieving intelligent closed-loop scheduling.

[0021] This method significantly improves resource utilization efficiency and product consistency during container floor processing through high-precision thermal conductivity characteristic identification and scheduling optimization mechanisms. Fourier transform is used to extract low-frequency thermal conductivity trends, aiding in early detection of heat conduction anomalies and enhancing system response sensitivity. The introduction of linear programming algorithms enables global optimization of material and energy consumption control parameters, reducing energy consumption and material waste. Feedback control strategies enhance operational stability, and differential prediction mechanisms improve adjustment accuracy. Finally, parameter templates constructed through historical log analysis effectively support the continuity and adaptability of processing strategies, improving the overall line's operational efficiency and intelligence level.

[0022] S1 includes: acquiring real-time data on thermal conductivity and local temperature gradient at the microscale during the base plate processing using a sensor array, storing this data as a time-series dataset to obtain raw heat conduction data; performing frequency domain transformation on the raw heat conduction data using Fourier transform to extract low-frequency domain features, resulting in a frequency domain feature dataset, where frequency components with frequencies below a preset threshold in the frequency domain data obtained after Fourier transform are used as low-frequency domain features; if the amplitude of low-frequency components in the frequency domain feature dataset exceeds the preset threshold, the low-frequency components are weighted to generate a weighted low-frequency feature set; performing an inverse Fourier transform on the weighted low-frequency feature set to reconstruct the spatiotemporal variation trend of thermal conductivity, obtaining a spatiotemporal variation distribution; based on the spatiotemporal variation distribution, using a k-means clustering algorithm to divide the thermal conductivity into regions, obtaining thermal conductivity characteristic partitions for different areas during processing; if abnormal regions exist in the thermal conductivity characteristic partitions, verification is performed using local temperature gradient data to determine the thermal conduction stability of the abnormal regions; based on the thermal conduction stability of the abnormal regions, an interpolation algorithm is used to generate a continuous distribution of the spatiotemporal variation trend of thermal conductivity, obtaining the final heat conduction trend.

[0023] In the specific implementation of this method, firstly, in step S1, a high-density sensor array deployed in the processing area of ​​the base plate is used to collect thermal conductivity and local temperature gradient data in real time. Each sensor samples at a fixed time interval, set to 1 second, with a sampling frequency of 1 Hz, ensuring sufficient timeliness and continuity of the sampled data. The sensors are spaced 10 mm apart and evenly distributed along the processing path to ensure that the thermal conduction characteristics of the processing area are fully covered. The thermal conductivity and local temperature gradient values ​​are stored separately in the form of numerical pairs and organized in chronological order into a time-series dataset, forming the raw thermal conduction data, which serves as the basic input for subsequent analysis.

[0024] The raw heat conduction data was then subjected to frequency domain analysis using a standard Fourier transform calculation process. This involved inputting the thermal conductivity time series recorded by each sensor node into the Fourier analysis flow. During the transformation, each sensor node output a set of complex frequency domain data, where each frequency component contained corresponding amplitude and phase information. The system extracted frequency components with frequencies less than 0.1 Hz from each set of frequency domain data. This frequency threshold was determined based on the statistical average of the periodic changes in heat conduction from historical process data analysis. 0.1 Hz corresponds to a slow-changing process with a period of 10 seconds, suitable for reflecting the macroscopic fluctuation trend of thermal conductivity. The extracted low-frequency components represent the long-term thermal behavior patterns during the heat conduction process.

[0025] The amplitude of low-frequency components is assessed using an amplitude threshold, which is set to 1.5 times the average amplitude of the low-frequency components at that sensor node. For example, if the average amplitude of all low-frequency components at a node is 4, then its corresponding amplitude threshold is 6. When the amplitude of a low-frequency component exceeds this threshold, it indicates significant thermal conductivity fluctuations at that frequency. The system selects this frequency component and assigns it a weighting factor. The weighting factor is fixed at 1.2, and its purpose is to enhance the influence of this component in the reconstruction process, thereby highlighting the importance of the heat conduction behavior it represents. After all frequency components that meet the conditions have undergone weighting processing, they are then input as a whole into the inverse Fourier transform module for reconstruction from the frequency domain to the time domain.

[0026] The result of the inverse Fourier transform is a thermal conductivity numerical matrix with time as the horizontal axis and spatial location as the vertical axis, where each data point represents the thermal conductivity value at a specific sensor location at a given time. Integrating this type of data from all sensor nodes forms a three-dimensional dataset, encompassing time, space, and thermal conductivity numerical dimensions, representing the spatiotemporal distribution of thermal conductivity variations throughout the entire processing area during the processing phase.

[0027] Subsequently, the system performs clustering processing on the spatial distribution of thermal conductivity using the k-means clustering algorithm. First, the number of cluster centers, k, is set. This value is determined by the actual number of physical partitions in the base plate processing area. For example, if a base plate with a width of 500 mm and a length of 1000 mm is divided into 5 thermal characteristic regions, then k is set to 5. The k-means clustering calculation process includes: first, initializing 5 random centroids; second, calculating the Euclidean distance between each data point and the 5 centroids, and assigning it to the nearest centroid; third, updating the centroid of each category to the average value of all its data points; fourth, repeating steps two and three until the cluster centers no longer change. The final result is the cluster label to which each data point belongs, thus completing the thermal conductivity partitioning.

[0028] After completing the thermal conductivity zoning, the system evaluates the changes in thermal conductivity within each region. The evaluation method involves calculating the rate of change of the thermal conductivity of all sensor nodes in that region over three consecutive time points. This is achieved by dividing the difference between the current value and the previous value by the previous value. If this ratio exceeds 0.1 (10%) for three consecutive times, the region is considered to have an abnormal thermal conductivity. This threshold represents the upper limit of the standard error for heat conduction fluctuations during the manufacturing process, determined based on large sample data and validated through multiple process trials. If an abnormal region is found, the system proceeds to the verification phase.

[0029] During the verification phase, the system retrieves the temperature gradient changes of all sensors in the area at the corresponding time points. The calculation method is the same: subtract the previous value from the current value and divide by the previous value. If the rate of change also exceeds 10%, the area is confirmed as an abnormal thermal conductivity region. Only when both the thermal conductivity and temperature gradient meet the abnormal change criteria is the area confirmed as a region with unstable thermal conduction, ensuring the accuracy of the judgment.

[0030] For confirmed anomalous regions, the system performs interpolation to complete their spatiotemporal thermal conductivity data. The interpolation algorithm is linear interpolation, meaning that valid data from two time points before and after the data point are used as the start and end points for interpolation, and missing data points are calculated linearly based on the time position. All data points in discontinuous anomalous regions are completed using this method until complete and continuous spatiotemporal trend data of thermal conductivity are generated, forming the final heat conduction trend.

[0031] S2 includes: acquiring real-time heat flux data during processing based on the adjusted parameter combination; generating a continuous heat flux distribution using data smoothing to obtain a smoothed heat flux dataset; extracting heat flux anomalies from the smoothed heat flux dataset; if the heat flux anomalies exceed a preset fluctuation threshold, generating a corrected heat flux distribution using an interpolation algorithm to obtain a corrected heat flux dataset; calculating the heat flux uniformity index of the processing area using the corrected heat flux dataset to obtain a heat flux uniformity distribution; acquiring heat conduction deviation data of the processing area based on the heat flux uniformity distribution; optimizing the power allocation of the processing equipment using a gradient descent algorithm to obtain optimized power parameters based on the heat conduction deviation data exceeding a preset deviation threshold; generating a control instruction set for the processing equipment from the optimized power parameters; predicting the heat flux change trend using time series analysis to obtain a predicted heat flux distribution; adjusting the operating parameters of the processing equipment based on the predicted heat flux distribution to generate real-time control signals to obtain a stable processing state; and collecting new thermal conductivity data from the stable processing state, processing the data using a mean filtering algorithm to obtain an updated thermal conductivity dataset.

