A blast furnace throat temperature estimation method and device based on multi-time scale fusion

Abstract

The application discloses a blast furnace throat temperature estimation method and device based on multi-time scale fusion, comprising: obtaining blast furnace equipment parameters, charge parameters, distribution matrix, blast furnace main parameters and blast furnace throat temperature measurement point data; taking the blast furnace equipment parameters, charge parameters and distribution matrix as inputs, calculating the shape of each layer of the blast furnace and the mineral coke ratio parameters of each layer; according to the multi-time scale characteristics of the internal reaction of the blast furnace, the blast furnace main parameters, the blast furnace throat temperature measurement point data and the mineral coke ratio parameters are divided to obtain high and low frequency data sets, and feature importance screening is performed respectively; the low frequency data set after feature screening is used to train a support vector regressor, the low frequency blast furnace throat temperature is estimated, the low frequency blast furnace throat temperature estimation value is expanded and fused into the high frequency data set after screening; based on the fused high frequency data set, an integrated method of a gated recurrent unit network is used to train a blast furnace throat temperature estimation model, which is used for high frequency real-time estimation of the temperature at the corresponding position of the blast furnace throat.

Description

Technical Field

[0001] This application relates to the field of energy and power engineering technology, and in particular to a method and apparatus for estimating blast furnace throat temperature based on multi-timescale fusion. Background Technology

[0002] With modernization, society's demand for both the quantity and quality of steel is constantly increasing. Steel production and quality levels have become important indicators of a country's economic strength and industrialization level. Blast furnace ironmaking is the main method of steel production worldwide, accounting for 70% of total energy consumption in steel production. A blast furnace is a large-scale countercurrent moving bed chemical reactor, involving complex mass transfer phenomena and chemical reactions, and is a typical "black box" container. Furthermore, the internal environment of a blast furnace is extremely harsh; high temperature, high pressure, and high dust levels make many parameters difficult to detect directly or extremely costly to measure, thus hindering the formation of closed-loop control for blast furnace operation.

[0003] Accurately detecting and rationally controlling the distribution and development of gas flow is a necessary prerequisite for the control of the blast furnace ironmaking process. The distribution of gas flow is directly related to important indicators such as blast furnace operating conditions, energy and raw material consumption, iron production and quality, and blast furnace lifespan. Different gas flow distributions will also result in different operating conditions inside the blast furnace. Therefore, timely understanding of the gas flow conditions at the charge surface and rationally controlling its distribution are of great significance for improving the quality and output of blast furnace products.

[0004] The temperature measurement at the blast furnace throat directly reflects the gas flow distribution and is a characteristic parameter of the gas flow state at the furnace top. Monitoring the throat temperature is one of the important aspects of monitoring the blast furnace ironmaking process. In production practice, the throat cross-shaped temperature measuring device based on thermocouple sensors is the most classic and widely used method. The throat cross-shaped temperature measurement has high accuracy, good real-time performance, and good dynamic performance. However, due to the presence of high-temperature gas in direct contact with the temperature sensor, significant dust, and harsh high-pressure environment at the throat, on the one hand, the collected data is often unusable directly; on the other hand, sensors frequently malfunction, making it difficult to detect accurate temperature data. Furthermore, the long shutdown and maintenance cycle of the blast furnace affects the staff's judgment of the actual operating status of the blast furnace, thus bringing risks to the monitoring of blast furnace operation and seriously affecting the efficient and low-consumption operation of the blast furnace. Summary of the Invention

[0005] The purpose of this application is to provide a method and apparatus for estimating blast furnace throat temperature based on multi-timescale fusion, so as to solve the problem that the throat temperature is difficult to measure accurately in the prior art.

[0006] According to a first aspect of the embodiments of this application, a method for estimating the throat temperature of a blast furnace based on multi-timescale fusion is provided, comprising:

[0007] Acquire blast furnace equipment parameters, burden parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data;

[0008] Using the blast furnace equipment parameters, furnace charge parameters, and charge distribution matrix as inputs, the shape of the material surface of each layer inside the blast furnace is calculated according to the charge distribution rule, the principle of equal volume, and the descent rule. The ore-coke ratio parameter of each layer is then calculated based on the material surface shape.

[0009] Based on the multi-timescale characteristics of the internal reaction of the blast furnace, the main parameters of the blast furnace, the temperature measurement point data of the furnace throat, and the ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets, and feature importance is filtered to obtain the high-frequency datasets and low-frequency datasets after feature filtering.

