A method for evaluating the sintering quality of an artificial heat-conducting film

Abstract

The application provides a kind of artificial heat-conducting film sintering quality evaluation detection method, method contains the construction of a personalized database comprising different materials and different process conditions corresponding to the electrical properties of artificial heat-conducting film;Through the analysis of electrical properties data, the performance fluctuation parameters of different sintering production equipment and the same sintering production equipment in different production periods are identified;Multiple types of sensors are arranged on the artificial heat-conducting film sintering production equipment, and the key parameters in the sintering production equipment are collected;Through data fusion algorithm, multiple types of key parameters are fused to obtain a fusion dataset;According to the performance fluctuation parameters, the sintering production equipment is adjusted;The key parameters in the fusion dataset are input into the sintering quality evaluation model, compared with the standard parameters of the personalized database, and the comparison result is obtained.The application realizes intelligent adjustment of sintering process through quality evaluation model and automatic feedback mechanism, improves product quality and production efficiency.

Description

Technical Field

[0001] This invention relates to the field of artificial graphite thermal conductive film preparation technology, and in particular to a detection method for evaluating the sintering quality of artificial thermal conductive films. Background Technology

[0002] The main raw material used in artificial graphite thermal conductive film is polyimide (PI), and the main preparation process includes PI slitting, PI rewinding, carbonization, graphitization, and calendering. The most critical processes are carbonization and graphitization. Carbonization and graphitization are heat treatment processes, with graphitization requiring a high temperature of 3000℃. Infrared non-contact temperature measurement is used, but this method is limited by the furnace environment and the product environment, resulting in slightly lower reliability and some deviation from the set temperature curve. (For example, poor exhaust in some equipment or varying concentrations of smoke in the furnace can cause measurement errors, as can different levels of contamination in the observation window). Deviations between different equipment and the cumulative deviations of a single equipment at different times result in poor product consistency, severely impacting the pass rate of some high-end sensitive products and failing to meet requirements. The key measurement method for the quality of artificial graphite film is thermal conductivity. This test requires professional testing personnel, specialized equipment, and a specialized environment, and its testing efficiency is extremely low. Artificial graphite thermal conductive film has excellent electrical and thermal conductivity, and its thermal and electrical conductivity are directly related to the quality of graphitization. Currently, the measuring instruments for testing electrical conductivity are reliable, efficient, and require relatively low skill from the measurement personnel. By establishing a comparative database of the thermal and electrical conductivity of products with varying masses and formulations, a standard for judging product conductivity can be established to guide the quality level of graphitized products. This method achieves rapid and efficient judgment. N samples of the same material, batch, and pretreatment are prepared as product monitoring samples (eliminating fluctuations in material and monitoring equipment). For products sintered from the same material, one sample is prepared for each sintering. SPC analysis is established by measuring the monitoring samples to quickly identify equipment anomalies or trends of deterioration in equipment condition, allowing engineers to make timely corrections. However, the current equipment capability judgment in this process has certain shortcomings. Even when all equipment indicators are qualified, product defects occur intermittently. Furthermore, there are significant differences between products from different equipment, making it impossible to identify quality risks caused by individual equipment for highly sensitive products.

[0003] Existing technology one, application number: CN202411558285.0, discloses a method for quality testing and analysis of graphene thermally conductive film materials. This method tests the appearance, thermal conductivity, physical properties, chemical properties, mechanical properties, and electrical properties of the graphene thermally conductive film, analyzes the test data, and constructs corresponding performance evaluation indices to comprehensively assess the material quality of the graphene thermally conductive film. This comprehensive testing of the material quality of the graphene thermally conductive film from multiple dimensions aims to improve the quality assessment system of graphene thermally conductive films. While this method is beneficial for better ensuring the material quality of graphene thermally conductive films and generating material quality test reports, allowing companies to identify problems in the production process and provide a basis for improving production processes and optimizing product design, the data analysis based on constructed evaluation indices has several drawbacks. First, the evaluation indices lack the ability to update autonomously, leading to errors in the data analysis results. Second, its intelligent analysis capabilities are poor and cannot be adapted to the current intelligent production of graphene thermally conductive films.

[0004] Prior art two, application number: CN 202411222530.0, discloses an improved method for measuring the thermal conductivity of a thermally conductive film based on a steady-state electrothermal method. This method uses an improved steady-state electrothermal method to obtain the in-plane thermal conductivity of the thermally conductive film surface based on the temperature distribution under different power levels. The measurement error is less than 3.0%, the uncertainty is reduced to 0.5%, and the response time is in the millisecond range. While this provides useful guidance for accurately evaluating the thermal conductivity of materials and offers technical support for the thermal management engineering applications of graphite film materials, its calculation process is complex, requires numerous parameters, and places high demands on professional computing capabilities.

[0005] Prior art three, application number CN 202411480722.1, discloses a graphene thermal conductive film material defect screening system based on image recognition. Based on images of graphene thermal conductive films, it acquires the surface thickness non-uniformity coefficient and composition distribution non-uniformity coefficient to analyze the surface non-uniformity; it also acquires information on surface pores, cracks, wrinkles, ripples, and edge defects to analyze surface defects; and it acquires information on surface foreign matter, oxidation, and corrosion to analyze surface impurities. Furthermore, it comprehensively evaluates the surface quality of the graphene thermal conductive film, determines whether the surface quality is acceptable, screens defective products, and generates a quality inspection report. While this system assesses the appearance of the graphene thermal conductive film from multiple dimensions to ensure the surface quality of the material, image recognition-based screening places high demands on image capture quality, increasing the cost of image recognition to some extent.

[0006] Current technologies 1, 2, and 3 suffer from limited detection methods and cumbersome operations, resulting in inaccurate test results. Therefore, this invention provides a detection method for evaluating the sintering quality of artificial thermally conductive films. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides a method for evaluating the sintering quality of artificial thermally conductive films, characterized by comprising the following steps:

[0008] Construct a personalized database containing electrical conductivity data of artificial thermal conductive films under different materials and process conditions; through analysis of electrical conductivity data, identify the performance fluctuation parameters of different sintering production equipment and the same sintering production equipment at different production periods;

[0009] Multiple types of sensors are deployed on the artificial thermal conductive film sintering production equipment to collect key parameters such as temperature, pressure, and time. The multiple types of key parameters are fused using a data fusion algorithm to obtain a fused dataset. The sintering production equipment is then adjusted based on performance fluctuation parameters.

[0010] Key parameters from the fusion dataset are input into the sintering quality assessment model and compared with standard parameters from the personalized database to obtain comparison results. The comparison results are used to identify which key parameters deviate from the standard range and are automatically fed back to the sintering production equipment for corresponding adjustments.

[0011] Optionally, the process of analyzing conductivity data includes the following steps:

[0012] A hybrid storage architecture for a personalized database is constructed, in which the basic data layer stores static data on material properties and sintering production equipment parameters; the process data layer stores dynamic data during the sintering process; and the performance data layer stores electrical conductivity test results and related indicators.

[0013] Data cleaning is performed on material data, process data, equipment data, and electrical conductivity data, and corresponding classification indexes, attribute indexes, numerical indexes, and time indexes are established according to key fields;

[0014] Based on the index, key patterns related to conductivity are extracted from the personalized database, and data points that deviate from the normal pattern are identified. The performance fluctuation parameters corresponding to the data points are decomposed into influencing factors at different levels, and the changes in conductivity are decomposed into microstructural changes at the material level.

[0015] Optionally, the process of obtaining the fused dataset includes the following steps:

[0016] The key parameters such as temperature, pressure and time collected by different sensors are normalized, and the influence of each key parameter on the sintering quality is evaluated by calculating the covariance matrix. The weight allocation is dynamically adjusted using an adaptive weight adjustment factor according to the real-time status of the sintering production equipment.

[0017] The normalized key parameters are weighted and fused according to dynamic weights to generate a preliminary fused dataset. Outliers in the fused dataset are removed, and the fused dataset with outliers removed is smoothed.

[0018] The optimized fusion dataset is standardized to meet the input requirements of the sintering quality assessment model. The standardized fusion dataset is then output to the sintering quality assessment model for comparison with the standard parameters of the personalized database.

[0019] Optionally, the process of constructing a sintering quality assessment model includes the following steps:

[0020] The multi-type sensor data of the sintering production equipment is mapped into a multi-dimensional state space. Based on the historical data of different materials and process conditions in the personalized database, dynamic weights are assigned to each key parameter. The action space adopts a composite structure, including two levels: macroscopic adjustment and microscopic optimization. The reward function adopts a hierarchical design.

[0021] The adaptive exploration strategy dynamically adjusts the exploration rate based on the uncertainty of the current state; at the same time, the strategy gradient optimizes multiple objectives such as sintering quality, equipment energy consumption and production efficiency, realizing the combination of action value assessment and multi-objective optimization of strategy gradient; and receives sensor data in real time and updates the adaptive exploration strategy.

[0022] The adjusted sintering quality still did not meet expectations. The sintering quality assessment model automatically triggered a self-correction mechanism to re-evaluate the definition of the state space and action space and optimize the reward function.

[0023] Optionally, the reward function may employ a hierarchical design, which includes the following steps:

[0024] The basic layer reward is designed by selecting key parameters for sintering quality and calculating the deviation between the actual value and the standard value of each key parameter; parameters with large deviations are penalized to ensure that the model tends to control the parameters within the standard range; and the optimization layer reward is designed to provide additional incentives for long-term quality improvement through historical performance trend analysis.

[0025] The basic layer rewards and optimization layer rewards are combined to obtain a comprehensive reward function; to adapt to different material and process conditions, the dynamic weights are adjusted based on historical data and the current state.