[0032] In the implementation of this method, the system first uses the material addition ratio and temperature control parameters optimized in the previous stage as heat input conditions to guide the heat treatment operation in the current processing cycle. Based on this, the processing equipment initiates a real-time heat flow data acquisition process. Sensors are deployed at all key heating and cooling points within the base plate processing area, with each sensor acquiring heat flow values ​​once per second. All data are recorded in timestamp order, forming a heat flow time series. Since actual heat flow signals are often accompanied by random fluctuations caused by mechanical vibration, electrical interference, and other factors, data smoothing is necessary. The data smoothing process employs a fixed-window moving average algorithm with a window length of 5. This means that data from every 5 consecutive time points is grouped together, and their arithmetic mean is calculated to replace the original value at the intermediate point, thereby eliminating short-term jumps and obtaining continuous heat flow distribution data. The processed results are then used to construct a smoothed heat flow dataset. This window length of 5 is the optimal value determined by comparing the impact of different window lengths on signal noise reduction and response speed.

[0033] In the smoothed heat flux dataset, the system compares the heat flux variation amplitude point by point to detect outliers. The outlier determination criterion is whether the difference between the current heat flux value and its previous time point is greater than a preset fluctuation threshold. This fluctuation threshold is 16, determined by statistically analyzing the range of heat flux data collected under normal operating conditions, and set to twice its standard deviation to ensure coverage of more than 95% of the normal variation range. If any heat flux data point differs from its previous time point by more than 16, it is determined to be an outlier. All outliers are individually marked and do not directly participate in subsequent calculations, but are instead used for interpolation correction.

[0034] The interpolation correction method is linear interpolation. For each outlier, the system finds its nearest two non-outlier points, denoted as the start and end points. The start point time is T1, corresponding to a heat flux value of V1, and the end point time is T2, corresponding to a heat flux value of V2. The outlier time is T. The formula for calculating the correction value is as follows: , This represents the corrected value for outliers. Through this calculation process, outliers are replaced with their expected values ​​under a linear trend, thus obtaining the corrected heat flux dataset.

[0035] Next, the system divides the corrected heat flux dataset by processing area, with each processing area set to 100 square millimeters, based on the physical deployment area of ​​the sensors. For the heat flux data collected in each area, its heat flux uniformity index is calculated. This index is defined as the standard deviation of all heat flux data in that area. The standard deviation is calculated by averaging the squares of the differences between all data points and the average heat flux value of the area, and then taking the square root. The smaller the result, the more uniform the heat distribution. The standard deviation values ​​of all areas are normalized, with the maximum value set to 1, and the remaining values ​​scaled proportionally to obtain the heat flux uniformity distribution.

[0036] Using heat flux uniformity as a basis for judgment, the system further calculates the heat conduction deviation data for each region, that is, the difference between the current heat flux uniformity value and the ideal target value for that region. The ideal target value is set to 5, which is the average value obtained by statistically analyzing the standard deviation of heat flux uniformity that yields the highest product pass rate in previous heat treatment results. The system sets the heat conduction deviation threshold to 0.5, meaning that if the absolute value of the heat flux uniformity value of a certain processing region minus 5 is greater than 0.5, then that region is determined to be a region with excessive heat conduction deviation and needs to be adjusted.

[0037] To eliminate bias, the system employs a gradient descent algorithm to optimize the power parameters of the processing equipment. The optimization objective is to minimize the sum of squares of heat flux deviations across all processing regions. In each iteration, the system calculates the heat flux distribution for each region based on the current power parameters, then calculates the deviation value for each region and calculates its sum of squares as the loss value. The partial derivative of this loss value with respect to the current power parameters is then calculated to determine the gradient direction, and the power value is updated in the opposite direction. The learning rate is set to 0.01, a balance value determined through comparing convergence speed and stability across multiple sets of experiments. The update formula is the current power value minus the learning rate multiplied by the gradient. The updated power is used for the next round of heat flux calculation and bias evaluation. This process is repeated until the difference between the loss values ​​of two consecutive rounds is less than 0.001, at which point convergence is considered achieved, and the current power parameter is output as the optimal power.

[0038] Based on the optimized power parameters, the system generates a control instruction set for each processing device. Each control instruction includes the device number, target power value, power adjustment start time, and duration. All instructions are sorted by time to form a time series input to the heat flow prediction module for predicting the heat flow trend in the next stage. The prediction method uses heat flow data from the past 30 seconds as a sample, employing a first-order slope calculation method. The difference between the average heat flow in the last 10 seconds and the average heat flow in the first 10 seconds is divided by the time interval to obtain the heat flow slope. The predicted heat flow distribution for the next 10 seconds is calculated by adding the current average heat flow to the slope multiplied by the prediction time. The formula for calculating the heat flow slope is: ,in, Indicates the slope of heat flow. This represents the average heat flow data over the past 30 seconds, specifically the first 10 seconds. This represents the mean of heat flow data in the last 10 seconds of the past 30 seconds. The predicted heat flow distribution for the next 10 seconds is calculated by adding the current mean heat flow to the slope and multiplying by the prediction time. The specific formula is as follows: ,in, This represents the predicted heat flux value at time t+Δt, where Δt represents the prediction duration, Δt∈(0,10] seconds, and t represents time. Indicates the slope of heat flow. This represents the average heat flow data over the last 10 seconds of the past 30 seconds.

[0039] The predicted heat flow distribution is used to adjust equipment operating parameters, including the heating power of the heater, the flow rate of the cooler, and the operating time of each piece of equipment. Control signals are adjusted by aligning them with the predicted heat flow trend. These control signals, including power magnitude, command execution time, and duration, are sent to the equipment control terminal in real time to ensure that actual heat flow changes match the predicted trend, thereby achieving a stable processing state.

[0040] Once the equipment reaches a stable state, the system initiates the next round of thermal conductivity data acquisition, collecting thermal conductivity and thermal diffusivity data every second. The collected data is processed by a mean filtering algorithm, which calculates the average value for every three time points to replace the intermediate value, thereby removing short-term noise and forming an updated thermal conductivity dataset, which serves as the basic data input for the next round of scheduling optimization.

[0041] S3 includes extracting scheduling instructions for heating and cooling resources from the resource allocation sequence, dividing the scheduling periods using a time series segmentation method to obtain a segmented scheduling dataset; calculating the resource utilization rate index for each period based on the segmented scheduling dataset; if the resource utilization rate index is lower than a preset threshold, adjusting the scheduling period allocation using a greedy algorithm to obtain an optimized scheduling period; obtaining resource consumption data for each period from the optimized scheduling period, processing the consumption data using a mean filtering method to obtain a smoothed resource consumption dataset; analyzing the resource fluctuation characteristics of each period based on the smoothed resource consumption dataset; if the fluctuation characteristics exceed a preset fluctuation threshold, generating a corrected resource consumption distribution using an interpolation method to obtain a corrected consumption dataset; extracting resource scheduling deviation data from the corrected consumption dataset, calculating the statistical characteristics of the deviation data to obtain a resource scheduling deviation distribution; adjusting the allocation ratio of heating and cooling resources based on the resource scheduling deviation distribution, generating a real-time resource control signal to obtain a stable resource scheduling state; collecting new resource allocation data from the stable resource scheduling state, processing it using a data standardization method to obtain an updated resource allocation dataset.

