Process modification method and system based on product evaluation

Abstract

The application relates to the field of material processing and manufacturing technology, and discloses a process correction method and system based on product evaluation, which comprises the following steps: collecting and preprocessing raw material characteristics, production process and product performance data, cleaning, standardizing and filling in missing values; constructing an adaptability evaluation model based on an information entropy minimization method, and optimizing process parameters; optimizing the process parameters by using a group intelligence optimization algorithm; and performing real-time adaptability evaluation on new raw materials by using a support vector machine model; and dynamically adjusting production process parameters according to the evaluation results; the system comprises a data collection module, a data processing module, an evaluation module, an optimization module, a real-time evaluation module and a dynamic adjustment module. The real-time adaptability evaluation technical scheme based on the support vector machine is adopted, the dynamic adaptation of raw material characteristics and production process is realized, the flexibility and accuracy of production are effectively improved, and the accuracy of production raw material adaptation is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of materials processing and manufacturing technology, specifically to a process modification method and system based on product evaluation. Background Technology

[0002] With the rapid development of the manufacturing industry, the precision control and automation of the production process have become important indicators for evaluating the quality of modern production systems. Especially in the field of high-precision manufacturing, the compatibility between raw materials and processes directly affects product quality and production efficiency.

[0003] In existing production processes, the suitability assessment of raw materials typically relies on manual experience or pre-defined rule systems. These methods generally determine whether raw materials are suitable for the current production process through manual testing, sampling, or laboratory analysis. Process parameters are also often fixed, set based on historical experience and standard operating procedures.

[0004] However, existing raw material compatibility assessment technologies, relying on manual judgment and static rules, suffer from inaccurate assessments of material property compatibility and delayed adjustments. When raw material properties change, timely adjustments are difficult, impacting production efficiency and product quality. Secondly, the lack of intelligent data-driven assessment makes it difficult for existing technologies to conduct in-depth analysis and decision-making in a big data environment. This makes it difficult to rely on real-time data support for production process optimization, easily leading to resource waste and production instability. Therefore, this invention provides a process correction method and system based on product evaluation to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a process correction method and system based on product evaluation, which solves the problem that in existing production processes, process parameter adjustments mainly rely on manual experience and lack a systematic evaluation and optimization mechanism based on product quality feedback.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a process modification method based on product evaluation, comprising the following steps:

[0007] Collect and preprocess raw material characteristic data, production process parameters, and product performance data, and clean, standardize, and fill in missing values ​​for the processed raw material data;

[0008] Based on the processed raw material data, a raw material and production process compatibility assessment model is constructed using the information entropy minimization method, and process parameters are optimized.

[0009] Based on the evaluation results of the raw material and production process compatibility assessment model, the process parameters are optimized using a swarm intelligence optimization algorithm.

[0010] The support vector machine model is used to evaluate the adaptability of new raw material information in real time and determine whether it meets the requirements of the current production process.

[0011] Based on real-time evaluation results, production process parameters are dynamically adjusted to ensure the best match between the process and the characteristics of raw materials.

[0012] Preferably, the step of collecting and preprocessing raw material characteristic data, production process parameters, and product performance data includes:

[0013] The system collects data on the physicochemical properties of raw materials and relevant parameters of the production process in real time during the production process, and inputs the collected data into the data processing system.

[0014] The collected raw material data is denoised by using a filtering algorithm to remove outliers.

[0015] Missing values ​​are imputed using either mean imputation or nearest neighbor imputation.

[0016] Standardize all data to ensure that all feature values ​​are within the same range.

[0017] Preferably, the step of constructing a raw material and production process compatibility assessment model and optimizing process parameters using the information entropy minimization method includes:

[0018] Define the compatibility evaluation index between raw materials and production processes, and use the information entropy method to quantify and analyze multiple process parameters;

[0019] Based on the principle of minimizing information entropy, the optimal combination of raw materials and process parameters is selected to reduce information uncertainty and optimize the production process.

[0020] The process parameters are optimized using computer programs, and the optimal parameter combination is achieved through iterative algorithms.

