A method for controlling and regulating a closed ferrochrome furnace

Abstract

The present invention relates to the technical field of chemical composition determination in a ferrochrome furnace. The present invention provides a method for controlling and regulating a closed ferrochrome furnace. Aiming at the problem of incorrect furnace control caused by inaccurate determination of ferrochrome alloy components, the present invention proposes the following solution: historical data including materials, reactions, and operating efficiency are acquired and processed. Then, an associated data set is constructed based on the reaction batches, data relationships are mined, and efficient operating data is screened and stored in a graph database to form a knowledge base. The expected operating efficiency is received and compared with the real-time collected data. When the efficiency is low, chemical material information is extracted and substituted into a digital twin model to verify the accuracy. Parameters are then optimized through dynamic time warping and genetic algorithms to form a continuous optimization closed loop, achieving precise control and stable operation.

Description

Technical Field

[0001] The invention relates to the technical field of chemical composition determination in a ferrochrome furnace, and in particular to a control and adjustment method for a closed ferrochrome furnace. Background Art

[0002] A closed ferrochrome furnace is an industrial electric furnace used to produce ferrochrome alloy (a key raw material for stainless steel manufacturing). Its core structure is a fully enclosed design. The raw materials in the furnace (such as chromium ore, coke and limestone) are heated by electric energy and a carbon thermal reduction reaction is carried out at a high temperature (about 1600-1800℃) to oxidize the chromium. The carbon reacts with the molten metal to produce liquid ferrochrome and carbon monoxide gas. Compared to traditional open furnaces, closed furnaces can efficiently recover combustible carbon monoxide as a secondary energy source, significantly reducing exhaust emissions and energy consumption. The enclosed environment also reduces heat loss and impurity intrusion, ensuring a uniform alloy composition and low impurity content.

[0003] During the operation of a closed ferrochrome furnace, chemical materials (such as chromium ore, coke, and flux) undergo a high-temperature carbothermal reduction reaction to produce ferrochrome and carbon monoxide. Real-time monitoring of parameters such as furnace temperature, pressure, gas composition, and slag and iron conditions is required to dynamically adjust system operations. For example, if the carbon monoxide concentration generated by the reaction is abnormal or the slag viscosity is too high, the raw material ratio needs to be optimized (such as increasing the coke ratio to enhance the reducing power) or the power supply needs to be adjusted to stabilize the furnace temperature. If the furnace pressure is too high or the risk of gas escape increases, the sealing system needs to be calibrated or the exhaust gas recovery device needs to be optimized to efficiently utilize combustible gases and reduce environmental pollution. Furthermore, by analyzing the alloy composition (such as the chromium / iron ratio and impurity content), feedback can be provided to adjust the flux addition amount or reaction time to ensure a balance between product quality and energy efficiency.

[0004] In the regulation of closed ferrochrome furnaces, inaccurate measurement of ferrochrome alloy composition (such as chromium / iron ratio and impurity element content) will trigger logical chain technical pain points: First, deviations in composition data will mislead decisions on raw material proportions (such as the amount of coke or flux added) and power supply adjustment. For example, excessive or insufficient carbon reducing agent will directly affect the efficiency of carbon thermal reduction, leading to an imbalance in the reaction in the furnace. Second, distorted composition feedback masks the true furnace conditions (such as abnormal slag-iron separation or local overburning), forcing operators to rely on empirical compensation measures (such as blindly increasing smelting time or repeatedly adjusting parameters), which in turn leads to a surge in energy consumption, fluctuations in furnace pressure, and even increased lining erosion. The continued output of incorrect alloy composition will also cause quality defects in downstream stainless steel products (such as decreased corrosion resistance), forming a systemic risk from process control to end products. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a closed ferrochrome furnace control and adjustment method to solve the problem of errors in the control and operation adjustment of the closed ferrochrome furnace caused by inaccurate measurement of the ferrochrome alloy composition (such as the chromium / iron ratio and the content of impurity elements).

[0006] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0007] The present invention provides a closed ferrochrome furnace control and adjustment method, comprising:

[0008] Step S101, acquiring historical data of a closed ferrochrome furnace, the historical data of the closed ferrochrome furnace including historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace;

[0009] Step S102, establishing a correlation relationship between historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace based on time information to obtain a closed ferrochrome furnace correlation data set;

[0010] Step S103: Filtering out data from the closed ferrochrome furnace associated data set to obtain data with an operating efficiency higher than a preset efficiency value, thereby obtaining a first closed ferrochrome furnace data set; retrieving chemical material data corresponding to the first closed ferrochrome furnace data set; and storing the chemical material data corresponding to the first closed ferrochrome furnace data set in a pre-built knowledge base, thereby obtaining a closed ferrochrome furnace chemical material knowledge base;

[0011] Step S104: receiving the expected operating efficiency of the closed ferrochrome furnace, collecting real-time operating efficiency data of the closed ferrochrome furnace, and comparing the real-time operating efficiency data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace. If the real-time operating efficiency data of the closed ferrochrome furnace is lower than the expected operating efficiency of the closed ferrochrome furnace, retrieving chemical material information within the closed ferrochrome furnace from the real-time operating efficiency data of the closed ferrochrome furnace to obtain a second closed ferrochrome furnace data set.

[0012] In step S105, the second closed ferrochrome furnace data group is matched with the closed ferrochrome furnace chemical material knowledge base to obtain a chemical material information matching result, and the chemical material information matching result is sent to the chemical material adding device. The operating efficiency of the closed ferrochrome furnace after the chemical material adding device adds the chemical material in the closed ferrochrome furnace is collected. If it is still lower than the expected operating efficiency of the closed ferrochrome furnace, the chemical material parameters in the chemical material information matching result are optimized using a genetic algorithm to obtain optimized chemical material parameters. The optimized chemical material parameters are used as execution data in the closed ferrochrome furnace adjustment process.

[0013] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S101 includes:

[0014] Acquire structured and unstructured data through embedded high-temperature sensors in the furnace, PLC system logs, and laboratory offline test reports;

[0015] Embedded high-temperature sensors in the furnace include infrared thermometers and laser gas analyzers;

[0016] Structured data includes temperature, pressure, and current, while unstructured data includes slag images and spectral analysis results;

[0017] Convert chemical raw material compositions from historical chemical reaction data in a closed ferrochrome furnace into mass percentage format.

[0018] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S102 includes:

[0019] Using the reaction batch as the time axis, a preset dynamic time warping algorithm is used to align data streams with different sampling frequencies, including second-level temperature data and minute-level gas composition data, to construct a time and event correlation matrix.

[0020] Extract time series features and mark the start and end time points of abnormal events. Based on time series features, mark the start and end time points of abnormal events to form a multidimensional correlation data set;

[0021] Through the time and event correlation matrix and multidimensional correlation data set, the historical chemical reaction data in the closed ferrochrome furnace and the historical operating efficiency data of the closed ferrochrome furnace are established based on time information correlation.

[0022] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S103 includes:

[0023] Set the threshold of the preset efficiency value data, the threshold is the closed ferrochrome furnace operating efficiency value ≥ 85%;

[0024] A graph database is used to store the chemical material data corresponding to the first closed ferrochrome furnace data group. The chemical material data corresponding to the first closed ferrochrome furnace data group includes chemical raw material ratios, reaction stages, efficiency indicators, edge relationship definition conditions, and chemical reaction result data.

