A cabinet anti-sagging adaptive temperature control curve generation method and system

Abstract

The present application relates to a kind of cabinet anti-sagging self-adaptive temperature control curve generation method and system, comprising: extracting paint flow tendency index from potential stress concentration area, using neural network model to predict the risk distribution of sagging, and obtaining risk assessment map;From the coordinated curve version, select the key time node, obtain the corresponding material layer thermal behavior data, use support vector machine model to verify the deformation synchronism, and obtain the verification result;According to the verification result, if the synchronism is insufficient, then integrate additional thermal coupling constraint, re-simulate the entire heating process, and determine the final self-adaptive temperature control curve.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method and system for generating an adaptive temperature control curve to prevent cabinet dripping. Background Technology

[0002] In the field of customized home furnishing manufacturing, the spraying process of cabinet surfaces is a crucial step in determining the final aesthetics and durability. During the curing process, the paint may experience a downward flow due to gravity, resulting in "sagging," which severely impacts product quality. Therefore, precise temperature control to prevent sagging is a core requirement for improving production quality. Currently, commonly used temperature control methods often set based on the average temperature of a single material or the overall structure, failing to fully consider the complex structure of cabinets, which is composed of multiple materials. When heated, the expansion and deformation of different parts of this structure are not synchronized, making it difficult for a single temperature profile to simultaneously meet the curing requirements of all material layers. This can easily lead to stress concentration or localized overheating at the joints between layers, indirectly exacerbating the paint's tendency to flow.

[0003] The deep-seated technical challenge of this problem stems first from the differentiated thermal response characteristics exhibited by the various material layers of the cabinet during heating. For example, the wood fibers of the substrate layer, the polymer of the edge banding layer, and the surface coating layer all react differently to temperature changes and undergo varying physical transformations. This difference is not isolated; it directly leads to the second core challenge: during the heating process of the composite structure, the thermal behavior of each layer generates complex interactions and constraints. For instance, the slower heat conduction of the edge banding delays the curing process of the coating in its adjacent areas, while the rapid thermal expansion of the substrate compresses the coating that is not yet fully cured. This cross-layer, dynamic thermo-mechanical coupling makes predicting and controlling overall deformation exceptionally difficult. Specifically on the production line, when an oven heats a cabinet door panel composed of a solid wood substrate, PVC edge banding, and polyester coating using a fixed program, coating accumulation at the edge banding joints or excessively thin coating in the center of the panel often occurs. This is precisely due to the failure to coordinate the asynchronous deformation of each material layer during heating.

[0004] Therefore, how to generate a temperature control curve that can adapt to the characteristics of the composite structure of the cabinet and coordinate the differentiated thermal behavior of its various material layers, so that the heating process can ensure good curing of the coating and effectively suppress sagging caused by structural deformation incoordination, has become the key issue to break through the current process bottleneck and achieve high-quality coating. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for generating an adaptive temperature control curve for preventing paint dripping in kitchen cabinets, in order to solve the technical problems of deformation interaction, stress concentration and paint dripping risk caused by the difference in thermal conductivity and expansion coefficient of material layers in the composite structure of kitchen cabinets during the heating process.

[0006] The technical solution of the present invention is as follows:

[0007] This invention provides a method for generating an adaptive temperature control curve to prevent cabinet dripping, mainly including:

[0008] Data on the material layers of the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer, are acquired. Temperature distribution during the heating process is collected in real time using sensors to obtain initial thermal response characteristics. Based on these characteristics, finite element analysis is used to simulate the deformation interaction of each material layer during heating. If the deformation difference exceeds a preset threshold, simulation parameters are adjusted to identify potential stress concentration areas. Paint flow tendency indicators are extracted from these potential stress concentration areas, and a neural network model is used to predict the distribution of sagging risks, resulting in a risk assessment map. A preliminary draft temperature control curve is generated based on the risk assessment map. Iterative optimization algorithms refine the curve segments. If local overheating indicators exceed the threshold, the draft is revised to determine a coordinated curve version. Key time points are selected from the coordinated curve version, and corresponding material layer thermal behavior data are acquired. A support vector machine model is used to verify deformation synchronicity, yielding verification results. Based on the verification results, if synchronicity is insufficient, additional thermo-mechanical coupling constraints are incorporated, and the entire heating process is re-simulated to determine the final adaptive temperature control curve.

[0009] This invention provides a cabinet anti-sagging adaptive temperature control curve generation system, mainly comprising: a data acquisition module for acquiring material layer data of the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer, and collecting the temperature distribution during the heating process in real time through sensors to obtain the initial thermal response characteristics; a finite element simulation module for simulating the deformation interaction of each material layer during heating using finite element analysis based on the initial thermal response characteristics, and adjusting the simulation parameters if the deformation difference exceeds a preset threshold to determine potential stress concentration areas; and a sagging risk assessment module for extracting paint flow tendency indicators from potential stress concentration areas and using a neural network model to predict the sagging risk distribution. The system generates a risk assessment map; a temperature control curve generation module generates a preliminary draft temperature control curve based on the risk assessment map, refines the curve segments through iterative optimization algorithms, and backtracks to modify the draft if the local overheating index exceeds the threshold, thus determining the coordinated curve version; a deformation synchronicity verification module selects key time nodes from the coordinated curve version, obtains the corresponding material layer thermal behavior data, verifies deformation synchronicity using a support vector machine model, and obtains the verification results; an adaptive temperature control curve determination module, based on the verification results, determines the final adaptive temperature control curve by incorporating additional thermo-mechanical coupling constraints and resimulating the entire heating process if synchronicity is insufficient.

[0010] The beneficial effects of this application are as follows: A method and system for generating an adaptive temperature control curve to prevent sagging in kitchen cabinets. During use, temperature distribution is collected in real time by sensors, and deformation interaction is simulated using finite element analysis to accurately identify potential stress concentration areas. A neural network model is then used to predict the distribution of sagging risk, forming a risk assessment map. Based on this, the invention generates and optimizes the temperature control curve, incorporates thermo-mechanical coupling constraints, verifies deformation synchronization through support vector machines to ensure the coordination of the heating process, re-simulates the entire heating process, and determines the final adaptive temperature control curve. The core innovation of this invention lies in the generation of the adaptive temperature control curve, which effectively suppresses sagging risk, improves production quality stability, and provides reliable assurance for the heating process of composite structures. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating a specific embodiment of the method for generating an adaptive temperature control curve to prevent dripping in a kitchen cabinet according to the present invention. Figure 2 This is a schematic diagram of a method for generating an adaptive temperature control curve to prevent dripping in a kitchen cabinet according to the present invention; Figure 3 This is another schematic diagram of a method for generating an adaptive temperature control curve for preventing dripping in a cabinet according to the present invention; Figure 4 This is a schematic diagram of the structure of a cabinet anti-drip adaptive temperature control curve generation system according to the present invention. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described embodiments are merely some embodiments of the invention, and not all embodiments. The components of the embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0013] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0014] It should be noted that 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0015] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0016] A specific embodiment of the method and system for generating an anti-drip adaptive temperature control curve for kitchen cabinets according to the present invention:

[0017] like Figures 1-3 This embodiment of a method for generating an adaptive temperature control curve to prevent cabinet dripping may specifically include:

[0018] S101. Obtain material layer data of the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer, and obtain the initial thermal response characteristics by collecting the temperature distribution during the heating process in real time through sensors.

[0019] Data is collected from the cabinet composite structure using a sensor system, recording the thermal conductivity and expansion coefficient of each material layer to obtain a preliminary physical parameter dataset. Based on this dataset, the thermal conductivity and expansion coefficient are categorized using preset threshold ranges to determine the performance range of each material layer under different temperature conditions. If the categorized performance range exceeds the preset threshold range, outliers are removed using a data filtering module, resulting in a corrected set of physical parameters. For this corrected set, combined with real-time temperature distribution data, a support vector machine algorithm is used to predict the thermal response value and determine potential trends during heating. Based on the predicted trends in thermal response values, the correlation between temperature distribution and material layers during heating is analyzed to obtain the thermal stress distribution results for each layer. Using the thermal stress distribution results, combined with environmental factors, the stability of the composite structure is assessed, and the thermal response characteristics of key areas are determined. If the thermal response characteristics of key areas do not conform to the preset safety range, the temperature distribution acquisition frequency is adjusted using a data mapping module to obtain a more accurate thermal response dataset.

[0020] In one possible implementation, the thermal conductivity rate and expansion coefficient are classified according to a preset threshold range based on a preliminary physical parameter dataset to determine the performance range of each material layer under different temperature conditions. The preset threshold range is a critical value set for high thermal conductivity / low thermal conductivity and high expansion / low expansion, used to distinguish the substrate, edge banding and coating layer.

[0021] For example, in a laboratory environment or during the pre-production stage, sensors record the thermal conductivity and coefficient of thermal expansion of the material within the range of 20°C to 180°C (typical coating curing range). The temperature axis is divided into several performance intervals (e.g., heating interval, isothermal interval, cooling interval), and the stability of the material's physical parameters within each interval is determined.

