A sintering process optimization method, device and equipment of a resistance sheet and a storage medium

Abstract

This disclosure provides a method, apparatus, device, and storage medium for optimizing the sintering process of a resistor sheet. The method includes inputting a combination of parameters to be optimized for the sintering process into a preset prediction model; determining specific predicted values ​​for several target macroscopic performance parameters of the parameter combination to be optimized; obtaining a combination of sintering process parameters based on a preset optimization objective function and the specific predicted values; acquiring locally encrypted data; establishing a precise nonlinear mapping relationship between the sintering process parameters and the target macroscopic electrical performance parameters based on the locally encrypted data; performing a local fine-grained search on the combination of sintering process parameters based on the precise nonlinear mapping relationship; determining the optimal combination of sintering process parameters; and optimizing the sintering process of the resistor sheet based on the optimal combination of sintering process parameters. This achieves rapid and low-cost process optimization.

Description

Technical Field

[0001] This disclosure relates to the field of industrial artificial intelligence technology, and in particular to a method, apparatus, equipment and storage medium for optimizing the sintering process of a resistor sheet. Background Technology

[0002] Zinc oxide resistance elements (ZO) are polycrystalline electronic functional ceramics. Their microstructure, composed of multiple grains and grain boundaries, exhibits nonlinear current / voltage characteristics and excellent energy absorption capabilities, thus protecting electrical and electronic equipment from lightning strikes or surge voltage damage. The composition of ZO resistance elements is complex. Besides the main component ZnO, oxides such as Bi₂O₃, Sb₂O₃, and Co₂O₃ are added. These additives only form effective secondary phases during sintering, contributing to the nonlinearity of the ZO resistance element. The sintering temperature, holding time, heating rate, and cooling rate of the ZO resistance element sintering process all ultimately affect its macroscopic electrical properties. Among these, sintering temperature has the greatest impact on the performance of ZnO resistance elements. However, when multiple process variables such as sintering temperature, holding time, and heating / cooling rates change simultaneously, it is difficult to directly predict how the electrical properties of the ZnO resistance element will change. To obtain the desired macroscopic electrical properties of zinc oxide resistance elements, traditional sintering process optimization relies on numerous repeated experiments. This requires multiple trials and errors to adjust parameters such as sintering temperature, time, and heating / cooling rates, consuming significant manpower and resources, resulting in high costs and long cycles. Therefore, finding a way to reduce the number of experiments required for sintering process adjustments while rapidly and effectively obtaining the desired macroscopic electrical properties of ZnO resistance elements is of great practical significance. Summary of the Invention

[0003] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, equipment, and storage medium for optimizing the sintering process of resistor sheets.

[0004] This disclosure provides an optimized method for the sintering process of a resistor sheet, the method comprising: The combination of parameters to be optimized in the sintering process is input into a preset prediction model, and the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized are determined according to the prediction model. Based on the preset optimization objective function and the specific predicted value, a combination of sintering process parameters is obtained. Local encrypted data is obtained based on the combination of sintering process parameters. Based on the local encrypted data, a precise nonlinear mapping relationship between the sintering process parameters and the target macroscopic electrical performance parameters is established. Based on the precise nonlinear mapping relationship, a local fine search is performed on the combination of sintering process parameters to determine the optimal combination of sintering process parameters, and the sintering process of the resistor sheet is optimized according to the optimal combination of sintering process parameters.

[0005] The method provided in this disclosure involves inputting a combination of parameters to be optimized into a preset prediction model, and determining specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized based on the prediction model, including: The combination of parameters to be optimized is converted into a new format. The converted format is then input into a pre-defined prediction model, and forward propagation is performed to generate the original predicted values ​​of the target macroscopic performance parameters. The original predicted values ​​are backpropagated based on the prediction model to convert them into specific predicted values ​​for the target macroscopic performance parameters.

[0006] The method provided in this disclosure, wherein the prediction model includes: A training dataset is constructed based on the macroscopic electrical performance parameters of the resistor obtained from several experiments using a pre-set set of discrete process parameters. Match process parameter pairs that interact with each other from the process knowledge base in the training dataset, construct interaction term features based on the process parameter pairs, and combine the interaction term features with the original process parameter features extracted from the discrete process parameter sample set to form an enhanced feature set. A positive model is obtained by training the enhanced feature set and the preset training strategy. The training dataset is structurally reversed to obtain an inverted dataset, and an inverted model is trained based on the inverted dataset. The prediction outputs of the forward model and the reverse model are used as new training features, and a prediction model is trained and derived based on the new training features.

[0007] The method provided in this disclosure, based on a preset optimization objective function and the specific predicted value, obtains a combination of sintering process parameters, including: Analyze the combination of parameters to be optimized, determine the optimization objective, and match the corresponding preset optimization objective function from the preset function table according to the optimization objective; Substitute the specific predicted value into the preset optimization objective function to calculate the objective function value of the parameter combination to be optimized, and calculate the combination score of the combination to be optimized based on the objective function value. A new combination of parameters to be optimized is generated based on the combination of parameters to be optimized. The score of the new combination of parameters to be optimized is calculated. The combination of parameters to be optimized with the highest score is taken as the sintering process parameter combination.

[0008] The method provided in this disclosure, which obtains locally encrypted data based on the combination of sintering process parameters, and establishes a precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters based on the locally encrypted data, includes: Taking the sintering process parameter combination as the center point, based on the physical and engineering characteristics of each parameter in the sintering process parameters, a first domain range of the corresponding parameter is determined, and local encrypted data of the sintering process parameter combination is determined according to the first domain range; Based on the local encrypted data input, a preset relational mapping model is established, which determines the precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters.