[0026] When the sintering quality assessment model triggers the self-correction mechanism, the reward function is re-evaluated and optimized. Based on changes in the current production environment, the definitions of key parameters and action space are adjusted; dynamic weights are recalculated based on the latest data; and the weight coefficients of the basic layer reward and the optimization layer reward are adjusted based on the optimization results.

[0027] Optionally, the process of combining action value assessment and policy gradient multi-objective optimization includes the following steps:

[0028] Sintering quality, equipment energy consumption, and production efficiency are identified as the core optimization objectives. The multi-objective optimization problem is transformed into a unified evaluation system, and different objectives are integrated into a comprehensive optimization function through weighting. Action value assessment is used to quantify the contribution of each action to the multi-objective optimization. The action value assessment at the macro and micro levels is combined to form a composite evaluation system.

[0029] The multi-objective optimization problem is decomposed into multiple sub-objectives, each corresponding to an independent policy gradient optimization path. By weighting, the policy gradients of each sub-objective are merged into a comprehensive gradient to guide the policy update direction. The weights of each sub-objective are dynamically adjusted according to the current state and objective priority.

[0030] The exploration rate is dynamically adjusted based on the current uncertainty and optimization progress; sensor data is received in real time to update the exploration strategy.

[0031] Optionally, the process of dynamically adjusting the exploration rate and updating the exploration strategy includes the following steps:

[0032] The current state is evaluated through multidimensional uncertainty quantification; the target conflict index is used to measure the mutual influence between various optimization objectives, while the historical data deviation reflects the degree of deviation between the current observation and the historical trend; a dynamic uncertainty score is generated through comprehensive calculation.

[0033] Nonlinear optimization progress evaluation is used to conduct in-depth analysis of optimization progress; sub-objective convergence speed is used to evaluate the optimization efficiency of each sub-objective, while the strategy stability index measures the robustness of the current strategy under different states, dynamically identifies bottlenecks in the optimization process, and adjusts the exploration strategy.

[0034] Based on the uncertainty score and optimization progress assessment results, the exploration rate is adjusted according to the target priority weight and environmental complexity factor.

[0035] Optionally, the process of re-evaluating the definitions of the state space and action space includes the following steps:

[0036] Based on the current quality deviation and optimization objectives, the contribution of each parameter in the multidimensional state space is reassessed, and key parameters that have a significant impact on quality are identified. According to real-time sensor data and historical performance trends, new dynamic weights are assigned to key parameters. External environmental variables and equipment status are incorporated into the state space.

[0037] Based on the severity of the current quality deviation, redefine the magnitude and priority of macroscopic adjustment actions; control actions for local parameter fine-tuning of local temperature and pressure gradient adjustment;

[0038] Based on the current distribution characteristics of quality deviations, the weights of key parameters and standard range deviations are readjusted; a dynamic threshold mechanism is introduced to dynamically adjust the standard range based on historical performance trends; and additional incentives are provided for long-term quality improvement through historical performance trend analysis.

[0039] Optionally, the process of dynamically adjusting the standard range based on historical performance trends includes the following steps:

[0040] Based on historical performance data, identify the regular characteristics of its changes over time and extract key features from historical performance trends; use trend models to predict the possible range of changes in key parameters in the future.

[0041] Compare the current actual values ​​of key parameters with historical performance trends to assess the degree of deviation.

[0042] Based on real-time performance evaluation results and trend prediction information, the threshold range is dynamically adjusted; if the current parameter value is higher than the historical trend prediction value, the upper limit of the threshold is increased; if it is lower than the prediction value, the lower limit of the threshold is decreased.

[0043] Optionally, the process of comparing the current actual values ​​of key parameters with historical performance trends includes the following steps:

[0044] Align the actual values ​​of current key parameters with historical performance trends over time; standardize the actual values ​​of current key parameters and historical performance trends to eliminate differences in dimensions and units; extract key features from historical performance trends as a benchmark for comparison.

[0045] Match the actual values ​​of current key parameters with key features of historical performance trends to identify performance differences across different dimensions; calculate the absolute difference between the actual values ​​of current key parameters and the predicted values ​​of historical performance trends to reflect the magnitude of the deviation; compare the absolute deviation with the fluctuation range of historical performance trends to calculate the relative deviation rate, reflecting the severity of the deviation.

[0046] Analyze the consistency between the current trend of the actual value of key parameters and the historical performance trend to determine whether they deviate from the expected direction; based on multi-dimensional indicators such as absolute deviation and relative deviation rate, comprehensively score the degree of deviation of the current key parameters; and classify the degree of deviation into different levels according to the comprehensive score results.

[0047] This invention constructs a personalized database to collect conductivity performance data under different material and process conditions, forming a standardized database. Through data analysis, it identifies performance fluctuation parameters of different sintering production equipment and the same equipment at different production stages. It identifies patterns in the performance fluctuations of sintering production equipment, providing a reference for quality assessment and equipment adjustment. Multi-type sensor data acquisition and fusion involves deploying multiple types of sensors on the sintering production equipment to collect key parameters such as temperature, pressure, and time in real time. A data fusion algorithm integrates these key parameters into a unified fusion dataset, improving data integrity and accuracy. Adjustments to the sintering production equipment are made based on performance fluctuation parameters to ensure optimal operation. Sintering quality assessment and feedback adjustment involve inputting the key parameters from the fusion dataset into a sintering quality assessment model and comparing them with standard parameters in the personalized database. Key parameters deviating from the standard range are identified, providing quantitative evidence for quality assessment. Based on the comparison results, automatic feedback is provided to the sintering production equipment for corresponding adjustments, ensuring stable sintering quality.

[0048] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0049] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0050] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0051] Figure 1 This is a flowchart of the detection method for evaluating the sintering quality of the artificial thermally conductive film in Embodiment 1 of the present invention;

[0052] Figure 2 This is a flowchart illustrating the process of analyzing conductivity data in Embodiment 2 of the present invention;

[0053] Figure 3 This is a process diagram of obtaining the fused dataset in Embodiment 3 of the present invention;

[0054] Figure 4This is a process diagram of constructing the sintering quality evaluation model in Embodiment 4 of the present invention;

[0055] Figure 5 This is a flowchart illustrating the hierarchical design of the reward function in Embodiment 5 of the present invention.

[0056] Figure 6 This is a process diagram illustrating the multi-objective optimization combination of action value assessment and strategy gradient in Embodiment 6 of the present invention;

[0057] Figure 7 This is a flowchart illustrating the process of dynamically adjusting the exploration rate and updating the exploration strategy in Embodiment 7 of the present invention.

[0058] Figure 8 This is a flowchart illustrating the process of re-evaluating the definitions of the state space and action space in Embodiment 8 of the present invention;

[0059] Figure 9 This is a flowchart illustrating the process of dynamically adjusting the standard range based on historical performance trends in Embodiment 9 of the present invention.

[0060] Figure 10 This is a flowchart illustrating the process of comparing the actual values ​​of current key parameters with historical performance trends in Embodiment 10 of the present invention. Detailed Implementation

[0061] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0062] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0063] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0064] Example 1: As Figure 1 As shown, this embodiment of the invention provides a detection method for evaluating the sintering quality of artificial thermally conductive films, comprising the following steps:

[0065] S100: Construct a personalized database containing electrical conductivity data of artificial thermal conductive films under different materials and process conditions; through analysis of electrical conductivity data, identify the performance fluctuation parameters of different sintering production equipment and the same sintering production equipment at different production periods;

[0066] S200: Multiple types of sensors are deployed on the artificial thermal conductive film sintering production equipment to collect key parameters such as temperature, pressure, and time in the sintering production equipment; the multiple types of key parameters are fused through a data fusion algorithm to obtain a fused dataset; the sintering production equipment is adjusted according to the performance fluctuation parameters;

[0067] S300: Input the key parameters in the fusion dataset into the sintering quality assessment model, compare them with the standard parameters in the personalized database, and obtain the comparison results; among them, the comparison results will automatically feed back to the sintering production equipment to identify which key parameters deviate from the standard range, and make corresponding adjustments.

[0068] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first constructs a personalized database containing electrical conductivity data of artificial thermal conductive films under different materials and process conditions; through analysis of electrical conductivity data, it identifies the performance fluctuation parameters of different sintering production equipment and the same sintering production equipment at different production periods; secondly, it deploys multiple types of sensors on the artificial thermal conductive film sintering production equipment to collect key parameters such as temperature, pressure, and time in the sintering production equipment; it fuses multiple types of key parameters through a data fusion algorithm to obtain a fused dataset; it adjusts the sintering production equipment according to the performance fluctuation parameters; finally, it inputs the key parameters in the fused dataset into the sintering quality evaluation model and compares them with the standard parameters in the personalized database to obtain the comparison results; among them, the comparison results automatically feed back to the sintering production equipment which key parameters deviate from the standard range, and makes corresponding adjustments. Step S100 of the above scheme involves constructing a personalized database, collecting conductivity data under different material and process conditions, and forming a standardized database. Through data analysis, it identifies performance fluctuation parameters of different sintering production equipment and the same equipment at different production stages. It identifies the patterns of performance fluctuations in sintering production equipment, providing a reference for quality assessment and adjustment of the equipment. The significance achieved is: providing reliable data support for quality assessment and process optimization; and providing a basis for maintenance and process improvement of sintering production equipment by identifying performance fluctuations, reducing uncertainties in production. Step S200 involves multi-type sensor data acquisition and fusion. Multiple types of sensors are deployed on the sintering production equipment to collect key parameters such as temperature, pressure, and time in real time. Through data fusion algorithms, these multiple key parameters are integrated into a unified fusion dataset, improving data integrity and accuracy. Adjustments are made to the sintering production equipment based on performance fluctuation parameters to ensure optimal operation. Significance Achieved: To achieve comprehensive monitoring of the sintering process, ensuring key parameters remain within reasonable ranges; to reduce errors from individual sensors and improve data reliability through data fusion; and to adjust sintering production equipment based on performance fluctuation parameters, enhancing the stability and consistency of equipment operation. Step S300, Sintering Quality Assessment and Feedback Adjustment, inputs key parameters from the fused dataset into the sintering quality assessment model and compares them with standard parameters from the personalized database; identifies key parameters deviating from the standard range, providing quantitative evidence for quality assessment; and automatically feeds back the comparison results to the sintering production equipment for corresponding adjustments, ensuring stable sintering quality. Significance Achieved: To ensure the quality of the sintering process meets standards through real-time assessment and feedback, reducing the defect rate; and to achieve automated adjustment of the production process, improving production efficiency and consistency.