[0022] In one possible implementation, if the performance range after classification exceeds a preset threshold range, an outlier is removed by a data filtering module to obtain a corrected set of physical parameters. The preset threshold range functions as validity filtering, referring to the allowable deviation range or safety boundary used to determine whether the real-time data collected by the sensor has caused unreasonable jumps due to interference or sensor malfunction. The data filtering module specifically comprises: a buffer for temporarily storing the raw physical parameter dataset collected by the sensor in real time; a comparator for comparing the collected real-time data with the preset classification performance range in real time; and a logic processor for executing "determination-execution" logic, i.e., initiating the removal procedure when data falls outside the threshold range. The data filtering module can remove noise interference through methods such as static deviation removal, dynamic smoothing filtering, or SVM (Support Vector Machine) assisted determination, and obtain a corrected set of physical parameters to provide an accurate basis for subsequent thermal response prediction. When frequent anomalies cause data discrepancies, it will also work with the data mapping module to adjust the sensor's acquisition frequency to improve data accuracy in key areas.

[0023] In one possible implementation, before prediction, the system first prepares high-quality input data through pre-processing steps, using a modified set of physical parameters (thermal conductivity and coefficient of thermal expansion of each material layer) and real-time monitored temperature distribution data. The target variable is set as the thermal response value, i.e., the physical state feedback of the material after heating. Then, based on the modified set of physical parameters and the real-time monitored temperature distribution data, a support vector machine algorithm is used to predict the thermal response value and determine the potential changing trend during the heating process.

[0024] Furthermore, based on the predicted trends in thermal response values, the correlation between temperature distribution and material layers during the heating process is analyzed. This involves using real-time monitored temperature distribution data as an independent variable and spatiotemporally matching it with the predicted response values ​​of different material layers (substrate, edge banding, coating). For example, the coupling relationship between the rate of temperature rise of the edge banding layer (PVC) and the heat release during coating curing is analyzed when oven heat reaches it. By comparing the response curves of different material layers under the same thermal environment, the delay time of heat conduction is calculated. This correlation analysis can identify which layers delay the curing of adjacent coatings due to slow thermal conductivity (such as edge banding). Environmental factors (such as ambient humidity and wind speed) are introduced during the analysis to dynamically adjust the weighting of the influence of temperature distribution on the stability of each material layer.

[0025] After obtaining the aforementioned correlations, the system transforms them into thermal stress distributions at each layer through physical modeling and numerical calculations: based on the specific expansion coefficients of each material layer, it calculates the asynchronous deformation caused by temperature correlations. For example, when the substrate (wood) expands rapidly while the coating layer is not fully cured, shear stress will be generated at the interface between the two. Using thermal response correlation data, combined with a modified set of physical parameters (thermal conductivity, expansion coefficient), a thermo-mechanical coupling equation for each material layer is constructed. The analyzed temperature correlation trends are then transformed into thermal loads applied to the model. By calculating the changes in internal forces at each layer under thermal loads, a complete thermal stress distribution result is output. The system then evaluates the overall stability of the composite structure based on the thermal stress distribution results and automatically identifies the point with the highest stress value, thus determining the thermal response characteristics of the key region. Thermal response characteristics refer to the set of physical feedback exhibited by the region under specific temperature conditions, including dynamic data such as deformation, heating rate, heat conduction delay time, and real-time expansion displacement.

[0026] The specific steps for assessing the stability of the composite structure are as follows: The system evaluates the stability of the cabinet composite structure through cross-analysis of multi-dimensional data. Specifically, the first step is to map the calculated thermal stress distribution results of each layer onto the physical structural model of the cabinet. The second step is to introduce environmental influencing factors (such as the ambient temperature, humidity, and wind speed of the current production line) as correction parameters. For example, the expansion coefficient of the wood substrate will change nonlinearly under high humidity conditions, and the system adjusts the stability value accordingly. The third step is to compare the shear stress between each layer with the adhesive force during the material curing process to determine whether there is a risk of interlayer delamination, uncontrolled deformation, or paint sagging.

[0027] In one possible implementation, if the thermal response characteristics of a critical area do not conform to a preset safety range, the data mapping module adjusts the temperature distribution sampling frequency to obtain a more accurate thermal response data set. The preset safety range is a standard used to measure whether the thermal response characteristics of the critical area are acceptable. The specific value varies depending on the material combination, but in practice, it is usually defined as deformation synchronization error <5% or thermal stress value < 60% of the material's yield strength. If the monitored thermal response characteristics (such as deformation) exceed this safety boundary, the system determines it to be inconsistent and enters the frequency adjustment stage. That is, when the deviation value approaches or exceeds the safety range, the data mapping module triggers an encryption command to increase the sensor's sampling frequency from the conventional (e.g., 1Hz) to 10Hz or higher. By obtaining a more accurate thermal response data set with higher time resolution, more detailed boundary conditions are provided for subsequent steps such as finite element simulation (S102) and sagging prediction (S103), thereby achieving precise locking of potential stress concentration areas.

[0028] The data mapping module, the core of the system's adaptive adjustment, consists of three parts: 1. Feature comparison unit: receives real-time thermal response data from key areas and compares it with the safe range in real time. 2. Mapping algorithm engine: establishes a mathematical function mapping relationship between the deviation value and the acquisition frequency. 3. Frequency controller: directly issues commands to the sensor hardware to change its sampling period.

[0029] S102. Based on the initial thermal response characteristics, finite element analysis is used to simulate the deformation interaction of each material layer during heating. If the deformation difference exceeds the preset threshold, the simulation parameters are adjusted to determine the potential stress concentration area.

[0030] Acquire initial thermal response characteristic data of the material layers. Establish a thermo-mechanical coupling model of the material layers using the finite element method. Apply a heating load to the model and calculate the deformation of each material layer. Determine whether the difference in deformation between the material layers exceeds a preset threshold. If the difference exceeds the preset threshold, adjust the simulation parameters of the material layers. Based on the adjusted simulation parameters, determine the locations of potential stress concentration regions.

[0031] In one possible implementation, a thermo-coupling model of the material layers is constructed to simulate the interaction of the cabinet composite structure (such as a wood substrate, PVC edge banding, and paint layer) when heated. Specific steps include: establishing a geometric model of the multi-layered composite structure based on the actual physical structure of the cabinet, defining the spatial location and thickness of the substrate, edge banding, and paint layer; assigning corresponding physical properties to each geometric layer, including the thermal conductivity rate and coefficient of thermal expansion obtained from step S101; discretizing the model using finite element meshing technology, and refining the mesh at the interfaces of different materials (such as edge banding joints) to improve computational accuracy; and finally, setting contact and constraints, i.e., defining the contact relationships between the material layers (such as fixed or sliding constraints) to simulate the interlayer bonding strength.

[0032] The process involves applying a heating load to the model and calculating the deformation of each material layer. Specifically, the temperature distribution data from the initial thermal response characteristics is applied as a thermal load to the model boundary. Temperature typically changes over time, simulating heating, isothermal, and cooling stages. Further thermo-mechanical coupling calculations are performed: first, the temperature field distribution within the model is calculated using the heat conduction equation. Based on the calculated temperature field, the temperature change is converted into thermal strain using the thermoelastic constitutive equation. Then, combining the elastic modulus and Poisson's ratio of each layer, the displacement vector of each node, i.e., the deformation of each material layer, is obtained by solving a system of linear finite element equations.

[0033] Furthermore, if the deformation difference exceeds a preset threshold, i.e., exceeds the deformation difference rate (e.g., interlayer relative displacement exceeds 0.1-0.5 mm) or the strain mismatch threshold, the system will automatically backtrack and recalibrate the model properties based on the corrected physical parameter set in S101. The interaction stiffness or thermal response damping of the material layers will be adjusted to make the simulation results more closely match the actual acquired initial thermal response characteristics. Based on the adjusted accurate model, the system identifies risk points through the following logic: scanning the thermal stress values ​​across the entire model field to identify the region where stress changes most drastically with spatial location (maximum gradient); and finding locations where the deformation displacement vector directions of different material layers contradict each other or have extremely large differences in magnitude. The system will automatically extract the spatial coordinates of these high-stress or high-difference points and define them as potential stress concentration areas.

[0034] S103. Extract coating flow tendency index from potential stress concentration areas, use neural network model to predict sagging risk distribution, and obtain risk assessment map.

[0035] By acquiring stress concentration data from potential areas, automated tools are used to perform preliminary data cleaning, such as noise reduction, outlier removal, and formatting, to obtain a pre-processed stress dataset. Based on this dataset, key indicators related to flow tendency are extracted, and a pre-established feature screening method is used to determine a set of indicators highly correlated with sagging risk. A neural network model is constructed for training and prediction of this indicator set to obtain the distribution probability value of sagging risk, generating intermediate results for risk distribution. From these intermediate results, combined with regional analysis data, a grid generation technique is used to map the distribution probability values ​​to specific regions, resulting in regionalized risk distribution data. This regional analysis data refers to the spatial location attributes of the cabinet composite structure, including the cabinet's geometric boundaries, the precise coordinate range of different material layers (such as substrate and edge banding), and the elevation / planar distribution information under gravity. Common grid generation techniques are mainly divided into structured grid methods (such as single-block, multi-block structured grids, and body-fitted grids), unstructured grids (such as Delaunay triangulation, leading-edge advancement, and quad / octree grids), and hybrid grids. If the probability values ​​of certain areas in the regionalized risk distribution data exceed a preset threshold, these areas are marked as high-risk, generating a labeled risk distribution dataset. Using this labeled risk distribution dataset, visualization techniques are employed to generate a risk assessment map, determining the final spillway risk assessment result.