[0009] The method provided in this disclosure performs a local fine-grained search on the sintering process parameter combination based on the precise nonlinear mapping relationship to determine the optimal sintering process parameter combination, and optimizes the sintering process of the resistor sheet according to the optimal sintering process parameter combination, including: Taking the sintering process parameter combination as the center, the second domain range of each parameter in the sintering process parameter combination is determined according to the process controllability accuracy; Using the precise nonlinear mapping relationship as a high-precision objective function, a local fine search is performed on the sintering process parameters within the second domain. The search results are sorted based on the high-precision objective function value, and the sintering process parameter combination with the optimal function value is selected to obtain the optimal sintering process parameter combination. The set values ​​of the optimal sintering process parameter combination are input into the tunnel kiln control system to execute the firing experiment; Monitor the actual process curve during the firing experiment, verify the fitting degree between the actual process curve and the set value of the optimal sintering process parameter combination, measure the electrical performance of the resistance sheet obtained from the firing experiment, verify the conformity with the expected target, solidify the process specification based on the verification results, and complete the optimization.

[0010] This disclosure also provides an apparatus for optimizing the sintering process of a resistor sheet, the apparatus comprising: The prediction module is used to input the combination of parameters to be optimized into a preset prediction model, and determine the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized based on the prediction model. A module is established to obtain a combination of sintering process parameters based on a preset optimization objective function and the specific predicted value, acquire local encrypted data based on the combination of sintering process parameters, and establish a precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters based on the local encrypted data. The optimization module is used to perform a local fine search on the combination of sintering process parameters based on the precise nonlinear mapping relationship, determine the optimal combination of sintering process parameters, and optimize the sintering process of the resistor sheet according to the optimal combination of sintering process parameters.

[0011] The apparatus provided in this disclosure, wherein the prediction module is specifically used for: The combination of parameters to be optimized is converted into a new format. The converted format is then input into a pre-defined prediction model, and forward propagation is performed to generate the original predicted values ​​of the target macroscopic performance parameters. The original predicted values ​​are backpropagated based on the prediction model to convert them into specific predicted values ​​for the target macroscopic performance parameters.

[0012] The apparatus provided in this disclosure, wherein the prediction model is specifically used for: A training dataset is constructed based on the macroscopic electrical performance parameters of the resistor obtained from several experiments using a pre-set set of discrete process parameters. Match process parameter pairs that interact with each other from the process knowledge base in the training dataset, construct interaction term features based on the process parameter pairs, and combine the interaction term features with the original process parameter features extracted from the discrete process parameter sample set to form an enhanced feature set. A positive model is obtained by training the enhanced feature set and the preset training strategy. The training dataset is structurally reversed to obtain an inverted dataset, and an inverted model is trained based on the inverted dataset. The prediction outputs of the forward model and the reverse model are used as new training features, and a prediction model is trained and derived based on the new training features.

[0013] The apparatus provided in this disclosure, wherein the establishment module is specifically used for: Analyze the combination of parameters to be optimized, determine the optimization objective, and match the corresponding preset optimization objective function from the preset function table according to the optimization objective; Substitute the specific predicted value into the preset optimization objective function to calculate the objective function value of the parameter combination to be optimized, and calculate the combination score of the combination to be optimized based on the objective function value. A new combination of parameters to be optimized is generated based on the combination of parameters to be optimized. The score of the new combination of parameters to be optimized is calculated. The combination of parameters to be optimized with the highest score is taken as the sintering process parameter combination.

[0014] The apparatus provided in this disclosure, wherein the establishment module is specifically used for: Taking the sintering process parameter combination as the center point, based on the physical and engineering characteristics of each parameter in the sintering process parameters, a first domain range of the corresponding parameter is determined, and local encrypted data of the sintering process parameter combination is determined according to the first domain range; Based on the local encrypted data input, a preset relational mapping model is established, which determines the precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters.

[0015] The optimization module in the apparatus provided in this disclosure is specifically used for: Taking the sintering process parameter combination as the center, the second domain range of each parameter in the sintering process parameter combination is determined according to the process controllability accuracy; Using the precise nonlinear mapping relationship as a high-precision objective function, a local fine search is performed on the sintering process parameters within the second domain. The search results are sorted based on the high-precision objective function value, and the sintering process parameter combination with the optimal function value is selected to obtain the optimal sintering process parameter combination. The set values ​​of the optimal sintering process parameter combination are input into the tunnel kiln control system to execute the firing experiment; Monitor the actual process curve during the firing experiment, verify the fitting degree between the actual process curve and the set value of the optimal sintering process parameter combination, measure the electrical performance of the resistance sheet obtained from the firing experiment, verify the conformity with the expected target, solidify the process specification based on the verification results, and complete the optimization.

[0016] This disclosure also provides an electronic device, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the sintering process optimization method for resistor sheets provided in this disclosure.

[0017] This disclosure also provides a computer-readable storage medium storing a computer program for executing the sintering process optimization method for resistors as provided in this disclosure.

