Method, device and equipment for rapidly calculating and predicting soft melting zone of blast furnace and medium

Abstract

The invention provides a rapid calculation and prediction method, device, equipment and medium for a soft melting zone of a blast furnace, and the method comprises the following steps: collecting blast furnace operation data, carrying out standardization processing, dynamically establishing a geometric model, constructing simulation models of grids with different resolutions, and carrying out hierarchical solving based on a historical sample matching condition to realize efficient prediction. When the matching degree is insufficient, the first resolution grid is adopted to quickly calculate and generate first interpolation data, when the matching degree is between basic matching and complete matching, second interpolation data is acquired in real time through the matching sample, the first or second interpolation data is mapped to the second resolution grid to be initialized, accurate calculation is carried out, and the first interpolation data and the second interpolation data are obtained. Fast and accurate calculation adopts a gas-solid single-core interaction parallel partitioning strategy to accelerate solution, a matching working condition result is directly reused when the matching degree meets complete matching, a calculation result and a standardized parameter are synchronously stored in a historical sample library, and data clustering matching and visual output are realized; the method effectively supports blast furnace smelting state monitoring, operation optimization and stable and smooth operation.

Description

Technical Field

[0001] This invention relates to the field of iron and steel smelting, and in particular to a method, apparatus, equipment and medium for rapid calculation and prediction of the softening zone of a blast furnace. Background Technology

[0002] As the core equipment in ironmaking, the blast furnace involves numerous processes and complex data, encompassing multiphase flow, heat transfer, chemical reactions, and phase transformations. The results are dynamically variable, exhibiting nonlinearity and large time delays. Furthermore, it operates under a closed, high-temperature, and high-pressure environment, lacking quantitative data and real-time monitoring methods. Traditional operation relies on manual experience, leading to delayed furnace condition adjustments, high fuel consumption, and difficulty in achieving precise control. The softening zone is a critical area for gas flow distribution and temperature and pressure fields, directly affecting blast furnace stability, output, and energy consumption. However, due to measurement difficulties, traditional experience-based operation cannot meet the demands for high efficiency, low carbon emissions, and intelligent operation.

[0003] Computational Fluid Dynamics (CFD) simulation models are considered an effective tool for understanding the "black box" of blast furnaces, enabling dynamic analysis of the softening zone state. However, existing CFD models suffer from significant drawbacks such as low computational efficiency, strong initialization dependence, and poor adaptability to operating conditions. Overall, current technologies struggle to achieve rapid prediction of the softening zone while maintaining accuracy, failing to overcome the challenge of online analysis of the blast furnace "black box" and provide theoretical support for real-time monitoring and operation optimization. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for rapid calculation and prediction of the softening zone in blast furnaces, in order to solve the problem that existing technologies cannot achieve rapid prediction of the softening zone while ensuring accuracy.

[0005] This application proposes a rapid calculation and prediction method for the softening zone of a blast furnace. The method includes: acquiring real-time operating data of the blast furnace under the current operating conditions, and standardizing the real-time operating data to generate standardized input parameters; establishing a geometric model of the blast furnace's inner contour based on the standardized input parameters, and dividing the geometric model into multiple sets of meshes with different resolutions, the multiple sets of meshes including at least a first resolution mesh and a second resolution mesh, the first resolution mesh being used for initial trial calculations, and the second resolution mesh being used for precise calculations; constructing a predictive calculation model for the softening zone, selecting different resolution meshes, and dynamically loading the standardized input parameters through a standardized interface to obtain calculation examples for different resolution meshes; and selecting a hierarchical solution strategy based on the matching situation between the standardized input parameters corresponding to the current operating conditions and historical samples: when the matching degree is insufficient to meet the basic matching threshold, a "coarse mesh trial calculation" is adopted. The collaborative solution strategy of "data mapping initialization - fine mesh accurate calculation" is adopted; when the matching degree is between the basic matching threshold and the complete matching threshold, the collaborative solution strategy of "historical sample matching - data mapping initialization - fine mesh accurate calculation" is adopted; when the matching degree meets the complete matching threshold, the real-time prediction strategy of "historical sample matching - matching result reuse" is adopted; the basic matching threshold and the complete matching threshold are set values ​​obtained based on the running data under historical working conditions; for the calculation and solution of grid cases with different resolutions, the spatial block parallel partitioning strategy of gas-solid single core interaction is adopted to accelerate the iterative solution process, and the calculation result corresponding to the current working condition is obtained after the calculation converges; the standardized input parameters and accurate calculation results of the current working condition are synchronously stored in the historical sample library for working condition data accumulation, clustering and matching query, and the multi-physics field distribution results in the furnace are visualized and displayed.

[0006] In one embodiment of this application, establishing a blast furnace inner contour geometric model based on the standardized input parameters includes: constructing a basic inner contour of the blast furnace body based on static geometric parameters in the standardized input parameters, wherein the static geometric parameters include the blast furnace segment height and diameter; generating dynamically changing feature structures in the basic inner contour by combining dynamic geometric parameters in the standardized input parameters, wherein the dynamic geometric parameters include the shape of the material surface position, the size of the swirling zone contour, the morphology of the dead material column, and the height of the slag and iron molten metal surface; and integrating the basic inner contour and the feature structures to generate a complete computational domain for predicting the softening zone, wherein the complete computational domain for predicting the softening zone covers the axial spatial range from the furnace throat to the hearth.

[0007] In one embodiment of this application, a predictive calculation model for the soft melt zone is constructed, including: establishing a model framework based on a preset dense-phase gas-solid two-phase flow decoupling processing method, with the following specific steps: generating a corresponding background grid based on the selected resolution grid, used to store the mass, momentum, and energy data of different phase components; performing flow tracking calculations of furnace charge particles based on the viscous flow method, recording the spatial location, batch, and attribute information of the furnace charge, and mapping the results of the flow tracking calculations to the background grid nodes; using a porous media model to calculate the gas flow resistance and heat transfer within the particle bed, wherein the resistance coefficient of the flow resistance is dynamically updated based on the raw material attributes and material distribution information in the standardized input parameters; setting source terms to realize the transfer of heat transfer, chemical reaction, and phase change data between the gas-solid-liquid phases, wherein the results of the source term calculation are used for the coupling of mass, momentum, and energy between phases.

[0008] In one embodiment of this application, the hierarchical solution strategy selection includes: for standardized input parameters under the current working condition, screening historical samples based on preset rules and obtaining the best matching sample, comprehensively calculating the matching degree between the two, and determining the solution strategy according to the relationship between the matching degree and the preset matching threshold; wherein, determining the solution strategy according to the relationship between the matching degree and the preset matching threshold includes: when the matching degree is less than the basic matching threshold, using sparse first-resolution grid cases to perform initial trial calculations, quickly obtaining convergence results as first interpolation data, mapping them as data sources to dense second-resolution grid cases as the initial field, and then performing precise calculations on the second-resolution grid cases; if the matching degree is greater than or equal to the basic matching threshold and less than the perfect matching threshold, then obtaining the matching results in real time through historical sample queries as second interpolation data, and mapping them as data sources to dense second-resolution grid cases as the initial field, and then performing precise calculations on the second-resolution grid cases; if the matching degree is greater than or equal to the basic matching threshold and less than the perfect matching threshold, then obtaining the matching results in real time through historical sample queries as second interpolation data, and then mapping them as second interpolation data. The data source is mapped to a dense second-resolution grid as the initial field, and then the precise calculation of the second-resolution grid is carried out. If the matching degree is greater than or equal to the complete matching threshold, the matching result is obtained in real time through historical sample query as the final precise calculation result of the current working condition, omitting the process of interpolation data acquisition, data mapping initialization and fine grid precise calculation. Before mapping the acquired first interpolation data or second interpolation data to a dense second-resolution grid as the initial field, it is necessary to detect whether there is a spatial position deviation between the data source and the current geometric model. If there is a deviation and it exceeds the preset range, the coordinate scaling correction of the data source is performed based on the blast furnace axial segmentation parameters, the corrected data source is mapped to the second-resolution grid, and the mapped data is set as the initial field of the second-resolution grid.

[0009] In one embodiment of this application, a spatial block parallel partitioning strategy with gas-solid single-core interaction is adopted to accelerate the iterative solution process, including: each iteration of the solution includes decoupling the velocity, pressure, temperature, and concentration information of the furnace charge flow and the gas flow, and then realizing the transfer and coupling solution of gas-solid phase data based on the background grid and source terms; during multiple iterations of the solution process, the residuals of the calculation equations and the fluctuations of variables are monitored until the preset convergence criteria are reached respectively, so as to obtain stable calculation results of the coupled solution; wherein, the iterative solution process includes dividing the grid space into blocks based on the preset gas-solid data single-core interaction principle, and following the principle of allocating gas phase and solid phase data in the same spatial location to the same calculation process.

[0010] In one embodiment of this application, storing the standardized input parameters and precise calculation results of the current operating condition to a historical sample database includes: extracting a set of key input parameters for the current operating condition, sorting each parameter based on a preset influence weight, and generating a multidimensional feature vector according to the parameter priority sequence. The set of key input parameters includes geometric structure, raw material properties, material distribution parameters, air supply parameters, furnace wall boundaries, and iron tapping parameters. A classification accuracy standard is set based on the preset influence weight, and historical samples are classified according to the multidimensional feature vector and the set accuracy standard. The multiphysics distribution results of the current operating condition are associated with the input feature vector, stored in the form of key-value pairs in the historical sample database of the operating condition, clustered, and retrieval identification information is added.

