An intelligent ore blending method for sintering process based on digital twin

Abstract

The present invention discloses an intelligent ore blending method for the sintering process based on digital twins, belonging to the field of industrial process optimization modeling. First, conduct a demand analysis on the digital twins of the sintering process and equipment, and use digital twin technology to establish a digital twin model for the entire process of multi-process sintering production. Then, starting from the actual business and process mechanism of sintering production, construct a twin system platform for simulation, prediction optimization, and test verification; Next, utilize the iron ore sintering knowledge base and historical operation data, and adopt the XGBoost algorithm and improved grey wolf optimization algorithm to construct an ore blending optimization module for the sintering process. Finally, deploy the proposed intelligent ore blending method on the twin system platform to realize real-time optimization of the raw material ratio in the sintering process, and display the optimization results on the front-end page.

Description

Technical Field

[0001] The present invention relates to an intelligent optimization method for raw material proportions in a sintering process, and specifically to an intelligent ore proportioning method for a sintering process based on digital twins. Background Art

[0002] The sintering production process is long and multi-step. From a control perspective, the sintering process is complex, nonlinear, time-varying, and uncertain, making it a typical complex controlled object. The quality of the blended ore and sintered ore batching directly affects the output and quality of metallurgical production and is closely related to the company's economic interests. How steel companies can rationally select and effectively utilize iron ore resources to ensure product quality at every stage while reducing steel production costs and thus enhancing their competitiveness has become a pressing issue. For a long time, sintering batching has largely been manually controlled by operators based on experience. With the deepening understanding of sintering raw material properties, the improvement of sintering equipment, the adoption of related process technologies, and the introduction and implementation of automatic control concepts, the application of intelligent optimized ore batching systems has become a key means to achieve "high quality, high yield, and low consumption" in sintering production, and has become a major focus for sintering plants both domestically and internationally to improve their technical level.

[0003] In the steelmaking production process, the first step is sintering batching, which includes ore blending and sintering. Sintering batching is the process of mixing ore, flux, and solid fuel in specific proportions, followed by thorough stirring and mixing. Sintering batching affects the sinter's iron grade, alkalinity, silica content, calcium oxide content, and impurity sulfur content. Improper control of these components can affect the sinter's strength, drum strength, alkalinity, and reducibility. During ore blending, the raw materials are controlled based on the feed quality requirements. Blended ore, the raw material for sintering batching, is the result of thoroughly mixing various ore raw materials. This affects the sinter's composition, particularly its iron grade and silica content. It is crucial for steel companies to rationally select and effectively utilize iron ore resources to ensure product quality at every stage while reducing steel production costs and enhancing their competitiveness.

[0004] There are some difficulties in the smelting production process of major domestic steel plants. The types of batching raw materials are numerous, and there are significant differences in physical properties and chemical composition contents. On the premise of strictly following the process production requirements, the batching plan is particularly important. Since the enterprise is composed of many functional departments, the optimization of the material formula involves the procurement department, the technical department, and the production department. There are differences in the assessment indicators within each department, and their respective batching optimization goals are different, resulting in difficulties in the collaborative optimization of the material formula among multiple departments. To design the optimal batching optimization plan before ironmaking, steel enterprises need to closely combine the three links of procurement, ore blending, and output evaluation, and use the method of combining big data, expert systems, and experimental data to optimize the material formula. However, there are a series of problems in simply mixing materials through manual experience, including difficulties in obtaining real-time material information, complex energy and mass flow coupling between multi-level batching, and difficulties in fine and intelligent optimized batching. These problems lead to the inability to effectively break down the barriers between links, resulting in the failure of the mixing ratio. Summary of the Invention

[0005] In order to overcome the deficiencies of the prior art, the present invention proposes an intelligent ore blending method for the sintering process based on digital twin. Starting from the actual business and process mechanism of sintering production, the present invention adopts advanced sensing, data mining and processing, data modeling, regulation and optimization and other technical means to build a digital twin system of the sintering production line based on the industrial Internet platform, construct an intelligent ore blending method for the sintering process, realize the reasonable procurement of raw materials, real-time ratio decision-making of materials before ironmaking, effective evaluation of output, reduce production costs, reduce the emission of harmful elements, and improve the quality and output of the sintering process.