[0018] The technical solution provided in this disclosure has the following advantages compared with the prior art: The sintering process optimization method for resistive wafers provided in this disclosure involves inputting the combination of sintering process parameters to be optimized into a prediction model, obtaining specific predicted values ​​of the target macroscopic performance parameters, and selecting sintering process parameter combinations based on the predicted values ​​through a preset optimization objective function. Local refinement is then performed around the sintering process parameter combinations to establish a precise nonlinear mapping relationship. This precise nonlinear mapping relationship is used to perform a local fine-grained search to determine the final optimal sintering process parameter combination, guiding the optimization of the sintering process. By establishing a quantitative mapping relationship between process parameters and electrical performance, the traditional trial-and-error method is replaced. This significantly reduces the number of physical experiments required to obtain the target performance, shortens the R&D cycle, reduces labor and material costs, and achieves rapid and accurate optimization of the sintering process. Attached Figure Description

[0019] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0020] Figure 1 A schematic flowchart illustrating the optimized sintering process of the resistor sheet provided in this embodiment of the present disclosure; Figure 2 A flowchart of the preset training strategy provided in the embodiments of this disclosure; Figure 3 A schematic diagram of the structure of the sintering process optimization device for the resistor sheet provided in the embodiments of this disclosure; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0021] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0022] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0023] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0024] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0025] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0026] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0027] To address the aforementioned problems, this disclosure provides an optimization method for the sintering process of resistor sheets. The method will be described below with reference to specific embodiments.

[0028] Figure 1 This is a flowchart illustrating a method for optimizing the sintering process of a resistor sheet according to an embodiment of the present disclosure. This method can be executed by a device for optimizing the sintering process of a resistor sheet, wherein the device can be implemented using software and / or hardware, and is generally integrated into an electronic device.

[0029] Example 1: This embodiment of the present disclosure provides a method for optimizing the sintering process of a resistor sheet, the method comprising: S101: Input the combination of parameters to be optimized in the sintering process into a preset prediction model, and determine the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized based on the prediction model. S102: Based on the preset optimization objective function and the specific predicted value, a combination of sintering process parameters is obtained; based on the combination of sintering process parameters, local encrypted data is obtained; and based on the local encrypted data, a precise nonlinear mapping relationship between the sintering process parameters and the target macroscopic electrical performance parameters is established. S103: Based on the precise nonlinear mapping relationship, perform a local fine search on the combination of sintering process parameters to determine the optimal combination of sintering process parameters, and optimize the sintering process of the resistor sheet according to the optimal combination of sintering process parameters.

[0030] In this embodiment, the combination of parameters to be optimized refers to a set of input variables or features that need to be optimized. These include sintering temperature, holding time, heating rate, and cooling rate.

[0031] In this embodiment, the input to the prediction model is a vector of combined parameters to be optimized after format conversion; its output is the initial prediction of the corresponding target macroscopic performance.

[0032] In this embodiment, the specific predicted value is the final prediction result obtained after backpropagation, which has the same units and dimensions as the target macroscopic performance parameter. For example, the varistor voltage = 737.5V and the nonlinear coefficient = 45 are obtained from -1.33 through a similar inverse transformation.

[0033] In this embodiment, the combination of parameters to be optimized is converted into a format and then input into the prediction model to perform forward propagation to obtain the original predicted value. It is then converted into the specific predicted value of the target macroscopic performance parameter through backpropagation.

[0034] In this embodiment, the sintering process parameter combination is the one with the highest score selected from all evaluated combinations of parameters to be optimized after the entire search process is completed. This is the final output of the optimization process and will be used to guide actual production or experiments.

[0035] In this embodiment, a preset optimization objective function is matched according to the optimization target, the specific predicted value is substituted into the calculation objective function value, a new combination to be optimized is generated through iteration and the score is calculated, and the combination with the highest score is selected as the sintering process parameter combination.

[0036] In this embodiment, the locally encrypted data is a new process parameter combination dataset generated by systematically and densely sampling within a multidimensional parameter subspace composed of the first domain range of each parameter.

[0037] In this embodiment, the precise nonlinear mapping relationship is revealed and expressed by the relational mapping model, which is a mathematical model or response surface that shows how small changes in process parameters nonlinearly and precisely affect the final performance within a first domain.

[0038] In this embodiment, the first neighborhood range is determined based on the combination of sintering process parameters and the physical and engineering characteristics of the parameters, and local encrypted data is generated. The encrypted data is then input into the relational mapping model to determine the precise nonlinear mapping relationship between the process parameters and the macroscopic electrical properties.

[0039] In this embodiment, the local fine-grained search involves using denser sampling or optimization algorithms within the second neighborhood to find the point that optimizes the objective function value. For example, within the range of [1162, 1168]℃, [2.18, 2.22]h, and [20.3, 20.7]%, a grid search is performed with step sizes of 0.5℃, 0.01h, and 0.1%, or an optimization algorithm is used to find the optimal combination.

[0040] In this embodiment, the optimal sintering process parameter combination is the set of parameters with the highest high-precision objective function value found through a local fine-grained search within the second domain. It represents the theoretically achievable production setpoints with the best performance prediction, taking into full account the actual production control precision. For example, after searching, the combination [sintering temperature = 1165.5℃, holding time = 2.20h, oxygen content = 20.5%] was found to have the highest predicted performance score, and this combination was determined as the optimal sintering process parameter combination.

[0041] In this embodiment, the second neighborhood range is determined based on the controllable precision of the process, with the combination of sintering process parameters as the center. The precise nonlinear mapping relationship is used as the high-precision objective function. A local fine search is performed in the second neighborhood to select the sintering process parameter combination with the optimal function value. The optimal sintering process parameter combination is obtained, and a sintering experiment is performed to monitor the process curve and electrical performance. The post-curing process procedure is then verified.