[0069] In summary, this embodiment achieves precise monitoring and optimization of the sintering process of artificial thermally conductive films through three steps: constructing a personalized database, acquiring and fusing data from multiple types of sensors, and evaluating and adjusting sintering quality. Its significance lies primarily in: providing reliable evidence for quality assessment through data accumulation and analysis; achieving comprehensive monitoring of the sintering process through multi-type sensors and data fusion; and realizing intelligent adjustment of the sintering process through a quality assessment model and automatic feedback mechanism, thereby improving product quality and production efficiency. This method can significantly improve production efficiency and product quality, and has significant practical application value.

[0070] Example 2: As Figure 2 As shown, based on Example 1, the process for analyzing conductivity data provided in this embodiment of the invention includes the following steps:

[0071] S101: Construct a hybrid storage architecture for a personalized database, in which the basic data layer stores static data such as material properties and sintering production equipment parameters; the process data layer stores dynamic data during the sintering process, such as temperature, pressure, and time series; and the performance data layer stores electrical conductivity test results and related indicators.

[0072] S102: Perform data cleaning on material data, process data, equipment data, and electrical conductivity data, and establish corresponding classification indexes, attribute indexes, numerical indexes, and time indexes according to key fields;

[0073] S103: Extract key patterns related to conductivity from the personalized database based on the index, identify data points that deviate from the normal pattern; decompose the performance fluctuation parameters corresponding to the data points into influencing factors at different levels, and decompose the changes in conductivity into changes in the microstructure at the material level.

[0074] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first establishes a hybrid storage architecture for a personalized database. The basic data layer stores static data such as material properties and sintering production equipment parameters; the process data layer stores dynamic data during the sintering process, such as temperature, pressure, and time series; and the performance data layer stores conductivity test results and related indicators. Secondly, the material data, process data, equipment data, and conductivity data are cleaned, and corresponding classification indexes, attribute indexes, numerical indexes, and time indexes are established according to key fields. Finally, key patterns related to conductivity are extracted from the personalized database based on the indexes, and data points that deviate from the normal patterns are identified. The performance fluctuation parameters corresponding to the data points are decomposed into influencing factors at different levels, and the changes in conductivity are decomposed into microstructural changes at the material level. Step S101 of the above scheme establishes a hybrid storage architecture for the personalized database. By dividing the data into a basic data layer, a process data layer, and a performance data layer, it achieves classified management of static, dynamic, and performance data, improving the logical clarity of data storage and retrieval efficiency. Combining the advantages of relational and non-relational databases, it can simultaneously process structured data (such as process parameters) and unstructured data (such as production logs and images), meeting diverse data storage needs. Through hierarchical design, the logical relationships between different data layers are clarified, providing a foundation for data correlation analysis and pattern recognition. Significance: Layered storage and hybrid architecture enable the database to efficiently store and manage massive amounts of data, providing a solid foundation for analysis; by clarifying the relationships between data, it creates conditions for multi-dimensional analysis of conductivity performance and identification of influencing factors; the design of the dynamic data layer enables the database to record changes in the sintering process in real time, providing support for real-time monitoring and dynamic adjustment of the production process. Step S102, data cleaning and index building, removes duplicate, missing, and outlier data through cleaning, ensuring data integrity, consistency, and accuracy, and providing a high-quality data foundation for analysis. By building categorical, attribute, numerical, and time indexes for key fields, the retrieval speed and query efficiency are significantly improved, especially when processing large-scale data. Index building not only improves retrieval efficiency but also enhances the correlation between data, facilitating pattern recognition and influencing factor decomposition. Significance: High-quality data is a prerequisite for the reliability of analysis results; data cleaning ensures the scientific validity and credibility of the analysis results; index building enables the database to quickly respond to complex queries, improving the overall system performance and user experience; through different types of indexes, data can be analyzed from multiple dimensions, supporting multi-factor correlation analysis of conductivity performance.Step S103, Pattern Recognition and Influencing Factor Decomposition, uses a dynamic pattern recognition algorithm to extract key patterns related to conductivity performance from massive amounts of data, such as periodic fluctuations, nonlinear trends, and abnormal patterns, providing a scientific basis for performance analysis. Identifying data points that deviate from normal patterns allows for timely detection of anomalies in the production process, providing clues for fault diagnosis and performance optimization. Decomposing performance fluctuation parameters into influencing factors at different levels, such as microstructural changes at the material level, clarifies the root causes of conductivity changes and provides direction for process optimization. Significance: Through pattern recognition and influencing factor decomposition, the inherent laws governing conductivity changes can be deeply explored, supporting the scientific management of the production process; the identification of anomalies enables timely detection of problems in the production process, providing a basis for dynamic adjustment and fault prevention of sintering production equipment; by clarifying influencing factors, scientific suggestions can be provided for optimizing and adjusting process parameters, thereby improving the conductivity performance and production quality of the product.

[0075] In summary, this embodiment achieves efficient data management and logical association through a hybrid storage architecture; improves data quality and retrieval efficiency through data cleaning and index construction, creating conditions for multi-dimensional analysis; and decomposes influencing factors through pattern recognition, deeply exploring the intrinsic laws governing changes in conductivity performance, providing a scientific basis for optimizing and adjusting the production process. These three interconnected steps constitute a complete process for conductivity performance data analysis, not only improving the efficiency and accuracy of data analysis but also providing strong support for the scientific management and process optimization of production. This embodies a deep integration of technology and practice, possessing significant practical application value and innovative significance.

[0076] Example 3: As Figure 3 As shown, based on Example 1, the process for obtaining the fused dataset provided in this embodiment of the invention includes the following steps:

[0077] S201: Normalize the key parameters such as temperature, pressure and time collected by different sensors, and evaluate the influence of each key parameter on the sintering quality by calculating the covariance matrix; dynamically adjust the weight allocation using an adaptive weight adjustment factor according to the real-time status of the sintering production equipment.

[0078] The formula for normalization is:

[0079]

[0080] In the formula, x ij μ represents the original value of the j-th key parameter in the i-th fused dataset; j σ represents the mean of the j-th key parameter; j x represents the standard deviation of the j-th key parameter; ij′This represents the normalized parameter values; n represents the total number of data sets to be merged.

[0081] Align the time series from different sensors for the time series T of the j-th key parameter. j The interpolation function is represented as:

[0082] S j (t)=a j t 3 +b j t 2 +c j t+d j

[0083] In the formula, S j (t) represents the interpolation result of the j-th key parameter at time t; a j ,b j ,c j ,d j Indicates the interpolation coefficients;

[0084] Covariance matrix expression:

[0085] C′=C+δ·C 2

[0086]

[0087] In the formula, C represents the original covariance matrix. δ represents the higher-order covariance coefficients; eigenvalue decomposition is performed on the expanded covariance matrix C′, and a regularization term is introduced to prevent overfitting.

[0088] C′=VΛ′V T +ηI

[0089]

[0090] In the formula, Λ′ represents the expanded eigenvalue matrix, η represents the regularization coefficient, and I represents the identity matrix;

[0091] Importance weights w of each key parameter j To indicate:

[0092]

[0093] In the formula, λ j This represents the feature value of the j-th key parameter;

[0094] Adaptive weight adjustment factor representation:

[0095] w j′ =w j ·(1+α j ·ΔPj )

[0096] ΔP j =|x ij′ -μ j′ |

[0097]

[0098] In the formula, ΔP j μ represents the real-time fluctuation amplitude of the j-th key parameter; j′ α represents the real-time mean of the j-th parameter; j Indicates the adjustment factor;

[0099] S202: The normalized key parameters are weighted and fused according to dynamic weights to generate a preliminary fused dataset. Outliers in the fused dataset are removed, and the fused dataset with outliers removed is smoothed.

[0100] in,

[0101]

[0102] In the formula, F i This represents the fusion value of the i-th fused dataset;

[0103] The nonlinear fusion function is extended to represent:

[0104]

[0105] In the formula, β1, β2, and β3 represent nonlinear fusion coefficients, which are determined by minimizing the fusion error; This indicates the introduction of trigonometric function terms to capture periodic changes; (F i -μ F ) 3 This indicates the introduction of a cubic polynomial term, used to capture higher-order nonlinear relationships.

[0106] The nonlinear fusion coefficients β1, β2, and β3 are made dynamic, adjusting to changes in data distribution. The formula for the dynamic coefficients is as follows:

[0107]

[0108] In the formula, γ k Indicates the dynamic adjustment factor; λ k σ is the eigenvalue of the k-th nonlinear term; j′ This represents the real-time standard deviation of the j-th key parameter;

[0109] Formula for removing outliers from a fused dataset:

[0110]

[0111] In the formula, ∑ represents the covariance matrix of the fused dataset; if Then remove the data point, where The 99th percentile of the chi-square distribution;

[0112] Formula for smoothing fused datasets:

[0113]

[0114] In the formula, K represents the smoothed value of the i-th fused dataset; i Indicates Kalman gain, R is the estimation error covariance of the previous time step;

[0115] S203: Standardize the optimized fusion dataset to meet the input requirements of the sintering quality assessment model, and output the standardized fusion dataset to the sintering quality assessment model for comparison with the standard parameters of the personalized database;

[0116] Fusion dataset standardization F i″ formula:

[0117]

[0118] In the formula, This represents the mean of the smoothed and merged dataset; This represents the standard deviation of the smoothed and merged dataset.