[0006] The present invention is realized by the following technical solutions:

[0007] An intelligent ore blending method for the sintering process based on digital twin, comprising the following steps:

[0008] 1) Analyze the requirements of the digital twin bodies of the sintering process and equipment, study the digital replication of the geometric attributes of the sintering process and equipment and the multi-time and multi-scale modeling of the operation mechanism, use digital twin technology to establish a digital twin model of the entire sintering production process, and input the operation data of the real physical model, including process data and product data, into the digital twin model to realize the virtual-real interaction between the physical model and the twin model; on this basis, carry out the construction of a material formula optimization module based on digital twin;

[0009] 2) Adopt advanced sensing, data mining and processing, data modeling, regulation and optimization, establish a sintering material formula optimization module on the basis of the digital twin model of the entire sintering production process, form a visual virtual sintering production system, and realize parallel operation with the on-site physical production line, providing a twin system platform for simulation, monitoring and diagnosis, prediction and optimization, and test verification for the safe and efficient operation of sintering production;

[0010] 3) Generate virtual and real data through digital twin technology to expand the database; at the same time, use data mining and database management technologies to achieve data cleaning, data dimensionality reduction, and data association; construct a sintering process knowledge base;

[0011] 4) Use the XGBoost model and improved grey wolf optimization algorithm to construct a batching optimization module for the sintering process, and propose an improved grey wolf optimization algorithm for solving the model. The improvements include two aspects: convergence factor adjustment based on the sigmoid function and individual update based on the differential mutation strategy

[0012] 5) Upload the built batching optimization module to the digital twin system to optimize the raw material ratio of the sintering process in real time and display it through the front end.

[0013] In the said step 1),

[0014] First, based on the analysis of the sintering process and the requirements of the equipment digital twin body, combined with the existing measurement methods, complete the acquisition of the geometric structure, spatial movement, and geometric association attributes of the sintering process and equipment entity objects. Based on the spatial geometric structure of the sintering process and equipment entity objects, combined with the spatial movement laws of the sintering process and equipment entity objects, use 3D reconstruction tools to realize the reconstruction of the spatial geometric models of the sintering process and equipment.

[0015] Then, use digital twin model fusion technology to establish a digital twin model with the geometric attributes and operating mechanisms of the sintering process / equipment. On this basis, use the information interaction technology between the sintering process and equipment twin models and the entity to endow the twin model with the operating state information and various operation information of the entity object, so that the digital twin model can map the actual object throughout the life cycle. At the same time, the digital twin model can also be continuously iteratively optimized according to the data transmitted by the entity to improve the mapping accuracy of the entity; predict various operating states and material change situations through the digital twin model, and guide the actual sintering production process according to the prediction information, thus constituting a truly digital sintering process and equipment digital twin body.

[0016] The establishment of the digital twin model for the entire sintering production process described in the said step 1) includes: (1) Efficient data transmission: To ensure the real-time performance of the system and improve the real-time following performance of the digital twin system, efficient transmission technology is required to shorten the time for the system to transmit data and reduce system latency; (2) Multi-temporal and multi-spatial scale modeling of multi-source heterogeneous data.

[0017] Step 4) includes: First, using the iron ore knowledge base, determine the raw material parameters related to the physical properties of sinter; then, use XGBoost to construct a mapping function between the raw material parameters and the physical properties, and use it as the optimization objective function; then, set a series of constraint conditions according to the chemical composition and process parameters of sintering, and establish a multi-objective optimization model; finally, propose an improved grey wolf optimization algorithm to solve the optimization model.

[0018] In the described step 4), based on the digital twin platform, the staff can calculate the material formula set by themselves. Through the backend batching optimization module, calculate the performance results of the current formula ratio, and verify the results of their own solutions. The calculation results of the material formula are displayed in the front-end digital model interface.