[0042] The working principle and beneficial effects of this embodiment are as follows: Input the combination of sintering process parameters to be optimized into the prediction model to obtain specific predicted values ​​of the target macroscopic performance parameters. Based on the predicted values, select sintering process parameter combinations through a preset optimization objective function. Local refinement is performed around the sintering process parameter combinations to establish a precise nonlinear mapping relationship. A local fine-grained search is performed using this precise nonlinear mapping relationship to determine the final optimal sintering process parameter combination, guiding the optimization of the sintering process. A quantitative mapping relationship between process parameters and electrical performance is established through a machine learning model, replacing the traditional trial-and-error method. This significantly reduces the number of physical experiments required to obtain the target performance, shortens the R&D cycle, reduces manpower and material costs, and achieves rapid and accurate optimization of the sintering process.

[0043] Example 2: The method provided in this embodiment of the present disclosure inputs the combination of parameters to be optimized into a preset prediction model, and determines the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized according to the prediction model, including: The combination of parameters to be optimized is converted into a new format. The converted format is then input into a pre-defined prediction model, and forward propagation is performed to generate the original predicted values ​​of the target macroscopic performance parameters. The original predicted values ​​are backpropagated based on the prediction model to convert them into specific predicted values ​​for the target macroscopic performance parameters.

[0044] In this embodiment, format conversion involves standardizing, normalizing, or encoding the original parameters into a standard numerical format that the prediction model can directly process. For example, the parameters [1150, 2.5, 150, 21, 0.95] are normalized using the mean and standard deviation determined during model training, and converted into a numerical vector similar to [0.87, -0.12, 1.05, 0.33, 0.18].

[0045] In this embodiment, forward propagation is the computation process in which data is passed from the model input layer to the output layer layer by layer. For example, the normalized vector input network is calculated through multiple hidden layers in sequence to finally obtain the original value of the output layer [2.75, -1.33].

[0046] In this embodiment, the target macroscopic performance parameters are the key performance indicators that ultimately need to be predicted and optimized. They are usually quantities with clear physical meaning and units, including voltage gradient, leakage current and nonlinear coefficient.

[0047] In this embodiment, the raw predicted values ​​are the values ​​directly output after the model's forward propagation, without final transformation. These values ​​are typically standardized during model training and have a fixed scaling and offset relationship with the true performance parameters. For example, the model output [2.75, -1.33] is not directly equal to the varistor voltage and nonlinear coefficient, but rather a transformation of them.

[0048] In this embodiment, backpropagation specifically refers to the post-processing transformation, specifically gradient backpropagation during non-training phases. It is an inverse transformation, the opposite of the standardization / normalization process in forward propagation. Its purpose is to de-standardize / de-normalize the model's original predictions, restoring them to concrete predictions with actual physical units and magnitudes.

[0049] The working principle and beneficial effects of this embodiment are as follows: After the combination of parameters to be optimized is converted into a format, it is input into the prediction model to perform forward propagation to obtain the original predicted value, and then converted into the specific predicted value of the target macroscopic performance parameter through back propagation. By combining forward propagation and back propagation, the original predicted value is converted into a macroscopic performance parameter that can be directly measured, thereby improving the prediction accuracy and decision applicability, and improving the efficiency of parameter optimization.

[0050] Example 3: The method provided in this embodiment of the present disclosure, wherein the prediction model includes: A training dataset is constructed based on the macroscopic electrical performance parameters of the resistor obtained from several experiments using a pre-set set of discrete process parameters. Match process parameter pairs that interact with each other from the process knowledge base in the training dataset, construct interaction term features based on the process parameter pairs, and combine the interaction term features with the original process parameter features extracted from the discrete process parameter sample set to form an enhanced feature set. A positive model is obtained by training the enhanced feature set and the preset training strategy. The training dataset is structurally reversed to obtain an inverted dataset, and an inverted model is trained based on the inverted dataset. The prediction outputs of the forward model and the reverse model are used as new training features, and a prediction model is trained and derived based on the new training features.

[0051] In this embodiment, the experimental design employs a uniform design, ensuring a uniform spatial distribution of sample points. The sample size can be freely set (e.g., 200 samples). Samples are generated by minimizing bias, making it suitable for high-dimensional spaces with no level limitation. It can handle continuous or discrete variables, avoiding model bias caused by local clustering. It can cover four variables: sintering temperature (900-1200ºC), holding time (0-10h), heating rate (0.5-8ºC / min), and cooling rate (0.5-8ºC / min). Zinc oxide resistance elements are sintered using an industrial tunnel kiln with adjusted sintering conditions. After subsequent processing, the macroscopic electrical properties of the resistance elements are measured. The DC varistor voltage and leakage current are measured using a DC varistor voltage parameter measuring instrument. The voltage gradient of the resistance element is obtained by dividing the DC varistor voltage by the height of the resistance element. The nonlinear coefficient of the resistance element is then measured using a nonlinear coefficient measuring instrument. 100-300 experimental data points are collected.

[0052] In this embodiment, the discrete process parameter sample set is a list of pre-designed process conditions in different combinations.

[0053] In this embodiment, the macroscopic electrical performance parameters are key performance indicators obtained by testing the resistor sheet fired according to the above sample parameters through physical experiments.

[0054] In this embodiment, the training dataset is a set of data consisting of combinations of process parameters and their corresponding measured electrical performance parameters.