[0010] A support vector regressor is trained using the low-frequency dataset after feature filtering. The trained support vector regressor model is then used to estimate the low-frequency furnace throat temperature. The estimated low-frequency furnace throat temperature is then expanded based on random oversampling and fused into the filtered high-frequency dataset to obtain the fused high-frequency dataset.

[0011] Based on the fused high-frequency dataset, an ensemble method using gated recurrent unit networks is used to train a blast furnace throat temperature estimation model.

[0012] The blast furnace throat temperature estimation model is used to perform high-frequency real-time estimation of the temperature at the corresponding location in the throat.

[0013] Optionally, the blast furnace equipment parameters, burden parameters, and burden distribution matrix are used as inputs to calculate the surface shape of each layer inside the blast furnace based on the burden distribution pattern, the principle of equal volume, and the descent pattern, including:

[0014] Based on the blast furnace equipment parameters, calculate the coordinate position of the tip of the pile formed when the furnace charge reaches the surface inside the blast furnace.

[0015] Using the material surface formed by the previous inclination angle as the new initial material surface shape, calculate the material surface shape function under the second to last chute inclination angles in sequence according to the material matrix, complete the material loop of one material matrix, and repeat this step to obtain the material surface shape of each layer.

[0016] Optionally, the coordinate position includes an abscissa and a ordinate, which are calculated based on the principle that the volume of the single-ring furnace charge in the charge distribution matrix is ​​equal to the volume between the shapes of the two consecutive charge surfaces.

[0017] Optionally, the coke ratio parameters of each layer are calculated based on the shape of the material surface, including:

[0018] Based on the material surface shape, the calculation formulas for the coke ratio parameters of each layer are as follows:

[0019]

[0020] Where b represents the distance between a point and the centerline of the blast furnace, χ(b) η Let represent the shape function of the material surface of the ηth layer, M represent the selected layer number, η and M are positive integers, the subscript o represents the ore layer, and c represents the coke layer.

[0021] Optionally, based on the multi-timescale characteristics of the reaction inside the blast furnace, the blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets, including:

[0022] Based on the multi-timescale characteristics of the internal reaction of the blast furnace, a high-frequency dataset and a low-frequency dataset are constructed with two different sampling frequencies. The high-frequency dataset includes the main parameters of the blast furnace and the temperature measurement point data of the furnace throat, while the low-frequency dataset includes the main parameters of the blast furnace and the temperature measurement point data of the furnace throat and the ore-coke ratio parameter.

[0023] Optionally, feature importance filtering can be performed separately to obtain high-frequency and low-frequency datasets after feature filtering, including:

[0024] Based on the data characteristics and distribution of high-frequency and low-frequency datasets, feature importance screening was performed separately. For high-frequency datasets, feature importance scores based on random forests were used for feature screening; for low-frequency datasets, feature screening was performed based on expert experience and literature.

[0025] Optionally, based on the fused high-frequency dataset, an ensemble method using gated recurrent unit networks is employed to train a blast furnace throat temperature estimation model, including:

[0026] The fused high-frequency dataset is divided into a training set and a validation set. By using different training sets and validation sets, gated recurrent unit network base models with different parameters are trained.

[0027] Based on the basic model of all gated cyclic unit networks, a model for estimating the blast furnace throat temperature is obtained.

[0028] According to a second aspect of the embodiments of this application, a blast furnace throat temperature estimation device based on multi-timescale fusion is provided, comprising:

[0029] The acquisition module is used to acquire blast furnace equipment parameters, furnace charge parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data;

[0030] The calculation module is used to take the blast furnace equipment parameters, furnace charge parameters and charge distribution matrix as input, calculate the material surface shape of each layer inside the blast furnace according to the charge distribution law, the equal volume principle and the descent law, and calculate the ore-coke ratio parameter of each layer according to the material surface shape;

[0031] The partitioning and filtering module is used to partition the blast furnace main parameters, throat temperature measurement point data and ore-coke ratio parameters into high-frequency datasets and low-frequency datasets based on the multi-timescale characteristics of the reaction inside the blast furnace, and to perform feature importance filtering to obtain the feature-filtered high-frequency datasets and low-frequency datasets.

[0032] The estimation expansion and fusion module is used to train a support vector regressor using the low-frequency dataset after feature selection, estimate the low-frequency furnace throat temperature using the trained support vector regressor model, expand the low-frequency furnace throat temperature estimate based on random oversampling, and fuse it into the selected high-frequency dataset to obtain the fused high-frequency dataset.