[0025] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S104 includes:

[0026] Receive the expected operating efficiency of the closed ferrochrome furnace set by the user, which includes the unit energy consumption target and the production rate, convert the expected operating efficiency of the closed ferrochrome furnace into a standardized efficiency indicator set, and store it in the server-side dynamic configuration library;

[0027] Deploy edge computing nodes to synchronously collect real-time data streams from furnace temperature, pressure, current, power, and output metering equipment, perform time series alignment and remove outliers, and generate a real-time operating efficiency data set with consistent timestamps.

[0028] The dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace is calculated based on the sliding window statistical method. If the deviation exceeds the allowable fluctuation range of the closed ferrochrome furnace operating efficiency by ±3% for three consecutive sampling periods, the chemical material information extraction instruction is activated;

[0029] Based on the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding period are traced back from the real-time operation efficiency data set. The chemical material parameters include coke feed amount, flux ratio, raw material grade, furnace pressure and gas composition.

[0030] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S104 includes:

[0031] According to the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding time period will be traced back from the real-time operation efficiency dataset as the data to be evaluated, and the data to be evaluated will be substituted into the preset digital twin model of the closed ferrochrome furnace to obtain the digital twin image data of the closed ferrochrome furnace. The digital twin image data of the closed ferrochrome furnace will be compared with the expected operating efficiency of the closed ferrochrome furnace. If the result of comparing the digital twin image data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace is consistent with the dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace calculated based on the sliding window statistics method, then the chemical material parameters of the corresponding time period traced back from the real-time operation efficiency dataset are accurate.

[0032] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S105 includes:

[0033] Dynamic time warping was used to calculate the multidimensional similarity between the second closed ferrochrome furnace data set and the historical data in the knowledge base, to screen matching cases and extract the material parameters of the matching cases as initial control recommendation data.

[0034] Genetic algorithm parameter optimization:

[0035] Coding design: Encoding material parameters into binary gene strings;

[0036] The initial control recommendation data is set as the optimization target, and new parameter combinations are iteratively generated using tournament selection, two-point crossover, and Gaussian mutation strategies. The feasibility of the new parameter combination is verified through rapid simulation of the digital twin, and the optimal solution set of the new parameter combination is output;

[0037] The optimal solution set of the new parameter combination is sent to the raw material conveying PLC system, and the recovery of the operating efficiency of the closed ferrochrome furnace is monitored in real time. If the standard is still not met, incremental learning of the knowledge base is triggered, and the new parameters and results are fed back to the knowledge base to form a continuous optimization closed loop.

[0038] Furthermore, in the closed ferrochrome furnace control and adjustment method of the present invention, step S105 includes:

[0039] The new parameter combination generated by the genetic algorithm is mapped to the input interface of the digital twin according to the preset rules, and the current furnace state data is simultaneously loaded as the initial condition of the simulation;

[0040] Based on the preset carbothermal reduction reaction mechanism model, multi-physics field coupling calculations were initiated within the twin to simulate the dynamic process within the furnace under the new parameters. The predicted values ​​of indicators such as chromium / iron ratio, impurity concentration, energy efficiency, and furnace pressure fluctuation curve were output, and simulation results were obtained.

[0041] Compare the simulation results with the preset process constraints, eliminate the parameter combinations with the risk of exceeding the standards in the new parameter combinations, sort the comparison results by priority weight, and generate an executable parameter sequence;

[0042] After the executable parameter sequence is sent to the actual furnace control, the operating data is collected in real time and the deviation is analyzed with the prediction results of the closed ferrochrome furnace digital twin model; if the deviation continues to exceed the tolerance range, the twin parameter self-correction is triggered.

[0043] Beneficial effects of the present invention:

[0044] By building a complete data-driven intelligent control system, the problem of closed ferrochrome furnace control errors caused by inaccurate determination of ferrochrome alloy composition has been effectively solved. The precise control and stable operation of the closed ferrochrome furnace have been achieved, ensuring that the production process can proceed according to the expected goals and improving product quality and production efficiency.

[0045] Using reaction batches as a benchmark, a dynamic time warping algorithm is used to align data streams of varying frequencies (e.g., second-level temperature data and minute-level gas composition data) to construct a correlation matrix and multidimensional dataset. This operation uncovers potential relationships between data, effectively integrating seemingly fragmented data across time. This lays a solid foundation for accurately analyzing the relationship between ferrochrome alloy composition and operating efficiency, enabling technicians to gain a deeper understanding of the inherent connection between furnace chemical reactions and operating efficiency, and helping to identify key factors influencing production.

[0046] The system monitors the efficiency recovery of the closed ferrochrome furnace in real time. If efficiency falls below target, incremental learning of the knowledge base is triggered, feeding new parameters and results back into the knowledge base. This allows the knowledge base to be continuously enriched and improved, giving the system the ability to self-learn and improve. When similar problems arise in the future, more accurate control decisions can be made based on more extensive empirical data.

[0047] After the executable parameter sequence is delivered to the actual furnace control, real-time operational data is collected and analyzed for deviations against the predictions of the closed ferrochrome furnace's digital twin model. If the deviation persists beyond the tolerance range, the twin's parameters are self-calibrated. This feedback mechanism continuously adjusts the digital twin's parameters to better simulate the actual furnace's operation, improving the match between simulation and reality. This in turn further optimizes control decisions, achieving continuous optimization of the closed ferrochrome furnace's control, and continuously improving control accuracy and production results.

[0048] In summary, the present invention, through the above series of technical means and operating procedures, brings significant beneficial effects to the control and regulation of closed ferrochrome furnaces in many aspects, and improves the stability, controllability and efficiency of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on the drawings without any creative work.

[0050] Figure 1 A schematic flow chart of a closed ferrochrome furnace control and adjustment method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. The technical solutions provided by each embodiment of the present invention are described in detail below in conjunction with the drawings.

[0052] In order to better understand the purpose of the present invention, the present invention is described in further detail below.

[0053] See also Figure 1 The present invention provides a closed ferrochrome furnace control and adjustment method, comprising:

[0054] Step S101, acquiring historical data of a closed ferrochrome furnace, the historical data of the closed ferrochrome furnace including historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace;

[0055] Multi-channel data collection: Using embedded high-temperature sensors within the furnace (such as infrared thermometers and laser gas analyzers), PLC system logs, and laboratory offline test reports, we capture both structured data (temperature, pressure, and current) and unstructured data (slag images and spectral analysis results). This multi-source data collection approach ensures a comprehensive understanding of the operating conditions of the closed ferrochrome furnace, providing rich information for subsequent analysis.

[0056] Data format processing: Convert the chemical raw material composition in the historical chemical reaction data in the closed ferrochrome furnace into mass percentage format, unify the data format, facilitate subsequent data processing and analysis, and eliminate errors and inconveniences that may be caused by data format differences.

[0057] Step S102, establishing a correlation relationship between historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace based on time information to obtain a closed ferrochrome furnace correlation data set;

[0058] Timeline alignment: Using the reaction batch as the timeline, a pre-set dynamic time warping algorithm is used to align data streams with different sampling frequencies, such as second-level temperature data and minute-level gas composition data, to construct a time and event correlation matrix. This step unifies data from different sources and frequencies along the time dimension, enabling analysis of the various data within the same timeframe and uncovering the inherent connections between the data.