[0036] In one possible implementation, based on the pre-processed stress dataset, key indicators related to flow tendency are extracted. A pre-established feature screening method is used to determine a set of indicators highly related to sagging risk. These key indicators typically include: local shear stress (internal forces generated by potential stress concentration areas determined by S102); gravity component projection (gravity vector distribution of the paint on the cabinet facade or corners); viscosity change rate prediction (the consistency index of the paint that decreases with increasing temperature); and surface tension gradient (uneven surface tension of the paint due to local overheating).

[0037] To improve the prediction accuracy of the neural network model, the system needs to filter out the variables that contribute most to sag risk from massive amounts of raw data. A pre-established feature screening method is used, which is a screening mechanism based on statistical correlation or sensitivity analysis. The specific implementation logic is as follows: mathematical methods (such as Pearson correlation coefficient or mutual information method) are used to calculate the correlation strength between each candidate indicator and historical sag samples. Redundant and low-contribution indicators are eliminated, retaining the set of indicators highly correlated with sag risk. The input variables that finally enter the neural network model are then selected to ensure that the model can output accurate sag risk distribution probability values.

[0038] In one possible implementation, the neural network model is constructed for training and prediction on a set of indicators to obtain the probability distribution value of sagging risk and generate intermediate results of risk distribution. The neural network model is mainly used to handle the complex mapping relationship between nonlinear, multi-dimensional physical indicators and sagging phenomena. Specifically, the input layer consists of the set of indicators determined in the preceding steps, typically including local shear stress, gravity components, coating viscosity change rate, and surface tension gradient. The hidden layer constructs a multi-layer deep sensing structure, using a nonlinear activation function (such as ReLU) to capture the nonlinear contribution of each physical indicator to the coating flow during heating. The output layer sets the output dimension to match the risk prediction requirements, typically using a Softmax or Sigmoid activation function to convert the output result into a probabilistic form. Model training utilizes pre-collected process sample data labeled "sagging / no sagging" for supervised learning, continuously optimizing network weights through backpropagation to establish a prediction model for "physical stress-temperature-sagging risk."

[0039] Specifically, the real-time extracted potential area stress dataset and key indicators are input into the trained model. The neural network calculates a weighted average of the input indicators and outputs a value between 0 and 1, representing the probability of dripping under specific physical conditions. For example, a probability value of 0.85 indicates a very high probability of dripping at that location. Further, the system preliminarily correlates the probability values ​​predicted by the model for all points with their corresponding physical coordinates, forming a raw distribution list containing "coordinate-probability" pairs. At this stage, while the data contains risk information, it has not yet achieved precise spatial alignment with the actual physical areas of the cabinet (such as edge banding and panels). The intermediate result refers to the raw risk dataset with probabilistic features, before undergoing area mapping and gridding processing.

[0040] In one possible implementation, if the probability value of certain areas in the regionalized risk distribution data exceeds a preset threshold, those areas are marked as high-risk, generating a risk distribution dataset with the labels. The preset threshold typically needs to be dynamically set based on the specific paint type (e.g., polyester paint, UV paint), coating thickness, and the cabinet placement angle. However, in actual industrial implementation, this threshold is usually set according to the following logic:

[0041] Risk probability threshold: Generally set between 0.7 and 0.85 (70%-85%). When the neural network predicts that the probability of sagging in a certain area exceeds this value, the system considers that the physical stress and temperature distribution at that location are sufficient to disrupt the surface tension balance of the coating.

[0042] Industry experience calibration: This threshold can also be reverse calibrated using a high-quality threshold segment in the historical process log (S109).

[0043] Specifically, when the regionalized risk distribution data exceeds the aforementioned threshold, the system will perform the following operations: mark the corresponding coordinate area as a "red zone" on the digital twin model or risk assessment map. These marked areas will directly become the key monitoring targets of the iterative optimization algorithm in step S104. If the local overheating index of these high-risk areas remains higher than the value in the subsequently generated temperature control curve, the system will force a retrospective modification of the draft temperature control.

[0044] S104. For the risk assessment map, generate a preliminary draft temperature control curve, refine the curve segments through iterative optimization algorithms, and if the local overheating index is higher than the threshold, backtrack and modify the draft to determine the coordinated curve version.

[0045] By leveraging risk assessment data and map analysis, temperature distribution information for key areas is obtained to establish a preliminary temperature control curve framework. Based on this framework, an iterative optimization algorithm is used to refine the curve segments, resulting in adjusted curve details. For these adjusted details, temperature index data for local areas is acquired to determine if overheating exceeding preset thresholds exists. If a local area's temperature index exceeds the preset threshold, a backtracking adjustment mechanism is triggered to recalculate the curve segments and determine the corrected curve shape.

[0046] In one possible implementation, the risk assessment data is used to obtain temperature distribution information of key areas from map analysis to determine a preliminary temperature control curve framework. The risk assessment data consists of probability values ​​generated by the system using a neural network and mapped to the physical coordinates of the cabinet through grid partitioning technology. The system then performs in-depth analysis of the map to extract structured risk assessment data, which includes parameters such as spatial coordinates, sag risk probability values, gravity sensitivity weights, and stress levels. The key areas in the temperature distribution information refer to specific spatial segments marked as high-risk or with probability values ​​exceeding a threshold on the risk assessment map. The temperature distribution information specifically includes: local real-time temperature values, i.e., the current surface temperature collected by sensors at these risk points; a temperature gradient, i.e., the rate of temperature change over time in this area (ΔT / Δt), a key indicator affecting the sharp decrease in paint viscosity; interlayer temperature difference, i.e., the temperature conduction gradient between the substrate (wood), edge banding (PVC), and paint layer, reflecting the uneven distribution of heat at the material interfaces; and a thermal hysteresis parameter, i.e., the feedback delay time describing the key area's response to oven ambient temperature adjustment.

[0047] After acquiring the above data, the system establishes key turning points (such as preheating, solvent evaporation, curing, and cooling stages) on the timeline based on the temperature gradient requirements of critical areas. Temperature constraint boundaries are set using the "sagging risk probability" from the risk assessment data as a reverse constraint. For example, if a certain segment is predicted to have an extremely high sagging risk, the framework will automatically limit the peak temperature of that segment or slow down the heating rate. Combining the initial thermal response characteristics of the material layers, the basic power output level of each heating zone in the oven is determined, forming a preliminary draft of the temperature control curve.

[0048] In one possible implementation, based on the initial temperature control curve framework, an iterative optimization algorithm is used to refine the curve segments to obtain adjusted curve details. Specifically, the system divides the initially determined temperature control curve framework into several small time segments (e.g., each step is 10-30 seconds) according to the time axis. Each segment corresponds to a different spatial position and heating intensity of the cabinet on the production line.

[0049] In one possible implementation, the temperature index data of local areas is obtained for the adjusted curve details to determine whether there is overheating exceeding a preset threshold. If the temperature index of a local area exceeds the preset threshold, a backtracking adjustment mechanism is triggered to recalculate the curve segment and determine the corrected curve shape. Specifically, for the high-risk areas marked in S103 (such as edge sealing joints and facade coatings), a local optimization target is set for that period. The target is usually to reduce the temperature rise slope or peak temperature of the area as much as possible while ensuring the coating is fully cured. The algorithm will fine-tune (perturb) the temperature setpoint of the segment. For example, reduce the output power of the infrared heater or adjust the frequency of the circulating fan during that period. The system simulates the physical performance of the key areas under the new temperature parameters. The local overheating index is calculated, which is usually weighted by the following parameters: temperature gradient deviation (the degree of proximity of the local temperature to the material's safe upper limit) and heat accumulation rate (the cumulative effect of heat absorbed by the material during that period). If the index is higher than the preset threshold, it is determined that the refinement scheme still has the risk of sagging or material damage. At this point, the system triggers a backtracking adjustment mechanism, automatically returning to the previous time node to readjust the power parameters or extend the heating time of that segment. If the indicators meet the requirements, the optimization parameters for the current segment are saved, and optimization begins for the next time segment. When all curve segments have completed local optimization and the local overheating indicators at all points are below the threshold, the algorithm stops iterating.

[0050] Through the above iterative processing, the originally smooth curve is transformed into a temperature control sequence with fine fluctuations and compensation characteristics, such as avoidance strategies: actively lowering the temperature at high-risk sagging times; and compensation strategies: mitigating thermal stress by gradually increasing the temperature during periods of large material expansion differences. This refined curve is called the coordinated curve version, which is an important input data for subsequent S105 verification of deformation synchronization.