[0021] Preferably, the step of optimizing process parameters using a swarm intelligence optimization algorithm includes:

[0022] Based on the evaluation results of the raw material and production process compatibility assessment model, a multi-objective optimization problem is constructed, with objectives including minimizing scrap rate, maximizing product quality, and improving production efficiency.

[0023] The particle swarm optimization algorithm is used to solve multiple objective functions and obtain a set of optimal process parameters;

[0024] The production process is adjusted based on the optimization results to ensure that all target constraints are met.

[0025] Preferably, the step of using a support vector machine model to perform real-time adaptability evaluation of new raw material information includes:

[0026] The new raw material characteristic data is input into the trained support vector machine model for real-time classification and judgment to determine whether the raw material is suitable for the current production process.

[0027] The support vector machine model classifies new raw materials as positive or negative based on the sample data in the training set and outputs the fit result.

[0028] If the real-time evaluation results indicate that the raw materials are not suitable for the current process, the system will automatically prompt and adjust the process parameters.

[0029] Preferably, the step of dynamically adjusting the production process parameters based on real-time evaluation results includes:

[0030] Based on the real-time adaptability assessment results, the production process parameters are adjusted in conjunction with the optimization objectives;

[0031] During the adjustment process, the process parameter values ​​are dynamically adjusted to ensure optimal matching with the characteristics of raw materials and changes in the production environment.

[0032] The adjusted process parameters are fed back to the production line in real time for actual production operations, and the effects are tracked and feedback is provided.

[0033] Preferably, the formula for quantifying multiple process parameters using the information entropy method is as follows:

[0034] ;

[0035] in, It is the entropy value; The probability distribution for each process parameter; For the first A specific value for a process parameter; This represents the total number of process parameters.

[0036] Preferably, the objective function for multi-objective optimization using the swarm intelligence optimization algorithm is:

[0037] ;

[0038] in, This refers to the quantity of scrap. Total production quantity; The value is the objective function value.

[0039] Preferably, the function for real-time adaptation evaluation of new raw material information using the support vector machine model is:

[0040] ;

[0041] in, The hyperplane normal vector; For penalty parameters; These are slack variables; This represents the total number of training data.

[0042] It also provides a process correction system based on product evaluation, including the following modules:

[0043] The data collection module is used to collect raw material characteristic data, production process parameters, and product performance data in real time.

[0044] The data processing module is used to clean, standardize, and fill in missing values ​​for the collected raw material data.

[0045] The evaluation module is used to construct a raw material and production process compatibility evaluation model based on the processed data using the information entropy minimization method.

[0046] The optimization module is used to optimize production process parameters based on the results of the adaptability evaluation model using swarm intelligence algorithms.

[0047] The real-time evaluation module is used to evaluate the suitability of new raw materials using a support vector machine model.

[0048] The dynamic adjustment module is used to adjust production process parameters based on real-time evaluation results to ensure the best match between the process and the characteristics of raw materials.

[0049] This invention provides a process modification method and system based on product evaluation. It has the following beneficial effects:

[0050] 1. This invention employs a real-time adaptability evaluation technique based on Support Vector Machine (SVM) to achieve dynamic adaptation between raw material characteristics and production processes. This technique allows for real-time monitoring of raw material adaptability during production and automated adjustments, effectively improving production flexibility and accuracy. Compared to existing technologies that rely on manual judgment or fixed rules, this invention avoids human error and significantly improves the automation level of the production process and the accuracy of raw material adaptation.

[0051] 2. This invention employs a dynamic process parameter adjustment mechanism, adjusting production parameters based on SVM output results to ensure the matching of raw materials and processes. This solution not only optimizes process control and reduces human intervention but also rapidly adapts to external changes based on real-time feedback, ensuring production stability. Compared to traditional production methods, existing technologies lack this adaptive adjustment function. This invention overcomes the bottleneck of fixed processes, enhancing process flexibility and system intelligence.

[0052] 3. This invention employs a multi-source data fusion technology, combining raw material characteristics, process equipment status, and environmental factors to further improve the accuracy and real-time performance of adaptability assessment. This comprehensive analysis provides more precise decision support for process adjustments, avoiding the limitations of relying on a single data source. Compared to existing single-factor-based assessment methods, this invention achieves more comprehensive production monitoring and adjustment, enhancing the level of intelligent decision-making in the process.

