Spraying and baking finish house flash-drying baking temperature optimization method based on intelligent data control

Abstract

The invention, which relates to the technical field of coating process temperature control, discloses an intelligent data-controlled spray-baking finish house flash-drying baking temperature optimization method comprising the following steps: acquiring wind speed data and temperature data of a target object grid point by grid point; performing space-time fitting on the wind speed data in a space neighborhood to generate a wind speed matrix, and performing numerical difference processing on the wind speed matrix to generate a vortex intensity matrix; judging whether the wind speed data is larger than a preset judgment threshold value or not according to a sliding window, if yes, counting the time sequence proportion of the wind speed data in the sliding window, and multiplying the time sequence proportion by a wind speed completion coefficient to obtain a stagnation factor matrix; according to the scheme, multi-threshold hierarchical segmentation is carried out on the stagnation factor matrix, and the vortex intensity matrix screening processing is combined, so that accurate identification of the real stagnation area in the air flow field in the spray-baking finish house is realized.

Description

Technical Field

[0001] This invention relates to the field of coating process temperature control technology, specifically to an intelligent data-controlled method for optimizing the flash-drying baking temperature of a spray booth. Background Technology

[0002] In modern industrial coating processes, flash drying in spray booths is a key step in improving coating quality and production efficiency. Spray booths typically use hot air to quickly dry the sprayed paint film, maintaining its gloss, adhesion, and uniformity while shortening the overall baking time. As coating processes become more refined and automated, the uniformity of temperature and airflow distribution within the spray booth has an increasingly significant impact on baking quality.

[0003] In existing technologies, to address the aforementioned issues, sensor networks are typically used to collect wind speed and temperature data. Empirical formulas are then used to calculate heater power or damper opening in each area, achieving a degree of zoned control. The advantage of these methods is that they provide some local temperature regulation capability and obtain basic flow field information through simple statistical analysis, thus avoiding stagnation zones to some extent. However, the empirical formulas used in these methods neglect the characteristics of wind speed changes over time and the data completion process when data is missing. This can easily lead to misjudging transient low-speed fluctuations as persistent stagnation zones, resulting in overheating or insufficient compensation in local temperature control. Furthermore, this inadequacy in stagnation zone identification makes it difficult for temperature control strategies to dynamically adapt to changes in the local flow field, and local temperature deviations are difficult to correct in a timely manner, thereby reducing the overall accuracy and uniformity of temperature control during the flash-drying baking process in paint spray booths. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent data-controlled method for optimizing the flash-drying baking temperature of a spray painting booth, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention discloses a method for optimizing the flash-drying baking temperature of a spray paint booth using intelligent data control, comprising the following steps:

[0007] Obtain wind speed and temperature data for the target object point by point;

[0008] The wind speed data is spatiotemporally fitted within a spatial neighborhood to generate a wind speed matrix, and the wind speed matrix is ​​numerically differencing to generate a vortex intensity matrix.

[0009] If the wind speed data is greater than a preset threshold, the time-series proportion of the wind speed data within the sliding window is calculated, and the time-series proportion is multiplied by the wind speed completion coefficient to obtain the stagnation factor matrix.

[0010] The wind speed completion coefficient is generated by weighted interpolation of the wind speed data in the spatial neighborhood.

[0011] The preset discrimination threshold is obtained by performing distribution statistics on the wind speed data contained in the corresponding grid points within the sliding window;

[0012] The stagnation factor matrix is ​​segmented using a multi-threshold hierarchical method and then sieved using the eddy intensity matrix to generate stagnation zone data. Boundary segmentation is then performed based on the stagnation zone data and the wind speed matrix. Finally, a fast convolution estimation is performed on the stagnation factor matrix, the eddy intensity matrix, and the temperature data within the boundary segmentation results to generate a temperature optimization scheme.

[0013] Secondly, this invention discloses an intelligent data-controlled flash-drying baking temperature optimization system for spray-paint booths, comprising:

[0014] The data acquisition module is used to acquire wind speed and temperature data of the target object on a grid-by-grid basis;

[0015] The wind speed data processing module is used to perform spatiotemporal fitting on the wind speed data in the spatial neighborhood to generate a wind speed matrix, and to perform numerical difference processing on the wind speed matrix to generate a vortex intensity matrix.

[0016] The stagnation analysis module is used to determine whether the wind speed data is greater than a preset discrimination threshold by sliding window. If so, the time-series proportion of the wind speed data in the sliding window is counted, and the time-series proportion is multiplied by the wind speed completion coefficient to obtain the stagnation factor matrix.

[0017] The temperature optimization scheme generation module is used to perform multi-threshold hierarchical segmentation on the stagnation factor matrix and combine it with the vortex strength matrix for screening to generate stagnation zone data. Based on the stagnation zone data and the wind speed matrix, boundary segmentation is performed, and then the stagnation factor matrix, the vortex strength matrix and the temperature data are rapidly convolved and estimated within the boundary segmentation results to generate a temperature optimization scheme.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0019] 1. This solution achieves accurate identification of the real stagnant areas in the airflow field inside the spray booth by performing multi-threshold hierarchical segmentation of the stagnation factor matrix and combining it with vortex strength matrix screening. By introducing wind speed completion coefficient and wind speed matrix to calculate the confidence of grid points and removing the hierarchical judgment results of low confidence points, it effectively eliminates misjudgments caused by transient disturbances or sensor noise, thereby ensuring the high reliability and traceability of stagnation area data. This enables the zoned temperature control strategy to be precisely adjusted based on the real stagnation areas, achieving the dual technical effects of improving baking quality and optimizing energy efficiency.

[0020] 2. This scheme uses stagnation zone data to extract an initial seed set, which can ensure that potential stagnation zones are accurately located. The adjacency matrix is ​​used to evaluate the wind speed consistency of adjacent grids, so as to achieve reasonable connectivity constraints between partitions. The thermal diffusivity coefficient matrix is ​​used to generate a cost matrix, so that the boundary division not only considers the flow field characteristics, but also takes into account the heat transfer efficiency. Finally, the boundary segmentation is completed through graph cut iterative optimization, ensuring that the segmentation result of each region conforms to the actual wind field and thermal diffusivity distribution, and realizing refined management of partition temperature control. Attached Figure Description

[0021] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0022] Figure 1 This is a flowchart illustrating the steps of an intelligent data-controlled method for optimizing the flash-drying baking temperature in a spray painting booth according to the present invention.

