Papermaking drying process energy efficiency monitoring identification method based on comprehensive production state
By combining clustering and data processing modules with stacked algorithm models, a powerful regression and classification prediction model is formed, which solves the problem of insufficient accuracy and prediction precision in energy efficiency monitoring during paper drying, and achieves more efficient energy efficiency assessment and management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-10
- Publication Date
- 2026-03-31
AI Technical Summary
Existing technologies lack a comprehensive approach to energy efficiency monitoring in the paper drying process, resulting in insufficient identification accuracy and energy efficiency level prediction precision. Furthermore, the models perform poorly when faced with complex nonlinear relationships.
A comprehensive production status-based approach is adopted, which groups and preprocesses the data through clustering and data processing modules. It combines multiple base models with a stacking algorithm model to form a powerful regression and classification prediction model. The Taylor expansion method is used to generate an energy efficiency coordinate map for energy efficiency level assessment.
It significantly improves the accuracy and prediction precision of energy efficiency monitoring, enhances the environmental adaptability and flexibility of the model, provides more accurate energy efficiency assessment results, and optimizes overall energy efficiency management.
Smart Images
Figure CN121765451A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of paper drying monitoring technology, specifically relating to a method for monitoring and identifying energy efficiency in the paper drying process based on comprehensive production status. Background Technology
[0002] Paper drying is a core process in the pulp and paper industry, directly impacting final paper quality and production efficiency. However, the drying process typically involves high energy consumption, placing a significant burden on the environment. Therefore, accurately monitoring and evaluating the energy efficiency of the paper drying process is crucial for improving production efficiency, reducing energy consumption, and minimizing environmental impact.
[0003] Currently, there are some research findings in the field of energy efficiency monitoring, including statistical analysis-based methods and some simple machine learning models. However, these methods often focus on single indicators or local optimization, neglecting the correlation between various factors and their impact on overall energy efficiency. Furthermore, most existing methods lack approaches that comprehensively consider multiple factors for specific industries (such as the paper industry).
[0004] Random Forest (RF) and Support Vector Machine (SVM) are commonly used foundational machine learning models, widely applied in classification and regression tasks. Random Forest, by integrating the predictions of multiple decision trees, possesses strong robustness and high-dimensional data processing capabilities, making it particularly suitable for tasks with complex features. Support Vector Machine, by constructing a maximum margin hyperplane, effectively addresses high-dimensional and nonlinear problems, excelling especially in few-shot learning and complex pattern recognition. While these foundational models perform well in traditional tasks, stacked models (stacked generalization models), as an ensemble learning method, can further improve prediction accuracy when facing complex nonlinear relationships between data. By combining the outputs of multiple foundational learners, stacked models, through a hierarchical structure, combine different types of learners to form a more powerful prediction framework, better capturing complex patterns in the data and thus enhancing the model's generalization ability.
[0005] Nevertheless, there is still room for improvement in the overall performance, recognition accuracy, energy efficiency level prediction accuracy, and environmental adaptability of current paper drying process models. Summary of the Invention
[0006] To address the shortcomings of existing technologies, the present invention aims to provide a method for monitoring and identifying energy efficiency in the paper drying process based on comprehensive production status, thereby solving the aforementioned technical problems in the prior art.
[0007] The objective of this invention can be achieved through the following technical solutions: A method for monitoring and identifying energy efficiency in the paper drying process based on comprehensive production status includes a clustering and data processing module, a stacked algorithm model for processing regression tasks, an energy efficiency level evaluation module, and a stacked algorithm model for processing classification tasks. The clustering and data processing module: clusters the input data, then preprocesses the grouped data to identify and classify potential production states in the data; The stacked algorithm model for processing regression tasks introduces a stacked algorithm to fuse multiple base models to form a powerful regression prediction model. The energy efficiency rating module: Based on the relationship between paper machine production output and steam flow, it uses Taylor expansion to analyze energy efficiency data, generates an energy efficiency coordinate graph, and evaluates and labels the data based on the energy efficiency coordinate graph; The stacked algorithm model for processing classification tasks uses a stacked algorithm to fuse multiple base model classifiers to form a powerful classification prediction model. This combines the advantages of various algorithms to improve the overall model performance and obtain the optimal multi-level energy efficiency prediction results for the paper drying process.
