Blast furnace temperature closed-loop control method and device

Abstract

The invention relates to the technical field of blast furnace smelting, and discloses a blast furnace temperature closed-loop control method and device, and the method comprises the steps: obtaining blast furnace data corresponding to each time point in a preset smelting period; obtaining a blast furnace data pool based on the blast furnace data at each time point by adopting a multi-feature extraction method; a plurality of prediction models are adopted, and the predicted furnace temperature of each time point in the preset time period is obtained based on the blast furnace data pool; and performing heat adjustment on the blast furnace based on the blast furnace data at each time point in the preset smelting period and the predicted furnace temperature at each time point in the preset smelting period. According to the method, multiple features are extracted from blast furnace data by adopting multiple feature extraction methods, prediction is performed by adopting multiple prediction models, prediction errors of a single model are reduced, the reliability and accuracy of furnace temperature prediction are improved, heat adjustment is performed on the blast furnace based on the predicted furnace temperature, and the stability of the furnace temperature is maintained; accurate closed-loop control over the furnace temperature of the blast furnace is achieved, and the intelligent level is improved.

Description

Technical Field

[0001] The invention relates to the technical field of blast furnace smelting, and in particular to a blast furnace temperature closed-loop control method and device. Background Art

[0002] Blast furnace ironmaking is one of the core processes of the steel industry. Its production process is complex and highly dependent on the stability of the furnace temperature. The blast furnace temperature is the result of the smooth operation of the furnace conditions. At the same time, the stability of the furnace temperature in turn affects the smooth operation of the blast furnace. Therefore, if the furnace temperature trend can be predicted, control measures can be taken in advance to ensure the stability of the furnace temperature, thereby ensuring the stable operation of the blast furnace and realizing closed-loop control of the blast furnace temperature.

[0003] At present, operators predict the trend of furnace temperature by observing various characteristic indicators of blast furnaces, and intervene by adjusting the amount of coal to stabilize the furnace temperature. However, excessive reliance on the experience of operators cannot meet the development requirements of modern blast furnace intelligence. The error between the predicted furnace temperature and the actual furnace temperature is large, making it difficult to accurately achieve closed-loop control of blast furnace temperature. Summary of the invention

[0004] In view of this, the present invention provides a blast furnace temperature closed-loop control method and device to solve the problem that the blast furnace temperature prediction error is large, resulting in the inability to accurately achieve blast furnace temperature closed-loop control.

[0005] In a first aspect, the present invention provides a closed-loop control method for blast furnace temperature, the method comprising: Obtain the blast furnace data corresponding to each time point in a preset smelting cycle, where the preset smelting cycle ends at the current time point and is divided into multiple time points according to a preset time step; Using a variety of feature extraction methods, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool; Using multiple prediction models, prediction is performed based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in the preset time period. The preset time period is divided into multiple time points according to the preset time step; Based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period, the heat of the blast furnace is adjusted.

[0006] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention provides a rich historical data basis by acquiring blast furnace data at all time points in a preset smelting cycle, and adopts multiple feature extraction methods to obtain a blast furnace data pool based on the blast furnace data corresponding to each time point. Various features can be extracted from the original blast furnace data, which can better reflect the working status of the blast furnace. Taking into account that different prediction models have their own advantages in data processing, using multiple prediction models for prediction can effectively reduce the prediction error of a single model, improve the reliability and accuracy of furnace temperature prediction, adjust the heat of the blast furnace based on the predicted furnace temperature, maintain the stability of the furnace temperature, and realize accurate closed-loop control of the blast furnace temperature. Compared with relying on the experience of operators to predict and control the furnace temperature, the level of intelligence is improved.

[0007] In an optional embodiment, multiple feature extraction methods are used to extract features based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool, including: Perform correlation analysis on the blast furnace data corresponding to each time point to obtain the first blast furnace data pool; Perform variance analysis on the blast furnace data corresponding to each time point to obtain the second blast furnace data pool; Using the recursive feature elimination method, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain the third blast furnace data pool; A decision tree regression model is used to extract features based on the blast furnace data corresponding to each time point to obtain the fourth blast furnace data pool; The first blast furnace data pool, the second blast furnace data pool, the third blast furnace data pool and the fourth blast furnace data pool are taken as a union to obtain a blast furnace data pool.

[0008] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention can extract blast furnace features from multiple angles by combining multiple feature extraction methods, and the features extracted by multiple methods are taken as a union, which can complement their respective advantages and improve the quality of the overall data, thereby ensuring that the blast furnace data pool contains rich information, and can provide more comprehensive and accurate input for subsequent prediction models, thereby improving the accuracy of furnace temperature prediction.

[0009] In an optional embodiment, multiple prediction models are used to predict based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in the preset time period, including: Based on the blast furnace data corresponding to each time point in the preset smelting cycle, the decision tree random forest model, CatBoost gradient boosting model, XGBoost regression model and neural network model are pre-trained to obtain the optimal prediction model; The furnace temperature in the blast furnace data at the current time point is used as the initial furnace temperature, and the blast furnace data pool is predicted using the optimal prediction model to obtain the predicted furnace temperature at the next time point after the current time point; The predicted furnace temperature at the next time point after the current time point is used as the new initial furnace temperature, and the above process of predicting the furnace temperature is repeated until the preset time period is completed to obtain the predicted furnace temperature corresponding to each time point in the preset time period.

[0010] The closed-loop control method for blast furnace temperature provided by the embodiment of the present invention performs pre-training based on multiple prediction models, and uses the optimal prediction model with the best prediction performance to predict the furnace temperature, thereby improving the accuracy of furnace temperature prediction and reducing prediction errors. Through recursive prediction at each time point, it can reflect changes in the production process in real time and provide dynamic furnace temperature prediction results.