[0011] In one embodiment of this application, the visualization output of the multiphysics distribution results in the furnace includes: extracting the remelting zone structure data based on the calculation results, including the remelting zone distribution shape and location, as well as the distribution information of each remelting zone and coke window; extracting the velocity, temperature, pressure, and concentration distribution data of the gas phase and solid phase in the furnace based on the calculation results, and generating corresponding cloud maps in the furnace; integrating the remelting zone information and the velocity, temperature, pressure, and concentration data of each phase, and performing 2D and 3D visualization display through comparison tables, distribution curves, distribution bar charts, distribution cloud maps, and related change trend charts.

[0012] This application provides a rapid calculation and prediction device for the softening zone of a blast furnace. The device includes: a data processing module for acquiring real-time operating data of the blast furnace under its current operating conditions and standardizing the real-time operating data to generate standardized input parameters; a geometric model construction module for constructing a geometric model of the blast furnace's inner contour based on the standardized input parameters, and dividing the geometric model into multiple sets of meshes with different resolutions, the multiple sets of meshes including at least a first resolution mesh and a second resolution mesh, the first resolution mesh being used for initial trial calculations to quickly obtain an initial field for accurate calculation, and the second resolution mesh being used for accurate result calculation to finally obtain accurate calculation results; and a calculation model construction module for constructing a predictive calculation model for the softening zone and dynamically loading the standardized input parameters into the multiple sets of meshes. In this system, computational examples with different resolution grids are constructed. The data solving module is used to select a hierarchical solving strategy based on the standardized input parameters corresponding to the current operating condition and the matching situation of historical samples. This includes three solving strategies: coarse grid trial calculation - data mapping initialization - fine grid accurate calculation, historical sample matching - data mapping initialization - fine grid accurate calculation, and historical sample matching - matching result reuse. It also combines a gas-solid single-core interactive spatial block parallel partitioning strategy to accelerate the iterative solving process of coarse and fine grids and quickly obtain the accurate calculation results corresponding to the current operating condition. The storage and output module is used to store the standardized input parameters and accurate calculation results of the current operating condition and store them in the form of key-value pairs in the operating condition sample database for rapid retrieval and matching of subsequent operating conditions, and synchronously visualize the multi-physics distribution results in the furnace.

[0013] This application provides an electronic device, which includes: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, enable the electronic device to implement the rapid calculation and prediction method for blast furnace softening zone as described above.

[0014] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the rapid calculation and prediction method for the blast furnace softening zone as described above.

[0015] The beneficial effects of the present invention are as follows: The present invention proposes a rapid calculation and prediction method for the blast furnace softening zone. This method improves the efficiency and intelligence level of blast furnace softening zone prediction by constructing a multi-resolution grid collaborative calculation framework and a dynamic working condition matching mechanism, and by adopting a hierarchical solution strategy and an efficient parallel solution method to accelerate the iterative solution process. Specifically, this method utilizes standardized input parameters to establish a geometric model and divide it into multiple sets of meshes. It combines rapid trial calculations using an initial low-resolution mesh to obtain a superior initial field, which is then used for accurate solutions on a high-resolution mesh. This effectively balances computational speed and result accuracy, overcoming the contradiction between efficiency and accuracy inherent in traditional single-mesh systems. Through intelligent matching and interpolation mapping strategies using historical sample operating data, it achieves adaptive switching between different hierarchical solution strategies. When the matching degree is insufficient, it rapidly generates the initial field through trial calculations using a low-resolution mesh; when the matching degree meets the standard, it obtains the initial field in real time and can even directly reuse historical results, significantly reducing redundant computational resource consumption. Its spatial block parallel strategy for gas-solid single-core interaction further accelerates iterative solutions on both low-resolution and high-resolution meshes, optimizing the communication consumption performance bottleneck between the gas-solid two-phase flow computation cores. Furthermore, the historical data system with synchronous storage and cluster analysis not only provides a dynamic sample library for operating condition matching but also enables real-time monitoring and decision support of the multi-physics field distribution within the furnace through visual output. This significantly enhances the predictive ability and control accuracy of the blast furnace operation process, providing efficient and reliable technical support for the intelligent optimization of complex metallurgical systems. Attached Figure Description

[0016] 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. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] In the attached diagram:

[0018] Figure 1 This is a schematic diagram illustrating the implementation environment of a rapid calculation and prediction method for the softening zone of a blast furnace provided in an embodiment of the present invention.

[0019] Figure 2 This is a flowchart of a rapid calculation and prediction method for the softening zone of a blast furnace provided in one embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of the geometric model of the inner contour of a blast furnace provided in one embodiment of the present invention;

[0021] Figure 4a This is a schematic diagram of a first-resolution grid division provided in one embodiment of the present invention;

[0022] Figure 4b This is a schematic diagram of a second resolution grid division provided in one embodiment of the present invention;

[0023] Figure 5a This is a schematic diagram showing the distribution structure of ore-coke in a blast furnace, as well as the shape and location distribution of the softening zone, provided in one embodiment of the present invention.

[0024] Figure 5b This is a visualization of the flow velocity distribution of blast furnace gas at different locations provided in one embodiment of the present invention.

[0025] Figure 5c This is a visualization of the temperature distribution cloud map of blast furnace gas at different locations provided in one embodiment of the present invention;

[0026] Figure 5d This is a visualization of the gas utilization rate distribution cloud map at different locations inside the blast furnace provided in one embodiment of the present invention;

[0027] Figure 6 This is a schematic diagram showing the long-term and historical trends of the calculation results of the soft fusion strip prediction calculation model provided in one embodiment of the present invention.

[0028] Figure 7 This is a schematic diagram illustrating the steps of a rapid calculation and prediction method for the softening zone of a blast furnace, as shown in an exemplary embodiment of this application.

[0029] Figure 8 This is a block diagram illustrating a rapid calculation and prediction apparatus for the blast furnace softening zone, as shown in an exemplary embodiment of this application.

[0030] Figure 9 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0031] The attached figures are labeled as follows:

[0032] h1 - Distance from the center line of the taphole to the center line of the tuyeres; h2 - Height of the furnace belly; h3 - Height of the furnace waist; h4 - Height of the furnace body; h5 - Height of the furnace throat; hr - Distance from the center line of the tuyeres to the lower edge of the furnace belly; d1 - Diameter of the furnace hearth; d2 - Diameter of the furnace waist; d3 - Diameter of the furnace throat; a - Angle of the dead material column. Detailed Implementation

[0033] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0034] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0035] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0036] First, it's important to clarify that coupling and decoupling are mathematical concepts used to describe the relationships between two or more variables. In specific problems, coupling means that two or more variables are interconnected, and a change in one will significantly affect the others. Decoupling, on the other hand, means treating the two or more coupled variables separately and in stages to better understand their relationship. In control system theory, decoupling is a common control strategy used to reduce the mutual influence between system variables and improve the system's stability and reliability.

[0037] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating the implementation environment of a rapid calculation and prediction method for the softening zone of a blast furnace provided in an embodiment of the present invention.

[0038] like Figure 1 As shown, the implementation environment of the rapid calculation and prediction method for the blast furnace softening zone includes a data acquisition module 101 and a computer device 102. The two interact with each other through industrial communication protocols (such as OPC UA and MQTT) to collaboratively complete the real-time prediction of the softening zone state.

[0039] The data acquisition module 101 incorporates a multi-source sensor array, including thermocouples for monitoring the temperature field, sensors for monitoring gas phase pressure, flow meters for monitoring gas phase flow rate, weighing instruments for monitoring solid phase weight, level detectors for monitoring material surface distribution, chemical analyzers for monitoring raw material composition, and cooling water flow meters for monitoring furnace wall heat exchange. This allows for real-time acquisition of key physical quantities relevant to blast furnace operation. The module includes built-in signal conditioning units for filtering and analog-to-digital conversion to preprocess the raw data, remove outliers, and perform standardization (such as normalization and discrete parameter encoding) to generate standardized input parameters. Data is transmitted to a computer via an industrial bus or wireless network, supporting continuous data acquisition under dynamic operating conditions and providing a visual interface for maintenance personnel to monitor the blast furnace's operating status.

[0040] Computer device 102, as the core computing unit, integrates a blast furnace geometric modeling engine, a multi-resolution mesh generation module, a CFD calculation engine, and a visualization analysis platform, and is responsible for triggering and executing the prediction process. After the data acquisition module 101 uploads standardized input parameters, the computer device first constructs a blast furnace inner contour model based on the geometric parameters and divides it into a multi-level structure containing a first-resolution mesh (coarse mesh) and a second-resolution mesh (fine mesh). The coarse mesh is used for rapid initialization calculations, and the fine mesh inherits the results of the coarse mesh through interpolation mapping. Iteratively solving the spatial block parallel partitioning strategy of gas-solid single-core interaction, loading the source terms of the control equations (gas phase flow equation, solid phase transport equation, and gas-solid coupling source terms), and generating multi-physics field distribution information (such as the soft melt zone contour, velocity field, temperature field, and pressure field) through decoupling and coupling iteration.