[0033] The training module is used to train a blast furnace throat temperature estimation model based on the fused high-frequency dataset using an ensemble method of gated recurrent unit networks.

[0034] The estimation module is used to use the blast furnace throat temperature estimation model to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat.

[0035] According to a third aspect of the embodiments of this application, an electronic device is provided, the electronic device comprising:

[0036] One or more processors;

[0037] Storage device for storing one or more programs.

[0038] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described in the first aspect.

[0039] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that the program, when executed by a processor, implements the method described in the first aspect.

[0040] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0041] As can be seen from the above embodiments, this application combines the corresponding motion laws inside the blast furnace to establish a mathematical model of the blast furnace burden surface, calculates the ore-coke ratio parameters of each layer inside the furnace, and obtains high-frequency and low-frequency datasets based on the multi-timescale characteristics of the reaction inside the blast furnace. It then uses machine learning and ensemble learning methods to achieve high-frequency real-time estimation of the temperature at the corresponding position of the furnace throat, thus realizing online monitoring of the temperature inside the blast furnace throat.

[0042] When the cross temperature measuring device in the blast furnace throat malfunctions, this invention uses modern methods to estimate the blast furnace throat temperature in real time, which can effectively help on-site personnel to more accurately judge the distribution of blast furnace gas flow, and has important practical significance for ensuring smooth flow inside the blast furnace.

[0043] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0044] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0045] Figure 1 This is a flowchart illustrating a blast furnace throat temperature estimation method based on multi-timescale fusion according to an exemplary embodiment.

[0046] Figure 2 This is a schematic diagram illustrating the acquisition of high-frequency and low-frequency datasets according to an exemplary embodiment.

[0047] Figure 3 This is a flowchart illustrating a blast furnace throat temperature estimation process based on multi-timescale fusion, according to an exemplary embodiment.

[0048] Figure 4 This is a graph showing the results of blast furnace throat temperature estimation using four methods according to an exemplary embodiment.

[0049] Figure 5 This is a graph showing the blast furnace throat temperature estimation error results of four methods according to an exemplary embodiment.

[0050] Figure 6 This is a block diagram illustrating a blast furnace throat temperature estimation device based on multi-timescale fusion according to an exemplary embodiment.

[0051] Figure 7 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation

[0052] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0053] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0054] The blast furnace ironmaking process exhibits significant multi-scale characteristics. The average residence times of the solid, liquid, and gaseous phases within the blast furnace vary considerably. For example, coke has an average residence time of 1-4 weeks, or even longer, while iron ore and the liquid phase have an average residence time of 5-10 hours, and the gaseous phase has an average residence time of 10-20 seconds. Theoretically, each sub-scale structure has its own control mechanism, contributing to the system's behavior. Therefore, capturing information at these different scales is crucial. To capture the complex dynamics of the blast furnace, it is necessary to construct a multi-timescale model that reflects the multi-scale characteristics exhibited during blast furnace ironmaking.

[0055] Figure 1 This is a flowchart illustrating a blast furnace throat temperature estimation method based on multi-timescale fusion according to an exemplary embodiment, such as... Figure 1 As shown, the method may include the following steps:

[0056] S1: Obtain blast furnace equipment parameters, burden parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data;

[0057] S2: Using the blast furnace equipment parameters, furnace charge parameters, and charge distribution matrix as inputs, calculate the material surface shape of each layer inside the blast furnace according to the charge distribution rule, equal volume principle, and descent rule, and calculate the ore-coke ratio parameter of each layer according to the material surface shape;

[0058] S3: Based on the multi-timescale characteristics of the reaction inside the blast furnace, the main parameters of the blast furnace, the temperature measurement point data of the furnace throat, and the ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets, and feature importance is filtered to obtain the high-frequency datasets and low-frequency datasets after feature filtering.

[0059] S4: Train a support vector regressor using the low-frequency dataset after feature filtering, estimate the low-frequency furnace throat temperature using the trained support vector regressor model, expand the low-frequency furnace throat temperature estimate based on random oversampling, and fuse it into the filtered high-frequency dataset to obtain the fused high-frequency dataset.

[0060] S5: Based on the fused high-frequency dataset, an ensemble method using gated recurrent unit networks is used to train a blast furnace throat temperature estimation model.

[0061] S6: Use the blast furnace throat temperature estimation model to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat.