[0055] In this embodiment, the process knowledge base is a rule base that stores expert knowledge or historical research conclusions in the field, indicating which process parameters may have significant interactions. For example, sintering temperature and holding time have a synergistic effect on grain growth, and oxygen content and heating / cooling rates have a combined effect on porosity.

[0056] In this embodiment, the process parameter pairs are combinations of parameters that may interact, identified from the process parameters of the training dataset based on a knowledge base.

[0057] In this embodiment, the interaction feature is a new feature constructed to quantify the above-mentioned interaction, which is usually the product of the pairing parameters.

[0058] In this embodiment, the enhanced feature set is a new feature set formed by merging the original process parameter features with the newly constructed interaction term features. This ensures that the model input not only contains independent parameters but also information about the coupling relationships between parameters.

[0059] In this embodiment, such as Figure 2As shown, the preset training strategy uses 10-fold cross-validation to calculate the mean square error (MSE), root mean square error (RMSE), and coefficient of determination (R²). For each macroscopic electrical performance parameter within the combination of process parameters to be optimized, the strategies are evaluated individually, calculating the voltage gradient, leakage current, and nonlinearity coefficient separately to avoid random bias from a single partition. A scatter plot of the predicted results versus the actual values ​​is visualized to observe for systematic biases (such as models generally overestimating leakage current). With the goal of minimizing the required voltage gradient, leakage current, and maximizing the nonlinearity coefficient, the cross-validation metrics of different models are compared (e.g., R² > 0.8 is considered acceptable). Finally, the algorithm with the best generalization ability on the electrical performance parameters is selected, resulting in the preset training strategy.

[0060] In this embodiment, the input to the forward model is the enhanced feature set, and the output is the predicted macroscopic electrical performance parameters.

[0061] In this embodiment, structural inversion involves swapping the roles of the input and output in the original training dataset.

[0062] In this embodiment, the reverse dataset is a dataset composed of macroscopic electrical performance parameters as input and process parameters as output.

[0063] In this embodiment, the inverse model takes a reverse dataset as input and outputs a predicted combination of process parameters. Its function is to deduce the required process given a target performance.

[0064] In this embodiment, the new training features are generated by first predicting the performance of a given combination of process parameters using a forward model, and then using a backward model to infer a set of process parameters based on this predicted performance. The original process parameters, the performance predicted by the forward model, and the process parameters inferred by the backward model are combined into a new set of more informative features.

[0065] In this embodiment, the final prediction model is an integrated prediction model trained using the newly constructed training features as input and the actual macroscopic electrical performance as output.

[0066] The working principle and beneficial effects of this embodiment are as follows: A training set is constructed based on a preset discrete process parameter sample set and the macroscopic performance parameters of the resistor sheet after experimentation. Interaction parameter pairs are matched from the process knowledge base to construct interaction term features and form an enhanced feature set. A forward model and a backward model are trained, and the prediction outputs of the two models are merged as new training features to obtain the final prediction model. By integrating interaction features and forward and backward models, the nonlinear relationship between process and performance is accurately captured, improving the prediction accuracy and generalization ability of complex process parameters and optimizing product performance design.

[0067] Example 4: The method provided in this embodiment of the present disclosure obtains a combination of sintering process parameters based on a preset optimization objective function and the specific predicted value, including: Analyze the combination of parameters to be optimized, determine the optimization objective, and match the corresponding preset optimization objective function from the preset function table according to the optimization objective; Substitute the specific predicted value into the preset optimization objective function to calculate the objective function value of the parameter combination to be optimized, and calculate the combination score of the combination to be optimized based on the objective function value. A new combination of parameters to be optimized is generated based on the combination of parameters to be optimized. The score of the new combination of parameters to be optimized is calculated. The combination of parameters to be optimized with the highest score is taken as the sintering process parameter combination.

[0068] In this embodiment, the optimization objective is the desired final performance state. It indicates the direction of optimization.

[0069] In this embodiment, the preset function table is a predefined function library or rule library that stores standard calculation formulas or algorithms for different optimization objectives.

[0070] In this embodiment, the preset optimization objective function is a specific mathematical function matched and instantiated from the function table according to the current optimization objective. It is a formula that maps specific predicted values ​​to a scalar score; a higher score indicates that the set of parameters is better.

[0071] In this embodiment, the objective function value is the original value calculated directly by substituting the specific predicted values ​​of a set of parameters into a preset optimization objective function. It directly reflects the degree to which the set of parameters satisfies the objective function.

[0072] In this embodiment, the combined score is typically the objective function value itself, or a value obtained after standardization and normalization, with the aim of ensuring fair comparison of scores across different rounds or search strategies. is a scalar used to rank all candidate parameter combinations.

[0073] The working principle and beneficial effects of this embodiment are as follows: based on the optimization target, a preset optimization objective function is matched, the specific predicted value is substituted into the calculation objective function value, a new combination to be optimized is generated through iteration and the score is calculated, the combination with the highest score is selected as the sintering process parameter combination, the objective function is automatically matched and iteratively optimized, the optimal process parameter combination is quickly locked, the efficiency and accuracy of process design are improved, and the product performance is ensured to meet the standards.

[0074] Example 5: The method provided in this embodiment of the present disclosure, which obtains locally encrypted data based on the combination of sintering process parameters, and establishes a precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters based on the locally encrypted data, includes: Taking the sintering process parameter combination as the center point, based on the physical and engineering characteristics of each parameter in the sintering process parameters, a first domain range of the corresponding parameter is determined, and local encrypted data of the sintering process parameter combination is determined according to the first domain range; Based on the local encrypted data input, a preset relational mapping model is established, which determines the precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters.