[0041] It should be noted that during the calculation process, the computer equipment synchronously calls the historical data storage module to extract the feature vectors of the matching working condition samples and the calculation results, comparing them with the current deviation values ​​(such as the position error of the remelting zone and the difference in flow velocity distribution) to help optimize the prediction accuracy. Finally, the prediction results are integrated into visual information through a remelting zone structure distribution map, gas phase and solid phase velocity, pressure, temperature, and concentration cloud maps, as well as data tables, distribution curves, and trend change graphs, and output to a local or cloud platform to support remote monitoring and production decision optimization. This environment significantly improves prediction efficiency through multi-resolution grid collaborative computing and historical sample library matching strategies, providing reliable support for real-time control of blast furnace smelting conditions.

[0042] like Figure 2 As shown, in an exemplary embodiment, the rapid calculation and prediction method for the blast furnace softening zone includes at least steps S210 to S250, which are described in detail below:

[0043] Step S210: Obtain real-time operating data of the blast furnace under current operating conditions, and perform standardized processing on the real-time operating data to generate standardized input parameters.

[0044] In one embodiment of this application, multi-source heterogeneous data is collected in real time through a blast furnace control system and sensor network, including production-related data such as geometric parameters, raw material properties, charging parameters, air supply parameters, and furnace wall parameters. Geometric parameters include static structural data such as blast furnace diameter and segment height; raw material properties include ore / coke composition (FeO, FeO, etc.), density, particle size distribution, and bulk density; charging parameters include charge position, ore batch weight, coke batch weight, charging angle, and chute speed; air supply parameters include air volume, temperature, humidity, pulverized coal injection rate, and oxygen enrichment rate; and furnace wall parameters include cooling water flow rate, temperature, thermocouple temperature, and slag thickness.

[0045] After collecting the raw data, it is first preprocessed, including but not limited to cleaning and dynamic parameter calculation. The specific processing flow includes at least the following steps: outlier removal, data smoothing, and dynamic parameter calculation. Outlier removal uses moving average filtering or random forest anomaly detection on data with large fluctuations, such as thermocouple temperature, to remove outliers exceeding reasonable ranges (e.g., temperatures >1000℃ or <20℃), and replaces the outlier data with neighboring values. Data smoothing uses exponential smoothing or sliding window averaging to reduce noise on data with large fluctuations, such as feed flow rate and air volume. Dynamic parameter calculation uses chemical analysis data or online detection equipment to dynamically solve parameters such as the iron oxide (FeO, FeO) content of the ore using the mass conservation equation.

[0046] Then, the preprocessed data is further standardized, which includes at least the following: normalization of continuous parameters, that is, mapping continuous variables such as air volume and temperature to the [0,1] interval; encoding of discrete parameters, that is, using One-Hot encoding or segmented mapping for material distribution mode (upright, inverted, mixed distribution); and weighting of key factors, that is, performing hierarchical standardization of parameters according to their influence weight (geometric structure > raw material properties > material distribution parameters > air supply parameters > furnace wall boundary), and performing more refined standardization division for high-weight factors (such as raw material properties).

[0047] Finally, the standardized data is transmitted to the CFD model module via OPC UA or MQTT interface in JSON / XML format to ensure input compatibility.

[0048] One-hot encoding is a method for converting discrete parameters (especially categorical variables) into a numerical form that is easier for machine learning algorithms to process. JSON (JavaScript Object Notation) represents a lightweight data exchange format; XML (eXtensible Markup Language) represents an extensible markup language; OPC UA (Open Platform Communications Unified Architecture) is an open communication protocol for industrial automation; and MQTT (Message Queuing Telemetry Transport) is a lightweight messaging protocol based on a publish / subscribe model.

[0049] Step S220: Establish a geometric model of the inner contour of the blast furnace based on standardized input parameters, and divide the geometric model into multiple sets of meshes with different resolutions. The multiple sets of meshes include at least a first resolution mesh and a second resolution mesh. The first resolution mesh is used for initial trial calculations, and the second resolution mesh is used for precise calculations.

[0050] In one embodiment of this application, a geometric model of the blast furnace inner contour is established based on standardized input parameters, including: constructing a basic inner contour of the blast furnace body based on static geometric parameters in the standardized input parameters, the static geometric parameters including the blast furnace segment height and diameter; generating dynamically changing feature structures in the basic inner contour by combining dynamic geometric parameters in the standardized input parameters, the dynamic geometric parameters including the shape of the material surface position, the size of the swirling zone contour, the morphology of the dead material column, and the height of the slag and iron molten metal surface; and integrating the basic inner contour and feature structures to generate a complete computational domain for predicting the softening zone, the complete computational domain for predicting the softening zone covering the axial spatial range from the furnace throat to the hearth.

[0051] In one specific embodiment of this application, based on static geometric parameters such as blast furnace segment height and diameter in the standardized input parameters, the basic inner contour of the blast furnace body is first constructed using a 3D modeling engine. This contour defines the geometric boundaries of key areas such as the throat, body, waist, and hearth in a parametric manner to ensure consistency with the actual blast furnace structure. Subsequently, combined with dynamic geometric parameters such as the shape of the material surface position, the size of the swirling zone contour, the morphology of the dead material column, and the slag-iron liquid level height in the standardized input parameters, a parametric modeling method is used to dynamically generate feature structures in the basic inner contour: the shape of the material surface position is generated through real-time material level detection or calculation using a material distribution model; the swirling zone contour is dynamically adjusted based on airflow distribution characteristics; the morphology of the dead material column is modeled by inferring from historical material stacking states; and the slag-iron liquid level height is mapped in real time based on hearth liquid level sensor data or slag-iron reserve model results. Finally, Boolean operations are used to spatially fuse the basic inner contour with the dynamic feature structure to generate a complete computational domain covering the entire axial space from the furnace throat to the hearth. This computational domain not only retains the static structural framework of the blast furnace, but also embeds dynamic geometric features that reflect the current working conditions, such as the undulation of the material surface and the expansion of the swirling zone. This provides a high-fidelity modeling foundation for subsequent prediction of the softening zone, which includes real physical boundary conditions and real-time state information. This significantly improves the adaptability of the computational domain to complex metallurgical processes and the reliability of prediction.

[0052] Figure 3 This is a schematic diagram of the geometric model of the inner contour of a blast furnace provided in one embodiment of the present invention.

[0053] In one specific embodiment of this application, the inner contour of the blast furnace is established based on real-time collected geometric modeling information, such as static data like the height and diameter of blast furnace sections, and a computational geometric model of the softening zone is established by integrating dynamic data such as the shape of the material surface, the size of the swirling zone, the morphology of the dead material column, and the slag-iron molten surface. Figure 3 As shown.

[0054] In one embodiment of this application, based on blast furnace geometric parameters such as furnace diameter, hearth radius, and throat height from standardized input parameters, a 3D modeling engine is used to construct a geometric model of the blast furnace's internal contour. This model accurately reproduces the internal spatial structure and boundary conditions of the blast furnace. Subsequently, a multi-level meshing strategy is adopted to generate multiple mesh systems for the geometric model, including a first-resolution mesh (coarse mesh) and a second-resolution mesh (fine mesh). The first-resolution mesh covers the entire computational domain with a sparser mesh density, used to quickly complete initial rapid calculations and generate approximate solutions for the gas-solid coupling field with low computational resource consumption, serving as the mapping data source for the initial field of the second-resolution mesh. The second-resolution mesh employs mesh refinement processing in key areas such as the soft melting zone transition area and gas distribution channels, and combines the interpolated data obtained from the first-resolution mesh results to map and generate the initial field, initiating the high-precision calculation process. This multi-level mesh collaborative mechanism shortens the convergence time through rapid calculations with the coarse mesh, while relying on the local fine-scale modeling of the fine mesh to ensure the calculation accuracy of key physical fields such as temperature gradient and gas phase flow velocity, ultimately achieving a balance between computational efficiency and result reliability.

[0055] Figure 4a This is a schematic diagram of a first-resolution grid division provided in one embodiment of the present invention; Figure 4b This is a schematic diagram of a second resolution grid division provided in one embodiment of the present invention.

[0056] In one specific embodiment of this application, two sets of meshes with different numbers of meshes are used to divide the established geometric model. Since the soft melt band simulation model is a 2D axisymmetric model, the number of coarse meshes is 10,000 (e.g., ...) to ensure computational accuracy and balance computational efficiency. Figure 4a As shown), the number of fine grids is 60,000 (e.g. Figure 4b (As shown).

[0057] Step S230: Construct a predictive calculation model for the soft fusion zone, select different resolution grids and dynamically load standardized input parameters through a standardized interface to obtain calculation examples for different resolution grids.

[0058] In one embodiment of this application, the construction of a predictive calculation model for the remelting zone includes: constructing the predictive calculation model for the remelting zone, which includes: establishing a model framework based on a preset dense-phase gas-solid two-phase flow decoupling processing method, the specific steps of which are as follows: generating a corresponding background grid based on the selected resolution grid, which is used to store the mass, momentum and energy data of different phase components; performing flow tracking calculation of furnace charge particles based on the viscous flow method, recording the spatial position, batch and attribute information of the furnace charge, and mapping the results of the flow tracking calculation to the background grid nodes; using a porous media model to calculate the gas flow resistance and heat transfer in the particle bed, the resistance coefficient of the flow resistance is dynamically updated based on the raw material attributes and material distribution information in the standardized input parameters; setting source terms to realize the heat transfer, chemical reaction and phase change data transfer between the gas-solid-liquid phases, wherein the results of the source term calculation are used for the coupling of mass, momentum and energy between phases.