[0062] As can be seen from the above technical solutions, this application combines the corresponding motion laws inside the blast furnace to establish a mathematical model of the blast furnace burden surface, calculates the ore-coke ratio parameters of each layer inside the furnace, and obtains high-frequency and low-frequency datasets based on the multi-timescale characteristics of the reaction inside the blast furnace. By using machine learning and ensemble learning methods, high-frequency real-time estimation of the temperature at the corresponding position of the furnace throat is achieved, realizing online monitoring of the temperature inside the blast furnace throat. This provides a stronger guarantee for on-site personnel to correctly judge the gas flow distribution and ensure the smooth operation of the blast furnace.

[0063] In the specific implementation of S1: acquire blast furnace equipment parameters, furnace charge parameters, charging matrix, blast furnace main parameters and furnace throat temperature measurement point data;

[0064] Specifically, the blast furnace equipment parameters include the total height of the blast furnace, the throat height, the belly height, the body height, the waist height, the throat radius, the body inclination angle, the belly radius, the waist inclination angle, the central throat length h0, the throttle valve opening S, the chute length l, the chute tilting distance b, the chute rotation speed ω, the chute friction coefficient μ, and the material line depth H.

[0065] Specifically, the furnace charge parameters include the average particle size D of the coke. o,c Average density ρ of coke, inner angle of repose of coke and outer corner

[0066] Specifically, the fabric matrix includes the chute inclination angle β, the number of rotations corresponding to each inclination angle, and the volume of single-ring coke.

[0067] Specifically, the main parameters of the blast furnace include oxygen enrichment rate (%), carbon monoxide content in the gas (%), hydrogen content in the gas (%), carbon dioxide content in the gas (%), and permeability index (m). 3 / h·kPa), cold air flow rate (m³) 3 / h), oxygen-enriched flow rate (m 3 / h), furnace top pressure (kPa), total differential pressure (kPa), theoretical combustion temperature (°C), furnace belly gas index (m / min), hot and cold air temperatures (°C), cold and hot air pressure (kPa), furnace top temperature (°C), and blast humidity (g / m).

[0068] In the specific implementation of S2: the blast furnace equipment parameters, burden parameters, and burden distribution matrix are taken as inputs. The shape of the material surface of each layer inside the blast furnace is calculated according to the burden distribution pattern, the principle of equal volume, and the descent pattern. The ore-coke ratio parameter of each layer is calculated based on the material surface shape. This step may include the following sub-steps:

[0069] S21: Based on the blast furnace equipment parameters, calculate the coordinate position of the tip of the pile formed when the furnace charge reaches the surface inside the blast furnace;

[0070] Specifically, based on the known blast furnace equipment parameters of chute length, chute inclination angle, central tube length and material line height, the abscissa of the pile tip formed by the furnace charge reaching the material surface in the blast furnace radius direction is calculated through force analysis. The ordinate is calculated based on the principle that the volume of the single ring furnace charge in the charge distribution matrix is ​​equal to the volume between the shapes of the two material surfaces.

[0071] S22: Take the material surface formed by the previous inclination angle as the new initial material surface shape, calculate the material surface shape function under the second to last chute inclination angles in sequence according to the material matrix, complete the material loop of one material matrix, repeat this step to obtain the material surface shape of each layer.

[0072] S23: Based on the shape of the material surface, the calculation formula for the coke ratio parameter of each layer is as follows:

[0073]

[0074] Where b represents the distance between a point and the centerline of the blast furnace, χ(b) η Let represent the shape function of the material surface of the ηth layer, M represent the selected layer number, η and M are positive integers, the subscript o represents the ore layer, and c represents the coke layer.

[0075] In the specific implementation of S3: Based on the multi-timescale characteristics of the reaction inside the blast furnace, the blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets, and feature importance filtering is performed on each to obtain the feature-filtered high-frequency dataset and low-frequency dataset; this step may include the following sub-steps:

[0076] S31: Based on the multi-timescale characteristics of the internal reaction of the blast furnace, a high-frequency dataset and a low-frequency dataset are constructed with two different sampling frequencies. The high-frequency dataset includes the main parameters of the blast furnace and the temperature measurement point data of the furnace throat, while the low-frequency dataset includes the main parameters of the blast furnace and the temperature measurement point data of the furnace throat and the ore-coke ratio parameter.

[0077] Specifically, the high-frequency dataset is sampled at a high time frequency and has rich time-scale information; the low-frequency dataset not only includes blast furnace main parameters and throat temperature measurement point data, but also ore-coke ratio parameters, and has rich feature information.