[0119] The working principle and beneficial effects of the above technical solution are as follows: First, this embodiment normalizes key parameters such as temperature, pressure, and time collected by different sensors. The influence of each key parameter on sintering quality is assessed by calculating the covariance matrix. Based on the real-time status of the sintering production equipment, an adaptive weighting adjustment factor dynamically adjusts the weight allocation. Second, the normalized key parameters are weighted and fused according to the dynamic weights to generate a preliminary fused dataset. Outliers in the fused dataset are removed, and the dataset with outliers removed is smoothed. Finally, the optimized fused dataset is standardized to meet the input requirements of the sintering quality assessment model. The standardized fused dataset is then output to the sintering quality assessment model for comparison with standard parameters in a personalized database. Step S201 of the above solution, normalization and dynamic weight adjustment, normalizes key parameters such as temperature, pressure, and time collected by different sensors, unifying data of different dimensions to the same scale. The normalization formula ensures that the mean of the data is 0 and the standard deviation is 1, thereby eliminating the influence of dimensional differences on subsequent analysis. The normalized data is comparable, facilitating weighted fusion and covariance analysis. Significance: This lays the foundation for the fusion of multi-source data, ensuring the fairness and consistency of data from different sensors. Covariance matrix and eigenvalue decomposition assess the impact of key parameters on sintering quality by calculating the covariance matrix. Introducing higher-order covariance and regularization terms prevents overfitting and ensures model stability; the covariance matrix reveals the correlation between parameters, while eigenvalue decomposition extracts the importance of key parameters. Significance: This provides a basis for dynamic weight allocation, ensuring that important parameters dominate the fusion process. Adaptive weight adjustment dynamically adjusts the weight allocation based on the real-time status of the sintering production equipment using an adaptive weight adjustment factor; the weights change dynamically with data fluctuations, capturing changes in real-time production status. Significance: This improves the flexibility and adaptability of fused data, ensuring the model can respond promptly to changes in the production environment. Step S202, weighted fusion and outlier handling, weights the normalized key parameters according to dynamic weights to generate a preliminary fused dataset; the fused dataset integrates multi-source information, more comprehensively reflecting sintering quality. Significance: This provides a high-quality data foundation for outlier detection and smoothing. Nonlinear fusion extension, by introducing trigonometric and cubic polynomial terms, captures periodic and higher-order nonlinear relationships in the data; it enhances the expressive power of the fusion model, enabling more accurate description of complex data characteristics. Significance: Improves the accuracy and reliability of fused data, providing richer information for sintering quality evaluation. Outlier removal, by calculating the chi-square distance, removes outliers from the fused dataset, eliminating noisy data and ensuring the purity of the fused data. Significance: Improves data reliability and avoids interference from outliers in the evaluation results.Smoothing is performed using Kalman filtering to smooth the fused dataset, reducing data fluctuations. The smoothed data is more stable and better reflects the long-term trend of sintering quality. Significance: Provides high-quality data for standardization and model input. Step S203 Standardization and Model Input: The optimized fused dataset is standardized to meet the input requirements of the sintering quality assessment model. Standardized data has a uniform scale and distribution, facilitating model processing and analysis. Significance: Ensures the standardization of model input, improving the accuracy and comparability of assessment results. Model Input and Comparison: The standardized fused dataset is output to the sintering quality assessment model and compared with the standard parameters of the personalized database. Through comparison, deviations and anomalies in sintering quality can be quickly identified. Significance: Provides a scientific basis for optimizing and adjusting the production process, ensuring the stability and consistency of sintering quality.

[0120] In summary, the entire fusion dataset generation process in this embodiment, through a series of techniques such as normalization, weighted fusion, outlier handling, smoothing, and standardization, transforms multi-source heterogeneous data into a high-quality, high-reliability fusion dataset. This not only improves data availability and analysis efficiency but also provides a solid data foundation for accurate evaluation and optimization of sintering quality. Its significance lies in achieving intelligent processing of production data, promoting refined management and high-quality development of sintering production; enhanced information density: through multi-source data fusion and optimization, a high-information-density dataset containing key parameters such as temperature, pressure, and time is generated, providing comprehensive data support for sintering quality evaluation; dynamic adaptability: through dynamic weight adjustment and nonlinear fusion, the algorithm can adapt to real-time changes in the state of sintering production equipment, improving the accuracy and reliability of the fusion dataset; anomaly handling capability: through outlier detection based on Mahalanobis distance and Kalman filtering, the algorithm can effectively reduce noise interference, ensuring the high quality of the fusion dataset; standardized output: through standardization processing, the fusion dataset can be directly input into the sintering quality evaluation model, simplifying the analysis process.

[0121] Example 4: Figure 4 As shown, based on Example 1, the process of constructing a sintering quality evaluation model provided in this embodiment of the invention includes the following steps:

[0122] S301: The multi-type sensor data of the sintering production equipment is mapped into a multi-dimensional state space. Based on the historical data of different materials and process conditions in the personalized database, dynamic weights are assigned to each key parameter. The action space adopts a composite structure, including two levels: macro-adjustment (such as overall temperature setting) and micro-optimization (such as local pressure fine-tuning). The reward function adopts a hierarchical design.

[0123] Among them, macro-level adjustments are used to quickly respond to significant quality deviations, while micro-level optimizations are used for refined control; the reward function adopts a hierarchical design, with the basic layer reward based on the degree of deviation of key parameters from the standard range, and the optimization layer reward providing additional incentives for long-term quality improvement through historical performance trend analysis.

[0124] S302: The adaptive exploration strategy dynamically adjusts the exploration rate based on the uncertainty of the current state; at the same time, the strategy gradient optimizes multiple objectives such as sintering quality, equipment energy consumption and production efficiency, realizing the combination of action value assessment and multi-objective optimization of strategy gradient; it receives sensor data in real time and updates the adaptive exploration strategy.

[0125] S303: The adjusted sintering quality still did not meet expectations. The sintering quality assessment model automatically triggered a self-correction mechanism to re-evaluate the definition of the state space and action space and optimize the reward function.

[0126] The working principle and beneficial effects of the above technical solution are as follows: First, this embodiment maps the multi-type sensor data of the sintering production equipment into a multi-dimensional state space. Based on the historical data of different materials and process conditions in the personalized database, dynamic weights are assigned to each key parameter. The action space adopts a composite structure, including two levels: macro-adjustment (such as overall temperature setting) and micro-optimization (such as local pressure fine-tuning). Among them, macro-adjustment is used to quickly respond to significant quality deviations, while micro-optimization is used for refined control. The reward function adopts a hierarchical design. The basic layer reward is based on the degree of deviation of key parameters from the standard range, while the optimization layer reward provides additional incentives for long-term quality improvement through historical performance trend analysis. Second, the adaptive exploration strategy dynamically adjusts the exploration rate according to the uncertainty of the current state. At the same time, the strategy gradient optimizes multiple objectives such as sintering quality, equipment energy consumption, and production efficiency, realizing the combination of action value evaluation and multi-objective optimization of strategy gradient. Sensor data is received in real time and the adaptive exploration strategy is updated. Finally, if the adjusted sintering quality still does not meet expectations, the sintering quality evaluation model automatically triggers a self-correction mechanism to re-evaluate the definition of the state space and action space and optimize the reward function. The above scheme's steps include: S301 Design of the state space, action space, and reward function; Multi-dimensional state space mapping: By mapping multi-type sensor data into a multi-dimensional state space, the model can comprehensively capture key parameters in sintering production, providing a high-quality data foundation for decision-making. Dynamic weight allocation: Based on historical data, dynamic weights are assigned to key parameters, enabling the model to adapt to different material and process conditions, improving its generalization ability; Composite action space design: The combination of macro-adjustment and micro-optimization allows the model to quickly respond to significant quality deviations while achieving refined control, improving the flexibility and efficiency of decision-making; Hierarchical reward function: The basic layer reward ensures key parameters are within the standard range, while the optimization layer reward incentivizes long-term quality improvement, providing the model with a clear optimization objective. Significance achieved: Data-driven accurate modeling: Through multi-dimensional state space and dynamic weight allocation, the model can more accurately reflect the actual situation of sintering production; Flexible and efficient decision-making capability: The composite action space design enables the model to meet the needs of different scenarios, improving the adaptability and efficiency of decision-making; Multi-objective optimization-oriented hierarchical reward function: The hierarchical reward function provides the model with an optimization objective that balances short-term quality and long-term performance, ensuring the model's sustainable development. Step S302: Adaptive exploration strategy and multi-objective optimization. The adaptive exploration strategy dynamically adjusts the exploration rate based on the uncertainty of the current state, enabling the model to learn efficiently in unknown environments while avoiding resource waste caused by over-exploration. The multi-objective optimization combines multiple objectives such as sintering quality, equipment energy consumption, and production efficiency through policy gradient optimization, enabling the model to achieve multi-objective collaborative optimization in complex scenarios. The real-time update mechanism receives sensor data in real time and updates the adaptive exploration strategy, enabling the model to quickly adapt to changes in the production environment and improve real-time decision-making capabilities.Significance Achieved: The adaptive exploration strategy minimizes resource consumption while ensuring learning efficiency; multi-objective collaborative optimization, through the combination of multi-objective optimization, enables the model to improve sintering quality while considering equipment energy consumption and production efficiency, maximizing overall benefits; real-time adaptability and the real-time update mechanism allow the model to quickly respond to changes in the production environment, improving the model's practicality and robustness. Step S303 Self-calibration Mechanism Triggering and Optimization: When the adjusted sintering quality still does not meet expectations, the model automatically triggers the self-calibration mechanism, re-evaluates the definition of the state space and action space, and optimizes the reward function; dynamic optimization capability: Through the self-calibration mechanism, the model can continuously optimize its performance in actual production, improving long-term stability and robustness. Significance Achieved: Dynamic optimization and self-improvement: The self-calibration mechanism enables the model to continuously optimize its performance in practical applications, improving long-term stability and robustness; continuous improvement and adaptive enhancement: Through dynamic optimization of the state space, action space, and reward function, the model can better adapt to complex and ever-changing production environments, ensuring long-term efficient operation.