[0051] The preset threshold refers to a local overheating threshold or a critical temperature rise rate threshold, which is typically set as the critical temperature at which the coating undergoes a "drastic change in thermo-rheological properties" (for example, the viscosity of some polyester coatings drops sharply above 65°C). The retrospective adjustment mechanism is a non-linear logical feedback program, its core being "breaking the current linear sequence and reconstructing the affected time window." Specifically, the system first determines which time segment (curve segment) caused the local overheating. Due to the thermal inertia of the temperature control process, overheating at the current moment is often caused by excessive heating power in the previous stage. The mechanism automatically retrospectively traces back 1-2 time points before the fault point. The priority of "preventing sagging" is elevated above "production efficiency," forcibly limiting the upper limit of energy output in the retrospective segment.

[0052] When the backtracking is triggered, the system regenerates the curve shape according to the following steps: The corresponding heater output power parameter (load) in the original draft is set as variable P. A hard constraint is set as "local overheat index ≤ preset threshold". An iterative optimization algorithm (such as gradient descent or heuristic search) is used to search for a new P' value within the constraint range. The new thermal response trend is calculated to verify the temperature peak of the key area under this power. If the temperature of a sensitive section is reduced, the system automatically calculates whether the heating time needs to be extended in the subsequent "safe section" to ensure that the total curing energy (heat accumulation) of the coating meets the standard. Spline interpolation or polynomial fitting methods are used for smoothing to ensure a smooth transition between the adjusted power jump points and avoid impacting the production line inverter. Finally, a multi-dimensional array containing time, temperature, and heater frequency is output, which is the coordinated curve version.

[0053] This process ensures that the temperature control curve is not a rigid setpoint, but an adaptive sequence generated through "simulation detection - over-limit backtracking - recalculation compensation". It fundamentally solves the problem of sagging caused by local heat accumulation in traditional processes.

[0054] Furthermore, by analyzing the corrected curve shape, overall coordination assessment data is obtained to determine whether the control requirements of the risk assessment are met. Based on the overall coordination assessment data, the curve version is finally calibrated to obtain a temperature control curve result that meets the needs of map analysis. The final calibrated temperature control curve result is then obtained, and combined with the dynamically updated data from the risk assessment, a curve application scheme for continuous monitoring is determined.

[0055] In one possible implementation, the final calibrated temperature control curve result is obtained, and combined with dynamically updated risk assessment data, a continuous monitoring curve application scheme is determined. Specifically, the system not only issues the temperature control curve but also sets dynamic fluctuation values ​​for each node on the curve based on risk assessment data. For example, in low-risk areas, a larger temperature fluctuation range is allowed. In high-risk areas (marked as red zones where sag is likely), the application scheme will forcibly narrow the monitoring value, triggering intervention if it deviates by 1-2°C. Based on the dynamic updates of the risk assessment, the application scheme includes multiple contingency curves, such as a normal mode: operating according to the calibrated main curve; and a compensation mode: if an increase in risk data is detected (e.g., slower drying due to ambient humidity), the application scheme automatically switches to a finely adjusted curve segment, mitigating the risk by reducing conveyor belt speed or increasing local airflow. The scheme also specifies the monitoring frequency and recording logic, aligning the real-time temperature curve execution with production quality indicators, providing a data source for generating complete process logs and subsequent alarm mechanisms.

[0056] The finalized continuous monitoring application scheme includes: 1. Control parameter set: heating zone power, conveyor line speed, exhaust pressure, etc. 2. Monitoring priority map: focusing on monitoring potential stress concentration areas and high sagging risk areas identified in S102 / S103. 3. Dynamic correction rules: how the system should correct curve parameters online when real-time thermal behavior data differs from the prediction model. This application scheme will eventually be integrated into the continuous monitoring system. If log data shows abnormal fluctuations in a certain indicator (exceeding the preset value), the system will trigger: an alarm mechanism to send a notification signal; strategy update, extracting high-risk category data, and reversing the value adjustment strategy of the monitoring system to achieve self-evolution of the algorithm. In this way, the application scheme realizes the transformation from "open-loop production" to "closed-loop intelligent monitoring," ensuring that each cabinet door panel can be cured under optimal temperature control. S105: Select key time nodes from the coordinated curve version, obtain the corresponding material layer thermal behavior data, and use a support vector machine model to verify deformation synchronization and obtain verification results.

[0057] Key time points are determined based on the coordinated curve version. Thermal behavior data of the material layer is acquired using these key time points. Deformation indices are calculated based on the thermal behavior data. A support vector machine (SVM) model is used to process the deformation indices and the time point sequence. If the output of the SVM model is higher than a preset threshold, deformation synchronization is considered valid. Verification result data is generated based on the judgment result. The verification result data is used to update the curve version.

[0058] In one possible implementation, the system determines key time points based on the inflection points of physical characteristics in the coordinated curve version. These include: the point of drastic temperature change, selected as the time point with the largest heating slope (highest ΔT / Δt), where the difference in thermal expansion between materials is most easily amplified; the phase change / curing critical point, selected based on material layer data, at the time point when the coating enters the curing exothermic stage or the sealing strip reaches its glass transition temperature; and the peak temperature point, the highest temperature point in each segment of the temperature control curve, used to verify structural stability under extreme thermal loads.

[0059] In one possible implementation, the process of converting thermal behavior data into deformation index calculations is as follows: Using the thermo-mechanical coupling model established in S102, the temperature load at key time points is input, and the displacement vector d of each node in each material layer is extracted. Relative variation calculation: The displacement difference between adjacent material layers (such as coating layer and sealing layer) at the same coordinate point is calculated. Normalizing the deformation difference yields a quantified deformation index (e.g., relative displacement per unit temperature change). This index reflects the degree of coordination between interlayer deformations; the smaller the value, the better the synchronicity.

[0060] In one possible implementation, a Support Vector Machine (SVM) model is used to process the deformation index and time point sequence. If the output of the SVM model is higher than a preset threshold, deformation synchronization is determined to be valid. Specifically, the extracted deformation index is used as a feature vector, and the corresponding time point sequence is used as a time feature, both input into a pre-trained SVM classification model. Internally, the model maps these data to a high-dimensional space using a kernel function to find the optimal hyperplane that can distinguish between synchronization and desynchronization. The output of the SVM is a confidence score or classification probability value. This value represents the reliability of maintaining synchronous deformation among different material layers under the current thermal conditions. A preset threshold is set as the classifier's decision threshold (e.g., 0.8 or 80%). If the output score > 0.8, deformation synchronization is determined to be valid; if the output score < 0.8, synchronization is considered insufficient, and subsequent step S106 correction needs to be triggered.

[0061] In one possible implementation, the verification result data generated based on the judgment result is a structured report set containing the following data: synchronization judgment conclusion (binary identifier), key point offset record (recorded at key time nodes, showing the specific displacement deviation values ​​between each layer), time-response correlation chart (showing the trend of deformation index changing over time during the entire temperature control curve operation), and risk area coordinates (if the verification fails, the data will clearly indicate which physical location of the cabinet, such as the upper left corner of the edge banding, has experienced synchronization failure, providing precise positioning for adding additional thermal coupling constraints to S106).

[0062] S106. Based on the verification results, if the synchronization is insufficient, additional thermo-coupling constraints are incorporated, the entire heating process is re-simulated, and the final adaptive temperature control curve is determined.

[0063] The verification results data are acquired, and initial screening is performed for cases of insufficient synchronization. If the detected synchronization is below a preset threshold, such as a confidence level of 0.8 (or 80%), the subsequent process of introducing thermal coupling constraints is triggered, identifying the specific intervals of insufficient synchronization. Based on the intervals of insufficient synchronization, additional thermal coupling constraints are introduced. These constraints are parameterized using a pre-established thermal model to determine the thermal distribution data after the constraints are introduced. Using the thermal distribution data, the simulation of the heating process is re-executed, dynamically recording the thermal changes at each time interval to obtain temperature fluctuation information during the simulation. Using this temperature fluctuation information, combined with an adaptive adjustment mechanism, a preliminary temperature control curve is constructed. If the fluctuation amplitude exceeds a preset range, the curve is smoothed to obtain the pre-adjusted curve data. Based on the pre-adjusted curve data, the matching degree between thermal coupling and the heating process is analyzed. If the matching degree does not meet the preset standard, secondary optimization is performed on local nodes of the curve to determine the optimized temperature control curve structure. Obtain the optimized temperature control curve structure, and verify the curve as a whole in conjunction with the final determined target requirements to determine whether it meets the comprehensive requirements of synchronization and thermal distribution, and output the final adaptive temperature control curve.

[0064] In one possible implementation, based on the identifier of the interval with insufficient synchronicity, additional thermo-mechanical coupling constraints are introduced. These constraints are parameterized using a pre-established thermodynamic model to determine the thermodynamic distribution data after the constraints are introduced. The additional thermo-mechanical coupling constraints are artificially introduced physical limiting parameters to correct asynchronous deformation. Specifically, they include: 1. Interface displacement continuity constraint: This mandates that the displacement difference between adjacent material layer contact surfaces must be limited to the micrometer level within a specific time period. 2. Local energy conservation equilibrium constraint: For high-risk areas, this limits the amount of heat inflow per unit time to force a leveling of the temperature difference between the substrate and the coating. 3. Mechanical property boundary constraint: This introduces the material's yield strength limit as a hard upper limit during model simulation to prevent the simulated deformation from exceeding the physical reality.