[0078] S32: Based on the data characteristics and distribution of high-frequency and low-frequency datasets, feature importance screening is performed separately. For high-frequency datasets, feature importance scores based on random forests are used for feature screening; for low-frequency datasets, feature screening is performed based on expert experience and literature.

[0079] Specifically, based on the feature importance scores of random forests, the specific steps include:

[0080] ① First, assume there are p trees in the forest. For the i-th decision tree, randomly sample from all sample sets, use the randomly sampled samples to build the tree, and calculate the mean squared error between the out-of-bag data and the true value, denoted as .

[0081] ②Then the variable X is randomly replaced. j The observed values ​​can also be reconstructed by adding random noise interference and calculating the data error of the out-of-bag data, denoted as .

[0082] ③ Then, calculate the difference between the two data errors, and after standardization, obtain the variable X. j The importance score of a variable in a random forest is obtained by taking its contribution in the i-th decision tree, i.e., its importance, and then averaging the contribution across all trees. The formula is as follows:

[0083]

[0084] VIM j Represents variable X j Importance score.

[0085] ④ Finally, based on the importance scores, the top L feature variables with the highest importance scores are selected as the filtered features in descending order.

[0086] In the specific implementation of S4: a support vector regressor is trained using the low-frequency dataset after feature selection, and the low-frequency furnace throat temperature is estimated using the trained support vector regressor model. The low-frequency furnace throat temperature estimate is expanded based on random oversampling and fused into the selected high-frequency dataset to obtain the fused high-frequency dataset.

[0087] Specifically, a support vector regressor is trained using the low-frequency dataset after feature selection. The calculation process of support vector regression is expressed as follows:

[0088] ① Given a blast furnace training sample D={(x1,y1),(x2,y2),…,(x m ,y m )},y i ∈R, the support vector regression problem can be formalized as:

[0089]

[0090] Where C is the regularization constant, and ι is the ∈-insensitive loss function:

[0091]

[0092] ② Introduce slack variable ξ i and The support vector regression problem can be rewritten as:

[0093]

[0094]

[0095] Specifically, the furnace throat temperature is estimated based on support vector regression, and the low-frequency furnace throat temperature estimate is randomly oversampled. Random oversampling is a method that randomly copies and repeats minority class samples to eventually achieve a relatively balanced minority and majority classes. After random oversampling, the sample is added to the high-frequency dataset.

[0096] In the specific implementation of S5: Based on the fused high-frequency dataset, an ensemble method using gated recurrent unit networks is used to train a blast furnace throat temperature estimation model; this step may include the following sub-steps:

[0097] S51: Divide the fused high-frequency dataset into training and validation sets, and train multiple gated recurrent unit network base models;

[0098] Specifically, the fused high-frequency dataset is divided into k parts, one of which is selected as the validation set, and the remaining k-1 parts are used as the training set. This method yields k different training and validation sets. Using these k different training and validation sets, k gated recurrent unit network base models with different parameters are trained.

[0099] S52: Based on the basic model of all gated cyclic unit networks, the blast furnace throat temperature estimation model is obtained.

[0100] Specifically, the mean of the k basic model throat temperature estimates is used as the final high-frequency throat temperature estimate, which can be expressed as follows:

[0101]

[0102]

[0103] Where t represents the estimated time, x t Let u represent the feature at time t of the high-frequency dataset. k This represents the k-th gated recurrent unit network base model after training. Let represent the high-frequency throat temperature estimate calculated by the k-th gated recurrent unit network base model after training at time t. u represents the final blast furnace throat temperature estimation model. The final high-frequency throat temperature estimate is obtained by calculating the mean of the throat temperature estimates from the k base models. This represents the estimated temperature of the blast furnace throat at time t.

[0104] Specifically, the computation process of the gated recurrent unit network is expressed as follows:

[0105] Gated cyclic unit networks introduce a gating mechanism to control how information is updated, including updating the gating gate. t and reset door r t The update gate controls how much information needs to be retained from historical states to determine the current state, and how much new information needs to be accepted from candidate states as the current state h. t ,Right now

[0106] h t =z t ⊙h t-1 +(1-z t )⊙g(x t ,h t-1 ;θ)

[0107] Among them, z t For updating the gate:

[0108] z t =σ(W z x t +U z h t-1 +b z )

[0109] Among them, the function g(x) t ,h t-1 ;θ) is defined as

[0110]

[0111] in Let r represent the candidate state at the current moment. t This is a reset gate used to control candidate states. Does the calculation depend on the state h from the previous time step? t-1 Its expression is as follows:

[0112] r t =σ(W r x t +U r h t-1 +b r )

[0113] In summary, the state update method for gated recurrent networks is as follows:

[0114]

[0115] Specifically, the output of the blast furnace throat temperature estimation model is the average of the throat temperature estimates output by all base models.