[0019] The beneficial effects of the present invention are as follows:

[0020] 1. The combination of digital twin technology and the existing batching method realizes the reasonable procurement of raw materials, the real-time ratio decision of materials before ironmaking, and the effective evaluation of outputs through material formula optimization and process parameter design simulation, thereby improving the sintering production efficiency, reducing production costs, reducing the emission of harmful elements, and improving product quality.

[0021] 2. An improved grey wolf optimization algorithm is proposed to solve the model, which can improve the convergence and speed of model solution and meet the requirements of real-time stability in the industrial field. Description of the Drawings

[0022] Figure 1 It is a flow schematic diagram of the intelligent ore blending method for the sintering process based on digital twin.

[0023] Figure 2 Schematic diagram of the heterogeneous element perception and transmission process.

[0024] Figure 3 Preprocessing flow chart of the iron ore knowledge database.

[0025] Figure 4 Schematic diagram of the technical route of sintering intelligent ore blending.

[0026] Figure 5 Mapping relationship diagram between sintering batching and the test indexes of finished ore.

[0027] Figure 6 Sintering batching optimization module based on intelligent optimization algorithm. Detailed Embodiment

[0028] The present invention will be further described in detail below with reference to the drawings and embodiments.

[0029] Figure 1A process schematic diagram of the intelligent ore blending method for the sintering process based on digital twin is provided. The process of the present invention can be summarized as follows: First, according to the physical entity model of the real world, the sintering production process is digitalized, visualized and embodied by using computer simulation technology, a digital twin model of the sintering process is built, and a corresponding digital twin verification platform is established; Then, an iron ore knowledge base is constructed through the mechanism knowledge of the sintering process, and a sintering burdening optimization model is established by using machine learning models and intelligent optimization algorithms; Subsequently, based on the digital twin platform and the burdening optimization model, the staff can calculate the material formula set by themselves, and through the backend burdening optimization model, calculate the performance results of the current formula ratio to verify the results of their own solutions; Finally, the calculation results of the material formula are displayed on the front-end digital model interface, the model results are debugged, and verified in the actual industrial site.

[0030] The following provides the specific steps of the intelligent ore blending method for the sintering process based on digital twin, including the following steps:

[0031] 1. Construction of the digital twin model of the sintering process. Based on the requirement analysis of the digital twin of the sintering process / equipment, combined with the existing measurement methods, the geometric attributes such as the geometric structure, spatial movement, and geometric correlation of the sintering process / equipment entity object are obtained. Based on the spatial geometric structure of the sintering process / equipment entity object, combined with the spatial movement law of the sintering process / equipment entity object, using 3D reconstruction tools, the reconstruction of the spatial geometric model of the sintering process / equipment is realized, which can accurately describe the operation status of key processes and main equipment and the changes in material flow in the actual sintering process in all directions at time scales such as seconds, minutes, hours, etc. and spatial scales such as equipment, processes, and production lines, and comprehensively depict the accurate relationship between the output and input in the actual sintering process. At the same time, aiming at the problem of limited computing resources in the reconstruction process of the sintering process / equipment entity model, the rendering optimization of the spatial geometric model of the sintering process / equipment is carried out. Considering the mutual connection of equipment in each stage of the sintering process, the spatial geometric models of each equipment in the sintering process are matched and connected, so as to realize the digital and accurate reproduction of the geometric attributes of the sintering process / equipment. It mainly includes:

[0032] 1.1 Digital replication technology for the geometric properties of the sintering process / equipment. Aiming at the problem of the intricate spatial geometric properties of the sintering process / equipment entity objects, a mapping and surveying method based on CAD technology is adopted. According to the real geometric properties of the entity objects at each stage of the sintering process / equipment, through specific measurement means, the geometric properties of the equipment entity objects in the real three-dimensional space are mapped to the two-dimensional space. To truly replicate the appearance design of the sintering process / equipment, CCD imaging technology is used to obtain the image information of the sintering process equipment, providing materials for the solid space geometric model of the sintering process / equipment. The spatial geometric attribute information of the sintering process / equipment entity objects is diverse and the data structure is complex, while the 3D reconstruction tools and rendering optimization tools have different functional focuses. Currently, commonly used 3D reconstruction tools include: Maya, Solidworks, 3Dsmax, etc. Maya is generally used for 3D reconstruction in film and television animation. Solidworks focuses on structural design and parameter design with a single function. While 3Dsmax can not only perform structural design and parameter design on specific objects, but also carry out scene modeling and simulate the motion state of objects, which is suitable for the 3D reconstruction of the spatial geometric properties of the digital twin body.

[0033] 1.2 Multi-temporal and multi-scale modeling technology for the operating mechanism of the sintering process / equipment. The sintering process is one of the agglomeration technologies for producing artificial iron ore in the sintering system. By melting concentrate powder at high temperature into ore lumps, it is used as the raw material for blast furnace sintering. The batching process of sintering is to mix various prepared raw materials (concentrate, ore powder, fuel, flux, return ore, iron-containing production waste, etc.) in a certain proportion through batching, mixing and granulation to obtain qualified sintering materials. Since the batching method directly affects the quality of sintered ore, by modeling the sintering batching process, the sintering digital twin model can map the operating conditions of the entire sintering batching process, obtain the physical properties of different batching methods before sintering, and lay a foundation for revealing the relationship between the quality of the final sintering materials and the batching method.

[0034] 1.3 Information interaction technology between the sintering process / equipment twin model and the entity. Using database technology, key information such as proportioning parameters, ignition temperature, exhaust fan power, FeO content, drum strength, etc. during the sintering process is dynamically transmitted to the sintering digital twin model. For large-capacity data such as sintering cross-section videos, industrial Ethernet is used for transmission. Aiming at the complex physical and chemical reactions in the sintering process, based on the sintering multiphase thermodynamics model, the real-time generation mechanism and production mechanism of sintering raw materials becoming sintered ore are accurately characterized. Based on the large amount of information generated during the sintering process, using multi-dimensional feature extraction methods, the deep features of the reaction sintering quality are mined, and the sintering digital twin model is updated throughout its life cycle, enabling the digital twin model to be synchronized with the real sintering process and having the mapping and evolution capabilities at multiple levels and scales. Realize the real-time interaction between the sintering digital twin model and the sintering system.

[0035] 2 Construction of the Twin Verification Platform. The twin verification platform collects sensor data, automated system data, and enterprise business system data, and integrates the data through technologies such as edge computing and protocol parsing, and then stores, backs up, and manages the data in the cloud of the data platform. In the digital twin system model architecture of the sintering production line, the twin data platform is located in the mapping layer of the association relationship between the physical space and the virtual space, and plays a role in connecting the digital twin application layer, the physical space entity model layer, and the virtual space multi-dimensional model layer. The main steps are as follows:

[0036] 2.1 The data platform collects heterogeneous data from data sources such as physical workshops, information systems, virtual models, and service systems, and uploads the information data encapsulated by communication protocols (such as Modbus TCP / IP, Modbus RTU, etc.) through various communication carriers (such as optical fibers, 4G, 5G wireless networks, etc.) to the multi-source concurrent data perception control intelligent front end, realizing the perception of heterogeneous elements of the digital twin system of the sintering production line;

[0037] 2.2 Build a self-organizing perception real-time network to achieve real-time, high-speed, and highly reliable transmission of multi-source concurrent data. The data platform supports integration with on-site automation, supports more than 10 open industrial control protocols, and at the same time can transmit sensor data into the industrial Internet through 4G and Ethernet, and achieve data interconnection with some brand industrial control systems. Due to the large amount of data collected by the digital twin system of the ironmaking production line, relevant data covering all elements, all processes, and all services such as massive sensor data, virtual model data, simulation data, and service system data in the physical workshop need to be transmitted with high efficiency and low latency to ensure the real-time following performance of the digital twin system. To meet the above requirements, a self-organizing perception real-time network - mobile ad hoc network is adopted to enable the heterogeneous elements, information systems, virtual model data, and service platform data in the physical workshop to achieve interconnection and interoperability through various industrial buses or wireless networks. The overall architecture and implementation process are as Figure 2 shown.