[0059] In one embodiment of this application, a predictive calculation model for the softening zone is constructed based on a preset dense-phase gas-solid two-phase flow decoupling processing method. The specific implementation process is as follows: First, a background grid is generated by moving and backing up the selected resolution grid (such as the first resolution grid or the second resolution grid). This background grid serves as a global data container, used to store the mass, momentum, and energy field data of the gas and solid phase components, and achieves efficient transmission of multi-physics information through a node interpolation algorithm. Subsequently, a viscous flow method is used to discretize and track the furnace charge particles, recording the spatial location, batch affiliation, and attribute parameters such as particle size and density of the particles. The tracking results are dynamically mapped to the background grid nodes to form a spatial distribution feature map of the particle bed. On this basis, a porous media model is introduced to simulate the gas flow resistance and heat transfer behavior inside the particle bed. The flow resistance coefficient is dynamically updated in real time based on the Ergun formula according to the raw material attributes such as bulk density, porosity, particle size, and sphericity in the standardized input parameters, so as to accurately reflect the influence of different raw material ratios and packing on the gas flow distribution. Finally, decoupling and coupling interactions between the gas and solid phases are achieved through source term settings: the mass source term is used to consider the mass exchange between the gas and solid phases caused by chemical reactions within the blast furnace; the momentum source term is used to realize the momentum coupling of the gas-solid flow; and the energy source term is used to consider the heat transfer between the gas and solid phases and the heat of chemical reactions within the furnace. This model framework, through decoupling and coupling strategies, balances computational efficiency with the integrity of the physical processes, providing high-precision numerical support for predicting the multiphase flow and heat transfer characteristics of the blast furnace softening zone.

[0060] In one embodiment of this application, considering that the softening zone is a comprehensive reflection of the control of the top charging method of the blast furnace, the air supply regulation of the lower tuyeres, the furnace wall cooling operation system, and the iron tapping control of the hearth, and that there are complex changes such as multiphase cross-flow, heat transfer, chemical reactions, and phase transformations within the furnace, the established softening zone calculation model is a CFD simulation model involving high temperature, high pressure, multiphase, multistate, and multi-field strongly coupled solutions. Modeling based on conventional multiphase flow methods suffers from problems such as high computational difficulty and poor robustness of the program solution.

[0061] In a specific embodiment of this application, for multi-process and multi-field coupled solutions, a decoupling method suitable for dense phase gas-solid two-phase flow is proposed to construct a soft melting zone simulation model: the calculation model is divided into multiple sub-modules, including silo unloading model, furnace top charging model, furnace body mirror model, air volume distribution model, swirling zone calculation model, furnace wall heat exchange model, hearth erosion model, dead material column model, slag and iron storage model, chemical reaction model, gas-solid sub-model, etc. The silo unloading model, furnace top charging model, and furnace body mirror model realize the tracking calculation of raw materials entering the furnace, and obtain information on furnace charge flow and mixing; the air volume distribution model and swirl zone calculation model realize the calculation of tuyer air supply operation, and obtain information on the state of the gas entering the furnace; the furnace wall heat transfer model realizes the calculation of the heat transfer state of the furnace wall, and obtains information on the thickness of the slag skin and the heat flow of the furnace wall; the hearth erosion model, dead material column model, and slag-iron storage model calculate the state of molten iron and refractory materials in the hearth, and obtain information such as the outline of the dead material column and the position of the slag-iron liquid surface; the chemical reaction model mainly considers the iron ore reduction reaction, uses the three-interface unreacted shrinkage core model to describe the reaction kinetics of iron ore particles, and uses the Ranz-Marshall equation to describe the convective heat transfer between the gas and solid phases, realizes the calculation of the iron ore reduction process in the furnace, and obtains information such as the heat of reaction, ore composition, and state in the furnace; finally, the gas-solid sub-model realizes the access, coupling, and integrated calculation of multi-process parameters, and constructs the soft melting zone calculation model.

[0062] The decoupling method applicable to dense phase gas-solid two-phase flow is based on "charge tracking + background grid + porous media + source term setting" to solve complex processes such as cross flow of gas, charge, and droplets, dynamic heat transfer, chemical reaction and phase change. The furnace charge tracking technology uses a viscous flow method to track and calculate the raw materials entering the furnace, replacing the computationally intensive and complex Discrete Element Method (DEM) simulation. This technology records the spatial, flow, and property information of the corresponding furnace charge, including but not limited to its location, batch, type, composition, particle size, and porosity. A background grid is used to map the corresponding furnace charge information onto a preset grid for storage and calculation. The porous media model considers the influence of the furnace charge on the gas flow distribution within the furnace through the charge's resistance characteristics, thereby achieving momentum coupling between the gas and solid phases after gas-solid decoupling. The characteristic parameters of the porous media, such as the viscous drag coefficient and inertial drag coefficient, can be obtained experimentally or theoretically solved and corrected based on the test data of the raw materials entering the furnace using the Ergun formula. The source term is used to realize the coupling of mass, momentum, and energy between the gas and solid phases. The mass source term considers the mass exchange between the gas and solid phases caused by chemical reactions within the blast furnace, the momentum source term realizes the momentum coupling of the gas-solid flow, and the energy source term considers the heat transfer and chemical reaction heat between the gas and solid phases within the furnace.

[0063] It should be noted that the Ranz-Marshall equation is a classical empirical formula describing the convective heat transfer coefficient in gas-solid two-phase flow; the Ergun equation is a classical semi-empirical formula describing the relationship between pressure drop and flow velocity in porous media. Both are widely used in flow heat transfer calculations in chemical, metallurgical, and energy fields (such as blast furnace burden, fixed-bed reactors, and filter media).

[0064] It is understandable that the CFD-based computational model of the soft melt zone can explain the correlation between different input parameters and the smelting results at the mechanistic level. In one embodiment of this application, for the aforementioned coarse and fine meshes, preprocessed real-time data is accessed through a standardized interface to achieve dynamic updates of raw material properties and dynamic loading of boundary conditions. It should be noted that the two mesh examples are identical in all calculation settings except for the number of meshes, and are used for subsequent initialization and accurate calculations for this working condition.

[0065] Step S240 involves selecting a tiered solution strategy based on the standardized input parameters corresponding to the current operating condition and the matching status of historical samples: when the matching degree is insufficient to meet the basic matching threshold, a collaborative solution strategy of "coarse grid trial calculation - data mapping initialization - fine grid accurate calculation" is adopted; when the matching degree is between the basic matching threshold and the perfect matching threshold, a collaborative solution strategy of "historical sample matching - data mapping initialization - fine grid accurate calculation" is adopted; when the matching degree meets the perfect matching threshold, a real-time prediction strategy of "historical sample matching - matching result reuse" is adopted. The basic matching threshold and the perfect matching threshold are set values ​​obtained based on the running data under historical operating conditions. For the calculation and solution of calculation examples with different resolution grids, a spatial block parallel partitioning strategy of gas-solid single-core interaction is adopted to accelerate the iterative solution process. After the calculation converges, the calculation result corresponding to the current operating condition is obtained.

[0066] In one embodiment of this application, the hierarchical solution strategy selection includes: for standardized input parameters under the current working condition, screening historical samples based on preset rules to obtain the best matching sample, comprehensively calculating the matching degree between the two, and determining the solution strategy based on the relationship between the matching degree and the preset matching threshold; wherein, determining the solution strategy based on the relationship between the matching degree and the preset matching threshold includes: when the matching degree is less than the basic matching threshold, using sparse first-resolution grid cases to perform initial trial calculations, quickly obtaining convergence results as first interpolation data, mapping them as data sources to dense second-resolution grid cases as the initial field, and then performing precise calculations on the second-resolution grid cases; if the matching degree is greater than or equal to the basic matching threshold and less than the perfect matching threshold, then obtaining the matching results in real time through historical sample queries as second interpolation data. The data source is mapped to a dense second-resolution grid as the initial field, and then the precise calculation of the second-resolution grid is carried out. If the matching degree is greater than or equal to the complete matching threshold, the matching result is obtained in real time through historical sample query as the final precise calculation result of the current working condition, omitting the process of interpolation data acquisition, data mapping initialization and fine grid precise calculation. Before mapping the acquired first or second interpolation data to a dense second-resolution grid as the initial field, it is necessary to check whether there is a spatial position deviation between the data source and the current geometric model. If there is a deviation and it exceeds the preset range, the coordinate scaling correction of the data source is performed based on the blast furnace axial segmentation parameters. The corrected data source is mapped to the second-resolution grid, and the mapped data is set as the initial field of the second-resolution grid.