[0042] In this step, the system first extracts all scheduling instructions involving heating and cooling resources from the resource allocation sequence formed in the previous stage. Each scheduling instruction includes five pieces of information: resource category, target equipment number, resource start time, resource end time, and resource usage per unit time. The system sorts all scheduling instructions in chronological order and divides the total scheduling time into several equal-length scheduling periods, each with a fixed length of 10 seconds. This time length is determined based on the response speed of the processing equipment and the scheduling control cycle, ensuring that resource usage, status monitoring, and feedback adjustment are completed within each period. The system maps each scheduling instruction to its corresponding scheduling period according to the time range it covers, and statistically analyzes the usage time, frequency, and amount of various resources within each scheduling period, ultimately generating a segmented scheduling dataset. Each data segment contains all resource usage events and their parameter information within that period.

[0043] Subsequently, the system calculates the resource utilization rate index based on the resource usage data for each time period. The utilization rate is defined as the cumulative time that the equipment actually uses resources within that time period divided by the total duration of that time period. For example, in a 10-second time period, if a heater is actually powered on for 6 seconds, its utilization rate is 0.6. The threshold for the resource utilization rate index is set at 0.6, based on the statistical results of resource utilization efficiency and energy consumption ratio in each equipment operation. This value ensures a balance between economic efficiency and resource use. When the resource utilization rate for a certain time period is lower than 0.6, the system initiates a greedy algorithm to optimize the resource allocation order within that time period. The greedy algorithm selects the task with the highest resource consumption efficiency per unit time from all tasks to be scheduled within that time period, prioritizes it for the current time period, and sequentially fills the time period with tasks until the total resource usage reaches the capacity limit of that time period or all tasks have been scheduled. After execution, an optimized scheduling schedule is output.

[0044] Next, the system re-extracts the actual consumption data of heating and cooling resources for each scheduling period based on the optimized scheduling schedule, and constructs a complete resource consumption time series in chronological order. Due to potential instantaneous peaks or noise interference during data acquisition, the system applies a sliding window mean filtering algorithm to smooth the resource consumption time series. Specifically, a sliding window is formed using five adjacent time points, and the average resource consumption value within that window is calculated and assigned to the middle time point of the window. The window length of 5 is set based on the relationship between the time constant of heat load change and the sensor sampling period, effectively eliminating high-frequency interference while preserving the overall trend. After processing, a smoothed resource consumption dataset is formed.

[0045] Resource fluctuation characteristics are analyzed on a smoothed dataset. Fluctuation is determined by whether the rate of change in resource consumption between two adjacent time points exceeds a set threshold. The rate of change is calculated as the current resource consumption value minus the previous time point's consumption value, then divided by the previous time point's consumption value. The fluctuation threshold is set at 10%, meaning that if the absolute value of the rate of change exceeds 0.1, the time period is considered to have abnormal resource fluctuations. This threshold is determined based on the standard error of resource flow stability under normal operating conditions; exceeding this value may lead to uneven heat treatment or control response misalignment. For time periods determined to have abnormal fluctuations, the system uses linear interpolation for correction. The correction method involves selecting a stable data point before and after the fluctuation segment as the interpolation boundary, and generating intermediate point resource consumption values ​​proportionally according to the time interval, thus obtaining a continuous and stable corrected resource consumption dataset.

[0046] Then, the system extracts resource scheduling deviation data for each time period from the corrected resource consumption dataset. The deviation value is defined as the actual resource consumption during that time period minus the target resource consumption in the original scheduling plan. The deviation values ​​for all time periods form a deviation sequence. The system performs statistical processing on this sequence, extracting four feature indicators: maximum deviation, minimum deviation, average deviation, and standard deviation. Based on these, a resource scheduling deviation distribution is constructed to assess the accuracy and balance of the current resource scheduling. If the standard deviation exceeds 0.2, it indicates a severe concentration of deviations. The threshold of 0.2 is a critical value determined through the analysis of the impact on thermal uniformity; exceeding this threshold will have a substantial impact on the thermal field uniformity of the base plate.

[0047] When deviations exceed limits, the system optimizes the allocation ratio of heating and cooling resources. The adjustment strategy, based on a constant total resource consumption, reduces resources during periods of high deviation and low efficiency, and increases resources during periods of low deviation and high efficiency. The adjustment range is proportional to the deviation value of each period relative to the total deviation. The optimized resource allocation scheme is converted into control signals, which include resource type, target output value, start time, and duration. The system adjusts the equipment operating status in real time based on these control signals. After real-time control is executed, the system continuously monitors changes in resource consumption to determine whether a stable resource scheduling state has been reached. The criteria for this determination are a resource fluctuation rate below 5% for five consecutive periods and a scheduling deviation standard deviation consistently below 0.2.

[0048] After reaching a stable scheduling state, the system collects all resource allocation data from the current round as the resource allocation data for the new round. This data includes three types of information for each device in each time period: resource type, total resource consumption, and average consumption per unit time, forming a complete resource allocation dataset. To facilitate subsequent model identification and cross-period comparison, the system performs standardization processing on this dataset. The standardization method subtracts the mean of each field from the entire dataset and then divides by the standard deviation of that field, transforming it into standard normal distribution data with a mean of 0 and a variance of 1. This ultimately forms the updated resource allocation dataset for subsequent iterative optimization or model prediction.

[0049] S4 includes extracting real-time device status data from the resource allocation sequence, processing the status data using a data standardization method to obtain a standardized status dataset; calculating the matching degree between the device status and the device capacity constraint based on the standardized status dataset, and generating an initial error signal if the matching degree is lower than a preset threshold; optimizing the initial error signal using the least squares method to obtain an optimized error signal; calculating the adjustment gain of the resource allocation ratio using a proportional controller based on the optimized error signal to obtain adjustment gain data; updating the resource allocation sequence using the adjustment gain data to generate a temporary scheduling sequence; verifying the matching degree between the device status data and the device capacity constraint based on the temporary scheduling sequence, and iteratively adjusting the gain data and updating the temporary scheduling sequence if the matching degree does not reach a preset threshold; and generating a final scheduling instruction based on the finally matched temporary scheduling sequence.

[0050] In the specific implementation of this step, the system first extracts real-time operating status data of all devices currently executing resource scheduling tasks from the current resource allocation sequence. The status data for each device includes five parameters: current power output value, actual resource consumption per unit time, instantaneous load rate, response time latency, and current internal temperature change rate. Each parameter is collected in real-time by the device's built-in status monitoring module at a sampling period of 1 second and stored in the device's status cache. To ensure data comparability and dimensional consistency in subsequent calculations, the system applies standardization processing to these five parameters. The processing method involves subtracting the historical mean from each raw data point and then dividing by its historical standard deviation to obtain the standardized value. The historical mean and standard deviation are calculated using sampled data of similar parameters from the device's past 100 normal processing cycles. After standardization, the standardized status data of all devices constitute a standardized status dataset in a unified format, which is then used by the matching degree calculation module.

[0051] Matching degree is calculated for each device based on the standardized state dataset. Matching degree is defined as the degree of consistency between the device's current state and its allowable capacity. Specifically, it is calculated by dividing the standardized current state value by the standardized maximum device capacity value, multiplying the result by 100, and converting it to a percentage. The maximum device capacity value is the continuous operating limit parameter specified by the device manufacturer and is also standardized. For example, if a device's current standardized power is 0.65 and its standardized capacity is 1, then the matching degree for that device is 65%. The system presets a lower limit threshold for matching degree of 80%. This value is determined by analyzing the minimum effective load data under stable operating conditions of the device throughout history. When the matching degree falls below this value, the system considers the device to be in a resource scheduling mismatch state.

