A method and system for evaluating and predicting blast furnace hearth activity based on big data
Through the blast furnace furnace activity evaluation and prediction method based on big data, the blast furnace furnace activity evaluation and insufficient prediction in the existing technology is solved, and efficient blast furnace furnace activity evaluation and accurate prediction are achieved. Blast furnace production provides stable data support and operational advice.
Patent Information
- Application Number
- CN202210412850.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-19
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2042-04-19
AI Technical Summary
The prior art has problems such as difficulty in implementing the activity evaluation of blast furnace cylinders, low data utilization, inability to predict the state of the furnace cylinder in advance, and feedback on blast furnace operation suggestions.
Using the blast furnace furnace activity evaluation and prediction method based on big data, the blast furnace furnace activity evaluation model is established by collecting and preprocessing the blast furnace production historical data, and establishing a prediction model through deep learning to achieve the activity of blast furnace furnace cylinder. Accurate evaluation and accurate prediction.
Accurate evaluation and accurate prediction of blast furnace cylinder activity are achieved, providing guarantees for the high yield, low consumption and forward direction of blast furnace, avoiding waste of data resources, and improving the quality of blast furnace production data.
Smart Images

Figure CN114819587B_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of data processing technology, and specifically relates to a method and system for evaluating and predicting the activity of a blast furnace hearth based on big data. Background Art
[0002] The state of the blast furnace hearth has an important impact on the "high quality, low consumption, high yield, long life and high efficiency" of blast furnace production. For example: high-quality molten iron depends on the full coupling reaction between slag and iron in the hearth; injecting fuel to reduce the consumption of coke will increase the "skeleton" load in the hearth; blast furnace operators expect stable and efficient production, which requires the blast furnace hearth to provide sufficient, stable, and uniformly distributed heat and reducing gas along the radial and circumferential directions of the hearth, and a large amount of water must be able to be discharged smoothly from the hearth; pig iron production must be competitive, and the life of the blast furnace must be extended. The hearth is the most difficult area to maintain in the blast furnace body, and it is also the key part that determines the life of the blast furnace. Slag and iron eventually converge in the hearth, and the activity of the hearth is the focus of blast furnace production. In other words, the active state of the blast furnace hearth is the basis for the stable and smooth operation of the blast furnace, and quantifying the activity of the blast furnace hearth is of great significance in guiding blast furnace production operations.
[0003] At present, the evaluation method of blast furnace hearth activity based on slag iron flow resistance coefficient has certain implementation difficulties and limitations, such as difficulty in sampling coke column in the hearth, inability to directly measure related parameters, many empirical coefficients of the model, poor correlation with blast furnace operation and blast furnace conditions, etc. The activity of blast furnace hearth is characterized by complexity, variability, and nonlinearity. Its influencing factors are intricate and have time lag. The evaluation method of blast furnace hearth activity based on temperature data has low data utilization and fails to fully mine valuable information in blast furnace production data, resulting in a waste of data resources. Although it can evaluate the current hearth activity status, it cannot predict the change of hearth status in advance and feedback blast furnace operation suggestions. In addition, the algorithm and mechanism are not well integrated, and the verification is insufficient, which is somewhat different from the actual production situation of the blast furnace.
[0004] In view of this, a blast furnace hearth activity evaluation and prediction method and system based on big data are provided, which realizes accurate evaluation and precise prediction of blast furnace hearth activity, and provides guarantee for high production, low consumption and smooth operation of the blast furnace. Summary of the invention
[0005] 1. Technical issues to be resolved
[0006] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides a blast furnace hearth activity evaluation and prediction method and system based on big data.