[0075] In this embodiment, physical properties and engineering properties refer to the inherent laws and practical constraints that determine the allowable fluctuation range of each process parameter. Physical properties are fundamental limitations determined by materials science and chemical reaction kinetics. For example, below a certain temperature, the reaction cannot proceed completely, while above a certain temperature, the material will decompose or melt. Engineering properties are practical constraints determined by production equipment, control precision, and process stability. For example, the temperature control precision of the kiln, the adjustment resolution of the gas flow meter, and the inherent batch-to-batch fluctuations.

[0076] In this embodiment, the first domain range is a relatively loose security exploration range defined for each individual parameter in the optimal parameter combination based on the above characteristics. This range is used for preliminary local data encryption and is intended to cover the entire area where the parameters may fluctuate reasonably.

[0077] In this embodiment, the relational mapping model takes as input combinations of process parameters from locally encrypted data. The output is an accurate prediction of the target macroscopic electrical performance parameters for these parameter combinations. For example, Gaussian process regression or high-order polynomial regression can be used as the relational mapping model. By inputting 50 sets of locally encrypted data, the model learns the complex relationships derived from this input.

[0078] The working principle and beneficial effects of this embodiment are as follows: Centering on the combination of sintering process parameters, a first neighborhood range is determined based on the physical and engineering characteristics of the parameters, and locally encrypted data is generated. This encrypted data is then input into a relational mapping model to determine the precise nonlinear mapping relationship between process parameters and macroscopic electrical performance. Establishing a precise process-performance mapping relationship within the optimal parameter neighborhood guides high-precision process fine-tuning, improves product performance stability and consistency, and avoids blind optimization.

[0079] Example 6: The method provided in this embodiment of the present disclosure performs a local fine search on the combination of sintering process parameters based on the precise nonlinear mapping relationship to determine the optimal combination of sintering process parameters, and optimizes the sintering process of the resistor sheet according to the optimal combination of sintering process parameters, including: Taking the sintering process parameter combination as the center, the second domain range of each parameter in the sintering process parameter combination is determined according to the process controllability accuracy; Using the precise nonlinear mapping relationship as a high-precision objective function, a local fine search is performed on the sintering process parameters within the second domain. The search results are sorted based on the high-precision objective function value, and the sintering process parameter combination with the optimal function value is selected to obtain the optimal sintering process parameter combination. The set values ​​of the optimal sintering process parameter combination are input into the tunnel kiln control system to execute the firing experiment; Monitor the actual process curve during the firing experiment, verify the fitting degree between the actual process curve and the set value of the optimal sintering process parameter combination, measure the electrical performance of the resistance sheet obtained from the firing experiment, verify the conformity with the expected target, solidify the process specification based on the verification results, and complete the optimization.

[0080] In this embodiment, process controllable precision refers to the parameter control precision that the production equipment (such as a tunnel kiln) and control system can actually achieve and maintain during stable operation. For example, the kiln temperature control system can stably control the temperature within ±3℃ of the set value during long-term operation; the gas mass flow meter can control the oxygen content within ±0.5% of the set value.

[0081] In this embodiment, the second domain range is a narrower, more practical fine-tuning search interval determined around the central optimum based on process controllable precision. This range represents the parameter fluctuation boundary that can be stably achieved in production. For example, if the center point is 1165℃ and the controllable precision is ±3℃, then the second domain range might be defined as [1162℃, 1168℃]. This range is much smaller than the first domain range ([1150, 1180]℃).

[0082] In this embodiment, the high-precision objective function is the precise nonlinear mapping relationship obtained in the previous step. At this stage, it is directly used as a scoring function to evaluate the quality of the parameters. Because it has extremely high accuracy within a local region, it can provide sensitive and accurate evaluation of minute parameter changes.

[0083] In this embodiment, the actual process curve is the curve showing the change of parameters such as temperature and gas flow rate over time, as actually recorded by the sensors, when the optimal combination of sintering process parameters is input into the kiln control system for actual firing experiments. For example, the set temperature is 1165.5℃, but the actual curve may show fluctuations between 1163℃ and 1167℃.

[0084] In this embodiment, the goodness of fit is the degree of agreement between the actual process curve and the set value. It is used to evaluate the control stability of the production equipment. For example, the mean and standard deviation of the actual temperature curve are calculated and compared with the set value of 1165.5℃ and the controllable accuracy of ±3℃.

[0085] In this embodiment, the electrical performance is obtained by testing the actual macroscopic electrical performance parameters of the experimentally fired resistor sheet.

[0086] In this embodiment, the degree of conformity to the expected target is evaluated by comparing the measured electrical performance with the optimization target set at the beginning of the optimization process to assess whether the target has been met.

[0087] In this embodiment, the verification result is a comprehensive evaluation conclusion of the goodness of fit and conformity. It determines whether to accept the set of parameters. For example, if the verification result shows that the process curve control is stable, the goodness of fit is >95%, and the electrical performance of the product meets all expected targets, the verification is passed.

[0088] In this embodiment, the curing process specification is based on the verification results, and the optimal combination of sintering process parameters and its allowed second range are formally written into the production operation manual or formula library as the standard production process for this model of product.