[0127] In summary, this embodiment provides the model with data-driven, accurate modeling and flexible, efficient decision-making capabilities, laying the foundation for the model. By combining adaptive exploration strategies and multi-objective optimization, the model's learning efficiency and comprehensive optimization capabilities are improved. Through a self-correction mechanism, the model achieves dynamic optimization and self-improvement, ensuring its long-term stability and robustness. These three steps together constitute the complete construction process of the sintering quality assessment model, enabling it to achieve efficient and intelligent decision optimization in complex and ever-changing sintering production environments.

[0128] Example 5: Figure 5 As shown, based on Example 4, the reward function provided in this embodiment of the invention adopts a hierarchical design process, which includes the following steps:

[0129] S3011: Design the basic layer reward, select key parameters for sintering quality, calculate the deviation between the actual value and the standard value of each key parameter; penalize parameters with large deviations to ensure that the model tends to control the parameters within the standard range; design the optimization layer reward by providing additional incentives for long-term quality improvement through historical performance trend analysis.

[0130] Here, we assume there are N key parameters, and the actual value of the j-th key parameter is x. j The standard value is Deviation is

[0131] Base layer reward R base Represented as:

[0132]

[0133] In the formula, w′ j It is the dynamic weight of the j-th key parameter, which is dynamically adjusted based on historical data and process conditions; the negative sign indicates that the larger the deviation, the smaller the reward (the greater the penalty);

[0134] Assume that within the time window T, the historical average value of the key parameter j is... The current value is x j ;

[0135] Optimization layer reward R opt It can be represented as:

[0136]

[0137] In the formula, α j It is the long-term optimization weight of the j-th key parameter, which is dynamically adjusted according to historical performance trends; the positive sign indicates that the reward increases when the current value is better than the historical average.

[0138] S3012: Combines the basic layer reward and the optimization layer reward to obtain a comprehensive reward function; to adapt to different material and process conditions, the dynamic weight is adjusted according to historical data and the current state;

[0139] R = R base +β·R opt

[0140] In the formula, β is the weighting coefficient of the optimization layer reward, which is used to balance short-term quality control and long-term performance improvement;

[0141] To adapt to different materials and process conditions, dynamic weight w′ j and α j Adjustments need to be made based on historical data and the current situation;

[0142]

[0143] In the formula, It is the historical variance of the j-th key parameter, reflecting its volatility; It represents the historical average change trend of the j-th key parameter;

[0144] S3013: When the sintering quality assessment model triggers the self-correction mechanism, the reward function is re-evaluated and optimized. Based on changes in the current production environment, the definitions of key parameters and action space are adjusted; dynamic weights are recalculated based on the latest data; and the weight coefficients of the basic layer reward and the optimization layer reward are adjusted based on the optimization results.

[0145] The working principle and beneficial effects of the above technical solution are as follows: Firstly, the basic layer reward is designed, selecting key parameters for sintering quality and calculating the deviation between the actual value and the standard value of each key parameter. Parameters with large deviations are penalized to ensure the model tends to control the parameters within the standard range. Through historical performance trend analysis, additional incentives are provided for long-term quality improvement, and an optimization layer reward is designed. Secondly, the basic layer reward and optimization layer reward are combined to obtain a comprehensive reward function. To adapt to different materials and process conditions, the dynamic weights are adjusted based on historical data and the current state. Finally, when the sintering quality assessment model triggers a self-correction mechanism, the reward function is re-evaluated and optimized, adjusting the definitions of key parameters and action space according to changes in the current production environment. The dynamic weights are recalculated based on the latest data. Based on the optimization results, the weight coefficients of the basic layer reward and optimization layer reward are adjusted. Step S3011 of the above solution, the design of the basic layer reward and optimization layer reward, by calculating the deviation between the actual value and the standard value of key parameters and penalizing parameters with large deviations, ensures that the model tends to control the parameters within the standard range, effectively controlling short-term quality fluctuations. Significance: The introduction of the basic layer reward provides the model with a clear optimization objective, namely, reducing the deviation of key parameters, thereby ensuring the stability and consistency of sintering quality. The adjustment of dynamic weights further enhances the model's adaptability, enabling it to flexibly adjust the penalty intensity according to different process conditions and historical data. The optimization layer reward provides additional incentives for long-term quality improvement by comparing the current value with the historical average; when the current value is better than the historical average, the reward increases, thereby driving the model to continuously optimize its long-term performance. Significance: The introduction of the optimization layer reward compensates for the limitations of the basic layer reward, focusing not only on short-term quality fluctuations but also on long-term performance improvement; the design helps to promote continuous improvement and innovation in the sintering process. Step S3012, the construction of the comprehensive reward function, combines the basic layer reward and the optimization layer reward. The comprehensive reward function can simultaneously consider short-term quality control and long-term performance improvement; by dynamically adjusting the weight coefficients, the model can find a balance point between different objectives according to actual needs. Significance: The design of the comprehensive reward function enables the model to flexibly cope with complex production environments, ensuring the stability of sintering quality and providing impetus for long-term optimization. The introduction of dynamic weights further enhances the model's adaptability and robustness. Step S3013: Triggering the self-calibration mechanism and optimizing the reward function. When the sintering quality assessment model triggers the self-calibration mechanism, the reward function will be re-evaluated and optimized according to changes in the current production environment. By adjusting the definitions of key parameters and action space, and recalculating dynamic weights, the model can quickly adapt to new production conditions. Significance: The introduction of the self-calibration mechanism significantly improves the model's adaptability and intelligence level; it can not only cope with changes in the production environment but also dynamically adjust the optimization strategy based on the latest data, ensuring that the sintering quality is always at its best.

[0146] In summary, the hierarchical design of the reward function in this embodiment achieves the dual goals of short-term quality control and long-term performance improvement through the combination of basic layer rewards and optimization layer rewards. The introduction of dynamic weights enables the model to flexibly adapt to different process conditions and production environments, while the self-correction mechanism further enhances the model's intelligence and adaptability. This not only improves the stability and consistency of sintering quality but also provides strong support for continuous process optimization and innovation. The hierarchical structure of the reward function design process balances short-term quality control and long-term performance improvement. Basic layer rewards ensure that key parameters are within standard ranges, while optimization layer rewards incentivize long-term quality improvement. Through dynamic weight adjustment and the self-correction mechanism, the reward function can adapt to complex and ever-changing production environments, providing the model with clear optimization objectives.

[0147] Example 6: As Figure 6 As shown, based on Example 4, the process of combining action value evaluation and policy gradient multi-objective optimization provided by this embodiment of the invention includes the following steps:

[0148] S3021: Sintering quality, equipment energy consumption, and production efficiency are identified as core optimization objectives. The multi-objective optimization problem is transformed into a unified evaluation system. Different objectives are integrated into a comprehensive optimization function through weighting. Action value assessment is used to quantify the contribution of each action to multi-objective optimization. The action value assessment at the macro and micro levels is combined to form a composite evaluation system.

[0149] S3022: Decompose the multi-objective optimization problem into multiple sub-objectives, each sub-objective corresponding to an independent policy gradient optimization path; by weighting, merge the policy gradients of each sub-objective into a comprehensive gradient to guide the policy update direction; dynamically adjust the weights of each sub-objective according to the current state and objective priority.

[0150] S3023: Dynamically adjust the exploration rate based on the uncertainty of the current state and the progress of optimization; receive sensor data in real time and update the exploration strategy.

[0151] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first determines sintering quality, equipment energy consumption, and production efficiency as core optimization objectives, transforming the multi-objective optimization problem into a unified evaluation system. Different objectives are integrated into a comprehensive optimization function through weighting. Action value assessment is used to quantify the contribution of each action to multi-objective optimization, combining macro- and micro-level action value assessments to form a composite evaluation system. Secondly, the multi-objective optimization problem is decomposed into multiple sub-objectives, each corresponding to an independent policy gradient optimization path. Through weighting, the policy gradients of each sub-objective are merged into a comprehensive gradient to guide the policy update direction. The weights of each sub-objective are dynamically adjusted based on the current state and objective priority. Finally, the exploration rate is dynamically adjusted based on the uncertainty of the current state and the optimization progress. Sensor data is received in real time to update the exploration strategy. Step S3021 of the above scheme involves constructing a multi-objective optimization framework and evaluating action value. This integrates sintering quality, equipment energy consumption, and production efficiency into a single comprehensive optimization function, avoiding local optima problems that may arise from single-objective optimization. By combining macro- and micro-level action value evaluation, it ensures both the efficiency of global optimization and the precision of local control. The weights of each objective are dynamically adjusted based on the current production status and priority, ensuring the optimization process can flexibly adapt to the needs of different scenarios. The significance achieved is that a unified evaluation system enables comprehensive optimization of multiple objectives, significantly improving the overall performance of sintering production. The composite evaluation system allows the model to simultaneously focus on global and local optimization, improving the precision and adaptability of control. The dynamic weight mechanism allows the optimization process to be flexibly adjusted according to actual needs, enhancing the model's practicality and robustness. Step S3022, multi-objective optimization of policy gradients, decomposes the multi-objective optimization problem into multiple sub-objectives, simplifying the complexity of the optimization problem. Weighted fusion of the policy gradients of each sub-objective into a single comprehensive gradient ensures the uniformity and consistency of the optimization direction. The weights of each sub-objective are dynamically adjusted based on the current state and objective priority, enabling the optimization process to adapt to the needs of different scenarios. Significance achieved: Objective decomposition and gradient fusion enable efficient solutions to multi-objective optimization problems, improving model convergence speed and optimization performance; dynamic weight adjustment ensures balance and coordination among sub-objectives, avoiding conflicts that may arise from single-objective optimization; multi-objective optimization using policy gradients allows the model to dynamically adjust its optimization strategy according to actual needs, significantly enhancing the model's intelligence level. Step S3023, the introduction of an adaptive exploration strategy, dynamically adjusts the exploration rate based on the uncertainty of the current state and the optimization progress, ensuring a balance between exploration and utilization in the optimization process; by receiving sensor data in real time and updating the exploration strategy, the model can quickly respond to changes in the production environment.Significance achieved: Dynamically adjusting the exploration rate enables the model to increase exploration when uncertainty is high in order to discover better actions; it reduces exploration when optimization tends to stabilize in order to improve efficiency; the real-time feedback mechanism enables the model to quickly adapt to changes in the production environment, significantly improving the model's real-time performance and adaptability; the introduction of the adaptive exploration strategy enables the optimization process to maintain high efficiency and stability in complex and ever-changing production environments, enhancing the model's robustness.