[0067] In one embodiment of this application, the standardized input parameters of the current working condition are screened and matched with the corresponding parameters in the historical sample library according to preset rules, and the comprehensive matching degree is calculated. If the matching degree does not reach the preset basic matching threshold, a coarse grid (first resolution grid) is used to conduct a rapid trial calculation to generate the first interpolation data as the initial field; if the matching degree meets the preset basic matching threshold but does not meet the preset complete matching threshold, the calculation result corresponding to the matching sample is directly extracted as the second interpolation data; if the matching degree meets the preset complete matching threshold, the calculation result corresponding to the matching sample is directly extracted as the accurate result output of the current working condition. Subsequently, the acquired first or second interpolation data is used as the initial field mapping data source of the second resolution grid example, and the spatial position deviation between the data source and the current geometric model is detected. If there is a deviation and it exceeds the preset range, the source data is corrected by coordinate scaling based on the blast furnace axial segmentation parameters to eliminate the physical field distribution offset caused by geometric structure differences. Finally, the corrected physical field distribution data is mapped to the second resolution grid and set as the initial field of the grid example to start the accurate calculation process. This strategy, through hierarchical matching and dynamic correction mechanisms, balances computational efficiency and accuracy, making it suitable for the real-time requirements of CFD simulation of blast furnace softening zones.

[0068] The basic matching threshold and the complete matching threshold are both set values ​​obtained based on operating data under historical conditions. This application does not impose any restrictions on the setting method or its specific value.

[0069] In one specific embodiment of this application, different hierarchical solution strategies are adopted for different working conditions, as follows:

[0070] First, based on the standardized input parameters of the current operating conditions (such as raw material properties, fabric parameters, and air supply parameters), historical samples are screened and matched according to preset rules. For example, the best historical matching sample is found using a weighted distance metric, and the comprehensive matching degree value is calculated to determine whether the matching degree meets the preset matching threshold condition. For example, with the current ore grade of 60% and the air supply volume of 5000 Nm³, 3 / min, oxygen enrichment rate of 3.5%, air supply temperature of 1240℃, and material distribution data: ore batch weight 105 tons / batch, coke batch weight 18.2 tons / batch, coke briquettes 1.95 tons / batch, material distribution matrix is ​​as follows: Under certain conditions, the weights of five major categories of factors, namely geometric structure, raw material properties, fabric parameters, air supply parameters, and furnace wall boundary, are determined by random forest to be 0.25, 0.25, 0.2, 0.2, and 0.1 respectively. The weighted distance is calculated based on the weights of each factor, and the historical best-matched sample is quickly located as sample N after matching to the nearest cluster. The matching degree of the current working condition and the input parameters of the historical working condition sample N is M, the preset basic matching threshold is X₁, and the complete matching threshold is X₂, where both X₁ and X₂ are greater than 0 and less than 1, and X₁ < X₂. For example, X₁ = 0.80 and X₂ = 0.98.

[0071] When there is no historical sample (M = 0) or the matching degree between the current working condition and the historical sample is insufficient, i.e., M < X₁, a collaborative solution strategy of "coarse grid trial calculation - data mapping initialization - fine grid accurate calculation" is adopted. For the coarse grid (with about 10,000 grid numbers) case, initialization settings are carried out through the initialization method Standard Initialization, and rapid calculation is carried out. After the calculation converges, the preliminary results of the velocity field, temperature field, pressure field, concentration field, etc. (.ip file or other formats) are output as the first interpolation data. The calculation consumption in this stage is less, and the obtained calculation results are relatively close to the final solution, providing a basis for the subsequent fine grid case (with about 60,000 grid numbers) calculation. Subsequently, the first interpolation data is mapped to the fine grid case through the linear interpolation method as the initial field data of the fine grid. Since the coarse grid and the fine grid cases have the same geometric structure at this time and there is no geometric deviation, no additional scaling preprocessing is required during the interpolation process, and the interpolation mapping error is controlled within ±2% to ensure the calculation accuracy. Finally, based on the mapped initial field, the accurate calculation of the fine grid is started. In the process of rapid trial calculation of the coarse grid and accurate calculation of the fine grid, a CPU parallel partition strategy of gas-solid data single-core interactive space partitioning is adopted to accelerate the iterative solution. The gas-solid data at the same position is assigned to the same process, reducing the communication time between cores and significantly improving the calculation efficiency. The calculation time-consuming under this hierarchical solution strategy is about 4 hours.

[0072] It should be noted that Standard Initialization (standard initialization) is a commonly used initial condition setting method in computational fluid dynamics (CFD) and numerical simulation, which is used to provide a reasonable initial guess value for the iterative solution process, thereby starting the calculation and accelerating convergence.

[0073] When the matching degree between the current working condition and the historical sample is between the basic matching threshold and the complete matching threshold, that is, X1 ≤ M < X2, a collaborative solution strategy of "historical sample matching - data mapping initialization - fine grid accurate calculation" is adopted. For the historical best-matching sample N of the current working condition, directly obtain the calculation result of sample N as the second interpolation data, map it to the fine grid case through linear interpolation method, and use it as the initial field data of the fine grid case, and directly start the fine grid accurate calculation. Under this hierarchical solution strategy, the coarse grid calculation process is omitted, and the calculation time is about 1.5 hours, meeting the on-site application requirements.

[0074] When the matching degree between the current working condition and the historical sample meets the complete matching threshold, that is, M ≥ X2, a real-time prediction strategy of "historical sample matching - reuse of matching results" is adopted. That is, for the historical best-matching sample N of the current working condition, directly obtain the calculation result of sample N as the accurate calculation result of the current working condition. Under this hierarchical solution strategy, the calculation processes of the coarse grid and the fine grid are omitted, and the calculation time is less than 1 minute, and real-time calculation prediction can be basically realized.

[0075] It should be noted that when querying historical samples that match the current working condition from historical data, usually the current working condition data is divided into different data dimensions, and feature vectors are generated for the information of each data dimension respectively. The data dimensions include but are not limited to key factors such as geometric structure, raw material properties, cloth parameters, air supply parameters, and furnace wall boundaries. The feature vectors of each dimension can be generated based on the quantization values of standardized input parameters or the values after normalization processing. In the matching process, various methods can be used to achieve rapid positioning and screening of historical samples, such as weighted distance matching, similarity measurement, clustering analysis matching, machine learning model matching, etc. The above methods can be flexibly selected according to actual needs, and this application does not limit the specific matching method, but only emphasizes the efficient and accurate matching of historical samples through multi-dimensional feature vectors and dynamic weight adjustment, so as to provide high-quality initial field data for the initialization calculation of the current working condition and significantly shorten the iterative solution time.

[0076] In an embodiment of the present application, a spatial block parallel partitioning strategy of gas-solid single-core interaction is adopted to accelerate the iterative solution process, including: each iterative solution includes decoupling and solving the velocity, pressure, temperature, and concentration information of the burden flow and gas flow, and then realizing the transfer and coupling solution of gas-solid phase data based on the background grid and source terms; in the process of multiple iterative solutions, monitor the residual of the calculation equation and the fluctuation of the monitored variables until they reach the preset convergence criteria respectively to obtain a stable calculation result of the coupling solution; among them, the iterative solution process includes dividing the grid space into blocks based on the preset gas-solid data single-core interaction principle, and following the principle of allocating gas-phase and solid-phase data at the same spatial position to the same calculation process.

[0077] In one specific embodiment of this application, the specific implementation process of accelerating iterative solution using a gas-solid single-core interactive spatial block parallel partitioning strategy is as follows:

[0078] In each iteration, the physical field information such as velocity, pressure, and temperature fields of the furnace charge flow and the gas flow are first decoupled, and updated values ​​for each field are generated. Then, based on the pre-built background grid data structure, the source term module completes the data transfer and coupling between the gas and solid phases: the momentum exchange term and heat transfer term in the gas phase flow calculation results are synchronized to the solid phase transport equation, while the drag coefficient, heat transfer coefficient, and chemical reaction of the solid phase particle bed are fed back to the gas phase control equation, realizing bidirectional coupled iteration.

[0079] During the multi-step iteration process, the calculation residuals (such as mass conservation error and energy balance error) and key monitoring variables (such as the root position of the softening zone and the average top temperature of the furnace top gas) are continuously monitored. When the residuals drop to the preset threshold (such as relative error < 1e-5) and the fluctuation range of the monitored variables is less than the set range (such as the fluctuation deviation of the average top temperature of the furnace top < 5℃), the calculation is determined to be converged and the iteration is terminated.

[0080] Then, in order to improve the efficiency of parallel computing, the grid domain is further divided into spatial blocks according to the preset gas-solid data single-core interaction principle: the blast furnace computing domain is divided into axial and radial blocks to ensure that gas phase grid nodes and solid phase particle tracking data in the same spatial location are assigned to the same computing process and processed directly by local storage and computing units, avoiding delays caused by cross-process communication.

[0081] It is understood that the strategy proposed in this embodiment significantly reduces the communication overhead in distributed computing through spatial locality optimization, while ensuring the integrity of the gas-solid coupling physical process, and ultimately achieves a significant improvement in the efficiency of parallel iterative solution.