[0007] (II) Technical solution
[0008] In order to achieve the above objectives, this application adopts the following technical solutions:
[0009] In a first aspect, the present application provides a method for evaluating and predicting the activity of a blast furnace hearth based on big data, comprising:
[0010] S1. Collect blast furnace production history data, and manage the blast furnace production history data through data preprocessing to obtain blast furnace hearth activity application data;
[0011] S2, determining a calculation method for the activity of a blast furnace hearth based on the blast furnace hearth activity application data, establishing a blast furnace hearth activity evaluation model by integrating the calculation method, and performing a preliminary evaluation of the blast furnace hearth activity by using the evaluation model;
[0012] S3, obtaining a blast furnace condition stability index based on the blast furnace hearth activity application data, revising a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and finally evaluating and grading the blast furnace hearth activity based on the revised blast furnace hearth activity evaluation model;
[0013] S4. Obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by performing deep learning on the influencing parameters, and predict the blast furnace hearth activity through the prediction model.
[0014] Optionally, also include:
[0015] S5. Acquire real-time data of blast furnace production, learn and train a blast furnace hearth activity prediction model based on the real-time data of blast furnace production, and update the neural network hyperparameters and influencing parameters of the prediction model through self-learning.
[0016] Optionally, S1 includes:
[0017] The blast furnace historical data includes blast furnace raw fuel data, blast furnace operation data, blast furnace smelting status data and slag iron discharge data;
[0018] The preprocessing includes data missing value processing, outlier processing, data frequency alignment and time lag processing.
[0019] Optionally, S2 includes:
[0020] S2-1. Based on the blast furnace hearth activity application data, obtain the correlation between the calculation method of each blast furnace hearth activity and the furnace condition parameter, and determine the calculation method of the blast furnace hearth activity based on the correlation;
[0021] The furnace condition parameters include coke ratio, air permeability, blast kinetic energy, gas utilization rate, and heat load;
[0022] S2-2, modifying the coefficients of the determined calculation method through machine learning, and establishing a blast furnace hearth activity evaluation model by integrating the modified calculation method;
[0023] S2-3. Obtain a blast furnace hearth activity index based on a blast furnace hearth activity evaluation model, and conduct a preliminary evaluation of the blast furnace hearth activity using the blast furnace hearth activity index.
[0024] Optionally, S3 includes:
[0025] S3-1. Obtain blast furnace condition parameters based on blast furnace hearth activity application data, and obtain blast furnace condition stability index by integrating blast furnace condition parameters;
[0026] The blast furnace condition parameters include blast furnace economic index, blast furnace forward index, and furnace erosion index;
[0027] S3-2, based on the blast furnace condition stability index, checking the accuracy of the blast furnace hearth activity evaluation model, obtaining the test result, and correcting the blast furnace hearth activity evaluation model based on the test result;
[0028] S3-3. Based on the revised blast furnace hearth activity evaluation model, the final blast furnace hearth activity index is obtained, and the blast furnace hearth activity is graded based on the blast furnace condition stability index and the final blast furnace hearth activity index.
[0029] Optionally, S3-3 includes:
[0030] If the blast furnace condition stability index and the final blast furnace hearth activity index are both higher than the corresponding preset values, the blast furnace hearth activity is at the first level;
[0031] If the blast furnace condition stability index is lower than the preset value and the final blast furnace hearth activity index is higher than the preset value, or the blast furnace condition stability index is higher than the preset value and the final blast furnace hearth activity index is lower than the preset value, the blast furnace hearth activity is at the second level;
[0032] If the blast furnace condition stability index and the final blast furnace hearth activity index are both lower than the corresponding preset values, the blast furnace hearth activity is at the third level.
[0033] Optionally, S4 includes:
[0034] Based on the acquired influencing parameters, a prediction model for blast furnace hearth activity is established by determining the neural network hyperparameters;
[0035] The neural network hyperparameters include the number of hidden layers, the number of neuron nodes in each layer, the learning rate, the activation function, the dropout rate and the number of iterations.