[0023] Figure 2 A schematic diagram of the process for generating stagnation zone data provided by the present invention;

[0024] Figure 3 This is a schematic diagram of the process for generating the thermal diffusivity matrix provided by the present invention.

[0025] Figure 4 A schematic diagram of the process for generating boundary segmentation results provided by the present invention;

[0026] Figure 5 This invention provides a schematic diagram of the module functions of an intelligent data-controlled flash-drying baking temperature optimization system for spray booths. Detailed Implementation

[0027] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0028] Application Overview:

[0029] In the flash-drying baking process of spray paint booths, the lack of modeling of the time dynamic characteristics and data completion mechanism in the wind speed data processing process leads to transient low-speed fluctuations being misjudged as continuous stagnation zones. This results in problems such as excessive or insufficient local temperature control. The essence of this problem is that existing methods rely solely on static empirical formulas to determine wind speed thresholds, failing to effectively distinguish between transient fluctuations and true stagnation states. At the same time, they ignore the correlation of spatial neighborhood data, making it difficult for temperature control strategies to dynamically adapt to changes in the flow field. This results in local temperature deviations that are difficult to correct in a timely manner, directly affecting key performance indicators such as coating gloss, adhesion, and uniformity.

[0030] For example, during the operation of the paint spraying booth in an automotive body painting production line, when multiple painting robots are working simultaneously, complex airflow disturbances occur in local areas due to equipment obstruction. The wind speed data collected by the sensor network shows intermittent gaps and instantaneous fluctuations. At this time, the existing control system judges the stagnation area based on a fixed empirical formula, identifying the instantaneous low-speed fluctuations caused by robot movement as a continuous stagnation area, and erroneously triggering excessive compensation of heater power. At the same time, because the data gap area is not spatially interpolated to complete, the temperature control compensation in adjacent areas is insufficient. This misjudgment causes the paint film in the transition area of ​​the curved surface of the car body to be locally over-baked or not fully cured, resulting in a decrease in coating adhesion and a deterioration in surface leveling, which in turn affects the coating quality stability of subsequent processes.

[0031] If this problem persists, the coupling imbalance between the temperature field and the flow field in the spray booth will be further aggravated, causing the identification results of stagnant areas to deviate from the actual physical characteristics of the flow field. The temperature optimization scheme will be difficult to accurately match the local flow field changes. The resulting temperature control deviation will cause the solvent evaporation rate to be mismatched during the paint film drying process, resulting in uneven stress distribution inside the coating film. Ultimately, this will lead to the accumulation of microstructural defects in the paint film, reducing the consistency of the overall coating quality and the reliability of the process.

[0032] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0033] Example 1:

[0034] Please see Figure 1 - Figure 4A method for optimizing the flash-drying baking temperature of a spray paint booth using intelligent data control, comprising the following steps:

[0035] Obtain wind speed and temperature data for the target object point by point;

[0036] The wind speed data is spatiotemporally fitted within the spatial neighborhood to generate a wind speed matrix, and the wind speed matrix is ​​numerically differencing to generate a vortex intensity matrix.

[0037] The sliding window is used to determine whether the wind speed data is greater than the preset threshold. If so, the time-series proportion of the wind speed data in the sliding window is counted, and the time-series proportion is multiplied by the wind speed completion coefficient to obtain the stagnation factor matrix.

[0038] Among them, the wind speed completion coefficient is generated by weighted interpolation of wind speed data within the spatial neighborhood;

[0039] The preset discrimination threshold is obtained by statistically analyzing the distribution of wind speed data contained in the corresponding grid points within the sliding window;

[0040] The stagnation factor matrix is ​​segmented by multiple thresholds and screened by combining it with the eddy intensity matrix to generate stagnation zone data. Boundary segmentation is then performed based on the stagnation zone data and the wind speed matrix. Finally, a fast convolution estimation is performed on the stagnation factor matrix, eddy intensity matrix, and temperature data within the boundary segmentation results to generate a temperature optimization scheme.

[0041] Spatiotemporal fitting refers to the process of modeling wind speed data in a spatial neighborhood to generate a wind speed matrix. It can be achieved using multinomial regression or spline interpolation, for example, using cubic spline functions to fit the wind speed time series data of grid points.

[0042] The wind speed completion coefficient refers to the weighted coefficient obtained by weighted interpolation of wind speed data in the spatial neighborhood. Furthermore, it can be implemented by inverse distance weighting or radial basis function interpolation.

[0043] The preset discrimination threshold is the threshold used to determine whether the wind speed is below the critical value. It can be achieved by calculating the fixed percentile of the wind speed data within the sliding window. For example, the 90th percentile can be used independently as the threshold setting basis for each grid point.

[0044] Multi-threshold hierarchical segmentation refers to the operation of applying multiple thresholds to the stagnation factor matrix for hierarchical judgment. It can be implemented by using a preset fixed threshold set or clustering algorithm, such as dividing the stagnation region based on experience by setting three threshold levels: high, medium and low.

[0045] In some of the embodiments described above in this application, a stagnation factor matrix is ​​proposed to identify airflow stagnation areas. However, in its implementation, there is a lack of dynamic analysis of the coupling relationship between temperature gradient and stagnation area, which makes it difficult for temperature optimization schemes to accurately capture changes in heat transfer requirements. Consequently, it is difficult to correct temperature deviations in a timely manner when there are fluctuations in the local flow field, thus affecting baking uniformity.

[0046] In this regard, this application further proposes that after obtaining the stagnation factor matrix, the process also includes generating the gradient field matrix, specifically including:

[0047] The temperature data is subjected to central difference calculation between grid points and adjacent grid points, and the central difference calculation results are subjected to weighted interpolation smoothing within the spatial neighborhood of the grid points to generate a smooth temperature difference vector.

[0048] The smoothed temperature difference vector and the stagnation factor matrix are coupled and analyzed according to grid points and time series to generate the gradient field matrix.

[0049] Among them, central difference calculation refers to estimating the local temperature change rate through numerical differentiation methods. It can be implemented using the symmetric difference formula in the finite difference method, for example, by using the temperature values ​​of adjacent grid points to perform linear combination calculations.

[0050] Weighted interpolation smoothing can be understood as applying spatial filtering to the difference results to suppress noise interference. It can be implemented using kernel-based local regression smoothing or adaptive weighted interpolation methods.

[0051] Coupling analysis specifically establishes a relationship between the smoothed temperature difference vector and the stagnation factor matrix in the spatiotemporal dimension, which can be achieved using dynamic time warping algorithms or sliding window correlation analysis.