[0052] When the matching degree of any device is below 80%, the system records the difference between its target matching degree and the actual matching degree as an initial error signal. For example, if the target matching degree is 80% and the actual matching degree is 65%, the error signal is 15%. The error signals of all devices are combined to form an initial error signal vector. The system then calls a least squares optimization algorithm to optimize this error signal vector. The optimization process is as follows: First, the error signal of each device is squared, and then all squared values ​​are summed to obtain the total sum of squared errors. Second, the system introduces an error weight adjustment factor and iteratively updates the error weight of each device, aiming to minimize the total sum of squared errors. The final output is an optimized error signal vector, where each value represents the recommendation strength for the current resource allocation adjustment.

[0053] The optimized error signal is input to the proportional controller for gain calculation. The proportional controller employs a proportional adjustment strategy, calculating the resource adjustment gain value for each device based on the product of each optimized error signal value and a proportional coefficient. The proportional coefficient is uniformly set to 0.5, a value determined through joint evaluation experiments of system adjustment sensitivity and stability, ensuring that the resource allocation response does not oscillate and quickly approaches the target state. For example, if a device's optimized error signal is 0.2, multiplying it by the proportional coefficient 0.5 yields a gain value of 0.1, indicating that its resource allocation needs to be increased by 10%. The system superimposes the adjustment gain values ​​of all devices into the original resource allocation ratio, generating a new resource allocation scheme and constructing a temporary scheduling sequence. Each record in the temporary scheduling sequence includes the device number, adjusted resource allocation ratio, target resource output value, scheduling start time, and duration.

[0054] Next, the system re-executes device status data acquisition based on the temporary scheduling sequence, and repeats the standardization process and matching degree calculation to verify whether the adjusted resource allocation ensures that the matching degree of all devices is higher than 80%. If the verification result shows that the matching degree of all devices reaches or exceeds 80%, the temporary scheduling sequence is deemed valid, and the system stops iterating. If there are still devices with a matching degree below 80%, the system enters the iterative optimization phase. The iteration method is to dynamically adjust the proportional coefficient value based on the current proportional coefficient and the trend of the error signal change in the previous round. The specific adjustment range is set to 0.1. In unstable trends, the proportional coefficient is reduced to avoid over-adjustment, and in steady-state trends, the proportional coefficient is appropriately increased to accelerate convergence. The new proportional coefficient is used to recalculate the gain value and generate a new round of temporary scheduling sequence, continuing the status verification process. This iteration process is set to a maximum of 10 execution rounds. If the matching degree threshold cannot be reached within 10 rounds, the system marks it as a resource scheduling anomaly and enters the manual intervention process.

[0055] Once the system confirms that the matching degree of all devices reaches or exceeds 80%, meaning the matching meets the set conditions, the system transforms the temporary scheduling sequence generated in the current round into a final scheduling instruction set. The final scheduling instruction includes the device number, target resource type (e.g., heating or cooling), target output intensity (e.g., kilowatts or liters per second), resource adjustment gain value, instruction issuance time, and task duration. The instructions are packaged in a standard communication format and transmitted to the corresponding device control terminal via fieldbus or real-time Ethernet. The device execution module adjusts its resource allocation strategy according to the received instructions, achieving closed-loop adaptive control of the scheduling logic. This entire process achieves dynamic matching of resource allocation and device capabilities, ensuring the continuous and stable operation of the container floor processing system.

[0056] S5 includes: parsing the processing equipment operating parameters contained in the final scheduling command; processing the parameter data using a data standardization method to obtain a standardized operating parameter set; calculating the current heat flux density of the processing equipment based on the standardized operating parameter set; analyzing the thermal conductivity of the base plate using a Fourier heat conduction model to obtain initial heat flux distribution data; if the initial heat flux distribution data does not match the preset heat flux uniformity threshold, generating a heat flux deviation signal; calculating the control cycle gain using a proportional-integral-derivative controller to obtain gain adjustment data; updating the processing equipment operating parameters using the gain adjustment data to generate a temporary operating parameter set; verifying the correction effect of the temporary operating parameter set on the thermal conductivity of the base plate using finite element analysis to obtain temporary heat flux distribution data; if the temporary heat flux distribution data still does not reach the preset heat flux uniformity threshold, iteratively optimizing the control cycle gain; updating the temporary operating parameter set to obtain an optimized operating parameter set; recalculating the thermal conductivity of the base plate using heat flux density analysis based on the optimized operating parameter set; verifying the thermal bridge elimination effect to obtain final heat flux distribution data; and generating the final operating parameter command for the processing equipment using the final heat flux distribution data to determine the real-time correction result of the thermal conductivity of the base plate.

[0057] In the specific implementation of this step, the system first receives the final scheduling instruction generated from the previous process and parses each of the processing equipment operating parameters contained in the instruction. Each set of operating parameters includes six items: equipment number, target heating power, target cooling flow rate, action time length, action area number, and control response delay time. After parsing, the system performs standardization processing on the above six operating parameters in sequence. The standardization calculation adopts a fixed format, that is, each operating parameter value is subtracted from its historical mean and then divided by its standard deviation. The historical mean and standard deviation are obtained by statistically analyzing the data of corresponding parameters of similar equipment in the most recent 100 batches of processing tasks, ensuring that the standardization results are representative and stable. The standardized operating parameters are uniformly integrated into a standardized operating parameter set, which serves as the basic input for subsequent calculations.

[0058] Based on a standardized set of operating parameters, the system calculates the heat flux density per unit area for each device, taking into account its actual physical control area. The heat flux density is calculated by multiplying the standardized heating power value by the device efficiency factor and then dividing by the area of ​​the control area, expressed in watts per square meter (W / m²). The device efficiency factor is the energy conversion efficiency coefficient obtained through the device's factory calibration, ranging from 0.85 to 0.95, with the specific value depending on the device model. The area is provided by the design drawings and registered in the system. For example, if the device's target power is 800 watts, the efficiency factor is 0.9, and the control area is 0.2 square meters, then the corresponding heat flux density is 3600 W / m².

[0059] After calculation, the system uses heat flux density as an input variable, combined with the thermal conductivity, thickness, and processing boundary temperature of the base plate material, and inputs it into the Fourier heat conduction model. The Fourier heat conduction model employs a steady-state two-dimensional conduction framework, and, considering different heat sources and boundary conditions, performs mesh generation and heat transfer path solving for the entire base plate area, generating an initial heat flux distribution map under the current state. The system uses this map as the basis for the base plate's thermal energy distribution, and statistically analyzes the mean, extreme values, and variance of heat flux density in all mesh regions. These data are then compared with a heat flux uniformity threshold to determine whether the heat conduction is uniform under the current state.

[0060] The heat flux uniformity threshold is set at ±15%, meaning the criterion is whether the difference between the heat flux density of each grid cell and the average heat flux density of the entire area is within ±15% of the average. This threshold is determined by statistically analyzing the heat treatment uniformity distribution range of historically qualified base plate products to ensure a balance between cost control and ensuring heat treatment consistency. If any cell in the test results exceeds this fluctuation threshold, the system determines that the current operating parameters have failed to meet the target thermal field consistency requirements and correction is required.

[0061] The first step in the calibration process is to calculate the deviation signal. The system defines the percentage difference between the actual heat flux density and the target heat flux density of each defective grid cell as the heat flux deviation signal. This signal is stored in vector form, indicating the overheating or undercooling state of different regions. The system then calls a proportional-integral-derivative (PID) controller to process these deviation signals, sequentially calculating the proportional, integral, and derivative terms. The proportional term is the current deviation signal value multiplied by a proportional coefficient, which is fixed at 0.5. This value was determined through response curve adjustment experiments to ensure a moderate adjustment speed that does not cause system oscillations. The integral term is calculated by multiplying the sum of the deviation signals at the current time point and the previous 10 time points by an integral coefficient, which is set to 0.05. The integration period is 10 seconds to ensure that accumulated deviations are not ignored. The derivative term is calculated by subtracting the deviation signal at the previous time point from the current deviation signal and multiplying by a derivative coefficient, which is set to 0.1 to offset the inertial effect caused by the deviation change trend. The three terms are added together to form the control loop gain, which serves as the basis for adjustment.

