A data-driven optimization method for smelting control

Abstract

The present invention relates to a data-driven optimization method for smelting control, belonging to the technical field of smelting control. The method includes the steps of: classifying operating conditions through clustering analysis using historical operation data of the current furnace, mining association rules within each cluster, and establishing a parameter rule base; obtaining the operating condition category of the current furnace temperature through the parameter rule base according to the production data of the furnace operation; making a real-time prediction of the change of the furnace temperature according to the current furnace temperature characteristic data and the operating condition category; selecting the control parameter combination with the optimal energy consumption from the parameter dataset according to the current operating condition category and the predicted furnace temperature data; monitoring the furnace temperature data in real time and performing operating condition identification, adjusting the control parameter combination of the smelting operation through furnace temperature prediction, and giving an early warning through the degree of furnace temperature fluctuation. The present invention predicts the furnace temperature through operating condition identification, and then screens the control parameter combination according to the prediction result and the operating condition category, so as to adjust and control the smelting process to achieve energy-saving optimization.

Description

Technical Field

[0001] The present invention belongs to the technical field of smelting control, and particularly relates to a data-driven smelting control optimization method. Background Art

[0002] With the continuous advancement of the automation and intelligentization processes of industrial production, the optimization of the smelting process has become a key area for improving energy utilization efficiency and reducing production costs. Traditional smelting processes usually rely on experience and manual adjustment. The control system's regulation of furnace temperature mostly depends on static or simple rules, which often leads to energy waste and insufficient control accuracy.

[0003] Currently, the analysis and prediction based on historical data have become an effective means to improve the efficiency of the smelting process. However, in the existing technologies, there is a lack of efficient methods for classifying working conditions in detail and characterizing states. The model for predicting furnace temperature only relies on historical data and cannot accurately reflect the complex relationships among all variables. At the same time, when screening the optimal control parameters, there is a lack of in-depth exploration of the complex associations between working condition categories and control parameters, resulting in insufficient consideration of the diversity of all working conditions, and the control effect is not ideal, thus affecting the energy-saving effect. Summary of the Invention

[0004] To solve the above problems existing in the prior art, the present invention provides a data-driven smelting control optimization method. Through working condition identification, furnace temperature prediction is carried out, and then according to the prediction results and working condition categories, a control parameter combination is screened, so as to regulate the smelting process to achieve energy-saving optimization.

[0005] The object of the present invention can be achieved by the following technical solutions:

[0006] A data-driven smelting control optimization method, comprising the following steps:

[0007] Establish a parameter rule base: Classify working conditions by using clustering analysis on the historical operation data of the current furnace, and conduct association rule mining within each cluster to establish a parameter rule base;

[0008] Obtain working conditions: Obtain the working condition category of the current furnace temperature according to the production data of the furnace operation through the parameter rule base;

[0009] Predict the furnace temperature trend: According to the current furnace temperature characteristic data and working condition category, conduct real-time prediction on the change of the furnace temperature;

[0010] Obtain control parameters: Select the control parameter combination with the optimal energy consumption from the parameter dataset according to the current working condition category and the predicted furnace temperature data;

[0011] Monitor operation anomalies: Monitor the furnace temperature data in real time and conduct working condition identification, adjust the control parameter combination of the smelting operation through furnace temperature prediction, and at the same time give an early warning through the degree of furnace temperature fluctuation.

[0012] Further, the historical operation data includes sensor timing data, process parameters, and equipment status logs, and the production data includes real-time furnace operation data and process parameters, where the process parameters include the heating time, furnace pressure, air-fuel ratio, and heating power set for process production.

[0013] Further, the establishment of the parameter rule base includes the following steps:

[0014] Data preprocessing: Perform binning operations on continuous control parameters to discretize them, remove noise and anomalies in the timing data, and enhance the data.

[0015] Operating condition identification: Identify the operating condition category of the current furnace temperature change through clustering analysis by mining relevant feature data related to the furnace temperature change.

[0016] Rule mining: Construct an association rule mining algorithm in each operating condition category, set the confidence level and minimum support, and generate an association rule with the operating condition category as the result to construct the parameter rule base.

[0017] Further, the operating condition identification includes the following steps:

[0018] Feature selection: Divide the data into grids and calculate the mutual information, obtain the MIC value through different grid segmentation methods, and select the top 30 features with the strongest relationship with the target variable to construct the furnace temperature feature data.