[0052] This scheme directly obtains local temperature change rate information by performing central difference calculation on temperature data between grid points and adjacent grid points. Since there is an inherent physical relationship between temperature gradient and heat transfer intensity, this step can effectively compensate for the limitation of wind speed data in representing heat energy distribution. On this basis, by performing weighted interpolation smoothing on the central difference calculation results in the spatial neighborhood of the grid points, and using spatial continuity constraints to suppress data jitter, the generated smooth temperature difference vector more accurately represents the true temperature gradient distribution. Furthermore, the smooth temperature difference vector and the stagnation factor matrix are coupled and analyzed according to grid points and time series, which correlates the dynamic characteristics of temperature change with the characteristics of airflow stagnation in the spatiotemporal dimension. The generated gradient field matrix can dynamically indicate the compensation direction of the heat exchange efficiency reduction region, thereby achieving precise control of local temperature deviation in subsequent optimization.

[0053] As a specific implementation method, the scheme of this application is implemented as follows: the central difference calculation can be specifically implemented by applying the five-point stencil difference method at the grid points, wherein the difference weights are dynamically adjusted according to the grid topology; the weighted interpolation smoothing process can be implemented by using a spatial filter based on the Gaussian kernel function to perform a weighted average of the difference results in the neighborhood; the coupling analysis can be specifically implemented by constructing a grid point-level time series correlation model to dynamically match and calculate the smoothed temperature difference vector and the stagnation factor matrix within a sliding time window.

[0054] Through the above technical solution, this application can dynamically capture changes in heat transfer demand and correct temperature deviations in a timely manner when there are fluctuations in the local flow field, thereby improving the accuracy and uniformity of temperature control during the flash drying process of the spray painting booth.

[0055] In practical applications, some embodiments of this application propose to perform multi-threshold hierarchical segmentation of the stagnation factor matrix and combine it with the vortex strength matrix for screening to generate accurate stagnation zone data. However, in its implementation, relying solely on a single threshold or simple statistical determination of the stagnation zone makes it difficult to effectively distinguish between transient low-speed fluctuations and continuous stagnation states. This leads to stagnation zone identification being easily affected by data noise, resulting in misjudgment points. Consequently, the temperature optimization scheme may experience local overheating or insufficient compensation due to distortion of the basic data.

[0056] To address this, this application further proposes a multi-threshold hierarchical segmentation of the stagnation factor matrix, combined with vortex strength matrix screening, to generate stagnation zone data, specifically including:

[0057] A multi-threshold set is applied to the stagnation factor matrix for stratified determination, and the confidence level of the stratified determination results is calculated based on the wind speed completion coefficient and the wind speed matrix at each grid point.

[0058] The multi-threshold set is obtained by performing distribution statistics on the stagnation factor matrix;

[0059] Determine whether the confidence score calculation result is less than the preset confidence score threshold; otherwise, remove the stratification judgment result corresponding to the grid point and generate stagnation zone data.

[0060] Among them, the multi-threshold set refers to a set of dynamic thresholds obtained by performing distribution statistics on the stagnation factor matrix, which can be implemented by generalization methods such as percentile statistics or kernel density estimation.

[0061] Hierarchical determination can be understood as the process of hierarchical identification of the stagnant factor matrix based on a set of multiple thresholds. It can be implemented by parallel methods such as interval division or step-by-step threshold comparison.

[0062] The wind speed completion coefficient refers to the compensation parameter generated by weighted interpolation of wind speed data in a spatial neighborhood. It can be implemented using spatial interpolation techniques such as inverse distance weighting or kriging interpolation.

[0063] Confidence calculation refers to the reliability assessment based on the wind speed completion coefficient and the wind speed matrix at grid points. It can be implemented using generalization methods such as weighted product or fuzzy logic.

[0064] The confidence threshold is a preset reference value used to determine the screening results. It can be set based on historical data statistics or experience.

[0065] This scheme first applies a multi-threshold set to the stagnation factor matrix for hierarchical determination, achieving graded identification of stagnation regions based on data distribution characteristics, laying a structured foundation for subsequent reliability assessment. Then, confidence levels are calculated for each grid point based on the wind speed completion coefficient and wind speed matrix, dynamically fusing flow field information and data integrity to accurately assess the credibility of the stagnation determination at each grid point. Finally, by judging whether the calculated confidence level is less than a preset confidence threshold, the hierarchical determination results corresponding to the corresponding grid points are eliminated, filtering out low-reliability determination points. This organic combination of multi-threshold hierarchical and confidence screening mechanisms ensures that the threshold settings closely align with the actual data distribution characteristics, enabling precise differentiation between transient low-speed fluctuations and continuous stagnation states, effectively eliminating data noise interference, and generating stagnation region data that truly reflects continuous stagnation states, providing accurate and reliable input for subsequent boundary segmentation and temperature optimization.

[0066] As a preferred embodiment, the solution of this application is implemented as follows: In the monitoring system of the spray painting booth, the stagnation factor matrix is ​​first classified and determined through a multi-threshold set, which is statistically obtained based on the historical distribution data of the stagnation factor matrix; subsequently, the system calculates the confidence level of each grid point based on the wind speed completion coefficient (generated through spatial neighborhood weighted interpolation) and the wind speed matrix; for grid points with a confidence level lower than the threshold, the system automatically removes their classification and determination results, retaining only the stagnation area data with high confidence for subsequent processing. This processing unit can specifically be an embedded microcontroller, responsible for executing the above determination and screening logic, and interacting with the sensor network of the spray painting booth via a fieldbus.

[0067] Through the above scheme, this application can effectively distinguish between transient low-speed fluctuations and continuous stagnation, reduce misjudgment points in stagnation zone identification, and avoid local overheating or insufficient compensation caused by distortion of basic data, thereby improving the accuracy and uniformity of temperature optimization during the flash drying process of the spray painting booth.

[0068] In some of the embodiments described above in this application, data on stagnant areas are proposed to accurately screen stagnant areas in the spray booth. However, in the implementation process, the dynamic changes in heat diffusion characteristics are not considered, which makes it difficult for the temperature optimization scheme to accurately adapt to the heat conduction process, resulting in local temperature deviations that are difficult to be effectively corrected.

[0069] In this regard, this application further proposes that after generating the stagnation zone data, it also includes generating a thermal diffusivity matrix, specifically including:

[0070] The temperature data is differentially processed within adjacent sliding windows, and the results of the differential processing are combined with the temperature data within the sliding windows to form an observation vector;

[0071] The gradient field matrix and wind speed matrix are weighted and calculated point by point in the grid to generate a coefficient matrix;

[0072] The thermal diffusivity matrix is ​​obtained by constructing and solving a regularized inverse problem on the observation vector and coefficient matrix.