[0082] In one specific embodiment of this application, to address the complexity and timeliness requirements of CFD simulation calculations for the softening zone of blast furnaces, a decoupling method suitable for large-particle-size, gas-solid two-phase flow problems is proposed. This method combines a spatial block parallel partitioning strategy with gas-solid single-core interaction, a multi-resolution grid hierarchical solution strategy, and a dynamic historical sample matching query strategy, significantly improving the efficiency of parallel iterative solution. The specific implementation is as follows:

[0083] This embodiment proposes a fast calculation method for the softening zone of a blast furnace based on CFD. This method combines a gas-solid single-core interactive spatial block parallel partitioning strategy, a multi-resolution grid hierarchical solution strategy, and a dynamic historical sample matching query strategy, significantly improving the iterative solution efficiency. Specifically:

[0084] Gas-solid single-core interactive parallel partitioning strategy: In the grid parallel partitioning stage, gas phase and solid phase data in the same spatial location are allocated to the same computing process, reducing the communication overhead between gas phase and solid phase data transmission cores, reducing parallel computing latency, and improving parallel computing efficiency.

[0085] Multi-resolution grid hierarchical solution strategy: Two sets of grids, coarse (10,000 grids) and fine (60,000 grids), are used for collaborative computation. When the matching degree between the current working condition and historical samples is insufficient, a rapid trial calculation is performed using the coarse grid. The convergence result of the coarse grid is interpolated and mapped to the initial field of the fine grid, replacing the calculation process after ordinary initialization using the fine grid, thus improving the efficiency of iterative solution. When the matching degree between the current working condition and historical samples is between the basic matching threshold and the perfect matching threshold, the best historical matching working condition result is obtained in real time by combining the historical sample database and interpolated and mapped to the initial field of the fine grid, omitting the rapid calculation process of the coarse grid, further shortening the iterative solution process and improving the efficiency of iterative solution. When the matching degree between the current working condition and historical samples meets the perfect matching threshold, the best historical matching working condition result is directly reused as the accurate calculation result of the current working condition, directly omitting the model iterative solution process, and finally achieving real-time calculation and prediction.

[0086] Historical sample dynamic matching query strategy: Based on cluster analysis (such as weighted K-means), the current working condition parameters (such as raw materials, fabric, air supply, etc.) are matched with the sample library for similarity. The best historical matching sample is identified and found. The corresponding standardized input parameters and calculation results are obtained, and the sample library is dynamically updated to cover more working conditions.

[0087] The convergence times of different strategies are compared below under 16-core conditions:

[0088] Traditional Metis parallel partitioning strategy: computation begins after fine mesh initialization throughout, with a single computation convergence time of approximately 20 hours;

[0089] Gas-solid single-core interactive parallel partitioning strategy: Calculation is performed after fine mesh initialization throughout the process, and the convergence time of a single calculation is shortened to 12 hours;

[0090] The gas-solid single-core interactive parallel partitioning strategy + multi-resolution grid collaborative solution strategy: coarse grid fast trial calculation - data mapping initialization - fine grid accurate calculation, the single calculation convergence time is reduced to 4 hours; historical sample matching - data mapping initialization - fine grid accurate calculation, the single calculation convergence time is further shortened to 1.5 hours; historical sample matching - matching result reuse, the single calculation time is shortened to within 1 minute, basically realizing real-time solution.

[0091] This method significantly improves computational efficiency by optimizing initialization settings, parallel computing strategies, and sample library learning, providing real-time support for blast furnace smelting status monitoring and operation optimization.

[0092] Therefore, by reasonably and quickly obtaining the initial field and combining it with the proposed efficient parallel partitioning strategy, the number of iteration steps and calculation cycle can be significantly shortened, the solution process can be accelerated, and rapid calculation and prediction of soft fused zones can be achieved.

[0093] Step S250: Synchronously store the standardized input parameters and accurate calculation results of the current working condition to the historical sample library for working condition data accumulation, clustering and matching query, and visualize the multi-physics field distribution results in the furnace.

[0094] In one embodiment of this application, storing standardized input parameters and accurate calculation results of the current operating condition to a historical sample database includes: extracting a set of key input parameters for the current operating condition, sorting each parameter based on a preset influence weight, generating a multi-dimensional feature vector according to the parameter priority sequence, wherein the set of key input parameters includes geometric structure, raw material properties, material distribution parameters, air supply parameters, furnace wall boundaries, and iron tapping parameters; setting a classification accuracy standard based on the preset influence weight, and classifying historical samples according to the multi-dimensional feature vector and the set accuracy standard; associating the multiphysics distribution results of the current operating condition with the input feature vector, storing them in the form of key-value pairs in the historical sample database of the operating condition, performing clustering and grouping, and adding retrieval identification information.

[0095] In one specific embodiment of this application, the process of storing the standardized input parameters and calculation results of the current operating condition is as follows:

[0096] First, a set of key input parameters is extracted from the current operating conditions. These include geometric parameters such as furnace segment dimensions, raw material properties such as particle size distribution and bulk density, material distribution boundary parameters such as material surface shape and batch distribution, air supply boundary parameters such as air volume and air temperature, furnace wall boundary parameters such as cooling water flow rate and cooling water inlet and outlet temperatures, and tapping parameters such as tapping cycle and taphole status. Based on preset influence weights (e.g., weighted by principal component analysis or expert experience), each parameter is ranked by importance, generating a multi-dimensional feature vector. This feature vector uses a parameter priority sequence as a basis to uniformly encode numerical parameters such as temperature and pressure, as well as discrete parameters such as raw material type codes, into a standardized format.

[0097] Subsequently, the feature vectors are classified according to the preset classification accuracy requirements (such as temperature error tolerance ±5℃, pressure error ±1kPa), and fuzzy clustering or K-means algorithm is used to divide the working condition categories to ensure that similar working conditions are grouped into the same subset.

[0098] Finally, the multiphysics distribution results of the current working condition, such as the soft melt zone profile, velocity field, and temperature field, are mapped to the corresponding feature vectors using key-value pairs. These mappings are then stored in a historical sample database of the working condition in structured data formats such as JSON or HDF5. During storage, timestamps, working condition tags, cluster numbers, and other retrieval identifiers are added synchronously. Indexing optimization techniques (such as B+ tree indexes and inverted indexes) are used to improve the retrieval efficiency of subsequent working condition matching, providing data support for rapid sample matching and prediction model optimization under dynamic working conditions.

[0099] In one embodiment of this application, the results of multiphysics distribution in the furnace are visualized and displayed, including: extracting the structure data of the remelting zone based on the calculation results, including the distribution shape and location of the remelting zone, as well as the specific distribution information of each remelting zone and coke window; extracting the velocity, temperature, pressure, and concentration distribution data of the gas phase and solid phase in the furnace based on the calculation results, and generating corresponding cloud maps in the furnace; integrating the remelting zone information and the velocity, temperature, pressure, and concentration data of each phase, and displaying them in 2D and 3D through comparison tables, distribution curves, distribution bar charts, distribution cloud maps, and related change trend charts.

[0100] In a specific embodiment of this application, the specific implementation process for visualizing and displaying the multiphysics field distribution results inside the furnace is as follows:

[0101] Based on the CFD calculation results, the structural data of the remelting zone is extracted, including its distribution shape (such as inverted V, V or W) and axial distribution position. The geometric features of the remelting zone and the focal window (such as number of layers, height position, length, thickness and distribution of intercalation layers) are analyzed layer by layer. A three-dimensional contour model of the remelting zone is generated by a three-dimensional reconstruction algorithm.

[0102] Simultaneously extract velocity, temperature, pressure, and concentration field data of the gas phase (gas flow) and solid phase (burden particles) in the furnace, and use interpolation algorithms to generate continuously distributed cloud maps (e.g., velocity cloud map shows the gas flow velocity gradient, temperature cloud map reflects heat energy distribution, pressure cloud map presents gas flow resistance characteristics, and concentration cloud map tracks the spatiotemporal changes of gas components such as CO2 or H2). Integrate the remelting zone structure information and multi-physics field data, and display them through multi-dimensional visualization methods: In the 2D view, use comparison tables to present the differences in remelting zone shape parameters under different operating conditions, and use distribution curves (e.g., axial temperature distribution curve, radial pressure gradient curve) and bar charts (e.g., statistics of the thickness of each remelting zone layer) to intuitively compare key indicators; In the 3D view, combine three-dimensional dynamic rendering technology to overlay the remelting zone outline with the gas-solid phase field cloud map, supporting rotation, scaling, and slicing operations, and reveal the dynamic evolution law through related change trend charts (e.g., remelting zone displacement trajectory under time series, gas flow velocity fluctuation curve).

[0103] Ultimately, all visualization results are integrated and displayed through an interactive interface (such as a web interface or a large screen system), supporting users to switch between observation modes from multiple dimensions. Combined with historical operating data, it provides visual prompts for abnormal status warnings and operational strategy optimization suggestions, providing an intuitive and scientific basis for real-time monitoring and decision-making in blast furnace smelting.

[0104] Figure 5a This is a schematic diagram showing the distribution structure of ore-coke in a blast furnace, as well as the shape and location distribution of the softening zone, provided in one embodiment of the present invention. Figure 5b This is a visualization of the flow velocity distribution of blast furnace gas at different locations provided in one embodiment of the present invention. Figure 5c This is a visualization of the temperature distribution cloud map of blast furnace gas at different locations provided in one embodiment of the present invention; Figure 5d This is a visualization of the gas utilization rate distribution at different locations within a blast furnace, provided in one embodiment of the present invention. Figure 6 This is a long-term and historical trend of the calculation results of the soft fusion band prediction calculation model provided in one embodiment of the present invention.