[0053] 4. This invention further enhances the predictability of production processes by introducing the simulation capabilities of digital twin models. This technology enables virtual simulation before actual operation, predicting in advance the impact of process adjustments on finished product quality, thereby reducing potential problems during production. Compared to existing technologies that lack predictive control, this invention solves the problem of lag in traditional process adjustments, making production management more efficient and precise. Attached Figure Description

[0054] Figure 1 This is a flowchart of the method steps of the present invention;

[0055] Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0056] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see the appendix Figure 1 This invention provides a process correction method based on product evaluation, comprising the following steps:

[0058] S1. Collect and preprocess raw material characteristic data, production process parameters and product performance data, and clean, standardize and fill missing values ​​in the processed raw material data.

[0059] S2. Based on the processed raw material data, a raw material and production process compatibility assessment model is constructed using the information entropy minimization method, and process parameters are optimized.

[0060] S3. Based on the evaluation results of the raw material and production process compatibility assessment model, the process parameters are optimized using a swarm intelligence optimization algorithm;

[0061] S4. Use a support vector machine model to evaluate the adaptability of new raw material information in real time and determine whether it meets the requirements of the current production process.

[0062] S5. Based on real-time evaluation results, dynamically adjust production process parameters to ensure the best match between process and raw material characteristics.

[0063] For step S1, in this embodiment, by collecting and preprocessing raw material characteristic data, production process parameters, and product performance data, it can be ensured that the subsequent evaluation and optimization process is based on sufficient and high-quality data. In actual operation, the collected data often comes from multiple different data sources, including but not limited to production equipment, sensors, quality control systems, and historical production data.

[0064] In this embodiment, the "raw material characteristic data" includes information on the physical and chemical properties of the raw materials, such as hardness, density, moisture content, and chemical composition; the "production process parameters" include temperature, pressure, production speed, and equipment configuration; and the "product performance data" covers the evaluation results of the final product quality, such as strength, toughness, and corrosion resistance. The accuracy of these data directly affects the effectiveness of subsequent models; therefore, it is necessary to ensure the accuracy, timeliness, and completeness of the data during collection.

[0065] In terms of data preprocessing, this embodiment employs the following methods to ensure data quality and usability:

[0066] Data cleaning: During data collection, factors such as equipment malfunctions, sensor errors, or environmental interference may introduce noise or outliers into the raw data. Therefore, data cleaning is necessary. Denoising algorithms such as median filtering or mean filtering are typically used for data preprocessing to remove these outliers. For data points with significant errors, algorithms can be used to identify and correct them to avoid impacting subsequent analysis results.

[0067] Missing value imputation: During the production process, missing data may occur due to measurement errors or sensor malfunctions. For these missing values, imputation methods generally include mean-based imputation or nearest-neighbor imputation. Mean-based imputation is simple and intuitive, suitable for situations with few missing data points. Nearest-neighbor imputation, on the other hand, better preserves the original distribution characteristics of the data, making it particularly suitable for situations with significant data loss.

[0068] For example, if a characteristic such as hardness is missing from raw material data, it can be imputed by calculating the average of that characteristic in historical data or by using values ​​from nearby raw material samples. This helps reduce bias caused by missing data.

[0069] Data standardization: Different production parameters and raw material characteristics typically have different dimensions and numerical ranges. To prevent certain features from dominating the analysis, data standardization is necessary. Standardization usually involves transforming data according to its mean and standard deviation, ensuring that the data for each feature is distributed within the same scale. Standardized data is more suitable for subsequent modeling and optimization work.

[0070] In some embodiments, the Z-score normalization method is used to transform all features. Specifically, the normalization formula is as follows:

[0071] ;

[0072] in, The standardized value. The original data values, The mean of the features, The standard deviation of the feature is denoted as . Using this method, all feature data will be transformed into a standard normal distribution with a mean of 0 and a standard deviation of 1, avoiding the influence of different units.