[0036] In a second aspect, the present application provides a blast furnace hearth activity evaluation and prediction system based on big data, comprising:
[0037] The data management module is specifically used to collect the blast furnace production history data, and process the blast furnace production history data through data preprocessing to obtain the blast furnace hearth activity application data;
[0038] A hearth activity evaluation module, specifically used to determine a calculation method for blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity evaluation model by integrating the calculation method, and perform a preliminary evaluation of the blast furnace hearth activity through the evaluation model;
[0039] A hearth activity correction module, specifically used to obtain a blast furnace condition stability index based on blast furnace hearth activity application data, correct a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and perform a final evaluation and grade classification of the blast furnace hearth activity based on the corrected blast furnace hearth activity evaluation model;
[0040] A hearth activity prediction module, specifically used to obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by deep learning the influencing parameters, and predict the blast furnace hearth activity through the prediction model;
[0041] The hearth activity update module is specifically used to learn and train the blast furnace hearth activity prediction model through self-learning based on the newly acquired real-time blast furnace production data, and to update the neural network hyperparameters and influencing parameters of the prediction model in real time or regularly.
[0042] (III) Beneficial effects
[0043] First, by strengthening the management and application of blast furnace production data, the waste of data resources is avoided, the quality of blast furnace production data is improved, and a stable data source is provided for the evaluation, prediction and feedback of blast furnace hearth activity.
[0044] Second, the blast furnace hearth activity evaluation system, which deeply integrates blast furnace production data and big data algorithms, can meet the actual production conditions of the blast furnace to the greatest extent and realize accurate evaluation of the blast furnace hearth activity.
[0045] Third, the blast furnace hearth activity prediction model is continuously updated through self-learning, achieving long-term and accurate prediction of the blast furnace hearth activity. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] The present application is described with the aid of the following drawings:
[0047] Figure 1It is a flowchart of a blast furnace hearth activity evaluation and prediction method based on big data;
[0048] Figure 2 Schematic diagram of the module structure of the blast furnace hearth activity evaluation and prediction system based on big data. DETAILED DESCRIPTION
[0049] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below in conjunction with the accompanying drawings through specific implementation methods. It is understood that the specific embodiments described below are only used to explain the relevant inventions, rather than to limit the invention. It should also be noted that the embodiments and features in the embodiments of this application can be combined with each other in the absence of conflict; for ease of description, only the parts related to the invention are shown in the accompanying drawings.
[0050] Embodiment 1 provides a method for evaluating and predicting the activity of a blast furnace hearth based on big data, such as Figure 1 As shown, the specific steps are as follows:
[0051] S1. Collect blast furnace production history data, and manage the blast furnace production history data through data preprocessing to obtain blast furnace hearth activity application data.
[0052] In this embodiment, the blast furnace historical data includes blast furnace raw fuel data, blast furnace operation data, blast furnace smelting status data and slag and iron emission data.
[0053] In this embodiment, preprocessing includes data missing value processing, outlier processing, data frequency alignment and time lag processing.
[0054] S2. Determine a calculation method for the activity of a blast furnace hearth based on the application data of the activity of the blast furnace hearth, establish a blast furnace hearth activity evaluation model by integrating the calculation method, and perform a preliminary evaluation of the activity of the blast furnace hearth through the evaluation model.
[0055] S3. Obtain a blast furnace condition stability index based on blast furnace hearth activity application data, correct a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and make a final evaluation and grade classification of the blast furnace hearth activity based on the corrected blast furnace hearth activity evaluation model.
[0056] S4. Obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by performing deep learning on the influencing parameters, and predict the blast furnace hearth activity through the prediction model.
[0057] S5. Acquire real-time data of blast furnace production, learn and train a blast furnace hearth activity prediction model based on the real-time data of blast furnace production, and update the neural network hyperparameters and influencing parameters of the prediction model through self-learning.
[0058] In this embodiment, the neural network hyperparameters include the number of hidden layers, the number of neuron nodes in each layer, the learning rate, the activation function, the dropout rate, and the number of iterations.
[0059] The method in Example 1 conducts a preliminary evaluation of the blast furnace hearth activity through a blast furnace hearth activity evaluation model, conducts a final evaluation and classification of the blast furnace hearth activity through a revised blast furnace hearth activity evaluation model, and predicts the blast furnace hearth activity through a blast furnace hearth activity prediction model, thereby achieving accurate evaluation and precise prediction of the blast furnace hearth activity, and providing a guarantee for the high quality, high yield, low consumption and smooth operation of the blast furnace.