[0105] In one specific embodiment of this application, the model calculation results and collected raw production data are displayed in real time on a client or web page, visualized through quantitative data, tables, curves, 2D / 3D images, reports, etc. In one specific embodiment of this application, a web-based web page is used for display. The ore-coke bed distribution structure within the blast furnace, as well as the shape and location distribution of the softening zone, are shown below. Figure 5a As shown in the figure; the velocity distribution cloud map of the gas flow at different locations inside the blast furnace is as follows. Figure 5b As shown; Temperature distribution cloud map of gas at different locations inside the blast furnace. Figure 5c As shown in the figure; the distribution cloud map of gas utilization rate at different locations inside the blast furnace is as follows. Figure 5d As shown. Furthermore, the long-term and historical trends of the calculation results from its remelting zone prediction model can be further displayed, as shown in the following example. Figure 6 As shown.

[0106] The case results show that the model in this application has high computational accuracy and can truly reflect the dynamic changes in the actual smelting state of the blast furnace.

[0107] Figure 7 This is a schematic diagram illustrating the steps of a rapid calculation and prediction method for the softening zone of a blast furnace, as shown in an exemplary embodiment of this application.

[0108] In one specific embodiment of this application, such as Figure 7As shown, the complete process of a rapid calculation and prediction method for the blast furnace softening zone based on computational fluid dynamics (CFD) is presented. Its core objective is to achieve efficient prediction and visualization analysis of the blast furnace softening zone state by standardizing the processing of real-time running data and using a multi-resolution grid hierarchical solution strategy.

[0109] The process begins with data acquisition and standardization. Key physical quantities such as temperature field, gas pressure, and material distribution during blast furnace operation are obtained through multi-source sensors. Standardized input parameters are generated through filtering, normalization, and discrete parameter encoding, providing a unified data foundation for subsequent modeling.

[0110] The process then proceeds to the geometric modeling and mesh generation stage. Based on standardized input parameters, a geometric model of the blast furnace's inner contour is constructed, and a multi-level structure is created, comprising a first-resolution mesh (coarse mesh) and a second-resolution mesh (fine mesh). The coarse mesh is used for rapid trial calculations to obtain optimal initial field mapping data, while the fine mesh inherits the initial field through interpolation mapping for high-precision calculations. In the CFD model construction phase, raw material properties and boundary conditions are dynamically loaded through a standardized interface. Combined with a gas-solid single-core interactive spatial block parallel partitioning strategy, efficient iterative calculations involving coupled multi-physics fields (velocity field, temperature field, and pressure field) are achieved.

[0111] The key to the process lies in determining the matching degree between the current operating conditions and historical samples, and selecting the solution strategy: when the matching degree does not reach the basic matching threshold, a coarse-grid example is used to quickly generate the first interpolation data; if the matching degree is between the basic matching threshold and the perfect matching threshold, the second interpolation data is obtained in real time based on the historical sample matching; the first or second interpolation data is mapped to a fine-grid example as the initial field, and then accurate calculation is performed; if the matching degree meets the perfect matching threshold, the operating condition results are directly reused as the accurate calculation results. Both the coarse-grid and fine-grid example solutions employ a spatial block parallel partitioning strategy with gas-solid single-core interaction to accelerate the iterative solution process, and finally, accurate calculation results are output after the convergence conditions are met.

[0112] The output stage not only includes multiphysics distribution results (such as the remelting zone outline and gas-solid phase field cloud map), but also updates the historical sample database of operating conditions through data accumulation and clustering, providing sample support for subsequent matching queries. Finally, all calculation results are visualized through an interface, presenting the remelting zone structure, velocity / temperature / pressure distribution, and dynamic trends in 2D / 3D form, supporting remote monitoring and production decision optimization. This workflow significantly improves the efficiency and accuracy of blast furnace remelting zone calculation and prediction through a multi-resolution grid hierarchical solution strategy and a gas-solid single-core interactive parallel partitioning strategy, providing reliable technical support for intelligent control of metallurgical processes.

[0113] Finally, it should be noted that this application proposes a rapid calculation and prediction method for the blast furnace softening zone. This method significantly improves the calculation and prediction speed and accuracy by constructing a softening zone prediction model based on computational fluid dynamics (CFD) and systematically optimizing key aspects affecting computational efficiency (initialization settings and iterative solution process). In the initialization phase, a multi-resolution grid hierarchical solution strategy is adopted: approximate solutions are quickly generated using a first-resolution grid example as the first interpolation data, or the best matching samples are selected based on clustering analysis results from a historical sample library (based on multi-dimensional feature vectors and dynamic matching algorithms) and their results are used as the second interpolation data. The first or second interpolation data is then mapped to a second-resolution grid example as the initial field, significantly shortening the convergence steps and convergence time of the second-resolution grid example. During the iterative calculation of both coarse and fine grid examples, a spatial block parallel partitioning strategy with gas-solid single-core interaction is introduced. Localized data allocation reduces inter-process communication overhead, and combined with a multi-physics decoupling-coupling iteration mechanism, efficient parallel processing of large-scale computational tasks is achieved. With the continuous accumulation of historical sample databases and the improvement of classification density, when the coverage of operating conditions reaches the range of changes in on-site operations, the system can directly reuse calculation results or quickly correct and generate approximate solutions based on highly matched samples, thereby compressing the prediction response time to the minute level and achieving near real-time or even real-time online prediction capabilities. This method continuously optimizes model performance through a dynamic learning mechanism, providing a scientific basis for monitoring the morphology of the softening zone, controlling gas distribution, and adjusting operating parameters during blast furnace smelting while ensuring computational accuracy. This effectively supports the intelligent operation and long-term stable operation of the production system.

[0114] Figure 8 This is a block diagram illustrating a rapid calculation and prediction apparatus for the softening zone of a blast furnace, as shown in an exemplary embodiment of this application. The apparatus can be applied to... Figure 1 The implementation environment shown is illustrated. This device can also be applied to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the device is applicable.

[0115] like Figure 8 As shown, the exemplary blast furnace softening zone rapid calculation and prediction device includes: a data processing module 810, a geometric model construction module 820, a calculation model construction module 830, a data solving module 840, and a storage and output module 850.

[0116] The data processing module 810 is used to acquire real-time operating data of the blast furnace under the current operating conditions, and to standardize the real-time operating data to generate standardized input parameters. The geometric model construction module 820 is used to construct a geometric model of the inner contour of the blast furnace based on the standardized input parameters, and to divide the geometric model into multiple sets of meshes with different resolutions. The multiple sets of meshes include at least a first resolution mesh and a second resolution mesh. The first resolution mesh is used for initial trial calculations to quickly obtain the initial field for accurate calculation, and the second resolution mesh is used for accurate result calculation to finally obtain accurate calculation results. The calculation model construction module 830 is used to construct a predictive calculation model for the softening zone, and to dynamically load the standardized input parameters into multiple sets of meshes to construct calculation examples with different resolution meshes. The data solving module 840 is used to select a hierarchical solving strategy based on the standardized input parameters corresponding to the current operating condition and the matching situation of historical samples. It includes three solving strategies: coarse grid trial calculation - data mapping initialization - fine grid accurate calculation, historical sample matching - data mapping initialization - fine grid accurate calculation, and historical sample matching - matching result reuse. It also combines a gas-solid single-core interactive spatial block parallel partitioning strategy to accelerate the iterative solving process of coarse and fine grids and quickly obtain the accurate calculation results corresponding to the current operating condition. The storage and output module 850 is used to store the standardized input parameters and accurate calculation results of the current operating condition and store them in the form of key-value pairs to the operating condition sample database for subsequent rapid retrieval and matching of operating conditions, and synchronously visualize the multi-physics field distribution results in the furnace.

[0117] It should be noted that the rapid calculation and prediction device for the blast furnace softening zone provided in the above embodiments and the rapid calculation and prediction method for the blast furnace softening zone provided in the above embodiments belong to the same concept. The specific operation methods of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the rapid calculation and prediction device for the blast furnace softening zone provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0118] Embodiments of this application also provide an electronic device, including: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, enable the electronic device to implement the rapid calculation and prediction method for the blast furnace softening zone provided in the above embodiments.

[0119] Figure 9 A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 9The computer system 900 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0120] like Figure 9 As shown, the computer system 900 includes a Central Processing Unit (CPU) 901, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 902 or programs loaded from Storage Unit 908 into Random Access Memory (RAM) 903, such as performing the methods described in the above embodiments. The RAM 903 also stores various programs and data required for system operation. The CPU 901, ROM 902, and RAM 903 are interconnected via a bus 904. An Input / Output (I / O) interface 905 is also connected to the bus 904.

[0121] The following components are connected to I / O interface 905: an input section 906 including a keyboard, mouse, etc.; an output section 907 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to I / O interface 905 as needed. Removable media 911, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 910 as needed so that computer programs read from them can be installed into storage section 908 as needed.

[0122] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 909, and / or installed from removable medium 911. When the computer program is executed by central processing unit (CPU) 901, it performs various functions defined in the system of this application.

[0123] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0125] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0126] Another aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a computer's processor, causes the computer to perform the rapid calculation and prediction method for the blast furnace softening zone as described above. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0127] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the rapid calculation and prediction method for the blast furnace softening zone provided in the various embodiments described above.