[0073] During data preprocessing, all steps must be flexibly adjusted based on the specific production scenario and data characteristics. For example, certain high-precision production processes may require more detailed missing value imputation and noise removal strategies, while other simpler production processes may be handled with simpler methods. The effectiveness of data preprocessing directly impacts the accuracy and efficiency of subsequent evaluation model construction; therefore, this step requires meticulous operation to ensure that the processed data quality meets the required standards.

[0074] Specifically, successful data preprocessing is crucial for subsequent steps, as it determines whether the raw material and production process compatibility assessment model receives accurate input data. Errors or biases in the data will affect the model's predictions, leading to ineffective subsequent process adjustments, increased production costs, and extended production cycles.

[0075] For step S2, in this embodiment, the information entropy minimization method is used to construct a raw material and production process compatibility evaluation model based on the processed raw material data. Process parameters are then optimized to achieve higher production efficiency and product quality. Information entropy, as a tool for measuring uncertainty, can help identify the optimal match between raw materials and process parameters, thereby enabling precise process adjustments.

[0076] In this embodiment, the raw material data is first cleaned and standardized, and then multiple process parameters are quantitatively analyzed using the information entropy minimization method. The principle of information entropy minimization is to select the combination of process parameters that can reduce system uncertainty in order to achieve the optimal production process.

[0077] Generally, information entropy minimization methods can effectively reduce uncertainty in complex systems when dealing with multiple variables, helping the system make more precise process adjustment choices. Information entropy methods can not only assess the compatibility between raw materials and processes, but also provide guidance for optimization objectives, ensuring balance under multi-objective conditions.

[0078] In one possible implementation, the information entropy minimization method quantifies process parameters using the following formula:

[0079] ;

[0080] in, Represents entropy; This represents the probability distribution of process parameters; This represents the total number of process parameters; This specifies the possible values ​​for each process parameter. In practical applications, the distribution of process parameters is obtained from historical production data and used to optimize existing production processes.

[0081] Alternatively, this embodiment can also employ an extended form of the entropy model to handle more complex multidimensional data, especially when multiple process variables are involved. For example, a weighted entropy model can be used, where the weight of each process parameter is adjusted according to its magnitude of influence on the final product quality:

[0082] ;

[0083] in, The weighted entropy value; This represents the probability distribution of process parameters; The weight of the process parameter indicates the importance of that process parameter to product quality; Possible values ​​for each process parameter; This represents the total number of process parameters. The weighted entropy model allows for more precise control over the impact of each process parameter during optimization, ensuring more efficient production while meeting quality requirements.

[0084] In other embodiments, information gain can also be used to assist in selecting optimal process parameters. In this case, by calculating the gain of each process parameter, it can help identify which parameter combinations maximize production efficiency and product quality while reducing entropy. For example, the objective function based on information gain is as follows:

[0085] ;

[0086] in, For information gain, The entropy value of the initial state. This represents the optimized entropy value. A larger information gain value indicates a smaller impact of current process parameters on the uncertainty of the system.

[0087] In practice, the information entropy minimization method can help identify the most suitable combination of process parameters for current production needs. This method not only improves the consistency of product quality but also reduces additional costs and scrap rates caused by mismatched raw materials during production.

[0088] To ensure the model operates efficiently in practical applications, this embodiment also employs computer-programmed automated optimization algorithms. These algorithms enable rapid processing of multidimensional process data and adjustment of production process parameters based on the optimization results of the objective function. This not only reduces manual intervention and improves efficiency but also ensures the accuracy and stability of process adjustments.

[0089] Ultimately, by employing the information entropy minimization method and its optimization model, the stability and adaptability of the production process can be significantly improved while ensuring both quality and efficiency. Implementing this method not only enhances production efficiency but also helps manufacturers provide more precise process adjustment solutions when facing environments with high complexity of raw materials and processes.

[0090] For step S3, in this embodiment, a swarm intelligence optimization algorithm is used to further optimize the production process parameters based on the evaluation results of the raw material and production process compatibility assessment model. Swarm intelligence optimization algorithms are optimization techniques that simulate the behavior of groups in nature, such as particle swarm optimization (PSO), which can find the optimal combination of process parameters under multi-objective optimization conditions. This process can effectively improve production efficiency, reduce scrap rates, and ensure the quality of the final product.