[0019] Density clustering: Dynamically select the minimum number of samples according to the sampling period of the sliding window, and use the selected features and the minimum number of samples to perform clustering using the HDBSCAN algorithm.

[0020] Operating condition definition: Reduce the dimension of the clustered data through UMAP, visualize the data points after dimension reduction, represent each data point with a different color to represent different clusters, and each cluster represents an operating condition state of the furnace temperature, and define the operating condition category.

[0021] Further, the rule mining includes the following steps:

[0022] Construct a transaction database: Convert the data in each operating condition category into a transaction format, represent each sample as a group item, and use the features of each sample as items.

[0023] Mine association rules: Use the association rule mining algorithm to mine the data in each operating condition category, set the minimum support and minimum confidence as the thresholds for rule screening, and output the association rules that meet the conditions.

[0024] Update association rules: Set a sliding window to add new operation data according to the operation characteristics of smelting, use the incremental learning method to update the association rules, and adjust the support and confidence levels.

[0025] Further, the furnace temperature trend prediction includes the following steps:

[0026] Build a model: Use a long short-term memory neural network as the furnace temperature prediction model, including an input layer, an LSTM layer, a fully connected layer, and an output layer;

[0027] Model training: Preprocess the input variables to convert them into time series data and divide them into training sets, validation sets, and test sets together with the corresponding furnace temperature values. Input the training data into the LSTM model and train the network through backpropagation and gradient descent algorithms;

[0028] Real-time prediction: Obtain the furnace temperature characteristic data and the working condition category of the furnace temperature in real time, input the obtained data into the trained furnace temperature prediction model, and perform real-time prediction of the furnace temperature.

[0029] Further, the input variables of the input layer include furnace temperature characteristic data and working condition category, and the output layer includes a single node representing the predicted furnace temperature.

[0030] Further, the obtaining of control parameters includes the following steps:

[0031] Data merging: Match the control parameter combinations corresponding to all working condition category data from historical operation data and establish a parameter data set;

[0032] Energy consumption calculation: Calculate the comprehensive energy consumption data during the operation of each group of control parameter combinations, including heating energy consumption, cooling energy consumption, and other energy consumptions, and store the energy consumption data together with its corresponding working condition category;

[0033] Parameter selection: Calculate the working condition categories similar to the current working condition category, and limit a group of relevant working condition categories by the difference between the predicted furnace temperature and the target furnace temperature. Select the control parameter combination with the minimum energy consumption in this group of working condition categories.

[0034] Further, in the operation anomaly monitoring, judging the stability of the furnace temperature change includes the following steps:

[0035] Preset a time window and monitor the furnace temperature data within the time window;

[0036] Calculate the deviation between the average furnace temperature and the target furnace temperature, and at the same time calculate the standard deviation of the furnace temperature data as the furnace temperature fluctuation value;

[0037] Set judgment thresholds for different levels of average deviation and fluctuation values. When both the average deviation and the fluctuation value exceed the set judgment thresholds, give a warning prompt according to the exceeding intervals set by the levels.

[0038] Furthermore, it also includes the steps:

[0039] Operation data update: Collect all the operation data of the current furnace, and according to the judgment of furnace temperature fluctuation, mark the combination of equipment control parameters in the stable state, use the collected data to update the parameter rule base, and use the marked control parameter combination to update the parameter data set.

[0040] The beneficial effects of the present invention are as follows:

[0041] The present invention realizes energy-saving optimization in the smelting process by adjusting the furnace temperature control parameters. First, cluster analysis is performed on historical data to establish working condition categories characterized by furnace temperature. Then, association rules are mined in each working condition category, and the condition data type and numerical range are used as the state representations of the working condition categories. Then, the working condition categories and furnace temperature characteristic data are input into the prediction model to obtain the predicted furnace temperature, thereby improving the accuracy of furnace temperature prediction. Finally, according to the predicted furnace temperature and working condition categories, eligible similar working condition categories are screened, and the optimal control parameters are obtained through energy consumption calculation, and the smelting process is continuously regulated to achieve energy-saving optimization.