[0062] The control gain is applied to the power, flow rate, and time parameters in the operating parameter set to form a new temporary operating parameter set. This parameter set is input to the finite element analysis module, which performs fine mesh generation on the base plate, with each element having an area of ​​100 square millimeters. Based on the new operating parameters, the heat source intensity and boundary conditions of each mesh element are reset. The system performs heat conduction path analysis on each element, simulates the heat diffusion process, and generates an updated temporary heat flux distribution map. The system again verifies the uniformity of heat flux density in all elements using ±15% as a judgment threshold. If the verification still does not meet the conditions, controller parameter adjustment and iteration are performed. The iteration method is as follows: if the maximum deviation signal in the previous round does not decrease significantly, the proportional coefficient is reduced by 0.05; if it decreases significantly, the integral coefficient is increased by 0.01 to accelerate convergence. The system iterates for a maximum of 10 rounds or automatically terminates when the heat flux density fluctuation in all elements is controlled within ±10%.

[0063] Finally, based on the optimized set of operating parameters, the system re-executes the heat flux density simulation and evaluates the thermal bridge elimination effect. The criterion is that the difference in heat flux density between all adjacent grid cells must not exceed 20%. 20% is the critical value for thermal bridge generation detected by thermal imaging in the base plate structure design; exceeding this value may lead to localized thermal stress concentration. If all areas meet the requirements, the system takes the operating parameters of this round as the final optimization result. The system then converts this into control signals to generate the final operating parameter instructions for the processing equipment, including the equipment number, final thermal power output value, cooling flow rate, duration, target area number, and response delay compensation time. All parameters are packaged into a standardized control instruction format and sent to the equipment control module in real time via the industrial communication bus to achieve real-time closed-loop correction of the base plate's thermal conductivity, ensuring consistent, stable, and efficient heat flux in all areas during processing.

[0064] S6 includes acquiring continuous monitoring data of the equipment's operating status through sensors, and generating performance index data using data standardization processing; calculating the cumulative deviation value using time series analysis based on the performance index data to obtain a deviation value sequence; if any value in the deviation value sequence exceeds a preset threshold, optimizing the performance index data using a Kalman filter algorithm to generate an optimized deviation value sequence; analyzing the system response characteristics using Fourier transform based on the optimized deviation value sequence to obtain response characteristic parameters; if the response characteristic parameters do not reach a preset stability threshold, adjusting the equipment operating parameters using an adaptive control algorithm to generate a temporary operating parameter set; verifying the stability of the processing process using heat flux density analysis based on the temporary operating parameter set to obtain stability state data; and generating the final operating parameter command for the processing equipment using the stability state data to determine the stability state of the processing process.

[0065] In the specific implementation of this step, the system first collects equipment operating status data through multi-channel real-time sensors deployed on each processing device. The collected data includes six types of parameters: equipment thermal power output, current change rate, cooling medium flow rate, working area temperature change rate, operating response time, and vibration amplitude. Each type of data is continuously sampled at a 1-second interval to form continuous monitoring data. The system inputs the above collected data as the raw operating status dataset into the data processing module. To eliminate dimensional and statistical distribution differences between different physical quantities, the system first performs standardization processing on the dataset. The standardization method is to subtract the mean of each item from the past 100 batches of the raw data and then divide by the standard deviation of that item. All means and standard deviations are derived from confirmed normal and stable operating condition records in the database, ensuring that the standardized data has strong statistical representativeness and comparative value. After standardization, a performance index dataset is generated, which serves as the basic data source for subsequent deviation analysis.

[0066] Next, the system performs time series analysis on the performance index dataset. The time series analysis steps include calculating the difference between the standardized index data at each time point and the historical reference value to obtain the instantaneous deviation value. Then, using a 30-second sliding window, the instantaneous deviation values ​​within that time period are summed point by point to generate the cumulative deviation value for that window, thus forming a deviation value sequence. The reference value is obtained by statistically analyzing the median value of the corresponding equipment index in all qualified processing batches. The focus of the cumulative deviation value analysis is to determine whether it exceeds the system's set stability tolerance. The system sets a deviation threshold of 5%, meaning that if the ratio of the cumulative deviation value to the reference value within any time window exceeds 5%, the equipment's operating status is considered unstable during that time period. This 5% setting is based on the standard deviation range of thousands of stable operating sample data collected in previous batch processing tasks, and its setting balances detection sensitivity and false alarm control.

[0067] If the system determines that a deviation value exceeds the limit, it initiates the Kalman filter algorithm to further optimize the performance index data. The Kalman filter consists of a prediction step and an update step. In the prediction step, the system estimates the current state based on the performance state at the previous time step and predicts its covariance. In the update step, the system collects the actual measured values ​​at the current time step and calculates a weighted average between the predicted and observed values ​​using the Kalman gain to obtain a more accurate state estimate. The Kalman gain is calculated using the ratio of the prediction covariance to the measurement error covariance, which is estimated from historical sampling errors. The final filtering result generates a new sequence of optimized deviation values.

[0068] Subsequently, the system performs Fourier transform analysis on the optimized deviation value sequence to obtain system response characteristic parameters. The Fourier transform process consists of three steps: signal preprocessing, discrete Fourier transform execution, and spectrum extraction. The system focuses on extracting two key indicators: the dominant frequency and the dominant frequency phase delay. The dominant frequency is the frequency value corresponding to the maximum amplitude in the spectrum; the phase delay is the phase difference of the dominant frequency signal relative to the ideal response. The system presets a dominant frequency threshold of 0.05 Hz, which indicates a slow system response; the phase delay threshold is set to 15 degrees, which indicates significant hysteresis in the control response. Both thresholds are determined based on long-term statistical results of frequency domain characteristics during stable processing and have clear engineering application significance.

[0069] If the Fourier analysis results show that the system response characteristics do not meet the above threshold requirements, the system will activate the adaptive control algorithm. This algorithm adjusts the equipment operating parameters based on the current deviation trend and frequency domain response, including heating power, cooling flow rate, and control response time. The system uses 20% of the current deviation value as the adjustment benchmark and increases or decreases these three parameters accordingly. For example, if the current main frequency is low and the deviation continues to increase positively, the system increases the heating power by 10% and shortens the control response time by 5%; if the phase delay exceeds the set limit, the control response delay time is reduced by 5%. All adjustments form a new set of temporary operating parameters.

[0070] The heat flux density distribution in the processing area is calculated using a temporary set of operating parameters. First, the processing area is divided into several equal-area grid cells, each with an area of ​​100 square millimeters. Then, the heat flux density value per unit time for each cell is recalculated based on the adjusted heating and cooling parameters. The system analyzes the maximum and minimum differences in heat flux density across the entire area to determine if the thermal field stability condition is met. The system sets a heat flux density fluctuation threshold of ±10%, meaning that if the deviation of the heat flux density of any cell from the global average does not exceed 10%, it is considered to have reached a uniform heat flux state.