[0091] In this embodiment, the swarm intelligence optimization algorithm optimizes multiple process objectives, such as minimizing scrap rate, maximizing product quality, and improving production efficiency, by evaluating the compatibility between raw material characteristics and process requirements. The optimization process does not rely on human experience but is automated through a computer program, greatly improving the accuracy and efficiency of the production process.

[0092] Specifically, swarm intelligence optimization algorithms construct a multi-objective optimization problem, aiming to simultaneously optimize multiple process objectives. Common objectives include:

[0093] Minimize scrap rate: By adjusting process parameters, reduce the amount of scrap generated during production, thereby reducing production costs and resource waste.

[0094] Maximize product quality: Improve product consistency and quality by optimizing process parameters to ensure compliance with customer requirements.

[0095] Improve production efficiency: By adjusting process parameters, increase production speed, thereby shortening the production cycle.

[0096] In some embodiments, the objective function is:

[0097] ;

[0098] in, The objective function value; This refers to the quantity of scrap. Let be the total production quantity. The optimization objective of this function is to minimize the scrap rate, ensuring that resources are used most efficiently.

[0099] As an alternative, swarm intelligence algorithms can optimize multiple objective functions using particle swarm optimization (PSO). PSO finds the optimal solution by simulating the motion of particles in a search space. In each iteration, each particle updates its position and shares information with other particles, gradually approaching the global optimum.

[0100] Specifically, the particle swarm optimization algorithm uses the following formula:

[0101] ;

[0102] ;

[0103] in, Represents particles At any moment speed; Represents particles At the present moment Location; The optimal position for the particle itself; The optimal position for the entire group; For inertial weights, and For learning factors; and It is a random number; Inertial weight; Indicates the first The particle in the first The speed of time; Indicates the first The particle in the first The position at that moment; Indicates the first The position of each particle in the search space.

[0104] In this embodiment, the particle swarm optimization algorithm can adjust the weight of each process parameter and find a balance among multiple process objectives. For example, the optimized process parameters can effectively reduce the scrap rate while improving product quality and production efficiency.

[0105] Specifically, when swarm intelligence optimization algorithms are executed in production processes, they gradually approach the optimal combination of process parameters through continuous iteration and local search. This optimization process can find the best parameters for process adjustment under the constraints of multiple objectives, thereby meeting different production requirements.

[0106] Furthermore, the advantage of swarm intelligence algorithms lies in their ability to adapt to dynamically changing production environments. For different characteristics of raw materials, the algorithm adjusts its optimization objectives in real time to cope with potential production fluctuations. This characteristic ensures the broad adaptability of process modification methods in practical applications, maintaining excellent performance under various production conditions.

[0107] In practical applications, it can be combined with other optimization algorithms, such as genetic algorithms (GA) or simulated annealing (SA), and used in conjunction with swarm intelligence optimization algorithms to further improve optimization accuracy. For example, when dealing with multiple complex objectives and constraints, genetic algorithms can complement particle swarm optimization algorithms, increasing the breadth and depth of the search space, thereby enhancing the overall optimization effect.

[0108] In addition, in some embodiments, swarm intelligence optimization algorithms can be combined with deep learning methods, utilizing neural networks to model the complex nonlinear relationships between raw materials and process parameters. This combination enables the optimization process to be more intelligent, further improving production efficiency and quality.

[0109] For step S4, in this embodiment, a Support Vector Machine (SVM) model is used to perform real-time adaptability evaluation of the new raw material information to ensure that the characteristics of the raw materials meet the requirements of the current production process. Through the classification capabilities of the SVM model, we can automatically determine whether the newly introduced raw materials are suitable for the current production process, and then adjust process parameters or production strategies to ensure the quality of the final product and the stability of production.

[0110] Generally, SVM models can find the optimal classification hyperplane in high-dimensional space, classifying the fit between raw material characteristics and production processes. In this way, we can achieve automated and precise process adjustments, reduce human intervention, and improve production efficiency and product consistency.