[0116] In the specific implementation of S6: the blast furnace throat temperature estimation model is used to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat.

[0117] Figure 2 A schematic diagram illustrating the acquisition of high-frequency and low-frequency datasets in this embodiment of the invention is provided. A blast furnace burden distribution model is established using blast furnace equipment parameters and a charge distribution matrix, and the ore-coke ratio parameter is calculated. Based on the multi-timescale characteristics of the blast furnace, this model is fused with the blast furnace master parameters according to sampling frequency, and divided into high-frequency and low-frequency datasets, with feature importance filtering performed separately for each. Figure 3 A flowchart of the blast furnace throat temperature estimation based on multi-timescale fusion in this embodiment of the invention is provided. The high-frequency and low-frequency datasets are obtained after feature filtering. A support vector regressor is trained using the low-frequency dataset to estimate the low-frequency throat temperature, and the estimated low-frequency throat temperature values ​​are randomly oversampled and fused into the high-frequency dataset. Finally, an ensemble method based on a gated recurrent unit network is used to train the fused high-frequency dataset to achieve real-time high-frequency estimation of the temperature at the corresponding location in the throat.

[0118] In this embodiment, blast furnace master parameter data, throat temperature measurement point data, and blast furnace charging matrix for a specific month were selected from a blast furnace database in China. A total of 4320 sets of blast furnace master parameter data and throat temperature measurement point data were used; the first 3840 sets were used for training, and the last 480 sets were used for testing. Based on the established charge layer distribution model and according to the sampling frequency, 3840 sets of high-frequency training datasets, 480 sets of high-frequency test datasets, 160 sets of low-frequency training datasets, and 20 sets of low-frequency test datasets were obtained.

[0119] To conduct a more comprehensive analysis and discussion, four methods were used to estimate the furnace throat temperature:

[0120] (1) Furnace temperature estimation method based on support vector regression: After feature selection, the high-frequency dataset is trained using support vector regression, and the furnace throat temperature is estimated by the trained support vector regression model.

[0121] (2) A method for estimating blast furnace throat temperature based on support vector regression and multi-timescale fusion: The low-frequency dataset after feature selection is used as input, and support vector regression is used to obtain the low-frequency estimate of the blast furnace throat temperature. The low-frequency throat temperature estimate is expanded based on random oversampling and fused into the high-frequency dataset. The fused high-frequency dataset is used as input, and support vector regression is used for training. The final high-frequency throat temperature estimate is obtained through the trained support vector regression model.

[0122] (3) Furnace temperature estimation method based on gated recurrent unit network integration method: After feature selection, the high-frequency dataset is trained using the gated recurrent unit network integration method, and the furnace throat temperature is estimated by the trained model.

[0123] (4) The method proposed in this embodiment of the invention: The low-frequency dataset after feature filtering is used as input, and support vector regression is used to obtain the low-frequency estimate of the blast furnace throat temperature. The low-frequency throat temperature estimate is expanded based on random oversampling and fused into the high-frequency dataset. The fused high-frequency dataset is used as input, and the model is trained based on the gated recurrent unit network ensemble method. The final high-frequency throat temperature estimate is obtained through the trained model.

[0124] The estimation performance of the model is evaluated using statistical indicators such as MAPE (Mean Absolute Percentage Error), MAE (Mean Absolute Error), and RMSE (Root Mean Square Error). The calculation methods for these three indicators are shown below:

[0125]

[0126]

[0127]

[0128] Where y(i) is the measured value of the furnace throat temperature. T represents the estimated value of the furnace throat temperature measurement, and T is the sample size.

[0129] The temperature estimation results of the four methods are as follows Figure 4 As shown, the error results between the temperature estimation and measurement values ​​obtained by the four methods are as follows: Figure 5 As shown. SVR is a furnace temperature estimation method based on support vector regression, SVR_MTSF is a furnace throat temperature estimation method based on support vector regression and multi-timescale fusion, I-GRU is an ensemble learning method based on gated recurrent unit networks, and I-GRU_MTSF is the method proposed in this invention. The evaluation index results are as follows:

[0130]

[0131]

[0132] like Figure 4 and Figure 5 As shown, the proposed method BGRU_MTSF is significantly closer to the measured value than the other three methods, and its error value is significantly lower than that of the other three methods. The overall trend of the results of the proposed method is consistent with the measured temperature. By fusing information from low-frequency data into high-frequency data and utilizing the integration method of gated cyclic unit networks, better temperature estimation results and higher accuracy are obtained, proving the effectiveness and reliability of the proposed method.