[0042] In the process of obtaining the working conditions of the present invention, it is realized based on the established parameter rule base. The parameter rule base is constructed by clustering the relevant data of furnace temperature changes and using association rules to mine associated data for each working condition as conditions. Through the working condition categories, the internal differences of furnace temperature changes are more accurately grasped and the working condition changes are refined, realizing high-precision working condition division, establishing the basic conditions for furnace temperature prediction and adjustment of control parameters. Then, a set of working condition categories is selected through the prediction results of furnace temperature and working condition categories to obtain the control parameter combination, realizing data-driven dynamic adaptation. This process mines the association relationship between data based on the big data of smelting, tightly combines the furnace temperature state and related characteristics together, and then uses furnace temperature prediction to fine-tune the working condition categories and change the control parameter combination to achieve the purpose of energy-saving optimization while meeting the target furnace temperature. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings.

[0044] Figure 1 It is a flow chart of a data-driven smelting control optimization method provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following describes in detail the specific embodiments, structures, features and effects of the present invention with reference to the accompanying drawings and preferred embodiments.

[0046] This embodiment provides a data-driven optimization method for smelting control, as Figure 1 shown, which includes the following steps:

[0047] S1. Establish a parameter rule base: Collect the historical operation data of the current furnace, classify the working conditions through clustering analysis after preprocessing, and mine association rules within each cluster to establish a parameter rule base associated with the working condition categories, including the following steps:

[0048] Among them, the historical operation data includes sensor time series data, process parameters, and equipment status logs, etc. The sensor time series data are the status parameters real-time monitored by the sensors installed during the smelting process, such as furnace temperature, gas flow rate, current, voltage, etc. These time series data reflect the dynamic characteristics during the smelting process; process parameters such as the heating time, furnace pressure, air-fuel ratio, heating power, etc. set by the process production are important factors affecting the smelting process; the equipment status logs record the operation status of the equipment, such as equipment startup, shutdown, fault alarm and other information.

[0049] S11. Data preprocessing: Perform binning operations on continuous control parameters to make them discrete, remove the noise and anomalies in the time series data, and enhance the data.

[0050] In this embodiment, OPTICS adaptive binning is adopted for continuous control parameters, a 30% window overlap rate is reserved to ensure the effectiveness of the association rules, STL decomposition is integrated for denoising of working condition characteristics, and data enhancement is realized through a SAE autoencoder.

[0051] It should be noted that the purpose of discretization is to convert continuous data into discrete intervals for subsequent analysis. The OPTICS adaptive binning method is specifically used, which can adapt to the data distribution and maintain efficient association rule mining capabilities. OPTICS (Ordering Points To Identify the Clustering Structure) is a density-based clustering algorithm that can handle the non-uniform distribution of data. STL (Seasonal and Trend decomposition using Loess) decomposition is a method for decomposing time series data, which decomposes the original time series data into three parts: trend, seasonality, and residuals. Through STL decomposition, the long-term trend and seasonal fluctuations in the smelting process can be separated, and then the noise and abnormal data caused by these factors can be removed. For the data in the smelting process, the SAE (Sparse Autoencoder) autoencoder can perform non-linear mapping on the input data, compress the original data into a low-dimensional space through compression, and then decode and reconstruct it; during the training process, SAE learns the main features of the data, removes noise or outliers from the reconstructed data to achieve data denoising and enhancement, and at the same time, through the autoencoder, missing values can be compensated and abnormal data can be corrected, thereby enhancing the reliability and accuracy of the data.

[0052] S12. Working condition identification: By mining the relevant feature data related to the furnace temperature change for clustering analysis, identify the working condition categories of the current furnace temperature change, including the following steps:

[0053] S121. Feature selection: Divide the data into grids and calculate the mutual information. Obtain the MIC value through different grid segmentation methods, and select the top 30 features with the strongest relationship with the target variable to form the furnace temperature feature data;

[0054] It can be understood that MIC (Maximal Information Coefficient) is a statistic for measuring the correlation between two variables. MIC aims to capture the non-linear and complex relationships between variables. The calculation process of MIC is to select an appropriate grid segmentation, calculate the mutual information, and maximize the mutual information value, and finally obtain a numerical value reflecting the strength of the relationship between the two variables.

[0055] It should be noted that in this embodiment, the purpose of working condition classification is to better grasp the furnace temperature change situation for furnace temperature prediction. Therefore, the target variable selected when initially calculating the MIC value is the furnace temperature.

[0056] S122. Density-based clustering: Dynamically select the minimum number of samples according to the sampling period of the sliding window, and use the selected features and the minimum number of samples to perform clustering using the HDBSCAN algorithm.