[0008] Furthermore, the solution provided by the clustering and data processing module includes data clustering and data collection and preprocessing; Among them, A1, data clustering: A101. Collect historical production data from the production line and compile a historical production dataset for the drying section. A102. Perform feature analysis and variance calculation on the collected data, and select the variable data that are strongly correlated with the energy efficiency of the drying section. A103. Based on the process analysis and mathematical analysis of the drying section, it was found that the quantitative data before sizing and the winding speed data are strongly correlated with the production status category and production energy efficiency. The hierarchical clustering algorithm was used to cluster the quantitative data before sizing and the winding speed data according to the production status. A2. Data Collection and Preprocessing: A201. Based on the 3σ principle and box plot method, outlier removal is performed on the calculated data in A102 to obtain the calculated data; A202. After performing LOESS local weighted regression and standardization on the calculated data, the data is split into training and test sets.
[0009] Furthermore, in A201, the data removed includes variable data with high collinearity, low correlation in the drying section, and low variance.
[0010] Furthermore, the stacked algorithm model processes the regression task module, and the optimization process is expressed as follows: ; in This represents the prediction result of the meta-learner model; This represents the prediction result of the nth base learner; This represents the weight of the prediction result of the nth base learner.
[0011] Furthermore, the optimization method for the regression task module of the stacked algorithm model includes optimization of the data feature extraction method and model training; The optimization of the data feature extraction method specifically includes the following: Selection of base learners: Random forest algorithm, support vector machine algorithm, and class boosting algorithm are selected as base learners for training and prediction to obtain the prediction results of each base model; Merging features: The predictions from these base models are merged as new features to form a new training dataset; The model training specifically includes the following: Feature reconstruction: The features extracted by the base learners are combined and used as new features in the training set, which are then input into the meta-learner training task; Evaluate model performance: Use the test set to evaluate the predictive performance of the base learner model and the meta-learner model optimized by the stacking algorithm.
[0012] Furthermore, the energy efficiency rating module obtains the relationship between paper machine production output and steam flow rate, and uses the Taylor expansion method to draw an energy efficiency coordinate graph to analyze the energy efficiency status of the drying section. The relationship between paper machine production output X and steam flow rate Y is expressed as follows: ; in and This indicates the current paper machine output and the benchmark paper machine output; and This indicates the current steam flow rate and benchmark steam flow rate of the drying section.
[0013] Furthermore, the judgment method of the energy efficiency rating module includes energy efficiency index calculation and energy efficiency rating characterization; The energy efficiency index calculation: using the collected basic data, the energy efficiency index of the drying section is calculated; The energy efficiency rating is characterized by: based on the relationship between paper machine production output and steam flow, using Taylor expansion to draw an energy efficiency coordinate graph, analyzing the energy efficiency rating of the drying section, and labeling the production data with "I", "II", and "III" energy efficiency rating labels.
[0014] Furthermore, the stacked algorithm model's classification task processing module calculates the accuracy, precision, and recall of different classifiers as follows: Accuracy: ; Accuracy: ; Recall rate: ; Where TP indicates that the true class of the sample is positive, and the model identifies it as positive; FP indicates that the true class of the sample is negative, but the model identifies it as positive. FN indicates that the true class of the sample is positive, but the model identifies it as negative. TN indicates that the true class of the sample is negative, and the model identifies it as negative.
[0015] Furthermore, the application method of the stacked algorithm model for processing classification tasks includes the selection of base model classifiers and the calculation and comparison of the accuracy, precision, and recall of each classifier; The selection of the base model classifier: classifiers are constructed using random forest algorithm, support vector machine algorithm and class boosting algorithm respectively, and the base model classifiers are stacked to construct a meta-classifier; The calculation and comparison of the accuracy, precision, and recall of each classifier: calculate the accuracy, precision, and recall of each classifier, select the classifier with the best performance, and obtain the best prediction result.