[0152] In summary, this embodiment achieves a combination of action value assessment and strategy gradient multi-objective optimization, realizing the following overall significance: Comprehensiveness and Refinement: Through a unified and composite assessment system, comprehensive optimization and refined control of sintering quality, equipment energy consumption, and production efficiency are achieved. Efficiency and Intelligence: Through objective decomposition, gradient fusion, and dynamic weight adjustment, efficient solution and intelligent adjustment of multi-objective optimization problems are achieved. Real-time Performance and Adaptability: Through adaptive exploration strategies and real-time feedback mechanisms, the model can quickly respond to changes in the production environment, significantly improving its real-time performance and adaptability. Robustness and Stability: Through dynamic weight mechanisms and self-correction mechanisms, the model can maintain high efficiency and stability in complex and ever-changing production environments, enhancing its robustness. This not only significantly improves the performance and practicality of the sintering quality assessment model but also provides an innovative solution to multi-objective optimization problems in complex industrial scenarios. This embodiment constructs a hierarchical and highly adaptable sintering quality assessment model by combining action value evaluation and strategy gradient multi-objective optimization. Through the combination of multi-objective optimization framework, action value evaluation, strategy gradient optimization, adaptive exploration strategy and self-correction mechanism, it achieves comprehensive optimization of sintering quality, equipment energy consumption and production efficiency, which has significant technical creativity and practicality.

[0153] Example 7: Figure 7 As shown, based on Example 6, the process of dynamically adjusting the exploration rate and updating the exploration strategy provided by this embodiment of the invention includes the following steps:

[0154] S30231: The current state is evaluated through multidimensional uncertainty quantification; the target conflict index is used to measure the mutual influence between various optimization objectives, while the historical data deviation reflects the degree of deviation between the current observation and the historical trend; a dynamic uncertainty score is generated through comprehensive calculation.

[0155] S30232: Nonlinear optimization progress evaluation is used to conduct in-depth analysis of optimization progress; the sub-objective convergence speed is used to evaluate the optimization efficiency of each sub-objective, while the strategy stability index measures the robustness of the current strategy under different states, dynamically identifies bottlenecks in the optimization process, and adjusts the exploration strategy.

[0156] S30233: Based on uncertainty scoring and optimization progress assessment results, adjust the exploration rate according to the target priority weight and environmental complexity factor.

[0157] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first evaluates the current state through multidimensional uncertainty quantification; the target conflict index is used to measure the mutual influence between various optimization objectives, while the historical data deviation reflects the degree of deviation between the current observation and the historical trend; through comprehensive calculation, a dynamic uncertainty score is generated; secondly, nonlinear optimization progress evaluation is used to conduct in-depth analysis of optimization progress; the sub-objective convergence speed is used to evaluate the optimization efficiency of each sub-objective, while the strategy stability index measures the robustness of the current strategy under different states, dynamically identifies bottlenecks in the optimization process, and adjusts the exploration strategy; finally, based on the uncertainty score and optimization progress evaluation results, the exploration rate is adjusted according to the target priority weight and environmental complexity factor. Step S30231 of the above solution evaluates the current state through multidimensional uncertainty quantification; the target conflict index quantifies the mutual influence between various optimization objectives and identifies the competitive or cooperative relationship between objectives; the historical data deviation evaluates the degree of deviation between the current observation and the historical trend, reflecting the impact of environmental dynamics or data noise; the dynamic uncertainty score integrates the above indicators to generate a quantitative value reflecting the uncertainty of the current state, providing a basis for adjusting the exploration rate. Significance Achieved: Precise state perception, through multi-dimensional uncertainty quantification, enables comprehensive perception of the complexity and dynamism of the current state, avoiding misjudgments caused by local information bias; dynamic adaptability provides a scientific basis for adjusting the exploration rate, ensuring the system can flexibly respond to environmental changes and improving optimization robustness. Step S30232 employs nonlinear optimization progress evaluation to conduct in-depth analysis of optimization progress, assesses the optimization efficiency of each sub-objective by evaluating sub-objective convergence speed, and identifies bottlenecks or inefficient links in the optimization process; the strategy stability index measures the robustness of the current strategy under different states, ensuring the reliability of the strategy in dynamic environments; optimization bottleneck identification dynamically identifies bottlenecks in the optimization process, providing direction for adjusting the exploration strategy. Significance Achieved: Improved optimization efficiency: By identifying differences in sub-objective convergence speed, inefficient links can be optimized in a targeted manner, accelerating the overall optimization process; Enhanced strategy robustness: By evaluating strategy stability, the system can avoid strategy failure due to environmental changes, ensuring the continuity of the optimization process; Dynamic strategy adjustment provides a scientific basis for updating the exploration strategy, ensuring the strategy can adapt to constantly changing optimization needs. Step S30233 adjusts the exploration rate based on the uncertainty score and optimization progress assessment results. The target priority weight dynamically adjusts the exploration rate according to the importance of each target, giving priority to satisfying high-priority targets. The environmental complexity factor flexibly adjusts the exploration rate according to the dynamics and complexity of the environment to avoid ineffective or excessive exploration. The exploration rate adjustment takes into account the above factors to dynamically balance exploration and utilization, ensuring the efficiency and effectiveness of the optimization process.Significance achieved: Optimized resource allocation: By dynamically adjusting the exploration rate, limited resources can be concentrated in the direction most in need of exploration, improving optimization efficiency; Approaching the global optimum: By balancing exploration and utilization, it can avoid getting trapped in local optima and gradually approach the global optimum; Enhanced environmental adaptability: By adjusting the exploration rate according to the environmental complexity factor, it can maintain efficient operation in complex and ever-changing environments.

[0158] In summary, this embodiment achieves precise state perception through multidimensional uncertainty quantification, comprehensively assessing the current state and providing a scientific basis for decision-making; efficient optimization progress is evaluated through nonlinear optimization progress assessment, identifying bottlenecks and improving optimization efficiency, ensuring the robustness of the strategy; and dynamic balance between exploration and utilization optimizes resource allocation through adaptive exploration rate adjustment, gradually approaching the global optimum. This not only improves the efficiency and effectiveness of multi-objective optimization but also demonstrates its powerful application value in complex environments through dynamic adaptability and robustness.

[0159] Example 8: As Figure 8 As shown, based on Example 4, the process of re-evaluating the definition of the state space and action space provided in this embodiment of the invention includes the following steps:

[0160] S3031: Based on the current quality deviation and optimization objectives, re-evaluate the contribution of each parameter in the multi-dimensional state space, identify key parameters that have a significant impact on quality; assign new dynamic weights to key parameters based on real-time sensor data and historical performance trends; and incorporate external environmental variables and equipment status into the state space.

[0161] S3032: Redefine the magnitude and priority of macroscopic adjustment actions based on the severity of the current quality deviation; control actions for local parameter fine-tuning of local temperature and pressure gradient adjustment;

[0162] S3033: Based on the current distribution characteristics of quality deviations, readjust the weights of key parameters and standard range deviations; introduce a dynamic threshold mechanism to dynamically adjust the standard range based on historical performance trends; and provide additional incentives for long-term quality improvement through historical performance trend analysis.