[0057] It can be understood that HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise) is a density-based clustering algorithm. Its core idea is to perform clustering based on the local density differences of the data. Different from the traditional K-Means, HDBSCAN does not require specifying the number of clusters in advance and can handle clusters of arbitrary shapes and noise points.

[0058] It should be noted that HDBSCAN determines whether a data point belongs to a certain cluster by calculating the local density of each data point, including the core distance and the reachability distance. First, calculate the "connection" relationship between points through the reachability distance and generate a minimum spanning tree, which ensures that all points are connected through the shortest reachability distance; gradually form different cluster structures based on the decreasing order of the reachability distance, and then construct a hierarchical clustering tree by merging and splitting clusters. Determine the final cluster structure by calculating the stability of clusters at different levels; finally, traverse the hierarchical tree and automatically select the cluster splitting point based on the stability threshold to determine the range of each cluster.

[0059] S123. Operating condition definition: Use UMAP to reduce the dimension of the clustered data, visualize the reduced data points, represent different clusters with different colors for each data point, and each cluster represents a furnace temperature operating condition state to define the operating condition categories.

[0060] Among them, UMAP (Uniform Manifold Approximation and Projection) is a dimensionality reduction technique widely used to map high-dimensional data into a low-dimensional space, usually used for data visualization and feature learning. Its main advantage is that it can retain the local structure of the data while retaining the global structure, making the reduced data have good visualization effects in two-dimensional or three-dimensional space.

[0061] S13. Rule mining: Build an association rule mining algorithm in each operating condition category, set the confidence level and the minimum support degree, and generate an association rule construction parameter rule base with the operating condition category as the result, including the following steps:

[0062] S131. Build a transaction database: Convert the data in each operating condition category into a transaction format, represent each sample as a group item, and use the features of each sample as items;

[0063] It should be noted that the sample features include at least one of the 30 features related to the furnace temperature change, such as the air-fuel ratio, the cooling water flow rate, etc. Through the mining of association rules for each type of working condition, the final conditional data types and value ranges are different, which are also the feature manifestations of different furnace temperature working conditions.

[0064] S132. Mine association rules: Use association rule mining algorithms (such as Apriori, FP-Growth, Eclat, etc.) to mine the data of each working condition category (cluster), set the minimum support and minimum confidence as the thresholds for rule screening, and output the association rules that meet the conditions.

[0065] In this embodiment, the initial minimum support is set to 0.15 and the minimum confidence is set to 0.8. They can be adjusted in subsequent applications to meet the needs of data mining. In addition, the lift can be set to 1.5 to limit the positive association between the premise and the result of the rule.

[0066] S133. Update association rules: According to the operating characteristics of smelting, set a sliding window to add new operating data, use the incremental learning method to update the association rules, and adjust the support and confidence.

[0067] It can be understood that by monitoring the operating state of the furnace, continuously collecting state data, adding the collected data to the historical operating data through a sliding window, and using the incremental learning method to process the new data to complete the update of the association rules, the consumption of computing resources can be reduced and the efficiency can be improved.

[0068] It should be noted that through the historical operating data, the working condition categories of furnace temperature change are defined by clustering the data related to furnace temperature change. Then, through association analysis in each type of working condition category, the parameters associated with the working condition are mined from the historical operating data and a parameter rule base is established. Using the associated parameters as the rule antecedent and the working condition category as the rule consequent, further, the working condition category of the current furnace temperature can be obtained through the real-time data in the production process. The mining of association rules is mainly to further associate the furnace temperature working conditions with the relevant data, so as to more accurately grasp the internal differences of furnace temperature changes and refine the working condition changes, and improve the accuracy of furnace temperature prediction.

[0069] S2. Obtain working conditions: Obtain the working condition category of the current furnace temperature according to the real-time operating data of the furnace and the process parameters through the parameter rule base.

[0070] It should be noted that the parameter rule base contains association rules based on operating condition categories. During application, real-time production data is used to match the antecedents (conditions) of the association rules to obtain the operating condition categories. Further, under the premise of each operating condition category, the data type composition and specific parameters of the antecedents are different. The production data includes real-time data of the furnace operation and process parameters. The real-time data is specifically sensor time-series data, and the process parameters include heating time, furnace pressure, air-fuel ratio, heating power, etc. set in the process production.