[0016] The energy efficiency monitoring and identification method for the paper drying process based on comprehensive production status includes the following steps: S1. Clustering and preprocessing are performed on the data from the stacked algorithm model processing regression task module, energy efficiency level evaluation module, and stacked algorithm model processing classification task module through the clustering and data processing module; S2. The regression task module uses a stacked algorithm to fuse multiple base models to form a powerful regression prediction model. This combines the advantages of various algorithms to improve the overall model performance and enhance the accuracy of steam flow prediction. S3. Based on the relationship between paper machine production output and steam flow, the energy efficiency evaluation module uses Taylor expansion to analyze energy efficiency data, generates an energy efficiency coordinate graph, and labels the production data with "I", "II", and "III" energy efficiency levels according to the energy efficiency coordinate graph. S4. By using a stacking algorithm to fuse multiple base model classifiers to form a powerful classification prediction model, the advantages of various different algorithms are combined to improve the overall model performance, so as to obtain the optimal multi-level energy efficiency prediction results for the paper drying process.
[0017] The beneficial effects of this invention are: 1. This invention ensures the stacking algorithm model through clustering and data processing modules, effectively grouping and preprocessing the input data to enable accurate analysis of data under different production conditions. This method's efficient execution in regression tasks, energy efficiency rating, and classification tasks ensures the model maintains high accuracy even when facing complex and nonlinear relationships, thereby significantly improving the reliability of energy efficiency monitoring.
[0018] 2. The stacked algorithm model regression task module in this invention integrates multiple base models and makes full use of the advantages of each model to form a powerful regression prediction model. It combines the advantages of different algorithms, significantly improving the recognition accuracy. The optimized model can better discover potential patterns in the data and improve prediction performance. This ensemble learning method not only improves the overall model performance, but also enhances the prediction accuracy of key indicators such as steam flow.
[0019] 3. The energy efficiency rating module of this invention generates an energy efficiency coordinate graph and performs detailed classification by refining the evaluation standards. This provides more accurate and reliable evaluation results. This method helps enterprises accurately identify and improve various energy efficiency indicators, providing them with more precise and reliable energy efficiency evaluation results, enabling them to clearly identify and improve various energy efficiency indicators, thereby optimizing overall energy efficiency management.
[0020] 4. The stacked algorithm model classification task module in this invention integrates multiple base model classifiers to form a powerful classification prediction model. It divides production data into multiple states, enabling the effective capture of production characteristics under different states. This design makes the system more flexible and adaptable in practical applications, allowing it to react quickly to different production conditions. By combining the advantages of different algorithms, it ensures the model's efficient and stable operation in complex environments.
[0021] 5. By integrating clustering and data processing, stacked algorithm regression and classification task modules, and energy efficiency rating module, this invention significantly improves the overall performance, recognition accuracy, energy efficiency rating precision, and environmental adaptability of the model, ensuring high efficiency, accuracy, and stability throughout the entire process from data input to energy efficiency rating output.
[0022] 6. The method of the present invention is not only applicable to the paper industry, but can also be extended to other fields such as textile manufacturing. By deploying corresponding monitoring systems, energy efficiency management optimization can be achieved in multiple industries and scenarios, creating greater economic value for enterprises. Attached Figure Description
[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0024] Figure 1 This is a schematic diagram of the overall process structure of an embodiment of the present invention; Figure 2 This is a comparison chart of the model prediction performance of the stacked algorithm model in the regression task module of this invention (in the first production state). Figure 3 This is an energy efficiency rating analysis diagram of the energy efficiency rating module in this embodiment of the invention; Figure 4 This is a graph showing the energy efficiency recognition performance of each classifier in all production states of the stacked algorithm model processing classification task module in this embodiment of the invention. Detailed Implementation
[0025] The technical solutions of 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.