[0071] If the system confirms that the thermal stability condition has been met, it encapsulates the current set of operating parameters into a final operating parameter instruction. This instruction contains six items: equipment number, final thermal power value, cooling flow rate, response time adjustment, target area number, and duration of action. The instruction is formatted by the system according to the industrial communication protocol and sent to the equipment control port as the officially executed control command. Ultimately, this achieves stable thermal field processing control based on multi-source sensor feedback, frequency domain response analysis, and adaptive regulation, ensuring that the container floor maintains uniform and reliable thermal conductivity during high-intensity continuous processing tasks.

[0072] S7 includes extracting optimized log data from stability confirmation data, removing outliers using data cleaning methods to obtain a cleaned log dataset; grouping historical adjustment records using K-means clustering algorithm based on the cleaned log dataset to generate a categorized adjustment parameter set; calculating the trade-off between resource consumption and heat conduction uniformity using a multi-objective optimization algorithm based on the categorized adjustment parameter set to obtain an optimized pre-adjustment parameter set; smoothing parameters by linear interpolation if any parameter in the optimized pre-adjustment parameter set exceeds a preset threshold to obtain a smoothed pre-adjustment parameter set; analyzing the distribution characteristics of microscale heat conduction non-uniformity using thermal conduction simulation based on the smoothed pre-adjustment parameter set to obtain heat conduction distribution data; generating a reference parameter template for the processing equipment using data mapping method based on the heat conduction distribution data to determine the pre-adjustment scheme; and comparing the reference parameter template with historical processing data using parameter verification methods to determine the applicability of the pre-adjustment scheme.

[0073] In implementing this step, the system first extracts optimization log data from the stability confirmation data. This data includes the operating parameters and performance indicators of each device under stable conditions during processing, specifically including six key data items: thermal power setting value, cooling system flow rate, response time adjustment value, thermal conductivity, heat flux density, and processing completion time. The system removes outlier data through a data cleaning process using a box plot method. The upper quartile, lower quartile, and interquartile range are calculated for each parameter. Values ​​exceeding the upper quartile plus 1.5 times the interquartile range or falling below the lower quartile minus 1.5 times the interquartile range are considered outliers and removed. The data after outlier removal constitutes the cleaned log dataset, providing high-quality input for subsequent clustering and modeling.

[0074] The cleaned log dataset was then standardized to convert all parameters into a standard normal distribution with a mean of zero and a standard deviation of one, eliminating differences between different dimensions and units. Next, the K-means clustering algorithm was used to divide all historical adjustment records into several classes to extract typical adjustment patterns. The value of K was set to 5, selected through repeated trials within the range of 1 to 10 using the silhouette coefficient evaluation method, maximizing inter-class differences and minimizing intra-class differences. Each class represents a typical combination of scheduling states or processing conditions, forming a set of classification adjustment parameters for subsequent optimization calculations.

[0075] Based on the categorized parameter set, the system performs multi-objective optimization, aiming to simultaneously minimize resource consumption and maximize heat conduction uniformity during processing. Resource consumption is calculated as the sum of heating and cooling system energy consumption, multiplied by the equipment's rated power, processing time, and cooling flow rate, respectively, multiplied by the fluid's heat capacity and temperature difference. Heat conduction uniformity is defined as the reciprocal of the standard deviation of the heat flux density in each cell of the processing base plate; a smaller standard deviation indicates higher uniformity. The system uses a linear weighting function to combine the two objectives, with resource consumption weighted at 0.6 and heat conduction uniformity weighted at 0.4. This weighting is based on the importance ranking of energy saving and product consistency requirements for processing enterprises. The final output is the optimized pre-adjustment parameter set for each typical scheduling scenario.

[0076] If any parameter in the optimization results is found to exceed the safe operating range of the equipment, such as thermal power exceeding 5000 watts, cooling flow rate exceeding 50 liters per minute, or response delay exceeding 2 seconds, the system will perform linear interpolation on these out-of-limit parameters to ensure parameter continuity and smoothness. The interpolation process uses two adjacent valid parameter points as endpoints to calculate the linear corresponding value of the current out-of-limit parameter within the interval, thereby correcting it to a smooth pre-adjusted parameter set within a reasonable range to avoid system oscillation or overload caused by parameter abrupt changes.

[0077] Subsequently, the system inputs the smoothed pre-adjusted parameter set into the thermal simulation module for heat conduction distribution simulation analysis. This simulation employs a heat conduction model based on the finite element method, dividing the container floor into equilateral mesh elements with a side length of 1 mm. Corresponding heat inputs and boundary conditions are applied to each element, and the heat conduction control equations are solved to obtain the heat flux density values ​​of each mesh element, ultimately generating heat conduction distribution data. This heat conduction distribution data establishes a one-to-one correspondence between the action area of ​​each parameter and the control area of ​​the processing equipment through mapping logic, forming a reference parameter template for the processing equipment. Each template contains 6 parameter items and their corresponding equipment numbers and control area locations.

[0078] After the reference parameter template is constructed, its applicability is verified. The verification method involves comparing the parameters in the template with the actual parameters of 50 historical batches of processing tasks. The comparison indicators include: thermal power setpoint, cooling flow rate, control response time, final processing time, and heat flux uniformity. If at least 30 batches of matched tasks meet the requirements that all parameter deviations are less than 10% and the standard deviation of heat flux density distribution is less than 5%, the system determines that the template has good applicability and can be directly called as a preset parameter template for future batches of processing tasks, realizing the optimal strategy extracted from historical stable operating conditions to guide the parameters of future scheduling tasks.

[0079] S8 includes extracting feature vectors from spatiotemporal trend data, performing dimensionality reduction on the feature vectors using principal component analysis to obtain a dimensionality-reduced feature dataset; establishing a mapping model between processing parameters and spatiotemporal trends using regression analysis based on the dimensionality-reduced feature dataset to obtain a parameter mapping model; adjusting the model weights using gradient descent algorithm if the prediction error of the parameter mapping model exceeds a preset threshold to obtain an optimized parameter mapping model; generating processing parameter configurations for a new batch of base plates using data interpolation based on the optimized parameter mapping model to obtain a processing parameter configuration set; extracting key control parameters from the processing parameter configuration set, verifying the stability of the parameter configurations during processing using simulation analysis to obtain stability verification data; adjusting the key control parameters using constraint optimization to obtain an adjusted processing parameter set based on the adjusted processing parameter set; and generating a parameter update template for the processing equipment using data storage methods based on the adjusted processing parameter set to determine the final processing parameter template.

[0080] In implementing the aforementioned steps, the system first extracts information constituting the characteristics of thermal performance changes from the spatiotemporal trend data generated by the preceding steps. These characteristics include four items: the rate of change of thermal conductivity over time, the temporal standard deviation of the temperature gradient, the change in heat flux density per unit time, and the spatial distribution ratio of areas with uneven thermal conduction. These four characteristic parameters are quantitative indicators obtained by long-term data acquisition through a sensor array, followed by Fourier analysis, filtering, and thermal conductivity characteristic determination. The rate of change of thermal conductivity over time is determined by the ratio of the change in thermal conductivity per unit time to the time interval; the standard deviation of the temperature gradient is obtained by sampling the temperature gradient at each time point and calculating its standard deviation; the change in heat flux density is the difference between the maximum and minimum heat flux density during processing; and the ratio of areas with uneven thermal conduction is the proportion of areas with significant fluctuations in thermal conductivity to the total processed area. Subsequently, principal component analysis is used to perform dimensionality reduction on the above characteristic parameter set. The specific steps are as follows: First, construct the covariance matrix of the feature vectors. Then, solve for the eigenvalues ​​and corresponding eigenvectors of the matrix. Sort all the eigenvalues ​​in descending order and calculate the cumulative contribution rate. If the cumulative contribution rate of the first two eigenvalues ​​exceeds 90%, then retain the first two eigenvectors as principal components to form the dimensionality-reduced feature dataset, thereby reducing the complexity of subsequent modeling and avoiding feature redundancy.