[0011] In an optional embodiment, based on the blast furnace data corresponding to each time point in the preset smelting cycle, the decision tree random forest model, the CatBoost gradient boosting model, the XGBoost regression model and the neural network model are pre-trained respectively to obtain the optimal prediction model, including: The decision tree random forest model, CatBoost gradient boosting model, XGBoost regression model and neural network model are used to predict the blast furnace data corresponding to any time point in the preset smelting cycle except the current time point and the starting time point, and the predicted furnace temperature corresponding to each model at the time point is obtained; For any model, the prediction error of the model is determined based on the furnace temperature in the blast furnace data at the time point and the predicted furnace temperature corresponding to the model at the time point; The model with the smallest prediction error among multiple models is determined as the optimal prediction model.

[0012] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention takes into account that different prediction models have their own advantages in processing nonlinear relationships and time series data with different characteristics. Therefore, when predicting the furnace temperature at each time point, the prediction errors of multiple prediction models are compared and the optimal model is selected, which can significantly improve the accuracy of furnace temperature prediction and reduce prediction errors.

[0013] In an optional embodiment, the blast furnace data includes a material rate, and based on the blast furnace data corresponding to each time point in a preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period, heat adjustment is performed on the blast furnace, including: When the material speed at two consecutive time points in the preset smelting cycle is greater than the preset material speed threshold, reducing the oxygen in the blast furnace; or, If the material speed does not exceed the preset material speed threshold at two consecutive time points in the preset smelting cycle, determine whether to adjust the coal quantity of the blast furnace.

[0014] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention monitors the blast furnace material rate in real time. When the material rate at two consecutive time points exceeds a preset material rate threshold, the oxygen in the furnace is reduced in time to adjust the furnace temperature. When the blast furnace material rate does not show the above situation, it is determined whether to adjust the coal amount to stabilize the furnace temperature, thereby realizing closed-loop control of the blast furnace temperature.

[0015] In an optional implementation, when the material speed does not exceed the preset material speed threshold at two consecutive time points in the preset smelting cycle, determining whether to adjust the coal quantity of the blast furnace includes: For any time point in the preset time period, when the predicted furnace temperature corresponding to the time point is within the preset furnace temperature range, the amount of coal in the blast furnace is not adjusted; or, When the predicted furnace temperature corresponding to a time point is not within the preset furnace temperature range, the coal injection flow rate is determined, and the coal quantity is adjusted based on the blast furnace coal quantity corresponding to the previous time point of the time point.

[0016] The closed-loop control method for blast furnace temperature provided by the embodiment of the present invention is based on the predicted furnace temperature at any time point. When the predicted furnace temperature at that time point is within the preset furnace temperature range, it is considered that the current coal quantity setting is appropriate and does not need to be adjusted, thereby maintaining the stability of the reaction in the furnace and avoiding unnecessary fluctuations. When the predicted furnace temperature is not within the preset furnace temperature range, the coal injection flow rate is determined and adjusted based on the blast furnace coal amount at the previous time point, thereby stabilizing the furnace temperature, realizing closed-loop control of the blast furnace temperature, and improving the stability of the production process.

[0017] In an optional implementation, when the predicted furnace temperature corresponding to the time point is not within the preset furnace temperature range, determining the coal injection flow rate includes: Obtain the batch coal weight, hourly batch number and silicon-coal ratio of the blast furnace; The coal injection flow rate is determined based on the predicted furnace temperature at a time point, the predicted furnace temperature at the next time point, the batch coal weight, the number of hourly batches and the silicon-coal ratio.

[0018] The closed-loop control method for blast furnace temperature provided by the embodiment of the present invention obtains the batch coal weight, hourly batch number and silicon-coal ratio of the blast furnace, combines the predicted furnace temperature at any time point and the next time point of that time point, ensures that the coal injection flow rate is jointly determined based on the latest production situation, thereby stabilizing the furnace temperature, realizing closed-loop control of the blast furnace temperature, and improving the stability of the production process.

[0019] In an optional embodiment, the method further includes: Determine the amount of blast furnace coal corresponding to a time point; When the blast furnace coal quantity corresponding to two consecutive time points in the preset time period is higher than the preset upper limit of the coal quantity and the predicted furnace temperature corresponding to two consecutive time points is lower than the preset lower limit of the furnace temperature, the blast furnace coal quantity will not be adjusted and the heat will be compensated by reducing the ore batch size.

[0020] The closed-loop control method for blast furnace temperature provided by an embodiment of the present invention can effectively avoid the problems caused by excessive adjustment by adopting a method of reducing ore batches to compensate for heat when the blast furnace coal quantity and predicted furnace temperature at two consecutive time points do not meet the requirements, while optimizing the heat balance in the furnace and improving production stability.

[0021] In an optional embodiment, the blast furnace data includes abnormal furnace condition data and supplementary heat data, and the method further includes: Based on the abnormal furnace condition data and supplementary heat data in the blast furnace data corresponding to each time point in the preset smelting cycle, blast furnace heat adjustment suggestions are generated.

[0022] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention monitors and analyzes occasional data in real time, thereby identifying situations that require adjustment, generating adjustment suggestions, and improving production stability.

[0023] In a second aspect, the present invention provides a blast furnace temperature closed-loop control device, the device comprising: An acquisition module is used to acquire blast furnace data corresponding to each time point in a preset smelting cycle, where the preset smelting cycle ends at the current time point and is divided into multiple time points according to a preset time step; An extraction module is used to extract features based on the blast furnace data corresponding to each time point using a variety of feature extraction methods to obtain a blast furnace data pool; A prediction module is used to use multiple prediction models to make predictions based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in a preset time period, and the preset time period is divided into multiple time points according to a preset time step; The first adjustment module is used to adjust the heat of the blast furnace based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period.