[0038] 2.3 Build a distributed global database to achieve global consistency and interface unification of system data and peripheral data, and achieve high-efficiency real-time storage, retrieval, and parallel computing of data; among them, distributed stream processing is adopted for data with high real-time requirements to achieve high-throughput and low-latency streaming data operations.

[0039] 3. Construction of the iron ore sintering performance knowledge base. This process mainly includes the iron ore knowledge base and the chemical composition knowledge base. For example, the inherent properties of iron ore powder are important factors affecting the formation ability of calcium ferrite. The type, particle size, compactness, alkalinity, and chemical composition (including CaO, MgO, SiO2, and Al2O3, etc.) of iron ore powder directly affect the mineral phase composition and distribution uniformity of sinter. Moreover, the inherent properties of iron ore powder are important factors affecting the formation ability of calcium ferrite. Iron ores are mainly magnetite ores and hematite ores. Magnetite sintering is more complex than hematite sintering because the unique spinel structure of magnetite often dissolves different impurities, and the types of gangue minerals also vary greatly. Al2O3 has a relatively large impact on the low-temperature reduction degradation rate of sinter. A high Al2O3 content will reduce the bonding phase, thereby reducing the strength of sinter, and a high Al2O3 content will increase the stress in the magnetite formed during the reduction process. In addition, based on the relevant results of the virtual-real data fusion of the aforementioned batching optimization, further combining the digital twin technology and embedding data mining and online processing and analysis methods can realize functions such as data cleaning, data dimensionality reduction, and data correlation analysis, providing support for more efficient online rapid optimization decision-making of batching. The specific idea is as Figure 3 shown.

[0040] 4. Construction of the batching optimization module for the sintering process. Conduct research on the batching production site, process the production site data, construct a single-process material formula optimization model, and construct a multi-process material formula optimization model. By combining the actual production needs, model the entire batching process, construct a formula optimization model, and realize the design of the optimized batching plan. The specific process is as Figure 4 shown. The steps for constructing this material optimization model are as follows:

[0041] 4.1 Determination of the objective function. Taking the raw material ratio of sintering, the operation parameters of the sintering process, and the state parameters of the sintering process as the model inputs, and the test indexes of sinter as the model outputs, construct a soft sensor model, which is used as the optimization objective function, as Figure 5 shown.

[0042] 4.2 Constraint conditions: The cost of sinter is calculated based on the real-time market fluctuations of various raw materials to calculate the cost per ton of sinter. The iron grade (TFe) affects the coke ratio and the molten iron output; the alkalinity has an important impact on the mineral structure after liquid solidification. High alkalinity is beneficial to the formation of calcium ferrite and dicalcium silicate, and low alkalinity tends to form vitreous slag phase; when the SiO2 content is relatively high, more bonding phase of sinter is generated, but too much SiO2 content will generate a large amount of silicate, reducing the strength of sinter. Therefore, it is necessary to constrain the chemical composition of sinter.

[0043] 5. Design of the improved grey wolf optimization algorithm

[0044] There are many constraints in sintering burden distribution. Besides the constraints of its own ratio, more component constraints need to be considered. Both the objective function and the constraint function are non-linear relationships, and non-linear optimization methods can be used. Swarm intelligence optimization has good robustness and can better achieve linear optimization solutions. The constraints of sinter ore include component constraint conditions and ratio constraint conditions. For each type of sinter ore raw material, the range of the sinter ore ratio needs to be set, and this value is often predicted by the operator. According to the differential evolution algorithm, the basis exchange operation and the entering column selection are continuously carried out to continuously reduce the objective function value. When the optimal solution discrimination criterion is met, the iteration ends and the optimal solution is obtained. The specific optimization scheme is as Figure 6 shown. The improved grey wolf optimization algorithm is as follows:

[0045] The basic grey wolf algorithm is a new type of intelligent optimization algorithm proposed by drawing on the ideas of the predation behavior of wolf packs and the division of labor in the social leadership hierarchy in nature. In a small grey wolf group, there are three optimal individuals, namely the α-wolf, β-wolf, and δ-wolf, who are at the upper layer of the pyramid, and other wolves obey the commands of these wolves.