[0081] Subsequently, based on the dimensionality-reduced feature dataset, the system uses a multinomial regression method to establish a mapping relationship between processing parameters and heat conduction trends. The processing parameters include six items: heating power, cooling flow rate, processing cycle time, heat flow equilibrium target value, response delay control upper limit, and unit processing energy consumption limit. Specifically, heating power is the output electrical power of the equipment's heating module, measured in watts, and its value is determined according to the upper limit of the equipment specifications; cooling flow rate is the volume of coolant in the fluid system per unit time, measured in liters per minute; processing cycle time is the time required to complete the processing of one base plate, measured in seconds; the heat flow equilibrium target value is the upper limit set for the standard deviation of heat flow in the processing area, measured in heat flux density units; the response delay upper limit is the time difference between the controller's response and the execution command, measured in seconds; and the energy consumption limit is the upper limit of the total energy consumption for a single processing cycle, measured in kilojoules. The multinomial regression model uses the dimensionality-reduced principal components as independent variables and the above six processing parameters as dependent variables, employing least squares fitting to establish the parameter mapping model.

[0082] During model training, the model's predictive performance is evaluated using five-fold cross-validation, and the average relative error of each fold is statistically analyzed. If the average error of any validation subset exceeds 5%, the system triggers an error optimization process, using a gradient descent-based model weight adjustment strategy to gradually correct the regression coefficients until all prediction errors are below a set threshold. The 5% threshold is determined by adding one standard deviation to the average error of 100 historical model cross-validations, ensuring that the error is controlled within an acceptable range.

[0083] After obtaining the optimized parameter mapping model, the system uses this model to predict the thermal performance input features of the new batch of base plates and generate a preliminary set of processing parameter configurations. Since the prediction results may have parameter jumps or sparsity in different feature intervals, the system uses a linear interpolation method to complete the configuration. Specifically, the interpolation method uses any two adjacent legal predicted parameter pairs as endpoints to construct a linear interval and uniformly fill the intermediate points, ultimately obtaining a processing parameter configuration set containing six consecutive parameters.

[0084] Next, three key control parameters—heating power, cooling flow rate, and response delay time—were extracted from the parameter configuration set and loaded into the thermal simulation module for simulation analysis. Specifically, a simulation mesh model with dimensions identical to the actual base plate was constructed in the finite element simulation environment. The aforementioned parameters were then loaded to simulate the heat conduction process, evaluating the heat flow distribution at each time point. If the heat flow density fluctuation at any simulation mesh point exceeded 10%, the parameter configuration was considered unstable. This 10% threshold was defined by the upper limit of the allowable heat flow equilibrium fluctuation range for the equipment, derived from the maximum tolerance in measured data that would not cause quality defects.

[0085] If the simulation results are unstable, the system immediately invokes a constrained optimization algorithm to reconfigure the parameters. The constraints are: heating power not exceeding 5000 watts, cooling flow rate not exceeding 50 liters per minute, and response delay not exceeding 2 seconds. The optimization objectives are to minimize processing energy consumption and maximize thermal uniformity. The two objectives are summed in a linear weighted manner with weighting coefficients of 0.6 and 0.4, respectively. The parameter combination is iteratively optimized using a particle swarm optimization algorithm until the thermal uniformity requirement is met and the energy consumption is kept below the target value.

[0086] Finally, the system generates a structured template based on the adjusted parameters. The template records the value of each parameter, the corresponding equipment module number, the area of ​​action number, and the update time. After generation, it is uploaded to the equipment management system as the final processing parameter template, realizing the standardized deployment and automated calling of parameters, and ensuring that subsequent batches of base plate processing continue to maintain high stability and low energy consumption.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for scheduling production resources for container flooring, characterized in that, The method comprises the following steps: S1, collecting the thermal conductivity data and local temperature gradient distribution data in the bottom plate processing process through a sensor array, processing the data to extract low-frequency domain features by using Fourier transform method, and obtaining the spatial and temporal uneven change trend of thermal conductivity; S2, judging whether the current processing state deviates from the preset thermal conductivity threshold according to the spatial and temporal uneven change trend, if deviating, starting the linear programming algorithm by setting the constraint condition and defining the objective function to optimize the material addition ratio and the temperature control parameter, and determining the adjusted parameter combination; S3, obtaining the allocation demand of heating and cooling resources from the adjusted parameter combination, limiting the variable range and solving the iteration process according to the allocation demand, and obtaining the resource allocation sequence; S4, obtaining the real-time equipment state data according to the resource allocation sequence, and performing optimal solution verification and parameter sensitivity analysis, judging whether the sequence matches the equipment capacity, if not matching, updating the sequence through error signal calculation and controller gain adjustment iteration until matching, and obtaining the final scheduling instruction; S5, according to the final scheduling instruction, updating the running parameters of the processing equipment in the feedback control cycle by using the integral term accumulation and the differential term prediction mechanism, and determining the real-time correction effect of the bottom plate thermal conductivity performance; S6, obtaining the cumulative deviation value of the performance index through the real-time correction effect, and performing system response monitoring and stability analysis, judging whether the cumulative deviation value is lower than the preset threshold, and obtaining the stability confirmation of the processing process; S7, extracting the optimization log data from the stability confirmation, grouping the historical adjustment records by using the resource consumption minimization and multi-objective trade-off method, and determining the reference parameter template for future processing; S8, obtaining state variable feedback and closed-loop gain optimization data for the reference parameter template, judging whether the template is suitable for new batch of bottom plate processing, if not suitable, re-optimizing the material addition ratio and the temperature control parameter by starting the linear programming algorithm through the constraint condition setting and the objective function definition, and obtaining the updated spatial and temporal change trend.

2. The method of claim 1, wherein: The S1 comprises: real-time collecting the microscale thermal conductivity and local temperature gradient data in the bottom plate processing process through a sensor array, storing as a time series data set, and obtaining original heat conduction data; performing frequency domain conversion on the original heat conduction data by using Fourier transform, extracting low-frequency domain features, and obtaining a frequency domain feature data set, wherein the frequency components with a frequency lower than a preset threshold in the frequency domain data obtained after Fourier transform are taken as low-frequency domain features; if the amplitude of the low-frequency component in the frequency domain feature data set exceeds the preset threshold, performing weighted processing on the low-frequency component to generate a weighted low-frequency feature set; performing inverse Fourier transform on the weighted low-frequency feature set to reconstruct the spatial and temporal change trend of thermal conductivity, and obtaining the spatial and temporal change distribution; according to the spatial and temporal change distribution, using a k-means clustering algorithm to divide the thermal conductivity into regions, and obtaining the thermal conductivity characteristic partition of different regions in the processing process; if there is an abnormal region in the thermal conductivity characteristic partition, verifying the abnormal region by using the local temperature gradient data to judge the heat conduction stability of the abnormal region; using an interpolation algorithm to generate a continuous distribution of the spatial and temporal change trend of thermal conductivity according to the heat conduction stability of the abnormal region, and obtaining the final heat conduction trend.