[0024] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the blast furnace temperature closed-loop control method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0025] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the blast furnace temperature closed-loop control method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0027] Figure 1 is a flow chart of a closed-loop control method for blast furnace temperature according to an embodiment of the present invention; Figure 2 is a structural block diagram of a blast furnace temperature closed-loop control device according to an embodiment of the present invention; Figure 3 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0029] By predicting the temperature trend of the blast furnace, control measures can be taken in advance to ensure the stability rate of the furnace temperature, thereby ensuring the stable and smooth operation of the blast furnace and realizing closed-loop control of the blast furnace temperature. At present, the operator predicts the trend of the furnace temperature by observing the various characterization indicators of the blast furnace, and intervenes by adjusting the amount of coal to stabilize the furnace temperature. However, the transitional reliance on the experience of the operator cannot meet the development requirements of the intelligentization of modern blast furnaces. The error between the predicted furnace temperature and the actual furnace temperature is large, and it is difficult to accurately realize the closed-loop control of the blast furnace temperature. The closed-loop control method of blast furnace temperature provided by the embodiment of the present invention improves the reliability and accuracy of furnace temperature prediction, adjusts the heat of the blast furnace based on the predicted furnace temperature, maintains the stability of the furnace temperature, realizes accurate closed-loop control of the blast furnace temperature, and improves the level of intelligence at the same time.

[0030] According to an embodiment of the present invention, an embodiment of a blast furnace temperature closed-loop control method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] In this embodiment, a closed-loop control method for blast furnace temperature is provided, which can be used for operating terminals such as computers, etc. Figure 1 : is a flow chart of a closed-loop control method for blast furnace temperature according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps: Step S101, obtain the blast furnace data corresponding to each time point in the preset smelting cycle, the preset smelting cycle takes the current time point as the end point, and is divided into multiple time points according to the preset time step. Specifically, the material on the furnace top needs to act on the furnace temperature after a period of time. Assuming that the period is 7 hours, the current time point is 17 o'clock, and the preset time step is 1 hour, the preset smelting cycle is 10 o'clock to 17 o'clock, and the corresponding time points are 10 o'clock, 11 o'clock, 12 o'clock, 13 o'clock, 14 o'clock, 15 o'clock, 16 o'clock and 17 o'clock, and the blast furnace data corresponding to the above 8 time points are obtained. The blast furnace data at any time point includes data of ten features, namely furnace temperature, material speed, slag iron information, hourly coal ratio, hourly fuel ratio, raw fuel quality, abnormal furnace condition data, supplementary heat data, technical parameters and operating parameters. Among them, the material speed includes 6 batches of material speed and 3 batches of material speed, based on the material running time of 6 consecutive batches of material and 3 consecutive batches of material, and considering the time acting on the furnace temperature, to assist in prejudging the furnace temperature. Taking the material speed of 6 batches of materials as an example, considering the preset smelting cycle at the same time, the signal of the probe ruler raising the ruler closest to the current time point and the 7th probe ruler raising signal before are determined, and the time corresponding to the two signals is subtracted as the material speed of the 6 batches of materials. Similarly, the material speed of 3 batches of materials can be obtained. For example, assuming that the probe ruler is raised once every hour for material discharge, if it is necessary to determine the material speed of 3 batches of materials at 14 o'clock in the preset smelting cycle from 10 o'clock to 17 o'clock, it is only necessary to determine the time difference between the probe ruler raising signal at 14 o'clock and the 3 probe ruler raising signals before 14 o'clock, that is, the time difference 3 between the probe ruler raising signal at 14 o'clock and the probe ruler raising signal at 11 o'clock, and the time difference 3 is used as the material speed of 3 batches of materials at 14 o'clock. The slag iron information includes the iron output of a single furnace, the slag output of a single furnace, the molten iron composition and the slag composition. Assuming that the current time point is 17:00, if the furnace temperature is predicted for the time points after 17:00, the slag and iron information corresponding to the three time points before 17:00 has a greater impact on the furnace temperature prediction, so only the slag and iron information from 14:00 to 17:00 can be obtained. The hourly coal ratio and the hourly fuel ratio are key indicators for predicting the furnace temperature. Since the hourly fuel ratio and the hourly coal ratio need to be calculated in combination with the batching results and the 6 batch material rates, the batching results need to be obtained, including the theoretical coke ratio, slag amount, coke batch, ore batch, theoretical basicity, theoretical iron amount, theoretical slag ratio, sintering ratio, pellet ratio and lump ore ratio. The hourly coal ratio at each time point can be calculated by the following formula (1), and the hourly fuel ratio at each time point can be calculated by the following formula (2). Since the coal can act on the furnace temperature after about 3 hours, only the hourly coal ratio from 14:00 to 17:00 can be obtained.

[0032] (1) In the formula, Indicates Hourly coal ratio at a time point; Indicates Hourly coal quantity at a time point; Indicates 6 batches of material speed at each time point; Indicates the theoretical amount of iron.

[0033] (2) In the formula, Indicates Hourly fuel ratio at a time point; Indicates Hourly coal ratio at a time point; Indicates the theoretical focal ratio.

[0034] The raw material quality in the above blast furnace data includes coke M40 crushing strength, coke CSR (Coke Strength after Reaction), sulfur content in coke, drum strength of sintered ore, ferrous oxide content in sintered ore, basicity of sintered ore, calorific value of pulverized coal, fixed carbon, moisture, etc. The abnormal furnace condition data in the above blast furnace data only have values ​​when abnormal conditions occur, and are null in other cases. Abnormal conditions include sliding material, hanging material, low material line, slag skin falling off, and top pressure peaking. Among them, when the material feeding speed is greater than the preset upper limit of the material feeding speed, it is considered that the material slip occurs, and the material feeding speed = (the height of the probe when the probe is lifted - the height of the probe when the probe floats) / (the time when the probe is lifted - the time when the probe floats); when the height of the probe when the probe is lifted is greater than the preset upper limit of the material line, it is considered that the material line is low; when the material speed of two consecutive batches of materials is less than the preset lower limit of the material speed, it is considered that the material is suspended; when the furnace temperature corresponding to two consecutive time points is less than the preset lower limit of the furnace temperature and the heat load is greater than the preset upper limit of the heat load, it is considered that the slag skin has fallen off; when the difference between the actual value of the blast furnace top pressure and the set value of the blast furnace top pressure is greater than the preset upper limit of the pressure deviation, it is considered that the top pressure peaks. For the supplementary heat data in the above blast furnace data, it is a null value under normal circumstances, and it is only occasionally supplemented when the heat is insufficient. It includes data such as net coke, moisture coke, and light materials. For the technical parameters in the above blast furnace data, including theoretical combustion temperature, blast kinetic energy, heat load, direct reduction carbon, and heat loss in the high temperature zone, since the above data corresponding to the three time points before the current time point have a greater impact on the furnace temperature prediction, only the technical parameters from 14:00 to 17:00 can be obtained. For the operating parameters in the above blast furnace data, including the hourly average and standard deviation of blast furnace coal volume, air supply volume, cooling wall temperature, static pressure, cross temperature measurement and other data, the action time is also considered to be 3 hours, so only the operating parameters from 14:00 to 17:00 can be obtained.