[0071] S3. Furnace temperature trend prediction: Establish a furnace temperature prediction model, and based on the current furnace temperature characteristic data and the operating condition category of the furnace temperature, conduct real-time prediction of the change of the furnace temperature, including the following steps:

[0072] S31. Model construction: Use a long short-term memory neural network (LSTM) as the furnace temperature prediction model, including an input layer, an LSTM layer, a fully connected layer, and an output layer;

[0073] The input variables of the input layer include furnace temperature characteristic data and operating condition categories. The LSTM layer uses LSTM units to process time-series data, and one or more LSTM layers can be set according to the characteristics of the data. The number of units in the LSTM layer can be adjusted according to the actual situation, and generally 64, 128, or 256 units are used. The fully connected layer (Dense layer) is set after the LSTM layer, and feature fusion and further learning are carried out by adding one or more fully connected layers. The output layer is usually a single node, representing the predicted furnace temperature.

[0074] S32. Model training: The input variables are preprocessed and converted into time-series data, and are divided into a training set, a validation set, and a test set together with the corresponding furnace temperature values. The training data is input into the LSTM model, and the network is trained through backpropagation and gradient descent algorithms;

[0075] S33. Real-time prediction: Real-time obtain the furnace temperature characteristic data and the operating condition category of the furnace temperature, and input the obtained data into the trained furnace temperature prediction model for real-time prediction of the furnace temperature.

[0076] It should be noted that in this embodiment, by adding the identified operating condition category of the current furnace temperature to the input of the model, the change characteristics of the furnace temperature operating condition are more refinedly reflected in the prediction model, thereby improving the accuracy of the furnace temperature prediction and providing a basis for the later energy-saving optimization control.

[0077] S4. Obtain control parameters: Select the control parameter combination with the optimal energy consumption from the parameter dataset according to the current operating condition category and the predicted furnace temperature data, including the following steps:

[0078] S41. Data Merging: Match the control parameter combinations corresponding to all working condition category data from historical operation data to establish a parameter data set.

[0079] Based on the working condition categories in the parameter rule base, supplement the corresponding control parameter combinations for each working condition using time series, including heating power, cooling flow rate, furnace gas flow rate, furnace charge input, etc., to construct the mapping relationship between the working conditions and control parameters, thereby establishing a complete data set.

[0080] S42. Energy Consumption Calculation: Calculate the comprehensive energy consumption data during the operation of each set of control parameter combinations, including heating energy consumption, cooling energy consumption, and other energy consumptions, and store the energy consumption data together with its corresponding working condition category.

[0081] It can be understood that for each set of control parameter combinations, calculate the comprehensive energy consumption based on the corresponding operation time, furnace temperature change, external conditions, etc. Other energy consumptions include the influence of furnace gas flow rate, furnace charge input, etc. on energy consumption.

[0082] S43. Parameter Selection: Calculate the working condition categories similar to the current working condition category, and limit a set of relevant working condition categories by predicting the difference between the furnace temperature and the target furnace temperature, and select the control parameter combination with the minimum energy consumption in this set of working condition categories.

[0083] Specifically, the similar working condition categories are obtained by calculating the similarity of the working conditions and screening with a similarity threshold, then define the deviation range between the predicted furnace temperature and the target furnace temperature, extract the samples with excessive deviation from the similar working conditions, and select the working condition category with the minimum energy consumption from the remaining samples to obtain the control parameters.

[0084] It should be noted that for the working condition categories obtained by clustering using the HDBSCAN algorithm, set the metric distance during the operation of HDBSCAN. Clusters with smaller distances are close in the feature space. Therefore, a series of working condition categories can be searched by calculating the center points between adjacent clusters, thereby providing some control parameter combinations for the control parameter selection. During the process of obtaining the working conditions, it is relatively accurate to determine the working condition type through the conditional data of the association rules. The purpose is to accurately reflect the state of the furnace temperature through the working condition categories. In this step, a set of similar working conditions is searched based on the current working condition category to provide a basis for the selection of control parameters. By further limiting the similar working condition categories according to the furnace temperature deviation, the working condition categories that do not meet the furnace temperature adjustment are excluded, and the control parameter combination is obtained through energy consumption comparison among the remaining working condition category samples, so as to adjust the smelting in advance and in line with the furnace temperature working condition to achieve energy-saving optimization.