[0060] Embodiment 2 provides a method for evaluating and predicting the activity of a blast furnace hearth based on big data, and the specific method steps are as follows:
[0061] S1. Collect blast furnace production history data, and manage the blast furnace production history data through data preprocessing to obtain blast furnace hearth activity application data.
[0062] Based on S1 above, it should be noted that:
[0063] An optional processing method for obtaining the blast furnace hearth activity application data is to obtain the blast furnace production history data of the most recent two years, and to process the obtained blast furnace production history data through data preprocessing, and then obtain the blast furnace hearth activity application data.
[0064] In this embodiment, for the preprocessing of blast furnace production historical data, the blast furnace production historical data can be converted, cleaned and integrated through big data technology to improve the quality of blast furnace production historical data and realize the connection and call of data between different processes.
[0065] In this embodiment, in order to better understand the preprocessing of blast furnace production history data, the specific operation process includes: for the problems existing in blast furnace production history data, the main processing means include filling or eliminating missing values, replacing or eliminating outliers, aligning the frequency of data of different processes, etc. The above operation process can not only improve the data quality, but also solve the problem of data sample imbalance and data inability to directly match caused by different data collection cycles of different processes.
[0066] S2. Determine a calculation method for the activity of a blast furnace hearth based on the application data of the activity of the blast furnace hearth, establish a blast furnace hearth activity evaluation model by integrating the calculation method, and perform a preliminary evaluation of the activity of the blast furnace hearth through the evaluation model.
[0067] Based on the above S2, it should be noted that in order to better understand the real-time evaluation of the activity of the blast furnace hearth, the specific method steps are as follows:
[0068] S2-1. Based on the blast furnace hearth activity application data, obtain the correlation between the calculation method of each blast furnace hearth activity and the furnace condition parameters, and determine the calculation method of the blast furnace hearth activity based on the correlation.
[0069] The furnace condition parameters include coke ratio, air permeability, blast kinetic energy, gas utilization rate, and heat load.
[0070] Based on the above S2-1, in order to better understand, we can analyze the correlation between the blast furnace hearth activity calculation method and the furnace condition parameters based on the maximum information coefficient MIC (MaximalInformation Coefficient), and screen out the blast furnace hearth activity calculation method with a maximum information coefficient MIC higher than 0.3.
[0071] S2-2. The coefficients of the determined calculation method are corrected through machine learning, and a blast furnace hearth activity evaluation model is established by integrating the corrected calculation method.
[0072] Based on the above S2-2, for better understanding, if the blast furnace hearth activity calculation method determined by screening contains empirical coefficients, the empirical coefficients in the blast furnace hearth activity calculation method are corrected based on the least squares method.
[0073] S2-3. Obtain a blast furnace hearth activity index based on a blast furnace hearth activity evaluation model, and conduct a preliminary evaluation of the blast furnace hearth activity using the blast furnace hearth activity index.
[0074] Based on S2-2 and S2-3 above, for a better understanding, the determined blast furnace hearth activity calculation method can be integrated based on principal component analysis to obtain the blast furnace hearth activity comprehensive index. The specific method steps are as follows:
[0075] 1) Perform normal conversion and standardization on the data of blast furnace hearth activity calculation method;
[0076] 2) Based on principal component analysis, the results of the calculation method of the hearth activity of each blast furnace are converted into the same number of principal components, and the characteristic root λ of each principal component is calculated;
[0077] 3) The weight of the i-th principal component is equal to the ratio of the characteristic root λ of the i-th principal component to the sum of the characteristic roots (i.e. w i =λ i / λ 和 );
[0078] 4) The principal components are weighted and summed according to their weights, and finally the comprehensive index of blast furnace hearth activity is calculated.
[0079] S3. Obtain a blast furnace condition stability index based on blast furnace hearth activity application data, correct a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and make a final evaluation and grade classification of the blast furnace hearth activity based on the corrected blast furnace hearth activity evaluation model.