[0035] It should be noted that the blast furnace data in step S101 can be stored in SQL software, and the blast furnace data can be matched with each time point in the preset smelting cycle according to the Cartesian product to form a data block.

[0036] Step S102, using a variety of feature extraction methods, based on the blast furnace data corresponding to each time point to extract features, to obtain a blast furnace data pool. Specifically, from the blast furnace data corresponding to 10:00 to 17:00, various features useful for predicting furnace temperature can be extracted to form a blast furnace data pool, which can better reflect the working status of the blast furnace.

[0037] Step S103, using multiple prediction models, based on the blast furnace data pool to predict, to obtain the predicted furnace temperature corresponding to each time point in the preset time period, the preset time period is divided into multiple time points according to the preset time step. Specifically, assuming that the preset smelting cycle is from 10:00 to 17:00, the preset time step is 1 hour, and the preset time period can be 18:00, 19:00 and 20:00. Considering that different prediction models have different advantages, multiple prediction models are used for prediction to obtain the predicted furnace temperatures at 18:00, 19:00 and 20:00, which can improve the accuracy of furnace temperature prediction.

[0038] Step S104, based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period, the blast furnace is heat-adjusted. Specifically, based on the historical blast furnace data and the predicted furnace temperature at a future time point, the blast furnace is heat-adjusted to maintain the furnace temperature, thereby realizing closed-loop control of the blast furnace temperature.

[0039] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention provides a rich historical data basis by acquiring blast furnace data at all time points in a preset smelting cycle, and adopts multiple feature extraction methods to obtain a blast furnace data pool based on the blast furnace data corresponding to each time point. Various features can be extracted from the original blast furnace data, which can better reflect the working status of the blast furnace. Taking into account that different prediction models have their own advantages in data processing, using multiple prediction models for prediction can effectively reduce the prediction error of a single model, improve the reliability and accuracy of furnace temperature prediction, adjust the heat of the blast furnace based on the predicted furnace temperature, maintain the stability of the furnace temperature, and realize accurate closed-loop control of the blast furnace temperature. Compared with relying on the experience of operators to predict and control the furnace temperature, the level of intelligence is improved.

[0040] In this embodiment, a blast furnace temperature closed-loop control method is provided, which can be used for the above-mentioned operation terminal. The method specifically includes the following steps: Step S201, obtain the blast furnace data corresponding to each time point in the preset smelting cycle, the preset smelting cycle ends at the current time point and is divided into multiple time points according to the preset time step. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0041] Step S202, using a variety of feature extraction methods, extracting features based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool.

[0042] Specifically, the above step S202 includes: Step S2021, performing correlation analysis on the blast furnace data corresponding to each time point to obtain the first blast furnace data pool. Specifically, for the blast furnace data corresponding to any time point, referring to step S101, the blast furnace data includes data of ten characteristics, and the correlation coefficient between the data of different characteristics and the furnace temperature is calculated, such as the Pearson correlation coefficient or the Spearman correlation coefficient. Assuming that the absolute value of the correlation coefficient between the material speed and the furnace temperature in the blast furnace data at any time point is greater than a certain threshold, the material speed data is retained. By performing correlation analysis on the data of all characteristics in the blast furnace data corresponding to each time point, the first blast furnace data pool highly correlated with the predicted furnace temperature is screened.

[0043] Step S2022, performing variance analysis on the blast furnace data corresponding to each time point to obtain a second blast furnace data pool. Specifically, for each feature of the blast furnace data, the variance of the feature data at all time points of the preset smelting cycle is calculated. If the variance is greater than a preset threshold, the feature data is retained to obtain a second blast furnace data pool. By calculating the variance of each feature of the blast furnace data, the degree of change at different time points can be measured.

[0044] Step S2023, using a recursive feature elimination method, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain a third blast furnace data pool. Specifically, the recursive feature elimination method is a feature selection method that gradually reduces the number of features by repeatedly building a model and removing the least important features until a preset number is reached. First, determine the number of features to be retained, such as retaining the 5 most important features. Then, merge the blast furnace data at all time points into a data set. Then, select a suitable machine learning model as a basic model, such as linear regression, support vector machine, etc. Then, perform the first round of fitting, use all the features in the data set to train the basic model, and score the importance of the basic model, such as furnace temperature: 0.8, material speed: 0.7, slag iron information: 0.6, hourly coal ratio: 0.5, hourly fuel ratio: 0.4, raw fuel quality: 0.3, abnormal furnace condition data: 0.2, supplementary heat data: 0.1, technical parameters: 0.05, operating parameters: 0.01, and determine the least important features according to the score, that is, operating parameters: 0.01. The operating parameter feature is removed, and the model is retrained with the remaining 9 features in the updated data set. The new least important feature is determined again based on the importance score of the model. The above process is repeated until there are 5 features in the data set, and the third blast furnace data pool is obtained. Through the recursive feature elimination method, the least important features in the blast furnace data can be gradually removed to obtain features that are highly correlated with the predicted furnace temperature.

[0045] Step S2024, using a decision tree regression model, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain a fourth blast furnace data pool. Specifically, the blast furnace data at all time points are input into the decision tree regression model to obtain the importance values ​​corresponding to the ten features of the blast furnace data, which are used to indicate the contribution of each feature to the furnace temperature. The importance values ​​corresponding to the ten features are sorted in descending order, assuming that the top five most important features are selected as the fourth blast furnace data pool. By using a decision tree regression model, features highly correlated with the predicted furnace temperature can be effectively screened.