[0089] The working principle and beneficial effects of this embodiment are as follows: Centering on the combination of sintering process parameters, a second neighborhood range is determined based on the controllable precision of the process. Using a precise nonlinear mapping relationship as a high-precision objective function, a local fine-tuning search is performed within the second neighborhood to select the sintering process parameter combination with the optimal function value. This yields the optimal sintering process parameter combination. A sintering experiment is then performed, and the process curves and electrical properties are monitored. The process specifications are then verified and solidified. This allows for precise fine-tuning and verification of process parameters within a controllable precision range, ensuring that actual production strictly adheres to the optimal settings, improving product performance compliance and process stability, and solidifying the optimal process.

[0090] To achieve the above embodiments, this disclosure also proposes an optimized apparatus for the sintering process of resistor sheets.

[0091] Figure 3 This is a schematic diagram of the structure of the resistive sheet sintering process optimization device provided in an embodiment of this disclosure. This device 200 can be implemented by software and / or hardware, and is generally integrated into an electronic device. For example... Figure 3 As shown, the device 200 includes: a prediction module 201, a building module 202, and an optimization module 203, wherein, The prediction module 201 is used to input the combination of parameters to be optimized into a preset prediction model, and determine the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized according to the prediction model. Module 202 is used to obtain a combination of sintering process parameters based on a preset optimization objective function and the specific predicted value, obtain locally encrypted data based on the combination of sintering process parameters, and establish a precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters based on the locally encrypted data. The optimization module 203 is used to perform a local fine search on the combination of sintering process parameters based on the precise nonlinear mapping relationship, determine the optimal combination of sintering process parameters, and optimize the sintering process of the resistor sheet according to the optimal combination of sintering process parameters.

[0092] The apparatus provided in this disclosure, wherein the prediction module 201 is specifically used for: The combination of parameters to be optimized is converted into a new format. The converted format is then input into a pre-defined prediction model, and forward propagation is performed to generate the original predicted values ​​of the target macroscopic performance parameters. The original predicted values ​​are backpropagated based on the prediction model to convert them into specific predicted values ​​for the target macroscopic performance parameters.

[0093] The apparatus provided in this disclosure embodiment, wherein the prediction model 201 is specifically used for: A training dataset is constructed based on the macroscopic electrical performance parameters of the resistor obtained from several experiments using a pre-set set of discrete process parameters. Match process parameter pairs that interact with each other from the process knowledge base in the training dataset, construct interaction term features based on the process parameter pairs, and combine the interaction term features with the original process parameter features extracted from the discrete process parameter sample set to form an enhanced feature set. A positive model is obtained by training the enhanced feature set and the preset training strategy. The training dataset is structurally reversed to obtain an inverted dataset, and an inverted model is trained based on the inverted dataset. The prediction outputs of the forward model and the reverse model are used as new training features, and a prediction model is trained and derived based on the new training features.

[0094] The apparatus provided in this disclosure embodiment, wherein the establishment module 202 is specifically used for: Analyze the combination of parameters to be optimized, determine the optimization objective, and match the corresponding preset optimization objective function from the preset function table according to the optimization objective; Substitute the specific predicted value into the preset optimization objective function to calculate the objective function value of the parameter combination to be optimized, and calculate the combination score of the combination to be optimized based on the objective function value. A new combination of parameters to be optimized is generated based on the combination of parameters to be optimized. The score of the new combination of parameters to be optimized is calculated. The combination of parameters to be optimized with the highest score is taken as the sintering process parameter combination.

[0095] The apparatus provided in this disclosure embodiment, wherein the establishment module 202 is specifically used for: Taking the sintering process parameter combination as the center point, based on the physical and engineering characteristics of each parameter in the sintering process parameters, a first domain range of the corresponding parameter is determined, and local encrypted data of the sintering process parameter combination is determined according to the first domain range; Based on the local encrypted data input, a preset relational mapping model is established, which determines the precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters.

[0096] The optimization module 203 in the apparatus provided in this embodiment is specifically used for: Taking the sintering process parameter combination as the center, the second domain range of each parameter in the sintering process parameter combination is determined according to the process controllability accuracy; Using the precise nonlinear mapping relationship as a high-precision objective function, a local fine search is performed on the sintering process parameters within the second domain. The search results are sorted based on the high-precision objective function value, and the sintering process parameter combination with the optimal function value is selected to obtain the optimal sintering process parameter combination. The set values ​​of the optimal sintering process parameter combination are input into the tunnel kiln control system to execute the firing experiment; Monitor the actual process curve during the firing experiment, verify the fitting degree between the actual process curve and the set value of the optimal sintering process parameter combination, measure the electrical performance of the resistance sheet obtained from the firing experiment, verify the conformity with the expected target, solidify the process specification based on the verification results, and complete the optimization.

[0097] The resistive sheet sintering process optimization apparatus provided in this disclosure can execute the resistive sheet sintering process optimization method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects of the method.

[0098] To implement the above embodiments, this disclosure also proposes a computer program product, including a computer program / instructions, which, when executed by a processor, implements the sintering process optimization method for the resistor sheet in the above embodiments.

[0099] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure.

[0100] The following is a detailed reference. Figure 4The diagram illustrates a structural schematic suitable for implementing the electronic device 300 in the embodiments of this disclosure. The electronic device 300 in the embodiments of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0101] like Figure 4 As shown, the electronic device 300 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a memory 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processor 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0102] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0103] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from memory 308, or installed from ROM 302. When the computer program is executed by processor 301, it performs the functions defined in the method for optimizing the sintering process of resistor sheets according to embodiments of this disclosure.