[0059] Feature extraction and dataset construction: Extract time series features and mark the start and end time points of abnormal events to form a multidimensional correlated dataset. This approach not only preserves the time series information of the data but also highlights key events, providing more valuable features for subsequent data analysis and model building.

[0060] Step S103: Filtering out data from the closed ferrochrome furnace associated data set to obtain data with an operating efficiency higher than a preset efficiency value, thereby obtaining a first closed ferrochrome furnace data set; retrieving chemical material data corresponding to the first closed ferrochrome furnace data set; and storing the chemical material data corresponding to the first closed ferrochrome furnace data set in a pre-built knowledge base, thereby obtaining a closed ferrochrome furnace chemical material knowledge base;

[0061] High-efficiency data screening: Set a data threshold for a preset efficiency value (e.g., a closed ferrochrome furnace operating efficiency value ≥ 85%) and filter data with operating efficiencies exceeding this threshold from the closed ferrochrome furnace associated data set to obtain the first closed ferrochrome furnace data set. This filtered high-efficiency data represents the optimal operating conditions of the closed ferrochrome furnace and provides a reference standard for subsequent regulation.

[0062] Data Storage: A graph database is used to store the chemical material data corresponding to the first closed ferrochrome furnace data set, including chemical raw material ratios, reaction stages, efficiency indicators, edge relationship definition conditions, and chemical reaction results. A graph database clearly represents complex relationships between data, facilitating query, analysis, and reasoning, and helping to quickly obtain historical empirical data relevant to current operating conditions.

[0063] Step S104: receiving the expected operating efficiency of the closed ferrochrome furnace, collecting real-time operating efficiency data of the closed ferrochrome furnace, and comparing the real-time operating efficiency data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace. If the real-time operating efficiency data of the closed ferrochrome furnace is lower than the expected operating efficiency of the closed ferrochrome furnace, retrieving chemical material information within the closed ferrochrome furnace from the real-time operating efficiency data of the closed ferrochrome furnace to obtain a second closed ferrochrome furnace data set.

[0064] Expected Efficiency Processing: This step receives the user's desired operating efficiency for a closed ferrochrome furnace (e.g., target specific energy consumption, production rate), converts it into a standardized set of efficiency indicators, and stores it in the server-side dynamic configuration library. This step transforms user requirements into a standardized data format that the system can process, providing a foundation for subsequent comparison and analysis.

[0065] Real-time data collection and processing: Edge computing nodes are deployed to synchronously collect real-time data streams from furnace temperature, pressure, current, power, and output metering equipment. Time alignment and outlier removal are performed to generate a real-time operational efficiency dataset with consistent timestamps. This real-time data collection and processing provides a timely reflection of the current operating status of the closed ferrochrome furnace.

[0066] Deviation Analysis and Command Activation: Using a sliding window statistical method, the system calculates the dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value. If the deviation exceeds the permissible fluctuation range of ±3% for three consecutive sampling periods, the chemical material information extraction command is activated. This dynamic deviation analysis can promptly detect abnormal changes in operating efficiency and trigger the extraction of chemical material information that may affect composition and control.

[0067] Data tracing and accuracy verification: Based on the activation instruction time, chemical material parameters (such as coke charge, flux ratio, raw material grade, furnace pressure, and gas composition) for the corresponding period are traced back from the real-time operating efficiency dataset. These parameters are then inserted into the pre-set digital twin model of the closed ferrochrome furnace as the evaluation data, generating digital twin image data and comparing it with the expected operating efficiency. If the comparison results are consistent with the dynamic deviation calculated using the sliding window statistical method, the traced chemical material parameters are accurate. This step verifies the data accuracy through the digital twin model, providing reliable data support for subsequent control.

[0068] In step S105, the second closed ferrochrome furnace data group is matched with the closed ferrochrome furnace chemical material knowledge base to obtain a chemical material information matching result, and the chemical material information matching result is sent to the chemical material adding device. The operating efficiency of the closed ferrochrome furnace after the chemical material adding device adds the chemical material in the closed ferrochrome furnace is collected. If it is still lower than the expected operating efficiency of the closed ferrochrome furnace, the chemical material parameters in the chemical material information matching result are optimized using a genetic algorithm to obtain optimized chemical material parameters. The optimized chemical material parameters are used as execution data in the closed ferrochrome furnace adjustment process.

[0069] Generating Initial Control Recommendations: Dynamic Time Warping (DTW) calculates the multidimensional similarity between the second closed ferrochrome furnace data set (i.e., the chemical material parameter data set obtained retrospectively) and historical data in the knowledge base. Matching cases is then screened and the material parameters of these matching cases are extracted as initial control recommendation data. By matching with historical data, successful control experiences from the past can be leveraged to provide preliminary solutions for current control.

[0070] Genetic algorithm parameter optimization:

[0071] Coding design: Encode material parameters into binary gene strings to facilitate genetic algorithm operation.

[0072] Generating New Parameter Combinations: Initial control recommendations are set as optimization targets, and new parameter combinations are iteratively generated using tournament selection, two-point crossover, and Gaussian mutation strategies. Genetic algorithms simulate natural evolutionary processes to continuously search for optimal parameter combinations to improve control effectiveness.

[0073] Feasibility Verification and Solution Output: Digital twins rapidly simulate and verify the feasibility of new parameter combinations, outputting the optimal solution set for these new parameter combinations. The digital twin simulates the actual operation of a closed ferrochrome furnace, allowing for virtual testing of new parameter combinations to ensure their feasibility and effectiveness in real-world applications.

[0074] Control execution and feedback optimization:

[0075] Parameter Distribution and Operation Monitoring: The optimal solution for the new parameter combination is distributed to the raw material handling PLC system, and the efficiency recovery of the closed ferrochrome furnace is monitored in real time. This step applies the optimized parameters to actual production and monitors the results in real time.

[0076] Knowledge base incremental learning and twin calibration: If operating efficiency still falls short of the target, incremental learning of the knowledge base is triggered, feeding new parameters and results back into the knowledge base. Simultaneously, after the executable parameter sequence is delivered to the actual furnace control, real-time operational data is collected and analyzed for deviations compared to the predicted results of the closed ferrochrome furnace digital twin model. If the deviation persists beyond the tolerance range, twin parameter self-calibration is triggered. This feedback optimization mechanism continuously updates the knowledge base and digital twin model, improving the accuracy and adaptability of control.

[0077] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S101 includes:

[0078] Acquire structured and unstructured data through embedded high-temperature sensors in the furnace, PLC system logs, and laboratory offline test reports;

[0079] Embedded high-temperature sensors in the furnace include infrared thermometers and laser gas analyzers;

[0080] Structured data includes temperature, pressure, and current, while unstructured data includes slag images and spectral analysis results;

[0081] Convert chemical raw material compositions from historical chemical reaction data in a closed ferrochrome furnace into mass percentage format.

[0082] Embedded high-temperature sensors within the furnace: Infrared thermometers and laser gas analyzers are used as embedded high-temperature sensors within the furnace. The infrared thermometer accurately measures the furnace temperature, a key parameter that directly impacts the chemical reaction process and product quality of ferrochrome alloys. The laser gas analyzer monitors the furnace gas composition in real time, identifying changes in the concentrations of gases such as carbon monoxide and carbon dioxide. The concentrations of these gases are closely correlated with the progress of the carbothermal reduction reaction and are crucial for determining whether the reaction is proceeding normally.