[0046] Step S2025, taking the union of the first blast furnace data pool, the second blast furnace data pool, the third blast furnace data pool and the fourth blast furnace data pool to obtain a blast furnace data pool. Specifically, all features in the four blast furnace data pools are merged together to form a comprehensive blast furnace data pool, ensuring that features extracted from different angles and methods are included, and providing a more comprehensive and accurate input for subsequent prediction models, so that the furnace temperature can be accurately predicted based on the blast furnace data pool. Optionally, during the merging process, if some features exist in multiple data pools, they only need to be retained once to avoid duplication.

[0047] In some optional implementations, after the blast furnace data pool is obtained in step S202, the operator can add features to the blast furnace data pool according to his own experience. For example, the blast furnace data pool does not include the feature of raw fuel quality, but the operator believes that this feature has a greater impact on the furnace temperature, and the feature can be manually added to the blast furnace data pool.

[0048] Step S203, using multiple prediction models, predicting based on the blast furnace data pool, and obtaining the predicted furnace temperature corresponding to each time point in a preset time period, where the preset time period is divided into multiple time points according to a preset time step.

[0049] Specifically, the above step S203 includes: Step S2031, based on the blast furnace data corresponding to each time point in the preset smelting cycle, the decision tree random forest model, the CatBoost gradient boosting model, the XGBoost regression model and the neural network model are pre-trained respectively to obtain the optimal prediction model.

[0050] In some optional implementations, the above step S2031 includes: Step a1, using decision tree random forest model, CatBoost gradient boosting model, XGBoost regression model and neural network model, predict the blast furnace data corresponding to any time point except the current time point and the starting time point in the preset smelting cycle, and obtain the predicted furnace temperature corresponding to each model at the time point. Specifically, assuming that the preset smelting cycle is 10 to 17 o'clock, based on the blast furnace data from 10 to 17 o'clock, for any time point in the other six time points except 10 and 17 o'clock, the above four prediction models are used to predict respectively, and the predicted furnace temperature is obtained. Assuming that the furnace temperature at 12 o'clock is predicted, the blast furnace data at 10 and 11 o'clock are input to the above four models for prediction, and 4 predicted furnace temperatures predicted by the four models at 12 o'clock are obtained. It should be noted that when furnace temperature prediction is performed at any time point, the blast furnace data at the time point before the time point is input into the model, and the blast furnace data at the time point after the time point is not used. The above four prediction models are all existing models, and the prediction process is not repeated here.

[0051] Step a2, for any model, based on the furnace temperature in the blast furnace data at the time point and the predicted furnace temperature corresponding to the model at the time point, determine the prediction error of the model. Specifically, assume that the furnace temperature is predicted at 14:00, 15:00 and 16:00 through step a1. For any model, assume that the model is a neural network model, and the furnace temperatures predicted by the neural network model at 14:00, 15:00 and 16:00 are 2000℃, 2100℃ and 2200℃ respectively. The furnace temperatures in the blast furnace data corresponding to 14:00, 15:00 and 16:00 are 2050℃, 20905 and 2210℃ respectively. The average value of the difference between the predicted furnace temperature and the actual furnace temperature at the three time points is taken as the prediction error of the neural network model.

[0052] Step a3, the model with the smallest prediction error among multiple models is determined as the optimal prediction model. Specifically, the prediction errors of the four prediction models can be determined respectively through the above step a2, and the prediction model corresponding to the minimum prediction error is used as the optimal prediction model, that is, the model shows the best prediction performance on the blast furnace data input to the model, which can significantly improve the accuracy of furnace temperature prediction and reduce prediction error. It should be noted that since these four models have different prediction advantages on different data, the optimal prediction model is not fixed, and it is necessary to determine the optimal prediction model applicable to the current blast furnace data through step S2031 before each furnace temperature prediction, so as to fit the operation of the current blast furnace and improve the accuracy of furnace temperature prediction.

[0053] Step S2032, taking the furnace temperature in the blast furnace data at the current time point as the initial furnace temperature, using the optimal prediction model to predict the blast furnace data pool, and obtaining the predicted furnace temperature at the next time point of the current time point. Specifically, assuming that the current time point is 17:00, taking the furnace temperature of 2000°C in the blast furnace data at 17:00 as the initial furnace temperature, using the optimal prediction model obtained in the above step S2031 to predict the furnace temperature at 18:00.

[0054] Step S2033, taking the predicted furnace temperature at the next time point of the current time point as the new initial furnace temperature, repeating the above process of predicting the furnace temperature until the preset time period is completed, and obtaining the predicted furnace temperature corresponding to each time point in the preset time period. Specifically, the preset time period is generally 3 hours. When the current time point is 17:00, each time point in the preset time period is 18:00, 19:00 and 20:00. Taking the predicted furnace temperature at 18:00 as the initial furnace temperature for the next prediction, repeating step S2032, obtaining the predicted furnace temperature at 19:00, thereby obtaining the predicted furnace temperature corresponding to each time point in the preset time period. By pre-training based on multiple prediction models, the optimal prediction model with the best prediction performance is used to predict the furnace temperature, thereby improving the accuracy of the furnace temperature prediction and reducing the prediction error. By recursive prediction at each time point, the changes in the production process can be reflected in real time, and dynamic furnace temperature prediction results can be provided.

[0055] Step S204, adjusting the heat of the blast furnace based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period.

[0056] Specifically, the above step S204 includes: Step S2041, when the material speed at two consecutive time points in the preset smelting cycle is greater than the preset material speed threshold, reduce the oxygen in the blast furnace. Specifically, in the daily closed-loop control process of the blast furnace temperature, the material speed may be too fast. If the operator still adopts the method of adding coal, the furnace temperature does not change significantly. Therefore, the material speed can be judged. If the material speed at two consecutive time points in the preset smelting cycle is greater than the preset material speed threshold, the furnace temperature is no longer adjusted by coal, but the furnace temperature is stabilized by reducing the oxygen in the furnace. Optionally, the material speed can adopt a 6-batch material speed or a 3-batch material speed.