[0133] As can be seen from the table, the proposed method has higher accuracy than the other three control methods, with MAPE of 0.030, MAE of 1.658, and RMSE of 2.001, all of which are the best among the four methods, thus providing evidence for its superiority. Figure 4 , Figure 5 The conclusions demonstrate that the temperature estimation method proposed in this invention is superior to other methods.

[0134] Corresponding to the aforementioned embodiment of a blast furnace throat temperature estimation method based on multi-timescale fusion, this application also provides an embodiment of a blast furnace throat temperature estimation device based on multi-timescale fusion.

[0135] Figure 6 This is a block diagram illustrating a blast furnace throat temperature estimation device based on multi-timescale fusion, according to an exemplary embodiment. (Refer to...) Figure 6 The device includes:

[0136] Module 1 is used to acquire blast furnace equipment parameters, furnace charge parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data;

[0137] Calculation module 2 is used to take the blast furnace equipment parameters, furnace charge parameters and charge distribution matrix as input, calculate the material surface shape of each layer inside the blast furnace according to the charge distribution law, the principle of equal volume and the descent law, and calculate the ore-coke ratio parameter of each layer according to the material surface shape;

[0138] The partitioning and filtering module 3 is used to partition the blast furnace main parameters, throat temperature measurement point data and ore-coke ratio parameters into high-frequency datasets and low-frequency datasets based on the multi-timescale characteristics of the reaction inside the blast furnace, and to perform feature importance filtering to obtain the feature-filtered high-frequency datasets and low-frequency datasets.

[0139] The estimation expansion and fusion module 4 is used to train a support vector regressor through the low-frequency dataset after feature screening, estimate the low-frequency furnace throat temperature using the trained support vector regressor model, expand the low-frequency furnace throat temperature estimate based on random oversampling, and fuse it into the screened high-frequency dataset to obtain the fused high-frequency dataset.

[0140] Training module 5 is used to train a blast furnace throat temperature estimation model based on the fused high-frequency dataset using an ensemble method of gated recurrent unit networks.

[0141] The estimation module 6 is used to use the blast furnace throat temperature estimation model to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat.

[0142] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0143] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0144] Accordingly, this application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the blast furnace throat temperature estimation method based on multi-timescale fusion as described above. Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities, which is a blast furnace throat temperature estimation device based on multi-timescale fusion provided in an embodiment of the present invention. Except for... Figure 7 In addition to the processor and memory shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0145] Accordingly, this application also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the blast furnace throat temperature estimation method based on multi-timescale fusion as described above. The computer-readable storage medium can be an internal storage unit of any data-processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of a wind turbine, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0146] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this application are indicated by the claims.

[0147] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for estimating blast furnace throat temperature based on multi-timescale fusion, characterized in that, include: Acquire blast furnace equipment parameters, burden parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data; Using the blast furnace equipment parameters, furnace charge parameters, and charge distribution matrix as inputs, the shape of the material surface of each layer inside the blast furnace is calculated according to the charge distribution rule, the principle of equal volume, and the descent rule. The ore-coke ratio parameter of each layer is then calculated based on the material surface shape. Based on the multi-timescale characteristics of the internal reaction of the blast furnace, the main parameters of the blast furnace, the temperature measurement point data of the furnace throat, and the ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets, and feature importance is filtered to obtain the high-frequency datasets and low-frequency datasets after feature filtering. A support vector regressor is trained using the low-frequency dataset after feature filtering. The trained support vector regressor model is then used to estimate the low-frequency furnace throat temperature. The estimated low-frequency furnace throat temperature is then expanded based on random oversampling and fused into the filtered high-frequency dataset to obtain the fused high-frequency dataset. Based on the fused high-frequency dataset, an ensemble method using gated recurrent unit networks is used to train a blast furnace throat temperature estimation model. The blast furnace throat temperature estimation model is used to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat. Specifically, based on the multi-timescale characteristics of the internal reaction of the blast furnace, the blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets. This includes: constructing high-frequency datasets and low-frequency datasets with two different sampling frequencies based on the multi-timescale characteristics of the internal reaction of the blast furnace. The high-frequency dataset contains blast furnace master parameters and throat temperature measurement point data, while the low-frequency dataset contains blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters. Specifically, based on the fused high-frequency dataset, an ensemble method of gated recurrent unit networks is used to train a blast furnace throat temperature estimation model, including: dividing the fused high-frequency dataset into a training set and a validation set, and training a gated recurrent unit network base model with different parameters through different training sets and validation sets; and obtaining a blast furnace throat temperature estimation model based on all gated recurrent unit network base models.