[0111] In this embodiment, the Support Vector Machine (SVM) model is trained using historical data to correlate the characteristics of raw materials (such as hardness, density, and chemical composition) with process parameters (such as temperature, pressure, and speed). After training, the SVM model can identify the compatibility of raw materials with the current production process and evaluate new raw material samples in real time. If the raw material is compatible with the current process, the SVM will return a positive result; otherwise, it will return a negative result, indicating the need to adjust the production process.

[0112] Specifically, support vector machines classify new input data by defining a decision function, which can be expressed as:

[0113] ;

[0114] in, Let be the decision function; The feature vector of the raw materials; Let be the normal vector of the hyperplane; For bias terms; Vector dot product. If If so, the raw material is deemed suitable for the current production process; if If the raw materials are not suitable, they are deemed unsuitable. This process ensures that every batch of raw materials in the production process can be accurately evaluated, avoiding production risks caused by incompatible raw materials.

[0115] As an alternative, SVM models improve classification performance by optimizing the hyperplane margin (maximum margin classification), making the model more robust in practical applications. For example, this can be achieved by choosing an appropriate penalty parameter. This allows for adjustments to the tolerance for misclassification, thereby balancing the model's complexity and generalization ability, and avoiding overfitting or underfitting.

[0116] In one possible implementation, SVM can also utilize kernel functions to handle the nonlinear relationship between raw materials and process characteristics. Specifically, the radial basis function (RBF) is widely used in extensions of SVM, and its kernel function formula is as follows:

[0117] ;

[0118] in, Kernel function, representing the sample and Similarity in high-dimensional mapping spaces; This is the input feature vector of the raw material sample; To compare sample vectors; The Euclidean distance between the characteristics of the raw materials; It is an exponential function; These are the parameters of the kernel function. By using kernel functions, SVM can achieve non-linear classification in high-dimensional spaces, thereby enhancing the model's adaptability to complex data.

[0119] In some implementations, the SVM model may suffer from data imbalance, meaning that the number of raw material samples for certain categories is relatively small. This can lead to bias in the classification model. In this case, the imbalance can be addressed by adjusting the class weights. Specifically, the weights of rarer classes can be increased during training, causing the model to give more weight to these classes when classifying. For example, the objective function of the SVM can be adjusted as follows:

[0120] ;

[0121] in, For penalty parameters; These are slack variables; Let be the normal vector of the hyperplane; This represents the total number of training data points. In cases of class imbalance, modifying the parameters in the objective function allows the model to focus more on the minority class, thereby improving classification accuracy.

[0122] In this embodiment, support vector machines can also be integrated with other machine learning algorithms to improve the model's classification performance. For example, by integrating algorithms such as decision trees or random forests, the robustness of fitness assessment in complex production environments can be further enhanced. Ensemble learning methods can combine the advantages of multiple models, making fitness assessment more stable and accurate.

[0123] In step S5, in this embodiment, the current production process parameters are dynamically adjusted based on the support vector machine model's assessment of raw material suitability. This adjustment mechanism enables closed-loop coupling and adaptive optimization between process parameters and raw material properties, thereby meeting the accuracy requirements of real-time control and improving the overall system response capability.

[0124] In this embodiment, once the SVM model determines that the current raw materials are not compatible with the established process, the process parameter adjustment mechanism is automatically triggered.

[0125] The process parameters include, but are not limited to, temperature setting, pressing pressure, holding time, and material conveying speed. The specific adjustment values ​​are based on the output values ​​of the decision function of the support vector machine. The difference between the parameter and the set threshold is calculated. Specifically, the parameter adjustment amount... The calculation can be performed using the following formula:

[0126] ;

[0127] in, Indicates the first The adjustment increment of each process parameter; In order to be with the first The values ​​of the sensitivity factors related to the process parameters are determined based on historical data and process impact analysis. This represents the target discrimination value corresponding to the SVM evaluation output; The value of the decision function output by the SVM model in the preceding steps.

[0128] Under normal circumstances, this adjustment strategy can achieve flexible matching of process conditions without changing the raw materials, so that the performance differences of the raw materials are compensated by the process parameters, thereby improving the stability and adaptability of the system.

[0129] As an alternative, process boundary constraints can be introduced during parameter adjustment to ensure that parameter variations do not exceed set limits. For example, when adjusting temperature parameters, the maximum adjustment range should be limited by the equipment's load-bearing capacity and the material's thermal stability curve to avoid inducing nonlinear behavior in the system or causing abnormal energy consumption.