[0057] Alternatively, in step S2042, if the material speed does not exceed the preset material speed threshold at two consecutive time points in the preset smelting cycle, it is determined whether to adjust the coal quantity of the blast furnace.

[0058] In some optional implementations, the above step S2042 includes: Step b1, for any time point in the preset time period, if the predicted furnace temperature corresponding to the time point is within the preset furnace temperature range, the amount of blast furnace coal is not adjusted. Specifically, if there are no two consecutive time points in the preset smelting cycle where the material speed is greater than the preset material speed threshold, the amount of coal continues to be used to adjust the furnace temperature. At this time, the predicted furnace temperature at any time point in the preset time period can be judged. If the predicted furnace temperature is within the preset furnace temperature range, the furnace temperature is stable at this time and no furnace temperature adjustment is required.

[0059] Alternatively, in step b2, when the predicted furnace temperature corresponding to the time point is not within the preset furnace temperature range, the coal injection flow rate is determined, and the coal quantity is adjusted based on the blast furnace coal quantity corresponding to the previous time point of the time point.

[0060] In some optional implementations, the above step b2 includes: Step c1, obtain the batch coal weight, hourly batch number and silicon-coal ratio of the blast furnace. Specifically, if the predicted furnace temperature is not within the preset furnace temperature range, it is necessary to stabilize the furnace temperature by adjusting the amount of coal based on the blast furnace coal amount at the previous time point at this time point. First, obtain the batch coal weight, hourly batch number and silicon-coal ratio. Among them, the batch coal weight indicates the amount of coal added each time the material is discharged, the hourly batch number indicates the number of times the material is added per hour, and the silicon-coal ratio indicates the proportional relationship between the silicon content and the amount of coal.

[0061] Step c2, based on the predicted furnace temperature at the time point, the predicted furnace temperature at the next time point, the batch coal weight, the number of hourly batches, and the silicon-coal ratio, the coal injection flow rate is determined. Specifically, the coal injection flow rate at any time point is determined by the following formula (3), so as to adjust the current coal amount to achieve the predicted furnace temperature at the next time point after the time point.

[0062] (3) In the formula, Indicates Coal injection flow rate at a time point; Indicates The predicted furnace temperature at the next time point from the previous time point; Indicates Predicted furnace temperature at a time point; It represents the silicon-coal ratio, which is usually 4, meaning 0.1% silicon corresponds to 4kg coal; Indicates the weight of the batch of iron; Indicates the number of hourly material batches.

[0063] Step S2043, determine the blast furnace coal quantity corresponding to the time point. Specifically, assuming that the time point is 18:00 in the preset time period, after determining whether to adjust the coal quantity based on the predicted furnace temperature at 18:00, if the coal quantity does not need to be adjusted, obtain the blast furnace coal quantity at 18:00, if the coal quantity needs to be adjusted, calculate the corresponding coal injection flow rate, and the operator adds coal powder to the blast furnace based on the blast furnace coal quantity at 17:00 according to the value of the coal injection flow rate, and then determine the blast furnace coal quantity after adding the coal powder.

[0064] Step S2044, when the blast furnace coal quantity corresponding to two consecutive time points in the preset time period is higher than the preset coal quantity upper limit and the predicted furnace temperature corresponding to two consecutive time points is lower than the preset furnace temperature lower limit, the blast furnace coal quantity is not adjusted, and the shrinking ore batch is adopted to compensate for the heat. Specifically, due to the conflict between the furnace temperature adjustment and the economic index in the daily blast furnace temperature closed-loop control process, when the coal quantity is sufficient and the furnace temperature rise is not obvious, the operator will not choose to de-load and maintain the furnace temperature, because de-load means that the coke ratio rises and the cost rises. Therefore, the blast furnace coal quantity at each time point in the preset time period is continuously determined by the above-mentioned step S2043. When the blast furnace coal quantity at two consecutive time points in the preset time period is higher than the preset coal quantity upper limit, and the predicted furnace temperature at two consecutive time points is lower than the preset furnace temperature lower limit, the coal quantity is no longer adjusted, and the shrinking ore batch is adopted to compensate for the heat, that is, by changing the addition of fuel and ore, etc., the furnace temperature is stabilized, the closed-loop control of the blast furnace temperature is realized, and the stability of the production process is improved.

[0065] Step S205, based on the abnormal furnace condition data and supplementary heat data in the blast furnace data corresponding to each time point in the preset smelting cycle, generate a blast furnace heat adjustment suggestion. Specifically, since the abnormal furnace condition data and supplementary heat data in the blast furnace data are sporadic data, they are usually eliminated when the feature extraction is performed in step S202, and cannot participate in the subsequent furnace temperature prediction process. Therefore, the abnormal furnace condition data and supplementary heat data can be analyzed separately to generate corresponding heat adjustment suggestions. When low material line, slag skin falling off and / or top pressure peaking occur, the furnace temperature is no longer adjusted by adjusting the amount of coal, but net coke is added to supplement heat to stabilize the furnace temperature. When hanging material occurs, the furnace temperature is no longer adjusted by adjusting the amount of coal, and the air supply volume is reduced by 1000m3 / min. Optionally, the air supply volume reduction value is only an example and is not limited. By real-time monitoring and analysis of sporadic data, the situation that needs to be adjusted is identified, and adjustment suggestions are generated to improve production stability.

[0066] In some optional implementations, the predicted furnace temperature, actual furnace temperature and blast furnace coal quantity at each time point obtained through the above steps can be displayed in real time in the application software in the operation terminal. At the same time, the blast furnace data at each time point and the change range of each feature in the blast furnace data are displayed in real time, and abnormal furnace condition data and compensation heat data are timely warned, and the blast furnace heat adjustment suggestions generated based on the abnormal furnace condition data and compensation heat data are displayed and reminded.