[0046] The process of these grey wolves chasing and surrounding their prey can be abstracted into a mathematical model, as shown in the following formula:

[0047]

[0048] Among them, X represents the position of the grey wolf in space, t is the number of iterations, X p is the position of the prey, A and C are coefficient variables, r 1 and r 2 are random numbers between 0 and 1 respectively. θ is the convergence factor, and its update formula is a decreasing function, t max is the maximum number of iterations, and the specific formula is as follows:

[0049]

[0050] Then, the positions of other grey wolf individuals are updated according to the positions of α, β, and δ respectively.

[0051] 5.1 Convergence factor adjustment based on the sigmoid function

[0052] According to the mechanism analysis of the previous grey wolf algorithm, the magnitude of the A value represents the range when the grey wolves surround the prey. The larger the A value, the larger the surrounding circle, and vice versa, the smaller the surrounding circle, which represents the global exploration and local fine search capabilities of the grey wolves. The value of A changes with the convergence factor θ changing. This indicates that the convergence factor affects the global search and local search capabilities of the grey wolf algorithm. However, in the basic GWO, the convergence factor decreases linearly, and this search strategy is difficult to adapt to the actual situation during the actual optimization process. In order to enable the grey wolves to search in a relatively wide range at the beginning stage and be able to perform fine search within a very small range at the end stage, the sigmod function is used here to control the value of the convergence factor, which can make the global exploration and local search capabilities of the algorithm stronger. The update strategy of its convergence factor is shown in the following formula. Adopting this strategy is beneficial to both accelerating the convergence speed and enabling the algorithm to obtain the optimal value at the end of the iteration.

[0053]

[0054] 5.2 Individual update based on differential mutation strategy

[0055] In the basic GWO algorithm, it can be seen from Equation (6) that the update of other grey wolf individuals in the population is determined by the positions of the three types of wolves, namely α, β, and δ. If these three types of wolves fall into the vicinity of the local optimal solution, it will lead to a reduction in the diversity of other wolf individuals, resulting in the premature phenomenon of the algorithm, being unable to break out of the local optimal surrounding circle, and the solution effect will become poor. Therefore, to solve this problem, a mutation operation operator can be introduced to enable the algorithm to avoid this situation of falling into the local optimum. Common mutation operators include Gaussian mutation, Cauchy mutation, etc.

[0056] Inspired by the differential evolution algorithm, the present invention uses a differential mutation operator to adjust the positions of other wolves. That is, the position update is performed by randomly selecting the current grey wolf individual, the optimal grey wolf individual, and a randomly selected grey wolf individual for random differential selection, and its expression is as follows:

[0057] .

[0058] 6. Based on the digital twin intelligent ore blending method platform, staff can calculate the material formula they set, and through the backend batching optimization model, calculate the performance results of the current formula ratio to verify the results of their ideas. The calculation results of the material formula are displayed in the front-end digital model interface.

[0059] Taking the burdening and sintering of a certain steel enterprise in a certain month as an example, the effectiveness of the intelligent ore blending method is verified. The upper and lower limits of the chemical composition requirements of the raw materials for burdening and sintering can be seen in Table 1. In addition, due to the influence of the burdening inventory factor, it is also necessary to balance the use of each ore type as much as possible to reduce the warehousing cost. The requirements for the proportion of each raw material in the preliminary burdening can be seen in Table 2, and the raw material costs can be seen in Table 3.

[0060]

[0061]

[0062]

[0063] 。

[0064] It can be seen from the optimization results in Table 4 that the results optimized by the intelligent ore blending method meet the requirements of the actual factory, reducing the cost per ton of sinter ore from 557.50 to 546.25, bringing greater economic benefits to the enterprise.