[0083] PLC system logs: PLC (Programmable Logic Controller) systems play a core control role in industrial production. Their logs record various operations and status information during system operation, including equipment start and stop times, and changes in component operating parameters. By analyzing PLC system logs, we can obtain time-series data on the operation of closed ferrochrome furnaces, understand equipment operational stability, and the synergistic relationships between various parameters, providing a crucial basis for comprehensively evaluating furnace performance.

[0084] Offline laboratory testing reports: Offline laboratory testing enables detailed compositional analysis and performance testing of furnace samples. For example, by analyzing slag and ferrochrome samples, we can accurately determine the content of various elements, impurities, and physical properties of the alloy. Although acquired offline, this data provides highly accurate compositional and performance information, complementing real-time monitoring data to provide a more comprehensive understanding of the chemical reactions within the closed ferrochrome furnace and product quality.

[0085] Data type classification:

[0086] Structured Data: Structured data specifically includes temperature, pressure, and current. This data has a well-defined format and meaning, enabling straightforward numerical calculations and analysis. Temperature reflects the thermal state within the furnace, pressure fluctuations reflect the intensity of the reaction and the physical equilibrium within the furnace, and current, which is related to electrical energy input, directly influences heating efficiency and reaction rate. Accurate acquisition and analysis of this structured data facilitates the development of mathematical models linking these parameters, leading to a better understanding and control of the operation of closed ferrochrome furnaces.

[0087] Unstructured data: Unstructured data includes slag images and spectral analysis results. Slag images can intuitively reflect the slag's morphology, color, and other characteristics. Image analysis technology can be used to obtain information such as the slag's particle size distribution and composition distribution, thereby inferring the progress of the reaction within the furnace and the slag-iron separation effect. Spectral analysis results can provide information on the chemical composition and structure of the materials within the furnace. By analyzing the spectrum, the existence form and content of various elements can be determined, providing important support for the precise control of the composition of ferrochrome alloys. Although the processing of unstructured data is relatively complex, it contains rich information and plays an irreplaceable role in deeply understanding the chemical reaction mechanisms within closed ferrochrome furnaces.

[0088] Unified data format:

[0089] The chemical raw material compositions in historical chemical reaction data from a closed ferrochrome furnace are converted to mass percentage format. In actual production, chemical raw material compositions may be expressed in a variety of ways, which can complicate data comparison, analysis, and model building. Converting these to mass percentage format eliminates these differences in data representation, making data from different sources comparable and facilitating comprehensive analysis and processing. This helps establish unified data standards, improves data availability and the accuracy of analytical results, and provides a reliable data foundation for the subsequent development of data-based regulatory strategies.

[0090] Specifically, in the closed ferrochrome furnace control and adjustment method of the present invention, step S102 includes:

[0091] Using the reaction batch as the time axis, a preset dynamic time warping algorithm is used to align data streams with different sampling frequencies, including second-level temperature data and minute-level gas composition data, to construct a time and event correlation matrix.

[0092] Extract time series features and mark the start and end time points of abnormal events. Based on time series features, mark the start and end time points of abnormal events to form a multidimensional correlation data set;

[0093] Through the time and event correlation matrix and multidimensional correlation data set, the historical chemical reaction data in the closed ferrochrome furnace and the historical operating efficiency data of the closed ferrochrome furnace are established based on time information correlation.

[0094] Align the data streams based on reaction batches:

[0095] Reasonable Timeline Selection: The reaction batch was chosen as the timeline benchmark because each batch of ferrochrome production processes is relatively independent and complete. Within a reaction batch, changes in various data (such as temperature and gas composition) are closely related to the production operations and chemical reactions within that batch. Using the reaction batch as the benchmark effectively integrates data generated at different time points within the same batch, facilitating analysis of various phenomena and patterns within that batch's production process.

[0096] The role of the dynamic time warping algorithm: In actual production, different types of data often have inconsistent sampling frequencies due to the different characteristics and requirements of the monitoring equipment. For example, second-level temperature data can quickly reflect the instantaneous changes in the temperature inside the furnace, while minute-level gas composition data focuses more on the macroscopic trend of gas composition changes. The preset dynamic time warping algorithm can align data streams with different sampling frequencies in time without changing the essential characteristics of the data. This is like aligning music with different rhythms, allowing different data to be compared and analyzed on the same time scale, thereby exploring potential correlations between them. The time and event association matrix constructed in this way intuitively shows the corresponding relationship between different data in the time dimension, providing a clear framework for subsequent data processing and analysis.

[0097] Extracting time series features and marking abnormal events:

[0098] The significance of time series features: Extracting time series features is a crucial tool for gaining a deeper understanding of the operating patterns of closed ferrochrome furnaces. These features can include temperature trends (increasing, decreasing, or stable), fluctuations in gas composition, and the periodicity of various parameters. These features reflect the normal dynamics of ferrochrome production, helping technicians understand the equipment's operating characteristics and the inherent patterns of chemical reactions. For example, analyzing the temperature rise slope can determine the initial reaction rate; observing the periodicity of gas composition changes can reveal the cyclical nature of certain reactions, providing a basis for optimizing production processes.

[0099] The importance of marking abnormal events: Marking the start and end time points of abnormal events is to focus on and record special situations in the production process. Abnormal events may include sudden increases or decreases in furnace temperature, abnormal fluctuations in gas composition, and sudden increases in furnace pressure. These abnormal conditions will not only affect the quality of ferrochrome alloys, but may also cause damage to equipment. Accurately marking the start and end time points of abnormal events can quickly locate the time period when the problem occurred in subsequent analysis, and combine with other relevant data to deeply explore the causes of abnormal events so that appropriate measures can be taken to prevent and resolve them. Based on these time series features and abnormal event markers, the multidimensional associated data set formed contains rich information, not only time dimension information, but also covers the relationship between different parameters and records of abnormal situations, providing multi-dimensional data support for comprehensive analysis of the operating status of closed ferrochrome furnaces.

[0100] Establish data association relationships:

[0101] The association relationship is constructed by using a time-based association matrix and a multidimensional association dataset to establish a temporal association between historical chemical reaction data within the closed ferrochrome furnace and its historical operating efficiency data. This involves temporally correlating and linking data reflecting chemical reactions, such as temperature and gas composition, with operating efficiency data (such as output and energy consumption). For example, at a specific point in time, the corresponding operating efficiency data, combined with chemical reaction data such as temperature and gas composition, can be analyzed to identify which chemical reaction factors influence operating efficiency and how. This establishment of an association relationship enables technicians to grasp the production process holistically, eliminating the need to view chemical reactions and operating efficiency in isolation and instead gaining a deeper understanding of the inherent connection between them.

[0102] The application value of correlations: This temporal correlation provides powerful support for subsequent data analysis and decision-making. In actual production, technicians can use historical data correlations to predict operating efficiency under different operating conditions and adjust production parameters in advance to optimize the production process, improve product quality, and reduce energy consumption. Furthermore, when operating efficiency anomalies occur, correlations can be used to quickly trace the chemical reaction factors that may have caused the problem, allowing targeted investigation and resolution, thereby improving production stability and reliability.

[0103] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S103 includes:

[0104] Set the threshold of the preset efficiency value data, the threshold is the closed ferrochrome furnace operating efficiency value ≥ 85%;

[0105] A graph database is used to store the chemical material data corresponding to the first closed ferrochrome furnace data group. The chemical material data corresponding to the first closed ferrochrome furnace data group includes chemical raw material ratios, reaction stages, efficiency indicators, edge relationship definition conditions, and chemical reaction result data.