[0073] Among them, the difference operation within adjacent sliding windows refers to the extraction of the rate of change of temperature data in continuous sliding windows over a time series, which can be achieved using forward difference, central difference, or higher-order difference methods.

[0074] The formation of the observation vector can be understood as combining the difference operation result with the original temperature data within the sliding window through vector concatenation or weighted fusion.

[0075] Weighted calculation per grid point refers to the independent fusion operation of the gradient field matrix and wind speed matrix for each grid point in the spatial dimension. It can be implemented by linear weighting, exponential decay weighting, or adaptive weighting based on flow field characteristics.

[0076] The construction and solution of the regularization inverse problem can be understood as the process of solving ill-conditioned equations by introducing constraint terms, which can be implemented using Tikhonov regularization, total variational regularization or sparse regularization methods.

[0077] Specifically, the proposed solution first performs a difference operation on the temperature data of adjacent sliding windows to extract the temperature change rate information and combines it with the original temperature data within the window to form an observation vector. This observation vector serves as a dynamic input model that fully characterizes the time-varying characteristics of the temperature field. Subsequently, based on the temperature gradient distribution reflected by the gradient field matrix and the flow field velocity characteristics represented by the wind speed matrix, point-by-point weighted calculations are performed at the spatial grid point scale to generate a coefficient matrix that accurately describes the local thermal diffusion physical mechanism. Finally, the observation vector and coefficient matrix are substituted into a regularized inverse problem framework, and the thermal diffusion coefficient matrix with spatial resolution is obtained by solving this optimization problem. This process effectively balances the data fitting accuracy and the stability of the solution through regularization constraints, thereby establishing an accurate mapping relationship between thermal diffusion characteristics and dynamic changes in the flow field.

[0078] As a specific implementation method, the scheme of this application is implemented as follows: the temperature data is subjected to a central difference method to perform a difference operation in adjacent sliding windows. The difference operation result and the temperature data in the window are concatenated to form an observation vector; the grid-point weighted calculation of the gradient field matrix and the wind speed matrix adopts an adaptive weighting strategy based on the Reynolds number of the flow field, wherein the weight coefficient is dynamically adjusted according to the local flow velocity; the solution of the regularization inverse problem adopts the Tikhonov regularization method, and the regularization parameters are determined by the L-curve criterion. The final obtained thermal diffusivity coefficient matrix is ​​stored in the embedded controller in a gridded data structure. The controller can be a microprocessor with an ARM Cortex-M7 architecture, which is used to update the thermal conduction parameters of the temperature optimization scheme in real time.

[0079] Through the above technical solution, this application realizes dynamic modeling of the heat diffusion characteristics in the spray baking booth, so that the temperature optimization scheme can accurately adapt to the differences in heat conduction process caused by local flow field changes, effectively avoid local temperature deviation caused by the distortion of heat diffusion characteristics, thereby improving the accuracy and uniformity of temperature control during flash drying baking.

[0080] In some of the embodiments described above in this application, boundary segmentation based on stagnation zone data and wind speed matrix is ​​proposed to accurately identify stagnation zone boundaries. However, in its implementation, the boundaries of stagnation zone data generated by relying solely on multi-threshold layering and confidence calculation are blurry and it is difficult to effectively combine the spatial consistency of wind speed and thermal diffusion characteristics, resulting in inaccurate boundary segmentation, which in turn affects the pertinence and effectiveness of temperature optimization.

[0081] In response, this application further proposes boundary segmentation based on stagnation zone data and wind speed matrix, specifically including:

[0082] Based on the stratification results, the stagnation zone boundaries are extracted from the stagnation zone data to generate an initial seed set;

[0083] The wind speed matrix is ​​evaluated for consistency in speed magnitude and direction between adjacent grid points to generate an adjacency matrix;

[0084] Based on the stratification results, the thermal diffusivity matrix is ​​normalized and inverted within the grid points to generate the cost matrix.

[0085] The initial seed set, adjacency matrix, and cost matrix are subjected to iterative optimization using graph cut with connectivity constraints to perform boundary segmentation until the preset number of iterations is reached, generating the boundary segmentation result.

[0086] Among them, stagnation zone boundary extraction refers to the process of identifying the boundary between stagnation and non-stagnation zones from the hierarchical determination results. It can be implemented by gradient-based edge detection algorithms or region growth boundary tracking methods.

[0087] The initial seed set can be understood as the initial point set representing the potential location of the boundary, which can be implemented by generating boundary pixels through morphological dilation operation or extracting contour points through threshold segmentation.

[0088] The assessment of the consistency of speed magnitude and direction refers to quantifying the similarity of wind speed vectors between adjacent grid points, which can be achieved by calculating the cosine of the vector angle or by normalizing the magnitude of the speed difference.

[0089] An adjacency matrix can be understood as a structured representation describing the connection relationships between grid points. It can be implemented using a weighted graph based on Euclidean distance or a binary connectivity matrix based on a similarity threshold.

[0090] Intra-regional normalization refers to the standardization of thermal diffusivity data within a specific stratified region of stagnation. This can be achieved using Z-score standardization or regional range normalization.

[0091] Inversion mapping can be understood as performing a numerical inverse mapping operation on the normalized thermal diffusivity, which can be implemented using a linear inversion function or a nonlinear decay mapping function.

[0092] The cost matrix is ​​a quantitative index matrix that comprehensively reflects the difficulty of boundary segmentation. It can be implemented by weighted fusion of multi-source data or by constructing a cost function based on a physical model.

[0093] Iterative optimization of graph cut with connectivity constraints can be understood as a computational process of boundary optimization under the condition of ensuring the connectivity of the region. It can be implemented by using the maximum flow minimum cut algorithm combined with connectivity verification mechanism or the connected component constraint optimization method in graph theory.