[0080] Based on the above S3, it should be noted that in order to better understand the modification of the blast furnace hearth activity evaluation model, an optional method is as follows:
[0081] S3-1. Acquire blast furnace condition parameters based on blast furnace hearth activity application data, and acquire blast furnace condition stability index by integrating blast furnace condition parameters.
[0082] In this embodiment, the blast furnace condition parameters include blast furnace economic index, blast furnace forward index, and furnace erosion index;
[0083] Based on the above S3-1, an optional processing method is: based on the principles of high production, high quality, smooth operation, low consumption and long life of the blast furnace, screen out blast furnace economic indicators and blast furnace smooth operation indicators.
[0084] In this embodiment, the blast furnace economic indicators include coke ratio and molten iron output; the blast furnace forward indicators include air permeability, blast kinetic energy, gas utilization rate, heat load, and molten iron [Si] content; the hearth erosion indicators include hearth erosion thickness, hearth side wall water temperature difference, and furnace bottom center temperature.
[0085] In this embodiment, the weight of each blast furnace condition parameter is calculated by factor analysis, and the blast furnace condition stability index is calculated based on the weight of the blast furnace condition parameters.
[0086] S3-2. Verify the accuracy of the blast furnace hearth activity evaluation model based on the blast furnace condition stability index, obtain the verification result, and modify the blast furnace hearth activity evaluation model based on the verification result.
[0087] Based on the above S3-2, an optional processing method is: based on the blast furnace hearth activity application data, the blast furnace hearth activity index and the blast furnace condition stability index are correlated through the maximum information coefficient to verify the accuracy of the blast furnace hearth activity evaluation model.
[0088] In this embodiment, the factors affecting the operation status of the blast furnace are divided into two aspects: blast furnace hearth activity and other factors, and the blast furnace operation status is characterized by a blast furnace condition stability index.
[0089] In this embodiment, when the blast furnace condition stability index is 20% higher than the average level, it means that the blast furnace condition operation state is stable, and at this time, the comprehensive index of furnace activity should also be higher than the average level.
[0090] In this embodiment, when the blast furnace condition stability index is 20% lower than the average level, it means that the blast furnace condition operation state is poor, and at this time, the comprehensive index of furnace activity should also be lower than the average level.
[0091] In this embodiment, when the blast furnace condition stability index is within ±20% of the average level, it represents that the blast furnace condition operation state is normalized, and at this time, the comprehensive index of furnace activity is not corrected.
[0092] S3-3. Based on the revised blast furnace hearth activity evaluation model, the final blast furnace hearth activity index is obtained, and the blast furnace hearth activity is graded based on the blast furnace condition stability index and the final blast furnace hearth activity index.
[0093] Based on the above S3-3, an optional processing method is: use K-means clustering algorithm to perform cluster analysis on the blast furnace hearth comprehensive activity index and blast furnace condition stability index, divide the blast furnace hearth activity into three levels: excellent, good, and poor, and obtain the data interval corresponding to each level.
[0094] In this embodiment, if the blast furnace condition stability index and the final blast furnace hearth activity index are both higher than the corresponding preset values, the blast furnace hearth activity is at the first level, that is, the excellent level.
[0095] Among them, the preset value is not a fixed value and can be determined according to the specific furnace conditions.
[0096] In this embodiment, if the blast furnace condition stability index is lower than the preset value and the final blast furnace hearth activity index is higher than the preset value, or the blast furnace condition stability index is higher than the preset value and the final blast furnace hearth activity index is lower than the preset value, then the blast furnace hearth activity is at the second level, which is a good level.
[0097] In this embodiment, if the blast furnace condition stability index and the final blast furnace hearth activity index are both lower than the corresponding preset values, the blast furnace hearth activity is at the third level, that is, the poor level.
[0098] S4. Obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by performing deep learning on the influencing parameters, and predict the blast furnace hearth activity through the prediction model.