2. The method according to claim 1, characterized in that, Using the blast furnace equipment parameters, burden parameters, and burden distribution matrix as input, the shape of the burden surface in each layer inside the blast furnace is calculated based on the burden distribution pattern, the principle of equal volume, and the descent pattern, including: Based on the blast furnace equipment parameters, calculate the coordinate position of the tip of the pile formed when the furnace charge reaches the surface inside the blast furnace. Using the material surface formed by the previous inclination angle as the new initial material surface shape, calculate the material surface shape function under the second to last chute inclination angles in sequence according to the material matrix, complete the material loop of one material matrix, repeat the steps, and obtain the material surface shape of each layer.

3. The method according to claim 2, characterized in that, The coordinate position includes an abscissa and a ordinate, which are calculated based on the principle that the volume of the single-ring furnace charge in the charge distribution matrix is ​​equal to the volume between the shapes of the two charge surfaces.

4. The method according to claim 1, characterized in that, The coke ratio parameters for each layer are calculated based on the shape of the material surface, including: Based on the material surface shape, the calculation formulas for the coke ratio parameters of each layer are as follows: ； in, This represents the distance between a point and the centerline of the blast furnace. Indicates the first The shape function of the material surface of the layer, Indicates the selected number of material layers. and A positive integer, with index Indicates the ore layer. This indicates the coke layer.

5. The method according to claim 1, characterized in that, Feature importance filtering was performed separately to obtain high-frequency and low-frequency datasets, including: Based on the data characteristics and distribution of high-frequency and low-frequency datasets, feature importance screening was performed separately. For high-frequency datasets, feature importance scores based on random forests were used for feature screening; for low-frequency datasets, feature screening was performed based on expert experience and literature.

6. A blast furnace throat temperature estimation device based on multi-timescale fusion, characterized in that, include: The acquisition module is used to acquire blast furnace equipment parameters, furnace charge parameters, charging matrix, blast furnace main parameters, and throat temperature measurement point data; The calculation module is used to take the blast furnace equipment parameters, furnace charge parameters and charge distribution matrix as input, calculate the material surface shape of each layer inside the blast furnace according to the charge distribution law, the equal volume principle and the descent law, and calculate the ore-coke ratio parameter of each layer according to the material surface shape; The partitioning and filtering module is used to partition the blast furnace main parameters, throat temperature measurement point data and ore-coke ratio parameters into high-frequency datasets and low-frequency datasets based on the multi-timescale characteristics of the reaction inside the blast furnace, and to perform feature importance filtering to obtain the feature-filtered high-frequency datasets and low-frequency datasets. The estimation expansion and fusion module is used to train a support vector regressor using the low-frequency dataset after feature selection, estimate the low-frequency furnace throat temperature using the trained support vector regressor model, expand the low-frequency furnace throat temperature estimate based on random oversampling, and fuse it into the selected high-frequency dataset to obtain the fused high-frequency dataset. The training module is used to train a blast furnace throat temperature estimation model based on the fused high-frequency dataset using an ensemble method of gated recurrent unit networks. The estimation module is used to use the blast furnace throat temperature estimation model to perform high-frequency real-time estimation of the temperature at the corresponding position of the throat. Specifically, based on the multi-timescale characteristics of the internal reaction of the blast furnace, the blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters are divided into high-frequency datasets and low-frequency datasets. This includes: constructing high-frequency datasets and low-frequency datasets with two different sampling frequencies based on the multi-timescale characteristics of the internal reaction of the blast furnace. The high-frequency dataset contains blast furnace master parameters and throat temperature measurement point data, while the low-frequency dataset contains blast furnace master parameters, throat temperature measurement point data, and ore-coke ratio parameters. Specifically, based on the fused high-frequency dataset, an ensemble method of gated recurrent unit networks is used to train a blast furnace throat temperature estimation model, including: dividing the fused high-frequency dataset into a training set and a validation set, and training a gated recurrent unit network base model with different parameters through different training sets and validation sets; and obtaining a blast furnace throat temperature estimation model based on all gated recurrent unit network base models.

7. An electronic device, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-5.

Images