[0026] like Figure 1 As shown, this embodiment of the invention provides a method for monitoring and identifying energy efficiency in the paper drying process based on comprehensive production status, including a clustering and data processing module, a stacked algorithm model for processing regression tasks, an energy efficiency level evaluation module, and a stacked algorithm model for processing classification tasks. Clustering and Data Processing Module: Clusters and groups the input data, and then preprocesses the grouped data to identify and classify potential production states in the data.
[0027] Responsible for processing and clustering the data from the stacked algorithm model's regression task module, energy efficiency rating module, and classification task module to ensure that the energy efficiency rating monitoring model for the paper drying process maintains optimal performance.
[0028] Specifically, this includes data clustering and data collection and preprocessing; the solution for the clustering and data processing module includes the following steps: A1. Data Clustering: A101. Collect historical production data from the production line, including but not limited to key parameters such as steam flow rate, drying temperature, papermaking output, pre-sizing amount, and winding speed, and compile a historical production dataset for the drying section. A102. Perform feature analysis and variance calculation on the collected data to select variables that are strongly correlated with the energy efficiency of the drying section; use statistical methods (such as analysis of variance) to perform feature analysis on each variable and select variables that are related to energy efficiency to reduce model complexity and improve prediction efficiency. A103. Based on the process analysis and mathematical analysis of the drying section, it was found that the quantitative data before sizing and the winding speed data are strongly correlated with the production status category and production energy efficiency. The hierarchical clustering algorithm was used to cluster the quantitative data before sizing and the winding speed data according to the production status.
[0029] A2. Data Collection and Preprocessing: A201. Based on the 3σ principle and box plot method, outlier removal is performed on the calculated data in S102. The removed data includes variables with high collinearity, low correlation in the drying section, and low variance, thereby obtaining the calculated data. A202. After performing LOESS local weighted regression and standardization on the calculated data, the data is split into training and test sets.
[0030] Clustering groups similar data, allowing each group to be processed using the most suitable model, thus improving processing accuracy and effectiveness. Different models handle specific categories of data, reducing computational complexity and time consumption, and overall accelerating processing speed.
[0031] The stacked algorithm model for handling regression tasks employs a stacked algorithm to fuse multiple base models into a powerful regression prediction model. This combines the advantages of various algorithms to improve the overall model performance and enhance the accuracy of steam flow prediction. The optimization process can be represented as follows: ; in, This represents the prediction result of the meta-learner model; This represents the prediction result of the nth base learner; This represents the weight of the prediction result of the nth base learner.
[0032] Stacked models (stacked generalization models), as an ensemble learning method, can further improve prediction accuracy by combining the outputs of multiple base learners. Through a hierarchical structure, stacked models combine different types of learners to form a more powerful prediction framework, which can better capture complex patterns in the data, thereby enhancing the model's generalization ability.
[0033] like Figure 2 As shown, the optimization method for the regression task module of the stacked algorithm model includes optimization of data feature extraction method and model training.
[0034] Optimization of data feature extraction methods: Selection of base learners: Random forest algorithm, support vector machine algorithm, and class boosting algorithm are selected as base learners for training and prediction to obtain the prediction results of each base model; Merging features: The predictions from these base models are merged as new features to form a new training dataset.
[0035] Model training: Feature reconstruction: The features extracted by the base learners are combined and used as new features in the training set, which are then input into the meta-learner training task; Evaluate model performance: Use the test set to evaluate the predictive performance of the base learner model and the meta-learner model optimized by the stacked generalization algorithm.
[0036] Energy efficiency rating module: Based on the relationship between paper machine production output and steam flow, the energy efficiency data is analyzed using the Taylor expansion method to generate an energy efficiency coordinate graph. Based on the energy efficiency coordinate graph, the data is evaluated and labeled with energy efficiency rating. The relationship between the paper machine's output and steam flow rate can be expressed as follows:
[0037] in and This refers to the current papermaking output and benchmark papermaking output of the papermaking machine. and This represents the current steam flow rate and benchmark steam flow rate of the drying section.