[0099] Based on the above S4, in order to better understand the prediction of blast furnace hearth activity, the specific method steps are as follows:
[0100] S4-1. Based on the application data of blast furnace hearth activity, the blast furnace parameters are transformed by constructing feature operators to obtain derived features. Taking the blast furnace hearth activity index as the target, based on the derived features and blast furnace parameters, the influencing parameters of the blast furnace hearth activity are determined through feature screening.
[0101] In this embodiment, the constructed feature operators include but are not limited to reciprocal conversion, exponential conversion, tangent conversion, S-type conversion, and polynomial conversion.
[0102] In this embodiment, the feature screening method includes but is not limited to the Pearson correlation coefficient and the maximum information coefficient.
[0103] Based on the above S4-1, an optional method is: taking the blast furnace hearth activity index as the target, based on expert experience, through a stepwise regression algorithm to eliminate irrelevant or redundant blast furnace condition parameters, and screen out the key influencing parameters affecting the blast furnace hearth activity, so as to reduce the blast furnace hearth activity prediction model training time and further improve the accuracy of the prediction model.
[0104] S4-2. Based on the determined influencing parameters, the neural network hyperparameters are determined through grid search and cross validation, and then a blast furnace hearth activity prediction model is established, and the blast furnace hearth activity is predicted based on the blast furnace hearth activity prediction model.
[0105] In this embodiment, the activity of the blast furnace hearth can be predicted one hour in advance based on the blast furnace hearth activity prediction model.
[0106] Based on S4-2 above, an optional method is: based on the screened influencing parameters of blast furnace hearth activity, split the data corresponding to the influencing parameters into training set, validation set and test set in a ratio of 7:2:1; and build a deep neural network, determine the number of hidden layers of the neural network, the number of neuron nodes in each layer, the learning rate, the activation function, the discard rate, the number of iterations and other hyperparameters through grid search and cross-validation, and then establish a blast furnace hearth activity prediction model, and make a high-precision prediction of the blast furnace hearth activity comprehensive index based on the prediction model.
[0107] S5. Acquire real-time data of blast furnace production, learn and train a blast furnace hearth activity prediction model based on the real-time data of blast furnace production, and update the neural network hyperparameters and influencing parameters of the prediction model through self-learning.
[0108] Based on the above S5, in the specific processing process, the self-update cycle of the blast furnace hearth activity prediction model is set, and the blast furnace hearth activity application data is updated by obtaining the real-time data of blast furnace production; the blast furnace hearth activity prediction model is re-learned and trained using the updated blast furnace hearth activity application data, and the neural network hyperparameters and hearth activity influencing parameters are regularly updated.
[0109] The method in Example 2 conducts a preliminary evaluation of the blast furnace hearth activity through a blast furnace hearth activity evaluation model, conducts a final evaluation and classification of the blast furnace hearth activity through a revised blast furnace hearth activity evaluation model, and predicts the blast furnace hearth activity through a blast furnace hearth activity prediction model, thereby achieving accurate evaluation and precise prediction of the blast furnace hearth activity, and providing a guarantee for high quality, high yield, low consumption and smooth operation of the blast furnace.
[0110] Embodiment 3 provides a method for calculating the activity of a blast furnace hearth, and the specific calculation method includes:
[0111] The first calculation method: Among them, A is the activity of blast furnace condition, T0 is the average temperature of the thermocouple on the side wall of the furnace, T C It is the average temperature of the thermocouples at the center of each layer of the furnace bottom.
[0112] Second calculation method: Among them, Y is the daily molten iron output of the blast furnace, and D is the number of times the blast furnace taps iron per day.
[0113] The third calculation method: Among them, t p is the physical heat of molten iron, and [Si] is the silicon content of molten iron.
[0114] The fourth calculation method: Among them, DMT is the dead material column temperature in the furnace core, t C It is the average temperature of thermocouples at each layer of the furnace side wall.
[0115] In this embodiment, the above four calculation methods are preferred methods for calculating the activity of the blast furnace hearth, but are not limited to the above four calculation methods.