[0065] The pre-established thermodynamic model is an upgraded version of the S102 basic model, and its construction steps are as follows: Based on the geometric model, a simple heat conduction model and an elasticity constitutive model are strongly coupled through a system of partial differential equations (PDEs). The model needs to define time-varying boundary conditions and be able to receive the temperature control curve output from S104 as a driving source. For the "insufficient synchronicity intervals" identified in S105, the computational mesh for this local area is refined using second- or third-order subdivisions to ensure the accuracy of constraint application. The parameterization of constraints through the pre-established thermodynamic model is the process of transforming abstract physical constraints into computer-calcifiable numerical values. Specifically, the additional thermodynamic coupling constraints are transformed into mathematical operators (such as penalty functions or Lagrange multipliers). The interval identifiers (time period and spatial coordinates) determined in S105 are mapped to independent variables in the model. For example, a functional relationship is established between time t and coordinates (x, y, z) and the Yodon intensity coefficient C. During the solution of the linear equations of the thermodynamic model, the stiffness matrix or heat conduction matrix is ​​adjusted according to the parameterized constraints. By resolving the equations, the full-field temperature and strain distributions after introducing constraints are obtained. These data will serve as the basis for resimulating the heating process (determining the final adaptive temperature control curve).

[0066] In one possible implementation, the temperature control curve is initially constructed using temperature fluctuation information combined with an adaptive adjustment mechanism. If the fluctuation amplitude exceeds a preset range, the curve is smoothed to obtain the preliminarily adjusted curve data. The temperature fluctuation information includes the instantaneous deviation between the predicted temperature of key areas (such as the sealing seam) in the model and the original S104 curve setting value after applying displacement continuity constraints or energy balance constraints. This fluctuation reflects the extent of temperature adjustment required to ensure deformation synchronization, including the frequency, amplitude (temperature difference), and trend of the fluctuation over time. The adaptive adjustment mechanism is an automated correction algorithm based on feedback control logic. Its working principle is as follows: the mechanism receives the aforementioned temperature fluctuation information in real time and attempts to find a new set of heating power parameters so that the actual thermal distribution can meet the new constraints introduced in S106. It not only adjusts the temperature at the current time point but also automatically predicts and adjusts the heating rate for subsequent time periods based on thermal inertia. Goal-oriented: The ultimate goal of the mechanism is to balance the needs of heating efficiency, solidification energy accumulation and deformation synchronization, thereby initially constructing an adaptive temperature control curve that includes constraints.

[0067] Further analysis reveals that if the fluctuation range exceeds the preset range (e.g., ±3% to ±5% of the set value) or the temperature rise rate jump does not exceed 0.5℃ / s, then a five-point cubic smoothing method or Gaussian filtering algorithm is typically used to process the temperature data point sequence. For example, the system automatically weakens those abrupt temperature jumps, transforming them into transition segments with gentler slopes. Through smoothing, high-frequency noise and physically abrupt changes in the curve are eliminated, ultimately determining the final shape of the adaptive temperature control curve.

[0068] In one possible implementation, the matching degree between thermo-coupling and the heating process is analyzed based on the initially adjusted curve data. If the matching degree does not meet the preset standard, secondary optimization is performed on the local nodes of the curve to determine the optimized temperature control curve structure. Specifically, the system compares the expected heat distribution data after introducing thermo-coupling constraints with the actual heating capacity that the heater can achieve under physical limits. It analyzes whether the instantaneous power required to meet deformation synchronization (thermo-coupling requirements) at a specific time node is within the linear range of the heating system's power output. By simulating the heating process, the displacement vector of each material layer under the current curve is calculated. If the direction and rate of displacement of each layer are highly consistent (i.e., the shear stress tends to a minimum), the matching degree is considered high. If the matching degree does not meet the preset standard, such as the matching degree coefficient... If over 92% of the simulated points meet the thermo-coupling constraints, then secondary optimization is performed on local nodes of the curve. Specifically, for the specific time node with the lowest contribution to the matching degree (i.e., the point with the most obvious conflict), small positive or negative perturbations are made to the temperature amplitude or the time span before and after that node without changing the overall temperature control framework. If the thermo-coupling requirements conflict with the heating efficiency, the secondary optimization will find the optimal balance between the two through a penalty function and readjust the curvature of that segment of the curve. After completing the local fine-tuning, the model is verified again until all indicators fall within the preset standard range.

[0069] Finally, the final adaptive temperature control curve determined by the system must simultaneously meet the following comprehensive requirements:

[0070] Ultimate goals: 1. Production quality goal: Completely suppress sagging and ensure the smoothness of the coating surface (curing effect indicators meet standards). 2. Structural stability goal: No delamination or cracking between material layers, and macroscopic deformation controlled within the process error range.

[0071] Comprehensive requirements for synchronicity and thermal distribution: 1. Deformation coordination requirement: To achieve synchronized dynamic displacement of the substrate, edge banding, and coating layer throughout the heating process, eliminating shear forces at the interfaces. 2. Thermal energy distribution requirement: To ensure that heat can penetrate into the composite structure uniformly and rhythmically, satisfying the heat accumulation required for coating curing while avoiding stress concentration caused by local overheating.

[0072] This embodiment of a method for generating an adaptive temperature control curve to prevent cabinet dripping, taking oven operation as an example, may further include:

[0073] S201. Extract application parameters from the final adaptive temperature control curve, transmit these parameters to the oven system on the production line, obtain real-time feedback data, and determine if the feedback deviation exceeds the threshold. Then, dynamically fine-tune the curve to obtain an optimized execution plan.

[0074] Specifically, application parameters are extracted from the temperature control curve. Pre-established analytical tools, such as feature extraction algorithms, decompose the curve data to obtain key control points and corresponding parameter values. These application parameters include temperature setpoint, time step, heating / cooling rate, heater power allocation ratio, and auxiliary equipment status. Key control points refer to important time points affecting the coating curing quality and material deformation synchronization, such as the preheating end point, solvent evaporation peak point, and isothermal curing start point. The extracted application parameters are transmitted to the oven equipment using a transmission system. During transmission, a verification mechanism confirms data integrity and ensures the parameters have accurately reached the target equipment. During oven operation, real-time feedback data is acquired. A data acquisition module continuously monitors the operating status, including temperature and time, to assess the stability and validity of the collected data. If the deviation from the expected value exceeds a preset threshold, a deviation judgment logic is triggered. Comparative analysis determines the specific range and direction of the deviation. Based on the deviation judgment results, a dynamic adjustment strategy is implemented, modifying the key control points in the curve parameters and generating adjusted control commands. By adjusting the curve parameters, the optimization scheme is updated, and instructions are sent to the oven equipment to complete the closed-loop control of the execution process, obtaining improved results in real-time operation. After obtaining the improvement results, the feedback data in the execution process is continuously monitored. If the deviation still exceeds the preset threshold, the dynamic adjustment steps are repeated until the control requirements are met, thus determining the final operational stability.

[0075] In one possible implementation, the validity judgment of the collected data mainly addresses whether the data truly reflects the physical state, eliminating dirty data caused by sensor malfunctions or environmental interference. Specifically, this includes: 1. Physical range check: Based on the performance range determined in S101, a reasonable physical range for the temperature sensor is set (e.g., 0°C to 250°C). If the data shows a negative value or a value far exceeding the upper limit of the heating capacity, it is deemed invalid. 2. Abrupt slope detection: The temperature difference ΔT between adjacent sampling points is calculated. Due to the inertia of heat conduction, the temperature cannot jump by tens of degrees within milliseconds. If ΔT / Δt exceeds the preset physical threshold, the system determines that the data is noise generated by electromagnetic interference and considers it invalid. 3. Logical consistency verification: Compare the data from multiple sensors within the same heating zone. If the reading deviation between the main sensor and the compensation sensor exceeds 20%, the validity of the current sampling point is questionable.

[0076] The stability assessment of collected data focuses on evaluating whether the production process is under control. Specifically, it includes: 1. Standard deviation and variance analysis: In the isothermal section of the temperature control curve, the acquisition module calculates the variance of the temperature data within a certain time window. If the variance continues to increase, it indicates that the heating compensation system is oscillating, indicating a decrease in stability. 2. Trend deviation analysis: The real-time acquired temperature curve is compared in real-time with the adaptive temperature control curve determined by S106. The real-time data is smoothed using a moving average method, and the residual between the real-time data and the target curve is calculated. If the residual continues to increase unidirectionally over multiple periods, the process is considered unstable. 3. Signal-to-noise ratio (SNR) assessment: This assesses the ratio of the effective temperature control component to the random fluctuation component in the original signal. A high SNR indicates stable and reliable data, which can be used as input for subsequent updates to the risk map.