[0128] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A rapid calculation and prediction method for the softening zone of a blast furnace, characterized in that, The method includes: Acquire real-time operating data of the blast furnace under its current operating conditions, and perform standardized processing on the real-time operating data to generate standardized input parameters; Based on the standardized input parameters, a geometric model of the blast furnace inner contour is established, and multiple sets of meshes with different resolutions are divided for the geometric model. The multiple sets of meshes include at least a first resolution mesh and a second resolution mesh. The first resolution mesh is used for initial trial calculations, and the second resolution mesh is used for precise calculations. A predictive computational model for soft fusion bands is constructed. Different resolution grids are selected and the standardized input parameters are dynamically loaded through a standardized interface to obtain computational examples for different resolution grids. A tiered solution strategy is selected based on the standardized input parameters corresponding to the current operating condition and the matching status of historical samples: when the matching degree is less than the basic matching threshold, a collaborative solution strategy of "coarse grid trial calculation - data mapping initialization - fine grid accurate calculation" is adopted; when the matching degree is between the basic matching threshold and the perfect matching threshold, a collaborative solution strategy of "historical sample matching - data mapping initialization - fine grid accurate calculation" is adopted; when the matching degree meets the perfect matching threshold, a real-time prediction strategy of "historical sample matching - matching result reuse" is adopted; the basic matching threshold and the perfect matching threshold are set values ​​obtained based on the operating data under historical operating conditions; For the computational solutions of different resolution grid examples, a spatial block parallel partitioning strategy with gas-solid single-core interaction is adopted to accelerate the iterative solution process. After the computation converges, the calculation results corresponding to the current working condition are obtained. The standardized input parameters and accurate calculation results of the current working condition are synchronously stored in the historical sample library for working condition data accumulation, clustering and matching queries, and the results of multi-physics field distribution in the furnace are visualized and displayed.

2. The rapid calculation and prediction method for the softening zone of a blast furnace according to claim 1, characterized in that, Based on the standardized input parameters, a geometric model of the blast furnace inner contour is established, including: Based on the static geometric parameters in the standardized input parameters, the basic inner contour of the blast furnace body is constructed. The static geometric parameters include the blast furnace segment height and diameter. Combining the dynamic geometric parameters in the standardized input parameters, a dynamically changing feature structure is generated in the basic inner contour. The dynamic geometric parameters include the shape of the material surface position, the size of the swirling zone contour, the morphology of the dead material column, and the height of the molten slag and iron surface. By integrating the basic inner contour with the feature structure, a complete computational domain for predicting the remelting zone is generated, which covers the axial spatial range from the furnace throat to the furnace hearth.

3. The rapid calculation and prediction method for the softening zone of a blast furnace according to claim 1, characterized in that, Construct a prediction calculation model for soft fusion tape, including: A model framework is established based on a pre-defined method for decoupling dense-phase gas-solid two-phase flow. The specific steps are as follows: A corresponding background grid is generated based on the selected resolution grid to store the mass, momentum and energy data of different phase components. The flow tracking calculation of furnace charge particles is performed based on the viscous flow method, the spatial location, batch and attribute information of the furnace charge are recorded, and the results of the flow tracking calculation are mapped to the background grid nodes. A porous media model is used to calculate the gas flow resistance and heat transfer within the particle bed. The resistance coefficient of the flow resistance is dynamically updated based on the raw material properties and fabric information in the standardized input parameters. Source terms are set up to realize the transfer of heat transfer, chemical reaction and phase change data between gas-solid-liquid phases. The results of source term calculations are used for the coupling of mass, momentum and energy between phases.

4. The rapid calculation and prediction method for the blast furnace softening zone according to claim 1, characterized in that, The selection of the hierarchical solution strategy includes: For the standardized input parameters under the current working conditions, historical samples are screened based on preset rules to obtain the best matching sample, and the matching degree between the two is calculated. The solution strategy is determined based on the relationship between the matching degree and the preset matching threshold. The solution strategy determined based on the relationship between the matching degree and the preset matching threshold includes the following steps: When the matching degree is less than the basic matching threshold, an initial trial calculation is performed using sparse first-resolution grid instances to quickly obtain convergence results as the first interpolation data. This data is then mapped to dense second-resolution grid instances as the initial field, and precise calculations are performed on the second-resolution grid instances. If the matching degree is greater than or equal to the basic matching threshold but less than the perfect matching threshold, the matching results are obtained in real time through historical sample queries as the second interpolation data. This data is then mapped to dense second-resolution grid instances as the initial field, and precise calculations are performed on the second-resolution grid instances. If the matching degree is greater than or equal to the perfect matching threshold, the matching results are obtained in real time through historical sample queries as the final precise calculation result for the current working condition, omitting the processes of interpolation data acquisition, data mapping initialization, and fine-grid precise calculation. Before mapping the acquired first or second interpolated data as a data source to a dense second-resolution grid instance as the initial field, it is necessary to detect whether there is a spatial positional deviation between the data source and the current geometric model. If there is a deviation and it exceeds a preset range, the data source is scaled and corrected based on the blast furnace axial segmentation parameters. The corrected data source is then mapped to the second-resolution grid instance, and the mapped data is set as the initial field of the second-resolution grid instance.

5. The rapid calculation and prediction method for the softening zone of a blast furnace according to claim 1, characterized in that, A spatial block-parallel partitioning strategy with gas-solid single-core interaction is adopted to accelerate the iterative solution process, including: Each iteration of the solution involves decoupling the velocity, pressure, temperature, and concentration information of the furnace charge flow and the gas flow, and then realizing the transfer and coupling solution of gas-solid phase data based on the background grid and source terms. During multiple iterative solutions, the residuals of the computational equations and the fluctuations of variables are monitored until the preset convergence criteria are met, so as to obtain stable computational results of the coupled solution. The iterative solution process includes dividing the grid space into blocks based on the preset single-core interaction principle of gas-solid data, and assigning gas phase and solid phase data at the same spatial location to the same calculation process.

6. The rapid calculation and prediction method for the softening zone of a blast furnace according to claim 1, characterized in that, The standardized input parameters and accurate calculation results of the current operating condition are stored in the historical sample database, including: Extract the key input parameter set of the current working condition, sort each parameter according to the preset influence weight, and generate a multi-dimensional feature vector according to the parameter priority sequence. The key input parameter set includes geometric structure, raw material properties, material distribution parameters, air supply parameters, furnace wall boundary and iron tapping parameters. The classification accuracy standard is set based on the preset influence weight, and historical samples are classified according to the multidimensional feature vector and the set accuracy standard. The multiphysics distribution results of the current working condition are associated with the input feature vector and stored in the working condition historical sample database in the form of key-value pairs. Clustering and grouping are then performed, and retrieval identification information is added.

7. The rapid calculation and prediction method for the softening zone of a blast furnace according to any one of claims 1-6, characterized in that, The results of the multiphysics field distribution within the furnace are visualized and displayed, including: Based on the calculation results, extract the structural data of the remelted band, including the distribution shape and location of the remelted band, as well as the distribution information of each layer of remelted band and focal window; Based on the calculation results, the velocity, temperature, pressure, and concentration distribution data of the gas phase and solid phase in the furnace were extracted respectively, and the corresponding cloud map in the furnace was generated. The information on the softening zone, the velocity, temperature, pressure, and concentration data of each phase are integrated and displayed in 2D and 3D through comparison tables, distribution curves, distribution bar charts, distribution cloud maps, and related trend charts.

8. A rapid calculation and prediction device for the softening zone of a blast furnace, characterized in that, The device includes: The data processing module is used to acquire real-time operating data of the blast furnace under the current operating conditions, and to standardize the real-time operating data to generate standardized input parameters. The geometric model construction module is used to construct a geometric model of the inner contour of the blast furnace based on the standardized input parameters, and divide the geometric model into multiple sets of meshes with different resolutions. The multiple sets of meshes include at least a first resolution mesh and a second resolution mesh. The first resolution mesh is used for initial trial calculations to quickly obtain the initial field for accurate calculation, and the second resolution mesh is used for accurate result calculation to finally obtain accurate calculation results. The computational model building module is used to build a computational model for predicting soft fusion bands and dynamically load the standardized input parameters into the multiple sets of grids to build computational examples with different resolution grids. The data solving module is used to select a hierarchical solving strategy based on the standardized input parameters corresponding to the current working condition and the matching situation of historical samples. It includes three solving strategies: coarse mesh trial calculation - data mapping initialization - fine mesh accurate calculation, historical sample matching - data mapping initialization - fine mesh accurate calculation, and historical sample matching - matching result reuse. It also combines a gas-solid single-core interactive spatial block parallel partitioning strategy to accelerate the coarse and fine mesh iterative solving process and quickly obtain the accurate calculation results corresponding to the current working condition. The storage and output module is used to store the standardized input parameters and accurate calculation results of the current working condition, and store them in the working condition sample database in the form of key-value pairs for subsequent rapid retrieval and matching of working conditions, and synchronously visualize the multi-physics field distribution results in the furnace.

9. An electronic device, characterized in that, The electronic device includes: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the electronic device to implement the rapid calculation and prediction method for the blast furnace softening zone as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by the computer's processor, causes the computer to perform the rapid calculation and prediction method for the blast furnace softening zone as described in any one of claims 1 to 7.

Images