[0065] The above-described embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the appended claims.

Claims

1. An intelligent ore blending method for sintering process based on digital twin, characterized in that, It includes the following steps: 1) Analyze the sintering process and the requirements of the equipment digital twin. Study the digital replication of the geometric properties of the sintering process and equipment and the multi - spatio - temporal scale modeling of the operation mechanism. Use digital twin technology to establish a digital twin model for the entire sintering production process and input the operation data of the real physical model, including process data and product data, into the digital twin model to achieve the virtual - real interaction between the physical model and the twin model. On this basis, build a material formula optimization module based on digital twin; 2) Adopt advanced sensing, data mining and processing, data modeling, and regulation optimization. Based on the digital twin model of the entire sintering production process, establish a sintering material formula optimization module to form a visual virtual sintering production system, which runs in parallel with the on - site physical production line, providing a twin system platform for simulation, monitoring and diagnosis, prediction and optimization, and test verification for the safe and efficient operation of sintering production; 3) Generate virtual - real data through digital twin technology to expand the database. At the same time, use data mining and database management technology to achieve data cleaning, data dimensionality reduction, and data association; Construct a sintering process knowledge base; 4) Use the XGBoost model and an improved grey wolf optimization algorithm to construct a sintering process batching optimization module, and propose an improved grey wolf optimization algorithm for solving the model. The improvement includes two aspects: convergence factor adjustment based on the sigmoid function and individual update based on the differential mutation strategy; 5) Upload the built batching optimization module to the digital twin system to optimize the raw material ratio in the sintering process in real - time and display it through the front - end.

2. The intelligent ore blending method for the sintering process based on digital twin according to claim 1, wherein In step 1), first, based on the analysis of the sintering process and the requirements of the equipment digital twin, combined with existing measurement methods, obtain the geometric structure, spatial motion, and geometric correlation attributes of the sintering process and equipment entity objects. Based on the spatial geometric structure of the sintering process and equipment entity objects, combined with the spatial motion law of the sintering process and equipment entity objects, use 3D reconstruction tools to realize the reconstruction of the spatial geometric model of the sintering process and equipment; Then, use digital twin model fusion technology to establish a digital twin model with the geometric properties and operation mechanism of the sintering process / equipment. On this basis, use the information interaction technology between the sintering process / equipment twin model and the entity to endow the twin model with the operation state information and various operation information of the entity object, so that the digital twin model can map the actual object throughout its life cycle. At the same time, the digital twin model can also be continuously iteratively optimized according to the data transmitted by the entity to improve the mapping accuracy of the entity. Predict various operation states and material change situations through the digital twin model, and guide the actual sintering production process according to the prediction information, thus constituting a truly digital sintering process and equipment digital twin.

3. The intelligent ore blending method for the sintering process based on digital twin according to claim 1, wherein, In step 1), the establishment of the digital twin model for the entire sintering production process includes: (1) Efficient data transmission: To ensure the real - time performance of the system and improve the real - time following performance of the digital twin system, an efficient transmission technology is required to shorten the time for the system to transmit data and reduce system latency; (2) Multi - spatio - temporal scale modeling of multi - source heterogeneous data.

4. The intelligent ore blending method for the sintering process based on digital twin according to claim 1, wherein, Step 4) includes: First, using the iron ore knowledge base, determine the raw material parameters related to the physical properties of sinter; Then, use XGBoost to construct the mapping function between the raw material parameters and the physical properties, and use it as the optimization objective function; Next, set a series of constraint conditions according to the chemical composition and process parameters of sintering to establish a multi-objective optimization model; Finally, an improved grey wolf optimization algorithm is proposed to solve the optimization model.

5. The intelligent ore blending method for sintering process based on digital twin according to claim 1, wherein In step 4), based on the digital twin platform, the staff calculates the material formula set by themselves, and through the backend batching optimization module, calculates the performance results of the current formula ratio to verify the results of their own plan. The calculation results of the material formula are displayed in the front-end digital model interface.

Images