[0106] Set the threshold for the preset efficiency value:

[0107] The necessity of setting a threshold: Operating efficiency varies within the production process of closed ferrochrome furnaces. To filter representative and valuable historical data from this extensive dataset, a reasonable preset efficiency threshold is necessary. This threshold acts like a sieve, filtering out data with high operating efficiency and relatively ideal production processes. By analyzing and utilizing this high-quality data, we can better understand the various parameters and conditions of closed ferrochrome furnaces operating at high efficiency, providing a sound reference for subsequent regulation and control.

[0108] Considerations for choosing 85% as the threshold: Setting the threshold at ≥85% for the operating efficiency of a closed ferrochrome furnace is the result of comprehensive consideration of multiple factors. On the one hand, this value ensures that the first closed ferrochrome furnace data set selected represents a relatively high level of production, and the production process corresponding to these data is likely to have good performance in terms of product quality, energy consumption, and other aspects. On the other hand, this threshold is also reasonable and feasible. It will not result in too little data being selected due to the threshold being too high, making it impossible to form an effective reference data set, nor will it result in uneven quality of the selected data due to the threshold being too low, making it lose its reference value.

[0109] Using graph database to store data:

[0110] Reason for choosing a graph database: A graph database is a database system specifically designed to handle data with complex relationships. In this case, the chemical material data corresponding to the first closed ferrochrome furnace data set includes multiple aspects, such as chemical raw material ratios, reaction stages, and efficiency indicators. These data are complexly interrelated. For example, different chemical raw material ratios may lead to different reaction stages, which in turn affects efficiency indicators and chemical reaction results. Traditional database structures (such as relational databases) may face certain difficulties in handling such complex relationships. However, graph databases can represent the relationships between these data using an intuitive graphical structure, making the connections between the data clear at a glance.

[0111] The advantages of graph database storage include: The graph database structure clearly displays the causal relationships and mutual influences between stored chemical raw material ratios, reaction stages, efficiency indicators, edge definition conditions, and chemical reaction results. For example, a graph database allows users to intuitively see which chemical raw material ratios achieve the highest efficiency indicators at a specific reaction stage, as well as the chemical reaction results this ratio will trigger. This highly visual and associative storage method not only facilitates data query and retrieval but also helps technical personnel gain a deeper understanding of the inherent connections between data, providing strong support for developing appropriate control strategies. When controlling a closed ferrochrome furnace, users can quickly find historical data cases similar to the current operating conditions in the graph database, referencing information such as chemical raw material ratios and reaction conditions to make more accurate decisions.

[0112] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S104 includes:

[0113] Receive the expected operating efficiency of the closed ferrochrome furnace set by the user, which includes the unit energy consumption target and the production rate, convert the expected operating efficiency of the closed ferrochrome furnace into a standardized efficiency indicator set, and store it in the server-side dynamic configuration library;

[0114] Deploy edge computing nodes to synchronously collect real-time data streams from furnace temperature, pressure, current, power, and output metering equipment, perform time series alignment and remove outliers, and generate a real-time operating efficiency data set with consistent timestamps.

[0115] The dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace is calculated based on the sliding window statistical method. If the deviation exceeds the allowable fluctuation range of the closed ferrochrome furnace operating efficiency by ±3% for three consecutive sampling periods, the chemical material information extraction instruction is activated;

[0116] Based on the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding period are traced back from the real-time operation efficiency data set. The chemical material parameters include coke feed amount, flux ratio, raw material grade, furnace pressure and gas composition.

[0117] Expected operational efficiency processing:

[0118] Receiving and Standardizing: Receive user-specified expected operating efficiencies for closed ferrochrome furnaces, including key indicators such as unit energy consumption targets and production rates. These expected operating efficiencies are converted into a standardized set of efficiency metrics. Expected values ​​set by different users may vary in terms of units and magnitudes. Standardization unifies the data format and range, ensuring consistency and comparability in subsequent comparisons and analyses. These standardized metrics are then stored in a server-side dynamic configuration library, readily accessible by the system and serving as a benchmark for real-time operating efficiency comparisons.

[0119] Real-time operation efficiency data collection and processing:

[0120] Edge computing node deployment and data collection: Edge computing nodes are deployed to synchronously collect real-time data streams from furnace temperature, pressure, current, power, and output metering equipment. Located close to the data source, edge computing nodes enable rapid processing at the source, reducing data transmission latency and improving the real-time and accuracy of data collection. These real-time data streams, collected from multiple key locations, comprehensively reflect the operating status of the closed ferrochrome furnace.

[0121] Timing alignment and outlier removal: Due to different sources, collected real-time data may have different sampling frequencies and starting points. Timing alignment ensures accurate temporal alignment between different data streams, providing a foundation for subsequent unified analysis. Furthermore, outliers are removed. These may be erroneous data caused by sensor failure, interference, and other factors. If left unaddressed, they can seriously affect the accuracy and reliability of data analysis. This series of processing generates a real-time operational efficiency dataset with consistent timestamps, providing high-quality data support for real-time operational efficiency calculations.

[0122] Dynamic deviation calculation and instruction activation:

[0123] Application of the sliding window statistical method: This method is used to calculate the dynamic deviation between the real-time efficiency of a closed ferrochrome furnace and its expected value. The sliding window statistical method is a commonly used data analysis method. It uses a fixed-length time window, which slides across time series data, to calculate statistical characteristics (such as mean and variance) of the data within the window. In this invention, this method can dynamically monitor the deviation between the real-time efficiency and the expected value, promptly capturing fluctuations in operating efficiency.

[0124] Deviation Detection and Command Activation: A trigger condition is set when the deviation of three consecutive sampling periods exceeds the permissible fluctuation range of ±3% in the operating efficiency of the sealed ferrochrome furnace. When this condition is met, it indicates that the current operating efficiency has significantly deviated from the expected value, indicating a potential problem affecting production. At this point, the chemical material information extraction command is activated. This trigger condition takes into account the normal fluctuation range of the production process and the sensitivity to abnormal conditions. This prevents frequent command triggering due to small fluctuations while ensuring timely detection and resolution of abnormal conditions that could affect production.

[0125] Chemical material parameter traceability:

[0126] Parameter tracing basis: Based on the activation time of the chemical material information extraction command, chemical material parameters for the corresponding period are traced back from the real-time operating efficiency dataset. This is because when operating efficiency anomalies occur, it is likely that changes in chemical material parameters (such as coke feed, flux ratio, raw material grade, furnace pressure, and gas composition) have affected the reaction process and ultimately operating efficiency. This tracing back can determine the specific chemical material parameters at the time of the efficiency anomaly, providing a basis for subsequent problem analysis and parameter adjustments.

[0127] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S104 includes:

[0128] According to the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding time period will be traced back from the real-time operation efficiency dataset as the data to be evaluated, and the data to be evaluated will be substituted into the preset digital twin model of the closed ferrochrome furnace to obtain the digital twin image data of the closed ferrochrome furnace. The digital twin image data of the closed ferrochrome furnace will be compared with the expected operating efficiency of the closed ferrochrome furnace. If the result of comparing the digital twin image data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace is consistent with the dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace calculated based on the sliding window statistics method, then the chemical material parameters of the corresponding time period traced back from the real-time operation efficiency dataset are accurate.