[0163] The working principle and beneficial effects of the above technical solution are as follows: First, based on the current quality deviation and optimization target, this embodiment re-evaluates the contribution of each parameter in the multi-dimensional state space and identifies key parameters that have a significant impact on quality; based on real-time sensor data and historical performance trends, new dynamic weights are assigned to key parameters; external environmental variables and equipment status are incorporated into the state space; second, based on the severity of the current quality deviation, the magnitude and priority of macroscopic adjustment actions are redefined; control actions for local parameter local temperature fine-tuning and pressure gradient adjustment are implemented; finally, based on the distribution characteristics of the current quality deviation, the weights of the deviations between key parameters and the standard range are readjusted; a dynamic threshold mechanism is introduced to dynamically adjust the standard range based on historical performance trends; and through historical performance trend analysis, additional incentives are provided for long-term quality improvement. The above scheme includes steps S3031: re-evaluation of the state space; parameter contribution analysis, which identifies key parameters that significantly impact quality and precisely focuses on the factors contributing most to quality deviations, improving the state space's relevance; dynamic weight adjustment, which assigns new dynamic weights to key parameters based on real-time sensor data and historical performance trends, ensuring the state space more accurately reflects the current production status; and inclusion of environmental and equipment status, which incorporates external environmental variables (such as temperature and humidity) and equipment status (such as aging level) into the state space, enhancing the model's comprehensiveness and adaptability. Significance: Precise perception: By optimizing the state space, the model can more accurately perceive the current production status, providing a more reliable basis for action decisions; improved adaptability: Including environmental and equipment status enables the model to better cope with complex and changing production environments, improving its robustness; efficiency optimization: Focusing on key parameters reduces interference from redundant information, improving the model's operating efficiency. Step S3032: Redefining the Action Space. Macro-level adjustment actions are optimized by redefining the magnitude and priority of macro-level adjustments based on the severity of the current quality deviation, ensuring rapid response to significant deviations. Micro-level optimization actions are expanded by adding control actions for local parameters (such as local temperature fine-tuning and pressure gradient adjustment), improving the fine-grained control capability of the action space. Significance: Rapid response: By optimizing macro-level adjustments, the model can quickly correct significant quality deviations, reducing production losses. Fine-grained control: Expanding micro-level optimization actions allows the model to finely adjust local parameters, improving the stability and consistency of overall quality. Multi-level control: By combining macro-level and micro-level actions, a multi-level, multi-dimensional control strategy is achieved, improving the overall performance of the model. Step S3033: Optimizing the Reward Function. Deviation weight adjustment: Based on the distribution characteristics of the current quality deviation, the weights of key parameters and standard range deviations are readjusted to ensure the reward function more accurately reflects the severity of quality deviations. Dynamic threshold mechanism: A dynamic threshold mechanism is introduced to dynamically adjust the standard range based on historical performance trends, avoiding reward deviations caused by fixed thresholds.Long-term performance incentives, through historical performance trend analysis, provide additional incentives for long-term quality improvements (such as stability and consistency), guiding the model to focus on long-term optimization goals. Significance: Accurate evaluation: By optimizing bias weights and introducing dynamic thresholds, the reward function can more accurately assess the current quality state, providing more reliable feedback to the model; Long-term optimization: Through additional incentives for long-term performance, it guides the model to focus on improving long-term quality while quickly responding to short-term biases; Multi-objective balancing: By dynamically adjusting the reward function, it ensures that the model can balance multiple objectives such as quality, energy consumption, and efficiency, improving its overall performance.

[0164] In summary, this embodiment significantly improves the overall performance of the model through precise perception, rapid response, and refined control; enhances the model's adaptability to complex and changing environments by incorporating environmental and equipment states and dynamically adjusting weights and thresholds; guides the model to focus on long-term quality improvement while rapidly responding to short-term deviations through long-term performance incentives and multi-objective balancing; and reduces production losses and improves production efficiency and product quality by optimizing control strategies and reward functions. This not only demonstrates the creative value of technological innovation but also provides a more efficient and reliable intelligent solution for sintering production.

[0165] Example 9: As Figure 9 As shown, based on Example 8, the process of dynamically adjusting the standard range according to historical performance trends provided in this embodiment of the invention includes the following steps:

[0166] S30331: Based on historical performance data, identify the regular characteristics of its changes over time, extract key features from historical performance trends, and use trend models to predict the possible range of changes in key parameters in the future.

[0167] S30332: Set an initial threshold range based on the statistical characteristics of historical performance data as a benchmark for dynamic adjustment; compare the actual values ​​of current key parameters with historical performance trends to assess the degree of deviation.

[0168] S30333: Based on real-time performance evaluation results and trend prediction information, dynamically adjust the threshold range; if the current parameter value is higher than the historical trend prediction value, increase the upper limit of the threshold; if it is lower than the prediction value, decrease the lower limit of the threshold.

[0169] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first identifies the regular characteristics of historical performance data over time and extracts key features from historical performance trends; it then uses a trend model to predict the possible range of changes in key parameters over a future period; secondly, it sets an initial threshold range based on the statistical characteristics of historical performance data as a benchmark for dynamic adjustment; it compares the actual value of the current key parameter with the historical performance trend to assess the degree of deviation; finally, based on the real-time performance evaluation results and combined with trend prediction information, it dynamically adjusts the threshold range; if the current parameter value is higher than the historical trend prediction value, the upper limit of the threshold is increased; if it is lower than the prediction value, the lower limit of the threshold is decreased. Step S30331 of the above solution involves identifying the regular characteristics of key parameters over time (such as periodicity, trends, and fluctuations) by analyzing historical performance data, providing a theoretical basis for dynamic adjustment; and trend prediction using a trend model to predict the possible range of changes in key parameters over a future period, providing a forward-looking basis for adjusting the dynamic threshold. Significance Achieved: Enhanced Predictive Ability: Through trend prediction, the model can anticipate changes in key parameters in advance, reducing misjudgments caused by sudden changes; Enhanced Adaptability: Based on the regularity of historical data, the model can better adapt to complex and ever-changing production environments, improving robustness. Step S30332: Set an initial threshold range based on the statistical characteristics of historical performance data as a benchmark for dynamic adjustment; Compare the actual values ​​of current key parameters with historical performance trends to assess their deviation; Initial threshold setting: Based on the statistical characteristics of historical performance data (such as mean, standard deviation, etc.), set an initial threshold range to provide a benchmark for dynamic adjustment; Real-time Performance Evaluation: By comparing the actual values ​​of current key parameters with historical performance trends, assess their deviation and provide real-time basis for dynamic adjustment. Significance Achieved: Ensured Threshold Rationality: The initial threshold range, set based on the statistical characteristics of historical data, can reflect the normal range of changes in key parameters, avoiding overly broad or narrow threshold settings; Real-time Monitoring and Feedback: Through real-time performance evaluation, the model can promptly detect abnormal changes in key parameters, providing timely feedback for dynamic adjustment. Step S30333, dynamic threshold adjustment, dynamically adjusts the threshold range based on real-time performance evaluation results and trend prediction information to reflect changes in the current production status. Threshold smoothing, through dynamic adjustment, avoids abrupt changes in the threshold range, ensuring its stability and continuity. The significance is that dynamic thresholds can more accurately reflect the actual range of changes in key parameters, reducing the possibility of misjudgments and omissions; through dynamic adjustment, the model can adapt to changes in the production environment, improving robustness and practicality; the dynamic threshold mechanism can guide the model to focus on long-term performance improvement, achieving continuous improvement in production quality.

[0170] In summary, this embodiment, through historical performance trend analysis and dynamic threshold adjustment, enables the model to more accurately reflect the changing patterns of key parameters and adapt to complex and ever-changing production environments. Real-time performance evaluation and dynamic adjustment mechanisms can promptly detect and correct abnormal changes in key parameters, ensuring the stability of the production process. Through trend prediction and dynamic threshold adjustment, the model can focus on long-term performance improvement, providing technical support for continuous improvement of production quality. Combining historical performance trends with dynamic threshold adjustment overcomes the limitations of traditional fixed thresholds, demonstrating the creativity and practicality of the technical solution. This not only improves the model's performance but also provides new ideas and methods for the intelligent optimization of production systems, possessing significant technical value and practical significance.

[0171] Example 10: As Figure 10 As shown in Example 9, the process of comparing the actual value of the current key parameter with the historical performance trend provided by this embodiment of the invention includes the following steps:

[0172] S303321: Align the actual values ​​of current key parameters with historical performance trends over time; standardize the actual values ​​of current key parameters and historical performance trends to eliminate differences in dimensions and units; extract key features from historical performance trends as a benchmark for comparison.

[0173] S303322: Match the actual values ​​of current key parameters with key features of historical performance trends to identify performance differences across different dimensions; calculate the absolute difference between the actual values ​​of current key parameters and the predicted values ​​of historical performance trends to reflect the magnitude of the deviation; compare the absolute deviation with the fluctuation range of historical performance trends to calculate the relative deviation rate to reflect the severity of the deviation.

[0174] S303323: Analyze the consistency between the current actual value change trend of key parameters and the historical performance trend to determine whether it deviates from the expected direction; based on multi-dimensional indicators such as absolute deviation and relative deviation rate, comprehensively score the degree of deviation of the current key parameters; according to the comprehensive score results, classify the degree of deviation into different levels.

[0175] The working principle and beneficial effects of the above technical solution are as follows: First, this embodiment aligns the actual value of the current key parameter with the historical performance trend over time; it then standardizes the actual value of the current key parameter and the historical performance trend to eliminate differences in dimensions and units; key features are extracted from the historical performance trend as a benchmark for comparison; second, the actual value of the current key parameter is matched with the key features of the historical performance trend to identify performance differences in different dimensions; the absolute difference between the actual value of the current key parameter and the predicted value of the historical performance trend is calculated to reflect the magnitude of the deviation; the absolute deviation is compared with the fluctuation range of the historical performance trend to calculate the relative deviation rate, reflecting the severity of the deviation; finally, the consistency between the actual value change trend of the current key parameter and the historical performance trend is analyzed to determine whether it deviates from the expected direction; based on multi-dimensional indicators such as the absolute deviation and the relative deviation rate, a comprehensive score is given for the degree of deviation of the actual value of the current key parameter; based on the comprehensive score result, the degree of deviation is divided into different levels. Step S303321, data alignment and preprocessing, ensures that the actual value of the current key parameter and the historical performance trend are on the same time scale, avoiding comparison errors caused by time deviation. Significance: Provides an accurate time benchmark for comparative analysis, enhancing the reliability of evaluation results. Standardization eliminates the differences in dimensions and units between the actual values ​​of current key parameters and historical performance trends, making them comparable. Significance: Improves the fairness and consistency of comparisons, avoiding misjudgments caused by different data units. Key feature extraction provides clear reference for matching and deviation assessment, enhancing the relevance of the assessment. Step S303322 Feature Matching and Deviation Quantification: Comprehensively reveals the differences between the actual values ​​of current key parameters and historical trends, providing multi-dimensional information for deviation assessment; intuitively quantifies the gap between the actual values ​​of current key parameters and historical trends, providing basic data for assessment; further assesses the severity of deviation based on the historical fluctuation range, avoiding underestimation of the degree of deviation due to large historical fluctuations. Step S303323 Trend Consistency Analysis and Comprehensive Assessment: Assesses whether the actual values ​​of current key parameters are consistent with historical trends from a trend perspective, providing a basis for deviation cause analysis; through comprehensive assessment of multi-dimensional indicators, avoids the limitations of a single indicator, improving the comprehensiveness and accuracy of the assessment results; provides clear grading criteria for dynamic adjustment, facilitating the adoption of corresponding adjustment strategies for different degrees of deviation.