[0085] S5. Operation Abnormality Monitoring: Real-time monitor the furnace temperature data and perform working condition identification, adjust the control parameter combination of the smelting operation through furnace temperature prediction, and at the same time set a furnace temperature fluctuation threshold to judge the stability of the furnace temperature change, and give an early warning when it exceeds the set threshold.

[0086] Among them, determining the stability of the furnace temperature change includes the following steps:

[0087] S51. Preset a time window and monitor the furnace temperature data within the time window;

[0088] S52. Calculate the deviation between the average furnace temperature and the target furnace temperature, and at the same time calculate the standard deviation of the furnace temperature data as the furnace temperature fluctuation value;

[0089] S53. Set judgment thresholds for different levels of average deviation and fluctuation value. When both the average deviation and the fluctuation value exceed the set judgment thresholds, give a warning prompt according to the exceeding intervals set by the levels.

[0090] It should be noted that the energy-saving optimization in this embodiment is based on predicting the furnace temperature. By selecting the optimal energy consumption control parameter combination through different furnace temperature working condition categories, timely fine-tuning the melting process control parameters, thereby continuously optimizing the energy consumption of melting. At the same time, a fault tolerance mechanism is set for the optimization process through warning prompts.

[0091] S6. Update operation data: Collect all the operation data of the current furnace, and according to the judgment of the furnace temperature fluctuation, mark the equipment control parameter combination in the stable state, use the collected data to update the parameter rule base, and use the marked control parameter combination to update the parameter data set.

[0092] It should be noted that the data in the furnace operation process is used to fill the historical operation data, so as to establish a more perfect and accurate parameter rule base. At the same time, since in the process of obtaining the control parameters, the parameter data set is established based on time series to complete data mapping, and there is a situation of furnace temperature lag in data matching. Therefore, by marking the furnace temperature in the subsequent data collection to update the parameter data set, the accuracy of obtaining the control parameters is improved, and the influence brought by the furnace temperature lag is reduced.

[0093] The present invention realizes energy-saving optimization in the melting process by adjusting the furnace temperature control parameters. First, perform cluster analysis on historical data to establish working condition categories characterized by furnace temperature. Then, mine association rules in each working condition category, use the conditional data type and numerical range as the state representation of the working condition category. Then, input the working condition category and furnace temperature characteristic data into the prediction model to obtain the predicted furnace temperature, thereby improving the accuracy of furnace temperature prediction. Finally, according to the predicted furnace temperature and the working condition category, screen out the similar working condition categories that meet the conditions, calculate the optimal control parameters through energy consumption calculation, and continuously adjust the melting process to achieve energy-saving optimization.

[0094] In the process of obtaining the working conditions of the present invention, it is realized based on the established parameter rule base. The parameter rule base is constructed by clustering the relevant data of the furnace temperature change and then using the association rule to mine the associated data for each working condition as conditions. By the working condition category, the internal differences of the furnace temperature change are grasped more accurately and the working condition change is refined, so as to realize the high-precision working condition division, establish the basic conditions for the furnace temperature prediction and the adjustment of the control parameters. Then, a set of working condition categories is selected through the prediction result of the furnace temperature and the working condition category to obtain the control parameter combination, realizing the data-driven dynamic adaptation. This process mines the association relationship between the data based on the big data of smelting, closely combines the furnace temperature state and the relevant characteristics, and then uses the furnace temperature prediction to fine-tune the working condition category and change the control parameter combination to achieve the purpose of energy-saving optimization under the condition of meeting the target furnace temperature.