[0116] In this embodiment, an optional processing procedure is: based on the blast furnace hearth activity application data, obtain the correlation between the above four calculation methods and the furnace condition parameters, and determine the calculation method of the blast furnace hearth activity based on the obtained correlation.
[0117] Based on the four calculation methods in Example 3, a calculation method for the activity of the blast furnace hearth can be determined, the coefficient of the determined calculation method can be corrected through machine learning, and a blast furnace hearth activity evaluation model can be established by integrating the corrected calculation methods, thereby achieving a scientific evaluation of the blast furnace hearth activity.
[0118] Embodiment 4 provides a blast furnace hearth activity evaluation and prediction system based on big data, such as Figure 2 As shown in the figure, the evaluation and prediction system specifically includes:
[0119] The data management module is specifically used to collect the blast furnace production history data, and process the blast furnace production history data through data preprocessing to obtain the blast furnace hearth activity application data;
[0120] A hearth activity evaluation module, specifically used to determine a calculation method for blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity evaluation model by integrating the calculation method, and perform a preliminary evaluation of the blast furnace hearth activity through the evaluation model;
[0121] A hearth activity correction module, specifically used to obtain a blast furnace condition stability index based on blast furnace hearth activity application data, correct a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and perform a final evaluation and grade classification of the blast furnace hearth activity based on the corrected blast furnace hearth activity evaluation model;
[0122] A hearth activity prediction module, specifically used to obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by deep learning the influencing parameters, and predict the blast furnace hearth activity through the prediction model;
[0123] The hearth activity update module is specifically used to learn and train the blast furnace hearth activity prediction model through self-learning based on the newly acquired real-time blast furnace production data, and to update the neural network hyperparameters and influencing parameters of the prediction model in real time or regularly.
[0124] Based on the data management module, hearth activity evaluation module, hearth activity correction module, hearth activity prediction module and hearth activity update module in Example 4, accurate evaluation, precise prediction and dynamic feedback of blast furnace hearth activity can be achieved, providing a guarantee for high quality, high yield, low consumption and smooth operation of the blast furnace.
[0125] It should be noted that in the claims, any figure marks between brackets should not be understood as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "one" or "an" preceding a component does not exclude the presence of multiple such components. In addition, it should be noted that in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" and the like refers to the specific features, structures, materials or characteristics described in conjunction with the embodiment or example included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0126] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments after knowing the basic creative concept. Therefore, the claims should be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.
[0127] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include these modifications and variations.
Claims
1. A method for evaluating and predicting the activity of a blast furnace hearth based on big data, characterized in that: include: S1. Collect blast furnace production history data, and manage the blast furnace production history data through data preprocessing to obtain blast furnace hearth activity application data; S2, determining a calculation method for the activity of a blast furnace hearth based on the blast furnace hearth activity application data, establishing a blast furnace hearth activity evaluation model by integrating the calculation method, and performing a preliminary evaluation of the blast furnace hearth activity by using the evaluation model; S3, obtaining a blast furnace condition stability index based on the blast furnace hearth activity application data, revising a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and finally evaluating and grading the blast furnace hearth activity based on the revised blast furnace hearth activity evaluation model; S4. Obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by performing deep learning on the influencing parameters, and predict the blast furnace hearth activity through the prediction model.
2. The evaluation and prediction method according to claim 1, characterized in that: Also includes: S5. Acquire real-time data of blast furnace production, learn and train a blast furnace hearth activity prediction model based on the real-time data of blast furnace production, and update the neural network hyperparameters and influencing parameters of the prediction model through self-learning.
3. The evaluation and prediction method according to claim 1, characterized in that: S1 includes: The blast furnace historical data includes blast furnace raw fuel data, blast furnace operation data, blast furnace smelting status data and slag and iron discharge data; The preprocessing includes data missing value processing, outlier processing, data frequency alignment and time lag processing.