[0038] like Figure 3 As shown, the judgment method of the energy efficiency rating module includes energy efficiency index calculation and energy efficiency rating characterization; Energy efficiency index calculation: Using the collected basic data, calculate the energy efficiency index of the drying section to fully understand the objective and detailed situation of various energy efficiency indicators of the drying section.
[0039] Energy efficiency rating: Based on the relationship between paper machine production output and steam flow, the energy efficiency coordinate graph is drawn using the Taylor expansion method to effectively classify and manage the energy efficiency of the drying section. The energy efficiency level of the drying section is analyzed, and the production data is labeled with "I", "II" and "III" energy efficiency level labels, enabling enterprises to more clearly identify the level of energy efficiency and take corresponding improvement measures.
[0040] By implementing energy efficiency rating systems, companies can more easily identify their energy efficiency levels and take corresponding improvement measures for different levels, thereby optimizing overall energy efficiency management. The three-tiered energy efficiency rating system refines the energy efficiency evaluation standards, making the assessment results more accurate and facilitating the identification of specific areas and directions for energy efficiency optimization.
[0041] The stacked algorithm model for classification tasks uses a stacked algorithm to fuse multiple base model classifiers into a powerful classification prediction model. This combines the advantages of different algorithms to improve the overall model performance and achieve optimal prediction results. The calculation process for accuracy, precision, and recall is as follows: ; ; ; The classifications are True Positive (TP), False Positive (FP), False Negative (FN), and True Negative (TN). TP: The sample's true class is positive, and the model classifies it as positive. FP: The sample's true class is negative, but the model classifies it as positive. FN: The sample's true class is positive, but the model classifies it as negative. TN: The sample's true class is negative, and the model classifies it as negative.
[0042] Machine learning algorithms can quickly adapt to changes in the characteristics of production data in the drying section, ensuring the identification of different energy efficiency levels and demonstrating strong environmental adaptability.
[0043] like Figure 4 As shown, the application method of the stacked algorithm model for processing classification tasks includes the selection of base model classifiers and the calculation and comparison of the accuracy, precision, and recall of each classifier. Selection of base model classifiers: classifiers are constructed using random forest algorithm, support vector machine algorithm, and class boosting algorithm respectively. The base model classifiers are stacked to construct a meta-classifier. Calculate and compare the accuracy, precision, and recall of each classifier: Calculate the accuracy, precision, and recall of each classifier, and select the classifier with the best performance to obtain the best prediction results.
[0044] The method for monitoring and identifying energy efficiency in the paper drying process based on comprehensive production status includes the following steps: S1. The clustering and data processing module preprocesses and clusters the data from the stacked algorithm model processing regression task module, energy efficiency level evaluation module, and stacked algorithm model processing classification task module. S2. The regression task module uses a stacked algorithm to fuse multiple base models to form a powerful regression prediction model. This combines the advantages of various algorithms to improve the overall model performance and enhance the accuracy of steam flow prediction. S3. Based on the relationship between paper machine production output and steam flow, the energy efficiency evaluation module uses Taylor expansion to analyze energy efficiency data, generates an energy efficiency coordinate graph, and labels the production data with "I", "II", and "III" energy efficiency levels according to the energy efficiency coordinate graph. S4. The classification task module, which uses a stacked algorithm model, compares the accuracy, precision, and recall of different classifiers. Based on real-time feedback and model performance, it selects the classifier with the best performance to obtain the optimal prediction result, thereby achieving the model's performance in identifying energy efficiency status.