[0130] In one possible implementation, parameter tuning can employ a step-by-step optimization strategy to reduce transient instability caused by a large, one-time adjustment. Under this strategy, parameter tuning is performed in stages, and corrections are made at each stage based on feedback results, forming a closed-loop control.

[0131] In some embodiments, actual online material measurement data, process equipment status parameters, and environmental fluctuation factors are all incorporated into the adjustment judgment. Specifically, multi-source feature encoding is input into the fusion module, and features are extracted through an improved converter structure to assist in determining the direction and magnitude of parameter adjustment.

[0132] In addition, a priority scheduling strategy can be configured in the system implementation to dynamically sort the priority of process parameter adjustments, so that key process variables (such as pressure and temperature) receive higher adjustment weights, while secondary variables (such as speed and flow) are adjusted appropriately according to the system operating status.

[0133] Parameter adjustment strategies can also be integrated into digital twin models. By comparing and simulating the virtual production line with the real production process, the impact of parameter changes on the performance of finished products can be predicted in real time. This allows for feedforward judgment capabilities before adjustment, providing a more targeted basis for subsequent control.

[0134] The product evaluation-based process correction system described below corresponds to and can be referenced in relation to the product evaluation-based process correction method described above.

[0135] Please see the appendix Figure 2 The present invention also provides a process correction system based on product evaluation, comprising:

[0136] The data collection module is used to collect raw material characteristic data, production process parameters, and product performance data in real time.

[0137] The data processing module is used to clean, standardize, and fill in missing values ​​for the collected raw material data.

[0138] The evaluation module is used to construct a raw material and production process compatibility evaluation model based on the processed data using the information entropy minimization method.

[0139] The optimization module is used to optimize production process parameters based on the results of the adaptability evaluation model using swarm intelligence algorithms.

[0140] The real-time evaluation module is used to evaluate the suitability of new raw materials using a support vector machine model.

[0141] The dynamic adjustment module is used to adjust production process parameters based on real-time evaluation results to ensure the best match between the process and the characteristics of raw materials.

[0142] The data collection module integrates sensors, production monitoring equipment, and product quality inspection tools to capture key data points in the process in real time. It supports multiple data formats and interface types, enabling seamless integration with existing production systems and ensuring the accuracy and real-time nature of data acquisition.

[0143] For the data processing module, data preprocessing techniques are employed, such as removing noisy data, imputing missing values, and standardizing numerical ranges, to ensure high data quality and facilitate subsequent analysis. This module can perform data correction based on rule engines or machine learning algorithms to improve the automation and accuracy of data processing.

[0144] The evaluation module uses information entropy theory to quantitatively assess the compatibility between raw materials and process parameters, maximizing production efficiency and minimizing resource waste. The module can automatically adjust the weights and parameters of the evaluation model according to different production environments to adapt to different types of production tasks.

[0145] The optimization module identifies the optimal combination of multiple process parameters to maximize production efficiency and reduce energy consumption and resource waste. This module can collaborate with the production scheduling system to update and adjust production plans in real time, thereby improving the flexibility and efficiency of the production process.

[0146] The real-time evaluation module assesses the suitability of new raw materials as they enter the production process, ensuring material compatibility during production. The support vector machine model responds quickly to real-time data streams, providing accurate predictions to guide immediate adjustments to the production line.

[0147] The dynamic adjustment module automatically adjusts the parameters of production equipment, such as temperature, pressure, and speed, to maintain production stability and product quality. It supports both manual intervention and automatic adjustment modes, allowing for flexible switching based on production requirements. Furthermore, it features intelligent learning capabilities to progressively improve the accuracy of process parameter adjustments.