[0077] If the data is deemed valid after evaluation, it proceeds directly to the data filtering module for cleaning and is then used to calculate local or deformation indicators. If the data is invalid or unstable, a data compensation mechanism is triggered, using linear interpolation with valid data before and after the invalid data to fill in any gaps. Alternatively, a frequency adjustment may be triggered; if stability deteriorates, the data mapping module will increase the sampling frequency to capture the causes of fluctuations with higher-precision samples. Finally, the unstable state is written to the process log, which serves as the basis for determining abnormal conditions and triggering alarms.

[0078] In one possible implementation, if the collected real-time feedback data is found to deviate from the expected value beyond a preset threshold, deviation judgment logic is triggered. The specific range and direction of the deviation are determined through comparative analysis. For example, at a specific production time point t, the expected value includes the target temperature to be reached at that time. The system considers the expected thermal stress level and estimated material deformation. Preset thresholds can include absolute temperature difference thresholds, such as ±1℃ to ±3℃, and a rate of change (slope) threshold. For the heating phase, if the real-time heating rate is faster or slower than the expected value by more than 0.2℃ / s, it will be considered a deviation even if the current temperature has not exceeded the limit. This threshold is derived inversely based on the performance range of S101 and the probability of sagging risk of S103. In the high-risk area (red zone), this value will automatically shrink to a smaller value. If a deviation from the expected value exceeds the preset threshold, the deviation judgment logic is triggered, and the system uses multi-dimensional comparative analysis to qualitatively and quantitatively determine the deviation.

[0079] A. Determine the direction of the deviation: 1. Positive deviation: Real-time temperature > expected value. Indicates localized heat accumulation; risk: excessively low coating viscosity, increased risk of sagging. 2. Negative deviation: Real-time temperature < expected value. Indicates insufficient heat; risk: incomplete curing, asynchronous deformation.

[0080] B. Determining the specific range of the deviation: 1. Time range determination: Through sliding window analysis, determine whether the deviation is time noise (brief jumps) or persistent deviation (trend drift). If it is persistent, the system will mark the length of the affected curve segment. 2. Spatial range determination (coordinate comparison): Utilize multi-point sensor feedback from the data acquisition module to compare deviation data at different locations (such as the center and edge of the door panel). By constructing a deviation heatmap, determine whether the deviation covers the entire cabinet surface or is limited to a specific potential stress concentration area.

[0081] In one possible implementation, based on the deviation judgment result, a dynamic adjustment strategy is implemented to modify the key control points in the curve parameters and generate adjusted control commands. For example, if the temperature at the corner of the cabinet (critical area) is found to be 3°C higher than normal, the dynamic adjustment strategy is implemented as follows: 1. Control point modification: Modify the control points in this section... The temperature was lowered from 75℃ to 72℃, and the subsequent isothermal curing point was extended by 15 seconds. 2. Instruction Generation: The PLC received instruction to change the value of register number 403, reducing the heater output current by 8% and the conveyor belt inverter frequency by 0.5Hz. The system achieved a closed-loop automated response from deviation detection to instruction issuance, ensuring the flexibility of the production process and the stability of quality.

[0082] S202. Based on the optimized implementation plan, simulate the curing effect of the coating, use finite element analysis to reconfirm the sagging suppression level, and determine the production quality indicators.

[0083] Specifically, by collecting multi-dimensional data during the coating curing process, an initial curing state dataset is constructed to obtain a preliminary description of the curing process. Based on the curing state dataset, the finite element method (FEM) is used to simulate and calculate the sagging phenomenon during the curing process, determining the distribution of sagging suppression effects. For the distribution of sagging suppression effects, stress and deformation data in key areas are obtained to identify potential weak points during the curing process. If the stress data in a key area exceeds a preset threshold, the curing parameters for that area are locally adjusted to obtain an optimized parameter configuration. Using the optimized parameter configuration, the FEM simulation is rerun to obtain updated sagging suppression effect data. Based on the updated sagging suppression effect data, the differences before and after adjustment are compared to determine the final production quality standard configuration. Using the final production quality standard configuration, control instructions for the curing process are generated, outputting an execution plan suitable for the production environment.

[0084] In one possible implementation, an initial curing state dataset is constructed by collecting multi-dimensional data during the coating curing process, including thermodynamic data, optical / morphological data, and time-related data, to obtain a preliminary description of the curing process. Specifically, the data acquisition module in S107 converts the voltage / current signals acquired by the sensors into physical quantities (such as temperature and displacement). The aforementioned multi-dimensional data are aligned according to a unified timestamp to form a multi-column matrix. For example, at the 50th second, the temperature is 60°C, the solvent evaporation rate is 30%, and the surface tension is X. This is the initial curing state dataset. Using pre-established analytical tools (such as feature extraction algorithms), key stages (such as wet film leveling stage, gel point, and fully dry stage) are identified from the dataset. The parameters in the dataset are calculated using empirical formulas or physical models (such as an n-order reaction kinetic model) to finally generate a multi-dimensional trend chart or state parameter table describing the curing characteristics of that stage.

[0085] In one possible implementation, based on the cured state dataset, the finite element analysis method is used to simulate the sagging phenomenon during the curing process and determine the distribution of sagging suppression effect. First, the system uses the application parameters (control point temperature, time step) obtained from S107 decomposition and the adaptive temperature control curve determined in S106 as initial excitation loads, and inputs them into the finite element model. The core of the simulation process lies in simulating the dynamic behavior of the coating liquid film under complex stress.

[0086] 1. Viscosity field evolution simulation: Based on the coating composition data, calculate the function of coating viscosity evolution over time under the current temperature control curve. 2. Force balance calculation: Calculate the vertical gravitational component acting on the coating in each mesh element of the model. Surface tension And the interfacial shear stress caused by uneven thermal expansion of the sheet metal. 3. Flow displacement solution: Using a simplified version of the Navier-Stokes equations, calculate the microscopic displacement L generated by the coating liquid film within a specific time period.

[0087] Furthermore, the system calculates the sagging suppression rate by comparing data from the unoptimized state with the current optimized curve state. The suppression effect is typically defined by the following indicators: 1. Liquid film stability: whether the predicted displacement L remains within the allowable leveling range (usually <0.5mm); 2. Thickness uniformity deviation: after simulated curing, observe the difference in coating thickness between the top and bottom of the facade. Finally, determining the distribution of the sagging suppression effect involves visualizing the calculation results: summarizing the suppression indicators of tens of thousands of finite element nodes across the entire site, and generating effect distribution maps of varying color depths on the 3D model surface of the cabinet. For example, a high suppression zone (green) represents that the temperature control curve perfectly offsets the effects of stress and gravity, and the coating state is extremely stable. An edge risk zone (yellow / red) represents that under the current curve, certain specific geometric locations (such as corners and opening edges) still have a slight tendency to flow.

[0088] In one possible implementation, the distribution of the sagging suppression effect is analyzed by acquiring stress and deformation data in key areas to identify potential weak points during the curing process. For example, when the internal shear stress of the coating layer... When the yield value of the coating exceeds the specified temperature, and the displacement vector continuously shifts upward or downward, the coordinate point is identified as a "sagging weak point." Comparing the displacement difference between the substrate and the coating layer, if the relative displacement deviation at a critical time point exceeds the allowable synchronization threshold (e.g., exceeding the tensile strength limit after coating film formation), the location is identified as a "cracking or peeling weak link." Finite element analysis cloud maps are used to identify areas with the densest stress contour lines, typically found at right-angle turns, edge banding seams, or opening edges of cabinets. These areas, due to their complex geometry and uneven heat absorption, are naturally weak points in a physical sense. If the deformation of a certain area shows a significant step (excessive gradient) compared to its surrounding area, the system considers that this area will generate significant internal stress during thermal expansion and contraction, marking it as a "structural weak point." During the curing process, the coating experiences a low-viscosity window (rapid heating but before curing begins). The system checks which areas experience the least stress release during this period. If the suppression effect distribution value continuously decreases during this window, the corresponding spatial location is the weak point.

[0089] In one possible implementation, if the stress data in a critical area exceeds a preset threshold, the curing parameters for that area are locally adjusted to obtain an optimized parameter configuration. The preset threshold typically includes a yield strength limit (set to 70%-80% of the yield strength) and an interfacial shear stress threshold (set between 1.5 MPa and 2.5 MPa). The curing parameters are control variables that directly affect the physicochemical changes of the coating, mainly referring to parameters such as local heating power, temperature gradient, residence time, and cooling rate. The local adjustment involves targeted intervention at the coordinate points identified in S107, with the specific steps as follows:

[0090] A. Power redistribution (spatial dimension adjustment): If the stress in a corner area exceeds the standard, the system will reduce the duty cycle of the corresponding heating unit in that area to reduce heat input and thus reduce the instantaneous thermal expansion at that point; if the surrounding area is affected by the cooling at that point, the system will fine-tune the output of the adjacent heating zone to maintain the overall thermal field balance.