[0129] Data preparation:

[0130] Acquiring data to be evaluated: Based on the activation time of the chemical material information extraction command, chemical material parameters for the corresponding period are traced back from the real-time operating efficiency dataset. These parameters include coke feed, flux ratio, raw material grade, furnace pressure, and gas composition, and are used as the data to be evaluated. This step is based on the suspicion that chemical material parameters may be the cause of previous operating efficiency deviations, so these parameters are obtained for further analysis and verification.

[0131] Digital twin model applications:

[0132] Model Substitution and Data Generation: The data to be evaluated is inserted into a pre-set digital twin model of a closed ferrochrome furnace. This digital twin model is a virtual representation of the actual physical system of the closed ferrochrome furnace. Built based on physical principles, mathematical models, and extensive historical data, it simulates the operation of the closed ferrochrome furnace under various operating conditions. By inputting the data to be evaluated into the model, the model operates according to its internal algorithms and logic, generating digital twin image data of the closed ferrochrome furnace. This data reflects the theoretical operating status and performance of the closed ferrochrome furnace under given chemical material parameters, including predicted values ​​for ferrochrome alloy composition, output, energy consumption, and other indicators.

[0133] Comparative verification:

[0134] Comparison process and judgment basis: The digital twin image data of the closed ferrochrome furnace is compared with the expected operating efficiency of the closed ferrochrome furnace. The expected operating efficiency is the target value set by the user and represents the ideal production operation state. At the same time, the dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value calculated based on the sliding window statistical method is reviewed. If the results of the comparison between the two are consistent, that is, the operating efficiency predicted by the digital twin model based on the chemical material parameters to be evaluated is consistent with the deviation between the actual monitored real-time efficiency and the expected value, then it can be considered that the chemical material parameters of the corresponding period traced back from the real-time operating efficiency data set are accurate.

[0135] In summary, this part of the technical content provides accuracy assurance for reverse traceability of chemical material parameters through comparative verification through the introduction of a digital twin model, improves the technical logic of the entire closed ferrochrome furnace control and adjustment method, and helps to improve the stability and controllability of the production process.

[0136] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S105 includes:

[0137] Dynamic time warping was used to calculate the multidimensional similarity between the second closed ferrochrome furnace data set and the historical data in the knowledge base, to screen matching cases and extract the material parameters of the matching cases as initial control recommendation data.

[0138] Genetic algorithm parameter optimization:

[0139] Coding design: Encoding material parameters into binary gene strings;

[0140] The initial control recommendation data is set as the optimization target, and new parameter combinations are iteratively generated using tournament selection, two-point crossover, and Gaussian mutation strategies. The feasibility of the new parameter combination is verified through rapid simulation of the digital twin, and the optimal solution set of the new parameter combination is output;

[0141] The optimal solution set of the new parameter combination is sent to the raw material conveying PLC system, and the recovery of the operating efficiency of the closed ferrochrome furnace is monitored in real time. If the standard is still not met, incremental learning of the knowledge base is triggered, and the new parameters and results are fed back to the knowledge base to form a continuous optimization closed loop.

[0142] Get initial control based on similarity matching:

[0143] Data comparison and similarity calculation: Dynamic time warping is used to calculate the multidimensional similarity between the second closed ferrochrome furnace data set and the historical data in the knowledge base. The second closed ferrochrome furnace data set is the chemical material information within the closed ferrochrome furnace, obtained by backtracking from the real-time operating efficiency dataset in step S104 when the real-time operating efficiency of the closed ferrochrome furnace falls below the expected operating efficiency. The knowledge base stores a large amount of historical operating data, which records the operation of the closed ferrochrome furnace under different operating conditions in the past. The dynamic time warping algorithm can handle the expansion and distortion of time series data on the time axis, thereby calculating the similarity between the two data sets across multiple dimensions. In this way, historical cases similar to the current operating conditions can be found.

[0144] Matching Case Screening and Parameter Extraction: Matching cases are selected based on the calculated similarity. These matching cases represent successful or relatively ideal operating conditions in similar situations in the past. Material parameters from these matching cases are extracted as initial control recommendations, providing a preliminary reference for current control. Because historical data contains successful experiences under different operating conditions, this matching method can quickly draw on past effective practices, providing a foundation for solving current problems.

[0145] Genetic algorithm parameter optimization:

[0146] Encoding design: Material parameters are encoded as binary gene strings. This is a common operation in genetic algorithms. By converting continuous material parameters (such as coke feed amount and flux ratio) into binary codes, the genetic algorithm can operate more easily. Each bit in the binary gene string can be considered a gene, and different gene combinations represent different parameter values. This encoding method enables the genetic algorithm to utilize the principles of genetic evolution in biology to search and optimize parameters.

[0147] Generating New Parameter Combinations: Initial control recommendations are set as the optimization target, and new parameter combinations are iteratively generated using tournament selection, two-point crossover, and Gaussian mutation strategies. The tournament selection strategy randomly selects a certain number of individuals (e.g., several gene strings) from the population and selects the individuals with the best fitness as parents to produce the next generation. Two-point crossover randomly selects two crossover points on the parent gene string and swaps the gene segments between them to generate new gene combinations. Gaussian mutation randomly mutates certain positions in the gene string according to a Gaussian distribution, introducing new genetic features. Through these operations, new parameter combinations are continuously generated iteratively, gradually searching for optimal solutions.

[0148] Feasibility Verification and Output of the Optimal Solution Set: The feasibility of new parameter combinations is verified through rapid simulation using the digital twin. The digital twin is a virtual model of the actual operation of a closed ferrochrome furnace, capable of simulating chemical reactions, physical processes, and operating efficiency within the furnace under different parameter combinations. The newly generated parameter combination is input into the digital twin for simulation. If the simulation results show that the parameter combination enables the closed ferrochrome furnace to operate within a reasonable range and meets production requirements, the parameter combination is considered feasible. After multiple rounds of iteration and verification, the optimal solution set for the new parameter combination is output. This optimal solution set includes a series of relatively optimal parameter combinations under the current conditions, optimized by the genetic algorithm and verified by the digital twin.

[0149] Control execution and feedback optimization:

[0150] Parameter Distribution and Operation Monitoring: The optimal solution for the new parameter combination is distributed to the raw material delivery PLC system. The PLC system controls parameters such as raw material delivery volume and delivery time. Transmitting the optimized parameter combination to the PLC system enables precise control of raw material input into the closed ferrochrome furnace. Simultaneously, the efficiency recovery of the closed ferrochrome furnace is monitored in real time, and the effectiveness of control measures is verified through actual operating data.

[0151] Incremental learning and continuous optimization of the knowledge base: If the operating efficiency of the closed ferrochrome furnace still does not meet the expected standards after adjustments, incremental learning of the knowledge base is triggered. This means that the newly tried parameters and actual operating results are fed back into the knowledge base, and the knowledge base is updated and expanded. This way, when similar problems are encountered in the future, the knowledge base will have more cases to refer to, and the genetic algorithm can also optimize based on more abundant data, thus forming a closed loop of continuous optimization. In this way, the system can continuously learn and improve, gradually enhancing the control and regulation capabilities of the closed ferrochrome furnace to adapt to different production needs and changing operating conditions.