[0094] Specifically, the proposed solution first extracts the boundaries of stagnation zone data based on the stratification determination results to generate an initial seed set, providing a high-confidence starting point for boundary segmentation. Then, it performs velocity consistency evaluation on the wind speed matrix between adjacent grid points to generate an adjacency matrix. This adjacency matrix effectively captures the spatial continuity characteristics of airflow, avoiding misjudging transient low-speed fluctuations as stagnation zone boundaries. Simultaneously, it normalizes the thermal diffusivity coefficient matrix within the region based on the stratification determination results and inversely maps it to generate a cost matrix. This processing combines the physical properties of thermal diffusivity, eliminating regional differences through normalization and utilizing inverse mapping to improve thermal diffusivity. The cost of segmenting good regions is reduced, thereby guiding the boundary to shift towards the real stagnation zone with low heat transfer efficiency. Finally, the initial seed set, adjacency matrix, and cost matrix are input into the graph cut iterative optimization process with connectivity constraints. This optimization process adjusts the boundary position step by step based on graph theory principles. In each iteration, the spatial consistency of wind speed, heat diffusion characteristics, and initial boundary confidence are comprehensively considered. The connectivity constraints ensure the physical rationality of the segmented regions until the preset number of iterations is reached to generate accurate boundary segmentation results. The results fully preserve the physical connectivity characteristics of the stagnation zone, providing a reliable basis for subsequent temperature optimization.

[0095] As a specific implementation method, the scheme of this application is implemented as follows: In the stagnation zone boundary extraction stage, the Canny edge detection algorithm is used to process the layer judgment results and generate a boundary contour with sub-pixel accuracy as the initial seed set; in the wind speed consistency evaluation stage, the cosine similarity of the wind speed vectors between adjacent grid points is calculated, and when the similarity is greater than 0.85, a high connection weight is set in the adjacency matrix; in the cost matrix generation stage, the heat diffusion coefficient matrix in each stagnation zone layer region is standardized by Z-score and numerically inverted by a linear function; in the boundary segmentation stage, the Boykov-Kolmogorov maximum flow algorithm is used to implement graph cut optimization with connectivity constraints, and the connectivity of the segmented region is verified after each iteration. When the boundary position change is less than a threshold for three consecutive iterations, the optimization process is terminated early.

[0096] Through the above technical solution, this application effectively solves the problem of blurred boundaries of the stagnation zone, realizes accurate identification of the stagnation zone boundary, and enables the temperature optimization scheme to accurately control the actual stagnation zone, avoiding local overheating or insufficient compensation caused by boundary misjudgment, and significantly improving the uniformity and control accuracy of temperature distribution during the flash drying process of the spray painting booth.

[0097] Specifically, in some of the embodiments described above in this application, boundary segmentation is proposed based on stagnation zone data and wind speed matrix to generate boundary segmentation results. However, in the implementation process, after the boundary segmentation results are generated, the multi-source data in the segmentation area is not weighted and optimized or processed in real time by convolution. This results in the temperature optimization scheme not being able to fully adapt to the dynamic changes in the local flow field, making it difficult to correct local temperature deviations in a timely manner.

[0098] To address this, this application further proposes a fast convolution estimation of the stagnation factor matrix, eddy current matrix, and temperature data within the boundary segmentation results, generating a temperature optimization scheme that specifically includes:

[0099] Within the boundary segmentation results, the thermal diffusivity matrix, stagnation factor matrix, eddy current matrix and temperature data are normalized and weighted to generate temperature compensation.

[0100] A convex optimization problem is constructed and solved for the temperature compensation amount to obtain the optimization weight vector;

[0101] The temperature compensation amount and the optimization weight vector are convolved and optimized in real time to generate a temperature optimization scheme.

[0102] Among them, normalization weighting refers to the standardization of multi-source data to eliminate differences in units and integrate key information. It can be achieved by methods such as z-score standardization or min-max normalization.

[0103] The construction and solution of convex optimization problems can be understood as the process of dynamically determining the optimal weight allocation under constraints, which can be implemented using convex optimization algorithms such as the interior point method or the gradient projection method.

[0104] Real-time convolution and secondary optimization specifically integrate spatial correlations through convolution operations and combine iterative fine-tuning. It can accelerate convolution calculations using fast Fourier transform and optimize using sequential quadratic programming.

[0105] Specifically, the proposed scheme first normalizes and weights the thermal diffusivity matrix, stagnation factor matrix, eddy current matrix, and temperature data within the boundary segmentation results to generate a temperature compensation quantity. This step eliminates the influence of dimensions and weightedly fuses multi-source information to form a comprehensive compensation index reflecting local needs. Subsequently, a convex optimization problem is constructed based on this temperature compensation quantity, and the optimization weight vector is obtained by solving it. This optimization weight vector dynamically adjusts the weight ratio of each compensation quantity under preset constraints to ensure that the scheme adapts to real-time flow field conditions. Finally, the temperature compensation quantity and the optimization weight vector are convolved in real time and subjected to secondary optimization. The convolution operation integrates spatial neighborhood correlations, and the secondary optimization refines the results to generate the final temperature optimization scheme. The above steps form a closed-loop optimization mechanism. The temperature compensation quantity serves as the input to the convex optimization problem, and the optimization weight vector directly drives the convolution process. The three work together to realize a complete chain from multi-source data fusion to dynamic weight adjustment and spatial response optimization, ensuring that the temperature optimization scheme can accurately match the flow field characteristics within the boundary segmentation region.

[0106] As a preferred embodiment, the solution of this application is specifically implemented as follows: the normalization weighting process can be specifically standardized by calculating the mean and standard deviation of each data matrix for z-score standardization, and linear weighted fusion with configurable weight coefficients is adopted; the convex optimization problem can be constructed as a quadratic programming problem to minimize the compensation error, and an open-source convex optimization solver is called for efficient solution; the real-time convolution operation can be converted to the frequency domain and spatial correlation integration is achieved by using fast Fourier transform, and the secondary optimization improves the convergence accuracy of the solution by iteratively adjusting the optimization weight vector.

[0107] Through the above scheme, this application realizes the real-time adaptation of the temperature optimization scheme to the dynamic changes of the local flow field, effectively corrects the local temperature deviation during the flash drying process of the spray painting booth, and improves the uniformity of temperature distribution and control accuracy.

[0108] In some of the embodiments described above in this application, a normalized weighted average of the thermal diffusivity matrix, stagnation factor matrix, vortex intensity matrix, and temperature data is proposed to generate a temperature compensation amount within the boundary segmentation result. However, in its implementation, the time spectrum characteristics of the vortex intensity matrix are not considered, making it difficult to accurately distinguish between transient low-speed fluctuations and continuous stagnation zones. This makes it difficult for the temperature compensation amount to dynamically respond to the instantaneous changes in the flow field, resulting in deviations in local temperature control and affecting the accuracy and adaptability of the temperature optimization scheme.