[0091] B. Node timing fine-tuning (adjustment of the time dimension):

[0092] Slower slope: The system slows down the heating rate in this region by modifying key control points on the adaptive temperature control curve. For example, the original 1.5℃ / s is adjusted to 1.0℃ / s, giving the substrate and coating more time to release internal stress synchronously. Added staged isothermal sections: At temperature points where stress rises sharply, a short isothermal holding period is inserted to utilize the material's creep properties to eliminate stress concentration.

[0093] C. Closed-loop simulation verification: The adjusted parameters will be input into the finite element analysis model again. The system will rerun the simulation to check whether the stress data at that coordinate point has fallen within the preset value. If it still exceeds the limit, the backtracking mechanism will be triggered, and it may even be necessary to return to S106 to introduce stronger thermo-mechanical coupling constraints.

[0094] S203. Select high-quality threshold segments from production quality indicators, generate complete process logs, and determine if the logs show abnormalities. If so, trigger an alarm mechanism to obtain a continuous monitoring framework.

[0095] Specifically, key indicators are extracted from production quality-related data. Preliminary screening of indicator values ​​is conducted to determine the upper and lower limits of the high-quality threshold range, resulting in a preliminary defined quality standard range. Based on this preliminary quality standard range, real-time data streams from the production process are acquired to construct a complete process log recording framework, identifying detailed operational information for each time period. The log data content within this framework is extracted, and through data cleaning and formatting, a structured log dataset is obtained. From this structured log dataset, indicator fluctuations within each time period are analyzed. If an indicator value exceeds a preset threshold range, it is considered an abnormal situation. For identified abnormal situations, a pre-established alarm mechanism is triggered, automatically sending a notification signal to obtain the specific time point and related data of the anomaly trigger. Based on the specific time point and related data of the anomaly trigger, combined with the operating rules of the continuous monitoring system, a support vector machine algorithm is used to classify abnormal trends and determine the potential risk level. From the classification results of potential risk levels, high-risk category data is extracted to update the threshold adjustment strategy of the monitoring system and determine the focus of subsequent monitoring.

[0096] In one possible implementation, key indicators are obtained from production quality-related data. Preliminary screening of these indicators determines the upper and lower limits of the high-quality threshold range, resulting in a preliminary quality standard interval. Specifically, the system retrieves historical production log data that were rated as "superior quality," and uses cluster analysis or statistical filtering to remove noise from abnormal fluctuations, retaining data samples representing the highest quality level. The selected superior quality indicators are then fitted with a normal distribution, and the mean is calculated. ) and standard deviation ( Determine the upper and lower limits: The upper limit is usually set as μ + kσ. For example, the upper limit of sagging must be lower than the critical value that triggers visually visible defects. The lower limit is usually set as μ - kσ. For example, the degree of cure must be higher than the minimum percentage that ensures material strength. This range of upper and lower limits is the high-quality threshold range, representing that products produced within this range have a very high yield guarantee. Among them, the key indicators are the core physical quantities for evaluating the success or failure of cabinet coating. They are usually extracted from the curing state dataset constructed by S108 and mainly include surface sagging, degree of cure, interlayer adhesion / stress residue, and surface gloss uniformity.

[0097] In one possible implementation, the system analyzes the fluctuations of indicators within each time period from a structured log dataset. This involves scanning the indicators by time slices, and if an indicator value exceeds a preset threshold range, it is considered an abnormal situation. The preset threshold range has two levels: 1. A quality red line threshold, which is the aforementioned high-quality threshold range; exceeding this range indicates a potential product downgrade. 2. A process stability threshold, limiting fluctuations during the execution of the temperature control curve; for example, the deviation between the real-time temperature and the adaptive curve must not exceed ±2℃. Abnormal situation determination is based on various methods. For example, if the sagging risk indicator suddenly jumps above the upper limit within a certain time period (e.g., t=200s), the system immediately marks it as abnormal. If the indicator is within the threshold but shows a trend towards the upper or lower limit for five consecutive time periods, the system uses a sliding window algorithm to predict an impending abnormality.

[0098] In one possible implementation, the step of classifying abnormal trends and determining potential risk levels using a support vector machine algorithm based on the specific time point of the anomaly trigger and related data, combined with the operating rules of the continuous monitoring system, includes the following operating rules:

[0099] 1. Data Alignment Rules: Real-time collected data such as temperature, displacement, and gloss must be strictly aligned with the timestamps of the standard temperature control curve to ensure that every second of data analyzed has a comparable expected baseline value.

[0100] 2. Multi-level linkage triggering rules:

[0101] Level 1 Trigger (Deviation Alert): When the indicator deviates from the upper or lower limit of the high-quality threshold, it is immediately recorded and added to the observation sequence. Level 2 Trigger (Classification Intervention): If the deviation continues to exceed the set step size (e.g., 5 consecutive data collection points), the SVM algorithm is forcibly started for risk assessment.

[0102] 3. Data backtracking rules: Once an anomaly is identified, the system must automatically retrieve the complete process parameters (heating power, ambient humidity, etc.) for the 60 seconds prior to that time point for root cause analysis.

[0103] 4. Threshold adaptive update rule: Based on the quality results of the produced finished products, the high-quality threshold of the monitoring system is periodically corrected in reverse to ensure that the rule evolves with changes in the production environment.

[0104] The Support Vector Machine (SVM) algorithm is used to classify abnormal trends and determine the potential risk level. Specifically, the system extracts data features before and after the abnormal time point as the input vector for the SVM, including deviation amplitude (the absolute magnitude of the indicator deviating from the threshold), rate of change (slope) (how fast the abnormal indicator rises or falls), spatiotemporal correlation (whether the anomaly occurs at a single point or simultaneously in multiple key areas of the cabinet), and environmental interference coefficient (the difference between the current production line's real-time power fluctuation and the expected power). The SVM then compares the above feature data with historically known failure modes using a pre-trained classifier. The SVM output is typically a category label, which, combined with indicator scores, classifies the risk into three levels: high, medium, and low risk.

[0105] For example, when t=1508, the sagging risk rate suddenly increased from 0.1 to 0.4. The slope of this curve segment, the corresponding local temperature rise rate, and the key point displacement difference were extracted as the input vector for the SVM. The SVM mapped the features to a high-dimensional space and found that they fell into the "thermorheologically abrupt change" classification cluster. The SVM output category was "medium risk" with a probability of 0.85. The system judged this as a potential sagging risk and automatically reduced the power output of the heating zone in that segment.

[0106] In this way, the monitoring system can identify subtle trends that are "not bad yet, but are getting worse," thus completing self-correction before the risk truly turns into a quality problem.

[0107] like Figure 4As shown, this invention provides a cabinet anti-sagging adaptive temperature control curve generation system, mainly comprising: a data acquisition module, used to acquire material layer data of the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer, and to collect the temperature distribution during the heating process in real time through sensors to obtain the initial thermal response characteristics; a finite element simulation module, used to simulate the deformation interaction of each material layer during heating using finite element analysis based on the initial thermal response characteristics, and to adjust the simulation parameters if the deformation difference exceeds a preset threshold to determine the potential stress concentration area; and a sagging risk assessment module, used to extract the paint flow tendency index from the potential stress concentration area and use a neural network model to predict the sagging risk. The system generates a risk assessment map based on the distribution of heat and temperature. A temperature control curve generation module generates a preliminary draft temperature control curve based on the risk assessment map, refines the curve segments using an iterative optimization algorithm, and backtracks to modify the draft if local overheating indicators exceed a threshold, thus determining the final coordinated curve version. A deformation synchronicity verification module selects key time nodes from the coordinated curve version, obtains corresponding material layer thermal behavior data, and verifies deformation synchronicity using a support vector machine model to obtain verification results. An adaptive temperature control curve determination module, based on the verification results, determines if synchronicity is insufficient, incorporates additional thermo-mechanical coupling constraints, re-simulates the entire heating process, and determines the final adaptive temperature control curve. Furthermore, the aforementioned cabinet anti-drip adaptive temperature control curve generation system also includes:

[0108] The dynamic execution optimization module extracts application parameters from the final adaptive temperature control curve, transmits these parameters to the oven system on the production line, obtains real-time feedback data, and dynamically fine-tunes the curve if the feedback deviation exceeds a threshold to obtain an optimized execution plan. The production quality confirmation module simulates the coating curing effect based on the optimized execution plan, uses finite element analysis to reconfirm the sagging suppression level, and determines the production quality indicators. The process monitoring module selects high-quality threshold segments from the production quality indicators, generates a complete process log, and triggers an alarm mechanism if the log shows abnormalities, thus obtaining a continuous monitoring framework.

[0109] If the technical solution of this application involves the collection, processing, or application of personal information, the relevant products have, before implementing any personal information processing activities, fully and clearly informed individuals of the processing rules in accordance with the "Personal Information Protection Law of the People's Republic of China" and other current laws and regulations, and obtained their voluntary and explicit consent. If sensitive personal information is involved, the product has obtained the individual's separate consent before processing, and such consent is given in an explicit manner. For example, prominent signs are set up in the area where information collection devices such as cameras are located, clearly indicating "Entering is considered as consent to the collection of personal information"; or through pop-ups, checkboxes, user-initiated uploads, etc., under the premise of clearly listing the processor's identity, processing purpose, processing method, and information type, the user actively completes the authorization operation. The above mechanisms ensure that all personal information processing activities are based on legal authorization and fully comply with national compliance requirements regarding personal information protection.