[0045] Example 1: Energy Efficiency Monitoring of Drying Section In the drying section production line, the clustering and data processing model first receives production data. Addressing the issue of multiple production states in actual production, a hierarchical clustering algorithm is used to divide the data into seven production states. Subsequently, the data under different production states undergoes preprocessing, including data matching and outlier removal, to resolve issues such as missing data, outliers, and noise. Next, base models are built using algorithms such as random forest, support vector machine, and stacked generalization to train and predict the data. Then, the stacked generalization algorithm is used to merge these base model predictions as new features, combining them into a new training dataset, which is input into the meta-model training task to build a steam flow prediction model. Then, based on the relationship between paper machine production output and steam flow, Taylor expansion is used to construct the "I," "II," and "III" energy efficiency level labels for the drying section. Finally, the stacked generalization algorithm is again used to construct an energy efficiency level evaluation model for the drying section that does not distinguish between production states, thus achieving energy efficiency level monitoring of the paper drying process under all production states.
[0046] Example 2: Application by textile manufacturers An energy efficiency monitoring system was deployed in the dyeing and finishing process of a textile production line, installing sensors and metering equipment to record key parameters such as temperature, humidity, and energy consumption in real time. Based on the energy efficiency assessment results, different batches of fabric were categorized and managed to optimize the production process and ensure stability. Simultaneously, process parameters were adjusted using the monitoring data to reduce energy consumption and improve production efficiency. By monitoring and adjusting process parameters in real time, textile manufacturers successfully reduced energy consumption and improved production efficiency, demonstrating the broad applicability of this method in other industries.
[0047] These two embodiments demonstrate the practical application of an energy efficiency monitoring and identification method for the paper drying process based on comprehensive production status. Through the collaborative work of clustering and data processing modules, stacked algorithm model regression task processing modules, energy efficiency level evaluation modules, and stacked algorithm model classification task processing modules, it achieves efficient analysis and processing of production data, providing strong support for energy efficiency monitoring applications on various production lines.
[0048] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.
Claims
1. A method for monitoring and identifying the energy efficiency of a paper drying process based on the integrated production state, characterized by, The clustering and data processing module, the stacked algorithm model processing regression task module, the energy efficiency level evaluation module, and the stacked algorithm model processing classification task module are included. The clustering and data processing module clusters and groups input data, pre-processes the grouped data, and identifies and classifies potential production states in the data. The stacked algorithm model processing regression task module introduces a stacked algorithm to fuse multiple base models to form a powerful regression prediction model. The energy efficiency level evaluation module analyzes energy efficiency data using Taylor expansion based on the relationship between paper machine production and steam flow, generates an energy efficiency coordinate graph, and evaluates and identifies the energy efficiency level of the data based on the energy efficiency coordinate graph. The stacked algorithm model processing classification task module uses a stacked algorithm to fuse multiple base model classifiers to form a powerful classification prediction model, combines the advantages of multiple algorithms to improve the performance of the overall model, and obtains optimal multi-stage energy efficiency prediction results for the paper drying process.
2. The method for monitoring and identifying the energy efficiency of a paper drying process based on the integrated production state according to claim 1, characterized in that, The clustering and data processing module includes data clustering, data collection, and preprocessing. A1, data clustering: A101, collect production line historical production data, and organize the drying department historical production data set; A102, perform feature analysis and variance calculation on the collected data, and select variable data with strong correlation with drying department energy efficiency; A103, based on drying department process analysis and mathematical analysis, it is concluded that the sizing quantity before sizing and the winding speed data have strong correlation with production state category and production energy efficiency, and hierarchical clustering algorithm is used to cluster the sizing quantity before sizing and the winding speed data. A2, data collection and preprocessing: A201, based on the 3σ principle and box plot method, the calculated data in A102 is excluded to obtain the calculated data; A202, after LOESS local weighted regression processing and standardization operation, the data is divided into training set and test set.
3. The method for monitoring and identifying the energy efficiency of a paper drying process based on the integrated production state according to claim 2, characterized in that, In A201, the excluded data includes high collinearity, drying department low correlation, and low variance variable data.
4. The method for monitoring and identifying the energy efficiency of a papermaking drying process based on the integrated production state according to claim 1, characterized in that, The optimization process of the stacked algorithm model processing regression task module is represented as follows: ； wherein represents the prediction result of the meta-learner model; denotes the prediction result of the nth base learner; denotes the weight of the prediction result of the nth base learner.