[0109] To address this, this application further proposes a normalized and weighted approach to the thermal diffusivity matrix, stagnation factor matrix, eddy current matrix, and temperature data within the boundary segmentation results, generating a temperature compensation quantity that specifically includes:

[0110] Within the boundary segmentation results, the thermal diffusivity matrix, stagnation factor matrix, time spectrum data, eddy intensity matrix and temperature data are normalized and weighted to generate temperature compensation.

[0111] The time spectrum data is generated by performing a short-time Fourier transform on the vortex intensity matrix within a sliding window.

[0112] Among them, time spectrum data refers to dynamic data that characterizes the time spectrum characteristics of the vortex intensity matrix, which can be realized by short-time Fourier transform, wavelet transform or Hilbert-Huang transform.

[0113] Among them, performing a short-time Fourier transform on the vortex intensity matrix within a sliding window refers to analyzing the spectral distribution of the vortex intensity matrix within a local time window, which can use a Hamming window, Hanning window, or Gaussian window as the window function.

[0114] Specifically, normalized weighting refers to the weighted fusion of data with different dimensions after standardization. It can be achieved using min-max normalization, Z-score normalization, or wavelet normalization methods.

[0115] Specifically, this application first generates time-spectrum data by performing a short-time Fourier transform on the vortex intensity matrix within a sliding window. This time-spectrum data can characterize the instantaneous changes in vortex intensity. Subsequently, within the boundary segmentation results, the thermal diffusivity matrix, stagnation factor matrix, time-spectrum data, vortex intensity matrix, and temperature data are normalized to eliminate dimensional differences. Finally, the normalized data are weighted and fused to generate a temperature compensation quantity. This process, through the introduction of time-spectrum data, enables the temperature compensation quantity to dynamically respond to instantaneous changes in the flow field, thereby accurately distinguishing between transient low-speed fluctuations and persistent stagnation zones, and avoiding temperature control deviations caused by misjudgment.

[0116] As a preferred embodiment, the solution of this application is specifically implemented as follows: In the temperature control system of the spray painting booth, the processing unit specifically adopts an ARM Cortex-M7 microcontroller with digital signal processing capabilities to perform short-time Fourier transform operations; in the sliding window processing, a Hanning window is used as the window function for time-frequency analysis; the normalization weighting process is implemented through an embedded software algorithm, which standardizes the input data and then performs fusion calculations based on preset weight coefficients to generate the temperature compensation amount.

[0117] Through the above scheme, this application can accurately distinguish between transient low-speed fluctuations and continuous stagnation zones, enabling the temperature compensation amount to dynamically respond to the instantaneous changes in the flow field, effectively avoiding overheating or insufficient compensation in local temperature control, thereby improving the accuracy and adaptability of the temperature optimization scheme.

[0118] In some of the embodiments described above in this application, a convex optimization problem is proposed to be constructed for the temperature compensation amount and solved to obtain the optimization weight vector for generating a temperature optimization scheme. However, in its implementation, the objective function is constructed only based on the temperature compensation amount, ignoring the influence of heat loss. This results in the optimization weight vector not accurately reflecting the actual heat loss, causing insufficient or excessive local temperature compensation, which in turn leads to uneven temperature distribution during baking, affecting the drying quality of the paint film and production efficiency.

[0119] In response, this application further proposes to construct and solve a convex optimization problem for the temperature compensation amount, obtaining the optimization weight vector, which specifically includes:

[0120] The objective function is constructed and solved using temperature compensation and heat loss data as inputs to obtain the optimized weight vector of the boundary segmentation result;

[0121] Among them, the heat loss data is generated by estimating the heat diffusivity matrix and the gradient field matrix through heat flow.

[0122] Among them, the temperature compensation amount refers to the value generated within the boundary segmentation result to quantify the temperature adjustment requirements. It can be calculated by normalizing the weighted average method to calculate the thermal diffusivity matrix, stagnation factor matrix, eddy current matrix and temperature data.

[0123] Heat loss data refers to data that reflects the characteristics of heat energy loss. It can be estimated and generated in real time based on the thermal diffusivity matrix and the gradient field matrix using Fourier's law of heat conduction.

[0124] The objective function is a mathematical expression used for convex optimization problems, which can be constructed as a quadratic programming form that includes temperature compensation and heat loss data.

[0125] The optimization weight vector is the decision vector that guides the adjustment of the temperature compensation amount, which can be obtained by solving the objective function;

[0126] The thermal diffusivity matrix is ​​a spatial distribution matrix that describes the thermal conductivity of a material. It can be calculated by performing differential operations on temperature data within adjacent sliding windows and combining it with the wind speed matrix.

[0127] The gradient field matrix is ​​a matrix that captures the spatial distribution characteristics of temperature gradients. It can be generated by performing central difference calculation and weighted interpolation smoothing on temperature data.

[0128] Heat flow estimation refers to the method of calculating the rate of heat transfer, which can be derived using a physical model based on the thermal diffusivity and temperature gradient.

[0129] Specifically, the proposed solution integrates heat loss data into the objective function construction process, enabling the optimization weight vector to simultaneously respond to both temperature compensation requirements and dynamic changes in actual heat loss. The temperature compensation amount is directly related to the temperature deviation that needs adjustment within the boundary segmentation region, while the heat loss data quantifies the energy loss in that region due to heat conduction and temperature gradients. During the solution process, the objective function dynamically balances the relationship between the two to avoid optimization bias caused by neglecting heat loss. The generation of heat loss data relies on heat flow estimation using the thermal diffusivity matrix and the gradient field matrix. The thermal diffusivity matrix accurately describes the heat conduction capacity, while the gradient field matrix captures the temperature gradient distribution characteristics. These two are coupled based on the physical mechanism that heat flow is proportional to the thermal diffusivity and temperature gradient, ensuring high accuracy of the heat loss data. This mechanism allows the objective function to respond in real-time to changes in the thermophysical properties within the paint booth, thus providing a reliable basis for the optimization weight vector and ultimately achieving precise adjustment of the temperature compensation amount.

[0130] As a specific implementation method, the solution of this application is implemented as follows: Within the boundary segmentation result, the temperature compensation amount is generated by normalizing and weighting the thermal diffusivity matrix, stagnation factor matrix, time spectrum data, eddy current matrix, and temperature data; the heat loss data is calculated by the heat flow estimation module, which receives the thermal diffusivity matrix and gradient field matrix as inputs and uses Fourier's law of heat conduction for real-time estimation; the objective function is constructed as a quadratic function that minimizes the weighted sum of the temperature compensation amount and the heat loss data, and the optimization weight vector is solved by the interior point method; wherein, the heat flow estimation module can be specifically implemented as a software module running on an ARM Cortex-M7 microcontroller, which performs grid-by-grid point operations on the input matrix and outputs heat loss data for constructing the objective function.