[0152] Specifically, the closed ferrochrome furnace control and adjustment method of the present invention, step S105 includes:

[0153] The new parameter combination generated by the genetic algorithm is mapped to the input interface of the digital twin according to the preset rules, and the current furnace state data is simultaneously loaded as the initial condition of the simulation;

[0154] Based on the preset carbothermal reduction reaction mechanism model, multi-physics field coupling calculations were initiated within the twin to simulate the dynamic process within the furnace under the new parameters. The predicted values ​​of indicators such as chromium / iron ratio, impurity concentration, energy efficiency, and furnace pressure fluctuation curve were output, and simulation results were obtained.

[0155] Compare the simulation results with the preset process constraints, eliminate the parameter combinations with the risk of exceeding the standards in the new parameter combinations, sort the comparison results by priority weight, and generate an executable parameter sequence;

[0156] After the executable parameter sequence is sent to the actual furnace control, the operating data is collected in real time and the deviation is analyzed with the prediction results of the closed ferrochrome furnace digital twin model; if the deviation continues to exceed the tolerance range, the twin parameter self-correction is triggered.

[0157] Digital twin parameter injection and initial condition setting:

[0158] Parameter Mapping and Loading: The new parameter combinations generated by the genetic algorithm are mapped to the digital twin's input interface according to pre-set rules. This step enables the digital twin to simulate the operation of the closed ferrochrome furnace based on the newly generated parameters. Simultaneously, the current furnace state data is loaded as initial simulation conditions. Because the actual furnace state (such as temperature distribution and pressure) will affect subsequent simulation results, only by accurately setting these initial conditions can the digital twin's simulation more closely resemble reality, providing a reliable foundation for subsequent accurate evaluation of the effectiveness of the new parameter combinations.

[0159] Simulation calculation based on mechanism model:

[0160] Multi-physics coupling calculations are initiated within the twin, based on a pre-set carbothermal reduction reaction mechanism model. Within a closed ferrochrome furnace, multiple physical fields interact, such as heat transfer, mass transfer, and chemical reaction kinetics, all of which influence and constrain each other. Multi-physics coupling calculations enable a comprehensive and accurate simulation of the dynamic processes within the furnace under new parameters, simulating the temporal evolution of various physical and chemical quantities within the furnace.

[0161] Output Index Predictions: After the above calculations, output index predictions include chromium / iron ratio, impurity concentration, energy efficiency, and furnace pressure fluctuation curves. These indicators are key factors in measuring the operating performance of a closed ferrochrome furnace. By simulating their predicted values, we can understand in advance the possible furnace conditions under new parameter combinations, such as whether the ferrochrome alloy composition meets requirements, whether energy consumption is within a reasonable range, and whether the furnace pressure is stable, thereby making a preliminary assessment of the feasibility of the new parameter combination.

[0162] Comparison and processing of simulation results and process constraints:

[0163] Eliminate risk of exceeding standards: Simulation results are compared with pre-set process constraints, which are standards set based on production requirements and equipment performance, such as the lower limit of chromium content, a safe range for furnace pressure, and an upper limit for impurity concentration. If this comparison reveals that a new parameter combination causes certain indicators to exceed the process constraints (i.e., poses a risk of exceeding standards), these parameter combinations with risk of exceeding standards are eliminated to ensure that the final parameter combination meets basic production requirements, guaranteeing product quality and equipment safety.

[0164] Results are sorted to generate an executable parameter sequence: After elimination, the comparison results are sorted by priority. These priority weights may be determined based on key factors in the production process. For example, a focus on product quality (chromium / iron ratio, impurity concentration, etc.) may give higher weights to parameters related to these indicators. This sorting generates an executable parameter sequence. The parameter combinations in this sequence are arranged by priority, while meeting process constraints, facilitating the subsequent selection of the appropriate parameter combination for actual furnace control.

[0165] Control execution, deviation analysis, and twin correction:

[0166] Parameter Distribution and Operational Data Collection: After the executable parameter sequence is distributed to the actual furnace control, the optimal parameter combination obtained through the previous series of screening and sorting is applied to the actual operation of the closed ferrochrome furnace to achieve precise furnace control. Simultaneously, real-time operational data is collected to provide timely information on the actual furnace operation under the new parameter combination. This operational data will be used for subsequent comparative analysis with the digital twin model's prediction results.

[0167] Deviation Analysis and Twin Parameter Self-Correction: A deviation analysis is performed between the real-time collected operating data and the predicted results of the closed ferrochrome furnace digital twin model. If the deviation continues to exceed the tolerance range, it indicates that the digital twin simulation differs significantly from the actual situation. This may be due to unaccounted factors in the actual furnace operation or inaccurate parameter settings of the digital twin itself. In this case, it is necessary to trigger the twin parameter self-correlation. By adjusting the relevant parameters of the digital twin, it can better simulate the actual furnace operation, thereby improving the accuracy of subsequent simulation and control.

[0168] The present invention effectively solves the problem of closed ferrochrome furnace control errors caused by inaccurate determination of ferrochrome alloy composition by constructing a data-driven intelligent control system, as follows:

[0169] Multi-source data acquisition and preprocessing: Comprehensively collect historical data through multiple channels, covering materials, reactions, and operating efficiency, and perform format conversion and cleaning to provide high-quality data support for subsequent analysis, ensure the accuracy of the data foundation, and avoid component determination deviations caused by data errors.

[0170] Construction of Correlated Datasets: Based on reaction batches, we utilize a dynamic time warping algorithm to align data streams of different frequencies, construct correlation matrices and multidimensional datasets, and explore potential relationships between data, laying the foundation for accurate analysis of the relationship between ferrochrome alloy composition and operating efficiency.

[0171] Knowledge base establishment: Filter efficient operation data, extract chemical material data and store it in the graph database to form a knowledge base, providing historical experience reference for real-time control, and helping to make accurate raw material ratios and parameter adjustment decisions when facing similar working conditions.

[0172] Real-time monitoring and comparison: Receive expected operating efficiency, collect real-time operating efficiency data and compare. When the efficiency is lower than expected, use the sliding window statistical method to judge and activate the chemical material information extraction instruction to promptly discover potential composition determination and control problems.

[0173] Data accuracy verification: Substitute the traceable chemical material parameters into the digital twin model and compare them with the expected operating efficiency to verify the parameter accuracy and avoid control errors caused by incorrect parameters.

[0174] Intelligent optimization and control: Dynamic time warping is used to screen matching cases to obtain initial control suggestions. Chemical material parameters are optimized through genetic algorithms. After verification through digital twin simulation, the optimal solution set is obtained and sent to the PLC system for execution, ensuring the scientific nature and effectiveness of the control instructions.

[0175] Feedback and continuous optimization: Real-time monitoring of operating efficiency. When it fails to meet the standards, incremental learning of the knowledge base is triggered, and new parameters and results are fed back to the knowledge base. At the same time, the actual operating data is compared with the predicted results of the digital twin model. If the deviation exceeds the range, the twin parameters are corrected, forming a continuous optimization closed loop and continuously improving the control accuracy.

[0176] Through the above steps, the present invention forms a complete and closed-loop control system from data collection, analysis, decision-making to execution and feedback, effectively overcoming the control problems caused by inaccurate measurement of ferrochrome alloy composition, and realizing precise control and stable operation of the closed ferrochrome furnace.