[0104] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0105] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0106] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0107] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the aforementioned method for optimizing the sintering process of the resistor sheet.

[0108] Electronic devices can be programmed with computer program code in one or more programming languages ​​or combinations thereof to perform the operations of this disclosure. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0109] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0110] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0111] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0112] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0113] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0114] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0115] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. An optimized method for the sintering process of a resistor sheet, characterized in that, include: The combination of parameters to be optimized in the sintering process is input into a preset prediction model, and the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized are determined according to the prediction model. Based on the preset optimization objective function and the specific predicted value, a combination of sintering process parameters is obtained. Local encrypted data is obtained based on the combination of sintering process parameters. Based on the local encrypted data, a precise nonlinear mapping relationship between the sintering process parameters and the target macroscopic electrical performance parameters is established. Based on the precise nonlinear mapping relationship, a local fine search is performed on the combination of sintering process parameters to determine the optimal combination of sintering process parameters, and the sintering process of the resistor sheet is optimized according to the optimal combination of sintering process parameters.

2. The method according to claim 1, characterized in that, The combination of parameters to be optimized is input into a preset prediction model, and the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized are determined according to the prediction model, including: The combination of parameters to be optimized is converted into a new format. The converted format is then input into a pre-defined prediction model, and forward propagation is performed to generate the original predicted values ​​of the target macroscopic performance parameters. The original predicted values ​​are backpropagated based on the prediction model to convert them into specific predicted values ​​for the target macroscopic performance parameters.

3. The method according to claim 2, characterized in that, The prediction model includes: A training dataset is constructed based on the macroscopic electrical performance parameters of the resistor obtained from several experiments using a pre-set set of discrete process parameters. Match process parameter pairs that interact with each other from the process knowledge base in the training dataset, construct interaction term features based on the process parameter pairs, and combine the interaction term features with the original process parameter features extracted from the discrete process parameter sample set to form an enhanced feature set. A positive model is obtained by training the enhanced feature set and the preset training strategy. The training dataset is structurally reversed to obtain an inverted dataset, and an inverted model is trained based on the inverted dataset. The prediction outputs of the forward model and the reverse model are used as new training features, and a prediction model is trained and derived based on the new training features.

4. The method according to claim 1, characterized in that, Based on the preset optimization objective function and the specific predicted values, a combination of sintering process parameters is obtained, including: Analyze the combination of parameters to be optimized, determine the optimization objective, and match the corresponding preset optimization objective function from the preset function table according to the optimization objective; Substitute the specific predicted value into the preset optimization objective function to calculate the objective function value of the parameter combination to be optimized, and calculate the combination score of the combination to be optimized based on the objective function value. A new combination of parameters to be optimized is generated based on the combination of parameters to be optimized. The score of the new combination of parameters to be optimized is calculated. The combination of parameters to be optimized with the highest score is taken as the sintering process parameter combination.

5. The method according to claim 1, characterized in that, Based on the sintering process parameter combination, locally encrypted data is obtained. Based on this locally encrypted data, a precise nonlinear mapping relationship is established between the sintering process parameters and the target macroscopic electrical performance parameters, including: Taking the sintering process parameter combination as the center point, based on the physical and engineering characteristics of each parameter in the sintering process parameters, a first domain range of the corresponding parameter is determined, and local encrypted data of the sintering process parameter combination is determined according to the first domain range; Based on the local encrypted data input, a preset relational mapping model is established, which determines the precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters.

6. The method according to claim 1, characterized in that, Based on the precise nonlinear mapping relationship, a local fine search is performed on the combination of sintering process parameters to determine the optimal combination of sintering process parameters. The sintering process of the resistor sheet is then optimized based on this optimal combination of sintering process parameters, including: Taking the sintering process parameter combination as the center, the second domain range of each parameter in the sintering process parameter combination is determined according to the process controllability accuracy; Using the precise nonlinear mapping relationship as a high-precision objective function, a local fine search is performed on the sintering process parameters within the second domain. The search results are sorted based on the high-precision objective function value, and the sintering process parameter combination with the optimal function value is selected to obtain the optimal sintering process parameter combination. The set values ​​of the optimal sintering process parameter combination are input into the tunnel kiln control system to execute the firing experiment; Monitor the actual process curve during the firing experiment, verify the fitting degree between the actual process curve and the set value of the optimal sintering process parameter combination, measure the electrical performance of the resistance sheet obtained from the firing experiment, verify the conformity with the expected target, solidify the process specification based on the verification results, and complete the optimization.

7. A device for optimizing the sintering process of a resistor sheet, characterized in that, The device includes: The prediction module is used to input the combination of parameters to be optimized into a preset prediction model, and determine the specific predicted values ​​of several target macroscopic performance parameters of the combination of parameters to be optimized based on the prediction model. A module is established to obtain a combination of sintering process parameters based on a preset optimization objective function and the specific predicted value, acquire local encrypted data based on the combination of sintering process parameters, and establish a precise nonlinear mapping relationship between sintering process parameters and target macroscopic electrical performance parameters based on the local encrypted data. The optimization module is used to perform a local fine search on the combination of sintering process parameters based on the precise nonlinear mapping relationship, determine the optimal combination of sintering process parameters, and optimize the sintering process of the resistor sheet according to the optimal combination of sintering process parameters.

8. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the sintering process optimization method for the resistor sheet as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, It stores a computer program / instruction thereon, which, when executed by a processor, implements the steps of the method described in any one of claims 1-6.

10. A computer program product, comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 6.

Images