[0131] Through the above technical solution, this application enables the optimized weight vector to accurately reflect the actual heat loss, avoid insufficient or excessive local temperature compensation, thereby effectively solving the problem of uneven temperature distribution during baking and improving the drying quality and production efficiency of the paint film.

[0132] In practical applications, some embodiments of this application propose to perform real-time convolution and secondary optimization on the temperature compensation amount and optimization weight vector to generate a temperature optimization scheme. However, in this process, the convolution operation fails to effectively integrate the real-time dynamic characteristics of the wind speed matrix and the thermal conduction response, making it difficult for the generated temperature optimization scheme to accurately adapt to local flow field changes. It is prone to temperature compensation deviation due to wind speed fluctuations or data loss, which in turn causes local overheating or insufficient drying of the coating during flash drying baking, affecting the overall temperature control uniformity and coating quality.

[0133] To address this, this application further proposes real-time convolution and secondary optimization of the temperature compensation amount and the optimization weight vector to generate a temperature optimization scheme, specifically including:

[0134] The temperature compensation amount, the optimization weight vector, and the conduction kernel are convolved to generate the pre-response matrix;

[0135] The conduction kernel is generated by fitting the conduction response to the wind speed matrix and temperature data in an offline manner.

[0136] The pre-response matrix is ​​modified by combining it with the wind speed matrix to generate a modified response matrix. The modified response matrix, temperature compensation amount, and optimization weight vector are then structurally combined to generate a temperature optimization scheme.

[0137] Among them, the conduction kernel refers to the physical model parameter that characterizes the influence of wind speed distribution on heat diffusion. It can be implemented by offline fitting method based on historical flow field and temperature data, such as fitting the nonlinear conduction response relationship between wind speed matrix and temperature data through Gaussian process regression.

[0138] The pre-response matrix can be understood as an initial estimation matrix that reflects the temperature compensation requirements and flow field conduction characteristics. It can be realized by linearly combining the temperature compensation amount, optimization weight vector and conduction kernel using convolution operation, for example, by accelerating the convolution calculation process through cyclic matrix decomposition.

[0139] The corrected response matrix refers to the response matrix after dynamic adjustment based on real-time wind speed. It can be implemented using a weighted correction method based on the wind speed matrix, such as vector compensation of the pre-response matrix according to the wind speed gradient direction.

[0140] Specifically, the proposed solution generates a conduction kernel by fitting the conduction response of the wind speed matrix and temperature data offline. This conduction kernel accurately characterizes the influence of wind speed distribution on heat diffusion. Subsequently, the temperature compensation amount, optimization weight vector, and conduction kernel are convolved to generate a pre-response matrix, which simultaneously reflects the temperature compensation requirement and the flow field conduction characteristics. Based on this, the pre-response matrix is ​​dynamically corrected using the real-time wind speed matrix to generate a corrected response matrix, compensating for heat transfer deviations caused by instantaneous changes in wind speed. Finally, by structurally combining the corrected response matrix, temperature compensation amount, and optimization weight vector, the various elements are systematically integrated to generate a temperature optimization scheme that conforms to both physical constraints and optimization objectives, thereby achieving precise adaptation to local flow field changes.

[0141] As a specific implementation method, the conduction kernel can collect historical spray booth operation data and use a kernel ridge regression model to fit the conduction response relationship between the wind speed matrix and temperature data. In the real-time processing stage, the temperature compensation amount and the optimization weight vector can be used with the conduction kernel to generate a pre-response matrix through a two-dimensional convolution operation. This convolution operation can be implemented in the frequency domain using a fast Fourier transform. Subsequently, the pre-response matrix is ​​vector-corrected according to the current wind speed matrix, for example, by enhancing the response values ​​corresponding to regions where the wind speed direction is consistent with the temperature gradient direction. Finally, the corrected response matrix, temperature compensation amount, and optimization weight vector are structurally integrated through a convex combination method to generate a temperature optimization scheme.

[0142] Through the above solution, this application effectively solves the problem that the convolution operation fails to integrate the dynamic characteristics of wind speed, enabling the temperature optimization scheme to accurately adapt to local flow field changes, avoiding temperature compensation deviations caused by wind speed fluctuations or data loss, thereby preventing local overheating or insufficient drying of the coating film during flash drying and baking, and significantly improving the uniformity of temperature control and coating quality in the spray booth.

[0143] Example 2:

[0144] Please see Figure 5 A smart data-controlled flash-drying baking temperature optimization system for spray paint booths, comprising:

[0145] The data acquisition module is used to acquire wind speed and temperature data of the target object on a grid-by-grid basis;

[0146] The wind speed data processing module is used to perform spatiotemporal fitting on wind speed data within a spatial neighborhood, generate a wind speed matrix, and perform numerical difference processing on the wind speed matrix to generate a vortex intensity matrix.

[0147] The stagnation analysis module is used to determine whether the wind speed data exceeds the preset threshold by using a sliding window. If so, it calculates the time-series proportion of the wind speed data within the sliding window and multiplies the time-series proportion by the wind speed completion coefficient to obtain the stagnation factor matrix.

[0148] The temperature optimization scheme generation module is used to perform multi-threshold hierarchical segmentation of the stagnation factor matrix and sieve processing in combination with the eddy strength matrix to generate stagnation zone data. Based on the stagnation zone data and the wind speed matrix, boundary segmentation is performed. Then, within the boundary segmentation results, the stagnation factor matrix, eddy strength matrix and temperature data are rapidly convolved and estimated to generate a temperature optimization scheme.

[0149] This embodiment has the same technical effects as Embodiment 1.

[0150] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. The data mentioned in this application, when used for calculations, have undergone normalization and other preprocessing to achieve dimensional uniformity.