[0110] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A method for generating an adaptive temperature control curve to prevent cabinet dripping, characterized in that, The method includes: Data on the material layers of the cabinet composite structure are obtained, including the thermal conductivity and expansion coefficient of each layer. The temperature distribution during the heating process is collected in real time by sensors to obtain the initial thermal response characteristics. Based on the initial thermal response characteristics, finite element analysis is used to simulate the deformation interaction of each material layer during heating. If the deformation difference exceeds a preset threshold, the simulation parameters are adjusted to determine the potential stress concentration area. The flow tendency index of the coating is extracted from the potential stress concentration area, and the distribution of sagging risk is predicted by a neural network model to obtain a risk assessment map. For the risk assessment map, a preliminary draft of the temperature control curve is generated. The curve segments are refined through iterative optimization algorithms. If the local overheating index is higher than the threshold, the draft is backtracked and modified to determine the coordinated version of the curve. Key time points were selected from the coordinated curve version, and corresponding material layer thermal behavior data were obtained. The deformation synchronization was verified using a support vector machine model, and the verification results were obtained. Based on the verification results, if the synchronization is insufficient, additional thermo-coupling constraints are incorporated, the entire heating process is re-simulated, and the final adaptive temperature control curve is determined.

2. The method for generating an adaptive temperature control curve for preventing dripping in a cabinet according to claim 1, characterized in that, The acquisition of material layer data for the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer, and the acquisition of initial thermal response characteristics by real-time acquisition of temperature distribution during the heating process using sensors, including: Data is collected on the cabinet composite structure using a sensor system, and the thermal conductivity and expansion coefficient of each material layer are recorded to obtain a preliminary dataset of physical parameters. Based on the preliminary physical parameter dataset, the thermal conductivity and expansion coefficient are classified using a preset threshold range to determine the performance range of each material layer under different temperature conditions. If the performance range after classification exceeds the preset threshold range, the outlier is removed through the data filtering module to obtain the corrected set of physical parameters. Based on the corrected set of physical parameters and combined with real-time monitored temperature distribution data, a support vector machine algorithm is used to predict the thermal response value and determine the potential trend of change during the heating process. Based on the predicted trend of thermal response values, the correlation between temperature distribution and material layers during the heating process is analyzed to obtain the thermal stress distribution results of each layer. The stability of the composite structure is assessed by combining the results of thermal stress distribution with environmental influencing factors, and the thermal response characteristics of key areas are determined. If the thermal response characteristics of the critical area do not match the preset safety range, the temperature distribution sampling frequency is adjusted through the data mapping module to obtain a more accurate thermal response dataset.

3. The method for generating an adaptive temperature control curve for preventing cabinet dripping according to claim 1, characterized in that, The process involves using finite element analysis to simulate the deformation interaction of each material layer during heating, based on the initial thermal response characteristics. If the deformation difference exceeds a preset threshold, the simulation parameters are adjusted to determine potential stress concentration areas. This includes: Obtain the initial thermal response characteristic data of the material layer; A thermo-mechanical coupling model of the material layer is established using the finite element method; Apply a heating load to the model and calculate the deformation of each material layer; Determine whether the difference in deformation between each material layer exceeds a preset threshold; If the difference exceeds the preset threshold, the simulation parameters of the material layer will be adjusted. Based on the adjusted simulation parameters, the locations of potential stress concentration areas are determined.

4. The method for generating an adaptive temperature control curve for preventing cabinet dripping according to claim 1, characterized in that... The process of extracting coating flow tendency indicators from potential stress concentration areas, using a neural network model to predict sagging risk distribution, and obtaining a risk assessment map includes: By acquiring stress concentration data from potential areas, and using automated tools to perform preliminary cleaning and formatting of the data, a preliminary stress dataset is obtained. Based on the pre-processed stress dataset, key indicators related to flow tendency are extracted, and a set of indicators highly correlated with sagging risk is determined using a pre-established feature screening method. For the set of indicators, a neural network model is constructed for training and prediction to obtain the probability distribution value of the risk of spillage and generate intermediate results of the risk distribution. From the intermediate results of risk distribution, combined with regional analysis data, grid partitioning technology is used to map the distribution probability values ​​to specific regions, thus obtaining regionalized risk distribution data; If the probability value of certain regions in the regionalized risk distribution data exceeds a preset threshold, then the regions are marked as high-risk, and a risk distribution dataset with the labels is generated. Using a labeled risk distribution dataset, a risk assessment map is generated using visualization techniques to determine the final risk assessment results for spillage.

5. The method for generating an adaptive temperature control curve for preventing cabinet dripping according to claim 1, characterized in that, The process involves generating a preliminary draft temperature control curve based on the risk assessment map, refining the curve segments through iterative optimization algorithms, and retrospectively modifying the draft if local overheating indicators exceed a threshold to determine the final, coordinated curve version. This includes: By using risk assessment data, temperature distribution information for key areas is obtained from map analysis to determine a preliminary framework for temperature control curves. Based on the initial temperature control curve framework, an iterative optimization algorithm is used to refine the curve segments, resulting in the adjusted curve details. For the adjusted curve details, obtain the temperature index data of the local area to determine whether there is overheating exceeding the preset threshold; If the temperature index in a local area exceeds the preset threshold, a backtracking adjustment mechanism is triggered to recalculate the curve segment and determine the corrected curve shape.

6. The method for generating an adaptive temperature control curve for preventing cabinet dripping according to claim 1, characterized in that, The process involves selecting key time points from the coordinated curve version, obtaining corresponding material layer thermal behavior data, and using a support vector machine model to verify deformation synchronization. The verification results include: Determine key time points based on the coordinated curve version; Thermal behavior data of the material layer are obtained at key time points; Deformation indexes are calculated based on thermal behavior data; Support vector machine model is used to process deformation index and time series; If the output of the support vector machine model is higher than the preset threshold, then the deformation synchronization is determined to be valid. Generate verification result data based on the judgment results; The validation results are used to update the curve version.

7. The method for generating an adaptive temperature control curve for preventing cabinet dripping according to claim 1, characterized in that, Based on the verification results, if the synchronization is insufficient, additional thermo-coupling constraints are incorporated, the entire heating process is re-simulated, and the final adaptive temperature control curve is determined, including: The verification result data is obtained, and a preliminary screening is performed for cases of insufficient synchronization. If the synchronization is detected to be lower than the preset threshold, the subsequent process of introducing thermal coupling constraints is triggered to obtain the specific interval identifier of insufficient synchronization. Based on the interval identifier with insufficient synchronization, additional thermo-coupling constraints are introduced. The constraints are parameterized by a pre-established thermo-mechanical model to determine the thermo-distribution data after the constraints are introduced. Using thermal distribution data, the simulation of the heating process is re-executed, and the thermal changes at each time interval are dynamically recorded to obtain temperature fluctuation information during the simulation process; By using temperature fluctuation information and combining it with an adaptive adjustment mechanism, a preliminary temperature control curve is constructed. If the fluctuation amplitude exceeds the preset range, the curve is smoothed to obtain the preliminarily adjusted curve data. Based on the preliminary adjusted curve data, the matching degree between thermodynamic coupling and heating process is analyzed. If the matching degree does not meet the preset standard, the local nodes of the curve are optimized a second time to determine the optimized temperature control curve structure. Obtain the optimized temperature control curve structure, and verify the curve as a whole in conjunction with the final determined target requirements to determine whether it meets the comprehensive requirements of synchronization and thermal distribution, and output the final adaptive temperature control curve.

8. A cabinet anti-drip adaptive temperature control curve generation system, characterized in that, The system includes: The data acquisition module is used to acquire material layer data of the cabinet composite structure, including the thermal conductivity and expansion coefficient of each layer. It also uses sensors to collect the temperature distribution during the heating process in real time to obtain the initial thermal response characteristics. The finite element simulation module is used to simulate the deformation interaction of each material layer during heating based on the initial thermal response characteristics. If the deformation difference exceeds a preset threshold, the simulation parameters are adjusted to determine the potential stress concentration area. The sag risk assessment module is used to extract coating flow tendency indicators from potential stress concentration areas, use a neural network model to predict the sag risk distribution, and obtain a risk assessment map. The temperature control curve generation module is used to generate a preliminary draft temperature control curve based on the risk assessment map. The curve segments are refined through iterative optimization algorithms. If the local overheating index is higher than the threshold, the draft is backtracked and modified to determine the coordinated curve version. The deformation synchronization verification module is used to select key time nodes from the coordinated curve version, obtain the corresponding material layer thermal behavior data, and use a support vector machine model to verify the deformation synchronization and obtain the verification results. The adaptive temperature control curve determination module is used to determine the final adaptive temperature control curve by incorporating additional thermo-coupling constraints and resimulating the entire heating process based on the verification results if the synchronization is insufficient.

Images