4. The evaluation and prediction method according to claim 1, characterized in that S2 include: S2-1. Based on the blast furnace hearth activity application data, obtain the correlation between the calculation method of each blast furnace hearth activity and the furnace condition parameter, and determine the calculation method of the blast furnace hearth activity based on the correlation; The furnace condition parameters include coke ratio, air permeability, blast kinetic energy, gas utilization rate, and heat load; S2-2, modifying the coefficients of the determined calculation method through machine learning, and establishing a blast furnace hearth activity evaluation model by integrating the modified calculation method; S2-3. Obtain a blast furnace hearth activity index based on a blast furnace hearth activity evaluation model, and conduct a preliminary evaluation of the blast furnace hearth activity using the blast furnace hearth activity index.
5. The evaluation and prediction method according to claim 4, characterized in that S3 include: S3-1. Obtain blast furnace condition parameters based on blast furnace hearth activity application data, and obtain blast furnace condition stability index by integrating blast furnace condition parameters; The blast furnace condition parameters include blast furnace economic index, blast furnace forward index, and furnace erosion index; S3-2, based on the blast furnace condition stability index, checking the accuracy of the blast furnace hearth activity evaluation model, obtaining the test result, and correcting the blast furnace hearth activity evaluation model based on the test result; S3-3. Based on the revised blast furnace hearth activity evaluation model, the final blast furnace hearth activity index is obtained, and the blast furnace hearth activity is graded based on the blast furnace condition stability index and the final blast furnace hearth activity index.
6. The evaluation and prediction method according to claim 5, characterized in that: S3-3 includes: If the blast furnace condition stability index and the final blast furnace hearth activity index are both higher than the corresponding preset values, the blast furnace hearth activity is at the first level; If the blast furnace condition stability index is lower than the preset value and the final blast furnace hearth activity index is higher than the preset value, or the blast furnace condition stability index is higher than the preset value and the final blast furnace hearth activity index is lower than the preset value, the blast furnace hearth activity is at the second level; If the blast furnace condition stability index and the final blast furnace hearth activity index are both lower than the corresponding preset values, the blast furnace hearth activity is at the third level.
7. The evaluation and prediction method according to claim 1, characterized in that S4 include: Based on the acquired influencing parameters, a prediction model for blast furnace hearth activity is established by determining the neural network hyperparameters; The neural network hyperparameters include the number of hidden layers, the number of neuron nodes in each layer, the learning rate, the activation function, the dropout rate and the number of iterations.
8. A blast furnace hearth activity evaluation and prediction system based on big data, characterized in that: include: The data management module is specifically used to collect the blast furnace production history data, and process the blast furnace production history data through data preprocessing to obtain the blast furnace hearth activity application data; A hearth activity evaluation module, specifically used to determine a calculation method for blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity evaluation model by integrating the calculation method, and perform a preliminary evaluation of the blast furnace hearth activity through the evaluation model; A hearth activity correction module, specifically used to obtain a blast furnace condition stability index based on blast furnace hearth activity application data, correct a blast furnace hearth activity evaluation model based on the blast furnace condition stability index, and perform a final evaluation and grade classification of the blast furnace hearth activity based on the corrected blast furnace hearth activity evaluation model; A hearth activity prediction module, specifically used to obtain influencing parameters of blast furnace hearth activity based on blast furnace hearth activity application data, establish a blast furnace hearth activity prediction model by deep learning the influencing parameters, and predict the blast furnace hearth activity through the prediction model; The hearth activity update module is specifically used to learn and train the blast furnace hearth activity prediction model through self-learning based on the newly acquired real-time blast furnace production data, and to update the neural network hyperparameters and influencing parameters of the prediction model in real time or regularly.
Citation Information
Patent Citations
Method for quantitative evaluation of activeness of blast furnace hearth
CN104388613A
Method for implementing efficient and low-consumption smelting of 4000m<3> blast furnace by aid of scanning radar
CN107641669A
Cited By
Machine learning-based dynamic regulation and control method and system for utilization ratio of sintered return mine cold-pressed block blast furnace
CN121832299A