[0095] The above is only the preferred embodiment of the present invention, and it is not intended to limit the present invention in any form. Although the present invention has been disclosed above with the preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to make equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A data-driven smelting control optimization method, characterized in that: The following steps are involved: Establish a parameter rule base: Use cluster analysis to classify the operating conditions based on the historical operating data of the current furnace, and mine association rules within each cluster to establish a parameter rule base; Working condition acquisition: according to the production data of the furnace operation, the working condition category of the current furnace temperature is obtained through the parameter rule library; Furnace temperature trend prediction: Based on the current furnace temperature characteristic data and working condition category, the change of furnace temperature is predicted in real time; Obtain control parameters: Select the control parameter combination with the best energy consumption from the parameter data set according to the current operating condition category and predicted furnace temperature data; Abnormal operation monitoring: Real-time monitoring of furnace temperature data and identification of working conditions, and adjustment of control parameter combination of smelting operation through furnace temperature prediction, while early warning through furnace temperature fluctuation degree; The historical operation data includes sensor timing data, process parameters and equipment status logs, and the production data includes furnace operation real-time data and process parameters, wherein the process parameters include heating time, furnace pressure, air-fuel ratio and heating power set in the process production; The step of establishing a parameter rule base comprises the following steps: Data preprocessing: binning continuous control parameters to discretize them, remove noise and anomalies in time series data, and enhance the data; Working condition identification: By mining the characteristic data related to the furnace temperature change and performing cluster analysis, the working condition category of the current furnace temperature change can be identified; Rule mining: construct an association rule mining algorithm in each working condition category, set the confidence level and minimum support level, and generate association rules with working condition categories as the result to build a parameter rule base; The obtaining of control parameters comprises the following steps: Data merging: Match the control parameter combinations corresponding to all operating condition category data from historical operation data to establish parameter data sets; Energy consumption calculation: Calculate the comprehensive energy consumption data of each set of control parameter combinations during operation, including heating energy consumption, cooling energy consumption and other energy consumption, and store the energy consumption data together with its corresponding operating condition category; Parameter selection: Calculate similar operating conditions based on the current operating condition, and define a group of related operating conditions by predicting the difference between the furnace temperature and the target furnace temperature, and select the control parameter combination with the lowest energy consumption in this group of operating conditions.

2. A data-driven smelting control optimization method according to claim 1, characterized in that: The working condition identification comprises the following steps: Feature selection: The data is divided into grids and mutual information is calculated. The MIC value is obtained through different grid segmentation methods. The top 30 features with the strongest relationship with the target variable are selected to construct the furnace temperature feature data; Density clustering: Dynamically select the minimum number of samples according to the sampling period of the sliding window, and use the selected features and the minimum number of samples to perform clustering using the HDBSCAN algorithm; Working condition definition: Use UMAP to reduce the dimension of the clustered data and visualize the data points after dimension reduction. Use different colors to represent different clusters for each data point. Each cluster represents a furnace temperature working condition and defines the working condition category.

3. A data-driven smelting control optimization method according to claim 2, characterized in that: The rule mining comprises the following steps: Construct a transaction database: convert the data in each working condition category into a transaction format, with each sample represented as a group item and the features of each sample as items; Mining association rules: Use the association rule mining algorithm to mine the data of each working condition category, set the minimum support and minimum confidence as the threshold for rule screening, and output the association rules that meet the conditions; Update association rules: According to the operation characteristics of smelting, set the sliding window to add new operation data, use the incremental learning method, update the association rules, and adjust the support and confidence.

4. The data-driven smelting control optimization method according to claim 1, characterized in that: The furnace temperature trend prediction comprises the following steps: Model construction: Long short-term memory neural network is used as the furnace temperature prediction model, including input layer, LSTM layer, fully connected layer and output layer; Model training: The input variables are converted into time series data through preprocessing and divided into training set, validation set and test set with the corresponding furnace temperature values. The training data is input into the LSTM model, and the network is trained through back propagation and gradient descent algorithms. Real-time prediction: Obtain furnace temperature characteristic data and furnace temperature operating condition category in real time, input the acquired data into the trained furnace temperature prediction model, and perform real-time prediction of furnace temperature.

5. A data-driven smelting control optimization method according to claim 4, characterized in that: The input variables of the input layer include furnace temperature characteristic data and operating condition categories, and the output layer includes a single node representing the predicted furnace temperature.

6. The data-driven smelting control optimization method according to claim 1, characterized in that: In the abnormal operation monitoring, early warning based on the furnace temperature fluctuation degree includes the following steps: Preset time window and monitor furnace temperature data within the time window; Calculate the deviation between the mean furnace temperature and the target furnace temperature, and calculate the standard deviation of the furnace temperature data as the furnace temperature fluctuation value; Set judgment thresholds for mean deviation and fluctuation value at different levels. When both mean deviation and fluctuation value exceed the set judgment threshold, an early warning will be issued according to the exceeded range set for the level.

7. The data-driven smelting control optimization method according to claim 1, characterized in that: Also includes the steps: Operation data update: Collect all the operation data of the current furnace, and mark the equipment control parameter combination in the stable state according to the judgment of furnace temperature fluctuation, update the parameter rule base with the collected data, and update the parameter data set with the marked control parameter combination.

Images