5. The method for monitoring and identifying the energy efficiency of a paper drying process based on the integrated production state according to claim 1, characterized in that, The optimization method of the stacked algorithm model processing regression task module includes data feature extraction method optimization and model training. The data feature extraction method optimization specifically includes the following: Selection of base learners: Random forest algorithm, support vector machine algorithm, and category boosting algorithm are selected as base learners for training and prediction to obtain the prediction results of each base model; Merging features: merging the prediction results of these base models as new features to form a new training data set; The model training specifically includes the following: Feature reconstruction: using the merged features extracted by the base learners as new features of the training set, input into the meta-learner training task; Evaluate model performance: use the test set to evaluate the prediction performance of the base learner model and the meta-learner model optimized by the stacked algorithm.
6. The method for monitoring and identifying the energy efficiency of a paper drying process based on the integrated production state according to claim 5, characterized in that, The energy efficiency grade evaluation module obtains the relationship between the paper machine production and the steam flow, adopts Taylor expansion method to draw the energy efficiency coordinate graph, analyzes the energy efficiency state of the drying section, and the relationship between the paper machine production X and the steam flow Y is expressed as: ； wherein and represents the current paper machine production and the benchmark paper machine production; and Qd and Qb represent the current steam flow and the benchmark steam flow of the drying section.
7. The method according to claim 6, characterized in that, The judgment method of the energy efficiency grade evaluation module includes energy efficiency index calculation and energy efficiency grade characterization; The energy efficiency index calculation: the energy efficiency index of the drying section is calculated by using the collected basic data; The energy efficiency grade characterization: based on the relationship between the paper machine production and the steam flow, the Taylor expansion method is used to draw the energy efficiency coordinate graph, the energy efficiency grade of the drying section is analyzed, and the production data is characterized by the energy efficiency grade labels "I", "II" and "III".
8. The method for monitoring and identifying the energy efficiency of a papermaking drying process based on the integrated production state according to claim 1, characterized in that, The calculation process of the accuracy, precision and recall of different classifiers in the stacked algorithm model processing classification task module is as follows: Accuracy: ; Precision: ; Recall rate: ; TP represents that the true class of the sample is positive, and the model identifies it as positive; FP represents that the true class of the sample is negative, but the model identifies it as positive; FN represents that the true class of the sample is positive, but the model identifies it as negative; TN represents that the true class of the sample is negative, and the model identifies it as negative.
9. The method according to claim 8, characterized in that, The application method of the stacked algorithm model processing classification task module includes the selection of base model classifiers and the calculation and comparison of the accuracy, precision and recall of each classifier; The selection of base model classifiers: the random forest algorithm, support vector machine algorithm and category boosting algorithm are used to construct classifiers, and the base model classifiers are combined into meta-classifiers through the stacking algorithm; The calculation and comparison of the accuracy, precision and recall of each classifier: the accuracy, precision and recall of each classifier are calculated, and the classifier with the best performance is selected to obtain the optimal prediction result.
10. The method according to any one of claims 1 to 9, characterized in that The following steps are included: S1, the data of the stacked algorithm model processing regression task module, the energy efficiency grade evaluation module and the stacked algorithm model processing classification task module are clustered and preprocessed by the clustering and data processing module; S2, the stacked algorithm model processing regression task module adopts the stacking algorithm to fuse multiple base models to form a powerful regression prediction model, combines the advantages of multiple different algorithms to improve the performance of the overall model, and improves the accuracy of steam flow prediction; S3, the energy efficiency grade evaluation module analyzes the energy efficiency data based on the relationship between the paper machine production and the steam flow, generates the energy efficiency coordinate graph by using the Taylor expansion method, and labels the production data with the energy efficiency grade labels "I", "II" and "III" according to the energy efficiency coordinate graph; S4, the stacked algorithm fuses multiple base model classifiers to form a powerful classification prediction model, combines the advantages of multiple different algorithms to improve the performance of the overall model, and obtains the optimal multi-stage energy efficiency prediction result of the paper drying process.