[0067] The blast furnace temperature closed-loop control method provided by the embodiment of the present invention provides a rich historical data basis by acquiring blast furnace data at all time points in a preset smelting cycle, and adopts multiple feature extraction methods to obtain a blast furnace data pool based on the blast furnace data corresponding to each time point. Various features can be extracted from the original blast furnace data, which can better reflect the working status of the blast furnace. Taking into account that different prediction models have their own advantages in data processing, using multiple prediction models for prediction can effectively reduce the prediction error of a single model, improve the reliability and accuracy of furnace temperature prediction, adjust the heat of the blast furnace based on the predicted furnace temperature, maintain the stability of the furnace temperature, and realize accurate closed-loop control of the blast furnace temperature. Compared with relying on the experience of operators to predict and control the furnace temperature, the level of intelligence is improved.

[0068] In this embodiment, a blast furnace temperature closed-loop control device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0069] This embodiment provides a blast furnace temperature closed-loop control device, such as Figure 2 As shown, including: The acquisition module 201 is used to acquire the blast furnace data corresponding to each time point in a preset smelting cycle. The preset smelting cycle ends at the current time point and is divided into multiple time points according to a preset time step.

[0070] The extraction module 202 is used to use a variety of feature extraction methods to perform feature extraction based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool.

[0071] The prediction module 203 is used to use multiple prediction models to make predictions based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in a preset time period, and the preset time period is divided into multiple time points according to a preset time step.

[0072] The first adjustment module 204 is used to adjust the heat of the blast furnace based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period.

[0073] In some optional implementations, the extraction module 202 includes: The first extraction unit is used to perform correlation analysis on the blast furnace data corresponding to each time point to obtain a first blast furnace data pool.

[0074] The second extraction unit is used to perform variance analysis on the blast furnace data corresponding to each time point to obtain a second blast furnace data pool.

[0075] The third extraction unit is used to adopt a recursive feature elimination method to perform feature extraction based on the blast furnace data corresponding to each time point to obtain a third blast furnace data pool.

[0076] The fourth extraction unit is used to use a decision tree regression model to perform feature extraction based on the blast furnace data corresponding to each time point to obtain a fourth blast furnace data pool.

[0077] The merging unit is used to take the union of the first blast furnace data pool, the second blast furnace data pool, the third blast furnace data pool and the fourth blast furnace data pool to obtain a blast furnace data pool.

[0078] In some optional implementations, the prediction module 203 includes: The pre-training unit is used to pre-train the decision tree random forest model, CatBoost gradient boosting model, XGBoost regression model and neural network model respectively based on the blast furnace data corresponding to each time point in the preset smelting cycle to obtain the optimal prediction model.

[0079] The first prediction unit is used to use the furnace temperature in the blast furnace data at a current time point as the initial furnace temperature, and use the optimal prediction model to predict the blast furnace data pool to obtain the predicted furnace temperature at the next time point after the current time point.

[0080] The second prediction unit is used to use the predicted furnace temperature at the next time point after the current time point as the new initial furnace temperature, repeat the above process of predicting the furnace temperature until the preset time period is completed, and obtain the predicted furnace temperature corresponding to each time point in the preset time period.

[0081] In some optional embodiments, the pre-training unit includes: The prediction subunit is used to use the decision tree random forest model, CatBoost gradient boosting model, XGBoost regression model and neural network model to predict the blast furnace data corresponding to any time point in the preset smelting cycle except the current time point and the starting time point, and obtain the predicted furnace temperature corresponding to each model at the time point.

[0082] The first determination subunit is used to determine the prediction error of the model for any model based on the furnace temperature in the blast furnace data at a time point and the predicted furnace temperature corresponding to the model at the time point.

[0083] The second determination subunit is used to determine the model with the smallest prediction error among the multiple models as the optimal prediction model.

[0084] In some optional embodiments, the blast furnace data includes material speed, and the first adjustment module 204 includes: The first adjustment unit is used to reduce the oxygen in the blast furnace when the material speed at two consecutive time points in a preset smelting cycle is greater than a preset material speed threshold.

[0085] Alternatively, the second adjustment unit is used to determine whether to adjust the coal quantity of the blast furnace when the material speed does not exceed a preset material speed threshold at two consecutive time points in a preset smelting cycle.

[0086] In some optional implementations, the second adjustment unit includes: The first adjustment subunit is used for not adjusting the blast furnace coal amount at any time point in the preset time period when the predicted furnace temperature corresponding to the time point is within the preset furnace temperature range.

[0087] Alternatively, the second adjustment subunit is used to determine the coal injection flow rate when the predicted furnace temperature corresponding to the time point is not within the preset furnace temperature range, and adjust the coal quantity based on the blast furnace coal quantity corresponding to the previous time point of the time point.

[0088] In some optional implementations, the second adjustment subunit includes: Acquisition component, used to obtain the batch coal weight, hourly batch number and silicon-coal ratio of the blast furnace.

[0089] A determination component is used to determine the coal injection flow rate based on the predicted furnace temperature at a time point, the predicted furnace temperature at the next time point, the batch coal weight, the number of hourly batches, and the silicon-coal ratio.

[0090] In some optional embodiments, the device further comprises: The determination module is used to determine the amount of blast furnace coal corresponding to a time point.

[0091] The second adjustment module is used to not adjust the blast furnace coal quantity and to compensate heat by reducing the ore batch when the blast furnace coal quantity corresponding to two consecutive time points in a preset time period is higher than the preset coal quantity upper limit and the predicted furnace temperature corresponding to two consecutive time points is lower than the preset furnace temperature lower limit.

[0092] In some optional embodiments, the blast furnace data includes abnormal furnace condition data and supplementary heat data, and the device further includes: A generation module is used to generate blast furnace heat adjustment suggestions based on abnormal furnace condition data and supplementary heat data in the blast furnace data corresponding to each time point in a preset smelting cycle.

[0093] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0094] The blast furnace temperature closed-loop control device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0095] The embodiment of the present invention also provides a computer device having the above Figure 2 The blast furnace temperature closed-loop control device shown.

[0096] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 3 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 3 A processor 10 is taken as an example.