[0176] In summary, this embodiment ensures the accuracy of comparative analysis and avoids misjudgments caused by data bias through time alignment, standardization, and key feature extraction. From feature matching to deviation quantification, trend consistency analysis, and comprehensive evaluation, it covers all aspects of deviation assessment, ensuring the comprehensiveness of the evaluation results. Based on real-time performance data and historical trend information, it can dynamically assess the degree of deviation, providing a real-time basis for dynamic adjustments. Through deviation level classification, it provides clear guidance for practical applications, facilitating appropriate adjustment measures for different degrees of deviation. This not only improves the accuracy and comprehensiveness of deviation assessment but also provides a scientific basis for dynamically adjusting the standard range, possessing significant technical value and practical significance.

[0177] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of equivalents of this invention, this invention is also intended to include these modifications and variations.

Claims

1. A method for evaluating the sintering quality of an artificial thermally conductive film, characterized in that, Includes the following steps: Construct a personalized database containing electrical conductivity data of artificial thermal conductive films under different materials and process conditions; through analysis of electrical conductivity data, identify the performance fluctuation parameters of different sintering production equipment and the same sintering production equipment at different production periods; Multiple types of sensors are deployed on the artificial heat-conducting film sintering production equipment to collect key parameters such as temperature, pressure and time in the sintering production equipment. Multiple types of key parameters are fused using a data fusion algorithm to obtain a fused dataset; the sintering production equipment is then adjusted based on performance fluctuation parameters. Key parameters from the fusion dataset are input into the sintering quality assessment model and compared with standard parameters from the personalized database to obtain comparison results. The comparison results are used to identify which key parameters deviate from the standard range and are automatically fed back to the sintering production equipment for corresponding adjustments. The process of constructing a sintering quality assessment model includes the following steps: The multi-type sensor data of the sintering production equipment is mapped into a multi-dimensional state space. Based on the historical data of different materials and process conditions in the personalized database, dynamic weights are assigned to each key parameter. The action space adopts a composite structure, including two levels: macroscopic adjustment and microscopic optimization. The reward function adopts a hierarchical design. Among them, macro-level adjustments are used to quickly respond to significant quality deviations, while micro-level optimizations are used for refined control; the reward function adopts a hierarchical design, with the basic layer reward based on the degree of deviation of key parameters from the standard range, and the optimization layer reward providing additional incentives for long-term quality improvement through historical performance trend analysis. The adaptive exploration strategy dynamically adjusts the exploration rate based on the uncertainty of the current state; at the same time, the strategy gradient optimizes multiple objectives such as sintering quality, equipment energy consumption and production efficiency, realizing the combination of action value assessment and multi-objective optimization of strategy gradient; and receives sensor data in real time and updates the adaptive exploration strategy. The adjusted sintering quality still did not meet expectations. The sintering quality assessment model automatically triggered a self-correction mechanism to re-evaluate the definition of the state space and action space and optimize the reward function.

2. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 1, characterized in that, The process of analyzing conductivity data includes the following steps: A hybrid storage architecture for a personalized database is constructed, in which the basic data layer stores static data on material properties and sintering production equipment parameters; and the process data layer stores dynamic data during the sintering process. The performance data layer stores the conductivity test results and related indicators; Data cleaning was performed on material data, process data, equipment data, and electrical conductivity data, and corresponding classification indexes, attribute indexes, numerical indexes, and time indexes were established according to key fields; Based on the index, key patterns related to conductivity are extracted from the personalized database, and data points that deviate from the normal pattern are identified. The performance fluctuation parameters corresponding to the data points are decomposed into influencing factors at different levels, and the changes in conductivity are decomposed into changes in the microstructure at the material level.

3. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 1, characterized in that, The process of obtaining the fused dataset includes the following steps: The temperature, pressure and time key parameters collected by different sensors are normalized, and the influence of each key parameter on the sintering quality is evaluated by calculating the covariance matrix. The weight allocation is dynamically adjusted using an adaptive weight adjustment factor based on the real-time status of the sintering production equipment. The normalized key parameters are weighted and fused according to dynamic weights to generate a preliminary fused dataset. Outliers in the fused dataset are removed, and the fused dataset with outliers removed is smoothed. The optimized fusion dataset is standardized to meet the input requirements of the sintering quality assessment model. The standardized fusion dataset is then output to the sintering quality assessment model for comparison with the standard parameters of the personalized database.

4. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 1, characterized in that, The process of using a hierarchical design for the reward function includes the following steps: Design the base layer reward, select the key parameters for sintering quality, and calculate the deviation between the actual value and the standard value of each key parameter; By analyzing historical performance trends, additional incentives are provided for long-term quality improvement through design optimization layer rewards; The basic layer rewards and optimization layer rewards are combined to obtain a comprehensive reward function; to adapt to different material and process conditions, the dynamic weights are adjusted based on historical data and the current state. When the sintering quality assessment model triggers the self-correction mechanism, the reward function is re-evaluated and optimized, and the definitions of key parameters and action space are adjusted according to changes in the current production environment; dynamic weights are recalculated based on the latest data. Based on the optimization results, adjust the weight coefficients of the basic layer rewards and the optimization layer rewards.

5. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 1, characterized in that, The process of combining action value assessment and policy gradient multi-objective optimization includes the following steps: Sintering quality, equipment energy consumption, and production efficiency are identified as the core optimization objectives. The multi-objective optimization problem is transformed into a unified evaluation system, and different objectives are integrated into a comprehensive optimization function through weighting. Action value assessment is used to quantify the contribution of each action to multi-objective optimization, combining macro and micro-level action value assessments to form a composite assessment system. The multi-objective optimization problem is decomposed into multiple sub-objectives, each corresponding to an independent policy gradient optimization path. By weighting, the policy gradients of each sub-objective are merged into a comprehensive gradient to guide the policy update direction. The weights of each sub-objective are dynamically adjusted according to the current state and objective priority. The exploration rate is dynamically adjusted based on the current uncertainty and optimization progress; sensor data is received in real time to update the exploration strategy.

6. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 5, characterized in that, The process of dynamically adjusting the exploration rate and updating the exploration strategy includes the following steps: The current state is evaluated through multidimensional uncertainty quantification; the target conflict index is used to measure the mutual influence between various optimization objectives, while the historical data deviation reflects the degree of deviation between the current observation and the historical trend; a dynamic uncertainty score is generated through comprehensive calculation. Nonlinear optimization progress evaluation is used to conduct in-depth analysis of optimization progress; sub-objective convergence speed is used to evaluate the optimization efficiency of each sub-objective, while the strategy stability index measures the robustness of the current strategy under different states, dynamically identifies bottlenecks in the optimization process, and adjusts the exploration strategy. Based on the uncertainty score and optimization progress assessment results, the exploration rate is adjusted according to the target priority weight and environmental complexity factor.

7. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 1, characterized in that, The process of re-evaluating the definitions of state space and action space includes the following steps: Based on the current quality deviation and optimization objectives, the contribution of each parameter in the multidimensional state space is reassessed, and key parameters that have a significant impact on quality are identified. According to real-time sensor data and historical performance trends, new dynamic weights are assigned to key parameters. External environmental variables and equipment status are incorporated into the state space. Based on the severity of the current quality deviation, redefine the magnitude and priority of macroscopic adjustment actions; control actions for local parameter fine-tuning of local temperature and pressure gradient adjustment; Based on the current distribution characteristics of quality deviations, the weights of key parameters and standard range deviations are readjusted; a dynamic threshold mechanism is introduced to dynamically adjust the standard range based on historical performance trends; and additional incentives are provided for long-term quality improvement through historical performance trend analysis.

8. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 7, characterized in that, The process of dynamically adjusting the standard range based on historical performance trends includes the following steps: Based on historical performance data, identify the regular characteristics of its changes over time and extract key features from historical performance trends. Use trend models to predict the possible range of changes in key parameters over a future period; Compare the current actual values ​​of key parameters with historical performance trends to assess the degree of deviation. Based on real-time performance evaluation results and trend prediction information, the threshold range is dynamically adjusted; if the current parameter value is higher than the historical trend prediction value, the upper limit of the threshold is increased. If the value is lower than the predicted value, the lower limit of the threshold is lowered.

9. The detection method for evaluating the sintering quality of artificial thermally conductive films as described in claim 8, characterized in that, The process of comparing the current actual values ​​of key parameters with historical performance trends includes the following steps: Align the actual values ​​of current key parameters with historical performance trends over time; standardize the actual values ​​of current key parameters and historical performance trends to eliminate differences in dimensions and units. Key features are extracted from historical performance trends to serve as a benchmark for comparison; Match the actual values ​​of current key parameters with key features of historical performance trends to identify performance differences across different dimensions; calculate the absolute difference between the actual values ​​of current key parameters and the predicted values ​​of historical performance trends to reflect the magnitude of the deviation; compare the absolute deviation with the fluctuation range of historical performance trends to calculate the relative deviation rate to reflect the severity of the deviation. Analyze the consistency between the current actual value change trend of key parameters and the historical performance trend to determine whether it deviates from the expected direction; based on multi-dimensional indicators such as absolute deviation and relative deviation rate, comprehensively score the degree of deviation of the current actual value of key parameters. Based on the comprehensive scoring results, the degree of deviation is divided into different levels.

Images