[0148] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0149] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A process modification method based on product evaluation, characterized in that, Includes the following steps: Collect and preprocess raw material characteristic data, production process parameters, and product performance data, and clean, standardize, and fill in missing values ​​for the processed raw material data; Based on the processed raw material data, a raw material and production process compatibility assessment model is constructed using the information entropy minimization method, and process parameters are optimized. Based on the evaluation results of the raw material and production process compatibility assessment model, the process parameters are optimized using a swarm intelligence optimization algorithm. The support vector machine model is used to evaluate the adaptability of new raw material information in real time and determine whether it meets the requirements of the current production process. Based on real-time evaluation results, production process parameters are dynamically adjusted to ensure the best match between the process and the characteristics of raw materials. The steps of constructing a raw material and production process compatibility assessment model and optimizing process parameters using the information entropy minimization method include: Define the compatibility evaluation index between raw materials and production processes, and use the information entropy method to quantify and analyze multiple process parameters; Based on the principle of minimizing information entropy, the optimal combination of raw materials and process parameters is selected to reduce information uncertainty and optimize the production process. The process parameters are optimized using computer programs, and the optimal parameter combination is achieved through iterative algorithms. The formula for quantifying multiple process parameters using the information entropy method is as follows: Where H(X) is the entropy value; p(x) i The probability distribution for each process parameter; x i Let be the specific value of the i-th process parameter; n is the total number of process parameters; The objective function for multi-objective optimization using swarm intelligence optimization algorithm is: Among them, R waste R represents the quantity of scrap. total f1 represents the total production quantity; f1 represents the objective function value. The function for real-time adaptation evaluation of new raw material information using the support vector machine model is: Where w is the hyperplane normal vector; C is the penalty parameter; ξ i is a slack variable; m is the total number of training data.

2. The process modification method based on product evaluation according to claim 1, characterized in that, The steps for collecting and preprocessing raw material characteristic data, production process parameters, and product performance data include: The system collects data on the physicochemical properties of raw materials and relevant parameters of the production process in real time during the production process, and inputs the collected data into the data processing system. The collected raw material data is denoised by using a filtering algorithm to remove outliers. Missing values ​​are imputed using either mean imputation or nearest neighbor imputation. Standardize all data to ensure that all feature values ​​are within the same range.

3. The process modification method based on product evaluation according to claim 1, characterized in that, The steps for optimizing process parameters using swarm intelligence optimization algorithms include: Based on the evaluation results of the raw material and production process compatibility assessment model, a multi-objective optimization problem is constructed, with objectives including minimizing scrap rate, maximizing product quality, and improving production efficiency. The particle swarm optimization algorithm is used to solve multiple objective functions and obtain a set of optimal process parameters; The production process is adjusted based on the optimization results to ensure that all target constraints are met.

4. The process modification method based on product evaluation according to claim 1, characterized in that, The steps for using a support vector machine model to perform real-time adaptability evaluation of new raw material information include: The new raw material characteristic data is input into the trained support vector machine model for real-time classification and judgment to determine whether the raw material is suitable for the current production process. The support vector machine model classifies new raw materials as positive or negative based on the sample data in the training set and outputs the fit result. If the real-time evaluation results indicate that the raw materials are not suitable for the current process, the system will automatically prompt and adjust the process parameters.

5. The process modification method based on product evaluation according to claim 1, characterized in that, The step of dynamically adjusting production process parameters based on real-time evaluation results includes: Based on the real-time adaptability assessment results, the production process parameters are adjusted in conjunction with the optimization objectives; During the adjustment process, the process parameter values ​​are dynamically adjusted to ensure optimal matching with the characteristics of raw materials and changes in the production environment. The adjusted process parameters are fed back to the production line in real time for actual production operations, and the effects are tracked and feedback is provided.

6. A product-evaluation-based process correction system, applied to the product-evaluation-based process correction method according to any one of claims 1-5, characterized in that, Includes the following modules: The data collection module is used to collect raw material characteristic data, production process parameters, and product performance data in real time. The data processing module is used to clean, standardize, and fill in missing values ​​for the collected raw material data. The evaluation module is used to construct a raw material and production process compatibility evaluation model based on the processed data using the information entropy minimization method. The optimization module is used to optimize production process parameters based on the results of the adaptability evaluation model using swarm intelligence algorithms. The real-time evaluation module is used to evaluate the suitability of new raw materials using a support vector machine model. The dynamic adjustment module is used to adjust production process parameters based on real-time evaluation results to ensure the best match between the process and the characteristics of raw materials.

Images