[0097] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0098] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0099] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0100] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0101] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0102] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device may be a touch screen.

[0103] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0104] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0105] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the present invention.

Claims

1. A blast furnace temperature closed-loop control method, characterized in that: The method comprises: Acquire blast furnace data corresponding to each time point in a preset smelting cycle, wherein the preset smelting cycle ends at the current time point and is divided into multiple time points according to a preset time step; Using a variety of feature extraction methods, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool; Using multiple prediction models, prediction is performed based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in a preset time period, wherein the preset time period is divided into multiple time points according to a preset time step; Based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period, the heat of the blast furnace is adjusted.

2. The method according to claim 1, characterized in that The method adopts a plurality of feature extraction methods to extract features based on the blast furnace data corresponding to each time point to obtain a blast furnace data pool, including: Perform correlation analysis on the blast furnace data corresponding to each time point to obtain the first blast furnace data pool; Perform variance analysis on the blast furnace data corresponding to each time point to obtain the second blast furnace data pool; Using the recursive feature elimination method, feature extraction is performed based on the blast furnace data corresponding to each time point to obtain the third blast furnace data pool; A decision tree regression model is used to extract features based on the blast furnace data corresponding to each time point to obtain the fourth blast furnace data pool; The first blast furnace data pool, the second blast furnace data pool, the third blast furnace data pool and the fourth blast furnace data pool are taken as a union to obtain a blast furnace data pool.

3. The method according to claim 1, characterized in that The method adopts multiple prediction models to predict based on the blast furnace data pool to obtain the predicted furnace temperature corresponding to each time point in the preset time period, including: Based on the blast furnace data corresponding to each time point in the preset smelting cycle, the decision tree random forest model, the CatBoost gradient boosting model, the XGBoost regression model and the neural network model are pre-trained to obtain the optimal prediction model; The furnace temperature in the blast furnace data at the current time point is used as the initial furnace temperature, and the blast furnace data pool is predicted using the optimal prediction model to obtain the predicted furnace temperature at the next time point after the current time point; The predicted furnace temperature at the next time point after the current time point is used as the new initial furnace temperature, and the above process of predicting the furnace temperature is repeated until the preset time period is completed to obtain the predicted furnace temperature corresponding to each time point in the preset time period.

4. The method according to claim 3, characterized in that Based on the blast furnace data corresponding to each time point in the preset smelting cycle, the decision tree random forest model, the CatBoost gradient boosting model, the XGBoost regression model and the neural network model are pre-trained respectively to obtain the optimal prediction model, including: A decision tree random forest model, a CatBoost gradient boosting model, an XGBoost regression model, and a neural network model are used to predict the blast furnace data corresponding to any time point in the preset smelting cycle except the current time point and the starting time point, and the predicted furnace temperature corresponding to each model at the time point is obtained; For any model, based on the furnace temperature in the blast furnace data at the time point and the predicted furnace temperature corresponding to the model at the time point, determining the prediction error of the model; The model with the smallest prediction error among multiple models is determined as the optimal prediction model.

5. The method according to claim 3, characterized in that: The blast furnace data includes a material rate, and heat adjustment of the blast furnace is performed based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period, including: When the material speed at two consecutive time points in the preset smelting cycle is greater than the preset material speed threshold, reducing the oxygen in the blast furnace; or, In the case that the material speed does not exceed the preset material speed threshold at two consecutive time points in the preset smelting cycle, it is determined whether to adjust the coal quantity of the blast furnace.

6. The method according to claim 5, characterized in that In the case that the material speed does not exceed the preset material speed threshold at two consecutive time points in the preset smelting cycle, determining whether to adjust the coal quantity of the blast furnace includes: For any time point in the preset time period, if the predicted furnace temperature corresponding to the time point is within the preset furnace temperature range, the coal quantity of the blast furnace is not adjusted; or, When the predicted furnace temperature corresponding to the time point is not within the preset furnace temperature range, the coal injection flow rate is determined, and the coal quantity is adjusted based on the blast furnace coal quantity corresponding to the previous time point of the time point.

7. The method according to claim 6, characterized in that When the predicted furnace temperature corresponding to the time point is not within the preset furnace temperature range, determining the coal injection flow rate includes: Obtain the batch coal weight, hourly batch number and silicon-coal ratio of the blast furnace; The coal injection flow rate is determined based on the predicted furnace temperature at the time point, the predicted furnace temperature at the next time point, the batch coal weight, the number of hourly batches, and the silicon-coal ratio.

8. The method according to claim 6, characterized in that The method further comprises: Determine the amount of blast furnace coal corresponding to the time point; When the blast furnace coal quantity corresponding to two consecutive time points in the preset time period is higher than the preset upper limit of the coal quantity and the predicted furnace temperature corresponding to two consecutive time points is lower than the preset lower limit of the furnace temperature, the blast furnace coal quantity is not adjusted and the heat is compensated by reducing the ore batch.

9. The method according to claim 1, characterized in that: The blast furnace data includes abnormal furnace condition data and supplementary heat data, and the method further includes: Based on the abnormal furnace condition data and supplementary heat data in the blast furnace data corresponding to each time point in the preset smelting cycle, a blast furnace heat adjustment suggestion is generated.

10. A blast furnace temperature closed-loop control device, characterized in that: The device comprises: An acquisition module, used to acquire blast furnace data corresponding to each time point in a preset smelting cycle, wherein the preset smelting cycle ends at the current time point and is divided into multiple time points according to a preset time step; An extraction module is used to extract features based on the blast furnace data corresponding to each time point using a variety of feature extraction methods to obtain a blast furnace data pool; A prediction module, for using a plurality of prediction models to make predictions based on the blast furnace data pool to obtain a predicted furnace temperature corresponding to each time point in a preset time period, wherein the preset time period is divided into a plurality of time points according to a preset time step; The first adjustment module is used to adjust the heat of the blast furnace based on the blast furnace data corresponding to each time point in the preset smelting cycle and the predicted furnace temperature corresponding to each time point in the preset time period.

Images