Claims

1. A closed ferrochrome furnace control and adjustment method, characterized in that: include: Step S101, acquiring historical data of a closed ferrochrome furnace, the historical data of the closed ferrochrome furnace including historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace; Step S102, establishing a correlation relationship between historical material data in the closed ferrochrome furnace, historical chemical reaction data in the closed ferrochrome furnace, and historical operating efficiency data of the closed ferrochrome furnace based on time information to obtain a closed ferrochrome furnace correlation data set; Step S103: Filtering out data from the closed ferrochrome furnace associated data set to obtain data with an operating efficiency higher than a preset efficiency value, thereby obtaining a first closed ferrochrome furnace data set; retrieving chemical material data corresponding to the first closed ferrochrome furnace data set; and storing the chemical material data corresponding to the first closed ferrochrome furnace data set in a pre-built knowledge base, thereby obtaining a closed ferrochrome furnace chemical material knowledge base; Step S104: receiving the expected operating efficiency of the closed ferrochrome furnace, collecting real-time operating efficiency data of the closed ferrochrome furnace, and comparing the real-time operating efficiency data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace. If the real-time operating efficiency data of the closed ferrochrome furnace is lower than the expected operating efficiency of the closed ferrochrome furnace, retrieving chemical material information within the closed ferrochrome furnace from the real-time operating efficiency data of the closed ferrochrome furnace to obtain a second closed ferrochrome furnace data set. In step S105, the second closed ferrochrome furnace data group is matched with the closed ferrochrome furnace chemical material knowledge base to obtain a chemical material information matching result, and the chemical material information matching result is sent to the chemical material adding device. The operating efficiency of the closed ferrochrome furnace after the chemical material adding device adds the chemical material in the closed ferrochrome furnace is collected. If it is still lower than the expected operating efficiency of the closed ferrochrome furnace, the chemical material parameters in the chemical material information matching result are optimized using a genetic algorithm to obtain optimized chemical material parameters. The optimized chemical material parameters are used as execution data in the closed ferrochrome furnace adjustment process.

2. A closed ferrochrome furnace control and adjustment method according to claim 1, characterized in that: The step S101 includes: Acquire structured and unstructured data through embedded high-temperature sensors in the furnace, PLC system logs, and laboratory offline test reports; Embedded high-temperature sensors in the furnace include infrared thermometers and laser gas analyzers; Structured data includes temperature, pressure, and current, while unstructured data includes slag images and spectral analysis results; Convert chemical raw material compositions from historical chemical reaction data in a closed ferrochrome furnace into mass percentage format.

3. A closed ferrochrome furnace control and adjustment method according to claim 1, characterized in that: The step S102 includes: Using the reaction batch as the time axis, a preset dynamic time warping algorithm is used to align data streams with different sampling frequencies, including second-level temperature data and minute-level gas composition data, to construct a time and event correlation matrix. Extract time series features and mark the start and end time points of abnormal events. Based on time series features, mark the start and end time points of abnormal events to form a multidimensional correlation data set; Through the time and event correlation matrix and multidimensional correlation data set, the historical chemical reaction data in the closed ferrochrome furnace and the historical operating efficiency data of the closed ferrochrome furnace are established based on time information correlation.

4. A closed ferrochrome furnace control and adjustment method according to claim 1, characterized in that: The step S103 includes: Set the threshold of the data of the preset efficiency value, and the data threshold of the preset efficiency value is 85%; A graph database is used to store the chemical material data corresponding to the first closed ferrochrome furnace data group. The chemical material data corresponding to the first closed ferrochrome furnace data group includes chemical raw material ratios, reaction stages, efficiency indicators, edge relationship definition conditions, and chemical reaction result data.

5. A closed ferrochrome furnace control and adjustment method according to claim 1, characterized in that: The step S104 includes: Receive the expected operating efficiency of the closed ferrochrome furnace set by the user, which includes the unit energy consumption target and the production rate, convert the expected operating efficiency of the closed ferrochrome furnace into a standardized efficiency indicator set, and store it in the server-side dynamic configuration library; Deploy edge computing nodes to synchronously collect real-time data streams from furnace temperature, pressure, current, power, and output metering equipment, perform time series alignment and remove outliers, and generate a real-time operating efficiency data set with consistent timestamps. The dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace is calculated based on the sliding window statistical method. If the deviation exceeds the allowable fluctuation range of the closed ferrochrome furnace operating efficiency by ±3% for three consecutive sampling periods, the chemical material information extraction instruction is activated; Based on the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding period are traced back from the real-time operation efficiency data set. The chemical material parameters include coke feed amount, flux ratio, raw material grade, furnace pressure and gas composition.

6. A closed ferrochrome furnace control and adjustment method according to claim 5, characterized in that: The step S104 includes: According to the time point of activating the chemical material information extraction instruction, the chemical material parameters of the corresponding time period will be traced back from the real-time operation efficiency dataset as the data to be evaluated, and the data to be evaluated will be substituted into the preset digital twin model of the closed ferrochrome furnace to obtain the digital twin image data of the closed ferrochrome furnace. The digital twin image data of the closed ferrochrome furnace will be compared with the expected operating efficiency of the closed ferrochrome furnace. If the result of comparing the digital twin image data of the closed ferrochrome furnace with the expected operating efficiency of the closed ferrochrome furnace is consistent with the dynamic deviation between the real-time efficiency of the closed ferrochrome furnace and the expected value of the closed ferrochrome furnace calculated based on the sliding window statistics method, then the chemical material parameters of the corresponding time period traced back from the real-time operation efficiency dataset are accurate.

7. A closed ferrochrome furnace control and adjustment method according to claim 1, characterized in that: The step S105 includes: Dynamic time warping was used to calculate the multidimensional similarity between the second closed ferrochrome furnace data set and the historical data in the knowledge base, to screen matching cases and extract the material parameters of the matching cases as initial control recommendation data. Genetic algorithm parameter optimization: Coding design: Encoding material parameters into binary gene strings; The initial control recommendation data is set as the optimization target, and new parameter combinations are iteratively generated using tournament selection, two-point crossover, and Gaussian mutation strategies. The feasibility of the new parameter combination is verified through rapid simulation of the digital twin, and the optimal solution set of the new parameter combination is output; The optimal solution set of the new parameter combination is sent to the raw material conveying PLC system, and the recovery of the operating efficiency of the closed ferrochrome furnace is monitored in real time. If the standard is still not met, incremental learning of the knowledge base is triggered, and the new parameters and results are fed back to the knowledge base to form a continuous optimization closed loop.

8. A closed ferrochrome furnace control and adjustment method according to claim 7, characterized in that: The step S105 includes: The new parameter combination generated by the genetic algorithm is mapped to the input interface of the digital twin according to the preset rules, and the current furnace state data is simultaneously loaded as the initial condition of the simulation; Based on the preset carbothermal reduction reaction mechanism model, multi-physics field coupling calculations were initiated within the twin to simulate the dynamic process within the furnace under the new parameters. The predicted values ​​of indicators such as chromium / iron ratio, impurity concentration, energy efficiency, and furnace pressure fluctuation curve were output, and simulation results were obtained. Compare the simulation results with the preset process constraints, eliminate the parameter combinations with the risk of exceeding the standards in the new parameter combinations, sort the comparison results by priority weight, and generate an executable parameter sequence; After the executable parameter sequence is sent to the actual furnace control, the operating data is collected in real time and the deviation is analyzed with the prediction results of the closed ferrochrome furnace digital twin model; if the deviation continues to exceed the tolerance range, the twin parameter self-correction is triggered.

Images