3. The method of claim 1, wherein: The S2 comprises: According to the adjusted parameter combination, real-time heat flow data in the processing process is obtained, continuous heat flow distribution is generated by adopting data smoothing processing, and a smooth heat flow data set is obtained; From the smooth heat flow data set, a heat flow abnormal point is extracted, if the heat flow abnormal point exceeds a preset fluctuation threshold, a corrected heat flow distribution is generated by an interpolation algorithm, and a corrected heat flow data set is obtained; Using the corrected heat flow data set, the heat flow uniformity index of the processing area is calculated, and the heat flow uniformity distribution is obtained; According to the heat flow uniformity distribution, the heat conduction deviation data of the processing area is obtained, if the heat conduction deviation data exceeds a preset deviation threshold, the power distribution of the processing equipment is optimized by a gradient descent algorithm, and the optimized power parameter is obtained; From the optimized power parameter, a control instruction set of the processing equipment is generated, and a time series analysis is used to predict the heat flow trend, and a predicted heat flow distribution is obtained; According to the predicted heat flow distribution, the operating parameters of the processing equipment are adjusted, real-time control signals are generated, and a stable processing state is obtained; From the stable processing state, new heat conduction performance data is collected, and a mean filter algorithm is used to process the data, and an updated heat conduction performance data set is obtained.

4. The method of claim 1, wherein: The S3 comprises: The scheduling instructions of heating and cooling resources are extracted from the resource allocation sequence, the scheduling time period is divided by using a time series segmentation method, and a segmented scheduling data set is obtained; According to the segmented scheduling data set, the resource utilization rate index of each period is calculated, if the resource utilization rate index is lower than a preset threshold, the scheduling time period distribution is adjusted by using a greedy algorithm, and an optimized scheduling time period is obtained; From the optimized scheduling time period, the resource consumption data of each period is obtained, and a mean filtering method is used to process the consumption data, and a smooth resource consumption data set is obtained; According to the smooth resource consumption data set, the resource fluctuation characteristics of each period are analyzed, if the fluctuation characteristics exceed a preset fluctuation threshold, a corrected resource consumption distribution is generated by using an interpolation method, and a corrected consumption data set is obtained; From the corrected consumption data set, resource scheduling deviation data is extracted, the statistical characteristics of the deviation data are calculated, and a resource scheduling deviation distribution is obtained; According to the resource scheduling deviation distribution, the allocation ratio of heating and cooling resources is adjusted, real-time resource control signals are generated, and a stable resource scheduling state is obtained; From the stable resource scheduling state, new resource allocation data is collected, and a data standardization method is used to process, and an updated resource allocation data set is obtained.

5. The method of claim 1, wherein: The S4 comprises: Real-time equipment state data is extracted from the resource allocation sequence, and a data standardization method is used to process the state data, and a standardized state data set is obtained; According to the standardized state data set, the matching degree of the device state and the device capacity constraint is calculated, if the matching degree is lower than a preset threshold, an initial error signal is generated; The initial error signal is optimized by using a least square method, and an optimized error signal is obtained; According to the optimized error signal, a proportional controller is used to calculate the adjustment gain of the resource allocation ratio, and an adjustment gain data is obtained; The resource allocation sequence is updated by using the adjustment gain data, and a temporary scheduling sequence is generated; According to the temporary scheduling sequence, the matching of the device state data and the device capacity constraint is verified, if the matching does not reach a preset threshold, the adjustment gain data is iteratively adjusted, and the temporary scheduling sequence is updated; The final scheduling instruction is generated by using the finally matched temporary scheduling sequence.

6. The method of claim 1, wherein: The S5 comprises: Through the final scheduling instruction, the processing equipment operation parameter contained in the instruction is parsed, the parameter data is processed by using a data standardization method, and a standardized operation parameter set is obtained; According to the standardized operation parameter set, the current heat flow density of the processing equipment is calculated, the bottom plate heat conduction performance is analyzed by using a Fourier heat conduction model, and initial heat flow distribution data is obtained; If the initial heat flow distribution data does not match the preset heat flow uniformity threshold, a heat flow deviation signal is generated, a proportional-integral-derivative controller is used to calculate a control cycle gain, and gain adjustment data is obtained; The processing equipment operation parameter is updated by using the gain adjustment data, a temporary operation parameter set is generated, the correction effect of the temporary operation parameter set on the bottom plate heat conduction performance is verified by using finite element analysis, and temporary heat flow distribution data is obtained; If the temporary heat flow distribution data still does not reach the preset heat flow uniformity threshold, the control cycle gain is iteratively optimized, the temporary operation parameter set is updated, and an optimized operation parameter set is obtained; According to the optimized operation parameter set, the bottom plate heat conduction performance is recalculated by using heat flow density analysis, the heat bridge elimination effect is verified, and final heat flow distribution data is obtained; Through the final heat flow distribution data, the final operation parameter instruction of the processing equipment is generated, and the real-time correction result of the bottom plate heat conduction performance is determined.

7. The method of claim 1, wherein: The S6 comprises: Continuous monitoring data of the equipment operation state is obtained by using the processing equipment sensor, and performance index data is generated by using data standardization processing; According to the performance index data, an accumulated deviation value is calculated by using a time series analysis method, and a deviation value sequence is obtained; If any value in the deviation value sequence exceeds a preset threshold, the performance index data is optimized by using a Kalman filtering algorithm, and an optimized deviation value sequence is generated; By using the optimized deviation value sequence, the response characteristics of the system are analyzed by using Fourier transform, and response characteristic parameters are obtained; If the response characteristic parameters do not reach a preset stability threshold, the equipment operation parameter is adjusted by using an adaptive control algorithm, and a temporary operation parameter set is generated; According to the temporary operation parameter set, the stability of the processing process is verified by using heat flow density analysis, and stability state data is obtained; Through the stability state data, the final operation parameter instruction of the processing equipment is generated, and the stability state of the processing process is determined.

8. The method of claim 1, wherein: The S7 comprises: Optimized log data is extracted from the stability confirmation data, and abnormal values are removed by using a data cleaning method, to obtain a cleaned log data set; According to the cleaned log data set, a K-means clustering algorithm is used to group historical adjustment records, and a classified adjustment parameter set is generated; By using the classified adjustment parameter set, a multi-objective optimization algorithm is used to calculate a trade-off scheme of resource consumption and heat conduction uniformity, and an optimized pre-adjustment parameter set is obtained; If any parameter in the optimized pre-adjustment parameter set exceeds a preset threshold, the parameter is smoothed by using a linear interpolation method, and a smoothed pre-adjustment parameter set is obtained; According to the smoothed pre-adjustment parameter set, the distribution characteristics of microscopic scale heat conduction unevenness are analyzed by using heat conduction simulation, and heat conduction distribution data is obtained; By using the heat conduction distribution data, a reference parameter template of the processing equipment is generated by using a data mapping method, and a pre-adjustment scheme is determined; According to the reference parameter template, a parameter verification method is used to compare historical processing data, and the applicability of the pre-adjustment scheme is judged.

9. The method of claim 1, wherein: The S8 comprises: extracting a feature vector from the spatio-temporal variation trend data, performing dimension reduction processing on the feature vector by using a principal component analysis method, and obtaining a dimension-reduced feature data set; establishing a mapping model of the processing parameters and the spatio-temporal variation trend by using a regression analysis method according to the dimension-reduced feature data set, and obtaining a parameter mapping model; if a prediction error of the parameter mapping model exceeds a preset threshold, adjusting model weights by using a gradient descent algorithm to obtain an optimized parameter mapping model; generating processing parameter configurations of a new batch of substrates by using a data interpolation method according to the optimized parameter mapping model, and obtaining a processing parameter configuration set.

10. The method of claim 9, wherein: The S8 further comprises: extracting key control parameters from the processing parameter configuration set, verifying stability of the parameter configurations in a processing process by using a simulation analysis method, and obtaining stability verification data; if a fluctuation amplitude of the stability verification data exceeds a preset threshold, adjusting the key control parameters by using a constraint optimization method to obtain an adjusted processing parameter set; generating a parameter update template of a processing device by using a data storage method according to the adjusted processing parameter set, and determining a final processing parameter template.

Images