[0151] 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 optimizing the flash-drying baking temperature of a spray paint booth using intelligent data control, characterized in that, Includes the following steps: Obtain wind speed and temperature data for the target object point by point; The wind speed data is spatiotemporally fitted within a spatial neighborhood to generate a wind speed matrix, and the wind speed matrix is ​​numerically differencing to generate a vortex intensity matrix. If the wind speed data is greater than a preset threshold, the time-series proportion of the wind speed data within the sliding window is calculated, and the time-series proportion is multiplied by the wind speed completion coefficient to obtain the stagnation factor matrix. The wind speed completion coefficient is generated by weighted interpolation of the wind speed data in the spatial neighborhood. The preset discrimination threshold is obtained by performing percentile statistics on the wind speed data contained in the corresponding grid points within the sliding window; The stagnation factor matrix is ​​segmented using a multi-threshold hierarchical method and then sieved using the eddy intensity matrix to generate stagnation zone data. Boundary segmentation is then performed based on the stagnation zone data and the wind speed matrix. Finally, a fast convolution estimation is performed on the stagnation factor matrix, the eddy intensity matrix, and the temperature data within the boundary segmentation results to generate a temperature optimization scheme.

2. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 1, characterized in that: After obtaining the stagnation factor matrix, the process also includes generating the gradient field matrix, specifically including: The temperature data is subjected to central difference calculation between grid points and adjacent grid points, and the central difference calculation results are subjected to weighted interpolation smoothing processing in the spatial neighborhood of the grid points to generate a smooth temperature difference vector. The smoothed temperature difference vector and the stagnation factor matrix are coupled and analyzed according to grid points and time series to generate a gradient field matrix.

3. The method for optimizing the flash-drying baking temperature of a spray booth with intelligent data control according to claim 2, characterized in that: The stagnation factor matrix is ​​subjected to multi-threshold hierarchical segmentation, and sieved in conjunction with the eddy strength matrix to generate stagnation zone data. Specifically, this includes: The stagnation factor matrix is ​​subjected to stratified determination using a multi-threshold set, and the confidence level of the stratified determination results is calculated based on the wind speed completion coefficient and the wind speed matrix at grid points. The multi-threshold set is obtained by performing distribution statistics on the stagnation factor matrix; Determine whether the confidence score calculation result is less than the preset confidence score threshold; otherwise, remove the stratification judgment result corresponding to the grid point and generate stagnation zone data.

4. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 3, characterized in that: After generating the stagnation zone data, the process also includes generating the thermal diffusivity matrix, specifically including: The temperature data is subjected to a difference operation within adjacent sliding windows, and the difference operation result is combined with the temperature data within the sliding window to form an observation vector; The gradient field matrix and the wind speed matrix are weighted and calculated point by point to generate a coefficient matrix; The thermal diffusivity matrix is ​​obtained by constructing and solving a regularized inverse problem on the observation vector and the coefficient matrix.

5. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 4, characterized in that: Boundary segmentation based on the stagnation zone data and the wind speed matrix specifically includes: Based on the hierarchical determination results, the stagnation zone boundaries are extracted from the stagnation zone data to generate an initial seed set; The wind speed matrix is ​​evaluated for consistency in speed magnitude and direction between adjacent grid points to generate an adjacency matrix; Based on the stratification determination results, the thermal diffusivity matrix is ​​normalized and inverted within the region according to the grid points to generate the cost matrix. The initial seed set, the adjacency matrix, and the cost matrix are subjected to boundary segmentation through graph cut iterative optimization with connectivity constraints until a preset number of iterations is reached, generating the boundary segmentation result.

6. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 5, characterized in that: Within the boundary segmentation result, a fast convolution estimation is performed on the stagnation factor matrix, the vortex intensity matrix, and the temperature data to generate a temperature optimization scheme, specifically including: Within the boundary segmentation result, the thermal diffusivity matrix, the stagnation factor matrix, the eddy current matrix, and the temperature data are normalized and weighted to generate a temperature compensation amount. A convex optimization problem is constructed and solved for the temperature compensation amount to obtain the optimization weight vector; The temperature compensation amount and the optimization weight vector are subjected to real-time convolution and secondary optimization to generate a temperature optimization scheme.

7. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 6, characterized in that: Within the boundary segmentation result, the thermal diffusivity matrix, the stagnation factor matrix, the eddy current matrix, and the temperature data are normalized and weighted to generate a temperature compensation amount, specifically including: Within the boundary segmentation result, the thermal diffusivity matrix, the stagnation factor matrix, the time spectrum data, the eddy current matrix, and the temperature data are normalized and weighted to generate a temperature compensation amount. The time spectrum data is generated by performing a short-time Fourier transform on the vortex intensity matrix within a sliding window.

8. The method for optimizing the flash-drying baking temperature of a spray painting booth with intelligent data control according to claim 6, characterized in that: A convex optimization problem is constructed and solved for the temperature compensation amount, resulting in the optimization weight vector, which specifically includes: Using the temperature compensation amount and heat loss data as input, an objective function is constructed and solved to obtain the optimized weight vector of the boundary segmentation result; The heat loss data is generated by estimating the heat diffusivity coefficient matrix and the gradient field matrix using heat flow estimation.

9. The method for optimizing the flash-drying baking temperature of a spray booth with intelligent data control according to claim 6, characterized in that: The temperature compensation amount and the optimization weight vector are subjected to real-time convolution and secondary optimization to generate a temperature optimization scheme, specifically including: The temperature compensation amount, the optimization weight vector, and the conduction kernel are used to generate a pre-response matrix through convolution; The conduction kernel is generated by fitting the conduction response of the wind speed matrix and the temperature data in an offline state. The pre-response matrix is ​​modified in combination with the wind speed matrix to generate a modified response matrix. The modified response matrix, the temperature compensation amount, and the optimization weight vector are then structurally combined to generate a temperature optimization scheme.

10. A smart data-controlled system for optimizing the flash-drying and baking temperature of a spray painting booth, characterized in that, include: The data acquisition module is used to acquire wind speed and temperature data of the target object on a grid-by-grid basis; The wind speed data processing module is used to perform spatiotemporal fitting on the wind speed data in the spatial neighborhood to generate a wind speed matrix, and to perform numerical difference processing on the wind speed matrix to generate a vortex intensity matrix. The stagnation analysis module is used to determine whether the wind speed data is greater than a preset discrimination threshold by sliding window. If so, the time-series proportion of the wind speed data in the sliding window is counted, and the time-series proportion is multiplied by the wind speed completion coefficient to obtain the stagnation factor matrix. The temperature optimization scheme generation module is used to perform multi-threshold hierarchical segmentation on the stagnation factor matrix and combine it with the vortex strength matrix for screening to generate stagnation zone data. Based on the stagnation zone data and the wind speed matrix, boundary segmentation is performed, and then the stagnation factor matrix, the vortex strength matrix and the temperature data are rapidly convolved and estimated within the boundary segmentation results to generate a temperature optimization scheme.

Images